OpenCloudOS-Kernel/mm/page-writeback.c

2848 lines
85 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* mm/page-writeback.c
*
* Copyright (C) 2002, Linus Torvalds.
* Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra
*
* Contains functions related to writing back dirty pages at the
* address_space level.
*
* 10Apr2002 Andrew Morton
* Initial version
*/
#include <linux/kernel.h>
#include <linux/export.h>
#include <linux/spinlock.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/slab.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/init.h>
#include <linux/backing-dev.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/blkdev.h>
#include <linux/mpage.h>
[PATCH] mm: tracking shared dirty pages Tracking of dirty pages in shared writeable mmap()s. The idea is simple: write protect clean shared writeable pages, catch the write-fault, make writeable and set dirty. On page write-back clean all the PTE dirty bits and write protect them once again. The implementation is a tad harder, mainly because the default backing_dev_info capabilities were too loosely maintained. Hence it is not enough to test the backing_dev_info for cap_account_dirty. The current heuristic is as follows, a VMA is eligible when: - its shared writeable (vm_flags & (VM_WRITE|VM_SHARED)) == (VM_WRITE|VM_SHARED) - it is not a 'special' mapping (vm_flags & (VM_PFNMAP|VM_INSERTPAGE)) == 0 - the backing_dev_info is cap_account_dirty mapping_cap_account_dirty(vma->vm_file->f_mapping) - f_op->mmap() didn't change the default page protection Page from remap_pfn_range() are explicitly excluded because their COW semantics are already horrid enough (see vm_normal_page() in do_wp_page()) and because they don't have a backing store anyway. mprotect() is taught about the new behaviour as well. However it overrides the last condition. Cleaning the pages on write-back is done with page_mkclean() a new rmap call. It can be called on any page, but is currently only implemented for mapped pages, if the page is found the be of a VMA that accounts dirty pages it will also wrprotect the PTE. Finally, in fs/buffers.c:try_to_free_buffers(); remove clear_page_dirty() from under ->private_lock. This seems to be safe, since ->private_lock is used to serialize access to the buffers, not the page itself. This is needed because clear_page_dirty() will call into page_mkclean() and would thereby violate locking order. [dhowells@redhat.com: Provide a page_mkclean() implementation for NOMMU] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Hugh Dickins <hugh@veritas.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-26 14:30:57 +08:00
#include <linux/rmap.h>
#include <linux/percpu.h>
#include <linux/smp.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/syscalls.h>
#include <linux/buffer_head.h> /* __set_page_dirty_buffers */
#include <linux/pagevec.h>
#include <linux/timer.h>
#include <linux/sched/rt.h>
#include <linux/sched/signal.h>
mm: vmscan: fix do_try_to_free_pages() livelock This patch is based on KOSAKI's work and I add a little more description, please refer https://lkml.org/lkml/2012/6/14/74. Currently, I found system can enter a state that there are lots of free pages in a zone but only order-0 and order-1 pages which means the zone is heavily fragmented, then high order allocation could make direct reclaim path's long stall(ex, 60 seconds) especially in no swap and no compaciton enviroment. This problem happened on v3.4, but it seems issue still lives in current tree, the reason is do_try_to_free_pages enter live lock: kswapd will go to sleep if the zones have been fully scanned and are still not balanced. As kswapd thinks there's little point trying all over again to avoid infinite loop. Instead it changes order from high-order to 0-order because kswapd think order-0 is the most important. Look at 73ce02e9 in detail. If watermarks are ok, kswapd will go back to sleep and may leave zone->all_unreclaimable =3D 0. It assume high-order users can still perform direct reclaim if they wish. Direct reclaim continue to reclaim for a high order which is not a COSTLY_ORDER without oom-killer until kswapd turn on zone->all_unreclaimble= . This is because to avoid too early oom-kill. So it means direct_reclaim depends on kswapd to break this loop. In worst case, direct-reclaim may continue to page reclaim forever when kswapd sleeps forever until someone like watchdog detect and finally kill the process. As described in: http://thread.gmane.org/gmane.linux.kernel.mm/103737 We can't turn on zone->all_unreclaimable from direct reclaim path because direct reclaim path don't take any lock and this way is racy. Thus this patch removes zone->all_unreclaimable field completely and recalculates zone reclaimable state every time. Note: we can't take the idea that direct-reclaim see zone->pages_scanned directly and kswapd continue to use zone->all_unreclaimable. Because, it is racy. commit 929bea7c71 (vmscan: all_unreclaimable() use zone->all_unreclaimable as a name) describes the detail. [akpm@linux-foundation.org: uninline zone_reclaimable_pages() and zone_reclaimable()] Cc: Aaditya Kumar <aaditya.kumar.30@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Nick Piggin <npiggin@gmail.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Bob Liu <lliubbo@gmail.com> Cc: Neil Zhang <zhangwm@marvell.com> Cc: Russell King - ARM Linux <linux@arm.linux.org.uk> Reviewed-by: Michal Hocko <mhocko@suse.cz> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Lisa Du <cldu@marvell.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:36 +08:00
#include <linux/mm_inline.h>
#include <trace/events/writeback.h>
mm: vmscan: fix do_try_to_free_pages() livelock This patch is based on KOSAKI's work and I add a little more description, please refer https://lkml.org/lkml/2012/6/14/74. Currently, I found system can enter a state that there are lots of free pages in a zone but only order-0 and order-1 pages which means the zone is heavily fragmented, then high order allocation could make direct reclaim path's long stall(ex, 60 seconds) especially in no swap and no compaciton enviroment. This problem happened on v3.4, but it seems issue still lives in current tree, the reason is do_try_to_free_pages enter live lock: kswapd will go to sleep if the zones have been fully scanned and are still not balanced. As kswapd thinks there's little point trying all over again to avoid infinite loop. Instead it changes order from high-order to 0-order because kswapd think order-0 is the most important. Look at 73ce02e9 in detail. If watermarks are ok, kswapd will go back to sleep and may leave zone->all_unreclaimable =3D 0. It assume high-order users can still perform direct reclaim if they wish. Direct reclaim continue to reclaim for a high order which is not a COSTLY_ORDER without oom-killer until kswapd turn on zone->all_unreclaimble= . This is because to avoid too early oom-kill. So it means direct_reclaim depends on kswapd to break this loop. In worst case, direct-reclaim may continue to page reclaim forever when kswapd sleeps forever until someone like watchdog detect and finally kill the process. As described in: http://thread.gmane.org/gmane.linux.kernel.mm/103737 We can't turn on zone->all_unreclaimable from direct reclaim path because direct reclaim path don't take any lock and this way is racy. Thus this patch removes zone->all_unreclaimable field completely and recalculates zone reclaimable state every time. Note: we can't take the idea that direct-reclaim see zone->pages_scanned directly and kswapd continue to use zone->all_unreclaimable. Because, it is racy. commit 929bea7c71 (vmscan: all_unreclaimable() use zone->all_unreclaimable as a name) describes the detail. [akpm@linux-foundation.org: uninline zone_reclaimable_pages() and zone_reclaimable()] Cc: Aaditya Kumar <aaditya.kumar.30@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Nick Piggin <npiggin@gmail.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Bob Liu <lliubbo@gmail.com> Cc: Neil Zhang <zhangwm@marvell.com> Cc: Russell King - ARM Linux <linux@arm.linux.org.uk> Reviewed-by: Michal Hocko <mhocko@suse.cz> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Lisa Du <cldu@marvell.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:36 +08:00
#include "internal.h"
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
/*
* Sleep at most 200ms at a time in balance_dirty_pages().
*/
#define MAX_PAUSE max(HZ/5, 1)
/*
* Try to keep balance_dirty_pages() call intervals higher than this many pages
* by raising pause time to max_pause when falls below it.
*/
#define DIRTY_POLL_THRESH (128 >> (PAGE_SHIFT - 10))
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
/*
* Estimate write bandwidth at 200ms intervals.
*/
#define BANDWIDTH_INTERVAL max(HZ/5, 1)
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
#define RATELIMIT_CALC_SHIFT 10
/*
* After a CPU has dirtied this many pages, balance_dirty_pages_ratelimited
* will look to see if it needs to force writeback or throttling.
*/
static long ratelimit_pages = 32;
/* The following parameters are exported via /proc/sys/vm */
/*
* Start background writeback (via writeback threads) at this percentage
*/
int dirty_background_ratio = 10;
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
/*
* dirty_background_bytes starts at 0 (disabled) so that it is a function of
* dirty_background_ratio * the amount of dirtyable memory
*/
unsigned long dirty_background_bytes;
/*
* free highmem will not be subtracted from the total free memory
* for calculating free ratios if vm_highmem_is_dirtyable is true
*/
int vm_highmem_is_dirtyable;
/*
* The generator of dirty data starts writeback at this percentage
*/
int vm_dirty_ratio = 20;
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
/*
* vm_dirty_bytes starts at 0 (disabled) so that it is a function of
* vm_dirty_ratio * the amount of dirtyable memory
*/
unsigned long vm_dirty_bytes;
/*
* The interval between `kupdate'-style writebacks
*/
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */
EXPORT_SYMBOL_GPL(dirty_writeback_interval);
/*
* The longest time for which data is allowed to remain dirty
*/
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */
/*
* Flag that makes the machine dump writes/reads and block dirtyings.
*/
int block_dump;
/*
* Flag that puts the machine in "laptop mode". Doubles as a timeout in jiffies:
* a full sync is triggered after this time elapses without any disk activity.
*/
int laptop_mode;
EXPORT_SYMBOL(laptop_mode);
/* End of sysctl-exported parameters */
struct wb_domain global_wb_domain;
/* consolidated parameters for balance_dirty_pages() and its subroutines */
struct dirty_throttle_control {
#ifdef CONFIG_CGROUP_WRITEBACK
struct wb_domain *dom;
struct dirty_throttle_control *gdtc; /* only set in memcg dtc's */
#endif
struct bdi_writeback *wb;
struct fprop_local_percpu *wb_completions;
unsigned long avail; /* dirtyable */
unsigned long dirty; /* file_dirty + write + nfs */
unsigned long thresh; /* dirty threshold */
unsigned long bg_thresh; /* dirty background threshold */
unsigned long wb_dirty; /* per-wb counterparts */
unsigned long wb_thresh;
unsigned long wb_bg_thresh;
unsigned long pos_ratio;
};
/*
* Length of period for aging writeout fractions of bdis. This is an
* arbitrarily chosen number. The longer the period, the slower fractions will
* reflect changes in current writeout rate.
*/
#define VM_COMPLETIONS_PERIOD_LEN (3*HZ)
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
#ifdef CONFIG_CGROUP_WRITEBACK
#define GDTC_INIT(__wb) .wb = (__wb), \
.dom = &global_wb_domain, \
.wb_completions = &(__wb)->completions
#define GDTC_INIT_NO_WB .dom = &global_wb_domain
#define MDTC_INIT(__wb, __gdtc) .wb = (__wb), \
.dom = mem_cgroup_wb_domain(__wb), \
.wb_completions = &(__wb)->memcg_completions, \
.gdtc = __gdtc
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
static bool mdtc_valid(struct dirty_throttle_control *dtc)
{
return dtc->dom;
}
static struct wb_domain *dtc_dom(struct dirty_throttle_control *dtc)
{
return dtc->dom;
}
static struct dirty_throttle_control *mdtc_gdtc(struct dirty_throttle_control *mdtc)
{
return mdtc->gdtc;
}
static struct fprop_local_percpu *wb_memcg_completions(struct bdi_writeback *wb)
{
return &wb->memcg_completions;
}
static void wb_min_max_ratio(struct bdi_writeback *wb,
unsigned long *minp, unsigned long *maxp)
{
unsigned long this_bw = wb->avg_write_bandwidth;
unsigned long tot_bw = atomic_long_read(&wb->bdi->tot_write_bandwidth);
unsigned long long min = wb->bdi->min_ratio;
unsigned long long max = wb->bdi->max_ratio;
/*
* @wb may already be clean by the time control reaches here and
* the total may not include its bw.
*/
if (this_bw < tot_bw) {
if (min) {
min *= this_bw;
mm/page-writeback.c: avoid potential division by zero in wb_min_max_ratio() Patch series "use div64_ul() instead of div_u64() if the divisor is unsigned long". We were first inspired by commit b0ab99e7736a ("sched: Fix possible divide by zero in avg_atom () calculation"), then refer to the recently analyzed mm code, we found this suspicious place. 201 if (min) { 202 min *= this_bw; 203 do_div(min, tot_bw); 204 } And we also disassembled and confirmed it: /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 201 0xffffffff811c37da <__wb_calc_thresh+234>: xor %r10d,%r10d 0xffffffff811c37dd <__wb_calc_thresh+237>: test %rax,%rax 0xffffffff811c37e0 <__wb_calc_thresh+240>: je 0xffffffff811c3800 <__wb_calc_thresh+272> /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 202 0xffffffff811c37e2 <__wb_calc_thresh+242>: imul %r8,%rax /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 203 0xffffffff811c37e6 <__wb_calc_thresh+246>: mov %r9d,%r10d ---> truncates it to 32 bits here 0xffffffff811c37e9 <__wb_calc_thresh+249>: xor %edx,%edx 0xffffffff811c37eb <__wb_calc_thresh+251>: div %r10 0xffffffff811c37ee <__wb_calc_thresh+254>: imul %rbx,%rax 0xffffffff811c37f2 <__wb_calc_thresh+258>: shr $0x2,%rax 0xffffffff811c37f6 <__wb_calc_thresh+262>: mul %rcx 0xffffffff811c37f9 <__wb_calc_thresh+265>: shr $0x2,%rdx 0xffffffff811c37fd <__wb_calc_thresh+269>: mov %rdx,%r10 This series uses div64_ul() instead of div_u64() if the divisor is unsigned long, to avoid truncation to 32-bit on 64-bit platforms. This patch (of 3): The variables 'min' and 'max' are unsigned long and do_div truncates them to 32 bits, which means it can test non-zero and be truncated to zero for division. Fix this issue by using div64_ul() instead. Link: http://lkml.kernel.org/r/20200102081442.8273-2-wenyang@linux.alibaba.com Fixes: 693108a8a667 ("writeback: make bdi->min/max_ratio handling cgroup writeback aware") Signed-off-by: Wen Yang <wenyang@linux.alibaba.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Qian Cai <cai@lca.pw> Cc: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 08:29:23 +08:00
min = div64_ul(min, tot_bw);
}
if (max < 100) {
max *= this_bw;
mm/page-writeback.c: avoid potential division by zero in wb_min_max_ratio() Patch series "use div64_ul() instead of div_u64() if the divisor is unsigned long". We were first inspired by commit b0ab99e7736a ("sched: Fix possible divide by zero in avg_atom () calculation"), then refer to the recently analyzed mm code, we found this suspicious place. 201 if (min) { 202 min *= this_bw; 203 do_div(min, tot_bw); 204 } And we also disassembled and confirmed it: /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 201 0xffffffff811c37da <__wb_calc_thresh+234>: xor %r10d,%r10d 0xffffffff811c37dd <__wb_calc_thresh+237>: test %rax,%rax 0xffffffff811c37e0 <__wb_calc_thresh+240>: je 0xffffffff811c3800 <__wb_calc_thresh+272> /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 202 0xffffffff811c37e2 <__wb_calc_thresh+242>: imul %r8,%rax /usr/src/debug/kernel-4.9.168-016.ali3000/linux-4.9.168-016.ali3000.alios7.x86_64/mm/page-writeback.c: 203 0xffffffff811c37e6 <__wb_calc_thresh+246>: mov %r9d,%r10d ---> truncates it to 32 bits here 0xffffffff811c37e9 <__wb_calc_thresh+249>: xor %edx,%edx 0xffffffff811c37eb <__wb_calc_thresh+251>: div %r10 0xffffffff811c37ee <__wb_calc_thresh+254>: imul %rbx,%rax 0xffffffff811c37f2 <__wb_calc_thresh+258>: shr $0x2,%rax 0xffffffff811c37f6 <__wb_calc_thresh+262>: mul %rcx 0xffffffff811c37f9 <__wb_calc_thresh+265>: shr $0x2,%rdx 0xffffffff811c37fd <__wb_calc_thresh+269>: mov %rdx,%r10 This series uses div64_ul() instead of div_u64() if the divisor is unsigned long, to avoid truncation to 32-bit on 64-bit platforms. This patch (of 3): The variables 'min' and 'max' are unsigned long and do_div truncates them to 32 bits, which means it can test non-zero and be truncated to zero for division. Fix this issue by using div64_ul() instead. Link: http://lkml.kernel.org/r/20200102081442.8273-2-wenyang@linux.alibaba.com Fixes: 693108a8a667 ("writeback: make bdi->min/max_ratio handling cgroup writeback aware") Signed-off-by: Wen Yang <wenyang@linux.alibaba.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Qian Cai <cai@lca.pw> Cc: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 08:29:23 +08:00
max = div64_ul(max, tot_bw);
}
}
*minp = min;
*maxp = max;
}
#else /* CONFIG_CGROUP_WRITEBACK */
#define GDTC_INIT(__wb) .wb = (__wb), \
.wb_completions = &(__wb)->completions
#define GDTC_INIT_NO_WB
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
#define MDTC_INIT(__wb, __gdtc)
static bool mdtc_valid(struct dirty_throttle_control *dtc)
{
return false;
}
static struct wb_domain *dtc_dom(struct dirty_throttle_control *dtc)
{
return &global_wb_domain;
}
static struct dirty_throttle_control *mdtc_gdtc(struct dirty_throttle_control *mdtc)
{
return NULL;
}
static struct fprop_local_percpu *wb_memcg_completions(struct bdi_writeback *wb)
{
return NULL;
}
static void wb_min_max_ratio(struct bdi_writeback *wb,
unsigned long *minp, unsigned long *maxp)
{
*minp = wb->bdi->min_ratio;
*maxp = wb->bdi->max_ratio;
}
#endif /* CONFIG_CGROUP_WRITEBACK */
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
/*
* In a memory zone, there is a certain amount of pages we consider
* available for the page cache, which is essentially the number of
* free and reclaimable pages, minus some zone reserves to protect
* lowmem and the ability to uphold the zone's watermarks without
* requiring writeback.
*
* This number of dirtyable pages is the base value of which the
* user-configurable dirty ratio is the effictive number of pages that
* are allowed to be actually dirtied. Per individual zone, or
* globally by using the sum of dirtyable pages over all zones.
*
* Because the user is allowed to specify the dirty limit globally as
* absolute number of bytes, calculating the per-zone dirty limit can
* require translating the configured limit into a percentage of
* global dirtyable memory first.
*/
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
/**
* node_dirtyable_memory - number of dirtyable pages in a node
* @pgdat: the node
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
*
* Return: the node's number of pages potentially available for dirty
* page cache. This is the base value for the per-node dirty limits.
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
*/
static unsigned long node_dirtyable_memory(struct pglist_data *pgdat)
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
{
unsigned long nr_pages = 0;
int z;
for (z = 0; z < MAX_NR_ZONES; z++) {
struct zone *zone = pgdat->node_zones + z;
if (!populated_zone(zone))
continue;
nr_pages += zone_page_state(zone, NR_FREE_PAGES);
}
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
/*
* Pages reserved for the kernel should not be considered
* dirtyable, to prevent a situation where reclaim has to
* clean pages in order to balance the zones.
*/
nr_pages -= min(nr_pages, pgdat->totalreserve_pages);
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
nr_pages += node_page_state(pgdat, NR_INACTIVE_FILE);
nr_pages += node_page_state(pgdat, NR_ACTIVE_FILE);
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
return nr_pages;
}
static unsigned long highmem_dirtyable_memory(unsigned long total)
{
#ifdef CONFIG_HIGHMEM
int node;
unsigned long x = 0;
int i;
for_each_node_state(node, N_HIGH_MEMORY) {
for (i = ZONE_NORMAL + 1; i < MAX_NR_ZONES; i++) {
struct zone *z;
unsigned long nr_pages;
if (!is_highmem_idx(i))
continue;
z = &NODE_DATA(node)->node_zones[i];
if (!populated_zone(z))
continue;
nr_pages = zone_page_state(z, NR_FREE_PAGES);
/* watch for underflows */
nr_pages -= min(nr_pages, high_wmark_pages(z));
nr_pages += zone_page_state(z, NR_ZONE_INACTIVE_FILE);
nr_pages += zone_page_state(z, NR_ZONE_ACTIVE_FILE);
x += nr_pages;
}
}
/*
* Unreclaimable memory (kernel memory or anonymous memory
* without swap) can bring down the dirtyable pages below
* the zone's dirty balance reserve and the above calculation
* will underflow. However we still want to add in nodes
* which are below threshold (negative values) to get a more
* accurate calculation but make sure that the total never
* underflows.
*/
if ((long)x < 0)
x = 0;
/*
* Make sure that the number of highmem pages is never larger
* than the number of the total dirtyable memory. This can only
* occur in very strange VM situations but we want to make sure
* that this does not occur.
*/
return min(x, total);
#else
return 0;
#endif
}
/**
* global_dirtyable_memory - number of globally dirtyable pages
*
* Return: the global number of pages potentially available for dirty
* page cache. This is the base value for the global dirty limits.
*/
static unsigned long global_dirtyable_memory(void)
{
unsigned long x;
x = global_zone_page_state(NR_FREE_PAGES);
/*
* Pages reserved for the kernel should not be considered
* dirtyable, to prevent a situation where reclaim has to
* clean pages in order to balance the zones.
*/
x -= min(x, totalreserve_pages);
mm, vmscan: move LRU lists to node This moves the LRU lists from the zone to the node and related data such as counters, tracing, congestion tracking and writeback tracking. Unfortunately, due to reclaim and compaction retry logic, it is necessary to account for the number of LRU pages on both zone and node logic. Most reclaim logic is based on the node counters but the retry logic uses the zone counters which do not distinguish inactive and active sizes. It would be possible to leave the LRU counters on a per-zone basis but it's a heavier calculation across multiple cache lines that is much more frequent than the retry checks. Other than the LRU counters, this is mostly a mechanical patch but note that it introduces a number of anomalies. For example, the scans are per-zone but using per-node counters. We also mark a node as congested when a zone is congested. This causes weird problems that are fixed later but is easier to review. In the event that there is excessive overhead on 32-bit systems due to the nodes being on LRU then there are two potential solutions 1. Long-term isolation of highmem pages when reclaim is lowmem When pages are skipped, they are immediately added back onto the LRU list. If lowmem reclaim persisted for long periods of time, the same highmem pages get continually scanned. The idea would be that lowmem keeps those pages on a separate list until a reclaim for highmem pages arrives that splices the highmem pages back onto the LRU. It potentially could be implemented similar to the UNEVICTABLE list. That would reduce the skip rate with the potential corner case is that highmem pages have to be scanned and reclaimed to free lowmem slab pages. 2. Linear scan lowmem pages if the initial LRU shrink fails This will break LRU ordering but may be preferable and faster during memory pressure than skipping LRU pages. Link: http://lkml.kernel.org/r/1467970510-21195-4-git-send-email-mgorman@techsingularity.net Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Hillf Danton <hillf.zj@alibaba-inc.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-29 06:45:31 +08:00
x += global_node_page_state(NR_INACTIVE_FILE);
x += global_node_page_state(NR_ACTIVE_FILE);
mm/page-writeback.c: fix dirty_balance_reserve subtraction from dirtyable memory Tejun reported stuttering and latency spikes on a system where random tasks would enter direct reclaim and get stuck on dirty pages. Around 50% of memory was occupied by tmpfs backed by an SSD, and another disk (rotating) was reading and writing at max speed to shrink a partition. : The problem was pretty ridiculous. It's a 8gig machine w/ one ssd and 10k : rpm harddrive and I could reliably reproduce constant stuttering every : several seconds for as long as buffered IO was going on on the hard drive : either with tmpfs occupying somewhere above 4gig or a test program which : allocates about the same amount of anon memory. Although swap usage was : zero, turning off swap also made the problem go away too. : : The trigger conditions seem quite plausible - high anon memory usage w/ : heavy buffered IO and swap configured - and it's highly likely that this : is happening in the wild too. (this can happen with copying large files : to usb sticks too, right?) This patch (of 2): The dirty_balance_reserve is an approximation of the fraction of free pages that the page allocator does not make available for page cache allocations. As a result, it has to be taken into account when calculating the amount of "dirtyable memory", the baseline to which dirty_background_ratio and dirty_ratio are applied. However, currently the reserve is subtracted from the sum of free and reclaimable pages, which is non-sensical and leads to erroneous results when the system is dominated by unreclaimable pages and the dirty_balance_reserve is bigger than free+reclaimable. In that case, at least the already allocated cache should be considered dirtyable. Fix the calculation by subtracting the reserve from the amount of free pages, then adding the reclaimable pages on top. [akpm@linux-foundation.org: fix CONFIG_HIGHMEM build] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Tested-by: Tejun Heo <tj@kernel.org> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-30 06:05:39 +08:00
if (!vm_highmem_is_dirtyable)
x -= highmem_dirtyable_memory(x);
return x + 1; /* Ensure that we never return 0 */
}
/**
* domain_dirty_limits - calculate thresh and bg_thresh for a wb_domain
* @dtc: dirty_throttle_control of interest
*
* Calculate @dtc->thresh and ->bg_thresh considering
* vm_dirty_{bytes|ratio} and dirty_background_{bytes|ratio}. The caller
* must ensure that @dtc->avail is set before calling this function. The
* dirty limits will be lifted by 1/4 for PF_LESS_THROTTLE (ie. nfsd) and
* real-time tasks.
*/
static void domain_dirty_limits(struct dirty_throttle_control *dtc)
{
const unsigned long available_memory = dtc->avail;
struct dirty_throttle_control *gdtc = mdtc_gdtc(dtc);
unsigned long bytes = vm_dirty_bytes;
unsigned long bg_bytes = dirty_background_bytes;
/* convert ratios to per-PAGE_SIZE for higher precision */
unsigned long ratio = (vm_dirty_ratio * PAGE_SIZE) / 100;
unsigned long bg_ratio = (dirty_background_ratio * PAGE_SIZE) / 100;
unsigned long thresh;
unsigned long bg_thresh;
struct task_struct *tsk;
/* gdtc is !NULL iff @dtc is for memcg domain */
if (gdtc) {
unsigned long global_avail = gdtc->avail;
/*
* The byte settings can't be applied directly to memcg
* domains. Convert them to ratios by scaling against
* globally available memory. As the ratios are in
* per-PAGE_SIZE, they can be obtained by dividing bytes by
* number of pages.
*/
if (bytes)
ratio = min(DIV_ROUND_UP(bytes, global_avail),
PAGE_SIZE);
if (bg_bytes)
bg_ratio = min(DIV_ROUND_UP(bg_bytes, global_avail),
PAGE_SIZE);
bytes = bg_bytes = 0;
}
if (bytes)
thresh = DIV_ROUND_UP(bytes, PAGE_SIZE);
else
thresh = (ratio * available_memory) / PAGE_SIZE;
if (bg_bytes)
bg_thresh = DIV_ROUND_UP(bg_bytes, PAGE_SIZE);
else
bg_thresh = (bg_ratio * available_memory) / PAGE_SIZE;
Revert "mm/page-writeback.c: print a warning if the vm dirtiness settings are illogical" This reverts commit 0f6d24f87856 ("mm/page-writeback.c: print a warning if the vm dirtiness settings are illogical") because it causes false positive warnings during OOM situations as noticed by Tetsuo Handa: Node 0 active_anon:3525940kB inactive_anon:8372kB active_file:216kB inactive_file:1872kB unevictable:0kB isolated(anon):0kB isolated(file):0kB mapped:2504kB dirty:52kB writeback:0kB shmem:8660kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 636928kB writeback_tmp:0kB unstable:0kB all_unreclaimable? yes Node 0 DMA free:14848kB min:284kB low:352kB high:420kB active_anon:992kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:15988kB managed:15904kB mlocked:0kB kernel_stack:0kB pagetables:24kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 2687 3645 3645 Node 0 DMA32 free:53004kB min:49608kB low:62008kB high:74408kB active_anon:2712648kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:3129216kB managed:2773132kB mlocked:0kB kernel_stack:96kB pagetables:5096kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 0 958 958 Node 0 Normal free:17140kB min:17684kB low:22104kB high:26524kB active_anon:812300kB inactive_anon:8372kB active_file:1228kB inactive_file:1868kB unevictable:0kB writepending:52kB present:1048576kB managed:981224kB mlocked:0kB kernel_stack:3520kB pagetables:8552kB bounce:0kB free_pcp:120kB local_pcp:120kB free_cma:0kB lowmem_reserve[]: 0 0 0 0 [...] Out of memory: Kill process 8459 (a.out) score 999 or sacrifice child Killed process 8459 (a.out) total-vm:4180kB, anon-rss:88kB, file-rss:0kB, shmem-rss:0kB oom_reaper: reaped process 8459 (a.out), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB vm direct limit must be set greater than background limit. The problem is that both thresh and bg_thresh will be 0 if available_memory is less than 4 pages when evaluating global_dirtyable_memory. While this might be worked around the whole point of the warning is dubious at best. We do rely on admins to do sensible things when changing tunable knobs. Dirty memory writeback knobs are not any special in that regards so revert the warning rather than adding more hacks to work this around. Debugged by Yafang Shao. Link: http://lkml.kernel.org/r/20171127091939.tahb77nznytcxw55@dhcp22.suse.cz Fixes: 0f6d24f87856 ("mm/page-writeback.c: print a warning if the vm dirtiness settings are illogical") Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Yafang Shao <laoar.shao@gmail.com> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-30 08:10:58 +08:00
if (bg_thresh >= thresh)
bg_thresh = thresh / 2;
tsk = current;
if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) {
MM: increase safety margin provided by PF_LESS_THROTTLE When nfsd is exporting a filesystem over NFS which is then NFS-mounted on the local machine there is a risk of deadlock. This happens when there are lots of dirty pages in the NFS filesystem and they cause NFSD to be throttled, either in throttle_vm_writeout() or in balance_dirty_pages(). To avoid this problem the PF_LESS_THROTTLE flag is set for NFSD threads and it provides a 25% increase to the limits that affect NFSD. Any process writing to an NFS filesystem will be throttled well before the number of dirty NFS pages reaches the limit imposed on NFSD, so NFSD will not deadlock on pages that it needs to write out. At least it shouldn't. All processes are allowed a small excess margin to avoid performing too many calculations: ratelimit_pages. ratelimit_pages is set so that if a thread on every CPU uses the entire margin, the total will only go 3% over the limit, and this is much less than the 25% bonus that PF_LESS_THROTTLE provides, so this margin shouldn't be a problem. But it is. The "total memory" that these 3% and 25% are calculated against are not really total memory but are "global_dirtyable_memory()" which doesn't include anonymous memory, just free memory and page-cache memory. The "ratelimit_pages" number is based on whatever the global_dirtyable_memory was on the last CPU hot-plug, which might not be what you expect, but is probably close to the total freeable memory. The throttle threshold uses the global_dirtable_memory at the moment when the throttling happens, which could be much less than at the last CPU hotplug. So if lots of anonymous memory has been allocated, thus pushing out lots of page-cache pages, then NFSD might end up being throttled due to dirty NFS pages because the "25%" bonus it gets is calculated against a rather small amount of dirtyable memory, while the "3%" margin that other processes are allowed to dirty without penalty is calculated against a much larger number. To remove this possibility of deadlock we need to make sure that the margin granted to PF_LESS_THROTTLE exceeds that rate-limit margin. Simply adding ratelimit_pages isn't enough as that should be multiplied by the number of cpus. So add "global_wb_domain.dirty_limit / 32" as that more accurately reflects the current total over-shoot margin. This ensures that the number of dirty NFS pages never gets so high that nfsd will be throttled waiting for them to be written. Link: http://lkml.kernel.org/r/87futgowwv.fsf@notabene.neil.brown.name Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-21 07:58:53 +08:00
bg_thresh += bg_thresh / 4 + global_wb_domain.dirty_limit / 32;
thresh += thresh / 4 + global_wb_domain.dirty_limit / 32;
}
dtc->thresh = thresh;
dtc->bg_thresh = bg_thresh;
/* we should eventually report the domain in the TP */
if (!gdtc)
trace_global_dirty_state(bg_thresh, thresh);
}
/**
* global_dirty_limits - background-writeback and dirty-throttling thresholds
* @pbackground: out parameter for bg_thresh
* @pdirty: out parameter for thresh
*
* Calculate bg_thresh and thresh for global_wb_domain. See
* domain_dirty_limits() for details.
*/
void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty)
{
struct dirty_throttle_control gdtc = { GDTC_INIT_NO_WB };
gdtc.avail = global_dirtyable_memory();
domain_dirty_limits(&gdtc);
*pbackground = gdtc.bg_thresh;
*pdirty = gdtc.thresh;
}
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
/**
* node_dirty_limit - maximum number of dirty pages allowed in a node
* @pgdat: the node
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
*
* Return: the maximum number of dirty pages allowed in a node, based
* on the node's dirtyable memory.
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
*/
static unsigned long node_dirty_limit(struct pglist_data *pgdat)
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
{
unsigned long node_memory = node_dirtyable_memory(pgdat);
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
struct task_struct *tsk = current;
unsigned long dirty;
if (vm_dirty_bytes)
dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE) *
node_memory / global_dirtyable_memory();
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
else
dirty = vm_dirty_ratio * node_memory / 100;
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk))
dirty += dirty / 4;
return dirty;
}
/**
* node_dirty_ok - tells whether a node is within its dirty limits
* @pgdat: the node to check
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
*
* Return: %true when the dirty pages in @pgdat are within the node's
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
* dirty limit, %false if the limit is exceeded.
*/
bool node_dirty_ok(struct pglist_data *pgdat)
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
{
unsigned long limit = node_dirty_limit(pgdat);
unsigned long nr_pages = 0;
nr_pages += node_page_state(pgdat, NR_FILE_DIRTY);
nr_pages += node_page_state(pgdat, NR_UNSTABLE_NFS);
nr_pages += node_page_state(pgdat, NR_WRITEBACK);
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
return nr_pages <= limit;
mm: try to distribute dirty pages fairly across zones The maximum number of dirty pages that exist in the system at any time is determined by a number of pages considered dirtyable and a user-configured percentage of those, or an absolute number in bytes. This number of dirtyable pages is the sum of memory provided by all the zones in the system minus their lowmem reserves and high watermarks, so that the system can retain a healthy number of free pages without having to reclaim dirty pages. But there is a flaw in that we have a zoned page allocator which does not care about the global state but rather the state of individual memory zones. And right now there is nothing that prevents one zone from filling up with dirty pages while other zones are spared, which frequently leads to situations where kswapd, in order to restore the watermark of free pages, does indeed have to write pages from that zone's LRU list. This can interfere so badly with IO from the flusher threads that major filesystems (btrfs, xfs, ext4) mostly ignore write requests from reclaim already, taking away the VM's only possibility to keep such a zone balanced, aside from hoping the flushers will soon clean pages from that zone. Enter per-zone dirty limits. They are to a zone's dirtyable memory what the global limit is to the global amount of dirtyable memory, and try to make sure that no single zone receives more than its fair share of the globally allowed dirty pages in the first place. As the number of pages considered dirtyable excludes the zones' lowmem reserves and high watermarks, the maximum number of dirty pages in a zone is such that the zone can always be balanced without requiring page cleaning. As this is a placement decision in the page allocator and pages are dirtied only after the allocation, this patch allows allocators to pass __GFP_WRITE when they know in advance that the page will be written to and become dirty soon. The page allocator will then attempt to allocate from the first zone of the zonelist - which on NUMA is determined by the task's NUMA memory policy - that has not exceeded its dirty limit. At first glance, it would appear that the diversion to lower zones can increase pressure on them, but this is not the case. With a full high zone, allocations will be diverted to lower zones eventually, so it is more of a shift in timing of the lower zone allocations. Workloads that previously could fit their dirty pages completely in the higher zone may be forced to allocate from lower zones, but the amount of pages that "spill over" are limited themselves by the lower zones' dirty constraints, and thus unlikely to become a problem. For now, the problem of unfair dirty page distribution remains for NUMA configurations where the zones allowed for allocation are in sum not big enough to trigger the global dirty limits, wake up the flusher threads and remedy the situation. Because of this, an allocation that could not succeed on any of the considered zones is allowed to ignore the dirty limits before going into direct reclaim or even failing the allocation, until a future patch changes the global dirty throttling and flusher thread activation so that they take individual zone states into account. Test results 15M DMA + 3246M DMA32 + 504 Normal = 3765M memory 40% dirty ratio 16G USB thumb drive 10 runs of dd if=/dev/zero of=disk/zeroes bs=32k count=$((10 << 15)) seconds nr_vmscan_write (stddev) min| median| max xfs vanilla: 549.747( 3.492) 0.000| 0.000| 0.000 patched: 550.996( 3.802) 0.000| 0.000| 0.000 fuse-ntfs vanilla: 1183.094(53.178) 54349.000| 59341.000| 65163.000 patched: 558.049(17.914) 0.000| 0.000| 43.000 btrfs vanilla: 573.679(14.015) 156657.000| 460178.000| 606926.000 patched: 563.365(11.368) 0.000| 0.000| 1362.000 ext4 vanilla: 561.197(15.782) 0.000|2725438.000|4143837.000 patched: 568.806(17.496) 0.000| 0.000| 0.000 Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Minchan Kim <minchan.kim@gmail.com> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Wu Fengguang <fengguang.wu@intel.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jan Kara <jack@suse.cz> Cc: Shaohua Li <shaohua.li@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:07:49 +08:00
}
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
int dirty_background_ratio_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp,
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
loff_t *ppos)
{
int ret;
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
if (ret == 0 && write)
dirty_background_bytes = 0;
return ret;
}
int dirty_background_bytes_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp,
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
loff_t *ppos)
{
int ret;
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
if (ret == 0 && write)
dirty_background_ratio = 0;
return ret;
}
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
int dirty_ratio_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp,
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
loff_t *ppos)
{
int old_ratio = vm_dirty_ratio;
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
int ret;
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
if (ret == 0 && write && vm_dirty_ratio != old_ratio) {
writeback_set_ratelimit();
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
vm_dirty_bytes = 0;
}
return ret;
}
int dirty_bytes_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp,
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
loff_t *ppos)
{
unsigned long old_bytes = vm_dirty_bytes;
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
int ret;
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
if (ret == 0 && write && vm_dirty_bytes != old_bytes) {
writeback_set_ratelimit();
mm: add dirty_background_bytes and dirty_bytes sysctls This change introduces two new sysctls to /proc/sys/vm: dirty_background_bytes and dirty_bytes. dirty_background_bytes is the counterpart to dirty_background_ratio and dirty_bytes is the counterpart to dirty_ratio. With growing memory capacities of individual machines, it's no longer sufficient to specify dirty thresholds as a percentage of the amount of dirtyable memory over the entire system. dirty_background_bytes and dirty_bytes specify quantities of memory, in bytes, that represent the dirty limits for the entire system. If either of these values is set, its value represents the amount of dirty memory that is needed to commence either background or direct writeback. When a `bytes' or `ratio' file is written, its counterpart becomes a function of the written value. For example, if dirty_bytes is written to be 8096, 8K of memory is required to commence direct writeback. dirty_ratio is then functionally equivalent to 8K / the amount of dirtyable memory: dirtyable_memory = free pages + mapped pages + file cache dirty_background_bytes = dirty_background_ratio * dirtyable_memory -or- dirty_background_ratio = dirty_background_bytes / dirtyable_memory AND dirty_bytes = dirty_ratio * dirtyable_memory -or- dirty_ratio = dirty_bytes / dirtyable_memory Only one of dirty_background_bytes and dirty_background_ratio may be specified at a time, and only one of dirty_bytes and dirty_ratio may be specified. When one sysctl is written, the other appears as 0 when read. The `bytes' files operate on a page size granularity since dirty limits are compared with ZVC values, which are in page units. Prior to this change, the minimum dirty_ratio was 5 as implemented by get_dirty_limits() although /proc/sys/vm/dirty_ratio would show any user written value between 0 and 100. This restriction is maintained, but dirty_bytes has a lower limit of only one page. Also prior to this change, the dirty_background_ratio could not equal or exceed dirty_ratio. This restriction is maintained in addition to restricting dirty_background_bytes. If either background threshold equals or exceeds that of the dirty threshold, it is implicitly set to half the dirty threshold. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andrea Righi <righi.andrea@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:31 +08:00
vm_dirty_ratio = 0;
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
return ret;
}
static unsigned long wp_next_time(unsigned long cur_time)
{
cur_time += VM_COMPLETIONS_PERIOD_LEN;
/* 0 has a special meaning... */
if (!cur_time)
return 1;
return cur_time;
}
static void wb_domain_writeout_inc(struct wb_domain *dom,
struct fprop_local_percpu *completions,
unsigned int max_prop_frac)
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
{
__fprop_inc_percpu_max(&dom->completions, completions,
max_prop_frac);
/* First event after period switching was turned off? */
if (unlikely(!dom->period_time)) {
/*
* We can race with other __bdi_writeout_inc calls here but
* it does not cause any harm since the resulting time when
* timer will fire and what is in writeout_period_time will be
* roughly the same.
*/
dom->period_time = wp_next_time(jiffies);
mod_timer(&dom->period_timer, dom->period_time);
}
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
/*
* Increment @wb's writeout completion count and the global writeout
* completion count. Called from test_clear_page_writeback().
*/
static inline void __wb_writeout_inc(struct bdi_writeback *wb)
{
struct wb_domain *cgdom;
inc_wb_stat(wb, WB_WRITTEN);
wb_domain_writeout_inc(&global_wb_domain, &wb->completions,
wb->bdi->max_prop_frac);
cgdom = mem_cgroup_wb_domain(wb);
if (cgdom)
wb_domain_writeout_inc(cgdom, wb_memcg_completions(wb),
wb->bdi->max_prop_frac);
}
void wb_writeout_inc(struct bdi_writeback *wb)
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
{
unsigned long flags;
local_irq_save(flags);
__wb_writeout_inc(wb);
local_irq_restore(flags);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
EXPORT_SYMBOL_GPL(wb_writeout_inc);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
/*
* On idle system, we can be called long after we scheduled because we use
* deferred timers so count with missed periods.
*/
static void writeout_period(struct timer_list *t)
{
struct wb_domain *dom = from_timer(dom, t, period_timer);
int miss_periods = (jiffies - dom->period_time) /
VM_COMPLETIONS_PERIOD_LEN;
if (fprop_new_period(&dom->completions, miss_periods + 1)) {
dom->period_time = wp_next_time(dom->period_time +
miss_periods * VM_COMPLETIONS_PERIOD_LEN);
mod_timer(&dom->period_timer, dom->period_time);
} else {
/*
* Aging has zeroed all fractions. Stop wasting CPU on period
* updates.
*/
dom->period_time = 0;
}
}
int wb_domain_init(struct wb_domain *dom, gfp_t gfp)
{
memset(dom, 0, sizeof(*dom));
spin_lock_init(&dom->lock);
timer_setup(&dom->period_timer, writeout_period, TIMER_DEFERRABLE);
dom->dirty_limit_tstamp = jiffies;
return fprop_global_init(&dom->completions, gfp);
}
#ifdef CONFIG_CGROUP_WRITEBACK
void wb_domain_exit(struct wb_domain *dom)
{
del_timer_sync(&dom->period_timer);
fprop_global_destroy(&dom->completions);
}
#endif
/*
* bdi_min_ratio keeps the sum of the minimum dirty shares of all
* registered backing devices, which, for obvious reasons, can not
* exceed 100%.
*/
static unsigned int bdi_min_ratio;
int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio)
{
int ret = 0;
spin_lock_bh(&bdi_lock);
if (min_ratio > bdi->max_ratio) {
ret = -EINVAL;
} else {
min_ratio -= bdi->min_ratio;
if (bdi_min_ratio + min_ratio < 100) {
bdi_min_ratio += min_ratio;
bdi->min_ratio += min_ratio;
} else {
ret = -EINVAL;
}
}
spin_unlock_bh(&bdi_lock);
return ret;
}
int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio)
{
int ret = 0;
if (max_ratio > 100)
return -EINVAL;
spin_lock_bh(&bdi_lock);
if (bdi->min_ratio > max_ratio) {
ret = -EINVAL;
} else {
bdi->max_ratio = max_ratio;
bdi->max_prop_frac = (FPROP_FRAC_BASE * max_ratio) / 100;
}
spin_unlock_bh(&bdi_lock);
return ret;
}
EXPORT_SYMBOL(bdi_set_max_ratio);
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
static unsigned long dirty_freerun_ceiling(unsigned long thresh,
unsigned long bg_thresh)
{
return (thresh + bg_thresh) / 2;
}
static unsigned long hard_dirty_limit(struct wb_domain *dom,
unsigned long thresh)
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
{
return max(thresh, dom->dirty_limit);
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
}
2015-09-30 01:04:26 +08:00
/*
* Memory which can be further allocated to a memcg domain is capped by
* system-wide clean memory excluding the amount being used in the domain.
*/
static void mdtc_calc_avail(struct dirty_throttle_control *mdtc,
unsigned long filepages, unsigned long headroom)
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
{
struct dirty_throttle_control *gdtc = mdtc_gdtc(mdtc);
2015-09-30 01:04:26 +08:00
unsigned long clean = filepages - min(filepages, mdtc->dirty);
unsigned long global_clean = gdtc->avail - min(gdtc->avail, gdtc->dirty);
unsigned long other_clean = global_clean - min(global_clean, clean);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
2015-09-30 01:04:26 +08:00
mdtc->avail = filepages + min(headroom, other_clean);
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
}
/**
* __wb_calc_thresh - @wb's share of dirty throttling threshold
* @dtc: dirty_throttle_context of interest
*
* Note that balance_dirty_pages() will only seriously take it as a hard limit
* when sleeping max_pause per page is not enough to keep the dirty pages under
* control. For example, when the device is completely stalled due to some error
* conditions, or when there are 1000 dd tasks writing to a slow 10MB/s USB key.
* In the other normal situations, it acts more gently by throttling the tasks
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
* more (rather than completely block them) when the wb dirty pages go high.
*
* It allocates high/low dirty limits to fast/slow devices, in order to prevent
* - starving fast devices
* - piling up dirty pages (that will take long time to sync) on slow devices
*
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
* The wb's share of dirty limit will be adapting to its throughput and
* bounded by the bdi->min_ratio and/or bdi->max_ratio parameters, if set.
*
* Return: @wb's dirty limit in pages. The term "dirty" in the context of
* dirty balancing includes all PG_dirty, PG_writeback and NFS unstable pages.
*/
static unsigned long __wb_calc_thresh(struct dirty_throttle_control *dtc)
{
struct wb_domain *dom = dtc_dom(dtc);
unsigned long thresh = dtc->thresh;
u64 wb_thresh;
unsigned long numerator, denominator;
unsigned long wb_min_ratio, wb_max_ratio;
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
/*
* Calculate this BDI's share of the thresh ratio.
*/
fprop_fraction_percpu(&dom->completions, dtc->wb_completions,
&numerator, &denominator);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
wb_thresh = (thresh * (100 - bdi_min_ratio)) / 100;
wb_thresh *= numerator;
wb_thresh = div64_ul(wb_thresh, denominator);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
wb_min_max_ratio(dtc->wb, &wb_min_ratio, &wb_max_ratio);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
wb_thresh += (thresh * wb_min_ratio) / 100;
if (wb_thresh > (thresh * wb_max_ratio) / 100)
wb_thresh = thresh * wb_max_ratio / 100;
return wb_thresh;
}
unsigned long wb_calc_thresh(struct bdi_writeback *wb, unsigned long thresh)
{
struct dirty_throttle_control gdtc = { GDTC_INIT(wb),
.thresh = thresh };
return __wb_calc_thresh(&gdtc);
}
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* setpoint - dirty 3
* f(dirty) := 1.0 + (----------------)
* limit - setpoint
*
* it's a 3rd order polynomial that subjects to
*
* (1) f(freerun) = 2.0 => rampup dirty_ratelimit reasonably fast
* (2) f(setpoint) = 1.0 => the balance point
* (3) f(limit) = 0 => the hard limit
* (4) df/dx <= 0 => negative feedback control
* (5) the closer to setpoint, the smaller |df/dx| (and the reverse)
* => fast response on large errors; small oscillation near setpoint
*/
static long long pos_ratio_polynom(unsigned long setpoint,
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
unsigned long dirty,
unsigned long limit)
{
long long pos_ratio;
long x;
x = div64_s64(((s64)setpoint - (s64)dirty) << RATELIMIT_CALC_SHIFT,
(limit - setpoint) | 1);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
pos_ratio = x;
pos_ratio = pos_ratio * x >> RATELIMIT_CALC_SHIFT;
pos_ratio = pos_ratio * x >> RATELIMIT_CALC_SHIFT;
pos_ratio += 1 << RATELIMIT_CALC_SHIFT;
return clamp(pos_ratio, 0LL, 2LL << RATELIMIT_CALC_SHIFT);
}
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
/*
* Dirty position control.
*
* (o) global/bdi setpoints
*
* We want the dirty pages be balanced around the global/wb setpoints.
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* When the number of dirty pages is higher/lower than the setpoint, the
* dirty position control ratio (and hence task dirty ratelimit) will be
* decreased/increased to bring the dirty pages back to the setpoint.
*
* pos_ratio = 1 << RATELIMIT_CALC_SHIFT
*
* if (dirty < setpoint) scale up pos_ratio
* if (dirty > setpoint) scale down pos_ratio
*
* if (wb_dirty < wb_setpoint) scale up pos_ratio
* if (wb_dirty > wb_setpoint) scale down pos_ratio
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* task_ratelimit = dirty_ratelimit * pos_ratio >> RATELIMIT_CALC_SHIFT
*
* (o) global control line
*
* ^ pos_ratio
* |
* | |<===== global dirty control scope ======>|
* 2.0 .............*
* | .*
* | . *
* | . *
* | . *
* | . *
* | . *
* 1.0 ................................*
* | . . *
* | . . *
* | . . *
* | . . *
* | . . *
* 0 +------------.------------------.----------------------*------------->
* freerun^ setpoint^ limit^ dirty pages
*
* (o) wb control line
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* ^ pos_ratio
* |
* | *
* | *
* | *
* | *
* | * |<=========== span ============>|
* 1.0 .......................*
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* | . *
* 1/4 ...............................................* * * * * * * * * * * *
* | . .
* | . .
* | . .
* 0 +----------------------.-------------------------------.------------->
* wb_setpoint^ x_intercept^
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* The wb control line won't drop below pos_ratio=1/4, so that wb_dirty can
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* be smoothly throttled down to normal if it starts high in situations like
* - start writing to a slow SD card and a fast disk at the same time. The SD
* card's wb_dirty may rush to many times higher than wb_setpoint.
* - the wb dirty thresh drops quickly due to change of JBOD workload
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*/
static void wb_position_ratio(struct dirty_throttle_control *dtc)
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
{
struct bdi_writeback *wb = dtc->wb;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long write_bw = wb->avg_write_bandwidth;
unsigned long freerun = dirty_freerun_ceiling(dtc->thresh, dtc->bg_thresh);
unsigned long limit = hard_dirty_limit(dtc_dom(dtc), dtc->thresh);
unsigned long wb_thresh = dtc->wb_thresh;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
unsigned long x_intercept;
unsigned long setpoint; /* dirty pages' target balance point */
unsigned long wb_setpoint;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
unsigned long span;
long long pos_ratio; /* for scaling up/down the rate limit */
long x;
dtc->pos_ratio = 0;
if (unlikely(dtc->dirty >= limit))
return;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
/*
* global setpoint
*
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* See comment for pos_ratio_polynom().
*/
setpoint = (freerun + limit) / 2;
pos_ratio = pos_ratio_polynom(setpoint, dtc->dirty, limit);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* The strictlimit feature is a tool preventing mistrusted filesystems
* from growing a large number of dirty pages before throttling. For
* such filesystems balance_dirty_pages always checks wb counters
* against wb limits. Even if global "nr_dirty" is under "freerun".
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* This is especially important for fuse which sets bdi->max_ratio to
* 1% by default. Without strictlimit feature, fuse writeback may
* consume arbitrary amount of RAM because it is accounted in
* NR_WRITEBACK_TEMP which is not involved in calculating "nr_dirty".
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
* Here, in wb_position_ratio(), we calculate pos_ratio based on
* two values: wb_dirty and wb_thresh. Let's consider an example:
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* total amount of RAM is 16GB, bdi->max_ratio is equal to 1%, global
* limits are set by default to 10% and 20% (background and throttle).
* Then wb_thresh is 1% of 20% of 16GB. This amounts to ~8K pages.
* wb_calc_thresh(wb, bg_thresh) is about ~4K pages. wb_setpoint is
* about ~6K pages (as the average of background and throttle wb
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* limits). The 3rd order polynomial will provide positive feedback if
* wb_dirty is under wb_setpoint and vice versa.
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* Note, that we cannot use global counters in these calculations
* because we want to throttle process writing to a strictlimit wb
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* much earlier than global "freerun" is reached (~23MB vs. ~2.3GB
* in the example above).
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*/
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (unlikely(wb->bdi->capabilities & BDI_CAP_STRICTLIMIT)) {
long long wb_pos_ratio;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
if (dtc->wb_dirty < 8) {
dtc->pos_ratio = min_t(long long, pos_ratio * 2,
2 << RATELIMIT_CALC_SHIFT);
return;
}
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
if (dtc->wb_dirty >= wb_thresh)
return;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
wb_setpoint = dirty_freerun_ceiling(wb_thresh,
dtc->wb_bg_thresh);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
if (wb_setpoint == 0 || wb_setpoint == wb_thresh)
return;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
wb_pos_ratio = pos_ratio_polynom(wb_setpoint, dtc->wb_dirty,
wb_thresh);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* Typically, for strictlimit case, wb_setpoint << setpoint
* and pos_ratio >> wb_pos_ratio. In the other words global
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* state ("dirty") is not limiting factor and we have to
* make decision based on wb counters. But there is an
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* important case when global pos_ratio should get precedence:
* global limits are exceeded (e.g. due to activities on other
* wb's) while given strictlimit wb is below limit.
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
*
* "pos_ratio * wb_pos_ratio" would work for the case above,
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* but it would look too non-natural for the case of all
* activity in the system coming from a single strictlimit wb
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* with bdi->max_ratio == 100%.
*
* Note that min() below somewhat changes the dynamics of the
* control system. Normally, pos_ratio value can be well over 3
* (when globally we are at freerun and wb is well below wb
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* setpoint). Now the maximum pos_ratio in the same situation
* is 2. We might want to tweak this if we observe the control
* system is too slow to adapt.
*/
dtc->pos_ratio = min(pos_ratio, wb_pos_ratio);
return;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
}
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
/*
* We have computed basic pos_ratio above based on global situation. If
* the wb is over/under its share of dirty pages, we want to scale
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* pos_ratio further down/up. That is done by the following mechanism.
*/
/*
* wb setpoint
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* f(wb_dirty) := 1.0 + k * (wb_dirty - wb_setpoint)
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* x_intercept - wb_dirty
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* := --------------------------
* x_intercept - wb_setpoint
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* The main wb control line is a linear function that subjects to
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* (1) f(wb_setpoint) = 1.0
* (2) k = - 1 / (8 * write_bw) (in single wb case)
* or equally: x_intercept = wb_setpoint + 8 * write_bw
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* For single wb case, the dirty pages are observed to fluctuate
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* regularly within range
* [wb_setpoint - write_bw/2, wb_setpoint + write_bw/2]
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* for various filesystems, where (2) can yield in a reasonable 12.5%
* fluctuation range for pos_ratio.
*
* For JBOD case, wb_thresh (not wb_dirty!) could fluctuate up to its
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
* own size, so move the slope over accordingly and choose a slope that
* yields 100% pos_ratio fluctuation on suddenly doubled wb_thresh.
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*/
if (unlikely(wb_thresh > dtc->thresh))
wb_thresh = dtc->thresh;
/*
* It's very possible that wb_thresh is close to 0 not because the
* device is slow, but that it has remained inactive for long time.
* Honour such devices a reasonable good (hopefully IO efficient)
* threshold, so that the occasional writes won't be blocked and active
* writes can rampup the threshold quickly.
*/
wb_thresh = max(wb_thresh, (limit - dtc->dirty) / 8);
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
/*
* scale global setpoint to wb's:
* wb_setpoint = setpoint * wb_thresh / thresh
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*/
Merge branch 'for-4.2/writeback' of git://git.kernel.dk/linux-block Pull cgroup writeback support from Jens Axboe: "This is the big pull request for adding cgroup writeback support. This code has been in development for a long time, and it has been simmering in for-next for a good chunk of this cycle too. This is one of those problems that has been talked about for at least half a decade, finally there's a solution and code to go with it. Also see last weeks writeup on LWN: http://lwn.net/Articles/648292/" * 'for-4.2/writeback' of git://git.kernel.dk/linux-block: (85 commits) writeback, blkio: add documentation for cgroup writeback support vfs, writeback: replace FS_CGROUP_WRITEBACK with SB_I_CGROUPWB writeback: do foreign inode detection iff cgroup writeback is enabled v9fs: fix error handling in v9fs_session_init() bdi: fix wrong error return value in cgwb_create() buffer: remove unusued 'ret' variable writeback: disassociate inodes from dying bdi_writebacks writeback: implement foreign cgroup inode bdi_writeback switching writeback: add lockdep annotation to inode_to_wb() writeback: use unlocked_inode_to_wb transaction in inode_congested() writeback: implement unlocked_inode_to_wb transaction and use it for stat updates writeback: implement [locked_]inode_to_wb_and_lock_list() writeback: implement foreign cgroup inode detection writeback: make writeback_control track the inode being written back writeback: relocate wb[_try]_get(), wb_put(), inode_{attach|detach}_wb() mm: vmscan: disable memcg direct reclaim stalling if cgroup writeback support is in use writeback: implement memcg writeback domain based throttling writeback: reset wb_domain->dirty_limit[_tstmp] when memcg domain size changes writeback: implement memcg wb_domain writeback: update wb_over_bg_thresh() to use wb_domain aware operations ...
2015-06-26 07:00:17 +08:00
x = div_u64((u64)wb_thresh << 16, dtc->thresh | 1);
wb_setpoint = setpoint * (u64)x >> 16;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
/*
* Use span=(8*write_bw) in single wb case as indicated by
* (thresh - wb_thresh ~= 0) and transit to wb_thresh in JBOD case.
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*
* wb_thresh thresh - wb_thresh
* span = --------- * (8 * write_bw) + ------------------ * wb_thresh
* thresh thresh
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
*/
span = (dtc->thresh - wb_thresh + 8 * write_bw) * (u64)x >> 16;
x_intercept = wb_setpoint + span;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
if (dtc->wb_dirty < x_intercept - span / 4) {
pos_ratio = div64_u64(pos_ratio * (x_intercept - dtc->wb_dirty),
Merge branch 'for-4.2/writeback' of git://git.kernel.dk/linux-block Pull cgroup writeback support from Jens Axboe: "This is the big pull request for adding cgroup writeback support. This code has been in development for a long time, and it has been simmering in for-next for a good chunk of this cycle too. This is one of those problems that has been talked about for at least half a decade, finally there's a solution and code to go with it. Also see last weeks writeup on LWN: http://lwn.net/Articles/648292/" * 'for-4.2/writeback' of git://git.kernel.dk/linux-block: (85 commits) writeback, blkio: add documentation for cgroup writeback support vfs, writeback: replace FS_CGROUP_WRITEBACK with SB_I_CGROUPWB writeback: do foreign inode detection iff cgroup writeback is enabled v9fs: fix error handling in v9fs_session_init() bdi: fix wrong error return value in cgwb_create() buffer: remove unusued 'ret' variable writeback: disassociate inodes from dying bdi_writebacks writeback: implement foreign cgroup inode bdi_writeback switching writeback: add lockdep annotation to inode_to_wb() writeback: use unlocked_inode_to_wb transaction in inode_congested() writeback: implement unlocked_inode_to_wb transaction and use it for stat updates writeback: implement [locked_]inode_to_wb_and_lock_list() writeback: implement foreign cgroup inode detection writeback: make writeback_control track the inode being written back writeback: relocate wb[_try]_get(), wb_put(), inode_{attach|detach}_wb() mm: vmscan: disable memcg direct reclaim stalling if cgroup writeback support is in use writeback: implement memcg writeback domain based throttling writeback: reset wb_domain->dirty_limit[_tstmp] when memcg domain size changes writeback: implement memcg wb_domain writeback: update wb_over_bg_thresh() to use wb_domain aware operations ...
2015-06-26 07:00:17 +08:00
(x_intercept - wb_setpoint) | 1);
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
} else
pos_ratio /= 4;
/*
* wb reserve area, safeguard against dirty pool underrun and disk idle
* It may push the desired control point of global dirty pages higher
* than setpoint.
*/
x_intercept = wb_thresh / 2;
if (dtc->wb_dirty < x_intercept) {
if (dtc->wb_dirty > x_intercept / 8)
pos_ratio = div_u64(pos_ratio * x_intercept,
dtc->wb_dirty);
else
pos_ratio *= 8;
}
dtc->pos_ratio = pos_ratio;
writeback: dirty position control bdi_position_ratio() provides a scale factor to bdi->dirty_ratelimit, so that the resulted task rate limit can drive the dirty pages back to the global/bdi setpoints. Old scheme is, | free run area | throttle area ----------------------------------------+----------------------------> thresh^ dirty pages New scheme is, ^ task rate limit | | * | * | * |[free run] * [smooth throttled] | * | * | * ..bdi->dirty_ratelimit..........* | . * | . * | . * | . * | . * +-------------------------------.-----------------------*------------> setpoint^ limit^ dirty pages The slope of the bdi control line should be 1) large enough to pull the dirty pages to setpoint reasonably fast 2) small enough to avoid big fluctuations in the resulted pos_ratio and hence task ratelimit Since the fluctuation range of the bdi dirty pages is typically observed to be within 1-second worth of data, the bdi control line's slope is selected to be a linear function of bdi write bandwidth, so that it can adapt to slow/fast storage devices well. Assume the bdi control line pos_ratio = 1.0 + k * (dirty - bdi_setpoint) where k is the negative slope. If targeting for 12.5% fluctuation range in pos_ratio when dirty pages are fluctuating in range [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2], we get slope k = - 1 / (8 * write_bw) Let pos_ratio(x_intercept) = 0, we get the parameter used in code: x_intercept = bdi_setpoint + 8 * write_bw The global/bdi slopes are nicely complementing each other when the system has only one major bdi (indicated by bdi_thresh ~= thresh): 1) slope of global control line => scaling to the control scope size 2) slope of main bdi control line => scaling to the writeout bandwidth so that - in memory tight systems, (1) becomes strong enough to squeeze dirty pages inside the control scope - in large memory systems where the "gravity" of (1) for pulling the dirty pages to setpoint is too weak, (2) can back (1) up and drive dirty pages to bdi_setpoint ~= setpoint reasonably fast. Unfortunately in JBOD setups, the fluctuation range of bdi threshold is related to memory size due to the interferences between disks. In this case, the bdi slope will be weighted sum of write_bw and bdi_thresh. Given equations span = x_intercept - bdi_setpoint k = df/dx = - 1 / span and the extremum values span = bdi_thresh dx = bdi_thresh we get df = - dx / span = - 1.0 That means, when bdi_dirty deviates bdi_thresh up, pos_ratio and hence task ratelimit will fluctuate by -100%. peter: use 3rd order polynomial for the global control line CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 06:04:18 +08:00
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
static void wb_update_write_bandwidth(struct bdi_writeback *wb,
unsigned long elapsed,
unsigned long written)
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
{
const unsigned long period = roundup_pow_of_two(3 * HZ);
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long avg = wb->avg_write_bandwidth;
unsigned long old = wb->write_bandwidth;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
u64 bw;
/*
* bw = written * HZ / elapsed
*
* bw * elapsed + write_bandwidth * (period - elapsed)
* write_bandwidth = ---------------------------------------------------
* period
*
* @written may have decreased due to account_page_redirty().
* Avoid underflowing @bw calculation.
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
*/
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
bw = written - min(written, wb->written_stamp);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
bw *= HZ;
if (unlikely(elapsed > period)) {
bw = div64_ul(bw, elapsed);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
avg = bw;
goto out;
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
bw += (u64)wb->write_bandwidth * (period - elapsed);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
bw >>= ilog2(period);
/*
* one more level of smoothing, for filtering out sudden spikes
*/
if (avg > old && old >= (unsigned long)bw)
avg -= (avg - old) >> 3;
if (avg < old && old <= (unsigned long)bw)
avg += (old - avg) >> 3;
out:
/* keep avg > 0 to guarantee that tot > 0 if there are dirty wbs */
avg = max(avg, 1LU);
if (wb_has_dirty_io(wb)) {
long delta = avg - wb->avg_write_bandwidth;
WARN_ON_ONCE(atomic_long_add_return(delta,
&wb->bdi->tot_write_bandwidth) <= 0);
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
wb->write_bandwidth = bw;
wb->avg_write_bandwidth = avg;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
}
static void update_dirty_limit(struct dirty_throttle_control *dtc)
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
{
struct wb_domain *dom = dtc_dom(dtc);
unsigned long thresh = dtc->thresh;
unsigned long limit = dom->dirty_limit;
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
/*
* Follow up in one step.
*/
if (limit < thresh) {
limit = thresh;
goto update;
}
/*
* Follow down slowly. Use the higher one as the target, because thresh
* may drop below dirty. This is exactly the reason to introduce
* dom->dirty_limit which is guaranteed to lie above the dirty pages.
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
*/
thresh = max(thresh, dtc->dirty);
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
if (limit > thresh) {
limit -= (limit - thresh) >> 5;
goto update;
}
return;
update:
dom->dirty_limit = limit;
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
}
static void domain_update_bandwidth(struct dirty_throttle_control *dtc,
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
unsigned long now)
{
struct wb_domain *dom = dtc_dom(dtc);
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
/*
* check locklessly first to optimize away locking for the most time
*/
if (time_before(now, dom->dirty_limit_tstamp + BANDWIDTH_INTERVAL))
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
return;
spin_lock(&dom->lock);
if (time_after_eq(now, dom->dirty_limit_tstamp + BANDWIDTH_INTERVAL)) {
update_dirty_limit(dtc);
dom->dirty_limit_tstamp = now;
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
}
spin_unlock(&dom->lock);
writeback: introduce smoothed global dirty limit The start of a heavy weight application (ie. KVM) may instantly knock down determine_dirtyable_memory() if the swap is not enabled or full. global_dirty_limits() and bdi_dirty_limit() will in turn get global/bdi dirty thresholds that are _much_ lower than the global/bdi dirty pages. balance_dirty_pages() will then heavily throttle all dirtiers including the light ones, until the dirty pages drop below the new dirty thresholds. During this _deep_ dirty-exceeded state, the system may appear rather unresponsive to the users. About "deep" dirty-exceeded: task_dirty_limit() assigns 1/8 lower dirty threshold to heavy dirtiers than light ones, and the dirty pages will be throttled around the heavy dirtiers' dirty threshold and reasonably below the light dirtiers' dirty threshold. In this state, only the heavy dirtiers will be throttled and the dirty pages are carefully controlled to not exceed the light dirtiers' dirty threshold. However if the threshold itself suddenly drops below the number of dirty pages, the light dirtiers will get heavily throttled. So introduce global_dirty_limit for tracking the global dirty threshold with policies - follow downwards slowly - follow up in one shot global_dirty_limit can effectively mask out the impact of sudden drop of dirtyable memory. It will be used in the next patch for two new type of dirty limits. Note that the new dirty limits are not going to avoid throttling the light dirtiers, but could limit their sleep time to 200ms. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-03-03 05:54:09 +08:00
}
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
/*
* Maintain wb->dirty_ratelimit, the base dirty throttle rate.
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
*
* Normal wb tasks will be curbed at or below it in long term.
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
* Obviously it should be around (write_bw / N) when there are N dd tasks.
*/
static void wb_update_dirty_ratelimit(struct dirty_throttle_control *dtc,
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long dirtied,
unsigned long elapsed)
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
{
struct bdi_writeback *wb = dtc->wb;
unsigned long dirty = dtc->dirty;
unsigned long freerun = dirty_freerun_ceiling(dtc->thresh, dtc->bg_thresh);
unsigned long limit = hard_dirty_limit(dtc_dom(dtc), dtc->thresh);
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
unsigned long setpoint = (freerun + limit) / 2;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long write_bw = wb->avg_write_bandwidth;
unsigned long dirty_ratelimit = wb->dirty_ratelimit;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
unsigned long dirty_rate;
unsigned long task_ratelimit;
unsigned long balanced_dirty_ratelimit;
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
unsigned long step;
unsigned long x;
unsigned long shift;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
/*
* The dirty rate will match the writeout rate in long term, except
* when dirty pages are truncated by userspace or re-dirtied by FS.
*/
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
dirty_rate = (dirtied - wb->dirtied_stamp) * HZ / elapsed;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
/*
* task_ratelimit reflects each dd's dirty rate for the past 200ms.
*/
task_ratelimit = (u64)dirty_ratelimit *
dtc->pos_ratio >> RATELIMIT_CALC_SHIFT;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
task_ratelimit++; /* it helps rampup dirty_ratelimit from tiny values */
/*
* A linear estimation of the "balanced" throttle rate. The theory is,
* if there are N dd tasks, each throttled at task_ratelimit, the wb's
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
* dirty_rate will be measured to be (N * task_ratelimit). So the below
* formula will yield the balanced rate limit (write_bw / N).
*
* Note that the expanded form is not a pure rate feedback:
* rate_(i+1) = rate_(i) * (write_bw / dirty_rate) (1)
* but also takes pos_ratio into account:
* rate_(i+1) = rate_(i) * (write_bw / dirty_rate) * pos_ratio (2)
*
* (1) is not realistic because pos_ratio also takes part in balancing
* the dirty rate. Consider the state
* pos_ratio = 0.5 (3)
* rate = 2 * (write_bw / N) (4)
* If (1) is used, it will stuck in that state! Because each dd will
* be throttled at
* task_ratelimit = pos_ratio * rate = (write_bw / N) (5)
* yielding
* dirty_rate = N * task_ratelimit = write_bw (6)
* put (6) into (1) we get
* rate_(i+1) = rate_(i) (7)
*
* So we end up using (2) to always keep
* rate_(i+1) ~= (write_bw / N) (8)
* regardless of the value of pos_ratio. As long as (8) is satisfied,
* pos_ratio is able to drive itself to 1.0, which is not only where
* the dirty count meet the setpoint, but also where the slope of
* pos_ratio is most flat and hence task_ratelimit is least fluctuated.
*/
balanced_dirty_ratelimit = div_u64((u64)task_ratelimit * write_bw,
dirty_rate | 1);
writeback: balanced_rate cannot exceed write bandwidth Add an upper limit to balanced_rate according to the below inequality. This filters out some rare but huge singular points, which at least enables more readable gnuplot figures. When there are N dd dirtiers, balanced_dirty_ratelimit = write_bw / N So it holds that balanced_dirty_ratelimit <= write_bw The singular points originate from dirty_rate in the below formular: balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate where dirty_rate = (number of page dirties in the past 200ms) / 200ms In the extreme case, if all dd tasks suddenly get blocked on something else and hence no pages are dirtied at all, dirty_rate will be 0 and balanced_dirty_ratelimit will be inf. This could happen in reality. Note that these huge singular points are not a real threat, since they are _guaranteed_ to be filtered out by the min(balanced_dirty_ratelimit, task_ratelimit) line in bdi_update_dirty_ratelimit(). task_ratelimit is based on the number of dirty pages, which will never _suddenly_ fly away like balanced_dirty_ratelimit. So any weirdly large balanced_dirty_ratelimit will be cut down to the level of task_ratelimit. There won't be tiny singular points though, as long as the dirty pages lie inside the dirty throttling region (above the freerun region). Because there the dd tasks will be throttled by balanced_dirty_pages() and won't be able to suddenly dirty much more pages than average. Acked-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-04 04:30:36 +08:00
/*
* balanced_dirty_ratelimit ~= (write_bw / N) <= write_bw
*/
if (unlikely(balanced_dirty_ratelimit > write_bw))
balanced_dirty_ratelimit = write_bw;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
/*
* We could safely do this and return immediately:
*
* wb->dirty_ratelimit = balanced_dirty_ratelimit;
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
*
* However to get a more stable dirty_ratelimit, the below elaborated
* code makes use of task_ratelimit to filter out singular points and
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
* limit the step size.
*
* The below code essentially only uses the relative value of
*
* task_ratelimit - dirty_ratelimit
* = (pos_ratio - 1) * dirty_ratelimit
*
* which reflects the direction and size of dirty position error.
*/
/*
* dirty_ratelimit will follow balanced_dirty_ratelimit iff
* task_ratelimit is on the same side of dirty_ratelimit, too.
* For example, when
* - dirty_ratelimit > balanced_dirty_ratelimit
* - dirty_ratelimit > task_ratelimit (dirty pages are above setpoint)
* lowering dirty_ratelimit will help meet both the position and rate
* control targets. Otherwise, don't update dirty_ratelimit if it will
* only help meet the rate target. After all, what the users ultimately
* feel and care are stable dirty rate and small position error.
*
* |task_ratelimit - dirty_ratelimit| is used to limit the step size
* and filter out the singular points of balanced_dirty_ratelimit. Which
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
* keeps jumping around randomly and can even leap far away at times
* due to the small 200ms estimation period of dirty_rate (we want to
* keep that period small to reduce time lags).
*/
step = 0;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* For strictlimit case, calculations above were based on wb counters
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
* and limits (starting from pos_ratio = wb_position_ratio() and up to
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate).
* Hence, to calculate "step" properly, we have to use wb_dirty as
* "dirty" and wb_setpoint as "setpoint".
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
*
* We rampup dirty_ratelimit forcibly if wb_dirty is low because
* it's possible that wb_thresh is close to zero due to inactivity
* of backing device.
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
*/
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (unlikely(wb->bdi->capabilities & BDI_CAP_STRICTLIMIT)) {
dirty = dtc->wb_dirty;
if (dtc->wb_dirty < 8)
setpoint = dtc->wb_dirty + 1;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
else
setpoint = (dtc->wb_thresh + dtc->wb_bg_thresh) / 2;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
}
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
if (dirty < setpoint) {
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
x = min3(wb->balanced_dirty_ratelimit,
balanced_dirty_ratelimit, task_ratelimit);
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
if (dirty_ratelimit < x)
step = x - dirty_ratelimit;
} else {
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
x = max3(wb->balanced_dirty_ratelimit,
balanced_dirty_ratelimit, task_ratelimit);
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
if (dirty_ratelimit > x)
step = dirty_ratelimit - x;
}
/*
* Don't pursue 100% rate matching. It's impossible since the balanced
* rate itself is constantly fluctuating. So decrease the track speed
* when it gets close to the target. Helps eliminate pointless tremors.
*/
shift = dirty_ratelimit / (2 * step + 1);
if (shift < BITS_PER_LONG)
step = DIV_ROUND_UP(step >> shift, 8);
else
step = 0;
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-27 05:53:24 +08:00
if (dirty_ratelimit < balanced_dirty_ratelimit)
dirty_ratelimit += step;
else
dirty_ratelimit -= step;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
wb->dirty_ratelimit = max(dirty_ratelimit, 1UL);
wb->balanced_dirty_ratelimit = balanced_dirty_ratelimit;
trace_bdi_dirty_ratelimit(wb, dirty_rate, task_ratelimit);
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
}
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
static void __wb_update_bandwidth(struct dirty_throttle_control *gdtc,
struct dirty_throttle_control *mdtc,
unsigned long start_time,
bool update_ratelimit)
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
{
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
struct bdi_writeback *wb = gdtc->wb;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
unsigned long now = jiffies;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long elapsed = now - wb->bw_time_stamp;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
unsigned long dirtied;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
unsigned long written;
lockdep_assert_held(&wb->list_lock);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
/*
* rate-limit, only update once every 200ms.
*/
if (elapsed < BANDWIDTH_INTERVAL)
return;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
dirtied = percpu_counter_read(&wb->stat[WB_DIRTIED]);
written = percpu_counter_read(&wb->stat[WB_WRITTEN]);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
/*
* Skip quiet periods when disk bandwidth is under-utilized.
* (at least 1s idle time between two flusher runs)
*/
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (elapsed > HZ && time_before(wb->bw_time_stamp, start_time))
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
goto snapshot;
if (update_ratelimit) {
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
domain_update_bandwidth(gdtc, now);
wb_update_dirty_ratelimit(gdtc, dirtied, elapsed);
/*
* @mdtc is always NULL if !CGROUP_WRITEBACK but the
* compiler has no way to figure that out. Help it.
*/
if (IS_ENABLED(CONFIG_CGROUP_WRITEBACK) && mdtc) {
domain_update_bandwidth(mdtc, now);
wb_update_dirty_ratelimit(mdtc, dirtied, elapsed);
}
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-13 00:51:31 +08:00
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
wb_update_write_bandwidth(wb, elapsed, written);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
snapshot:
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
wb->dirtied_stamp = dirtied;
wb->written_stamp = written;
wb->bw_time_stamp = now;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
}
void wb_update_bandwidth(struct bdi_writeback *wb, unsigned long start_time)
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
{
struct dirty_throttle_control gdtc = { GDTC_INIT(wb) };
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
__wb_update_bandwidth(&gdtc, NULL, start_time, false);
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
}
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
/*
* After a task dirtied this many pages, balance_dirty_pages_ratelimited()
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
* will look to see if it needs to start dirty throttling.
*
* If dirty_poll_interval is too low, big NUMA machines will call the expensive
* global_zone_page_state() too often. So scale it near-sqrt to the safety margin
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
* (the number of pages we may dirty without exceeding the dirty limits).
*/
static unsigned long dirty_poll_interval(unsigned long dirty,
unsigned long thresh)
{
if (thresh > dirty)
return 1UL << (ilog2(thresh - dirty) >> 1);
return 1;
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
static unsigned long wb_max_pause(struct bdi_writeback *wb,
unsigned long wb_dirty)
{
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
unsigned long bw = wb->avg_write_bandwidth;
writeback: fix negative bdi max pause Toralf runs trinity on UML/i386. After some time it hangs and the last message line is BUG: soft lockup - CPU#0 stuck for 22s! [trinity-child0:1521] It's found that pages_dirtied becomes very large. More than 1000000000 pages in this case: period = HZ * pages_dirtied / task_ratelimit; BUG_ON(pages_dirtied > 2000000000); BUG_ON(pages_dirtied > 1000000000); <--------- UML debug printf shows that we got negative pause here: ick: pause : -984 ick: pages_dirtied : 0 ick: task_ratelimit: 0 pause: + if (pause < 0) { + extern int printf(char *, ...); + printf("ick : pause : %li\n", pause); + printf("ick: pages_dirtied : %lu\n", pages_dirtied); + printf("ick: task_ratelimit: %lu\n", task_ratelimit); + BUG_ON(1); + } trace_balance_dirty_pages(bdi, Since pause is bounded by [min_pause, max_pause] where min_pause is also bounded by max_pause. It's suspected and demonstrated that the max_pause calculation goes wrong: ick: pause : -717 ick: min_pause : -177 ick: max_pause : -717 ick: pages_dirtied : 14 ick: task_ratelimit: 0 The problem lies in the two "long = unsigned long" assignments in bdi_max_pause() which might go negative if the highest bit is 1, and the min_t(long, ...) check failed to protect it falling under 0. Fix all of them by using "unsigned long" throughout the function. Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Reported-by: Toralf Förster <toralf.foerster@gmx.de> Tested-by: Toralf Förster <toralf.foerster@gmx.de> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Richard Weinberger <richard@nod.at> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-10-17 04:47:03 +08:00
unsigned long t;
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
/*
* Limit pause time for small memory systems. If sleeping for too long
* time, a small pool of dirty/writeback pages may go empty and disk go
* idle.
*
* 8 serves as the safety ratio.
*/
t = wb_dirty / (1 + bw / roundup_pow_of_two(1 + HZ / 8));
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
t++;
writeback: fix negative bdi max pause Toralf runs trinity on UML/i386. After some time it hangs and the last message line is BUG: soft lockup - CPU#0 stuck for 22s! [trinity-child0:1521] It's found that pages_dirtied becomes very large. More than 1000000000 pages in this case: period = HZ * pages_dirtied / task_ratelimit; BUG_ON(pages_dirtied > 2000000000); BUG_ON(pages_dirtied > 1000000000); <--------- UML debug printf shows that we got negative pause here: ick: pause : -984 ick: pages_dirtied : 0 ick: task_ratelimit: 0 pause: + if (pause < 0) { + extern int printf(char *, ...); + printf("ick : pause : %li\n", pause); + printf("ick: pages_dirtied : %lu\n", pages_dirtied); + printf("ick: task_ratelimit: %lu\n", task_ratelimit); + BUG_ON(1); + } trace_balance_dirty_pages(bdi, Since pause is bounded by [min_pause, max_pause] where min_pause is also bounded by max_pause. It's suspected and demonstrated that the max_pause calculation goes wrong: ick: pause : -717 ick: min_pause : -177 ick: max_pause : -717 ick: pages_dirtied : 14 ick: task_ratelimit: 0 The problem lies in the two "long = unsigned long" assignments in bdi_max_pause() which might go negative if the highest bit is 1, and the min_t(long, ...) check failed to protect it falling under 0. Fix all of them by using "unsigned long" throughout the function. Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Reported-by: Toralf Förster <toralf.foerster@gmx.de> Tested-by: Toralf Förster <toralf.foerster@gmx.de> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Richard Weinberger <richard@nod.at> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-10-17 04:47:03 +08:00
return min_t(unsigned long, t, MAX_PAUSE);
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
static long wb_min_pause(struct bdi_writeback *wb,
long max_pause,
unsigned long task_ratelimit,
unsigned long dirty_ratelimit,
int *nr_dirtied_pause)
{
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
long hi = ilog2(wb->avg_write_bandwidth);
long lo = ilog2(wb->dirty_ratelimit);
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
long t; /* target pause */
long pause; /* estimated next pause */
int pages; /* target nr_dirtied_pause */
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
/* target for 10ms pause on 1-dd case */
t = max(1, HZ / 100);
/*
* Scale up pause time for concurrent dirtiers in order to reduce CPU
* overheads.
*
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
* (N * 10ms) on 2^N concurrent tasks.
*/
if (hi > lo)
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
t += (hi - lo) * (10 * HZ) / 1024;
/*
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
* This is a bit convoluted. We try to base the next nr_dirtied_pause
* on the much more stable dirty_ratelimit. However the next pause time
* will be computed based on task_ratelimit and the two rate limits may
* depart considerably at some time. Especially if task_ratelimit goes
* below dirty_ratelimit/2 and the target pause is max_pause, the next
* pause time will be max_pause*2 _trimmed down_ to max_pause. As a
* result task_ratelimit won't be executed faithfully, which could
* eventually bring down dirty_ratelimit.
*
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
* We apply two rules to fix it up:
* 1) try to estimate the next pause time and if necessary, use a lower
* nr_dirtied_pause so as not to exceed max_pause. When this happens,
* nr_dirtied_pause will be "dancing" with task_ratelimit.
* 2) limit the target pause time to max_pause/2, so that the normal
* small fluctuations of task_ratelimit won't trigger rule (1) and
* nr_dirtied_pause will remain as stable as dirty_ratelimit.
*/
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
t = min(t, 1 + max_pause / 2);
pages = dirty_ratelimit * t / roundup_pow_of_two(HZ);
/*
* Tiny nr_dirtied_pause is found to hurt I/O performance in the test
* case fio-mmap-randwrite-64k, which does 16*{sync read, async write}.
* When the 16 consecutive reads are often interrupted by some dirty
* throttling pause during the async writes, cfq will go into idles
* (deadline is fine). So push nr_dirtied_pause as high as possible
* until reaches DIRTY_POLL_THRESH=32 pages.
*/
if (pages < DIRTY_POLL_THRESH) {
t = max_pause;
pages = dirty_ratelimit * t / roundup_pow_of_two(HZ);
if (pages > DIRTY_POLL_THRESH) {
pages = DIRTY_POLL_THRESH;
t = HZ * DIRTY_POLL_THRESH / dirty_ratelimit;
}
}
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
pause = HZ * pages / (task_ratelimit + 1);
if (pause > max_pause) {
t = max_pause;
pages = task_ratelimit * t / roundup_pow_of_two(HZ);
}
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
*nr_dirtied_pause = pages;
/*
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
* The minimal pause time will normally be half the target pause time.
*/
return pages >= DIRTY_POLL_THRESH ? 1 + t / 2 : t;
}
static inline void wb_dirty_limits(struct dirty_throttle_control *dtc)
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
{
struct bdi_writeback *wb = dtc->wb;
unsigned long wb_reclaimable;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* wb_thresh is not treated as some limiting factor as
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* dirty_thresh, due to reasons
* - in JBOD setup, wb_thresh can fluctuate a lot
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* - in a system with HDD and USB key, the USB key may somehow
* go into state (wb_dirty >> wb_thresh) either because
* wb_dirty starts high, or because wb_thresh drops low.
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* In this case we don't want to hard throttle the USB key
* dirtiers for 100 seconds until wb_dirty drops under
* wb_thresh. Instead the auxiliary wb control line in
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
* wb_position_ratio() will let the dirtier task progress
* at some rate <= (write_bw / 2) for bringing down wb_dirty.
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
*/
dtc->wb_thresh = __wb_calc_thresh(dtc);
dtc->wb_bg_thresh = dtc->thresh ?
div_u64((u64)dtc->wb_thresh * dtc->bg_thresh, dtc->thresh) : 0;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
/*
* In order to avoid the stacked BDI deadlock we need
* to ensure we accurately count the 'dirty' pages when
* the threshold is low.
*
* Otherwise it would be possible to get thresh+n pages
* reported dirty, even though there are thresh-m pages
* actually dirty; with m+n sitting in the percpu
* deltas.
*/
if (dtc->wb_thresh < 2 * wb_stat_error()) {
wb_reclaimable = wb_stat_sum(wb, WB_RECLAIMABLE);
dtc->wb_dirty = wb_reclaimable + wb_stat_sum(wb, WB_WRITEBACK);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
} else {
wb_reclaimable = wb_stat(wb, WB_RECLAIMABLE);
dtc->wb_dirty = wb_reclaimable + wb_stat(wb, WB_WRITEBACK);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
}
}
/*
* balance_dirty_pages() must be called by processes which are generating dirty
* data. It looks at the number of dirty pages in the machine and will force
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
* the caller to wait once crossing the (background_thresh + dirty_thresh) / 2.
* If we're over `background_thresh' then the writeback threads are woken to
* perform some writeout.
*/
static void balance_dirty_pages(struct bdi_writeback *wb,
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
unsigned long pages_dirtied)
{
struct dirty_throttle_control gdtc_stor = { GDTC_INIT(wb) };
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
struct dirty_throttle_control mdtc_stor = { MDTC_INIT(wb, &gdtc_stor) };
struct dirty_throttle_control * const gdtc = &gdtc_stor;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
struct dirty_throttle_control * const mdtc = mdtc_valid(&mdtc_stor) ?
&mdtc_stor : NULL;
struct dirty_throttle_control *sdtc;
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
unsigned long nr_reclaimable; /* = file_dirty + unstable_nfs */
long period;
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
long pause;
long max_pause;
long min_pause;
int nr_dirtied_pause;
writeback: balance_dirty_pages(): reduce calls to global_page_state Reducing the number of times balance_dirty_pages calls global_page_state reduces the cache references and so improves write performance on a variety of workloads. 'perf stats' of simple fio write tests shows the reduction in cache access. Where the test is fio 'write,mmap,600Mb,pre_read' on AMD AthlonX2 with 3Gb memory (dirty_threshold approx 600 Mb) running each test 10 times, dropping the fasted & slowest values then taking the average & standard deviation average (s.d.) in millions (10^6) 2.6.31-rc8 648.6 (14.6) +patch 620.1 (16.5) Achieving this reduction is by dropping clip_bdi_dirty_limit as it rereads the counters to apply the dirty_threshold and moving this check up into balance_dirty_pages where it has already read the counters. Also by rearrange the for loop to only contain one copy of the limit tests allows the pdflush test after the loop to use the local copies of the counters rather than rereading them. In the common case with no throttling it now calls global_page_state 5 fewer times and bdi_stat 2 fewer. Fengguang: This patch slightly changes behavior by replacing clip_bdi_dirty_limit() with the explicit check (nr_reclaimable + nr_writeback >= dirty_thresh) to avoid exceeding the dirty limit. Since the bdi dirty limit is mostly accurate we don't need to do routinely clip. A simple dirty limit check would be enough. The check is necessary because, in principle we should throttle everything calling balance_dirty_pages() when we're over the total limit, as said by Peter. We now set and clear dirty_exceeded not only based on bdi dirty limits, but also on the global dirty limit. The global limit check is added in place of clip_bdi_dirty_limit() for safety and not intended as a behavior change. The bdi limits should be tight enough to keep all dirty pages under the global limit at most time; occasional small exceeding should be OK though. The change makes the logic more obvious: the global limit is the ultimate goal and shall be always imposed. We may now start background writeback work based on outdated conditions. That's safe because the bdi flush thread will (and have to) double check the states. It reduces overall overheads because the test based on old states still have good chance to be right. [akpm@linux-foundation.org] fix uninitialized dirty_exceeded Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com> Cc: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Jens Axboe <axboe@kernel.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-08-12 05:17:37 +08:00
bool dirty_exceeded = false;
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
unsigned long task_ratelimit;
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
unsigned long dirty_ratelimit;
struct backing_dev_info *bdi = wb->bdi;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
bool strictlimit = bdi->capabilities & BDI_CAP_STRICTLIMIT;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
unsigned long start_time = jiffies;
for (;;) {
unsigned long now = jiffies;
unsigned long dirty, thresh, bg_thresh;
unsigned long m_dirty = 0; /* stop bogus uninit warnings */
unsigned long m_thresh = 0;
unsigned long m_bg_thresh = 0;
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
/*
* Unstable writes are a feature of certain networked
* filesystems (i.e. NFS) in which data may have been
* written to the server's write cache, but has not yet
* been flushed to permanent storage.
*/
nr_reclaimable = global_node_page_state(NR_FILE_DIRTY) +
global_node_page_state(NR_UNSTABLE_NFS);
gdtc->avail = global_dirtyable_memory();
gdtc->dirty = nr_reclaimable + global_node_page_state(NR_WRITEBACK);
domain_dirty_limits(gdtc);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
if (unlikely(strictlimit)) {
wb_dirty_limits(gdtc);
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
dirty = gdtc->wb_dirty;
thresh = gdtc->wb_thresh;
bg_thresh = gdtc->wb_bg_thresh;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
} else {
dirty = gdtc->dirty;
thresh = gdtc->thresh;
bg_thresh = gdtc->bg_thresh;
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
}
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
if (mdtc) {
2015-09-30 01:04:26 +08:00
unsigned long filepages, headroom, writeback;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
/*
* If @wb belongs to !root memcg, repeat the same
* basic calculations for the memcg domain.
*/
2015-09-30 01:04:26 +08:00
mem_cgroup_wb_stats(wb, &filepages, &headroom,
&mdtc->dirty, &writeback);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
mdtc->dirty += writeback;
2015-09-30 01:04:26 +08:00
mdtc_calc_avail(mdtc, filepages, headroom);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
domain_dirty_limits(mdtc);
if (unlikely(strictlimit)) {
wb_dirty_limits(mdtc);
m_dirty = mdtc->wb_dirty;
m_thresh = mdtc->wb_thresh;
m_bg_thresh = mdtc->wb_bg_thresh;
} else {
m_dirty = mdtc->dirty;
m_thresh = mdtc->thresh;
m_bg_thresh = mdtc->bg_thresh;
}
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
}
/*
* Throttle it only when the background writeback cannot
* catch-up. This avoids (excessively) small writeouts
* when the wb limits are ramping up in case of !strictlimit.
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
*
* In strictlimit case make decision based on the wb counters
* and limits. Small writeouts when the wb limits are ramping
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
* up are the price we consciously pay for strictlimit-ing.
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
*
* If memcg domain is in effect, @dirty should be under
* both global and memcg freerun ceilings.
*/
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
if (dirty <= dirty_freerun_ceiling(thresh, bg_thresh) &&
(!mdtc ||
m_dirty <= dirty_freerun_ceiling(m_thresh, m_bg_thresh))) {
unsigned long intv = dirty_poll_interval(dirty, thresh);
unsigned long m_intv = ULONG_MAX;
current->dirty_paused_when = now;
current->nr_dirtied = 0;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
if (mdtc)
m_intv = dirty_poll_interval(m_dirty, m_thresh);
current->nr_dirtied_pause = min(intv, m_intv);
break;
}
if (unlikely(!writeback_in_progress(wb)))
wb_start_background_writeback(wb);
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-27 00:06:56 +08:00
mem_cgroup_flush_foreign(wb);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
/*
* Calculate global domain's pos_ratio and select the
* global dtc by default.
*/
mm/page-writeback.c: add strictlimit feature The feature prevents mistrusted filesystems (ie: FUSE mounts created by unprivileged users) to grow a large number of dirty pages before throttling. For such filesystems balance_dirty_pages always check bdi counters against bdi limits. I.e. even if global "nr_dirty" is under "freerun", it's not allowed to skip bdi checks. The only use case for now is fuse: it sets bdi max_ratio to 1% by default and system administrators are supposed to expect that this limit won't be exceeded. The feature is on if a BDI is marked by BDI_CAP_STRICTLIMIT flag. A filesystem may set the flag when it initializes its BDI. The problematic scenario comes from the fact that nobody pays attention to the NR_WRITEBACK_TEMP counter (i.e. number of pages under fuse writeback). The implementation of fuse writeback releases original page (by calling end_page_writeback) almost immediately. A fuse request queued for real processing bears a copy of original page. Hence, if userspace fuse daemon doesn't finalize write requests in timely manner, an aggressive mmap writer can pollute virtually all memory by those temporary fuse page copies. They are carefully accounted in NR_WRITEBACK_TEMP, but nobody cares. To make further explanations shorter, let me use "NR_WRITEBACK_TEMP problem" as a shortcut for "a possibility of uncontrolled grow of amount of RAM consumed by temporary pages allocated by kernel fuse to process writeback". The problem was very easy to reproduce. There is a trivial example filesystem implementation in fuse userspace distribution: fusexmp_fh.c. I added "sleep(1);" to the write methods, then recompiled and mounted it. Then created a huge file on the mount point and run a simple program which mmap-ed the file to a memory region, then wrote a data to the region. An hour later I observed almost all RAM consumed by fuse writeback. Since then some unrelated changes in kernel fuse made it more difficult to reproduce, but it is still possible now. Putting this theoretical happens-in-the-lab thing aside, there is another thing that really hurts real world (FUSE) users. This is write-through page cache policy FUSE currently uses. I.e. handling write(2), kernel fuse populates page cache and flushes user data to the server synchronously. This is excessively suboptimal. Pavel Emelyanov's patches ("writeback cache policy") solve the problem, but they also make resolving NR_WRITEBACK_TEMP problem absolutely necessary. Otherwise, simply copying a huge file to a fuse mount would result in memory starvation. Miklos, the maintainer of FUSE, believes strictlimit feature the way to go. And eventually putting FUSE topics aside, there is one more use-case for strictlimit feature. Using a slow USB stick (mass storage) in a machine with huge amount of RAM installed is a well-known pain. Let's make simple computations. Assuming 64GB of RAM installed, existing implementation of balance_dirty_pages will start throttling only after 9.6GB of RAM becomes dirty (freerun == 15% of total RAM). So, the command "cp 9GB_file /media/my-usb-storage/" may return in a few seconds, but subsequent "umount /media/my-usb-storage/" will take more than two hours if effective throughput of the storage is, to say, 1MB/sec. After inclusion of strictlimit feature, it will be trivial to add a knob (e.g. /sys/devices/virtual/bdi/x:y/strictlimit) to enable it on demand. Manually or via udev rule. May be I'm wrong, but it seems to be quite a natural desire to limit the amount of dirty memory for some devices we are not fully trust (in the sense of sustainable throughput). [akpm@linux-foundation.org: fix warning in page-writeback.c] Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com> Cc: Jan Kara <jack@suse.cz> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 05:22:46 +08:00
if (!strictlimit)
wb_dirty_limits(gdtc);
dirty_exceeded = (gdtc->wb_dirty > gdtc->wb_thresh) &&
((gdtc->dirty > gdtc->thresh) || strictlimit);
wb_position_ratio(gdtc);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
sdtc = gdtc;
if (mdtc) {
/*
* If memcg domain is in effect, calculate its
* pos_ratio. @wb should satisfy constraints from
* both global and memcg domains. Choose the one
* w/ lower pos_ratio.
*/
if (!strictlimit)
wb_dirty_limits(mdtc);
dirty_exceeded |= (mdtc->wb_dirty > mdtc->wb_thresh) &&
((mdtc->dirty > mdtc->thresh) || strictlimit);
wb_position_ratio(mdtc);
if (mdtc->pos_ratio < gdtc->pos_ratio)
sdtc = mdtc;
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (dirty_exceeded && !wb->dirty_exceeded)
wb->dirty_exceeded = 1;
if (time_is_before_jiffies(wb->bw_time_stamp +
BANDWIDTH_INTERVAL)) {
spin_lock(&wb->list_lock);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
__wb_update_bandwidth(gdtc, mdtc, start_time, true);
spin_unlock(&wb->list_lock);
}
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-30 01:22:30 +08:00
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
/* throttle according to the chosen dtc */
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
dirty_ratelimit = wb->dirty_ratelimit;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
task_ratelimit = ((u64)dirty_ratelimit * sdtc->pos_ratio) >>
RATELIMIT_CALC_SHIFT;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
max_pause = wb_max_pause(wb, sdtc->wb_dirty);
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
min_pause = wb_min_pause(wb, max_pause,
task_ratelimit, dirty_ratelimit,
&nr_dirtied_pause);
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
if (unlikely(task_ratelimit == 0)) {
period = max_pause;
pause = max_pause;
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
goto pause;
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
period = HZ * pages_dirtied / task_ratelimit;
pause = period;
if (current->dirty_paused_when)
pause -= now - current->dirty_paused_when;
/*
* For less than 1s think time (ext3/4 may block the dirtier
* for up to 800ms from time to time on 1-HDD; so does xfs,
* however at much less frequency), try to compensate it in
* future periods by updating the virtual time; otherwise just
* do a reset, as it may be a light dirtier.
*/
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
if (pause < min_pause) {
trace_balance_dirty_pages(wb,
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
sdtc->thresh,
sdtc->bg_thresh,
sdtc->dirty,
sdtc->wb_thresh,
sdtc->wb_dirty,
dirty_ratelimit,
task_ratelimit,
pages_dirtied,
period,
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
min(pause, 0L),
start_time);
if (pause < -HZ) {
current->dirty_paused_when = now;
current->nr_dirtied = 0;
} else if (period) {
current->dirty_paused_when += period;
current->nr_dirtied = 0;
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
} else if (current->nr_dirtied_pause <= pages_dirtied)
current->nr_dirtied_pause += pages_dirtied;
break;
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
if (unlikely(pause > max_pause)) {
/* for occasional dropped task_ratelimit */
now += min(pause - max_pause, max_pause);
pause = max_pause;
}
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
pause:
trace_balance_dirty_pages(wb,
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
sdtc->thresh,
sdtc->bg_thresh,
sdtc->dirty,
sdtc->wb_thresh,
sdtc->wb_dirty,
dirty_ratelimit,
task_ratelimit,
pages_dirtied,
period,
pause,
start_time);
__set_current_state(TASK_KILLABLE);
wb->dirty_sleep = now;
io_schedule_timeout(pause);
current->dirty_paused_when = now + pause;
current->nr_dirtied = 0;
writeback: max, min and target dirty pause time Control the pause time and the call intervals to balance_dirty_pages() with three parameters: 1) max_pause, limited by bdi_dirty and MAX_PAUSE 2) the target pause time, grows with the number of dd tasks and is normally limited by max_pause/2 3) the minimal pause, set to half the target pause and is used to skip short sleeps and accumulate them into bigger ones The typical behaviors after patch: - if ever task_ratelimit is far below dirty_ratelimit, the pause time will remain constant at max_pause and nr_dirtied_pause will be fluctuating with task_ratelimit - in the normal cases, nr_dirtied_pause will remain stable (keep in the same pace with dirty_ratelimit) and the pause time will be fluctuating with task_ratelimit In summary, someone has to fluctuate with task_ratelimit, because task_ratelimit = nr_dirtied_pause / pause We normally prefer a stable nr_dirtied_pause, until reaching max_pause. The notable behavior changes are: - in stable workloads, there will no longer be sudden big trajectory switching of nr_dirtied_pause as concerned by Peter. It will be as smooth as dirty_ratelimit and changing proportionally with it (as always, assuming bdi bandwidth does not fluctuate across 2^N lines, otherwise nr_dirtied_pause will show up in 2+ parallel trajectories) - in the rare cases when something keeps task_ratelimit far below dirty_ratelimit, the smoothness can no longer be retained and nr_dirtied_pause will be "dancing" with task_ratelimit. This fixes a (not that destructive but still not good) bug that dirty_ratelimit gets brought down undesirably <= balanced_dirty_ratelimit is under estimated <= weakly executed task_ratelimit <= pause goes too large and gets trimmed down to max_pause <= nr_dirtied_pause (based on dirty_ratelimit) is set too large <= dirty_ratelimit being much larger than task_ratelimit - introduce min_pause to avoid small pause sleeps - when pause is trimmed down to max_pause, try to compensate it at the next pause time The "refactor" type of changes are: The max_pause equation is slightly transformed to make it slightly more efficient. We now scale target_pause by (N * 10ms) on 2^N concurrent tasks, which is effectively equal to the original scaling max_pause by (N * 20ms) because the original code does implicit target_pause ~= max_pause / 2. Based on the same implicit ratio, target_pause starts with 10ms on 1 dd. CC: Jan Kara <jack@suse.cz> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-12-01 01:08:55 +08:00
current->nr_dirtied_pause = nr_dirtied_pause;
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
/*
* This is typically equal to (dirty < thresh) and can also
* keep "1000+ dd on a slow USB stick" under control.
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
*/
if (task_ratelimit)
writeback: introduce max-pause and pass-good dirty limits The max-pause limit helps to keep the sleep time inside balance_dirty_pages() within MAX_PAUSE=200ms. The 200ms max sleep means per task rate limit of 8pages/200ms=160KB/s when dirty exceeded, which normally is enough to stop dirtiers from continue pushing the dirty pages high, unless there are a sufficient large number of slow dirtiers (eg. 500 tasks doing 160KB/s will still sum up to 80MB/s, exceeding the write bandwidth of a slow disk and hence accumulating more and more dirty pages). The pass-good limit helps to let go of the good bdi's in the presence of a blocked bdi (ie. NFS server not responding) or slow USB disk which for some reason build up a large number of initial dirty pages that refuse to go away anytime soon. For example, given two bdi's A and B and the initial state bdi_thresh_A = dirty_thresh / 2 bdi_thresh_B = dirty_thresh / 2 bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 Then A get blocked, after a dozen seconds bdi_thresh_A = 0 bdi_thresh_B = dirty_thresh bdi_dirty_A = dirty_thresh / 2 bdi_dirty_B = dirty_thresh / 2 The (bdi_dirty_B < bdi_thresh_B) test is now useless and the dirty pages will be effectively throttled by condition (nr_dirty < dirty_thresh). This has two problems: (1) we lose the protections for light dirtiers (2) balance_dirty_pages() effectively becomes IO-less because the (bdi_nr_reclaimable > bdi_thresh) test won't be true. This is good for IO, but balance_dirty_pages() loses an important way to break out of the loop which leads to more spread out throttle delays. DIRTY_PASSGOOD_AREA can eliminate the above issues. The only problem is, DIRTY_PASSGOOD_AREA needs to be defined as 2 to fully cover the above example while this patch uses the more conservative value 8 so as not to surprise people with too many dirty pages than expected. The max-pause limit won't noticeably impact the speed dirty pages are knocked down when there is a sudden drop of global/bdi dirty thresholds. Because the heavy dirties will be throttled below 160KB/s which is slow enough. It does help to avoid long dirty throttle delays and especially will make light dirtiers more responsive. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-20 12:18:42 +08:00
break;
/*
* In the case of an unresponding NFS server and the NFS dirty
* pages exceeds dirty_thresh, give the other good wb's a pipe
* to go through, so that tasks on them still remain responsive.
*
* In theory 1 page is enough to keep the consumer-producer
* pipe going: the flusher cleans 1 page => the task dirties 1
* more page. However wb_dirty has accounting errors. So use
* the larger and more IO friendly wb_stat_error.
*/
if (sdtc->wb_dirty <= wb_stat_error())
break;
if (fatal_signal_pending(current))
break;
}
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (!dirty_exceeded && wb->dirty_exceeded)
wb->dirty_exceeded = 0;
if (writeback_in_progress(wb))
return;
/*
* In laptop mode, we wait until hitting the higher threshold before
* starting background writeout, and then write out all the way down
* to the lower threshold. So slow writers cause minimal disk activity.
*
* In normal mode, we start background writeout at the lower
* background_thresh, to keep the amount of dirty memory low.
*/
writeback: IO-less balance_dirty_pages() As proposed by Chris, Dave and Jan, don't start foreground writeback IO inside balance_dirty_pages(). Instead, simply let it idle sleep for some time to throttle the dirtying task. In the mean while, kick off the per-bdi flusher thread to do background writeback IO. RATIONALS ========= - disk seeks on concurrent writeback of multiple inodes (Dave Chinner) If every thread doing writes and being throttled start foreground writeback, it leads to N IO submitters from at least N different inodes at the same time, end up with N different sets of IO being issued with potentially zero locality to each other, resulting in much lower elevator sort/merge efficiency and hence we seek the disk all over the place to service the different sets of IO. OTOH, if there is only one submission thread, it doesn't jump between inodes in the same way when congestion clears - it keeps writing to the same inode, resulting in large related chunks of sequential IOs being issued to the disk. This is more efficient than the above foreground writeback because the elevator works better and the disk seeks less. - lock contention and cache bouncing on concurrent IO submitters (Dave Chinner) With this patchset, the fs_mark benchmark on a 12-drive software RAID0 goes from CPU bound to IO bound, freeing "3-4 CPUs worth of spinlock contention". * "CPU usage has dropped by ~55%", "it certainly appears that most of the CPU time saving comes from the removal of contention on the inode_wb_list_lock" (IMHO at least 10% comes from the reduction of cacheline bouncing, because the new code is able to call much less frequently into balance_dirty_pages() and hence access the global page states) * the user space "App overhead" is reduced by 20%, by avoiding the cacheline pollution by the complex writeback code path * "for a ~5% throughput reduction", "the number of write IOs have dropped by ~25%", and the elapsed time reduced from 41:42.17 to 40:53.23. * On a simple test of 100 dd, it reduces the CPU %system time from 30% to 3%, and improves IO throughput from 38MB/s to 42MB/s. - IO size too small for fast arrays and too large for slow USB sticks The write_chunk used by current balance_dirty_pages() cannot be directly set to some large value (eg. 128MB) for better IO efficiency. Because it could lead to more than 1 second user perceivable stalls. Even the current 4MB write size may be too large for slow USB sticks. The fact that balance_dirty_pages() starts IO on itself couples the IO size to wait time, which makes it hard to do suitable IO size while keeping the wait time under control. Now it's possible to increase writeback chunk size proportional to the disk bandwidth. In a simple test of 50 dd's on XFS, 1-HDD, 3GB ram, the larger writeback size dramatically reduces the seek count to 1/10 (far beyond my expectation) and improves the write throughput by 24%. - long block time in balance_dirty_pages() hurts desktop responsiveness Many of us may have the experience: it often takes a couple of seconds or even long time to stop a heavy writing dd/cp/tar command with Ctrl-C or "kill -9". - IO pipeline broken by bumpy write() progress There are a broad class of "loop {read(buf); write(buf);}" applications whose read() pipeline will be under-utilized or even come to a stop if the write()s have long latencies _or_ don't progress in a constant rate. The current threshold based throttling inherently transfers the large low level IO completion fluctuations to bumpy application write()s, and further deteriorates with increasing number of dirtiers and/or bdi's. For example, when doing 50 dd's + 1 remote rsync to an XFS partition, the rsync progresses very bumpy in legacy kernel, and throughput is improved by 67% by this patchset. (plus the larger write chunk size, it will be 93% speedup). The new rate based throttling can support 1000+ dd's with excellent smoothness, low latency and low overheads. For the above reasons, it's much better to do IO-less and low latency pauses in balance_dirty_pages(). Jan Kara, Dave Chinner and me explored the scheme to let balance_dirty_pages() wait for enough writeback IO completions to safeguard the dirty limit. However it's found to have two problems: - in large NUMA systems, the per-cpu counters may have big accounting errors, leading to big throttle wait time and jitters. - NFS may kill large amount of unstable pages with one single COMMIT. Because NFS server serves COMMIT with expensive fsync() IOs, it is desirable to delay and reduce the number of COMMITs. So it's not likely to optimize away such kind of bursty IO completions, and the resulted large (and tiny) stall times in IO completion based throttling. So here is a pause time oriented approach, which tries to control the pause time in each balance_dirty_pages() invocations, by controlling the number of pages dirtied before calling balance_dirty_pages(), for smooth and efficient dirty throttling: - avoid useless (eg. zero pause time) balance_dirty_pages() calls - avoid too small pause time (less than 4ms, which burns CPU power) - avoid too large pause time (more than 200ms, which hurts responsiveness) - avoid big fluctuations of pause times It can control pause times at will. The default policy (in a followup patch) will be to do ~10ms pauses in 1-dd case, and increase to ~100ms in 1000-dd case. BEHAVIOR CHANGE =============== (1) dirty threshold Users will notice that the applications will get throttled once crossing the global (background + dirty)/2=15% threshold, and then balanced around 17.5%. Before patch, the behavior is to just throttle it at 20% dirtyable memory in 1-dd case. Since the task will be soft throttled earlier than before, it may be perceived by end users as performance "slow down" if his application happens to dirty more than 15% dirtyable memory. (2) smoothness/responsiveness Users will notice a more responsive system during heavy writeback. "killall dd" will take effect instantly. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-28 08:45:12 +08:00
if (laptop_mode)
return;
if (nr_reclaimable > gdtc->bg_thresh)
wb_start_background_writeback(wb);
}
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
static DEFINE_PER_CPU(int, bdp_ratelimits);
/*
* Normal tasks are throttled by
* loop {
* dirty tsk->nr_dirtied_pause pages;
* take a snap in balance_dirty_pages();
* }
* However there is a worst case. If every task exit immediately when dirtied
* (tsk->nr_dirtied_pause - 1) pages, balance_dirty_pages() will never be
* called to throttle the page dirties. The solution is to save the not yet
* throttled page dirties in dirty_throttle_leaks on task exit and charge them
* randomly into the running tasks. This works well for the above worst case,
* as the new task will pick up and accumulate the old task's leaked dirty
* count and eventually get throttled.
*/
DEFINE_PER_CPU(int, dirty_throttle_leaks) = 0;
/**
* balance_dirty_pages_ratelimited - balance dirty memory state
* @mapping: address_space which was dirtied
*
* Processes which are dirtying memory should call in here once for each page
* which was newly dirtied. The function will periodically check the system's
* dirty state and will initiate writeback if needed.
*
* On really big machines, get_writeback_state is expensive, so try to avoid
* calling it too often (ratelimiting). But once we're over the dirty memory
* limit we decrease the ratelimiting by a lot, to prevent individual processes
* from overshooting the limit by (ratelimit_pages) each.
*/
void balance_dirty_pages_ratelimited(struct address_space *mapping)
{
struct inode *inode = mapping->host;
struct backing_dev_info *bdi = inode_to_bdi(inode);
struct bdi_writeback *wb = NULL;
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
int ratelimit;
int *p;
if (!bdi_cap_account_dirty(bdi))
return;
if (inode_cgwb_enabled(inode))
wb = wb_get_create_current(bdi, GFP_KERNEL);
if (!wb)
wb = &bdi->wb;
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
ratelimit = current->nr_dirtied_pause;
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:28 +08:00
if (wb->dirty_exceeded)
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
ratelimit = min(ratelimit, 32 >> (PAGE_SHIFT - 10));
preempt_disable();
/*
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
* This prevents one CPU to accumulate too many dirtied pages without
* calling into balance_dirty_pages(), which can happen when there are
* 1000+ tasks, all of them start dirtying pages at exactly the same
* time, hence all honoured too large initial task->nr_dirtied_pause.
*/
p = this_cpu_ptr(&bdp_ratelimits);
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
if (unlikely(current->nr_dirtied >= ratelimit))
*p = 0;
else if (unlikely(*p >= ratelimit_pages)) {
*p = 0;
ratelimit = 0;
}
/*
* Pick up the dirtied pages by the exited tasks. This avoids lots of
* short-lived tasks (eg. gcc invocations in a kernel build) escaping
* the dirty throttling and livelock other long-run dirtiers.
*/
p = this_cpu_ptr(&dirty_throttle_leaks);
if (*p > 0 && current->nr_dirtied < ratelimit) {
unsigned long nr_pages_dirtied;
nr_pages_dirtied = min(*p, ratelimit - current->nr_dirtied);
*p -= nr_pages_dirtied;
current->nr_dirtied += nr_pages_dirtied;
}
preempt_enable();
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
if (unlikely(current->nr_dirtied >= ratelimit))
balance_dirty_pages(wb, current->nr_dirtied);
wb_put(wb);
}
EXPORT_SYMBOL(balance_dirty_pages_ratelimited);
/**
* wb_over_bg_thresh - does @wb need to be written back?
* @wb: bdi_writeback of interest
*
* Determines whether background writeback should keep writing @wb or it's
* clean enough.
*
* Return: %true if writeback should continue.
*/
bool wb_over_bg_thresh(struct bdi_writeback *wb)
{
struct dirty_throttle_control gdtc_stor = { GDTC_INIT(wb) };
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
struct dirty_throttle_control mdtc_stor = { MDTC_INIT(wb, &gdtc_stor) };
struct dirty_throttle_control * const gdtc = &gdtc_stor;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
struct dirty_throttle_control * const mdtc = mdtc_valid(&mdtc_stor) ?
&mdtc_stor : NULL;
/*
* Similar to balance_dirty_pages() but ignores pages being written
* as we're trying to decide whether to put more under writeback.
*/
gdtc->avail = global_dirtyable_memory();
gdtc->dirty = global_node_page_state(NR_FILE_DIRTY) +
global_node_page_state(NR_UNSTABLE_NFS);
domain_dirty_limits(gdtc);
if (gdtc->dirty > gdtc->bg_thresh)
return true;
writeback: Fix performance regression in wb_over_bg_thresh() Commit 947e9762a8dd ("writeback: update wb_over_bg_thresh() to use wb_domain aware operations") unintentionally changed this function's meaning from "are there more dirty pages than the background writeback threshold" to "are there more dirty pages than the writeback threshold". The background writeback threshold is typically half of the writeback threshold, so this had the effect of raising the number of dirty pages required to cause a writeback worker to perform background writeout. This can cause a very severe performance regression when a BDI uses BDI_CAP_STRICTLIMIT because balance_dirty_pages() and the writeback worker can now disagree on whether writeback should be initiated. For example, in a system having 1GB of RAM, a single spinning disk, and a "pass-through" FUSE filesystem mounted over the disk, application code mmapped a 128MB file on the disk and was randomly dirtying pages in that mapping. Because FUSE uses strictlimit and has a default max_ratio of only 1%, in balance_dirty_pages, thresh is ~200, bg_thresh is ~100, and the dirty_freerun_ceiling is the average of those, ~150. So, it pauses the dirtying processes when we have 151 dirty pages and wakes up a background writeback worker. But the worker tests the wrong threshold (200 instead of 100), so it does not initiate writeback and just returns. Thus, balance_dirty_pages keeps looping, sleeping and then waking up the worker who will do nothing. It remains stuck in this state until the few dirty pages that we have finally expire and we write them back for that reason. Then the whole process repeats, resulting in near-zero throughput through the FUSE BDI. The fix is to call the parameterized variant of wb_calc_thresh, so that the worker will do writeback if the bg_thresh is exceeded which was the behavior before the referenced commit. Fixes: 947e9762a8dd ("writeback: update wb_over_bg_thresh() to use wb_domain aware operations") Signed-off-by: Howard Cochran <hcochran@kernelspring.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> Cc: <stable@vger.kernel.org> # v4.2+ Tested-by Sedat Dilek <sedat.dilek@gmail.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-03-10 14:12:39 +08:00
if (wb_stat(wb, WB_RECLAIMABLE) >
wb_calc_thresh(gdtc->wb, gdtc->bg_thresh))
return true;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
if (mdtc) {
2015-09-30 01:04:26 +08:00
unsigned long filepages, headroom, writeback;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
2015-09-30 01:04:26 +08:00
mem_cgroup_wb_stats(wb, &filepages, &headroom, &mdtc->dirty,
&writeback);
mdtc_calc_avail(mdtc, filepages, headroom);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
domain_dirty_limits(mdtc); /* ditto, ignore writeback */
if (mdtc->dirty > mdtc->bg_thresh)
return true;
writeback: Fix performance regression in wb_over_bg_thresh() Commit 947e9762a8dd ("writeback: update wb_over_bg_thresh() to use wb_domain aware operations") unintentionally changed this function's meaning from "are there more dirty pages than the background writeback threshold" to "are there more dirty pages than the writeback threshold". The background writeback threshold is typically half of the writeback threshold, so this had the effect of raising the number of dirty pages required to cause a writeback worker to perform background writeout. This can cause a very severe performance regression when a BDI uses BDI_CAP_STRICTLIMIT because balance_dirty_pages() and the writeback worker can now disagree on whether writeback should be initiated. For example, in a system having 1GB of RAM, a single spinning disk, and a "pass-through" FUSE filesystem mounted over the disk, application code mmapped a 128MB file on the disk and was randomly dirtying pages in that mapping. Because FUSE uses strictlimit and has a default max_ratio of only 1%, in balance_dirty_pages, thresh is ~200, bg_thresh is ~100, and the dirty_freerun_ceiling is the average of those, ~150. So, it pauses the dirtying processes when we have 151 dirty pages and wakes up a background writeback worker. But the worker tests the wrong threshold (200 instead of 100), so it does not initiate writeback and just returns. Thus, balance_dirty_pages keeps looping, sleeping and then waking up the worker who will do nothing. It remains stuck in this state until the few dirty pages that we have finally expire and we write them back for that reason. Then the whole process repeats, resulting in near-zero throughput through the FUSE BDI. The fix is to call the parameterized variant of wb_calc_thresh, so that the worker will do writeback if the bg_thresh is exceeded which was the behavior before the referenced commit. Fixes: 947e9762a8dd ("writeback: update wb_over_bg_thresh() to use wb_domain aware operations") Signed-off-by: Howard Cochran <hcochran@kernelspring.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> Cc: <stable@vger.kernel.org> # v4.2+ Tested-by Sedat Dilek <sedat.dilek@gmail.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-03-10 14:12:39 +08:00
if (wb_stat(wb, WB_RECLAIMABLE) >
wb_calc_thresh(mdtc->wb, mdtc->bg_thresh))
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 06:23:35 +08:00
return true;
}
return false;
}
/*
* sysctl handler for /proc/sys/vm/dirty_writeback_centisecs
*/
int dirty_writeback_centisecs_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *length, loff_t *ppos)
{
unsigned int old_interval = dirty_writeback_interval;
int ret;
ret = proc_dointvec(table, write, buffer, length, ppos);
/*
* Writing 0 to dirty_writeback_interval will disable periodic writeback
* and a different non-zero value will wakeup the writeback threads.
* wb_wakeup_delayed() would be more appropriate, but it's a pain to
* iterate over all bdis and wbs.
* The reason we do this is to make the change take effect immediately.
*/
if (!ret && write && dirty_writeback_interval &&
dirty_writeback_interval != old_interval)
wakeup_flusher_threads(WB_REASON_PERIODIC);
return ret;
}
#ifdef CONFIG_BLOCK
void laptop_mode_timer_fn(struct timer_list *t)
{
struct backing_dev_info *backing_dev_info =
from_timer(backing_dev_info, t, laptop_mode_wb_timer);
wakeup_flusher_threads_bdi(backing_dev_info, WB_REASON_LAPTOP_TIMER);
}
/*
* We've spun up the disk and we're in laptop mode: schedule writeback
* of all dirty data a few seconds from now. If the flush is already scheduled
* then push it back - the user is still using the disk.
*/
void laptop_io_completion(struct backing_dev_info *info)
{
mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode);
}
/*
* We're in laptop mode and we've just synced. The sync's writes will have
* caused another writeback to be scheduled by laptop_io_completion.
* Nothing needs to be written back anymore, so we unschedule the writeback.
*/
void laptop_sync_completion(void)
{
struct backing_dev_info *bdi;
rcu_read_lock();
list_for_each_entry_rcu(bdi, &bdi_list, bdi_list)
del_timer(&bdi->laptop_mode_wb_timer);
rcu_read_unlock();
}
#endif
/*
* If ratelimit_pages is too high then we can get into dirty-data overload
* if a large number of processes all perform writes at the same time.
* If it is too low then SMP machines will call the (expensive)
* get_writeback_state too often.
*
* Here we set ratelimit_pages to a level which ensures that when all CPUs are
* dirtying in parallel, we cannot go more than 3% (1/32) over the dirty memory
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
* thresholds.
*/
void writeback_set_ratelimit(void)
{
struct wb_domain *dom = &global_wb_domain;
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
unsigned long background_thresh;
unsigned long dirty_thresh;
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
global_dirty_limits(&background_thresh, &dirty_thresh);
dom->dirty_limit = dirty_thresh;
writeback: per task dirty rate limit Add two fields to task_struct. 1) account dirtied pages in the individual tasks, for accuracy 2) per-task balance_dirty_pages() call intervals, for flexibility The balance_dirty_pages() call interval (ie. nr_dirtied_pause) will scale near-sqrt to the safety gap between dirty pages and threshold. The main problem of per-task nr_dirtied is, if 1k+ tasks start dirtying pages at exactly the same time, each task will be assigned a large initial nr_dirtied_pause, so that the dirty threshold will be exceeded long before each task reached its nr_dirtied_pause and hence call balance_dirty_pages(). The solution is to watch for the number of pages dirtied on each CPU in between the calls into balance_dirty_pages(). If it exceeds ratelimit_pages (3% dirty threshold), force call balance_dirty_pages() for a chance to set bdi->dirty_exceeded. In normal situations, this safeguarding condition is not expected to trigger at all. On the sqrt in dirty_poll_interval(): It will serve as an initial guess when dirty pages are still in the freerun area. When dirty pages are floating inside the dirty control scope [freerun, limit], a followup patch will use some refined dirty poll interval to get the desired pause time. thresh-dirty (MB) sqrt 1 16 2 22 4 32 8 45 16 64 32 90 64 128 128 181 256 256 512 362 1024 512 The above table means, given 1MB (or 1GB) gap and the dd tasks polling balance_dirty_pages() on every 16 (or 512) pages, the dirty limit won't be exceeded as long as there are less than 16 (or 512) concurrent dd's. So sqrt naturally leads to less overheads and more safe concurrent tasks for large memory servers, which have large (thresh-freerun) gaps. peter: keep the per-CPU ratelimit for safeguarding the 1k+ tasks case CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Reviewed-by: Andrea Righi <andrea@betterlinux.com> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 08:10:12 +08:00
ratelimit_pages = dirty_thresh / (num_online_cpus() * 32);
if (ratelimit_pages < 16)
ratelimit_pages = 16;
}
static int page_writeback_cpu_online(unsigned int cpu)
{
writeback_set_ratelimit();
return 0;
}
/*
* Called early on to tune the page writeback dirty limits.
*
* We used to scale dirty pages according to how total memory
* related to pages that could be allocated for buffers (by
* comparing nr_free_buffer_pages() to vm_total_pages.
*
* However, that was when we used "dirty_ratio" to scale with
* all memory, and we don't do that any more. "dirty_ratio"
* is now applied to total non-HIGHPAGE memory (by subtracting
* totalhigh_pages from vm_total_pages), and as such we can't
* get into the old insane situation any more where we had
* large amounts of dirty pages compared to a small amount of
* non-HIGHMEM memory.
*
* But we might still want to scale the dirty_ratio by how
* much memory the box has..
*/
void __init page_writeback_init(void)
{
BUG_ON(wb_domain_init(&global_wb_domain, GFP_KERNEL));
cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "mm/writeback:online",
page_writeback_cpu_online, NULL);
cpuhp_setup_state(CPUHP_MM_WRITEBACK_DEAD, "mm/writeback:dead", NULL,
page_writeback_cpu_online);
}
/**
* tag_pages_for_writeback - tag pages to be written by write_cache_pages
* @mapping: address space structure to write
* @start: starting page index
* @end: ending page index (inclusive)
*
* This function scans the page range from @start to @end (inclusive) and tags
* all pages that have DIRTY tag set with a special TOWRITE tag. The idea is
* that write_cache_pages (or whoever calls this function) will then use
* TOWRITE tag to identify pages eligible for writeback. This mechanism is
* used to avoid livelocking of writeback by a process steadily creating new
* dirty pages in the file (thus it is important for this function to be quick
* so that it can tag pages faster than a dirtying process can create them).
*/
void tag_pages_for_writeback(struct address_space *mapping,
pgoff_t start, pgoff_t end)
{
XA_STATE(xas, &mapping->i_pages, start);
unsigned int tagged = 0;
void *page;
xas_lock_irq(&xas);
xas_for_each_marked(&xas, page, end, PAGECACHE_TAG_DIRTY) {
xas_set_mark(&xas, PAGECACHE_TAG_TOWRITE);
if (++tagged % XA_CHECK_SCHED)
continue;
xas_pause(&xas);
xas_unlock_irq(&xas);
cond_resched();
xas_lock_irq(&xas);
}
xas_unlock_irq(&xas);
}
EXPORT_SYMBOL(tag_pages_for_writeback);
/**
* write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
* @mapping: address space structure to write
* @wbc: subtract the number of written pages from *@wbc->nr_to_write
* @writepage: function called for each page
* @data: data passed to writepage function
*
* If a page is already under I/O, write_cache_pages() skips it, even
* if it's dirty. This is desirable behaviour for memory-cleaning writeback,
* but it is INCORRECT for data-integrity system calls such as fsync(). fsync()
* and msync() need to guarantee that all the data which was dirty at the time
* the call was made get new I/O started against them. If wbc->sync_mode is
* WB_SYNC_ALL then we were called for data integrity and we must wait for
* existing IO to complete.
*
* To avoid livelocks (when other process dirties new pages), we first tag
* pages which should be written back with TOWRITE tag and only then start
* writing them. For data-integrity sync we have to be careful so that we do
* not miss some pages (e.g., because some other process has cleared TOWRITE
* tag we set). The rule we follow is that TOWRITE tag can be cleared only
* by the process clearing the DIRTY tag (and submitting the page for IO).
mm/page-writeback.c: fix range_cyclic writeback vs writepages deadlock We've recently seen a workload on XFS filesystems with a repeatable deadlock between background writeback and a multi-process application doing concurrent writes and fsyncs to a small range of a file. range_cyclic writeback Process 1 Process 2 xfs_vm_writepages write_cache_pages writeback_index = 2 cycled = 0 .... find page 2 dirty lock Page 2 ->writepage page 2 writeback page 2 clean page 2 added to bio no more pages write() locks page 1 dirties page 1 locks page 2 dirties page 1 fsync() .... xfs_vm_writepages write_cache_pages start index 0 find page 1 towrite lock Page 1 ->writepage page 1 writeback page 1 clean page 1 added to bio find page 2 towrite lock Page 2 page 2 is writeback <blocks> write() locks page 1 dirties page 1 fsync() .... xfs_vm_writepages write_cache_pages start index 0 !done && !cycled sets index to 0, restarts lookup find page 1 dirty find page 1 towrite lock Page 1 page 1 is writeback <blocks> lock Page 1 <blocks> DEADLOCK because: - process 1 needs page 2 writeback to complete to make enough progress to issue IO pending for page 1 - writeback needs page 1 writeback to complete so process 2 can progress and unlock the page it is blocked on, then it can issue the IO pending for page 2 - process 2 can't make progress until process 1 issues IO for page 1 The underlying cause of the problem here is that range_cyclic writeback is processing pages in descending index order as we hold higher index pages in a structure controlled from above write_cache_pages(). The write_cache_pages() caller needs to be able to submit these pages for IO before write_cache_pages restarts writeback at mapping index 0 to avoid wcp inverting the page lock/writeback wait order. generic_writepages() is not susceptible to this bug as it has no private context held across write_cache_pages() - filesystems using this infrastructure always submit pages in ->writepage immediately and so there is no problem with range_cyclic going back to mapping index 0. However: mpage_writepages() has a private bio context, exofs_writepages() has page_collect fuse_writepages() has fuse_fill_wb_data nfs_writepages() has nfs_pageio_descriptor xfs_vm_writepages() has xfs_writepage_ctx All of these ->writepages implementations can hold pages under writeback in their private structures until write_cache_pages() returns, and hence they are all susceptible to this deadlock. Also worth noting is that ext4 has it's own bastardised version of write_cache_pages() and so it /may/ have an equivalent deadlock. I looked at the code long enough to understand that it has a similar retry loop for range_cyclic writeback reaching the end of the file and then promptly ran away before my eyes bled too much. I'll leave it for the ext4 developers to determine if their code is actually has this deadlock and how to fix it if it has. There's a few ways I can see avoid this deadlock. There's probably more, but these are the first I've though of: 1. get rid of range_cyclic altogether 2. range_cyclic always stops at EOF, and we start again from writeback index 0 on the next call into write_cache_pages() 2a. wcp also returns EAGAIN to ->writepages implementations to indicate range cyclic has hit EOF. writepages implementations can then flush the current context and call wpc again to continue. i.e. lift the retry into the ->writepages implementation 3. range_cyclic uses trylock_page() rather than lock_page(), and it skips pages it can't lock without blocking. It will already do this for pages under writeback, so this seems like a no-brainer 3a. all non-WB_SYNC_ALL writeback uses trylock_page() to avoid blocking as per pages under writeback. I don't think #1 is an option - range_cyclic prevents frequently dirtied lower file offset from starving background writeback of rarely touched higher file offsets. #2 is simple, and I don't think it will have any impact on performance as going back to the start of the file implies an immediate seek. We'll have exactly the same number of seeks if we switch writeback to another inode, and then come back to this one later and restart from index 0. #2a is pretty much "status quo without the deadlock". Moving the retry loop up into the wcp caller means we can issue IO on the pending pages before calling wcp again, and so avoid locking or waiting on pages in the wrong order. I'm not convinced we need to do this given that we get the same thing from #2 on the next writeback call from the writeback infrastructure. #3 is really just a band-aid - it doesn't fix the access/wait inversion problem, just prevents it from becoming a deadlock situation. I'd prefer we fix the inversion, not sweep it under the carpet like this. #3a is really an optimisation that just so happens to include the band-aid fix of #3. So it seems that the simplest way to fix this issue is to implement solution #2 Link: http://lkml.kernel.org/r/20181005054526.21507-1-david@fromorbit.com Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Jan Kara <jack@suse.de> Cc: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-27 06:09:45 +08:00
*
* To avoid deadlocks between range_cyclic writeback and callers that hold
* pages in PageWriteback to aggregate IO until write_cache_pages() returns,
* we do not loop back to the start of the file. Doing so causes a page
* lock/page writeback access order inversion - we should only ever lock
* multiple pages in ascending page->index order, and looping back to the start
* of the file violates that rule and causes deadlocks.
*
* Return: %0 on success, negative error code otherwise
*/
int write_cache_pages(struct address_space *mapping,
struct writeback_control *wbc, writepage_t writepage,
void *data)
{
int ret = 0;
int done = 0;
mm/page-writeback.c: don't break integrity writeback on ->writepage() error write_cache_pages() is used in both background and integrity writeback scenarios by various filesystems. Background writeback is mostly concerned with cleaning a certain number of dirty pages based on various mm heuristics. It may not write the full set of dirty pages or wait for I/O to complete. Integrity writeback is responsible for persisting a set of dirty pages before the writeback job completes. For example, an fsync() call must perform integrity writeback to ensure data is on disk before the call returns. write_cache_pages() unconditionally breaks out of its processing loop in the event of a ->writepage() error. This is fine for background writeback, which had no strict requirements and will eventually come around again. This can cause problems for integrity writeback on filesystems that might need to clean up state associated with failed page writeouts. For example, XFS performs internal delayed allocation accounting before returning a ->writepage() error, where applicable. If the current writeback happens to be associated with an unmount and write_cache_pages() completes the writeback prematurely due to error, the filesystem is unmounted in an inconsistent state if dirty+delalloc pages still exist. To handle this problem, update write_cache_pages() to always process the full set of pages for integrity writeback regardless of ->writepage() errors. Save the first encountered error and return it to the caller once complete. This facilitates XFS (or any other fs that expects integrity writeback to process the entire set of dirty pages) to clean up its internal state completely in the event of persistent mapping errors. Background writeback continues to exit on the first error encountered. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.com Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 16:37:20 +08:00
int error;
struct pagevec pvec;
int nr_pages;
pgoff_t uninitialized_var(writeback_index);
pgoff_t index;
pgoff_t end; /* Inclusive */
pgoff_t done_index;
int range_whole = 0;
xa_mark_t tag;
pagevec_init(&pvec);
if (wbc->range_cyclic) {
writeback_index = mapping->writeback_index; /* prev offset */
index = writeback_index;
end = -1;
} else {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
index = wbc->range_start >> PAGE_SHIFT;
end = wbc->range_end >> PAGE_SHIFT;
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
range_whole = 1;
}
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) {
tag_pages_for_writeback(mapping, index, end);
tag = PAGECACHE_TAG_TOWRITE;
} else {
tag = PAGECACHE_TAG_DIRTY;
}
done_index = index;
while (!done && (index <= end)) {
int i;
nr_pages = pagevec_lookup_range_tag(&pvec, mapping, &index, end,
tag);
if (nr_pages == 0)
break;
for (i = 0; i < nr_pages; i++) {
struct page *page = pvec.pages[i];
writeback: make mapping->writeback_index to point to the last written page For range-cyclic writeback (e.g. kupdate), the writeback code sets a continuation point of the next writeback to mapping->writeback_index which is set the page after the last written page. This happens so that we evenly write the whole file even if pages in it get continuously redirtied. However, in some cases, sequential writer is writing in the middle of the page and it just redirties the last written page by continuing from that. For example with an application which uses a file as a big ring buffer we see: [1st writeback session] ... flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898514 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898522 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898530 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898538 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898546 + 8 kworker/0:1-11 4571: block_rq_issue: 8,0 W 0 () 94898514 + 40 >> flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898554 + 8 >> flush-8:0-2743 4571: block_rq_issue: 8,0 W 0 () 94898554 + 8 [2nd writeback session after 35sec] flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898562 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898570 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898578 + 8 ... kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94898562 + 640 kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94899202 + 72 ... flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899962 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899970 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899978 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899986 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899994 + 8 kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94899962 + 40 >> flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898554 + 8 >> flush-8:0-2743 4606: block_rq_issue: 8,0 W 0 () 94898554 + 8 So we seeked back to 94898554 after we wrote all the pages at the end of the file. This extra seek seems unnecessary. If we continue writeback from the last written page, we can avoid it and do not cause harm to other cases. The original intent of even writeout over the whole file is preserved and if the page does not get redirtied pagevec_lookup_tag() just skips it. As an exceptional case, when I/O error happens, set done_index to the next page as the comment in the code suggests. Tested-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-03-23 07:33:40 +08:00
done_index = page->index;
lock_page(page);
/*
* Page truncated or invalidated. We can freely skip it
* then, even for data integrity operations: the page
* has disappeared concurrently, so there could be no
* real expectation of this data interity operation
* even if there is now a new, dirty page at the same
* pagecache address.
*/
if (unlikely(page->mapping != mapping)) {
continue_unlock:
unlock_page(page);
continue;
}
if (!PageDirty(page)) {
/* someone wrote it for us */
goto continue_unlock;
}
if (PageWriteback(page)) {
if (wbc->sync_mode != WB_SYNC_NONE)
wait_on_page_writeback(page);
else
goto continue_unlock;
}
BUG_ON(PageWriteback(page));
if (!clear_page_dirty_for_io(page))
goto continue_unlock;
trace_wbc_writepage(wbc, inode_to_bdi(mapping->host));
mm/page-writeback.c: don't break integrity writeback on ->writepage() error write_cache_pages() is used in both background and integrity writeback scenarios by various filesystems. Background writeback is mostly concerned with cleaning a certain number of dirty pages based on various mm heuristics. It may not write the full set of dirty pages or wait for I/O to complete. Integrity writeback is responsible for persisting a set of dirty pages before the writeback job completes. For example, an fsync() call must perform integrity writeback to ensure data is on disk before the call returns. write_cache_pages() unconditionally breaks out of its processing loop in the event of a ->writepage() error. This is fine for background writeback, which had no strict requirements and will eventually come around again. This can cause problems for integrity writeback on filesystems that might need to clean up state associated with failed page writeouts. For example, XFS performs internal delayed allocation accounting before returning a ->writepage() error, where applicable. If the current writeback happens to be associated with an unmount and write_cache_pages() completes the writeback prematurely due to error, the filesystem is unmounted in an inconsistent state if dirty+delalloc pages still exist. To handle this problem, update write_cache_pages() to always process the full set of pages for integrity writeback regardless of ->writepage() errors. Save the first encountered error and return it to the caller once complete. This facilitates XFS (or any other fs that expects integrity writeback to process the entire set of dirty pages) to clean up its internal state completely in the event of persistent mapping errors. Background writeback continues to exit on the first error encountered. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.com Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 16:37:20 +08:00
error = (*writepage)(page, wbc, data);
if (unlikely(error)) {
/*
* Handle errors according to the type of
* writeback. There's no need to continue for
* background writeback. Just push done_index
* past this page so media errors won't choke
* writeout for the entire file. For integrity
* writeback, we must process the entire dirty
* set regardless of errors because the fs may
* still have state to clear for each page. In
* that case we continue processing and return
* the first error.
*/
if (error == AOP_WRITEPAGE_ACTIVATE) {
unlock_page(page);
mm/page-writeback.c: don't break integrity writeback on ->writepage() error write_cache_pages() is used in both background and integrity writeback scenarios by various filesystems. Background writeback is mostly concerned with cleaning a certain number of dirty pages based on various mm heuristics. It may not write the full set of dirty pages or wait for I/O to complete. Integrity writeback is responsible for persisting a set of dirty pages before the writeback job completes. For example, an fsync() call must perform integrity writeback to ensure data is on disk before the call returns. write_cache_pages() unconditionally breaks out of its processing loop in the event of a ->writepage() error. This is fine for background writeback, which had no strict requirements and will eventually come around again. This can cause problems for integrity writeback on filesystems that might need to clean up state associated with failed page writeouts. For example, XFS performs internal delayed allocation accounting before returning a ->writepage() error, where applicable. If the current writeback happens to be associated with an unmount and write_cache_pages() completes the writeback prematurely due to error, the filesystem is unmounted in an inconsistent state if dirty+delalloc pages still exist. To handle this problem, update write_cache_pages() to always process the full set of pages for integrity writeback regardless of ->writepage() errors. Save the first encountered error and return it to the caller once complete. This facilitates XFS (or any other fs that expects integrity writeback to process the entire set of dirty pages) to clean up its internal state completely in the event of persistent mapping errors. Background writeback continues to exit on the first error encountered. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.com Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 16:37:20 +08:00
error = 0;
} else if (wbc->sync_mode != WB_SYNC_ALL) {
ret = error;
writeback: make mapping->writeback_index to point to the last written page For range-cyclic writeback (e.g. kupdate), the writeback code sets a continuation point of the next writeback to mapping->writeback_index which is set the page after the last written page. This happens so that we evenly write the whole file even if pages in it get continuously redirtied. However, in some cases, sequential writer is writing in the middle of the page and it just redirties the last written page by continuing from that. For example with an application which uses a file as a big ring buffer we see: [1st writeback session] ... flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898514 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898522 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898530 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898538 + 8 flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898546 + 8 kworker/0:1-11 4571: block_rq_issue: 8,0 W 0 () 94898514 + 40 >> flush-8:0-2743 4571: block_bio_queue: 8,0 W 94898554 + 8 >> flush-8:0-2743 4571: block_rq_issue: 8,0 W 0 () 94898554 + 8 [2nd writeback session after 35sec] flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898562 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898570 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898578 + 8 ... kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94898562 + 640 kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94899202 + 72 ... flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899962 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899970 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899978 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899986 + 8 flush-8:0-2743 4606: block_bio_queue: 8,0 W 94899994 + 8 kworker/0:1-11 4606: block_rq_issue: 8,0 W 0 () 94899962 + 40 >> flush-8:0-2743 4606: block_bio_queue: 8,0 W 94898554 + 8 >> flush-8:0-2743 4606: block_rq_issue: 8,0 W 0 () 94898554 + 8 So we seeked back to 94898554 after we wrote all the pages at the end of the file. This extra seek seems unnecessary. If we continue writeback from the last written page, we can avoid it and do not cause harm to other cases. The original intent of even writeout over the whole file is preserved and if the page does not get redirtied pagevec_lookup_tag() just skips it. As an exceptional case, when I/O error happens, set done_index to the next page as the comment in the code suggests. Tested-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-03-23 07:33:40 +08:00
done_index = page->index + 1;
done = 1;
break;
}
mm/page-writeback.c: don't break integrity writeback on ->writepage() error write_cache_pages() is used in both background and integrity writeback scenarios by various filesystems. Background writeback is mostly concerned with cleaning a certain number of dirty pages based on various mm heuristics. It may not write the full set of dirty pages or wait for I/O to complete. Integrity writeback is responsible for persisting a set of dirty pages before the writeback job completes. For example, an fsync() call must perform integrity writeback to ensure data is on disk before the call returns. write_cache_pages() unconditionally breaks out of its processing loop in the event of a ->writepage() error. This is fine for background writeback, which had no strict requirements and will eventually come around again. This can cause problems for integrity writeback on filesystems that might need to clean up state associated with failed page writeouts. For example, XFS performs internal delayed allocation accounting before returning a ->writepage() error, where applicable. If the current writeback happens to be associated with an unmount and write_cache_pages() completes the writeback prematurely due to error, the filesystem is unmounted in an inconsistent state if dirty+delalloc pages still exist. To handle this problem, update write_cache_pages() to always process the full set of pages for integrity writeback regardless of ->writepage() errors. Save the first encountered error and return it to the caller once complete. This facilitates XFS (or any other fs that expects integrity writeback to process the entire set of dirty pages) to clean up its internal state completely in the event of persistent mapping errors. Background writeback continues to exit on the first error encountered. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.com Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 16:37:20 +08:00
if (!ret)
ret = error;
}
writeback: write_cache_pages doesn't terminate at nr_to_write <= 0 I noticed XFS writeback in 2.6.36-rc1 was much slower than it should have been. Enabling writeback tracing showed: flush-253:16-8516 [007] 1342952.351608: wbc_writepage: bdi 253:16: towrt=1024 skip=0 mode=0 kupd=0 bgrd=1 reclm=0 cyclic=1 more=0 older=0x0 start=0x0 end=0x0 flush-253:16-8516 [007] 1342952.351654: wbc_writepage: bdi 253:16: towrt=1023 skip=0 mode=0 kupd=0 bgrd=1 reclm=0 cyclic=1 more=0 older=0x0 start=0x0 end=0x0 flush-253:16-8516 [000] 1342952.369520: wbc_writepage: bdi 253:16: towrt=0 skip=0 mode=0 kupd=0 bgrd=1 reclm=0 cyclic=1 more=0 older=0x0 start=0x0 end=0x0 flush-253:16-8516 [000] 1342952.369542: wbc_writepage: bdi 253:16: towrt=-1 skip=0 mode=0 kupd=0 bgrd=1 reclm=0 cyclic=1 more=0 older=0x0 start=0x0 end=0x0 flush-253:16-8516 [000] 1342952.369549: wbc_writepage: bdi 253:16: towrt=-2 skip=0 mode=0 kupd=0 bgrd=1 reclm=0 cyclic=1 more=0 older=0x0 start=0x0 end=0x0 Writeback is not terminating in background writeback if ->writepage is returning with wbc->nr_to_write == 0, resulting in sub-optimal single page writeback on XFS. Fix the write_cache_pages loop to terminate correctly when this situation occurs and so prevent this sub-optimal background writeback pattern. This improves sustained sequential buffered write performance from around 250MB/s to 750MB/s for a 100GB file on an XFS filesystem on my 8p test VM. Cc:<stable@kernel.org> Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Wu Fengguang <fengguang.wu@intel.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-08-24 09:44:34 +08:00
/*
* We stop writing back only if we are not doing
* integrity sync. In case of integrity sync we have to
* keep going until we have written all the pages
* we tagged for writeback prior to entering this loop.
*/
if (--wbc->nr_to_write <= 0 &&
wbc->sync_mode == WB_SYNC_NONE) {
done = 1;
break;
mm: write_cache_pages integrity fix In write_cache_pages, nr_to_write is heeded even for data-integrity syncs, so the function will return success after writing out nr_to_write pages, even if that was not sufficient to guarantee data integrity. The callers tend to set it to values that could break data interity semantics easily in practice. For example, nr_to_write can be set to mapping->nr_pages * 2, however if a file has a single, dirty page, then fsync is called, subsequent pages might be concurrently added and dirtied, then write_cache_pages might writeout two of these newly dirty pages, while not writing out the old page that should have been written out. Fix this by ignoring nr_to_write if it is a data integrity sync. This is a data integrity bug. The reason this has been done in the past is to avoid stalling sync operations behind page dirtiers. "If a file has one dirty page at offset 1000000000000000 then someone does an fsync() and someone else gets in first and starts madly writing pages at offset 0, we want to write that page at 1000000000000000. Somehow." What we do today is return success after an arbitrary amount of pages are written, whether or not we have provided the data-integrity semantics that the caller has asked for. Even this doesn't actually fix all stall cases completely: in the above situation, if the file has a huge number of pages in pagecache (but not dirty), then mapping->nrpages is going to be huge, even if pages are being dirtied. This change does indeed make the possibility of long stalls lager, and that's not a good thing, but lying about data integrity is even worse. We have to either perform the sync, or return -ELINUXISLAME so at least the caller knows what has happened. There are subsequent competing approaches in the works to solve the stall problems properly, without compromising data integrity. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Chris Mason <chris.mason@oracle.com> Cc: Dave Chinner <david@fromorbit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-07 06:39:08 +08:00
}
}
pagevec_release(&pvec);
cond_resched();
}
mm/page-writeback.c: fix range_cyclic writeback vs writepages deadlock We've recently seen a workload on XFS filesystems with a repeatable deadlock between background writeback and a multi-process application doing concurrent writes and fsyncs to a small range of a file. range_cyclic writeback Process 1 Process 2 xfs_vm_writepages write_cache_pages writeback_index = 2 cycled = 0 .... find page 2 dirty lock Page 2 ->writepage page 2 writeback page 2 clean page 2 added to bio no more pages write() locks page 1 dirties page 1 locks page 2 dirties page 1 fsync() .... xfs_vm_writepages write_cache_pages start index 0 find page 1 towrite lock Page 1 ->writepage page 1 writeback page 1 clean page 1 added to bio find page 2 towrite lock Page 2 page 2 is writeback <blocks> write() locks page 1 dirties page 1 fsync() .... xfs_vm_writepages write_cache_pages start index 0 !done && !cycled sets index to 0, restarts lookup find page 1 dirty find page 1 towrite lock Page 1 page 1 is writeback <blocks> lock Page 1 <blocks> DEADLOCK because: - process 1 needs page 2 writeback to complete to make enough progress to issue IO pending for page 1 - writeback needs page 1 writeback to complete so process 2 can progress and unlock the page it is blocked on, then it can issue the IO pending for page 2 - process 2 can't make progress until process 1 issues IO for page 1 The underlying cause of the problem here is that range_cyclic writeback is processing pages in descending index order as we hold higher index pages in a structure controlled from above write_cache_pages(). The write_cache_pages() caller needs to be able to submit these pages for IO before write_cache_pages restarts writeback at mapping index 0 to avoid wcp inverting the page lock/writeback wait order. generic_writepages() is not susceptible to this bug as it has no private context held across write_cache_pages() - filesystems using this infrastructure always submit pages in ->writepage immediately and so there is no problem with range_cyclic going back to mapping index 0. However: mpage_writepages() has a private bio context, exofs_writepages() has page_collect fuse_writepages() has fuse_fill_wb_data nfs_writepages() has nfs_pageio_descriptor xfs_vm_writepages() has xfs_writepage_ctx All of these ->writepages implementations can hold pages under writeback in their private structures until write_cache_pages() returns, and hence they are all susceptible to this deadlock. Also worth noting is that ext4 has it's own bastardised version of write_cache_pages() and so it /may/ have an equivalent deadlock. I looked at the code long enough to understand that it has a similar retry loop for range_cyclic writeback reaching the end of the file and then promptly ran away before my eyes bled too much. I'll leave it for the ext4 developers to determine if their code is actually has this deadlock and how to fix it if it has. There's a few ways I can see avoid this deadlock. There's probably more, but these are the first I've though of: 1. get rid of range_cyclic altogether 2. range_cyclic always stops at EOF, and we start again from writeback index 0 on the next call into write_cache_pages() 2a. wcp also returns EAGAIN to ->writepages implementations to indicate range cyclic has hit EOF. writepages implementations can then flush the current context and call wpc again to continue. i.e. lift the retry into the ->writepages implementation 3. range_cyclic uses trylock_page() rather than lock_page(), and it skips pages it can't lock without blocking. It will already do this for pages under writeback, so this seems like a no-brainer 3a. all non-WB_SYNC_ALL writeback uses trylock_page() to avoid blocking as per pages under writeback. I don't think #1 is an option - range_cyclic prevents frequently dirtied lower file offset from starving background writeback of rarely touched higher file offsets. #2 is simple, and I don't think it will have any impact on performance as going back to the start of the file implies an immediate seek. We'll have exactly the same number of seeks if we switch writeback to another inode, and then come back to this one later and restart from index 0. #2a is pretty much "status quo without the deadlock". Moving the retry loop up into the wcp caller means we can issue IO on the pending pages before calling wcp again, and so avoid locking or waiting on pages in the wrong order. I'm not convinced we need to do this given that we get the same thing from #2 on the next writeback call from the writeback infrastructure. #3 is really just a band-aid - it doesn't fix the access/wait inversion problem, just prevents it from becoming a deadlock situation. I'd prefer we fix the inversion, not sweep it under the carpet like this. #3a is really an optimisation that just so happens to include the band-aid fix of #3. So it seems that the simplest way to fix this issue is to implement solution #2 Link: http://lkml.kernel.org/r/20181005054526.21507-1-david@fromorbit.com Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Jan Kara <jack@suse.de> Cc: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-27 06:09:45 +08:00
/*
* If we hit the last page and there is more work to be done: wrap
* back the index back to the start of the file for the next
* time we are called.
*/
if (wbc->range_cyclic && !done)
done_index = 0;
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
mapping->writeback_index = done_index;
return ret;
}
EXPORT_SYMBOL(write_cache_pages);
/*
* Function used by generic_writepages to call the real writepage
* function and set the mapping flags on error
*/
static int __writepage(struct page *page, struct writeback_control *wbc,
void *data)
{
struct address_space *mapping = data;
int ret = mapping->a_ops->writepage(page, wbc);
mapping_set_error(mapping, ret);
return ret;
}
/**
* generic_writepages - walk the list of dirty pages of the given address space and writepage() all of them.
* @mapping: address space structure to write
* @wbc: subtract the number of written pages from *@wbc->nr_to_write
*
* This is a library function, which implements the writepages()
* address_space_operation.
*
* Return: %0 on success, negative error code otherwise
*/
int generic_writepages(struct address_space *mapping,
struct writeback_control *wbc)
{
struct blk_plug plug;
int ret;
/* deal with chardevs and other special file */
if (!mapping->a_ops->writepage)
return 0;
blk_start_plug(&plug);
ret = write_cache_pages(mapping, wbc, __writepage, mapping);
blk_finish_plug(&plug);
return ret;
}
EXPORT_SYMBOL(generic_writepages);
int do_writepages(struct address_space *mapping, struct writeback_control *wbc)
{
int ret;
if (wbc->nr_to_write <= 0)
return 0;
while (1) {
if (mapping->a_ops->writepages)
ret = mapping->a_ops->writepages(mapping, wbc);
else
ret = generic_writepages(mapping, wbc);
if ((ret != -ENOMEM) || (wbc->sync_mode != WB_SYNC_ALL))
break;
cond_resched();
congestion_wait(BLK_RW_ASYNC, HZ/50);
}
return ret;
}
/**
* write_one_page - write out a single page and wait on I/O
* @page: the page to write
*
* The page must be locked by the caller and will be unlocked upon return.
*
* Note that the mapping's AS_EIO/AS_ENOSPC flags will be cleared when this
* function returns.
*
* Return: %0 on success, negative error code otherwise
*/
int write_one_page(struct page *page)
{
struct address_space *mapping = page->mapping;
int ret = 0;
struct writeback_control wbc = {
.sync_mode = WB_SYNC_ALL,
.nr_to_write = 1,
};
BUG_ON(!PageLocked(page));
wait_on_page_writeback(page);
if (clear_page_dirty_for_io(page)) {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
get_page(page);
ret = mapping->a_ops->writepage(page, &wbc);
if (ret == 0)
wait_on_page_writeback(page);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
put_page(page);
} else {
unlock_page(page);
}
if (!ret)
ret = filemap_check_errors(mapping);
return ret;
}
EXPORT_SYMBOL(write_one_page);
/*
* For address_spaces which do not use buffers nor write back.
*/
int __set_page_dirty_no_writeback(struct page *page)
{
if (!PageDirty(page))
return !TestSetPageDirty(page);
return 0;
}
/*
* Helper function for set_page_dirty family.
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
*
* Caller must hold lock_page_memcg().
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
*
* NOTE: This relies on being atomic wrt interrupts.
*/
void account_page_dirtied(struct page *page, struct address_space *mapping)
{
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:37 +08:00
struct inode *inode = mapping->host;
trace_writeback_dirty_page(page, mapping);
if (mapping_cap_account_dirty(mapping)) {
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:37 +08:00
struct bdi_writeback *wb;
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:37 +08:00
inode_attach_wb(inode, page);
wb = inode_to_wb(inode);
__inc_lruvec_page_state(page, NR_FILE_DIRTY);
__inc_zone_page_state(page, NR_ZONE_WRITE_PENDING);
__inc_node_page_state(page, NR_DIRTIED);
inc_wb_stat(wb, WB_RECLAIMABLE);
inc_wb_stat(wb, WB_DIRTIED);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
task_io_account_write(PAGE_SIZE);
current->nr_dirtied++;
this_cpu_inc(bdp_ratelimits);
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-27 00:06:56 +08:00
mem_cgroup_track_foreign_dirty(page, wb);
}
}
page_writeback: clean up mess around cancel_dirty_page() This patch replaces cancel_dirty_page() with a helper function account_page_cleaned() which only updates counters. It's called from truncate_complete_page() and from try_to_free_buffers() (hack for ext3). Page is locked in both cases, page-lock protects against concurrent dirtiers: see commit 2d6d7f982846 ("mm: protect set_page_dirty() from ongoing truncation"). Delete_from_page_cache() shouldn't be called for dirty pages, they must be handled by caller (either written or truncated). This patch treats final dirty accounting fixup at the end of __delete_from_page_cache() as a debug check and adds WARN_ON_ONCE() around it. If something removes dirty pages without proper handling that might be a bug and unwritten data might be lost. Hugetlbfs has no dirty pages accounting, ClearPageDirty() is enough here. cancel_dirty_page() in nfs_wb_page_cancel() is redundant. This is helper for nfs_invalidate_page() and it's called only in case complete invalidation. The mess was started in v2.6.20 after commits 46d2277c796f ("Clean up and make try_to_free_buffers() not race with dirty pages") and 3e67c0987d75 ("truncate: clear page dirtiness before running try_to_free_buffers()") first was reverted right in v2.6.20 in commit ecdfc9787fe5 ("Resurrect 'try_to_free_buffers()' VM hackery"), second in v2.6.25 commit a2b345642f53 ("Fix dirty page accounting leak with ext3 data=journal"). Custom fixes were introduced between these points. NFS in v2.6.23, commit 1b3b4a1a2deb ("NFS: Fix a write request leak in nfs_invalidate_page()"). Kludge in __delete_from_page_cache() in v2.6.24, commit 3a6927906f1b ("Do dirty page accounting when removing a page from the page cache"). Since v2.6.25 all of them are redundant. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Cc: Tejun Heo <tj@kernel.org> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-04-15 06:45:27 +08:00
/*
* Helper function for deaccounting dirty page without writeback.
*
* Caller must hold lock_page_memcg().
page_writeback: clean up mess around cancel_dirty_page() This patch replaces cancel_dirty_page() with a helper function account_page_cleaned() which only updates counters. It's called from truncate_complete_page() and from try_to_free_buffers() (hack for ext3). Page is locked in both cases, page-lock protects against concurrent dirtiers: see commit 2d6d7f982846 ("mm: protect set_page_dirty() from ongoing truncation"). Delete_from_page_cache() shouldn't be called for dirty pages, they must be handled by caller (either written or truncated). This patch treats final dirty accounting fixup at the end of __delete_from_page_cache() as a debug check and adds WARN_ON_ONCE() around it. If something removes dirty pages without proper handling that might be a bug and unwritten data might be lost. Hugetlbfs has no dirty pages accounting, ClearPageDirty() is enough here. cancel_dirty_page() in nfs_wb_page_cancel() is redundant. This is helper for nfs_invalidate_page() and it's called only in case complete invalidation. The mess was started in v2.6.20 after commits 46d2277c796f ("Clean up and make try_to_free_buffers() not race with dirty pages") and 3e67c0987d75 ("truncate: clear page dirtiness before running try_to_free_buffers()") first was reverted right in v2.6.20 in commit ecdfc9787fe5 ("Resurrect 'try_to_free_buffers()' VM hackery"), second in v2.6.25 commit a2b345642f53 ("Fix dirty page accounting leak with ext3 data=journal"). Custom fixes were introduced between these points. NFS in v2.6.23, commit 1b3b4a1a2deb ("NFS: Fix a write request leak in nfs_invalidate_page()"). Kludge in __delete_from_page_cache() in v2.6.24, commit 3a6927906f1b ("Do dirty page accounting when removing a page from the page cache"). Since v2.6.25 all of them are redundant. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Cc: Tejun Heo <tj@kernel.org> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-04-15 06:45:27 +08:00
*/
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
void account_page_cleaned(struct page *page, struct address_space *mapping,
struct bdi_writeback *wb)
page_writeback: clean up mess around cancel_dirty_page() This patch replaces cancel_dirty_page() with a helper function account_page_cleaned() which only updates counters. It's called from truncate_complete_page() and from try_to_free_buffers() (hack for ext3). Page is locked in both cases, page-lock protects against concurrent dirtiers: see commit 2d6d7f982846 ("mm: protect set_page_dirty() from ongoing truncation"). Delete_from_page_cache() shouldn't be called for dirty pages, they must be handled by caller (either written or truncated). This patch treats final dirty accounting fixup at the end of __delete_from_page_cache() as a debug check and adds WARN_ON_ONCE() around it. If something removes dirty pages without proper handling that might be a bug and unwritten data might be lost. Hugetlbfs has no dirty pages accounting, ClearPageDirty() is enough here. cancel_dirty_page() in nfs_wb_page_cancel() is redundant. This is helper for nfs_invalidate_page() and it's called only in case complete invalidation. The mess was started in v2.6.20 after commits 46d2277c796f ("Clean up and make try_to_free_buffers() not race with dirty pages") and 3e67c0987d75 ("truncate: clear page dirtiness before running try_to_free_buffers()") first was reverted right in v2.6.20 in commit ecdfc9787fe5 ("Resurrect 'try_to_free_buffers()' VM hackery"), second in v2.6.25 commit a2b345642f53 ("Fix dirty page accounting leak with ext3 data=journal"). Custom fixes were introduced between these points. NFS in v2.6.23, commit 1b3b4a1a2deb ("NFS: Fix a write request leak in nfs_invalidate_page()"). Kludge in __delete_from_page_cache() in v2.6.24, commit 3a6927906f1b ("Do dirty page accounting when removing a page from the page cache"). Since v2.6.25 all of them are redundant. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Cc: Tejun Heo <tj@kernel.org> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-04-15 06:45:27 +08:00
{
if (mapping_cap_account_dirty(mapping)) {
dec_lruvec_page_state(page, NR_FILE_DIRTY);
dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
dec_wb_stat(wb, WB_RECLAIMABLE);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
task_io_account_cancelled_write(PAGE_SIZE);
page_writeback: clean up mess around cancel_dirty_page() This patch replaces cancel_dirty_page() with a helper function account_page_cleaned() which only updates counters. It's called from truncate_complete_page() and from try_to_free_buffers() (hack for ext3). Page is locked in both cases, page-lock protects against concurrent dirtiers: see commit 2d6d7f982846 ("mm: protect set_page_dirty() from ongoing truncation"). Delete_from_page_cache() shouldn't be called for dirty pages, they must be handled by caller (either written or truncated). This patch treats final dirty accounting fixup at the end of __delete_from_page_cache() as a debug check and adds WARN_ON_ONCE() around it. If something removes dirty pages without proper handling that might be a bug and unwritten data might be lost. Hugetlbfs has no dirty pages accounting, ClearPageDirty() is enough here. cancel_dirty_page() in nfs_wb_page_cancel() is redundant. This is helper for nfs_invalidate_page() and it's called only in case complete invalidation. The mess was started in v2.6.20 after commits 46d2277c796f ("Clean up and make try_to_free_buffers() not race with dirty pages") and 3e67c0987d75 ("truncate: clear page dirtiness before running try_to_free_buffers()") first was reverted right in v2.6.20 in commit ecdfc9787fe5 ("Resurrect 'try_to_free_buffers()' VM hackery"), second in v2.6.25 commit a2b345642f53 ("Fix dirty page accounting leak with ext3 data=journal"). Custom fixes were introduced between these points. NFS in v2.6.23, commit 1b3b4a1a2deb ("NFS: Fix a write request leak in nfs_invalidate_page()"). Kludge in __delete_from_page_cache() in v2.6.24, commit 3a6927906f1b ("Do dirty page accounting when removing a page from the page cache"). Since v2.6.25 all of them are redundant. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Cc: Tejun Heo <tj@kernel.org> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-04-15 06:45:27 +08:00
}
}
/*
* For address_spaces which do not use buffers. Just tag the page as dirty in
* the xarray.
*
* This is also used when a single buffer is being dirtied: we want to set the
* page dirty in that case, but not all the buffers. This is a "bottom-up"
* dirtying, whereas __set_page_dirty_buffers() is a "top-down" dirtying.
*
mm: protect set_page_dirty() from ongoing truncation Tejun, while reviewing the code, spotted the following race condition between the dirtying and truncation of a page: __set_page_dirty_nobuffers() __delete_from_page_cache() if (TestSetPageDirty(page)) page->mapping = NULL if (PageDirty()) dec_zone_page_state(page, NR_FILE_DIRTY); dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); if (page->mapping) account_page_dirtied(page) __inc_zone_page_state(page, NR_FILE_DIRTY); __inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); which results in an imbalance of NR_FILE_DIRTY and BDI_RECLAIMABLE. Dirtiers usually lock out truncation, either by holding the page lock directly, or in case of zap_pte_range(), by pinning the mapcount with the page table lock held. The notable exception to this rule, though, is do_wp_page(), for which this race exists. However, do_wp_page() already waits for a locked page to unlock before setting the dirty bit, in order to prevent a race where clear_page_dirty() misses the page bit in the presence of dirty ptes. Upgrade that wait to a fully locked set_page_dirty() to also cover the situation explained above. Afterwards, the code in set_page_dirty() dealing with a truncation race is no longer needed. Remove it. Reported-by: Tejun Heo <tj@kernel.org> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-01-09 06:32:18 +08:00
* The caller must ensure this doesn't race with truncation. Most will simply
* hold the page lock, but e.g. zap_pte_range() calls with the page mapped and
* the pte lock held, which also locks out truncation.
*/
int __set_page_dirty_nobuffers(struct page *page)
{
lock_page_memcg(page);
if (!TestSetPageDirty(page)) {
struct address_space *mapping = page_mapping(page);
mm: __set_page_dirty_nobuffers() uses spin_lock_irqsave() instead of spin_lock_irq() During aio stress test, we observed the following lockdep warning. This mean AIO+numa_balancing is currently deadlockable. The problem is, aio_migratepage disable interrupt, but __set_page_dirty_nobuffers unintentionally enable it again. Generally, all helper function should use spin_lock_irqsave() instead of spin_lock_irq() because they don't know caller at all. other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&ctx->completion_lock)->rlock); <Interrupt> lock(&(&ctx->completion_lock)->rlock); *** DEADLOCK *** dump_stack+0x19/0x1b print_usage_bug+0x1f7/0x208 mark_lock+0x21d/0x2a0 mark_held_locks+0xb9/0x140 trace_hardirqs_on_caller+0x105/0x1d0 trace_hardirqs_on+0xd/0x10 _raw_spin_unlock_irq+0x2c/0x50 __set_page_dirty_nobuffers+0x8c/0xf0 migrate_page_copy+0x434/0x540 aio_migratepage+0xb1/0x140 move_to_new_page+0x7d/0x230 migrate_pages+0x5e5/0x700 migrate_misplaced_page+0xbc/0xf0 do_numa_page+0x102/0x190 handle_pte_fault+0x241/0x970 handle_mm_fault+0x265/0x370 __do_page_fault+0x172/0x5a0 do_page_fault+0x1a/0x70 page_fault+0x28/0x30 Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Larry Woodman <lwoodman@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Johannes Weiner <jweiner@redhat.com> Acked-by: David Rientjes <rientjes@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-02-07 04:04:24 +08:00
unsigned long flags;
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
if (!mapping) {
unlock_page_memcg(page);
return 1;
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
}
xa_lock_irqsave(&mapping->i_pages, flags);
mm: protect set_page_dirty() from ongoing truncation Tejun, while reviewing the code, spotted the following race condition between the dirtying and truncation of a page: __set_page_dirty_nobuffers() __delete_from_page_cache() if (TestSetPageDirty(page)) page->mapping = NULL if (PageDirty()) dec_zone_page_state(page, NR_FILE_DIRTY); dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); if (page->mapping) account_page_dirtied(page) __inc_zone_page_state(page, NR_FILE_DIRTY); __inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); which results in an imbalance of NR_FILE_DIRTY and BDI_RECLAIMABLE. Dirtiers usually lock out truncation, either by holding the page lock directly, or in case of zap_pte_range(), by pinning the mapcount with the page table lock held. The notable exception to this rule, though, is do_wp_page(), for which this race exists. However, do_wp_page() already waits for a locked page to unlock before setting the dirty bit, in order to prevent a race where clear_page_dirty() misses the page bit in the presence of dirty ptes. Upgrade that wait to a fully locked set_page_dirty() to also cover the situation explained above. Afterwards, the code in set_page_dirty() dealing with a truncation race is no longer needed. Remove it. Reported-by: Tejun Heo <tj@kernel.org> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-01-09 06:32:18 +08:00
BUG_ON(page_mapping(page) != mapping);
WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page));
account_page_dirtied(page, mapping);
__xa_set_mark(&mapping->i_pages, page_index(page),
mm: protect set_page_dirty() from ongoing truncation Tejun, while reviewing the code, spotted the following race condition between the dirtying and truncation of a page: __set_page_dirty_nobuffers() __delete_from_page_cache() if (TestSetPageDirty(page)) page->mapping = NULL if (PageDirty()) dec_zone_page_state(page, NR_FILE_DIRTY); dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); if (page->mapping) account_page_dirtied(page) __inc_zone_page_state(page, NR_FILE_DIRTY); __inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); which results in an imbalance of NR_FILE_DIRTY and BDI_RECLAIMABLE. Dirtiers usually lock out truncation, either by holding the page lock directly, or in case of zap_pte_range(), by pinning the mapcount with the page table lock held. The notable exception to this rule, though, is do_wp_page(), for which this race exists. However, do_wp_page() already waits for a locked page to unlock before setting the dirty bit, in order to prevent a race where clear_page_dirty() misses the page bit in the presence of dirty ptes. Upgrade that wait to a fully locked set_page_dirty() to also cover the situation explained above. Afterwards, the code in set_page_dirty() dealing with a truncation race is no longer needed. Remove it. Reported-by: Tejun Heo <tj@kernel.org> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-01-09 06:32:18 +08:00
PAGECACHE_TAG_DIRTY);
xa_unlock_irqrestore(&mapping->i_pages, flags);
unlock_page_memcg(page);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
if (mapping->host) {
/* !PageAnon && !swapper_space */
__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
}
return 1;
}
unlock_page_memcg(page);
return 0;
}
EXPORT_SYMBOL(__set_page_dirty_nobuffers);
/*
* Call this whenever redirtying a page, to de-account the dirty counters
* (NR_DIRTIED, WB_DIRTIED, tsk->nr_dirtied), so that they match the written
* counters (NR_WRITTEN, WB_WRITTEN) in long term. The mismatches will lead to
* systematic errors in balanced_dirty_ratelimit and the dirty pages position
* control.
*/
void account_page_redirty(struct page *page)
{
struct address_space *mapping = page->mapping;
if (mapping && mapping_cap_account_dirty(mapping)) {
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
struct inode *inode = mapping->host;
struct bdi_writeback *wb;
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
struct wb_lock_cookie cookie = {};
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
wb = unlocked_inode_to_wb_begin(inode, &cookie);
current->nr_dirtied--;
dec_node_page_state(page, NR_DIRTIED);
dec_wb_stat(wb, WB_DIRTIED);
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
unlocked_inode_to_wb_end(inode, &cookie);
}
}
EXPORT_SYMBOL(account_page_redirty);
/*
* When a writepage implementation decides that it doesn't want to write this
* page for some reason, it should redirty the locked page via
* redirty_page_for_writepage() and it should then unlock the page and return 0
*/
int redirty_page_for_writepage(struct writeback_control *wbc, struct page *page)
{
int ret;
wbc->pages_skipped++;
ret = __set_page_dirty_nobuffers(page);
account_page_redirty(page);
return ret;
}
EXPORT_SYMBOL(redirty_page_for_writepage);
/*
* Dirty a page.
*
* For pages with a mapping this should be done under the page lock
* for the benefit of asynchronous memory errors who prefer a consistent
* dirty state. This rule can be broken in some special cases,
* but should be better not to.
*
* If the mapping doesn't provide a set_page_dirty a_op, then
* just fall through and assume that it wants buffer_heads.
*/
int set_page_dirty(struct page *page)
{
struct address_space *mapping = page_mapping(page);
page = compound_head(page);
if (likely(mapping)) {
int (*spd)(struct page *) = mapping->a_ops->set_page_dirty;
mm: reclaim invalidated page ASAP invalidate_mapping_pages is very big hint to reclaimer. It means user doesn't want to use the page any more. So in order to prevent working set page eviction, this patch move the page into tail of inactive list by PG_reclaim. Please, remember that pages in inactive list are working set as well as active list. If we don't move pages into inactive list's tail, pages near by tail of inactive list can be evicted although we have a big clue about useless pages. It's totally bad. Now PG_readahead/PG_reclaim is shared. fe3cba17 added ClearPageReclaim into clear_page_dirty_for_io for preventing fast reclaiming readahead marker page. In this series, PG_reclaim is used by invalidated page, too. If VM find the page is invalidated and it's dirty, it sets PG_reclaim to reclaim asap. Then, when the dirty page will be writeback, clear_page_dirty_for_io will clear PG_reclaim unconditionally. It disturbs this serie's goal. I think it's okay to clear PG_readahead when the page is dirty, not writeback time. So this patch moves ClearPageReadahead. In v4, ClearPageReadahead in set_page_dirty has a problem which is reported by Steven Barrett. It's due to compound page. Some driver(ex, audio) calls set_page_dirty with compound page which isn't on LRU. but my patch does ClearPageRelcaim on compound page. In non-CONFIG_PAGEFLAGS_EXTENDED, it breaks PageTail flag. I think it doesn't affect THP and pass my test with THP enabling but Cced Andrea for double check. Signed-off-by: Minchan Kim <minchan.kim@gmail.com> Reported-by: Steven Barrett <damentz@liquorix.net> Reviewed-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Rik van Riel <riel@redhat.com> Acked-by: Mel Gorman <mel@csn.ul.ie> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-03-23 07:32:54 +08:00
/*
* readahead/lru_deactivate_page could remain
* PG_readahead/PG_reclaim due to race with end_page_writeback
* About readahead, if the page is written, the flags would be
* reset. So no problem.
* About lru_deactivate_page, if the page is redirty, the flag
* will be reset. So no problem. but if the page is used by readahead
* it will confuse readahead and make it restart the size rampup
* process. But it's a trivial problem.
*/
if (PageReclaim(page))
ClearPageReclaim(page);
[PATCH] BLOCK: Make it possible to disable the block layer [try #6] Make it possible to disable the block layer. Not all embedded devices require it, some can make do with just JFFS2, NFS, ramfs, etc - none of which require the block layer to be present. This patch does the following: (*) Introduces CONFIG_BLOCK to disable the block layer, buffering and blockdev support. (*) Adds dependencies on CONFIG_BLOCK to any configuration item that controls an item that uses the block layer. This includes: (*) Block I/O tracing. (*) Disk partition code. (*) All filesystems that are block based, eg: Ext3, ReiserFS, ISOFS. (*) The SCSI layer. As far as I can tell, even SCSI chardevs use the block layer to do scheduling. Some drivers that use SCSI facilities - such as USB storage - end up disabled indirectly from this. (*) Various block-based device drivers, such as IDE and the old CDROM drivers. (*) MTD blockdev handling and FTL. (*) JFFS - which uses set_bdev_super(), something it could avoid doing by taking a leaf out of JFFS2's book. (*) Makes most of the contents of linux/blkdev.h, linux/buffer_head.h and linux/elevator.h contingent on CONFIG_BLOCK being set. sector_div() is, however, still used in places, and so is still available. (*) Also made contingent are the contents of linux/mpage.h, linux/genhd.h and parts of linux/fs.h. (*) Makes a number of files in fs/ contingent on CONFIG_BLOCK. (*) Makes mm/bounce.c (bounce buffering) contingent on CONFIG_BLOCK. (*) set_page_dirty() doesn't call __set_page_dirty_buffers() if CONFIG_BLOCK is not enabled. (*) fs/no-block.c is created to hold out-of-line stubs and things that are required when CONFIG_BLOCK is not set: (*) Default blockdev file operations (to give error ENODEV on opening). (*) Makes some /proc changes: (*) /proc/devices does not list any blockdevs. (*) /proc/diskstats and /proc/partitions are contingent on CONFIG_BLOCK. (*) Makes some compat ioctl handling contingent on CONFIG_BLOCK. (*) If CONFIG_BLOCK is not defined, makes sys_quotactl() return -ENODEV if given command other than Q_SYNC or if a special device is specified. (*) In init/do_mounts.c, no reference is made to the blockdev routines if CONFIG_BLOCK is not defined. This does not prohibit NFS roots or JFFS2. (*) The bdflush, ioprio_set and ioprio_get syscalls can now be absent (return error ENOSYS by way of cond_syscall if so). (*) The seclvl_bd_claim() and seclvl_bd_release() security calls do nothing if CONFIG_BLOCK is not set, since they can't then happen. Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2006-10-01 02:45:40 +08:00
#ifdef CONFIG_BLOCK
if (!spd)
spd = __set_page_dirty_buffers;
#endif
return (*spd)(page);
}
if (!PageDirty(page)) {
if (!TestSetPageDirty(page))
return 1;
}
return 0;
}
EXPORT_SYMBOL(set_page_dirty);
/*
* set_page_dirty() is racy if the caller has no reference against
* page->mapping->host, and if the page is unlocked. This is because another
* CPU could truncate the page off the mapping and then free the mapping.
*
* Usually, the page _is_ locked, or the caller is a user-space process which
* holds a reference on the inode by having an open file.
*
* In other cases, the page should be locked before running set_page_dirty().
*/
int set_page_dirty_lock(struct page *page)
{
int ret;
lock_page(page);
ret = set_page_dirty(page);
unlock_page(page);
return ret;
}
EXPORT_SYMBOL(set_page_dirty_lock);
/*
* This cancels just the dirty bit on the kernel page itself, it does NOT
* actually remove dirty bits on any mmap's that may be around. It also
* leaves the page tagged dirty, so any sync activity will still find it on
* the dirty lists, and in particular, clear_page_dirty_for_io() will still
* look at the dirty bits in the VM.
*
* Doing this should *normally* only ever be done when a page is truncated,
* and is not actually mapped anywhere at all. However, fs/buffer.c does
* this when it notices that somebody has cleaned out all the buffers on a
* page without actually doing it through the VM. Can you say "ext3 is
* horribly ugly"? Thought you could.
*/
mm: speed up cancel_dirty_page() for clean pages Patch series "Speed up page cache truncation", v1. When rebasing our enterprise distro to a newer kernel (from 4.4 to 4.12) we have noticed a regression in bonnie++ benchmark when deleting files. Eventually we have tracked this down to a fact that page cache truncation got slower by about 10%. There were both gains and losses in the above interval of kernels but we have been able to identify that commit 83929372f629 ("filemap: prepare find and delete operations for huge pages") caused about 10% regression on its own. After some investigation it didn't seem easily possible to fix the regression while maintaining the THP in page cache functionality so we've decided to optimize the page cache truncation path instead to make up for the change. This series is a result of that effort. Patch 1 is an easy speedup of cancel_dirty_page(). Patches 2-6 refactor page cache truncation code so that it is easier to batch radix tree operations. Patch 7 implements batching of deletes from the radix tree which more than makes up for the original regression. This patch (of 7): cancel_dirty_page() does quite some work even for clean pages (fetching of mapping, locking of memcg, atomic bit op on page flags) so it accounts for ~2.5% of cost of truncation of a clean page. That is not much but still dumb for something we don't need at all. Check whether a page is actually dirty and avoid any work if not. Link: http://lkml.kernel.org/r/20171010151937.26984-2-jack@suse.cz Signed-off-by: Jan Kara <jack@suse.cz> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Andi Kleen <ak@linux.intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-16 09:37:11 +08:00
void __cancel_dirty_page(struct page *page)
{
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
struct address_space *mapping = page_mapping(page);
if (mapping_cap_account_dirty(mapping)) {
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
struct inode *inode = mapping->host;
struct bdi_writeback *wb;
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
struct wb_lock_cookie cookie = {};
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
lock_page_memcg(page);
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
wb = unlocked_inode_to_wb_begin(inode, &cookie);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
if (TestClearPageDirty(page))
account_page_cleaned(page, mapping, wb);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
unlocked_inode_to_wb_end(inode, &cookie);
unlock_page_memcg(page);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
} else {
ClearPageDirty(page);
}
}
mm: speed up cancel_dirty_page() for clean pages Patch series "Speed up page cache truncation", v1. When rebasing our enterprise distro to a newer kernel (from 4.4 to 4.12) we have noticed a regression in bonnie++ benchmark when deleting files. Eventually we have tracked this down to a fact that page cache truncation got slower by about 10%. There were both gains and losses in the above interval of kernels but we have been able to identify that commit 83929372f629 ("filemap: prepare find and delete operations for huge pages") caused about 10% regression on its own. After some investigation it didn't seem easily possible to fix the regression while maintaining the THP in page cache functionality so we've decided to optimize the page cache truncation path instead to make up for the change. This series is a result of that effort. Patch 1 is an easy speedup of cancel_dirty_page(). Patches 2-6 refactor page cache truncation code so that it is easier to batch radix tree operations. Patch 7 implements batching of deletes from the radix tree which more than makes up for the original regression. This patch (of 7): cancel_dirty_page() does quite some work even for clean pages (fetching of mapping, locking of memcg, atomic bit op on page flags) so it accounts for ~2.5% of cost of truncation of a clean page. That is not much but still dumb for something we don't need at all. Check whether a page is actually dirty and avoid any work if not. Link: http://lkml.kernel.org/r/20171010151937.26984-2-jack@suse.cz Signed-off-by: Jan Kara <jack@suse.cz> Acked-by: Mel Gorman <mgorman@suse.de> Reviewed-by: Andi Kleen <ak@linux.intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-16 09:37:11 +08:00
EXPORT_SYMBOL(__cancel_dirty_page);
/*
* Clear a page's dirty flag, while caring for dirty memory accounting.
* Returns true if the page was previously dirty.
*
* This is for preparing to put the page under writeout. We leave the page
* tagged as dirty in the xarray so that a concurrent write-for-sync
* can discover it via a PAGECACHE_TAG_DIRTY walk. The ->writepage
* implementation will run either set_page_writeback() or set_page_dirty(),
* at which stage we bring the page's dirty flag and xarray dirty tag
* back into sync.
*
* This incoherency between the page's dirty flag and xarray tag is
* unfortunate, but it only exists while the page is locked.
*/
int clear_page_dirty_for_io(struct page *page)
{
struct address_space *mapping = page_mapping(page);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
int ret = 0;
VM_BUG_ON_PAGE(!PageLocked(page), page);
VM: Fix nasty and subtle race in shared mmap'ed page writeback The VM layer (on the face of it, fairly reasonably) expected that when it does a ->writepage() call to the filesystem, it would write out the full page at that point in time. Especially since it had earlier marked the whole page dirty with "set_page_dirty()". But that isn't actually the case: ->writepage() does not actually write a page, it writes the parts of the page that have been explicitly marked dirty before, *and* that had not got written out for other reasons since the last time we told it they were dirty. That last caveat is the important one. Which _most_ of the time ends up being the whole page (since we had called "set_page_dirty()" on the page earlier), but if the filesystem had done any dirty flushing of its own (for example, to honor some internal write ordering guarantees), it might end up doing only a partial page IO (or none at all) when ->writepage() is actually called. That is the correct thing in general (since we actually often _want_ only the known-dirty parts of the page to be written out), but the shared dirty page handling had implicitly forgotten about these details, and had a number of cases where it was doing just the "->writepage()" part, without telling the low-level filesystem that the whole page might have been re-dirtied as part of being mapped writably into user space. Since most of the time the FS did actually write out the full page, we didn't notice this for a loong time, and this needed some really odd patterns to trigger. But it caused occasional corruption with rtorrent and with the Debian "apt" database, because both use shared mmaps to update the end result. This fixes it. Finally. After way too much hair-pulling. Acked-by: Nick Piggin <nickpiggin@yahoo.com.au> Acked-by: Martin J. Bligh <mbligh@google.com> Acked-by: Martin Michlmayr <tbm@cyrius.com> Acked-by: Martin Johansson <martin@fatbob.nu> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Andrei Popa <andrei.popa@i-neo.ro> Cc: High Dickins <hugh@veritas.com> Cc: Andrew Morton <akpm@osdl.org>, Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Segher Boessenkool <segher@kernel.crashing.org> Cc: David Miller <davem@davemloft.net> Cc: Arjan van de Ven <arjan@infradead.org> Cc: Gordon Farquharson <gordonfarquharson@gmail.com> Cc: Guillaume Chazarain <guichaz@yahoo.fr> Cc: Theodore Tso <tytso@mit.edu> Cc: Kenneth Cheng <kenneth.w.chen@intel.com> Cc: Tobias Diedrich <ranma@tdiedrich.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-30 02:00:58 +08:00
if (mapping && mapping_cap_account_dirty(mapping)) {
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
struct inode *inode = mapping->host;
struct bdi_writeback *wb;
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
struct wb_lock_cookie cookie = {};
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
VM: Fix nasty and subtle race in shared mmap'ed page writeback The VM layer (on the face of it, fairly reasonably) expected that when it does a ->writepage() call to the filesystem, it would write out the full page at that point in time. Especially since it had earlier marked the whole page dirty with "set_page_dirty()". But that isn't actually the case: ->writepage() does not actually write a page, it writes the parts of the page that have been explicitly marked dirty before, *and* that had not got written out for other reasons since the last time we told it they were dirty. That last caveat is the important one. Which _most_ of the time ends up being the whole page (since we had called "set_page_dirty()" on the page earlier), but if the filesystem had done any dirty flushing of its own (for example, to honor some internal write ordering guarantees), it might end up doing only a partial page IO (or none at all) when ->writepage() is actually called. That is the correct thing in general (since we actually often _want_ only the known-dirty parts of the page to be written out), but the shared dirty page handling had implicitly forgotten about these details, and had a number of cases where it was doing just the "->writepage()" part, without telling the low-level filesystem that the whole page might have been re-dirtied as part of being mapped writably into user space. Since most of the time the FS did actually write out the full page, we didn't notice this for a loong time, and this needed some really odd patterns to trigger. But it caused occasional corruption with rtorrent and with the Debian "apt" database, because both use shared mmaps to update the end result. This fixes it. Finally. After way too much hair-pulling. Acked-by: Nick Piggin <nickpiggin@yahoo.com.au> Acked-by: Martin J. Bligh <mbligh@google.com> Acked-by: Martin Michlmayr <tbm@cyrius.com> Acked-by: Martin Johansson <martin@fatbob.nu> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Andrei Popa <andrei.popa@i-neo.ro> Cc: High Dickins <hugh@veritas.com> Cc: Andrew Morton <akpm@osdl.org>, Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Segher Boessenkool <segher@kernel.crashing.org> Cc: David Miller <davem@davemloft.net> Cc: Arjan van de Ven <arjan@infradead.org> Cc: Gordon Farquharson <gordonfarquharson@gmail.com> Cc: Guillaume Chazarain <guichaz@yahoo.fr> Cc: Theodore Tso <tytso@mit.edu> Cc: Kenneth Cheng <kenneth.w.chen@intel.com> Cc: Tobias Diedrich <ranma@tdiedrich.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-30 02:00:58 +08:00
/*
* Yes, Virginia, this is indeed insane.
*
* We use this sequence to make sure that
* (a) we account for dirty stats properly
* (b) we tell the low-level filesystem to
* mark the whole page dirty if it was
* dirty in a pagetable. Only to then
* (c) clean the page again and return 1 to
* cause the writeback.
*
* This way we avoid all nasty races with the
* dirty bit in multiple places and clearing
* them concurrently from different threads.
*
* Note! Normally the "set_page_dirty(page)"
* has no effect on the actual dirty bit - since
* that will already usually be set. But we
* need the side effects, and it can help us
* avoid races.
*
* We basically use the page "master dirty bit"
* as a serialization point for all the different
* threads doing their things.
*/
if (page_mkclean(page))
set_page_dirty(page);
/*
* We carefully synchronise fault handlers against
* installing a dirty pte and marking the page dirty
mm: protect set_page_dirty() from ongoing truncation Tejun, while reviewing the code, spotted the following race condition between the dirtying and truncation of a page: __set_page_dirty_nobuffers() __delete_from_page_cache() if (TestSetPageDirty(page)) page->mapping = NULL if (PageDirty()) dec_zone_page_state(page, NR_FILE_DIRTY); dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); if (page->mapping) account_page_dirtied(page) __inc_zone_page_state(page, NR_FILE_DIRTY); __inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); which results in an imbalance of NR_FILE_DIRTY and BDI_RECLAIMABLE. Dirtiers usually lock out truncation, either by holding the page lock directly, or in case of zap_pte_range(), by pinning the mapcount with the page table lock held. The notable exception to this rule, though, is do_wp_page(), for which this race exists. However, do_wp_page() already waits for a locked page to unlock before setting the dirty bit, in order to prevent a race where clear_page_dirty() misses the page bit in the presence of dirty ptes. Upgrade that wait to a fully locked set_page_dirty() to also cover the situation explained above. Afterwards, the code in set_page_dirty() dealing with a truncation race is no longer needed. Remove it. Reported-by: Tejun Heo <tj@kernel.org> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-01-09 06:32:18 +08:00
* at this point. We do this by having them hold the
* page lock while dirtying the page, and pages are
* always locked coming in here, so we get the desired
* exclusion.
*/
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
wb = unlocked_inode_to_wb_begin(inode, &cookie);
VM: Fix nasty and subtle race in shared mmap'ed page writeback The VM layer (on the face of it, fairly reasonably) expected that when it does a ->writepage() call to the filesystem, it would write out the full page at that point in time. Especially since it had earlier marked the whole page dirty with "set_page_dirty()". But that isn't actually the case: ->writepage() does not actually write a page, it writes the parts of the page that have been explicitly marked dirty before, *and* that had not got written out for other reasons since the last time we told it they were dirty. That last caveat is the important one. Which _most_ of the time ends up being the whole page (since we had called "set_page_dirty()" on the page earlier), but if the filesystem had done any dirty flushing of its own (for example, to honor some internal write ordering guarantees), it might end up doing only a partial page IO (or none at all) when ->writepage() is actually called. That is the correct thing in general (since we actually often _want_ only the known-dirty parts of the page to be written out), but the shared dirty page handling had implicitly forgotten about these details, and had a number of cases where it was doing just the "->writepage()" part, without telling the low-level filesystem that the whole page might have been re-dirtied as part of being mapped writably into user space. Since most of the time the FS did actually write out the full page, we didn't notice this for a loong time, and this needed some really odd patterns to trigger. But it caused occasional corruption with rtorrent and with the Debian "apt" database, because both use shared mmaps to update the end result. This fixes it. Finally. After way too much hair-pulling. Acked-by: Nick Piggin <nickpiggin@yahoo.com.au> Acked-by: Martin J. Bligh <mbligh@google.com> Acked-by: Martin Michlmayr <tbm@cyrius.com> Acked-by: Martin Johansson <martin@fatbob.nu> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Andrei Popa <andrei.popa@i-neo.ro> Cc: High Dickins <hugh@veritas.com> Cc: Andrew Morton <akpm@osdl.org>, Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Segher Boessenkool <segher@kernel.crashing.org> Cc: David Miller <davem@davemloft.net> Cc: Arjan van de Ven <arjan@infradead.org> Cc: Gordon Farquharson <gordonfarquharson@gmail.com> Cc: Guillaume Chazarain <guichaz@yahoo.fr> Cc: Theodore Tso <tytso@mit.edu> Cc: Kenneth Cheng <kenneth.w.chen@intel.com> Cc: Tobias Diedrich <ranma@tdiedrich.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-30 02:00:58 +08:00
if (TestClearPageDirty(page)) {
dec_lruvec_page_state(page, NR_FILE_DIRTY);
dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
writeback: implement unlocked_inode_to_wb transaction and use it for stat updates The mechanism for detecting whether an inode should switch its wb (bdi_writeback) association is now in place. This patch build the framework for the actual switching. This patch adds a new inode flag I_WB_SWITCHING, which has two functions. First, the easy one, it ensures that there's only one switching in progress for a give inode. Second, it's used as a mechanism to synchronize wb stat updates. The two stats, WB_RECLAIMABLE and WB_WRITEBACK, aren't event counters but track the current number of dirty pages and pages under writeback respectively. As such, when an inode is moved from one wb to another, the inode's portion of those stats have to be transferred together; unfortunately, this is a bit tricky as those stat updates are percpu operations which are performed without holding any lock in some places. This patch solves the problem in a similar way as memcg. Each such lockless stat updates are wrapped in transaction surrounded by unlocked_inode_to_wb_begin/end(). During normal operation, they map to rcu_read_lock/unlock(); however, if I_WB_SWITCHING is asserted, mapping->tree_lock is grabbed across the transaction. In turn, the switching path sets I_WB_SWITCHING and waits for a RCU grace period to pass before actually starting to switch, which guarantees that all stat update paths are synchronizing against mapping->tree_lock. This patch still doesn't implement the actual switching. v3: Updated on top of the recent cancel_dirty_page() updates. unlocked_inode_to_wb_begin() now nests inside mem_cgroup_begin_page_stat() to match the locking order. v2: The i_wb access transaction will be used for !stat accesses too. Function names and comments updated accordingly. s/inode_wb_stat_unlocked_{begin|end}/unlocked_inode_to_wb_{begin|end}/ s/switch_wb/switch_wbs/ Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-29 02:50:53 +08:00
dec_wb_stat(wb, WB_RECLAIMABLE);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
ret = 1;
}
writeback: safer lock nesting lock_page_memcg()/unlock_page_memcg() use spin_lock_irqsave/restore() if the page's memcg is undergoing move accounting, which occurs when a process leaves its memcg for a new one that has memory.move_charge_at_immigrate set. unlocked_inode_to_wb_begin,end() use spin_lock_irq/spin_unlock_irq() if the given inode is switching writeback domains. Switches occur when enough writes are issued from a new domain. This existing pattern is thus suspicious: lock_page_memcg(page); unlocked_inode_to_wb_begin(inode, &locked); ... unlocked_inode_to_wb_end(inode, locked); unlock_page_memcg(page); If both inode switch and process memcg migration are both in-flight then unlocked_inode_to_wb_end() will unconditionally enable interrupts while still holding the lock_page_memcg() irq spinlock. This suggests the possibility of deadlock if an interrupt occurs before unlock_page_memcg(). truncate __cancel_dirty_page lock_page_memcg unlocked_inode_to_wb_begin unlocked_inode_to_wb_end <interrupts mistakenly enabled> <interrupt> end_page_writeback test_clear_page_writeback lock_page_memcg <deadlock> unlock_page_memcg Due to configuration limitations this deadlock is not currently possible because we don't mix cgroup writeback (a cgroupv2 feature) and memory.move_charge_at_immigrate (a cgroupv1 feature). If the kernel is hacked to always claim inode switching and memcg moving_account, then this script triggers lockup in less than a minute: cd /mnt/cgroup/memory mkdir a b echo 1 > a/memory.move_charge_at_immigrate echo 1 > b/memory.move_charge_at_immigrate ( echo $BASHPID > a/cgroup.procs while true; do dd if=/dev/zero of=/mnt/big bs=1M count=256 done ) & while true; do sync done & sleep 1h & SLEEP=$! while true; do echo $SLEEP > a/cgroup.procs echo $SLEEP > b/cgroup.procs done The deadlock does not seem possible, so it's debatable if there's any reason to modify the kernel. I suggest we should to prevent future surprises. And Wang Long said "this deadlock occurs three times in our environment", so there's more reason to apply this, even to stable. Stable 4.4 has minor conflicts applying this patch. For a clean 4.4 patch see "[PATCH for-4.4] writeback: safer lock nesting" https://lkml.org/lkml/2018/4/11/146 Wang Long said "this deadlock occurs three times in our environment" [gthelen@google.com: v4] Link: http://lkml.kernel.org/r/20180411084653.254724-1-gthelen@google.com [akpm@linux-foundation.org: comment tweaks, struct initialization simplification] Change-Id: Ibb773e8045852978f6207074491d262f1b3fb613 Link: http://lkml.kernel.org/r/20180410005908.167976-1-gthelen@google.com Fixes: 682aa8e1a6a1 ("writeback: implement unlocked_inode_to_wb transaction and use it for stat updates") Signed-off-by: Greg Thelen <gthelen@google.com> Reported-by: Wang Long <wanglong19@meituan.com> Acked-by: Wang Long <wanglong19@meituan.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-21 05:55:42 +08:00
unlocked_inode_to_wb_end(inode, &cookie);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-23 05:13:16 +08:00
return ret;
}
VM: Fix nasty and subtle race in shared mmap'ed page writeback The VM layer (on the face of it, fairly reasonably) expected that when it does a ->writepage() call to the filesystem, it would write out the full page at that point in time. Especially since it had earlier marked the whole page dirty with "set_page_dirty()". But that isn't actually the case: ->writepage() does not actually write a page, it writes the parts of the page that have been explicitly marked dirty before, *and* that had not got written out for other reasons since the last time we told it they were dirty. That last caveat is the important one. Which _most_ of the time ends up being the whole page (since we had called "set_page_dirty()" on the page earlier), but if the filesystem had done any dirty flushing of its own (for example, to honor some internal write ordering guarantees), it might end up doing only a partial page IO (or none at all) when ->writepage() is actually called. That is the correct thing in general (since we actually often _want_ only the known-dirty parts of the page to be written out), but the shared dirty page handling had implicitly forgotten about these details, and had a number of cases where it was doing just the "->writepage()" part, without telling the low-level filesystem that the whole page might have been re-dirtied as part of being mapped writably into user space. Since most of the time the FS did actually write out the full page, we didn't notice this for a loong time, and this needed some really odd patterns to trigger. But it caused occasional corruption with rtorrent and with the Debian "apt" database, because both use shared mmaps to update the end result. This fixes it. Finally. After way too much hair-pulling. Acked-by: Nick Piggin <nickpiggin@yahoo.com.au> Acked-by: Martin J. Bligh <mbligh@google.com> Acked-by: Martin Michlmayr <tbm@cyrius.com> Acked-by: Martin Johansson <martin@fatbob.nu> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Andrei Popa <andrei.popa@i-neo.ro> Cc: High Dickins <hugh@veritas.com> Cc: Andrew Morton <akpm@osdl.org>, Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Segher Boessenkool <segher@kernel.crashing.org> Cc: David Miller <davem@davemloft.net> Cc: Arjan van de Ven <arjan@infradead.org> Cc: Gordon Farquharson <gordonfarquharson@gmail.com> Cc: Guillaume Chazarain <guichaz@yahoo.fr> Cc: Theodore Tso <tytso@mit.edu> Cc: Kenneth Cheng <kenneth.w.chen@intel.com> Cc: Tobias Diedrich <ranma@tdiedrich.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-30 02:00:58 +08:00
return TestClearPageDirty(page);
}
EXPORT_SYMBOL(clear_page_dirty_for_io);
int test_clear_page_writeback(struct page *page)
{
struct address_space *mapping = page_mapping(page);
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-19 06:15:48 +08:00
struct mem_cgroup *memcg;
struct lruvec *lruvec;
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-30 05:50:48 +08:00
int ret;
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-19 06:15:48 +08:00
memcg = lock_page_memcg(page);
lruvec = mem_cgroup_page_lruvec(page, page_pgdat(page));
mm: don't use radix tree writeback tags for pages in swap cache File pages use a set of radix tree tags (DIRTY, TOWRITE, WRITEBACK, etc.) to accelerate finding the pages with a specific tag in the radix tree during inode writeback. But for anonymous pages in the swap cache, there is no inode writeback. So there is no need to find the pages with some writeback tags in the radix tree. It is not necessary to touch radix tree writeback tags for pages in the swap cache. Per Rik van Riel's suggestion, a new flag AS_NO_WRITEBACK_TAGS is introduced for address spaces which don't need to update the writeback tags. The flag is set for swap caches. It may be used for DAX file systems, etc. With this patch, the swap out bandwidth improved 22.3% (from ~1.2GB/s to ~1.48GBps) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. The improvement comes from the reduced contention on the swap cache radix tree lock. To test sequential swapping out, the test case uses 8 processes, which sequentially allocate and write to the anonymous pages until RAM and part of the swap device is used up. Details of comparison is as follow, base base+patch ---------------- -------------------------- %stddev %change %stddev \ | \ 2506952 ± 2% +28.1% 3212076 ± 7% vm-scalability.throughput 1207402 ± 7% +22.3% 1476578 ± 6% vmstat.swap.so 10.86 ± 12% -23.4% 8.31 ± 16% perf-profile.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list 10.82 ± 13% -33.1% 7.24 ± 14% perf-profile.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_zone_memcg 10.36 ± 11% -100.0% 0.00 ± -1% perf-profile.cycles-pp._raw_spin_lock_irqsave.__test_set_page_writeback.bdev_write_page.__swap_writepage.swap_writepage 10.52 ± 12% -100.0% 0.00 ± -1% perf-profile.cycles-pp._raw_spin_lock_irqsave.test_clear_page_writeback.end_page_writeback.page_endio.pmem_rw_page Link: http://lkml.kernel.org/r/1472578089-5560-1-git-send-email-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Tejun Heo <tj@kernel.org> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-10-08 07:59:30 +08:00
if (mapping && mapping_use_writeback_tags(mapping)) {
struct inode *inode = mapping->host;
struct backing_dev_info *bdi = inode_to_bdi(inode);
unsigned long flags;
xa_lock_irqsave(&mapping->i_pages, flags);
ret = TestClearPageWriteback(page);
if (ret) {
__xa_clear_mark(&mapping->i_pages, page_index(page),
PAGECACHE_TAG_WRITEBACK);
if (bdi_cap_account_writeback(bdi)) {
struct bdi_writeback *wb = inode_to_wb(inode);
dec_wb_stat(wb, WB_WRITEBACK);
__wb_writeout_inc(wb);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 14:25:50 +08:00
}
}
fs/fs-writeback.c: add a new writeback list for sync wait_sb_inodes() currently does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on. To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the sb when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for. Define a couple helpers to add/remove an inode from the writeback list and call them when the overall mapping is tagged for or cleared from writeback. Update wait_sb_inodes() to walk only the inodes under writeback due to the sync. With this change, filesystem sync times are significantly reduced for fs' with largely populated inode caches and otherwise no other work to do. For example, on a 16xcpu 2GHz x86-64 server, 10TB XFS filesystem with a ~10m entry inode cache, sync times are reduced from ~7.3s to less than 0.1s when the filesystem is fully clean. Link: http://lkml.kernel.org/r/1466594593-6757-2-git-send-email-bfoster@redhat.com Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Tested-by: Holger Hoffstätte <holger.hoffstaette@applied-asynchrony.com> Cc: Al Viro <viro@ZenIV.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-27 06:21:50 +08:00
if (mapping->host && !mapping_tagged(mapping,
PAGECACHE_TAG_WRITEBACK))
sb_clear_inode_writeback(mapping->host);
xa_unlock_irqrestore(&mapping->i_pages, flags);
} else {
ret = TestClearPageWriteback(page);
}
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-19 06:15:48 +08:00
/*
* NOTE: Page might be free now! Writeback doesn't hold a page
* reference on its own, it relies on truncation to wait for
* the clearing of PG_writeback. The below can only access
* page state that is static across allocation cycles.
*/
if (ret) {
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-19 06:15:48 +08:00
dec_lruvec_state(lruvec, NR_WRITEBACK);
dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
inc_node_page_state(page, NR_WRITTEN);
}
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-19 06:15:48 +08:00
__unlock_page_memcg(memcg);
return ret;
}
ext4: fix data integrity sync in ordered mode When we perform a data integrity sync we tag all the dirty pages with PAGECACHE_TAG_TOWRITE at start of ext4_da_writepages. Later we check for this tag in write_cache_pages_da and creates a struct mpage_da_data containing contiguously indexed pages tagged with this tag and sync these pages with a call to mpage_da_map_and_submit. This process is done in while loop until all the PAGECACHE_TAG_TOWRITE pages are synced. We also do journal start and stop in each iteration. journal_stop could initiate journal commit which would call ext4_writepage which in turn will call ext4_bio_write_page even for delayed OR unwritten buffers. When ext4_bio_write_page is called for such buffers, even though it does not sync them but it clears the PAGECACHE_TAG_TOWRITE of the corresponding page and hence these pages are also not synced by the currently running data integrity sync. We will end up with dirty pages although sync is completed. This could cause a potential data loss when the sync call is followed by a truncate_pagecache call, which is exactly the case in collapse_range. (It will cause generic/127 failure in xfstests) To avoid this issue, we can use set_page_writeback_keepwrite instead of set_page_writeback, which doesn't clear TOWRITE tag. Cc: stable@vger.kernel.org Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com> Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
2014-05-12 20:12:25 +08:00
int __test_set_page_writeback(struct page *page, bool keep_write)
{
struct address_space *mapping = page_mapping(page);
mm/gup/writeback: add callbacks for inaccessible pages With the introduction of protected KVM guests on s390 there is now a concept of inaccessible pages. These pages need to be made accessible before the host can access them. While cpu accesses will trigger a fault that can be resolved, I/O accesses will just fail. We need to add a callback into architecture code for places that will do I/O, namely when writeback is started or when a page reference is taken. This is not only to enable paging, file backing etc, it is also necessary to protect the host against a malicious user space. For example a bad QEMU could simply start direct I/O on such protected memory. We do not want userspace to be able to trigger I/O errors and thus the logic is "whenever somebody accesses that page (gup) or does I/O, make sure that this page can be accessed". When the guest tries to access that page we will wait in the page fault handler for writeback to have finished and for the page_ref to be the expected value. On s390x the function is not supposed to fail, so it is ok to use a WARN_ON on failure. If we ever need some more finegrained handling we can tackle this when we know the details. Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: David Hildenbrand <david@redhat.com> Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: John Hubbard <jhubbard@nvidia.com> Acked-by: Will Deacon <will@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Matthew Wilcox <willy@infradead.org> Cc: Ira Weiny <ira.weiny@intel.com> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Michal Hocko <mhocko@suse.com> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Vlastimil Babka <vbabka@suse.cz> Link: http://lkml.kernel.org/r/20200306132537.783769-3-imbrenda@linux.ibm.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 12:05:56 +08:00
int ret, access_ret;
lock_page_memcg(page);
mm: don't use radix tree writeback tags for pages in swap cache File pages use a set of radix tree tags (DIRTY, TOWRITE, WRITEBACK, etc.) to accelerate finding the pages with a specific tag in the radix tree during inode writeback. But for anonymous pages in the swap cache, there is no inode writeback. So there is no need to find the pages with some writeback tags in the radix tree. It is not necessary to touch radix tree writeback tags for pages in the swap cache. Per Rik van Riel's suggestion, a new flag AS_NO_WRITEBACK_TAGS is introduced for address spaces which don't need to update the writeback tags. The flag is set for swap caches. It may be used for DAX file systems, etc. With this patch, the swap out bandwidth improved 22.3% (from ~1.2GB/s to ~1.48GBps) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. The improvement comes from the reduced contention on the swap cache radix tree lock. To test sequential swapping out, the test case uses 8 processes, which sequentially allocate and write to the anonymous pages until RAM and part of the swap device is used up. Details of comparison is as follow, base base+patch ---------------- -------------------------- %stddev %change %stddev \ | \ 2506952 ± 2% +28.1% 3212076 ± 7% vm-scalability.throughput 1207402 ± 7% +22.3% 1476578 ± 6% vmstat.swap.so 10.86 ± 12% -23.4% 8.31 ± 16% perf-profile.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list 10.82 ± 13% -33.1% 7.24 ± 14% perf-profile.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_zone_memcg 10.36 ± 11% -100.0% 0.00 ± -1% perf-profile.cycles-pp._raw_spin_lock_irqsave.__test_set_page_writeback.bdev_write_page.__swap_writepage.swap_writepage 10.52 ± 12% -100.0% 0.00 ± -1% perf-profile.cycles-pp._raw_spin_lock_irqsave.test_clear_page_writeback.end_page_writeback.page_endio.pmem_rw_page Link: http://lkml.kernel.org/r/1472578089-5560-1-git-send-email-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Tejun Heo <tj@kernel.org> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-10-08 07:59:30 +08:00
if (mapping && mapping_use_writeback_tags(mapping)) {
XA_STATE(xas, &mapping->i_pages, page_index(page));
struct inode *inode = mapping->host;
struct backing_dev_info *bdi = inode_to_bdi(inode);
unsigned long flags;
xas_lock_irqsave(&xas, flags);
xas_load(&xas);
ret = TestSetPageWriteback(page);
if (!ret) {
fs/fs-writeback.c: add a new writeback list for sync wait_sb_inodes() currently does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on. To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the sb when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for. Define a couple helpers to add/remove an inode from the writeback list and call them when the overall mapping is tagged for or cleared from writeback. Update wait_sb_inodes() to walk only the inodes under writeback due to the sync. With this change, filesystem sync times are significantly reduced for fs' with largely populated inode caches and otherwise no other work to do. For example, on a 16xcpu 2GHz x86-64 server, 10TB XFS filesystem with a ~10m entry inode cache, sync times are reduced from ~7.3s to less than 0.1s when the filesystem is fully clean. Link: http://lkml.kernel.org/r/1466594593-6757-2-git-send-email-bfoster@redhat.com Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Tested-by: Holger Hoffstätte <holger.hoffstaette@applied-asynchrony.com> Cc: Al Viro <viro@ZenIV.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-27 06:21:50 +08:00
bool on_wblist;
on_wblist = mapping_tagged(mapping,
PAGECACHE_TAG_WRITEBACK);
xas_set_mark(&xas, PAGECACHE_TAG_WRITEBACK);
if (bdi_cap_account_writeback(bdi))
inc_wb_stat(inode_to_wb(inode), WB_WRITEBACK);
fs/fs-writeback.c: add a new writeback list for sync wait_sb_inodes() currently does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on. To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the sb when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for. Define a couple helpers to add/remove an inode from the writeback list and call them when the overall mapping is tagged for or cleared from writeback. Update wait_sb_inodes() to walk only the inodes under writeback due to the sync. With this change, filesystem sync times are significantly reduced for fs' with largely populated inode caches and otherwise no other work to do. For example, on a 16xcpu 2GHz x86-64 server, 10TB XFS filesystem with a ~10m entry inode cache, sync times are reduced from ~7.3s to less than 0.1s when the filesystem is fully clean. Link: http://lkml.kernel.org/r/1466594593-6757-2-git-send-email-bfoster@redhat.com Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Tested-by: Holger Hoffstätte <holger.hoffstaette@applied-asynchrony.com> Cc: Al Viro <viro@ZenIV.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-27 06:21:50 +08:00
/*
* We can come through here when swapping anonymous
* pages, so we don't necessarily have an inode to track
* for sync.
*/
if (mapping->host && !on_wblist)
sb_mark_inode_writeback(mapping->host);
}
if (!PageDirty(page))
xas_clear_mark(&xas, PAGECACHE_TAG_DIRTY);
ext4: fix data integrity sync in ordered mode When we perform a data integrity sync we tag all the dirty pages with PAGECACHE_TAG_TOWRITE at start of ext4_da_writepages. Later we check for this tag in write_cache_pages_da and creates a struct mpage_da_data containing contiguously indexed pages tagged with this tag and sync these pages with a call to mpage_da_map_and_submit. This process is done in while loop until all the PAGECACHE_TAG_TOWRITE pages are synced. We also do journal start and stop in each iteration. journal_stop could initiate journal commit which would call ext4_writepage which in turn will call ext4_bio_write_page even for delayed OR unwritten buffers. When ext4_bio_write_page is called for such buffers, even though it does not sync them but it clears the PAGECACHE_TAG_TOWRITE of the corresponding page and hence these pages are also not synced by the currently running data integrity sync. We will end up with dirty pages although sync is completed. This could cause a potential data loss when the sync call is followed by a truncate_pagecache call, which is exactly the case in collapse_range. (It will cause generic/127 failure in xfstests) To avoid this issue, we can use set_page_writeback_keepwrite instead of set_page_writeback, which doesn't clear TOWRITE tag. Cc: stable@vger.kernel.org Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com> Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
2014-05-12 20:12:25 +08:00
if (!keep_write)
xas_clear_mark(&xas, PAGECACHE_TAG_TOWRITE);
xas_unlock_irqrestore(&xas, flags);
} else {
ret = TestSetPageWriteback(page);
}
if (!ret) {
inc_lruvec_page_state(page, NR_WRITEBACK);
inc_zone_page_state(page, NR_ZONE_WRITE_PENDING);
}
unlock_page_memcg(page);
mm/gup/writeback: add callbacks for inaccessible pages With the introduction of protected KVM guests on s390 there is now a concept of inaccessible pages. These pages need to be made accessible before the host can access them. While cpu accesses will trigger a fault that can be resolved, I/O accesses will just fail. We need to add a callback into architecture code for places that will do I/O, namely when writeback is started or when a page reference is taken. This is not only to enable paging, file backing etc, it is also necessary to protect the host against a malicious user space. For example a bad QEMU could simply start direct I/O on such protected memory. We do not want userspace to be able to trigger I/O errors and thus the logic is "whenever somebody accesses that page (gup) or does I/O, make sure that this page can be accessed". When the guest tries to access that page we will wait in the page fault handler for writeback to have finished and for the page_ref to be the expected value. On s390x the function is not supposed to fail, so it is ok to use a WARN_ON on failure. If we ever need some more finegrained handling we can tackle this when we know the details. Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: David Hildenbrand <david@redhat.com> Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: John Hubbard <jhubbard@nvidia.com> Acked-by: Will Deacon <will@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Matthew Wilcox <willy@infradead.org> Cc: Ira Weiny <ira.weiny@intel.com> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Michal Hocko <mhocko@suse.com> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Vlastimil Babka <vbabka@suse.cz> Link: http://lkml.kernel.org/r/20200306132537.783769-3-imbrenda@linux.ibm.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 12:05:56 +08:00
access_ret = arch_make_page_accessible(page);
/*
* If writeback has been triggered on a page that cannot be made
* accessible, it is too late to recover here.
*/
VM_BUG_ON_PAGE(access_ret != 0, page);
return ret;
}
ext4: fix data integrity sync in ordered mode When we perform a data integrity sync we tag all the dirty pages with PAGECACHE_TAG_TOWRITE at start of ext4_da_writepages. Later we check for this tag in write_cache_pages_da and creates a struct mpage_da_data containing contiguously indexed pages tagged with this tag and sync these pages with a call to mpage_da_map_and_submit. This process is done in while loop until all the PAGECACHE_TAG_TOWRITE pages are synced. We also do journal start and stop in each iteration. journal_stop could initiate journal commit which would call ext4_writepage which in turn will call ext4_bio_write_page even for delayed OR unwritten buffers. When ext4_bio_write_page is called for such buffers, even though it does not sync them but it clears the PAGECACHE_TAG_TOWRITE of the corresponding page and hence these pages are also not synced by the currently running data integrity sync. We will end up with dirty pages although sync is completed. This could cause a potential data loss when the sync call is followed by a truncate_pagecache call, which is exactly the case in collapse_range. (It will cause generic/127 failure in xfstests) To avoid this issue, we can use set_page_writeback_keepwrite instead of set_page_writeback, which doesn't clear TOWRITE tag. Cc: stable@vger.kernel.org Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com> Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
2014-05-12 20:12:25 +08:00
EXPORT_SYMBOL(__test_set_page_writeback);
/*
* Wait for a page to complete writeback
*/
void wait_on_page_writeback(struct page *page)
{
if (PageWriteback(page)) {
trace_wait_on_page_writeback(page, page_mapping(page));
wait_on_page_bit(page, PG_writeback);
}
}
EXPORT_SYMBOL_GPL(wait_on_page_writeback);
mm: only enforce stable page writes if the backing device requires it Create a helper function to check if a backing device requires stable page writes and, if so, performs the necessary wait. Then, make it so that all points in the memory manager that handle making pages writable use the helper function. This should provide stable page write support to most filesystems, while eliminating unnecessary waiting for devices that don't require the feature. Before this patchset, all filesystems would block, regardless of whether or not it was necessary. ext3 would wait, but still generate occasional checksum errors. The network filesystems were left to do their own thing, so they'd wait too. After this patchset, all the disk filesystems except ext3 and btrfs will wait only if the hardware requires it. ext3 (if necessary) snapshots pages instead of blocking, and btrfs provides its own bdi so the mm will never wait. Network filesystems haven't been touched, so either they provide their own stable page guarantees or they don't block at all. The blocking behavior is back to what it was before 3.0 if you don't have a disk requiring stable page writes. Here's the result of using dbench to test latency on ext2: 3.8.0-rc3: Operation Count AvgLat MaxLat ---------------------------------------- WriteX 109347 0.028 59.817 ReadX 347180 0.004 3.391 Flush 15514 29.828 287.283 Throughput 57.429 MB/sec 4 clients 4 procs max_latency=287.290 ms 3.8.0-rc3 + patches: WriteX 105556 0.029 4.273 ReadX 335004 0.005 4.112 Flush 14982 30.540 298.634 Throughput 55.4496 MB/sec 4 clients 4 procs max_latency=298.650 ms As you can see, the maximum write latency drops considerably with this patch enabled. The other filesystems (ext3/ext4/xfs/btrfs) behave similarly, but see the cover letter for those results. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Acked-by: Steven Whitehouse <swhiteho@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Artem Bityutskiy <dedekind1@gmail.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@sandia.gov> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 08:42:51 +08:00
/**
* wait_for_stable_page() - wait for writeback to finish, if necessary.
* @page: The page to wait on.
*
* This function determines if the given page is related to a backing device
* that requires page contents to be held stable during writeback. If so, then
* it will wait for any pending writeback to complete.
*/
void wait_for_stable_page(struct page *page)
{
if (bdi_cap_stable_pages_required(inode_to_bdi(page->mapping->host)))
wait_on_page_writeback(page);
mm: only enforce stable page writes if the backing device requires it Create a helper function to check if a backing device requires stable page writes and, if so, performs the necessary wait. Then, make it so that all points in the memory manager that handle making pages writable use the helper function. This should provide stable page write support to most filesystems, while eliminating unnecessary waiting for devices that don't require the feature. Before this patchset, all filesystems would block, regardless of whether or not it was necessary. ext3 would wait, but still generate occasional checksum errors. The network filesystems were left to do their own thing, so they'd wait too. After this patchset, all the disk filesystems except ext3 and btrfs will wait only if the hardware requires it. ext3 (if necessary) snapshots pages instead of blocking, and btrfs provides its own bdi so the mm will never wait. Network filesystems haven't been touched, so either they provide their own stable page guarantees or they don't block at all. The blocking behavior is back to what it was before 3.0 if you don't have a disk requiring stable page writes. Here's the result of using dbench to test latency on ext2: 3.8.0-rc3: Operation Count AvgLat MaxLat ---------------------------------------- WriteX 109347 0.028 59.817 ReadX 347180 0.004 3.391 Flush 15514 29.828 287.283 Throughput 57.429 MB/sec 4 clients 4 procs max_latency=287.290 ms 3.8.0-rc3 + patches: WriteX 105556 0.029 4.273 ReadX 335004 0.005 4.112 Flush 14982 30.540 298.634 Throughput 55.4496 MB/sec 4 clients 4 procs max_latency=298.650 ms As you can see, the maximum write latency drops considerably with this patch enabled. The other filesystems (ext3/ext4/xfs/btrfs) behave similarly, but see the cover letter for those results. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Acked-by: Steven Whitehouse <swhiteho@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Artem Bityutskiy <dedekind1@gmail.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@sandia.gov> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 08:42:51 +08:00
}
EXPORT_SYMBOL_GPL(wait_for_stable_page);