UbiquitousOS
/
OpenCloudOS-Kernel
mirror of https://gitee.com/OpenCloudOS/OpenCloudOS-Kernel.git
11
0
Fork 0
Code Issues Projects Releases Wiki Activity
OpenCloudOS-Kernel/mm/workingset.c

1046 lines
35 KiB
C
Raw Blame History

// SPDX-License-Identifier: GPL-2.0
/*
* Workingset detection
*
* Copyright (C) 2013 Red Hat, Inc., Johannes Weiner
*/
#include <linux/memcontrol.h>
#include <linux/mm_inline.h>
#include <linux/writeback.h>
#include <linux/shmem_fs.h>
#include <linux/pagemap.h>
#include <linux/atomic.h>
#include <linux/module.h>
#include <linux/swap.h>
#include <linux/dax.h>
#include <linux/fs.h>
#include <linux/mm.h>
/*
* Double CLOCK lists
*
* Per node, two clock lists are maintained for file pages: the
* inactive and the active list. Freshly faulted pages start out at
* the head of the inactive list and page reclaim scans pages from the
* tail. Pages that are accessed multiple times on the inactive list
* are promoted to the active list, to protect them from reclaim,
* whereas active pages are demoted to the inactive list when the
* active list grows too big.
*
* fault ------------------------+
* |
* +--------------+ | +-------------+
* reclaim <- | inactive | <-+-- demotion | active | <--+
* +--------------+ +-------------+ |
* | |
* +-------------- promotion ------------------+
*
*
* Access frequency and refault distance
*
* A workload is thrashing when its pages are frequently used but they
* are evicted from the inactive list every time before another access
* would have promoted them to the active list.
*
* In cases where the average access distance between thrashing pages
* is bigger than the size of memory there is nothing that can be
* done - the thrashing set could never fit into memory under any
* circumstance.
*
* However, the average access distance could be bigger than the
* inactive list, yet smaller than the size of memory. In this case,
* the set could fit into memory if it weren't for the currently
* active pages - which may be used more, hopefully less frequently:
*
* +-memory available to cache-+
* | |
* +-inactive------+-active----+
* a b | c d e f g h i | J K L M N |
* +---------------+-----------+
*
* It is prohibitively expensive to accurately track access frequency
* of pages. But a reasonable approximation can be made to measure
* thrashing on the inactive list, after which refaulting pages can be
* activated optimistically to compete with the existing active pages.
*
* Approximating inactive page access frequency - Observations:
*
* 1. When a page is accessed for the first time, it is added to the
* head of the inactive list, slides every existing inactive page
* towards the tail by one slot, and pushes the current tail page
* out of memory.
*
* 2. When a page is accessed for the second time, it is promoted to
* the active list, shrinking the inactive list by one slot. This
* also slides all inactive pages that were faulted into the cache
* more recently than the activated page towards the tail of the
* inactive list.
*
* Thus:
*
* 1. The sum of evictions and activations between any two points in
* time indicate the minimum number of inactive pages accessed in
* between.
*
* 2. Moving one inactive page N page slots towards the tail of the
* list requires at least N inactive page accesses.
*
* Combining these:
*
* 1. When a page is finally evicted from memory, the number of
* inactive pages accessed while the page was in cache is at least
* the number of page slots on the inactive list.
*
* 2. In addition, measuring the sum of evictions and activations (E)
* at the time of a page's eviction, and comparing it to another
* reading (R) at the time the page faults back into memory tells
* the minimum number of accesses while the page was not cached.
* This is called the refault distance.
*
* Because the first access of the page was the fault and the second
* access the refault, we combine the in-cache distance with the
* out-of-cache distance to get the complete minimum access distance
* of this page:
*
* NR_inactive + (R - E)
*
* And knowing the minimum access distance of a page, we can easily
* tell if the page would be able to stay in cache assuming all page
* slots in the cache were available:
*
* NR_inactive + (R - E) <= NR_inactive + NR_active
*
* which can be further simplified to
*
* (R - E) <= NR_active
*
* Put into words, the refault distance (out-of-cache) can be seen as
* a deficit in inactive list space (in-cache). If the inactive list
* had (R - E) more page slots, the page would not have been evicted
* in between accesses, but activated instead. And on a full system,
* the only thing eating into inactive list space is active pages.
*
*
* Refaulting inactive pages
*
* All that is known about the active list is that the pages have been
* accessed more than once in the past. This means that at any given
* time there is actually a good chance that pages on the active list
* are no longer in active use.
*
* So when a refault distance of (R - E) is observed and there are at
* least (R - E) active pages, the refaulting page is activated
* optimistically in the hope that (R - E) active pages are actually
* used less frequently than the refaulting page - or even not used at
* all anymore.
*
* That means if inactive cache is refaulting with a suitable refault
* distance, we assume the cache workingset is transitioning and put
* pressure on the current active list.
*
* If this is wrong and demotion kicks in, the pages which are truly
* used more frequently will be reactivated while the less frequently
* used once will be evicted from memory.
*
* But if this is right, the stale pages will be pushed out of memory
* and the used pages get to stay in cache.
*
* Refaulting active pages
*
* If on the other hand the refaulting pages have recently been
* deactivated, it means that the active list is no longer protecting
* actively used cache from reclaim. The cache is NOT transitioning to
* a different workingset; the existing workingset is thrashing in the
* space allocated to the page cache.
*
*
* Implementation
*
* For each node's LRU lists, a counter for inactive evictions and
* activations is maintained (node->nonresident_age).
*
* On eviction, a snapshot of this counter (along with some bits to
* identify the node) is stored in the now empty page cache
* slot of the evicted page. This is called a shadow entry.
*
* On cache misses for which there are shadow entries, an eligible
* refault distance will immediately activate the refaulting page.
*/
#define WORKINGSET_SHIFT 1
#define EVICTION_SHIFT ((BITS_PER_LONG - BITS_PER_XA_VALUE) + \
WORKINGSET_SHIFT + NODES_SHIFT + \
MEM_CGROUP_ID_SHIFT)
#define EVICTION_MASK (~0UL >> EVICTION_SHIFT)
/*
* Eviction timestamps need to be able to cover the full range of
* actionable refaults. However, bits are tight in the xarray
* entry, and after storing the identifier for the lruvec there might
* not be enough left to represent every single actionable refault. In
* that case, we have to sacrifice granularity for distance, and group
* evictions into coarser buckets by shaving off lower timestamp bits.
*/
static unsigned int bucket_order __read_mostly;
/*
* MGLRU may reuse some bits of eviction for better accuracy, but
* LRU_GEN_WIDTH and LRU_REFS_WIDTH already took some place.
*/
#define LRU_GEN_EVICTION_SHIFT (EVICTION_SHIFT + LRU_GEN_WIDTH + LRU_REFS_WIDTH)
#define LRU_GEN_EVICTION_MASK (EVICTION_MASK >> (LRU_GEN_WIDTH + LRU_REFS_WIDTH))
#define LRU_GEN_EVICTION_GEN_MASK (LRU_GEN_EVICTION_MASK & ~(LRU_GEN_EVICTION_MASK << (LRU_GEN_WIDTH + LRU_REFS_WIDTH)))
static unsigned int lru_gen_bucket_order __read_mostly;
static void *pack_shadow(int memcgid, pg_data_t *pgdat, unsigned long eviction,
bool workingset)
{
eviction &= EVICTION_MASK;
eviction = (eviction << MEM_CGROUP_ID_SHIFT) | memcgid;
eviction = (eviction << NODES_SHIFT) | pgdat->node_id;
eviction = (eviction << WORKINGSET_SHIFT) | workingset;
return xa_mk_value(eviction);
}
static void unpack_shadow(void *shadow, int *memcgidp, pg_data_t **pgdat,
unsigned long *evictionp, bool *workingsetp)
{
unsigned long entry = xa_to_value(shadow);
int memcgid, nid;
bool workingset;
workingset = entry & ((1UL << WORKINGSET_SHIFT) - 1);
entry >>= WORKINGSET_SHIFT;
nid = entry & ((1UL << NODES_SHIFT) - 1);
entry >>= NODES_SHIFT;
memcgid = entry & ((1UL << MEM_CGROUP_ID_SHIFT) - 1);
entry >>= MEM_CGROUP_ID_SHIFT;
*memcgidp = memcgid;
*pgdat = NODE_DATA(nid);
*evictionp = entry;
*workingsetp = workingset;
}
#ifdef CONFIG_WORKINGSET_EVICT_EVAL
void workingset_evict_file(struct lruvec *lruvec, unsigned long nr_pages)
{
do {
atomic_long_add(nr_pages, &lruvec->evicted_file);
} while ((lruvec = parent_lruvec(lruvec)));
}
#endif
#ifdef CONFIG_LRU_GEN
static void *lru_gen_eviction(struct page *page)
{
int hist;
unsigned long token;
unsigned long min_seq;
struct lruvec *lruvec;
struct lru_gen_struct *lrugen;
int type = page_is_file_lru(page);
int delta = hpage_nr_pages(page);
int refs = page_lru_refs(page);
int tier = lru_tier_from_refs(refs);
struct mem_cgroup *memcg = page_memcg(page);
struct pglist_data *pgdat = page_pgdat(page);
/* XXX: kasong: during testing period, keep this BUG_ON to catch
* potential tailing page bug due to folio modification, or maybe
* just move this to hpage helpers */
VM_BUG_ON_PAGE(PageTail(page), page);
BUILD_BUG_ON(LRU_GEN_WIDTH + LRU_REFS_WIDTH > BITS_PER_LONG - EVICTION_SHIFT);
lruvec = mem_cgroup_lruvec(memcg, pgdat);
lrugen = &lruvec->lrugen;
min_seq = READ_ONCE(lrugen->min_seq[type]);
token = atomic_long_read(&lruvec->nonresident_age);
token >>= lru_gen_bucket_order;
token &= LRU_GEN_EVICTION_MASK;
token <<= (LRU_GEN_WIDTH + LRU_REFS_WIDTH);
token |= (((min_seq << LRU_REFS_WIDTH) | max(refs - 1, 0)) & LRU_GEN_EVICTION_GEN_MASK);
hist = lru_hist_from_seq(min_seq);
atomic_long_add(delta, &lrugen->evicted[hist][type][tier]);
/* keep the workingset ticking */
workingset_age_nonresident(lruvec, delta);
#ifdef CONFIG_WORKINGSET_EVICT_EVAL
workingset_evict_file(lruvec, delta);
#endif
return pack_shadow(mem_cgroup_id(memcg), pgdat, token, refs);
}
static void lru_gen_refault(struct page *page, void *shadow, unsigned long *eviction)
{
int hist, tier, refs;
int memcg_id;
bool workingset;
unsigned long token;
unsigned long min_seq;
struct lruvec *lruvec;
struct lru_gen_struct *lrugen;
struct mem_cgroup *memcg;
struct pglist_data *pgdat;
int type = page_is_file_lru(page);
int delta = hpage_nr_pages(page);
/* XXX: kasong: during testing period, keep this BUG_ON to catch
* potential tailing page bug due to folio modification, or maybe
* just move this to hpage helpers */
VM_BUG_ON_PAGE(PageTail(page), page);
unpack_shadow(shadow, &memcg_id, &pgdat, &token, &workingset);
*eviction = token >> (LRU_GEN_WIDTH + LRU_REFS_WIDTH);
*eviction <<= lru_gen_bucket_order;
token &= LRU_GEN_EVICTION_GEN_MASK;
if (pgdat != page_pgdat(page))
return;
rcu_read_lock();
memcg = page_memcg_rcu(page);
if (memcg_id != mem_cgroup_id(memcg))
goto unlock;
lruvec = mem_cgroup_lruvec(memcg, pgdat);
lrugen = &lruvec->lrugen;
min_seq = READ_ONCE(lrugen->min_seq[type]);
if ((token >> LRU_REFS_WIDTH) != (min_seq & (BIT(LRU_GEN_WIDTH) - 1)))
goto unlock;
hist = lru_hist_from_seq(min_seq);
/* see the comment in page_lru_refs() */
refs = (token & (BIT(LRU_REFS_WIDTH) - 1)) + workingset;
tier = lru_tier_from_refs(refs);
atomic_long_add(delta, &lrugen->refaulted[hist][type][tier]);
mod_lruvec_state(lruvec, WORKINGSET_REFAULT_BASE + type, delta);
/*
* Count the following two cases as stalls:
* 1. For pages accessed through page tables, hotter pages pushed out
* hot pages which refaulted immediately.
* 2. For pages accessed multiple times through file descriptors,
* numbers of accesses might have been out of the range.
*/
if (lru_gen_in_fault() || refs == BIT(LRU_REFS_WIDTH)) {
SetPageWorkingset(page);
mod_lruvec_state(lruvec, WORKINGSET_RESTORE_BASE + type, delta);
}
unlock:
rcu_read_unlock();
}
#else /* !CONFIG_LRU_GEN */
static void *lru_gen_eviction(struct page *page)
{
return NULL;
}
static void lru_gen_refault(struct page *page, void *shadow, unsigned long *eviction)
{
return;
}
#endif /* CONFIG_LRU_GEN */
/* Evaluate the effectively evicted file workingset size */
#ifdef CONFIG_WORKINGSET_EVICT_EVAL
/* Global aggregator work */
static struct delayed_work avgs_work;
/*
* If a page is evicted and never come back, either this page is really cold or it
* is deleted on disk.
*
* For cold page, it could take up all of memory until kswapd start to shrink it.
* For deleted page, the shadow will be gone too, so no refault.
*
* If a page comes back before it's shadow is released, that's a refault, which means
* file page reclaim have gone over-aggressive and that page would not have been evicted
* if all the page, include it self, stayed in memory.
*/
static void workingset_refault_evaluation(struct page *page, struct mem_cgroup *memcg,
struct pglist_data *pgdat, unsigned long eviction)
{
bool file = page_is_file_lru(page);
unsigned long refault, refault_distance;
unsigned long memcg_limit_pg, memcg_usage_pg;
unsigned long balanced_value;
unsigned long remain_distance;
struct lruvec *lruvec;
/*
* We only track file page, anon page can be simply obtained by reading
* swap device status.
*/
if (!file)
return;
/* If original memcg is gone, account this to new memcg the page being read from */
if (!memcg) {
memcg = page_memcg(page);
pgdat = page_pgdat(page);
}
lruvec = mem_cgroup_lruvec(memcg, pgdat);
/* XXX: memcg can be NULL, go through lruvec */
if (!memcg)
memcg = lruvec_memcg(lruvec);
if (!memcg)
return;
/*
* Calc the refault distance and limit by available memory.
*/
refault = atomic_long_read(&lruvec->nonresident_age);
if (lru_gen_enabled())
refault_distance = (refault - eviction) & LRU_GEN_EVICTION_MASK;
else
refault_distance = (refault - eviction) & EVICTION_MASK;
/* If limit not set, use total ram size */
memcg_limit_pg = memcg->memory.max;
if (memcg_limit_pg == PAGE_COUNTER_MAX)
memcg_limit_pg = totalram_pages();
/* Get the current memory usage of this cgroup */
if (mem_cgroup_is_root(memcg))
memcg_usage_pg = memcg_page_state(memcg, NR_FILE_PAGES) + memcg_page_state(memcg, NR_ANON_MAPPED);
else
memcg_usage_pg = page_counter_read(&memcg->memory);
remain_distance = memcg_limit_pg - memcg_usage_pg;
balanced_value = min(refault_distance, remain_distance);
do {
/*
* Not taking any lock, for better performance, may lead to some
* event got lost, but it's just a estimation anyway.
*/
WRITE_ONCE(lruvec->eval_count, lruvec->eval_count + 1);
WRITE_ONCE(lruvec->total_distance, lruvec->total_distance + balanced_value);
} while ((lruvec = parent_lruvec(lruvec)));
}
#define FSHIFT 12 /* 12 of bits of precision */
#define FIXED_P (1 << FSHIFT) /* fixed-point */
#define EVAL_FREQ (10 * HZ + 1 ) /* 10 sec intervals */
#define EXP_1 3467 /* 1/exp(10sec/1min), caculated with FIXED_P / math.exp(10.0/60) */
#define EXP_10 4028 /* 1/exp(10sec/10min), same as above */
#define EXP_30 4073 /* 1/exp(10sec/30min), same as above */
#define EXP_60 4084 /* 1/exp(10sec/60min), same as above */
/*
* a1 = a0 * e + a * (1 - e), same as calc_load.
*/
static inline unsigned long
calc_avg(unsigned long prev_val, unsigned long exp, unsigned long next_val)
{
unsigned long ret;
ret = prev_val * exp + next_val * (FIXED_P - exp);
if (next_val >= prev_val)
ret += FIXED_P-1;
return ret / FIXED_P;
}
/*
* Trigger every 10s (EVAL_FREQ).
*/
static void evict_eval_work(struct work_struct *work)
{
struct delayed_work *dwork;
struct mem_cgroup *memcg;
unsigned long memcg_limit_pg, memcg_usage_pg, memcg_avail_pg;
/*
* Three things to calculate here:
*
* memcg->workingset_valid_eviction_*: corrected reading of valid evictions
* (may plus activitions) of a memcg. Indicats the valid amount of memory ever get evicted.
* This value tends to get really large, so correct it in a way to limit it's size.
*
* lruvec->workingset_watermark_*: corrected max value of avg refault distance
* of a memcg.
*
* lruvec->workingset_refault_distance_*: avg refault distance
* of a memcg. Indicats the most recent refault distances.
*
*/
memcg = mem_cgroup_iter(NULL, NULL, NULL);
for (; memcg != NULL; memcg = mem_cgroup_iter(NULL, memcg, NULL)) {
int nid;
unsigned long lru_period_distance, lru_period_watermark;
unsigned long cg_eviction = 0;
unsigned long active_page = 0;
unsigned long remaining_watermark;
/* If limit not set, use total ram size */
memcg_limit_pg = memcg->memory.max;
if (memcg_limit_pg == PAGE_COUNTER_MAX)
memcg_limit_pg = totalram_pages();
/* Get the current memory usage of this cgroup */
if (mem_cgroup_is_root(memcg))
memcg_usage_pg = memcg_page_state(memcg, NR_FILE_PAGES) + memcg_page_state(memcg, NR_ANON_MAPPED);
else
memcg_usage_pg = page_counter_read(&memcg->memory);
/* Get the current available memory of this cgroup, neither valid_eviction or refault should exceed this */
memcg_avail_pg = memcg_limit_pg - memcg_usage_pg;
/* Calculate workingset_refault_distance_* */
remaining_watermark = memcg_avail_pg;
for_each_node_state(nid, N_MEMORY) {
struct lruvec *lruvec = &memcg->nodeinfo[nid]->lruvec;
/* If refault ever happend, update using new value, else keep momentum using last value */
if (lruvec->eval_count) {
lru_period_distance = READ_ONCE(lruvec->total_distance) / READ_ONCE(lruvec->eval_count);
WRITE_ONCE(lruvec->eval_count, 0);
WRITE_ONCE(lruvec->total_distance, 0);
} else {
lru_period_distance = 0;
}
if (lru_period_distance) {
lruvec->workingset_refault_distance_last = lru_period_distance;
lruvec->workingset_refault_distance_avg[0] = calc_avg(
lruvec->workingset_refault_distance_avg[0], EXP_1, lru_period_distance);
lruvec->workingset_refault_distance_avg[1] = calc_avg(
lruvec->workingset_refault_distance_avg[1], EXP_10, lru_period_distance);
lruvec->workingset_refault_distance_avg[2] = calc_avg(
lruvec->workingset_refault_distance_avg[2], EXP_30, lru_period_distance);
}
/* Make sure the watermark drops slowly, and consider caches are drained to 0 after 1 more hour of idle */
lru_period_watermark = max(lru_period_distance, lruvec->workingset_watermark_last);
lru_period_watermark = calc_avg(lru_period_watermark, EXP_60, 0);
/* Make sure all lruvec's refault distance doesn't add up to a value larger than total memory */
lru_period_watermark = min(remaining_watermark, lru_period_watermark);
lruvec->workingset_watermark_last = lru_period_watermark;
lruvec->workingset_watermark_avg[0] = calc_avg(
lruvec->workingset_watermark_avg[0], EXP_1, lru_period_watermark);
lruvec->workingset_watermark_avg[1] = calc_avg(
lruvec->workingset_watermark_avg[1], EXP_10, lru_period_watermark);
lruvec->workingset_watermark_avg[2] = calc_avg(
lruvec->workingset_watermark_avg[2], EXP_30, lru_period_watermark);
remaining_watermark -= lru_period_watermark;
/* For memcg's workingset_valid_eviction_* */
active_page += lruvec_page_state(lruvec, NR_ACTIVE_FILE) + lruvec_page_state(lruvec, NR_ACTIVE_ANON);
cg_eviction += atomic_long_read(&lruvec->evicted_file);
}
/* It could soon exceed memcg_avail_pg, balance using max memory and watermark */
cg_eviction = min(cg_eviction, memcg_avail_pg);
/*
* XXX: watermark is legacy only
*/
if (READ_ONCE(memcg->memory.watermark) > active_page)
cg_eviction = min(cg_eviction, READ_ONCE(memcg->memory.watermark) - active_page);
/* Make it a evict distance of about past hour */
memcg->workingset_valid_eviction_last = max(cg_eviction, memcg->workingset_eviction_last) - memcg->workingset_eviction_last;
memcg->workingset_eviction_last = calc_avg(memcg->workingset_eviction_last, EXP_60, cg_eviction);
memcg->workingset_valid_eviction_avg[0] = calc_avg(
memcg->workingset_valid_eviction_avg[0], EXP_1, memcg->workingset_valid_eviction_last);
memcg->workingset_valid_eviction_avg[1] = calc_avg(
memcg->workingset_valid_eviction_avg[1], EXP_10, memcg->workingset_valid_eviction_last);
memcg->workingset_valid_eviction_avg[2] = calc_avg(
memcg->workingset_valid_eviction_avg[2], EXP_30, memcg->workingset_valid_eviction_last);
}
dwork = to_delayed_work(work);
schedule_delayed_work(dwork, EVAL_FREQ + 1);
}
static int __init workingset_file_eval_init(void)
{
INIT_DELAYED_WORK(&avgs_work, evict_eval_work);
schedule_delayed_work(&avgs_work, EVAL_FREQ + 1);
return 0;
}
module_init(workingset_file_eval_init);
#else
static void workingset_refault_evaluation(struct page *page, struct mem_cgroup *memcg,
struct pglist_data *pgdat, unsigned long eviction)
{
return;
}
#endif /* CONFIG_WORKINGSET_EVICT_EVAL */
/**
* workingset_age_nonresident - age non-resident entries as LRU ages
* @memcg: the lruvec that was aged
* @nr_pages: the number of pages to count
*
* As in-memory pages are aged, non-resident pages need to be aged as
* well, in order for the refault distances later on to be comparable
* to the in-memory dimensions. This function allows reclaim and LRU
* operations to drive the non-resident aging along in parallel.
*/
void workingset_age_nonresident(struct lruvec *lruvec, unsigned long nr_pages)
{
/*
* Reclaiming a cgroup means reclaiming all its children in a
* round-robin fashion. That means that each cgroup has an LRU
* order that is composed of the LRU orders of its child
* cgroups; and every page has an LRU position not just in the
* cgroup that owns it, but in all of that group's ancestors.
*
* So when the physical inactive list of a leaf cgroup ages,
* the virtual inactive lists of all its parents, including
* the root cgroup's, age as well.
*/
do {
atomic_long_add(nr_pages, &lruvec->nonresident_age);
} while ((lruvec = parent_lruvec(lruvec)));
}
/**
* workingset_eviction - note the eviction of a page from memory
* @target_memcg: the cgroup that is causing the reclaim
* @page: the page being evicted
*
* Returns a shadow entry to be stored in @page->mapping->i_pages in place
* of the evicted @page so that a later refault can be detected.
*/
void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg)
{
struct pglist_data *pgdat = page_pgdat(page);
unsigned long eviction;
struct lruvec *lruvec;
int memcgid;
/* Page is fully exclusive and pins page->mem_cgroup */
VM_BUG_ON_PAGE(PageLRU(page), page);
VM_BUG_ON_PAGE(page_count(page), page);
VM_BUG_ON_PAGE(!PageLocked(page), page);
if (lru_gen_enabled())
return lru_gen_eviction(page);
lruvec = mem_cgroup_lruvec(target_memcg, pgdat);
/* XXX: target_memcg can be NULL, go through lruvec */
memcgid = mem_cgroup_id(lruvec_memcg(lruvec));
eviction = atomic_long_read(&lruvec->nonresident_age);
eviction >>= bucket_order;
workingset_age_nonresident(lruvec, thp_nr_pages(page));
#ifdef CONFIG_WORKINGSET_EVICT_EVAL
workingset_evict_file(lruvec, thp_nr_pages(page));
#endif
return pack_shadow(memcgid, pgdat, eviction, PageWorkingset(page));
}
/**
* workingset_refault - evaluate the refault of a previously evicted page
* @page: the freshly allocated replacement page
* @shadow: shadow entry of the evicted page
*
* Calculates and evaluates the refault distance of the previously
* evicted page in the context of the node and the memcg whose memory
* pressure caused the eviction.
*/
void workingset_refault(struct page *page, void *shadow)
{
bool file = page_is_file_lru(page);
struct mem_cgroup *eviction_memcg;
struct lruvec *eviction_lruvec;
unsigned long refault_distance;
unsigned long workingset_size;
struct pglist_data *pgdat;
struct mem_cgroup *memcg;
unsigned long eviction;
struct lruvec *lruvec;
unsigned long refault;
bool workingset;
int memcgid;
unpack_shadow(shadow, &memcgid, &pgdat, &eviction, &workingset);
eviction <<= bucket_order;
if (lru_gen_enabled()) {
lru_gen_refault(page, shadow, &eviction);
/* Simple diabled the workingset detection below, it's done by MGLRU already */
workingset = false;
}
rcu_read_lock();
/*
* Look up the memcg associated with the stored ID. It might
* have been deleted since the page's eviction.
*
* Note that in rare events the ID could have been recycled
* for a new cgroup that refaults a shared page. This is
* impossible to tell from the available data. However, this
* should be a rare and limited disturbance, and activations
* are always speculative anyway. Ultimately, it's the aging
* algorithm's job to shake out the minimum access frequency
* for the active cache.
*
* XXX: On !CONFIG_MEMCG, this will always return NULL; it
* would be better if the root_mem_cgroup existed in all
* configurations instead.
*/
eviction_memcg = mem_cgroup_from_id(memcgid);
if (!mem_cgroup_disabled() && !eviction_memcg)
goto out;
eviction_lruvec = mem_cgroup_lruvec(eviction_memcg, pgdat);
refault = atomic_long_read(&eviction_lruvec->nonresident_age);
/* For CONFIG_WORKINGSET_EVICT_EVAL */
workingset_refault_evaluation(page, eviction_memcg, pgdat, eviction);
/*
* Calculate the refault distance
*
* The unsigned subtraction here gives an accurate distance
* across nonresident_age overflows in most cases. There is a
* special case: usually, shadow entries have a short lifetime
* and are either refaulted or reclaimed along with the inode
* before they get too old. But it is not impossible for the
* nonresident_age to lap a shadow entry in the field, which
* can then result in a false small refault distance, leading
* to a false activation should this old entry actually
* refault again. However, earlier kernels used to deactivate
* unconditionally with *every* reclaim invocation for the
* longest time, so the occasional inappropriate activation
* leading to pressure on the active list is not a problem.
*/
if (lru_gen_enabled())
refault_distance = (refault - eviction) & LRU_GEN_EVICTION_MASK;
else
refault_distance = (refault - eviction) & EVICTION_MASK;
/*
* The activation decision for this page is made at the level
* where the eviction occurred, as that is where the LRU order
* during page reclaim is being determined.
*
* However, the cgroup that will own the page is the one that
* is actually experiencing the refault event.
*/
memcg = page_memcg(page);
lruvec = mem_cgroup_lruvec(memcg, pgdat);
inc_lruvec_state(lruvec, WORKINGSET_REFAULT_BASE + file);
/*
* Compare the distance to the existing workingset size. We
* don't activate pages that couldn't stay resident even if
* all the memory was available to the workingset. Whether
* workingset competition needs to consider anon or not depends
* on having swap.
*/
workingset_size = lruvec_page_state(eviction_lruvec, NR_ACTIVE_FILE);
if (!file) {
workingset_size += lruvec_page_state(eviction_lruvec,
NR_INACTIVE_FILE);
}
if (mem_cgroup_get_nr_swap_pages(memcg) > 0) {
workingset_size += lruvec_page_state(eviction_lruvec,
NR_ACTIVE_ANON);
if (file) {
workingset_size += lruvec_page_state(eviction_lruvec,
NR_INACTIVE_ANON);
}
}
if (refault_distance > workingset_size)
goto out;
SetPageActive(page);
workingset_age_nonresident(lruvec, thp_nr_pages(page));
inc_lruvec_state(lruvec, WORKINGSET_ACTIVATE_BASE + file);
/* Page was active prior to eviction */
if (workingset) {
SetPageWorkingset(page);
/* XXX: Move to lru_cache_add() when it supports new vs putback */
lru_note_cost_page(page);
inc_lruvec_state(lruvec, WORKINGSET_RESTORE_BASE + file);
}
out:
rcu_read_unlock();
}
/**
* workingset_activation - note a page activation
* @page: page that is being activated
*/
void workingset_activation(struct page *page)
{
struct mem_cgroup *memcg;
struct lruvec *lruvec;
rcu_read_lock();
/*
* Filter non-memcg pages here, e.g. unmap can call
* mark_page_accessed() on VDSO pages.
*
* XXX: See workingset_refault() - this should return
* root_mem_cgroup even for !CONFIG_MEMCG.
*/
memcg = page_memcg_rcu(page);
if (!mem_cgroup_disabled() && !memcg)
goto out;
lruvec = mem_cgroup_page_lruvec(page, page_pgdat(page));
workingset_age_nonresident(lruvec, thp_nr_pages(page));
out:
rcu_read_unlock();
}
/*
* Shadow entries reflect the share of the working set that does not
* fit into memory, so their number depends on the access pattern of
* the workload. In most cases, they will refault or get reclaimed
* along with the inode, but a (malicious) workload that streams
* through files with a total size several times that of available
* memory, while preventing the inodes from being reclaimed, can
* create excessive amounts of shadow nodes. To keep a lid on this,
* track shadow nodes and reclaim them when they grow way past the
* point where they would still be useful.
*/
static struct list_lru shadow_nodes;
void workingset_update_node(struct xa_node *node)
{
/*
* Track non-empty nodes that contain only shadow entries;
* unlink those that contain pages or are being freed.
*
* Avoid acquiring the list_lru lock when the nodes are
* already where they should be. The list_empty() test is safe
* as node->private_list is protected by the i_pages lock.
*/
VM_WARN_ON_ONCE(!irqs_disabled()); /* For __inc_lruvec_page_state */
if (node->count && node->count == node->nr_values) {
if (list_empty(&node->private_list)) {
list_lru_add(&shadow_nodes, &node->private_list);
__inc_lruvec_slab_state(node, WORKINGSET_NODES);
}
} else {
if (!list_empty(&node->private_list)) {
list_lru_del(&shadow_nodes, &node->private_list);
__dec_lruvec_slab_state(node, WORKINGSET_NODES);
}
}
}
static unsigned long count_shadow_nodes(struct shrinker *shrinker,
struct shrink_control *sc)
{
unsigned long max_nodes;
unsigned long nodes;
unsigned long pages;
nodes = list_lru_shrink_count(&shadow_nodes, sc);
/*
* Approximate a reasonable limit for the nodes
* containing shadow entries. We don't need to keep more
* shadow entries than possible pages on the active list,
* since refault distances bigger than that are dismissed.
*
* The size of the active list converges toward 100% of
* overall page cache as memory grows, with only a tiny
* inactive list. Assume the total cache size for that.
*
* Nodes might be sparsely populated, with only one shadow
* entry in the extreme case. Obviously, we cannot keep one
* node for every eligible shadow entry, so compromise on a
* worst-case density of 1/8th. Below that, not all eligible
* refaults can be detected anymore.
*
* On 64-bit with 7 xa_nodes per page and 64 slots
* each, this will reclaim shadow entries when they consume
* ~1.8% of available memory:
*
* PAGE_SIZE / xa_nodes / node_entries * 8 / PAGE_SIZE
*/
#ifdef CONFIG_MEMCG
if (sc->memcg) {
struct lruvec *lruvec;
int i;
lruvec = mem_cgroup_lruvec(sc->memcg, NODE_DATA(sc->nid));
for (pages = 0, i = 0; i < NR_LRU_LISTS; i++)
pages += lruvec_page_state_local(lruvec,
NR_LRU_BASE + i);
pages += lruvec_page_state_local(
lruvec, NR_SLAB_RECLAIMABLE_B) >> PAGE_SHIFT;
pages += lruvec_page_state_local(
lruvec, NR_SLAB_UNRECLAIMABLE_B) >> PAGE_SHIFT;
} else
#endif
pages = node_present_pages(sc->nid);
max_nodes = pages >> (XA_CHUNK_SHIFT - 3);
if (!nodes)
return SHRINK_EMPTY;
if (nodes <= max_nodes)
return 0;
return nodes - max_nodes;
}
static enum lru_status shadow_lru_isolate(struct list_head *item,
struct list_lru_one *lru,
spinlock_t *lru_lock,
void *arg) __must_hold(lru_lock)
{
struct xa_node *node = container_of(item, struct xa_node, private_list);
XA_STATE(xas, node->array, 0);
struct address_space *mapping;
int ret;
/*
* Page cache insertions and deletions synchroneously maintain
* the shadow node LRU under the i_pages lock and the
* lru_lock. Because the page cache tree is emptied before
* the inode can be destroyed, holding the lru_lock pins any
* address_space that has nodes on the LRU.
*
* We can then safely transition to the i_pages lock to
* pin only the address_space of the particular node we want
* to reclaim, take the node off-LRU, and drop the lru_lock.
*/
mapping = container_of(node->array, struct address_space, i_pages);
/* Coming from the list, invert the lock order */
if (!xa_trylock(&mapping->i_pages)) {
spin_unlock_irq(lru_lock);
ret = LRU_RETRY;
goto out;
}
list_lru_isolate(lru, item);
__dec_lruvec_slab_state(node, WORKINGSET_NODES);
spin_unlock(lru_lock);
/*
* The nodes should only contain one or more shadow entries,
* no pages, so we expect to be able to remove them all and
* delete and free the empty node afterwards.
*/
if (WARN_ON_ONCE(!node->nr_values))
goto out_invalid;
if (WARN_ON_ONCE(node->count != node->nr_values))
goto out_invalid;
mapping->nrexceptional -= node->nr_values;
xas.xa_node = xa_parent_locked(&mapping->i_pages, node);
xas.xa_offset = node->offset;
xas.xa_shift = node->shift + XA_CHUNK_SHIFT;
xas_set_update(&xas, workingset_update_node);
/*
* We could store a shadow entry here which was the minimum of the
* shadow entries we were tracking ...
*/
xas_store(&xas, NULL);
__inc_lruvec_slab_state(node, WORKINGSET_NODERECLAIM);
out_invalid:
xa_unlock_irq(&mapping->i_pages);
ret = LRU_REMOVED_RETRY;
out:
cond_resched();
spin_lock_irq(lru_lock);
return ret;
}
static unsigned long scan_shadow_nodes(struct shrinker *shrinker,
struct shrink_control *sc)
{
/* list_lru lock nests inside the IRQ-safe i_pages lock */
return list_lru_shrink_walk_irq(&shadow_nodes, sc, shadow_lru_isolate,
NULL);
}
static struct shrinker workingset_shadow_shrinker = {
.count_objects = count_shadow_nodes,
.scan_objects = scan_shadow_nodes,
.seeks = 0, /* ->count reports only fully expendable nodes */
.flags = SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE,
};
/*
* Our list_lru->lock is IRQ-safe as it nests inside the IRQ-safe
* i_pages lock.
*/
static struct lock_class_key shadow_nodes_key;
static int __init workingset_init(void)
{
unsigned int timestamp_bits;
unsigned int max_order;
int ret;
unsigned int lru_gen_timestamp_bits;
BUILD_BUG_ON(BITS_PER_LONG < EVICTION_SHIFT);
/*
* Calculate the eviction bucket size to cover the longest
* actionable refault distance, which is currently half of
* memory (totalram_pages/2). However, memory hotplug may add
* some more pages at runtime, so keep working with up to
* double the initial memory by using totalram_pages as-is.
*/
timestamp_bits = BITS_PER_LONG - EVICTION_SHIFT;
max_order = fls_long(totalram_pages() - 1);
if (max_order > timestamp_bits)
bucket_order = max_order - timestamp_bits;
pr_info("workingset: timestamp_bits=%d max_order=%d bucket_order=%u\n",
timestamp_bits, max_order, bucket_order);
BUILD_BUG_ON(BITS_PER_LONG < LRU_GEN_EVICTION_SHIFT);
lru_gen_timestamp_bits = BITS_PER_LONG - LRU_GEN_EVICTION_SHIFT;
if (max_order > lru_gen_timestamp_bits)
lru_gen_bucket_order = max_order - timestamp_bits;
pr_info("workingset: lru_gen_timestamp_bits=%d max_order=%d lru_gen_bucket_order=%u\n",
lru_gen_timestamp_bits, max_order, lru_gen_bucket_order);
ret = prealloc_shrinker(&workingset_shadow_shrinker);
if (ret)
goto err;
ret = __list_lru_init(&shadow_nodes, true, &shadow_nodes_key,
&workingset_shadow_shrinker);
if (ret)
goto err_list_lru;
register_shrinker_prepared(&workingset_shadow_shrinker);
return 0;
err_list_lru:
free_prealloced_shrinker(&workingset_shadow_shrinker);
err:
return ret;
}
module_init(workingset_init);
Reference in New Issue View Git Blame Copy Permalink