This patch introduces a runtime hot/cold data separation method
for f2fs, in order to improve the accuracy for data temperature
classification, reduce the garbage collection overhead after
long-term data updates.
Enhanced hot/cold data separation can record data block update
frequency as "age" of the extent per inode, and take use of the age
info to indicate better temperature type for data block allocation:
- It records total data blocks allocated since mount;
- When file extent has been updated, it calculate the count of data
blocks allocated since last update as the age of the extent;
- Before the data block allocated, it searches for the age info and
chooses the suitable segment for allocation.
Test and result:
- Prepare: create about 30000 files
* 3% for cold files (with cold file extension like .apk, from 3M to 10M)
* 50% for warm files (with random file extension like .FcDxq, from 1K
to 4M)
* 47% for hot files (with hot file extension like .db, from 1K to 256K)
- create(5%)/random update(90%)/delete(5%) the files
* total write amount is about 70G
* fsync will be called for .db files, and buffered write will be used
for other files
The storage of test device is large enough(128G) so that it will not
switch to SSR mode during the test.
Benefit: dirty segment count increment reduce about 14%
- before: Dirty +21110
- after: Dirty +18286
Signed-off-by: qixiaoyu1 <qixiaoyu1@xiaomi.com>
Signed-off-by: xiongping1 <xiongping1@xiaomi.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Current atomic write has three major issues like below.
- keeps the updates in non-reclaimable memory space and they are even
hard to be migrated, which is not good for contiguous memory
allocation.
- disk spaces used for atomic files cannot be garbage collected, so
this makes it difficult for the filesystem to be defragmented.
- If atomic write operations hit the threshold of either memory usage
or garbage collection failure count, All the atomic write operations
will fail immediately.
To resolve the issues, I will keep a COW inode internally for all the
updates to be flushed from memory, when we need to flush them out in a
situation like high memory pressure. These COW inodes will be tagged
as orphan inodes to be reclaimed in case of sudden power-cut or system
failure during atomic writes.
Signed-off-by: Daeho Jeong <daehojeong@google.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This adds a sysfs entry to call checkpoint during fsync() in order to avoid
long elapsed time to run roll-forward recovery when booting the device.
Default value doesn't enforce the limitation which is same as before.
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch enables f2fs_balance_fs_bg() to check all metadatas' dirty
threshold rather than just checking node block's, so that checkpoint()
from background can be triggered more frequently to avoid heaping up
too much dirty metadatas.
Threshold value by default:
race with foreground ops single type global
No 16MB 24MB
Yes 24MB 36MB
In addtion, let f2fs_balance_fs_bg() be aware of roll-forward sapce
as well as fsync().
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch supports to migrate swapfile in aligned write mode during
swapon in order to keep swapfile being aligned to section as much as
possible, then pinned swapfile will locates fully filled section which
may not affected by GC.
However, for the case that swapfile's size is not aligned to section
size, it will still leave last extent in file's tail as unaligned due
to its size is smaller than section size, like case #2.
case #1
xfs_io -f /mnt/f2fs/file -c "pwrite 0 4M" -c "fsync"
Before swapon:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..3047]: 1123352..1126399 3048 0x1000
1: [3048..7143]: 237568..241663 4096 0x1000
2: [7144..8191]: 245760..246807 1048 0x1001
After swapon:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..8191]: 249856..258047 8192 0x1001
Kmsg:
F2FS-fs (zram0): Swapfile (2) is not align to section:
1) creat(), 2) ioctl(F2FS_IOC_SET_PIN_FILE), 3) fallocate(2097152 * n)
case #2
xfs_io -f /mnt/f2fs/file -c "pwrite 0 3M" -c "fsync"
Before swapon:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..3047]: 246808..249855 3048 0x1000
1: [3048..6143]: 237568..240663 3096 0x1001
After swapon:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..4095]: 258048..262143 4096 0x1000
1: [4096..6143]: 238616..240663 2048 0x1001
Kmsg:
F2FS-fs (zram0): Swapfile: last extent is not aligned to section
F2FS-fs (zram0): Swapfile (2) is not align to section:
1) creat(), 2) ioctl(F2FS_IOC_SET_PIN_FILE), 3) fallocate(2097152 * n)
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Support to use address space of inner inode to cache compressed block,
in order to improve cache hit ratio of random read.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Restruct f2fs page private layout for below reasons:
There are some cases that f2fs wants to set a flag in a page to
indicate a specified status of page:
a) page is in transaction list for atomic write
b) page contains dummy data for aligned write
c) page is migrating for GC
d) page contains inline data for inline inode flush
e) page belongs to merkle tree, and is verified for fsverity
f) page is dirty and has filesystem/inode reference count for writeback
g) page is temporary and has decompress io context reference for compression
There are existed places in page structure we can use to store
f2fs private status/data:
- page.flags: PG_checked, PG_private
- page.private
However it was a mess when we using them, which may cause potential
confliction:
page.private PG_private PG_checked page._refcount (+1 at most)
a) -1 set +1
b) -2 set
c), d), e) set
f) 0 set +1
g) pointer set
The other problem is page.flags has no free slot, if we can avoid set
zero to page.private and set PG_private flag, then we use non-zero value
to indicate PG_private status, so that we may have chance to reclaim
PG_private slot for other usage. [1]
The other concern is f2fs has bad scalability in aspect of indicating
more page status.
So in this patch, let's restructure f2fs' page.private as below to
solve above issues:
Layout A: lowest bit should be 1
| bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... |
bit 0 PAGE_PRIVATE_NOT_POINTER
bit 1 PAGE_PRIVATE_ATOMIC_WRITE
bit 2 PAGE_PRIVATE_DUMMY_WRITE
bit 3 PAGE_PRIVATE_ONGOING_MIGRATION
bit 4 PAGE_PRIVATE_INLINE_INODE
bit 5 PAGE_PRIVATE_REF_RESOURCE
bit 6- f2fs private data
Layout B: lowest bit should be 0
page.private is a wrapped pointer.
After the change:
page.private PG_private PG_checked page._refcount (+1 at most)
a) 11 set +1
b) 101 set +1
c) 1001 set +1
d) 10001 set +1
e) set
f) 100001 set +1
g) pointer set +1
[1] https://lore.kernel.org/linux-f2fs-devel/20210422154705.GO3596236@casper.infradead.org/T/#u
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
With the default DPOLICY_BG discard thread is ioaware, which prevents
the discard thread from issuing the discard commands. On low RAM setups,
it is observed that these discard commands in the cache are consuming
high memory. This patch aims to relax the memory pressure on the system
due to f2fs pending discard cmds by changing the policy to DPOLICY_FORCE
based on the nm_i->ram_thresh configured.
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In f2fs_try_to_free_nids(), .nid_list_lock spinlock critical region will
increase as expected shrink number increase, to avoid spining other CPUs
for long time, we change to release nid caches with small batch each time
under .nid_list_lock coverage.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch corrects the SPDX License Identifier style in
header files related to F2FS File System support.
For C header files Documentation/process/license-rules.rst
mandates C-like comments (opposed to C source files where
C++ style should be used).
Changes made by using a script provided by Joe Perches here:
https://lkml.org/lkml/2019/2/7/46.
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
For all ordered cases in f2fs_wait_on_page_writeback(), we need to
check PageWriteback status, so let's clean up to relocate the check
into f2fs_wait_on_page_writeback().
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Remove the verbose license text from f2fs files and replace them with
SPDX tags. This does not change the license of any of the code.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Fsyncer will wait on all dnode pages of regular writeback before flushing,
if there are async dnode pages blocked by IO scheduler, it may decrease
fsync's performance.
In this patch, we choose to let f2fs_balance_fs_bg() to trigger checkpoint
to flush these dnode pages of regular, so async IO of dnode page can be
elimitnated, making fsyncer only need to wait for sync IO.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Enable in-memory inode checksum to protect metadata blocks from
in-memory scribbles when checking consistency, which has no
performance requirements.
Signed-off-by: Weichao Guo <guoweichao@huawei.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Nat entry set is used only in checkpoint(), and during checkpoint() we
won't flush new nat entry with unallocated address, so we don't need to
add new nat entry into nat set, then nat_entry_set::entry_cnt can
indicate actual entry count we need to flush in checkpoint().
Signed-off-by: Yunlei He <heyunlei@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
When setting COLD_BIT_SHIFT flag in node block, we only need to call
set_cold_node() in new_node_page() and recover_inode_page() during
node page initialization. So remove unneeded set_cold_node() in other
places.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
If fsck.f2fs changes crc, we have no way to recover some inode blocks by roll-
forward recovery. Let's relax the condition to recover them.
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
If some abnormal users try lots of atomic write operations, f2fs is able to
produce pinned pages in the main memory which affects system performance.
This patch limits that as 20% over total memory size, and if f2fs reaches
to the limit, it will drop all the inmemory pages.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
As Fan Li reported, there is no user traversing nid_list[ALLOC_NID_LIST]
which is used for tracking preallocated nids. Let's drop it, and only
track preallocated nids in free_nid_root radix-tree.
Reported-by: Fan Li <fanofcode.li@samsung.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The index of segment which the next nat block is in has only one different
bit than the current one, so to get the next nat address, we can simply
alter that one bit.
Signed-off-by: Fan Li <fanofcode.li@samsung.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Add braces around variables used within macros for those make sense
to do it. Many of the macros in f2fs already do this. What this commit
doesn't do is anything that changes line# as a result of adding braces,
which usually affects the binary via __LINE__.
Confirmed no diff in fs/f2fs/f2fs.ko before/after this commit on x86_64,
to make sure this has no functional change as well as there's been no
unexpected side effect due to callers' arithmetics within the existing
code.
Signed-off-by: Tomohiro Kusumi <tkusumi@tuxera.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
use a slightly simpler expression to calculate nat block with nid.
Signed-off-by: Fan Li <fanofcode.li@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
There are four places that getting the crc value in f2fs_checkpoint,
just add a new helper cur_cp_crc for them.
Signed-off-by: Kinglong Mee <kinglongmee@gmail.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
proc A: proc B:
- writeback_sb_inodes
- __writeback_single_inode
- do_writepages
- f2fs_write_node_pages
- f2fs_balance_fs_bg - write_checkpoint
- build_free_nids - flush_nat_entries
- __build_free_nids - __flush_nat_entry_set
- ra_meta_pages - get_next_nat_page
- current_nat_addr - set_to_next_nat
[do nat_bitmap checking] - f2fs_change_bit
For proc A, nat_bitmap and nat_bitmap_mir would be compared without lock_op and
nm_i->nat_tree_lock, while proc B is changing nat_bitmap/nat_bitmap_ver in cp.
So it is normal for nat_bitmap/nat_bitmap diffrence under such scenario.
This patch fix this by removing the monitoring point.
[Fix: 599a09b f2fs: check in-memory nat version bitmap]
Signed-off-by: Hou Pengyang <houpengyang@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Currently, if we call fsync after updating the xattr date belongs to the
file, f2fs needs to trigger checkpoint to keep xattr data consistent. But,
this policy cause low performance as checkpoint will block most foreground
operations and cause unneeded and unrelated IOs around checkpoint.
This patch will reuse regular file recovery policy for xattr node block,
so, we change to write xattr node block tagged with fsync flag to warm
area instead of cold area, and during recovery, we search warm node chain
for fsynced xattr block, and do the recovery.
So, for below application IO pattern, performance can be improved
obviously:
- touch file
- create/update/delete xattr entry in file
- fsync file
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch adds a mirror for nat version bitmap, and use it to detect
in-memory bitmap corruption which may be caused by bit-transition of
cache or memory overflow.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
During free nid allocation, in order to do preallocation, we will tag free
nid entry as allocated one and still leave it in free nid list, for other
allocators who want to grab free nids, it needs to traverse the free nid
list for lookup. It becomes overhead in scenario of allocating free nid
intensively by multithreads.
This patch splits free nid list to two list: {free,alloc}_nid_list, to
keep free nids and preallocated free nids separately, after that, traverse
latency will be gone, besides split nid_cnt for separate statistic.
Additionally, introduce __insert_nid_to_list and __remove_nid_from_list for
cleanup.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
[Jaegeuk Kim: modify f2fs_bug_on to avoid needless branches]
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
f2fs contained a number of endianness conversion bugs.
Also, one function should have been 'static'.
Found with sparse by running 'make C=2 CF=-D__CHECK_ENDIAN__ fs/f2fs/'
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch introduces spinlock to protect updating process of ckpt_flags
field in struct f2fs_checkpoint, it avoids incorrectly updating in race
condition.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
[Jaegeuk Kim: add __is_set_ckpt_flags likewise __set_ckpt_flags]
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Previously, we used cp_version only to detect recoverable dnodes.
In order to avoid same garbage cp_version, we needed to truncate the next
dnode during checkpoint, resulting in additional discard or data write.
If we can distinguish this by using crc in addition to cp_version, we can
remove this overhead.
There is backward compatibility concern where it changes node_footer layout.
So, this patch introduces a new checkpoint flag, CP_CRC_RECOVERY_FLAG, to
detect new layout. New layout will be activated only when this flag is set.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The readahead nat pages are more likely to be reclaimed quickly, so it'd better
to gather more free nids in advance.
And, let's keep some free nids as much as possible.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In write_begin, if storage supports stable_page, we don't need to wait for
writeback to update its contents.
This patch introduces to use wait_for_stable_page instead of
wait_on_page_writeback.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The sceanrio is:
1. create fully node blocks
2. flush node blocks
3. write inline_data for all the node blocks again
4. flush node blocks redundantly
So, this patch tries to flush inline_data when flushing node blocks.
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch exports a new sysfs entry 'dirty_nat_ratio' to control threshold
of dirty nat entries, if current ratio exceeds configured threshold,
checkpoint will be triggered in f2fs_balance_fs_bg for flushing dirty nats.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
When testing f2fs with xfstest, generic/251 is stuck for long time,
the case uses below serials to obtain fresh released space in device,
in order to prepare for following fstrim test.
1. rm -rf /mnt/dir
2. mkdir /mnt/dir/
3. cp -axT `pwd`/ /mnt/dir/
4. goto 1
During preparing step, all nat entries will be cached in nat cache,
most of them are dirty entries with invalid blkaddr, which means
nodes related to these entries have been truncated, and they could
be reused after the dirty entries been checkpointed.
However, there was no checkpoint been triggered, so nid allocators
(e.g. mkdir, creat) will run into long journey of iterating all NAT
pages, looking for free nids in alloc_nid->build_free_nids.
Here, in f2fs_balance_fs_bg we give another chance to do checkpoint
to flush nat entries for reusing them in free nid cache when dirty
entry count exceeds 10% of max count.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Use sbi->blocks_per_seg directly to avoid unnecessary calculation when using
1 << sbi->log_blocks_per_seg.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
After finishing building free nid cache, we will try to readahead
asynchronously 4 more pages for the next reloading, the count of
readahead nid pages is fixed.
In some case, like SMR drive, read less sectors with fixed count
each time we trigger RA may be low efficient, since we will face
high seeking overhead, so we'd better let user to configure this
parameter from sysfs in specific workload.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The periodic checkpoint can resolve the previous issue.
So, now we can use this again to improve the reported performance regression:
https://lkml.org/lkml/2015/10/8/20
This reverts commit 15bec0ff5a9ba6d203178fa8772259df6207942a.
Previously, we skip dentry block writes when wbc is SYNC_NONE with no memory
pressure and the number of dirty pages is pretty small.
But, we didn't skip for normal data writes, which gives us not much big impact
on overall performance.
Moreover, by skipping some data writes, kworker falls into infinite loop to try
to write blocks, when many dir inodes have only one dentry block.
So, this patch removes skipping data writes.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Introduce infra macro and data structure for rb-tree based extent cache:
Macros:
* EXT_TREE_VEC_SIZE: indicate vector size for gang lookup in extent tree.
* F2FS_MIN_EXTENT_LEN: indicate minimum length of extent managed in cache.
* EXTENT_CACHE_SHRINK_NUMBER: indicate number of extent in cache will be shrunk.
Basic data structures for extent cache:
* struct extent_tree: extent tree entry per inode.
* struct extent_node: extent info node linked in extent tree.
Besides, adding new extent cache related fields in f2fs_sb_info.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Changman Lee <cm224.lee@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In the normal case, the radix_tree_nodes are freed successfully.
But, when cp_error was detected, we should destroy them forcefully.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In do_recover_data, we find and update previous node pages after updating
its new block addresses.
After then, we call fill_node_footer without reset field, we erase its
cold bit so that this new cold node block is written to wrong log area.
This patch fixes not to miss its old flag.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch moves one member of struct nat_entry: _flag_ to struct node_info,
so _version_ in struct node_info and _flag_ which are unsigned char type will
merge to one 32-bit space in register/memory. So the size of nat_entry will be
reduced from 28 bytes to 24 bytes (for 64-bit machine, reduce its size from 40
bytes to 32 bytes) and then slab memory using by f2fs will be reduced.
changes from v2:
o update description of memory usage gain for 64-bit machine suggested by
Changman Lee.
changes from v1:
o introduce inline copy_node_info() to copy valid data from node info suggested
by Jaegeuk Kim, it can avoid bug.
Reviewed-by: Changman Lee <cm224.lee@samsung.com>
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Jaegeuk Kim