linux.git/fs/f2fs/extent_cache.c, branch v6.3

f2fs: fix to do sanity check on extent cache correctly

2023-02-09T17:17:30+00:00

In do_read_inode(), sanity check for extent cache should be called after
f2fs_init_read_extent_tree(), fix it.

Fixes: 72840cccc0a1 ("f2fs: allocate the extent_cache by default")
Signed-off-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: fix typos in comments

2023-02-07T18:39:28+00:00

This patch is to fix typos in f2fs files.

Signed-off-by: Jinyoung Choi 
Reviewed-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: add sysfs nodes to set last_age_weight

2023-02-07T18:39:19+00:00

Signed-off-by: qixiaoyu1 
Signed-off-by: xiongping1 
Reviewed-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: fix wrong calculation of block age

2023-02-06T03:30:05+00:00

Currently we wrongly calculate the new block age to
old * LAST_AGE_WEIGHT / 100.

Fix it to new * (100 - LAST_AGE_WEIGHT) / 100
                + old * LAST_AGE_WEIGHT / 100.

Signed-off-by: qixiaoyu1 
Signed-off-by: xiongping1 
Reviewed-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: add a f2fs_lookup_extent_cache_block helper

2023-01-06T23:13:00+00:00

All but three callers of f2fs_lookup_extent_cache just want the block
address.  Add a small helper to simplify them.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: let's avoid panic if extent_tree is not created

2023-01-03T16:59:06+00:00

This patch avoids the below panic.

pc : __lookup_extent_tree+0xd8/0x760
lr : f2fs_do_write_data_page+0x104/0x87c
sp : ffffffc010cbb3c0
x29: ffffffc010cbb3e0 x28: 0000000000000000
x27: ffffff8803e7f020 x26: ffffff8803e7ed40
x25: ffffff8803e7f020 x24: ffffffc010cbb460
x23: ffffffc010cbb480 x22: 0000000000000000
x21: 0000000000000000 x20: ffffffff22e90900
x19: 0000000000000000 x18: ffffffc010c5d080
x17: 0000000000000000 x16: 0000000000000020
x15: ffffffdb1acdbb88 x14: ffffff888759e2b0
x13: 0000000000000000 x12: ffffff802da49000
x11: 000000000a001200 x10: ffffff8803e7ed40
x9 : ffffff8023195800 x8 : ffffff802da49078
x7 : 0000000000000001 x6 : 0000000000000000
x5 : 0000000000000006 x4 : ffffffc010cbba28
x3 : 0000000000000000 x2 : ffffffc010cbb480
x1 : 0000000000000000 x0 : ffffff8803e7ed40
Call trace:
 __lookup_extent_tree+0xd8/0x760
 f2fs_do_write_data_page+0x104/0x87c
 f2fs_write_single_data_page+0x420/0xb60
 f2fs_write_cache_pages+0x418/0xb1c
 __f2fs_write_data_pages+0x428/0x58c
 f2fs_write_data_pages+0x30/0x40
 do_writepages+0x88/0x190
 __writeback_single_inode+0x48/0x448
 writeback_sb_inodes+0x468/0x9e8
 __writeback_inodes_wb+0xb8/0x2a4
 wb_writeback+0x33c/0x740
 wb_do_writeback+0x2b4/0x400
 wb_workfn+0xe4/0x34c
 process_one_work+0x24c/0x5bc
 worker_thread+0x3e8/0xa50
 kthread+0x150/0x1b4

Signed-off-by: Jaegeuk Kim

f2fs: should use a temp extent_info for lookup

2023-01-03T16:59:06+00:00

Otherwise, __lookup_extent_tree() will override the given extent_info which will
be used by caller.

Signed-off-by: Jaegeuk Kim

f2fs: don't mix to use union values in extent_info

2023-01-03T16:59:06+00:00

Let's explicitly use the defined values in block_age case only.

Signed-off-by: Jaegeuk Kim

f2fs: initialize extent_cache parameter

2023-01-03T16:59:05+00:00

This can avoid confusing tracepoint values.

Signed-off-by: Jaegeuk Kim

f2fs: add block_age-based extent cache

2022-12-12T22:53:56+00:00

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 
Signed-off-by: xiongping1 
Signed-off-by: Jaegeuk Kim