diff options
| author | Chao Yu <yuchao0@huawei.com> | 2020-05-29 17:29:47 +0800 |
|---|---|---|
| committer | Jaegeuk Kim <jaegeuk@kernel.org> | 2020-05-30 08:17:52 -0700 |
| commit | ca7f76e680745d3b8a386638045f85dac1c4b2f4 (patch) | |
| tree | c6889db95486ee70e3bee573a9dd4bff7136dc09 /fs/f2fs/segment.c | |
| parent | dc35d73a42291b2a68f5b9a12b2b095b82194c1f (diff) | |
f2fs: fix wrong discard space
Under heavy fsstress, we may triggle panic while issuing discard,
because __check_sit_bitmap() detects that discard command may earse
valid data blocks, the root cause is as below race stack described,
since we removed lock when flushing quota data, quota data writeback
may race with write_checkpoint(), so that it causes inconsistency in
between cached discard entry and segment bitmap.
- f2fs_write_checkpoint
- block_operations
- set_sbi_flag(sbi, SBI_QUOTA_SKIP_FLUSH)
- f2fs_flush_sit_entries
- add_discard_addrs
- __set_bit_le(i, (void *)de->discard_map);
- f2fs_write_data_pages
- f2fs_write_single_data_page
: inode is quota one, cp_rwsem won't be locked
- f2fs_do_write_data_page
- f2fs_allocate_data_block
- f2fs_wait_discard_bio
: discard entry has not been added yet.
- update_sit_entry
- f2fs_clear_prefree_segments
- f2fs_issue_discard
: add discard entry
In order to fix this, this patch uses node_write to serialize
f2fs_allocate_data_block and checkpoint.
Fixes: 435cbab95e39 ("f2fs: fix quota_sync failure due to f2fs_lock_op")
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Diffstat (limited to 'fs/f2fs/segment.c')
| -rw-r--r-- | fs/f2fs/segment.c | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index ebbadde6cbce..196f31503511 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -3107,6 +3107,14 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, type = CURSEG_COLD_DATA; } + /* + * We need to wait for node_write to avoid block allocation during + * checkpoint. This can only happen to quota writes which can cause + * the below discard race condition. + */ + if (IS_DATASEG(type)) + down_write(&sbi->node_write); + down_read(&SM_I(sbi)->curseg_lock); mutex_lock(&curseg->curseg_mutex); @@ -3172,6 +3180,9 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, up_read(&SM_I(sbi)->curseg_lock); + if (IS_DATASEG(type)) + up_write(&sbi->node_write); + if (put_pin_sem) up_read(&sbi->pin_sem); } |