Re: [f2fs-dev] [PATCH] f2fs: fix node_cnt race between extent node destroy and writeback

[Chao Yu <chao@kernel.org>]

On 4/19/2026 12:29 AM, Yongpeng Yang wrote:
> 
> On 4/18/26 8:51 AM, Chao Yu via Linux-f2fs-devel wrote:
>> On 4/17/26 21:26, Yongpeng Yang wrote:
>>>
>>> On 4/17/26 17:00, Chao Yu via Linux-f2fs-devel wrote:
>>>> On 4/3/26 22:40, Yongpeng Yang wrote:
>>>>> From: Yongpeng Yang <yangyongpeng@xiaomi.com>
>>>>>
>>>>> f2fs_destroy_extent_node() does not set FI_NO_EXTENT before clearing
>>>>> extent nodes. When called from f2fs_drop_inode() with I_SYNC set,
>>>>> concurrent kworker writeback can insert new extent nodes into the same
>>>>> extent tree, racing with the destroy and triggering f2fs_bug_on() in
>>>>> __destroy_extent_node(). The scenario is as follows:
>>>>>
>>>>> drop inode                            writeback
>>>>>     - iput
>>>>>      - f2fs_drop_inode  // I_SYNC set
>>>>>       - f2fs_destroy_extent_node
>>>>>        - __destroy_extent_node
>>>>>         - while (node_cnt) {
>>>>>            write_lock(&et->lock)
>>>>>            __free_extent_tree
>>>>>            write_unlock(&et->lock)
>>>>>                                           - __writeback_single_inode
>>>>>                                            - f2fs_outplace_write_data
>>>>>                                             -
>>>>> f2fs_update_read_extent_cache
>>>>>                                              -
>>>>> __update_extent_tree_range
>>>>>                                               // FI_NO_EXTENT not set,
>>>>>                                               // insert new extent node
>>>>>           } // node_cnt == 0, exit while
>>>>>         - f2fs_bug_on(node_cnt)  // node_cnt > 0
>>>>>
>>>>> Additionally, __update_extent_tree_range() only checks FI_NO_EXTENT for
>>>>> EX_READ type, leaving EX_BLOCK_AGE updates completely unprotected.
>>>>>
>>>>> This patch set FI_NO_EXTENT under et->lock in __destroy_extent_node(),
>>>>> consistent with other callers (__update_extent_tree_range and
>>>>> __drop_extent_tree) and check FI_NO_EXTENT for both EX_READ and
>>>>> EX_BLOCK_AGE tree.
>>>>
>>>> I suffered below test failure, then I bisect to this change.
>>>>
>>>>       generic/475  84s ... [failed, exit status 1]- output mismatch
>>>> (see /
>>>> share/git/fstests/results//generic/475.out.bad)
>>>>       --- tests/generic/475.out   2025-01-12 21:57:40.279440664 +0800
>>>>       +++ /share/git/fstests/results//generic/475.out.bad 2026-04-17
>>>> 12:08:28.000000000 +0800
>>>>       @@ -1,2 +1,6 @@
>>>>        QA output created by 475
>>>>        Silence is golden.
>>>>       +mount: /mnt/scratch_f2fs: mount system call failed: Structure
>>>> needs
>>>> cleaning.
>>>>       +       dmesg(1) may have more information after failed mount
>>>> system
>>>> call.
>>>>       +mount failed
>>>>       +(see /share/git/fstests/results//generic/475.full for details)
>>>>       ...
>>>>       (Run 'diff -u /share/git/fstests/tests/generic/475.out /share/git/
>>>> fstests/results//generic/475.out.bad'  to see the entire diff)
>>>>
>>>>
>>>>       generic/388  73s ... [failed, exit status 1]- output mismatch
>>>> (see /
>>>> share/git/fstests/results//generic/388.out.bad)
>>>>       --- tests/generic/388.out   2025-01-12 21:57:40.275440602 +0800
>>>>       +++ /share/git/fstests/results//generic/388.out.bad 2026-04-17
>>>> 11:58:05.000000000 +0800
>>>>       @@ -1,2 +1,6 @@
>>>>        QA output created by 388
>>>>        Silence is golden.
>>>>       +mount: /mnt/scratch_f2fs: mount system call failed: Structure
>>>> needs
>>>> cleaning.
>>>>       +       dmesg(1) may have more information after failed mount
>>>> system
>>>> call.
>>>>       +cycle mount failed
>>>>       +(see /share/git/fstests/results//generic/388.full for details)
>>>>       ...
>>>>       (Run 'diff -u /share/git/fstests/tests/generic/388.out /share/git/
>>>> fstests/results//generic/388.out.bad'  to see the entire diff)
>>>>
>>>>
>>>>       F2FS-fs (dm-0): sanity_check_extent_cache: inode (ino=1761) extent
>>>> info [220057, 57, 6] is incorrect, run fsck to fix
>>>>
>>>> I suspect we may miss any extent updates after we set FI_NO_EXTENT in
>>>> __destroy_extent_node(), result in failing in
>>>> sanity_check_extent_cache().
>>>>
>>>> Can we just relocate f2fs_bug_on(node_cnt) rather than complicated
>>>> change?
>>>> Thoughts?
>>>
>>> Oh, I overlooked largest extent. How about relocate
>>> f2fs_bug_on(node_cnt) to __destroy_extent_tree?
>>>
>>> static void __destroy_extent_tree(struct inode *inode, enum extent_type
>>> type)
>>>
>>>           /* free all extent info belong to this extent tree */
>>>           node_cnt = __destroy_extent_node(inode, type);
>>> +       f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
>>
>>       /* free all extent info belong to this extent tree */
>>       node_cnt = __destroy_extent_node(inode, type);
>>
>>       /* delete extent tree entry in radix tree */
>>       mutex_lock(&eti->extent_tree_lock);
>>       f2fs_bug_on(sbi, atomic_read(&et->node_cnt));  <---
>>
>> Oh, it has already checked node_cnt, so, maybe we can just remove the
>> check in
>> __destroy_extent_node()?
> 
> Yes. BTW, is it correct to remove the call to f2fs_destroy_extent_node()
> in f2fs_drop_inode()? It seems this call is unnecessary, since
> f2fs_evict_inode() will eventually delete all extent nodes properly.

I think it's fine to keep it according to original intention "destroy
extent_tree for the truncation case" introduced from 3e72f721390d
("f2fs: use extent_cache by default"). It helps the performance w/
in batch extent node release.

Thanks,

> 
> Thanks
> Yongpeng,
> 
>>
>> Thanks,
>>
>>
>>>
>>> Thanks
>>> Yongpeng,
>>>
>>>>
>>>> Thanks,
>>>>
>>>>>
>>>>> Fixes: 3fc5d5a182f6 ("f2fs: fix to shrink read extent node in batches")
>>>>> Cc: stable@vger.kernel.org
>>>>> Signed-off-by: Yongpeng Yang <yangyongpeng@xiaomi.com>
>>>>> ---
>>>>>     fs/f2fs/extent_cache.c | 17 ++++++++++-------
>>>>>     1 file changed, 10 insertions(+), 7 deletions(-)
>>>>>
>>>>> diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
>>>>> index 0ed84cc065a7..87169fd29d89 100644
>>>>> --- a/fs/f2fs/extent_cache.c
>>>>> +++ b/fs/f2fs/extent_cache.c
>>>>> @@ -119,9 +119,10 @@ static bool __may_extent_tree(struct inode
>>>>> *inode, enum extent_type type)
>>>>>         if (!__init_may_extent_tree(inode, type))
>>>>>             return false;
>>>>>     +    if (is_inode_flag_set(inode, FI_NO_EXTENT))
>>>>> +        return false;
>>>>> +
>>>>>         if (type == EX_READ) {
>>>>> -        if (is_inode_flag_set(inode, FI_NO_EXTENT))
>>>>> -            return false;
>>>>>             if (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
>>>>>                      !f2fs_sb_has_readonly(F2FS_I_SB(inode)))
>>>>>                 return false;
>>>>> @@ -644,6 +645,8 @@ static unsigned int __destroy_extent_node(struct
>>>>> inode *inode,
>>>>>           while (atomic_read(&et->node_cnt)) {
>>>>>             write_lock(&et->lock);
>>>>> +        if (!is_inode_flag_set(inode, FI_NO_EXTENT))
>>>>> +            set_inode_flag(inode, FI_NO_EXTENT);
>>>>>             node_cnt += __free_extent_tree(sbi, et, nr_shrink);
>>>>>             write_unlock(&et->lock);
>>>>>         }
>>>>> @@ -688,12 +691,12 @@ static void __update_extent_tree_range(struct
>>>>> inode *inode,
>>>>>           write_lock(&et->lock);
>>>>>     -    if (type == EX_READ) {
>>>>> -        if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
>>>>> -            write_unlock(&et->lock);
>>>>> -            return;
>>>>> -        }
>>>>> +    if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
>>>>> +        write_unlock(&et->lock);
>>>>> +        return;
>>>>> +    }
>>>>>     +    if (type == EX_READ) {
>>>>>             prev = et->largest;
>>>>>             dei.len = 0;
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Linux-f2fs-devel mailing list
>>>> Linux-f2fs-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
>>>
>>
>>
>>
>> _______________________________________________
>> Linux-f2fs-devel mailing list
>> Linux-f2fs-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
>