Re: [PATCH] ext4: wait for tail page commit before truncating a failed write

From: Jan Kara

Date: Fri Nov 13 2026 - 07:56:03 EST


On Thu 12-11-26 19:59:06, diasyzhangqing@xxxxxxxxx wrote:
> From: Qing Zhang <diasyzhang@xxxxxxxxxxx>
>
> In data=journal mode, a failed extending write can leave buffers beyond
> i_size attached to the committing transaction. When blocksize < PAGE_SIZE,
> invalidating these buffers in the partial tail folio returns -EBUSY,
> triggering a warning in ext4_journalled_invalidate_folio() and leaving
> invalidation incomplete.
>
> Call ext4_wait_for_tail_page_commit() before truncating the page cache
> in ext4_truncate_failed_write(), as is already done when shrinking a file
> via ext4_setattr(). The helper releases the folio lock before waiting
> for the commit to finish, then retries invalidation.
>
> Fixes: 53e872681fed ("ext4: fix deadlock in journal_unmap_buffer()")
> Reported-by: syzbot+f111db4cb2acd44eb2f4@xxxxxxxxxxxxxxxxxxxxxxxxx
> Closes: https://syzbot.appspot.com/bug?extid=f111db4cb2acd44eb2f4
> Signed-off-by: Qing Zhang <diasyzhang@xxxxxxxxxxx>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@xxxxxxx>

Honza

> ---
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
>
> fs/ext4/ext4.h | 1 +
> fs/ext4/inode.c | 2 +-
> fs/ext4/truncate.h | 2 ++
> 3 files changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index f38fec3167fb..505ec38fa508 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -3214,6 +3214,7 @@ int ext4_get_fc_inode_loc(struct super_block *sb, unsigned int ino,
> extern int ext4_inode_attach_jinode(struct inode *inode);
> extern int ext4_can_truncate(struct inode *inode);
> extern int ext4_truncate(struct inode *);
> +extern void ext4_wait_for_tail_page_commit(struct inode *inode);
> extern int ext4_break_layouts(struct inode *);
> extern int ext4_truncate_page_cache_block_range(struct inode *inode,
> loff_t start, loff_t end);
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index 10cd85d4c721..f59c4a5b7561 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -6845,7 +6845,7 @@ int ext4_sync_inode_metadata(struct inode *inode, struct writeback_control *wbc)
> * buffers that are attached to a folio straddling i_size and are undergoing
> * commit. In that case we have to wait for commit to finish and try again.
> */
> -static void ext4_wait_for_tail_page_commit(struct inode *inode)
> +void ext4_wait_for_tail_page_commit(struct inode *inode)
> {
> unsigned offset;
> journal_t *journal = EXT4_SB(inode->i_sb)->s_journal;
> diff --git a/fs/ext4/truncate.h b/fs/ext4/truncate.h
> index ce84aa2786c7..c10e4ab4242e 100644
> --- a/fs/ext4/truncate.h
> +++ b/fs/ext4/truncate.h
> @@ -17,6 +17,8 @@ static inline void ext4_truncate_failed_write(struct inode *inode)
> * We don't need to call ext4_break_layouts() because the blocks we
> * are truncating were never visible to userspace.
> */
> filemap_invalidate_lock(mapping);
> + if (ext4_should_journal_data(inode))
> + ext4_wait_for_tail_page_commit(inode);
> truncate_inode_pages(mapping, inode->i_size);
> ext4_truncate(inode);
> filemap_invalidate_unlock(mapping);
> --
> 2.43.0
>
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR