Syzbot reported that after nilfs2 reads a corrupted file system image and degrades to read-only, the BUG_ON check for the buffer delay flag in submit_bh_wbc() may fail, causing a kernel bug. This is because the buffer delay flag is not cleared when clearing the buffer state flags to discard a page/folio or a buffer head. So, fix this. This became necessary when the use of nilfs2's own page clear routine was expanded. This state inconsistency does not occur if the buffer is written normally by log writing. Signed-off-by: Ryusuke Konishi <konishi.ryusuke@xxxxxxxxx> Reported-by: syzbot+985ada84bf055a575c07@xxxxxxxxxxxxxxxxxxxxxxxxx Closes: https://syzkaller.appspot.com/bug?extid=985ada84bf055a575c07 Fixes: 8c26c4e2694a ("nilfs2: fix issue with flush kernel thread after remount in RO mode because of driver's internal error or metadata corruption") Cc: stable@xxxxxxxxxxxxxxx --- Andrew, please apply this as a bug fix. This fixes a kernel bug recently reported by Syzbot. Thanks, Ryusuke Konishi fs/nilfs2/page.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index 9c0b7cddeaae..5436eb0424bd 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -77,7 +77,8 @@ void nilfs_forget_buffer(struct buffer_head *bh) const unsigned long clear_bits = (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | - BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected) | + BIT(BH_Delay)); lock_buffer(bh); set_mask_bits(&bh->b_state, clear_bits, 0); @@ -406,7 +407,8 @@ void nilfs_clear_folio_dirty(struct folio *folio) const unsigned long clear_bits = (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | - BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected) | + BIT(BH_Delay)); bh = head; do { -- 2.43.0
