[v1] f2fs: add error checking in do_write_page()

[PATCH] f2fs: add error checking in do_write_page()

Posted by mason.zhang 2 months, 1 week ago

Otherwise, the filesystem may unaware of potential file corruption.

Signed-off-by: mason.zhang <masonzhang.linuxer@gmail.com>
---
 fs/f2fs/segment.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index ae1223ef648f..ce80ba226aed 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -3936,12 +3936,15 @@ static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
 	int seg_type = log_type_to_seg_type(type);
 	bool keep_order = (f2fs_lfs_mode(fio->sbi) &&
 				seg_type == CURSEG_COLD_DATA);
+	int err;
 
 	if (keep_order)
 		f2fs_down_read(&fio->sbi->io_order_lock);
 
-	if (f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
-			&fio->new_blkaddr, sum, type, fio)) {
+	err = f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
+			&fio->new_blkaddr, sum, type, fio);
+	if (unlikely(err)) {
+		f2fs_err(fio->sbi, "Failed to allocate data block(%d)", err);
 		if (fscrypt_inode_uses_fs_layer_crypto(folio->mapping->host))
 			fscrypt_finalize_bounce_page(&fio->encrypted_page);
 		folio_end_writeback(folio);
-- 
2.50.0

Re: [PATCH] f2fs: add error checking in do_write_page()

Posted by Chao Yu 2 months, 1 week ago

On 7/24/2025 11:59 PM, mason.zhang wrote:
> Otherwise, the filesystem may unaware of potential file corruption.
> 
> Signed-off-by: mason.zhang <masonzhang.linuxer@gmail.com>
> ---
>   fs/f2fs/segment.c | 7 +++++--
>   1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> index ae1223ef648f..ce80ba226aed 100644
> --- a/fs/f2fs/segment.c
> +++ b/fs/f2fs/segment.c
> @@ -3936,12 +3936,15 @@ static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
>   	int seg_type = log_type_to_seg_type(type);
>   	bool keep_order = (f2fs_lfs_mode(fio->sbi) &&
>   				seg_type == CURSEG_COLD_DATA);
> +	int err;
>   
>   	if (keep_order)
>   		f2fs_down_read(&fio->sbi->io_order_lock);
>   
> -	if (f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
> -			&fio->new_blkaddr, sum, type, fio)) {
> +	err = f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
> +			&fio->new_blkaddr, sum, type, fio);
> +	if (unlikely(err)) {
> +		f2fs_err(fio->sbi, "Failed to allocate data block(%d)", err);

How about using f2fs_err_ratelimited() to avoid too many logs?

Can you please dump more informations about inode, page, blkaddr...?

Thanks,

>   		if (fscrypt_inode_uses_fs_layer_crypto(folio->mapping->host))
>   			fscrypt_finalize_bounce_page(&fio->encrypted_page);
>   		folio_end_writeback(folio);