On Mar 24, 2024 / 20:13, Chao Yu wrote:
...
> Hi Shinichiro,
>
> Can you please check below diff? IIUC, for the case: f2fs_map_blocks()
> returns zero blkaddr in non-primary device, which is a verified valid
> block address, we'd better to check m_flags & F2FS_MAP_MAPPED instead
> of map.m_pblk != NULL_ADDR to decide whether tagging IOMAP_MAPPED flag
> or not.
>
> ---
> fs/f2fs/data.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index 6f66e3e4221a..41a56d4298c8 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -4203,7 +4203,7 @@ static int f2fs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
> if (WARN_ON_ONCE(map.m_pblk == COMPRESS_ADDR))
> return -EINVAL;
>
> - if (map.m_pblk != NULL_ADDR) {
> + if (map.m_flags & F2FS_MAP_MAPPED) {
> iomap->length = blks_to_bytes(inode, map.m_len);
> iomap->type = IOMAP_MAPPED;
> iomap->flags |= IOMAP_F_MERGED;
>
Thanks Chao, I confirmed that the diff above avoids the WARN and zbd/010
failure. From that point of view, it looks good.
One thing I noticed is that the commit message of 8d3c1fa3fa5ea ("f2fs:
don't rely on F2FS_MAP_* in f2fs_iomap_begin") says that f2fs_map_blocks()
might be setting F2FS_MAP_* flag on a hole, and that's why the commit
avoided the F2FS_MAP_MAPPED flag check. So I was not sure if it is the
right thing to reintroduce the flag check.
