Re: [PATCH] libfrog: fix the if condition in xfrog_bulk_req_v1_setup

["Darrick J. Wong" <djwong@kernel.org>]

On Fri, Jul 29, 2022 at 03:57:46PM +0800, Stephen Zhang wrote:
> when scanning all inodes in each ag, hdr->ino serves as a iterator to
> specify the ino to start scanning with.
> 
> After hdr->ino-- , we can get the last ino returned from the previous
> iteration.
> 
> But there are cases that hdr->ino-- is pointless, that is,the case when
> starting to scan inodes in each ag.
> 
> Hence the condition should be cvt_ino_to_agno(xfd, hdr->ino) ==0, which
> represents the start of scan in each ag,

Er, cvt_ino_to_agno extracts the AG number from an inumber;
cvt_ino_to_agino extracts the inumber within an AG.  Given your
description of the problem (not wanting hdr->ino to go backwards in the
inumber space when it's already at the start of an AG), I think you want
the latter here?

> instead of hdr->ino ==0, which represents the start of scan in ag 0 only.
> 
> Signed-off-by: Stephen Zhang <zhangshida@kylinos.cn>
> ---
>  libfrog/bulkstat.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/libfrog/bulkstat.c b/libfrog/bulkstat.c
> index 195f6ea0..77a385bb 100644
> --- a/libfrog/bulkstat.c
> +++ b/libfrog/bulkstat.c
> @@ -172,7 +172,7 @@ xfrog_bulk_req_v1_setup(
>  	if (!buf)
>  		return -errno;
>  
> -	if (hdr->ino)
> +	if (cvt_ino_to_agno(xfd, hdr->ino))

...because I think this change means that we go backwards for any inode
in AG 0, and we do not go backwards for any other AG.

--D

>  		hdr->ino--;
>  	bulkreq->lastip = (__u64 *)&hdr->ino,
>  	bulkreq->icount = hdr->icount,
> -- 
> 2.25.1
>