Re: [PATCH v1] xfs: Fix calculation of m_rtx_per_rbmblock

["Nirjhar Roy (IBM)" <nirjhar.roy.lists@gmail.com>]

On 1/19/26 17:16, Nirjhar Roy (IBM) wrote:
> m_rtx_per_rbmblock of struct xfs_mount should use the function
> xfs_rtbitmap_rtx_per_rbmblock() to calculate the number of rt
> extents tracked per bitmap file block instead of always
> calculating it as mp->m_blockwsize << XFS_NBWORDLOG.
> When metadir/rtgroups is enabled, the number of tracked extents
> per bitmap file block is slightly less than sizeof(fsblock) in
> bits, since some bytes are reserved by struct xfs_rtbuf_blkinfo.
>
> Signed-off-by: Nirjhar Roy (IBM) <nirjhar.roy.lists@gmail.com>
> ---
>   fs/xfs/libxfs/xfs_sb.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
> index 94c272a2ae26..b2b4d7f21a5d 100644
> --- a/fs/xfs/libxfs/xfs_sb.c
> +++ b/fs/xfs/libxfs/xfs_sb.c
> @@ -1268,7 +1268,7 @@ xfs_sb_mount_common(
>   	mp->m_agno_log = xfs_highbit32(sbp->sb_agcount - 1) + 1;
>   	mp->m_blockmask = sbp->sb_blocksize - 1;
>   	mp->m_blockwsize = xfs_rtbmblock_size(sbp) >> XFS_WORDLOG;
> -	mp->m_rtx_per_rbmblock = mp->m_blockwsize << XFS_NBWORDLOG;
> +	mp->m_rtx_per_rbmblock = xfs_rtbitmap_rtx_per_rbmblock(mp);

Okay, so there are some build failures with this change (reported by 
kernel test robot). It looks like xfs_rtbitmap_rtx_per_rbmblock() gets 
conditionally compiled and xfs_rtbmblock_size() takes care of the 
metadir enable/disablement. I will keep the present code unchanged and 
cancel this patch.

--NR

>   
>   	ags->blocks = mp->m_sb.sb_agblocks;
>   	ags->blklog = mp->m_sb.sb_agblklog;

-- 
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore