diff mbox series

xfs_scrub: don't use statvfs to collect data volume block counts

Message ID 20201002201834.GC49524@magnolia (mailing list archive)
State Superseded
Headers show
Series xfs_scrub: don't use statvfs to collect data volume block counts | expand

Commit Message

Darrick J. Wong Oct. 2, 2020, 8:18 p.m. UTC
From: Darrick J. Wong <darrick.wong@oracle.com>

The function scrub_scan_estimate_blocks naïvely uses the statvfs counts
to estimate the size and free blocks on the data volume.  Unfortunately,
it fails to account for the fact that statvfs can return the size and
free counts for the realtime volume if the root directory has the
rtinherit flag set, which leads to phase 7 reporting totally absurd
quantities.

The XFS_IOC_FSCOUNTS ioctl returns the size and free block count of both
volumes correctly, so use that instead.

Fixes: 604dd3345f35 ("xfs_scrub: filesystem counter collection functions")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 scrub/fscounters.c |    6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

Comments

Eric Sandeen Oct. 4, 2020, 10:31 p.m. UTC | #1
On 10/2/20 3:18 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> The function scrub_scan_estimate_blocks naïvely uses the statvfs counts
> to estimate the size and free blocks on the data volume.  Unfortunately,
> it fails to account for the fact that statvfs can return the size and
> free counts for the realtime volume if the root directory has the
> rtinherit flag set, which leads to phase 7 reporting totally absurd
> quantities.
> 
> The XFS_IOC_FSCOUNTS ioctl returns the size and free block count of both
> volumes correctly, so use that instead.
> 
> Fixes: 604dd3345f35 ("xfs_scrub: filesystem counter collection functions")
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  scrub/fscounters.c |    6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/scrub/fscounters.c b/scrub/fscounters.c
> index f9d64f8c008f..a2ca0b3f018c 100644
> --- a/scrub/fscounters.c
> +++ b/scrub/fscounters.c
> @@ -154,10 +154,8 @@ scrub_scan_estimate_blocks(
>  
>  	sfs.f_bfree += rb.resblks_avail;
>  
> -	*d_blocks = sfs.f_blocks;
> -	if (ctx->mnt.fsgeom.logstart > 0)
> -		*d_blocks += ctx->mnt.fsgeom.logblocks;
> -	*d_bfree = sfs.f_bfree;
> +	*d_blocks = ctx->mnt.fsgeom.datablocks;
> +	*d_bfree = fc.freedata;
>  	*r_blocks = ctx->mnt.fsgeom.rtblocks;
>  	*r_bfree = fc.freertx;
>  	*f_files = sfs.f_files;
> 

could this just use fc.freeino/fc.allocino too and drop statvfrsr altogether?
(or did I lose track of differences between the counters on the 2 interfaces...)

-Eric
diff mbox series

Patch

diff --git a/scrub/fscounters.c b/scrub/fscounters.c
index f9d64f8c008f..a2ca0b3f018c 100644
--- a/scrub/fscounters.c
+++ b/scrub/fscounters.c
@@ -154,10 +154,8 @@  scrub_scan_estimate_blocks(
 
 	sfs.f_bfree += rb.resblks_avail;
 
-	*d_blocks = sfs.f_blocks;
-	if (ctx->mnt.fsgeom.logstart > 0)
-		*d_blocks += ctx->mnt.fsgeom.logblocks;
-	*d_bfree = sfs.f_bfree;
+	*d_blocks = ctx->mnt.fsgeom.datablocks;
+	*d_bfree = fc.freedata;
 	*r_blocks = ctx->mnt.fsgeom.rtblocks;
 	*r_bfree = fc.freertx;
 	*f_files = sfs.f_files;