cluster-devel.redhat.com archive mirror
 help / color / mirror / Atom feed
* [Cluster-devel] [GFS2 PATCH] GFS2: Don't skip rgrps for Orlov's sake if reservation is active
@ 2012-11-12 18:31 Bob Peterson
  2012-11-12 18:41 ` Steven Whitehouse
  0 siblings, 1 reply; 2+ messages in thread
From: Bob Peterson @ 2012-11-12 18:31 UTC (permalink / raw)
  To: cluster-devel.redhat.com

Hi,

This patch changes function gfs2_inplace_reserve so that it does
not perform the Orlov allocator resource group skipping in cases
where we have an active reservation. Skipping an active rgrp is
improper and gets our block reservation counts out of sync.

Regards,

Bob Peterson
Red Hat File Systems

Signed-off-by: Bob Peterson <rpeterso@redhat.com> 
---
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 99a6197..61a4314 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1811,13 +1811,14 @@ int gfs2_inplace_reserve(struct gfs2_inode *ip, u32 requested, u32 aflags)
 	if (gfs2_rs_active(rs)) {
 		begin = rs->rs_rbm.rgd;
 		flags = 0; /* Yoda: Do or do not. There is no try */
-	} else if (ip->i_rgd && rgrp_contains_block(ip->i_rgd, ip->i_goal)) {
-		rs->rs_rbm.rgd = begin = ip->i_rgd;
 	} else {
-		rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
+		if (ip->i_rgd && rgrp_contains_block(ip->i_rgd, ip->i_goal))
+			rs->rs_rbm.rgd = begin = ip->i_rgd;
+		else
+			rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
+		if (S_ISDIR(ip->i_inode.i_mode) && (aflags & GFS2_AF_ORLOV))
+			skip = gfs2_orlov_skip(ip);
 	}
-	if (S_ISDIR(ip->i_inode.i_mode) && (aflags & GFS2_AF_ORLOV))
-		skip = gfs2_orlov_skip(ip);
 	if (rs->rs_rbm.rgd == NULL)
 		return -EBADSLT;
 



^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [Cluster-devel] [GFS2 PATCH] GFS2: Don't skip rgrps for Orlov's sake if reservation is active
  2012-11-12 18:31 [Cluster-devel] [GFS2 PATCH] GFS2: Don't skip rgrps for Orlov's sake if reservation is active Bob Peterson
@ 2012-11-12 18:41 ` Steven Whitehouse
  0 siblings, 0 replies; 2+ messages in thread
From: Steven Whitehouse @ 2012-11-12 18:41 UTC (permalink / raw)
  To: cluster-devel.redhat.com

Hi,

I don't think this makes sense... we will never have an active
reservation when the orlov allocator is in use, since it is only ever
used for new directory inodes,

Steve.

On Mon, 2012-11-12 at 13:31 -0500, Bob Peterson wrote:
> Hi,
> 
> This patch changes function gfs2_inplace_reserve so that it does
> not perform the Orlov allocator resource group skipping in cases
> where we have an active reservation. Skipping an active rgrp is
> improper and gets our block reservation counts out of sync.
> 
> Regards,
> 
> Bob Peterson
> Red Hat File Systems
> 
> Signed-off-by: Bob Peterson <rpeterso@redhat.com> 
> ---
> diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
> index 99a6197..61a4314 100644
> --- a/fs/gfs2/rgrp.c
> +++ b/fs/gfs2/rgrp.c
> @@ -1811,13 +1811,14 @@ int gfs2_inplace_reserve(struct gfs2_inode *ip, u32 requested, u32 aflags)
>  	if (gfs2_rs_active(rs)) {
>  		begin = rs->rs_rbm.rgd;
>  		flags = 0; /* Yoda: Do or do not. There is no try */
> -	} else if (ip->i_rgd && rgrp_contains_block(ip->i_rgd, ip->i_goal)) {
> -		rs->rs_rbm.rgd = begin = ip->i_rgd;
>  	} else {
> -		rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
> +		if (ip->i_rgd && rgrp_contains_block(ip->i_rgd, ip->i_goal))
> +			rs->rs_rbm.rgd = begin = ip->i_rgd;
> +		else
> +			rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
> +		if (S_ISDIR(ip->i_inode.i_mode) && (aflags & GFS2_AF_ORLOV))
> +			skip = gfs2_orlov_skip(ip);
>  	}
> -	if (S_ISDIR(ip->i_inode.i_mode) && (aflags & GFS2_AF_ORLOV))
> -		skip = gfs2_orlov_skip(ip);
>  	if (rs->rs_rbm.rgd == NULL)
>  		return -EBADSLT;
>  




^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2012-11-12 18:41 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-11-12 18:31 [Cluster-devel] [GFS2 PATCH] GFS2: Don't skip rgrps for Orlov's sake if reservation is active Bob Peterson
2012-11-12 18:41 ` Steven Whitehouse

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).