[Cluster-devel] [GFS2 PATCH] GFS2: Check rs_free with rd_rsspin protection

[Bob Peterson]

Hi,

For the last process to close a file opened for write, function
gfs2_rsqa_delete was deleting the file's inode's block reservation
out of the rgrp reservations tree. Then it was checking to make sure
rs_free was 0, but it was performing the check outside the protection
of rd_rsspin spin_lock. The rd_rsspin spin_lock protection is needed
to prevent a race between the process freeing the reservation and
another who is allocating a new set of blocks inside the same rgrp
for the same inode, thus changing its value.

Signed-off-by: Bob Peterson <rpeterso at redhat.com>

Regards,

Bob Peterson
Red Hat File Systems
---
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index fba38ca..86ccc015 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -658,6 +658,7 @@ void gfs2_rs_deltree(struct gfs2_blkreserv *rs)
 	if (rgd) {
 		spin_lock(&rgd->rd_rsspin);
 		__rs_deltree(rs);
+		BUG_ON(rs->rs_free);
 		spin_unlock(&rgd->rd_rsspin);
 	}
 }
@@ -671,10 +672,8 @@ void gfs2_rs_deltree(struct gfs2_blkreserv *rs)
 void gfs2_rsqa_delete(struct gfs2_inode *ip, atomic_t *wcount)
 {
 	down_write(&ip->i_rw_mutex);
-	if ((wcount == NULL) || (atomic_read(wcount) <= 1)) {
+	if ((wcount == NULL) || (atomic_read(wcount) <= 1))
 		gfs2_rs_deltree(&ip->i_res);
-		BUG_ON(ip->i_res.rs_free);
-	}
 	up_write(&ip->i_rw_mutex);
 	gfs2_qa_delete(ip, wcount);
 }