xfs: active inodes stat is broken
authorDave Chinner <dchinner@redhat.com>
Wed, 3 Dec 2014 22:42:40 +0000 (09:42 +1100)
committerDave Chinner <david@fromorbit.com>
Wed, 3 Dec 2014 22:42:40 +0000 (09:42 +1100)
vn_active only ever gets decremented, so it has a very large
negative number.  Make it track the inode count we currently have
allocated properly so we can easily track the size of the inode
cache via tools like PCP.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
fs/xfs/xfs_icache.c
fs/xfs/xfs_super.c

index b45f7b2..f7a742e 100644 (file)
@@ -65,6 +65,7 @@ xfs_inode_alloc(
                return NULL;
        }
 
+       XFS_STATS_INC(vn_active);
        ASSERT(atomic_read(&ip->i_pincount) == 0);
        ASSERT(!spin_is_locked(&ip->i_flags_lock));
        ASSERT(!xfs_isiflocked(ip));
@@ -130,6 +131,7 @@ xfs_inode_free(
        /* asserts to verify all state is correct here */
        ASSERT(atomic_read(&ip->i_pincount) == 0);
        ASSERT(!xfs_isiflocked(ip));
+       XFS_STATS_DEC(vn_active);
 
        call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback);
 }
index a0f58d9..45fa5b5 100644 (file)
@@ -1008,7 +1008,6 @@ xfs_fs_evict_inode(
        clear_inode(inode);
        XFS_STATS_INC(vn_rele);
        XFS_STATS_INC(vn_remove);
-       XFS_STATS_DEC(vn_active);
 
        xfs_inactive(ip);
 }