On Aug 30, 2009, at 12:59 PM, Trond Myklebust wrote:
On Sun, 2009-08-30 at 12:42 -0400, Christoph Hellwig wrote:
On Sun, Aug 30, 2009 at 12:34:33PM -0400, Chuck Lever wrote:
+static void nfs_kill_super(struct super_block *sb)
{
+ struct nfs_server *server = NFS_SB(sb);
+
+ dprintk("--> %s\n", __func__);
+
+#ifdef CONFIG_NFS_V4
+ if (server->nfs_client->rpc_ops->version == 4) {
+ nfs_super_return_all_delegations(sb);
+ nfs4_renewd_prepare_shutdown(server);
+ }
+#endif /* CONFIG_NFS_V4 */
bdi_unregister(&server->backing_dev_info);
This was previously not done for nfs4. If it is a bug-fixed that
should be documented in the patch description, and if not it needs
to be changed.
It has always been done, but it was in a separate function
(nfs4_kill_super()). I don't really see what we gain by inlining it
into
nfs_kill_super..
Actually nfs4_kill_super didn't unregister the bdi.
Also, I'm concerned about the growth of "if (version == X)" type
constructs. We shouldn't be looking at the version number in order to
figure out whether or not we're holding delegations, or whether or
not a
particular daemon thread is running.
AFAICS, in this case it should in any case be safe to call
nfs_super_return_all_delegations() (as long as CONFIG_NFS_V4 is
defined
- that we might want to fix). Ditto for
nfs4_renewd_prepare_shutdown().
That's all fine, but as I said in the cover letter, this patch may be
dropped since the file system is still nfs4 under the covers, so I
think nfs4_kill_super would be invoked anyway for "-t nfs -o vers=4".
Cheers
Trond
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
