public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Benny Halevy <bhalevy@panasas.com>
To: bfields@fieldses.org
Cc: linux-nfs@vger.kernel.org, pnfs@linux-nfs.org,
	Benny Halevy <bhalevy@panasas.com>,
	Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Subject: [PATCH] nfsd41: renew_client must be called under the state lock
Date: Thu, 20 Aug 2009 03:21:56 +0300	[thread overview]
Message-ID: <1250727716-28107-1-git-send-email-bhalevy@panasas.com> (raw)

nfsd4_sequence() should not renew the client state if the session was not
found or if there was a bad slot.  This will also avoid dereferencing a
null session pointer.

FIXME: until we work out the state locking so we can use a
spin lock to protect the cl_lru we need to take the state_lock
to renew the client.

Signed-off-by: Benny Halevy <bhalevy@panasas.com>
[nfsd41: Do not renew state on error]
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
---

Bruce, I'm not whether it is too late but this patch fixes
a bug in 2.6.31-rc that we've hit in the last Bakeathon.

Benny

 fs/nfsd/nfs4state.c |   30 +++++++++++++++++++-----------
 1 files changed, 19 insertions(+), 11 deletions(-)

diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 1dedde9..3e7b20a 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -1436,12 +1436,16 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	spin_lock(&sessionid_lock);
 	status = nfserr_badsession;
 	session = find_in_sessionid_hashtbl(&seq->sessionid);
-	if (!session)
-		goto out;
+	if (!session) {
+		spin_unlock(&sessionid_lock);
+		goto err;
+	}
 
 	status = nfserr_badslot;
-	if (seq->slotid >= session->se_fchannel.maxreqs)
-		goto out;
+	if (seq->slotid >= session->se_fchannel.maxreqs) {
+		spin_unlock(&sessionid_lock);
+		goto err;
+	}
 
 	slot = &session->se_slots[seq->slotid];
 	dprintk("%s: slotid %d\n", __func__, seq->slotid);
@@ -1454,10 +1458,12 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 		 * for nfsd4_proc_compound processing */
 		status = nfsd4_replay_cache_entry(resp, seq);
 		cstate->status = nfserr_replay_cache;
-		goto replay_cache;
-	}
-	if (status)
 		goto out;
+	}
+	if (status) {
+		spin_unlock(&sessionid_lock);
+		goto err;
+	}
 
 	/* Success! bump slot seqid */
 	slot->sl_inuse = true;
@@ -1470,15 +1476,17 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	cstate->slot = slot;
 	cstate->session = session;
 
-replay_cache:
-	/* Renew the clientid on success and on replay.
-	 * Hold a session reference until done processing the compound:
+	/* Hold a session reference until done processing the compound:
 	 * nfsd4_put_session called only if the cstate slot is set.
 	 */
-	renew_client(session->se_client);
 	nfsd4_get_session(session);
 out:
 	spin_unlock(&sessionid_lock);
+	/* Renew the clientid on success and on replay */
+	nfs4_lock_state();
+	renew_client(session->se_client);
+	nfs4_unlock_state();
+err:
 	dprintk("%s: return %d\n", __func__, ntohl(status));
 	return status;
 }
-- 
1.6.4


             reply	other threads:[~2009-08-20  0:48 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-08-20  0:21 Benny Halevy [this message]
2009-08-24 16:29 ` [PATCH] nfsd41: renew_client must be called under the state lock J. Bruce Fields
2009-08-25  7:06   ` Benny Halevy
2009-08-25 13:18     ` J. Bruce Fields
2009-08-25 17:02       ` Benny Halevy
2009-08-25 21:59         ` J. Bruce Fields
2009-08-26  8:20           ` [pnfs] " Boaz Harrosh
2009-08-26 21:02             ` J. Bruce Fields
2009-08-27 16:37               ` Benny Halevy
2009-08-27 21:16                 ` J. Bruce Fields
2009-08-27 21:38                   ` J. Bruce Fields
  -- strict thread matches above, loose matches on Subject: below --
2009-06-19  2:01 Benny Halevy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1250727716-28107-1-git-send-email-bhalevy@panasas.com \
    --to=bhalevy@panasas.com \
    --cc=Ricardo.Labiaga@netapp.com \
    --cc=bfields@fieldses.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=pnfs@linux-nfs.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox