From: Derrick Brashear <shadow@dementia.org>
Date: Thu, 28 Oct 2004 19:09:04 +0000 (+0000)
Subject: STABLE14-drop-h-lock-around-putconn-20041027
X-Git-Tag: openafs-devel-1_3_74~22
X-Git-Url: https://git.michaelhowe.org/gitweb/?a=commitdiff_plain;h=46239bd78681bb8b51237800f425691d5665835e;p=packages%2Fo%2Fopenafs.git

STABLE14-drop-h-lock-around-putconn-20041027


the problem with rx_PutConnection is h_FreeConnection uses the host global lock

====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================

the problem with rx_PutConnection is h_FreeConnection uses the host global lock
but it shouldn't matter (callback conn should be a client not server conn and
thus not h_FreeConnection on cleanup)

further, we should free the conns we GetConnection'd not the ones still in
the host struct at the end.


(cherry picked from commit 2fbfef7f9157aad571307564faae1ce7aae4e3be)
---

diff --git a/src/viced/callback.c b/src/viced/callback.c
index 90d684d72..fc45bb379 100644
--- a/src/viced/callback.c
+++ b/src/viced/callback.c
@@ -855,11 +855,20 @@ MultiBreakCallBack_r(struct cbstruct cba[], int ncbas,
 	struct host *hp;
 	hp = cba[i].hp;
 	if (hp && xhost != hp) {
-	    rx_PutConnection(hp->callback_rxcon);
 	    h_Release_r(hp);
 	}
     }
 
+    /* H_UNLOCK around this so h_FreeConnection does not deadlock.
+       h_FreeConnection should *never* be called on a callback connection,
+       but on 10/27/04 a deadlock occurred where it was, when we know why,
+       this should be reverted. -- shadow */
+    H_UNLOCK;
+    for (i = 0; i < j; i++) {
+	rx_PutConnection(conns[i]);
+    }
+    H_LOCK;
+
     return;
 }