[ ssic-linux-Bugs-2719607 ] memory leak in ipcname_gettotal() path

Bugs item #2719607, was opened at 2009-03-28 16:47
Message generated for change (Settings changed) made by rogertsang
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=405834&aid=2719607&group_id=32541

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: IPC
Group: default
>Status: Closed
Resolution: Fixed
Priority: 5
Private: No
Submitted By: Roger Tsang (rogertsang)
Assigned to: John Hughes (hughesj)
Summary: memory leak in ipcname_gettotal() path

Initial Comment:
https://sourceforge.net/mailarchive/forum.php?thread_name=49C8C9D9.6000204%40Calva.COM&forum_name=ssic-linux-devel

----------------------------------------------------------------------

Comment By: Roger Tsang (rogertsang)
Date: 2009-10-26 23:49

Message:
checked-in

----------------------------------------------------------------------

Comment By: Roger Tsang (rogertsang)
Date: 2009-04-28 09:56

Message:
node_id_pairs is a vector

----------------------------------------------------------------------

Comment By: Roger Tsang (rogertsang)
Date: 2009-04-28 09:54

Message:
There is also a memory corruption bug in this path.
When doing RPC the length of the struct node_id_pairs buffer is
incorrect.

--- linux.orig/cluster/ssi/ipc/namesvr_clnt.c
+++ linux/cluster/ssi/ipc/namesvr_clnt.c
@@ -188,7 +188,7 @@ again:
 {
        clusternode_t   server_node;
-       int             status, rval, len = 0;
+       int rval;
        extern clms_key_svc_t ipc_key_service;
 again:
        server_node = name_server_node;
@@ -203,13 +203,12 @@ again:
                rval = ipcname_gettotal(service, *node_id_pairs, sz);
        } else {
                ssi_procstate_t pstate;
+               int status, count = *sz;

-               if (*sz > 0)
-                       len = *sz * sizeof(struct ssi_nodeid_pair);
                ssi_procstate_get(&pstate);

-                                               node_id_pairs, &len, sz);
+                                               node_id_pairs, &count,
sz);
                if (status == -EAGAIN || status == -EREMOTE) {
                        /* Server is doing failover,dying, or isn't ready
yet.*/
                        clms_waitfor_key_service(0);
Index: linux/cluster/ssi/ipc/namesvr_svr.c
===================================================================
@@ -88,34 +88,33 @@ ripc_ipcname_getid(clusternode_t *node, 
  */
 void
-               int *len, int *sz)
+               ssi_procstate_t *pstate,
+               struct ssi_nodeid_pair **node_id_pairs, int *count,
+               int *sz)
 {
        ssi_procstate_t save_pstate;
-       int count = *sz;

        if (ipcname_failover_flag) {
                *rval = -EAGAIN;
                return;
        }

-       *len = 0;
        ssi_procstate_get(&save_pstate);
        ssi_procstate_set(pstate);
-       if (count > 0) {
-               if (*node_id_pairs == NULL) {
-                       *sz = 0;
+       if (*count > 0) {
+               if (!*node_id_pairs) {
+                       *count = *sz = 0;
                        goto done;
                }
        }

        *rval = ipcname_gettotal(service, *node_id_pairs, sz);
-       if (count > 0) {
-               if (count > *sz)
-                       count = *sz;
-               *len = count * sizeof(**node_id_pairs);
-       }
+       if (!*rval) {
+               if (*count && *sz < *count)
+                       *count = *sz;
+       } else
+               *count = 0;
 done:
        ssi_procstate_set(&save_pstate);
 }

----------------------------------------------------------------------

Comment By: Roger Tsang (rogertsang)
Date: 2009-03-28 16:49

Message:
fix attached

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=405834&aid=2719607&group_id=32541

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev