Fix potential crash when entire span request is received.

The variable index used in this scenario for accessing the dahdi_pvts was
wrong and was most likely copied from the several other places it is used
correctly.

(closes issue #15998)
Reported by: tsearle
Patches:
      dahdi_reset_crash.patch uploaded by tsearle (license 373)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@222393 65c4cc65-6c06-0410-ace0-fbb531ad65f3

diff --git a/channels/chan_dahdi.c b/channels/chan_dahdi.c
index f3a27d345c32891a8b9a5b015ecf35fcece0cbf8..100ce5f71a0ce0f94c93def0e2637c53b70c3e8f 100644 (file)
--- a/channels/chan_dahdi.c
+++ b/channels/chan_dahdi.c
@@ -9180,8 +9180,8 @@ static void *pri_dchannel(void *vpri)
                                                                pri_destroycall(pri->pri, pri->pvts[x]->call);
                                                                pri->pvts[x]->call = NULL;
                                                        }
-                                                       if (pri->pvts[chanpos]->realcall) 
-                                                               pri_hangup_all(pri->pvts[chanpos]->realcall, pri);
+                                                       if (pri->pvts[x]->realcall) 
+                                                               pri_hangup_all(pri->pvts[x]->realcall, pri);
                                                        else if (pri->pvts[x]->owner)
                                                                pri->pvts[x]->owner->_softhangup |= AST_SOFTHANGUP_DEV;
                                                        ast_mutex_unlock(&pri->pvts[x]->lock);