Re: Turning on mem leak detection

In nessusd 2.0.12 a leak can be fixed with the following patch. I was 
beating my head against the wall trying to figure out why cache_add_name() 
was showing up in my dmalloc logs with two different line numbers. I was 
staring at libnessus/arglists.c for a long time trying to figure out how two 
mallocs could possibly be happening. I eventually realized there was 
probably another function with the same name. Sure enough 
nessusd/pluginscheduler.c has the same function. Seems like the library and 
program could benefit here from some code reuse. Maybe modify the functions 
to take the cache as a parameter so there could still be separate caches, 
but the functions would remain the same for both caches. Anyway, the 
following patch fixes the leak.

BTW (Renaud), I'll probably have more of these as I'm trying to stabilize 
the growth of a kept alive nessusd session. Is there some other way you want 
me to send these to you?

Thanks,
Lance

==== nessusd/pluginscheduler.c#1 - nessusd/pluginscheduler.c ====
@@ -182,6 +183,8 @@
     nc->prev->next = nc->next;
    else
     cache[h].next = nc->next;
+
+    efree(&nc);
  }
}



_______________________________________________
Nessus mailing list
Nessus@list.nessus.org
http://mail.nessus.org/mailman/listinfo/nessus