Re: [checker] nmap: possible memory leak in nsi_new2

[Fyodor <fyodor () insecure org>]

On Tue, May 02, 2006 at 01:25:26AM -0700, Ted Kremenek wrote:
> Hello again,
>
> My tool flagged another potential memory leak in nsi_new2 when the  
> call to dup() fails:

Verified and fixed:

Index: nsock_iod.c
===================================================================
--- nsock_iod.c (revision 2656)
+++ nsock_iod.c (working copy)
@@ -95,7 +95,10 @@
     nsi->state = NSIOD_STATE_INITIAL;
   } else {
     nsi->sd = dup(sd);
-    if (nsi->sd == -1) return NULL;
+    if (nsi->sd == -1) {
+      free(nsi);
+      return NULL;
+    }
     nsock_unblock_socket(nsi->sd);
     nsi->state = NSIOD_STATE_UNKNOWN;
   }


(Freeing nsi is OK, even if it came from the nsp->free_iods list
rather than the safe_malloc directly).

Thanks,
Fyodor


_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev