On Fri, Sep 23, 2005 at 02:16:57PM +0000, Emmanuel Dreyfus wrote:
> Log Message:
> Correctly initialize NAT-T VID so that we don't free unallocated space
I don't understant what have been fixed....
> if (iph1->rmconf->nat_traversal)
> plist = isakmp_plist_append_natt_vids(plist, vid_natt);
> - else
> - vid_natt=NULL;
The "old way": declare the vid_natt, then: if nat_traversal, the first
thing which will be done by isakmp_plist_append_natt_vids() is to set
up to NULL all values.
If there is no nat_t, set up the first entry of the array to NULL
(which is enough to break the free loop at the end).
Why is this "better" to do the "new way" ????
On Fri, Sep 23, 2005 at 04:36:41PM +0200, VANHULLEBUS Yvan wrote:
> I don't understant what have been fixed....
A segmentation fault when my_identifier is badly set and NAT-T is on.
The call to isakmp_plist_append_natt_vids() which correctly initialize the
NAT VID may have been skipped in some situations (check the goto end
earlier in the code), leaving an uninitialized vid_natt to the code
that tries to free the VID array.
Now I run into another VID related SIGSEGV later in the code. Rontudjuuuu!
Get latest updates about Open Source Projects, Conferences and News.