ANNOUNCEMENT: Live Wireshark University & Allegro Packets online APAC Wireshark Training Session
April 17th, 2024 | 14:30-16:00 SGT (UTC+8) | Online

Wireshark-dev: Re: [Wireshark-dev] [Wireshark-bugs] [Bug 7814] Buildbot crash output: fuzz-2012

From: Evan Huus <eapache@xxxxxxxxx>
Date: Wed, 10 Oct 2012 07:24:54 -0400
On Wed, Oct 10, 2012 at 7:13 AM, Martin Mathieson
<martin.r.mathieson@xxxxxxxxxxxxxx> wrote:
>
>
> On Wed, Oct 10, 2012 at 5:19 AM, Jakub Zawadzki <darkjames-ws@xxxxxxxxxxxx>
> wrote:
>>
>> On Tue, Oct 09, 2012 at 10:53:41PM -0400, Martin Mathieson wrote:
>> > I am getting the same assertion, for every file that I try
>> > reload/refilter.
>>
>> Can you get errno number for me?
>>
>
> 12 (Cannot Allocate Memory?)

Probably caused by, from the man page for mprotect:
"Addresses in the range [addr, addr+len-1] are invalid for the address
space of the process, or specify one or more pages that are not
mapped."

Could this be caused by trying to mprotect the same region of memory
twice? That would never happen in the old allocator, but is certainly
a possibility now.

>>
>> > Is there a fix in the works?
>>
>> Nope, but I submitted alternative patch in
>>  [Bug 7775] Wireshark leaks memory when selecting packets
>>
>> > In the meantime, could someone advise which files to rollback to which
>> > versions so that trunk is again usable?
>>
>> r45388 epan/emem.[ch] epan/epan.c epan_dissect.h
>
>
> Thanks.  If you can't reproduce it there, I'm happy to try patches and let
> you know what happens.

I committed a work-around in revision 45444 that should fix the bug
(potentially at a small speed cost) until I have a chance to figure
out a proper solution. I'm unfortunately short on time for a few days,
so if someone else wants to take a crack at it feel free.

Evan