Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 29 Apr 2013 20:47:23 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: krb5-23 Format Failure in Latest Unstable Jumbo on MacOSX

On 29 Apr, 2013, at 20:40 , Erik Winkler <ewinkler@...ls.com> wrote:
> Testing the latest Jumbo on MacOS X shows an error for the macosx-x86-64 target when compiled for OpenMP.  See below:
> 
> ../run/john -test -format:krb5-23
> Benchmarking: Kerberos 5 db etype 23 rc4-hmac [32/64]... FAILED (cmp_all(1))
> 
> 
> The error does NOT occur when the OpenMP compile flag is disabled.
> 
> ../run/john -test -format:krb5-23
> Benchmarking: Kerberos 5 db etype 23 rc4-hmac [32/64]... DONE
> Raw:	538801 c/s real, 538801 c/s virtual
> 
> 
> Tested with Apple's gcc compiler (version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00)) and with gcc-4.7.2.

Thanks, this is known. I can't find any problem in the code and I have started to think it's actually a bug in OSX' Heimdal version of libkrb.

https://github.com/magnumripper/JohnTheRipper/issues/270

Actually, I haven't ever seen the problem when using --format=krb5-23, only when running a full --test. I get it with or without OMP but it's intermittent. When faced with the debugger, the coward bug crawled back under its stone until I gave up.

magnum

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.