Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 12 Feb 2013 10:45:04 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: formats failing test on big-endian

On 12 Feb, 2013, at 2:33 , magnum <john.magnum@...hmail.com> wrote:
> On 11 Feb, 2013, at 19:36 , magnum <john.magnum@...hmail.com> wrote:
> On 9 Feb, 2013, at 11:44 , magnum <john.magnum@...hmail.com> wrote:
>>> All Unicode problems are fixed. They are now fixed in a canonical way so there should be very few such bugs in the future. All unicode functions work with UTF16-LE regardless of host arch. That was always the idea but some fixes had been put backwards over time.
>>> 
>>> I fixed some other things too but these remains:
>> ...
>> 
>> All issues are fixed now, except openvms. That format still segfaults on my Sparc.
>> 
>> Disabling openvms, all formats pass self-test with and without any --encoding. With just one format left to fix, I hope someone can nail that one too!
> 
> I got a new issue with ppc32. The ssh format crashes with double free (or corruption). I don't seem to be able to get any debug info from the backtrace (wtf? I just can't get any). There's code in get_salt() that doesn't look right to me (openssl free functions called twice) but commenting it out does not help so maybe they can be used that way safely.


It sure looks like OpenSSL:
*** glibc detected *** ../run/john: double free or corruption (!prev): 0x107eb770 ***
======= Backtrace: =========
/lib/powerpc-linux-gnu/libc.so.6(+0x86ef4)[0xf695ef4]
/lib/powerpc-linux-gnu/libc.so.6(cfree+0x8c)[0xf69b4cc]
/usr/lib/powerpc-linux-gnu/libcrypto.so.1.0.0(CRYPTO_free+0x5c)[0xfdf4794]
/usr/lib/powerpc-linux-gnu/libcrypto.so.1.0.0(EVP_MD_CTX_cleanup+0xf8)[0xfe7cf94]
/usr/lib/powerpc-linux-gnu/libcrypto.so.1.0.0(EVP_BytesToKey+0x19c)[0xfe7ee28]
../run/john[0x100582fc]


We also have these failing BE, when enabled:

Benchmarking: Mozilla SHA-1 3DES [32/32]... FAILED (cmp_all(1))
Benchmarking: Kerberos 5 db etype 23 rc4-hmac [32/32]... FAILED (cmp_all(1))
Benchmarking: Kerberos 5 db etype 18 aes256-cts-hmac-sha1-96 [32/32]... FAILED (cmp_all(1))


magnum

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ