Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 16 Dec 2011 19:28:29 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: 1.7.9-jumbo

Jim -

On Thu, Dec 15, 2011 at 05:07:46PM -0600, jfoug wrote:
> Did you want me to try to localize the -O0 in john.c to just cygwin builds?
> I had not thought of that, until I was writing the reply.  It makes no
> difference at all on speed, and only a few bytes on size.  But for whatever
> reason, it places the code properly, or aligns the stack, or properly builds
> a 'something' that keeps those builds from coring.  The strange thing, is
> you can test all day long, and they work.  However, when you try to actually
> 'run' data (reading from a file of passwords, etc), it dies.
> 
> I may be the only one experiencing this issue, so it could be that I simply
> keep patching the official make (which is what I was doing before).  But I
> will work with whatever you want it to be.

Localizing -O0 for john.c to just the Cygwin build is not very
important.  It'd be better to find and fix the actual bug, eliminating
the need for the -O0 even for Cygwin.

Can you at least identify the instruction (and nearby instruction
sequence) where the process crashes?  And see why that instruction
crashes (what type of fault it is - e.g., NULL pointer dereference,
out of bounds access, unaligned SSE2 access, or something else).

Thanks,

Alexander

Powered by blists - more mailing lists

Your e-mail address:

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