Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <op.x3isi0iqzz6j51@1pqhgq1.dtn.com>
Date: Mon, 17 Aug 2015 11:10:02 -0500
From: JimF <jfoug@....net>
To: john-dev@...ts.openwall.com
Subject: Re: FMT_OMP_BAD

On Mon, 17 Aug 2015 11:01:46 -0500, Solar Designer <solar@...nwall.com>  
wrote:

> Kai, magnum -
> ....
> No, that's redundant.  Some formats do it, but they should actually be
> changed to use the simpler approach ...
>
> ... Unless this becomes wrong when we introduce FAST_FORMATS_OMP checks.
> For example, rawMD5_fmt_plug.c has:
>
> #include "formats.h"
>
> #if !FAST_FORMATS_OMP
> #undef _OPENMP
> #endif
>
> [...]
>
> #ifdef _OPENMP
>                 FMT_OMP | FMT_OMP_BAD |
> #endif
>

If that is the logic wanted for 'fast formats omp', then simply make that  
code a header, that is included last (or right before memdbg.h).  Within  
that header you would have things like:


#if !FAST_FORMATS_OMP
#undef _OPENMP
#undef FMT_OMP
#define FMT_OMP 0
#undef FMT_BAD_OMP
#define FMT_OMP_BAD 0
#endif


Then simply include that header file in formats where you want that  
logic.  Also, it may be that this header should be included FIRST, since  
the _OPENMP logic being set can make changes in certain headers.

It if ends up being the first thing, then the header may only need to  
contain:

#if !FAST_FORMATS_OMP
#undef _OPENMP
#endif

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.