john-users - Re: clang benchmarks

Openwall GNU/*/Linux 3.0 - a small security-enhanced Linux distro for servers

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Thu, 20 May 2010 05:47:09 +0400
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: clang benchmarks

On Wed, May 19, 2010 at 03:03:04PM +0200, bartavelle@...quise.net wrote:
> Le 18/05/2010 23:48, Solar Designer a ?crit :
> >Can you please give this a try?  If this solves the problem, then I'd
> >like to commit this fix into the official JtR tree.
> 
> It seems to compile properly. I didn't test the resulting code though. 
> The svn version of clang I used (trunk 103935) didn't mind these 
> extensions and worked with your vanilla code.

Thanks.  Then I think it'd be better to check for specific versions of
Clang, but I was not easily able to determine when support for "Labels
as Values" was introduced into Clang.  It appears to have been in LLVM
for years:

http://blog.llvm.org/2010/01/address-of-label-and-indirect-branches.html

"LLVM has long supported this extension by lowering it to a "correct"
but extremely inefficient form.  New in LLVM 2.7 is IR support for taking
the address of a label and jumping to it later, which allows
implementing this extension much more efficiently."

Release notes for older versions of LLVM (even from 2007) already list
"Labels as Values" as supported (presumably inefficiently).

> The ubuntu version here didn't work:
> clang version 1.1 (branches/release_27)

Thanks for the info.  I would have expected this version to support
"Labels as Values".  Maybe there's a relevant bug in it.

I guess I'll need to try Clang out myself eventually.

Alexander