Re: [GOLD] add new method for computing a build ID

[Mike Frysinger <vapier at gentoo dot org>]

On Wednesday 03 October 2012 15:49:38 David Miller wrote:
> From: Mike Frysinger <vapier@gentoo.org>
> > On Wednesday 03 October 2012 15:25:44 David Miller wrote:
> >> From: gpike@chromium.org
> >> > Motivation for the change: computing a build ID is currently a
> >> > sequential bottleneck. E.g., in the example I mentioned in my email
> >> > yesterday, a 24.5s link was spending 5.2s on the build ID computation.
> >> 
> >> Have you tried linking with OpenSSL and using their optimized MD5 and
> >> SHA1 variants?  I bet it makes a huge difference, some cpus even have
> >> SHA1 and MD5 instructions.
> > 
> > last i heard, openssl's license is incompatible with the GPL
> 
> I'm just saying, if the SHA1 and MD5 code we have in libiberty or
> wherever unnecessarily slow that guides the solution.  I'm not saying
> we have to use OpenSSL's code for the fix.
> 
> You don't even know if that's what the issue is.
> 
> It's a real scarecrow to bring up licensing issues when I'm merely
> trying to point out that we should first determine if we simply have
> an unreasonable slow set of crypto hashing functions.

i didn't say "don't do speed tests".  i was merely pointing out that openssl 
isn't something binutils may rely on.
-mike

Attachment: signature.asc
Description: This is a digitally signed message part.