This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] completely hookize CONDITIONAL_REGISTER_USAGE
- From: Richard Henderson <rth at redhat dot com>
- To: Nathan Froyd <froydnj at codesourcery dot com>
- Cc: gcc-patches at gcc dot gnu dot org, nickc at redhat dot com, bernds at codesourcery dot com, dave dot anglin at nrc-cnrc dot gc dot ca
- Date: Fri, 19 Nov 2010 10:37:29 -0800
- Subject: Re: [PATCH] completely hookize CONDITIONAL_REGISTER_USAGE
- References: <20101118205530.GL24469@nightcrawler>
On 11/18/2010 12:55 PM, Nathan Froyd wrote:
> The patch below hookizes CONDITIONAL_REGISTER_USAGE and hookizes all
> targets. This is pretty straightforward, but a couple of targets stood
> out as a bit unusual:
> - pa had an unusual definition of CONDITIONAL_REGISTER_USAGE for 32-bit
> and 64-bit targets. I think I got the logic right; Dave, would you
> mind double-checking my merging?
> - m32r had some logic for SUBTARGET_CONDITIONAL_REGISTER_USAGE, but
> AFAICS, SUBTARGET_CONDITIONAL_REGISTER_USAGE was never defined, so I
> eliminated it. Nick, is that OK?
> - bfin used CONDITIONAL_REGISTER_USAGE to call a
> conditional_register_usage, which didn't actually do the bulk of the
> work, but initialized some global rtxes. I took the liberty of
> munging that function into the new bfin_conditional_register_usage.
> Bernd, does that seem OK?
> Tested on x86_64-unknown-linux-gnu. Will build some cross-compilers to
> verify those targets as well. OK?
Ok modulo JDA's hppa comments.