This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, ivopts] Handle type-converted IVs better
- From: Roger Sayle <roger at eyesopen dot com>
- To: Adam Nemet <anemet at caviumnetworks dot com>
- Cc: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>, <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 21 Sep 2006 20:46:11 -0600 (MDT)
- Subject: Re: [PATCH, ivopts] Handle type-converted IVs better
Hi Adam,
On Thu, 21 Sep 2006, Adam Nemet wrote:
> > No where in the thread has it been mentioned that this is a regression
> > fix, so it isn't really suitable for mainline whilst we're in stage 3.
>
> I am sorry for not being explicit about it but as most ivopts
> miss-optimizations this is a code-quality regression from 3.4
> (pre-tree-ssa days).
I'll consider it for mainline, if you can describe/provide a testcase
where we've regressed and that is restored by your patch. I don't doubt
that you're right this is a regression, but I'm curious to see an actual
example with assembler from v3.4, mainline and mainline+patch. If it
can be made into a new dejagnu test for the testsuite, even better.
My apologies for the minor formality. There's no need to file a PR or
anything, but it needs a bit more supporting description.
Thanks in advance,
Roger
--