[PATCH] Fix PR rtl-optimization/27661 (reload problem)

Roger Sayle roger@eyesopen.com
Tue May 23 15:59:00 GMT 2006


Hi Ulrich,

On Tue, 23 May 2006, Ulrich Weigand wrote:
> ChangeLog:
>
> 	PR rtl-optimization/27661
> 	* reload.c (find_reloads): When reloading a VOIDmode constant
> 	as address due to an EXTRA_MEMORY_CONSTRAINT or 'o' constraint,
> 	use Pmode as mode of the reload register.
>
> testsuite/ChangeLog:
>
> 	PR rtl-optimization/27661
> 	* gcc.dg/pr27661.c: New test case.

The patch itself looks reasonable, but are you able to test it on
anything more than the s390*?  Once its been on mainline for a while
without problems, it looks to be safe enough for the branches, but
I was wondering if we could at least test it on a primary platform
or two, before committing to mainline.  I appreciate that the change
should only affect code that currently fails, but reload patches
make me nervous, especially during stage3.

I'll preapprove this patch for mainline, once its had a little more
testing.

Thanks for fixing this.

Roger
--



More information about the Gcc-patches mailing list