This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, rs6000] fix vsxcopy test for power9
- From: Will Schmidt <will_schmidt at vnet dot ibm dot com>
- To: Segher Boessenkool <segher at kernel dot crashing dot org>
- Cc: gcc-patches at gcc dot gnu dot org, David Edelsohn <dje dot gcc at gmail dot com>, Bill Schmidt <wschmidt at linux dot vnet dot ibm dot com>
- Date: Wed, 07 Feb 2018 16:59:23 -0600
- Subject: Re: [PATCH, rs6000] fix vsxcopy test for power9
- Authentication-results: sourceware.org; auth=none
- References: <1518022786.11602.266.camel@brimstone.rchland.ibm.com> <20180207180231.GF21977@gate.crashing.org>
- Reply-to: will_schmidt at vnet dot ibm dot com
On Wed, 2018-02-07 at 12:02 -0600, Segher Boessenkool wrote:
> Hi,
>
> On Wed, Feb 07, 2018 at 10:59:46AM -0600, Will Schmidt wrote:
> > Noted during review of test results on P9. The vsxcopy.c test is looking
> > for lxvd2x, stxvd2x instructions in generated code. For P9 targets, we will
> > instead generate lxv, stxv instructions.
> > Thus, update the test to handle that codegen as well.
> > Sniff-tested on P9.
> > OK for trunk?
>
> There are many insns that start with "lxv" (or "stxv"); you may want to
> use \m and \M, like:
>
> > -/* { dg-final { scan-assembler "lxvd2x" } } */
> > -/* { dg-final { scan-assembler "stxvd2x" } } */
> > +/* { dg-final { scan-assembler "lxvd2x|lxv" } } */
> > +/* { dg-final { scan-assembler "stxvd2x|stxv" } } */
>
> /* { dg-final { scan-assembler {\m(lxvd2x|lxv)\M} } } */
> /* { dg-final { scan-assembler {\m(stxvd2x|stxv)\M} } } */
>
> ("lxv" without anything more will already also match "lxvd2x", etc.)
> Okay with such a change (if it works ;-) )
updated sniff-test suggests to me that it does. :-)
I've used the \m\M bits before, but never surrounding an (a|b)
construct. something learned. :-)
Thanks,
-Will
>
>
> Segher
>