On Tue, Oct 20, 2015 at 5:38 PM, Andre Vieira <andre.simoesdiasvie...@arm.com> wrote: > > > On 20/10/15 17:25, Ramana Radhakrishnan wrote: >> >> On Tue, Oct 20, 2015 at 4:52 PM, Andre Vieira >> <andre.simoesdiasvie...@arm.com> wrote: >>> >>> Hi, >>> >>> This patch addresses PR-67948 by changing the xor-and.c test, initially >>> written for a simplify-rtx pattern, to make it pass post r228661 (see >>> https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00676.html). This test no >>> longer triggered the simplify-rtx pattern it was written for prior to >>> r228661, though other optimizations did lead to the same assembly the >>> test >>> checked for. The optimization added with r228661 matches the pattern used >>> in >>> the test and optimizes it to a better and still valid sequence. Being >>> unable >>> to easily change the test to trigger the original simplify-rtx pattern, I >>> chose to change it to pass with the new produced assembly sequence. >>> >>> This is correct because the transformation is valid and it yields a more >>> efficient pattern. However, as I pointed out before this test doesn't >>> test >>> the optimization it originally was intended for. >>> >>> Tested by running regression tests for armv6. >>> >>> Is this OK to commit? >>> >> >> Missing Changelog - please remember to put the PR number in the >> Changelog in the correct format i.e PR testsuite/67948. Ok with that. >> >> I suspect that the simplify-rtx.c is much less likely to trigger given >> your match.pd change, but it will be fun proving that. > > Ideally, I'd like to prove the simplify-rtx.c couldn't be triggered anymore. > Though I wouldn't know where to begin with that. I did spend a tiny effort > trying to trigger it with a version before the match.pd change, but with no > success.
>> >> >> regards >> ramana >> >> >>> Thanks, >>> Andre >> >> > Here is the ChangeLog (with the PR), sorry for that! > > gcc/testsuite/ChangeLog > 2015-10-15 Andre Vieira <andre.simoesdiasvie...@arm.com> > > PR testsuite/67948 > * gcc.target/arm/xor-and.c: check for eor instead of orr. Applied with a minor tweak to the changelog message. regards Ramana