Re: [PATCH][ARM] Disable store_minmaxsi pattern for arm_restrict_i

2014-10-01 Thread Kyrill Tkachov
Ping. https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00154.html Thanks, Kyrill On 02/09/14 16:34, Kyrill Tkachov wrote: Hi all, The store_minmaxsi produces a cmp + ite + 2 conditional stores and is thus inappropriate when the ARMv8-A IT block rules are in place. Previously we had disabled it

Re: [PATCH][ARM] Disable store_minmaxsi pattern for arm_restrict_i

2014-10-01 Thread Richard Earnshaw
On 02/09/14 16:34, Kyrill Tkachov wrote: Hi all, The store_minmaxsi produces a cmp + ite + 2 conditional stores and is thus inappropriate when the ARMv8-A IT block rules are in place. Previously we had disabled it for speed optimisations, but it should be disabled completely when

[PATCH][ARM] Disable store_minmaxsi pattern for arm_restrict_i

2014-09-02 Thread Kyrill Tkachov
Hi all, The store_minmaxsi produces a cmp + ite + 2 conditional stores and is thus inappropriate when the ARMv8-A IT block rules are in place. Previously we had disabled it for speed optimisations, but it should be disabled completely when -mrestrict-it is in effect. Ok for trunk and 4.9?