On 24 January 2017 at 17:55, Bernd Schmidt <bschm...@redhat.com> wrote: > On 01/24/2017 05:50 PM, Kyrill Tkachov wrote: >> >> >> Actually trying it out with an explicit -mcpu=cortex-a5 (so -O2 -S >> -mfpu=fp-armv8 -mcpu=cortex-a57 -mfloat-abi=hard) I get >> the test failing before and after the patch. The code generated is >> vcmp.f64 d0, d1 >> vmrs APSR_nzcv, FPSCR >> vmovvs.f64 d0, d1 >> bx lr >> >> whereas the desired (e.g. with -mcpu=cortex-a57) is: >> vcmp.f64 d0, d1 >> vmrs APSR_nzcv, FPSCR >> vselvs.f64 d0, d1, d0 >> bx lr > > > Yes, I've seen both of these generated with different options, but the patch > did not make a difference here either. > > For the moment I'll assume this was a false alarm, i.e. Christophe > misidentified the patch and something else went wrong. > Ha... the regression occurred between r 244818 and r 244816, and I read r 244816 ChangeLog too quickly and did not notice it was modifying ifcvt.c in addition to x86-only files.
So it's likely that it's your other patch for pr78634 that caused the regression I mentioned. Does it make more sense? Sorry for the probably wrong identification. It always takes some time for me to reproduce regressions manually because I can only keep logs/results of validations, the toolchains actually built are deleted once validation completes. Christophe > > Bernd
