Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-20 Thread Segher Boessenkool
Hi! On Wed, May 20, 2020 at 10:56:21AM -0700, Douglas B Rupp wrote: > >So, which spec are those 4 lines above from? > From -with-specs= in the configure command.  We use that > capability extensively for VxWorks. static const char *const driver_self_specs[] = { "%{fdump-final-insns:-fdump-fina

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-20 Thread Douglas B Rupp
Backporting the DRIVER_SELF_SPECS change fixed all the mdejagnu-cpu related tests. Thanks for your help on this. Consider this patch request withdrawn.  I'll resubmit one for the remaining failing tests for discussion. --Doug

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-20 Thread Douglas B Rupp
On 5/20/20 9:46 AM, Segher Boessenkool wrote: So, which spec are those 4 lines above from? From -with-specs= in the configure command.  We use that capability extensively for VxWorks. We have as DRIVER_SELF_SPECS (see rs6000.h) #define DRIVER_SELF_SPECS \ "%{mdejagnu-cpu=*: %Interesting.

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-20 Thread Segher Boessenkool
Hi~ On Tue, May 19, 2020 at 03:20:37PM -0700, Douglas B Rupp wrote: > On 5/19/20 3:03 PM, Segher Boessenkool wrote: > >This is a compile test, so it does not matter at all what the kernel is > >doing or not doing. And the -mcpu=power8 flag should enable altivec. > >Apparently it doesn't for you,

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Douglas B Rupp
Apologies on this test, the problem is -mdejagnu-cpu doesn't get matched by the spec logic If I pass -mcpu=power8, then indeed the test passes fine. I'm not sure how to fix this cleanly.

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Douglas B Rupp
On 5/19/20 3:03 PM, Segher Boessenkool wrote: This is a compile test, so it does not matter at all what the kernel is doing or not doing. And the -mcpu=power8 flag should enable altivec. Apparently it doesn't for you, but why is that? It's likely the specs file we use for the kernel we are

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Segher Boessenkool
On Tue, May 19, 2020 at 02:41:45PM -0700, Douglas B Rupp wrote: > > On 5/19/20 1:53 PM, Segher Boessenkool wrote: > >How can altivec not be supported when you use -mcpu=power8? > > > For the test in question: builtins-1-be-folded.c: > > The VxWorks header file altivec.h is included by builtins-1.

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Douglas B Rupp
On 5/19/20 1:53 PM, Segher Boessenkool wrote: How can altivec not be supported when you use -mcpu=power8? For the test in question: builtins-1-be-folded.c: The VxWorks header file altivec.h is included by builtins-1.fold.h and the error is: altivec.h:34:2: error: #error Use the "-maltivec"

Re: [PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Segher Boessenkool
Hi! On Tue, May 19, 2020 at 12:29:40PM -0700, Douglas B Rupp wrote: > The proposed patch adds some checking for vsx and altivec being > supported on several powerpc tests. > > For vxworks, we have to spec particular cpus, so these tests fail when > they should show be N/A. > diff --git gcc/tes

[PATCH] Tweak some powerpc tests for altivec_ok and vsx_ok

2020-05-19 Thread Douglas B Rupp
Greetings, The proposed patch adds some checking for vsx and altivec being supported on several powerpc tests. For vxworks, we have to spec particular cpus, so these tests fail when they should show be N/A. --Douglas Rupp, AdaCore diff --git gcc/testsuite/gcc.target/powerpc/builtins-1-b