On 7/26/24 12:07 PM, Carl Love wrote:
> On 7/24/24 11:47 AM, Segher Boessenkool wrote:
>>>>> +/* { dg-do run { target { int128 } && { power10_hw } } } */
>>>> Everything power10 is int128 always.
>>> OK, so don't need the power10_hw. Changed to just int128 for the target:
>> No, the other way around: you cannot run the code on machines without
>> these (ISA 3.1) instructions!
>
> OK, made it { dg-do run { target power10_hw } } */
>>
>> But p10 always satisfies the int128 predicate. Although, hrm, how
>> about -m32 :-)
>
> If I test with -m64 I get 8 passes:
>
> make -j 1 && make check-gcc RUNTESTFLAGS="-v -v
> powerpc.exp=vec-shift-double-runnable-int128.c --target_board=unix'{-m64}'"
> # of expected passes 8
>
> If I test with -m32 I get unsupported test,
>
> make -j 1 && make check-gcc RUNTESTFLAGS="-v -v
> powerpc.exp=vec-shift-double-runnable-int128.c --target_board=unix'{-m32}'"
> # of unsupported tests 1
>
> Looking further into the output the checks say:
> Checking pattern "hppa*-*-hpux*" with powerpc64le-unknown-linux-gnu
> compiler exited with status 1
> output is:
> cc1: error: '-m32' not supported in this configuration^M
>
> check_cached_effective_target power10_hw_available: returning 0 for
> unix/-m32
> is-effective-target: power10_hw 0
>
> Looks like the power10_hw is sufficient to prevent the test from running.
> Don't need to explicitly check for int128 as well.
That's only because you're testing on LE where -m32 is a not supported
configuration. If you tested on BE, you'd probably see a FAIL for -m32.
I think your initial target tests with int128 and power10_hw is actually
correct. We have an internal Power10 partition (ltcd97-lp7) running a
BE distro you can test this on.
Peter
