Re: [PATCH] testsuite: Skip intrinsics test if arm

[Richard Earnshaw <Richard.Earnshaw@foss.arm.com>]

On 15/01/2023 17:06, Torbjorn SVENSSON via Gcc-patches wrote:
> 
> 
> On 2023-01-12 16:03, Richard Earnshaw wrote:
>>
>>
>> On 19/09/2022 17:16, Torbjörn SVENSSON via Gcc-patches wrote:
>>> In the test case, it's clearly written that intrinsics is not
>>> implemented on arm*. A simple xfail does not help since there are
>>> link error and that would cause an UNRESOLVED testcase rather than
>>> XFAIL.
>>> By chaning to dg-skip-if, the entire test case is omitted.
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>>     * gcc.target/aarch64/advsimd-intrinsics/vld1x2.c: Replace
>>>     dg-xfail-if with gd-skip-if.
>>
>> Sorry for the delay reviewing this, I missed it at the time.
>>
>> My problem with your suggested solution is that if these intrinsics 
>> are ever added this test will not automatically pick this up as it 
>> will have been disabled.  I presume from the comment (and the body of 
>> the test that contains an #ifdef for aarch64) that this is expected to 
>> be a temporary issue rather than something permanent.
>>
>> So IMO I think it is correct to leave this as unresolved because the 
>> test cannot be built due to an issue with the compiler.
> 
> This patch has already been merged after Kyrill reviewed it back in 
> September.
> 
> Without this change, the log would be filled with warnings about missing 
> types. Maybe we could add some check that will enable the test only if 
> the types are known?
> Would that mitigate your concern?
> 
> Attached is the log from vld1x2.c on Cortex-A7 with -mfloat-abi=hard 
> -mfpu=neon.
> 
> When I look at the result of a run, I only look at the test cases that 
> are either FAIL (obviously), XPASS and UNRESOLVED. All other test cases 
> are in a "good" state from what I can tell. If there are a lot of test 
> cases in the UNRESOLVED state, that are not yet implemented year after 
> year, it makes it harder to identify those test cases that are of 
> interest. Right or wrong, that's why I suggested to remove it for the 
> list of test cases that should be working.
> 
> Let me know what you think.

Ah, OK.  Somehow I'd misplaced v2 of the patch, which is the version 
that got approved.

R.

> 
> Kind regards,
> Torbjörn
> 
>>
>> R.
>>
>>>
>>> Co-Authored-By: Yvan ROUX  <yvan.roux@foss.st.com>
>>> Signed-off-by: Torbjörn SVENSSON  <torbjorn.svensson@foss.st.com>
>>> ---
>>>   gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vld1x2.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git 
>>> a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vld1x2.c 
>>> b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vld1x2.c
>>> index 92a139bc523..f933102be47 100644
>>> --- a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vld1x2.c
>>> +++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vld1x2.c
>>> @@ -1,6 +1,6 @@
>>>   /* We haven't implemented these intrinsics for arm yet.  */
>>> -/* { dg-xfail-if "" { arm*-*-* } } */
>>>   /* { dg-do run } */
>>> +/* { dg-skip-if "unsupported" { arm*-*-* } } */
>>>   /* { dg-options "-O3" } */
>>>   #include <arm_neon.h>