Re: [ld/testsuite] Skip "Mixing PIC and non-PIC" testcase on ARM/AArch64 if when no -fpie or -fPIE

[Jiong Wang <jiong dot wang at foss dot arm dot com>]

H.J. Lu writes:

> On Wed, Mar 2, 2016 at 10:04 AM, Jiong Wang <jiong.wang@foss.arm.com> wrote:
>>
>>
>> On 02/03/16 17:40, H.J. Lu wrote:
>>>
>>> On Wed, Mar 2, 2016 at 9:36 AM, Jiong Wang <jiong.wang@foss.arm.com>
>>> wrote:
>>>>
>>>>
>>>> On 02/03/16 16:40, H.J. Lu wrote:
>>>>>
>>>>> On Wed, Mar 2, 2016 at 8:08 AM, Jiong Wang <jiong.wang@foss.arm.com>
>>>>> wrote:
>>>>>>>>>
>>>>>>>>> Have you looked at the testcase I added? Are there anything
>>>>>>>>> which are target specific?
>>>>>>>>
>>>>>>>>
>>>>>>>> I do have looked at the testcase, they do be purely C code.
>>>>>>>>
>>>>>>>> If the convention of generic is syntax generic instead of both syntax
>>>>>>>> and
>>>>>>>> sematics, I don't have further comment on this.
>>>>>>>>
>>>>>>>> Anyway, attached patch skips the non-pie version "Mixing PIC and
>>>>>>>> non-PIC"
>>>>>>>> testcase.
>>>>>>>>
>>>>>>>> Not sure if it's trivial enough to qualify obvious, so OK for master
>>>>>>>> branch?
>>>>>>>>
>>>>>>>> 2016-03-02  Jiong Wang  <jiong.wang@arm.com>
>>>>>>>>
>>>>>>>> ld/testsuite/
>>>>>>>>
>>>>>>>>      * ld-elf/shared.exp (mix_pic_and_non_pic): Only run on ARM and
>>>>>>>> AArch64
>>>>>>>>        when -fPIE or -fpie specified.
>>>>>>>
>>>>>>> That is wrong.  If you don't want to see FAIL, you can skip the whole.
>>>>>>> mix_pic_and_non_pic.  Please don't modify mix_pic_and_non_pic.
>>>>>>
>>>>>>
>>>>>> Then, how about the updated version?
>>>>>>
>>>>>> 2016-03-02  Jiong Wang  <jiong.wang@arm.com>
>>>>>>
>>>>>> ld/testsuite/
>>>>>>
>>>>>>      * ld-elf/shared.exp: Skip "Mixing PIC and non-PIC" on ARM and
>>>>>> AArch64
>>>>>>      if -fpie or -fPIE not specified.
>>>>>>
>>>>> Please use setup_xfail instead of skipping.
>>>>
>>>>
>>>> setup_xfail will actually update one global xfail related flag which
>>>> affect
>>>> the
>>>> pass/fail after it. so you must put "setup_xfail" accurately before the
>>>> fail
>>>> you
>>>> want to xfail or the pass you want to xpass.
>>>
>>> There is
>>>
>>> ld-elf/comm-data.exp:setup_xfail "arm*-*-*" "ld/13802"
>>>
>>> How does it work?
>>
>>
>> I think that's because it's using run_ld_link_test, while your
>> "mix_pic_and_non_pic"
>> is using run_cc_link_test.
>>
>> From ld.log, those compile steps in run_cc_link_test is treated as a
>> seperate tesing, thus
>> an internal PASS is generated. So if you put setup_xfail outside
>> mix_pic_and_non_pic, then
>> the flag setup will be wrongly taken by the later PASS and turns it into
>> XPASS.
>
> Have you tried?  I tried:
>
>
> diff --git a/ld/testsuite/ld-elf/shared.exp b/ld/testsuite/ld-elf/shared.exp
> index 0624f0d..69b9e38 100644
> --- a/ld/testsuite/ld-elf/shared.exp
> +++ b/ld/testsuite/ld-elf/shared.exp
> @@ -649,5 +649,6 @@ proc mix_pic_and_non_pic {cflags ldflags} {
>      }
>  }
>
> +setup_xfail "*-*-*"
>  mix_pic_and_non_pic "" ""
>  mix_pic_and_non_pic "-fPIE" "-pie"
>
> I only got
>
> XPASS: Build libpr19719a.so

That's exactly the problem I said.  The xfail flag has taken by the
following check, no?

With the use of "setup_xfail", what we are expecting is:

  PASS: Build libpr19719a.so
  XFAIL: Mixing PIC and non-PIC

not:

  XPASS: Build libpr19719a.so
  FAIL: Mixing PIC and non-PIC

-- 
Regards,
Jiong