[compiler-rt] r304673 - Mark sancov test as unsupported on Darwin

[Kostya Serebryany via llvm-commits]

On Tue, Jun 13, 2017 at 6:00 PM, Vedant Kumar via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

>
> On Jun 13, 2017, at 5:58 PM, Vedant Kumar via llvm-commits <
> llvm-commits at lists.llvm.org> wrote:
>
>
> On Jun 13, 2017, at 5:53 PM, George Karpenkov <ekarpenkov at apple.com>
> wrote:
>
> I’m confused, all libfuzzer tests pass on my machine.
> I suspect the issue is with older Darwin versions.
>
>
> This could be the case. The public Darwin bots may have older OSs and SDKs
> than the machine you're using.
>
> In that case, can we guard against older versions specifically, rather
> than disabling Darwin altogether?
>
>
> You could, but this would only make sense to me if there's at least one
> public Darwin bots on which this test passes, since otherwise the test
> would still not reliably provide coverage on Darwin.
>
> Also, why the error has only appeared now?
>
>
> Are you referring to the "error in backend: Global variable
> '__sancov_gen_' has an invalid section specifier" issue? This has been an
> issue since at least June 3rd, but we've been marking the tests as
> unsupported on Darwin to suppress the failures.
>
>
> Er, excuse me, I meant '__sancov_counters'. (While we're on the topic,
> could the fix here be using '__sancov_cnts'?)
>
probably. But I have no way to test it (other than by committing the
change)

>
> vedant
>
>
> vedant
>
>
> On Jun 13, 2017, at 4:10 PM, Kostya Serebryany <kcc at google.com> wrote:
>
> also: r305335
>
> On Wed, Jun 7, 2017 at 10:11 PM, George Karpenkov <ekarpenkov at apple.com>
> wrote:
>
>> Weird though, my primary testing platform is a mac… maybe a different
>> version? I’ll take a look.
>> Also section specifier did change between versions.
>>
>> On Jun 7, 2017, at 9:29 PM, Kostya Serebryany <kcc at google.com> wrote:
>>
>> +ekarpenkov@
>>
>> Will it help to use shorter name for this section on Mac?
>>
>> --kcc
>>
>> On Sat, Jun 3, 2017 at 6:56 PM, Vedant Kumar via llvm-commits <
>> llvm-commits at lists.llvm.org> wrote:
>>
>>> Author: vedantk
>>> Date: Sat Jun  3 20:56:44 2017
>>> New Revision: 304673
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=304673&view=rev
>>> Log:
>>> Mark sancov test as unsupported on Darwin
>>>
>>> This test has been failing on all Darwin bots since it was introduced:
>>> http://lab.llvm.org:8080/green/job/clang-stage1-configure-RA_check/32111
>>>
>>> fatal error: error in backend: Global variable '__sancov_gen_' has an
>>> invalid section specifier '__DATA,__sancov_counters': mach-o section
>>> specifier requires a section whose length is between 1 and 16 characters.
>>> Target: x86_64-apple-darwin15.6.0
>>>
>>> Modified:
>>>     compiler-rt/trunk/test/sanitizer_common/TestCases/sanitizer_
>>> coverage_inline8bit_counter.cc
>>>
>>> Modified: compiler-rt/trunk/test/sanitizer_common/TestCases/sanitizer_
>>> coverage_inline8bit_counter.cc
>>> URL: http://llvm.org/viewvc/llvm-project/compiler-rt/
>>> trunk/test/sanitizer_common/TestCases/sanitizer_coverage_
>>> inline8bit_counter.cc?rev=304673&r1=304672&r2=304673&view=diff
>>> ============================================================
>>> ==================
>>> --- compiler-rt/trunk/test/sanitizer_common/TestCases/sanitizer_coverage_inline8bit_counter.cc
>>> (original)
>>> +++ compiler-rt/trunk/test/sanitizer_common/TestCases/sanitizer_coverage_inline8bit_counter.cc
>>> Sat Jun  3 20:56:44 2017
>>> @@ -1,7 +1,7 @@
>>>  // Tests -fsanitize-coverage=inline-8bit-counters
>>>  //
>>>  // REQUIRES: has_sancovcc,stable-runtime
>>> -// UNSUPPORTED: i386-darwin
>>> +// UNSUPPORTED: i386-darwin, x86_64-darwin
>>>  //
>>>  // RUN: %clangxx -O0 %s -fsanitize-coverage=inline-8bit-counters 2>&1
>>>
>>>
>>>
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at lists.llvm.org
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>>
>>
>>
>>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170613/5062fd45/attachment.html>