[Bug testsuite/79455] New: c-c++-common/tsan/race_on_mutex.c fails on powerpcle starting with r244854 (where it was activated)

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79455

            Bug ID: 79455
           Summary: c-c++-common/tsan/race_on_mutex.c fails on powerpcle
                    starting with r244854 (where it was activated)
           Product: gcc
           Version: 7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: seurer at gcc dot gnu.org
  Target Milestone: ---

------------------------------------------------------------------------
r244854 | jakub | 2017-01-24 02:19:37 -0600 (Tue, 24 Jan 2017) | 2 lines

        * configure.tgt: Enable tsan and lsan on powerpc64{,le}-*-linux*.


When run on powerpcle the output is the same as on x86 except for one section,
the one with memset.

On x86 this part is:
  Previous write of size 1 at 0x000000601300 by thread T1:
    #0 pthread_mutex_init
/home/seurer/gcc/gcc-test/libsanitizer/tsan/tsan_interceptors.cc:1117
(libtsan.so.0+0x00000002c1ee)
    #1 Thread1
/home/seurer/gcc/gcc-test/gcc/testsuite/c-c++-common/tsan/race_on_mutex.c:12
(race_on_mutex.exe+0x000000400bf8)

And on power it is:
  Previous write of size 8 at 0x000010011600 by thread T1:
    #0 memset
/home/seurer/gcc/gcc-test/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:558
(libtsan.so.0+0x0000000360f4)
    #1 pthread_mutex_init <null> (libpthread.so.0+0x00000000b61c)
    #2 Thread1
/home/seurer/gcc/gcc-test/gcc/testsuite/c-c++-common/tsan/race_on_mutex.c:12
(race_on_mutex.exe+0x000010000cb4)

It fails when run as a test because it doesn't quite match the expected pattern

/* { dg-output "  Previous write of size 1 at .* by thread T1:(\n|\r\n|\r)" }
*/
/* { dg-output "    #0 pthread_mutex_init .* (.)*" } */
/* { dg-output "    #1 Thread1.* .*(race_on_mutex.c:12|\\?{2}:0) .*" } */

>From poking around in the test case and the tsan code it looks like the test is
detecting what it should and the difference in output is mostly "cosmetic".