[Bug go/78978] New: [7 regression] runtime/pprof FAILs on Solaris 2/x86

["ro at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78978

            Bug ID: 78978
           Summary: [7 regression] runtime/pprof FAILs on Solaris 2/x86
           Product: gcc
           Version: 7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: go
          Assignee: ian at airs dot com
          Reporter: ro at gcc dot gnu.org
                CC: cmang at google dot com
  Target Milestone: ---
              Host: i386-pc-solaris2.*
            Target: i386-pc-solaris2.*
             Build: i386-pc-solaris2.*

Some time ago, runtime/pprof started to FAIL on Solaris/x86 with as and ld:

ld.so.1: a.out: fatal: a.out: hardware capability (CA_SUNW_HW_1) unsupported:
0x4400000  [ AES SSSE3 ]
FAIL: runtime/pprof

The same issue has been fixed before for libgo.so by linking it with
-mclear-hwcap,
but that change isn't effective here since the runtime/pprofl/check a.out is
linked with

# At least for now, we need -static-libgo for this test, because
# otherwise we can't get the line numbers.
# Also use -fno-inline to get better results from the memory profiler.
runtime_pprof_check_GOCFLAGS = -static-libgo -fno-inline

Adding $(HWCAP_LDCAPS) here fixes the immediate failure, but leaves the general
issue of libgo.a in the presence of hwcaps.

I see two options:

* Build either all of libgo or at least aeshash.la with -Wa,-nH (as is done in
  libstdc++) to avoid generating hwcaps in the first place, or

* Make -static-libgo imply -mclear-hwcap.

  Rainer