[Bug d/88462] New: All D execution tests FAIL on Solaris/SPARC

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88462

            Bug ID: 88462
           Summary: All D execution tests FAIL on Solaris/SPARC
           Product: gcc
           Version: 9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: d
          Assignee: ibuclaw at gdcproject dot org
          Reporter: ro at gcc dot gnu.org
  Target Milestone: ---
            Target: sparc*-sun-solaris2.*

Even with the latest patch for PR d/88150 to use sections_elf_shared.d on
Solaris,
which allows the vast majority of gdc execution tests to PASS on Solaris
11.4/x86,
I have no such luck on Solaris 11.5/SPARC: all tests FAIL with

Aborting from local/libphobos/libdruntime/core/sync/mutex.d(95) Error:
pthread_mutex_init failed.

Thread 2 received signal SIGABRT, Aborted.
[Switching to Thread 1 (LWP 1)]
0xfec7e044 in __lwp_sigqueue () from /lib/libc.so.1
(gdb) where
#0  0xfec7e044 in __lwp_sigqueue () from /lib/libc.so.1
#1  0xfebb9898 in raise () from /lib/libc.so.1
#2  0xfeb8b1d0 in abort () from /lib/libc.so.1
#3  0x000c150c in core.internal.abort.abort(immutable(char)[],
immutable(char)[], uint) (msg=..., 
    filename=<error reading variable: Cannot access memory at address 0x5f>, 
    line=95)
    at
/vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/core/internal/abort.d:44
#4  0x000fe32c in core.sync.mutex.Mutex.this!(core.sync.mutex.Mutex).this(bool)
    (this=0x12de34 <core.thread.Thread._locks+44>, _unused_=true)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/core/sync/mutex.d:94
#5  0x000fdeac in core.sync.mutex.Mutex.this() (
    this=0x12de34 <core.thread.Thread._locks+44>)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/core/sync/mutex.d:63
#6  0x000c64d0 in core.thread.Thread.initLocks() ()
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/core/thread.d:1726
#7  0x000c69d8 in thread_init ()
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/core/thread.d:2022
#8  0x000a232c in gc_init ()
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/gc/proxy.d:56
#9  0x00080684 in rt_init ()
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/rt/dmain2.d:187
#10 0x000810c8 in runAll (this=0xffbfe904)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/rt/dmain2.d:485
#11 0x00081020 in tryExec (this=0xffbfe904, dg=...)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/rt/dmain2.d:461
#12 0x00080f2c in _d_run_main (argc=1, argv=0xffbfea34, 
    mainFunc=0x6b2bc <D main>)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/rt/dmain2.d:494
#13 0x0006b24c in main (argc=1, argv=0xffbfea34)
    at /vol/gcc/src/hg/trunk/solaris/libphobos/libdruntime/__entrypoint.di:44
#14 0x0006b014 in _start ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

            !pthread_mutex_init(cast(pthread_mutex_t*) &m_hndl, &attr) ||
                abort("Error: pthread_mutex_init failed.");

After much digging and head scratching, I found what's wrong:
pthread_mutex_init
expects the mutex to be long long (i.e. 8 byte) aligned.  I'd thought this
would happen automatically given the declaration in core/sys/posix/sys/types.d
with the ulong __pthread_mutex_data field which has a natural alignment of
64 bits.  Whatever I do, however, I only end up with the mutex being 4-byte
aligned:

* apply align(8): at the beginning of the pthread_mutex_t fields,

* apply align(8) to the struct pthread_mutex_t declaration, or

* apply align(8) to the m_hndl member of Class Mutex in core/sync/mutex.d.

To guard against incomplete dependencies that could lead to some parts not
being recompiled when they should, I've always recompiled all of libphobos to
be sure I didn't miss something.

I've no idea what I'm doing wrong here.