RE: threads and gc on SPARC/Solaris 2.8

["Boehm, Hans" <hans_boehm at hp dot com>]

My reading is that in 3.0, GC_clear_stack_inner is not used in the
multithreaded configuration, and it should thus be OK.

The trunk has a bug, as does my version.  GC_clear_stack_inner is called,
but not defined for SPARC + threads.  It looks to me like removing the /* &&
!defined(THREADS) */ condition in mach_dep.c should fix the problem.  I
think that was only a space optimization, which is now no longer correct.
I'll apply that fix to my version, but I'd appreciate it if someone with a
less modified tree could test it and check it in sooner than I will.

This does seem to imply that nobody has built the trunk on SPARC/Solaris for
a few months.

Hans

> -----Original Message-----
> From: Andrew Haley [mailto:aph@cambridge.redhat.com]
> Sent: Wednesday, June 27, 2001 11:55 AM
> To: java@gcc.gnu.org
> Cc: Hans-J. Boehm
> Subject: threads and gc on SPARC/Solaris 2.8
> 
> 
> In mach_dep.c we have:
> 
> #if defined(ASM_CLEAR_CODE) && !defined(THREADS)
> GC_clear_stack_inner: 
>    ... code to clear the stack.
> 
> GC_clear_stack_inner is defined nowhere else, and we depend on it.
> Yet it isn't compiled when THREADS is defined, which it always needs
> to be for Java.  However, comments say it is essential.  If I compile
> GC_clear_stack_inner and use it with threads, it faults with an
> illegal instruction.
> 
> What is going on here?  I don't get it.  Has anybody made gcj work on
> Sparc/Solaris?
> 
> Andrew.
>