Re: PATCH: Fix Solaris 2/x86 signal handling in libgcj

[Andrew Haley <aph at redhat dot com>]

On 07/12/2010 05:37 PM, Rainer Orth wrote:
> Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:
> 
>> Andrew Haley <aph at redhat dot com> writes:
>>
>> [Please keep me on the Cc: I'm not subscribed to java-patches and hadn't
>> seen your message until I checked the archives.  Thanks.]
> [...]
>> It took me some time to get MD_FALLBACK_FRAME_STATE_FOR right for all
>> Solaris 2/x86 versions (patch to be posted shortly), but once that is
> 
> I've now posted the patch here:
> 
> 	Support Solaris 9, 11/x86 in MD_FALLBACK_FRAME_STATE_FOR
> 	http://gcc.gnu.org/ml/gcc-patches/2010-07/msg00988.html
> 
>> With the patch
>>
>> 	http://gcc.gnu.org/ml/gcc-patches/2010-02/msg00994.html
>>
>> applied, the Throw_2 failures vanish.  I've yet to investigate the
>> PR16923 failure, though.
>>
>> Therefore, I'd like to apply the patch to mainline, 4.4. and 4.5
>> branches.  Ok?
> 
> Testing revealed that I need one additional change on the libjava side,
> though: unlike the previous version of my MD_FALLBACK_FRAME_STATE_FOR,
> the new one relies on the frame pointer since it has to do some stack
> walking, which conflicts with libjava being built with
> -fomit-frame-pointer by default.  To enable this, I need the Makefile.am
> fragment below, too.
> 
> As before, sucessfully bootstrapped on Solaris 10 and 11/x86, leaving
> only the 64-bit PR16923 failure.
> 
> Ok for mainline (and 4.4 and 4.5 branches after testing there), together
> with the original patch?

Yes, that looks reasonable enough.

Andrew.