This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PLEASE APPROVE: patch to fix PR 5879, a serious gcc 3.1regression
>>>>> "Richard" == Richard Henderson <rth@redhat.com> writes:
> On Tue, Mar 19, 2002 at 01:40:27PM +0000, Jason Merrill wrote:
>> Seems to me that the solution is to perform EH processing before the
>> sibcall pass; this also allows us to sibcallize the call to
>> _Unwind_SjLj_Unregister. Richard, what do you think?
> The EH processing can't handle CALL_PLACEHOLDER. I ran into too
> many problems when I did so, mostly with different exception regions
> inside the different code sequences. Were you actually able to
> bootstrap this patch with --enable-sjlj-exceptions?
Nope, it was just posted for discussion.
> A better solution is to do the first round of dead code elimination,
> and if there are still live exception regions, and we're using sjlj,
> then prevent all sibcalls.
OK. It would be nice to be able to sibcallize the call to
_Unwind_SjLj_Unregister, though.
Jason