This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: bogus calls cleanup
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: bogus calls cleanup
- From: Jan Hubicka <jh at suse dot cz>
- Date: Fri, 12 May 2000 09:05:11 +0200
- Cc: jh at suse dot cz, gcc-bugs at gcc dot gnu dot org
- References: <20000511193956.A11103@cygnus.com>
> Tue May 9 16:30:27 MET DST 2000 Jan Hubicka <jh@suse.cz>
>
> * calls.c (expand_call): Reorganize somewhat in order
> to avoid unnecesary recalculation inside loop and make
> tail call failure code cleaner.
>
> This broke sparc sibcalls. Note that initialize_argument_information
> is dependant on FUNCTION_INCOMING_ARG, which is of course pass-dependant.
>
> I'm going to revert the patch for now.
Oops... I see it depends on the sibbling calls.... Duh.
This is very unfortunate. Being able to initialze argument information
before doing the normal call or sibbling call dependent code is IMO important
otherwise we will be still detecting the sibcall failures later and I would
really love to have initialization code and emitting separate.
Modifying initialze_argument_information to fill two separate values
normal_call_reg and tail_call_reg sounds OK to you?
Honza
>
>
>
> r~