This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Bug in gcc CVS HEAD from a few days ago with alloca and stack alignment
- To: khendricks at ivey dot uwo dot ca
- Subject: Re: Bug in gcc CVS HEAD from a few days ago with alloca and stack alignment
- From: Geoff Keating <geoffk at geoffk dot org>
- Date: 11 Feb 2001 00:34:07 -0800
- CC: gcc-patches at gcc dot gnu dot org, David Edelsohn <dje at watson dot ibm dot com>
- References: <200102110033.TAA37292@mal-ach.watson.ibm.com> <01021022504800.16569@localhost>
Kevin B. Hendricks <khendricks@ivey.uwo.ca> writes:
> Hi David,
>
> Thanks!
>
> I changed STACK_BOUNDARY to 128 and rebuilt the entire gcc and now everything
> has started working including the gcj closure code and the gij interpreter.
>
> Will you please commit a proper fix? Franz's mail machine bounced my last
> mail so I don't think he knows about this problem.
The proper fix is to find out what part of the compiler is failing to
honour ABI_STACK_BOUNDARY and fix that. I don't think changing
STACK_BOUNDARY to 128 gives any significant benefits for optimisation as yet,
since there aren't yet any operations that need that alignment.
> > So it looks to me that this is some incorrect accomodation made >
> for eABI, i.e., someone trying to use the sysv4 configuration for
> eABI > instead of creating a real cross-compiler.
I believe it's to allow linking of sysv4 code with eabi code,
primarily to save on multilibs. STACK_BOUNDARY is the stack alignment
that the compiler can assume, ABI_STACK_BOUNDARY is the stack
alignment that the compiler should preserve.
--
- Geoffrey Keating <geoffk@geoffk.org>