This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: g77 testsuite failures under hpux 10.20
- To: law at cygnus dot com
- Subject: Re: g77 testsuite failures under hpux 10.20
- From: "John David Anglin" <dave at hiauly1 dot hia dot nrc dot ca>
- Date: Thu, 6 Jul 2000 23:42:58 -0400 (EDT)
- Cc: gcc-bugs at gcc dot gnu dot org
>
> In message <200007062319.TAA11417@hiauly1.hia.nrc.ca>you write:
> > The obstack_grow call has walked over the rtl.
> OK. This is odd since I don't think the RTL should have been allocated
> in an obstack to begin with since RTL items are part of the garbage
> collected memory, not obstacked memory.
It appears that allocation in obstacked memory occurs if ggc_p is zero
(cf, ggc.h:27):
/* Language-specific code defines this variable to be either one (if
it wants garbage collection), or zero (if it does not). */
extern int ggc_p;
This could explain why the problem is g77 specific. Of course, there
is still a problem as to why obstacked memory is being overwritten.
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)