This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] Fix memory overflow in mainline (take 2)
- From: Roger Sayle <roger at www dot eyesopen dot com>
- To: Gerald Pfeifer <gerald at pfeifer dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Tue, 18 Feb 2003 06:38:24 -0700 (MST)
- Subject: Re: [PATCH] Fix memory overflow in mainline (take 2)
On Tue, 18 Feb 2003, Gerald Pfeifer wrote:
> On Mon, 17 Feb 2003, Roger Sayle wrote:
> > 2003-02-17 Roger Sayle <firstname.lastname@example.org>
> > * sbitmap.c (sbitmap_resize): New function.
> > * sbitmap.h (sbitmap_resize): Prototype here.
> > * recog.c (split_all_insns): Use sbitmap_resize.
> This does not solve the bootstrap comparison error I'm seeing on
> Or do we now have _two_ different breakages in our tree? :-(
Yep, we have two different breakages. This patch was originally
developed to solve the building StrictMath.lo segmentation fault.
However, yesterday morning Jan applied a patch as "obvious" that
alleviates the failure yesterday morning. Unfortunately, his change
is still unsafe (reading beyond allocated memory) but we no longer
segmentation fault during a Linux bootstrap. Hence this memory
overrun still needs to be fixed, lest it bite another platform or
when compiling user code.
I haven't seen the bootstrap comparison failures you're seeing on
i386-unknown-freebsd4.7, on either i686-pc-linux-gnu or i686-pc-cygwin.
Can you determine whether its i386 specific or caused by FreeBSD's