This is the mail archive of the
mailing list for the GCC project.
Re: [COMMITTED][PATCH/RFC] force_reg() doesn't propagate MEM_POINTER flag to REG_POINTER
- From: David Daney <ddaney at avtrex dot com>
- To: Peter Bergner <bergner at vnet dot ibm dot com>
- Cc: Roger Sayle <roger at eyesopen dot com>, Anton Blanchard <anton at samba dot org>, gcc-patches at gcc dot gnu dot org
- Date: Thu, 19 Oct 2006 18:14:29 -0700
- Subject: Re: [COMMITTED][PATCH/RFC] force_reg() doesn't propagate MEM_POINTER flag to REG_POINTER
- References: <Pine.LNX.email@example.com> <firstname.lastname@example.org>
Peter Bergner wrote:
On Thu, 2006-09-21 at 19:39 -0600, Roger Sayle wrote:
On Thu, 21 Sep 2006, Peter Bergner wrote:
* explow.c (force_reg): Set REG_POINTER flag according to
As PR28690 is a regression, this is OK for mainline. I think its
reasonable to commit this patch on its own to see if it creates
problems (for example, HPPA is sensitive to REG_POINTER changes),
independently of the rest of your fix. I don't expect any problems
from the change itself, which makes sense, but from latent bugs
that may be exposed by the RTL differences.
Thanks for fixing this,
Ok, I committed this to mainline by itself. Thanks.
This causes: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29519
Which leads to a complete failure of libgcj to do anything useful on