This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: ping [RFC] [patch] fix PR32893 - forcing alignment >= STACK_BOUNDARY
- From: "H.J. Lu" <hjl at lucon dot org>
- To: Dorit Nuzman <DORIT at il dot ibm dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Andrew Pinski <pinskia at gmail dot com>, Mark Mitchell <mark at codesourcery dot com>
- Date: Mon, 22 Oct 2007 06:36:21 -0700
- Subject: Re: ping [RFC] [patch] fix PR32893 - forcing alignment >= STACK_BOUNDARY
- References: <OFD2A8D94D.CA1972FB-ON8825737C.001A08DB-8825737C.001BFB56@il.ibm.com>
On Sun, Oct 21, 2007 at 09:05:38PM -0800, Dorit Nuzman wrote:
>
> So I'm not sure how to proceed with PR32893 (see details and proposed patch
> below).
>
> I didn't yet get a response from Pinsky on whether his fix for PR16660
> would also fix this PR,
> and recently H.J. marked this PR as dependent on the newly opened PR33721.
> It seems like there's a more general issue here that needs to be solved,
> and the simple fix I suggested below in the vectorizer is only a workaround
> (that would also involve a lot of testcase changes). I believe (hope...)
> that a fix to either PR16660 or PR33721 would also fix this PR, but I'm not
> sure what's going on with these PRs (what are the prospects of either of
> these PRs to get resolved in the near future?). In the meantime we are
> generating wrong code in the vectorizer...
We are working on a proposal to properly fix the stack alignment issue.
We will start a project for gcc 4.4.
H.J.