This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/51895] [4.7 Regression] ICE in simplify_subreg
- From: "jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Fri, 20 Jan 2012 09:07:54 +0000
- Subject: [Bug middle-end/51895] [4.7 Regression] ICE in simplify_subreg
- Auto-submitted: auto-generated
- References: <bug-51895-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-01-20 09:07:54 UTC ---
(In reply to comment #10)
> > I wonder why it does this, instead of just using type S, and if it really has
> > to for some reason, why it can't at least make sure it has the same TYPE_MODE.
> > Changing a TImode argument to a BLKmode argument doesn't look at least like a
> > good optimization.
>
> I concur, BLKmode means spilling to memory at some point, so this looks like a
> clear pessimization to me.
Sure, but I believe just that eipa_sra is what is causing this pessimization by
doing a bad decision.
Anyway, if you prefer one of the other 3 patches (for now?), which one it is?
Or some other place to handle it?