On Thu, Mar 17, 2016 at 01:32:15PM +, Joern Rennecke wrote:
> > Note that during register allocation / reload, REGNO equivalence is
> > generally wrong,
> as it fails to distinguish between the frame pointer and a hard register that
> has the
> same regno as the frame pointer which has been
On 03/17/2016 12:16 PM, Jakub Jelinek wrote:
Thus, I've reverted the patch (kept the testcase), and after some
discussions on IRC bootstrapped/regtested on x86_64-linux and i686-linux
following version, which right now should change behavior just for the i?86
case and nothing else, so shouldn't
On Thu, Mar 17, 2016 at 11:07:03PM +1030, Alan Modra wrote:
> On Thu, Mar 17, 2016 at 12:16:58PM +0100, Jakub Jelinek wrote:
> > the rs6000 backend for whatever strange reason I haven't understood
> > really wants pointer equality instead of REGNO comparison (even when the
> > modes match), one
On 03/16/2016 01:22 PM, Jakub Jelinek wrote:
So, this is what we've converged to on IRC and passed bootstrap/regtest
on x86_64-linux and i686-linux. Is this ok for trunk?
The explanation was a bit confusing at first, but I think this looks
reasonable. The assert worries me, but triggering it
Hi!
The following testcase is miscompiled on ia32, because
a peephole2 calls replace_rtx trying to replace SImode %ecx with
SImode %edx, but replace_rtx (unlike e.g. simplify_replace_rtx
or validate_replace_rtx), in addition to modifying the rtxes
in place (fine in this case) only does pointer
On Thu, Mar 17, 2016 at 12:16:58PM +0100, Jakub Jelinek wrote:
> the rs6000 backend for whatever strange reason I haven't understood
> really wants pointer equality instead of REGNO comparison (even when the
> modes match), one (reg:DI 12) should be replaced, another (reg:DI 12)
> should not.
By
On Wed, Mar 16, 2016 at 01:59:29PM +0100, Bernd Schmidt wrote:
> On 03/16/2016 01:22 PM, Jakub Jelinek wrote:
> >So, this is what we've converged to on IRC and passed bootstrap/regtest
> >on x86_64-linux and i686-linux. Is this ok for trunk?
>
> The explanation was a bit confusing at first, but
Hi!
On Wed, Mar 16, 2016 at 05:48:33PM -0500, Segher Boessenkool wrote:
> On Wed, Mar 16, 2016 at 01:59:29PM +0100, Bernd Schmidt wrote:
> > On 03/16/2016 01:22 PM, Jakub Jelinek wrote:
> > >So, this is what we've converged to on IRC and passed bootstrap/regtest
> > >on x86_64-linux and
On Thu, 2016-03-17 at 12:16 +0100, Jakub Jelinek wrote:
> Thus, I've reverted the patch (kept the testcase), and after some
> discussions on IRC bootstrapped/regtested on x86_64-linux and i686
> -linux following version, which right now should change behavior just
> for the i?86 case and nothing
