[PATCH] [gcc, combine] Backport to GCC 5.0 branch "PR46164: Don't combine the insns if a volatile register is contained".

[PATCH] [gcc, combine] Backport to GCC 5.0 branch "PR46164: Don't combine the insns if a volatile register is contained".

[Hale Wang]

> -----Original Message-----
> From: Terry Guo [mailto:flameroc@gmail.com]
> Sent: Wednesday, April 22, 2015 3:24 PM
> To: Segher Boessenkool
> Cc: law@redhat.com; Hale Wang; GCC Patches; Richard Sandiford
> Subject: Re: Ping^3 : [PATCH] [gcc, combine] PR46164: Don't combine the
> insns if a volatile register is contained.
> 
> On Wed, Apr 22, 2015 at 10:30 AM, Segher Boessenkool
> <segher@kernel.crashing.org> wrote:
> > On Wed, Apr 22, 2015 at 10:21:43AM +0800, Terry Guo wrote:
> >> gcc/ChangeLog:
> >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> >>                     Terry Guo  <terry.guo@arm.com>
> >>
> >>        PR rtl-optimization/64818
> >>        * combine.c (can_combine_p): Don't combine user-specified register
> if
> >>        it is in an asm input.
> >>
> >> gcc/testsuite/ChangeLog:
> >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> >>                     Terry Guo  <terry.guo@arm.com>
> >>
> >>        PR rtl-optimization/64818
> >>        * gcc.target/arm/pr64818.c: New.
> >
> > This is okay for trunk, if it has been bootstrapped and regression tested.
> >
> > Thanks,
> >
> >
> > Segher
> 
> Thanks Segher. The patch is tested with bootstrap and regression test for
> x86_64. No problem found. Committed as revision 222306.
> 

This patch applies cleanly on GCC 5.0 branch. Bootstrap and regression test are OK for X86_64.

Can we backport this patch to GCC 5.0 branch?

Best Regards,
Hale

> BR,
> Terry

Re: [PATCH] [gcc, combine] Backport to GCC 5.0 branch "PR46164: Don't combine the insns if a volatile register is contained".

[Segher Boessenkool]

On Thu, May 14, 2015 at 01:56:54PM +0800, Hale Wang wrote:
> > >> gcc/ChangeLog:
> > >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> > >>                     Terry Guo  <terry.guo@arm.com>
> > >>
> > >>        PR rtl-optimization/64818
> > >>        * combine.c (can_combine_p): Don't combine user-specified register
> > if
> > >>        it is in an asm input.
> > >>
> > >> gcc/testsuite/ChangeLog:
> > >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> > >>                     Terry Guo  <terry.guo@arm.com>
> > >>
> > >>        PR rtl-optimization/64818
> > >>        * gcc.target/arm/pr64818.c: New.

> This patch applies cleanly on GCC 5.0 branch. Bootstrap and regression test are OK for X86_64.
> 
> Can we backport this patch to GCC 5.0 branch?

It should be perfectly safe, and it's a pretty nasty bug.  But it is
technically not a regression (or is it?), so I'll defer to the release
managers.


Segher

RE: [PATCH] [gcc, combine] Backport to GCC 5.0 branch "PR46164: Don't combine the insns if a volatile register is contained".

[Hale Wang]

> -----Original Message-----
> From: Segher Boessenkool [mailto:segher@kernel.crashing.org]
> Sent: Thursday, May 14, 2015 9:29 PM
> To: Hale Wang
> Cc: law@redhat.com; GCC Patches; Richard Sandiford; 'Terry Guo'
> Subject: Re: [PATCH] [gcc, combine] Backport to GCC 5.0 branch "PR46164:
> Don't combine the insns if a volatile register is contained".
> 
> On Thu, May 14, 2015 at 01:56:54PM +0800, Hale Wang wrote:
> > > >> gcc/ChangeLog:
> > > >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> > > >>                     Terry Guo  <terry.guo@arm.com>
> > > >>
> > > >>        PR rtl-optimization/64818
> > > >>        * combine.c (can_combine_p): Don't combine user-specified
> > > >> register
> > > if
> > > >>        it is in an asm input.
> > > >>
> > > >> gcc/testsuite/ChangeLog:
> > > >> 2015-04-22 Hale Wang <hale.wang@arm.com>
> > > >>                     Terry Guo  <terry.guo@arm.com>
> > > >>
> > > >>        PR rtl-optimization/64818
> > > >>        * gcc.target/arm/pr64818.c: New.
> 
> > This patch applies cleanly on GCC 5.0 branch. Bootstrap and regression
test
> are OK for X86_64.
> >
> > Can we backport this patch to GCC 5.0 branch?
> 
> It should be perfectly safe, and it's a pretty nasty bug.  But it is
technically not
> a regression (or is it?), so I'll defer to the release managers.
> 

Yes, I agree it is not a regression.
> 
> Segher