On Sat, Dec 10, 2016 at 10:37:38AM -0600, Segher Boessenkool wrote:
> On Fri, Dec 09, 2016 at 04:23:44PM +0100, Dominik Vogt wrote:
> > 0001-*
> >
> > Deal with mode expanding zero_extracts in change_zero_ext. The
> > patch looks good to me, but not sure whether endianness is
> > handled pr
On Sat, Dec 10, 2016 at 10:37:38AM -0600, Segher Boessenkool wrote:
> Okay for trunk with those changes (did you bootstrap+regcheck this
> already?) Thanks,
Oh, and a changelog :-)
Segher
On Fri, Dec 09, 2016 at 04:23:44PM +0100, Dominik Vogt wrote:
> 0001-*
>
> Deal with mode expanding zero_extracts in change_zero_ext. The
> patch looks good to me, but not sure whether endianness is
> handled properly. Is the second argument of gen_rtx_SUBREG
> correct?
> >From 600ed3da
On Fri, Dec 09, 2016 at 08:32:01PM +0100, Dominik Vogt wrote:
> > > This is a work in progress with the goal of fixing the first
> > > problem and similar ones by calling simplify_set after
> > > change_zero_ext to get rid of the overly complex code. That
> > > works fine in principle, but
On Fri, Dec 09, 2016 at 12:25:04PM -0600, Segher Boessenkool wrote:
> Hi Dominik,
>
> On Fri, Dec 09, 2016 at 04:23:44PM +0100, Dominik Vogt wrote:
> > 0001-*
> >
> > Deal with mode expanding zero_extracts in change_zero_ext. The
> > patch looks good to me, but not sure whether endianness is
Hi Dominik,
On Fri, Dec 09, 2016 at 04:23:44PM +0100, Dominik Vogt wrote:
> 0001-*
>
> Deal with mode expanding zero_extracts in change_zero_ext. The
> patch looks good to me, but not sure whether endianness is
> handled properly. Is the second argument of gen_rtx_SUBREG
> correct?
Thi
There are several situations that combine.c:change_zero_ext does not
handle well yet. One of them is
(and:SI (subreg:SI (zero_extract:DI (reg:DI) ...) ...)
(with const_int operands to "and" and "zero_extract")
=>
(and:SI (subreg:SI (and:DI (lshiftrt:DI ...)))
with two nested "and"s. Anoth
