On Mon, Mar 24, 2014 at 11:32:39AM +, Marcus Shawcroft wrote:
> On 21 March 2014 14:55, Alan Lawrence wrote:
> > This patch fixes a bug whereby a vector like V8QImode {1,0,1,0,1,0,1,0} can
> > result in an instruction like
> >
> > movi v1.4h, 0x1
> >
> > whereas on bigendian this should be
> >
On 21 March 2014 14:55, Alan Lawrence wrote:
> This patch fixes a bug whereby a vector like V8QImode {1,0,1,0,1,0,1,0} can
> result in an instruction like
>
> movi v1.4h, 0x1
>
> whereas on bigendian this should be
>
> movi v1.4h, 0x1, lsl 8
>
> Regression tested on aarch64_be-none-elf: no changes
This patch fixes a bug whereby a vector like V8QImode {1,0,1,0,1,0,1,0} can
result in an instruction like
movi v1.4h, 0x1
whereas on bigendian this should be
movi v1.4h, 0x1, lsl 8
Regression tested on aarch64_be-none-elf: no changes in libstdc++, newlib; no
regressions in gcc or g++ and FAI