On Fri, 12 Feb 2021 at 17:02, Richard Earnshaw
<richard.earns...@foss.arm.com> wrote:
>
> On 12/02/2021 14:20, Christophe Lyon via Gcc-patches wrote:
> > This test forces -march=armv8.1-m.main, which supports only Thumb mode.
> > However, if the toolchain is not configured --with-thumb, the test
> > fails with:
> > error: target CPU does not support ARM mode
> >
> > Adding -mthumb to dg-options fixes the problem.
> >
>
> Hmm, this sounds like a different problem. The driver, these days, is
> supposed to detect when the architecture only supports thumb and to add
> that automatically when calling cc1 etc. So why isn't that working in
> this case?
>
> (See gcc/common/config/arm/arm-common.c:arm_target_thumb_only).
>
Well, I've just rebuilt --with-mode=arm --with-cpu=cortex-a9
--target=arm-none-eabi
and when I compile
$ arm-none-eabi-gcc hello.c -march=armv8.1-m.main -S
cc1: error: target CPU does not support ARM mode
I put a breakpoint in arm_target_thumb_only and it did not trigger.
> R.
>
> > Committed as obvious.
> >
> > 2021-02-12 Christophe Lyon <christophe.l...@linaro.org>
> >
> > PR target/98931
> > gcc/testsuite/
> > * gcc.target/arm/pr98931.c: Add -mthumb
> > ---
> > gcc/testsuite/gcc.target/arm/pr98931.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/gcc/testsuite/gcc.target/arm/pr98931.c
> > b/gcc/testsuite/gcc.target/arm/pr98931.c
> > index 313876a..66070ad 100644
> > --- a/gcc/testsuite/gcc.target/arm/pr98931.c
> > +++ b/gcc/testsuite/gcc.target/arm/pr98931.c
> > @@ -1,6 +1,6 @@
> > /* { dg-do assemble } */
> > /* { dg-skip-if "avoid conflicting multilib options" { *-*-* } { "-marm"
> > "-mcpu=*" } } */
> > -/* { dg-options "-march=armv8.1-m.main -O3
> > --param=max-completely-peeled-insns=1300 --save-temps" } */
> > +/* { dg-options "-march=armv8.1-m.main -O3
> > --param=max-completely-peeled-insns=1300 --save-temps -mthumb" } */
> >
> > extern long long a[][20][26][26][22];
> >
> >
>
