On 19/06/2025 14:25, Alexandre Oliva wrote: > On Jun 19, 2025, Alexandre Oliva <ol...@adacore.com> wrote: > >> Or maybe the requirements for this testcase should be stated as >> arm_arch_v7? I'd have to add arm_arch_v7 to >> check_effective_target_arm_arch_FUNC_ok et al, if there aren't reasons >> why it's not there, but I'd be happy to do that, and use dg-add-options >> instead of an explicit -march=armv7, like pr117675.c, if that would be >> preferred. > > It turned out that the regression report I got from linaro was related > https://linaro.atlassian.net/browse/GNU-1599 > > While testing a configuration with -mfloat-abi=hard, -marmv7 is rejected > because it doesn't have an FPU. > > Requiring thumb2 wouldn't have avoided this failure, but requiring armv7 > is a proper fix for both problems. > > (sorry it me took so long; I'm still busy busy with other > fp2sp-elimination-disabled-during-lra problems affecting arm) > > > Without stating the architecture version required by the test, test > runs with options that are incompatible with the required > architecture version fail, e.g. -mfloat-abi=hard. > > armv7 was not covered by the long list of arm variants in > target-supports.exp, so add it, and use it for the effective target > requirement and for the option. > > Tested on arm-eabi, and also on arm-vx7r2 with gcc-14. Ok to install? > > > for gcc/testsuite/ChangeLog > > PR rtl-optimization/120424 > * lib/target-supports.exp (arm arches): Add arm_arch_v7. > * g++.target/arm/pr120424.C: Require armv7 support. Use > dg-add-options arm_arch_v7 instead of explicit -march=armv7. > --- > gcc/testsuite/g++.target/arm/pr120424.C | 4 +++- > gcc/testsuite/lib/target-supports.exp | 1 + > 2 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/gcc/testsuite/g++.target/arm/pr120424.C > b/gcc/testsuite/g++.target/arm/pr120424.C > index 4d0e49013c04a..40295ac80da93 100644 > --- a/gcc/testsuite/g++.target/arm/pr120424.C > +++ b/gcc/testsuite/g++.target/arm/pr120424.C > @@ -1,5 +1,7 @@ > /* { dg-do compile } */ > -/* { dg-options "-march=armv7 -O2 -fstack-clash-protection > -fnon-call-exceptions" } */ > +/* { dg-require-effective-target arm_arch_v7_ok } */ > +/* { dg-options "-O2 -fstack-clash-protection -fnon-call-exceptions" } */ > +/* { dg-add-options arm_arch_v7 } */ > /* { dg-final { scan-assembler-not {#-8} } } */ > /* LRA register elimination gets confused when register spilling > causes arm_frame_pointer_required to switch from false to true, and > diff --git a/gcc/testsuite/lib/target-supports.exp > b/gcc/testsuite/lib/target-supports.exp > index dfffe3adfbdd0..858fa1787f19c 100644 > --- a/gcc/testsuite/lib/target-supports.exp > +++ b/gcc/testsuite/lib/target-supports.exp > @@ -6073,6 +6073,7 @@ foreach { armfunc armflag armdefs } { > v6z_arm "-march=armv6z+fp -marm" "__ARM_ARCH_6Z__ && !__thumb__" > v6z_thumb "-march=armv6z+fp -mthumb -mfloat-abi=softfp" > "__ARM_ARCH_6Z__ && __thumb__" > v6m "-march=armv6-m -mthumb -mfloat-abi=soft" __ARM_ARCH_6M__ > + v7 "-march=armv7" __ARM_ARCH_7__
Similarly to the line directly below, this needs to be 'armv7+fp', otherwise this won't work on targets that use the hard float abi and use -mfpu=auto (pretty much the default these days). Otherwise, OK. R. > v7a "-march=armv7-a+fp" __ARM_ARCH_7A__ > v7a_arm "-march=armv7-a+fp -marm" "__ARM_ARCH_7A__ && !__thumb__" > v7a_fp_hard "-march=armv7-a+fp -mfpu=auto -mfloat-abi=hard" > __ARM_ARCH_7A__ > >
