paperchalice wrote:
> With this in mind I get: (ffp-model=precise ; default) clang
> --target=arm-none-eabihf -mcpu=cortex-a53 -mfloat-abi=hard -S -o - file.c
> -emit-llvm -ffp-model=precise | grep eabi
>
> > !2 = !{i32 2, !"arm-eabi-fp-denormal", i32 1}
> > !3 = !{i32 2, !"arm-eabi-fp-exceptions", i32 1}
> > !4 = !{i32 2, !"arm-eabi-fp-number-model", i32 3}
>
> (ffp-model=strict) clang --target=arm-none-eabihf -mcpu=cortex-a53
> -mfloat-abi=hard -S -o - file.c -emit-llvm -ffp-model=strict | grep eabi
>
> > !2 = !{i32 2, !"arm-eabi-fp-denormal", i32 1}
> > !3 = !{i32 2, !"arm-eabi-fp-number-model", i32 3}
>
When testing with `-ffp-model=strict`, clang emits `warning: overriding
currently unsupported use of floating point exceptions on this target` and sets
exception model to `ignore` forcely.
https://github.com/llvm/llvm-project/pull/161106
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
