Re: [Minios-devel] [UNIKRAFT PATCHv5 22/46] arch/arm64: Avoid using the floating-point and Advanced SIMD registers

[Simon Kuenzer <simon.kuenzer@xxxxxxxxx>]

Hi,

On 10.08.2018 09:08, Wei Chen wrote:
> From: Wei Chen <Wei.Chen@xxxxxxx>
>
> On Arm64, sometimes, the GCC will use floating-point and Advanced
> SIMD registers to pass parameters. For example, the va_list will
> use the SIMD&FP registers (like q0, q1) to store parameters, no
> matter you are using floating-point or not.
>
> So, we use the GCC -mgeneral-regs-only flag to force GCC to use
> generic registers only util the floating-point and Advanced SIMD
> registers are required actually.
>
> Signed-off-by: Wei Chen <Wei.Chen@xxxxxxx>
> ---
>   arch/arm/arm64/Makefile.uk | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/arch/arm/arm64/Makefile.uk b/arch/arm/arm64/Makefile.uk
> index d4ba7e3..82a159f 100644
> --- a/arch/arm/arm64/Makefile.uk
> +++ b/arch/arm/arm64/Makefile.uk
> @@ -1,6 +1,6 @@
> -ASFLAGS  += -D__ARM_64__
> -CFLAGS   += -D__ARM_64__ -fms-extensions
> -CXXFLAGS += -D__ARM_64__ -fms-extensions
> +ASFLAGS  += -D__ARM_64__ -mgeneral-regs-only
> +CFLAGS   += -D__ARM_64__ -fms-extensions -mgeneral-regs-only
> +CXXFLAGS += -D__ARM_64__ -fms-extensions -mgeneral-regs-only
Could you maybe also add a short TODO comment and explain that we are 
using non-floating point for now.
Other form this, this patch is fine ;-)

>   
>   # GCC support -mcpu=native for arm64 from 6.0
>   ifeq ($(CONFIG_MARCH_ARM64_NATIVE),y)
_______________________________________________
Minios-devel mailing list
Minios-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/minios-devel