Re: [PATCH v2 1/3] ARM: use .fpu assembler directives instead of assembler arguments

[Russell King - ARM Linux admin]

On Sun, Apr 19, 2020 at 02:35:49PM +0200, Stefan Agner wrote:
> Explicit FPU selection has been introduced in commit 1a6be26d5b1a
> ("[ARM] Enable VFP to be built when non-VFP capable CPUs are selected")
> to make use of assembler mnemonics for VFP instructions.
> 
> However, clang currently does not support passing assembler flags
> like this and errors out with:
> clang-10: error: the clang compiler does not support '-Wa,-mfpu=softvfp+vfp'
> 
> Make use of the .fpu assembler directives to select the floating point
> hardware selectively. Also use the new unified assembler language
> mnemonics. This allows to build these procedures with Clang.
> 
> Link: https://github.com/ClangBuiltLinux/linux/issues/762
> Signed-off-by: Stefan Agner <stefan@xxxxxxxx>
> ---
> Changes in v2:
> - Add link in commit message
> 
>  arch/arm/vfp/Makefile |  2 --
>  arch/arm/vfp/vfphw.S  | 30 +++++++++++++++++++-----------
>  2 files changed, 19 insertions(+), 13 deletions(-)
> 
> diff --git a/arch/arm/vfp/Makefile b/arch/arm/vfp/Makefile
> index 9975b63ac3b0..749901a72d6d 100644
> --- a/arch/arm/vfp/Makefile
> +++ b/arch/arm/vfp/Makefile
> @@ -8,6 +8,4 @@
>  # ccflags-y := -DDEBUG
>  # asflags-y := -DDEBUG
>  
> -KBUILD_AFLAGS	:=$(KBUILD_AFLAGS:-msoft-float=-Wa,-mfpu=softvfp+vfp -mfloat-abi=soft)
> -
>  obj-y		+= vfpmodule.o entry.o vfphw.o vfpsingle.o vfpdouble.o
> diff --git a/arch/arm/vfp/vfphw.S b/arch/arm/vfp/vfphw.S
> index b2e560290860..e214007a20a2 100644
> --- a/arch/arm/vfp/vfphw.S
> +++ b/arch/arm/vfp/vfphw.S
> @@ -258,11 +258,13 @@ vfp_current_hw_state_address:
>  
>  ENTRY(vfp_get_float)
>  	tbl_branch r0, r3, #3
> -	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	mrc	p10, 0, r0, c\dr, c0, 0	@ fmrs	r0, s0
> +	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,

Apart from the extraneous comma above, this looks fine, thanks.

> +1:	vmov	r0, s\dr
>  	ret	lr
>  	.org	1b + 8
> -1:	mrc	p10, 0, r0, c\dr, c0, 4	@ fmrs	r0, s1
> +	.endr
> +	.irp	dr,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31
> +1:	vmov	r0, s\dr
>  	ret	lr
>  	.org	1b + 8
>  	.endr
> @@ -271,10 +273,12 @@ ENDPROC(vfp_get_float)
>  ENTRY(vfp_put_float)
>  	tbl_branch r1, r3, #3
>  	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	mcr	p10, 0, r0, c\dr, c0, 0	@ fmsr	r0, s0
> +1:	vmov	s\dr, r0
>  	ret	lr
>  	.org	1b + 8
> -1:	mcr	p10, 0, r0, c\dr, c0, 4	@ fmsr	r0, s1
> +	.endr
> +	.irp	dr,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31
> +1:	vmov	s\dr, r0
>  	ret	lr
>  	.org	1b + 8
>  	.endr
> @@ -282,15 +286,17 @@ ENDPROC(vfp_put_float)
>  
>  ENTRY(vfp_get_double)
>  	tbl_branch r0, r3, #3
> +	.fpu	vfpv2
>  	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	fmrrd	r0, r1, d\dr
> +1:	vmov	r0, r1, d\dr
>  	ret	lr
>  	.org	1b + 8
>  	.endr
>  #ifdef CONFIG_VFPv3
>  	@ d16 - d31 registers
> -	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	mrrc	p11, 3, r0, r1, c\dr	@ fmrrd	r0, r1, d\dr
> +	.fpu	vfpv3
> +	.irp	dr,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31
> +1:	vmov	r0, r1, d\dr
>  	ret	lr
>  	.org	1b + 8
>  	.endr
> @@ -304,15 +310,17 @@ ENDPROC(vfp_get_double)
>  
>  ENTRY(vfp_put_double)
>  	tbl_branch r2, r3, #3
> +	.fpu	vfpv2
>  	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	fmdrr	d\dr, r0, r1
> +1:	vmov	d\dr, r0, r1
>  	ret	lr
>  	.org	1b + 8
>  	.endr
>  #ifdef CONFIG_VFPv3
> +	.fpu	vfpv3
>  	@ d16 - d31 registers
> -	.irp	dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
> -1:	mcrr	p11, 3, r0, r1, c\dr	@ fmdrr	r0, r1, d\dr
> +	.irp	dr,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31
> +1:	vmov	d\dr, r0, r1
>  	ret	lr
>  	.org	1b + 8
>  	.endr
> -- 
> 2.25.1
> 
> 

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up