Re: [PATCH 4/8] s390/entry: workaround llvm's IAS limitations

[Nathan Chancellor <nathan@kernel.org>]

On Thu, May 12, 2022 at 07:24:25PM +0200, Heiko Carstens wrote:
> On Wed, May 11, 2022 at 10:30:05AM -0700, Nathan Chancellor wrote:
> > Hi Heiko,
> > 
> > On Wed, May 11, 2022 at 02:05:28PM +0200, Heiko Carstens wrote:
> > > llvm's integrated assembler cannot handle immediate values which are
> > > calculated with two local labels:
> > > 
> > > <instantiation>:3:13: error: invalid operand for instruction
> > >  clgfi %r14,.Lsie_done - .Lsie_gmap
> > > 
> > > Workaround this by adding clang specific code which reads the specific
> > > value from memory. Since this code is within the hot paths of the kernel
> > > and adds an additional memory reference, keep the original code, and add
> > > ifdef'ed code.
> > > 
> > > Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
> > > ---
> > >  arch/s390/kernel/entry.S | 11 +++++++++++
> > >  1 file changed, 11 insertions(+)
> > > 
> > > diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
> > > index e1664b45090f..ff7a75078e93 100644
> > > --- a/arch/s390/kernel/entry.S
> > > +++ b/arch/s390/kernel/entry.S
> > > @@ -171,8 +171,19 @@ _LPP_OFFSET	= __LC_LPP
> > >  	.macro OUTSIDE reg,start,end,outside_label
> > >  	larl	%r14,\start
> > >  	slgrk	%r14,\reg,%r14
> > > +#ifdef CONFIG_CC_IS_CLANG
> > 
> > I intend to put this series through my build and boot test matrix later
> > today but one fly by comment in the meantime. Should this be
> > CONFIG_AS_IS_LLVM if this is an integrated assembler limitation, rather
> > than a clang one?
> 
> Yes, that makes a lot of sense. Considering that I will drop the
> previous patch within this series, the new version looks like:
> 
> From fe4fb0b014378d84ae517deaea338577b2ea6ae0 Mon Sep 17 00:00:00 2001
> From: Heiko Carstens <hca@linux.ibm.com>
> Date: Sat, 7 May 2022 15:00:40 +0200
> Subject: [PATCH 3/7] s390/entry: workaround llvm's IAS limitations
> 
> llvm's integrated assembler cannot handle immediate values which are
> calculated with two local labels:
> 
> <instantiation>:3:13: error: invalid operand for instruction
>  clgfi %r14,.Lsie_done - .Lsie_gmap
> 
> Workaround this by adding clang specific code which reads the specific
> value from memory. Since this code is within the hot paths of the kernel
> and adds an additional memory reference, keep the original code, and add
> ifdef'ed code.
> 
> Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
> ---
>  arch/s390/kernel/entry.S | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
> index a6b45eaa3450..f2f30bfba1e9 100644
> --- a/arch/s390/kernel/entry.S
> +++ b/arch/s390/kernel/entry.S
> @@ -172,9 +172,19 @@ _LPP_OFFSET	= __LC_LPP
>  	lgr	%r14,\reg
>  	larl	%r13,\start
>  	slgr	%r14,%r13
> -	lghi	%r13,\end - \start
> -	clgr	%r14,%r13
> +#ifdef CONFIG_AS_IS_LLVM
> +	clgfrl	%r14,.Lrange_size\@
> +#else
> +	clgfi	%r14,\end - \start
> +#endif
>  	jhe	\outside_label
> +#ifdef CONFIG_CC_IS_CLANG

I think this one also wants to be CONFIG_AS_IS_LLVM, right?

Other than that, seems fine to me, although I have no knowledge of s390
assembly so that statement probably means next to nothing :)

Cheers,
Nathan

> +	.section .rodata, "a"
> +	.align 4
> +.Lrange_size\@:
> +	.long	\end - \start
> +	.previous
> +#endif
>  	.endm
>  
>  	.macro SIEEXIT
> -- 
> 2.32.0