Re: [PATCH v4 04/11] s390/percpu: Let MVIY_PERCPU() calculate alternative displacement

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] s390/percpu: incomplete refactoring in arch_raw_cpu_ptr()
--

commit 7a99372f1f5c810f33cd0d340f3c23eb3cf46f5f
Author: Heiko Carstens <hca@linux.ibm.com>

s390/percpu: Let MVIY_PERCPU() calculate alternative displacement

This commit updates the percpu inline assemblies to calculate the alternative
lowcore displacement directly within the assembly string using LC_ALT_ADDR.
This removes the need to repeatedly pass the same alternative displacement
values as constraints, making the code more readable.

> diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
> index 9888b38dedab2..3bc448c4c2dd7 100644
> --- a/arch/s390/include/asm/percpu.h
> +++ b/arch/s390/include/asm/percpu.h
[ ... ]
> -#define AG_ALT(disp, dispalt, reg)						\
> -	ALTERNATIVE("	ag	" reg ", " disp	   "(%%r0)\n",			\
> -		    "	ag	" reg ", " dispalt "(%%r0)\n",			\
> +#define AG_ALT(disp, reg)							\
> +	ALTERNATIVE("	ag	" reg ", " disp			"(%%r0)\n",	\
> +		    "	ag	" reg ", " disp "+" LC_ALT_ADDR "(%%r0)\n",	\
>  		    ALT_FEATURE(MFEATURE_LOWCORE))

[Severity: Low]
Does this refactoring miss arch_raw_cpu_ptr() which is defined earlier in
the same file?

Looking at arch_raw_cpu_ptr(), it still uses an ALTERNATIVE block that passes
LOWCORE_ALT_ADDRESS manually:

arch/s390/include/asm/percpu.h:arch_raw_cpu_ptr()
	asm_inline volatile(
	ALTERNATIVE("ag		%[__ptr__],%[offzero](%%r0)\n",
		    "ag		%[__ptr__],%[offalt](%%r0)\n",
		    ALT_FEATURE(MFEATURE_LOWCORE))
	: [__ptr__] "+d" (tcp_ptr__)
	: [offzero] "i" (lc_percpu),
	  [offalt] "i" (lc_percpu + LOWCORE_ALT_ADDRESS),

Could this be updated to use the new LC_ALT_ADDR calculation strategy for
consistency?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260921155806.2447506-1-hca@linux.ibm.com?part=4