Re: [PATCH v16 13/16] arm64/delay: enable testing smp_cond_load_relaxed_timeout()

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] arm64: barrier: missing exports for smp_cond_load_*_timeout() backends
--

commit bfc8166431acdeebcd1866a53cbd230aa1bcd738
Author: Ankur Arora <ankur.a.arora@oracle.com>

arm64/delay: enable testing smp_cond_load_relaxed_timeout()

This commit enables the barrier tests to be built as a module by exporting
the underlying delay and timer functions when KUnit is enabled.

> diff --git a/arch/arm64/lib/delay.c b/arch/arm64/lib/delay.c
> index f08eacef2f0a4..20bc2894b65b9 100644
> --- a/arch/arm64/lib/delay.c
> +++ b/arch/arm64/lib/delay.c
[ ... ]
> @@ -30,6 +31,7 @@ u64 notrace __delay_cycles(void)
>  	guard(preempt_notrace)();
>  	return __arch_counter_get_cntvct_stable();
>  }
> +EXPORT_SYMBOL_IF_KUNIT(__delay_cycles);

[Severity: Medium]
Does exporting __delay_cycles() only for KUnit cause unresolved symbol errors
for standard loadable modules?

If a standard non-test module uses atomic_cond_read_acquire_timeout() or
smp_cond_load_relaxed_timeout() on ARM64, the macros expand to inline barrier
functions like __cmpwait_relaxed_timeout(), which call __delay_cycles().

Since this symbol is only exported when CONFIG_KUNIT is enabled, will standard
production builds fail to build or load modules using these APIs?

>  
>  void __delay(unsigned long cycles)
> diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
> index cc9a8b399004d..227447b19b143 100644
> --- a/drivers/clocksource/arm_arch_timer.c
> +++ b/drivers/clocksource/arm_arch_timer.c
[ ... ]
> @@ -893,6 +894,7 @@ bool arch_timer_evtstrm_available(void)
>  	 */
>  	return cpumask_test_cpu(raw_smp_processor_id(), &evtstrm_available);
>  }
> +EXPORT_SYMBOL_IF_KUNIT(arch_timer_evtstrm_available);

[Severity: Medium]
Similarly, does restricting the export of arch_timer_evtstrm_available() to
KUnit builds break standard modules?

The inline cpu_poll_relax() backend relies on this function. Standard modules
using the new timeout APIs might fail to resolve this symbol when compiled
without CONFIG_KUNIT. Should these be exported unconditionally using standard
EXPORT_SYMBOL_GPL() instead?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261016070837.440748-1-ankur.a.arora@oracle.com?part=13