Re: [PATCH v15 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] Conditional export of symbols used by generic inline macros breaks standard module usage.
--

commit 062222b285463a87a0b87a1f19a5112a409a0599
Author: Ankur Arora <ankur.a.arora@oracle.com>

arm64/delay: enable testing smp_cond_load_relaxed_timeout()

This enables the barrier tests to be built as a module by exporting the 
internal delay and timer functions for KUnit.

> 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]
Since smp_cond_load_relaxed_timeout() is an inline macro that expands to call
cpu_poll_relax() and __cmpwait_relaxed_timeout(), and those rely on
__delay_cycles(), does restricting this export to KUnit builds break standard
loadable modules?

If a standard kernel module (built without CONFIG_KUNIT) tries to use the
generic smp_cond_load_relaxed_timeout() API, it seems it will fail to link
with a modpost error because this mandatory symbol isn't exported. Would a
standard EXPORT_SYMBOL_GPL() be more appropriate here to ensure the generic
barrier API remains usable by all modules?

> diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
> index 4adf756423de9..91ab8552ead83 100644
> --- a/drivers/clocksource/arm_arch_timer.c
> +++ b/drivers/clocksource/arm_arch_timer.c

[ ... ]

> @@ -897,6 +898,7 @@ 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]
Similar to __delay_cycles() above, arch_timer_evtstrm_available() is called
by the inline expansion of smp_cond_load_relaxed_timeout(). Does this
conditional export render the generic synchronization API unusable for
non-KUnit modules that might want to use it?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831202251.305046-1-ankur.a.arora@oracle.com?part=13