Re: [PATCH v5 10/21] KVM: selftests: Verify interrupts are received when IRQ affinity changes in IRQ test

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The test incorrectly assumes contiguous CPU topologies when generating random CPU IDs and ignores `fflush()` errors, leading to silent failures and buffered write accumulation.
--

commit a578bd3957ea3e342b8ed83cc29e669f1b6658b8
Author: David Matlack <dmatlack@google.com>

KVM: selftests: Verify interrupts are received when IRQ affinity changes in IRQ test

This commit adds the '-a' flag to tools/testing/selftests/kvm/irq_test.c to
randomly affinitize the device's host IRQ to different physical CPUs
throughout the test. This stresses the kernel's ability to maintain correct
interrupt routing and delivery.

> diff --git a/tools/testing/selftests/kvm/irq_test.c b/tools/testing/selftests/kvm/irq_test.c
> index 5b42d8abd1ea0..90cff165c3512 100644
> --- a/tools/testing/selftests/kvm/irq_test.c
> +++ b/tools/testing/selftests/kvm/irq_test.c
[ ... ]
> @@ -187,12 +199,23 @@ int main(int argc, char **argv)
[ ... ]
>  	for (i = 0; i < nr_irqs; i++) {
>  		struct kvm_vcpu *vcpu = vcpus[i % nr_vcpus];
>  		struct timespec start;
>  
>  		kvm_route_msi(vm, gsi, vcpu, vector);
>  
> +		if (irq_affinity_fp) {
> +			irq_cpu = kvm_random_u64(&kvm_rng) % get_nprocs();
> +			write_proc_irq_smp_affinity_list(irq_affinity_fp, irq, irq_cpu);
> +		}

[Severity: Medium]
Will using get_nprocs() here generate invalid CPU IDs on systems with
sparse or non-contiguous CPU topologies?

For example, if only CPUs 0 and 2 are online, get_nprocs() returns 2,
which means the modulo operation could generate an irq_cpu of 1, targeting
an offline or invalid CPU.

Writing an invalid CPU ID fails with -EINVAL, but it looks like
write_proc_irq_smp_affinity_list() uses fprintf(), which only buffers the
data in userspace.

If fflush() returns an error that goes unchecked, could the glibc buffer
retain the unwritten malformed string? Since irq_affinity_fp is reused
across loop iterations, this might cause all subsequent affinity changes to
accumulate and silently fail.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260604020143.748245-1-jrhilke@google.com?part=10