[PATCH v8 19/20] KVM: selftests: Make number of vCPUs configurable in IRQ test
From: Sean Christopherson
Date: Fri Jun 26 2026 - 17:41:54 EST
From: David Matlack <dmatlack@xxxxxxxxxx>
Extend the eventfd IRQ test with a '-v' flag to allow the user to
configure the number of vCPUs to create and run (versus only ever using a
single vCPU).
Update the routing logic to play nice with 32 bit IDs, enable x2APIC format
in KVM (to enable 32-bit ID routing), and disable KVM's x2APIC broadcast
quirk so that targeting vCPU 255 doesn't blast the interrupt to all vCPUs
when in x2APIC mode.
Signed-off-by: David Matlack <dmatlack@xxxxxxxxxx>
Co-developed-by: Josh Hilke <jrhilke@xxxxxxxxxx>
Signed-off-by: Josh Hilke <jrhilke@xxxxxxxxxx>
Co-developed-by: Sean Christopherson <seanjc@xxxxxxxxxx>
Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx>
---
tools/testing/selftests/kvm/irq_test.c | 16 +++++++++++++---
1 file changed, 13 insertions(+), 3 deletions(-)
diff --git a/tools/testing/selftests/kvm/irq_test.c b/tools/testing/selftests/kvm/irq_test.c
index c0ff6e60b167..d2c745c54960 100644
--- a/tools/testing/selftests/kvm/irq_test.c
+++ b/tools/testing/selftests/kvm/irq_test.c
@@ -124,7 +124,8 @@ static void kvm_route_msi(struct kvm_vm *vm, u32 gsi, struct kvm_vcpu *vcpu,
.entry = {
.gsi = gsi,
.type = KVM_IRQ_ROUTING_MSI,
- .u.msi.address_lo = 0xFEE00000 | (vcpu->id << 12),
+ .u.msi.address_lo = 0xFEE00000 | (vcpu->id & GENMASK(7, 0)) << 12,
+ .u.msi.address_hi = vcpu->id & GENMASK(31, 8),
.u.msi.data = use_nmi ? NMI_VECTOR | (4 << 8) : vector,
},
};
@@ -157,7 +158,7 @@ static const char *probe_iommu_type(void)
static void help(const char *name)
{
- printf("Usage: %s [-a] [-d <segment:bus:device.function>] [-e] [-h] [-i nr_irqs] [-m] [-n] [-t iommu_type]\n", name);
+ printf("Usage: %s [-a] [-d <segment:bus:device.function>] [-e] [-h] [-i nr_irqs] [-m] [-n] [-t iommu_type] [-v nr_vcpus]\n", name);
printf("\n");
printf("Tests KVM interrupt routing and delivery via irqfd.\n");
printf("-a Affine the device's host IRQ to a random physical CPU\n");
@@ -167,6 +168,7 @@ static void help(const char *name)
printf("-m Pin target vCPU to random physical CPU before triggering interrupt\n");
printf("-n Deliver 50 percent of IRQs as non-maskable interrupts\n");
printf("-t Override the IOMMU type to use (vfio_type1_iommu or iommufd)\n");
+ printf("-v Number of vCPUS to run\n");
printf("\n");
exit(KSFT_FAIL);
}
@@ -203,7 +205,7 @@ int main(int argc, char **argv)
struct kvm_vm *vm;
int irq, irq_cpu;
- while ((c = getopt(argc, argv, "ad:ehi:mnt:")) != -1) {
+ while ((c = getopt(argc, argv, "ad:ehi:mnt:v:")) != -1) {
switch (c) {
case 'a':
irq_affinity = true;
@@ -226,6 +228,11 @@ int main(int argc, char **argv)
case 't':
iommu_type = optarg;
break;
+ case 'v':
+ nr_vcpus = atoi_positive("Number of vCPUS", optarg);
+ TEST_ASSERT(nr_vcpus <= KVM_MAX_VCPUS,
+ "KVM selftests support at most %u vCPUs", KVM_MAX_VCPUS);
+ break;
case 'h':
default:
help(argv[0]);
@@ -235,6 +242,9 @@ int main(int argc, char **argv)
TEST_REQUIRE(kvm_arch_has_default_irqchip());
vm = vm_create_with_vcpus(nr_vcpus, guest_code, vcpus);
+ vm_enable_cap(vm, KVM_CAP_X2APIC_API, KVM_X2APIC_API_USE_32BIT_IDS |
+ KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK);
+
vm_install_exception_handler(vm, vector, guest_irq_handler);
vm_install_exception_handler(vm, NMI_VECTOR, guest_nmi_handler);
--
2.55.0.rc0.799.gd6f94ed593-goog