The arch_timer and vgic_irq kselftests assume that they can create a vgic-v3, using the library function vgic_v3_setup() which aborts with a test failure if it is not possible to do so. Since vgic-v3 can only be instantiated on systems where the host has GICv3 this leads to false positives on older systems where that is not the case.
Fix this by changing vgic_v3_setup() to return an error if the vgic can't be instantiated and have the callers skip if this happens. We could also exit flagging a skip in vgic_v3_setup() but this would prevent future test cases conditionally deciding which GIC to use or generally doing more complex output.
Signed-off-by: Mark Brown broonie@kernel.org --- tools/testing/selftests/kvm/aarch64/arch_timer.c | 7 ++++++- tools/testing/selftests/kvm/aarch64/vgic_irq.c | 5 +++++ tools/testing/selftests/kvm/lib/aarch64/vgic.c | 4 +++- 3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/kvm/aarch64/arch_timer.c b/tools/testing/selftests/kvm/aarch64/arch_timer.c index 9ad38bd360a4..791d38404652 100644 --- a/tools/testing/selftests/kvm/aarch64/arch_timer.c +++ b/tools/testing/selftests/kvm/aarch64/arch_timer.c @@ -366,6 +366,7 @@ static struct kvm_vm *test_vm_create(void) { struct kvm_vm *vm; unsigned int i; + int ret; int nr_vcpus = test_args.nr_vcpus;
vm = vm_create_default_with_vcpus(nr_vcpus, 0, 0, guest_code, NULL); @@ -382,7 +383,11 @@ static struct kvm_vm *test_vm_create(void)
ucall_init(vm, NULL); test_init_timer_irq(vm); - vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA); + ret = vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA); + if (ret < 0) { + pr_info("Failed to create vgic-v3, skipping\n"); + exit(KSFT_SKIP); + }
/* Make all the test's cmdline args visible to the guest */ sync_global_to_guest(vm, test_args); diff --git a/tools/testing/selftests/kvm/aarch64/vgic_irq.c b/tools/testing/selftests/kvm/aarch64/vgic_irq.c index e6c7d7f8fbd1..8c6b61b8e6aa 100644 --- a/tools/testing/selftests/kvm/aarch64/vgic_irq.c +++ b/tools/testing/selftests/kvm/aarch64/vgic_irq.c @@ -761,6 +761,11 @@ static void test_vgic(uint32_t nr_irqs, bool level_sensitive, bool eoi_split)
gic_fd = vgic_v3_setup(vm, 1, nr_irqs, GICD_BASE_GPA, GICR_BASE_GPA); + if (gic_fd < 0) { + pr_info("Failed to create vgic-v3, skipping\n"); + exit(KSFT_SKIP); + } +
vm_install_exception_handler(vm, VECTOR_IRQ_CURRENT, guest_irq_handlers[args.eoi_split][args.level_sensitive]); diff --git a/tools/testing/selftests/kvm/lib/aarch64/vgic.c b/tools/testing/selftests/kvm/lib/aarch64/vgic.c index b3a0fca0d780..647c18733e1b 100644 --- a/tools/testing/selftests/kvm/lib/aarch64/vgic.c +++ b/tools/testing/selftests/kvm/lib/aarch64/vgic.c @@ -52,7 +52,9 @@ int vgic_v3_setup(struct kvm_vm *vm, unsigned int nr_vcpus, uint32_t nr_irqs, nr_vcpus, nr_vcpus_created);
/* Distributor setup */ - gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false); + gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true); + if (gic_fd == -1) + return -1;
kvm_device_access(gic_fd, KVM_DEV_ARM_VGIC_GRP_NR_IRQS, 0, &nr_irqs, true);
On 1/25/22 12:28 PM, Mark Brown wrote:
The arch_timer and vgic_irq kselftests assume that they can create a vgic-v3, using the library function vgic_v3_setup() which aborts with a test failure if it is not possible to do so. Since vgic-v3 can only be instantiated on systems where the host has GICv3 this leads to false positives on older systems where that is not the case.
Fix this by changing vgic_v3_setup() to return an error if the vgic can't be instantiated and have the callers skip if this happens. We could also exit flagging a skip in vgic_v3_setup() but this would prevent future test cases conditionally deciding which GIC to use or generally doing more complex output.
Signed-off-by: Mark Brown broonie@kernel.org
tools/testing/selftests/kvm/aarch64/arch_timer.c | 7 ++++++- tools/testing/selftests/kvm/aarch64/vgic_irq.c | 5 +++++ tools/testing/selftests/kvm/lib/aarch64/vgic.c | 4 +++- 3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/kvm/aarch64/arch_timer.c b/tools/testing/selftests/kvm/aarch64/arch_timer.c index 9ad38bd360a4..791d38404652 100644 --- a/tools/testing/selftests/kvm/aarch64/arch_timer.c +++ b/tools/testing/selftests/kvm/aarch64/arch_timer.c @@ -366,6 +366,7 @@ static struct kvm_vm *test_vm_create(void) { struct kvm_vm *vm; unsigned int i;
- int ret; int nr_vcpus = test_args.nr_vcpus;
vm = vm_create_default_with_vcpus(nr_vcpus, 0, 0, guest_code, NULL); @@ -382,7 +383,11 @@ static struct kvm_vm *test_vm_create(void) ucall_init(vm, NULL); test_init_timer_irq(vm);
- vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA);
- ret = vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA);
- if (ret < 0) {
pr_info("Failed to create vgic-v3, skipping\n");exit(KSFT_SKIP);- }
/* Make all the test's cmdline args visible to the guest */ sync_global_to_guest(vm, test_args); diff --git a/tools/testing/selftests/kvm/aarch64/vgic_irq.c b/tools/testing/selftests/kvm/aarch64/vgic_irq.c index e6c7d7f8fbd1..8c6b61b8e6aa 100644 --- a/tools/testing/selftests/kvm/aarch64/vgic_irq.c +++ b/tools/testing/selftests/kvm/aarch64/vgic_irq.c @@ -761,6 +761,11 @@ static void test_vgic(uint32_t nr_irqs, bool level_sensitive, bool eoi_split) gic_fd = vgic_v3_setup(vm, 1, nr_irqs, GICD_BASE_GPA, GICR_BASE_GPA);
- if (gic_fd < 0) {
pr_info("Failed to create vgic-v3, skipping\n");exit(KSFT_SKIP);- }
vm_install_exception_handler(vm, VECTOR_IRQ_CURRENT, guest_irq_handlers[args.eoi_split][args.level_sensitive]); diff --git a/tools/testing/selftests/kvm/lib/aarch64/vgic.c b/tools/testing/selftests/kvm/lib/aarch64/vgic.c index b3a0fca0d780..647c18733e1b 100644 --- a/tools/testing/selftests/kvm/lib/aarch64/vgic.c +++ b/tools/testing/selftests/kvm/lib/aarch64/vgic.c @@ -52,7 +52,9 @@ int vgic_v3_setup(struct kvm_vm *vm, unsigned int nr_vcpus, uint32_t nr_irqs, nr_vcpus, nr_vcpus_created); /* Distributor setup */
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false);
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true);
Assume the flag change is intentional. The reason isn't obvious in the change log
- if (gic_fd == -1)
return -1;kvm_device_access(gic_fd, KVM_DEV_ARM_VGIC_GRP_NR_IRQS, 0, &nr_irqs, true);
Looks good to me otherwise - thanks for adding the skips
Reviewed-by: Shuah Khan skhan@linuxfoundation.org
thanks, -- Shuah
On Tue, Jan 25, 2022 at 07:28:51PM +0000, Mark Brown wrote:
The arch_timer and vgic_irq kselftests assume that they can create a vgic-v3, using the library function vgic_v3_setup() which aborts with a test failure if it is not possible to do so. Since vgic-v3 can only be instantiated on systems where the host has GICv3 this leads to false positives on older systems where that is not the case.
Fix this by changing vgic_v3_setup() to return an error if the vgic can't be instantiated and have the callers skip if this happens. We could also exit flagging a skip in vgic_v3_setup() but this would prevent future test cases conditionally deciding which GIC to use or generally doing more complex output.
Signed-off-by: Mark Brown broonie@kernel.org
tools/testing/selftests/kvm/aarch64/arch_timer.c | 7 ++++++- tools/testing/selftests/kvm/aarch64/vgic_irq.c | 5 +++++ tools/testing/selftests/kvm/lib/aarch64/vgic.c | 4 +++- 3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/kvm/aarch64/arch_timer.c b/tools/testing/selftests/kvm/aarch64/arch_timer.c index 9ad38bd360a4..791d38404652 100644 --- a/tools/testing/selftests/kvm/aarch64/arch_timer.c +++ b/tools/testing/selftests/kvm/aarch64/arch_timer.c @@ -366,6 +366,7 @@ static struct kvm_vm *test_vm_create(void) { struct kvm_vm *vm; unsigned int i;
- int ret; int nr_vcpus = test_args.nr_vcpus;
vm = vm_create_default_with_vcpus(nr_vcpus, 0, 0, guest_code, NULL); @@ -382,7 +383,11 @@ static struct kvm_vm *test_vm_create(void) ucall_init(vm, NULL); test_init_timer_irq(vm);
- vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA);
- ret = vgic_v3_setup(vm, nr_vcpus, 64, GICD_BASE_GPA, GICR_BASE_GPA);
- if (ret < 0) {
pr_info("Failed to create vgic-v3, skipping\n");exit(KSFT_SKIP);- }
/* Make all the test's cmdline args visible to the guest */ sync_global_to_guest(vm, test_args); diff --git a/tools/testing/selftests/kvm/aarch64/vgic_irq.c b/tools/testing/selftests/kvm/aarch64/vgic_irq.c index e6c7d7f8fbd1..8c6b61b8e6aa 100644 --- a/tools/testing/selftests/kvm/aarch64/vgic_irq.c +++ b/tools/testing/selftests/kvm/aarch64/vgic_irq.c @@ -761,6 +761,11 @@ static void test_vgic(uint32_t nr_irqs, bool level_sensitive, bool eoi_split) gic_fd = vgic_v3_setup(vm, 1, nr_irqs, GICD_BASE_GPA, GICR_BASE_GPA);
- if (gic_fd < 0) {
pr_info("Failed to create vgic-v3, skipping\n");exit(KSFT_SKIP);- }
vm_install_exception_handler(vm, VECTOR_IRQ_CURRENT, guest_irq_handlers[args.eoi_split][args.level_sensitive]); diff --git a/tools/testing/selftests/kvm/lib/aarch64/vgic.c b/tools/testing/selftests/kvm/lib/aarch64/vgic.c index b3a0fca0d780..647c18733e1b 100644 --- a/tools/testing/selftests/kvm/lib/aarch64/vgic.c +++ b/tools/testing/selftests/kvm/lib/aarch64/vgic.c @@ -52,7 +52,9 @@ int vgic_v3_setup(struct kvm_vm *vm, unsigned int nr_vcpus, uint32_t nr_irqs, nr_vcpus, nr_vcpus_created); /* Distributor setup */
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false);
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true);
- if (gic_fd == -1)
return -1;kvm_device_access(gic_fd, KVM_DEV_ARM_VGIC_GRP_NR_IRQS, 0, &nr_irqs, true); -- 2.30.2
kvmarm mailing list kvmarm@lists.cs.columbia.edu https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
Reviewed-by: Ricardo Koller ricarkol@google.com
On Tue, 25 Jan 2022 19:28:51 +0000, Mark Brown broonie@kernel.org wrote:
The arch_timer and vgic_irq kselftests assume that they can create a vgic-v3, using the library function vgic_v3_setup() which aborts with a test failure if it is not possible to do so. Since vgic-v3 can only be instantiated on systems where the host has GICv3 this leads to false positives on older systems where that is not the case.
Fix this by changing vgic_v3_setup() to return an error if the vgic can't be instantiated and have the callers skip if this happens. We could also exit flagging a skip in vgic_v3_setup() but this would prevent future test cases conditionally deciding which GIC to use or generally doing more complex output.
Signed-off-by: Mark Brown broonie@kernel.org
tools/testing/selftests/kvm/aarch64/arch_timer.c | 7 ++++++- tools/testing/selftests/kvm/aarch64/vgic_irq.c | 5 +++++ tools/testing/selftests/kvm/lib/aarch64/vgic.c | 4 +++- 3 files changed, 14 insertions(+), 2 deletions(-)
[...]
diff --git a/tools/testing/selftests/kvm/lib/aarch64/vgic.c b/tools/testing/selftests/kvm/lib/aarch64/vgic.c index b3a0fca0d780..647c18733e1b 100644 --- a/tools/testing/selftests/kvm/lib/aarch64/vgic.c +++ b/tools/testing/selftests/kvm/lib/aarch64/vgic.c @@ -52,7 +52,9 @@ int vgic_v3_setup(struct kvm_vm *vm, unsigned int nr_vcpus, uint32_t nr_irqs, nr_vcpus, nr_vcpus_created); /* Distributor setup */
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false);
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true);
So you now only test whether it is possible to create a virtual GICv3, but don't actually create it. How does this work?
M.
On Wed, Jan 26, 2022 at 09:07:48AM +0000, Marc Zyngier wrote:
Mark Brown broonie@kernel.org wrote:
/* Distributor setup */
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false);
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true);
So you now only test whether it is possible to create a virtual GICv3, but don't actually create it. How does this work?
Oh, that's rather obscure in the API - so the file descriptor returned if the test flag is specified can't actually be used?
On Wed, 26 Jan 2022 12:52:22 +0000, Mark Brown broonie@kernel.org wrote:
On Wed, Jan 26, 2022 at 09:07:48AM +0000, Marc Zyngier wrote:
Mark Brown broonie@kernel.org wrote:
/* Distributor setup */
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, false);
- gic_fd = kvm_create_device(vm, KVM_DEV_TYPE_ARM_VGIC_V3, true);
So you now only test whether it is possible to create a virtual GICv3, but don't actually create it. How does this work?
Oh, that's rather obscure in the API - so the file descriptor returned if the test flag is specified can't actually be used?
No file descriptor is returned at all from the kernel, so you should always get -1 as populated by kvm_create_device().
See virt/kvm/kvm_main.c::kvm_ioctl_create_device().
M.
On Wed, Jan 26, 2022 at 01:05:01PM +0000, Marc Zyngier wrote:
Mark Brown broonie@kernel.org wrote:
So you now only test whether it is possible to create a virtual GICv3, but don't actually create it. How does this work?
Oh, that's rather obscure in the API - so the file descriptor returned if the test flag is specified can't actually be used?
No file descriptor is returned at all from the kernel, so you should always get -1 as populated by kvm_create_device().
See virt/kvm/kvm_main.c::kvm_ioctl_create_device().
Ugh, right - the descriptor is returned separately to the return code but is then mapped onto the return code by the library since the library just aborts on error in the non-test case and in the test case the kernel returns an inverted boolean which gets passed straight through. Like I say the API seems a bit confusing here.
aOn Wed, 26 Jan 2022 13:21:17 +0000, Mark Brown broonie@kernel.org wrote:
[1 <text/plain; us-ascii (7bit)>] On Wed, Jan 26, 2022 at 01:05:01PM +0000, Marc Zyngier wrote:
Mark Brown broonie@kernel.org wrote:
So you now only test whether it is possible to create a virtual GICv3, but don't actually create it. How does this work?
Oh, that's rather obscure in the API - so the file descriptor returned if the test flag is specified can't actually be used?
No file descriptor is returned at all from the kernel, so you should always get -1 as populated by kvm_create_device().
See virt/kvm/kvm_main.c::kvm_ioctl_create_device().
Ugh, right - the descriptor is returned separately to the return code but is then mapped onto the return code by the library since the library just aborts on error in the non-test case and in the test case the kernel returns an inverted boolean which gets passed straight through. Like I say the API seems a bit confusing here.
That's a very... "kind" way of putting it.
M.
linux-kselftest-mirror@lists.linaro.org
-
Marc Zyngier -
Mark Brown -
Ricardo Koller -
Shuah Khan