Re: [PATCH] selftests: KVM: Gracefully handle missing vCPU features

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Aug 18, 2021 at 09:29:40PM +0000, Oliver Upton wrote:
> An error of ENOENT for the KVM_ARM_VCPU_INIT ioctl indicates that one of
> the requested feature flags is not supported by the kernel/hardware.
> Detect the case when KVM doesn't support the requested features and skip
> the test rather than failing it.
> 
> Cc: Andrew Jones <drjones@xxxxxxxxxx>
> Signed-off-by: Oliver Upton <oupton@xxxxxxxxxx>
> ---
> Applies to 5.14-rc6. Tested by running all selftests on an Ampere Mt.
> Jade system.
> 
>  .../testing/selftests/kvm/lib/aarch64/processor.c | 15 ++++++++++++++-
>  1 file changed, 14 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/kvm/lib/aarch64/processor.c b/tools/testing/selftests/kvm/lib/aarch64/processor.c
> index 632b74d6b3ca..b1064a0c5e62 100644
> --- a/tools/testing/selftests/kvm/lib/aarch64/processor.c
> +++ b/tools/testing/selftests/kvm/lib/aarch64/processor.c
> @@ -216,6 +216,7 @@ void aarch64_vcpu_setup(struct kvm_vm *vm, int vcpuid, struct kvm_vcpu_init *ini
>  {
>  	struct kvm_vcpu_init default_init = { .target = -1, };
>  	uint64_t sctlr_el1, tcr_el1;
> +	int ret;
>  
>  	if (!init)
>  		init = &default_init;
> @@ -226,7 +227,19 @@ void aarch64_vcpu_setup(struct kvm_vm *vm, int vcpuid, struct kvm_vcpu_init *ini
>  		init->target = preferred.target;
>  	}
>  
> -	vcpu_ioctl(vm, vcpuid, KVM_ARM_VCPU_INIT, init);
> +	ret = _vcpu_ioctl(vm, vcpuid, KVM_ARM_VCPU_INIT, init);
> +
> +	/*
> +	 * Missing kernel feature support should result in skipping the test,
> +	 * not failing it.
> +	 */
> +	if (ret && errno == ENOENT) {
> +		print_skip("requested vCPU features not supported; skipping test.");

", skipping test" will already be appended by print_skip().

> +		exit(KSFT_SKIP);
> +	}
> +
> +	TEST_ASSERT(!ret, "KVM_ARM_VCPU_INIT failed, rc: %i errno: %i (%s)",
> +		    ret, errno, strerror(errno));
>  
>  	/*
>  	 * Enable FP/ASIMD to avoid trapping when accessing Q0-Q15
> -- 
> 2.33.0.rc1.237.g0d66db33f3-goog
>

I think I'd rather try to keep exit(KSFT_SKIP)'s out of the lib code. It'd
be better if the test gets to decide whether to skip or not. How about
moving this check+skip into the test instead?

Thanks,
drew

_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm



[Index of Archives]     [Linux KVM]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux