Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 24 May 2023 16:50:54 -0700
In-Reply-To: <20230420104622.12504-8-ljrcore@126.com>
Mime-Version: 1.0
References: <20230420104622.12504-1-ljrcore@126.com> <20230420104622.12504-8-ljrcore@126.com>
Message-ID: <ZG6i3sqOcZDg/UCG@google.com>
Subject: Re: [PATCH v2 7/7] KVM: selftests: Test pmu event filter with
 incompatible kvm_pmu_event_filter
From:   Sean Christopherson <seanjc@google.com>
To:     Jinrong Liang <ljr.kernel@gmail.com>
Cc:     Like Xu <like.xu.linux@gmail.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Jonathan Corbet <corbet@lwn.net>,
        Shuah Khan <shuah@kernel.org>,
        Aaron Lewis <aaronlewis@google.com>,
        David Matlack <dmatlack@google.com>,
        Vishal Annapurve <vannapurve@google.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Bagas Sanjaya <bagasdotme@gmail.com>,
        Jinrong Liang <cloudliang@tencent.com>,
        linux-kselftest@vger.kernel.org, linux-doc@vger.kernel.org,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="us-ascii"
Precedence: bulk

On Thu, Apr 20, 2023, Jinrong Liang wrote:
> From: Jinrong Liang <cloudliang@tencent.com>
> 
> From: Jinrong Liang <cloudliang@tencent.com>
> 
> Add test to verify the behavior of the pmu event filter when an
> incomplete kvm_pmu_event_filter structure is used. By running the
> test, we can ensure that the pmu event filter correctly handles
> incomplete structures and does not allow events to be counted when
> they should not be.
> 
> Signed-off-by: Jinrong Liang <cloudliang@tencent.com>
> ---
>  .../kvm/x86_64/pmu_event_filter_test.c        | 23 +++++++++++++++++++
>  1 file changed, 23 insertions(+)
> 
> diff --git a/tools/testing/selftests/kvm/x86_64/pmu_event_filter_test.c b/tools/testing/selftests/kvm/x86_64/pmu_event_filter_test.c
> index 9be4c6f8fb7e..a6b6e0d086ae 100644
> --- a/tools/testing/selftests/kvm/x86_64/pmu_event_filter_test.c
> +++ b/tools/testing/selftests/kvm/x86_64/pmu_event_filter_test.c
> @@ -881,6 +881,24 @@ static bool fixed_ctr_is_allowed(uint8_t idx, uint32_t action, uint32_t bitmap)
>  		(action == KVM_PMU_EVENT_DENY && !(bitmap & BIT_ULL(idx)));
>  }
>  
> +struct incompatible_pmu_event_filter {
> +	__u32 action;
> +	__u32 nevents;
> +	__u32 fixed_counter_bitmap;
> +};
> +
> +static uint64_t test_incompatible_filter(struct kvm_vcpu *vcpu, uint32_t action,
> +					 uint32_t bitmap)
> +{
> +	struct incompatible_pmu_event_filter err_f;
> +
> +	err_f.action = action;
> +	err_f.fixed_counter_bitmap = bitmap;
> +	ioctl((vcpu->vm)->fd, KVM_SET_PMU_EVENT_FILTER, &err_f.action);

This is completely busted.  It "passes" by luck, not because it's a valid test.
The size of the argument is embedded in the IOCTL number itself, which means that
unless glibc is being very nice and using a macro + typeof + sizeof to sanity check
things, which I highly doubt is the case, this ioctl() is passing random stack data,
a.k.a. garbage, to KVM.

In short, drop this patch.