Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Thu, 5 May 2022 19:14:48 +0000
From:   Sean Christopherson <seanjc@google.com>
To:     Ben Gardon <bgardon@google.com>
Cc:     linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        Paolo Bonzini <pbonzini@redhat.com>,
        Peter Xu <peterx@redhat.com>,
        David Matlack <dmatlack@google.com>,
        Jim Mattson <jmattson@google.com>,
        David Dunn <daviddunn@google.com>,
        Jing Zhang <jingzhangos@google.com>,
        Junaid Shahid <junaids@google.com>
Subject: Re: [PATCH v7 10/11] KVM: selftests: Test disabling NX hugepages on
 a VM
Message-ID: <YnQiKJcufscYYq/j@google.com>
References: <20220503183045.978509-1-bgardon@google.com>
 <20220503183045.978509-11-bgardon@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220503183045.978509-11-bgardon@google.com>
Precedence: bulk

On Tue, May 03, 2022, Ben Gardon wrote:
> +	if (disable_nx_huge_pages) {
> +		/*
> +		 * Cannot run the test without NX huge pages if the kernel
> +		 * does not support it.
> +		 */
> +		if (!kvm_check_cap(KVM_CAP_VM_DISABLE_NX_HUGE_PAGES))
> +			return;
> +
> +		r = __vm_disable_nx_huge_pages(vm);
> +		if (reboot_permissions) {
> +			TEST_ASSERT(!r, "Disabling NX huge pages should succeed if process has reboot permissions");
> +		} else {
> +			TEST_ASSERT(r == -EPERM, "This process should not have permission to disable NX huge pages");

This is wrong, the return value on ioctl() failure is -1, the error code is
in errno and it's a positive value.

LOL, but it passes because EPERM == 1, hilarious.  To avoid confusion:

			TEST_ASSERT(r == -1 && errno == EPERM,
				    "This process should not have permission to disable NX huge pages");

> diff --git a/tools/testing/selftests/kvm/x86_64/nx_huge_pages_test.sh b/tools/testing/selftests/kvm/x86_64/nx_huge_pages_test.sh
> index 60bfed8181b9..c21c1f639141 100755
> --- a/tools/testing/selftests/kvm/x86_64/nx_huge_pages_test.sh
> +++ b/tools/testing/selftests/kvm/x86_64/nx_huge_pages_test.sh
> @@ -16,6 +16,8 @@ HUGE_PAGES=$(sudo cat /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages)
>  
>  set +e
>  
> +NXECUTABLE="$(dirname $0)/nx_huge_pages_test"
> +
>  (
>  	set -e
>  
> @@ -24,7 +26,15 @@ set +e
>  	sudo echo 100 > /sys/module/kvm/parameters/nx_huge_pages_recovery_period_ms
>  	sudo echo 3 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
>  
> -	"$(dirname $0)"/nx_huge_pages_test 887563923
> +	# Test with reboot permissions
> +	sudo setcap cap_sys_boot+ep $NXECUTABLE

This leaves cap_sys_boot set on the executable if the script is run as root.

Probably this?  It's moderately user friendly without going too crazy on error
handling.

	# Test with reboot permissions
	if [ $(whoami) != "root" ] ; then
		sudo setcap cap_sys_boot+ep $NXECUTABLE
	fi
	$NXECUTABLE 887563923 1

	# Test without reboot permissions
	if [ $(whoami) != "root" ] ; then
		sudo setcap cap_sys_boot-ep $NXECUTABLE
		$NXECUTABLE 887563923 0
	fi

> +	$NXECUTABLE 887563923 1
> +
> +	# Test without reboot permissions
> +	if [ $(whoami) != "root" ] ; then
> +		sudo setcap cap_sys_boot-ep $NXECUTABLE
> +		$NXECUTABLE 887563923 0

I would much prefer a proper flag, not a magic 0 vs. 1.  

> +	fi
>  )
>  RET=$?
>  
> -- 
> 2.36.0.464.gb9c8b46e94-goog
>