LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Vitaly Kuznetsov <vkuznets@redhat.com>
To: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>
Cc: Sean Christopherson <seanjc@google.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>,
	linux-kernel@vger.kernel.org
Subject: [PATCH 2/2] KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN
Date: Mon, 22 Nov 2021 18:58:18 +0100	[thread overview]
Message-ID: <20211122175818.608220-3-vkuznets@redhat.com> (raw)
In-Reply-To: <20211122175818.608220-1-vkuznets@redhat.com>

Commit 63f5a1909f9e ("KVM: x86: Alert userspace that KVM_SET_CPUID{,2}
after KVM_RUN is broken") officially deprecated KVM_SET_CPUID{,2} ioctls
after first successful KVM_RUN and promissed to make this sequence forbiden
in 5.16. It's time to fulfil the promise.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 arch/x86/kvm/mmu/mmu.c | 20 +++-----------------
 arch/x86/kvm/x86.c     | 27 +++++++++++++++++++++++++++
 2 files changed, 30 insertions(+), 17 deletions(-)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 3be9beea838d..669e86688cbf 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -5032,24 +5032,10 @@ void kvm_mmu_after_set_cpuid(struct kvm_vcpu *vcpu)
 	kvm_mmu_reset_context(vcpu);
 
 	/*
-	 * KVM does not correctly handle changing guest CPUID after KVM_RUN, as
-	 * MAXPHYADDR, GBPAGES support, AMD reserved bit behavior, etc.. aren't
-	 * tracked in kvm_mmu_page_role.  As a result, KVM may miss guest page
-	 * faults due to reusing SPs/SPTEs.  Alert userspace, but otherwise
-	 * sweep the problem under the rug.
-	 *
-	 * KVM's horrific CPUID ABI makes the problem all but impossible to
-	 * solve, as correctly handling multiple vCPU models (with respect to
-	 * paging and physical address properties) in a single VM would require
-	 * tracking all relevant CPUID information in kvm_mmu_page_role.  That
-	 * is very undesirable as it would double the memory requirements for
-	 * gfn_track (see struct kvm_mmu_page_role comments), and in practice
-	 * no sane VMM mucks with the core vCPU model on the fly.
+	 * Changing guest CPUID after KVM_RUN is forbidden, see the comment in
+	 * kvm_arch_vcpu_ioctl().
 	 */
-	if (vcpu->arch.last_vmentry_cpu != -1) {
-		pr_warn_ratelimited("KVM: KVM_SET_CPUID{,2} after KVM_RUN may cause guest instability\n");
-		pr_warn_ratelimited("KVM: KVM_SET_CPUID{,2} will fail after KVM_RUN starting with Linux 5.16\n");
-	}
+	KVM_BUG_ON(vcpu->arch.last_vmentry_cpu != -1, vcpu->kvm);
 }
 
 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 5a403d92833f..3cfaccc24efb 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -5125,6 +5125,25 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
 		struct kvm_cpuid cpuid;
 
 		r = -EFAULT;
+
+		/*
+		 * KVM does not correctly handle changing guest CPUID after KVM_RUN, as
+		 * MAXPHYADDR, GBPAGES support, AMD reserved bit behavior, etc.. aren't
+		 * tracked in kvm_mmu_page_role.  As a result, KVM may miss guest page
+		 * faults due to reusing SPs/SPTEs.  Alert userspace, but otherwise
+		 * sweep the problem under the rug.
+		 *
+		 * KVM's horrific CPUID ABI makes the problem all but impossible to
+		 * solve, as correctly handling multiple vCPU models (with respect to
+		 * paging and physical address properties) in a single VM would require
+		 * tracking all relevant CPUID information in kvm_mmu_page_role.  That
+		 * is very undesirable as it would double the memory requirements for
+		 * gfn_track (see struct kvm_mmu_page_role comments), and in practice
+		 * no sane VMM mucks with the core vCPU model on the fly.
+		 */
+		if (vcpu->arch.last_vmentry_cpu != -1)
+			goto out;
+
 		if (copy_from_user(&cpuid, cpuid_arg, sizeof(cpuid)))
 			goto out;
 		r = kvm_vcpu_ioctl_set_cpuid(vcpu, &cpuid, cpuid_arg->entries);
@@ -5135,6 +5154,14 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
 		struct kvm_cpuid2 cpuid;
 
 		r = -EFAULT;
+
+		/*
+		 * KVM_SET_CPUID{,2} after KVM_RUN is forbidded, see the comment in
+		 * KVM_SET_CPUID case above.
+		 */
+		if (vcpu->arch.last_vmentry_cpu != -1)
+			goto out;
+
 		if (copy_from_user(&cpuid, cpuid_arg, sizeof(cpuid)))
 			goto out;
 		r = kvm_vcpu_ioctl_set_cpuid2(vcpu, &cpuid,
-- 
2.33.1


  parent reply	other threads:[~2021-11-22 17:58 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-22 17:58 [PATCH 0/2] " Vitaly Kuznetsov
2021-11-22 17:58 ` [PATCH 1/2] KVM: selftests: Avoid KVM_SET_CPUID2 after KVM_RUN in hyperv_features test Vitaly Kuznetsov
2021-11-22 17:58 ` Vitaly Kuznetsov [this message]
2021-11-26 12:20   ` [PATCH 2/2] KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN Paolo Bonzini
2021-12-27 17:32     ` Igor Mammedov
2022-01-02 17:31       ` Paolo Bonzini
2022-01-03  8:04         ` Vitaly Kuznetsov
2022-01-03  9:40           ` Igor Mammedov
2022-01-03 12:56             ` Vitaly Kuznetsov
2022-01-05  8:17               ` Igor Mammedov
2022-01-05  9:12                 ` Vitaly Kuznetsov
2022-01-05  9:10               ` Paolo Bonzini
2022-01-05 10:09                 ` Vitaly Kuznetsov
2022-01-07  9:02                   ` Vitaly Kuznetsov
2022-01-07 18:15                     ` Paolo Bonzini
2022-01-11  8:00                       ` Igor Mammedov
2022-01-12 13:58                         ` Vitaly Kuznetsov
2022-01-12 18:39                           ` Paolo Bonzini
2022-01-13  9:27                             ` Vitaly Kuznetsov
2022-01-13 14:28                               ` Maxim Levitsky
2022-01-13 14:36                                 ` Vitaly Kuznetsov
2022-01-13 14:41                                   ` Maxim Levitsky
2022-01-13 14:59                                     ` Vitaly Kuznetsov
2022-01-13 16:26                                       ` Sean Christopherson
2022-01-13 16:30                                         ` Maxim Levitsky
2022-01-13 22:33             ` Sean Christopherson
2022-01-14  8:28               ` Maxim Levitsky
2022-01-14 16:08                 ` Sean Christopherson
2022-01-14  8:55               ` Igor Mammedov
2022-01-14  9:31                 ` Vitaly Kuznetsov
2022-01-14 11:22                   ` Igor Mammedov
2022-01-14 12:25                     ` Vitaly Kuznetsov
2022-01-14 17:00                       ` Sean Christopherson
2022-01-17  9:55                         ` Vitaly Kuznetsov
2022-01-17 11:20                           ` Paolo Bonzini
2022-01-17 13:02                             ` Vitaly Kuznetsov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211122175818.608220-3-vkuznets@redhat.com \
    --to=vkuznets@redhat.com \
    --cc=jmattson@google.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=seanjc@google.com \
    --cc=wanpengli@tencent.com \
    --subject='Re: [PATCH 2/2] KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).