Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: Sean Christopherson <seanjc@google.com>
To: Sean Christopherson <seanjc@google.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
	 Chao Gao <chao.gao@intel.com>, Borislav Petkov <bp@alien8.de>,
	Xin Li <xin@zytor.com>,  Dapeng Mi <dapeng1.mi@linux.intel.com>,
	Francesco Lavra <francescolavra.fl@gmail.com>,
	 Manali Shukla <Manali.Shukla@amd.com>
Subject: [PATCH v2 06/32] KVM: SVM: Kill the VM instead of the host if MSR interception is buggy
Date: Tue, 10 Jun 2025 15:57:11 -0700	[thread overview]
Message-ID: <20250610225737.156318-7-seanjc@google.com> (raw)
In-Reply-To: <20250610225737.156318-1-seanjc@google.com>

WARN and kill the VM instead of panicking the host if KVM attempts to set
or query MSR interception for an unsupported MSR.  Accessing the MSR
interception bitmaps only meaningfully affects post-VMRUN behavior, and
KVM_BUG_ON() is guaranteed to prevent the current vCPU from doing VMRUN,
i.e. there is no need to panic the entire host.

Opportunistically move the sanity checks about their use to index into the
MSRPM, e.g. so that bugs only WARN and terminate the VM, as opposed to
doing that _and_ generating an out-of-bounds load.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 arch/x86/kvm/svm/svm.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index c75977ca600b..7e39b9df61f1 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -824,11 +824,12 @@ static bool msr_write_intercepted(struct kvm_vcpu *vcpu, u32 msr)
 				      to_svm(vcpu)->msrpm;
 
 	offset    = svm_msrpm_offset(msr);
+	if (KVM_BUG_ON(offset == MSR_INVALID, vcpu->kvm))
+		return false;
+
 	bit_write = 2 * (msr & 0x0f) + 1;
 	tmp       = msrpm[offset];
 
-	BUG_ON(offset == MSR_INVALID);
-
 	return test_bit(bit_write, &tmp);
 }
 
@@ -854,12 +855,13 @@ static void set_msr_interception_bitmap(struct kvm_vcpu *vcpu, u32 *msrpm,
 		write = 0;
 
 	offset    = svm_msrpm_offset(msr);
+	if (KVM_BUG_ON(offset == MSR_INVALID, vcpu->kvm))
+		return;
+
 	bit_read  = 2 * (msr & 0x0f);
 	bit_write = 2 * (msr & 0x0f) + 1;
 	tmp       = msrpm[offset];
 
-	BUG_ON(offset == MSR_INVALID);
-
 	read  ? clear_bit(bit_read,  &tmp) : set_bit(bit_read,  &tmp);
 	write ? clear_bit(bit_write, &tmp) : set_bit(bit_write, &tmp);
 
-- 
2.50.0.rc0.642.g800a2b2222-goog


  parent reply	other threads:[~2025-06-10 22:57 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-10 22:57 [PATCH v2 00/32] KVM: x86: Clean up MSR interception code Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 01/32] KVM: SVM: Disable interception of SPEC_CTRL iff the MSR exists for the guest Sean Christopherson
2025-06-11  4:38   ` Binbin Wu
2025-06-11  7:14     ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 02/32] KVM: SVM: Allocate IOPM pages after initial setup in svm_hardware_setup() Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 03/32] KVM: SVM: Don't BUG if setting up the MSR intercept bitmaps fails Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 04/32] KVM: SVM: Tag MSR bitmap initialization helpers with __init Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 05/32] KVM: SVM: Use ARRAY_SIZE() to iterate over direct_access_msrs Sean Christopherson
2025-06-10 22:57 ` Sean Christopherson [this message]
2025-06-11  2:16   ` [PATCH v2 06/32] KVM: SVM: Kill the VM instead of the host if MSR interception is buggy Mi, Dapeng
2025-06-10 22:57 ` [PATCH v2 07/32] KVM: x86: Use non-atomic bit ops to manipulate "shadow" MSR intercepts Sean Christopherson
2025-06-11  6:38   ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 08/32] KVM: SVM: Massage name and param of helper that merges vmcb01 and vmcb12 MSRPMs Sean Christopherson
2025-06-11  2:22   ` Mi, Dapeng
2025-06-10 22:57 ` [PATCH v2 09/32] KVM: SVM: Clean up macros related to architectural MSRPM definitions Sean Christopherson
2025-06-11  6:09   ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 10/32] KVM: nSVM: Use dedicated array of MSRPM offsets to merge L0 and L1 bitmaps Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 11/32] KVM: nSVM: Omit SEV-ES specific passthrough MSRs from L0+L1 bitmap merge Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 12/32] KVM: nSVM: Don't initialize vmcb02 MSRPM with vmcb01's "always passthrough" Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 13/32] KVM: SVM: Add helpers for accessing MSR bitmap that don't rely on offsets Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 14/32] KVM: SVM: Implement and adopt VMX style MSR intercepts APIs Sean Christopherson
2025-06-11  7:31   ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 15/32] KVM: SVM: Pass through GHCB MSR if and only if VM is an SEV-ES guest Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 16/32] KVM: SVM: Drop "always" flag from list of possible passthrough MSRs Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 17/32] KVM: x86: Move definition of X2APIC_MSR() to lapic.h Sean Christopherson
2025-06-11  2:29   ` Mi, Dapeng
2025-06-10 22:57 ` [PATCH v2 18/32] KVM: VMX: Manually recalc all MSR intercepts on userspace MSR filter change Sean Christopherson
2025-06-11  6:52   ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 19/32] KVM: SVM: " Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 20/32] KVM: x86: Rename msr_filter_changed() => recalc_msr_intercepts() Sean Christopherson
2025-06-11  7:09   ` Binbin Wu
2025-06-10 22:57 ` [PATCH v2 21/32] KVM: SVM: Rename init_vmcb_after_set_cpuid() to make it intercepts specific Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 22/32] KVM: SVM: Fold svm_vcpu_init_msrpm() into its sole caller Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 23/32] KVM: SVM: Merge "after set CPUID" intercept recalc helpers Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 24/32] KVM: SVM: Drop explicit check on MSRPM offset when emulating SEV-ES accesses Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 25/32] KVM: SVM: Move svm_msrpm_offset() to nested.c Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 26/32] KVM: SVM: Store MSRPM pointer as "void *" instead of "u32 *" Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 27/32] KVM: nSVM: Access MSRPM in 4-byte chunks only for merging L0 and L1 bitmaps Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 28/32] KVM: SVM: Return -EINVAL instead of MSR_INVALID to signal out-of-range MSR Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 29/32] KVM: nSVM: Merge MSRPM in 64-bit chunks on 64-bit kernels Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 30/32] KVM: SVM: Add a helper to allocate and initialize permissions bitmaps Sean Christopherson
2025-06-10 22:57 ` [PATCH v2 31/32] KVM: x86: Simplify userspace filter logic when disabling MSR interception Sean Christopherson
2025-06-11  2:35   ` Mi, Dapeng
2025-06-10 22:57 ` [PATCH v2 32/32] KVM: selftests: Verify KVM disable interception (for userspace) on filter change Sean Christopherson
2025-06-24 19:38 ` [PATCH v2 00/32] KVM: x86: Clean up MSR interception code Sean Christopherson
2025-06-25 12:03 ` Manali Shukla

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=20250610225737.156318-7-seanjc@google.com \
    --to=seanjc@google.com \
    --cc=Manali.Shukla@amd.com \
    --cc=bp@alien8.de \
    --cc=chao.gao@intel.com \
    --cc=dapeng1.mi@linux.intel.com \
    --cc=francescolavra.fl@gmail.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=xin@zytor.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox