[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20251108004524.1600006-2-yosry.ahmed@linux.dev>
Date: Sat, 8 Nov 2025 00:45:19 +0000
From: Yosry Ahmed <yosry.ahmed@...ux.dev>
To: Sean Christopherson <seanjc@...gle.com>
Cc: Paolo Bonzini <pbonzini@...hat.com>,
Jim Mattson <jmattson@...gle.com>,
Maxim Levitsky <mlevitsk@...hat.com>,
kvm@...r.kernel.org,
linux-kernel@...r.kernel.org,
Yosry Ahmed <yosry.ahmed@...ux.dev>,
Matteo Rizzo <matteorizzo@...gle.com>,
evn@...gle.com
Subject: [PATCH 1/6] KVM: SVM: Mark VMCB_LBR dirty when MSR_IA32_DEBUGCTLMSR is updated
Clear the VMCB_LBR clean bit when MSR_IA32_DEBUGCTLMSR is updated, as
the only valid bit is DEBUGCTLMSR_LBR.
The history is complicated, it was correctly cleared for L1 before
commit 1d5a1b5860ed ("KVM: x86: nSVM: correctly virtualize LBR msrs when
L2 is running"), then the latter relied on svm_update_lbrv() to clear
it, but it only did so for L2. Go back to clearing it directly in
svm_set_msr().
Fixes: 1d5a1b5860ed ("KVM: x86: nSVM: correctly virtualize LBR msrs when L2 is running")
Reported-by: Matteo Rizzo <matteorizzo@...gle.com>
Reported-by: evn@...gle.com
Co-developed-by: Jim Mattson <jmattson@...gle.com>
Signed-off-by: Jim Mattson <jmattson@...gle.com>
Signed-off-by: Yosry Ahmed <yosry.ahmed@...ux.dev>
---
arch/x86/kvm/svm/svm.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 55bd7aa5cd743..d25c56b30b4e2 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -3009,7 +3009,11 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr)
if (data & DEBUGCTL_RESERVED_BITS)
return 1;
+ if (svm_get_lbr_vmcb(svm)->save.dbgctl == data)
+ break;
+
svm_get_lbr_vmcb(svm)->save.dbgctl = data;
+ vmcb_mark_dirty(svm->vmcb, VMCB_LBR);
svm_update_lbrv(vcpu);
break;
case MSR_VM_HSAVE_PA:
--
2.51.2.1041.gc1ab5b90ca-goog
Powered by blists - more mailing lists