diff mbox series

[v2] KVM: x86: Ignore MSR_AMD64_TW_CFG access

Message ID 1ce85d9c7c9e9632393816cf19c902e0a3f411f1.1697731406.git.maciej.szmigiero@oracle.com
State New
Headers
Series [v2] KVM: x86: Ignore MSR_AMD64_TW_CFG access |

Commit Message

Maciej S. Szmigiero Oct. 19, 2023, 4:06 p.m. UTC 
  From: "Maciej S. Szmigiero" <maciej.szmigiero@oracle.com>

Hyper-V enabled Windows Server 2022 KVM VM cannot be started on Zen1 Ryzen
since it crashes at boot with SYSTEM_THREAD_EXCEPTION_NOT_HANDLED +
STATUS_PRIVILEGED_INSTRUCTION (in other words, because of an unexpected #GP
in the guest kernel).

This is because Windows tries to set bit 8 in MSR_AMD64_TW_CFG and can't
handle receiving a #GP when doing so.

Give this MSR the same treatment that commit 2e32b7190641
("x86, kvm: Add MSR_AMD64_BU_CFG2 to the list of ignored MSRs") gave
MSR_AMD64_BU_CFG2 under justification that this MSR is baremetal-relevant
only.
Although apparently it was then needed for Linux guests, not Windows as in
this case.

With this change, the aforementioned guest setup is able to finish booting
successfully.

This issue can be reproduced either on a Summit Ridge Ryzen (with
just "-cpu host") or on a Naples EPYC (with "-cpu host,stepping=1" since
EPYC is ordinarily stepping 2).

Signed-off-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
---

Changes from v1:
    Rename MSR_AMD64_BU_CFG to MSR_AMD64_TW_CFG since Tom says that's the
    proper name of that MSR for Zen.

 arch/x86/include/asm/msr-index.h | 1 +
 arch/x86/kvm/x86.c               | 2 ++
 2 files changed, 3 insertions(+)

Comments

Sean Christopherson Oct. 20, 2023, 10:56 p.m. UTC | #1 
On Thu, 19 Oct 2023 18:06:57 +0200, Maciej S. Szmigiero wrote:
> Hyper-V enabled Windows Server 2022 KVM VM cannot be started on Zen1 Ryzen
> since it crashes at boot with SYSTEM_THREAD_EXCEPTION_NOT_HANDLED +
> STATUS_PRIVILEGED_INSTRUCTION (in other words, because of an unexpected #GP
> in the guest kernel).
> 
> This is because Windows tries to set bit 8 in MSR_AMD64_TW_CFG and can't
> handle receiving a #GP when doing so.
> 
> [...]

Applied to kvm-x86 misc, thanks!  I added a paragraph at the end of the
changelog to capture the gist of the discussion on why we agreed that having
KVM eat MSR accesses is the least awful option.  I also tagged this for stable.

Paolo, holler if you want to grab this for v6.6 and I'll drop my copy.

[1/1] KVM: x86: Ignore MSR_AMD64_TW_CFG access
      https://github.com/kvm-x86/linux/commit/2770d4722036

--
https://github.com/kvm-x86/linux/tree/next
diff mbox series

Patch

diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h
index 1d111350197f..8bcbebb56b8f 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -553,6 +553,7 @@ 
 #define MSR_AMD64_CPUID_FN_1		0xc0011004
 #define MSR_AMD64_LS_CFG		0xc0011020
 #define MSR_AMD64_DC_CFG		0xc0011022
+#define MSR_AMD64_TW_CFG		0xc0011023
 
 #define MSR_AMD64_DE_CFG		0xc0011029
 #define MSR_AMD64_DE_CFG_LFENCE_SERIALIZE_BIT	 1
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 9f18b06bbda6..fd1b099b0964 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3641,6 +3641,7 @@  int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 	case MSR_AMD64_PATCH_LOADER:
 	case MSR_AMD64_BU_CFG2:
 	case MSR_AMD64_DC_CFG:
+	case MSR_AMD64_TW_CFG:
 	case MSR_F15H_EX_CFG:
 		break;
 
@@ -4065,6 +4066,7 @@  int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 	case MSR_AMD64_BU_CFG2:
 	case MSR_IA32_PERF_CTL:
 	case MSR_AMD64_DC_CFG:
+	case MSR_AMD64_TW_CFG:
 	case MSR_F15H_EX_CFG:
 	/*
 	 * Intel Sandy Bridge CPUs must support the RAPL (running average power