diff mbox series

[v2,4/6] x86/bugs: Use a static branch to guard IBPB on vCPU switch

Message ID 20250227012712.3193063-5-yosry.ahmed@linux.dev (mailing list archive)
State New
Headers show
Series IBPB cleanups and a fixup | expand

Commit Message

Yosry Ahmed Feb. 27, 2025, 1:27 a.m. UTC 
Instead of using X86_FEATURE_USE_IBPB to guard the IBPB execution in KVM
when a new vCPU is loaded, introduce a static branch, similar to
switch_mm_*_ibpb.

This makes it obvious in spectre_v2_user_select_mitigation() what
exactly is being toggled, instead of the unclear X86_FEATURE_USE_IBPB
(which will be shortly removed). It also provides more fine-grained
control, making it simpler to change/add paths that control the IBPB in
the vCPU switch path without affecting other IBPBs.

Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
---
 arch/x86/include/asm/nospec-branch.h | 2 ++
 arch/x86/kernel/cpu/bugs.c           | 5 +++++
 arch/x86/kvm/svm/svm.c               | 2 +-
 arch/x86/kvm/vmx/vmx.c               | 2 +-
 4 files changed, 9 insertions(+), 2 deletions(-)

Comments

Sean Christopherson Feb. 27, 2025, 1:42 a.m. UTC | #1 
On Thu, Feb 27, 2025, Yosry Ahmed wrote:
> Instead of using X86_FEATURE_USE_IBPB to guard the IBPB execution in KVM
> when a new vCPU is loaded, introduce a static branch, similar to
> switch_mm_*_ibpb.
> 
> This makes it obvious in spectre_v2_user_select_mitigation() what
> exactly is being toggled, instead of the unclear X86_FEATURE_USE_IBPB
> (which will be shortly removed). It also provides more fine-grained
> control, making it simpler to change/add paths that control the IBPB in
> the vCPU switch path without affecting other IBPBs.
> 
> Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
> Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
> ---

Acked-by: Sean Christopherson <seanjc@google.com>
diff mbox series

Patch

diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
index 7cbb76a2434b9..36bc395bdef32 100644
--- a/arch/x86/include/asm/nospec-branch.h
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -552,6 +552,8 @@  DECLARE_STATIC_KEY_FALSE(switch_to_cond_stibp);
 DECLARE_STATIC_KEY_FALSE(switch_mm_cond_ibpb);
 DECLARE_STATIC_KEY_FALSE(switch_mm_always_ibpb);
 
+DECLARE_STATIC_KEY_FALSE(switch_vcpu_ibpb);
+
 DECLARE_STATIC_KEY_FALSE(mds_idle_clear);
 
 DECLARE_STATIC_KEY_FALSE(switch_mm_cond_l1d_flush);
diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
index 1d7afc40f2272..7f904d0b0b04f 100644
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -113,6 +113,10 @@  DEFINE_STATIC_KEY_FALSE(switch_mm_cond_ibpb);
 /* Control unconditional IBPB in switch_mm() */
 DEFINE_STATIC_KEY_FALSE(switch_mm_always_ibpb);
 
+/* Control IBPB on vCPU load */
+DEFINE_STATIC_KEY_FALSE(switch_vcpu_ibpb);
+EXPORT_SYMBOL_GPL(switch_vcpu_ibpb);
+
 /* Control MDS CPU buffer clear before idling (halt, mwait) */
 DEFINE_STATIC_KEY_FALSE(mds_idle_clear);
 EXPORT_SYMBOL_GPL(mds_idle_clear);
@@ -1365,6 +1369,7 @@  spectre_v2_user_select_mitigation(void)
 	/* Initialize Indirect Branch Prediction Barrier */
 	if (boot_cpu_has(X86_FEATURE_IBPB)) {
 		setup_force_cpu_cap(X86_FEATURE_USE_IBPB);
+		static_branch_enable(&switch_vcpu_ibpb);
 
 		spectre_v2_user_ibpb = mode;
 		switch (cmd) {
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 57222c3b56592..a73875ffbc3df 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -1566,7 +1566,7 @@  static void svm_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
 		sd->current_vmcb = svm->vmcb;
 
 		if (!cpu_feature_enabled(X86_FEATURE_IBPB_ON_VMEXIT) &&
-		    cpu_feature_enabled(X86_FEATURE_USE_IBPB))
+		    static_branch_likely(&switch_vcpu_ibpb))
 			indirect_branch_prediction_barrier();
 	}
 	if (kvm_vcpu_apicv_active(vcpu))
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 042b7a88157b0..956f9143a3c4c 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -1477,7 +1477,7 @@  void vmx_vcpu_load_vmcs(struct kvm_vcpu *vcpu, int cpu,
 		 * performs IBPB on nested VM-Exit (a single nested transition
 		 * may switch the active VMCS multiple times).
 		 */
-		if (cpu_feature_enabled(X86_FEATURE_USE_IBPB) &&
+		if (static_branch_likely(&switch_vcpu_ibpb) &&
 		    (!buddy || WARN_ON_ONCE(buddy->vmcs != prev)))
 			indirect_branch_prediction_barrier();
 	}