Skip to contentSkip to conversation

kvm/x86: Fix PT "host mode"

PATCH In review Alexander Shishkin 2 participants 3 messages Aug 23, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 kvm/x86: Fix PT "host mode"

Regardless of the "pt_mode", the kvm driver installs its interrupt handler for Intel PT, which always overrides the native handler, causing data loss inside kvm guests, while we're expecting to trace them.

Fix this by only installing kvm's perf_guest_cbs if pt_mode is set to guest tracing.

Fixes ff9d07a0e7ce7 ("KVM: Implement perf callbacks for guest sampling")
---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          |  6 ++++++
 arch/x86/kvm/x86.c              | 10 ++++++++--
 3 files changed, 15 insertions(+), 2 deletions(-)
arch/x86/include/asm/kvm_host.h +1 -0
index 55efbacfc244..84a1ed067f35 100644
@@ -1408,6 +1408,7 @@ struct kvm_x86_init_ops {
 	int (*disabled_by_bios)(void);
 	int (*check_processor_compatibility)(void);
 	int (*hardware_setup)(void);
+	int (*intel_pt_enabled)(void);
 
 	struct kvm_x86_ops *runtime_ops;
 };
arch/x86/kvm/vmx/vmx.c +6 -0
index 4bceb5ca3a89..0c239aa3532a 100644
@@ -7943,11 +7943,17 @@ static __init int hardware_setup(void)
 	return r;
 }
 
+static int vmx_intel_pt_enabled(void)
+{
+	return vmx_pt_mode_is_host_guest();
+}
+
 static struct kvm_x86_init_ops vmx_init_ops __initdata = {
 	.cpu_has_kvm_support = cpu_has_kvm_support,
 	.disabled_by_bios = vmx_disabled_by_bios,
 	.check_processor_compatibility = vmx_check_processor_compat,
 	.hardware_setup = hardware_setup,
+	.intel_pt_enabled = vmx_intel_pt_enabled,
 
 	.runtime_ops = &vmx_x86_ops,
 };
arch/x86/kvm/x86.c +8 -2
index 9b6bca616929..3ba0001e7388 100644
@@ -268,6 +268,8 @@ static struct kmem_cache *x86_fpu_cache;
 
 static struct kmem_cache *x86_emulator_cache;
 
+static int __read_mostly intel_pt_enabled;
+
 /*
  * When called, it means the previous get/set msr reached an invalid msr.
  * Return true if we want to ignore/silent this failed msr access.
@@ -8194,7 +8196,10 @@ int kvm_arch_init(void *opaque)
 
 	kvm_timer_init();
 
-	perf_register_guest_info_callbacks(&kvm_guest_cbs);
+	if (ops->intel_pt_enabled && ops->intel_pt_enabled()) {
+		perf_register_guest_info_callbacks(&kvm_guest_cbs);
+		intel_pt_enabled = 1;
+	}
 
 	if (boot_cpu_has(X86_FEATURE_XSAVE)) {
 		host_xcr0 = xgetbv(XCR_XFEATURE_ENABLED_MASK);
@@ -8229,7 +8234,8 @@ void kvm_arch_exit(void)
 		clear_hv_tscchange_cb();
 #endif
 	kvm_lapic_exit();
-	perf_unregister_guest_info_callbacks(&kvm_guest_cbs);
+	if (intel_pt_enabled)
+		perf_unregister_guest_info_callbacks(&kvm_guest_cbs);
 
 	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
 		cpufreq_unregister_notifier(&kvmclock_cpufreq_notifier_block,
Signature · 1 line
2.32.0
#1 lore raw
Quoted 6 lines · Alexander Shishkin

Regardless of the "pt_mode", the kvm driver installs its interrupt handler for Intel PT, which always overrides the native handler, causing data loss inside kvm guests, while we're expecting to trace them.

Fix this by only installing kvm's perf_guest_cbs if pt_mode is set to guest tracing.

Uh, regardless of the correctness of such a change (spoiler alert), making an enormous leap from "one thing is wrong" to "nuke it all!" needs way more justfication/explanation. Or more realistically, such a leap should be a good indication that the proposed change is not correct.

Quoted 2 lines
Fixes ff9d07a0e7ce7 ("KVM: Implement perf callbacks for guest sampling")

This should be another clue that the fix isn't correct. That patch is from 2010, Intel PT was announced in 2013 and merged in 2019.

Quoted 8 lines
---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          |  6 ++++++
 arch/x86/kvm/x86.c              | 10 ++++++++--
 3 files changed, 15 insertions(+), 2 deletions(-)

...

Quoted 14 lines
arch/x86/kvm/x86.c +2 -0
index 9b6bca616929..3ba0001e7388 100644
@@ -268,6 +268,8 @@ static struct kmem_cache *x86_fpu_cache;
 
 static struct kmem_cache *x86_emulator_cache;
 
+static int __read_mostly intel_pt_enabled;
+
 /*
  * When called, it means the previous get/set msr reached an invalid msr.
  * Return true if we want to ignore/silent this failed msr access.

arch/x86/kvm/x86.c int kvm_arch_init(void *opaque) · line 8199

Quoted 5 lines
arch/x86/kvm/x86.c +1 -1
@@ -8194,7 +8196,10 @@ int kvm_arch_init(void *opaque)
 
 	kvm_timer_init();
 
-	perf_register_guest_info_callbacks(&kvm_guest_cbs);
+	if (ops->intel_pt_enabled && ops->intel_pt_enabled()) r

This is not remotely correct. vmx.c's "pt_mode", which is queried via this path, is modified by hardware_setup(), a.k.a. kvm_x86_ops.hardware_setup(), which runs _after_ this code. And as alluded to above, these are generic perf callbacks, installing them if and only if Intel PT is enabled in a specific mode completely breaks "regular" perf.

I'll post a small series, there's a bit of code massage needed to fix this properly. The PMI handler can also be optimized to avoid a retpoline when PT is not exposed to the guest.

Quoted 19 lines
arch/x86/kvm/x86.c +5 -1
+		perf_register_guest_info_callbacks(&kvm_guest_cbs);
+		intel_pt_enabled = 1;
+	}
 
 	if (boot_cpu_has(X86_FEATURE_XSAVE)) {
 		host_xcr0 = xgetbv(XCR_XFEATURE_ENABLED_MASK);
@@ -8229,7 +8234,8 @@ void kvm_arch_exit(void)
 		clear_hv_tscchange_cb();
 #endif
 	kvm_lapic_exit();
-	perf_unregister_guest_info_callbacks(&kvm_guest_cbs);
+	if (intel_pt_enabled)
+		perf_unregister_guest_info_callbacks(&kvm_guest_cbs);
 
 	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
 		cpufreq_unregister_notifier(&kvmclock_cpufreq_notifier_block,
Signature · 1 line
2.32.0
#2 lore raw
Quoted 7 lines · Sean Christopherson
Quoted 2 lines · Alexander Shishkin
Fixes ff9d07a0e7ce7 ("KVM: Implement perf callbacks for guest sampling")

This should be another clue that the fix isn't correct. That patch is from 2010,

Right, this should have been 8479e04e7d6b1 ("KVM: x86: Inject PMI for KVM guest") instead.

Quoted 1 line

Intel PT was announced in 2013 and merged in 2019.

Technically, 2019 is when kvm started breaking host PT.

Quoted 5 lines

This is not remotely correct. vmx.c's "pt_mode", which is queried via this path, is modified by hardware_setup(), a.k.a. kvm_x86_ops.hardware_setup(), which runs _after_ this code. And as alluded to above, these are generic perf callbacks, installing them if and only if Intel PT is enabled in a specific mode completely breaks "regular" perf.

I see your point, the callchain code will catch fire.

Quoted 3 lines

I'll post a small series, there's a bit of code massage needed to fix this properly. The PMI handler can also be optimized to avoid a retpoline when PT is not exposed to the guest.

The actual PMU handler also needs to know that kvm won't be needing it so it can call the regular PT handler.

One could unset cbs->handle_intel_pt_intr() or one could have it return different things depending on whether it was actually taken in kvm. But both are rather disgusting.

Regards,
--
Alex