* [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP
@ 2021-04-09 4:18 Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 2/3] KVM: X86: Count attempted/successful directed yield Wanpeng Li
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Wanpeng Li @ 2021-04-09 4:18 UTC (permalink / raw)
To: linux-kernel, kvm
Cc: Paolo Bonzini, Sean Christopherson, Vitaly Kuznetsov, Wanpeng Li,
Jim Mattson, Joerg Roedel
From: Wanpeng Li <wanpengli@tencent.com>
Enable PV TLB shootdown when !CONFIG_SMP doesn't make sense. Let's
move it inside CONFIG_SMP. In addition, we can avoid define and
alloc __pv_cpu_mask when !CONFIG_SMP and get rid of 'alloc' variable
in kvm_alloc_cpumask.
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
---
v1 -> v2:
* shuffle things around a bit more
arch/x86/kernel/kvm.c | 118 +++++++++++++++++++++++---------------------------
1 file changed, 55 insertions(+), 63 deletions(-)
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index 5e78e01..224a7a1 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -451,6 +451,10 @@ static void __init sev_map_percpu_data(void)
}
}
+#ifdef CONFIG_SMP
+
+static DEFINE_PER_CPU(cpumask_var_t, __pv_cpu_mask);
+
static bool pv_tlb_flush_supported(void)
{
return (kvm_para_has_feature(KVM_FEATURE_PV_TLB_FLUSH) &&
@@ -458,10 +462,6 @@ static bool pv_tlb_flush_supported(void)
kvm_para_has_feature(KVM_FEATURE_STEAL_TIME));
}
-static DEFINE_PER_CPU(cpumask_var_t, __pv_cpu_mask);
-
-#ifdef CONFIG_SMP
-
static bool pv_ipi_supported(void)
{
return kvm_para_has_feature(KVM_FEATURE_PV_SEND_IPI);
@@ -574,6 +574,49 @@ static void kvm_smp_send_call_func_ipi(const struct cpumask *mask)
}
}
+static void kvm_flush_tlb_others(const struct cpumask *cpumask,
+ const struct flush_tlb_info *info)
+{
+ u8 state;
+ int cpu;
+ struct kvm_steal_time *src;
+ struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
+
+ cpumask_copy(flushmask, cpumask);
+ /*
+ * We have to call flush only on online vCPUs. And
+ * queue flush_on_enter for pre-empted vCPUs
+ */
+ for_each_cpu(cpu, flushmask) {
+ src = &per_cpu(steal_time, cpu);
+ state = READ_ONCE(src->preempted);
+ if ((state & KVM_VCPU_PREEMPTED)) {
+ if (try_cmpxchg(&src->preempted, &state,
+ state | KVM_VCPU_FLUSH_TLB))
+ __cpumask_clear_cpu(cpu, flushmask);
+ }
+ }
+
+ native_flush_tlb_others(flushmask, info);
+}
+
+static __init int kvm_alloc_cpumask(void)
+{
+ int cpu;
+
+ if (!kvm_para_available() || nopv)
+ return 0;
+
+ if (pv_tlb_flush_supported() || pv_ipi_supported())
+ for_each_possible_cpu(cpu) {
+ zalloc_cpumask_var_node(per_cpu_ptr(&__pv_cpu_mask, cpu),
+ GFP_KERNEL, cpu_to_node(cpu));
+ }
+
+ return 0;
+}
+arch_initcall(kvm_alloc_cpumask);
+
static void __init kvm_smp_prepare_boot_cpu(void)
{
/*
@@ -611,33 +654,8 @@ static int kvm_cpu_down_prepare(unsigned int cpu)
local_irq_enable();
return 0;
}
-#endif
-
-static void kvm_flush_tlb_others(const struct cpumask *cpumask,
- const struct flush_tlb_info *info)
-{
- u8 state;
- int cpu;
- struct kvm_steal_time *src;
- struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
-
- cpumask_copy(flushmask, cpumask);
- /*
- * We have to call flush only on online vCPUs. And
- * queue flush_on_enter for pre-empted vCPUs
- */
- for_each_cpu(cpu, flushmask) {
- src = &per_cpu(steal_time, cpu);
- state = READ_ONCE(src->preempted);
- if ((state & KVM_VCPU_PREEMPTED)) {
- if (try_cmpxchg(&src->preempted, &state,
- state | KVM_VCPU_FLUSH_TLB))
- __cpumask_clear_cpu(cpu, flushmask);
- }
- }
- native_flush_tlb_others(flushmask, info);
-}
+#endif
static void __init kvm_guest_init(void)
{
@@ -653,12 +671,6 @@ static void __init kvm_guest_init(void)
pv_ops.time.steal_clock = kvm_steal_clock;
}
- if (pv_tlb_flush_supported()) {
- pv_ops.mmu.flush_tlb_others = kvm_flush_tlb_others;
- pv_ops.mmu.tlb_remove_table = tlb_remove_table;
- pr_info("KVM setup pv remote TLB flush\n");
- }
-
if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
apic_set_eoi_write(kvm_guest_apic_eoi_write);
@@ -668,6 +680,12 @@ static void __init kvm_guest_init(void)
}
#ifdef CONFIG_SMP
+ if (pv_tlb_flush_supported()) {
+ pv_ops.mmu.flush_tlb_others = kvm_flush_tlb_others;
+ pv_ops.mmu.tlb_remove_table = tlb_remove_table;
+ pr_info("KVM setup pv remote TLB flush\n");
+ }
+
smp_ops.smp_prepare_boot_cpu = kvm_smp_prepare_boot_cpu;
if (pv_sched_yield_supported()) {
smp_ops.send_call_func_ipi = kvm_smp_send_call_func_ipi;
@@ -734,7 +752,7 @@ static uint32_t __init kvm_detect(void)
static void __init kvm_apic_init(void)
{
-#if defined(CONFIG_SMP)
+#ifdef CONFIG_SMP
if (pv_ipi_supported())
kvm_setup_pv_ipi();
#endif
@@ -794,32 +812,6 @@ static __init int activate_jump_labels(void)
}
arch_initcall(activate_jump_labels);
-static __init int kvm_alloc_cpumask(void)
-{
- int cpu;
- bool alloc = false;
-
- if (!kvm_para_available() || nopv)
- return 0;
-
- if (pv_tlb_flush_supported())
- alloc = true;
-
-#if defined(CONFIG_SMP)
- if (pv_ipi_supported())
- alloc = true;
-#endif
-
- if (alloc)
- for_each_possible_cpu(cpu) {
- zalloc_cpumask_var_node(per_cpu_ptr(&__pv_cpu_mask, cpu),
- GFP_KERNEL, cpu_to_node(cpu));
- }
-
- return 0;
-}
-arch_initcall(kvm_alloc_cpumask);
-
#ifdef CONFIG_PARAVIRT_SPINLOCKS
/* Kick a cpu by its apicid. Used to wake up a halted vcpu */
--
2.7.4
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH v2 2/3] KVM: X86: Count attempted/successful directed yield
2021-04-09 4:18 [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Wanpeng Li
@ 2021-04-09 4:18 ` Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 3/3] KVM: X86: Do not yield to self Wanpeng Li
2021-04-17 12:34 ` [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Paolo Bonzini
2 siblings, 0 replies; 4+ messages in thread
From: Wanpeng Li @ 2021-04-09 4:18 UTC (permalink / raw)
To: linux-kernel, kvm
Cc: Paolo Bonzini, Sean Christopherson, Vitaly Kuznetsov, Wanpeng Li,
Jim Mattson, Joerg Roedel
From: Wanpeng Li <wanpengli@tencent.com>
To analyze some performance issues with lock contention and scheduling,
it is nice to know when directed yield are successful or failing.
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
---
v1 -> v2:
* rename new vcpu stat
* account success instead of ignore
arch/x86/include/asm/kvm_host.h | 2 ++
arch/x86/kvm/x86.c | 24 ++++++++++++++++++------
2 files changed, 20 insertions(+), 6 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 44f8930..5af7411 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1126,6 +1126,8 @@ struct kvm_vcpu_stat {
u64 halt_poll_success_ns;
u64 halt_poll_fail_ns;
u64 nested_run;
+ u64 directed_yield_attempted;
+ u64 directed_yield_successful;
};
struct x86_instruction_info;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 16fb395..f08e9b4 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -246,6 +246,8 @@ struct kvm_stats_debugfs_item debugfs_entries[] = {
VCPU_STAT("halt_poll_success_ns", halt_poll_success_ns),
VCPU_STAT("halt_poll_fail_ns", halt_poll_fail_ns),
VCPU_STAT("nested_run", nested_run),
+ VCPU_STAT("directed_yield_attempted", directed_yield_attempted),
+ VCPU_STAT("directed_yield_successful", directed_yield_successful),
VM_STAT("mmu_shadow_zapped", mmu_shadow_zapped),
VM_STAT("mmu_pte_write", mmu_pte_write),
VM_STAT("mmu_pde_zapped", mmu_pde_zapped),
@@ -8211,21 +8213,31 @@ void kvm_apicv_init(struct kvm *kvm, bool enable)
}
EXPORT_SYMBOL_GPL(kvm_apicv_init);
-static void kvm_sched_yield(struct kvm *kvm, unsigned long dest_id)
+static void kvm_sched_yield(struct kvm_vcpu *vcpu, unsigned long dest_id)
{
struct kvm_vcpu *target = NULL;
struct kvm_apic_map *map;
+ vcpu->stat.directed_yield_attempted++;
+
rcu_read_lock();
- map = rcu_dereference(kvm->arch.apic_map);
+ map = rcu_dereference(vcpu->kvm->arch.apic_map);
if (likely(map) && dest_id <= map->max_apic_id && map->phys_map[dest_id])
target = map->phys_map[dest_id]->vcpu;
rcu_read_unlock();
- if (target && READ_ONCE(target->ready))
- kvm_vcpu_yield_to(target);
+ if (!target || !READ_ONCE(target->ready))
+ goto no_yield;
+
+ if (kvm_vcpu_yield_to(target) <= 0)
+ goto no_yield;
+
+ vcpu->stat.directed_yield_successful++;
+
+no_yield:
+ return;
}
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
@@ -8272,7 +8284,7 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
break;
kvm_pv_kick_cpu_op(vcpu->kvm, a0, a1);
- kvm_sched_yield(vcpu->kvm, a1);
+ kvm_sched_yield(vcpu, a1);
ret = 0;
break;
#ifdef CONFIG_X86_64
@@ -8290,7 +8302,7 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
if (!guest_pv_has(vcpu, KVM_FEATURE_PV_SCHED_YIELD))
break;
- kvm_sched_yield(vcpu->kvm, a0);
+ kvm_sched_yield(vcpu, a0);
ret = 0;
break;
default:
--
2.7.4
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH v2 3/3] KVM: X86: Do not yield to self
2021-04-09 4:18 [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 2/3] KVM: X86: Count attempted/successful directed yield Wanpeng Li
@ 2021-04-09 4:18 ` Wanpeng Li
2021-04-17 12:34 ` [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Paolo Bonzini
2 siblings, 0 replies; 4+ messages in thread
From: Wanpeng Li @ 2021-04-09 4:18 UTC (permalink / raw)
To: linux-kernel, kvm
Cc: Paolo Bonzini, Sean Christopherson, Vitaly Kuznetsov, Wanpeng Li,
Jim Mattson, Joerg Roedel
From: Wanpeng Li <wanpengli@tencent.com>
If the target is self we do not need to yield, we can avoid malicious
guest to play this.
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
---
v1 -> v2:
* update comments
arch/x86/kvm/x86.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index f08e9b4..ce9a1d2 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8231,6 +8231,10 @@ static void kvm_sched_yield(struct kvm_vcpu *vcpu, unsigned long dest_id)
if (!target || !READ_ONCE(target->ready))
goto no_yield;
+ /* Ignore requests to yield to self */
+ if (vcpu == target)
+ goto no_yield;
+
if (kvm_vcpu_yield_to(target) <= 0)
goto no_yield;
--
2.7.4
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP
2021-04-09 4:18 [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 2/3] KVM: X86: Count attempted/successful directed yield Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 3/3] KVM: X86: Do not yield to self Wanpeng Li
@ 2021-04-17 12:34 ` Paolo Bonzini
2 siblings, 0 replies; 4+ messages in thread
From: Paolo Bonzini @ 2021-04-17 12:34 UTC (permalink / raw)
To: Wanpeng Li, linux-kernel, kvm
Cc: Sean Christopherson, Vitaly Kuznetsov, Wanpeng Li, Jim Mattson,
Joerg Roedel
On 09/04/21 06:18, Wanpeng Li wrote:
> From: Wanpeng Li <wanpengli@tencent.com>
>
> Enable PV TLB shootdown when !CONFIG_SMP doesn't make sense. Let's
> move it inside CONFIG_SMP. In addition, we can avoid define and
> alloc __pv_cpu_mask when !CONFIG_SMP and get rid of 'alloc' variable
> in kvm_alloc_cpumask.
>
> Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
> ---
> v1 -> v2:
> * shuffle things around a bit more
>
> arch/x86/kernel/kvm.c | 118 +++++++++++++++++++++++---------------------------
> 1 file changed, 55 insertions(+), 63 deletions(-)
>
> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
> index 5e78e01..224a7a1 100644
> --- a/arch/x86/kernel/kvm.c
> +++ b/arch/x86/kernel/kvm.c
> @@ -451,6 +451,10 @@ static void __init sev_map_percpu_data(void)
> }
> }
>
> +#ifdef CONFIG_SMP
> +
> +static DEFINE_PER_CPU(cpumask_var_t, __pv_cpu_mask);
> +
> static bool pv_tlb_flush_supported(void)
> {
> return (kvm_para_has_feature(KVM_FEATURE_PV_TLB_FLUSH) &&
> @@ -458,10 +462,6 @@ static bool pv_tlb_flush_supported(void)
> kvm_para_has_feature(KVM_FEATURE_STEAL_TIME));
> }
>
> -static DEFINE_PER_CPU(cpumask_var_t, __pv_cpu_mask);
> -
> -#ifdef CONFIG_SMP
> -
> static bool pv_ipi_supported(void)
> {
> return kvm_para_has_feature(KVM_FEATURE_PV_SEND_IPI);
> @@ -574,6 +574,49 @@ static void kvm_smp_send_call_func_ipi(const struct cpumask *mask)
> }
> }
>
> +static void kvm_flush_tlb_others(const struct cpumask *cpumask,
> + const struct flush_tlb_info *info)
> +{
> + u8 state;
> + int cpu;
> + struct kvm_steal_time *src;
> + struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
> +
> + cpumask_copy(flushmask, cpumask);
> + /*
> + * We have to call flush only on online vCPUs. And
> + * queue flush_on_enter for pre-empted vCPUs
> + */
> + for_each_cpu(cpu, flushmask) {
> + src = &per_cpu(steal_time, cpu);
> + state = READ_ONCE(src->preempted);
> + if ((state & KVM_VCPU_PREEMPTED)) {
> + if (try_cmpxchg(&src->preempted, &state,
> + state | KVM_VCPU_FLUSH_TLB))
> + __cpumask_clear_cpu(cpu, flushmask);
> + }
> + }
> +
> + native_flush_tlb_others(flushmask, info);
> +}
> +
> +static __init int kvm_alloc_cpumask(void)
> +{
> + int cpu;
> +
> + if (!kvm_para_available() || nopv)
> + return 0;
> +
> + if (pv_tlb_flush_supported() || pv_ipi_supported())
> + for_each_possible_cpu(cpu) {
> + zalloc_cpumask_var_node(per_cpu_ptr(&__pv_cpu_mask, cpu),
> + GFP_KERNEL, cpu_to_node(cpu));
> + }
> +
> + return 0;
> +}
> +arch_initcall(kvm_alloc_cpumask);
> +
> static void __init kvm_smp_prepare_boot_cpu(void)
> {
> /*
> @@ -611,33 +654,8 @@ static int kvm_cpu_down_prepare(unsigned int cpu)
> local_irq_enable();
> return 0;
> }
> -#endif
> -
> -static void kvm_flush_tlb_others(const struct cpumask *cpumask,
> - const struct flush_tlb_info *info)
> -{
> - u8 state;
> - int cpu;
> - struct kvm_steal_time *src;
> - struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
> -
> - cpumask_copy(flushmask, cpumask);
> - /*
> - * We have to call flush only on online vCPUs. And
> - * queue flush_on_enter for pre-empted vCPUs
> - */
> - for_each_cpu(cpu, flushmask) {
> - src = &per_cpu(steal_time, cpu);
> - state = READ_ONCE(src->preempted);
> - if ((state & KVM_VCPU_PREEMPTED)) {
> - if (try_cmpxchg(&src->preempted, &state,
> - state | KVM_VCPU_FLUSH_TLB))
> - __cpumask_clear_cpu(cpu, flushmask);
> - }
> - }
>
> - native_flush_tlb_others(flushmask, info);
> -}
> +#endif
>
> static void __init kvm_guest_init(void)
> {
> @@ -653,12 +671,6 @@ static void __init kvm_guest_init(void)
> pv_ops.time.steal_clock = kvm_steal_clock;
> }
>
> - if (pv_tlb_flush_supported()) {
> - pv_ops.mmu.flush_tlb_others = kvm_flush_tlb_others;
> - pv_ops.mmu.tlb_remove_table = tlb_remove_table;
> - pr_info("KVM setup pv remote TLB flush\n");
> - }
> -
> if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
> apic_set_eoi_write(kvm_guest_apic_eoi_write);
>
> @@ -668,6 +680,12 @@ static void __init kvm_guest_init(void)
> }
>
> #ifdef CONFIG_SMP
> + if (pv_tlb_flush_supported()) {
> + pv_ops.mmu.flush_tlb_others = kvm_flush_tlb_others;
> + pv_ops.mmu.tlb_remove_table = tlb_remove_table;
> + pr_info("KVM setup pv remote TLB flush\n");
> + }
> +
> smp_ops.smp_prepare_boot_cpu = kvm_smp_prepare_boot_cpu;
> if (pv_sched_yield_supported()) {
> smp_ops.send_call_func_ipi = kvm_smp_send_call_func_ipi;
> @@ -734,7 +752,7 @@ static uint32_t __init kvm_detect(void)
>
> static void __init kvm_apic_init(void)
> {
> -#if defined(CONFIG_SMP)
> +#ifdef CONFIG_SMP
> if (pv_ipi_supported())
> kvm_setup_pv_ipi();
> #endif
> @@ -794,32 +812,6 @@ static __init int activate_jump_labels(void)
> }
> arch_initcall(activate_jump_labels);
>
> -static __init int kvm_alloc_cpumask(void)
> -{
> - int cpu;
> - bool alloc = false;
> -
> - if (!kvm_para_available() || nopv)
> - return 0;
> -
> - if (pv_tlb_flush_supported())
> - alloc = true;
> -
> -#if defined(CONFIG_SMP)
> - if (pv_ipi_supported())
> - alloc = true;
> -#endif
> -
> - if (alloc)
> - for_each_possible_cpu(cpu) {
> - zalloc_cpumask_var_node(per_cpu_ptr(&__pv_cpu_mask, cpu),
> - GFP_KERNEL, cpu_to_node(cpu));
> - }
> -
> - return 0;
> -}
> -arch_initcall(kvm_alloc_cpumask);
> -
> #ifdef CONFIG_PARAVIRT_SPINLOCKS
>
> /* Kick a cpu by its apicid. Used to wake up a halted vcpu */
>
Queued all three, thanks.
Paolo
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2021-04-17 12:34 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-09 4:18 [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 2/3] KVM: X86: Count attempted/successful directed yield Wanpeng Li
2021-04-09 4:18 ` [PATCH v2 3/3] KVM: X86: Do not yield to self Wanpeng Li
2021-04-17 12:34 ` [PATCH v2 1/3] x86/kvm: Don't bother __pv_cpu_mask when !CONFIG_SMP Paolo Bonzini
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).