Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1472301 > unrolled thread
| Started by | Wanpeng Li <kernellwp@gmail.com> |
|---|---|
| First post | 2016-08-30 10:20 +0200 |
| Last post | 2016-09-02 13:00 +0200 |
| Articles | 4 — 2 participants |
Back to article view | Back to linux.kernel
[PATCH 1/2] KVM: lapic: fix preemption timer backward when TSC backward Wanpeng Li <kernellwp@gmail.com> - 2016-08-30 10:20 +0200
[PATCH 2/2] KVM: nVMX: make emulated nested preemption timer pinned Wanpeng Li <kernellwp@gmail.com> - 2016-08-30 10:20 +0200
Re: [PATCH 2/2] KVM: nVMX: make emulated nested preemption timer pinned Paolo Bonzini <pbonzini@redhat.com> - 2016-09-02 13:00 +0200
Re: [PATCH 1/2] KVM: lapic: fix preemption timer backward when TSC backward Paolo Bonzini <pbonzini@redhat.com> - 2016-09-02 13:00 +0200
| From | Wanpeng Li <kernellwp@gmail.com> |
|---|---|
| Date | 2016-08-30 10:20 +0200 |
| Subject | [PATCH 1/2] KVM: lapic: fix preemption timer backward when TSC backward |
| Message-ID | <sbMqZ-2gM-3@gated-at.bofh.it> |
From: Wanpeng Li <wanpeng.li@hotmail.com>
TSC_OFFSET will be adjusted if discovers TSC backward during vCPU load.
The preemption timer which will leverage guest tsc to reprogram its
preemption timer value is also reprogrammed if vCPU is scheded in to
a different pCPU. However, current implementation reprogram preemption
timer before TSC_OFFSET is adjusted to the right value, this will result
in preemption timer also backward and fire prematurity.
This patch fix it by adjusting TSC_OFFSET before reprogramming preemption
timer if TSC backward.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Yunhong Jiang <yunhong.jiang@intel.com>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
---
arch/x86/kvm/x86.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 19f9f9e..699f872 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2743,16 +2743,16 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
if (tsc_delta < 0)
mark_tsc_unstable("KVM discovered backwards TSC");
- if (kvm_lapic_hv_timer_in_use(vcpu) &&
- kvm_x86_ops->set_hv_timer(vcpu,
- kvm_get_lapic_tscdeadline_msr(vcpu)))
- kvm_lapic_switch_to_sw_timer(vcpu);
if (check_tsc_unstable()) {
u64 offset = kvm_compute_tsc_offset(vcpu,
vcpu->arch.last_guest_tsc);
kvm_x86_ops->write_tsc_offset(vcpu, offset);
vcpu->arch.tsc_catchup = 1;
}
+ if (kvm_lapic_hv_timer_in_use(vcpu) &&
+ kvm_x86_ops->set_hv_timer(vcpu,
+ kvm_get_lapic_tscdeadline_msr(vcpu)))
+ kvm_lapic_switch_to_sw_timer(vcpu);
/*
* On a host with synchronized TSC, there is no need to update
* kvmclock on vcpu->cpu migration
--
1.9.1
[toc] | [next] | [standalone]
| From | Wanpeng Li <kernellwp@gmail.com> |
|---|---|
| Date | 2016-08-30 10:20 +0200 |
| Subject | [PATCH 2/2] KVM: nVMX: make emulated nested preemption timer pinned |
| Message-ID | <sbMqZ-2gM-9@gated-at.bofh.it> |
| In reply to | #1472301 |
From: Wanpeng Li <wanpeng.li@hotmail.com>
Commit 61abdbe0bc ("kvm: x86: make lapic hrtimer pinned") pins the emulated
lapic timer. This patch does the same for the emulated nested preemption
timer to avoid vmexit an unrelated vCPU and the latency of kicking IPI to
another vCPU.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Yunhong Jiang <yunhong.jiang@intel.com>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
---
arch/x86/kvm/vmx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5cede40..090a3f8 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -7013,7 +7013,7 @@ static int handle_vmon(struct kvm_vcpu *vcpu)
vmx->nested.vmcs02_num = 0;
hrtimer_init(&vmx->nested.preemption_timer, CLOCK_MONOTONIC,
- HRTIMER_MODE_REL);
+ HRTIMER_MODE_REL_PINNED);
vmx->nested.preemption_timer.function = vmx_preemption_timer_fn;
vmx->nested.vmxon = true;
--
1.9.1
[toc] | [prev] | [next] | [standalone]
| From | Paolo Bonzini <pbonzini@redhat.com> |
|---|---|
| Date | 2016-09-02 13:00 +0200 |
| Subject | Re: [PATCH 2/2] KVM: nVMX: make emulated nested preemption timer pinned |
| Message-ID | <scUmu-6SO-13@gated-at.bofh.it> |
| In reply to | #1472302 |
On 30/08/2016 10:14, Wanpeng Li wrote:
> From: Wanpeng Li <wanpeng.li@hotmail.com>
>
> Commit 61abdbe0bc ("kvm: x86: make lapic hrtimer pinned") pins the emulated
> lapic timer. This patch does the same for the emulated nested preemption
> timer to avoid vmexit an unrelated vCPU and the latency of kicking IPI to
> another vCPU.
>
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Cc: Radim Krčmář <rkrcmar@redhat.com>
> Cc: Yunhong Jiang <yunhong.jiang@intel.com>
> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
> ---
> arch/x86/kvm/vmx.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 5cede40..090a3f8 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -7013,7 +7013,7 @@ static int handle_vmon(struct kvm_vcpu *vcpu)
> vmx->nested.vmcs02_num = 0;
>
> hrtimer_init(&vmx->nested.preemption_timer, CLOCK_MONOTONIC,
> - HRTIMER_MODE_REL);
> + HRTIMER_MODE_REL_PINNED);
> vmx->nested.preemption_timer.function = vmx_preemption_timer_fn;
>
> vmx->nested.vmxon = true;
>
Queued for 4.9, thanks.
Paolo
[toc] | [prev] | [next] | [standalone]
| From | Paolo Bonzini <pbonzini@redhat.com> |
|---|---|
| Date | 2016-09-02 13:00 +0200 |
| Subject | Re: [PATCH 1/2] KVM: lapic: fix preemption timer backward when TSC backward |
| Message-ID | <scUmu-6SO-7@gated-at.bofh.it> |
| In reply to | #1472301 |
On 30/08/2016 10:14, Wanpeng Li wrote:
> From: Wanpeng Li <wanpeng.li@hotmail.com>
>
> TSC_OFFSET will be adjusted if discovers TSC backward during vCPU load.
> The preemption timer which will leverage guest tsc to reprogram its
> preemption timer value is also reprogrammed if vCPU is scheded in to
> a different pCPU. However, current implementation reprogram preemption
> timer before TSC_OFFSET is adjusted to the right value, this will result
> in preemption timer also backward and fire prematurity.
>
> This patch fix it by adjusting TSC_OFFSET before reprogramming preemption
> timer if TSC backward.
>
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Cc: Radim Krčmář <rkrcmar@redhat.com>
> Cc: Yunhong Jiang <yunhong.jiang@intel.com>
> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
> ---
> arch/x86/kvm/x86.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 19f9f9e..699f872 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -2743,16 +2743,16 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> if (tsc_delta < 0)
> mark_tsc_unstable("KVM discovered backwards TSC");
>
> - if (kvm_lapic_hv_timer_in_use(vcpu) &&
> - kvm_x86_ops->set_hv_timer(vcpu,
> - kvm_get_lapic_tscdeadline_msr(vcpu)))
> - kvm_lapic_switch_to_sw_timer(vcpu);
> if (check_tsc_unstable()) {
> u64 offset = kvm_compute_tsc_offset(vcpu,
> vcpu->arch.last_guest_tsc);
> kvm_x86_ops->write_tsc_offset(vcpu, offset);
> vcpu->arch.tsc_catchup = 1;
> }
> + if (kvm_lapic_hv_timer_in_use(vcpu) &&
> + kvm_x86_ops->set_hv_timer(vcpu,
> + kvm_get_lapic_tscdeadline_msr(vcpu)))
> + kvm_lapic_switch_to_sw_timer(vcpu);
> /*
> * On a host with synchronized TSC, there is no need to update
> * kvmclock on vcpu->cpu migration
>
Queued for 4.8, thanks.
Paolo
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web