From: Wanpeng Li <[email protected]>
This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
vCPU is preempted), we found great regression caused by this commit.
Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
records/s with this commit.
Host Guest score
vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
Exit from aggressive wait-early mechanism can result in yield premature and
incur extra scheduling latency in over-subscribe scenario.
kvm optimizes:
[1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
[2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
Tested-by: [email protected]
Cc: Peter Zijlstra <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Waiman Long <[email protected]>
Cc: Paolo Bonzini <[email protected]>
Cc: Radim Krčmář <[email protected]>
Cc: [email protected]
Cc: [email protected]
Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
Signed-off-by: Wanpeng Li <[email protected]>
---
kernel/locking/qspinlock_paravirt.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
index 89bab07..e84d21a 100644
--- a/kernel/locking/qspinlock_paravirt.h
+++ b/kernel/locking/qspinlock_paravirt.h
@@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
if ((loop & PV_PREV_CHECK_MASK) != 0)
return false;
- return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
+ return READ_ONCE(prev->state) != vcpu_running;
}
/*
--
2.7.4
On 09/09/19 12:56, Waiman Long wrote:
> On 9/9/19 2:40 AM, Wanpeng Li wrote:
>> From: Wanpeng Li <[email protected]>
>>
>> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
>> vCPU is preempted), we found great regression caused by this commit.
>>
>> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
>> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
>> records/s with this commit.
>>
>> Host Guest score
>>
>> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
>> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
>> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
>> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
>>
>> Exit from aggressive wait-early mechanism can result in yield premature and
>> incur extra scheduling latency in over-subscribe scenario.
>>
>> kvm optimizes:
>> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
>> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
>>
>> Tested-by: [email protected]
>> Cc: Peter Zijlstra <[email protected]>
>> Cc: Thomas Gleixner <[email protected]>
>> Cc: Ingo Molnar <[email protected]>
>> Cc: Waiman Long <[email protected]>
>> Cc: Paolo Bonzini <[email protected]>
>> Cc: Radim Krčmář <[email protected]>
>> Cc: [email protected]
>> Cc: [email protected]
>> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
>> Signed-off-by: Wanpeng Li <[email protected]>
>> ---
>> kernel/locking/qspinlock_paravirt.h | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
>> index 89bab07..e84d21a 100644
>> --- a/kernel/locking/qspinlock_paravirt.h
>> +++ b/kernel/locking/qspinlock_paravirt.h
>> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
>> if ((loop & PV_PREV_CHECK_MASK) != 0)
>> return false;
>>
>> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
>> + return READ_ONCE(prev->state) != vcpu_running;
>> }
>>
>> /*
>
> There are several possibilities for this performance regression:
>
> 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
> cacheline contention issue depending on how that callback is implemented.
Unlikely, it is a single percpu read.
> 2) KVM may set the preempt flag for a short period whenver an vmexit
> happens even if a vmenter is executed shortly after. In this case, we
> may want to use a more durable vcpu suspend flag that indicates the vcpu
> won't get a real vcpu back for a longer period of time.
It sets it for exits to userspace, but they shouldn't really happen on a
properly-configured system.
However, it's easy to test this theory:
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 2e302e977dac..feb6c75a7a88 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3368,26 +3368,28 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
int idx;
- if (vcpu->preempted)
+ if (vcpu->preempted) {
vcpu->arch.preempted_in_kernel = !kvm_x86_ops->get_cpl(vcpu);
- /*
- * Disable page faults because we're in atomic context here.
- * kvm_write_guest_offset_cached() would call might_fault()
- * that relies on pagefault_disable() to tell if there's a
- * bug. NOTE: the write to guest memory may not go through if
- * during postcopy live migration or if there's heavy guest
- * paging.
- */
- pagefault_disable();
- /*
- * kvm_memslots() will be called by
- * kvm_write_guest_offset_cached() so take the srcu lock.
- */
- idx = srcu_read_lock(&vcpu->kvm->srcu);
- kvm_steal_time_set_preempted(vcpu);
- srcu_read_unlock(&vcpu->kvm->srcu, idx);
- pagefault_enable();
+ /*
+ * Disable page faults because we're in atomic context here.
+ * kvm_write_guest_offset_cached() would call might_fault()
+ * that relies on pagefault_disable() to tell if there's a
+ * bug. NOTE: the write to guest memory may not go through if
+ * during postcopy live migration or if there's heavy guest
+ * paging.
+ */
+ pagefault_disable();
+ /*
+ * kvm_memslots() will be called by
+ * kvm_write_guest_offset_cached() so take the srcu lock.
+ */
+ idx = srcu_read_lock(&vcpu->kvm->srcu);
+ kvm_steal_time_set_preempted(vcpu);
+ srcu_read_unlock(&vcpu->kvm->srcu, idx);
+ pagefault_enable();
+ }
+
kvm_x86_ops->vcpu_put(vcpu);
vcpu->arch.last_host_tsc = rdtsc();
/*
Wanpeng, can you try?
Paolo
> Perhaps you can add a lock event counter to count the number of
> wait_early events caused by vcpu_is_preempted() being true to see if it
> really cause a lot more wait_early than without the vcpu_is_preempted()
> call.
>
> I have no objection to this, I just want to find out the root cause of it.
>
> Cheers,
> Longman
>
On 9/9/19 2:40 AM, Wanpeng Li wrote:
> From: Wanpeng Li <[email protected]>
>
> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
> vCPU is preempted), we found great regression caused by this commit.
>
> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
> records/s with this commit.
>
> Host Guest score
>
> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
>
> Exit from aggressive wait-early mechanism can result in yield premature and
> incur extra scheduling latency in over-subscribe scenario.
>
> kvm optimizes:
> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
>
> Tested-by: [email protected]
> Cc: Peter Zijlstra <[email protected]>
> Cc: Thomas Gleixner <[email protected]>
> Cc: Ingo Molnar <[email protected]>
> Cc: Waiman Long <[email protected]>
> Cc: Paolo Bonzini <[email protected]>
> Cc: Radim Krčmář <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
> Signed-off-by: Wanpeng Li <[email protected]>
> ---
> kernel/locking/qspinlock_paravirt.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
> index 89bab07..e84d21a 100644
> --- a/kernel/locking/qspinlock_paravirt.h
> +++ b/kernel/locking/qspinlock_paravirt.h
> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
> if ((loop & PV_PREV_CHECK_MASK) != 0)
> return false;
>
> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
> + return READ_ONCE(prev->state) != vcpu_running;
> }
>
> /*
There are several possibilities for this performance regression:
1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
cacheline contention issue depending on how that callback is implemented.
2) KVM may set the preempt flag for a short period whenver an vmexit
happens even if a vmenter is executed shortly after. In this case, we
may want to use a more durable vcpu suspend flag that indicates the vcpu
won't get a real vcpu back for a longer period of time.
Perhaps you can add a lock event counter to count the number of
wait_early events caused by vcpu_is_preempted() being true to see if it
really cause a lot more wait_early than without the vcpu_is_preempted()
call.
I have no objection to this, I just want to find out the root cause of it.
Cheers,
Longman
On Mon, 9 Sep 2019 at 19:06, Paolo Bonzini <[email protected]> wrote:
>
> On 09/09/19 12:56, Waiman Long wrote:
> > On 9/9/19 2:40 AM, Wanpeng Li wrote:
> >> From: Wanpeng Li <[email protected]>
> >>
> >> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
> >> vCPU is preempted), we found great regression caused by this commit.
> >>
> >> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
> >> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
> >> records/s with this commit.
> >>
> >> Host Guest score
> >>
> >> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
> >> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
> >> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
> >> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
> >>
> >> Exit from aggressive wait-early mechanism can result in yield premature and
> >> incur extra scheduling latency in over-subscribe scenario.
> >>
> >> kvm optimizes:
> >> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
> >> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
> >>
> >> Tested-by: [email protected]
> >> Cc: Peter Zijlstra <[email protected]>
> >> Cc: Thomas Gleixner <[email protected]>
> >> Cc: Ingo Molnar <[email protected]>
> >> Cc: Waiman Long <[email protected]>
> >> Cc: Paolo Bonzini <[email protected]>
> >> Cc: Radim Krčmář <[email protected]>
> >> Cc: [email protected]
> >> Cc: [email protected]
> >> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
> >> Signed-off-by: Wanpeng Li <[email protected]>
> >> ---
> >> kernel/locking/qspinlock_paravirt.h | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
> >> index 89bab07..e84d21a 100644
> >> --- a/kernel/locking/qspinlock_paravirt.h
> >> +++ b/kernel/locking/qspinlock_paravirt.h
> >> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
> >> if ((loop & PV_PREV_CHECK_MASK) != 0)
> >> return false;
> >>
> >> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
> >> + return READ_ONCE(prev->state) != vcpu_running;
> >> }
> >>
> >> /*
> >
> > There are several possibilities for this performance regression:
> >
> > 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
> > cacheline contention issue depending on how that callback is implemented.
>
> Unlikely, it is a single percpu read.
>
> > 2) KVM may set the preempt flag for a short period whenver an vmexit
> > happens even if a vmenter is executed shortly after. In this case, we
> > may want to use a more durable vcpu suspend flag that indicates the vcpu
> > won't get a real vcpu back for a longer period of time.
>
> It sets it for exits to userspace, but they shouldn't really happen on a
> properly-configured system.
>
> However, it's easy to test this theory:
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 2e302e977dac..feb6c75a7a88 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -3368,26 +3368,28 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
> {
> int idx;
>
> - if (vcpu->preempted)
> + if (vcpu->preempted) {
> vcpu->arch.preempted_in_kernel = !kvm_x86_ops->get_cpl(vcpu);
>
> - /*
> - * Disable page faults because we're in atomic context here.
> - * kvm_write_guest_offset_cached() would call might_fault()
> - * that relies on pagefault_disable() to tell if there's a
> - * bug. NOTE: the write to guest memory may not go through if
> - * during postcopy live migration or if there's heavy guest
> - * paging.
> - */
> - pagefault_disable();
> - /*
> - * kvm_memslots() will be called by
> - * kvm_write_guest_offset_cached() so take the srcu lock.
> - */
> - idx = srcu_read_lock(&vcpu->kvm->srcu);
> - kvm_steal_time_set_preempted(vcpu);
> - srcu_read_unlock(&vcpu->kvm->srcu, idx);
> - pagefault_enable();
> + /*
> + * Disable page faults because we're in atomic context here.
> + * kvm_write_guest_offset_cached() would call might_fault()
> + * that relies on pagefault_disable() to tell if there's a
> + * bug. NOTE: the write to guest memory may not go through if
> + * during postcopy live migration or if there's heavy guest
> + * paging.
> + */
> + pagefault_disable();
> + /*
> + * kvm_memslots() will be called by
> + * kvm_write_guest_offset_cached() so take the srcu lock.
> + */
> + idx = srcu_read_lock(&vcpu->kvm->srcu);
> + kvm_steal_time_set_preempted(vcpu);
> + srcu_read_unlock(&vcpu->kvm->srcu, idx);
> + pagefault_enable();
> + }
> +
> kvm_x86_ops->vcpu_put(vcpu);
> vcpu->arch.last_host_tsc = rdtsc();
> /*
>
> Wanpeng, can you try?
Yes, there is no difference for the score.
Wanpeng
On Mon, 9 Sep 2019 at 18:56, Waiman Long <[email protected]> wrote:
>
> On 9/9/19 2:40 AM, Wanpeng Li wrote:
> > From: Wanpeng Li <[email protected]>
> >
> > This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
> > vCPU is preempted), we found great regression caused by this commit.
> >
> > Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
> > The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
> > records/s with this commit.
> >
> > Host Guest score
> >
> > vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
> > vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
> > vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
> > vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
> >
> > Exit from aggressive wait-early mechanism can result in yield premature and
> > incur extra scheduling latency in over-subscribe scenario.
> >
> > kvm optimizes:
> > [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
> > [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
> >
> > Tested-by: [email protected]
> > Cc: Peter Zijlstra <[email protected]>
> > Cc: Thomas Gleixner <[email protected]>
> > Cc: Ingo Molnar <[email protected]>
> > Cc: Waiman Long <[email protected]>
> > Cc: Paolo Bonzini <[email protected]>
> > Cc: Radim Krčmář <[email protected]>
> > Cc: [email protected]
> > Cc: [email protected]
> > Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
> > Signed-off-by: Wanpeng Li <[email protected]>
> > ---
> > kernel/locking/qspinlock_paravirt.h | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
> > index 89bab07..e84d21a 100644
> > --- a/kernel/locking/qspinlock_paravirt.h
> > +++ b/kernel/locking/qspinlock_paravirt.h
> > @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
> > if ((loop & PV_PREV_CHECK_MASK) != 0)
> > return false;
> >
> > - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
> > + return READ_ONCE(prev->state) != vcpu_running;
> > }
> >
> > /*
>
> There are several possibilities for this performance regression:
>
> 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
> cacheline contention issue depending on how that callback is implemented.
>
> 2) KVM may set the preempt flag for a short period whenver an vmexit
> happens even if a vmenter is executed shortly after. In this case, we
> may want to use a more durable vcpu suspend flag that indicates the vcpu
> won't get a real vcpu back for a longer period of time.
>
> Perhaps you can add a lock event counter to count the number of
> wait_early events caused by vcpu_is_preempted() being true to see if it
> really cause a lot more wait_early than without the vcpu_is_preempted()
> call.
pv_wait_again:1:179
pv_wait_early:1:189429
pv_wait_head:1:263
pv_wait_node:1:189429
pv_vcpu_is_preempted:1:45588
=========sleep 5============
pv_wait_again:1:181
pv_wait_early:1:202574
pv_wait_head:1:267
pv_wait_node:1:202590
pv_vcpu_is_preempted:1:46336
The sampling period is 5s, 6% of wait_early events caused by
vcpu_is_preempted() being true.
Wanpeng
On 9/10/19 6:56 AM, Wanpeng Li wrote:
> On Mon, 9 Sep 2019 at 18:56, Waiman Long <[email protected]> wrote:
>> On 9/9/19 2:40 AM, Wanpeng Li wrote:
>>> From: Wanpeng Li <[email protected]>
>>>
>>> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
>>> vCPU is preempted), we found great regression caused by this commit.
>>>
>>> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
>>> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
>>> records/s with this commit.
>>>
>>> Host Guest score
>>>
>>> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
>>> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
>>> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
>>> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
>>>
>>> Exit from aggressive wait-early mechanism can result in yield premature and
>>> incur extra scheduling latency in over-subscribe scenario.
>>>
>>> kvm optimizes:
>>> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
>>> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
>>>
>>> Tested-by: [email protected]
>>> Cc: Peter Zijlstra <[email protected]>
>>> Cc: Thomas Gleixner <[email protected]>
>>> Cc: Ingo Molnar <[email protected]>
>>> Cc: Waiman Long <[email protected]>
>>> Cc: Paolo Bonzini <[email protected]>
>>> Cc: Radim Krčmář <[email protected]>
>>> Cc: [email protected]
>>> Cc: [email protected]
>>> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
>>> Signed-off-by: Wanpeng Li <[email protected]>
>>> ---
>>> kernel/locking/qspinlock_paravirt.h | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
>>> index 89bab07..e84d21a 100644
>>> --- a/kernel/locking/qspinlock_paravirt.h
>>> +++ b/kernel/locking/qspinlock_paravirt.h
>>> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
>>> if ((loop & PV_PREV_CHECK_MASK) != 0)
>>> return false;
>>>
>>> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
>>> + return READ_ONCE(prev->state) != vcpu_running;
>>> }
>>>
>>> /*
>> There are several possibilities for this performance regression:
>>
>> 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
>> cacheline contention issue depending on how that callback is implemented.
>>
>> 2) KVM may set the preempt flag for a short period whenver an vmexit
>> happens even if a vmenter is executed shortly after. In this case, we
>> may want to use a more durable vcpu suspend flag that indicates the vcpu
>> won't get a real vcpu back for a longer period of time.
>>
>> Perhaps you can add a lock event counter to count the number of
>> wait_early events caused by vcpu_is_preempted() being true to see if it
>> really cause a lot more wait_early than without the vcpu_is_preempted()
>> call.
> pv_wait_again:1:179
> pv_wait_early:1:189429
> pv_wait_head:1:263
> pv_wait_node:1:189429
> pv_vcpu_is_preempted:1:45588
> =========sleep 5============
> pv_wait_again:1:181
> pv_wait_early:1:202574
> pv_wait_head:1:267
> pv_wait_node:1:202590
> pv_vcpu_is_preempted:1:46336
>
> The sampling period is 5s, 6% of wait_early events caused by
> vcpu_is_preempted() being true.
6% isn't that high. However, when one vCPU voluntarily releases its
vCPU, all the subsequently waiters in the queue will do the same. It is
a cascading effect. Perhaps we wait early too aggressive with the
original patch.
I also look up the email chain of the original commit. The patch
submitter did not provide any performance data to support this change.
The patch just looked reasonable at that time. So there was no
objection. Given that we now have hard evidence that this was not a good
idea. I think we should revert it.
Reviewed-by: Waiman Long <[email protected]>
Thanks,
Longman
On 11/09/19 06:25, Waiman Long wrote:
> On 9/10/19 6:56 AM, Wanpeng Li wrote:
>> On Mon, 9 Sep 2019 at 18:56, Waiman Long <[email protected]> wrote:
>>> On 9/9/19 2:40 AM, Wanpeng Li wrote:
>>>> From: Wanpeng Li <[email protected]>
>>>>
>>>> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
>>>> vCPU is preempted), we found great regression caused by this commit.
>>>>
>>>> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
>>>> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
>>>> records/s with this commit.
>>>>
>>>> Host Guest score
>>>>
>>>> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
>>>> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
>>>> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
>>>> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
>>>>
>>>> Exit from aggressive wait-early mechanism can result in yield premature and
>>>> incur extra scheduling latency in over-subscribe scenario.
>>>>
>>>> kvm optimizes:
>>>> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
>>>> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
>>>>
>>>> Tested-by: [email protected]
>>>> Cc: Peter Zijlstra <[email protected]>
>>>> Cc: Thomas Gleixner <[email protected]>
>>>> Cc: Ingo Molnar <[email protected]>
>>>> Cc: Waiman Long <[email protected]>
>>>> Cc: Paolo Bonzini <[email protected]>
>>>> Cc: Radim Krčmář <[email protected]>
>>>> Cc: [email protected]
>>>> Cc: [email protected]
>>>> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
>>>> Signed-off-by: Wanpeng Li <[email protected]>
>>>> ---
>>>> kernel/locking/qspinlock_paravirt.h | 2 +-
>>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
>>>> index 89bab07..e84d21a 100644
>>>> --- a/kernel/locking/qspinlock_paravirt.h
>>>> +++ b/kernel/locking/qspinlock_paravirt.h
>>>> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
>>>> if ((loop & PV_PREV_CHECK_MASK) != 0)
>>>> return false;
>>>>
>>>> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
>>>> + return READ_ONCE(prev->state) != vcpu_running;
>>>> }
>>>>
>>>> /*
>>> There are several possibilities for this performance regression:
>>>
>>> 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
>>> cacheline contention issue depending on how that callback is implemented.
>>>
>>> 2) KVM may set the preempt flag for a short period whenver an vmexit
>>> happens even if a vmenter is executed shortly after. In this case, we
>>> may want to use a more durable vcpu suspend flag that indicates the vcpu
>>> won't get a real vcpu back for a longer period of time.
>>>
>>> Perhaps you can add a lock event counter to count the number of
>>> wait_early events caused by vcpu_is_preempted() being true to see if it
>>> really cause a lot more wait_early than without the vcpu_is_preempted()
>>> call.
>> pv_wait_again:1:179
>> pv_wait_early:1:189429
>> pv_wait_head:1:263
>> pv_wait_node:1:189429
>> pv_vcpu_is_preempted:1:45588
>> =========sleep 5============
>> pv_wait_again:1:181
>> pv_wait_early:1:202574
>> pv_wait_head:1:267
>> pv_wait_node:1:202590
>> pv_vcpu_is_preempted:1:46336
>>
>> The sampling period is 5s, 6% of wait_early events caused by
>> vcpu_is_preempted() being true.
>
> 6% isn't that high. However, when one vCPU voluntarily releases its
> vCPU, all the subsequently waiters in the queue will do the same. It is
> a cascading effect. Perhaps we wait early too aggressive with the
> original patch.
>
> I also look up the email chain of the original commit. The patch
> submitter did not provide any performance data to support this change.
> The patch just looked reasonable at that time. So there was no
> objection. Given that we now have hard evidence that this was not a good
> idea. I think we should revert it.
>
> Reviewed-by: Waiman Long <[email protected]>
>
> Thanks,
> Longman
>
Queued, thanks.
Paolo
On Wed, 11 Sep 2019 at 21:04, Paolo Bonzini <[email protected]> wrote:
>
> On 11/09/19 06:25, Waiman Long wrote:
> > On 9/10/19 6:56 AM, Wanpeng Li wrote:
> >> On Mon, 9 Sep 2019 at 18:56, Waiman Long <[email protected]> wrote:
> >>> On 9/9/19 2:40 AM, Wanpeng Li wrote:
> >>>> From: Wanpeng Li <[email protected]>
> >>>>
> >>>> This patch reverts commit 75437bb304b20 (locking/pvqspinlock: Don't wait if
> >>>> vCPU is preempted), we found great regression caused by this commit.
> >>>>
> >>>> Xeon Skylake box, 2 sockets, 40 cores, 80 threads, three VMs, each is 80 vCPUs.
> >>>> The score of ebizzy -M can reduce from 13000-14000 records/s to 1700-1800
> >>>> records/s with this commit.
> >>>>
> >>>> Host Guest score
> >>>>
> >>>> vanilla + w/o kvm optimizes vanilla 1700-1800 records/s
> >>>> vanilla + w/o kvm optimizes vanilla + revert 13000-14000 records/s
> >>>> vanilla + w/ kvm optimizes vanilla 4500-5000 records/s
> >>>> vanilla + w/ kvm optimizes vanilla + revert 14000-15500 records/s
> >>>>
> >>>> Exit from aggressive wait-early mechanism can result in yield premature and
> >>>> incur extra scheduling latency in over-subscribe scenario.
> >>>>
> >>>> kvm optimizes:
> >>>> [1] commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts)
> >>>> [2] commit 266e85a5ec9 (KVM: X86: Boost queue head vCPU to mitigate lock waiter preemption)
> >>>>
> >>>> Tested-by: [email protected]
> >>>> Cc: Peter Zijlstra <[email protected]>
> >>>> Cc: Thomas Gleixner <[email protected]>
> >>>> Cc: Ingo Molnar <[email protected]>
> >>>> Cc: Waiman Long <[email protected]>
> >>>> Cc: Paolo Bonzini <[email protected]>
> >>>> Cc: Radim Krčmář <[email protected]>
> >>>> Cc: [email protected]
> >>>> Cc: [email protected]
> >>>> Fixes: 75437bb304b20 (locking/pvqspinlock: Don't wait if vCPU is preempted)
> >>>> Signed-off-by: Wanpeng Li <[email protected]>
> >>>> ---
> >>>> kernel/locking/qspinlock_paravirt.h | 2 +-
> >>>> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h
> >>>> index 89bab07..e84d21a 100644
> >>>> --- a/kernel/locking/qspinlock_paravirt.h
> >>>> +++ b/kernel/locking/qspinlock_paravirt.h
> >>>> @@ -269,7 +269,7 @@ pv_wait_early(struct pv_node *prev, int loop)
> >>>> if ((loop & PV_PREV_CHECK_MASK) != 0)
> >>>> return false;
> >>>>
> >>>> - return READ_ONCE(prev->state) != vcpu_running || vcpu_is_preempted(prev->cpu);
> >>>> + return READ_ONCE(prev->state) != vcpu_running;
> >>>> }
> >>>>
> >>>> /*
> >>> There are several possibilities for this performance regression:
> >>>
> >>> 1) Multiple vcpus calling vcpu_is_preempted() repeatedly may cause some
> >>> cacheline contention issue depending on how that callback is implemented.
> >>>
> >>> 2) KVM may set the preempt flag for a short period whenver an vmexit
> >>> happens even if a vmenter is executed shortly after. In this case, we
> >>> may want to use a more durable vcpu suspend flag that indicates the vcpu
> >>> won't get a real vcpu back for a longer period of time.
> >>>
> >>> Perhaps you can add a lock event counter to count the number of
> >>> wait_early events caused by vcpu_is_preempted() being true to see if it
> >>> really cause a lot more wait_early than without the vcpu_is_preempted()
> >>> call.
> >> pv_wait_again:1:179
> >> pv_wait_early:1:189429
> >> pv_wait_head:1:263
> >> pv_wait_node:1:189429
> >> pv_vcpu_is_preempted:1:45588
> >> =========sleep 5============
> >> pv_wait_again:1:181
> >> pv_wait_early:1:202574
> >> pv_wait_head:1:267
> >> pv_wait_node:1:202590
> >> pv_vcpu_is_preempted:1:46336
> >>
> >> The sampling period is 5s, 6% of wait_early events caused by
> >> vcpu_is_preempted() being true.
> >
> > 6% isn't that high. However, when one vCPU voluntarily releases its
> > vCPU, all the subsequently waiters in the queue will do the same. It is
> > a cascading effect. Perhaps we wait early too aggressive with the
> > original patch.
> >
> > I also look up the email chain of the original commit. The patch
> > submitter did not provide any performance data to support this change.
> > The patch just looked reasonable at that time. So there was no
> > objection. Given that we now have hard evidence that this was not a good
> > idea. I think we should revert it.
> >
> > Reviewed-by: Waiman Long <[email protected]>
> >
> > Thanks,
> > Longman
> >
>
> Queued, thanks.
Didn't see it in yesterday's updated kvm/queue. :)
Wanpeng