Re: [PATCH v3 0/3] KVM: Yield to IPI target if necessary

From: Nadav Amit
Date: Tue Jun 11 2019 - 21:42:21 EST

> On Jun 11, 2019, at 6:18 PM, Wanpeng Li <kernellwp@xxxxxxxxx> wrote:
> On Wed, 12 Jun 2019 at 00:57, Nadav Amit <nadav.amit@xxxxxxxxx> wrote:
>>> On Jun 11, 2019, at 3:02 AM, Wanpeng Li <kernellwp@xxxxxxxxx> wrote:
>>> On Tue, 11 Jun 2019 at 09:48, Nadav Amit <nadav.amit@xxxxxxxxx> wrote:
>>>>> On Jun 10, 2019, at 6:45 PM, Wanpeng Li <kernellwp@xxxxxxxxx> wrote:
>>>>> On Tue, 11 Jun 2019 at 09:11, Sean Christopherson
>>>>> <sean.j.christopherson@xxxxxxxxx> wrote:
>>>>>> On Mon, Jun 10, 2019 at 04:34:20PM +0200, Radim KrÄmÃÅ wrote:
>>>>>>> 2019-05-30 09:05+0800, Wanpeng Li:
>>>>>>>> The idea is from Xen, when sending a call-function IPI-many to vCPUs,
>>>>>>>> yield if any of the IPI target vCPUs was preempted. 17% performance
>>>>>>>> increasement of ebizzy benchmark can be observed in an over-subscribe
>>>>>>>> environment. (w/ kvm-pv-tlb disabled, testing TLB flush call-function
>>>>>>>> IPI-many since call-function is not easy to be trigged by userspace
>>>>>>>> workload).
>>>>>>> Have you checked if we could gain performance by having the yield as an
>>>>>>> extension to our PV IPI call?
>>>>>>> It would allow us to skip the VM entry/exit overhead on the caller.
>>>>>>> (The benefit of that might be negligible and it also poses a
>>>>>>> complication when splitting the target mask into several PV IPI
>>>>>>> hypercalls.)
>>>>>> Tangetially related to splitting PV IPI hypercalls, are there any major
>>>>>> hurdles to supporting shorthand? Not having to generate the mask for
>>>>>> ->send_IPI_allbutself and ->kvm_send_ipi_all seems like an easy to way
>>>>>> shave cycles for affected flows.
>>>>> Not sure why shorthand is not used for native x2apic mode.
>>>> Why do you say so? native_send_call_func_ipi() checks if allbutself
>>>> shorthand should be used and does so (even though the check can be more
>>>> efficient - Iâm looking at that code right nowâ)
>>> Please continue to follow the apic/x2apic driver. Just apic_flat set
>> Indeed - I was sure by the name that it does it correctly. Thatâs stupid.
>> Iâll add it to the patch-set I am working on (TLB shootdown improvements),
>> if you donât mind.
> Original for hotplug cpu safe.
> Not sure shortcut native support is acceptable, I will play my
> kvm_send_ipi_allbutself and kvm_send_ipi_all. :)

Yes, I saw these threads before. But I think the test in
native_send_call_func_ipi() should take care of it.

Iâll recheck.