[0/3] KVM: X86: Single target IPI fastpath enhancement

Message

Wanpeng Li March 26, 2020, 2:19 a.m. UTC

The original single target IPI fastpath patch forgot to filter the 
ICR destination shorthand field. Multicast IPI is not suitable for 
this feature since wakeup the multiple sleeping vCPUs will extend 
the interrupt disabled time, it especially worse in the over-subscribe 
and VM has a little bit more vCPUs scenario. Let's narrow it down to 
single target IPI. In addition, this patchset micro-optimize virtual 
IPI emulation sequence for fastpath.

Wanpeng Li (3):
  KVM: X86: Delay read msr data iff writes ICR MSR
  KVM: X86: Narrow down the IPI fastpath to single target IPI
  KVM: X86: Micro-optimize IPI fastpath delay

 arch/x86/kvm/lapic.c |  4 ++--
 arch/x86/kvm/lapic.h |  1 +
 arch/x86/kvm/x86.c   | 14 +++++++++++---
 3 files changed, 14 insertions(+), 5 deletions(-)

Comments

Paolo Bonzini March 26, 2020, 9:46 a.m. UTC | #1

On 26/03/20 03:19, Wanpeng Li wrote:
> The original single target IPI fastpath patch forgot to filter the 
> ICR destination shorthand field. Multicast IPI is not suitable for 
> this feature since wakeup the multiple sleeping vCPUs will extend 
> the interrupt disabled time, it especially worse in the over-subscribe 
> and VM has a little bit more vCPUs scenario. Let's narrow it down to 
> single target IPI. In addition, this patchset micro-optimize virtual 
> IPI emulation sequence for fastpath.
> 
> Wanpeng Li (3):
>   KVM: X86: Delay read msr data iff writes ICR MSR
>   KVM: X86: Narrow down the IPI fastpath to single target IPI
>   KVM: X86: Micro-optimize IPI fastpath delay
> 
>  arch/x86/kvm/lapic.c |  4 ++--
>  arch/x86/kvm/lapic.h |  1 +
>  arch/x86/kvm/x86.c   | 14 +++++++++++---
>  3 files changed, 14 insertions(+), 5 deletions(-)
> 

Queued 2 for 5.6 and 1-3 for 5.7, thanks.

Paolo