RE: [RFC PATCH 1/4] KVM: x86: Split the APIC from the rest of IRQCHIP.

From: "Wu, Feng" <feng.wu@intel.com>
To: Andrew Honig <ahonig@google.com>
Cc: Steve Rutherford <srutherford@google.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	"kvm@vger.kernel.org" <kvm@vger.kernel.org>,
	"Wu, Feng" <feng.wu@intel.com>
Subject: RE: [RFC PATCH 1/4] KVM: x86: Split the APIC from the rest of IRQCHIP.
Date: Fri, 15 May 2015 01:28:08 +0000	[thread overview]
Message-ID: <E959C4978C3B6342920538CF579893F00252AA6F@SHSMSX104.ccr.corp.intel.com> (raw)
In-Reply-To: <CAKB9nXuzkeYZAiyTcvu=QgqN9TQB0kCr3-8CAvkREN-7Z=ieNw@mail.gmail.com>

> -----Original Message-----
> From: Andrew Honig [mailto:ahonig@google.com]
> Sent: Friday, May 15, 2015 3:29 AM
> To: Wu, Feng
> Cc: Steve Rutherford; Paolo Bonzini; kvm@vger.kernel.org
> Subject: Re: [RFC PATCH 1/4] KVM: x86: Split the APIC from the rest of IRQCHIP.
> 
> >
> > BTW, what is the purpose of this series. If I understand it correctly, you only
> want to
> > use the in-kernel lapic and leave the others (pic, ioapic) in userspace, what is
> the
> > benefit of it?
> 
> The purpose is to achieve the security benefit of removing some of the
> interrupt handling into userspace, without incurring a significant
> performance penalty.  If you move the entire IRQCHIP into userspace,
> we've seen perf impacts from 15-200% depending on the workload.  With
> this patch series, we're seeing perf penalty <1% on our tests (TCP_RR
> latency, TCP throughput, and Disk I/O).  See
> (https://lwn.net/Articles/619332/)

Good to know this, thanks for the sharing!

Thanks,
Feng

> 
> >
> > Thanks,
> > Feng
> >
> >>