From mboxrd@z Thu Jan  1 00:00:00 1970
From: Andi Kleen <ak@linux.intel.com>
Subject: Re: [PATCH] KVM: VMX: Cache IA32_DEBUGCTL in memory
Date: Wed, 29 Nov 2017 12:56:24 -0800
Message-ID: <20171129205624.GD3070@tassilo.jf.intel.com>
References: <1511935673-7371-1-git-send-email-wanpeng.li@hotmail.com>
 <87fu8xj69o.fsf@linux.intel.com>
 <CALMp9eSmqa8Wgvcwa=58CvBndfKQ6N_Uf8HTpMPe8WC_gM-iGA@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Wanpeng Li <kernellwp@gmail.com>,
        LKML <linux-kernel@vger.kernel.org>,
        kvm list <kvm@vger.kernel.org>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Radim =?utf-8?B?S3LEjW3DocWZ?= <rkrcmar@redhat.com>,
        Wanpeng Li <wanpeng.li@hotmail.com>
To: Jim Mattson <jmattson@google.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <CALMp9eSmqa8Wgvcwa=58CvBndfKQ6N_Uf8HTpMPe8WC_gM-iGA@mail.gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: kvm.vger.kernel.org

On Wed, Nov 29, 2017 at 11:05:46AM -0800, Jim Mattson wrote:
> An alternative is to give the L1 guest read permission for this MSR in
> the MSR permission bitmaps. It's still going to be ~80 cycles, but
> that's better than the cost of a VM-exit/VM-entry round-trip.

It's a useful optimization, 80 cycles is 80 cycles.

The cache invalidation could likely be really simple, like:
have a global counter
always check the counter before and after and don't use the cache
if they don't match.
change KDB etc. to increase the counter.

-Andi