From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Date: Thu, 29 Mar 2012 15:26:59 +0000 Subject: Re: Question about removing memslots Message-Id: <4F747F43.9040601@redhat.com> List-Id: References: <1332919486.2882.88.camel@pasglop> <4F72DBE2.6060909@redhat.com> <20120329141553.3bc61de3.yoshikawa.takuya@oss.ntt.co.jp> <4F742EEC.3020501@redhat.com> <20120330002105.a66858d82bdd1802cae7b401@gmail.com> In-Reply-To: <20120330002105.a66858d82bdd1802cae7b401@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Takuya Yoshikawa Cc: Takuya Yoshikawa , Benjamin Herrenschmidt , kvm@vger.kernel.org, Alexander Graf , kvm-ppc@vger.kernel.org, Scott Wood , Alex Williamson On 03/29/2012 05:21 PM, Takuya Yoshikawa wrote: > On Thu, 29 Mar 2012 11:44:12 +0200 > Avi Kivity wrote: > > > > Even without using reverse mapping we can restrict that flush easily: > > > > > > http://www.spinics.net/lists/kvm/msg68695.html > > > [PATCH] KVM: Avoid zapping unrelated shadows in __kvm_set_memory_region() > > > > > > This would be better than using reverse mapping because we do not have so > > > many shadow pages when we are in a tight loop like you mensioned. > > > > > > Anyway we could easily see tens of milliseconds difference by eliminating > > > unrelated flush. > > > > Hm, the patch uses ->slot_bitmap which we might want to kill if we > > increase the number of slots dramatically, as some people want to do. > > > > btw, what happened to that patch, did it just get ignored on the list? > > I did not get any comments, maybe because it was during around your vacation. Care to refresh it? I think it's worthwhile. And please ping me (or Marcelo, or others) if you get no reviews. -- error compiling committee.c: too many arguments to function