From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [RFC] KVM: Balloon support for device assignment Date: Sun, 21 Feb 2010 15:25:32 +0200 Message-ID: <4B81344C.30809@redhat.com> References: <1266399807-4498-1-git-send-email-borove@il.ibm.com> <1266399807-4498-2-git-send-email-borove@il.ibm.com> <4B7BC47D.4030808@redhat.com> <20100217175237.GI11999@il.ibm.com> <4B80F485.50704@redhat.com> <20100221131732.GC7677@il.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Eran Borovik , kvm@vger.kernel.org, Ben-Ami Yassour , Gerd Hoffmann To: Muli Ben-Yehuda Return-path: Received: from mx1.redhat.com ([209.132.183.28]:50808 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751984Ab0BUNZg (ORCPT ); Sun, 21 Feb 2010 08:25:36 -0500 In-Reply-To: <20100221131732.GC7677@il.ibm.com> Sender: kvm-owner@vger.kernel.org List-ID: On 02/21/2010 03:17 PM, Muli Ben-Yehuda wrote: > On Sun, Feb 21, 2010 at 10:53:25AM +0200, Avi Kivity wrote: > > >> I'm not worried about transition cost (that's for whoever makes the >> transition to pay) but about how the APIs would translate to the new >> way of doing things. >> >> For example, it might be done using mmu notifiers. We might use the >> existing madvise(MADV_DONTNEED) as the entry point (though I don't >> know how we would indicate deflation). >> > I guess we could extend mmu notifiers to notify on any PTE transition, > but that seems like a lot of VM complexity for a little gain, given > the simple KVM-internal alternatives. > mmu notifiers already notify on any pte transition. -- error compiling committee.c: too many arguments to function