From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] KVM: MMU: fix huge page adapted on non-PAE host Date: Tue, 29 May 2012 17:23:34 +0300 Message-ID: <4FC4DBE6.1040703@redhat.com> References: <4FC316E3.6080607@linux.vnet.ibm.com> <4FC35A15.6080000@redhat.com> <4FC363EE.6060204@linux.vnet.ibm.com> <4FC36E85.4010909@redhat.com> <4FC37600.1060301@linux.vnet.ibm.com> <4FC37A18.10809@redhat.com> <4FC38084.40409@linux.vnet.ibm.com> <4FC38362.6010802@redhat.com> <20120528143221.GF4016@redhat.com> <4FC38E48.9080701@redhat.com> <20120528144439.GH4016@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Xiao Guangrong , Marcelo Tosatti , LKML , KVM To: Andrea Arcangeli Return-path: In-Reply-To: <20120528144439.GH4016@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On 05/28/2012 05:44 PM, Andrea Arcangeli wrote: > On Mon, May 28, 2012 at 05:40:08PM +0300, Avi Kivity wrote: >> Yes, I see it now. Adjusting mask is incorrect since we won't have the >> same adjustment on release. I'll apply the patch for 3.5. > > Sounds great to me. One thing I'm not sure about is about the real need of > the mmio check vs a stright get_page (we shouldn't ever get to a gup-fast > succeeding if this was a mmio region?) but I don't see chances that it > could hurt either. This is a device assignment mmio region. gpu_fast() can succeed is we happen to have a struct page for it (pci hole) or fail if we don't (64-bit BAR). See hva_to_pfn() (and indeed this is the reason we use pfns instead of struct pages). -- error compiling committee.c: too many arguments to function