From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH 16/22] KVM: MMU: Track page fault data in struct vcpu Date: Tue, 27 Apr 2010 15:58:36 +0300 Message-ID: <4BD6DF7C.1090203@redhat.com> References: <1272364712-17425-1-git-send-email-joerg.roedel@amd.com> <1272364712-17425-17-git-send-email-joerg.roedel@amd.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , kvm@vger.kernel.org, linux-kernel@vger.kernel.org To: Joerg Roedel Return-path: Received: from mx1.redhat.com ([209.132.183.28]:36563 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754046Ab0D0M6j (ORCPT ); Tue, 27 Apr 2010 08:58:39 -0400 In-Reply-To: <1272364712-17425-17-git-send-email-joerg.roedel@amd.com> Sender: kvm-owner@vger.kernel.org List-ID: On 04/27/2010 01:38 PM, Joerg Roedel wrote: > This patch introduces two fields in vcpu_arch for x86: > > * fault_address > * fault_error_code > > This will be used to correctly propagate page faults back > into the guest when we could have either an ordinary page > fault or a nested page fault. In the case of a nested page > fault the fault-address is different from the original > address that should be walked. So we need to keep track > about the real fault-address. > We could also remove the current path of the error_code to > the fault. But this change is too invasive and outside the > scope of this patch set. It will be changed and tested > seperatly. > > Signed-off-by: Joerg Roedel > --- > arch/x86/include/asm/kvm_host.h | 3 +++ > arch/x86/kvm/paging_tmpl.h | 4 ++++ > 2 files changed, 7 insertions(+), 0 deletions(-) > > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h > index d9dfc8c..8426870 100644 > --- a/arch/x86/include/asm/kvm_host.h > +++ b/arch/x86/include/asm/kvm_host.h > @@ -298,6 +298,9 @@ struct kvm_vcpu_arch { > /* Used for two dimensional paging emulation */ > struct kvm_mmu nested_mmu; > > + unsigned long fault_address; > Probably a problem on i386. How does npt handle faults when the guest is using pae paging and the host (in our case the guest...) isn't? I see it uses exit_info_2 for the address, which is a u64. So we probably need to upgrade gva_t to a u64. Please send this as a separate patch, and test on i386 hosts. > + int fault_error_code; > unsigned. Maybe put the two in a struct, easier to pass around. -- error compiling committee.c: too many arguments to function