From mboxrd@z Thu Jan 1 00:00:00 1970 From: Junaid Shahid Subject: Re: [PATCH v3 7/8] kvm: x86: mmu: Lockless access tracking for Intel CPUs without EPT A bits. Date: Mon, 19 Dec 2016 19:36:44 -0800 Message-ID: <2941755.AtdGOR38c0@js-desktop.mtv.corp.google.com> References: <4157789.R9cn7kUSZu@js-desktop.mtv.corp.google.com> <1942485779.4450132.1481984369832.JavaMail.zimbra@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Cc: Xiao Guangrong , kvm@vger.kernel.org, andreslc@google.com, pfeiner@google.com To: Paolo Bonzini Return-path: Received: from mail-pg0-f51.google.com ([74.125.83.51]:36797 "EHLO mail-pg0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752111AbcLTDgq (ORCPT ); Mon, 19 Dec 2016 22:36:46 -0500 Received: by mail-pg0-f51.google.com with SMTP id f188so67560036pgc.3 for ; Mon, 19 Dec 2016 19:36:46 -0800 (PST) In-Reply-To: <1942485779.4450132.1481984369832.JavaMail.zimbra@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Saturday, December 17, 2016 09:19:29 AM Paolo Bonzini wrote: > >=20 > > Yes, I guess we can do that since if the restored PTE doesn=E2=80=99= t satisfy the > > access we are just going to get another fault anyway. >=20 > Please do it as a follow up, since it complicates the logic a bit. >=20 > .... > >=20 > > Ok. So you mean that vmx.c should just pass VMX_EPT_RWX_MASK here a= nd > > VMX_EPT_MISCONFIG_WX_VALUE for the mmio mask and then mmu.c should = add in > > SPTE_SPECIAL_MASK before storing these values in shadow_acc_track_m= ask and > > shadow_mmio_mask? >=20 > I think I agree, but we can do this too as a separate follow-up clean= up patch. >=20 Sure. I=E2=80=99ll defer these to follow-up patches. What about the cha= nge to move the pgtable walk inside the retry loop in fast_page_fault? = Should I update the current patch-set to do that or should we defer tha= t to a later patch as well? Thanks, Junaid