From mboxrd@z Thu Jan  1 00:00:00 1970
From: Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [CFT PATCH 0/2] KVM: support XSAVES usage in the host
Date: Mon, 24 Nov 2014 18:53:40 +0100
Message-ID: <547370A4.3070707@redhat.com>
References: <1416594678-13011-1-git-send-email-pbonzini@redhat.com> <DE9192FA-6A22-4AEB-88C2-AE17DA503685@gmail.com> <547318E5.8070408@redhat.com> <DAC65E0B-06CC-4B34-85B2-72AD5F86826B@gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	kvm list <kvm@vger.kernel.org>,
	Nadav Amit <namit@cs.technion.ac.il>,
	Wanpeng Li <wanpeng.li@linux.intel.com>
To: Nadav Amit <nadav.amit@gmail.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <DAC65E0B-06CC-4B34-85B2-72AD5F86826B@gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: kvm.vger.kernel.org



On 24/11/2014 16:28, Nadav Amit wrote:
>=20
> Since kvm_load_guest_fpu is called before the guest_fpu is ever store=
d, there are 2 more problems that currently cause #GP:
> 1. XCOMP_BV[63] =3D 0
> 2. XSTATE_BV sets a bit (including bit 63) that is not set in XCOMP_B=
V (XCOMP_BV is initialised to zero).
>=20
> [see SDM 13.11 "OPERATION OF XRSTORS=E2=80=9D]
>=20
> Once I initialise XCOMP_BV to (1ull << 63) | XSTATE_BV, the guest run=
s successfully.
> I have not checked any other qemu functionality that might be affecte=
d by the patch.

I posted patches that assume that QEMU calls KVM_SET_XSAVE early enough=
=2E
 If this is not the case, can you cook up and post a patch to
kvm_arch_vcpu_init that fixes the remaining problem?

Paolo