From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:43006) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VHZGh-0000as-OX for qemu-devel@nongnu.org; Thu, 05 Sep 2013 09:06:48 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VHZGZ-0004IO-AF for qemu-devel@nongnu.org; Thu, 05 Sep 2013 09:06:39 -0400 Received: from mail-ea0-x22e.google.com ([2a00:1450:4013:c01::22e]:38910) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VHZGZ-0004IK-2m for qemu-devel@nongnu.org; Thu, 05 Sep 2013 09:06:31 -0400 Received: by mail-ea0-f174.google.com with SMTP id z15so888972ead.33 for ; Thu, 05 Sep 2013 06:06:30 -0700 (PDT) Sender: Paolo Bonzini From: Paolo Bonzini Date: Thu, 5 Sep 2013 15:06:22 +0200 Message-Id: <1378386382-415-3-git-send-email-pbonzini@redhat.com> In-Reply-To: <1378386382-415-1-git-send-email-pbonzini@redhat.com> References: <1378386382-415-1-git-send-email-pbonzini@redhat.com> Subject: [Qemu-devel] [PATCH uq/master 2/2] KVM: make XSAVE support more robust List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: ehabkost@redhat.com, gleb@redhat.com, kvm@vger.kernel.org QEMU moves state from CPUArchState to struct kvm_xsave and back when it invokes the KVM_*_XSAVE ioctls. Because it doesn't treat the XSAVE region as an opaque blob, it might be impossible to set some state on the destination if migrating to an older version. This patch blocks migration if it finds that unsupported bits are set in the XSTATE_BV header field. To make this work robustly, QEMU should only report in env->xstate_bv those fields that will actually end up in the migration stream. Signed-off-by: Paolo Bonzini --- target-i386/kvm.c | 3 ++- target-i386/machine.c | 4 ++++ 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/target-i386/kvm.c b/target-i386/kvm.c index 749aa09..df08a4b 100644 --- a/target-i386/kvm.c +++ b/target-i386/kvm.c @@ -1291,7 +1291,8 @@ static int kvm_get_xsave(X86CPU *cpu) sizeof env->fpregs); memcpy(env->xmm_regs, &xsave->region[XSAVE_XMM_SPACE], sizeof env->xmm_regs); - env->xstate_bv = *(uint64_t *)&xsave->region[XSAVE_XSTATE_BV]; + env->xstate_bv = *(uint64_t *)&xsave->region[XSAVE_XSTATE_BV] & + XSTATE_SUPPORTED; memcpy(env->ymmh_regs, &xsave->region[XSAVE_YMMH_SPACE], sizeof env->ymmh_regs); return 0; diff --git a/target-i386/machine.c b/target-i386/machine.c index dc81cde..9e2cfcf 100644 --- a/target-i386/machine.c +++ b/target-i386/machine.c @@ -278,6 +278,10 @@ static int cpu_post_load(void *opaque, int version_id) CPUX86State *env = &cpu->env; int i; + if (env->xstate_bv & ~XSTATE_SUPPORTED) { + return -EINVAL; + } + /* * Real mode guest segments register DPL should be zero. * Older KVM version were setting it wrongly. -- 1.8.3.1