From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:37390) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XowW6-0004nx-6u for qemu-devel@nongnu.org; Thu, 13 Nov 2014 10:41:06 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XowW1-0007oW-0A for qemu-devel@nongnu.org; Thu, 13 Nov 2014 10:41:01 -0500 Received: from mail-wi0-x22c.google.com ([2a00:1450:400c:c05::22c]:55577) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XowW0-0007oN-QL for qemu-devel@nongnu.org; Thu, 13 Nov 2014 10:40:56 -0500 Received: by mail-wi0-f172.google.com with SMTP id bs8so8729290wib.5 for ; Thu, 13 Nov 2014 07:40:56 -0800 (PST) Sender: Paolo Bonzini From: Paolo Bonzini Date: Thu, 13 Nov 2014 16:40:26 +0100 Message-Id: <1415893228-25823-12-git-send-email-pbonzini@redhat.com> In-Reply-To: <1415893228-25823-1-git-send-email-pbonzini@redhat.com> References: <1415893228-25823-1-git-send-email-pbonzini@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Subject: [Qemu-devel] [PULL 11/13] kvmclock: Add comment explaining why we need cpu_clean_all_dirty() List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Andrey Korolyov , =?UTF-8?q?Marcin=20Gibu=C3=85=E2=80=9Aa?= , Eduardo Habkost , Marcelo Tosatti From: Eduardo Habkost Try to explain why commit 317b0a6d8ba44e9bf8f9c3dbd776c4536843d82c needed a cpu_clean_all_dirty() call just after calling cpu_synchronize_all_states(). Signed-off-by: Eduardo Habkost Cc: Andrey Korolyov Cc: Marcin Gibuła Cc: Marcelo Tosatti Cc: Paolo Bonzini Signed-off-by: Paolo Bonzini --- hw/i386/kvm/clock.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/hw/i386/kvm/clock.c b/hw/i386/kvm/clock.c index 1ac60d6..58be2bd 100644 --- a/hw/i386/kvm/clock.c +++ b/hw/i386/kvm/clock.c @@ -127,7 +127,21 @@ static void kvmclock_vm_state_change(void *opaque, int running, } cpu_synchronize_all_states(); + /* In theory, the cpu_synchronize_all_states() call above wouldn't + * affect the rest of the code, as the VCPU state inside CPUState + * is supposed to always match the VCPU state on the kernel side. + * + * In practice, calling cpu_synchronize_state() too soon will load the + * kernel-side APIC state into X86CPU.apic_state too early, APIC state + * won't be reloaded later because CPUState.vcpu_dirty==true, and + * outdated APIC state may be migrated to another host. + * + * The real fix would be to make sure outdated APIC state is read + * from the kernel again when necessary. While this is not fixed, we + * need the cpu_clean_all_dirty() call below. + */ cpu_clean_all_dirty(); + ret = kvm_vm_ioctl(kvm_state, KVM_GET_CLOCK, &data); if (ret < 0) { fprintf(stderr, "KVM_GET_CLOCK failed: %s\n", strerror(ret)); -- 1.8.3.1