From patchwork Thu Oct 29 08:22:06 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Li, Liang Z" X-Patchwork-Id: 537724 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [IPv6:2001:4830:134:3::11]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 7D440140D97 for ; Thu, 29 Oct 2015 19:27:47 +1100 (AEDT) Received: from localhost ([::1]:42621 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZriYj-0006Rk-5U for incoming@patchwork.ozlabs.org; Thu, 29 Oct 2015 04:27:45 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:48507) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZriYE-0005e6-1O for qemu-devel@nongnu.org; Thu, 29 Oct 2015 04:27:15 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZriYC-0007lz-46 for qemu-devel@nongnu.org; Thu, 29 Oct 2015 04:27:13 -0400 Received: from mga09.intel.com ([134.134.136.24]:32862) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZriYB-0007li-QU for qemu-devel@nongnu.org; Thu, 29 Oct 2015 04:27:12 -0400 Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by orsmga102.jf.intel.com with ESMTP; 29 Oct 2015 01:27:12 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.20,213,1444719600"; d="scan'208";a="590173924" Received: from ll.sh.intel.com (HELO localhost) ([10.239.13.27]) by FMSMGA003.fm.intel.com with ESMTP; 29 Oct 2015 01:27:10 -0700 From: Liang Li To: qemu-devel@nongnu.org Date: Thu, 29 Oct 2015 16:22:06 +0800 Message-Id: <1446106927-15490-2-git-send-email-liang.z.li@intel.com> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1446106927-15490-1-git-send-email-liang.z.li@intel.com> References: <1446106927-15490-1-git-send-email-liang.z.li@intel.com> X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 134.134.136.24 Cc: pbonzini@redhat.com, mtosatti@redhat.com, m.gibula@beyond.pl, Liang Li Subject: [Qemu-devel] [RESEND 1/2] kvmclock: use a light weight interface to update env->tsc. X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org The commit 317b0a6d8 fixed an issue which caused by the outdated env->tsc value, but the fix lead to cpu_synchronize_all_states called twice during live migration. The cpu_synchronize_all_states takes about 130us for a VM which has 4 vcpus, it's a bit expensive. This patch use a light weight interface to update the env->tsc value, it only takes about 20us to update the env->tsc. Signed-off-by: Liang Li --- hw/i386/kvm/clock.c | 26 +++++++++----------------- target-i386/kvm.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++ target-i386/kvm_i386.h | 1 + 3 files changed, 56 insertions(+), 17 deletions(-) diff --git a/hw/i386/kvm/clock.c b/hw/i386/kvm/clock.c index efdf165..2e69cdb 100644 --- a/hw/i386/kvm/clock.c +++ b/hw/i386/kvm/clock.c @@ -17,7 +17,7 @@ #include "qemu/host-utils.h" #include "sysemu/sysemu.h" #include "sysemu/kvm.h" -#include "sysemu/cpus.h" +#include "kvm_i386.h" #include "hw/sysbus.h" #include "hw/kvm/clock.h" @@ -125,22 +125,14 @@ static void kvmclock_vm_state_change(void *opaque, int running, return; } - 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(); - + CPU_FOREACH(cpu) { + ret = kvm_get_tsc(cpu); + if (ret < 0) { + fprintf(stderr, "KVM_GET_MSRS failed: %s\n", strerror(ret)); + abort(); + return; + } + } ret = kvm_vm_ioctl(kvm_state, KVM_GET_CLOCK, &data); if (ret < 0) { fprintf(stderr, "KVM_GET_CLOCK failed: %s\n", strerror(ret)); diff --git a/target-i386/kvm.c b/target-i386/kvm.c index 64046cb..eae90e1 100644 --- a/target-i386/kvm.c +++ b/target-i386/kvm.c @@ -111,6 +111,52 @@ bool kvm_allows_irq0_override(void) return !kvm_irqchip_in_kernel() || kvm_has_gsi_routing(); } +int kvm_get_tsc(CPUState *cs) +{ + X86CPU *cpu = X86_CPU(cs); + CPUX86State *env = &cpu->env; + struct { + struct kvm_msrs info; + struct kvm_msr_entry entries[1]; + } msr_data; + struct kvm_msr_entry *msrs = msr_data.entries; + int ret, i, n; + + n = 0; + + if (!env->tsc_valid) { + msrs[n++].index = MSR_IA32_TSC; + env->tsc_valid = !runstate_is_running(); + } + + if (n == 0) { + return 0; + } + + msr_data.info = (struct kvm_msrs) { + .nmsrs = n, + }; + + ret = kvm_vcpu_ioctl(CPU(cpu), KVM_GET_MSRS, &msr_data); + if (ret < 0) { + return ret; + } + + for (i = 0; i < ret; i++) { + uint32_t index = msrs[i].index; + switch (index) { + case MSR_IA32_TSC: + env->tsc = msrs[i].data; + break; + default: + break; + } + } + + return 0; +} + + static struct kvm_cpuid2 *try_get_cpuid(KVMState *s, int max) { struct kvm_cpuid2 *cpuid; diff --git a/target-i386/kvm_i386.h b/target-i386/kvm_i386.h index e557e94..141fab5 100644 --- a/target-i386/kvm_i386.h +++ b/target-i386/kvm_i386.h @@ -17,6 +17,7 @@ bool kvm_allows_irq0_override(void); bool kvm_has_smm(void); void kvm_arch_reset_vcpu(X86CPU *cs); void kvm_arch_do_init_vcpu(X86CPU *cs); +int kvm_get_tsc(CPUState *cs); int kvm_device_pci_assign(KVMState *s, PCIHostDeviceAddress *dev_addr, uint32_t flags, uint32_t *dev_id);