From patchwork Mon Jul 4 15:32:28 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Glauber Costa X-Patchwork-Id: 942432 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by demeter2.kernel.org (8.14.4/8.14.4) with ESMTP id p64FmMTe000394 for ; Mon, 4 Jul 2011 15:48:23 GMT Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758078Ab1GDPnK (ORCPT ); Mon, 4 Jul 2011 11:43:10 -0400 Received: from mx1.redhat.com ([209.132.183.28]:30096 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757795Ab1GDPk5 (ORCPT ); Mon, 4 Jul 2011 11:40:57 -0400 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id p64FemBU031734 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 4 Jul 2011 11:40:49 -0400 Received: from virtlab1.virt.bos.redhat.com (virtlab1.virt.bos.redhat.com [10.16.72.21]) by int-mx09.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id p64FedRT032414; Mon, 4 Jul 2011 11:40:48 -0400 From: Glauber Costa To: kvm@vger.kernel.org Cc: linux-kernel@vger.kernel.org, Rik van Riel , Jeremy Fitzhardinge , Peter Zijlstra , Avi Kivity , Anthony Liguori , Eric B Munson Subject: [PATCH v5 9/9] KVM-GST: KVM Steal time registration Date: Mon, 4 Jul 2011 11:32:28 -0400 Message-Id: <1309793548-16714-10-git-send-email-glommer@redhat.com> In-Reply-To: <1309793548-16714-1-git-send-email-glommer@redhat.com> References: <1309793548-16714-1-git-send-email-glommer@redhat.com> X-Scanned-By: MIMEDefang 2.68 on 10.5.11.22 Sender: kvm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org X-Greylist: IP, sender and recipient auto-whitelisted, not delayed by milter-greylist-4.2.6 (demeter2.kernel.org [140.211.167.43]); Mon, 04 Jul 2011 15:48:23 +0000 (UTC) This patch implements the kvm bits of the steal time infrastructure. The most important part of it, is the steal time clock. It is an continuous clock that shows the accumulated amount of steal time since vcpu creation. It is supposed to survive cpu offlining/onlining. Signed-off-by: Glauber Costa CC: Rik van Riel CC: Jeremy Fitzhardinge CC: Peter Zijlstra CC: Avi Kivity CC: Anthony Liguori CC: Eric B Munson Tested-by: Eric B Munson Acked-by: Rik van Riel --- Documentation/kernel-parameters.txt | 4 ++ arch/x86/include/asm/kvm_para.h | 1 + arch/x86/kernel/kvm.c | 73 +++++++++++++++++++++++++++++++++++ arch/x86/kernel/kvmclock.c | 2 + 4 files changed, 80 insertions(+), 0 deletions(-) diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index fd248a31..a722574 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1737,6 +1737,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted. no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page fault handling. + no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting. + steal time is computed, but won't influence scheduler + behaviour + nolapic [X86-32,APIC] Do not enable or use the local APIC. nolapic_timer [X86-32,APIC] Do not use the local APIC timer. diff --git a/arch/x86/include/asm/kvm_para.h b/arch/x86/include/asm/kvm_para.h index c484ba8..35d732d 100644 --- a/arch/x86/include/asm/kvm_para.h +++ b/arch/x86/include/asm/kvm_para.h @@ -94,6 +94,7 @@ struct kvm_vcpu_pv_apf_data { extern void kvmclock_init(void); extern int kvm_register_clock(char *txt); +extern void kvm_disable_steal_time(void); /* This instruction is vmcall. On non-VT architectures, it will generate a diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c index 33c07b0..58331c2 100644 --- a/arch/x86/kernel/kvm.c +++ b/arch/x86/kernel/kvm.c @@ -51,6 +51,15 @@ static int parse_no_kvmapf(char *arg) early_param("no-kvmapf", parse_no_kvmapf); +static int steal_acc = 1; +static int parse_no_stealacc(char *arg) +{ + steal_acc = 0; + return 0; +} + +early_param("no-steal-acc", parse_no_stealacc); + struct kvm_para_state { u8 mmu_queue[MMU_QUEUE_SIZE]; int mmu_queue_len; @@ -58,6 +67,8 @@ struct kvm_para_state { static DEFINE_PER_CPU(struct kvm_para_state, para_state); static DEFINE_PER_CPU(struct kvm_vcpu_pv_apf_data, apf_reason) __aligned(64); +static DEFINE_PER_CPU(struct kvm_steal_time, steal_time) __aligned(64); +static int has_steal_clock = 0; static struct kvm_para_state *kvm_para_state(void) { @@ -441,6 +452,21 @@ static void __init paravirt_ops_setup(void) #endif } +static void kvm_register_steal_time(void) +{ + int cpu = smp_processor_id(); + struct kvm_steal_time *st = &per_cpu(steal_time, cpu); + + if (!has_steal_clock) + return; + + memset(st, 0, sizeof(*st)); + + wrmsrl(MSR_KVM_STEAL_TIME, (__pa(st) | KVM_MSR_ENABLED)); + printk(KERN_INFO "kvm-stealtime: cpu %d, msr %lx\n", + cpu, __pa(st)); +} + void __cpuinit kvm_guest_cpu_init(void) { if (!kvm_para_available()) @@ -457,6 +483,9 @@ void __cpuinit kvm_guest_cpu_init(void) printk(KERN_INFO"KVM setup async PF for cpu %d\n", smp_processor_id()); } + + if (has_steal_clock) + kvm_register_steal_time(); } static void kvm_pv_disable_apf(void *unused) @@ -483,6 +512,31 @@ static struct notifier_block kvm_pv_reboot_nb = { .notifier_call = kvm_pv_reboot_notify, }; +static u64 kvm_steal_clock(int cpu) +{ + u64 steal; + struct kvm_steal_time *src; + int version; + + src = &per_cpu(steal_time, cpu); + do { + version = src->version; + rmb(); + steal = src->steal; + rmb(); + } while ((version & 1) || (version != src->version)); + + return steal; +} + +void kvm_disable_steal_time(void) +{ + if (!has_steal_clock) + return; + + wrmsr(MSR_KVM_STEAL_TIME, 0, 0); +} + #ifdef CONFIG_SMP static void __init kvm_smp_prepare_boot_cpu(void) { @@ -500,6 +554,7 @@ static void __cpuinit kvm_guest_cpu_online(void *dummy) static void kvm_guest_cpu_offline(void *dummy) { + kvm_disable_steal_time(); kvm_pv_disable_apf(NULL); apf_task_wake_all(); } @@ -548,6 +603,11 @@ void __init kvm_guest_init(void) if (kvm_para_has_feature(KVM_FEATURE_ASYNC_PF)) x86_init.irqs.trap_init = kvm_apf_trap_init; + if (kvm_para_has_feature(KVM_FEATURE_STEAL_TIME)) { + has_steal_clock = 1; + pv_time_ops.steal_clock = kvm_steal_clock; + } + #ifdef CONFIG_SMP smp_ops.smp_prepare_boot_cpu = kvm_smp_prepare_boot_cpu; register_cpu_notifier(&kvm_cpu_notifier); @@ -555,3 +615,16 @@ void __init kvm_guest_init(void) kvm_guest_cpu_init(); #endif } + +static __init int activate_jump_labels(void) +{ + if (has_steal_clock) { + jump_label_inc(¶virt_steal_enabled); + if (steal_acc) + jump_label_inc(¶virt_steal_rq_enabled); + } + + return 0; +} +arch_initcall(activate_jump_labels); + diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index 6389a6b..c1a0188 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -160,6 +160,7 @@ static void __cpuinit kvm_setup_secondary_clock(void) static void kvm_crash_shutdown(struct pt_regs *regs) { native_write_msr(msr_kvm_system_time, 0, 0); + kvm_disable_steal_time(); native_machine_crash_shutdown(regs); } #endif @@ -167,6 +168,7 @@ static void kvm_crash_shutdown(struct pt_regs *regs) static void kvm_shutdown(void) { native_write_msr(msr_kvm_system_time, 0, 0); + kvm_disable_steal_time(); native_machine_shutdown(); }