qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Liang Li <liang.z.li@intel.com>
To: qemu-devel@nongnu.org
Cc: Liang Li <liang.z.li@intel.com>,
	ehabkost@redhat.com, mst@redhat.com, mtosatti@redhat.com,
	quintela@redhat.com, m.gibula@beyond.pl, amit.shah@redhat.com,
	pbonzini@redhat.com, rth@twiddle.net
Subject: [Qemu-devel] [v2 1/2] kvmclock: add a new function to update env->tsc.
Date: Thu,  5 Nov 2015 11:51:03 +0800	[thread overview]
Message-ID: <1446695464-27116-2-git-send-email-liang.z.li@intel.com> (raw)
In-Reply-To: <1446695464-27116-1-git-send-email-liang.z.li@intel.com>

The commit 317b0a6d8 fixed an issue which caused by the outdated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.

Synchronize the whole CPU context just for updating env->tsc is too
wasting, this patch use a new function to update the env->tsc.
Comparing to 'cpu_synchronize_all_states()', it only takes about 20us.

Signed-off-by: Liang Li <liang.z.li@intel.com>
---
 hw/i386/kvm/clock.c    | 18 ++----------------
 target-i386/kvm.c      | 45 +++++++++++++++++++++++++++++++++++++++++++++
 target-i386/kvm_i386.h |  1 +
 3 files changed, 48 insertions(+), 16 deletions(-)

diff --git a/hw/i386/kvm/clock.c b/hw/i386/kvm/clock.c
index efdf165..0593a3f 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"
 
@@ -125,21 +125,7 @@ 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-         * outdated APIC state may be migrated to another host.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-         * The real fix would be to make sure outdated APIC state is read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();
+        kvm_synchronize_all_tsc();
 
         ret = kvm_vm_ioctl(kvm_state, KVM_GET_CLOCK, &data);
         if (ret < 0) {
diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index 64046cb..2a9953b 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -111,6 +111,51 @@ bool kvm_allows_irq0_override(void)
     return !kvm_irqchip_in_kernel() || kvm_has_gsi_routing();
 }
 
+static 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;
+    int ret;
+
+    if (env->tsc_valid) {
+        return 0;
+    }
+
+    msr_data.info.nmsrs = 1;
+    msr_data.entries[0].index = MSR_IA32_TSC;
+    env->tsc_valid = !runstate_is_running();
+
+    ret = kvm_vcpu_ioctl(CPU(cpu), KVM_GET_MSRS, &msr_data);
+    if (ret < 0) {
+        return ret;
+    }
+
+    env->tsc = msr_data.entries[0].data;
+    return 0;
+}
+
+static inline void do_kvm_synchronize_tsc(void *arg)
+{
+    CPUState *cpu = arg;
+
+    kvm_get_tsc(cpu);
+}
+
+void kvm_synchronize_all_tsc(void)
+{
+    CPUState *cpu;
+
+    if (kvm_enabled()) {
+        CPU_FOREACH(cpu) {
+            run_on_cpu(cpu, do_kvm_synchronize_tsc, cpu);
+        }
+    }
+}
+
 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..c1b312b 100644
--- a/target-i386/kvm_i386.h
+++ b/target-i386/kvm_i386.h
@@ -15,6 +15,7 @@
 
 bool kvm_allows_irq0_override(void);
 bool kvm_has_smm(void);
+void kvm_synchronize_all_tsc(void);
 void kvm_arch_reset_vcpu(X86CPU *cs);
 void kvm_arch_do_init_vcpu(X86CPU *cs);
 
-- 
1.9.1

  reply	other threads:[~2015-11-05  3:56 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-05  3:51 [Qemu-devel] [v2 0/2] Optimize the env->tsc update operation Liang Li
2015-11-05  3:51 ` Liang Li [this message]
2015-11-05  3:51 ` [Qemu-devel] [v2 2/2] Revert "Introduce cpu_clean_all_dirty" Liang Li
2015-11-05 10:28 ` [Qemu-devel] [v2 0/2] Optimize the env->tsc update operation Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1446695464-27116-2-git-send-email-liang.z.li@intel.com \
    --to=liang.z.li@intel.com \
    --cc=amit.shah@redhat.com \
    --cc=ehabkost@redhat.com \
    --cc=m.gibula@beyond.pl \
    --cc=mst@redhat.com \
    --cc=mtosatti@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=rth@twiddle.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).