Release 4.14 arch/x86/kernel/kvmclock.c
/* KVM paravirtual clock driver. A clocksource implementation
Copyright (C) 2008 Glauber de Oliveira Costa, Red Hat Inc.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <linux/clocksource.h>
#include <linux/kvm_para.h>
#include <asm/pvclock.h>
#include <asm/msr.h>
#include <asm/apic.h>
#include <linux/percpu.h>
#include <linux/hardirq.h>
#include <linux/memblock.h>
#include <linux/sched.h>
#include <linux/sched/clock.h>
#include <asm/x86_init.h>
#include <asm/reboot.h>
#include <asm/kvmclock.h>
static int kvmclock __ro_after_init = 1;
static int msr_kvm_system_time = MSR_KVM_SYSTEM_TIME;
static int msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK;
static u64 kvm_sched_clock_offset;
static int parse_no_kvmclock(char *arg)
{
kvmclock = 0;
return 0;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 17 | 100.00% | 1 | 100.00% |
Total | 17 | 100.00% | 1 | 100.00% |
early_param("no-kvmclock", parse_no_kvmclock);
/* The hypervisor will put information about time periodically here */
static struct pvclock_vsyscall_time_info *hv_clock;
static struct pvclock_wall_clock wall_clock;
struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void)
{
return hv_clock;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Andrew Lutomirski | 12 | 100.00% | 1 | 100.00% |
Total | 12 | 100.00% | 1 | 100.00% |
EXPORT_SYMBOL_GPL(pvclock_pvti_cpu0_va);
/*
* The wallclock is the time of day when we booted. Since then, some time may
* have elapsed since the hypervisor wrote the data. So we try to account for
* that with system time
*/
static void kvm_get_wallclock(struct timespec *now)
{
struct pvclock_vcpu_time_info *vcpu_time;
int low, high;
int cpu;
low = (int)__pa_symbol(&wall_clock);
high = ((u64)__pa_symbol(&wall_clock) >> 32);
native_write_msr(msr_kvm_wall_clock, low, high);
cpu = get_cpu();
vcpu_time = &hv_clock[cpu].pvti;
pvclock_read_wallclock(&wall_clock, vcpu_time, now);
put_cpu();
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 54 | 62.07% | 3 | 33.33% |
Marcelo Tosatti | 13 | 14.94% | 2 | 22.22% |
Gerd Hoffmann | 11 | 12.64% | 1 | 11.11% |
David Vrabel | 6 | 6.90% | 1 | 11.11% |
Tiejun Chen | 2 | 2.30% | 1 | 11.11% |
Ingo Molnar | 1 | 1.15% | 1 | 11.11% |
Total | 87 | 100.00% | 9 | 100.00% |
static int kvm_set_wallclock(const struct timespec *now)
{
return -ENODEV;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 9 | 56.25% | 1 | 20.00% |
David Vrabel | 4 | 25.00% | 1 | 20.00% |
Ingo Molnar | 1 | 6.25% | 1 | 20.00% |
Jason Gunthorpe | 1 | 6.25% | 1 | 20.00% |
Gerd Hoffmann | 1 | 6.25% | 1 | 20.00% |
Total | 16 | 100.00% | 5 | 100.00% |
static u64 kvm_clock_read(void)
{
struct pvclock_vcpu_time_info *src;
u64 ret;
int cpu;
preempt_disable_notrace();
cpu = smp_processor_id();
src = &hv_clock[cpu].pvti;
ret = pvclock_clocksource_read(src);
preempt_enable_notrace();
return ret;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 18 | 36.00% | 1 | 16.67% |
Marcelo Tosatti | 13 | 26.00% | 2 | 33.33% |
Gerd Hoffmann | 12 | 24.00% | 1 | 16.67% |
Avi Kivity | 5 | 10.00% | 1 | 16.67% |
Thomas Gleixner | 2 | 4.00% | 1 | 16.67% |
Total | 50 | 100.00% | 6 | 100.00% |
static u64 kvm_clock_get_cycles(struct clocksource *cs)
{
return kvm_clock_read();
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Magnus Damm | 14 | 93.33% | 1 | 50.00% |
Thomas Gleixner | 1 | 6.67% | 1 | 50.00% |
Total | 15 | 100.00% | 2 | 100.00% |
static u64 kvm_sched_clock_read(void)
{
return kvm_clock_read() - kvm_sched_clock_offset;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Radim Krčmář | 13 | 92.86% | 1 | 50.00% |
Thomas Gleixner | 1 | 7.14% | 1 | 50.00% |
Total | 14 | 100.00% | 2 | 100.00% |
static inline void kvm_sched_clock_init(bool stable)
{
if (!stable) {
pv_time_ops.sched_clock = kvm_clock_read;
clear_sched_clock_stable();
return;
}
kvm_sched_clock_offset = kvm_clock_read();
pv_time_ops.sched_clock = kvm_sched_clock_read;
printk(KERN_INFO "kvm-clock: using sched offset of %llu cycles\n",
kvm_sched_clock_offset);
BUILD_BUG_ON(sizeof(kvm_sched_clock_offset) >
sizeof(((struct pvclock_vcpu_time_info *)NULL)->system_time));
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Radim Krčmář | 65 | 95.59% | 1 | 50.00% |
Peter Zijlstra | 3 | 4.41% | 1 | 50.00% |
Total | 68 | 100.00% | 2 | 100.00% |
/*
* If we don't do that, there is the possibility that the guest
* will calibrate under heavy load - thus, getting a lower lpj -
* and execute the delays themselves without load. This is wrong,
* because no delay loop can finish beforehand.
* Any heuristics is subject to fail, because ultimately, a large
* poll of guests can be running and trouble each other. So we preset
* lpj here
*/
static unsigned long kvm_get_tsc_khz(void)
{
struct pvclock_vcpu_time_info *src;
int cpu;
unsigned long tsc_khz;
cpu = get_cpu();
src = &hv_clock[cpu].pvti;
tsc_khz = pvclock_tsc_khz(src);
put_cpu();
return tsc_khz;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Marcelo Tosatti | 23 | 46.94% | 2 | 40.00% |
Eduardo Pereira Habkost | 14 | 28.57% | 1 | 20.00% |
Glauber de Oliveira Costa | 10 | 20.41% | 1 | 20.00% |
Tiejun Chen | 2 | 4.08% | 1 | 20.00% |
Total | 49 | 100.00% | 5 | 100.00% |
static void kvm_get_preset_lpj(void)
{
unsigned long khz;
u64 lpj;
khz = kvm_get_tsc_khz();
lpj = ((u64)khz * 1000);
do_div(lpj, HZ);
preset_lpj = lpj;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 40 | 95.24% | 1 | 50.00% |
Eduardo Pereira Habkost | 2 | 4.76% | 1 | 50.00% |
Total | 42 | 100.00% | 2 | 100.00% |
bool kvm_check_and_clear_guest_paused(void)
{
bool ret = false;
struct pvclock_vcpu_time_info *src;
int cpu = smp_processor_id();
if (!hv_clock)
return ret;
src = &hv_clock[cpu].pvti;
if ((src->flags & PVCLOCK_GUEST_STOPPED) != 0) {
src->flags &= ~PVCLOCK_GUEST_STOPPED;
pvclock_touch_watchdogs();
ret = true;
}
return ret;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Eric B Munson | 47 | 65.28% | 1 | 25.00% |
Marcelo Tosatti | 25 | 34.72% | 3 | 75.00% |
Total | 72 | 100.00% | 4 | 100.00% |
struct clocksource kvm_clock = {
.name = "kvm-clock",
.read = kvm_clock_get_cycles,
.rating = 400,
.mask = CLOCKSOURCE_MASK(64),
.flags = CLOCK_SOURCE_IS_CONTINUOUS,
};
EXPORT_SYMBOL_GPL(kvm_clock);
int kvm_register_clock(char *txt)
{
int cpu = smp_processor_id();
int low, high, ret;
struct pvclock_vcpu_time_info *src;
if (!hv_clock)
return 0;
src = &hv_clock[cpu].pvti;
low = (int)slow_virt_to_phys(src) | 1;
high = ((u64)slow_virt_to_phys(src) >> 32);
ret = native_write_msr_safe(msr_kvm_system_time, low, high);
printk(KERN_INFO "kvm-clock: cpu %d, msr %x:%x, %s\n",
cpu, high, low, txt);
return ret;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 41 | 41.41% | 1 | 14.29% |
Gerd Hoffmann | 17 | 17.17% | 1 | 14.29% |
Arjan Koers | 14 | 14.14% | 1 | 14.29% |
Marcelo Tosatti | 13 | 13.13% | 2 | 28.57% |
Jan Kiszka | 12 | 12.12% | 1 | 14.29% |
Dave Hansen | 2 | 2.02% | 1 | 14.29% |
Total | 99 | 100.00% | 7 | 100.00% |
static void kvm_save_sched_clock_state(void)
{
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Marcelo Tosatti | 7 | 100.00% | 1 | 100.00% |
Total | 7 | 100.00% | 1 | 100.00% |
static void kvm_restore_sched_clock_state(void)
{
kvm_register_clock("primary cpu clock, resume");
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Marcelo Tosatti | 13 | 100.00% | 1 | 100.00% |
Total | 13 | 100.00% | 1 | 100.00% |
#ifdef CONFIG_X86_LOCAL_APIC
static void kvm_setup_secondary_clock(void)
{
/*
* Now that the first cpu already had this clocksource initialized,
* we shouldn't fail.
*/
WARN_ON(kvm_register_clock("secondary cpu clock"));
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 14 | 82.35% | 1 | 50.00% |
Gerd Hoffmann | 3 | 17.65% | 1 | 50.00% |
Total | 17 | 100.00% | 2 | 100.00% |
#endif
/*
* After the clock is registered, the host will keep writing to the
* registered memory location. If the guest happens to shutdown, this memory
* won't be valid. In cases like kexec, in which you install a new kernel, this
* means a random memory location will be kept being written. So before any
* kind of shutdown from our side, we unregister the clock by writing anything
* that does not have the 'enable' bit set in the msr
*/
#ifdef CONFIG_KEXEC_CORE
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);
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 28 | 100.00% | 3 | 100.00% |
Total | 28 | 100.00% | 3 | 100.00% |
#endif
static void kvm_shutdown(void)
{
native_write_msr(msr_kvm_system_time, 0, 0);
kvm_disable_steal_time();
native_machine_shutdown();
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 23 | 100.00% | 3 | 100.00% |
Total | 23 | 100.00% | 3 | 100.00% |
void __init kvmclock_init(void)
{
struct pvclock_vcpu_time_info *vcpu_time;
unsigned long mem;
int size, cpu;
u8 flags;
size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
if (!kvm_para_available())
return;
if (kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE2)) {
msr_kvm_system_time = MSR_KVM_SYSTEM_TIME_NEW;
msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK_NEW;
} else if (!(kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE)))
return;
printk(KERN_INFO "kvm-clock: Using msrs %x and %x",
msr_kvm_system_time, msr_kvm_wall_clock);
mem = memblock_alloc(size, PAGE_SIZE);
if (!mem)
return;
hv_clock = __va(mem);
memset(hv_clock, 0, size);
if (kvm_register_clock("primary cpu clock")) {
hv_clock = NULL;
memblock_free(mem, size);
return;
}
if (kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE_STABLE_BIT))
pvclock_set_flags(PVCLOCK_TSC_STABLE_BIT);
cpu = get_cpu();
vcpu_time = &hv_clock[cpu].pvti;
flags = pvclock_read_flags(vcpu_time);
kvm_sched_clock_init(flags & PVCLOCK_TSC_STABLE_BIT);
put_cpu();
x86_platform.calibrate_tsc = kvm_get_tsc_khz;
x86_platform.calibrate_cpu = kvm_get_tsc_khz;
x86_platform.get_wallclock = kvm_get_wallclock;
x86_platform.set_wallclock = kvm_set_wallclock;
#ifdef CONFIG_X86_LOCAL_APIC
x86_cpuinit.early_percpu_clock_init =
kvm_setup_secondary_clock;
#endif
x86_platform.save_sched_clock_state = kvm_save_sched_clock_state;
x86_platform.restore_sched_clock_state = kvm_restore_sched_clock_state;
machine_ops.shutdown = kvm_shutdown;
#ifdef CONFIG_KEXEC_CORE
machine_ops.crash_shutdown = kvm_crash_shutdown;
#endif
kvm_get_preset_lpj();
clocksource_register_hz(&kvm_clock, NSEC_PER_SEC);
pv_info.name = "KVM";
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 114 | 43.02% | 6 | 28.57% |
Marcelo Tosatti | 67 | 25.28% | 3 | 14.29% |
Radim Krčmář | 41 | 15.47% | 1 | 4.76% |
Igor Mammedov | 10 | 3.77% | 2 | 9.52% |
Luiz Fernando N. Capitulino | 10 | 3.77% | 1 | 4.76% |
Feng Tang | 8 | 3.02% | 1 | 4.76% |
Prarit Bhargava | 6 | 2.26% | 1 | 4.76% |
John Stultz | 3 | 1.13% | 1 | 4.76% |
Gerd Hoffmann | 2 | 0.75% | 1 | 4.76% |
Thomas Gleixner | 2 | 0.75% | 2 | 9.52% |
Fernando Luis Vázquez Cao | 1 | 0.38% | 1 | 4.76% |
Dave Young | 1 | 0.38% | 1 | 4.76% |
Total | 265 | 100.00% | 21 | 100.00% |
int __init kvm_setup_vsyscall_timeinfo(void)
{
#ifdef CONFIG_X86_64
int cpu;
u8 flags;
struct pvclock_vcpu_time_info *vcpu_time;
unsigned int size;
if (!hv_clock)
return 0;
size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
cpu = get_cpu();
vcpu_time = &hv_clock[cpu].pvti;
flags = pvclock_read_flags(vcpu_time);
if (!(flags & PVCLOCK_TSC_STABLE_BIT)) {
put_cpu();
return 1;
}
put_cpu();
kvm_clock.archdata.vclock_mode = VCLOCK_PVCLOCK;
#endif
return 0;
}
Contributors
Person | Tokens | Prop | Commits | CommitProp |
Marcelo Tosatti | 91 | 89.22% | 2 | 50.00% |
Jan Kiszka | 8 | 7.84% | 1 | 25.00% |
Tiejun Chen | 3 | 2.94% | 1 | 25.00% |
Total | 102 | 100.00% | 4 | 100.00% |
Overall Contributors
Person | Tokens | Prop | Commits | CommitProp |
Glauber de Oliveira Costa | 461 | 40.62% | 8 | 18.60% |
Marcelo Tosatti | 287 | 25.29% | 6 | 13.95% |
Radim Krčmář | 122 | 10.75% | 1 | 2.33% |
Gerd Hoffmann | 50 | 4.41% | 1 | 2.33% |
Eric B Munson | 50 | 4.41% | 1 | 2.33% |
Jan Kiszka | 20 | 1.76% | 1 | 2.33% |
Eduardo Pereira Habkost | 16 | 1.41% | 1 | 2.33% |
Magnus Damm | 15 | 1.32% | 1 | 2.33% |
Arjan Koers | 14 | 1.23% | 1 | 2.33% |
Luiz Fernando N. Capitulino | 13 | 1.15% | 1 | 2.33% |
Andrew Lutomirski | 12 | 1.06% | 1 | 2.33% |
Thomas Gleixner | 10 | 0.88% | 3 | 6.98% |
Igor Mammedov | 10 | 0.88% | 2 | 4.65% |
David Vrabel | 10 | 0.88% | 1 | 2.33% |
Feng Tang | 8 | 0.70% | 1 | 2.33% |
Tiejun Chen | 7 | 0.62% | 1 | 2.33% |
Prarit Bhargava | 6 | 0.53% | 1 | 2.33% |
Avi Kivity | 5 | 0.44% | 1 | 2.33% |
Ingo Molnar | 5 | 0.44% | 2 | 4.65% |
Peter Zijlstra | 3 | 0.26% | 1 | 2.33% |
John Stultz | 3 | 0.26% | 1 | 2.33% |
Dave Young | 2 | 0.18% | 1 | 2.33% |
Dave Hansen | 2 | 0.18% | 1 | 2.33% |
Kees Cook | 1 | 0.09% | 1 | 2.33% |
Jason Gunthorpe | 1 | 0.09% | 1 | 2.33% |
Fernando Luis Vázquez Cao | 1 | 0.09% | 1 | 2.33% |
Adam Buchbinder | 1 | 0.09% | 1 | 2.33% |
Total | 1135 | 100.00% | 43 | 100.00% |
Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.