[mirror_ubuntu-artful-kernel.git] / 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;
}
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;
}
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();
}

static int kvm_set_wallclock(const struct timespec *now)
{
	return -1;
}

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;
}

static u64 kvm_clock_get_cycles(struct clocksource *cs)
{
	return kvm_clock_read();
}

static u64 kvm_sched_clock_read(void)
{
	return kvm_clock_read() - kvm_sched_clock_offset;
}

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));
}

/*
 * 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;
}

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;
}

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;
}

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;
}

static void kvm_save_sched_clock_state(void)
{
}

static void kvm_restore_sched_clock_state(void)
{
	kvm_register_clock("primary cpu clock, resume");
}

#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"));
}
#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);
}
#endif

static void kvm_shutdown(void)
{
	native_write_msr(msr_kvm_system_time, 0, 0);
	kvm_disable_steal_time();
	native_machine_shutdown();
}

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";
}

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;
}
Commit	Line	Data
790c73f6 GOC	1	/* KVM paravirtual clock driver. A clocksource implementation
	2	Copyright (C) 2008 Glauber de Oliveira Costa, Red Hat Inc.
	3
	4	This program is free software; you can redistribute it and/or modify
	5	it under the terms of the GNU General Public License as published by
	6	the Free Software Foundation; either version 2 of the License, or
	7	(at your option) any later version.
	8
	9	This program is distributed in the hope that it will be useful,
	10	but WITHOUT ANY WARRANTY; without even the implied warranty of
	11	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	12	GNU General Public License for more details.
	13
	14	You should have received a copy of the GNU General Public License
	15	along with this program; if not, write to the Free Software
	16	Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
	17	*/
	18
	19	#include <linux/clocksource.h>
	20	#include <linux/kvm_para.h>
f6e16d5a	21	#include <asm/pvclock.h>
790c73f6 GOC	22	#include <asm/msr.h>
	23	#include <asm/apic.h>
	24	#include <linux/percpu.h>
3b5d56b9	25	#include <linux/hardirq.h>
7069ed67	26	#include <linux/memblock.h>
0ad83caa	27	#include <linux/sched.h>
e6017571	28	#include <linux/sched/clock.h>
736decac TG	29
736decac TG	30	#include <asm/x86_init.h>
1e977aa1	31	#include <asm/reboot.h>
f4066c2b	32	#include <asm/kvmclock.h>
790c73f6	33
404f6aac	34	static int kvmclock __ro_after_init = 1;
838815a7 GC	35	static int msr_kvm_system_time = MSR_KVM_SYSTEM_TIME;
838815a7 GC	36	static int msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK;
a5a1d1c2	37	static u64 kvm_sched_clock_offset;
790c73f6 GOC	38
	39	static int parse_no_kvmclock(char *arg)
	40	{
	41	kvmclock = 0;
	42	return 0;
	43	}
	44	early_param("no-kvmclock", parse_no_kvmclock);
	45
	46	/* The hypervisor will put information about time periodically here */
3dc4f7cf	47	static struct pvclock_vsyscall_time_info *hv_clock;
f6e16d5a	48	static struct pvclock_wall_clock wall_clock;
790c73f6	49
dac16fba AL	50	struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void)
	51	{
	52	return hv_clock;
	53	}
f4066c2b	54	EXPORT_SYMBOL_GPL(pvclock_pvti_cpu0_va);
dac16fba	55
790c73f6 GOC	56	/*
	57	* The wallclock is the time of day when we booted. Since then, some time may
	58	* have elapsed since the hypervisor wrote the data. So we try to account for
	59	* that with system time
	60	*/
3565184e	61	static void kvm_get_wallclock(struct timespec *now)
790c73f6	62	{
f6e16d5a	63	struct pvclock_vcpu_time_info *vcpu_time;
790c73f6	64	int low, high;
7069ed67	65	int cpu;
790c73f6	66
a20316d2 GC	67	low = (int)__pa_symbol(&wall_clock);
a20316d2 GC	68	high = ((u64)__pa_symbol(&wall_clock) >> 32);
838815a7 GC	69
838815a7 GC	70	native_write_msr(msr_kvm_wall_clock, low, high);
790c73f6	71
c6338ce4	72	cpu = get_cpu();
7069ed67	73
3dc4f7cf	74	vcpu_time = &hv_clock[cpu].pvti;
3565184e	75	pvclock_read_wallclock(&wall_clock, vcpu_time, now);
7069ed67	76
c6338ce4	77	put_cpu();
790c73f6 GOC	78	}
790c73f6 GOC	79
3565184e	80	static int kvm_set_wallclock(const struct timespec *now)
790c73f6	81	{
f6e16d5a	82	return -1;
790c73f6 GOC	83	}
790c73f6 GOC	84
a5a1d1c2	85	static u64 kvm_clock_read(void)
790c73f6	86	{
f6e16d5a	87	struct pvclock_vcpu_time_info *src;
a5a1d1c2	88	u64 ret;
7069ed67	89	int cpu;
790c73f6	90
95ef1e52	91	preempt_disable_notrace();
7069ed67	92	cpu = smp_processor_id();
3dc4f7cf	93	src = &hv_clock[cpu].pvti;
f6e16d5a	94	ret = pvclock_clocksource_read(src);
95ef1e52	95	preempt_enable_notrace();
f6e16d5a	96	return ret;
790c73f6	97	}
f6e16d5a	98
a5a1d1c2	99	static u64 kvm_clock_get_cycles(struct clocksource *cs)
8e19608e MD	100	{
	101	return kvm_clock_read();
	102	}
	103
a5a1d1c2	104	static u64 kvm_sched_clock_read(void)
72c930dc RK	105	{
	106	return kvm_clock_read() - kvm_sched_clock_offset;
	107	}
	108
	109	static inline void kvm_sched_clock_init(bool stable)
	110	{
	111	if (!stable) {
	112	pv_time_ops.sched_clock = kvm_clock_read;
acb04058	113	clear_sched_clock_stable();
72c930dc RK	114	return;
	115	}
	116
	117	kvm_sched_clock_offset = kvm_clock_read();
	118	pv_time_ops.sched_clock = kvm_sched_clock_read;
72c930dc RK	119
	120	printk(KERN_INFO "kvm-clock: using sched offset of %llu cycles\n",
	121	kvm_sched_clock_offset);
	122
	123	BUILD_BUG_ON(sizeof(kvm_sched_clock_offset) >
	124	sizeof(((struct pvclock_vcpu_time_info *)NULL)->system_time));
	125	}
	126
0293615f GC	127	/*
	128	* If we don't do that, there is the possibility that the guest
	129	* will calibrate under heavy load - thus, getting a lower lpj -
	130	* and execute the delays themselves without load. This is wrong,
	131	* because no delay loop can finish beforehand.
	132	* Any heuristics is subject to fail, because ultimately, a large
	133	* poll of guests can be running and trouble each other. So we preset
	134	* lpj here
	135	*/
	136	static unsigned long kvm_get_tsc_khz(void)
	137	{
e93353c9	138	struct pvclock_vcpu_time_info *src;
7069ed67 MT	139	int cpu;
	140	unsigned long tsc_khz;
	141
c6338ce4	142	cpu = get_cpu();
3dc4f7cf	143	src = &hv_clock[cpu].pvti;
7069ed67	144	tsc_khz = pvclock_tsc_khz(src);
c6338ce4	145	put_cpu();
7069ed67	146	return tsc_khz;
0293615f GC	147	}
	148
	149	static void kvm_get_preset_lpj(void)
	150	{
0293615f GC	151	unsigned long khz;
	152	u64 lpj;
	153
e93353c9	154	khz = kvm_get_tsc_khz();
0293615f GC	155
	156	lpj = ((u64)khz * 1000);
	157	do_div(lpj, HZ);
	158	preset_lpj = lpj;
	159	}
	160
3b5d56b9 EM	161	bool kvm_check_and_clear_guest_paused(void)
	162	{
	163	bool ret = false;
	164	struct pvclock_vcpu_time_info *src;
7069ed67 MT	165	int cpu = smp_processor_id();
	166
	167	if (!hv_clock)
	168	return ret;
3b5d56b9	169
3dc4f7cf	170	src = &hv_clock[cpu].pvti;
3b5d56b9	171	if ((src->flags & PVCLOCK_GUEST_STOPPED) != 0) {
7069ed67	172	src->flags &= ~PVCLOCK_GUEST_STOPPED;
d63285e9	173	pvclock_touch_watchdogs();
3b5d56b9 EM	174	ret = true;
	175	}
	176
	177	return ret;
	178	}
3b5d56b9	179
f4066c2b	180	struct clocksource kvm_clock = {
790c73f6	181	.name = "kvm-clock",
8e19608e	182	.read = kvm_clock_get_cycles,
790c73f6 GOC	183	.rating = 400,
790c73f6 GOC	184	.mask = CLOCKSOURCE_MASK(64),
790c73f6 GOC	185	.flags = CLOCK_SOURCE_IS_CONTINUOUS,
790c73f6 GOC	186	};
f4066c2b	187	EXPORT_SYMBOL_GPL(kvm_clock);
790c73f6	188
ca3f1017	189	int kvm_register_clock(char *txt)
790c73f6 GOC	190	{
790c73f6 GOC	191	int cpu = smp_processor_id();
19b6a85b	192	int low, high, ret;
fe1140cc JK	193	struct pvclock_vcpu_time_info *src;
	194
	195	if (!hv_clock)
	196	return 0;
19b6a85b	197
fe1140cc	198	src = &hv_clock[cpu].pvti;
5dfd486c DH	199	low = (int)slow_virt_to_phys(src) \| 1;
5dfd486c DH	200	high = ((u64)slow_virt_to_phys(src) >> 32);
19b6a85b	201	ret = native_write_msr_safe(msr_kvm_system_time, low, high);
f6e16d5a GH	202	printk(KERN_INFO "kvm-clock: cpu %d, msr %x:%x, %s\n",
f6e16d5a GH	203	cpu, high, low, txt);
838815a7	204
19b6a85b	205	return ret;
790c73f6 GOC	206	}
790c73f6 GOC	207
b74f05d6 MT	208	static void kvm_save_sched_clock_state(void)
	209	{
	210	}
	211
	212	static void kvm_restore_sched_clock_state(void)
	213	{
	214	kvm_register_clock("primary cpu clock, resume");
	215	}
	216
b8ba5f10	217	#ifdef CONFIG_X86_LOCAL_APIC
148f9bb8	218	static void kvm_setup_secondary_clock(void)
790c73f6 GOC	219	{
	220	/*
	221	* Now that the first cpu already had this clocksource initialized,
	222	* we shouldn't fail.
	223	*/
f6e16d5a	224	WARN_ON(kvm_register_clock("secondary cpu clock"));
790c73f6	225	}
b8ba5f10	226	#endif
790c73f6	227
1e977aa1 GC	228	/*
	229	* After the clock is registered, the host will keep writing to the
	230	* registered memory location. If the guest happens to shutdown, this memory
	231	* won't be valid. In cases like kexec, in which you install a new kernel, this
	232	* means a random memory location will be kept being written. So before any
6a6256f9	233	* kind of shutdown from our side, we unregister the clock by writing anything
1e977aa1 GC	234	* that does not have the 'enable' bit set in the msr
1e977aa1 GC	235	*/
2965faa5	236	#ifdef CONFIG_KEXEC_CORE
1e977aa1 GC	237	static void kvm_crash_shutdown(struct pt_regs *regs)
1e977aa1 GC	238	{
838815a7	239	native_write_msr(msr_kvm_system_time, 0, 0);
d910f5c1	240	kvm_disable_steal_time();
1e977aa1 GC	241	native_machine_crash_shutdown(regs);
	242	}
	243	#endif
	244
	245	static void kvm_shutdown(void)
	246	{
838815a7	247	native_write_msr(msr_kvm_system_time, 0, 0);
d910f5c1	248	kvm_disable_steal_time();
1e977aa1 GC	249	native_machine_shutdown();
	250	}
	251
790c73f6 GOC	252	void __init kvmclock_init(void)
790c73f6 GOC	253	{
0ad83caa	254	struct pvclock_vcpu_time_info *vcpu_time;
7069ed67	255	unsigned long mem;
0ad83caa LC	256	int size, cpu;
0ad83caa LC	257	u8 flags;
ed55705d MT	258
ed55705d MT	259	size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
7069ed67	260
790c73f6 GOC	261	if (!kvm_para_available())
	262	return;
	263
838815a7 GC	264	if (kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE2)) {
	265	msr_kvm_system_time = MSR_KVM_SYSTEM_TIME_NEW;
	266	msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK_NEW;
	267	} else if (!(kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE)))
	268	return;
	269
	270	printk(KERN_INFO "kvm-clock: Using msrs %x and %x",
	271	msr_kvm_system_time, msr_kvm_wall_clock);
	272
ed55705d	273	mem = memblock_alloc(size, PAGE_SIZE);
7069ed67 MT	274	if (!mem)
	275	return;
	276	hv_clock = __va(mem);
07868fc6	277	memset(hv_clock, 0, size);
7069ed67	278
0d75de4a	279	if (kvm_register_clock("primary cpu clock")) {
7069ed67	280	hv_clock = NULL;
ed55705d	281	memblock_free(mem, size);
838815a7	282	return;
7069ed67	283	}
72c930dc RK	284
	285	if (kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE_STABLE_BIT))
	286	pvclock_set_flags(PVCLOCK_TSC_STABLE_BIT);
	287
	288	cpu = get_cpu();
	289	vcpu_time = &hv_clock[cpu].pvti;
	290	flags = pvclock_read_flags(vcpu_time);
	291
	292	kvm_sched_clock_init(flags & PVCLOCK_TSC_STABLE_BIT);
	293	put_cpu();
	294
838815a7	295	x86_platform.calibrate_tsc = kvm_get_tsc_khz;
a4497a86	296	x86_platform.calibrate_cpu = kvm_get_tsc_khz;
838815a7 GC	297	x86_platform.get_wallclock = kvm_get_wallclock;
838815a7 GC	298	x86_platform.set_wallclock = kvm_set_wallclock;
b8ba5f10	299	#ifdef CONFIG_X86_LOCAL_APIC
df156f90	300	x86_cpuinit.early_percpu_clock_init =
838815a7	301	kvm_setup_secondary_clock;
b8ba5f10	302	#endif
b74f05d6 MT	303	x86_platform.save_sched_clock_state = kvm_save_sched_clock_state;
b74f05d6 MT	304	x86_platform.restore_sched_clock_state = kvm_restore_sched_clock_state;
838815a7	305	machine_ops.shutdown = kvm_shutdown;
2965faa5	306	#ifdef CONFIG_KEXEC_CORE
838815a7	307	machine_ops.crash_shutdown = kvm_crash_shutdown;
1e977aa1	308	#endif
838815a7	309	kvm_get_preset_lpj();
b01cc1b0	310	clocksource_register_hz(&kvm_clock, NSEC_PER_SEC);
838815a7	311	pv_info.name = "KVM";
790c73f6	312	}
3dc4f7cf MT	313
	314	int __init kvm_setup_vsyscall_timeinfo(void)
	315	{
	316	#ifdef CONFIG_X86_64
	317	int cpu;
3dc4f7cf MT	318	u8 flags;
	319	struct pvclock_vcpu_time_info *vcpu_time;
	320	unsigned int size;
	321
fe1140cc JK	322	if (!hv_clock)
	323	return 0;
	324
ed55705d	325	size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
3dc4f7cf	326
c6338ce4	327	cpu = get_cpu();
3dc4f7cf MT	328
	329	vcpu_time = &hv_clock[cpu].pvti;
	330	flags = pvclock_read_flags(vcpu_time);
	331
	332	if (!(flags & PVCLOCK_TSC_STABLE_BIT)) {
c6338ce4	333	put_cpu();
3dc4f7cf MT	334	return 1;
	335	}
	336
c6338ce4	337	put_cpu();
3dc4f7cf MT	338
	339	kvm_clock.archdata.vclock_mode = VCLOCK_PVCLOCK;
	340	#endif
	341	return 0;
	342	}