[mirror_ubuntu-bionic-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/mem_encrypt.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;

/*
 * 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)slow_virt_to_phys(wall_clock);
	high = ((u64)slow_virt_to_phys(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 -ENODEV;
}

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();
}

static phys_addr_t __init kvm_memblock_alloc(phys_addr_t size,
					     phys_addr_t align)
{
	phys_addr_t mem;

	mem = memblock_alloc(size, align);
	if (!mem)
		return 0;

	if (sev_active()) {
		if (early_set_memory_decrypted((unsigned long)__va(mem), size))
			goto e_free;
	}

	return mem;
e_free:
	memblock_free(mem, size);
	return 0;
}

static void __init kvm_memblock_free(phys_addr_t addr, phys_addr_t size)
{
	if (sev_active())
		early_set_memory_encrypted((unsigned long)__va(addr), size);

	memblock_free(addr, size);
}

void __init kvmclock_init(void)
{
	struct pvclock_vcpu_time_info *vcpu_time;
	unsigned long mem, mem_wall_clock;
	int size, cpu, wall_clock_size;
	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;

	wall_clock_size = PAGE_ALIGN(sizeof(struct pvclock_wall_clock));
	mem_wall_clock = kvm_memblock_alloc(wall_clock_size, PAGE_SIZE);
	if (!mem_wall_clock)
		return;

	wall_clock = __va(mem_wall_clock);
	memset(wall_clock, 0, wall_clock_size);

	mem = kvm_memblock_alloc(size, PAGE_SIZE);
	if (!mem) {
		kvm_memblock_free(mem_wall_clock, wall_clock_size);
		wall_clock = NULL;
		return;
	}

	hv_clock = __va(mem);
	memset(hv_clock, 0, size);

	if (kvm_register_clock("primary cpu clock")) {
		hv_clock = NULL;
		kvm_memblock_free(mem, size);
		kvm_memblock_free(mem_wall_clock, wall_clock_size);
		wall_clock = NULL;
		return;
	}

	printk(KERN_INFO "kvm-clock: Using msrs %x and %x",
		msr_kvm_system_time, msr_kvm_wall_clock);

	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;
	}

	pvclock_set_pvti_cpu0_va(hv_clock);
	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	29
819aeee0	30	#include <asm/mem_encrypt.h>
736decac	31	#include <asm/x86_init.h>
1e977aa1	32	#include <asm/reboot.h>
f4066c2b	33	#include <asm/kvmclock.h>
790c73f6	34
404f6aac	35	static int kvmclock __ro_after_init = 1;
838815a7 GC	36	static int msr_kvm_system_time = MSR_KVM_SYSTEM_TIME;
838815a7 GC	37	static int msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK;
a5a1d1c2	38	static u64 kvm_sched_clock_offset;
790c73f6 GOC	39
	40	static int parse_no_kvmclock(char *arg)
	41	{
	42	kvmclock = 0;
	43	return 0;
	44	}
	45	early_param("no-kvmclock", parse_no_kvmclock);
	46
	47	/* The hypervisor will put information about time periodically here */
3dc4f7cf	48	static struct pvclock_vsyscall_time_info *hv_clock;
819aeee0	49	static struct pvclock_wall_clock *wall_clock;
790c73f6	50
790c73f6 GOC	51	/*
	52	* The wallclock is the time of day when we booted. Since then, some time may
	53	* have elapsed since the hypervisor wrote the data. So we try to account for
	54	* that with system time
	55	*/
3565184e	56	static void kvm_get_wallclock(struct timespec *now)
790c73f6	57	{
f6e16d5a	58	struct pvclock_vcpu_time_info *vcpu_time;
790c73f6	59	int low, high;
7069ed67	60	int cpu;
790c73f6	61
819aeee0 BS	62	low = (int)slow_virt_to_phys(wall_clock);
819aeee0 BS	63	high = ((u64)slow_virt_to_phys(wall_clock) >> 32);
838815a7 GC	64
838815a7 GC	65	native_write_msr(msr_kvm_wall_clock, low, high);
790c73f6	66
c6338ce4	67	cpu = get_cpu();
7069ed67	68
3dc4f7cf	69	vcpu_time = &hv_clock[cpu].pvti;
819aeee0	70	pvclock_read_wallclock(wall_clock, vcpu_time, now);
7069ed67	71
c6338ce4	72	put_cpu();
790c73f6 GOC	73	}
790c73f6 GOC	74
3565184e	75	static int kvm_set_wallclock(const struct timespec *now)
790c73f6	76	{
00875520	77	return -ENODEV;
790c73f6 GOC	78	}
790c73f6 GOC	79
a5a1d1c2	80	static u64 kvm_clock_read(void)
790c73f6	81	{
f6e16d5a	82	struct pvclock_vcpu_time_info *src;
a5a1d1c2	83	u64 ret;
7069ed67	84	int cpu;
790c73f6	85
95ef1e52	86	preempt_disable_notrace();
7069ed67	87	cpu = smp_processor_id();
3dc4f7cf	88	src = &hv_clock[cpu].pvti;
f6e16d5a	89	ret = pvclock_clocksource_read(src);
95ef1e52	90	preempt_enable_notrace();
f6e16d5a	91	return ret;
790c73f6	92	}
f6e16d5a	93
a5a1d1c2	94	static u64 kvm_clock_get_cycles(struct clocksource *cs)
8e19608e MD	95	{
	96	return kvm_clock_read();
	97	}
	98
a5a1d1c2	99	static u64 kvm_sched_clock_read(void)
72c930dc RK	100	{
	101	return kvm_clock_read() - kvm_sched_clock_offset;
	102	}
	103
	104	static inline void kvm_sched_clock_init(bool stable)
	105	{
	106	if (!stable) {
	107	pv_time_ops.sched_clock = kvm_clock_read;
acb04058	108	clear_sched_clock_stable();
72c930dc RK	109	return;
	110	}
	111
	112	kvm_sched_clock_offset = kvm_clock_read();
	113	pv_time_ops.sched_clock = kvm_sched_clock_read;
72c930dc RK	114
	115	printk(KERN_INFO "kvm-clock: using sched offset of %llu cycles\n",
	116	kvm_sched_clock_offset);
	117
	118	BUILD_BUG_ON(sizeof(kvm_sched_clock_offset) >
	119	sizeof(((struct pvclock_vcpu_time_info *)NULL)->system_time));
	120	}
	121
0293615f GC	122	/*
	123	* If we don't do that, there is the possibility that the guest
	124	* will calibrate under heavy load - thus, getting a lower lpj -
	125	* and execute the delays themselves without load. This is wrong,
	126	* because no delay loop can finish beforehand.
	127	* Any heuristics is subject to fail, because ultimately, a large
	128	* poll of guests can be running and trouble each other. So we preset
	129	* lpj here
	130	*/
	131	static unsigned long kvm_get_tsc_khz(void)
	132	{
e93353c9	133	struct pvclock_vcpu_time_info *src;
7069ed67 MT	134	int cpu;
	135	unsigned long tsc_khz;
	136
c6338ce4	137	cpu = get_cpu();
3dc4f7cf	138	src = &hv_clock[cpu].pvti;
7069ed67	139	tsc_khz = pvclock_tsc_khz(src);
c6338ce4	140	put_cpu();
7069ed67	141	return tsc_khz;
0293615f GC	142	}
	143
	144	static void kvm_get_preset_lpj(void)
	145	{
0293615f GC	146	unsigned long khz;
	147	u64 lpj;
	148
e93353c9	149	khz = kvm_get_tsc_khz();
0293615f GC	150
	151	lpj = ((u64)khz * 1000);
	152	do_div(lpj, HZ);
	153	preset_lpj = lpj;
	154	}
	155
3b5d56b9 EM	156	bool kvm_check_and_clear_guest_paused(void)
	157	{
	158	bool ret = false;
	159	struct pvclock_vcpu_time_info *src;
7069ed67 MT	160	int cpu = smp_processor_id();
	161
	162	if (!hv_clock)
	163	return ret;
3b5d56b9	164
3dc4f7cf	165	src = &hv_clock[cpu].pvti;
3b5d56b9	166	if ((src->flags & PVCLOCK_GUEST_STOPPED) != 0) {
7069ed67	167	src->flags &= ~PVCLOCK_GUEST_STOPPED;
d63285e9	168	pvclock_touch_watchdogs();
3b5d56b9 EM	169	ret = true;
	170	}
	171
	172	return ret;
	173	}
3b5d56b9	174
f4066c2b	175	struct clocksource kvm_clock = {
790c73f6	176	.name = "kvm-clock",
8e19608e	177	.read = kvm_clock_get_cycles,
790c73f6 GOC	178	.rating = 400,
790c73f6 GOC	179	.mask = CLOCKSOURCE_MASK(64),
790c73f6 GOC	180	.flags = CLOCK_SOURCE_IS_CONTINUOUS,
790c73f6 GOC	181	};
f4066c2b	182	EXPORT_SYMBOL_GPL(kvm_clock);
790c73f6	183
ca3f1017	184	int kvm_register_clock(char *txt)
790c73f6 GOC	185	{
790c73f6 GOC	186	int cpu = smp_processor_id();
19b6a85b	187	int low, high, ret;
fe1140cc JK	188	struct pvclock_vcpu_time_info *src;
	189
	190	if (!hv_clock)
	191	return 0;
19b6a85b	192
fe1140cc	193	src = &hv_clock[cpu].pvti;
5dfd486c DH	194	low = (int)slow_virt_to_phys(src) \| 1;
5dfd486c DH	195	high = ((u64)slow_virt_to_phys(src) >> 32);
19b6a85b	196	ret = native_write_msr_safe(msr_kvm_system_time, low, high);
f6e16d5a GH	197	printk(KERN_INFO "kvm-clock: cpu %d, msr %x:%x, %s\n",
f6e16d5a GH	198	cpu, high, low, txt);
838815a7	199
19b6a85b	200	return ret;
790c73f6 GOC	201	}
790c73f6 GOC	202
b74f05d6 MT	203	static void kvm_save_sched_clock_state(void)
	204	{
	205	}
	206
	207	static void kvm_restore_sched_clock_state(void)
	208	{
	209	kvm_register_clock("primary cpu clock, resume");
	210	}
	211
b8ba5f10	212	#ifdef CONFIG_X86_LOCAL_APIC
148f9bb8	213	static void kvm_setup_secondary_clock(void)
790c73f6 GOC	214	{
	215	/*
	216	* Now that the first cpu already had this clocksource initialized,
	217	* we shouldn't fail.
	218	*/
f6e16d5a	219	WARN_ON(kvm_register_clock("secondary cpu clock"));
790c73f6	220	}
b8ba5f10	221	#endif
790c73f6	222
1e977aa1 GC	223	/*
	224	* After the clock is registered, the host will keep writing to the
	225	* registered memory location. If the guest happens to shutdown, this memory
	226	* won't be valid. In cases like kexec, in which you install a new kernel, this
	227	* means a random memory location will be kept being written. So before any
6a6256f9	228	* kind of shutdown from our side, we unregister the clock by writing anything
1e977aa1 GC	229	* that does not have the 'enable' bit set in the msr
1e977aa1 GC	230	*/
2965faa5	231	#ifdef CONFIG_KEXEC_CORE
1e977aa1 GC	232	static void kvm_crash_shutdown(struct pt_regs *regs)
1e977aa1 GC	233	{
838815a7	234	native_write_msr(msr_kvm_system_time, 0, 0);
d910f5c1	235	kvm_disable_steal_time();
1e977aa1 GC	236	native_machine_crash_shutdown(regs);
	237	}
	238	#endif
	239
	240	static void kvm_shutdown(void)
	241	{
838815a7	242	native_write_msr(msr_kvm_system_time, 0, 0);
d910f5c1	243	kvm_disable_steal_time();
1e977aa1 GC	244	native_machine_shutdown();
	245	}
	246
819aeee0 BS	247	static phys_addr_t __init kvm_memblock_alloc(phys_addr_t size,
	248	phys_addr_t align)
	249	{
	250	phys_addr_t mem;
	251
	252	mem = memblock_alloc(size, align);
	253	if (!mem)
	254	return 0;
	255
	256	if (sev_active()) {
	257	if (early_set_memory_decrypted((unsigned long)__va(mem), size))
	258	goto e_free;
	259	}
	260
	261	return mem;
	262	e_free:
	263	memblock_free(mem, size);
	264	return 0;
	265	}
	266
	267	static void __init kvm_memblock_free(phys_addr_t addr, phys_addr_t size)
	268	{
	269	if (sev_active())
	270	early_set_memory_encrypted((unsigned long)__va(addr), size);
	271
	272	memblock_free(addr, size);
	273	}
	274
790c73f6 GOC	275	void __init kvmclock_init(void)
790c73f6 GOC	276	{
0ad83caa	277	struct pvclock_vcpu_time_info *vcpu_time;
819aeee0 BS	278	unsigned long mem, mem_wall_clock;
819aeee0 BS	279	int size, cpu, wall_clock_size;
0ad83caa	280	u8 flags;
ed55705d MT	281
ed55705d MT	282	size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
7069ed67	283
790c73f6 GOC	284	if (!kvm_para_available())
	285	return;
	286
838815a7 GC	287	if (kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE2)) {
	288	msr_kvm_system_time = MSR_KVM_SYSTEM_TIME_NEW;
	289	msr_kvm_wall_clock = MSR_KVM_WALL_CLOCK_NEW;
	290	} else if (!(kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE)))
	291	return;
	292
819aeee0 BS	293	wall_clock_size = PAGE_ALIGN(sizeof(struct pvclock_wall_clock));
	294	mem_wall_clock = kvm_memblock_alloc(wall_clock_size, PAGE_SIZE);
	295	if (!mem_wall_clock)
	296	return;
838815a7	297
819aeee0 BS	298	wall_clock = __va(mem_wall_clock);
	299	memset(wall_clock, 0, wall_clock_size);
	300
	301	mem = kvm_memblock_alloc(size, PAGE_SIZE);
	302	if (!mem) {
	303	kvm_memblock_free(mem_wall_clock, wall_clock_size);
	304	wall_clock = NULL;
7069ed67	305	return;
819aeee0 BS	306	}
819aeee0 BS	307
7069ed67	308	hv_clock = __va(mem);
07868fc6	309	memset(hv_clock, 0, size);
7069ed67	310
0d75de4a	311	if (kvm_register_clock("primary cpu clock")) {
7069ed67	312	hv_clock = NULL;
819aeee0 BS	313	kvm_memblock_free(mem, size);
	314	kvm_memblock_free(mem_wall_clock, wall_clock_size);
	315	wall_clock = NULL;
838815a7	316	return;
7069ed67	317	}
72c930dc	318
819aeee0 BS	319	printk(KERN_INFO "kvm-clock: Using msrs %x and %x",
	320	msr_kvm_system_time, msr_kvm_wall_clock);
	321
72c930dc RK	322	if (kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE_STABLE_BIT))
	323	pvclock_set_flags(PVCLOCK_TSC_STABLE_BIT);
	324
	325	cpu = get_cpu();
	326	vcpu_time = &hv_clock[cpu].pvti;
	327	flags = pvclock_read_flags(vcpu_time);
	328
	329	kvm_sched_clock_init(flags & PVCLOCK_TSC_STABLE_BIT);
	330	put_cpu();
	331
838815a7	332	x86_platform.calibrate_tsc = kvm_get_tsc_khz;
a4497a86	333	x86_platform.calibrate_cpu = kvm_get_tsc_khz;
838815a7 GC	334	x86_platform.get_wallclock = kvm_get_wallclock;
838815a7 GC	335	x86_platform.set_wallclock = kvm_set_wallclock;
b8ba5f10	336	#ifdef CONFIG_X86_LOCAL_APIC
df156f90	337	x86_cpuinit.early_percpu_clock_init =
838815a7	338	kvm_setup_secondary_clock;
b8ba5f10	339	#endif
b74f05d6 MT	340	x86_platform.save_sched_clock_state = kvm_save_sched_clock_state;
b74f05d6 MT	341	x86_platform.restore_sched_clock_state = kvm_restore_sched_clock_state;
838815a7	342	machine_ops.shutdown = kvm_shutdown;
2965faa5	343	#ifdef CONFIG_KEXEC_CORE
838815a7	344	machine_ops.crash_shutdown = kvm_crash_shutdown;
1e977aa1	345	#endif
838815a7	346	kvm_get_preset_lpj();
b01cc1b0	347	clocksource_register_hz(&kvm_clock, NSEC_PER_SEC);
838815a7	348	pv_info.name = "KVM";
790c73f6	349	}
3dc4f7cf MT	350
	351	int __init kvm_setup_vsyscall_timeinfo(void)
	352	{
	353	#ifdef CONFIG_X86_64
	354	int cpu;
3dc4f7cf MT	355	u8 flags;
	356	struct pvclock_vcpu_time_info *vcpu_time;
	357	unsigned int size;
	358
fe1140cc JK	359	if (!hv_clock)
	360	return 0;
	361
ed55705d	362	size = PAGE_ALIGN(sizeof(struct pvclock_vsyscall_time_info)*NR_CPUS);
3dc4f7cf	363
c6338ce4	364	cpu = get_cpu();
3dc4f7cf MT	365
	366	vcpu_time = &hv_clock[cpu].pvti;
	367	flags = pvclock_read_flags(vcpu_time);
	368
	369	if (!(flags & PVCLOCK_TSC_STABLE_BIT)) {
c6338ce4	370	put_cpu();
3dc4f7cf MT	371	return 1;
	372	}
	373
9f08890a	374	pvclock_set_pvti_cpu0_va(hv_clock);
c6338ce4	375	put_cpu();
3dc4f7cf MT	376
	377	kvm_clock.archdata.vclock_mode = VCLOCK_PVCLOCK;
	378	#endif
	379	return 0;
	380	}