[mirror_ubuntu-zesty-kernel.git] / drivers / hv / hv.c

/*
 * Copyright (c) 2009, Microsoft Corporation.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope 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., 59 Temple
 * Place - Suite 330, Boston, MA 02111-1307 USA.
 *
 * Authors:
 *   Haiyang Zhang <haiyangz@microsoft.com>
 *   Hank Janssen  <hjanssen@microsoft.com>
 *
 */
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/hyperv.h>
#include <linux/version.h>
#include <linux/interrupt.h>
#include <linux/clockchips.h>
#include <asm/hyperv.h>
#include <asm/mshyperv.h>
#include "hyperv_vmbus.h"

/* The one and only */
struct hv_context hv_context = {
	.synic_initialized	= false,
};

#define HV_TIMER_FREQUENCY (10 * 1000 * 1000) /* 100ns period */
#define HV_MAX_MAX_DELTA_TICKS 0xffffffff
#define HV_MIN_DELTA_TICKS 1

/*
 * hv_init - Main initialization routine.
 *
 * This routine must be called before any other routines in here are called
 */
int hv_init(void)
{

	memset(hv_context.synic_event_page, 0, sizeof(void *) * NR_CPUS);
	memset(hv_context.synic_message_page, 0,
	       sizeof(void *) * NR_CPUS);
	memset(hv_context.post_msg_page, 0,
	       sizeof(void *) * NR_CPUS);
	memset(hv_context.vp_index, 0,
	       sizeof(int) * NR_CPUS);
	memset(hv_context.event_dpc, 0,
	       sizeof(void *) * NR_CPUS);
	memset(hv_context.msg_dpc, 0,
	       sizeof(void *) * NR_CPUS);
	memset(hv_context.clk_evt, 0,
	       sizeof(void *) * NR_CPUS);

	if (!hv_is_hypercall_page_setup())
		return -ENOTSUPP;

	return 0;
}

/*
 * hv_cleanup - Cleanup routine.
 *
 * This routine is called normally during driver unloading or exiting.
 */
void hv_cleanup(bool crash)
{

}

/*
 * hv_post_message - Post a message using the hypervisor message IPC.
 *
 * This involves a hypercall.
 */
int hv_post_message(union hv_connection_id connection_id,
		  enum hv_message_type message_type,
		  void *payload, size_t payload_size)
{

	struct hv_input_post_message *aligned_msg;
	u64 status;

	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
		return -EMSGSIZE;

	aligned_msg = (struct hv_input_post_message *)
			hv_context.post_msg_page[get_cpu()];

	aligned_msg->connectionid = connection_id;
	aligned_msg->reserved = 0;
	aligned_msg->message_type = message_type;
	aligned_msg->payload_size = payload_size;
	memcpy((void *)aligned_msg->payload, payload, payload_size);

	status = hv_do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL);

	put_cpu();
	return status & 0xFFFF;
}

static int hv_ce_set_next_event(unsigned long delta,
				struct clock_event_device *evt)
{
	u64 current_tick;

	WARN_ON(!clockevent_state_oneshot(evt));

	hv_get_current_tick(current_tick);
	current_tick += delta;
	hv_init_timer(HV_X64_MSR_STIMER0_COUNT, current_tick);
	return 0;
}

static int hv_ce_shutdown(struct clock_event_device *evt)
{
	hv_init_timer(HV_X64_MSR_STIMER0_COUNT, 0);
	hv_init_timer_config(HV_X64_MSR_STIMER0_CONFIG, 0);

	return 0;
}

static int hv_ce_set_oneshot(struct clock_event_device *evt)
{
	union hv_timer_config timer_cfg;

	timer_cfg.enable = 1;
	timer_cfg.auto_enable = 1;
	timer_cfg.sintx = VMBUS_MESSAGE_SINT;
	hv_init_timer_config(HV_X64_MSR_STIMER0_CONFIG, timer_cfg.as_uint64);

	return 0;
}

static void hv_init_clockevent_device(struct clock_event_device *dev, int cpu)
{
	dev->name = "Hyper-V clockevent";
	dev->features = CLOCK_EVT_FEAT_ONESHOT;
	dev->cpumask = cpumask_of(cpu);
	dev->rating = 1000;
	/*
	 * Avoid settint dev->owner = THIS_MODULE deliberately as doing so will
	 * result in clockevents_config_and_register() taking additional
	 * references to the hv_vmbus module making it impossible to unload.
	 */

	dev->set_state_shutdown = hv_ce_shutdown;
	dev->set_state_oneshot = hv_ce_set_oneshot;
	dev->set_next_event = hv_ce_set_next_event;
}


int hv_synic_alloc(void)
{
	size_t size = sizeof(struct tasklet_struct);
	size_t ced_size = sizeof(struct clock_event_device);
	int cpu;

	hv_context.hv_numa_map = kzalloc(sizeof(struct cpumask) * nr_node_ids,
					 GFP_ATOMIC);
	if (hv_context.hv_numa_map == NULL) {
		pr_err("Unable to allocate NUMA map\n");
		goto err;
	}

	for_each_present_cpu(cpu) {
		hv_context.event_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
		if (hv_context.event_dpc[cpu] == NULL) {
			pr_err("Unable to allocate event dpc\n");
			goto err;
		}
		tasklet_init(hv_context.event_dpc[cpu], vmbus_on_event, cpu);

		hv_context.msg_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
		if (hv_context.msg_dpc[cpu] == NULL) {
			pr_err("Unable to allocate event dpc\n");
			goto err;
		}
		tasklet_init(hv_context.msg_dpc[cpu], vmbus_on_msg_dpc, cpu);

		hv_context.clk_evt[cpu] = kzalloc(ced_size, GFP_ATOMIC);
		if (hv_context.clk_evt[cpu] == NULL) {
			pr_err("Unable to allocate clock event device\n");
			goto err;
		}

		hv_init_clockevent_device(hv_context.clk_evt[cpu], cpu);

		hv_context.synic_message_page[cpu] =
			(void *)get_zeroed_page(GFP_ATOMIC);

		if (hv_context.synic_message_page[cpu] == NULL) {
			pr_err("Unable to allocate SYNIC message page\n");
			goto err;
		}

		hv_context.synic_event_page[cpu] =
			(void *)get_zeroed_page(GFP_ATOMIC);

		if (hv_context.synic_event_page[cpu] == NULL) {
			pr_err("Unable to allocate SYNIC event page\n");
			goto err;
		}

		hv_context.post_msg_page[cpu] =
			(void *)get_zeroed_page(GFP_ATOMIC);

		if (hv_context.post_msg_page[cpu] == NULL) {
			pr_err("Unable to allocate post msg page\n");
			goto err;
		}

		INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);
	}

	return 0;
err:
	return -ENOMEM;
}

static void hv_synic_free_cpu(int cpu)
{
	kfree(hv_context.event_dpc[cpu]);
	kfree(hv_context.msg_dpc[cpu]);
	kfree(hv_context.clk_evt[cpu]);
	if (hv_context.synic_event_page[cpu])
		free_page((unsigned long)hv_context.synic_event_page[cpu]);
	if (hv_context.synic_message_page[cpu])
		free_page((unsigned long)hv_context.synic_message_page[cpu]);
	if (hv_context.post_msg_page[cpu])
		free_page((unsigned long)hv_context.post_msg_page[cpu]);
}

void hv_synic_free(void)
{
	int cpu;

	kfree(hv_context.hv_numa_map);
	for_each_present_cpu(cpu)
		hv_synic_free_cpu(cpu);
}

/*
 * hv_synic_init - Initialize the Synthethic Interrupt Controller.
 *
 * If it is already initialized by another entity (ie x2v shim), we need to
 * retrieve the initialized message and event pages.  Otherwise, we create and
 * initialize the message and event pages.
 */
int hv_synic_init(unsigned int cpu)
{
	union hv_synic_simp simp;
	union hv_synic_siefp siefp;
	union hv_synic_sint shared_sint;
	union hv_synic_scontrol sctrl;
	u64 vp_index;

	/* Setup the Synic's message page */
	hv_get_simp(simp.as_uint64);
	simp.simp_enabled = 1;
	simp.base_simp_gpa = virt_to_phys(hv_context.synic_message_page[cpu])
		>> PAGE_SHIFT;

	hv_set_simp(simp.as_uint64);

	/* Setup the Synic's event page */
	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
	siefp.siefp_enabled = 1;
	siefp.base_siefp_gpa = virt_to_phys(hv_context.synic_event_page[cpu])
		>> PAGE_SHIFT;

	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);

	/* Setup the shared SINT. */
	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	shared_sint.as_uint64 = 0;
	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
	shared_sint.masked = false;
	shared_sint.auto_eoi = true;

	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	/* Enable the global synic bit */
	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	sctrl.enable = 1;

	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);

	hv_context.synic_initialized = true;

	/*
	 * Setup the mapping between Hyper-V's notion
	 * of cpuid and Linux' notion of cpuid.
	 * This array will be indexed using Linux cpuid.
	 */
	rdmsrl(HV_X64_MSR_VP_INDEX, vp_index);
	hv_context.vp_index[cpu] = (u32)vp_index;

	/*
	 * Register the per-cpu clockevent source.
	 */
	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
		clockevents_config_and_register(hv_context.clk_evt[cpu],
						HV_TIMER_FREQUENCY,
						HV_MIN_DELTA_TICKS,
						HV_MAX_MAX_DELTA_TICKS);
	return 0;
}

/*
 * hv_synic_clockevents_cleanup - Cleanup clockevent devices
 */
void hv_synic_clockevents_cleanup(void)
{
	int cpu;

	if (!(ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE))
		return;

	for_each_present_cpu(cpu)
		clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
}

/*
 * hv_synic_cleanup - Cleanup routine for hv_synic_init().
 */
int hv_synic_cleanup(unsigned int cpu)
{
	union hv_synic_sint shared_sint;
	union hv_synic_simp simp;
	union hv_synic_siefp siefp;
	union hv_synic_scontrol sctrl;
	struct vmbus_channel *channel, *sc;
	bool channel_found = false;
	unsigned long flags;

	if (!hv_context.synic_initialized)
		return -EFAULT;

	/*
	 * Search for channels which are bound to the CPU we're about to
	 * cleanup. In case we find one and vmbus is still connected we need to
	 * fail, this will effectively prevent CPU offlining. There is no way
	 * we can re-bind channels to different CPUs for now.
	 */
	mutex_lock(&vmbus_connection.channel_mutex);
	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
		if (channel->target_cpu == cpu) {
			channel_found = true;
			break;
		}
		spin_lock_irqsave(&channel->lock, flags);
		list_for_each_entry(sc, &channel->sc_list, sc_list) {
			if (sc->target_cpu == cpu) {
				channel_found = true;
				break;
			}
		}
		spin_unlock_irqrestore(&channel->lock, flags);
		if (channel_found)
			break;
	}
	mutex_unlock(&vmbus_connection.channel_mutex);

	if (channel_found && vmbus_connection.conn_state == CONNECTED)
		return -EBUSY;

	/* Turn off clockevent device */
	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE) {
		clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
		hv_ce_shutdown(hv_context.clk_evt[cpu]);
	}

	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	shared_sint.masked = 1;

	/* Need to correctly cleanup in the case of SMP!!! */
	/* Disable the interrupt */
	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	hv_get_simp(simp.as_uint64);
	simp.simp_enabled = 0;
	simp.base_simp_gpa = 0;

	hv_set_simp(simp.as_uint64);

	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
	siefp.siefp_enabled = 0;
	siefp.base_siefp_gpa = 0;

	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);

	/* Disable the global synic bit */
	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	sctrl.enable = 0;
	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);

	return 0;
}
Commit	Line	Data
3e7ee490	1	/*
3e7ee490 HJ	2	* Copyright (c) 2009, Microsoft Corporation.
	3	*
	4	* This program is free software; you can redistribute it and/or modify it
	5	* under the terms and conditions of the GNU General Public License,
	6	* version 2, as published by the Free Software Foundation.
	7	*
	8	* This program is distributed in the hope it will be useful, but WITHOUT
	9	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
	10	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
	11	* more details.
	12	*
	13	* You should have received a copy of the GNU General Public License along with
	14	* this program; if not, write to the Free Software Foundation, Inc., 59 Temple
	15	* Place - Suite 330, Boston, MA 02111-1307 USA.
	16	*
	17	* Authors:
	18	* Haiyang Zhang <haiyangz@microsoft.com>
	19	* Hank Janssen <hjanssen@microsoft.com>
	20	*
	21	*/
0a46618d HJ	22	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
0a46618d HJ	23
a0086dc5 GKH	24	#include <linux/kernel.h>
a0086dc5 GKH	25	#include <linux/mm.h>
5a0e3ad6	26	#include <linux/slab.h>
b7c947f0	27	#include <linux/vmalloc.h>
46a97191	28	#include <linux/hyperv.h>
83ba0c4f	29	#include <linux/version.h>
db11f12a	30	#include <linux/interrupt.h>
4061ed9e	31	#include <linux/clockchips.h>
407dd164	32	#include <asm/hyperv.h>
4061ed9e	33	#include <asm/mshyperv.h>
0f2a6619	34	#include "hyperv_vmbus.h"
3e7ee490	35
454f18a9	36	/* The one and only */
6a0aaa18 HZ	37	struct hv_context hv_context = {
6a0aaa18 HZ	38	.synic_initialized = false,
3e7ee490 HJ	39	};
3e7ee490 HJ	40
4061ed9e S	41	#define HV_TIMER_FREQUENCY (10 * 1000 * 1000) /* 100ns period */
	42	#define HV_MAX_MAX_DELTA_TICKS 0xffffffff
	43	#define HV_MIN_DELTA_TICKS 1
	44
3e189519	45	/*
d44890c8	46	* hv_init - Main initialization routine.
0831ad04 GKH	47	*
	48	* This routine must be called before any other routines in here are called
	49	*/
d44890c8	50	int hv_init(void)
3e7ee490	51	{
3e7ee490	52
14c1bf8a	53	memset(hv_context.synic_event_page, 0, sizeof(void ) NR_CPUS);
6a0aaa18	54	memset(hv_context.synic_message_page, 0,
14c1bf8a	55	sizeof(void ) NR_CPUS);
b29ef354 S	56	memset(hv_context.post_msg_page, 0,
b29ef354 S	57	sizeof(void ) NR_CPUS);
917ea427 S	58	memset(hv_context.vp_index, 0,
917ea427 S	59	sizeof(int) * NR_CPUS);
db11f12a S	60	memset(hv_context.event_dpc, 0,
db11f12a S	61	sizeof(void ) NR_CPUS);
d81274aa S	62	memset(hv_context.msg_dpc, 0,
d81274aa S	63	sizeof(void ) NR_CPUS);
4061ed9e S	64	memset(hv_context.clk_evt, 0,
4061ed9e S	65	sizeof(void ) NR_CPUS);
3e7ee490	66
c4398016	67	if (!hv_is_hypercall_page_setup())
b1d6b256	68	return -ENOTSUPP;
3e7ee490	69
5433e003	70	return 0;
3e7ee490 HJ	71	}
3e7ee490 HJ	72
3e189519	73	/*
d44890c8	74	* hv_cleanup - Cleanup routine.
0831ad04 GKH	75	*
	76	* This routine is called normally during driver unloading or exiting.
	77	*/
a9f61ca7	78	void hv_cleanup(bool crash)
3e7ee490	79	{
ca9357bd	80
3e7ee490 HJ	81	}
3e7ee490 HJ	82
3e189519	83	/*
d44890c8	84	* hv_post_message - Post a message using the hypervisor message IPC.
0831ad04 GKH	85	*
	86	* This involves a hypercall.
	87	*/
415f0a02	88	int hv_post_message(union hv_connection_id connection_id,
b8dfb264 HZ	89	enum hv_message_type message_type,
b8dfb264 HZ	90	void *payload, size_t payload_size)
3e7ee490	91	{
3e7ee490	92
b8dfb264	93	struct hv_input_post_message *aligned_msg;
a108393d	94	u64 status;
3e7ee490	95
b8dfb264	96	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
39594abc	97	return -EMSGSIZE;
3e7ee490	98
b8dfb264	99	aligned_msg = (struct hv_input_post_message *)
b29ef354	100	hv_context.post_msg_page[get_cpu()];
3e7ee490	101
b8dfb264	102	aligned_msg->connectionid = connection_id;
b29ef354	103	aligned_msg->reserved = 0;
b8dfb264 HZ	104	aligned_msg->message_type = message_type;
	105	aligned_msg->payload_size = payload_size;
	106	memcpy((void *)aligned_msg->payload, payload, payload_size);
3e7ee490	107
a108393d	108	status = hv_do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL);
3e7ee490	109
b29ef354	110	put_cpu();
a108393d	111	return status & 0xFFFF;
3e7ee490 HJ	112	}
3e7ee490 HJ	113
4061ed9e S	114	static int hv_ce_set_next_event(unsigned long delta,
	115	struct clock_event_device *evt)
	116	{
a5a1d1c2	117	u64 current_tick;
4061ed9e	118
bc609cb4	119	WARN_ON(!clockevent_state_oneshot(evt));
4061ed9e	120
72329152	121	hv_get_current_tick(current_tick);
4061ed9e	122	current_tick += delta;
72329152	123	hv_init_timer(HV_X64_MSR_STIMER0_COUNT, current_tick);
4061ed9e S	124	return 0;
	125	}
	126
bc609cb4 VK	127	static int hv_ce_shutdown(struct clock_event_device *evt)
bc609cb4 VK	128	{
72329152 S	129	hv_init_timer(HV_X64_MSR_STIMER0_COUNT, 0);
72329152 S	130	hv_init_timer_config(HV_X64_MSR_STIMER0_CONFIG, 0);
bc609cb4 VK	131
	132	return 0;
	133	}
	134
	135	static int hv_ce_set_oneshot(struct clock_event_device *evt)
4061ed9e S	136	{
	137	union hv_timer_config timer_cfg;
	138
bc609cb4 VK	139	timer_cfg.enable = 1;
	140	timer_cfg.auto_enable = 1;
	141	timer_cfg.sintx = VMBUS_MESSAGE_SINT;
72329152	142	hv_init_timer_config(HV_X64_MSR_STIMER0_CONFIG, timer_cfg.as_uint64);
bc609cb4 VK	143
bc609cb4 VK	144	return 0;
4061ed9e S	145	}
	146
	147	static void hv_init_clockevent_device(struct clock_event_device *dev, int cpu)
	148	{
	149	dev->name = "Hyper-V clockevent";
	150	dev->features = CLOCK_EVT_FEAT_ONESHOT;
	151	dev->cpumask = cpumask_of(cpu);
	152	dev->rating = 1000;
e086748c VK	153	/*
	154	* Avoid settint dev->owner = THIS_MODULE deliberately as doing so will
	155	* result in clockevents_config_and_register() taking additional
	156	* references to the hv_vmbus module making it impossible to unload.
	157	*/
4061ed9e	158
bc609cb4 VK	159	dev->set_state_shutdown = hv_ce_shutdown;
bc609cb4 VK	160	dev->set_state_oneshot = hv_ce_set_oneshot;
4061ed9e S	161	dev->set_next_event = hv_ce_set_next_event;
	162	}
	163
2608fb65 JW	164
	165	int hv_synic_alloc(void)
	166	{
	167	size_t size = sizeof(struct tasklet_struct);
4061ed9e	168	size_t ced_size = sizeof(struct clock_event_device);
2608fb65 JW	169	int cpu;
2608fb65 JW	170
9f01ec53 S	171	hv_context.hv_numa_map = kzalloc(sizeof(struct cpumask) * nr_node_ids,
	172	GFP_ATOMIC);
	173	if (hv_context.hv_numa_map == NULL) {
	174	pr_err("Unable to allocate NUMA map\n");
	175	goto err;
	176	}
	177
d74e2e80	178	for_each_present_cpu(cpu) {
2608fb65 JW	179	hv_context.event_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
	180	if (hv_context.event_dpc[cpu] == NULL) {
	181	pr_err("Unable to allocate event dpc\n");
	182	goto err;
	183	}
	184	tasklet_init(hv_context.event_dpc[cpu], vmbus_on_event, cpu);
	185
d81274aa S	186	hv_context.msg_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
	187	if (hv_context.msg_dpc[cpu] == NULL) {
	188	pr_err("Unable to allocate event dpc\n");
	189	goto err;
	190	}
	191	tasklet_init(hv_context.msg_dpc[cpu], vmbus_on_msg_dpc, cpu);
	192
4061ed9e S	193	hv_context.clk_evt[cpu] = kzalloc(ced_size, GFP_ATOMIC);
	194	if (hv_context.clk_evt[cpu] == NULL) {
	195	pr_err("Unable to allocate clock event device\n");
	196	goto err;
	197	}
9f01ec53	198
4061ed9e S	199	hv_init_clockevent_device(hv_context.clk_evt[cpu], cpu);
4061ed9e S	200
2608fb65 JW	201	hv_context.synic_message_page[cpu] =
	202	(void *)get_zeroed_page(GFP_ATOMIC);
	203
	204	if (hv_context.synic_message_page[cpu] == NULL) {
	205	pr_err("Unable to allocate SYNIC message page\n");
	206	goto err;
	207	}
	208
	209	hv_context.synic_event_page[cpu] =
	210	(void *)get_zeroed_page(GFP_ATOMIC);
	211
	212	if (hv_context.synic_event_page[cpu] == NULL) {
	213	pr_err("Unable to allocate SYNIC event page\n");
	214	goto err;
	215	}
b29ef354 S	216
	217	hv_context.post_msg_page[cpu] =
	218	(void *)get_zeroed_page(GFP_ATOMIC);
	219
	220	if (hv_context.post_msg_page[cpu] == NULL) {
	221	pr_err("Unable to allocate post msg page\n");
	222	goto err;
	223	}
9b1bf703 VK	224
9b1bf703 VK	225	INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);
2608fb65 JW	226	}
	227
	228	return 0;
	229	err:
	230	return -ENOMEM;
	231	}
	232
8712954d	233	static void hv_synic_free_cpu(int cpu)
2608fb65 JW	234	{
2608fb65 JW	235	kfree(hv_context.event_dpc[cpu]);
d81274aa	236	kfree(hv_context.msg_dpc[cpu]);
4061ed9e	237	kfree(hv_context.clk_evt[cpu]);
fdf91dae	238	if (hv_context.synic_event_page[cpu])
2608fb65 JW	239	free_page((unsigned long)hv_context.synic_event_page[cpu]);
	240	if (hv_context.synic_message_page[cpu])
	241	free_page((unsigned long)hv_context.synic_message_page[cpu]);
b29ef354 S	242	if (hv_context.post_msg_page[cpu])
b29ef354 S	243	free_page((unsigned long)hv_context.post_msg_page[cpu]);
2608fb65 JW	244	}
	245
	246	void hv_synic_free(void)
	247	{
	248	int cpu;
	249
9f01ec53	250	kfree(hv_context.hv_numa_map);
d74e2e80	251	for_each_present_cpu(cpu)
2608fb65 JW	252	hv_synic_free_cpu(cpu);
	253	}
	254
3e189519	255	/*
d44890c8	256	* hv_synic_init - Initialize the Synthethic Interrupt Controller.
0831ad04 GKH	257	*
	258	* If it is already initialized by another entity (ie x2v shim), we need to
	259	* retrieve the initialized message and event pages. Otherwise, we create and
	260	* initialize the message and event pages.
	261	*/
fdf5149c	262	int hv_synic_init(unsigned int cpu)
3e7ee490	263	{
eacb1b4d GKH	264	union hv_synic_simp simp;
eacb1b4d GKH	265	union hv_synic_siefp siefp;
b8dfb264	266	union hv_synic_sint shared_sint;
eacb1b4d	267	union hv_synic_scontrol sctrl;
917ea427	268	u64 vp_index;
a73e6b7c	269
a73e6b7c	270	/* Setup the Synic's message page */
a62c7499	271	hv_get_simp(simp.as_uint64);
f6feebe0	272	simp.simp_enabled = 1;
6a0aaa18	273	simp.base_simp_gpa = virt_to_phys(hv_context.synic_message_page[cpu])
a73e6b7c	274	>> PAGE_SHIFT;
3e7ee490	275
a62c7499	276	hv_set_simp(simp.as_uint64);
3e7ee490	277
a73e6b7c	278	/* Setup the Synic's event page */
f6feebe0 HZ	279	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
f6feebe0 HZ	280	siefp.siefp_enabled = 1;
6a0aaa18	281	siefp.base_siefp_gpa = virt_to_phys(hv_context.synic_event_page[cpu])
a73e6b7c HJ	282	>> PAGE_SHIFT;
a73e6b7c HJ	283
f6feebe0	284	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
0831ad04	285
0831ad04	286	/* Setup the shared SINT. */
b8dfb264	287	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	288
b8dfb264	289	shared_sint.as_uint64 = 0;
302a3c0f	290	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
b8dfb264	291	shared_sint.masked = false;
b0209501	292	shared_sint.auto_eoi = true;
3e7ee490	293
b8dfb264	294	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	295
454f18a9	296	/* Enable the global synic bit */
f6feebe0 HZ	297	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
f6feebe0 HZ	298	sctrl.enable = 1;
3e7ee490	299
f6feebe0	300	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
3e7ee490	301
6a0aaa18	302	hv_context.synic_initialized = true;
917ea427 S	303
	304	/*
	305	* Setup the mapping between Hyper-V's notion
	306	* of cpuid and Linux' notion of cpuid.
	307	* This array will be indexed using Linux cpuid.
	308	*/
	309	rdmsrl(HV_X64_MSR_VP_INDEX, vp_index);
	310	hv_context.vp_index[cpu] = (u32)vp_index;
3a28fa35	311
4061ed9e S	312	/*
	313	* Register the per-cpu clockevent source.
	314	*/
	315	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
	316	clockevents_config_and_register(hv_context.clk_evt[cpu],
	317	HV_TIMER_FREQUENCY,
	318	HV_MIN_DELTA_TICKS,
	319	HV_MAX_MAX_DELTA_TICKS);
fdf5149c	320	return 0;
3e7ee490 HJ	321	}
3e7ee490 HJ	322
e086748c VK	323	/*
	324	* hv_synic_clockevents_cleanup - Cleanup clockevent devices
	325	*/
	326	void hv_synic_clockevents_cleanup(void)
	327	{
	328	int cpu;
	329
	330	if (!(ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE))
	331	return;
	332
6ffc4b85	333	for_each_present_cpu(cpu)
e086748c VK	334	clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
	335	}
	336
3e189519	337	/*
d44890c8	338	* hv_synic_cleanup - Cleanup routine for hv_synic_init().
0831ad04	339	*/
fdf5149c	340	int hv_synic_cleanup(unsigned int cpu)
3e7ee490	341	{
b8dfb264	342	union hv_synic_sint shared_sint;
eacb1b4d GKH	343	union hv_synic_simp simp;
eacb1b4d GKH	344	union hv_synic_siefp siefp;
e72e7ac5	345	union hv_synic_scontrol sctrl;
ccbf4254 VK	346	struct vmbus_channel channel, sc;
	347	bool channel_found = false;
	348	unsigned long flags;
3e7ee490	349
6a0aaa18	350	if (!hv_context.synic_initialized)
fdf5149c	351	return -EFAULT;
3e7ee490	352
ccbf4254 VK	353	/*
	354	* Search for channels which are bound to the CPU we're about to
	355	* cleanup. In case we find one and vmbus is still connected we need to
	356	* fail, this will effectively prevent CPU offlining. There is no way
	357	* we can re-bind channels to different CPUs for now.
	358	*/
	359	mutex_lock(&vmbus_connection.channel_mutex);
	360	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
	361	if (channel->target_cpu == cpu) {
	362	channel_found = true;
	363	break;
	364	}
	365	spin_lock_irqsave(&channel->lock, flags);
	366	list_for_each_entry(sc, &channel->sc_list, sc_list) {
	367	if (sc->target_cpu == cpu) {
	368	channel_found = true;
	369	break;
	370	}
	371	}
	372	spin_unlock_irqrestore(&channel->lock, flags);
	373	if (channel_found)
	374	break;
	375	}
	376	mutex_unlock(&vmbus_connection.channel_mutex);
	377
	378	if (channel_found && vmbus_connection.conn_state == CONNECTED)
	379	return -EBUSY;
	380
e086748c	381	/* Turn off clockevent device */
6ffc4b85 VK	382	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE) {
6ffc4b85 VK	383	clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
bc609cb4	384	hv_ce_shutdown(hv_context.clk_evt[cpu]);
6ffc4b85	385	}
e086748c	386
b8dfb264	387	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	388
b8dfb264	389	shared_sint.masked = 1;
3e7ee490	390
7692fd4d	391	/* Need to correctly cleanup in the case of SMP!!! */
454f18a9	392	/* Disable the interrupt */
b8dfb264	393	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	394
a62c7499	395	hv_get_simp(simp.as_uint64);
f6feebe0 HZ	396	simp.simp_enabled = 0;
f6feebe0 HZ	397	simp.base_simp_gpa = 0;
3e7ee490	398
a62c7499	399	hv_set_simp(simp.as_uint64);
3e7ee490	400
f6feebe0 HZ	401	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
	402	siefp.siefp_enabled = 0;
	403	siefp.base_siefp_gpa = 0;
3e7ee490	404
f6feebe0	405	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
3e7ee490	406
e72e7ac5 VK	407	/* Disable the global synic bit */
	408	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	409	sctrl.enable = 0;
	410	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
fdf5149c VK	411
fdf5149c VK	412	return 0;
3e7ee490	413	}