[thirdparty/kernel/stable.git] / drivers / hv / hv.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright (c) 2009, Microsoft Corporation.
 *
 * 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/random.h>
#include <linux/clockchips.h>
#include <asm/mshyperv.h>
#include "hyperv_vmbus.h"

/* The one and only */
struct hv_context hv_context;

/*
 * If false, we're using the old mechanism for stimer0 interrupts
 * where it sends a VMbus message when it expires. The old
 * mechanism is used when running on older versions of Hyper-V
 * that don't support Direct Mode. While Hyper-V provides
 * four stimer's per CPU, Linux uses only stimer0.
 */
static bool direct_mode_enabled;
static int stimer0_irq;
static int stimer0_vector;

#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)
{
	hv_context.cpu_context = alloc_percpu(struct hv_per_cpu_context);
	if (!hv_context.cpu_context)
		return -ENOMEM;

	direct_mode_enabled = ms_hyperv.misc_features &
			HV_STIMER_DIRECT_MODE_AVAILABLE;
	return 0;
}

/*
 * 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;
	struct hv_per_cpu_context *hv_cpu;
	u64 status;

	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
		return -EMSGSIZE;

	hv_cpu = get_cpu_ptr(hv_context.cpu_context);
	aligned_msg = hv_cpu->post_msg_page;
	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);

	/* Preemption must remain disabled until after the hypercall
	 * so some other thread can't get scheduled onto this cpu and
	 * corrupt the per-cpu post_msg_page
	 */
	put_cpu_ptr(hv_cpu);

	return status & 0xFFFF;
}

/*
 * ISR for when stimer0 is operating in Direct Mode.  Direct Mode
 * does not use VMbus or any VMbus messages, so process here and not
 * in the VMbus driver code.
 */

static void hv_stimer0_isr(void)
{
	struct hv_per_cpu_context *hv_cpu;

	hv_cpu = this_cpu_ptr(hv_context.cpu_context);
	hv_cpu->clk_evt->event_handler(hv_cpu->clk_evt);
	add_interrupt_randomness(stimer0_vector, 0);
}

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

	WARN_ON(!clockevent_state_oneshot(evt));

	current_tick = hyperv_cs->read(NULL);
	current_tick += delta;
	hv_init_timer(0, current_tick);
	return 0;
}

static int hv_ce_shutdown(struct clock_event_device *evt)
{
	hv_init_timer(0, 0);
	hv_init_timer_config(0, 0);
	if (direct_mode_enabled)
		hv_disable_stimer0_percpu_irq(stimer0_irq);

	return 0;
}

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

	timer_cfg.as_uint64 = 0;
	timer_cfg.enable = 1;
	timer_cfg.auto_enable = 1;
	if (direct_mode_enabled) {
		/*
		 * When it expires, the timer will directly interrupt
		 * on the specified hardware vector/IRQ.
		 */
		timer_cfg.direct_mode = 1;
		timer_cfg.apic_vector = stimer0_vector;
		hv_enable_stimer0_percpu_irq(stimer0_irq);
	} else {
		/*
		 * When it expires, the timer will generate a VMbus message,
		 * to be handled by the normal VMbus interrupt handler.
		 */
		timer_cfg.direct_mode = 0;
		timer_cfg.sintx = VMBUS_MESSAGE_SINT;
	}
	hv_init_timer_config(0, 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)
{
	int cpu;
	struct hv_per_cpu_context *hv_cpu;

	/*
	 * First, zero all per-cpu memory areas so hv_synic_free() can
	 * detect what memory has been allocated and cleanup properly
	 * after any failures.
	 */
	for_each_present_cpu(cpu) {
		hv_cpu = per_cpu_ptr(hv_context.cpu_context, cpu);
		memset(hv_cpu, 0, sizeof(*hv_cpu));
	}

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

	for_each_present_cpu(cpu) {
		hv_cpu = per_cpu_ptr(hv_context.cpu_context, cpu);

		tasklet_init(&hv_cpu->msg_dpc,
			     vmbus_on_msg_dpc, (unsigned long) hv_cpu);

		hv_cpu->clk_evt = kzalloc(sizeof(struct clock_event_device),
					  GFP_KERNEL);
		if (hv_cpu->clk_evt == NULL) {
			pr_err("Unable to allocate clock event device\n");
			goto err;
		}
		hv_init_clockevent_device(hv_cpu->clk_evt, cpu);

		hv_cpu->synic_message_page =
			(void *)get_zeroed_page(GFP_ATOMIC);
		if (hv_cpu->synic_message_page == NULL) {
			pr_err("Unable to allocate SYNIC message page\n");
			goto err;
		}

		hv_cpu->synic_event_page = (void *)get_zeroed_page(GFP_ATOMIC);
		if (hv_cpu->synic_event_page == NULL) {
			pr_err("Unable to allocate SYNIC event page\n");
			goto err;
		}

		hv_cpu->post_msg_page = (void *)get_zeroed_page(GFP_ATOMIC);
		if (hv_cpu->post_msg_page == NULL) {
			pr_err("Unable to allocate post msg page\n");
			goto err;
		}

		INIT_LIST_HEAD(&hv_cpu->chan_list);
	}

	if (direct_mode_enabled &&
	    hv_setup_stimer0_irq(&stimer0_irq, &stimer0_vector,
				hv_stimer0_isr))
		goto err;

	return 0;
err:
	/*
	 * Any memory allocations that succeeded will be freed when
	 * the caller cleans up by calling hv_synic_free()
	 */
	return -ENOMEM;
}


void hv_synic_free(void)
{
	int cpu;

	for_each_present_cpu(cpu) {
		struct hv_per_cpu_context *hv_cpu
			= per_cpu_ptr(hv_context.cpu_context, cpu);

		kfree(hv_cpu->clk_evt);
		free_page((unsigned long)hv_cpu->synic_event_page);
		free_page((unsigned long)hv_cpu->synic_message_page);
		free_page((unsigned long)hv_cpu->post_msg_page);
	}

	kfree(hv_context.hv_numa_map);
}

/*
 * hv_synic_init - Initialize the Synthetic 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)
{
	struct hv_per_cpu_context *hv_cpu
		= per_cpu_ptr(hv_context.cpu_context, cpu);
	union hv_synic_simp simp;
	union hv_synic_siefp siefp;
	union hv_synic_sint shared_sint;
	union hv_synic_scontrol sctrl;

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

	hv_set_simp(simp.as_uint64);

	/* Setup the Synic's event page */
	hv_get_siefp(siefp.as_uint64);
	siefp.siefp_enabled = 1;
	siefp.base_siefp_gpa = virt_to_phys(hv_cpu->synic_event_page)
		>> PAGE_SHIFT;

	hv_set_siefp(siefp.as_uint64);

	/* Setup the shared SINT. */
	hv_get_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
	shared_sint.masked = false;
	if (ms_hyperv.hints & HV_DEPRECATING_AEOI_RECOMMENDED)
		shared_sint.auto_eoi = false;
	else
		shared_sint.auto_eoi = true;

	hv_set_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

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

	hv_set_synic_state(sctrl.as_uint64);

	/*
	 * Register the per-cpu clockevent source.
	 */
	if (ms_hyperv.features & HV_MSR_SYNTIMER_AVAILABLE)
		clockevents_config_and_register(hv_cpu->clk_evt,
						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_MSR_SYNTIMER_AVAILABLE))
		return;

	if (direct_mode_enabled)
		hv_remove_stimer0_irq(stimer0_irq);

	for_each_present_cpu(cpu) {
		struct hv_per_cpu_context *hv_cpu
			= per_cpu_ptr(hv_context.cpu_context, cpu);

		clockevents_unbind_device(hv_cpu->clk_evt, 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;

	hv_get_synic_state(sctrl.as_uint64);
	if (sctrl.enable != 1)
		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_MSR_SYNTIMER_AVAILABLE) {
		struct hv_per_cpu_context *hv_cpu
			= this_cpu_ptr(hv_context.cpu_context);

		clockevents_unbind_device(hv_cpu->clk_evt, cpu);
		hv_ce_shutdown(hv_cpu->clk_evt);
	}

	hv_get_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);

	shared_sint.masked = 1;

	/* Need to correctly cleanup in the case of SMP!!! */
	/* Disable the interrupt */
	hv_set_synint_state(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);

	hv_get_siefp(siefp.as_uint64);
	siefp.siefp_enabled = 0;
	siefp.base_siefp_gpa = 0;

	hv_set_siefp(siefp.as_uint64);

	/* Disable the global synic bit */
	sctrl.enable = 0;
	hv_set_synic_state(sctrl.as_uint64);

	return 0;
}
Commit	Line	Data
3b20eb23	1	// SPDX-License-Identifier: GPL-2.0-only
3e7ee490	2	/*
3e7ee490 HJ	3	* Copyright (c) 2009, Microsoft Corporation.
3e7ee490 HJ	4	*
3e7ee490 HJ	5	* Authors:
	6	* Haiyang Zhang <haiyangz@microsoft.com>
	7	* Hank Janssen <hjanssen@microsoft.com>
3e7ee490	8	*/
0a46618d HJ	9	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
0a46618d HJ	10
a0086dc5 GKH	11	#include <linux/kernel.h>
a0086dc5 GKH	12	#include <linux/mm.h>
5a0e3ad6	13	#include <linux/slab.h>
b7c947f0	14	#include <linux/vmalloc.h>
46a97191	15	#include <linux/hyperv.h>
83ba0c4f	16	#include <linux/version.h>
248e742a	17	#include <linux/random.h>
4061ed9e	18	#include <linux/clockchips.h>
4061ed9e	19	#include <asm/mshyperv.h>
0f2a6619	20	#include "hyperv_vmbus.h"
3e7ee490	21
454f18a9	22	/* The one and only */
a3cadf38	23	struct hv_context hv_context;
3e7ee490	24
248e742a MK	25	/*
	26	* If false, we're using the old mechanism for stimer0 interrupts
	27	* where it sends a VMbus message when it expires. The old
	28	* mechanism is used when running on older versions of Hyper-V
	29	* that don't support Direct Mode. While Hyper-V provides
	30	* four stimer's per CPU, Linux uses only stimer0.
	31	*/
	32	static bool direct_mode_enabled;
	33	static int stimer0_irq;
	34	static int stimer0_vector;
	35
4061ed9e S	36	#define HV_TIMER_FREQUENCY (10 * 1000 * 1000) /* 100ns period */
	37	#define HV_MAX_MAX_DELTA_TICKS 0xffffffff
	38	#define HV_MIN_DELTA_TICKS 1
	39
3e189519	40	/*
d44890c8	41	* hv_init - Main initialization routine.
0831ad04 GKH	42	*
	43	* This routine must be called before any other routines in here are called
	44	*/
d44890c8	45	int hv_init(void)
3e7ee490	46	{
37cdd991 SH	47	hv_context.cpu_context = alloc_percpu(struct hv_per_cpu_context);
	48	if (!hv_context.cpu_context)
	49	return -ENOMEM;
	50
248e742a	51	direct_mode_enabled = ms_hyperv.misc_features &
7dc9b6b8	52	HV_STIMER_DIRECT_MODE_AVAILABLE;
5433e003	53	return 0;
3e7ee490 HJ	54	}
3e7ee490 HJ	55
3e189519	56	/*
d44890c8	57	* hv_post_message - Post a message using the hypervisor message IPC.
0831ad04 GKH	58	*
	59	* This involves a hypercall.
	60	*/
415f0a02	61	int hv_post_message(union hv_connection_id connection_id,
b8dfb264 HZ	62	enum hv_message_type message_type,
b8dfb264 HZ	63	void *payload, size_t payload_size)
3e7ee490	64	{
b8dfb264	65	struct hv_input_post_message *aligned_msg;
37cdd991	66	struct hv_per_cpu_context *hv_cpu;
a108393d	67	u64 status;
3e7ee490	68
b8dfb264	69	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
39594abc	70	return -EMSGSIZE;
3e7ee490	71
37cdd991 SH	72	hv_cpu = get_cpu_ptr(hv_context.cpu_context);
37cdd991 SH	73	aligned_msg = hv_cpu->post_msg_page;
b8dfb264	74	aligned_msg->connectionid = connection_id;
b29ef354	75	aligned_msg->reserved = 0;
b8dfb264 HZ	76	aligned_msg->message_type = message_type;
	77	aligned_msg->payload_size = payload_size;
	78	memcpy((void *)aligned_msg->payload, payload, payload_size);
3e7ee490	79
a108393d	80	status = hv_do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL);
3e7ee490	81
13b9abfc MK	82	/* Preemption must remain disabled until after the hypercall
	83	* so some other thread can't get scheduled onto this cpu and
	84	* corrupt the per-cpu post_msg_page
	85	*/
	86	put_cpu_ptr(hv_cpu);
	87
a108393d	88	return status & 0xFFFF;
3e7ee490 HJ	89	}
3e7ee490 HJ	90
248e742a MK	91	/*
	92	* ISR for when stimer0 is operating in Direct Mode. Direct Mode
	93	* does not use VMbus or any VMbus messages, so process here and not
	94	* in the VMbus driver code.
	95	*/
	96
	97	static void hv_stimer0_isr(void)
	98	{
	99	struct hv_per_cpu_context *hv_cpu;
	100
	101	hv_cpu = this_cpu_ptr(hv_context.cpu_context);
	102	hv_cpu->clk_evt->event_handler(hv_cpu->clk_evt);
	103	add_interrupt_randomness(stimer0_vector, 0);
	104	}
	105
4061ed9e S	106	static int hv_ce_set_next_event(unsigned long delta,
	107	struct clock_event_device *evt)
	108	{
a5a1d1c2	109	u64 current_tick;
4061ed9e	110
bc609cb4	111	WARN_ON(!clockevent_state_oneshot(evt));
4061ed9e	112
e546d778	113	current_tick = hyperv_cs->read(NULL);
4061ed9e	114	current_tick += delta;
619a4c8b	115	hv_init_timer(0, current_tick);
4061ed9e S	116	return 0;
	117	}
	118
bc609cb4 VK	119	static int hv_ce_shutdown(struct clock_event_device *evt)
bc609cb4 VK	120	{
619a4c8b MK	121	hv_init_timer(0, 0);
619a4c8b MK	122	hv_init_timer_config(0, 0);
248e742a MK	123	if (direct_mode_enabled)
248e742a MK	124	hv_disable_stimer0_percpu_irq(stimer0_irq);
bc609cb4 VK	125
	126	return 0;
	127	}
	128
	129	static int hv_ce_set_oneshot(struct clock_event_device *evt)
4061ed9e	130	{
0aa67255	131	union hv_stimer_config timer_cfg;
4061ed9e	132
248e742a	133	timer_cfg.as_uint64 = 0;
bc609cb4 VK	134	timer_cfg.enable = 1;
bc609cb4 VK	135	timer_cfg.auto_enable = 1;
248e742a MK	136	if (direct_mode_enabled) {
	137	/*
	138	* When it expires, the timer will directly interrupt
	139	* on the specified hardware vector/IRQ.
	140	*/
	141	timer_cfg.direct_mode = 1;
	142	timer_cfg.apic_vector = stimer0_vector;
	143	hv_enable_stimer0_percpu_irq(stimer0_irq);
	144	} else {
	145	/*
	146	* When it expires, the timer will generate a VMbus message,
	147	* to be handled by the normal VMbus interrupt handler.
	148	*/
	149	timer_cfg.direct_mode = 0;
	150	timer_cfg.sintx = VMBUS_MESSAGE_SINT;
	151	}
619a4c8b	152	hv_init_timer_config(0, timer_cfg.as_uint64);
bc609cb4	153	return 0;
4061ed9e S	154	}
	155
	156	static void hv_init_clockevent_device(struct clock_event_device *dev, int cpu)
	157	{
	158	dev->name = "Hyper-V clockevent";
	159	dev->features = CLOCK_EVT_FEAT_ONESHOT;
	160	dev->cpumask = cpumask_of(cpu);
	161	dev->rating = 1000;
e086748c VK	162	/*
	163	* Avoid settint dev->owner = THIS_MODULE deliberately as doing so will
	164	* result in clockevents_config_and_register() taking additional
	165	* references to the hv_vmbus module making it impossible to unload.
	166	*/
4061ed9e	167
bc609cb4 VK	168	dev->set_state_shutdown = hv_ce_shutdown;
bc609cb4 VK	169	dev->set_state_oneshot = hv_ce_set_oneshot;
4061ed9e S	170	dev->set_next_event = hv_ce_set_next_event;
	171	}
	172
2608fb65 JW	173
	174	int hv_synic_alloc(void)
	175	{
2608fb65	176	int cpu;
f25a7ece MK	177	struct hv_per_cpu_context *hv_cpu;
	178
	179	/*
	180	* First, zero all per-cpu memory areas so hv_synic_free() can
	181	* detect what memory has been allocated and cleanup properly
	182	* after any failures.
	183	*/
	184	for_each_present_cpu(cpu) {
	185	hv_cpu = per_cpu_ptr(hv_context.cpu_context, cpu);
	186	memset(hv_cpu, 0, sizeof(*hv_cpu));
	187	}
2608fb65	188
6396bb22	189	hv_context.hv_numa_map = kcalloc(nr_node_ids, sizeof(struct cpumask),
597ff72f	190	GFP_KERNEL);
9f01ec53 S	191	if (hv_context.hv_numa_map == NULL) {
	192	pr_err("Unable to allocate NUMA map\n");
	193	goto err;
	194	}
	195
421b8f20	196	for_each_present_cpu(cpu) {
f25a7ece	197	hv_cpu = per_cpu_ptr(hv_context.cpu_context, cpu);
37cdd991	198
37cdd991 SH	199	tasklet_init(&hv_cpu->msg_dpc,
	200	vmbus_on_msg_dpc, (unsigned long) hv_cpu);
	201
	202	hv_cpu->clk_evt = kzalloc(sizeof(struct clock_event_device),
	203	GFP_KERNEL);
	204	if (hv_cpu->clk_evt == NULL) {
4061ed9e S	205	pr_err("Unable to allocate clock event device\n");
	206	goto err;
	207	}
37cdd991	208	hv_init_clockevent_device(hv_cpu->clk_evt, cpu);
9f01ec53	209
37cdd991	210	hv_cpu->synic_message_page =
2608fb65	211	(void *)get_zeroed_page(GFP_ATOMIC);
37cdd991	212	if (hv_cpu->synic_message_page == NULL) {
2608fb65 JW	213	pr_err("Unable to allocate SYNIC message page\n");
	214	goto err;
	215	}
	216
37cdd991 SH	217	hv_cpu->synic_event_page = (void *)get_zeroed_page(GFP_ATOMIC);
37cdd991 SH	218	if (hv_cpu->synic_event_page == NULL) {
2608fb65 JW	219	pr_err("Unable to allocate SYNIC event page\n");
	220	goto err;
	221	}
b29ef354	222
37cdd991 SH	223	hv_cpu->post_msg_page = (void *)get_zeroed_page(GFP_ATOMIC);
37cdd991 SH	224	if (hv_cpu->post_msg_page == NULL) {
b29ef354 S	225	pr_err("Unable to allocate post msg page\n");
	226	goto err;
	227	}
3c7630d3	228
37cdd991	229	INIT_LIST_HEAD(&hv_cpu->chan_list);
2608fb65 JW	230	}
2608fb65 JW	231
248e742a MK	232	if (direct_mode_enabled &&
	233	hv_setup_stimer0_irq(&stimer0_irq, &stimer0_vector,
	234	hv_stimer0_isr))
	235	goto err;
	236
2608fb65 JW	237	return 0;
2608fb65 JW	238	err:
57208632 MK	239	/*
	240	* Any memory allocations that succeeded will be freed when
	241	* the caller cleans up by calling hv_synic_free()
	242	*/
2608fb65 JW	243	return -ENOMEM;
	244	}
	245
2608fb65 JW	246
	247	void hv_synic_free(void)
	248	{
	249	int cpu;
	250
37cdd991 SH	251	for_each_present_cpu(cpu) {
	252	struct hv_per_cpu_context *hv_cpu
	253	= per_cpu_ptr(hv_context.cpu_context, cpu);
	254
57208632 MK	255	kfree(hv_cpu->clk_evt);
	256	free_page((unsigned long)hv_cpu->synic_event_page);
	257	free_page((unsigned long)hv_cpu->synic_message_page);
	258	free_page((unsigned long)hv_cpu->post_msg_page);
37cdd991 SH	259	}
37cdd991 SH	260
9f01ec53	261	kfree(hv_context.hv_numa_map);
2608fb65 JW	262	}
2608fb65 JW	263
3e189519	264	/*
68cb8117	265	* hv_synic_init - Initialize the Synthetic Interrupt Controller.
0831ad04 GKH	266	*
	267	* If it is already initialized by another entity (ie x2v shim), we need to
	268	* retrieve the initialized message and event pages. Otherwise, we create and
	269	* initialize the message and event pages.
	270	*/
76d36ab7	271	int hv_synic_init(unsigned int cpu)
3e7ee490	272	{
37cdd991 SH	273	struct hv_per_cpu_context *hv_cpu
37cdd991 SH	274	= per_cpu_ptr(hv_context.cpu_context, cpu);
eacb1b4d GKH	275	union hv_synic_simp simp;
eacb1b4d GKH	276	union hv_synic_siefp siefp;
b8dfb264	277	union hv_synic_sint shared_sint;
eacb1b4d	278	union hv_synic_scontrol sctrl;
a73e6b7c	279
a73e6b7c	280	/* Setup the Synic's message page */
155e4a2f	281	hv_get_simp(simp.as_uint64);
f6feebe0	282	simp.simp_enabled = 1;
37cdd991	283	simp.base_simp_gpa = virt_to_phys(hv_cpu->synic_message_page)
a73e6b7c	284	>> PAGE_SHIFT;
3e7ee490	285
155e4a2f	286	hv_set_simp(simp.as_uint64);
3e7ee490	287
a73e6b7c	288	/* Setup the Synic's event page */
8e307bf8	289	hv_get_siefp(siefp.as_uint64);
f6feebe0	290	siefp.siefp_enabled = 1;
37cdd991	291	siefp.base_siefp_gpa = virt_to_phys(hv_cpu->synic_event_page)
a73e6b7c HJ	292	>> PAGE_SHIFT;
a73e6b7c HJ	293
8e307bf8	294	hv_set_siefp(siefp.as_uint64);
0831ad04	295
0831ad04	296	/* Setup the shared SINT. */
619a4c8b	297	hv_get_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	298
302a3c0f	299	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
b8dfb264	300	shared_sint.masked = false;
7dc9b6b8	301	if (ms_hyperv.hints & HV_DEPRECATING_AEOI_RECOMMENDED)
6c248aad S	302	shared_sint.auto_eoi = false;
	303	else
	304	shared_sint.auto_eoi = true;
3e7ee490	305
619a4c8b	306	hv_set_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	307
454f18a9	308	/* Enable the global synic bit */
06d1d98a	309	hv_get_synic_state(sctrl.as_uint64);
f6feebe0	310	sctrl.enable = 1;
3e7ee490	311
06d1d98a	312	hv_set_synic_state(sctrl.as_uint64);
3e7ee490	313
4061ed9e S	314	/*
	315	* Register the per-cpu clockevent source.
	316	*/
7dc9b6b8	317	if (ms_hyperv.features & HV_MSR_SYNTIMER_AVAILABLE)
37cdd991	318	clockevents_config_and_register(hv_cpu->clk_evt,
4061ed9e S	319	HV_TIMER_FREQUENCY,
	320	HV_MIN_DELTA_TICKS,
	321	HV_MAX_MAX_DELTA_TICKS);
76d36ab7	322	return 0;
3e7ee490 HJ	323	}
3e7ee490 HJ	324
e086748c VK	325	/*
	326	* hv_synic_clockevents_cleanup - Cleanup clockevent devices
	327	*/
	328	void hv_synic_clockevents_cleanup(void)
	329	{
	330	int cpu;
	331
7dc9b6b8	332	if (!(ms_hyperv.features & HV_MSR_SYNTIMER_AVAILABLE))
e086748c VK	333	return;
e086748c VK	334
248e742a MK	335	if (direct_mode_enabled)
	336	hv_remove_stimer0_irq(stimer0_irq);
	337
37cdd991 SH	338	for_each_present_cpu(cpu) {
	339	struct hv_per_cpu_context *hv_cpu
	340	= per_cpu_ptr(hv_context.cpu_context, cpu);
	341
	342	clockevents_unbind_device(hv_cpu->clk_evt, cpu);
	343	}
e086748c VK	344	}
e086748c VK	345
3e189519	346	/*
d44890c8	347	* hv_synic_cleanup - Cleanup routine for hv_synic_init().
0831ad04	348	*/
76d36ab7	349	int hv_synic_cleanup(unsigned int cpu)
3e7ee490	350	{
b8dfb264	351	union hv_synic_sint shared_sint;
eacb1b4d GKH	352	union hv_synic_simp simp;
eacb1b4d GKH	353	union hv_synic_siefp siefp;
e72e7ac5	354	union hv_synic_scontrol sctrl;
523b9408 VK	355	struct vmbus_channel channel, sc;
	356	bool channel_found = false;
	357	unsigned long flags;
3e7ee490	358
a3cadf38 S	359	hv_get_synic_state(sctrl.as_uint64);
a3cadf38 S	360	if (sctrl.enable != 1)
76d36ab7	361	return -EFAULT;
3e7ee490	362
523b9408 VK	363	/*
	364	* Search for channels which are bound to the CPU we're about to
	365	* cleanup. In case we find one and vmbus is still connected we need to
	366	* fail, this will effectively prevent CPU offlining. There is no way
	367	* we can re-bind channels to different CPUs for now.
	368	*/
	369	mutex_lock(&vmbus_connection.channel_mutex);
	370	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
	371	if (channel->target_cpu == cpu) {
	372	channel_found = true;
	373	break;
	374	}
	375	spin_lock_irqsave(&channel->lock, flags);
	376	list_for_each_entry(sc, &channel->sc_list, sc_list) {
	377	if (sc->target_cpu == cpu) {
	378	channel_found = true;
	379	break;
	380	}
	381	}
	382	spin_unlock_irqrestore(&channel->lock, flags);
	383	if (channel_found)
	384	break;
	385	}
	386	mutex_unlock(&vmbus_connection.channel_mutex);
	387
	388	if (channel_found && vmbus_connection.conn_state == CONNECTED)
	389	return -EBUSY;
	390
e086748c	391	/* Turn off clockevent device */
7dc9b6b8	392	if (ms_hyperv.features & HV_MSR_SYNTIMER_AVAILABLE) {
37cdd991 SH	393	struct hv_per_cpu_context *hv_cpu
	394	= this_cpu_ptr(hv_context.cpu_context);
	395
	396	clockevents_unbind_device(hv_cpu->clk_evt, cpu);
	397	hv_ce_shutdown(hv_cpu->clk_evt);
6ffc4b85	398	}
e086748c	399
619a4c8b	400	hv_get_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	401
b8dfb264	402	shared_sint.masked = 1;
3e7ee490	403
7692fd4d	404	/* Need to correctly cleanup in the case of SMP!!! */
454f18a9	405	/* Disable the interrupt */
619a4c8b	406	hv_set_synint_state(VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	407
155e4a2f	408	hv_get_simp(simp.as_uint64);
f6feebe0 HZ	409	simp.simp_enabled = 0;
f6feebe0 HZ	410	simp.base_simp_gpa = 0;
3e7ee490	411
155e4a2f	412	hv_set_simp(simp.as_uint64);
3e7ee490	413
8e307bf8	414	hv_get_siefp(siefp.as_uint64);
f6feebe0 HZ	415	siefp.siefp_enabled = 0;
f6feebe0 HZ	416	siefp.base_siefp_gpa = 0;
3e7ee490	417
8e307bf8	418	hv_set_siefp(siefp.as_uint64);
3e7ee490	419
e72e7ac5	420	/* Disable the global synic bit */
e72e7ac5	421	sctrl.enable = 0;
06d1d98a	422	hv_set_synic_state(sctrl.as_uint64);
76d36ab7 VK	423
76d36ab7 VK	424	return 0;
3e7ee490	425	}