[linux-2.6-block.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,
	.hypercall_page		= NULL,
};

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

/*
 * query_hypervisor_info - Get version info of the windows hypervisor
 */
unsigned int host_info_eax;
unsigned int host_info_ebx;
unsigned int host_info_ecx;
unsigned int host_info_edx;

static int query_hypervisor_info(void)
{
	unsigned int eax;
	unsigned int ebx;
	unsigned int ecx;
	unsigned int edx;
	unsigned int max_leaf;
	unsigned int op;

	/*
	* Its assumed that this is called after confirming that Viridian
	* is present. Query id and revision.
	*/
	eax = 0;
	ebx = 0;
	ecx = 0;
	edx = 0;
	op = HVCPUID_VENDOR_MAXFUNCTION;
	cpuid(op, &eax, &ebx, &ecx, &edx);

	max_leaf = eax;

	if (max_leaf >= HVCPUID_VERSION) {
		eax = 0;
		ebx = 0;
		ecx = 0;
		edx = 0;
		op = HVCPUID_VERSION;
		cpuid(op, &eax, &ebx, &ecx, &edx);
		host_info_eax = eax;
		host_info_ebx = ebx;
		host_info_ecx = ecx;
		host_info_edx = edx;
	}
	return max_leaf;
}

/*
 * do_hypercall- Invoke the specified hypercall
 */
static u64 do_hypercall(u64 control, void *input, void *output)
{
	u64 input_address = (input) ? virt_to_phys(input) : 0;
	u64 output_address = (output) ? virt_to_phys(output) : 0;
	void *hypercall_page = hv_context.hypercall_page;
#ifdef CONFIG_X86_64
	u64 hv_status = 0;

	if (!hypercall_page)
		return (u64)ULLONG_MAX;

	__asm__ __volatile__("mov %0, %%r8" : : "r" (output_address) : "r8");
	__asm__ __volatile__("call *%3" : "=a" (hv_status) :
			     "c" (control), "d" (input_address),
			     "m" (hypercall_page));

	return hv_status;

#else

	u32 control_hi = control >> 32;
	u32 control_lo = control & 0xFFFFFFFF;
	u32 hv_status_hi = 1;
	u32 hv_status_lo = 1;
	u32 input_address_hi = input_address >> 32;
	u32 input_address_lo = input_address & 0xFFFFFFFF;
	u32 output_address_hi = output_address >> 32;
	u32 output_address_lo = output_address & 0xFFFFFFFF;

	if (!hypercall_page)
		return (u64)ULLONG_MAX;

	__asm__ __volatile__ ("call *%8" : "=d"(hv_status_hi),
			      "=a"(hv_status_lo) : "d" (control_hi),
			      "a" (control_lo), "b" (input_address_hi),
			      "c" (input_address_lo), "D"(output_address_hi),
			      "S"(output_address_lo), "m" (hypercall_page));

	return hv_status_lo | ((u64)hv_status_hi << 32);
#endif /* !x86_64 */
}

/*
 * hv_init - Main initialization routine.
 *
 * This routine must be called before any other routines in here are called
 */
int hv_init(void)
{
	int max_leaf;
	union hv_x64_msr_hypercall_contents hypercall_msr;
	void *virtaddr = NULL;

	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.clk_evt, 0,
	       sizeof(void *) * NR_CPUS);

	max_leaf = query_hypervisor_info();

	/*
	 * Write our OS ID.
	 */
	hv_context.guestid = generate_guest_id(0, LINUX_VERSION_CODE, 0);
	wrmsrl(HV_X64_MSR_GUEST_OS_ID, hv_context.guestid);

	/* See if the hypercall page is already set */
	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);

	virtaddr = __vmalloc(PAGE_SIZE, GFP_KERNEL, PAGE_KERNEL_EXEC);

	if (!virtaddr)
		goto cleanup;

	hypercall_msr.enable = 1;

	hypercall_msr.guest_physical_address = vmalloc_to_pfn(virtaddr);
	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);

	/* Confirm that hypercall page did get setup. */
	hypercall_msr.as_uint64 = 0;
	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);

	if (!hypercall_msr.enable)
		goto cleanup;

	hv_context.hypercall_page = virtaddr;

	return 0;

cleanup:
	if (virtaddr) {
		if (hypercall_msr.enable) {
			hypercall_msr.as_uint64 = 0;
			wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
		}

		vfree(virtaddr);
	}

	return -ENOTSUPP;
}

/*
 * hv_cleanup - Cleanup routine.
 *
 * This routine is called normally during driver unloading or exiting.
 */
void hv_cleanup(void)
{
	union hv_x64_msr_hypercall_contents hypercall_msr;

	/* Reset our OS id */
	wrmsrl(HV_X64_MSR_GUEST_OS_ID, 0);

	if (hv_context.hypercall_page) {
		hypercall_msr.as_uint64 = 0;
		wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
		vfree(hv_context.hypercall_page);
		hv_context.hypercall_page = NULL;
	}
}

/*
 * 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;
	u16 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 = do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL)
		& 0xFFFF;

	put_cpu();
	return status;
}


/*
 * hv_signal_event -
 * Signal an event on the specified connection using the hypervisor event IPC.
 *
 * This involves a hypercall.
 */
u16 hv_signal_event(void *con_id)
{
	u16 status;

	status = (do_hypercall(HVCALL_SIGNAL_EVENT, con_id, NULL) & 0xFFFF);

	return status;
}

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

	WARN_ON(evt->mode != CLOCK_EVT_MODE_ONESHOT);

	rdmsrl(HV_X64_MSR_TIME_REF_COUNT, current_tick);
	current_tick += delta;
	wrmsrl(HV_X64_MSR_STIMER0_COUNT, current_tick);
	return 0;
}

static void hv_ce_setmode(enum clock_event_mode mode,
			  struct clock_event_device *evt)
{
	union hv_timer_config timer_cfg;

	switch (mode) {
	case CLOCK_EVT_MODE_PERIODIC:
		/* unsupported */
		break;

	case CLOCK_EVT_MODE_ONESHOT:
		timer_cfg.enable = 1;
		timer_cfg.auto_enable = 1;
		timer_cfg.sintx = VMBUS_MESSAGE_SINT;
		wrmsrl(HV_X64_MSR_STIMER0_CONFIG, timer_cfg.as_uint64);
		break;

	case CLOCK_EVT_MODE_UNUSED:
	case CLOCK_EVT_MODE_SHUTDOWN:
		wrmsrl(HV_X64_MSR_STIMER0_COUNT, 0);
		wrmsrl(HV_X64_MSR_STIMER0_CONFIG, 0);
		break;
	case CLOCK_EVT_MODE_RESUME:
		break;
	}
}

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_mode = hv_ce_setmode;
	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_online_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.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;
		}
	}

	return 0;
err:
	return -ENOMEM;
}

static void hv_synic_free_cpu(int cpu)
{
	kfree(hv_context.event_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_online_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.
 */
void hv_synic_init(void *arg)
{
	u64 version;
	union hv_synic_simp simp;
	union hv_synic_siefp siefp;
	union hv_synic_sint shared_sint;
	union hv_synic_scontrol sctrl;
	u64 vp_index;

	int cpu = smp_processor_id();

	if (!hv_context.hypercall_page)
		return;

	/* Check the version */
	rdmsrl(HV_X64_MSR_SVERSION, version);

	/* Setup the Synic's message page */
	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
	simp.simp_enabled = 1;
	simp.base_simp_gpa = virt_to_phys(hv_context.synic_message_page[cpu])
		>> PAGE_SHIFT;

	wrmsrl(HV_X64_MSR_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;

	INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);

	/*
	 * 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;
}

/*
 * 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_online_cpu(cpu)
		clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
}

/*
 * hv_synic_cleanup - Cleanup routine for hv_synic_init().
 */
void hv_synic_cleanup(void *arg)
{
	union hv_synic_sint shared_sint;
	union hv_synic_simp simp;
	union hv_synic_siefp siefp;
	union hv_synic_scontrol sctrl;
	int cpu = smp_processor_id();

	if (!hv_context.synic_initialized)
		return;

	/* Turn off clockevent device */
	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
		hv_ce_setmode(CLOCK_EVT_MODE_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);

	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
	simp.simp_enabled = 0;
	simp.base_simp_gpa = 0;

	wrmsrl(HV_X64_MSR_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);
}
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 = {
	38	.synic_initialized = false,
	39	.hypercall_page = NULL,
3e7ee490 HJ	40	};
3e7ee490 HJ	41
4061ed9e S	42	#define HV_TIMER_FREQUENCY (10 * 1000 * 1000) /* 100ns period */
	43	#define HV_MAX_MAX_DELTA_TICKS 0xffffffff
	44	#define HV_MIN_DELTA_TICKS 1
	45
3e189519	46	/*
d44890c8	47	* query_hypervisor_info - Get version info of the windows hypervisor
0831ad04	48	*/
5fbebb2d S	49	unsigned int host_info_eax;
	50	unsigned int host_info_ebx;
	51	unsigned int host_info_ecx;
	52	unsigned int host_info_edx;
	53
d44890c8	54	static int query_hypervisor_info(void)
0831ad04 GKH	55	{
	56	unsigned int eax;
	57	unsigned int ebx;
	58	unsigned int ecx;
	59	unsigned int edx;
b8dfb264	60	unsigned int max_leaf;
0831ad04	61	unsigned int op;
3e7ee490	62
0831ad04 GKH	63	/*
	64	* Its assumed that this is called after confirming that Viridian
	65	* is present. Query id and revision.
	66	*/
	67	eax = 0;
	68	ebx = 0;
	69	ecx = 0;
	70	edx = 0;
f6feebe0	71	op = HVCPUID_VENDOR_MAXFUNCTION;
0831ad04	72	cpuid(op, &eax, &ebx, &ecx, &edx);
3e7ee490	73
b8dfb264	74	max_leaf = eax;
0831ad04	75
b8dfb264	76	if (max_leaf >= HVCPUID_VERSION) {
0831ad04 GKH	77	eax = 0;
	78	ebx = 0;
	79	ecx = 0;
	80	edx = 0;
f6feebe0	81	op = HVCPUID_VERSION;
0831ad04	82	cpuid(op, &eax, &ebx, &ecx, &edx);
5fbebb2d S	83	host_info_eax = eax;
	84	host_info_ebx = ebx;
	85	host_info_ecx = ecx;
	86	host_info_edx = edx;
0831ad04	87	}
b8dfb264	88	return max_leaf;
0831ad04	89	}
3e7ee490	90
3e189519	91	/*
d44890c8	92	* do_hypercall- Invoke the specified hypercall
0831ad04	93	*/
d44890c8	94	static u64 do_hypercall(u64 control, void input, void output)
3e7ee490	95	{
b8dfb264 HZ	96	u64 input_address = (input) ? virt_to_phys(input) : 0;
b8dfb264 HZ	97	u64 output_address = (output) ? virt_to_phys(output) : 0;
dec317fd	98	void *hypercall_page = hv_context.hypercall_page;
d7646eaa VK	99	#ifdef CONFIG_X86_64
	100	u64 hv_status = 0;
	101
	102	if (!hypercall_page)
	103	return (u64)ULLONG_MAX;
3e7ee490	104
b8dfb264 HZ	105	__asm__ __volatile__("mov %0, %%r8" : : "r" (output_address) : "r8");
	106	__asm__ __volatile__("call *%3" : "=a" (hv_status) :
	107	"c" (control), "d" (input_address),
	108	"m" (hypercall_page));
3e7ee490	109
b8dfb264	110	return hv_status;
3e7ee490 HJ	111
	112	#else
	113
b8dfb264 HZ	114	u32 control_hi = control >> 32;
	115	u32 control_lo = control & 0xFFFFFFFF;
	116	u32 hv_status_hi = 1;
	117	u32 hv_status_lo = 1;
b8dfb264 HZ	118	u32 input_address_hi = input_address >> 32;
b8dfb264 HZ	119	u32 input_address_lo = input_address & 0xFFFFFFFF;
b8dfb264 HZ	120	u32 output_address_hi = output_address >> 32;
b8dfb264 HZ	121	u32 output_address_lo = output_address & 0xFFFFFFFF;
d7646eaa VK	122
	123	if (!hypercall_page)
	124	return (u64)ULLONG_MAX;
3e7ee490	125
b8dfb264 HZ	126	__asm__ __volatile__ ("call *%8" : "=d"(hv_status_hi),
	127	"=a"(hv_status_lo) : "d" (control_hi),
	128	"a" (control_lo), "b" (input_address_hi),
	129	"c" (input_address_lo), "D"(output_address_hi),
	130	"S"(output_address_lo), "m" (hypercall_page));
3e7ee490	131
b8dfb264	132	return hv_status_lo \| ((u64)hv_status_hi << 32);
0831ad04	133	#endif /* !x86_64 */
3e7ee490 HJ	134	}
3e7ee490 HJ	135
3e189519	136	/*
d44890c8	137	* hv_init - Main initialization routine.
0831ad04 GKH	138	*
	139	* This routine must be called before any other routines in here are called
	140	*/
d44890c8	141	int hv_init(void)
3e7ee490	142	{
b8dfb264 HZ	143	int max_leaf;
	144	union hv_x64_msr_hypercall_contents hypercall_msr;
	145	void *virtaddr = NULL;
3e7ee490	146
14c1bf8a	147	memset(hv_context.synic_event_page, 0, sizeof(void ) NR_CPUS);
6a0aaa18	148	memset(hv_context.synic_message_page, 0,
14c1bf8a	149	sizeof(void ) NR_CPUS);
b29ef354 S	150	memset(hv_context.post_msg_page, 0,
b29ef354 S	151	sizeof(void ) NR_CPUS);
917ea427 S	152	memset(hv_context.vp_index, 0,
917ea427 S	153	sizeof(int) * NR_CPUS);
db11f12a S	154	memset(hv_context.event_dpc, 0,
db11f12a S	155	sizeof(void ) NR_CPUS);
4061ed9e S	156	memset(hv_context.clk_evt, 0,
4061ed9e S	157	sizeof(void ) NR_CPUS);
3e7ee490	158
d44890c8	159	max_leaf = query_hypervisor_info();
3e7ee490	160
83ba0c4f S	161	/*
	162	* Write our OS ID.
	163	*/
	164	hv_context.guestid = generate_guest_id(0, LINUX_VERSION_CODE, 0);
	165	wrmsrl(HV_X64_MSR_GUEST_OS_ID, hv_context.guestid);
a73e6b7c	166
454f18a9	167	/* See if the hypercall page is already set */
b8dfb264	168	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
3e7ee490	169
df3493e0	170	virtaddr = __vmalloc(PAGE_SIZE, GFP_KERNEL, PAGE_KERNEL_EXEC);
3e7ee490	171
98e08702	172	if (!virtaddr)
44939d37	173	goto cleanup;
3e7ee490	174
b8dfb264	175	hypercall_msr.enable = 1;
a73e6b7c	176
b8dfb264 HZ	177	hypercall_msr.guest_physical_address = vmalloc_to_pfn(virtaddr);
b8dfb264 HZ	178	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
a73e6b7c HJ	179
a73e6b7c HJ	180	/* Confirm that hypercall page did get setup. */
b8dfb264 HZ	181	hypercall_msr.as_uint64 = 0;
b8dfb264 HZ	182	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
a73e6b7c	183
98e08702	184	if (!hypercall_msr.enable)
44939d37	185	goto cleanup;
3e7ee490	186
b8dfb264	187	hv_context.hypercall_page = virtaddr;
a73e6b7c	188
5433e003	189	return 0;
3e7ee490	190
44939d37	191	cleanup:
b8dfb264 HZ	192	if (virtaddr) {
	193	if (hypercall_msr.enable) {
	194	hypercall_msr.as_uint64 = 0;
	195	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
3e7ee490 HJ	196	}
3e7ee490 HJ	197
b8dfb264	198	vfree(virtaddr);
3e7ee490	199	}
5433e003 S	200
5433e003 S	201	return -ENOTSUPP;
3e7ee490 HJ	202	}
3e7ee490 HJ	203
3e189519	204	/*
d44890c8	205	* hv_cleanup - Cleanup routine.
0831ad04 GKH	206	*
	207	* This routine is called normally during driver unloading or exiting.
	208	*/
d44890c8	209	void hv_cleanup(void)
3e7ee490	210	{
b8dfb264	211	union hv_x64_msr_hypercall_contents hypercall_msr;
3e7ee490	212
93e5bd06 S	213	/* Reset our OS id */
	214	wrmsrl(HV_X64_MSR_GUEST_OS_ID, 0);
	215
6a0aaa18	216	if (hv_context.hypercall_page) {
b8dfb264 HZ	217	hypercall_msr.as_uint64 = 0;
b8dfb264 HZ	218	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
6a0aaa18 HZ	219	vfree(hv_context.hypercall_page);
6a0aaa18 HZ	220	hv_context.hypercall_page = NULL;
3e7ee490	221	}
3e7ee490 HJ	222	}
3e7ee490 HJ	223
3e189519	224	/*
d44890c8	225	* hv_post_message - Post a message using the hypervisor message IPC.
0831ad04 GKH	226	*
	227	* This involves a hypercall.
	228	*/
415f0a02	229	int hv_post_message(union hv_connection_id connection_id,
b8dfb264 HZ	230	enum hv_message_type message_type,
b8dfb264 HZ	231	void *payload, size_t payload_size)
3e7ee490	232	{
3e7ee490	233
b8dfb264	234	struct hv_input_post_message *aligned_msg;
034469e6	235	u16 status;
3e7ee490	236
b8dfb264	237	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
39594abc	238	return -EMSGSIZE;
3e7ee490	239
b8dfb264	240	aligned_msg = (struct hv_input_post_message *)
b29ef354	241	hv_context.post_msg_page[get_cpu()];
3e7ee490	242
b8dfb264	243	aligned_msg->connectionid = connection_id;
b29ef354	244	aligned_msg->reserved = 0;
b8dfb264 HZ	245	aligned_msg->message_type = message_type;
	246	aligned_msg->payload_size = payload_size;
	247	memcpy((void *)aligned_msg->payload, payload, payload_size);
3e7ee490	248
d44890c8 HZ	249	status = do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL)
d44890c8 HZ	250	& 0xFFFF;
3e7ee490	251
b29ef354	252	put_cpu();
3e7ee490 HJ	253	return status;
	254	}
	255
	256
3e189519	257	/*
d44890c8 HZ	258	* hv_signal_event -
d44890c8 HZ	259	* Signal an event on the specified connection using the hypervisor event IPC.
0831ad04 GKH	260	*
	261	* This involves a hypercall.
	262	*/
1f42248d	263	u16 hv_signal_event(void *con_id)
3e7ee490	264	{
034469e6	265	u16 status;
3e7ee490	266
1f42248d S	267	status = (do_hypercall(HVCALL_SIGNAL_EVENT, con_id, NULL) & 0xFFFF);
1f42248d S	268
3e7ee490 HJ	269	return status;
	270	}
	271
4061ed9e S	272	static int hv_ce_set_next_event(unsigned long delta,
	273	struct clock_event_device *evt)
	274	{
	275	cycle_t current_tick;
	276
	277	WARN_ON(evt->mode != CLOCK_EVT_MODE_ONESHOT);
	278
	279	rdmsrl(HV_X64_MSR_TIME_REF_COUNT, current_tick);
	280	current_tick += delta;
	281	wrmsrl(HV_X64_MSR_STIMER0_COUNT, current_tick);
	282	return 0;
	283	}
	284
	285	static void hv_ce_setmode(enum clock_event_mode mode,
	286	struct clock_event_device *evt)
	287	{
	288	union hv_timer_config timer_cfg;
	289
	290	switch (mode) {
	291	case CLOCK_EVT_MODE_PERIODIC:
	292	/* unsupported */
	293	break;
	294
	295	case CLOCK_EVT_MODE_ONESHOT:
	296	timer_cfg.enable = 1;
	297	timer_cfg.auto_enable = 1;
	298	timer_cfg.sintx = VMBUS_MESSAGE_SINT;
	299	wrmsrl(HV_X64_MSR_STIMER0_CONFIG, timer_cfg.as_uint64);
	300	break;
	301
	302	case CLOCK_EVT_MODE_UNUSED:
	303	case CLOCK_EVT_MODE_SHUTDOWN:
	304	wrmsrl(HV_X64_MSR_STIMER0_COUNT, 0);
	305	wrmsrl(HV_X64_MSR_STIMER0_CONFIG, 0);
	306	break;
	307	case CLOCK_EVT_MODE_RESUME:
	308	break;
	309	}
	310	}
	311
	312	static void hv_init_clockevent_device(struct clock_event_device *dev, int cpu)
	313	{
	314	dev->name = "Hyper-V clockevent";
	315	dev->features = CLOCK_EVT_FEAT_ONESHOT;
	316	dev->cpumask = cpumask_of(cpu);
	317	dev->rating = 1000;
e086748c VK	318	/*
	319	* Avoid settint dev->owner = THIS_MODULE deliberately as doing so will
	320	* result in clockevents_config_and_register() taking additional
	321	* references to the hv_vmbus module making it impossible to unload.
	322	*/
4061ed9e S	323
	324	dev->set_mode = hv_ce_setmode;
	325	dev->set_next_event = hv_ce_set_next_event;
	326	}
	327
2608fb65 JW	328
	329	int hv_synic_alloc(void)
	330	{
	331	size_t size = sizeof(struct tasklet_struct);
4061ed9e	332	size_t ced_size = sizeof(struct clock_event_device);
2608fb65 JW	333	int cpu;
2608fb65 JW	334
9f01ec53 S	335	hv_context.hv_numa_map = kzalloc(sizeof(struct cpumask) * nr_node_ids,
	336	GFP_ATOMIC);
	337	if (hv_context.hv_numa_map == NULL) {
	338	pr_err("Unable to allocate NUMA map\n");
	339	goto err;
	340	}
	341
2608fb65 JW	342	for_each_online_cpu(cpu) {
	343	hv_context.event_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
	344	if (hv_context.event_dpc[cpu] == NULL) {
	345	pr_err("Unable to allocate event dpc\n");
	346	goto err;
	347	}
	348	tasklet_init(hv_context.event_dpc[cpu], vmbus_on_event, cpu);
	349
4061ed9e S	350	hv_context.clk_evt[cpu] = kzalloc(ced_size, GFP_ATOMIC);
	351	if (hv_context.clk_evt[cpu] == NULL) {
	352	pr_err("Unable to allocate clock event device\n");
	353	goto err;
	354	}
9f01ec53	355
4061ed9e S	356	hv_init_clockevent_device(hv_context.clk_evt[cpu], cpu);
4061ed9e S	357
2608fb65 JW	358	hv_context.synic_message_page[cpu] =
	359	(void *)get_zeroed_page(GFP_ATOMIC);
	360
	361	if (hv_context.synic_message_page[cpu] == NULL) {
	362	pr_err("Unable to allocate SYNIC message page\n");
	363	goto err;
	364	}
	365
	366	hv_context.synic_event_page[cpu] =
	367	(void *)get_zeroed_page(GFP_ATOMIC);
	368
	369	if (hv_context.synic_event_page[cpu] == NULL) {
	370	pr_err("Unable to allocate SYNIC event page\n");
	371	goto err;
	372	}
b29ef354 S	373
	374	hv_context.post_msg_page[cpu] =
	375	(void *)get_zeroed_page(GFP_ATOMIC);
	376
	377	if (hv_context.post_msg_page[cpu] == NULL) {
	378	pr_err("Unable to allocate post msg page\n");
	379	goto err;
	380	}
2608fb65 JW	381	}
	382
	383	return 0;
	384	err:
	385	return -ENOMEM;
	386	}
	387
8712954d	388	static void hv_synic_free_cpu(int cpu)
2608fb65 JW	389	{
2608fb65 JW	390	kfree(hv_context.event_dpc[cpu]);
4061ed9e	391	kfree(hv_context.clk_evt[cpu]);
fdf91dae	392	if (hv_context.synic_event_page[cpu])
2608fb65 JW	393	free_page((unsigned long)hv_context.synic_event_page[cpu]);
	394	if (hv_context.synic_message_page[cpu])
	395	free_page((unsigned long)hv_context.synic_message_page[cpu]);
b29ef354 S	396	if (hv_context.post_msg_page[cpu])
b29ef354 S	397	free_page((unsigned long)hv_context.post_msg_page[cpu]);
2608fb65 JW	398	}
	399
	400	void hv_synic_free(void)
	401	{
	402	int cpu;
	403
9f01ec53	404	kfree(hv_context.hv_numa_map);
2608fb65 JW	405	for_each_online_cpu(cpu)
	406	hv_synic_free_cpu(cpu);
	407	}
	408
3e189519	409	/*
d44890c8	410	* hv_synic_init - Initialize the Synthethic Interrupt Controller.
0831ad04 GKH	411	*
	412	* If it is already initialized by another entity (ie x2v shim), we need to
	413	* retrieve the initialized message and event pages. Otherwise, we create and
	414	* initialize the message and event pages.
	415	*/
302a3c0f	416	void hv_synic_init(void *arg)
3e7ee490	417	{
0831ad04	418	u64 version;
eacb1b4d GKH	419	union hv_synic_simp simp;
eacb1b4d GKH	420	union hv_synic_siefp siefp;
b8dfb264	421	union hv_synic_sint shared_sint;
eacb1b4d	422	union hv_synic_scontrol sctrl;
917ea427	423	u64 vp_index;
a73e6b7c	424
7692fd4d	425	int cpu = smp_processor_id();
3e7ee490	426
6a0aaa18	427	if (!hv_context.hypercall_page)
7692fd4d	428	return;
3e7ee490	429
454f18a9	430	/* Check the version */
a51ed7d6	431	rdmsrl(HV_X64_MSR_SVERSION, version);
3e7ee490	432
a73e6b7c	433	/* Setup the Synic's message page */
f6feebe0 HZ	434	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
f6feebe0 HZ	435	simp.simp_enabled = 1;
6a0aaa18	436	simp.base_simp_gpa = virt_to_phys(hv_context.synic_message_page[cpu])
a73e6b7c	437	>> PAGE_SHIFT;
3e7ee490	438
f6feebe0	439	wrmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
3e7ee490	440
a73e6b7c	441	/* Setup the Synic's event page */
f6feebe0 HZ	442	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
f6feebe0 HZ	443	siefp.siefp_enabled = 1;
6a0aaa18	444	siefp.base_siefp_gpa = virt_to_phys(hv_context.synic_event_page[cpu])
a73e6b7c HJ	445	>> PAGE_SHIFT;
a73e6b7c HJ	446
f6feebe0	447	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
0831ad04	448
0831ad04	449	/* Setup the shared SINT. */
b8dfb264	450	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	451
b8dfb264	452	shared_sint.as_uint64 = 0;
302a3c0f	453	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
b8dfb264	454	shared_sint.masked = false;
b0209501	455	shared_sint.auto_eoi = true;
3e7ee490	456
b8dfb264	457	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	458
454f18a9	459	/* Enable the global synic bit */
f6feebe0 HZ	460	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
f6feebe0 HZ	461	sctrl.enable = 1;
3e7ee490	462
f6feebe0	463	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
3e7ee490	464
6a0aaa18	465	hv_context.synic_initialized = true;
917ea427 S	466
	467	/*
	468	* Setup the mapping between Hyper-V's notion
	469	* of cpuid and Linux' notion of cpuid.
	470	* This array will be indexed using Linux cpuid.
	471	*/
	472	rdmsrl(HV_X64_MSR_VP_INDEX, vp_index);
	473	hv_context.vp_index[cpu] = (u32)vp_index;
3a28fa35 S	474
3a28fa35 S	475	INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);
4061ed9e S	476
	477	/*
	478	* Register the per-cpu clockevent source.
	479	*/
	480	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
	481	clockevents_config_and_register(hv_context.clk_evt[cpu],
	482	HV_TIMER_FREQUENCY,
	483	HV_MIN_DELTA_TICKS,
	484	HV_MAX_MAX_DELTA_TICKS);
7692fd4d	485	return;
3e7ee490 HJ	486	}
3e7ee490 HJ	487
e086748c VK	488	/*
	489	* hv_synic_clockevents_cleanup - Cleanup clockevent devices
	490	*/
	491	void hv_synic_clockevents_cleanup(void)
	492	{
	493	int cpu;
	494
	495	if (!(ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE))
	496	return;
	497
	498	for_each_online_cpu(cpu)
	499	clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
	500	}
	501
3e189519	502	/*
d44890c8	503	* hv_synic_cleanup - Cleanup routine for hv_synic_init().
0831ad04	504	*/
d44890c8	505	void hv_synic_cleanup(void *arg)
3e7ee490	506	{
b8dfb264	507	union hv_synic_sint shared_sint;
eacb1b4d GKH	508	union hv_synic_simp simp;
eacb1b4d GKH	509	union hv_synic_siefp siefp;
e72e7ac5	510	union hv_synic_scontrol sctrl;
7692fd4d	511	int cpu = smp_processor_id();
3e7ee490	512
6a0aaa18	513	if (!hv_context.synic_initialized)
3e7ee490	514	return;
3e7ee490	515
e086748c VK	516	/* Turn off clockevent device */
	517	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
	518	hv_ce_setmode(CLOCK_EVT_MODE_SHUTDOWN,
	519	hv_context.clk_evt[cpu]);
	520
b8dfb264	521	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	522
b8dfb264	523	shared_sint.masked = 1;
3e7ee490	524
7692fd4d	525	/* Need to correctly cleanup in the case of SMP!!! */
454f18a9	526	/* Disable the interrupt */
b8dfb264	527	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	528
f6feebe0 HZ	529	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
	530	simp.simp_enabled = 0;
	531	simp.base_simp_gpa = 0;
3e7ee490	532
f6feebe0	533	wrmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
3e7ee490	534
f6feebe0 HZ	535	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
	536	siefp.siefp_enabled = 0;
	537	siefp.base_siefp_gpa = 0;
3e7ee490	538
f6feebe0	539	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
3e7ee490	540
e72e7ac5 VK	541	/* Disable the global synic bit */
	542	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	543	sctrl.enable = 0;
	544	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
3e7ee490	545	}