[deliverable/linux.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)
{
#ifdef CONFIG_X86_64
	u64 hv_status = 0;
	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;

	__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;
	u64 input_address = (input) ? virt_to_phys(input) : 0;
	u32 input_address_hi = input_address >> 32;
	u32 input_address_lo = input_address & 0xFFFFFFFF;
	u64 output_address = (output) ? virt_to_phys(output) : 0;
	u32 output_address_hi = output_address >> 32;
	u32 output_address_lo = output_address & 0xFFFFFFFF;
	void *hypercall_page = hv_context.hypercall_page;

	__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;

	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;

	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);

	hv_synic_free_cpu(cpu);
}
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	{
530cf207	96	#ifdef CONFIG_X86_64
b8dfb264 HZ	97	u64 hv_status = 0;
	98	u64 input_address = (input) ? virt_to_phys(input) : 0;
	99	u64 output_address = (output) ? virt_to_phys(output) : 0;
dec317fd	100	void *hypercall_page = hv_context.hypercall_page;
3e7ee490	101
b8dfb264 HZ	102	__asm__ __volatile__("mov %0, %%r8" : : "r" (output_address) : "r8");
	103	__asm__ __volatile__("call *%3" : "=a" (hv_status) :
	104	"c" (control), "d" (input_address),
	105	"m" (hypercall_page));
3e7ee490	106
b8dfb264	107	return hv_status;
3e7ee490 HJ	108
	109	#else
	110
b8dfb264 HZ	111	u32 control_hi = control >> 32;
	112	u32 control_lo = control & 0xFFFFFFFF;
	113	u32 hv_status_hi = 1;
	114	u32 hv_status_lo = 1;
	115	u64 input_address = (input) ? virt_to_phys(input) : 0;
	116	u32 input_address_hi = input_address >> 32;
	117	u32 input_address_lo = input_address & 0xFFFFFFFF;
	118	u64 output_address = (output) ? virt_to_phys(output) : 0;
	119	u32 output_address_hi = output_address >> 32;
	120	u32 output_address_lo = output_address & 0xFFFFFFFF;
dec317fd	121	void *hypercall_page = hv_context.hypercall_page;
3e7ee490	122
b8dfb264 HZ	123	__asm__ __volatile__ ("call *%8" : "=d"(hv_status_hi),
	124	"=a"(hv_status_lo) : "d" (control_hi),
	125	"a" (control_lo), "b" (input_address_hi),
	126	"c" (input_address_lo), "D"(output_address_hi),
	127	"S"(output_address_lo), "m" (hypercall_page));
3e7ee490	128
b8dfb264	129	return hv_status_lo \| ((u64)hv_status_hi << 32);
0831ad04	130	#endif /* !x86_64 */
3e7ee490 HJ	131	}
3e7ee490 HJ	132
3e189519	133	/*
d44890c8	134	* hv_init - Main initialization routine.
0831ad04 GKH	135	*
	136	* This routine must be called before any other routines in here are called
	137	*/
d44890c8	138	int hv_init(void)
3e7ee490	139	{
b8dfb264 HZ	140	int max_leaf;
	141	union hv_x64_msr_hypercall_contents hypercall_msr;
	142	void *virtaddr = NULL;
3e7ee490	143
14c1bf8a	144	memset(hv_context.synic_event_page, 0, sizeof(void ) NR_CPUS);
6a0aaa18	145	memset(hv_context.synic_message_page, 0,
14c1bf8a	146	sizeof(void ) NR_CPUS);
b29ef354 S	147	memset(hv_context.post_msg_page, 0,
b29ef354 S	148	sizeof(void ) NR_CPUS);
917ea427 S	149	memset(hv_context.vp_index, 0,
917ea427 S	150	sizeof(int) * NR_CPUS);
db11f12a S	151	memset(hv_context.event_dpc, 0,
db11f12a S	152	sizeof(void ) NR_CPUS);
4061ed9e S	153	memset(hv_context.clk_evt, 0,
4061ed9e S	154	sizeof(void ) NR_CPUS);
3e7ee490	155
d44890c8	156	max_leaf = query_hypervisor_info();
3e7ee490	157
83ba0c4f S	158	/*
	159	* Write our OS ID.
	160	*/
	161	hv_context.guestid = generate_guest_id(0, LINUX_VERSION_CODE, 0);
	162	wrmsrl(HV_X64_MSR_GUEST_OS_ID, hv_context.guestid);
a73e6b7c	163
454f18a9	164	/* See if the hypercall page is already set */
b8dfb264	165	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
3e7ee490	166
df3493e0	167	virtaddr = __vmalloc(PAGE_SIZE, GFP_KERNEL, PAGE_KERNEL_EXEC);
3e7ee490	168
98e08702	169	if (!virtaddr)
44939d37	170	goto cleanup;
3e7ee490	171
b8dfb264	172	hypercall_msr.enable = 1;
a73e6b7c	173
b8dfb264 HZ	174	hypercall_msr.guest_physical_address = vmalloc_to_pfn(virtaddr);
b8dfb264 HZ	175	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
a73e6b7c HJ	176
a73e6b7c HJ	177	/* Confirm that hypercall page did get setup. */
b8dfb264 HZ	178	hypercall_msr.as_uint64 = 0;
b8dfb264 HZ	179	rdmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
a73e6b7c	180
98e08702	181	if (!hypercall_msr.enable)
44939d37	182	goto cleanup;
3e7ee490	183
b8dfb264	184	hv_context.hypercall_page = virtaddr;
a73e6b7c	185
5433e003	186	return 0;
3e7ee490	187
44939d37	188	cleanup:
b8dfb264 HZ	189	if (virtaddr) {
	190	if (hypercall_msr.enable) {
	191	hypercall_msr.as_uint64 = 0;
	192	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
3e7ee490 HJ	193	}
3e7ee490 HJ	194
b8dfb264	195	vfree(virtaddr);
3e7ee490	196	}
5433e003 S	197
5433e003 S	198	return -ENOTSUPP;
3e7ee490 HJ	199	}
3e7ee490 HJ	200
3e189519	201	/*
d44890c8	202	* hv_cleanup - Cleanup routine.
0831ad04 GKH	203	*
	204	* This routine is called normally during driver unloading or exiting.
	205	*/
d44890c8	206	void hv_cleanup(void)
3e7ee490	207	{
b8dfb264	208	union hv_x64_msr_hypercall_contents hypercall_msr;
3e7ee490	209
93e5bd06 S	210	/* Reset our OS id */
	211	wrmsrl(HV_X64_MSR_GUEST_OS_ID, 0);
	212
6a0aaa18	213	if (hv_context.hypercall_page) {
b8dfb264 HZ	214	hypercall_msr.as_uint64 = 0;
b8dfb264 HZ	215	wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
6a0aaa18 HZ	216	vfree(hv_context.hypercall_page);
6a0aaa18 HZ	217	hv_context.hypercall_page = NULL;
3e7ee490	218	}
3e7ee490 HJ	219	}
3e7ee490 HJ	220
3e189519	221	/*
d44890c8	222	* hv_post_message - Post a message using the hypervisor message IPC.
0831ad04 GKH	223	*
	224	* This involves a hypercall.
	225	*/
415f0a02	226	int hv_post_message(union hv_connection_id connection_id,
b8dfb264 HZ	227	enum hv_message_type message_type,
b8dfb264 HZ	228	void *payload, size_t payload_size)
3e7ee490	229	{
3e7ee490	230
b8dfb264	231	struct hv_input_post_message *aligned_msg;
034469e6	232	u16 status;
3e7ee490	233
b8dfb264	234	if (payload_size > HV_MESSAGE_PAYLOAD_BYTE_COUNT)
39594abc	235	return -EMSGSIZE;
3e7ee490	236
b8dfb264	237	aligned_msg = (struct hv_input_post_message *)
b29ef354	238	hv_context.post_msg_page[get_cpu()];
3e7ee490	239
b8dfb264	240	aligned_msg->connectionid = connection_id;
b29ef354	241	aligned_msg->reserved = 0;
b8dfb264 HZ	242	aligned_msg->message_type = message_type;
	243	aligned_msg->payload_size = payload_size;
	244	memcpy((void *)aligned_msg->payload, payload, payload_size);
3e7ee490	245
d44890c8 HZ	246	status = do_hypercall(HVCALL_POST_MESSAGE, aligned_msg, NULL)
d44890c8 HZ	247	& 0xFFFF;
3e7ee490	248
b29ef354	249	put_cpu();
3e7ee490 HJ	250	return status;
	251	}
	252
	253
3e189519	254	/*
d44890c8 HZ	255	* hv_signal_event -
d44890c8 HZ	256	* Signal an event on the specified connection using the hypervisor event IPC.
0831ad04 GKH	257	*
	258	* This involves a hypercall.
	259	*/
1f42248d	260	u16 hv_signal_event(void *con_id)
3e7ee490	261	{
034469e6	262	u16 status;
3e7ee490	263
1f42248d S	264	status = (do_hypercall(HVCALL_SIGNAL_EVENT, con_id, NULL) & 0xFFFF);
1f42248d S	265
3e7ee490 HJ	266	return status;
	267	}
	268
4061ed9e S	269	static int hv_ce_set_next_event(unsigned long delta,
	270	struct clock_event_device *evt)
	271	{
	272	cycle_t current_tick;
	273
	274	WARN_ON(evt->mode != CLOCK_EVT_MODE_ONESHOT);
	275
	276	rdmsrl(HV_X64_MSR_TIME_REF_COUNT, current_tick);
	277	current_tick += delta;
	278	wrmsrl(HV_X64_MSR_STIMER0_COUNT, current_tick);
	279	return 0;
	280	}
	281
	282	static void hv_ce_setmode(enum clock_event_mode mode,
	283	struct clock_event_device *evt)
	284	{
	285	union hv_timer_config timer_cfg;
	286
	287	switch (mode) {
	288	case CLOCK_EVT_MODE_PERIODIC:
	289	/* unsupported */
	290	break;
	291
	292	case CLOCK_EVT_MODE_ONESHOT:
	293	timer_cfg.enable = 1;
	294	timer_cfg.auto_enable = 1;
	295	timer_cfg.sintx = VMBUS_MESSAGE_SINT;
	296	wrmsrl(HV_X64_MSR_STIMER0_CONFIG, timer_cfg.as_uint64);
	297	break;
	298
	299	case CLOCK_EVT_MODE_UNUSED:
	300	case CLOCK_EVT_MODE_SHUTDOWN:
	301	wrmsrl(HV_X64_MSR_STIMER0_COUNT, 0);
	302	wrmsrl(HV_X64_MSR_STIMER0_CONFIG, 0);
	303	break;
	304	case CLOCK_EVT_MODE_RESUME:
	305	break;
	306	}
	307	}
	308
	309	static void hv_init_clockevent_device(struct clock_event_device *dev, int cpu)
	310	{
	311	dev->name = "Hyper-V clockevent";
	312	dev->features = CLOCK_EVT_FEAT_ONESHOT;
	313	dev->cpumask = cpumask_of(cpu);
	314	dev->rating = 1000;
e086748c VK	315	/*
	316	* Avoid settint dev->owner = THIS_MODULE deliberately as doing so will
	317	* result in clockevents_config_and_register() taking additional
	318	* references to the hv_vmbus module making it impossible to unload.
	319	*/
4061ed9e S	320
	321	dev->set_mode = hv_ce_setmode;
	322	dev->set_next_event = hv_ce_set_next_event;
	323	}
	324
2608fb65 JW	325
	326	int hv_synic_alloc(void)
	327	{
	328	size_t size = sizeof(struct tasklet_struct);
4061ed9e	329	size_t ced_size = sizeof(struct clock_event_device);
2608fb65 JW	330	int cpu;
	331
	332	for_each_online_cpu(cpu) {
	333	hv_context.event_dpc[cpu] = kmalloc(size, GFP_ATOMIC);
	334	if (hv_context.event_dpc[cpu] == NULL) {
	335	pr_err("Unable to allocate event dpc\n");
	336	goto err;
	337	}
	338	tasklet_init(hv_context.event_dpc[cpu], vmbus_on_event, cpu);
	339
4061ed9e S	340	hv_context.clk_evt[cpu] = kzalloc(ced_size, GFP_ATOMIC);
	341	if (hv_context.clk_evt[cpu] == NULL) {
	342	pr_err("Unable to allocate clock event device\n");
	343	goto err;
	344	}
	345	hv_init_clockevent_device(hv_context.clk_evt[cpu], cpu);
	346
2608fb65 JW	347	hv_context.synic_message_page[cpu] =
	348	(void *)get_zeroed_page(GFP_ATOMIC);
	349
	350	if (hv_context.synic_message_page[cpu] == NULL) {
	351	pr_err("Unable to allocate SYNIC message page\n");
	352	goto err;
	353	}
	354
	355	hv_context.synic_event_page[cpu] =
	356	(void *)get_zeroed_page(GFP_ATOMIC);
	357
	358	if (hv_context.synic_event_page[cpu] == NULL) {
	359	pr_err("Unable to allocate SYNIC event page\n");
	360	goto err;
	361	}
b29ef354 S	362
	363	hv_context.post_msg_page[cpu] =
	364	(void *)get_zeroed_page(GFP_ATOMIC);
	365
	366	if (hv_context.post_msg_page[cpu] == NULL) {
	367	pr_err("Unable to allocate post msg page\n");
	368	goto err;
	369	}
2608fb65 JW	370	}
	371
	372	return 0;
	373	err:
	374	return -ENOMEM;
	375	}
	376
8712954d	377	static void hv_synic_free_cpu(int cpu)
2608fb65 JW	378	{
2608fb65 JW	379	kfree(hv_context.event_dpc[cpu]);
4061ed9e	380	kfree(hv_context.clk_evt[cpu]);
fdf91dae	381	if (hv_context.synic_event_page[cpu])
2608fb65 JW	382	free_page((unsigned long)hv_context.synic_event_page[cpu]);
	383	if (hv_context.synic_message_page[cpu])
	384	free_page((unsigned long)hv_context.synic_message_page[cpu]);
b29ef354 S	385	if (hv_context.post_msg_page[cpu])
b29ef354 S	386	free_page((unsigned long)hv_context.post_msg_page[cpu]);
2608fb65 JW	387	}
	388
	389	void hv_synic_free(void)
	390	{
	391	int cpu;
	392
	393	for_each_online_cpu(cpu)
	394	hv_synic_free_cpu(cpu);
	395	}
	396
3e189519	397	/*
d44890c8	398	* hv_synic_init - Initialize the Synthethic Interrupt Controller.
0831ad04 GKH	399	*
	400	* If it is already initialized by another entity (ie x2v shim), we need to
	401	* retrieve the initialized message and event pages. Otherwise, we create and
	402	* initialize the message and event pages.
	403	*/
302a3c0f	404	void hv_synic_init(void *arg)
3e7ee490	405	{
0831ad04	406	u64 version;
eacb1b4d GKH	407	union hv_synic_simp simp;
eacb1b4d GKH	408	union hv_synic_siefp siefp;
b8dfb264	409	union hv_synic_sint shared_sint;
eacb1b4d	410	union hv_synic_scontrol sctrl;
917ea427	411	u64 vp_index;
a73e6b7c	412
7692fd4d	413	int cpu = smp_processor_id();
3e7ee490	414
6a0aaa18	415	if (!hv_context.hypercall_page)
7692fd4d	416	return;
3e7ee490	417
454f18a9	418	/* Check the version */
a51ed7d6	419	rdmsrl(HV_X64_MSR_SVERSION, version);
3e7ee490	420
a73e6b7c	421	/* Setup the Synic's message page */
f6feebe0 HZ	422	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
f6feebe0 HZ	423	simp.simp_enabled = 1;
6a0aaa18	424	simp.base_simp_gpa = virt_to_phys(hv_context.synic_message_page[cpu])
a73e6b7c	425	>> PAGE_SHIFT;
3e7ee490	426
f6feebe0	427	wrmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
3e7ee490	428
a73e6b7c	429	/* Setup the Synic's event page */
f6feebe0 HZ	430	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
f6feebe0 HZ	431	siefp.siefp_enabled = 1;
6a0aaa18	432	siefp.base_siefp_gpa = virt_to_phys(hv_context.synic_event_page[cpu])
a73e6b7c HJ	433	>> PAGE_SHIFT;
a73e6b7c HJ	434
f6feebe0	435	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
0831ad04	436
0831ad04	437	/* Setup the shared SINT. */
b8dfb264	438	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	439
b8dfb264	440	shared_sint.as_uint64 = 0;
302a3c0f	441	shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR;
b8dfb264	442	shared_sint.masked = false;
b0209501	443	shared_sint.auto_eoi = true;
3e7ee490	444
b8dfb264	445	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	446
454f18a9	447	/* Enable the global synic bit */
f6feebe0 HZ	448	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
f6feebe0 HZ	449	sctrl.enable = 1;
3e7ee490	450
f6feebe0	451	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
3e7ee490	452
6a0aaa18	453	hv_context.synic_initialized = true;
917ea427 S	454
	455	/*
	456	* Setup the mapping between Hyper-V's notion
	457	* of cpuid and Linux' notion of cpuid.
	458	* This array will be indexed using Linux cpuid.
	459	*/
	460	rdmsrl(HV_X64_MSR_VP_INDEX, vp_index);
	461	hv_context.vp_index[cpu] = (u32)vp_index;
3a28fa35 S	462
3a28fa35 S	463	INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);
4061ed9e S	464
	465	/*
	466	* Register the per-cpu clockevent source.
	467	*/
	468	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
	469	clockevents_config_and_register(hv_context.clk_evt[cpu],
	470	HV_TIMER_FREQUENCY,
	471	HV_MIN_DELTA_TICKS,
	472	HV_MAX_MAX_DELTA_TICKS);
7692fd4d	473	return;
3e7ee490 HJ	474	}
3e7ee490 HJ	475
e086748c VK	476	/*
	477	* hv_synic_clockevents_cleanup - Cleanup clockevent devices
	478	*/
	479	void hv_synic_clockevents_cleanup(void)
	480	{
	481	int cpu;
	482
	483	if (!(ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE))
	484	return;
	485
	486	for_each_online_cpu(cpu)
	487	clockevents_unbind_device(hv_context.clk_evt[cpu], cpu);
	488	}
	489
3e189519	490	/*
d44890c8	491	* hv_synic_cleanup - Cleanup routine for hv_synic_init().
0831ad04	492	*/
d44890c8	493	void hv_synic_cleanup(void *arg)
3e7ee490	494	{
b8dfb264	495	union hv_synic_sint shared_sint;
eacb1b4d GKH	496	union hv_synic_simp simp;
eacb1b4d GKH	497	union hv_synic_siefp siefp;
e72e7ac5	498	union hv_synic_scontrol sctrl;
7692fd4d	499	int cpu = smp_processor_id();
3e7ee490	500
6a0aaa18	501	if (!hv_context.synic_initialized)
3e7ee490	502	return;
3e7ee490	503
e086748c VK	504	/* Turn off clockevent device */
	505	if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE)
	506	hv_ce_setmode(CLOCK_EVT_MODE_SHUTDOWN,
	507	hv_context.clk_evt[cpu]);
	508
b8dfb264	509	rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	510
b8dfb264	511	shared_sint.masked = 1;
3e7ee490	512
7692fd4d	513	/* Need to correctly cleanup in the case of SMP!!! */
454f18a9	514	/* Disable the interrupt */
b8dfb264	515	wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64);
3e7ee490	516
f6feebe0 HZ	517	rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
	518	simp.simp_enabled = 0;
	519	simp.base_simp_gpa = 0;
3e7ee490	520
f6feebe0	521	wrmsrl(HV_X64_MSR_SIMP, simp.as_uint64);
3e7ee490	522
f6feebe0 HZ	523	rdmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
	524	siefp.siefp_enabled = 0;
	525	siefp.base_siefp_gpa = 0;
3e7ee490	526
f6feebe0	527	wrmsrl(HV_X64_MSR_SIEFP, siefp.as_uint64);
3e7ee490	528
e72e7ac5 VK	529	/* Disable the global synic bit */
	530	rdmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	531	sctrl.enable = 0;
	532	wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64);
	533
	534	hv_synic_free_cpu(cpu);
3e7ee490	535	}