vmbus_drv.c 26.2 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
/*
 * 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>
20
 *   K. Y. Srinivasan <kys@microsoft.com>
21
 *
22
 */
23 24
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

25 26 27 28 29
#include <linux/init.h>
#include <linux/module.h>
#include <linux/device.h>
#include <linux/interrupt.h>
#include <linux/sysctl.h>
30
#include <linux/slab.h>
31
#include <linux/acpi.h>
32
#include <linux/completion.h>
33
#include <linux/hyperv.h>
34
#include <linux/kernel_stat.h>
35
#include <linux/clockchips.h>
36
#include <linux/cpu.h>
37
#include <asm/hyperv.h>
38
#include <asm/hypervisor.h>
39
#include <asm/mshyperv.h>
40
#include "hyperv_vmbus.h"
41

42
static struct acpi_device  *hv_acpi_dev;
43

44
static struct tasklet_struct msg_dpc;
45
static struct completion probe_event;
46
static int irq;
47

48 49 50 51 52
struct resource hyperv_mmio = {
	.name  = "hyperv mmio",
	.flags = IORESOURCE_MEM,
};
EXPORT_SYMBOL_GPL(hyperv_mmio);
53

54 55 56 57 58 59 60 61
static int vmbus_exists(void)
{
	if (hv_acpi_dev == NULL)
		return -ENODEV;

	return 0;
}

62 63 64 65 66 67 68 69
#define VMBUS_ALIAS_LEN ((sizeof((struct hv_vmbus_device_id *)0)->guid) * 2)
static void print_alias_name(struct hv_device *hv_dev, char *alias_name)
{
	int i;
	for (i = 0; i < VMBUS_ALIAS_LEN; i += 2)
		sprintf(&alias_name[i], "%02x", hv_dev->dev_type.b[i/2]);
}

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
static u8 channel_monitor_group(struct vmbus_channel *channel)
{
	return (u8)channel->offermsg.monitorid / 32;
}

static u8 channel_monitor_offset(struct vmbus_channel *channel)
{
	return (u8)channel->offermsg.monitorid % 32;
}

static u32 channel_pending(struct vmbus_channel *channel,
			   struct hv_monitor_page *monitor_page)
{
	u8 monitor_group = channel_monitor_group(channel);
	return monitor_page->trigger_group[monitor_group].pending;
}

87 88 89 90 91 92 93 94
static u32 channel_latency(struct vmbus_channel *channel,
			   struct hv_monitor_page *monitor_page)
{
	u8 monitor_group = channel_monitor_group(channel);
	u8 monitor_offset = channel_monitor_offset(channel);
	return monitor_page->latency[monitor_group][monitor_offset];
}

95 96 97 98 99 100 101 102
static u32 channel_conn_id(struct vmbus_channel *channel,
			   struct hv_monitor_page *monitor_page)
{
	u8 monitor_group = channel_monitor_group(channel);
	u8 monitor_offset = channel_monitor_offset(channel);
	return monitor_page->parameter[monitor_group][monitor_offset].connectionid.u.id;
}

103 104 105 106 107 108 109 110 111 112 113
static ssize_t id_show(struct device *dev, struct device_attribute *dev_attr,
		       char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n", hv_dev->channel->offermsg.child_relid);
}
static DEVICE_ATTR_RO(id);

114 115 116 117 118 119 120 121 122 123 124
static ssize_t state_show(struct device *dev, struct device_attribute *dev_attr,
			  char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n", hv_dev->channel->state);
}
static DEVICE_ATTR_RO(state);

125 126 127 128 129 130 131 132 133 134 135
static ssize_t monitor_id_show(struct device *dev,
			       struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n", hv_dev->channel->offermsg.monitorid);
}
static DEVICE_ATTR_RO(monitor_id);

136 137 138 139 140 141 142 143 144 145 146 147
static ssize_t class_id_show(struct device *dev,
			       struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "{%pUl}\n",
		       hv_dev->channel->offermsg.offer.if_type.b);
}
static DEVICE_ATTR_RO(class_id);

148 149 150 151 152 153 154 155 156 157 158 159
static ssize_t device_id_show(struct device *dev,
			      struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "{%pUl}\n",
		       hv_dev->channel->offermsg.offer.if_instance.b);
}
static DEVICE_ATTR_RO(device_id);

160 161 162 163 164 165 166 167 168 169 170
static ssize_t modalias_show(struct device *dev,
			     struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	char alias_name[VMBUS_ALIAS_LEN + 1];

	print_alias_name(hv_dev, alias_name);
	return sprintf(buf, "vmbus:%s\n", alias_name);
}
static DEVICE_ATTR_RO(modalias);

171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
static ssize_t server_monitor_pending_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_pending(hv_dev->channel,
				       vmbus_connection.monitor_pages[1]));
}
static DEVICE_ATTR_RO(server_monitor_pending);

static ssize_t client_monitor_pending_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_pending(hv_dev->channel,
				       vmbus_connection.monitor_pages[1]));
}
static DEVICE_ATTR_RO(client_monitor_pending);
198

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
static ssize_t server_monitor_latency_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_latency(hv_dev->channel,
				       vmbus_connection.monitor_pages[0]));
}
static DEVICE_ATTR_RO(server_monitor_latency);

static ssize_t client_monitor_latency_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_latency(hv_dev->channel,
				       vmbus_connection.monitor_pages[1]));
}
static DEVICE_ATTR_RO(client_monitor_latency);

227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
static ssize_t server_monitor_conn_id_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_conn_id(hv_dev->channel,
				       vmbus_connection.monitor_pages[0]));
}
static DEVICE_ATTR_RO(server_monitor_conn_id);

static ssize_t client_monitor_conn_id_show(struct device *dev,
					   struct device_attribute *dev_attr,
					   char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);

	if (!hv_dev->channel)
		return -ENODEV;
	return sprintf(buf, "%d\n",
		       channel_conn_id(hv_dev->channel,
				       vmbus_connection.monitor_pages[1]));
}
static DEVICE_ATTR_RO(client_monitor_conn_id);

255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
static ssize_t out_intr_mask_show(struct device *dev,
				  struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info outbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
	return sprintf(buf, "%d\n", outbound.current_interrupt_mask);
}
static DEVICE_ATTR_RO(out_intr_mask);

static ssize_t out_read_index_show(struct device *dev,
				   struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info outbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
	return sprintf(buf, "%d\n", outbound.current_read_index);
}
static DEVICE_ATTR_RO(out_read_index);

static ssize_t out_write_index_show(struct device *dev,
				    struct device_attribute *dev_attr,
				    char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info outbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
	return sprintf(buf, "%d\n", outbound.current_write_index);
}
static DEVICE_ATTR_RO(out_write_index);

static ssize_t out_read_bytes_avail_show(struct device *dev,
					 struct device_attribute *dev_attr,
					 char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info outbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
	return sprintf(buf, "%d\n", outbound.bytes_avail_toread);
}
static DEVICE_ATTR_RO(out_read_bytes_avail);

static ssize_t out_write_bytes_avail_show(struct device *dev,
					  struct device_attribute *dev_attr,
					  char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info outbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
	return sprintf(buf, "%d\n", outbound.bytes_avail_towrite);
}
static DEVICE_ATTR_RO(out_write_bytes_avail);

static ssize_t in_intr_mask_show(struct device *dev,
				 struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info inbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
	return sprintf(buf, "%d\n", inbound.current_interrupt_mask);
}
static DEVICE_ATTR_RO(in_intr_mask);

static ssize_t in_read_index_show(struct device *dev,
				  struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info inbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
	return sprintf(buf, "%d\n", inbound.current_read_index);
}
static DEVICE_ATTR_RO(in_read_index);

static ssize_t in_write_index_show(struct device *dev,
				   struct device_attribute *dev_attr, char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info inbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
	return sprintf(buf, "%d\n", inbound.current_write_index);
}
static DEVICE_ATTR_RO(in_write_index);

static ssize_t in_read_bytes_avail_show(struct device *dev,
					struct device_attribute *dev_attr,
					char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info inbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
	return sprintf(buf, "%d\n", inbound.bytes_avail_toread);
}
static DEVICE_ATTR_RO(in_read_bytes_avail);

static ssize_t in_write_bytes_avail_show(struct device *dev,
					 struct device_attribute *dev_attr,
					 char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct hv_ring_buffer_debug_info inbound;

	if (!hv_dev->channel)
		return -ENODEV;
	hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
	return sprintf(buf, "%d\n", inbound.bytes_avail_towrite);
}
static DEVICE_ATTR_RO(in_write_bytes_avail);

/* Set up per device attributes in /sys/bus/vmbus/devices/<bus device> */
391 392
static struct attribute *vmbus_attrs[] = {
	&dev_attr_id.attr,
393
	&dev_attr_state.attr,
394
	&dev_attr_monitor_id.attr,
395
	&dev_attr_class_id.attr,
396
	&dev_attr_device_id.attr,
397
	&dev_attr_modalias.attr,
398 399
	&dev_attr_server_monitor_pending.attr,
	&dev_attr_client_monitor_pending.attr,
400 401
	&dev_attr_server_monitor_latency.attr,
	&dev_attr_client_monitor_latency.attr,
402 403
	&dev_attr_server_monitor_conn_id.attr,
	&dev_attr_client_monitor_conn_id.attr,
404 405 406 407 408 409 410 411 412 413
	&dev_attr_out_intr_mask.attr,
	&dev_attr_out_read_index.attr,
	&dev_attr_out_write_index.attr,
	&dev_attr_out_read_bytes_avail.attr,
	&dev_attr_out_write_bytes_avail.attr,
	&dev_attr_in_intr_mask.attr,
	&dev_attr_in_read_index.attr,
	&dev_attr_in_write_index.attr,
	&dev_attr_in_read_bytes_avail.attr,
	&dev_attr_in_write_bytes_avail.attr,
414 415 416 417
	NULL,
};
ATTRIBUTE_GROUPS(vmbus);

418 419 420 421 422 423
/*
 * vmbus_uevent - add uevent for our device
 *
 * This routine is invoked when a device is added or removed on the vmbus to
 * generate a uevent to udev in the userspace. The udev will then look at its
 * rule and the uevent generated here to load the appropriate driver
424 425 426 427
 *
 * The alias string will be of the form vmbus:guid where guid is the string
 * representation of the device guid (each byte of the guid will be
 * represented with two hex characters.
428 429 430 431
 */
static int vmbus_uevent(struct device *device, struct kobj_uevent_env *env)
{
	struct hv_device *dev = device_to_hv_device(device);
432 433
	int ret;
	char alias_name[VMBUS_ALIAS_LEN + 1];
434

435
	print_alias_name(dev, alias_name);
436 437
	ret = add_uevent_var(env, "MODALIAS=vmbus:%s", alias_name);
	return ret;
438 439
}

S
stephen hemminger 已提交
440
static const uuid_le null_guid;
441 442 443 444 445 446 447 448

static inline bool is_null_guid(const __u8 *guid)
{
	if (memcmp(guid, &null_guid, sizeof(uuid_le)))
		return false;
	return true;
}

449 450 451 452 453 454
/*
 * Return a matching hv_vmbus_device_id pointer.
 * If there is no match, return NULL.
 */
static const struct hv_vmbus_device_id *hv_vmbus_get_id(
					const struct hv_vmbus_device_id *id,
S
stephen hemminger 已提交
455
					const __u8 *guid)
456 457 458 459 460 461 462 463 464
{
	for (; !is_null_guid(id->guid); id++)
		if (!memcmp(&id->guid, guid, sizeof(uuid_le)))
			return id;

	return NULL;
}


465 466 467 468 469 470 471

/*
 * vmbus_match - Attempt to match the specified device to the specified driver
 */
static int vmbus_match(struct device *device, struct device_driver *driver)
{
	struct hv_driver *drv = drv_to_hv_drv(driver);
472
	struct hv_device *hv_dev = device_to_hv_device(device);
473

474 475
	if (hv_vmbus_get_id(drv->id_table, hv_dev->dev_type.b))
		return 1;
476

477
	return 0;
478 479
}

480 481 482 483 484 485 486 487
/*
 * vmbus_probe - Add the new vmbus's child device
 */
static int vmbus_probe(struct device *child_device)
{
	int ret = 0;
	struct hv_driver *drv =
			drv_to_hv_drv(child_device->driver);
488
	struct hv_device *dev = device_to_hv_device(child_device);
489
	const struct hv_vmbus_device_id *dev_id;
490

491
	dev_id = hv_vmbus_get_id(drv->id_table, dev->dev_type.b);
492
	if (drv->probe) {
493
		ret = drv->probe(dev, dev_id);
494
		if (ret != 0)
495 496
			pr_err("probe failed for device %s (%d)\n",
			       dev_name(child_device), ret);
497 498

	} else {
499 500
		pr_err("probe not set for driver %s\n",
		       dev_name(child_device));
501
		ret = -ENODEV;
502 503 504 505
	}
	return ret;
}

506 507 508 509 510
/*
 * vmbus_remove - Remove a vmbus device
 */
static int vmbus_remove(struct device *child_device)
{
511
	struct hv_driver *drv = drv_to_hv_drv(child_device->driver);
512
	struct hv_device *dev = device_to_hv_device(child_device);
513

514 515 516 517 518
	if (drv->remove)
		drv->remove(dev);
	else
		pr_err("remove not set for driver %s\n",
			dev_name(child_device));
519 520 521 522

	return 0;
}

523 524 525 526 527 528 529

/*
 * vmbus_shutdown - Shutdown a vmbus device
 */
static void vmbus_shutdown(struct device *child_device)
{
	struct hv_driver *drv;
530
	struct hv_device *dev = device_to_hv_device(child_device);
531 532 533 534 535 536 537 538


	/* The device may not be attached yet */
	if (!child_device->driver)
		return;

	drv = drv_to_hv_drv(child_device->driver);

539 540
	if (drv->shutdown)
		drv->shutdown(dev);
541 542 543 544

	return;
}

545 546 547 548 549 550

/*
 * vmbus_device_release - Final callback release of the vmbus child device
 */
static void vmbus_device_release(struct device *device)
{
551
	struct hv_device *hv_dev = device_to_hv_device(device);
552

553
	kfree(hv_dev);
554 555 556

}

557
/* The one and only one */
558 559 560 561 562 563 564
static struct bus_type  hv_bus = {
	.name =		"vmbus",
	.match =		vmbus_match,
	.shutdown =		vmbus_shutdown,
	.remove =		vmbus_remove,
	.probe =		vmbus_probe,
	.uevent =		vmbus_uevent,
565
	.dev_groups =		vmbus_groups,
566 567
};

568 569 570 571 572 573 574 575 576
struct onmessage_work_context {
	struct work_struct work;
	struct hv_message msg;
};

static void vmbus_onmessage_work(struct work_struct *work)
{
	struct onmessage_work_context *ctx;

577 578 579 580
	/* Do not process messages if we're in DISCONNECTED state */
	if (vmbus_connection.conn_state == DISCONNECTED)
		return;

581 582 583 584 585 586
	ctx = container_of(work, struct onmessage_work_context,
			   work);
	vmbus_onmessage(&ctx->msg);
	kfree(ctx);
}

587
static void hv_process_timer_expiration(struct hv_message *msg, int cpu)
588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614
{
	struct clock_event_device *dev = hv_context.clk_evt[cpu];

	if (dev->event_handler)
		dev->event_handler(dev);

	msg->header.message_type = HVMSG_NONE;

	/*
	 * Make sure the write to MessageType (ie set to
	 * HVMSG_NONE) happens before we read the
	 * MessagePending and EOMing. Otherwise, the EOMing
	 * will not deliver any more messages since there is
	 * no empty slot
	 */
	mb();

	if (msg->header.message_flags.msg_pending) {
		/*
		 * This will cause message queue rescan to
		 * possibly deliver another msg from the
		 * hypervisor
		 */
		wrmsrl(HV_X64_MSR_EOM, 0);
	}
}

615
static void vmbus_on_msg_dpc(unsigned long data)
G
Greg Kroah-Hartman 已提交
616 617 618 619 620
{
	int cpu = smp_processor_id();
	void *page_addr = hv_context.synic_message_page[cpu];
	struct hv_message *msg = (struct hv_message *)page_addr +
				  VMBUS_MESSAGE_SINT;
621
	struct onmessage_work_context *ctx;
G
Greg Kroah-Hartman 已提交
622 623 624 625 626 627

	while (1) {
		if (msg->header.message_type == HVMSG_NONE) {
			/* no msg */
			break;
		} else {
628 629
			ctx = kmalloc(sizeof(*ctx), GFP_ATOMIC);
			if (ctx == NULL)
G
Greg Kroah-Hartman 已提交
630
				continue;
631 632
			INIT_WORK(&ctx->work, vmbus_onmessage_work);
			memcpy(&ctx->msg, msg, sizeof(*msg));
633
			queue_work(vmbus_connection.work_queue, &ctx->work);
G
Greg Kroah-Hartman 已提交
634 635 636 637 638 639 640 641 642 643 644
		}

		msg->header.message_type = HVMSG_NONE;

		/*
		 * Make sure the write to MessageType (ie set to
		 * HVMSG_NONE) happens before we read the
		 * MessagePending and EOMing. Otherwise, the EOMing
		 * will not deliver any more messages since there is
		 * no empty slot
		 */
645
		mb();
G
Greg Kroah-Hartman 已提交
646 647 648 649 650 651 652 653 654 655 656 657

		if (msg->header.message_flags.msg_pending) {
			/*
			 * This will cause message queue rescan to
			 * possibly deliver another msg from the
			 * hypervisor
			 */
			wrmsrl(HV_X64_MSR_EOM, 0);
		}
	}
}

658
static void vmbus_isr(void)
G
Greg Kroah-Hartman 已提交
659 660 661 662 663
{
	int cpu = smp_processor_id();
	void *page_addr;
	struct hv_message *msg;
	union hv_synic_event_flags *event;
664
	bool handled = false;
G
Greg Kroah-Hartman 已提交
665

666 667
	page_addr = hv_context.synic_event_page[cpu];
	if (page_addr == NULL)
668
		return;
669 670 671

	event = (union hv_synic_event_flags *)page_addr +
					 VMBUS_MESSAGE_SINT;
672 673 674 675 676
	/*
	 * Check for events before checking for messages. This is the order
	 * in which events and messages are checked in Windows guests on
	 * Hyper-V, and the Windows team suggested we do the same.
	 */
G
Greg Kroah-Hartman 已提交
677

678 679
	if ((vmbus_proto_version == VERSION_WS2008) ||
		(vmbus_proto_version == VERSION_WIN7)) {
G
Greg Kroah-Hartman 已提交
680

681 682 683 684 685 686 687 688 689 690 691 692
		/* Since we are a child, we only need to check bit 0 */
		if (sync_test_and_clear_bit(0,
			(unsigned long *) &event->flags32[0])) {
			handled = true;
		}
	} else {
		/*
		 * Our host is win8 or above. The signaling mechanism
		 * has changed and we can directly look at the event page.
		 * If bit n is set then we have an interrup on the channel
		 * whose id is n.
		 */
693 694
		handled = true;
	}
695

696
	if (handled)
697
		tasklet_schedule(hv_context.event_dpc[cpu]);
698 699


700 701 702 703
	page_addr = hv_context.synic_message_page[cpu];
	msg = (struct hv_message *)page_addr + VMBUS_MESSAGE_SINT;

	/* Check if there are actual msgs to be processed */
704 705 706 707 708 709
	if (msg->header.message_type != HVMSG_NONE) {
		if (msg->header.message_type == HVMSG_TIMER_EXPIRED)
			hv_process_timer_expiration(msg, cpu);
		else
			tasklet_schedule(&msg_dpc);
	}
710 711
}

712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744
#ifdef CONFIG_HOTPLUG_CPU
static int hyperv_cpu_disable(void)
{
	return -ENOSYS;
}

static void hv_cpu_hotplug_quirk(bool vmbus_loaded)
{
	static void *previous_cpu_disable;

	/*
	 * Offlining a CPU when running on newer hypervisors (WS2012R2, Win8,
	 * ...) is not supported at this moment as channel interrupts are
	 * distributed across all of them.
	 */

	if ((vmbus_proto_version == VERSION_WS2008) ||
	    (vmbus_proto_version == VERSION_WIN7))
		return;

	if (vmbus_loaded) {
		previous_cpu_disable = smp_ops.cpu_disable;
		smp_ops.cpu_disable = hyperv_cpu_disable;
		pr_notice("CPU offlining is not supported by hypervisor\n");
	} else if (previous_cpu_disable)
		smp_ops.cpu_disable = previous_cpu_disable;
}
#else
static void hv_cpu_hotplug_quirk(bool vmbus_loaded)
{
}
#endif

745
/*
746 747 748
 * vmbus_bus_init -Main vmbus driver initialization routine.
 *
 * Here, we
749 750 751 752
 *	- initialize the vmbus driver context
 *	- invoke the vmbus hv main init routine
 *	- get the irq resource
 *	- retrieve the channel offers
753
 */
754
static int vmbus_bus_init(int irq)
755
{
756
	int ret;
757

758 759
	/* Hypervisor initialization...setup hypercall page..etc */
	ret = hv_init();
760
	if (ret != 0) {
761
		pr_err("Unable to initialize the hypervisor - 0x%x\n", ret);
762
		return ret;
763 764
	}

765
	tasklet_init(&msg_dpc, vmbus_on_msg_dpc, 0);
766

767
	ret = bus_register(&hv_bus);
768
	if (ret)
769
		goto err_cleanup;
770

771
	hv_setup_vmbus_irq(vmbus_isr);
772

773 774 775
	ret = hv_synic_alloc();
	if (ret)
		goto err_alloc;
776
	/*
777
	 * Initialize the per-cpu interrupt state and
778 779
	 * connect to the host.
	 */
780
	on_each_cpu(hv_synic_init, NULL, 1);
781
	ret = vmbus_connect();
782
	if (ret)
783
		goto err_alloc;
784

785
	hv_cpu_hotplug_quirk(true);
786
	vmbus_request_offers();
787

788
	return 0;
789

790 791
err_alloc:
	hv_synic_free();
792
	hv_remove_vmbus_irq();
793 794 795 796 797 798 799

	bus_unregister(&hv_bus);

err_cleanup:
	hv_cleanup();

	return ret;
800 801
}

802
/**
803 804 805 806
 * __vmbus_child_driver_register - Register a vmbus's driver
 * @drv: Pointer to driver structure you want to register
 * @owner: owner module of the drv
 * @mod_name: module name string
807 808
 *
 * Registers the given driver with Linux through the 'driver_register()' call
809
 * and sets up the hyper-v vmbus handling for this driver.
810 811
 * It will return the state of the 'driver_register()' call.
 *
812
 */
813
int __vmbus_driver_register(struct hv_driver *hv_driver, struct module *owner, const char *mod_name)
814
{
815
	int ret;
816

817
	pr_info("registering driver %s\n", hv_driver->name);
818

819 820 821 822
	ret = vmbus_exists();
	if (ret < 0)
		return ret;

823 824 825 826
	hv_driver->driver.name = hv_driver->name;
	hv_driver->driver.owner = owner;
	hv_driver->driver.mod_name = mod_name;
	hv_driver->driver.bus = &hv_bus;
827

828
	ret = driver_register(&hv_driver->driver);
829

830
	return ret;
831
}
832
EXPORT_SYMBOL_GPL(__vmbus_driver_register);
833

834
/**
835 836
 * vmbus_driver_unregister() - Unregister a vmbus's driver
 * @drv: Pointer to driver structure you want to un-register
837
 *
838 839
 * Un-register the given driver that was previous registered with a call to
 * vmbus_driver_register()
840
 */
841
void vmbus_driver_unregister(struct hv_driver *hv_driver)
842
{
843
	pr_info("unregistering driver %s\n", hv_driver->name);
844

845
	if (!vmbus_exists())
846
		driver_unregister(&hv_driver->driver);
847
}
848
EXPORT_SYMBOL_GPL(vmbus_driver_unregister);
849

850
/*
851
 * vmbus_device_create - Creates and registers a new child device
852
 * on the vmbus.
853
 */
S
stephen hemminger 已提交
854 855 856
struct hv_device *vmbus_device_create(const uuid_le *type,
				      const uuid_le *instance,
				      struct vmbus_channel *channel)
857
{
858
	struct hv_device *child_device_obj;
859

860 861
	child_device_obj = kzalloc(sizeof(struct hv_device), GFP_KERNEL);
	if (!child_device_obj) {
862
		pr_err("Unable to allocate device object for child device\n");
863 864 865
		return NULL;
	}

866
	child_device_obj->channel = channel;
867
	memcpy(&child_device_obj->dev_type, type, sizeof(uuid_le));
868
	memcpy(&child_device_obj->dev_instance, instance,
869
	       sizeof(uuid_le));
870 871 872 873 874


	return child_device_obj;
}

875
/*
876
 * vmbus_device_register - Register the child device
877
 */
878
int vmbus_device_register(struct hv_device *child_device_obj)
879
{
880
	int ret = 0;
881

882 883
	dev_set_name(&child_device_obj->device, "vmbus_%d",
		     child_device_obj->channel->id);
884

885
	child_device_obj->device.bus = &hv_bus;
886
	child_device_obj->device.parent = &hv_acpi_dev->dev;
887
	child_device_obj->device.release = vmbus_device_release;
888

889 890 891 892
	/*
	 * Register with the LDM. This will kick off the driver/device
	 * binding...which will eventually call vmbus_match() and vmbus_probe()
	 */
893
	ret = device_register(&child_device_obj->device);
894 895

	if (ret)
896
		pr_err("Unable to register child device\n");
897
	else
898
		pr_debug("child device %s registered\n",
899
			dev_name(&child_device_obj->device));
900 901 902 903

	return ret;
}

904
/*
905
 * vmbus_device_unregister - Remove the specified child device
906
 * from the vmbus.
907
 */
908
void vmbus_device_unregister(struct hv_device *device_obj)
909
{
910 911 912
	pr_debug("child device %s unregistered\n",
		dev_name(&device_obj->device));

913 914 915 916
	/*
	 * Kick off the process of unregistering the device.
	 * This will call vmbus_remove() and eventually vmbus_device_release()
	 */
917
	device_unregister(&device_obj->device);
918 919 920
}


921
/*
922 923
 * VMBUS is an acpi enumerated device. Get the the information we
 * need from DSDT.
924 925
 */

926
static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx)
927
{
928 929 930
	switch (res->type) {
	case ACPI_RESOURCE_TYPE_IRQ:
		irq = res->data.irq.interrupts[0];
G
Gerd Hoffmann 已提交
931
		break;
932

933
	case ACPI_RESOURCE_TYPE_ADDRESS64:
934 935
		hyperv_mmio.start = res->data.address64.address.minimum;
		hyperv_mmio.end = res->data.address64.address.maximum;
G
Gerd Hoffmann 已提交
936
		break;
937 938 939 940 941 942 943 944
	}

	return AE_OK;
}

static int vmbus_acpi_add(struct acpi_device *device)
{
	acpi_status result;
945
	int ret_val = -ENODEV;
946

947 948
	hv_acpi_dev = device;

949
	result = acpi_walk_resources(device->handle, METHOD_NAME__CRS,
950
					vmbus_walk_resources, NULL);
951

952 953 954 955 956 957 958 959 960 961 962 963 964
	if (ACPI_FAILURE(result))
		goto acpi_walk_err;
	/*
	 * The parent of the vmbus acpi device (Gen2 firmware) is the VMOD that
	 * has the mmio ranges. Get that.
	 */
	if (device->parent) {
		result = acpi_walk_resources(device->parent->handle,
					METHOD_NAME__CRS,
					vmbus_walk_resources, NULL);

		if (ACPI_FAILURE(result))
			goto acpi_walk_err;
965 966
		if (hyperv_mmio.start && hyperv_mmio.end)
			request_resource(&iomem_resource, &hyperv_mmio);
967
	}
968 969 970
	ret_val = 0;

acpi_walk_err:
971
	complete(&probe_event);
972
	return ret_val;
973 974 975 976
}

static const struct acpi_device_id vmbus_acpi_device_ids[] = {
	{"VMBUS", 0},
977
	{"VMBus", 0},
978 979 980 981 982 983 984 985 986 987 988 989
	{"", 0},
};
MODULE_DEVICE_TABLE(acpi, vmbus_acpi_device_ids);

static struct acpi_driver vmbus_acpi_driver = {
	.name = "vmbus",
	.ids = vmbus_acpi_device_ids,
	.ops = {
		.add = vmbus_acpi_add,
	},
};

990
static int __init hv_acpi_init(void)
991
{
992
	int ret, t;
993

994
	if (x86_hyper != &x86_hyper_ms_hyperv)
995 996
		return -ENODEV;

997 998 999 1000 1001
	init_completion(&probe_event);

	/*
	 * Get irq resources first.
	 */
1002 1003
	ret = acpi_bus_register_driver(&vmbus_acpi_driver);

1004 1005 1006
	if (ret)
		return ret;

1007 1008 1009 1010 1011
	t = wait_for_completion_timeout(&probe_event, 5*HZ);
	if (t == 0) {
		ret = -ETIMEDOUT;
		goto cleanup;
	}
1012 1013

	if (irq <= 0) {
1014 1015
		ret = -ENODEV;
		goto cleanup;
1016 1017
	}

1018 1019
	ret = vmbus_bus_init(irq);
	if (ret)
1020 1021 1022 1023 1024 1025
		goto cleanup;

	return 0;

cleanup:
	acpi_bus_unregister_driver(&vmbus_acpi_driver);
1026
	hv_acpi_dev = NULL;
1027
	return ret;
1028 1029
}

1030 1031
static void __exit vmbus_exit(void)
{
1032 1033
	int cpu;

1034
	vmbus_connection.conn_state = DISCONNECTED;
1035
	hv_remove_vmbus_irq();
1036 1037 1038
	vmbus_free_channels();
	bus_unregister(&hv_bus);
	hv_cleanup();
1039 1040
	for_each_online_cpu(cpu)
		smp_call_function_single(cpu, hv_synic_cleanup, NULL, 1);
1041
	acpi_bus_unregister_driver(&vmbus_acpi_driver);
1042
	hv_cpu_hotplug_quirk(false);
1043
	vmbus_disconnect();
1044 1045
}

1046

1047
MODULE_LICENSE("GPL");
1048

1049
subsys_initcall(hv_acpi_init);
1050
module_exit(vmbus_exit);