vmbus_drv.c 34.1 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 41
#include <linux/notifier.h>
#include <linux/ptrace.h>
42
#include <linux/screen_info.h>
43
#include <linux/kdebug.h>
44
#include "hyperv_vmbus.h"
45

46
static struct acpi_device  *hv_acpi_dev;
47

48
static struct completion probe_event;
49

50

51
static void hyperv_report_panic(struct pt_regs *regs)
52
{
53
	static bool panic_reported;
54

55 56 57 58 59 60 61 62
	/*
	 * We prefer to report panic on 'die' chain as we have proper
	 * registers to report, but if we miss it (e.g. on BUG()) we need
	 * to report it on 'panic'.
	 */
	if (panic_reported)
		return;
	panic_reported = true;
63 64 65 66 67 68 69 70 71 72 73

	wrmsrl(HV_X64_MSR_CRASH_P0, regs->ip);
	wrmsrl(HV_X64_MSR_CRASH_P1, regs->ax);
	wrmsrl(HV_X64_MSR_CRASH_P2, regs->bx);
	wrmsrl(HV_X64_MSR_CRASH_P3, regs->cx);
	wrmsrl(HV_X64_MSR_CRASH_P4, regs->dx);

	/*
	 * Let Hyper-V know there is crash data available
	 */
	wrmsrl(HV_X64_MSR_CRASH_CTL, HV_CRASH_CTL_CRASH_NOTIFY);
74 75 76 77 78 79 80 81 82 83
}

static int hyperv_panic_event(struct notifier_block *nb, unsigned long val,
			      void *args)
{
	struct pt_regs *regs;

	regs = current_pt_regs();

	hyperv_report_panic(regs);
84 85 86
	return NOTIFY_DONE;
}

87 88 89 90 91 92 93 94 95 96 97 98 99
static int hyperv_die_event(struct notifier_block *nb, unsigned long val,
			    void *args)
{
	struct die_args *die = (struct die_args *)args;
	struct pt_regs *regs = die->regs;

	hyperv_report_panic(regs);
	return NOTIFY_DONE;
}

static struct notifier_block hyperv_die_block = {
	.notifier_call = hyperv_die_event,
};
100 101 102 103
static struct notifier_block hyperv_panic_block = {
	.notifier_call = hyperv_panic_event,
};

104
struct resource *hyperv_mmio;
105

106 107 108 109 110 111 112 113
static int vmbus_exists(void)
{
	if (hv_acpi_dev == NULL)
		return -ENODEV;

	return 0;
}

114 115 116 117 118 119 120 121
#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]);
}

122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
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;
}

139 140 141 142 143 144 145 146
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];
}

147 148 149 150 151 152 153 154
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;
}

155 156 157 158 159 160 161 162 163 164 165
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);

166 167 168 169 170 171 172 173 174 175 176
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);

177 178 179 180 181 182 183 184 185 186 187
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);

188 189 190 191 192 193 194 195 196 197 198 199
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);

200 201 202 203 204 205 206 207 208 209 210 211
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);

212 213 214 215 216 217 218 219 220 221 222
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);

223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
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);
250

251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278
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);

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
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);

307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
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);

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478
static ssize_t channel_vp_mapping_show(struct device *dev,
				       struct device_attribute *dev_attr,
				       char *buf)
{
	struct hv_device *hv_dev = device_to_hv_device(dev);
	struct vmbus_channel *channel = hv_dev->channel, *cur_sc;
	unsigned long flags;
	int buf_size = PAGE_SIZE, n_written, tot_written;
	struct list_head *cur;

	if (!channel)
		return -ENODEV;

	tot_written = snprintf(buf, buf_size, "%u:%u\n",
		channel->offermsg.child_relid, channel->target_cpu);

	spin_lock_irqsave(&channel->lock, flags);

	list_for_each(cur, &channel->sc_list) {
		if (tot_written >= buf_size - 1)
			break;

		cur_sc = list_entry(cur, struct vmbus_channel, sc_list);
		n_written = scnprintf(buf + tot_written,
				     buf_size - tot_written,
				     "%u:%u\n",
				     cur_sc->offermsg.child_relid,
				     cur_sc->target_cpu);
		tot_written += n_written;
	}

	spin_unlock_irqrestore(&channel->lock, flags);

	return tot_written;
}
static DEVICE_ATTR_RO(channel_vp_mapping);

479
/* Set up per device attributes in /sys/bus/vmbus/devices/<bus device> */
480 481
static struct attribute *vmbus_attrs[] = {
	&dev_attr_id.attr,
482
	&dev_attr_state.attr,
483
	&dev_attr_monitor_id.attr,
484
	&dev_attr_class_id.attr,
485
	&dev_attr_device_id.attr,
486
	&dev_attr_modalias.attr,
487 488
	&dev_attr_server_monitor_pending.attr,
	&dev_attr_client_monitor_pending.attr,
489 490
	&dev_attr_server_monitor_latency.attr,
	&dev_attr_client_monitor_latency.attr,
491 492
	&dev_attr_server_monitor_conn_id.attr,
	&dev_attr_client_monitor_conn_id.attr,
493 494 495 496 497 498 499 500 501 502
	&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,
503
	&dev_attr_channel_vp_mapping.attr,
504 505 506 507
	NULL,
};
ATTRIBUTE_GROUPS(vmbus);

508 509 510 511 512 513
/*
 * 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
514 515 516 517
 *
 * 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.
518 519 520 521
 */
static int vmbus_uevent(struct device *device, struct kobj_uevent_env *env)
{
	struct hv_device *dev = device_to_hv_device(device);
522 523
	int ret;
	char alias_name[VMBUS_ALIAS_LEN + 1];
524

525
	print_alias_name(dev, alias_name);
526 527
	ret = add_uevent_var(env, "MODALIAS=vmbus:%s", alias_name);
	return ret;
528 529
}

530
static const uuid_le null_guid;
531

532
static inline bool is_null_guid(const uuid_le *guid)
533
{
534
	if (uuid_le_cmp(*guid, null_guid))
535 536 537 538
		return false;
	return true;
}

539 540 541 542 543 544
/*
 * 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,
545
					const uuid_le *guid)
546
{
547
	for (; !is_null_guid(&id->guid); id++)
548
		if (!uuid_le_cmp(id->guid, *guid))
549 550 551 552 553 554
			return id;

	return NULL;
}


555 556 557 558 559 560 561

/*
 * 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);
562
	struct hv_device *hv_dev = device_to_hv_device(device);
563

564
	if (hv_vmbus_get_id(drv->id_table, &hv_dev->dev_type))
565
		return 1;
566

567
	return 0;
568 569
}

570 571 572 573 574 575 576 577
/*
 * 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);
578
	struct hv_device *dev = device_to_hv_device(child_device);
579
	const struct hv_vmbus_device_id *dev_id;
580

581
	dev_id = hv_vmbus_get_id(drv->id_table, &dev->dev_type);
582
	if (drv->probe) {
583
		ret = drv->probe(dev, dev_id);
584
		if (ret != 0)
585 586
			pr_err("probe failed for device %s (%d)\n",
			       dev_name(child_device), ret);
587 588

	} else {
589 590
		pr_err("probe not set for driver %s\n",
		       dev_name(child_device));
591
		ret = -ENODEV;
592 593 594 595
	}
	return ret;
}

596 597 598 599 600
/*
 * vmbus_remove - Remove a vmbus device
 */
static int vmbus_remove(struct device *child_device)
{
601
	struct hv_driver *drv;
602
	struct hv_device *dev = device_to_hv_device(child_device);
603

604 605 606 607 608
	if (child_device->driver) {
		drv = drv_to_hv_drv(child_device->driver);
		if (drv->remove)
			drv->remove(dev);
	}
609 610 611 612

	return 0;
}

613 614 615 616 617 618 619

/*
 * vmbus_shutdown - Shutdown a vmbus device
 */
static void vmbus_shutdown(struct device *child_device)
{
	struct hv_driver *drv;
620
	struct hv_device *dev = device_to_hv_device(child_device);
621 622 623 624 625 626 627 628


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

	drv = drv_to_hv_drv(child_device->driver);

629 630
	if (drv->shutdown)
		drv->shutdown(dev);
631 632 633 634

	return;
}

635 636 637 638 639 640

/*
 * vmbus_device_release - Final callback release of the vmbus child device
 */
static void vmbus_device_release(struct device *device)
{
641
	struct hv_device *hv_dev = device_to_hv_device(device);
642
	struct vmbus_channel *channel = hv_dev->channel;
643

644 645
	hv_process_channel_removal(channel,
				   channel->offermsg.child_relid);
646
	kfree(hv_dev);
647 648 649

}

650
/* The one and only one */
651 652 653 654 655 656 657
static struct bus_type  hv_bus = {
	.name =		"vmbus",
	.match =		vmbus_match,
	.shutdown =		vmbus_shutdown,
	.remove =		vmbus_remove,
	.probe =		vmbus_probe,
	.uevent =		vmbus_uevent,
658
	.dev_groups =		vmbus_groups,
659 660
};

661 662 663 664 665 666 667 668 669
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;

670 671 672 673
	/* Do not process messages if we're in DISCONNECTED state */
	if (vmbus_connection.conn_state == DISCONNECTED)
		return;

674 675 676 677 678 679
	ctx = container_of(work, struct onmessage_work_context,
			   work);
	vmbus_onmessage(&ctx->msg);
	kfree(ctx);
}

680
static void hv_process_timer_expiration(struct hv_message *msg, int cpu)
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707
{
	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);
	}
}

708
void vmbus_on_msg_dpc(unsigned long data)
709 710 711 712 713
{
	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;
714 715
	struct vmbus_channel_message_header *hdr;
	struct vmbus_channel_message_table_entry *entry;
716
	struct onmessage_work_context *ctx;
717 718

	while (1) {
719
		if (msg->header.message_type == HVMSG_NONE)
720 721
			/* no msg */
			break;
722 723 724 725 726 727 728 729 730 731

		hdr = (struct vmbus_channel_message_header *)msg->u.payload;

		if (hdr->msgtype >= CHANNELMSG_COUNT) {
			WARN_ONCE(1, "unknown msgtype=%d\n", hdr->msgtype);
			goto msg_handled;
		}

		entry = &channel_message_table[hdr->msgtype];
		if (entry->handler_type	== VMHT_BLOCKING) {
732 733
			ctx = kmalloc(sizeof(*ctx), GFP_ATOMIC);
			if (ctx == NULL)
734
				continue;
735

736 737
			INIT_WORK(&ctx->work, vmbus_onmessage_work);
			memcpy(&ctx->msg, msg, sizeof(*msg));
738

739
			queue_work(vmbus_connection.work_queue, &ctx->work);
740 741
		} else
			entry->message_handler(hdr);
742

743
msg_handled:
744 745 746 747 748 749 750 751 752
		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
		 */
753
		mb();
754 755 756 757 758 759 760 761 762 763 764 765

		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);
		}
	}
}

766
static void vmbus_isr(void)
767 768 769 770 771
{
	int cpu = smp_processor_id();
	void *page_addr;
	struct hv_message *msg;
	union hv_synic_event_flags *event;
772
	bool handled = false;
773

774 775
	page_addr = hv_context.synic_event_page[cpu];
	if (page_addr == NULL)
776
		return;
777 778 779

	event = (union hv_synic_event_flags *)page_addr +
					 VMBUS_MESSAGE_SINT;
780 781 782 783 784
	/*
	 * 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.
	 */
785

786 787
	if ((vmbus_proto_version == VERSION_WS2008) ||
		(vmbus_proto_version == VERSION_WIN7)) {
788

789 790 791 792 793 794 795 796 797 798 799 800
		/* 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.
		 */
801 802
		handled = true;
	}
803

804
	if (handled)
805
		tasklet_schedule(hv_context.event_dpc[cpu]);
806 807


808 809 810 811
	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 */
812 813 814 815
	if (msg->header.message_type != HVMSG_NONE) {
		if (msg->header.message_type == HVMSG_TIMER_EXPIRED)
			hv_process_timer_expiration(msg, cpu);
		else
816
			tasklet_schedule(hv_context.msg_dpc[cpu]);
817
	}
818 819
}

820

821
/*
822 823 824
 * vmbus_bus_init -Main vmbus driver initialization routine.
 *
 * Here, we
825 826 827
 *	- initialize the vmbus driver context
 *	- invoke the vmbus hv main init routine
 *	- retrieve the channel offers
828
 */
829
static int vmbus_bus_init(void)
830
{
831
	int ret;
832

833 834
	/* Hypervisor initialization...setup hypercall page..etc */
	ret = hv_init();
835
	if (ret != 0) {
836
		pr_err("Unable to initialize the hypervisor - 0x%x\n", ret);
837
		return ret;
838 839
	}

840
	ret = bus_register(&hv_bus);
841
	if (ret)
842
		goto err_cleanup;
843

844
	hv_setup_vmbus_irq(vmbus_isr);
845

846 847 848
	ret = hv_synic_alloc();
	if (ret)
		goto err_alloc;
849
	/*
850
	 * Initialize the per-cpu interrupt state and
851 852
	 * connect to the host.
	 */
853
	on_each_cpu(hv_synic_init, NULL, 1);
854
	ret = vmbus_connect();
855
	if (ret)
856
		goto err_connect;
857

858 859
	if (vmbus_proto_version > VERSION_WIN7)
		cpu_hotplug_disable();
860 861 862 863

	/*
	 * Only register if the crash MSRs are available
	 */
864
	if (ms_hyperv.misc_features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE) {
865
		register_die_notifier(&hyperv_die_block);
866 867 868 869
		atomic_notifier_chain_register(&panic_notifier_list,
					       &hyperv_panic_block);
	}

870
	vmbus_request_offers();
871

872
	return 0;
873

874 875
err_connect:
	on_each_cpu(hv_synic_cleanup, NULL, 1);
876 877
err_alloc:
	hv_synic_free();
878
	hv_remove_vmbus_irq();
879 880 881 882 883 884 885

	bus_unregister(&hv_bus);

err_cleanup:
	hv_cleanup();

	return ret;
886 887
}

888
/**
889 890
 * __vmbus_child_driver_register() - Register a vmbus's driver
 * @hv_driver: Pointer to driver structure you want to register
891 892
 * @owner: owner module of the drv
 * @mod_name: module name string
893 894
 *
 * Registers the given driver with Linux through the 'driver_register()' call
895
 * and sets up the hyper-v vmbus handling for this driver.
896 897
 * It will return the state of the 'driver_register()' call.
 *
898
 */
899
int __vmbus_driver_register(struct hv_driver *hv_driver, struct module *owner, const char *mod_name)
900
{
901
	int ret;
902

903
	pr_info("registering driver %s\n", hv_driver->name);
904

905 906 907 908
	ret = vmbus_exists();
	if (ret < 0)
		return ret;

909 910 911 912
	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;
913

914
	ret = driver_register(&hv_driver->driver);
915

916
	return ret;
917
}
918
EXPORT_SYMBOL_GPL(__vmbus_driver_register);
919

920
/**
921
 * vmbus_driver_unregister() - Unregister a vmbus's driver
922 923
 * @hv_driver: Pointer to driver structure you want to
 *             un-register
924
 *
925 926
 * Un-register the given driver that was previous registered with a call to
 * vmbus_driver_register()
927
 */
928
void vmbus_driver_unregister(struct hv_driver *hv_driver)
929
{
930
	pr_info("unregistering driver %s\n", hv_driver->name);
931

932
	if (!vmbus_exists())
933
		driver_unregister(&hv_driver->driver);
934
}
935
EXPORT_SYMBOL_GPL(vmbus_driver_unregister);
936

937
/*
938
 * vmbus_device_create - Creates and registers a new child device
939
 * on the vmbus.
940
 */
941 942 943
struct hv_device *vmbus_device_create(const uuid_le *type,
				      const uuid_le *instance,
				      struct vmbus_channel *channel)
944
{
945
	struct hv_device *child_device_obj;
946

947 948
	child_device_obj = kzalloc(sizeof(struct hv_device), GFP_KERNEL);
	if (!child_device_obj) {
949
		pr_err("Unable to allocate device object for child device\n");
950 951 952
		return NULL;
	}

953
	child_device_obj->channel = channel;
954
	memcpy(&child_device_obj->dev_type, type, sizeof(uuid_le));
955
	memcpy(&child_device_obj->dev_instance, instance,
956
	       sizeof(uuid_le));
957 958 959 960 961


	return child_device_obj;
}

962
/*
963
 * vmbus_device_register - Register the child device
964
 */
965
int vmbus_device_register(struct hv_device *child_device_obj)
966
{
967
	int ret = 0;
968

969 970
	dev_set_name(&child_device_obj->device, "vmbus_%d",
		     child_device_obj->channel->id);
971

972
	child_device_obj->device.bus = &hv_bus;
973
	child_device_obj->device.parent = &hv_acpi_dev->dev;
974
	child_device_obj->device.release = vmbus_device_release;
975

976 977 978 979
	/*
	 * Register with the LDM. This will kick off the driver/device
	 * binding...which will eventually call vmbus_match() and vmbus_probe()
	 */
980
	ret = device_register(&child_device_obj->device);
981 982

	if (ret)
983
		pr_err("Unable to register child device\n");
984
	else
985
		pr_debug("child device %s registered\n",
986
			dev_name(&child_device_obj->device));
987 988 989 990

	return ret;
}

991
/*
992
 * vmbus_device_unregister - Remove the specified child device
993
 * from the vmbus.
994
 */
995
void vmbus_device_unregister(struct hv_device *device_obj)
996
{
997 998 999
	pr_debug("child device %s unregistered\n",
		dev_name(&device_obj->device));

1000 1001 1002 1003
	/*
	 * Kick off the process of unregistering the device.
	 * This will call vmbus_remove() and eventually vmbus_device_release()
	 */
1004
	device_unregister(&device_obj->device);
1005 1006 1007
}


1008
/*
1009
 * VMBUS is an acpi enumerated device. Get the information we
1010
 * need from DSDT.
1011
 */
1012
#define VTPM_BASE_ADDRESS 0xfed40000
1013
static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx)
1014
{
1015 1016 1017 1018 1019 1020
	resource_size_t start = 0;
	resource_size_t end = 0;
	struct resource *new_res;
	struct resource **old_res = &hyperv_mmio;
	struct resource **prev_res = NULL;

1021
	switch (res->type) {
1022 1023 1024 1025 1026 1027 1028 1029 1030

	/*
	 * "Address" descriptors are for bus windows. Ignore
	 * "memory" descriptors, which are for registers on
	 * devices.
	 */
	case ACPI_RESOURCE_TYPE_ADDRESS32:
		start = res->data.address32.address.minimum;
		end = res->data.address32.address.maximum;
Gerd Hoffmann's avatar
Gerd Hoffmann committed
1031
		break;
1032

1033
	case ACPI_RESOURCE_TYPE_ADDRESS64:
1034 1035
		start = res->data.address64.address.minimum;
		end = res->data.address64.address.maximum;
Gerd Hoffmann's avatar
Gerd Hoffmann committed
1036
		break;
1037 1038 1039 1040 1041

	default:
		/* Unused resource type */
		return AE_OK;

1042
	}
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
	/*
	 * Ignore ranges that are below 1MB, as they're not
	 * necessary or useful here.
	 */
	if (end < 0x100000)
		return AE_OK;

	new_res = kzalloc(sizeof(*new_res), GFP_ATOMIC);
	if (!new_res)
		return AE_NO_MEMORY;

	/* If this range overlaps the virtual TPM, truncate it. */
	if (end > VTPM_BASE_ADDRESS && start < VTPM_BASE_ADDRESS)
		end = VTPM_BASE_ADDRESS;

	new_res->name = "hyperv mmio";
	new_res->flags = IORESOURCE_MEM;
	new_res->start = start;
	new_res->end = end;

1063 1064 1065 1066
	/*
	 * Stick ranges from higher in address space at the front of the list.
	 * If two ranges are adjacent, merge them.
	 */
1067 1068 1069 1070 1071 1072
	do {
		if (!*old_res) {
			*old_res = new_res;
			break;
		}

1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084
		if (((*old_res)->end + 1) == new_res->start) {
			(*old_res)->end = new_res->end;
			kfree(new_res);
			break;
		}

		if ((*old_res)->start == new_res->end + 1) {
			(*old_res)->start = new_res->start;
			kfree(new_res);
			break;
		}

1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096
		if ((*old_res)->end < new_res->start) {
			new_res->sibling = *old_res;
			if (prev_res)
				(*prev_res)->sibling = new_res;
			*old_res = new_res;
			break;
		}

		prev_res = old_res;
		old_res = &(*old_res)->sibling;

	} while (1);
1097 1098 1099 1100

	return AE_OK;
}

1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
static int vmbus_acpi_remove(struct acpi_device *device)
{
	struct resource *cur_res;
	struct resource *next_res;

	if (hyperv_mmio) {
		for (cur_res = hyperv_mmio; cur_res; cur_res = next_res) {
			next_res = cur_res->sibling;
			kfree(cur_res);
		}
	}

	return 0;
}

1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194
/**
 * vmbus_allocate_mmio() - Pick a memory-mapped I/O range.
 * @new:		If successful, supplied a pointer to the
 *			allocated MMIO space.
 * @device_obj:		Identifies the caller
 * @min:		Minimum guest physical address of the
 *			allocation
 * @max:		Maximum guest physical address
 * @size:		Size of the range to be allocated
 * @align:		Alignment of the range to be allocated
 * @fb_overlap_ok:	Whether this allocation can be allowed
 *			to overlap the video frame buffer.
 *
 * This function walks the resources granted to VMBus by the
 * _CRS object in the ACPI namespace underneath the parent
 * "bridge" whether that's a root PCI bus in the Generation 1
 * case or a Module Device in the Generation 2 case.  It then
 * attempts to allocate from the global MMIO pool in a way that
 * matches the constraints supplied in these parameters and by
 * that _CRS.
 *
 * Return: 0 on success, -errno on failure
 */
int vmbus_allocate_mmio(struct resource **new, struct hv_device *device_obj,
			resource_size_t min, resource_size_t max,
			resource_size_t size, resource_size_t align,
			bool fb_overlap_ok)
{
	struct resource *iter;
	resource_size_t range_min, range_max, start, local_min, local_max;
	const char *dev_n = dev_name(&device_obj->device);
	u32 fb_end = screen_info.lfb_base + (screen_info.lfb_size << 1);
	int i;

	for (iter = hyperv_mmio; iter; iter = iter->sibling) {
		if ((iter->start >= max) || (iter->end <= min))
			continue;

		range_min = iter->start;
		range_max = iter->end;

		/* If this range overlaps the frame buffer, split it into
		   two tries. */
		for (i = 0; i < 2; i++) {
			local_min = range_min;
			local_max = range_max;
			if (fb_overlap_ok || (range_min >= fb_end) ||
			    (range_max <= screen_info.lfb_base)) {
				i++;
			} else {
				if ((range_min <= screen_info.lfb_base) &&
				    (range_max >= screen_info.lfb_base)) {
					/*
					 * The frame buffer is in this window,
					 * so trim this into the part that
					 * preceeds the frame buffer.
					 */
					local_max = screen_info.lfb_base - 1;
					range_min = fb_end;
				} else {
					range_min = fb_end;
					continue;
				}
			}

			start = (local_min + align - 1) & ~(align - 1);
			for (; start + size - 1 <= local_max; start += align) {
				*new = request_mem_region_exclusive(start, size,
								    dev_n);
				if (*new)
					return 0;
			}
		}
	}

	return -ENXIO;
}
EXPORT_SYMBOL_GPL(vmbus_allocate_mmio);

1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211
/**
 * vmbus_cpu_number_to_vp_number() - Map CPU to VP.
 * @cpu_number: CPU number in Linux terms
 *
 * This function returns the mapping between the Linux processor
 * number and the hypervisor's virtual processor number, useful
 * in making hypercalls and such that talk about specific
 * processors.
 *
 * Return: Virtual processor number in Hyper-V terms
 */
int vmbus_cpu_number_to_vp_number(int cpu_number)
{
	return hv_context.vp_index[cpu_number];
}
EXPORT_SYMBOL_GPL(vmbus_cpu_number_to_vp_number);

1212 1213 1214
static int vmbus_acpi_add(struct acpi_device *device)
{
	acpi_status result;
1215
	int ret_val = -ENODEV;
1216
	struct acpi_device *ancestor;
1217

1218 1219
	hv_acpi_dev = device;

1220
	result = acpi_walk_resources(device->handle, METHOD_NAME__CRS,
1221
					vmbus_walk_resources, NULL);
1222

1223 1224 1225
	if (ACPI_FAILURE(result))
		goto acpi_walk_err;
	/*
1226 1227
	 * Some ancestor of the vmbus acpi device (Gen1 or Gen2
	 * firmware) is the VMOD that has the mmio ranges. Get that.
1228
	 */
1229 1230 1231
	for (ancestor = device->parent; ancestor; ancestor = ancestor->parent) {
		result = acpi_walk_resources(ancestor->handle, METHOD_NAME__CRS,
					     vmbus_walk_resources, NULL);
1232 1233

		if (ACPI_FAILURE(result))
1234 1235 1236
			continue;
		if (hyperv_mmio)
			break;
1237
	}
1238 1239 1240
	ret_val = 0;

acpi_walk_err:
1241
	complete(&probe_event);
1242 1243
	if (ret_val)
		vmbus_acpi_remove(device);
1244
	return ret_val;
1245 1246 1247 1248
}

static const struct acpi_device_id vmbus_acpi_device_ids[] = {
	{"VMBUS", 0},
1249
	{"VMBus", 0},
1250 1251 1252 1253 1254 1255 1256 1257 1258
	{"", 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,
1259
		.remove = vmbus_acpi_remove,
1260 1261 1262
	},
};

1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
static void hv_kexec_handler(void)
{
	int cpu;

	hv_synic_clockevents_cleanup();
	vmbus_initiate_unload();
	for_each_online_cpu(cpu)
		smp_call_function_single(cpu, hv_synic_cleanup, NULL, 1);
	hv_cleanup();
};

1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
static void hv_crash_handler(struct pt_regs *regs)
{
	vmbus_initiate_unload();
	/*
	 * In crash handler we can't schedule synic cleanup for all CPUs,
	 * doing the cleanup for current CPU only. This should be sufficient
	 * for kdump.
	 */
	hv_synic_cleanup(NULL);
	hv_cleanup();
};

1286
static int __init hv_acpi_init(void)
1287
{
1288
	int ret, t;
1289

1290
	if (x86_hyper != &x86_hyper_ms_hyperv)
1291 1292
		return -ENODEV;

1293 1294 1295
	init_completion(&probe_event);

	/*
1296
	 * Get ACPI resources first.
1297
	 */
1298 1299
	ret = acpi_bus_register_driver(&vmbus_acpi_driver);

1300 1301 1302
	if (ret)
		return ret;

1303 1304 1305 1306 1307
	t = wait_for_completion_timeout(&probe_event, 5*HZ);
	if (t == 0) {
		ret = -ETIMEDOUT;
		goto cleanup;
	}
1308

1309
	ret = vmbus_bus_init();
1310
	if (ret)
1311 1312
		goto cleanup;

1313
	hv_setup_kexec_handler(hv_kexec_handler);
1314
	hv_setup_crash_handler(hv_crash_handler);
1315

1316 1317 1318 1319
	return 0;

cleanup:
	acpi_bus_unregister_driver(&vmbus_acpi_driver);
1320
	hv_acpi_dev = NULL;
1321
	return ret;
1322 1323
}

1324 1325
static void __exit vmbus_exit(void)
{
1326 1327
	int cpu;

1328
	hv_remove_kexec_handler();
1329
	hv_remove_crash_handler();
1330
	vmbus_connection.conn_state = DISCONNECTED;
1331
	hv_synic_clockevents_cleanup();
1332
	vmbus_disconnect();
1333
	hv_remove_vmbus_irq();
1334 1335
	for_each_online_cpu(cpu)
		tasklet_kill(hv_context.msg_dpc[cpu]);
1336
	vmbus_free_channels();
1337
	if (ms_hyperv.misc_features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE) {
1338
		unregister_die_notifier(&hyperv_die_block);
1339 1340 1341
		atomic_notifier_chain_unregister(&panic_notifier_list,
						 &hyperv_panic_block);
	}
1342 1343
	bus_unregister(&hv_bus);
	hv_cleanup();
1344 1345
	for_each_online_cpu(cpu) {
		tasklet_kill(hv_context.event_dpc[cpu]);
1346
		smp_call_function_single(cpu, hv_synic_cleanup, NULL, 1);
1347
	}
1348
	hv_synic_free();
1349
	acpi_bus_unregister_driver(&vmbus_acpi_driver);
1350 1351
	if (vmbus_proto_version > VERSION_WIN7)
		cpu_hotplug_enable();
1352 1353
}

1354

1355
MODULE_LICENSE("GPL");
1356

1357
subsys_initcall(hv_acpi_init);
1358
module_exit(vmbus_exit);