2 * Copyright (c) 2009, Microsoft Corporation.
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.
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
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.
18 * Haiyang Zhang <haiyangz@microsoft.com>
19 * Hank Janssen <hjanssen@microsoft.com>
20 * K. Y. Srinivasan <kys@microsoft.com>
23 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
25 #include <linux/init.h>
26 #include <linux/module.h>
27 #include <linux/device.h>
28 #include <linux/interrupt.h>
29 #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 <linux/sched/task_stack.h>
39 #include <asm/hyperv.h>
40 #include <asm/hypervisor.h>
41 #include <asm/mshyperv.h>
42 #include <linux/notifier.h>
43 #include <linux/ptrace.h>
44 #include <linux/screen_info.h>
45 #include <linux/kdebug.h>
46 #include <linux/efi.h>
47 #include <linux/random.h>
48 #include "hyperv_vmbus.h"
51 struct list_head node;
52 struct hv_vmbus_device_id id;
55 static struct acpi_device *hv_acpi_dev;
57 static struct completion probe_event;
59 static int hyperv_cpuhp_online;
61 static int hyperv_panic_event(struct notifier_block *nb, unsigned long val,
66 regs = current_pt_regs();
68 hyperv_report_panic(regs);
72 static int hyperv_die_event(struct notifier_block *nb, unsigned long val,
75 struct die_args *die = (struct die_args *)args;
76 struct pt_regs *regs = die->regs;
78 hyperv_report_panic(regs);
82 static struct notifier_block hyperv_die_block = {
83 .notifier_call = hyperv_die_event,
85 static struct notifier_block hyperv_panic_block = {
86 .notifier_call = hyperv_panic_event,
89 static const char *fb_mmio_name = "fb_range";
90 static struct resource *fb_mmio;
91 static struct resource *hyperv_mmio;
92 static DEFINE_SEMAPHORE(hyperv_mmio_lock);
94 static int vmbus_exists(void)
96 if (hv_acpi_dev == NULL)
102 #define VMBUS_ALIAS_LEN ((sizeof((struct hv_vmbus_device_id *)0)->guid) * 2)
103 static void print_alias_name(struct hv_device *hv_dev, char *alias_name)
106 for (i = 0; i < VMBUS_ALIAS_LEN; i += 2)
107 sprintf(&alias_name[i], "%02x", hv_dev->dev_type.b[i/2]);
110 static u8 channel_monitor_group(struct vmbus_channel *channel)
112 return (u8)channel->offermsg.monitorid / 32;
115 static u8 channel_monitor_offset(struct vmbus_channel *channel)
117 return (u8)channel->offermsg.monitorid % 32;
120 static u32 channel_pending(struct vmbus_channel *channel,
121 struct hv_monitor_page *monitor_page)
123 u8 monitor_group = channel_monitor_group(channel);
124 return monitor_page->trigger_group[monitor_group].pending;
127 static u32 channel_latency(struct vmbus_channel *channel,
128 struct hv_monitor_page *monitor_page)
130 u8 monitor_group = channel_monitor_group(channel);
131 u8 monitor_offset = channel_monitor_offset(channel);
132 return monitor_page->latency[monitor_group][monitor_offset];
135 static u32 channel_conn_id(struct vmbus_channel *channel,
136 struct hv_monitor_page *monitor_page)
138 u8 monitor_group = channel_monitor_group(channel);
139 u8 monitor_offset = channel_monitor_offset(channel);
140 return monitor_page->parameter[monitor_group][monitor_offset].connectionid.u.id;
143 static ssize_t id_show(struct device *dev, struct device_attribute *dev_attr,
146 struct hv_device *hv_dev = device_to_hv_device(dev);
148 if (!hv_dev->channel)
150 return sprintf(buf, "%d\n", hv_dev->channel->offermsg.child_relid);
152 static DEVICE_ATTR_RO(id);
154 static ssize_t state_show(struct device *dev, struct device_attribute *dev_attr,
157 struct hv_device *hv_dev = device_to_hv_device(dev);
159 if (!hv_dev->channel)
161 return sprintf(buf, "%d\n", hv_dev->channel->state);
163 static DEVICE_ATTR_RO(state);
165 static ssize_t monitor_id_show(struct device *dev,
166 struct device_attribute *dev_attr, char *buf)
168 struct hv_device *hv_dev = device_to_hv_device(dev);
170 if (!hv_dev->channel)
172 return sprintf(buf, "%d\n", hv_dev->channel->offermsg.monitorid);
174 static DEVICE_ATTR_RO(monitor_id);
176 static ssize_t class_id_show(struct device *dev,
177 struct device_attribute *dev_attr, char *buf)
179 struct hv_device *hv_dev = device_to_hv_device(dev);
181 if (!hv_dev->channel)
183 return sprintf(buf, "{%pUl}\n",
184 hv_dev->channel->offermsg.offer.if_type.b);
186 static DEVICE_ATTR_RO(class_id);
188 static ssize_t device_id_show(struct device *dev,
189 struct device_attribute *dev_attr, char *buf)
191 struct hv_device *hv_dev = device_to_hv_device(dev);
193 if (!hv_dev->channel)
195 return sprintf(buf, "{%pUl}\n",
196 hv_dev->channel->offermsg.offer.if_instance.b);
198 static DEVICE_ATTR_RO(device_id);
200 static ssize_t modalias_show(struct device *dev,
201 struct device_attribute *dev_attr, char *buf)
203 struct hv_device *hv_dev = device_to_hv_device(dev);
204 char alias_name[VMBUS_ALIAS_LEN + 1];
206 print_alias_name(hv_dev, alias_name);
207 return sprintf(buf, "vmbus:%s\n", alias_name);
209 static DEVICE_ATTR_RO(modalias);
211 static ssize_t server_monitor_pending_show(struct device *dev,
212 struct device_attribute *dev_attr,
215 struct hv_device *hv_dev = device_to_hv_device(dev);
217 if (!hv_dev->channel)
219 return sprintf(buf, "%d\n",
220 channel_pending(hv_dev->channel,
221 vmbus_connection.monitor_pages[1]));
223 static DEVICE_ATTR_RO(server_monitor_pending);
225 static ssize_t client_monitor_pending_show(struct device *dev,
226 struct device_attribute *dev_attr,
229 struct hv_device *hv_dev = device_to_hv_device(dev);
231 if (!hv_dev->channel)
233 return sprintf(buf, "%d\n",
234 channel_pending(hv_dev->channel,
235 vmbus_connection.monitor_pages[1]));
237 static DEVICE_ATTR_RO(client_monitor_pending);
239 static ssize_t server_monitor_latency_show(struct device *dev,
240 struct device_attribute *dev_attr,
243 struct hv_device *hv_dev = device_to_hv_device(dev);
245 if (!hv_dev->channel)
247 return sprintf(buf, "%d\n",
248 channel_latency(hv_dev->channel,
249 vmbus_connection.monitor_pages[0]));
251 static DEVICE_ATTR_RO(server_monitor_latency);
253 static ssize_t client_monitor_latency_show(struct device *dev,
254 struct device_attribute *dev_attr,
257 struct hv_device *hv_dev = device_to_hv_device(dev);
259 if (!hv_dev->channel)
261 return sprintf(buf, "%d\n",
262 channel_latency(hv_dev->channel,
263 vmbus_connection.monitor_pages[1]));
265 static DEVICE_ATTR_RO(client_monitor_latency);
267 static ssize_t server_monitor_conn_id_show(struct device *dev,
268 struct device_attribute *dev_attr,
271 struct hv_device *hv_dev = device_to_hv_device(dev);
273 if (!hv_dev->channel)
275 return sprintf(buf, "%d\n",
276 channel_conn_id(hv_dev->channel,
277 vmbus_connection.monitor_pages[0]));
279 static DEVICE_ATTR_RO(server_monitor_conn_id);
281 static ssize_t client_monitor_conn_id_show(struct device *dev,
282 struct device_attribute *dev_attr,
285 struct hv_device *hv_dev = device_to_hv_device(dev);
287 if (!hv_dev->channel)
289 return sprintf(buf, "%d\n",
290 channel_conn_id(hv_dev->channel,
291 vmbus_connection.monitor_pages[1]));
293 static DEVICE_ATTR_RO(client_monitor_conn_id);
295 static ssize_t out_intr_mask_show(struct device *dev,
296 struct device_attribute *dev_attr, char *buf)
298 struct hv_device *hv_dev = device_to_hv_device(dev);
299 struct hv_ring_buffer_debug_info outbound;
301 if (!hv_dev->channel)
303 hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
304 return sprintf(buf, "%d\n", outbound.current_interrupt_mask);
306 static DEVICE_ATTR_RO(out_intr_mask);
308 static ssize_t out_read_index_show(struct device *dev,
309 struct device_attribute *dev_attr, char *buf)
311 struct hv_device *hv_dev = device_to_hv_device(dev);
312 struct hv_ring_buffer_debug_info outbound;
314 if (!hv_dev->channel)
316 hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
317 return sprintf(buf, "%d\n", outbound.current_read_index);
319 static DEVICE_ATTR_RO(out_read_index);
321 static ssize_t out_write_index_show(struct device *dev,
322 struct device_attribute *dev_attr,
325 struct hv_device *hv_dev = device_to_hv_device(dev);
326 struct hv_ring_buffer_debug_info outbound;
328 if (!hv_dev->channel)
330 hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
331 return sprintf(buf, "%d\n", outbound.current_write_index);
333 static DEVICE_ATTR_RO(out_write_index);
335 static ssize_t out_read_bytes_avail_show(struct device *dev,
336 struct device_attribute *dev_attr,
339 struct hv_device *hv_dev = device_to_hv_device(dev);
340 struct hv_ring_buffer_debug_info outbound;
342 if (!hv_dev->channel)
344 hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
345 return sprintf(buf, "%d\n", outbound.bytes_avail_toread);
347 static DEVICE_ATTR_RO(out_read_bytes_avail);
349 static ssize_t out_write_bytes_avail_show(struct device *dev,
350 struct device_attribute *dev_attr,
353 struct hv_device *hv_dev = device_to_hv_device(dev);
354 struct hv_ring_buffer_debug_info outbound;
356 if (!hv_dev->channel)
358 hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound);
359 return sprintf(buf, "%d\n", outbound.bytes_avail_towrite);
361 static DEVICE_ATTR_RO(out_write_bytes_avail);
363 static ssize_t in_intr_mask_show(struct device *dev,
364 struct device_attribute *dev_attr, char *buf)
366 struct hv_device *hv_dev = device_to_hv_device(dev);
367 struct hv_ring_buffer_debug_info inbound;
369 if (!hv_dev->channel)
371 hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
372 return sprintf(buf, "%d\n", inbound.current_interrupt_mask);
374 static DEVICE_ATTR_RO(in_intr_mask);
376 static ssize_t in_read_index_show(struct device *dev,
377 struct device_attribute *dev_attr, char *buf)
379 struct hv_device *hv_dev = device_to_hv_device(dev);
380 struct hv_ring_buffer_debug_info inbound;
382 if (!hv_dev->channel)
384 hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
385 return sprintf(buf, "%d\n", inbound.current_read_index);
387 static DEVICE_ATTR_RO(in_read_index);
389 static ssize_t in_write_index_show(struct device *dev,
390 struct device_attribute *dev_attr, char *buf)
392 struct hv_device *hv_dev = device_to_hv_device(dev);
393 struct hv_ring_buffer_debug_info inbound;
395 if (!hv_dev->channel)
397 hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
398 return sprintf(buf, "%d\n", inbound.current_write_index);
400 static DEVICE_ATTR_RO(in_write_index);
402 static ssize_t in_read_bytes_avail_show(struct device *dev,
403 struct device_attribute *dev_attr,
406 struct hv_device *hv_dev = device_to_hv_device(dev);
407 struct hv_ring_buffer_debug_info inbound;
409 if (!hv_dev->channel)
411 hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
412 return sprintf(buf, "%d\n", inbound.bytes_avail_toread);
414 static DEVICE_ATTR_RO(in_read_bytes_avail);
416 static ssize_t in_write_bytes_avail_show(struct device *dev,
417 struct device_attribute *dev_attr,
420 struct hv_device *hv_dev = device_to_hv_device(dev);
421 struct hv_ring_buffer_debug_info inbound;
423 if (!hv_dev->channel)
425 hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound);
426 return sprintf(buf, "%d\n", inbound.bytes_avail_towrite);
428 static DEVICE_ATTR_RO(in_write_bytes_avail);
430 static ssize_t channel_vp_mapping_show(struct device *dev,
431 struct device_attribute *dev_attr,
434 struct hv_device *hv_dev = device_to_hv_device(dev);
435 struct vmbus_channel *channel = hv_dev->channel, *cur_sc;
437 int buf_size = PAGE_SIZE, n_written, tot_written;
438 struct list_head *cur;
443 tot_written = snprintf(buf, buf_size, "%u:%u\n",
444 channel->offermsg.child_relid, channel->target_cpu);
446 spin_lock_irqsave(&channel->lock, flags);
448 list_for_each(cur, &channel->sc_list) {
449 if (tot_written >= buf_size - 1)
452 cur_sc = list_entry(cur, struct vmbus_channel, sc_list);
453 n_written = scnprintf(buf + tot_written,
454 buf_size - tot_written,
456 cur_sc->offermsg.child_relid,
458 tot_written += n_written;
461 spin_unlock_irqrestore(&channel->lock, flags);
465 static DEVICE_ATTR_RO(channel_vp_mapping);
467 static ssize_t vendor_show(struct device *dev,
468 struct device_attribute *dev_attr,
471 struct hv_device *hv_dev = device_to_hv_device(dev);
472 return sprintf(buf, "0x%x\n", hv_dev->vendor_id);
474 static DEVICE_ATTR_RO(vendor);
476 static ssize_t device_show(struct device *dev,
477 struct device_attribute *dev_attr,
480 struct hv_device *hv_dev = device_to_hv_device(dev);
481 return sprintf(buf, "0x%x\n", hv_dev->device_id);
483 static DEVICE_ATTR_RO(device);
485 /* Set up per device attributes in /sys/bus/vmbus/devices/<bus device> */
486 static struct attribute *vmbus_dev_attrs[] = {
488 &dev_attr_state.attr,
489 &dev_attr_monitor_id.attr,
490 &dev_attr_class_id.attr,
491 &dev_attr_device_id.attr,
492 &dev_attr_modalias.attr,
493 &dev_attr_server_monitor_pending.attr,
494 &dev_attr_client_monitor_pending.attr,
495 &dev_attr_server_monitor_latency.attr,
496 &dev_attr_client_monitor_latency.attr,
497 &dev_attr_server_monitor_conn_id.attr,
498 &dev_attr_client_monitor_conn_id.attr,
499 &dev_attr_out_intr_mask.attr,
500 &dev_attr_out_read_index.attr,
501 &dev_attr_out_write_index.attr,
502 &dev_attr_out_read_bytes_avail.attr,
503 &dev_attr_out_write_bytes_avail.attr,
504 &dev_attr_in_intr_mask.attr,
505 &dev_attr_in_read_index.attr,
506 &dev_attr_in_write_index.attr,
507 &dev_attr_in_read_bytes_avail.attr,
508 &dev_attr_in_write_bytes_avail.attr,
509 &dev_attr_channel_vp_mapping.attr,
510 &dev_attr_vendor.attr,
511 &dev_attr_device.attr,
514 ATTRIBUTE_GROUPS(vmbus_dev);
517 * vmbus_uevent - add uevent for our device
519 * This routine is invoked when a device is added or removed on the vmbus to
520 * generate a uevent to udev in the userspace. The udev will then look at its
521 * rule and the uevent generated here to load the appropriate driver
523 * The alias string will be of the form vmbus:guid where guid is the string
524 * representation of the device guid (each byte of the guid will be
525 * represented with two hex characters.
527 static int vmbus_uevent(struct device *device, struct kobj_uevent_env *env)
529 struct hv_device *dev = device_to_hv_device(device);
531 char alias_name[VMBUS_ALIAS_LEN + 1];
533 print_alias_name(dev, alias_name);
534 ret = add_uevent_var(env, "MODALIAS=vmbus:%s", alias_name);
538 static const uuid_le null_guid;
540 static inline bool is_null_guid(const uuid_le *guid)
542 if (uuid_le_cmp(*guid, null_guid))
548 * Return a matching hv_vmbus_device_id pointer.
549 * If there is no match, return NULL.
551 static const struct hv_vmbus_device_id *hv_vmbus_get_id(struct hv_driver *drv,
554 const struct hv_vmbus_device_id *id = NULL;
555 struct vmbus_dynid *dynid;
557 /* Look at the dynamic ids first, before the static ones */
558 spin_lock(&drv->dynids.lock);
559 list_for_each_entry(dynid, &drv->dynids.list, node) {
560 if (!uuid_le_cmp(dynid->id.guid, *guid)) {
565 spin_unlock(&drv->dynids.lock);
572 return NULL; /* empty device table */
574 for (; !is_null_guid(&id->guid); id++)
575 if (!uuid_le_cmp(id->guid, *guid))
581 /* vmbus_add_dynid - add a new device ID to this driver and re-probe devices */
582 static int vmbus_add_dynid(struct hv_driver *drv, uuid_le *guid)
584 struct vmbus_dynid *dynid;
586 dynid = kzalloc(sizeof(*dynid), GFP_KERNEL);
590 dynid->id.guid = *guid;
592 spin_lock(&drv->dynids.lock);
593 list_add_tail(&dynid->node, &drv->dynids.list);
594 spin_unlock(&drv->dynids.lock);
596 return driver_attach(&drv->driver);
599 static void vmbus_free_dynids(struct hv_driver *drv)
601 struct vmbus_dynid *dynid, *n;
603 spin_lock(&drv->dynids.lock);
604 list_for_each_entry_safe(dynid, n, &drv->dynids.list, node) {
605 list_del(&dynid->node);
608 spin_unlock(&drv->dynids.lock);
612 * store_new_id - sysfs frontend to vmbus_add_dynid()
614 * Allow GUIDs to be added to an existing driver via sysfs.
616 static ssize_t new_id_store(struct device_driver *driver, const char *buf,
619 struct hv_driver *drv = drv_to_hv_drv(driver);
623 retval = uuid_le_to_bin(buf, &guid);
627 if (hv_vmbus_get_id(drv, &guid))
630 retval = vmbus_add_dynid(drv, &guid);
635 static DRIVER_ATTR_WO(new_id);
638 * store_remove_id - remove a PCI device ID from this driver
640 * Removes a dynamic pci device ID to this driver.
642 static ssize_t remove_id_store(struct device_driver *driver, const char *buf,
645 struct hv_driver *drv = drv_to_hv_drv(driver);
646 struct vmbus_dynid *dynid, *n;
650 retval = uuid_le_to_bin(buf, &guid);
655 spin_lock(&drv->dynids.lock);
656 list_for_each_entry_safe(dynid, n, &drv->dynids.list, node) {
657 struct hv_vmbus_device_id *id = &dynid->id;
659 if (!uuid_le_cmp(id->guid, guid)) {
660 list_del(&dynid->node);
666 spin_unlock(&drv->dynids.lock);
670 static DRIVER_ATTR_WO(remove_id);
672 static struct attribute *vmbus_drv_attrs[] = {
673 &driver_attr_new_id.attr,
674 &driver_attr_remove_id.attr,
677 ATTRIBUTE_GROUPS(vmbus_drv);
681 * vmbus_match - Attempt to match the specified device to the specified driver
683 static int vmbus_match(struct device *device, struct device_driver *driver)
685 struct hv_driver *drv = drv_to_hv_drv(driver);
686 struct hv_device *hv_dev = device_to_hv_device(device);
688 /* The hv_sock driver handles all hv_sock offers. */
689 if (is_hvsock_channel(hv_dev->channel))
692 if (hv_vmbus_get_id(drv, &hv_dev->dev_type))
699 * vmbus_probe - Add the new vmbus's child device
701 static int vmbus_probe(struct device *child_device)
704 struct hv_driver *drv =
705 drv_to_hv_drv(child_device->driver);
706 struct hv_device *dev = device_to_hv_device(child_device);
707 const struct hv_vmbus_device_id *dev_id;
709 dev_id = hv_vmbus_get_id(drv, &dev->dev_type);
711 ret = drv->probe(dev, dev_id);
713 pr_err("probe failed for device %s (%d)\n",
714 dev_name(child_device), ret);
717 pr_err("probe not set for driver %s\n",
718 dev_name(child_device));
725 * vmbus_remove - Remove a vmbus device
727 static int vmbus_remove(struct device *child_device)
729 struct hv_driver *drv;
730 struct hv_device *dev = device_to_hv_device(child_device);
732 if (child_device->driver) {
733 drv = drv_to_hv_drv(child_device->driver);
743 * vmbus_shutdown - Shutdown a vmbus device
745 static void vmbus_shutdown(struct device *child_device)
747 struct hv_driver *drv;
748 struct hv_device *dev = device_to_hv_device(child_device);
751 /* The device may not be attached yet */
752 if (!child_device->driver)
755 drv = drv_to_hv_drv(child_device->driver);
763 * vmbus_device_release - Final callback release of the vmbus child device
765 static void vmbus_device_release(struct device *device)
767 struct hv_device *hv_dev = device_to_hv_device(device);
768 struct vmbus_channel *channel = hv_dev->channel;
770 mutex_lock(&vmbus_connection.channel_mutex);
771 hv_process_channel_removal(channel,
772 channel->offermsg.child_relid);
773 mutex_unlock(&vmbus_connection.channel_mutex);
778 /* The one and only one */
779 static struct bus_type hv_bus = {
781 .match = vmbus_match,
782 .shutdown = vmbus_shutdown,
783 .remove = vmbus_remove,
784 .probe = vmbus_probe,
785 .uevent = vmbus_uevent,
786 .dev_groups = vmbus_dev_groups,
787 .drv_groups = vmbus_drv_groups,
790 struct onmessage_work_context {
791 struct work_struct work;
792 struct hv_message msg;
795 static void vmbus_onmessage_work(struct work_struct *work)
797 struct onmessage_work_context *ctx;
799 /* Do not process messages if we're in DISCONNECTED state */
800 if (vmbus_connection.conn_state == DISCONNECTED)
803 ctx = container_of(work, struct onmessage_work_context,
805 vmbus_onmessage(&ctx->msg);
809 static void hv_process_timer_expiration(struct hv_message *msg,
810 struct hv_per_cpu_context *hv_cpu)
812 struct clock_event_device *dev = hv_cpu->clk_evt;
814 if (dev->event_handler)
815 dev->event_handler(dev);
817 vmbus_signal_eom(msg, HVMSG_TIMER_EXPIRED);
820 void vmbus_on_msg_dpc(unsigned long data)
822 struct hv_per_cpu_context *hv_cpu = (void *)data;
823 void *page_addr = hv_cpu->synic_message_page;
824 struct hv_message *msg = (struct hv_message *)page_addr +
826 struct vmbus_channel_message_header *hdr;
827 const struct vmbus_channel_message_table_entry *entry;
828 struct onmessage_work_context *ctx;
829 u32 message_type = msg->header.message_type;
831 if (message_type == HVMSG_NONE)
835 hdr = (struct vmbus_channel_message_header *)msg->u.payload;
837 if (hdr->msgtype >= CHANNELMSG_COUNT) {
838 WARN_ONCE(1, "unknown msgtype=%d\n", hdr->msgtype);
842 entry = &channel_message_table[hdr->msgtype];
843 if (entry->handler_type == VMHT_BLOCKING) {
844 ctx = kmalloc(sizeof(*ctx), GFP_ATOMIC);
848 INIT_WORK(&ctx->work, vmbus_onmessage_work);
849 memcpy(&ctx->msg, msg, sizeof(*msg));
852 * The host can generate a rescind message while we
853 * may still be handling the original offer. We deal with
854 * this condition by ensuring the processing is done on the
857 switch (hdr->msgtype) {
858 case CHANNELMSG_RESCIND_CHANNELOFFER:
860 * If we are handling the rescind message;
861 * schedule the work on the global work queue.
863 schedule_work_on(vmbus_connection.connect_cpu,
867 case CHANNELMSG_OFFERCHANNEL:
868 atomic_inc(&vmbus_connection.offer_in_progress);
869 queue_work_on(vmbus_connection.connect_cpu,
870 vmbus_connection.work_queue,
875 queue_work(vmbus_connection.work_queue, &ctx->work);
878 entry->message_handler(hdr);
881 vmbus_signal_eom(msg, message_type);
886 * Direct callback for channels using other deferred processing
888 static void vmbus_channel_isr(struct vmbus_channel *channel)
890 void (*callback_fn)(void *);
892 callback_fn = READ_ONCE(channel->onchannel_callback);
893 if (likely(callback_fn != NULL))
894 (*callback_fn)(channel->channel_callback_context);
898 * Schedule all channels with events pending
900 static void vmbus_chan_sched(struct hv_per_cpu_context *hv_cpu)
902 unsigned long *recv_int_page;
905 if (vmbus_proto_version < VERSION_WIN8) {
906 maxbits = MAX_NUM_CHANNELS_SUPPORTED;
907 recv_int_page = vmbus_connection.recv_int_page;
910 * When the host is win8 and beyond, the event page
911 * can be directly checked to get the id of the channel
912 * that has the interrupt pending.
914 void *page_addr = hv_cpu->synic_event_page;
915 union hv_synic_event_flags *event
916 = (union hv_synic_event_flags *)page_addr +
919 maxbits = HV_EVENT_FLAGS_COUNT;
920 recv_int_page = event->flags;
923 if (unlikely(!recv_int_page))
926 for_each_set_bit(relid, recv_int_page, maxbits) {
927 struct vmbus_channel *channel;
929 if (!sync_test_and_clear_bit(relid, recv_int_page))
932 /* Special case - vmbus channel protocol msg */
938 /* Find channel based on relid */
939 list_for_each_entry_rcu(channel, &hv_cpu->chan_list, percpu_list) {
940 if (channel->offermsg.child_relid != relid)
943 switch (channel->callback_mode) {
945 vmbus_channel_isr(channel);
948 case HV_CALL_BATCHED:
949 hv_begin_read(&channel->inbound);
952 tasklet_schedule(&channel->callback_event);
960 static void vmbus_isr(void)
962 struct hv_per_cpu_context *hv_cpu
963 = this_cpu_ptr(hv_context.cpu_context);
964 void *page_addr = hv_cpu->synic_event_page;
965 struct hv_message *msg;
966 union hv_synic_event_flags *event;
967 bool handled = false;
969 if (unlikely(page_addr == NULL))
972 event = (union hv_synic_event_flags *)page_addr +
975 * Check for events before checking for messages. This is the order
976 * in which events and messages are checked in Windows guests on
977 * Hyper-V, and the Windows team suggested we do the same.
980 if ((vmbus_proto_version == VERSION_WS2008) ||
981 (vmbus_proto_version == VERSION_WIN7)) {
983 /* Since we are a child, we only need to check bit 0 */
984 if (sync_test_and_clear_bit(0, event->flags))
988 * Our host is win8 or above. The signaling mechanism
989 * has changed and we can directly look at the event page.
990 * If bit n is set then we have an interrup on the channel
997 vmbus_chan_sched(hv_cpu);
999 page_addr = hv_cpu->synic_message_page;
1000 msg = (struct hv_message *)page_addr + VMBUS_MESSAGE_SINT;
1002 /* Check if there are actual msgs to be processed */
1003 if (msg->header.message_type != HVMSG_NONE) {
1004 if (msg->header.message_type == HVMSG_TIMER_EXPIRED)
1005 hv_process_timer_expiration(msg, hv_cpu);
1007 tasklet_schedule(&hv_cpu->msg_dpc);
1010 add_interrupt_randomness(HYPERVISOR_CALLBACK_VECTOR, 0);
1015 * vmbus_bus_init -Main vmbus driver initialization routine.
1018 * - initialize the vmbus driver context
1019 * - invoke the vmbus hv main init routine
1020 * - retrieve the channel offers
1022 static int vmbus_bus_init(void)
1026 /* Hypervisor initialization...setup hypercall page..etc */
1029 pr_err("Unable to initialize the hypervisor - 0x%x\n", ret);
1033 ret = bus_register(&hv_bus);
1037 hv_setup_vmbus_irq(vmbus_isr);
1039 ret = hv_synic_alloc();
1043 * Initialize the per-cpu interrupt state and
1044 * connect to the host.
1046 ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "x86/hyperv:online",
1047 hv_synic_init, hv_synic_cleanup);
1050 hyperv_cpuhp_online = ret;
1052 ret = vmbus_connect();
1057 * Only register if the crash MSRs are available
1059 if (ms_hyperv.misc_features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE) {
1060 register_die_notifier(&hyperv_die_block);
1061 atomic_notifier_chain_register(&panic_notifier_list,
1062 &hyperv_panic_block);
1065 vmbus_request_offers();
1070 cpuhp_remove_state(hyperv_cpuhp_online);
1073 hv_remove_vmbus_irq();
1075 bus_unregister(&hv_bus);
1081 * __vmbus_child_driver_register() - Register a vmbus's driver
1082 * @hv_driver: Pointer to driver structure you want to register
1083 * @owner: owner module of the drv
1084 * @mod_name: module name string
1086 * Registers the given driver with Linux through the 'driver_register()' call
1087 * and sets up the hyper-v vmbus handling for this driver.
1088 * It will return the state of the 'driver_register()' call.
1091 int __vmbus_driver_register(struct hv_driver *hv_driver, struct module *owner, const char *mod_name)
1095 pr_info("registering driver %s\n", hv_driver->name);
1097 ret = vmbus_exists();
1101 hv_driver->driver.name = hv_driver->name;
1102 hv_driver->driver.owner = owner;
1103 hv_driver->driver.mod_name = mod_name;
1104 hv_driver->driver.bus = &hv_bus;
1106 spin_lock_init(&hv_driver->dynids.lock);
1107 INIT_LIST_HEAD(&hv_driver->dynids.list);
1109 ret = driver_register(&hv_driver->driver);
1113 EXPORT_SYMBOL_GPL(__vmbus_driver_register);
1116 * vmbus_driver_unregister() - Unregister a vmbus's driver
1117 * @hv_driver: Pointer to driver structure you want to
1120 * Un-register the given driver that was previous registered with a call to
1121 * vmbus_driver_register()
1123 void vmbus_driver_unregister(struct hv_driver *hv_driver)
1125 pr_info("unregistering driver %s\n", hv_driver->name);
1127 if (!vmbus_exists()) {
1128 driver_unregister(&hv_driver->driver);
1129 vmbus_free_dynids(hv_driver);
1132 EXPORT_SYMBOL_GPL(vmbus_driver_unregister);
1135 * vmbus_device_create - Creates and registers a new child device
1138 struct hv_device *vmbus_device_create(const uuid_le *type,
1139 const uuid_le *instance,
1140 struct vmbus_channel *channel)
1142 struct hv_device *child_device_obj;
1144 child_device_obj = kzalloc(sizeof(struct hv_device), GFP_KERNEL);
1145 if (!child_device_obj) {
1146 pr_err("Unable to allocate device object for child device\n");
1150 child_device_obj->channel = channel;
1151 memcpy(&child_device_obj->dev_type, type, sizeof(uuid_le));
1152 memcpy(&child_device_obj->dev_instance, instance,
1154 child_device_obj->vendor_id = 0x1414; /* MSFT vendor ID */
1157 return child_device_obj;
1161 * vmbus_device_register - Register the child device
1163 int vmbus_device_register(struct hv_device *child_device_obj)
1167 dev_set_name(&child_device_obj->device, "%pUl",
1168 child_device_obj->channel->offermsg.offer.if_instance.b);
1170 child_device_obj->device.bus = &hv_bus;
1171 child_device_obj->device.parent = &hv_acpi_dev->dev;
1172 child_device_obj->device.release = vmbus_device_release;
1175 * Register with the LDM. This will kick off the driver/device
1176 * binding...which will eventually call vmbus_match() and vmbus_probe()
1178 ret = device_register(&child_device_obj->device);
1181 pr_err("Unable to register child device\n");
1183 pr_debug("child device %s registered\n",
1184 dev_name(&child_device_obj->device));
1190 * vmbus_device_unregister - Remove the specified child device
1193 void vmbus_device_unregister(struct hv_device *device_obj)
1195 pr_debug("child device %s unregistered\n",
1196 dev_name(&device_obj->device));
1199 * Kick off the process of unregistering the device.
1200 * This will call vmbus_remove() and eventually vmbus_device_release()
1202 device_unregister(&device_obj->device);
1207 * VMBUS is an acpi enumerated device. Get the information we
1210 #define VTPM_BASE_ADDRESS 0xfed40000
1211 static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx)
1213 resource_size_t start = 0;
1214 resource_size_t end = 0;
1215 struct resource *new_res;
1216 struct resource **old_res = &hyperv_mmio;
1217 struct resource **prev_res = NULL;
1219 switch (res->type) {
1222 * "Address" descriptors are for bus windows. Ignore
1223 * "memory" descriptors, which are for registers on
1226 case ACPI_RESOURCE_TYPE_ADDRESS32:
1227 start = res->data.address32.address.minimum;
1228 end = res->data.address32.address.maximum;
1231 case ACPI_RESOURCE_TYPE_ADDRESS64:
1232 start = res->data.address64.address.minimum;
1233 end = res->data.address64.address.maximum;
1237 /* Unused resource type */
1242 * Ignore ranges that are below 1MB, as they're not
1243 * necessary or useful here.
1248 new_res = kzalloc(sizeof(*new_res), GFP_ATOMIC);
1250 return AE_NO_MEMORY;
1252 /* If this range overlaps the virtual TPM, truncate it. */
1253 if (end > VTPM_BASE_ADDRESS && start < VTPM_BASE_ADDRESS)
1254 end = VTPM_BASE_ADDRESS;
1256 new_res->name = "hyperv mmio";
1257 new_res->flags = IORESOURCE_MEM;
1258 new_res->start = start;
1262 * If two ranges are adjacent, merge them.
1270 if (((*old_res)->end + 1) == new_res->start) {
1271 (*old_res)->end = new_res->end;
1276 if ((*old_res)->start == new_res->end + 1) {
1277 (*old_res)->start = new_res->start;
1282 if ((*old_res)->start > new_res->end) {
1283 new_res->sibling = *old_res;
1285 (*prev_res)->sibling = new_res;
1291 old_res = &(*old_res)->sibling;
1298 static int vmbus_acpi_remove(struct acpi_device *device)
1300 struct resource *cur_res;
1301 struct resource *next_res;
1305 __release_region(hyperv_mmio, fb_mmio->start,
1306 resource_size(fb_mmio));
1310 for (cur_res = hyperv_mmio; cur_res; cur_res = next_res) {
1311 next_res = cur_res->sibling;
1319 static void vmbus_reserve_fb(void)
1323 * Make a claim for the frame buffer in the resource tree under the
1324 * first node, which will be the one below 4GB. The length seems to
1325 * be underreported, particularly in a Generation 1 VM. So start out
1326 * reserving a larger area and make it smaller until it succeeds.
1329 if (screen_info.lfb_base) {
1330 if (efi_enabled(EFI_BOOT))
1331 size = max_t(__u32, screen_info.lfb_size, 0x800000);
1333 size = max_t(__u32, screen_info.lfb_size, 0x4000000);
1335 for (; !fb_mmio && (size >= 0x100000); size >>= 1) {
1336 fb_mmio = __request_region(hyperv_mmio,
1337 screen_info.lfb_base, size,
1344 * vmbus_allocate_mmio() - Pick a memory-mapped I/O range.
1345 * @new: If successful, supplied a pointer to the
1346 * allocated MMIO space.
1347 * @device_obj: Identifies the caller
1348 * @min: Minimum guest physical address of the
1350 * @max: Maximum guest physical address
1351 * @size: Size of the range to be allocated
1352 * @align: Alignment of the range to be allocated
1353 * @fb_overlap_ok: Whether this allocation can be allowed
1354 * to overlap the video frame buffer.
1356 * This function walks the resources granted to VMBus by the
1357 * _CRS object in the ACPI namespace underneath the parent
1358 * "bridge" whether that's a root PCI bus in the Generation 1
1359 * case or a Module Device in the Generation 2 case. It then
1360 * attempts to allocate from the global MMIO pool in a way that
1361 * matches the constraints supplied in these parameters and by
1364 * Return: 0 on success, -errno on failure
1366 int vmbus_allocate_mmio(struct resource **new, struct hv_device *device_obj,
1367 resource_size_t min, resource_size_t max,
1368 resource_size_t size, resource_size_t align,
1371 struct resource *iter, *shadow;
1372 resource_size_t range_min, range_max, start;
1373 const char *dev_n = dev_name(&device_obj->device);
1377 down(&hyperv_mmio_lock);
1380 * If overlaps with frame buffers are allowed, then first attempt to
1381 * make the allocation from within the reserved region. Because it
1382 * is already reserved, no shadow allocation is necessary.
1384 if (fb_overlap_ok && fb_mmio && !(min > fb_mmio->end) &&
1385 !(max < fb_mmio->start)) {
1387 range_min = fb_mmio->start;
1388 range_max = fb_mmio->end;
1389 start = (range_min + align - 1) & ~(align - 1);
1390 for (; start + size - 1 <= range_max; start += align) {
1391 *new = request_mem_region_exclusive(start, size, dev_n);
1399 for (iter = hyperv_mmio; iter; iter = iter->sibling) {
1400 if ((iter->start >= max) || (iter->end <= min))
1403 range_min = iter->start;
1404 range_max = iter->end;
1405 start = (range_min + align - 1) & ~(align - 1);
1406 for (; start + size - 1 <= range_max; start += align) {
1407 shadow = __request_region(iter, start, size, NULL,
1412 *new = request_mem_region_exclusive(start, size, dev_n);
1414 shadow->name = (char *)*new;
1419 __release_region(iter, start, size);
1424 up(&hyperv_mmio_lock);
1427 EXPORT_SYMBOL_GPL(vmbus_allocate_mmio);
1430 * vmbus_free_mmio() - Free a memory-mapped I/O range.
1431 * @start: Base address of region to release.
1432 * @size: Size of the range to be allocated
1434 * This function releases anything requested by
1435 * vmbus_mmio_allocate().
1437 void vmbus_free_mmio(resource_size_t start, resource_size_t size)
1439 struct resource *iter;
1441 down(&hyperv_mmio_lock);
1442 for (iter = hyperv_mmio; iter; iter = iter->sibling) {
1443 if ((iter->start >= start + size) || (iter->end <= start))
1446 __release_region(iter, start, size);
1448 release_mem_region(start, size);
1449 up(&hyperv_mmio_lock);
1452 EXPORT_SYMBOL_GPL(vmbus_free_mmio);
1455 * vmbus_cpu_number_to_vp_number() - Map CPU to VP.
1456 * @cpu_number: CPU number in Linux terms
1458 * This function returns the mapping between the Linux processor
1459 * number and the hypervisor's virtual processor number, useful
1460 * in making hypercalls and such that talk about specific
1463 * Return: Virtual processor number in Hyper-V terms
1465 int vmbus_cpu_number_to_vp_number(int cpu_number)
1467 return hv_context.vp_index[cpu_number];
1469 EXPORT_SYMBOL_GPL(vmbus_cpu_number_to_vp_number);
1471 static int vmbus_acpi_add(struct acpi_device *device)
1474 int ret_val = -ENODEV;
1475 struct acpi_device *ancestor;
1477 hv_acpi_dev = device;
1479 result = acpi_walk_resources(device->handle, METHOD_NAME__CRS,
1480 vmbus_walk_resources, NULL);
1482 if (ACPI_FAILURE(result))
1485 * Some ancestor of the vmbus acpi device (Gen1 or Gen2
1486 * firmware) is the VMOD that has the mmio ranges. Get that.
1488 for (ancestor = device->parent; ancestor; ancestor = ancestor->parent) {
1489 result = acpi_walk_resources(ancestor->handle, METHOD_NAME__CRS,
1490 vmbus_walk_resources, NULL);
1492 if (ACPI_FAILURE(result))
1502 complete(&probe_event);
1504 vmbus_acpi_remove(device);
1508 static const struct acpi_device_id vmbus_acpi_device_ids[] = {
1513 MODULE_DEVICE_TABLE(acpi, vmbus_acpi_device_ids);
1515 static struct acpi_driver vmbus_acpi_driver = {
1517 .ids = vmbus_acpi_device_ids,
1519 .add = vmbus_acpi_add,
1520 .remove = vmbus_acpi_remove,
1524 static void hv_kexec_handler(void)
1526 hv_synic_clockevents_cleanup();
1527 vmbus_initiate_unload(false);
1528 vmbus_connection.conn_state = DISCONNECTED;
1529 /* Make sure conn_state is set as hv_synic_cleanup checks for it */
1531 cpuhp_remove_state(hyperv_cpuhp_online);
1535 static void hv_crash_handler(struct pt_regs *regs)
1537 vmbus_initiate_unload(true);
1539 * In crash handler we can't schedule synic cleanup for all CPUs,
1540 * doing the cleanup for current CPU only. This should be sufficient
1543 vmbus_connection.conn_state = DISCONNECTED;
1544 hv_synic_cleanup(smp_processor_id());
1548 static int __init hv_acpi_init(void)
1552 if (x86_hyper != &x86_hyper_ms_hyperv)
1555 init_completion(&probe_event);
1558 * Get ACPI resources first.
1560 ret = acpi_bus_register_driver(&vmbus_acpi_driver);
1565 t = wait_for_completion_timeout(&probe_event, 5*HZ);
1571 ret = vmbus_bus_init();
1575 hv_setup_kexec_handler(hv_kexec_handler);
1576 hv_setup_crash_handler(hv_crash_handler);
1581 acpi_bus_unregister_driver(&vmbus_acpi_driver);
1586 static void __exit vmbus_exit(void)
1590 hv_remove_kexec_handler();
1591 hv_remove_crash_handler();
1592 vmbus_connection.conn_state = DISCONNECTED;
1593 hv_synic_clockevents_cleanup();
1595 hv_remove_vmbus_irq();
1596 for_each_online_cpu(cpu) {
1597 struct hv_per_cpu_context *hv_cpu
1598 = per_cpu_ptr(hv_context.cpu_context, cpu);
1600 tasklet_kill(&hv_cpu->msg_dpc);
1602 vmbus_free_channels();
1604 if (ms_hyperv.misc_features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE) {
1605 unregister_die_notifier(&hyperv_die_block);
1606 atomic_notifier_chain_unregister(&panic_notifier_list,
1607 &hyperv_panic_block);
1609 bus_unregister(&hv_bus);
1611 cpuhp_remove_state(hyperv_cpuhp_online);
1613 acpi_bus_unregister_driver(&vmbus_acpi_driver);
1617 MODULE_LICENSE("GPL");
1619 subsys_initcall(hv_acpi_init);
1620 module_exit(vmbus_exit);