kvm_main.c 61.8 KB
Newer Older
A
Avi Kivity 已提交
1 2 3 4 5 6 7
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * This module enables machines with Intel VT-x extensions to run virtual
 * machines without emulation or binary translation.
 *
 * Copyright (C) 2006 Qumranet, Inc.
N
Nicolas Kaiser 已提交
8
 * Copyright 2010 Red Hat, Inc. and/or its affiliates.
A
Avi Kivity 已提交
9 10 11 12 13 14 15 16 17 18
 *
 * Authors:
 *   Avi Kivity   <avi@qumranet.com>
 *   Yaniv Kamay  <yaniv@qumranet.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */

19
#include "iodev.h"
A
Avi Kivity 已提交
20

21
#include <linux/kvm_host.h>
A
Avi Kivity 已提交
22 23 24 25 26 27 28 29 30 31 32
#include <linux/kvm.h>
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/percpu.h>
#include <linux/mm.h>
#include <linux/miscdevice.h>
#include <linux/vmalloc.h>
#include <linux/reboot.h>
#include <linux/debugfs.h>
#include <linux/highmem.h>
#include <linux/file.h>
33
#include <linux/syscore_ops.h>
A
Avi Kivity 已提交
34
#include <linux/cpu.h>
A
Alexey Dobriyan 已提交
35
#include <linux/sched.h>
36 37
#include <linux/cpumask.h>
#include <linux/smp.h>
38
#include <linux/anon_inodes.h>
39
#include <linux/profile.h>
40
#include <linux/kvm_para.h>
41
#include <linux/pagemap.h>
42
#include <linux/mman.h>
43
#include <linux/swap.h>
44
#include <linux/bitops.h>
45
#include <linux/spinlock.h>
46
#include <linux/compat.h>
47
#include <linux/srcu.h>
J
Joerg Roedel 已提交
48
#include <linux/hugetlb.h>
49
#include <linux/slab.h>
50 51
#include <linux/sort.h>
#include <linux/bsearch.h>
A
Avi Kivity 已提交
52

A
Avi Kivity 已提交
53 54 55
#include <asm/processor.h>
#include <asm/io.h>
#include <asm/uaccess.h>
56
#include <asm/pgtable.h>
A
Avi Kivity 已提交
57

58
#include "coalesced_mmio.h"
59
#include "async_pf.h"
60

61 62 63
#define CREATE_TRACE_POINTS
#include <trace/events/kvm.h>

A
Avi Kivity 已提交
64 65 66
MODULE_AUTHOR("Qumranet");
MODULE_LICENSE("GPL");

67 68 69
/*
 * Ordering of locks:
 *
70
 * 		kvm->lock --> kvm->slots_lock --> kvm->irq_lock
71 72
 */

73
DEFINE_RAW_SPINLOCK(kvm_lock);
74
LIST_HEAD(vm_list);
75

76
static cpumask_var_t cpus_hardware_enabled;
77 78
static int kvm_usage_count = 0;
static atomic_t hardware_enable_failed;
79

80 81
struct kmem_cache *kvm_vcpu_cache;
EXPORT_SYMBOL_GPL(kvm_vcpu_cache);
A
Avi Kivity 已提交
82

83 84
static __read_mostly struct preempt_ops kvm_preempt_ops;

85
struct dentry *kvm_debugfs_dir;
A
Avi Kivity 已提交
86

A
Avi Kivity 已提交
87 88
static long kvm_vcpu_ioctl(struct file *file, unsigned int ioctl,
			   unsigned long arg);
89 90 91 92
#ifdef CONFIG_COMPAT
static long kvm_vcpu_compat_ioctl(struct file *file, unsigned int ioctl,
				  unsigned long arg);
#endif
93 94
static int hardware_enable_all(void);
static void hardware_disable_all(void);
A
Avi Kivity 已提交
95

M
Marcelo Tosatti 已提交
96 97
static void kvm_io_bus_destroy(struct kvm_io_bus *bus);

98 99
bool kvm_rebooting;
EXPORT_SYMBOL_GPL(kvm_rebooting);
100

101 102
static bool largepages_enabled = true;

103 104 105
struct page *bad_page;
static pfn_t bad_pfn;

106 107
static struct page *hwpoison_page;
static pfn_t hwpoison_pfn;
108

109 110
static struct page *fault_page;
static pfn_t fault_pfn;
111

112
inline int kvm_is_mmio_pfn(pfn_t pfn)
B
Ben-Ami Yassour 已提交
113
{
114
	if (pfn_valid(pfn)) {
115
		int reserved;
116
		struct page *tail = pfn_to_page(pfn);
117 118
		struct page *head = compound_trans_head(tail);
		reserved = PageReserved(head);
119 120
		if (head != tail) {
			/*
121 122 123 124 125 126 127 128
			 * "head" is not a dangling pointer
			 * (compound_trans_head takes care of that)
			 * but the hugepage may have been splitted
			 * from under us (and we may not hold a
			 * reference count on the head page so it can
			 * be reused before we run PageReferenced), so
			 * we've to check PageTail before returning
			 * what we just read.
129
			 */
130 131 132
			smp_rmb();
			if (PageTail(tail))
				return reserved;
133 134
		}
		return PageReserved(tail);
135
	}
B
Ben-Ami Yassour 已提交
136 137 138 139

	return true;
}

A
Avi Kivity 已提交
140 141 142
/*
 * Switches to specified vcpu, until a matching vcpu_put()
 */
143
void vcpu_load(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
144
{
145 146
	int cpu;

A
Avi Kivity 已提交
147
	mutex_lock(&vcpu->mutex);
148 149 150 151 152 153 154 155
	if (unlikely(vcpu->pid != current->pids[PIDTYPE_PID].pid)) {
		/* The thread running this VCPU changed. */
		struct pid *oldpid = vcpu->pid;
		struct pid *newpid = get_task_pid(current, PIDTYPE_PID);
		rcu_assign_pointer(vcpu->pid, newpid);
		synchronize_rcu();
		put_pid(oldpid);
	}
156 157
	cpu = get_cpu();
	preempt_notifier_register(&vcpu->preempt_notifier);
158
	kvm_arch_vcpu_load(vcpu, cpu);
159
	put_cpu();
A
Avi Kivity 已提交
160 161
}

162
void vcpu_put(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
163
{
164
	preempt_disable();
165
	kvm_arch_vcpu_put(vcpu);
166 167
	preempt_notifier_unregister(&vcpu->preempt_notifier);
	preempt_enable();
A
Avi Kivity 已提交
168 169 170
	mutex_unlock(&vcpu->mutex);
}

171 172 173 174
static void ack_flush(void *_completed)
{
}

175
static bool make_all_cpus_request(struct kvm *kvm, unsigned int req)
176
{
177
	int i, cpu, me;
178 179
	cpumask_var_t cpus;
	bool called = true;
180 181
	struct kvm_vcpu *vcpu;

182
	zalloc_cpumask_var(&cpus, GFP_ATOMIC);
183

184
	me = get_cpu();
185
	kvm_for_each_vcpu(i, vcpu, kvm) {
186
		kvm_make_request(req, vcpu);
187
		cpu = vcpu->cpu;
188 189 190 191 192 193

		/* Set ->requests bit before we read ->mode */
		smp_mb();

		if (cpus != NULL && cpu != -1 && cpu != me &&
		      kvm_vcpu_exiting_guest_mode(vcpu) != OUTSIDE_GUEST_MODE)
194
			cpumask_set_cpu(cpu, cpus);
195
	}
196 197 198 199 200 201
	if (unlikely(cpus == NULL))
		smp_call_function_many(cpu_online_mask, ack_flush, NULL, 1);
	else if (!cpumask_empty(cpus))
		smp_call_function_many(cpus, ack_flush, NULL, 1);
	else
		called = false;
202
	put_cpu();
203
	free_cpumask_var(cpus);
204
	return called;
205 206
}

207
void kvm_flush_remote_tlbs(struct kvm *kvm)
208
{
209
	long dirty_count = kvm->tlbs_dirty;
210 211

	smp_mb();
212 213
	if (make_all_cpus_request(kvm, KVM_REQ_TLB_FLUSH))
		++kvm->stat.remote_tlb_flush;
214
	cmpxchg(&kvm->tlbs_dirty, dirty_count, 0);
215 216
}

217 218 219 220
void kvm_reload_remote_mmus(struct kvm *kvm)
{
	make_all_cpus_request(kvm, KVM_REQ_MMU_RELOAD);
}
221

R
Rusty Russell 已提交
222 223 224 225 226 227 228 229 230
int kvm_vcpu_init(struct kvm_vcpu *vcpu, struct kvm *kvm, unsigned id)
{
	struct page *page;
	int r;

	mutex_init(&vcpu->mutex);
	vcpu->cpu = -1;
	vcpu->kvm = kvm;
	vcpu->vcpu_id = id;
231
	vcpu->pid = NULL;
E
Eddie Dong 已提交
232
	init_waitqueue_head(&vcpu->wq);
233
	kvm_async_pf_vcpu_init(vcpu);
R
Rusty Russell 已提交
234 235 236 237 238 239 240 241

	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
	if (!page) {
		r = -ENOMEM;
		goto fail;
	}
	vcpu->run = page_address(page);

242
	r = kvm_arch_vcpu_init(vcpu);
R
Rusty Russell 已提交
243
	if (r < 0)
244
		goto fail_free_run;
R
Rusty Russell 已提交
245 246 247 248 249
	return 0;

fail_free_run:
	free_page((unsigned long)vcpu->run);
fail:
250
	return r;
R
Rusty Russell 已提交
251 252 253 254 255
}
EXPORT_SYMBOL_GPL(kvm_vcpu_init);

void kvm_vcpu_uninit(struct kvm_vcpu *vcpu)
{
256
	put_pid(vcpu->pid);
257
	kvm_arch_vcpu_uninit(vcpu);
R
Rusty Russell 已提交
258 259 260 261
	free_page((unsigned long)vcpu->run);
}
EXPORT_SYMBOL_GPL(kvm_vcpu_uninit);

262 263 264 265 266 267 268 269 270 271 272
#if defined(CONFIG_MMU_NOTIFIER) && defined(KVM_ARCH_WANT_MMU_NOTIFIER)
static inline struct kvm *mmu_notifier_to_kvm(struct mmu_notifier *mn)
{
	return container_of(mn, struct kvm, mmu_notifier);
}

static void kvm_mmu_notifier_invalidate_page(struct mmu_notifier *mn,
					     struct mm_struct *mm,
					     unsigned long address)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
273
	int need_tlb_flush, idx;
274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292

	/*
	 * When ->invalidate_page runs, the linux pte has been zapped
	 * already but the page is still allocated until
	 * ->invalidate_page returns. So if we increase the sequence
	 * here the kvm page fault will notice if the spte can't be
	 * established because the page is going to be freed. If
	 * instead the kvm page fault establishes the spte before
	 * ->invalidate_page runs, kvm_unmap_hva will release it
	 * before returning.
	 *
	 * The sequence increase only need to be seen at spin_unlock
	 * time, and not at spin_lock time.
	 *
	 * Increasing the sequence after the spin_unlock would be
	 * unsafe because the kvm page fault could then establish the
	 * pte after kvm_unmap_hva returned, without noticing the page
	 * is going to be freed.
	 */
293
	idx = srcu_read_lock(&kvm->srcu);
294
	spin_lock(&kvm->mmu_lock);
295

296
	kvm->mmu_notifier_seq++;
297
	need_tlb_flush = kvm_unmap_hva(kvm, address) | kvm->tlbs_dirty;
298 299 300 301
	/* we've to flush the tlb before the pages can be freed */
	if (need_tlb_flush)
		kvm_flush_remote_tlbs(kvm);

302 303
	spin_unlock(&kvm->mmu_lock);
	srcu_read_unlock(&kvm->srcu, idx);
304 305
}

306 307 308 309 310 311
static void kvm_mmu_notifier_change_pte(struct mmu_notifier *mn,
					struct mm_struct *mm,
					unsigned long address,
					pte_t pte)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
312
	int idx;
313

314
	idx = srcu_read_lock(&kvm->srcu);
315 316 317 318
	spin_lock(&kvm->mmu_lock);
	kvm->mmu_notifier_seq++;
	kvm_set_spte_hva(kvm, address, pte);
	spin_unlock(&kvm->mmu_lock);
319
	srcu_read_unlock(&kvm->srcu, idx);
320 321
}

322 323 324 325 326 327
static void kvm_mmu_notifier_invalidate_range_start(struct mmu_notifier *mn,
						    struct mm_struct *mm,
						    unsigned long start,
						    unsigned long end)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
328
	int need_tlb_flush = 0, idx;
329

330
	idx = srcu_read_lock(&kvm->srcu);
331 332 333 334 335 336 337
	spin_lock(&kvm->mmu_lock);
	/*
	 * The count increase must become visible at unlock time as no
	 * spte can be established without taking the mmu_lock and
	 * count is also read inside the mmu_lock critical section.
	 */
	kvm->mmu_notifier_count++;
338
	need_tlb_flush = kvm_unmap_hva_range(kvm, start, end);
339
	need_tlb_flush |= kvm->tlbs_dirty;
340 341 342
	/* we've to flush the tlb before the pages can be freed */
	if (need_tlb_flush)
		kvm_flush_remote_tlbs(kvm);
343 344 345

	spin_unlock(&kvm->mmu_lock);
	srcu_read_unlock(&kvm->srcu, idx);
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
}

static void kvm_mmu_notifier_invalidate_range_end(struct mmu_notifier *mn,
						  struct mm_struct *mm,
						  unsigned long start,
						  unsigned long end)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);

	spin_lock(&kvm->mmu_lock);
	/*
	 * This sequence increase will notify the kvm page fault that
	 * the page that is going to be mapped in the spte could have
	 * been freed.
	 */
	kvm->mmu_notifier_seq++;
362
	smp_wmb();
363 364
	/*
	 * The above sequence increase must be visible before the
365 366
	 * below count decrease, which is ensured by the smp_wmb above
	 * in conjunction with the smp_rmb in mmu_notifier_retry().
367 368 369 370 371 372 373 374 375 376 377 378
	 */
	kvm->mmu_notifier_count--;
	spin_unlock(&kvm->mmu_lock);

	BUG_ON(kvm->mmu_notifier_count < 0);
}

static int kvm_mmu_notifier_clear_flush_young(struct mmu_notifier *mn,
					      struct mm_struct *mm,
					      unsigned long address)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
379
	int young, idx;
380

381
	idx = srcu_read_lock(&kvm->srcu);
382 383
	spin_lock(&kvm->mmu_lock);

384
	young = kvm_age_hva(kvm, address);
385 386 387
	if (young)
		kvm_flush_remote_tlbs(kvm);

388 389 390
	spin_unlock(&kvm->mmu_lock);
	srcu_read_unlock(&kvm->srcu, idx);

391 392 393
	return young;
}

A
Andrea Arcangeli 已提交
394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
static int kvm_mmu_notifier_test_young(struct mmu_notifier *mn,
				       struct mm_struct *mm,
				       unsigned long address)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
	int young, idx;

	idx = srcu_read_lock(&kvm->srcu);
	spin_lock(&kvm->mmu_lock);
	young = kvm_test_age_hva(kvm, address);
	spin_unlock(&kvm->mmu_lock);
	srcu_read_unlock(&kvm->srcu, idx);

	return young;
}

410 411 412 413
static void kvm_mmu_notifier_release(struct mmu_notifier *mn,
				     struct mm_struct *mm)
{
	struct kvm *kvm = mmu_notifier_to_kvm(mn);
414 415 416
	int idx;

	idx = srcu_read_lock(&kvm->srcu);
417
	kvm_arch_flush_shadow(kvm);
418
	srcu_read_unlock(&kvm->srcu, idx);
419 420
}

421 422 423 424 425
static const struct mmu_notifier_ops kvm_mmu_notifier_ops = {
	.invalidate_page	= kvm_mmu_notifier_invalidate_page,
	.invalidate_range_start	= kvm_mmu_notifier_invalidate_range_start,
	.invalidate_range_end	= kvm_mmu_notifier_invalidate_range_end,
	.clear_flush_young	= kvm_mmu_notifier_clear_flush_young,
A
Andrea Arcangeli 已提交
426
	.test_young		= kvm_mmu_notifier_test_young,
427
	.change_pte		= kvm_mmu_notifier_change_pte,
428
	.release		= kvm_mmu_notifier_release,
429
};
430 431 432 433 434 435 436 437 438 439 440 441 442 443

static int kvm_init_mmu_notifier(struct kvm *kvm)
{
	kvm->mmu_notifier.ops = &kvm_mmu_notifier_ops;
	return mmu_notifier_register(&kvm->mmu_notifier, current->mm);
}

#else  /* !(CONFIG_MMU_NOTIFIER && KVM_ARCH_WANT_MMU_NOTIFIER) */

static int kvm_init_mmu_notifier(struct kvm *kvm)
{
	return 0;
}

444 445
#endif /* CONFIG_MMU_NOTIFIER && KVM_ARCH_WANT_MMU_NOTIFIER */

446 447 448 449 450 451
static void kvm_init_memslots_id(struct kvm *kvm)
{
	int i;
	struct kvm_memslots *slots = kvm->memslots;

	for (i = 0; i < KVM_MEM_SLOTS_NUM; i++)
452
		slots->id_to_index[i] = slots->memslots[i].id = i;
453 454
}

455
static struct kvm *kvm_create_vm(unsigned long type)
A
Avi Kivity 已提交
456
{
457 458
	int r, i;
	struct kvm *kvm = kvm_arch_alloc_vm();
A
Avi Kivity 已提交
459

460 461 462
	if (!kvm)
		return ERR_PTR(-ENOMEM);

463
	r = kvm_arch_init_vm(kvm, type);
464 465
	if (r)
		goto out_err_nodisable;
466 467 468 469 470

	r = hardware_enable_all();
	if (r)
		goto out_err_nodisable;

471 472
#ifdef CONFIG_HAVE_KVM_IRQCHIP
	INIT_HLIST_HEAD(&kvm->mask_notifier_list);
473
	INIT_HLIST_HEAD(&kvm->irq_ack_notifier_list);
474
#endif
A
Avi Kivity 已提交
475

476 477 478
	r = -ENOMEM;
	kvm->memslots = kzalloc(sizeof(struct kvm_memslots), GFP_KERNEL);
	if (!kvm->memslots)
479
		goto out_err_nosrcu;
480
	kvm_init_memslots_id(kvm);
481
	if (init_srcu_struct(&kvm->srcu))
482
		goto out_err_nosrcu;
M
Marcelo Tosatti 已提交
483 484 485
	for (i = 0; i < KVM_NR_BUSES; i++) {
		kvm->buses[i] = kzalloc(sizeof(struct kvm_io_bus),
					GFP_KERNEL);
486
		if (!kvm->buses[i])
M
Marcelo Tosatti 已提交
487 488
			goto out_err;
	}
489

490
	spin_lock_init(&kvm->mmu_lock);
491 492
	kvm->mm = current->mm;
	atomic_inc(&kvm->mm->mm_count);
G
Gregory Haskins 已提交
493
	kvm_eventfd_init(kvm);
S
Shaohua Li 已提交
494
	mutex_init(&kvm->lock);
495
	mutex_init(&kvm->irq_lock);
496
	mutex_init(&kvm->slots_lock);
I
Izik Eidus 已提交
497
	atomic_set(&kvm->users_count, 1);
498 499 500 501 502

	r = kvm_init_mmu_notifier(kvm);
	if (r)
		goto out_err;

503
	raw_spin_lock(&kvm_lock);
504
	list_add(&kvm->vm_list, &vm_list);
505
	raw_spin_unlock(&kvm_lock);
506

507
	return kvm;
508 509

out_err:
510 511
	cleanup_srcu_struct(&kvm->srcu);
out_err_nosrcu:
512 513
	hardware_disable_all();
out_err_nodisable:
M
Marcelo Tosatti 已提交
514 515
	for (i = 0; i < KVM_NR_BUSES; i++)
		kfree(kvm->buses[i]);
516
	kfree(kvm->memslots);
517
	kvm_arch_free_vm(kvm);
518
	return ERR_PTR(r);
519 520
}

521 522 523 524
/*
 * Avoid using vmalloc for a small buffer.
 * Should not be used when the size is statically known.
 */
525
void *kvm_kvzalloc(unsigned long size)
526 527 528 529 530 531 532
{
	if (size > PAGE_SIZE)
		return vzalloc(size);
	else
		return kzalloc(size, GFP_KERNEL);
}

533
void kvm_kvfree(const void *addr)
534 535 536 537 538 539 540
{
	if (is_vmalloc_addr(addr))
		vfree(addr);
	else
		kfree(addr);
}

541 542 543 544 545
static void kvm_destroy_dirty_bitmap(struct kvm_memory_slot *memslot)
{
	if (!memslot->dirty_bitmap)
		return;

546
	kvm_kvfree(memslot->dirty_bitmap);
547 548 549
	memslot->dirty_bitmap = NULL;
}

A
Avi Kivity 已提交
550 551 552 553 554 555
/*
 * Free any memory in @free but not in @dont.
 */
static void kvm_free_physmem_slot(struct kvm_memory_slot *free,
				  struct kvm_memory_slot *dont)
{
556 557
	if (!dont || free->rmap != dont->rmap)
		vfree(free->rmap);
A
Avi Kivity 已提交
558 559

	if (!dont || free->dirty_bitmap != dont->dirty_bitmap)
560
		kvm_destroy_dirty_bitmap(free);
A
Avi Kivity 已提交
561

562
	kvm_arch_free_memslot(free, dont);
M
Marcelo Tosatti 已提交
563

A
Avi Kivity 已提交
564
	free->npages = 0;
565
	free->rmap = NULL;
A
Avi Kivity 已提交
566 567
}

568
void kvm_free_physmem(struct kvm *kvm)
A
Avi Kivity 已提交
569
{
570
	struct kvm_memslots *slots = kvm->memslots;
571
	struct kvm_memory_slot *memslot;
572

573 574
	kvm_for_each_memslot(memslot, slots)
		kvm_free_physmem_slot(memslot, NULL);
A
Avi Kivity 已提交
575

576
	kfree(kvm->memslots);
A
Avi Kivity 已提交
577 578
}

579 580
static void kvm_destroy_vm(struct kvm *kvm)
{
M
Marcelo Tosatti 已提交
581
	int i;
582 583
	struct mm_struct *mm = kvm->mm;

584
	kvm_arch_sync_events(kvm);
585
	raw_spin_lock(&kvm_lock);
586
	list_del(&kvm->vm_list);
587
	raw_spin_unlock(&kvm_lock);
588
	kvm_free_irq_routing(kvm);
M
Marcelo Tosatti 已提交
589 590
	for (i = 0; i < KVM_NR_BUSES; i++)
		kvm_io_bus_destroy(kvm->buses[i]);
591
	kvm_coalesced_mmio_free(kvm);
592 593
#if defined(CONFIG_MMU_NOTIFIER) && defined(KVM_ARCH_WANT_MMU_NOTIFIER)
	mmu_notifier_unregister(&kvm->mmu_notifier, kvm->mm);
594 595
#else
	kvm_arch_flush_shadow(kvm);
596
#endif
597
	kvm_arch_destroy_vm(kvm);
598 599 600
	kvm_free_physmem(kvm);
	cleanup_srcu_struct(&kvm->srcu);
	kvm_arch_free_vm(kvm);
601
	hardware_disable_all();
602
	mmdrop(mm);
603 604
}

I
Izik Eidus 已提交
605 606 607 608 609 610 611 612 613 614 615 616 617 618
void kvm_get_kvm(struct kvm *kvm)
{
	atomic_inc(&kvm->users_count);
}
EXPORT_SYMBOL_GPL(kvm_get_kvm);

void kvm_put_kvm(struct kvm *kvm)
{
	if (atomic_dec_and_test(&kvm->users_count))
		kvm_destroy_vm(kvm);
}
EXPORT_SYMBOL_GPL(kvm_put_kvm);


619 620 621 622
static int kvm_vm_release(struct inode *inode, struct file *filp)
{
	struct kvm *kvm = filp->private_data;

G
Gregory Haskins 已提交
623 624
	kvm_irqfd_release(kvm);

I
Izik Eidus 已提交
625
	kvm_put_kvm(kvm);
A
Avi Kivity 已提交
626 627 628
	return 0;
}

629 630
/*
 * Allocation size is twice as large as the actual dirty bitmap size.
631
 * See x86's kvm_vm_ioctl_get_dirty_log() why this is needed.
632
 */
633 634
static int kvm_create_dirty_bitmap(struct kvm_memory_slot *memslot)
{
635
#ifndef CONFIG_S390
636
	unsigned long dirty_bytes = 2 * kvm_dirty_bitmap_bytes(memslot);
637

638
	memslot->dirty_bitmap = kvm_kvzalloc(dirty_bytes);
639 640 641
	if (!memslot->dirty_bitmap)
		return -ENOMEM;

642
#endif /* !CONFIG_S390 */
643 644 645
	return 0;
}

646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
static int cmp_memslot(const void *slot1, const void *slot2)
{
	struct kvm_memory_slot *s1, *s2;

	s1 = (struct kvm_memory_slot *)slot1;
	s2 = (struct kvm_memory_slot *)slot2;

	if (s1->npages < s2->npages)
		return 1;
	if (s1->npages > s2->npages)
		return -1;

	return 0;
}

/*
 * Sort the memslots base on its size, so the larger slots
 * will get better fit.
 */
static void sort_memslots(struct kvm_memslots *slots)
{
667 668
	int i;

669 670
	sort(slots->memslots, KVM_MEM_SLOTS_NUM,
	      sizeof(struct kvm_memory_slot), cmp_memslot, NULL);
671 672 673

	for (i = 0; i < KVM_MEM_SLOTS_NUM; i++)
		slots->id_to_index[slots->memslots[i].id] = i;
674 675
}

676 677 678 679
void update_memslots(struct kvm_memslots *slots, struct kvm_memory_slot *new)
{
	if (new) {
		int id = new->id;
680
		struct kvm_memory_slot *old = id_to_memslot(slots, id);
681
		unsigned long npages = old->npages;
682

683
		*old = *new;
684 685
		if (new->npages != npages)
			sort_memslots(slots);
686 687 688 689 690
	}

	slots->generation++;
}

A
Avi Kivity 已提交
691 692 693 694 695
/*
 * Allocate some memory and give it an address in the guest physical address
 * space.
 *
 * Discontiguous memory is allowed, mostly for framebuffers.
696
 *
697
 * Must be called holding mmap_sem for write.
A
Avi Kivity 已提交
698
 */
699 700 701
int __kvm_set_memory_region(struct kvm *kvm,
			    struct kvm_userspace_memory_region *mem,
			    int user_alloc)
A
Avi Kivity 已提交
702
{
703
	int r;
A
Avi Kivity 已提交
704
	gfn_t base_gfn;
705 706
	unsigned long npages;
	unsigned long i;
A
Avi Kivity 已提交
707 708
	struct kvm_memory_slot *memslot;
	struct kvm_memory_slot old, new;
709
	struct kvm_memslots *slots, *old_memslots;
A
Avi Kivity 已提交
710 711 712 713 714 715 716

	r = -EINVAL;
	/* General sanity checks */
	if (mem->memory_size & (PAGE_SIZE - 1))
		goto out;
	if (mem->guest_phys_addr & (PAGE_SIZE - 1))
		goto out;
717 718 719
	/* We can read the guest memory with __xxx_user() later on. */
	if (user_alloc &&
	    ((mem->userspace_addr & (PAGE_SIZE - 1)) ||
720 721 722
	     !access_ok(VERIFY_WRITE,
			(void __user *)(unsigned long)mem->userspace_addr,
			mem->memory_size)))
723
		goto out;
724
	if (mem->slot >= KVM_MEM_SLOTS_NUM)
A
Avi Kivity 已提交
725 726 727 728
		goto out;
	if (mem->guest_phys_addr + mem->memory_size < mem->guest_phys_addr)
		goto out;

729
	memslot = id_to_memslot(kvm->memslots, mem->slot);
A
Avi Kivity 已提交
730 731 732
	base_gfn = mem->guest_phys_addr >> PAGE_SHIFT;
	npages = mem->memory_size >> PAGE_SHIFT;

733 734 735 736
	r = -EINVAL;
	if (npages > KVM_MEM_MAX_NR_PAGES)
		goto out;

A
Avi Kivity 已提交
737 738 739 740 741
	if (!npages)
		mem->flags &= ~KVM_MEM_LOG_DIRTY_PAGES;

	new = old = *memslot;

742
	new.id = mem->slot;
A
Avi Kivity 已提交
743 744 745 746 747 748 749
	new.base_gfn = base_gfn;
	new.npages = npages;
	new.flags = mem->flags;

	/* Disallow changing a memory slot's size. */
	r = -EINVAL;
	if (npages && old.npages && npages != old.npages)
750
		goto out_free;
A
Avi Kivity 已提交
751 752 753 754

	/* Check for overlaps */
	r = -EEXIST;
	for (i = 0; i < KVM_MEMORY_SLOTS; ++i) {
755
		struct kvm_memory_slot *s = &kvm->memslots->memslots[i];
A
Avi Kivity 已提交
756

757
		if (s == memslot || !s->npages)
A
Avi Kivity 已提交
758 759 760
			continue;
		if (!((base_gfn + npages <= s->base_gfn) ||
		      (base_gfn >= s->base_gfn + s->npages)))
761
			goto out_free;
A
Avi Kivity 已提交
762 763 764 765
	}

	/* Free page dirty bitmap if unneeded */
	if (!(new.flags & KVM_MEM_LOG_DIRTY_PAGES))
A
Al Viro 已提交
766
		new.dirty_bitmap = NULL;
A
Avi Kivity 已提交
767 768 769 770

	r = -ENOMEM;

	/* Allocate if a slot is being created */
771 772 773
	if (npages && !old.npages) {
		new.user_alloc = user_alloc;
		new.userspace_addr = mem->userspace_addr;
774
#ifndef CONFIG_S390
775
		new.rmap = vzalloc(npages * sizeof(*new.rmap));
776
		if (!new.rmap)
777
			goto out_free;
778
#endif /* not defined CONFIG_S390 */
779 780
		if (kvm_arch_create_memslot(&new, npages))
			goto out_free;
A
Avi Kivity 已提交
781
	}
782

A
Avi Kivity 已提交
783 784
	/* Allocate page dirty bitmap if needed */
	if ((new.flags & KVM_MEM_LOG_DIRTY_PAGES) && !new.dirty_bitmap) {
785
		if (kvm_create_dirty_bitmap(&new) < 0)
786
			goto out_free;
787
		/* destroy any largepage mappings for dirty tracking */
A
Avi Kivity 已提交
788 789
	}

790
	if (!npages) {
791 792
		struct kvm_memory_slot *slot;

793
		r = -ENOMEM;
794 795
		slots = kmemdup(kvm->memslots, sizeof(struct kvm_memslots),
				GFP_KERNEL);
796 797
		if (!slots)
			goto out_free;
798 799 800
		slot = id_to_memslot(slots, mem->slot);
		slot->flags |= KVM_MEMSLOT_INVALID;

801
		update_memslots(slots, NULL);
802 803 804 805 806 807 808 809 810 811 812

		old_memslots = kvm->memslots;
		rcu_assign_pointer(kvm->memslots, slots);
		synchronize_srcu_expedited(&kvm->srcu);
		/* From this point no new shadow pages pointing to a deleted
		 * memslot will be created.
		 *
		 * validation of sp->gfn happens in:
		 * 	- gfn_to_hva (kvm_read_guest, gfn_to_pfn)
		 * 	- kvm_is_visible_gfn (mmu_check_roots)
		 */
813
		kvm_arch_flush_shadow(kvm);
814 815
		kfree(old_memslots);
	}
816

817 818 819 820
	r = kvm_arch_prepare_memory_region(kvm, &new, old, mem, user_alloc);
	if (r)
		goto out_free;

821
	/* map/unmap the pages in iommu page table */
822 823 824 825
	if (npages) {
		r = kvm_iommu_map_pages(kvm, &new);
		if (r)
			goto out_free;
826 827
	} else
		kvm_iommu_unmap_pages(kvm, &old);
828

829
	r = -ENOMEM;
830 831
	slots = kmemdup(kvm->memslots, sizeof(struct kvm_memslots),
			GFP_KERNEL);
832 833 834 835 836 837 838
	if (!slots)
		goto out_free;

	/* actual memory is freed via old in kvm_free_physmem_slot below */
	if (!npages) {
		new.rmap = NULL;
		new.dirty_bitmap = NULL;
839
		memset(&new.arch, 0, sizeof(new.arch));
840 841
	}

842
	update_memslots(slots, &new);
843 844 845
	old_memslots = kvm->memslots;
	rcu_assign_pointer(kvm->memslots, slots);
	synchronize_srcu_expedited(&kvm->srcu);
846

847
	kvm_arch_commit_memory_region(kvm, mem, old, user_alloc);
848

849 850 851 852 853 854 855
	/*
	 * If the new memory slot is created, we need to clear all
	 * mmio sptes.
	 */
	if (npages && old.base_gfn != mem->guest_phys_addr >> PAGE_SHIFT)
		kvm_arch_flush_shadow(kvm);

856 857 858
	kvm_free_physmem_slot(&old, &new);
	kfree(old_memslots);

A
Avi Kivity 已提交
859 860
	return 0;

861
out_free:
A
Avi Kivity 已提交
862 863 864
	kvm_free_physmem_slot(&new, &old);
out:
	return r;
865 866

}
867 868 869 870 871 872 873 874
EXPORT_SYMBOL_GPL(__kvm_set_memory_region);

int kvm_set_memory_region(struct kvm *kvm,
			  struct kvm_userspace_memory_region *mem,
			  int user_alloc)
{
	int r;

875
	mutex_lock(&kvm->slots_lock);
876
	r = __kvm_set_memory_region(kvm, mem, user_alloc);
877
	mutex_unlock(&kvm->slots_lock);
878 879
	return r;
}
880 881
EXPORT_SYMBOL_GPL(kvm_set_memory_region);

882 883 884 885
int kvm_vm_ioctl_set_memory_region(struct kvm *kvm,
				   struct
				   kvm_userspace_memory_region *mem,
				   int user_alloc)
886
{
887 888
	if (mem->slot >= KVM_MEMORY_SLOTS)
		return -EINVAL;
889
	return kvm_set_memory_region(kvm, mem, user_alloc);
A
Avi Kivity 已提交
890 891
}

892 893
int kvm_get_dirty_log(struct kvm *kvm,
			struct kvm_dirty_log *log, int *is_dirty)
A
Avi Kivity 已提交
894 895 896
{
	struct kvm_memory_slot *memslot;
	int r, i;
897
	unsigned long n;
A
Avi Kivity 已提交
898 899 900 901 902 903
	unsigned long any = 0;

	r = -EINVAL;
	if (log->slot >= KVM_MEMORY_SLOTS)
		goto out;

904
	memslot = id_to_memslot(kvm->memslots, log->slot);
A
Avi Kivity 已提交
905 906 907 908
	r = -ENOENT;
	if (!memslot->dirty_bitmap)
		goto out;

909
	n = kvm_dirty_bitmap_bytes(memslot);
A
Avi Kivity 已提交
910

911
	for (i = 0; !any && i < n/sizeof(long); ++i)
A
Avi Kivity 已提交
912 913 914 915 916 917
		any = memslot->dirty_bitmap[i];

	r = -EFAULT;
	if (copy_to_user(log->dirty_bitmap, memslot->dirty_bitmap, n))
		goto out;

918 919
	if (any)
		*is_dirty = 1;
A
Avi Kivity 已提交
920 921 922 923 924 925

	r = 0;
out:
	return r;
}

926 927 928 929 930
bool kvm_largepages_enabled(void)
{
	return largepages_enabled;
}

931 932 933 934 935 936
void kvm_disable_largepages(void)
{
	largepages_enabled = false;
}
EXPORT_SYMBOL_GPL(kvm_disable_largepages);

937 938
int is_error_page(struct page *page)
{
939
	return page == bad_page || page == hwpoison_page || page == fault_page;
940 941 942
}
EXPORT_SYMBOL_GPL(is_error_page);

943 944
int is_error_pfn(pfn_t pfn)
{
945
	return pfn == bad_pfn || pfn == hwpoison_pfn || pfn == fault_pfn;
946 947 948
}
EXPORT_SYMBOL_GPL(is_error_pfn);

949 950 951 952 953 954
int is_hwpoison_pfn(pfn_t pfn)
{
	return pfn == hwpoison_pfn;
}
EXPORT_SYMBOL_GPL(is_hwpoison_pfn);

955 956 957 958 959 960 961 962 963 964 965 966
int is_noslot_pfn(pfn_t pfn)
{
	return pfn == bad_pfn;
}
EXPORT_SYMBOL_GPL(is_noslot_pfn);

int is_invalid_pfn(pfn_t pfn)
{
	return pfn == hwpoison_pfn || pfn == fault_pfn;
}
EXPORT_SYMBOL_GPL(is_invalid_pfn);

I
Izik Eidus 已提交
967 968 969 970 971 972 973 974 975 976 977
static inline unsigned long bad_hva(void)
{
	return PAGE_OFFSET;
}

int kvm_is_error_hva(unsigned long addr)
{
	return addr == bad_hva();
}
EXPORT_SYMBOL_GPL(kvm_is_error_hva);

978 979 980 981
struct kvm_memory_slot *gfn_to_memslot(struct kvm *kvm, gfn_t gfn)
{
	return __gfn_to_memslot(kvm_memslots(kvm), gfn);
}
A
Avi Kivity 已提交
982
EXPORT_SYMBOL_GPL(gfn_to_memslot);
A
Avi Kivity 已提交
983

984 985
int kvm_is_visible_gfn(struct kvm *kvm, gfn_t gfn)
{
986
	struct kvm_memory_slot *memslot = gfn_to_memslot(kvm, gfn);
987

988 989 990
	if (!memslot || memslot->id >= KVM_MEMORY_SLOTS ||
	      memslot->flags & KVM_MEMSLOT_INVALID)
		return 0;
991

992
	return 1;
993 994 995
}
EXPORT_SYMBOL_GPL(kvm_is_visible_gfn);

J
Joerg Roedel 已提交
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
unsigned long kvm_host_page_size(struct kvm *kvm, gfn_t gfn)
{
	struct vm_area_struct *vma;
	unsigned long addr, size;

	size = PAGE_SIZE;

	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr))
		return PAGE_SIZE;

	down_read(&current->mm->mmap_sem);
	vma = find_vma(current->mm, addr);
	if (!vma)
		goto out;

	size = vma_kernel_pagesize(vma);

out:
	up_read(&current->mm->mmap_sem);

	return size;
}

1020
static unsigned long gfn_to_hva_many(struct kvm_memory_slot *slot, gfn_t gfn,
1021
				     gfn_t *nr_pages)
I
Izik Eidus 已提交
1022
{
1023
	if (!slot || slot->flags & KVM_MEMSLOT_INVALID)
I
Izik Eidus 已提交
1024
		return bad_hva();
1025 1026 1027 1028

	if (nr_pages)
		*nr_pages = slot->npages - (gfn - slot->base_gfn);

1029
	return gfn_to_hva_memslot(slot, gfn);
I
Izik Eidus 已提交
1030
}
1031 1032 1033

unsigned long gfn_to_hva(struct kvm *kvm, gfn_t gfn)
{
1034
	return gfn_to_hva_many(gfn_to_memslot(kvm, gfn), gfn, NULL);
1035
}
1036
EXPORT_SYMBOL_GPL(gfn_to_hva);
I
Izik Eidus 已提交
1037

1038
pfn_t get_fault_pfn(void)
1039 1040 1041 1042
{
	get_page(fault_page);
	return fault_pfn;
}
1043
EXPORT_SYMBOL_GPL(get_fault_pfn);
1044

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
int get_user_page_nowait(struct task_struct *tsk, struct mm_struct *mm,
	unsigned long start, int write, struct page **page)
{
	int flags = FOLL_TOUCH | FOLL_NOWAIT | FOLL_HWPOISON | FOLL_GET;

	if (write)
		flags |= FOLL_WRITE;

	return __get_user_pages(tsk, mm, start, 1, flags, page, NULL, NULL);
}

1056 1057 1058 1059 1060 1061 1062 1063 1064
static inline int check_user_page_hwpoison(unsigned long addr)
{
	int rc, flags = FOLL_TOUCH | FOLL_HWPOISON | FOLL_WRITE;

	rc = __get_user_pages(current, current->mm, addr, 1,
			      flags, NULL, NULL, NULL);
	return rc == -EHWPOISON;
}

1065 1066
static pfn_t hva_to_pfn(unsigned long addr, bool atomic, bool *async,
			bool write_fault, bool *writable)
A
Avi Kivity 已提交
1067
{
1068
	struct page *page[1];
1069
	int npages = 0;
1070
	pfn_t pfn;
A
Avi Kivity 已提交
1071

1072 1073 1074
	/* we can do it either atomically or asynchronously, not both */
	BUG_ON(atomic && async);

1075 1076 1077 1078 1079
	BUG_ON(!write_fault && !writable);

	if (writable)
		*writable = true;

1080
	if (atomic || async)
1081
		npages = __get_user_pages_fast(addr, 1, 1, page);
1082 1083

	if (unlikely(npages != 1) && !atomic) {
1084
		might_sleep();
1085 1086 1087 1088

		if (writable)
			*writable = write_fault;

1089 1090 1091 1092 1093 1094 1095 1096
		if (async) {
			down_read(&current->mm->mmap_sem);
			npages = get_user_page_nowait(current, current->mm,
						     addr, write_fault, page);
			up_read(&current->mm->mmap_sem);
		} else
			npages = get_user_pages_fast(addr, 1, write_fault,
						     page);
1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109

		/* map read fault as writable if possible */
		if (unlikely(!write_fault) && npages == 1) {
			struct page *wpage[1];

			npages = __get_user_pages_fast(addr, 1, 1, wpage);
			if (npages == 1) {
				*writable = true;
				put_page(page[0]);
				page[0] = wpage[0];
			}
			npages = 1;
		}
1110
	}
I
Izik Eidus 已提交
1111

1112 1113 1114
	if (unlikely(npages != 1)) {
		struct vm_area_struct *vma;

1115
		if (atomic)
1116
			return get_fault_pfn();
1117

1118
		down_read(&current->mm->mmap_sem);
1119 1120
		if (npages == -EHWPOISON ||
			(!async && check_user_page_hwpoison(addr))) {
1121
			up_read(&current->mm->mmap_sem);
1122 1123 1124 1125
			get_page(hwpoison_page);
			return page_to_pfn(hwpoison_page);
		}

1126
		vma = find_vma_intersection(current->mm, addr, addr+1);
1127

1128 1129 1130 1131 1132 1133 1134 1135
		if (vma == NULL)
			pfn = get_fault_pfn();
		else if ((vma->vm_flags & VM_PFNMAP)) {
			pfn = ((addr - vma->vm_start) >> PAGE_SHIFT) +
				vma->vm_pgoff;
			BUG_ON(!kvm_is_mmio_pfn(pfn));
		} else {
			if (async && (vma->vm_flags & VM_WRITE))
1136
				*async = true;
1137
			pfn = get_fault_pfn();
1138
		}
1139
		up_read(&current->mm->mmap_sem);
1140 1141
	} else
		pfn = page_to_pfn(page[0]);
1142

1143
	return pfn;
1144 1145
}

1146
pfn_t hva_to_pfn_atomic(unsigned long addr)
1147
{
1148
	return hva_to_pfn(addr, true, NULL, true, NULL);
1149 1150 1151
}
EXPORT_SYMBOL_GPL(hva_to_pfn_atomic);

1152 1153
static pfn_t __gfn_to_pfn(struct kvm *kvm, gfn_t gfn, bool atomic, bool *async,
			  bool write_fault, bool *writable)
1154 1155 1156
{
	unsigned long addr;

1157 1158 1159
	if (async)
		*async = false;

1160 1161 1162 1163 1164 1165
	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr)) {
		get_page(bad_page);
		return page_to_pfn(bad_page);
	}

1166
	return hva_to_pfn(addr, atomic, async, write_fault, writable);
1167 1168 1169 1170
}

pfn_t gfn_to_pfn_atomic(struct kvm *kvm, gfn_t gfn)
{
1171
	return __gfn_to_pfn(kvm, gfn, true, NULL, true, NULL);
1172 1173 1174
}
EXPORT_SYMBOL_GPL(gfn_to_pfn_atomic);

1175 1176
pfn_t gfn_to_pfn_async(struct kvm *kvm, gfn_t gfn, bool *async,
		       bool write_fault, bool *writable)
1177
{
1178
	return __gfn_to_pfn(kvm, gfn, false, async, write_fault, writable);
1179 1180 1181
}
EXPORT_SYMBOL_GPL(gfn_to_pfn_async);

1182 1183
pfn_t gfn_to_pfn(struct kvm *kvm, gfn_t gfn)
{
1184
	return __gfn_to_pfn(kvm, gfn, false, NULL, true, NULL);
1185
}
1186 1187
EXPORT_SYMBOL_GPL(gfn_to_pfn);

1188 1189 1190 1191 1192 1193 1194
pfn_t gfn_to_pfn_prot(struct kvm *kvm, gfn_t gfn, bool write_fault,
		      bool *writable)
{
	return __gfn_to_pfn(kvm, gfn, false, NULL, write_fault, writable);
}
EXPORT_SYMBOL_GPL(gfn_to_pfn_prot);

1195
pfn_t gfn_to_pfn_memslot(struct kvm_memory_slot *slot, gfn_t gfn)
1196 1197
{
	unsigned long addr = gfn_to_hva_memslot(slot, gfn);
1198
	return hva_to_pfn(addr, false, NULL, true, NULL);
1199 1200
}

1201 1202 1203 1204 1205 1206
int gfn_to_page_many_atomic(struct kvm *kvm, gfn_t gfn, struct page **pages,
								  int nr_pages)
{
	unsigned long addr;
	gfn_t entry;

1207
	addr = gfn_to_hva_many(gfn_to_memslot(kvm, gfn), gfn, &entry);
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
	if (kvm_is_error_hva(addr))
		return -1;

	if (entry < nr_pages)
		return 0;

	return __get_user_pages_fast(addr, nr_pages, 1, pages);
}
EXPORT_SYMBOL_GPL(gfn_to_page_many_atomic);

1218 1219
struct page *gfn_to_page(struct kvm *kvm, gfn_t gfn)
{
1220 1221 1222
	pfn_t pfn;

	pfn = gfn_to_pfn(kvm, gfn);
1223
	if (!kvm_is_mmio_pfn(pfn))
1224 1225
		return pfn_to_page(pfn);

1226
	WARN_ON(kvm_is_mmio_pfn(pfn));
1227 1228 1229

	get_page(bad_page);
	return bad_page;
A
Avi Kivity 已提交
1230
}
1231

A
Avi Kivity 已提交
1232 1233
EXPORT_SYMBOL_GPL(gfn_to_page);

1234 1235
void kvm_release_page_clean(struct page *page)
{
1236
	kvm_release_pfn_clean(page_to_pfn(page));
1237 1238 1239
}
EXPORT_SYMBOL_GPL(kvm_release_page_clean);

1240 1241
void kvm_release_pfn_clean(pfn_t pfn)
{
1242
	if (!kvm_is_mmio_pfn(pfn))
1243
		put_page(pfn_to_page(pfn));
1244 1245 1246
}
EXPORT_SYMBOL_GPL(kvm_release_pfn_clean);

1247
void kvm_release_page_dirty(struct page *page)
1248
{
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
	kvm_release_pfn_dirty(page_to_pfn(page));
}
EXPORT_SYMBOL_GPL(kvm_release_page_dirty);

void kvm_release_pfn_dirty(pfn_t pfn)
{
	kvm_set_pfn_dirty(pfn);
	kvm_release_pfn_clean(pfn);
}
EXPORT_SYMBOL_GPL(kvm_release_pfn_dirty);

void kvm_set_page_dirty(struct page *page)
{
	kvm_set_pfn_dirty(page_to_pfn(page));
}
EXPORT_SYMBOL_GPL(kvm_set_page_dirty);

void kvm_set_pfn_dirty(pfn_t pfn)
{
1268
	if (!kvm_is_mmio_pfn(pfn)) {
1269 1270 1271 1272
		struct page *page = pfn_to_page(pfn);
		if (!PageReserved(page))
			SetPageDirty(page);
	}
1273
}
1274 1275 1276 1277
EXPORT_SYMBOL_GPL(kvm_set_pfn_dirty);

void kvm_set_pfn_accessed(pfn_t pfn)
{
1278
	if (!kvm_is_mmio_pfn(pfn))
1279
		mark_page_accessed(pfn_to_page(pfn));
1280 1281 1282 1283 1284
}
EXPORT_SYMBOL_GPL(kvm_set_pfn_accessed);

void kvm_get_pfn(pfn_t pfn)
{
1285
	if (!kvm_is_mmio_pfn(pfn))
1286
		get_page(pfn_to_page(pfn));
1287 1288
}
EXPORT_SYMBOL_GPL(kvm_get_pfn);
1289

1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
static int next_segment(unsigned long len, int offset)
{
	if (len > PAGE_SIZE - offset)
		return PAGE_SIZE - offset;
	else
		return len;
}

int kvm_read_guest_page(struct kvm *kvm, gfn_t gfn, void *data, int offset,
			int len)
{
1301 1302
	int r;
	unsigned long addr;
1303

1304 1305 1306
	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr))
		return -EFAULT;
1307
	r = __copy_from_user(data, (void __user *)addr + offset, len);
1308
	if (r)
1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333
		return -EFAULT;
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_read_guest_page);

int kvm_read_guest(struct kvm *kvm, gpa_t gpa, void *data, unsigned long len)
{
	gfn_t gfn = gpa >> PAGE_SHIFT;
	int seg;
	int offset = offset_in_page(gpa);
	int ret;

	while ((seg = next_segment(len, offset)) != 0) {
		ret = kvm_read_guest_page(kvm, gfn, data, offset, seg);
		if (ret < 0)
			return ret;
		offset = 0;
		len -= seg;
		data += seg;
		++gfn;
	}
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_read_guest);

1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
int kvm_read_guest_atomic(struct kvm *kvm, gpa_t gpa, void *data,
			  unsigned long len)
{
	int r;
	unsigned long addr;
	gfn_t gfn = gpa >> PAGE_SHIFT;
	int offset = offset_in_page(gpa);

	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr))
		return -EFAULT;
1345
	pagefault_disable();
1346
	r = __copy_from_user_inatomic(data, (void __user *)addr + offset, len);
1347
	pagefault_enable();
1348 1349 1350 1351 1352 1353
	if (r)
		return -EFAULT;
	return 0;
}
EXPORT_SYMBOL(kvm_read_guest_atomic);

1354 1355 1356
int kvm_write_guest_page(struct kvm *kvm, gfn_t gfn, const void *data,
			 int offset, int len)
{
1357 1358
	int r;
	unsigned long addr;
1359

1360 1361 1362
	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr))
		return -EFAULT;
1363
	r = __copy_to_user((void __user *)addr + offset, data, len);
1364
	if (r)
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390
		return -EFAULT;
	mark_page_dirty(kvm, gfn);
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_write_guest_page);

int kvm_write_guest(struct kvm *kvm, gpa_t gpa, const void *data,
		    unsigned long len)
{
	gfn_t gfn = gpa >> PAGE_SHIFT;
	int seg;
	int offset = offset_in_page(gpa);
	int ret;

	while ((seg = next_segment(len, offset)) != 0) {
		ret = kvm_write_guest_page(kvm, gfn, data, offset, seg);
		if (ret < 0)
			return ret;
		offset = 0;
		len -= seg;
		data += seg;
		++gfn;
	}
	return 0;
}

1391 1392 1393 1394 1395 1396 1397 1398 1399
int kvm_gfn_to_hva_cache_init(struct kvm *kvm, struct gfn_to_hva_cache *ghc,
			      gpa_t gpa)
{
	struct kvm_memslots *slots = kvm_memslots(kvm);
	int offset = offset_in_page(gpa);
	gfn_t gfn = gpa >> PAGE_SHIFT;

	ghc->gpa = gpa;
	ghc->generation = slots->generation;
1400
	ghc->memslot = gfn_to_memslot(kvm, gfn);
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422
	ghc->hva = gfn_to_hva_many(ghc->memslot, gfn, NULL);
	if (!kvm_is_error_hva(ghc->hva))
		ghc->hva += offset;
	else
		return -EFAULT;

	return 0;
}
EXPORT_SYMBOL_GPL(kvm_gfn_to_hva_cache_init);

int kvm_write_guest_cached(struct kvm *kvm, struct gfn_to_hva_cache *ghc,
			   void *data, unsigned long len)
{
	struct kvm_memslots *slots = kvm_memslots(kvm);
	int r;

	if (slots->generation != ghc->generation)
		kvm_gfn_to_hva_cache_init(kvm, ghc, ghc->gpa);

	if (kvm_is_error_hva(ghc->hva))
		return -EFAULT;

1423
	r = __copy_to_user((void __user *)ghc->hva, data, len);
1424 1425 1426 1427 1428 1429 1430 1431
	if (r)
		return -EFAULT;
	mark_page_dirty_in_slot(kvm, ghc->memslot, ghc->gpa >> PAGE_SHIFT);

	return 0;
}
EXPORT_SYMBOL_GPL(kvm_write_guest_cached);

1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
int kvm_read_guest_cached(struct kvm *kvm, struct gfn_to_hva_cache *ghc,
			   void *data, unsigned long len)
{
	struct kvm_memslots *slots = kvm_memslots(kvm);
	int r;

	if (slots->generation != ghc->generation)
		kvm_gfn_to_hva_cache_init(kvm, ghc, ghc->gpa);

	if (kvm_is_error_hva(ghc->hva))
		return -EFAULT;

	r = __copy_from_user(data, (void __user *)ghc->hva, len);
	if (r)
		return -EFAULT;

	return 0;
}
EXPORT_SYMBOL_GPL(kvm_read_guest_cached);

1452 1453
int kvm_clear_guest_page(struct kvm *kvm, gfn_t gfn, int offset, int len)
{
1454 1455
	return kvm_write_guest_page(kvm, gfn, (const void *) empty_zero_page,
				    offset, len);
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477
}
EXPORT_SYMBOL_GPL(kvm_clear_guest_page);

int kvm_clear_guest(struct kvm *kvm, gpa_t gpa, unsigned long len)
{
	gfn_t gfn = gpa >> PAGE_SHIFT;
	int seg;
	int offset = offset_in_page(gpa);
	int ret;

        while ((seg = next_segment(len, offset)) != 0) {
		ret = kvm_clear_guest_page(kvm, gfn, offset, seg);
		if (ret < 0)
			return ret;
		offset = 0;
		len -= seg;
		++gfn;
	}
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_clear_guest);

1478 1479
void mark_page_dirty_in_slot(struct kvm *kvm, struct kvm_memory_slot *memslot,
			     gfn_t gfn)
A
Avi Kivity 已提交
1480
{
R
Rusty Russell 已提交
1481 1482
	if (memslot && memslot->dirty_bitmap) {
		unsigned long rel_gfn = gfn - memslot->base_gfn;
A
Avi Kivity 已提交
1483

1484 1485
		/* TODO: introduce set_bit_le() and use it */
		test_and_set_bit_le(rel_gfn, memslot->dirty_bitmap);
A
Avi Kivity 已提交
1486 1487 1488
	}
}

1489 1490 1491 1492 1493 1494 1495 1496
void mark_page_dirty(struct kvm *kvm, gfn_t gfn)
{
	struct kvm_memory_slot *memslot;

	memslot = gfn_to_memslot(kvm, gfn);
	mark_page_dirty_in_slot(kvm, memslot, gfn);
}

E
Eddie Dong 已提交
1497 1498 1499
/*
 * The vCPU has executed a HLT instruction with in-kernel mode enabled.
 */
1500
void kvm_vcpu_block(struct kvm_vcpu *vcpu)
1501
{
1502 1503 1504 1505 1506
	DEFINE_WAIT(wait);

	for (;;) {
		prepare_to_wait(&vcpu->wq, &wait, TASK_INTERRUPTIBLE);

1507
		if (kvm_arch_vcpu_runnable(vcpu)) {
1508
			kvm_make_request(KVM_REQ_UNHALT, vcpu);
1509
			break;
1510
		}
1511 1512
		if (kvm_cpu_has_pending_timer(vcpu))
			break;
1513 1514 1515
		if (signal_pending(current))
			break;

E
Eddie Dong 已提交
1516 1517
		schedule();
	}
1518

1519
	finish_wait(&vcpu->wq, &wait);
E
Eddie Dong 已提交
1520 1521
}

1522
#ifndef CONFIG_S390
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
/*
 * Kick a sleeping VCPU, or a guest VCPU in guest mode, into host kernel mode.
 */
void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
{
	int me;
	int cpu = vcpu->cpu;
	wait_queue_head_t *wqp;

	wqp = kvm_arch_vcpu_wq(vcpu);
	if (waitqueue_active(wqp)) {
		wake_up_interruptible(wqp);
		++vcpu->stat.halt_wakeup;
	}

	me = get_cpu();
	if (cpu != me && (unsigned)cpu < nr_cpu_ids && cpu_online(cpu))
		if (kvm_arch_vcpu_should_kick(vcpu))
			smp_send_reschedule(cpu);
	put_cpu();
}
1544
#endif /* !CONFIG_S390 */
1545

A
Avi Kivity 已提交
1546 1547
void kvm_resched(struct kvm_vcpu *vcpu)
{
1548 1549
	if (!need_resched())
		return;
A
Avi Kivity 已提交
1550 1551 1552 1553
	cond_resched();
}
EXPORT_SYMBOL_GPL(kvm_resched);

1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578
bool kvm_vcpu_yield_to(struct kvm_vcpu *target)
{
	struct pid *pid;
	struct task_struct *task = NULL;

	rcu_read_lock();
	pid = rcu_dereference(target->pid);
	if (pid)
		task = get_pid_task(target->pid, PIDTYPE_PID);
	rcu_read_unlock();
	if (!task)
		return false;
	if (task->flags & PF_VCPU) {
		put_task_struct(task);
		return false;
	}
	if (yield_to(task, 1)) {
		put_task_struct(task);
		return true;
	}
	put_task_struct(task);
	return false;
}
EXPORT_SYMBOL_GPL(kvm_vcpu_yield_to);

1579
void kvm_vcpu_on_spin(struct kvm_vcpu *me)
Z
Zhai, Edwin 已提交
1580
{
1581 1582 1583 1584 1585 1586
	struct kvm *kvm = me->kvm;
	struct kvm_vcpu *vcpu;
	int last_boosted_vcpu = me->kvm->last_boosted_vcpu;
	int yielded = 0;
	int pass;
	int i;
Z
Zhai, Edwin 已提交
1587

1588 1589 1590 1591 1592 1593 1594 1595 1596
	/*
	 * We boost the priority of a VCPU that is runnable but not
	 * currently running, because it got preempted by something
	 * else and called schedule in __vcpu_run.  Hopefully that
	 * VCPU is holding the lock that we need and will release it.
	 * We approximate round-robin by starting at the last boosted VCPU.
	 */
	for (pass = 0; pass < 2 && !yielded; pass++) {
		kvm_for_each_vcpu(i, vcpu, kvm) {
1597
			if (!pass && i <= last_boosted_vcpu) {
1598 1599 1600 1601 1602 1603 1604 1605
				i = last_boosted_vcpu;
				continue;
			} else if (pass && i > last_boosted_vcpu)
				break;
			if (vcpu == me)
				continue;
			if (waitqueue_active(&vcpu->wq))
				continue;
1606
			if (kvm_vcpu_yield_to(vcpu)) {
1607 1608 1609 1610 1611 1612
				kvm->last_boosted_vcpu = i;
				yielded = 1;
				break;
			}
		}
	}
Z
Zhai, Edwin 已提交
1613 1614 1615
}
EXPORT_SYMBOL_GPL(kvm_vcpu_on_spin);

1616
static int kvm_vcpu_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
1617 1618 1619 1620
{
	struct kvm_vcpu *vcpu = vma->vm_file->private_data;
	struct page *page;

1621
	if (vmf->pgoff == 0)
1622
		page = virt_to_page(vcpu->run);
A
Avi Kivity 已提交
1623
#ifdef CONFIG_X86
1624
	else if (vmf->pgoff == KVM_PIO_PAGE_OFFSET)
1625
		page = virt_to_page(vcpu->arch.pio_data);
1626 1627 1628 1629
#endif
#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	else if (vmf->pgoff == KVM_COALESCED_MMIO_PAGE_OFFSET)
		page = virt_to_page(vcpu->kvm->coalesced_mmio_ring);
A
Avi Kivity 已提交
1630
#endif
1631
	else
1632
		return kvm_arch_vcpu_fault(vcpu, vmf);
1633
	get_page(page);
1634 1635
	vmf->page = page;
	return 0;
1636 1637
}

1638
static const struct vm_operations_struct kvm_vcpu_vm_ops = {
1639
	.fault = kvm_vcpu_fault,
1640 1641 1642 1643 1644 1645 1646 1647
};

static int kvm_vcpu_mmap(struct file *file, struct vm_area_struct *vma)
{
	vma->vm_ops = &kvm_vcpu_vm_ops;
	return 0;
}

A
Avi Kivity 已提交
1648 1649 1650 1651
static int kvm_vcpu_release(struct inode *inode, struct file *filp)
{
	struct kvm_vcpu *vcpu = filp->private_data;

A
Al Viro 已提交
1652
	kvm_put_kvm(vcpu->kvm);
A
Avi Kivity 已提交
1653 1654 1655
	return 0;
}

1656
static struct file_operations kvm_vcpu_fops = {
A
Avi Kivity 已提交
1657 1658
	.release        = kvm_vcpu_release,
	.unlocked_ioctl = kvm_vcpu_ioctl,
1659 1660 1661
#ifdef CONFIG_COMPAT
	.compat_ioctl   = kvm_vcpu_compat_ioctl,
#endif
1662
	.mmap           = kvm_vcpu_mmap,
1663
	.llseek		= noop_llseek,
A
Avi Kivity 已提交
1664 1665 1666 1667 1668 1669 1670
};

/*
 * Allocates an inode for the vcpu.
 */
static int create_vcpu_fd(struct kvm_vcpu *vcpu)
{
1671
	return anon_inode_getfd("kvm-vcpu", &kvm_vcpu_fops, vcpu, O_RDWR);
A
Avi Kivity 已提交
1672 1673
}

1674 1675 1676
/*
 * Creates some virtual cpus.  Good luck creating more than one.
 */
1677
static int kvm_vm_ioctl_create_vcpu(struct kvm *kvm, u32 id)
1678 1679
{
	int r;
1680
	struct kvm_vcpu *vcpu, *v;
1681

1682
	vcpu = kvm_arch_vcpu_create(kvm, id);
R
Rusty Russell 已提交
1683 1684
	if (IS_ERR(vcpu))
		return PTR_ERR(vcpu);
1685

1686 1687
	preempt_notifier_init(&vcpu->preempt_notifier, &kvm_preempt_ops);

1688 1689
	r = kvm_arch_vcpu_setup(vcpu);
	if (r)
1690
		goto vcpu_destroy;
1691

S
Shaohua Li 已提交
1692
	mutex_lock(&kvm->lock);
1693 1694 1695 1696
	if (!kvm_vcpu_compatible(vcpu)) {
		r = -EINVAL;
		goto unlock_vcpu_destroy;
	}
1697 1698
	if (atomic_read(&kvm->online_vcpus) == KVM_MAX_VCPUS) {
		r = -EINVAL;
1699
		goto unlock_vcpu_destroy;
R
Rusty Russell 已提交
1700
	}
1701

1702 1703
	kvm_for_each_vcpu(r, v, kvm)
		if (v->vcpu_id == id) {
1704
			r = -EEXIST;
1705
			goto unlock_vcpu_destroy;
1706 1707 1708
		}

	BUG_ON(kvm->vcpus[atomic_read(&kvm->online_vcpus)]);
1709

R
Rusty Russell 已提交
1710
	/* Now it's all set up, let userspace reach it */
A
Al Viro 已提交
1711
	kvm_get_kvm(kvm);
A
Avi Kivity 已提交
1712
	r = create_vcpu_fd(vcpu);
1713 1714
	if (r < 0) {
		kvm_put_kvm(kvm);
1715
		goto unlock_vcpu_destroy;
1716 1717 1718 1719 1720 1721 1722
	}

	kvm->vcpus[atomic_read(&kvm->online_vcpus)] = vcpu;
	smp_wmb();
	atomic_inc(&kvm->online_vcpus);

	mutex_unlock(&kvm->lock);
R
Rusty Russell 已提交
1723
	return r;
1724

1725
unlock_vcpu_destroy:
1726
	mutex_unlock(&kvm->lock);
1727
vcpu_destroy:
1728
	kvm_arch_vcpu_destroy(vcpu);
1729 1730 1731
	return r;
}

A
Avi Kivity 已提交
1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
static int kvm_vcpu_ioctl_set_sigmask(struct kvm_vcpu *vcpu, sigset_t *sigset)
{
	if (sigset) {
		sigdelsetmask(sigset, sigmask(SIGKILL)|sigmask(SIGSTOP));
		vcpu->sigset_active = 1;
		vcpu->sigset = *sigset;
	} else
		vcpu->sigset_active = 0;
	return 0;
}

A
Avi Kivity 已提交
1743 1744
static long kvm_vcpu_ioctl(struct file *filp,
			   unsigned int ioctl, unsigned long arg)
A
Avi Kivity 已提交
1745
{
A
Avi Kivity 已提交
1746
	struct kvm_vcpu *vcpu = filp->private_data;
A
Al Viro 已提交
1747
	void __user *argp = (void __user *)arg;
1748
	int r;
1749 1750
	struct kvm_fpu *fpu = NULL;
	struct kvm_sregs *kvm_sregs = NULL;
A
Avi Kivity 已提交
1751

1752 1753
	if (vcpu->kvm->mm != current->mm)
		return -EIO;
1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765

#if defined(CONFIG_S390) || defined(CONFIG_PPC)
	/*
	 * Special cases: vcpu ioctls that are asynchronous to vcpu execution,
	 * so vcpu_load() would break it.
	 */
	if (ioctl == KVM_S390_INTERRUPT || ioctl == KVM_INTERRUPT)
		return kvm_arch_vcpu_ioctl(filp, ioctl, arg);
#endif


	vcpu_load(vcpu);
A
Avi Kivity 已提交
1766
	switch (ioctl) {
1767
	case KVM_RUN:
1768 1769 1770
		r = -EINVAL;
		if (arg)
			goto out;
1771
		r = kvm_arch_vcpu_ioctl_run(vcpu, vcpu->run);
1772
		trace_kvm_userspace_exit(vcpu->run->exit_reason, r);
A
Avi Kivity 已提交
1773 1774
		break;
	case KVM_GET_REGS: {
1775
		struct kvm_regs *kvm_regs;
A
Avi Kivity 已提交
1776

1777 1778 1779
		r = -ENOMEM;
		kvm_regs = kzalloc(sizeof(struct kvm_regs), GFP_KERNEL);
		if (!kvm_regs)
A
Avi Kivity 已提交
1780
			goto out;
1781 1782 1783
		r = kvm_arch_vcpu_ioctl_get_regs(vcpu, kvm_regs);
		if (r)
			goto out_free1;
A
Avi Kivity 已提交
1784
		r = -EFAULT;
1785 1786
		if (copy_to_user(argp, kvm_regs, sizeof(struct kvm_regs)))
			goto out_free1;
A
Avi Kivity 已提交
1787
		r = 0;
1788 1789
out_free1:
		kfree(kvm_regs);
A
Avi Kivity 已提交
1790 1791 1792
		break;
	}
	case KVM_SET_REGS: {
1793
		struct kvm_regs *kvm_regs;
A
Avi Kivity 已提交
1794

1795
		r = -ENOMEM;
1796 1797 1798
		kvm_regs = memdup_user(argp, sizeof(*kvm_regs));
		if (IS_ERR(kvm_regs)) {
			r = PTR_ERR(kvm_regs);
A
Avi Kivity 已提交
1799
			goto out;
1800
		}
1801
		r = kvm_arch_vcpu_ioctl_set_regs(vcpu, kvm_regs);
A
Avi Kivity 已提交
1802
		if (r)
1803
			goto out_free2;
A
Avi Kivity 已提交
1804
		r = 0;
1805 1806
out_free2:
		kfree(kvm_regs);
A
Avi Kivity 已提交
1807 1808 1809
		break;
	}
	case KVM_GET_SREGS: {
1810 1811 1812 1813 1814
		kvm_sregs = kzalloc(sizeof(struct kvm_sregs), GFP_KERNEL);
		r = -ENOMEM;
		if (!kvm_sregs)
			goto out;
		r = kvm_arch_vcpu_ioctl_get_sregs(vcpu, kvm_sregs);
A
Avi Kivity 已提交
1815 1816 1817
		if (r)
			goto out;
		r = -EFAULT;
1818
		if (copy_to_user(argp, kvm_sregs, sizeof(struct kvm_sregs)))
A
Avi Kivity 已提交
1819 1820 1821 1822 1823
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_SREGS: {
1824 1825 1826
		kvm_sregs = memdup_user(argp, sizeof(*kvm_sregs));
		if (IS_ERR(kvm_sregs)) {
			r = PTR_ERR(kvm_sregs);
A
Avi Kivity 已提交
1827
			goto out;
1828
		}
1829
		r = kvm_arch_vcpu_ioctl_set_sregs(vcpu, kvm_sregs);
A
Avi Kivity 已提交
1830 1831 1832 1833 1834
		if (r)
			goto out;
		r = 0;
		break;
	}
1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858
	case KVM_GET_MP_STATE: {
		struct kvm_mp_state mp_state;

		r = kvm_arch_vcpu_ioctl_get_mpstate(vcpu, &mp_state);
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(argp, &mp_state, sizeof mp_state))
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_MP_STATE: {
		struct kvm_mp_state mp_state;

		r = -EFAULT;
		if (copy_from_user(&mp_state, argp, sizeof mp_state))
			goto out;
		r = kvm_arch_vcpu_ioctl_set_mpstate(vcpu, &mp_state);
		if (r)
			goto out;
		r = 0;
		break;
	}
A
Avi Kivity 已提交
1859 1860 1861 1862
	case KVM_TRANSLATE: {
		struct kvm_translation tr;

		r = -EFAULT;
A
Al Viro 已提交
1863
		if (copy_from_user(&tr, argp, sizeof tr))
A
Avi Kivity 已提交
1864
			goto out;
1865
		r = kvm_arch_vcpu_ioctl_translate(vcpu, &tr);
A
Avi Kivity 已提交
1866 1867 1868
		if (r)
			goto out;
		r = -EFAULT;
A
Al Viro 已提交
1869
		if (copy_to_user(argp, &tr, sizeof tr))
A
Avi Kivity 已提交
1870 1871 1872 1873
			goto out;
		r = 0;
		break;
	}
J
Jan Kiszka 已提交
1874 1875
	case KVM_SET_GUEST_DEBUG: {
		struct kvm_guest_debug dbg;
A
Avi Kivity 已提交
1876 1877

		r = -EFAULT;
A
Al Viro 已提交
1878
		if (copy_from_user(&dbg, argp, sizeof dbg))
A
Avi Kivity 已提交
1879
			goto out;
J
Jan Kiszka 已提交
1880
		r = kvm_arch_vcpu_ioctl_set_guest_debug(vcpu, &dbg);
A
Avi Kivity 已提交
1881 1882 1883 1884 1885
		if (r)
			goto out;
		r = 0;
		break;
	}
A
Avi Kivity 已提交
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
	case KVM_SET_SIGNAL_MASK: {
		struct kvm_signal_mask __user *sigmask_arg = argp;
		struct kvm_signal_mask kvm_sigmask;
		sigset_t sigset, *p;

		p = NULL;
		if (argp) {
			r = -EFAULT;
			if (copy_from_user(&kvm_sigmask, argp,
					   sizeof kvm_sigmask))
				goto out;
			r = -EINVAL;
			if (kvm_sigmask.len != sizeof sigset)
				goto out;
			r = -EFAULT;
			if (copy_from_user(&sigset, sigmask_arg->sigset,
					   sizeof sigset))
				goto out;
			p = &sigset;
		}
1906
		r = kvm_vcpu_ioctl_set_sigmask(vcpu, p);
A
Avi Kivity 已提交
1907 1908
		break;
	}
A
Avi Kivity 已提交
1909
	case KVM_GET_FPU: {
1910 1911 1912 1913 1914
		fpu = kzalloc(sizeof(struct kvm_fpu), GFP_KERNEL);
		r = -ENOMEM;
		if (!fpu)
			goto out;
		r = kvm_arch_vcpu_ioctl_get_fpu(vcpu, fpu);
A
Avi Kivity 已提交
1915 1916 1917
		if (r)
			goto out;
		r = -EFAULT;
1918
		if (copy_to_user(argp, fpu, sizeof(struct kvm_fpu)))
A
Avi Kivity 已提交
1919 1920 1921 1922 1923
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_FPU: {
1924 1925 1926
		fpu = memdup_user(argp, sizeof(*fpu));
		if (IS_ERR(fpu)) {
			r = PTR_ERR(fpu);
A
Avi Kivity 已提交
1927
			goto out;
1928
		}
1929
		r = kvm_arch_vcpu_ioctl_set_fpu(vcpu, fpu);
A
Avi Kivity 已提交
1930 1931 1932 1933 1934
		if (r)
			goto out;
		r = 0;
		break;
	}
A
Avi Kivity 已提交
1935
	default:
1936
		r = kvm_arch_vcpu_ioctl(filp, ioctl, arg);
A
Avi Kivity 已提交
1937 1938
	}
out:
1939
	vcpu_put(vcpu);
1940 1941
	kfree(fpu);
	kfree(kvm_sregs);
A
Avi Kivity 已提交
1942 1943 1944
	return r;
}

1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
#ifdef CONFIG_COMPAT
static long kvm_vcpu_compat_ioctl(struct file *filp,
				  unsigned int ioctl, unsigned long arg)
{
	struct kvm_vcpu *vcpu = filp->private_data;
	void __user *argp = compat_ptr(arg);
	int r;

	if (vcpu->kvm->mm != current->mm)
		return -EIO;

	switch (ioctl) {
	case KVM_SET_SIGNAL_MASK: {
		struct kvm_signal_mask __user *sigmask_arg = argp;
		struct kvm_signal_mask kvm_sigmask;
		compat_sigset_t csigset;
		sigset_t sigset;

		if (argp) {
			r = -EFAULT;
			if (copy_from_user(&kvm_sigmask, argp,
					   sizeof kvm_sigmask))
				goto out;
			r = -EINVAL;
			if (kvm_sigmask.len != sizeof csigset)
				goto out;
			r = -EFAULT;
			if (copy_from_user(&csigset, sigmask_arg->sigset,
					   sizeof csigset))
				goto out;
		}
		sigset_from_compat(&sigset, &csigset);
		r = kvm_vcpu_ioctl_set_sigmask(vcpu, &sigset);
		break;
	}
	default:
		r = kvm_vcpu_ioctl(filp, ioctl, arg);
	}

out:
	return r;
}
#endif

A
Avi Kivity 已提交
1989 1990 1991 1992 1993
static long kvm_vm_ioctl(struct file *filp,
			   unsigned int ioctl, unsigned long arg)
{
	struct kvm *kvm = filp->private_data;
	void __user *argp = (void __user *)arg;
1994
	int r;
A
Avi Kivity 已提交
1995

1996 1997
	if (kvm->mm != current->mm)
		return -EIO;
A
Avi Kivity 已提交
1998 1999 2000 2001 2002 2003
	switch (ioctl) {
	case KVM_CREATE_VCPU:
		r = kvm_vm_ioctl_create_vcpu(kvm, arg);
		if (r < 0)
			goto out;
		break;
2004 2005 2006 2007 2008 2009 2010 2011 2012
	case KVM_SET_USER_MEMORY_REGION: {
		struct kvm_userspace_memory_region kvm_userspace_mem;

		r = -EFAULT;
		if (copy_from_user(&kvm_userspace_mem, argp,
						sizeof kvm_userspace_mem))
			goto out;

		r = kvm_vm_ioctl_set_memory_region(kvm, &kvm_userspace_mem, 1);
A
Avi Kivity 已提交
2013 2014 2015 2016 2017 2018 2019 2020
		if (r)
			goto out;
		break;
	}
	case KVM_GET_DIRTY_LOG: {
		struct kvm_dirty_log log;

		r = -EFAULT;
A
Al Viro 已提交
2021
		if (copy_from_user(&log, argp, sizeof log))
A
Avi Kivity 已提交
2022
			goto out;
2023
		r = kvm_vm_ioctl_get_dirty_log(kvm, &log);
A
Avi Kivity 已提交
2024 2025 2026 2027
		if (r)
			goto out;
		break;
	}
2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051
#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	case KVM_REGISTER_COALESCED_MMIO: {
		struct kvm_coalesced_mmio_zone zone;
		r = -EFAULT;
		if (copy_from_user(&zone, argp, sizeof zone))
			goto out;
		r = kvm_vm_ioctl_register_coalesced_mmio(kvm, &zone);
		if (r)
			goto out;
		r = 0;
		break;
	}
	case KVM_UNREGISTER_COALESCED_MMIO: {
		struct kvm_coalesced_mmio_zone zone;
		r = -EFAULT;
		if (copy_from_user(&zone, argp, sizeof zone))
			goto out;
		r = kvm_vm_ioctl_unregister_coalesced_mmio(kvm, &zone);
		if (r)
			goto out;
		r = 0;
		break;
	}
#endif
G
Gregory Haskins 已提交
2052 2053 2054 2055 2056 2057 2058 2059 2060
	case KVM_IRQFD: {
		struct kvm_irqfd data;

		r = -EFAULT;
		if (copy_from_user(&data, argp, sizeof data))
			goto out;
		r = kvm_irqfd(kvm, data.fd, data.gsi, data.flags);
		break;
	}
G
Gregory Haskins 已提交
2061 2062 2063 2064 2065 2066 2067 2068 2069
	case KVM_IOEVENTFD: {
		struct kvm_ioeventfd data;

		r = -EFAULT;
		if (copy_from_user(&data, argp, sizeof data))
			goto out;
		r = kvm_ioeventfd(kvm, &data);
		break;
	}
2070 2071 2072
#ifdef CONFIG_KVM_APIC_ARCHITECTURE
	case KVM_SET_BOOT_CPU_ID:
		r = 0;
2073
		mutex_lock(&kvm->lock);
2074 2075 2076 2077
		if (atomic_read(&kvm->online_vcpus) != 0)
			r = -EBUSY;
		else
			kvm->bsp_vcpu_id = arg;
2078
		mutex_unlock(&kvm->lock);
2079
		break;
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
#endif
#ifdef CONFIG_HAVE_KVM_MSI
	case KVM_SIGNAL_MSI: {
		struct kvm_msi msi;

		r = -EFAULT;
		if (copy_from_user(&msi, argp, sizeof msi))
			goto out;
		r = kvm_send_userspace_msi(kvm, &msi);
		break;
	}
2091
#endif
2092
	default:
2093
		r = kvm_arch_vm_ioctl(filp, ioctl, arg);
2094 2095
		if (r == -ENOTTY)
			r = kvm_vm_ioctl_assigned_device(kvm, ioctl, arg);
2096 2097 2098 2099 2100
	}
out:
	return r;
}

2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
#ifdef CONFIG_COMPAT
struct compat_kvm_dirty_log {
	__u32 slot;
	__u32 padding1;
	union {
		compat_uptr_t dirty_bitmap; /* one bit per page */
		__u64 padding2;
	};
};

static long kvm_vm_compat_ioctl(struct file *filp,
			   unsigned int ioctl, unsigned long arg)
{
	struct kvm *kvm = filp->private_data;
	int r;

	if (kvm->mm != current->mm)
		return -EIO;
	switch (ioctl) {
	case KVM_GET_DIRTY_LOG: {
		struct compat_kvm_dirty_log compat_log;
		struct kvm_dirty_log log;

		r = -EFAULT;
		if (copy_from_user(&compat_log, (void __user *)arg,
				   sizeof(compat_log)))
			goto out;
		log.slot	 = compat_log.slot;
		log.padding1	 = compat_log.padding1;
		log.padding2	 = compat_log.padding2;
		log.dirty_bitmap = compat_ptr(compat_log.dirty_bitmap);

		r = kvm_vm_ioctl_get_dirty_log(kvm, &log);
		if (r)
			goto out;
		break;
	}
	default:
		r = kvm_vm_ioctl(filp, ioctl, arg);
	}

out:
	return r;
}
#endif

2147
static int kvm_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
2148
{
2149 2150 2151 2152
	struct page *page[1];
	unsigned long addr;
	int npages;
	gfn_t gfn = vmf->pgoff;
2153 2154
	struct kvm *kvm = vma->vm_file->private_data;

2155 2156
	addr = gfn_to_hva(kvm, gfn);
	if (kvm_is_error_hva(addr))
2157
		return VM_FAULT_SIGBUS;
2158 2159 2160 2161

	npages = get_user_pages(current, current->mm, addr, 1, 1, 0, page,
				NULL);
	if (unlikely(npages != 1))
2162
		return VM_FAULT_SIGBUS;
2163 2164

	vmf->page = page[0];
2165
	return 0;
2166 2167
}

2168
static const struct vm_operations_struct kvm_vm_vm_ops = {
2169
	.fault = kvm_vm_fault,
2170 2171 2172 2173 2174 2175 2176 2177
};

static int kvm_vm_mmap(struct file *file, struct vm_area_struct *vma)
{
	vma->vm_ops = &kvm_vm_vm_ops;
	return 0;
}

2178
static struct file_operations kvm_vm_fops = {
2179 2180
	.release        = kvm_vm_release,
	.unlocked_ioctl = kvm_vm_ioctl,
2181 2182 2183
#ifdef CONFIG_COMPAT
	.compat_ioctl   = kvm_vm_compat_ioctl,
#endif
2184
	.mmap           = kvm_vm_mmap,
2185
	.llseek		= noop_llseek,
2186 2187
};

2188
static int kvm_dev_ioctl_create_vm(unsigned long type)
2189
{
2190
	int r;
2191 2192
	struct kvm *kvm;

2193
	kvm = kvm_create_vm(type);
2194 2195
	if (IS_ERR(kvm))
		return PTR_ERR(kvm);
2196 2197 2198 2199 2200 2201 2202
#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	r = kvm_coalesced_mmio_init(kvm);
	if (r < 0) {
		kvm_put_kvm(kvm);
		return r;
	}
#endif
2203 2204
	r = anon_inode_getfd("kvm-vm", &kvm_vm_fops, kvm, O_RDWR);
	if (r < 0)
A
Al Viro 已提交
2205
		kvm_put_kvm(kvm);
2206

2207
	return r;
2208 2209
}

2210 2211 2212
static long kvm_dev_ioctl_check_extension_generic(long arg)
{
	switch (arg) {
2213
	case KVM_CAP_USER_MEMORY:
2214
	case KVM_CAP_DESTROY_MEMORY_REGION_WORKS:
2215
	case KVM_CAP_JOIN_MEMORY_REGIONS_WORKS:
2216 2217 2218
#ifdef CONFIG_KVM_APIC_ARCHITECTURE
	case KVM_CAP_SET_BOOT_CPU_ID:
#endif
2219
	case KVM_CAP_INTERNAL_ERROR_DATA:
2220 2221 2222
#ifdef CONFIG_HAVE_KVM_MSI
	case KVM_CAP_SIGNAL_MSI:
#endif
2223
		return 1;
2224
#ifdef KVM_CAP_IRQ_ROUTING
2225
	case KVM_CAP_IRQ_ROUTING:
2226
		return KVM_MAX_IRQ_ROUTES;
2227
#endif
2228 2229 2230 2231 2232 2233
	default:
		break;
	}
	return kvm_dev_ioctl_check_extension(arg);
}

2234 2235 2236
static long kvm_dev_ioctl(struct file *filp,
			  unsigned int ioctl, unsigned long arg)
{
2237
	long r = -EINVAL;
2238 2239 2240

	switch (ioctl) {
	case KVM_GET_API_VERSION:
2241 2242 2243
		r = -EINVAL;
		if (arg)
			goto out;
2244 2245 2246
		r = KVM_API_VERSION;
		break;
	case KVM_CREATE_VM:
2247
		r = kvm_dev_ioctl_create_vm(arg);
2248
		break;
2249
	case KVM_CHECK_EXTENSION:
2250
		r = kvm_dev_ioctl_check_extension_generic(arg);
2251
		break;
2252 2253 2254 2255
	case KVM_GET_VCPU_MMAP_SIZE:
		r = -EINVAL;
		if (arg)
			goto out;
2256 2257 2258
		r = PAGE_SIZE;     /* struct kvm_run */
#ifdef CONFIG_X86
		r += PAGE_SIZE;    /* pio data page */
2259 2260 2261
#endif
#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
		r += PAGE_SIZE;    /* coalesced mmio ring page */
2262
#endif
2263
		break;
2264 2265 2266
	case KVM_TRACE_ENABLE:
	case KVM_TRACE_PAUSE:
	case KVM_TRACE_DISABLE:
2267
		r = -EOPNOTSUPP;
2268
		break;
A
Avi Kivity 已提交
2269
	default:
2270
		return kvm_arch_dev_ioctl(filp, ioctl, arg);
A
Avi Kivity 已提交
2271 2272 2273 2274 2275 2276 2277 2278
	}
out:
	return r;
}

static struct file_operations kvm_chardev_ops = {
	.unlocked_ioctl = kvm_dev_ioctl,
	.compat_ioctl   = kvm_dev_ioctl,
2279
	.llseek		= noop_llseek,
A
Avi Kivity 已提交
2280 2281 2282
};

static struct miscdevice kvm_dev = {
A
Avi Kivity 已提交
2283
	KVM_MINOR,
A
Avi Kivity 已提交
2284 2285 2286 2287
	"kvm",
	&kvm_chardev_ops,
};

2288
static void hardware_enable_nolock(void *junk)
2289 2290
{
	int cpu = raw_smp_processor_id();
2291
	int r;
2292

2293
	if (cpumask_test_cpu(cpu, cpus_hardware_enabled))
2294
		return;
2295

2296
	cpumask_set_cpu(cpu, cpus_hardware_enabled);
2297 2298 2299 2300 2301 2302 2303 2304 2305

	r = kvm_arch_hardware_enable(NULL);

	if (r) {
		cpumask_clear_cpu(cpu, cpus_hardware_enabled);
		atomic_inc(&hardware_enable_failed);
		printk(KERN_INFO "kvm: enabling virtualization on "
				 "CPU%d failed\n", cpu);
	}
2306 2307
}

2308 2309
static void hardware_enable(void *junk)
{
2310
	raw_spin_lock(&kvm_lock);
2311
	hardware_enable_nolock(junk);
2312
	raw_spin_unlock(&kvm_lock);
2313 2314 2315
}

static void hardware_disable_nolock(void *junk)
2316 2317 2318
{
	int cpu = raw_smp_processor_id();

2319
	if (!cpumask_test_cpu(cpu, cpus_hardware_enabled))
2320
		return;
2321
	cpumask_clear_cpu(cpu, cpus_hardware_enabled);
2322
	kvm_arch_hardware_disable(NULL);
2323 2324
}

2325 2326
static void hardware_disable(void *junk)
{
2327
	raw_spin_lock(&kvm_lock);
2328
	hardware_disable_nolock(junk);
2329
	raw_spin_unlock(&kvm_lock);
2330 2331
}

2332 2333 2334 2335 2336 2337
static void hardware_disable_all_nolock(void)
{
	BUG_ON(!kvm_usage_count);

	kvm_usage_count--;
	if (!kvm_usage_count)
2338
		on_each_cpu(hardware_disable_nolock, NULL, 1);
2339 2340 2341 2342
}

static void hardware_disable_all(void)
{
2343
	raw_spin_lock(&kvm_lock);
2344
	hardware_disable_all_nolock();
2345
	raw_spin_unlock(&kvm_lock);
2346 2347 2348 2349 2350 2351
}

static int hardware_enable_all(void)
{
	int r = 0;

2352
	raw_spin_lock(&kvm_lock);
2353 2354 2355 2356

	kvm_usage_count++;
	if (kvm_usage_count == 1) {
		atomic_set(&hardware_enable_failed, 0);
2357
		on_each_cpu(hardware_enable_nolock, NULL, 1);
2358 2359 2360 2361 2362 2363 2364

		if (atomic_read(&hardware_enable_failed)) {
			hardware_disable_all_nolock();
			r = -EBUSY;
		}
	}

2365
	raw_spin_unlock(&kvm_lock);
2366 2367 2368 2369

	return r;
}

A
Avi Kivity 已提交
2370 2371 2372 2373 2374
static int kvm_cpu_hotplug(struct notifier_block *notifier, unsigned long val,
			   void *v)
{
	int cpu = (long)v;

2375 2376 2377
	if (!kvm_usage_count)
		return NOTIFY_OK;

2378
	val &= ~CPU_TASKS_FROZEN;
A
Avi Kivity 已提交
2379
	switch (val) {
2380
	case CPU_DYING:
2381 2382 2383 2384
		printk(KERN_INFO "kvm: disabling virtualization on CPU%d\n",
		       cpu);
		hardware_disable(NULL);
		break;
2385
	case CPU_STARTING:
2386 2387
		printk(KERN_INFO "kvm: enabling virtualization on CPU%d\n",
		       cpu);
2388
		hardware_enable(NULL);
A
Avi Kivity 已提交
2389 2390 2391 2392 2393
		break;
	}
	return NOTIFY_OK;
}

2394

2395
asmlinkage void kvm_spurious_fault(void)
2396 2397 2398 2399
{
	/* Fault while not rebooting.  We want the trace. */
	BUG();
}
2400
EXPORT_SYMBOL_GPL(kvm_spurious_fault);
2401

2402
static int kvm_reboot(struct notifier_block *notifier, unsigned long val,
M
Mike Day 已提交
2403
		      void *v)
2404
{
2405 2406 2407 2408 2409 2410 2411 2412
	/*
	 * Some (well, at least mine) BIOSes hang on reboot if
	 * in vmx root mode.
	 *
	 * And Intel TXT required VMX off for all cpu when system shutdown.
	 */
	printk(KERN_INFO "kvm: exiting hardware virtualization\n");
	kvm_rebooting = true;
2413
	on_each_cpu(hardware_disable_nolock, NULL, 1);
2414 2415 2416 2417 2418 2419 2420 2421
	return NOTIFY_OK;
}

static struct notifier_block kvm_reboot_notifier = {
	.notifier_call = kvm_reboot,
	.priority = 0,
};

M
Marcelo Tosatti 已提交
2422
static void kvm_io_bus_destroy(struct kvm_io_bus *bus)
2423 2424 2425 2426
{
	int i;

	for (i = 0; i < bus->dev_count; i++) {
2427
		struct kvm_io_device *pos = bus->range[i].dev;
2428 2429 2430

		kvm_iodevice_destructor(pos);
	}
M
Marcelo Tosatti 已提交
2431
	kfree(bus);
2432 2433
}

2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484
int kvm_io_bus_sort_cmp(const void *p1, const void *p2)
{
	const struct kvm_io_range *r1 = p1;
	const struct kvm_io_range *r2 = p2;

	if (r1->addr < r2->addr)
		return -1;
	if (r1->addr + r1->len > r2->addr + r2->len)
		return 1;
	return 0;
}

int kvm_io_bus_insert_dev(struct kvm_io_bus *bus, struct kvm_io_device *dev,
			  gpa_t addr, int len)
{
	bus->range[bus->dev_count++] = (struct kvm_io_range) {
		.addr = addr,
		.len = len,
		.dev = dev,
	};

	sort(bus->range, bus->dev_count, sizeof(struct kvm_io_range),
		kvm_io_bus_sort_cmp, NULL);

	return 0;
}

int kvm_io_bus_get_first_dev(struct kvm_io_bus *bus,
			     gpa_t addr, int len)
{
	struct kvm_io_range *range, key;
	int off;

	key = (struct kvm_io_range) {
		.addr = addr,
		.len = len,
	};

	range = bsearch(&key, bus->range, bus->dev_count,
			sizeof(struct kvm_io_range), kvm_io_bus_sort_cmp);
	if (range == NULL)
		return -ENOENT;

	off = range - bus->range;

	while (off > 0 && kvm_io_bus_sort_cmp(&key, &bus->range[off-1]) == 0)
		off--;

	return off;
}

2485
/* kvm_io_bus_write - called under kvm->slots_lock */
M
Marcelo Tosatti 已提交
2486
int kvm_io_bus_write(struct kvm *kvm, enum kvm_bus bus_idx, gpa_t addr,
2487
		     int len, const void *val)
2488
{
2489
	int idx;
2490
	struct kvm_io_bus *bus;
2491 2492 2493 2494 2495 2496
	struct kvm_io_range range;

	range = (struct kvm_io_range) {
		.addr = addr,
		.len = len,
	};
2497 2498

	bus = srcu_dereference(kvm->buses[bus_idx], &kvm->srcu);
2499 2500 2501 2502 2503 2504 2505
	idx = kvm_io_bus_get_first_dev(bus, addr, len);
	if (idx < 0)
		return -EOPNOTSUPP;

	while (idx < bus->dev_count &&
		kvm_io_bus_sort_cmp(&range, &bus->range[idx]) == 0) {
		if (!kvm_iodevice_write(bus->range[idx].dev, addr, len, val))
2506
			return 0;
2507 2508 2509
		idx++;
	}

2510 2511
	return -EOPNOTSUPP;
}
2512

2513
/* kvm_io_bus_read - called under kvm->slots_lock */
M
Marcelo Tosatti 已提交
2514 2515
int kvm_io_bus_read(struct kvm *kvm, enum kvm_bus bus_idx, gpa_t addr,
		    int len, void *val)
2516
{
2517
	int idx;
2518
	struct kvm_io_bus *bus;
2519 2520 2521 2522 2523 2524
	struct kvm_io_range range;

	range = (struct kvm_io_range) {
		.addr = addr,
		.len = len,
	};
M
Marcelo Tosatti 已提交
2525

2526
	bus = srcu_dereference(kvm->buses[bus_idx], &kvm->srcu);
2527 2528 2529 2530 2531 2532 2533
	idx = kvm_io_bus_get_first_dev(bus, addr, len);
	if (idx < 0)
		return -EOPNOTSUPP;

	while (idx < bus->dev_count &&
		kvm_io_bus_sort_cmp(&range, &bus->range[idx]) == 0) {
		if (!kvm_iodevice_read(bus->range[idx].dev, addr, len, val))
2534
			return 0;
2535 2536 2537
		idx++;
	}

2538
	return -EOPNOTSUPP;
2539 2540
}

2541
/* Caller must hold slots_lock. */
2542 2543
int kvm_io_bus_register_dev(struct kvm *kvm, enum kvm_bus bus_idx, gpa_t addr,
			    int len, struct kvm_io_device *dev)
2544
{
M
Marcelo Tosatti 已提交
2545
	struct kvm_io_bus *new_bus, *bus;
2546

M
Marcelo Tosatti 已提交
2547
	bus = kvm->buses[bus_idx];
2548
	if (bus->dev_count > NR_IOBUS_DEVS - 1)
2549
		return -ENOSPC;
2550

2551 2552
	new_bus = kzalloc(sizeof(*bus) + ((bus->dev_count + 1) *
			  sizeof(struct kvm_io_range)), GFP_KERNEL);
M
Marcelo Tosatti 已提交
2553 2554
	if (!new_bus)
		return -ENOMEM;
2555 2556
	memcpy(new_bus, bus, sizeof(*bus) + (bus->dev_count *
	       sizeof(struct kvm_io_range)));
2557
	kvm_io_bus_insert_dev(new_bus, dev, addr, len);
M
Marcelo Tosatti 已提交
2558 2559 2560
	rcu_assign_pointer(kvm->buses[bus_idx], new_bus);
	synchronize_srcu_expedited(&kvm->srcu);
	kfree(bus);
2561 2562 2563 2564

	return 0;
}

2565
/* Caller must hold slots_lock. */
M
Marcelo Tosatti 已提交
2566 2567
int kvm_io_bus_unregister_dev(struct kvm *kvm, enum kvm_bus bus_idx,
			      struct kvm_io_device *dev)
2568
{
M
Marcelo Tosatti 已提交
2569 2570
	int i, r;
	struct kvm_io_bus *new_bus, *bus;
2571

2572
	bus = kvm->buses[bus_idx];
M
Marcelo Tosatti 已提交
2573
	r = -ENOENT;
2574 2575
	for (i = 0; i < bus->dev_count; i++)
		if (bus->range[i].dev == dev) {
M
Marcelo Tosatti 已提交
2576
			r = 0;
2577 2578
			break;
		}
M
Marcelo Tosatti 已提交
2579

2580
	if (r)
M
Marcelo Tosatti 已提交
2581
		return r;
2582 2583 2584 2585 2586 2587 2588 2589 2590 2591

	new_bus = kzalloc(sizeof(*bus) + ((bus->dev_count - 1) *
			  sizeof(struct kvm_io_range)), GFP_KERNEL);
	if (!new_bus)
		return -ENOMEM;

	memcpy(new_bus, bus, sizeof(*bus) + i * sizeof(struct kvm_io_range));
	new_bus->dev_count--;
	memcpy(new_bus->range + i, bus->range + i + 1,
	       (new_bus->dev_count - i) * sizeof(struct kvm_io_range));
M
Marcelo Tosatti 已提交
2592 2593 2594 2595 2596

	rcu_assign_pointer(kvm->buses[bus_idx], new_bus);
	synchronize_srcu_expedited(&kvm->srcu);
	kfree(bus);
	return r;
2597 2598
}

A
Avi Kivity 已提交
2599 2600 2601 2602
static struct notifier_block kvm_cpu_notifier = {
	.notifier_call = kvm_cpu_hotplug,
};

2603
static int vm_stat_get(void *_offset, u64 *val)
2604 2605 2606 2607
{
	unsigned offset = (long)_offset;
	struct kvm *kvm;

2608
	*val = 0;
2609
	raw_spin_lock(&kvm_lock);
2610
	list_for_each_entry(kvm, &vm_list, vm_list)
2611
		*val += *(u32 *)((void *)kvm + offset);
2612
	raw_spin_unlock(&kvm_lock);
2613
	return 0;
2614 2615 2616 2617
}

DEFINE_SIMPLE_ATTRIBUTE(vm_stat_fops, vm_stat_get, NULL, "%llu\n");

2618
static int vcpu_stat_get(void *_offset, u64 *val)
A
Avi Kivity 已提交
2619 2620 2621 2622 2623 2624
{
	unsigned offset = (long)_offset;
	struct kvm *kvm;
	struct kvm_vcpu *vcpu;
	int i;

2625
	*val = 0;
2626
	raw_spin_lock(&kvm_lock);
A
Avi Kivity 已提交
2627
	list_for_each_entry(kvm, &vm_list, vm_list)
2628 2629 2630
		kvm_for_each_vcpu(i, vcpu, kvm)
			*val += *(u32 *)((void *)vcpu + offset);

2631
	raw_spin_unlock(&kvm_lock);
2632
	return 0;
A
Avi Kivity 已提交
2633 2634
}

2635 2636
DEFINE_SIMPLE_ATTRIBUTE(vcpu_stat_fops, vcpu_stat_get, NULL, "%llu\n");

2637
static const struct file_operations *stat_fops[] = {
2638 2639 2640
	[KVM_STAT_VCPU] = &vcpu_stat_fops,
	[KVM_STAT_VM]   = &vm_stat_fops,
};
A
Avi Kivity 已提交
2641

2642
static int kvm_init_debug(void)
A
Avi Kivity 已提交
2643
{
2644
	int r = -EFAULT;
A
Avi Kivity 已提交
2645 2646
	struct kvm_stats_debugfs_item *p;

2647
	kvm_debugfs_dir = debugfs_create_dir("kvm", NULL);
2648 2649 2650 2651
	if (kvm_debugfs_dir == NULL)
		goto out;

	for (p = debugfs_entries; p->name; ++p) {
2652
		p->dentry = debugfs_create_file(p->name, 0444, kvm_debugfs_dir,
A
Avi Kivity 已提交
2653
						(void *)(long)p->offset,
2654
						stat_fops[p->kind]);
2655 2656 2657 2658 2659 2660 2661 2662 2663 2664
		if (p->dentry == NULL)
			goto out_dir;
	}

	return 0;

out_dir:
	debugfs_remove_recursive(kvm_debugfs_dir);
out:
	return r;
A
Avi Kivity 已提交
2665 2666 2667 2668 2669 2670 2671 2672
}

static void kvm_exit_debug(void)
{
	struct kvm_stats_debugfs_item *p;

	for (p = debugfs_entries; p->name; ++p)
		debugfs_remove(p->dentry);
2673
	debugfs_remove(kvm_debugfs_dir);
A
Avi Kivity 已提交
2674 2675
}

2676
static int kvm_suspend(void)
2677
{
2678
	if (kvm_usage_count)
2679
		hardware_disable_nolock(NULL);
2680 2681 2682
	return 0;
}

2683
static void kvm_resume(void)
2684
{
2685
	if (kvm_usage_count) {
2686
		WARN_ON(raw_spin_is_locked(&kvm_lock));
2687
		hardware_enable_nolock(NULL);
2688
	}
2689 2690
}

2691
static struct syscore_ops kvm_syscore_ops = {
2692 2693 2694 2695
	.suspend = kvm_suspend,
	.resume = kvm_resume,
};

2696 2697 2698 2699 2700 2701 2702 2703 2704 2705
static inline
struct kvm_vcpu *preempt_notifier_to_vcpu(struct preempt_notifier *pn)
{
	return container_of(pn, struct kvm_vcpu, preempt_notifier);
}

static void kvm_sched_in(struct preempt_notifier *pn, int cpu)
{
	struct kvm_vcpu *vcpu = preempt_notifier_to_vcpu(pn);

2706
	kvm_arch_vcpu_load(vcpu, cpu);
2707 2708 2709 2710 2711 2712 2713
}

static void kvm_sched_out(struct preempt_notifier *pn,
			  struct task_struct *next)
{
	struct kvm_vcpu *vcpu = preempt_notifier_to_vcpu(pn);

2714
	kvm_arch_vcpu_put(vcpu);
2715 2716
}

2717
int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
2718
		  struct module *module)
A
Avi Kivity 已提交
2719 2720
{
	int r;
Y
Yang, Sheng 已提交
2721
	int cpu;
A
Avi Kivity 已提交
2722

2723 2724
	r = kvm_arch_init(opaque);
	if (r)
2725
		goto out_fail;
2726 2727 2728 2729 2730 2731 2732 2733

	bad_page = alloc_page(GFP_KERNEL | __GFP_ZERO);

	if (bad_page == NULL) {
		r = -ENOMEM;
		goto out;
	}

2734 2735
	bad_pfn = page_to_pfn(bad_page);

2736 2737 2738 2739 2740 2741 2742 2743 2744
	hwpoison_page = alloc_page(GFP_KERNEL | __GFP_ZERO);

	if (hwpoison_page == NULL) {
		r = -ENOMEM;
		goto out_free_0;
	}

	hwpoison_pfn = page_to_pfn(hwpoison_page);

2745 2746 2747 2748 2749 2750 2751 2752 2753
	fault_page = alloc_page(GFP_KERNEL | __GFP_ZERO);

	if (fault_page == NULL) {
		r = -ENOMEM;
		goto out_free_0;
	}

	fault_pfn = page_to_pfn(fault_page);

2754
	if (!zalloc_cpumask_var(&cpus_hardware_enabled, GFP_KERNEL)) {
2755 2756 2757 2758
		r = -ENOMEM;
		goto out_free_0;
	}

2759
	r = kvm_arch_hardware_setup();
A
Avi Kivity 已提交
2760
	if (r < 0)
2761
		goto out_free_0a;
A
Avi Kivity 已提交
2762

Y
Yang, Sheng 已提交
2763 2764
	for_each_online_cpu(cpu) {
		smp_call_function_single(cpu,
2765
				kvm_arch_check_processor_compat,
2766
				&r, 1);
Y
Yang, Sheng 已提交
2767
		if (r < 0)
2768
			goto out_free_1;
Y
Yang, Sheng 已提交
2769 2770
	}

A
Avi Kivity 已提交
2771 2772
	r = register_cpu_notifier(&kvm_cpu_notifier);
	if (r)
2773
		goto out_free_2;
A
Avi Kivity 已提交
2774 2775
	register_reboot_notifier(&kvm_reboot_notifier);

2776
	/* A kmem cache lets us meet the alignment requirements of fx_save. */
2777 2778 2779
	if (!vcpu_align)
		vcpu_align = __alignof__(struct kvm_vcpu);
	kvm_vcpu_cache = kmem_cache_create("kvm_vcpu", vcpu_size, vcpu_align,
J
Joe Perches 已提交
2780
					   0, NULL);
2781 2782
	if (!kvm_vcpu_cache) {
		r = -ENOMEM;
2783
		goto out_free_3;
2784 2785
	}

2786 2787 2788 2789
	r = kvm_async_pf_init();
	if (r)
		goto out_free;

A
Avi Kivity 已提交
2790
	kvm_chardev_ops.owner = module;
2791 2792
	kvm_vm_fops.owner = module;
	kvm_vcpu_fops.owner = module;
A
Avi Kivity 已提交
2793 2794 2795

	r = misc_register(&kvm_dev);
	if (r) {
M
Mike Day 已提交
2796
		printk(KERN_ERR "kvm: misc device register failed\n");
2797
		goto out_unreg;
A
Avi Kivity 已提交
2798 2799
	}

2800 2801
	register_syscore_ops(&kvm_syscore_ops);

2802 2803 2804
	kvm_preempt_ops.sched_in = kvm_sched_in;
	kvm_preempt_ops.sched_out = kvm_sched_out;

2805 2806 2807 2808 2809
	r = kvm_init_debug();
	if (r) {
		printk(KERN_ERR "kvm: create debugfs files failed\n");
		goto out_undebugfs;
	}
2810

2811
	return 0;
A
Avi Kivity 已提交
2812

2813 2814
out_undebugfs:
	unregister_syscore_ops(&kvm_syscore_ops);
2815 2816
out_unreg:
	kvm_async_pf_deinit();
A
Avi Kivity 已提交
2817
out_free:
2818
	kmem_cache_destroy(kvm_vcpu_cache);
2819
out_free_3:
A
Avi Kivity 已提交
2820
	unregister_reboot_notifier(&kvm_reboot_notifier);
A
Avi Kivity 已提交
2821
	unregister_cpu_notifier(&kvm_cpu_notifier);
2822 2823
out_free_2:
out_free_1:
2824
	kvm_arch_hardware_unsetup();
2825 2826
out_free_0a:
	free_cpumask_var(cpus_hardware_enabled);
2827
out_free_0:
2828 2829
	if (fault_page)
		__free_page(fault_page);
2830 2831
	if (hwpoison_page)
		__free_page(hwpoison_page);
2832
	__free_page(bad_page);
2833
out:
2834
	kvm_arch_exit();
2835
out_fail:
A
Avi Kivity 已提交
2836 2837
	return r;
}
2838
EXPORT_SYMBOL_GPL(kvm_init);
A
Avi Kivity 已提交
2839

2840
void kvm_exit(void)
A
Avi Kivity 已提交
2841
{
2842
	kvm_exit_debug();
A
Avi Kivity 已提交
2843
	misc_deregister(&kvm_dev);
2844
	kmem_cache_destroy(kvm_vcpu_cache);
2845
	kvm_async_pf_deinit();
2846
	unregister_syscore_ops(&kvm_syscore_ops);
A
Avi Kivity 已提交
2847
	unregister_reboot_notifier(&kvm_reboot_notifier);
2848
	unregister_cpu_notifier(&kvm_cpu_notifier);
2849
	on_each_cpu(hardware_disable_nolock, NULL, 1);
2850
	kvm_arch_hardware_unsetup();
2851
	kvm_arch_exit();
2852
	free_cpumask_var(cpus_hardware_enabled);
2853
	__free_page(hwpoison_page);
2854
	__free_page(bad_page);
A
Avi Kivity 已提交
2855
}
2856
EXPORT_SYMBOL_GPL(kvm_exit);