x86.c 134.2 KB
Newer Older
1 2 3 4 5 6
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * derived from drivers/kvm/kvm_main.c
 *
 * Copyright (C) 2006 Qumranet, Inc.
B
Ben-Ami Yassour 已提交
7 8
 * Copyright (C) 2008 Qumranet, Inc.
 * Copyright IBM Corporation, 2008
9 10 11 12
 *
 * Authors:
 *   Avi Kivity   <avi@qumranet.com>
 *   Yaniv Kamay  <yaniv@qumranet.com>
B
Ben-Ami Yassour 已提交
13 14
 *   Amit Shah    <amit.shah@qumranet.com>
 *   Ben-Ami Yassour <benami@il.ibm.com>
15 16 17 18 19 20
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */

21
#include <linux/kvm_host.h>
22
#include "irq.h"
23
#include "mmu.h"
S
Sheng Yang 已提交
24
#include "i8254.h"
25
#include "tss.h"
26
#include "kvm_cache_regs.h"
27
#include "x86.h"
28

29
#include <linux/clocksource.h>
B
Ben-Ami Yassour 已提交
30
#include <linux/interrupt.h>
31 32 33
#include <linux/kvm.h>
#include <linux/fs.h>
#include <linux/vmalloc.h>
34
#include <linux/module.h>
35
#include <linux/mman.h>
36
#include <linux/highmem.h>
J
Joerg Roedel 已提交
37
#include <linux/iommu.h>
B
Ben-Ami Yassour 已提交
38
#include <linux/intel-iommu.h>
39
#include <linux/cpufreq.h>
A
Avi Kivity 已提交
40
#include <linux/user-return-notifier.h>
41
#include <linux/srcu.h>
42
#include <linux/slab.h>
43
#include <linux/perf_event.h>
A
Avi Kivity 已提交
44
#include <trace/events/kvm.h>
X
Xiao Guangrong 已提交
45

46 47
#define CREATE_TRACE_POINTS
#include "trace.h"
48

49
#include <asm/debugreg.h>
50
#include <asm/uaccess.h>
51
#include <asm/msr.h>
52
#include <asm/desc.h>
S
Sheng Yang 已提交
53
#include <asm/mtrr.h>
H
Huang Ying 已提交
54
#include <asm/mce.h>
55

56
#define MAX_IO_MSRS 256
57 58 59 60 61 62 63 64 65 66 67
#define CR0_RESERVED_BITS						\
	(~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
			  | X86_CR0_ET | X86_CR0_NE | X86_CR0_WP | X86_CR0_AM \
			  | X86_CR0_NW | X86_CR0_CD | X86_CR0_PG))
#define CR4_RESERVED_BITS						\
	(~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\
			  | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE	\
			  | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR	\
			  | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE))

#define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
H
Huang Ying 已提交
68 69 70 71

#define KVM_MAX_MCE_BANKS 32
#define KVM_MCE_CAP_SUPPORTED MCG_CTL_P

72 73 74 75 76 77 78 79 80
/* EFER defaults:
 * - enable syscall per default because its emulated by KVM
 * - enable LME and LMA per default on 64 bit KVM
 */
#ifdef CONFIG_X86_64
static u64 __read_mostly efer_reserved_bits = 0xfffffffffffffafeULL;
#else
static u64 __read_mostly efer_reserved_bits = 0xfffffffffffffffeULL;
#endif
81

82 83
#define VM_STAT(x) offsetof(struct kvm, stat.x), KVM_STAT_VM
#define VCPU_STAT(x) offsetof(struct kvm_vcpu, stat.x), KVM_STAT_VCPU
84

85
static void update_cr8_intercept(struct kvm_vcpu *vcpu);
86 87 88
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
				    struct kvm_cpuid_entry2 __user *entries);

89
struct kvm_x86_ops *kvm_x86_ops;
90
EXPORT_SYMBOL_GPL(kvm_x86_ops);
91

92 93 94
int ignore_msrs = 0;
module_param_named(ignore_msrs, ignore_msrs, bool, S_IRUGO | S_IWUSR);

A
Avi Kivity 已提交
95 96 97 98
#define KVM_NR_SHARED_MSRS 16

struct kvm_shared_msrs_global {
	int nr;
99
	u32 msrs[KVM_NR_SHARED_MSRS];
A
Avi Kivity 已提交
100 101 102 103 104
};

struct kvm_shared_msrs {
	struct user_return_notifier urn;
	bool registered;
105 106 107 108
	struct kvm_shared_msr_values {
		u64 host;
		u64 curr;
	} values[KVM_NR_SHARED_MSRS];
A
Avi Kivity 已提交
109 110 111 112 113
};

static struct kvm_shared_msrs_global __read_mostly shared_msrs_global;
static DEFINE_PER_CPU(struct kvm_shared_msrs, shared_msrs);

114
struct kvm_stats_debugfs_item debugfs_entries[] = {
115 116 117 118 119 120 121 122 123
	{ "pf_fixed", VCPU_STAT(pf_fixed) },
	{ "pf_guest", VCPU_STAT(pf_guest) },
	{ "tlb_flush", VCPU_STAT(tlb_flush) },
	{ "invlpg", VCPU_STAT(invlpg) },
	{ "exits", VCPU_STAT(exits) },
	{ "io_exits", VCPU_STAT(io_exits) },
	{ "mmio_exits", VCPU_STAT(mmio_exits) },
	{ "signal_exits", VCPU_STAT(signal_exits) },
	{ "irq_window", VCPU_STAT(irq_window_exits) },
124
	{ "nmi_window", VCPU_STAT(nmi_window_exits) },
125 126
	{ "halt_exits", VCPU_STAT(halt_exits) },
	{ "halt_wakeup", VCPU_STAT(halt_wakeup) },
A
Amit Shah 已提交
127
	{ "hypercalls", VCPU_STAT(hypercalls) },
128 129 130 131 132 133 134
	{ "request_irq", VCPU_STAT(request_irq_exits) },
	{ "irq_exits", VCPU_STAT(irq_exits) },
	{ "host_state_reload", VCPU_STAT(host_state_reload) },
	{ "efer_reload", VCPU_STAT(efer_reload) },
	{ "fpu_reload", VCPU_STAT(fpu_reload) },
	{ "insn_emulation", VCPU_STAT(insn_emulation) },
	{ "insn_emulation_fail", VCPU_STAT(insn_emulation_fail) },
135
	{ "irq_injections", VCPU_STAT(irq_injections) },
136
	{ "nmi_injections", VCPU_STAT(nmi_injections) },
A
Avi Kivity 已提交
137 138 139 140 141 142
	{ "mmu_shadow_zapped", VM_STAT(mmu_shadow_zapped) },
	{ "mmu_pte_write", VM_STAT(mmu_pte_write) },
	{ "mmu_pte_updated", VM_STAT(mmu_pte_updated) },
	{ "mmu_pde_zapped", VM_STAT(mmu_pde_zapped) },
	{ "mmu_flooded", VM_STAT(mmu_flooded) },
	{ "mmu_recycled", VM_STAT(mmu_recycled) },
A
Avi Kivity 已提交
143
	{ "mmu_cache_miss", VM_STAT(mmu_cache_miss) },
144
	{ "mmu_unsync", VM_STAT(mmu_unsync) },
145
	{ "remote_tlb_flush", VM_STAT(remote_tlb_flush) },
M
Marcelo Tosatti 已提交
146
	{ "largepages", VM_STAT(lpages) },
147 148 149
	{ NULL }
};

A
Avi Kivity 已提交
150 151 152 153 154
static void kvm_on_user_return(struct user_return_notifier *urn)
{
	unsigned slot;
	struct kvm_shared_msrs *locals
		= container_of(urn, struct kvm_shared_msrs, urn);
155
	struct kvm_shared_msr_values *values;
A
Avi Kivity 已提交
156 157

	for (slot = 0; slot < shared_msrs_global.nr; ++slot) {
158 159 160 161
		values = &locals->values[slot];
		if (values->host != values->curr) {
			wrmsrl(shared_msrs_global.msrs[slot], values->host);
			values->curr = values->host;
A
Avi Kivity 已提交
162 163 164 165 166 167
		}
	}
	locals->registered = false;
	user_return_notifier_unregister(urn);
}

168
static void shared_msr_update(unsigned slot, u32 msr)
A
Avi Kivity 已提交
169
{
170
	struct kvm_shared_msrs *smsr;
A
Avi Kivity 已提交
171 172
	u64 value;

173 174 175 176 177 178 179 180 181 182 183 184 185 186
	smsr = &__get_cpu_var(shared_msrs);
	/* only read, and nobody should modify it at this time,
	 * so don't need lock */
	if (slot >= shared_msrs_global.nr) {
		printk(KERN_ERR "kvm: invalid MSR slot!");
		return;
	}
	rdmsrl_safe(msr, &value);
	smsr->values[slot].host = value;
	smsr->values[slot].curr = value;
}

void kvm_define_shared_msr(unsigned slot, u32 msr)
{
A
Avi Kivity 已提交
187 188
	if (slot >= shared_msrs_global.nr)
		shared_msrs_global.nr = slot + 1;
189 190 191
	shared_msrs_global.msrs[slot] = msr;
	/* we need ensured the shared_msr_global have been updated */
	smp_wmb();
A
Avi Kivity 已提交
192 193 194 195 196 197 198 199
}
EXPORT_SYMBOL_GPL(kvm_define_shared_msr);

static void kvm_shared_msr_cpu_online(void)
{
	unsigned i;

	for (i = 0; i < shared_msrs_global.nr; ++i)
200
		shared_msr_update(i, shared_msrs_global.msrs[i]);
A
Avi Kivity 已提交
201 202
}

203
void kvm_set_shared_msr(unsigned slot, u64 value, u64 mask)
A
Avi Kivity 已提交
204 205 206
{
	struct kvm_shared_msrs *smsr = &__get_cpu_var(shared_msrs);

207
	if (((value ^ smsr->values[slot].curr) & mask) == 0)
A
Avi Kivity 已提交
208
		return;
209 210
	smsr->values[slot].curr = value;
	wrmsrl(shared_msrs_global.msrs[slot], value);
A
Avi Kivity 已提交
211 212 213 214 215 216 217 218
	if (!smsr->registered) {
		smsr->urn.on_user_return = kvm_on_user_return;
		user_return_notifier_register(&smsr->urn);
		smsr->registered = true;
	}
}
EXPORT_SYMBOL_GPL(kvm_set_shared_msr);

219 220 221 222 223 224 225 226
static void drop_user_return_notifiers(void *ignore)
{
	struct kvm_shared_msrs *smsr = &__get_cpu_var(shared_msrs);

	if (smsr->registered)
		kvm_on_user_return(&smsr->urn);
}

227 228 229
u64 kvm_get_apic_base(struct kvm_vcpu *vcpu)
{
	if (irqchip_in_kernel(vcpu->kvm))
230
		return vcpu->arch.apic_base;
231
	else
232
		return vcpu->arch.apic_base;
233 234 235 236 237 238 239 240 241
}
EXPORT_SYMBOL_GPL(kvm_get_apic_base);

void kvm_set_apic_base(struct kvm_vcpu *vcpu, u64 data)
{
	/* TODO: reserve bits check */
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_lapic_set_base(vcpu, data);
	else
242
		vcpu->arch.apic_base = data;
243 244 245
}
EXPORT_SYMBOL_GPL(kvm_set_apic_base);

246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
#define EXCPT_BENIGN		0
#define EXCPT_CONTRIBUTORY	1
#define EXCPT_PF		2

static int exception_class(int vector)
{
	switch (vector) {
	case PF_VECTOR:
		return EXCPT_PF;
	case DE_VECTOR:
	case TS_VECTOR:
	case NP_VECTOR:
	case SS_VECTOR:
	case GP_VECTOR:
		return EXCPT_CONTRIBUTORY;
	default:
		break;
	}
	return EXCPT_BENIGN;
}

static void kvm_multiple_exception(struct kvm_vcpu *vcpu,
268 269
		unsigned nr, bool has_error, u32 error_code,
		bool reinject)
270 271 272 273 274 275 276 277 278 279
{
	u32 prev_nr;
	int class1, class2;

	if (!vcpu->arch.exception.pending) {
	queue:
		vcpu->arch.exception.pending = true;
		vcpu->arch.exception.has_error_code = has_error;
		vcpu->arch.exception.nr = nr;
		vcpu->arch.exception.error_code = error_code;
280
		vcpu->arch.exception.reinject = reinject;
281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
		return;
	}

	/* to check exception */
	prev_nr = vcpu->arch.exception.nr;
	if (prev_nr == DF_VECTOR) {
		/* triple fault -> shutdown */
		set_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests);
		return;
	}
	class1 = exception_class(prev_nr);
	class2 = exception_class(nr);
	if ((class1 == EXCPT_CONTRIBUTORY && class2 == EXCPT_CONTRIBUTORY)
		|| (class1 == EXCPT_PF && class2 != EXCPT_BENIGN)) {
		/* generate double fault per SDM Table 5-5 */
		vcpu->arch.exception.pending = true;
		vcpu->arch.exception.has_error_code = true;
		vcpu->arch.exception.nr = DF_VECTOR;
		vcpu->arch.exception.error_code = 0;
	} else
		/* replace previous exception with a new one in a hope
		   that instruction re-execution will regenerate lost
		   exception */
		goto queue;
}

307 308
void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr)
{
309
	kvm_multiple_exception(vcpu, nr, false, 0, false);
310 311 312
}
EXPORT_SYMBOL_GPL(kvm_queue_exception);

313 314 315 316 317 318
void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr)
{
	kvm_multiple_exception(vcpu, nr, false, 0, true);
}
EXPORT_SYMBOL_GPL(kvm_requeue_exception);

319 320 321 322
void kvm_inject_page_fault(struct kvm_vcpu *vcpu, unsigned long addr,
			   u32 error_code)
{
	++vcpu->stat.pf_guest;
323
	vcpu->arch.cr2 = addr;
324 325 326
	kvm_queue_exception_e(vcpu, PF_VECTOR, error_code);
}

327 328 329 330 331 332
void kvm_inject_nmi(struct kvm_vcpu *vcpu)
{
	vcpu->arch.nmi_pending = 1;
}
EXPORT_SYMBOL_GPL(kvm_inject_nmi);

333 334
void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
{
335
	kvm_multiple_exception(vcpu, nr, true, error_code, false);
336 337 338
}
EXPORT_SYMBOL_GPL(kvm_queue_exception_e);

339 340 341 342 343 344
void kvm_requeue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
{
	kvm_multiple_exception(vcpu, nr, true, error_code, true);
}
EXPORT_SYMBOL_GPL(kvm_requeue_exception_e);

345 346 347 348 349
/*
 * Checks if cpl <= required_cpl; if true, return true.  Otherwise queue
 * a #GP and return false.
 */
bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl)
350
{
351 352 353 354
	if (kvm_x86_ops->get_cpl(vcpu) <= required_cpl)
		return true;
	kvm_queue_exception_e(vcpu, GP_VECTOR, 0);
	return false;
355
}
356
EXPORT_SYMBOL_GPL(kvm_require_cpl);
357

358 359 360 361 362 363 364 365 366
/*
 * Load the pae pdptrs.  Return true is they are all valid.
 */
int load_pdptrs(struct kvm_vcpu *vcpu, unsigned long cr3)
{
	gfn_t pdpt_gfn = cr3 >> PAGE_SHIFT;
	unsigned offset = ((cr3 & (PAGE_SIZE-1)) >> 5) << 2;
	int i;
	int ret;
367
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
368 369 370 371 372 373 374 375

	ret = kvm_read_guest_page(vcpu->kvm, pdpt_gfn, pdpte,
				  offset * sizeof(u64), sizeof(pdpte));
	if (ret < 0) {
		ret = 0;
		goto out;
	}
	for (i = 0; i < ARRAY_SIZE(pdpte); ++i) {
376
		if (is_present_gpte(pdpte[i]) &&
377
		    (pdpte[i] & vcpu->arch.mmu.rsvd_bits_mask[0][2])) {
378 379 380 381 382 383
			ret = 0;
			goto out;
		}
	}
	ret = 1;

384
	memcpy(vcpu->arch.pdptrs, pdpte, sizeof(vcpu->arch.pdptrs));
A
Avi Kivity 已提交
385 386 387 388
	__set_bit(VCPU_EXREG_PDPTR,
		  (unsigned long *)&vcpu->arch.regs_avail);
	__set_bit(VCPU_EXREG_PDPTR,
		  (unsigned long *)&vcpu->arch.regs_dirty);
389 390 391 392
out:

	return ret;
}
393
EXPORT_SYMBOL_GPL(load_pdptrs);
394

395 396
static bool pdptrs_changed(struct kvm_vcpu *vcpu)
{
397
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
398 399 400 401 402 403
	bool changed = true;
	int r;

	if (is_long_mode(vcpu) || !is_pae(vcpu))
		return false;

A
Avi Kivity 已提交
404 405 406 407
	if (!test_bit(VCPU_EXREG_PDPTR,
		      (unsigned long *)&vcpu->arch.regs_avail))
		return true;

408
	r = kvm_read_guest(vcpu->kvm, vcpu->arch.cr3 & ~31u, pdpte, sizeof(pdpte));
409 410
	if (r < 0)
		goto out;
411
	changed = memcmp(pdpte, vcpu->arch.pdptrs, sizeof(pdpte)) != 0;
412 413 414 415 416
out:

	return changed;
}

417
static int __kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
418
{
419 420
	cr0 |= X86_CR0_ET;

421
#ifdef CONFIG_X86_64
422 423
	if (cr0 & 0xffffffff00000000UL)
		return 1;
424 425 426
#endif

	cr0 &= ~CR0_RESERVED_BITS;
427

428 429
	if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD))
		return 1;
430

431 432
	if ((cr0 & X86_CR0_PG) && !(cr0 & X86_CR0_PE))
		return 1;
433 434 435

	if (!is_paging(vcpu) && (cr0 & X86_CR0_PG)) {
#ifdef CONFIG_X86_64
436
		if ((vcpu->arch.efer & EFER_LME)) {
437 438
			int cs_db, cs_l;

439 440
			if (!is_pae(vcpu))
				return 1;
441
			kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
442 443
			if (cs_l)
				return 1;
444 445
		} else
#endif
446 447
		if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.cr3))
			return 1;
448 449 450 451 452
	}

	kvm_x86_ops->set_cr0(vcpu, cr0);

	kvm_mmu_reset_context(vcpu);
453 454 455 456 457 458 459
	return 0;
}

void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
{
	if (__kvm_set_cr0(vcpu, cr0))
		kvm_inject_gp(vcpu, 0);
460
}
461
EXPORT_SYMBOL_GPL(kvm_set_cr0);
462

463
void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw)
464
{
465
	kvm_set_cr0(vcpu, kvm_read_cr0_bits(vcpu, ~0x0eul) | (msw & 0x0f));
466
}
467
EXPORT_SYMBOL_GPL(kvm_lmsw);
468

469
int __kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
470
{
471
	unsigned long old_cr4 = kvm_read_cr4(vcpu);
472 473
	unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE;

474 475
	if (cr4 & CR4_RESERVED_BITS)
		return 1;
476 477

	if (is_long_mode(vcpu)) {
478 479
		if (!(cr4 & X86_CR4_PAE))
			return 1;
480 481
	} else if (is_paging(vcpu) && (cr4 & X86_CR4_PAE)
		   && ((cr4 ^ old_cr4) & pdptr_bits)
482 483 484 485 486
		   && !load_pdptrs(vcpu, vcpu->arch.cr3))
		return 1;

	if (cr4 & X86_CR4_VMXE)
		return 1;
487 488

	kvm_x86_ops->set_cr4(vcpu, cr4);
489
	vcpu->arch.cr4 = cr4;
490
	kvm_mmu_reset_context(vcpu);
491 492 493 494 495 496 497 498

	return 0;
}

void kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
{
	if (__kvm_set_cr4(vcpu, cr4))
		kvm_inject_gp(vcpu, 0);
499
}
500
EXPORT_SYMBOL_GPL(kvm_set_cr4);
501

502
static int __kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
503
{
504
	if (cr3 == vcpu->arch.cr3 && !pdptrs_changed(vcpu)) {
505
		kvm_mmu_sync_roots(vcpu);
506
		kvm_mmu_flush_tlb(vcpu);
507
		return 0;
508 509
	}

510
	if (is_long_mode(vcpu)) {
511 512
		if (cr3 & CR3_L_MODE_RESERVED_BITS)
			return 1;
513 514
	} else {
		if (is_pae(vcpu)) {
515 516 517 518
			if (cr3 & CR3_PAE_RESERVED_BITS)
				return 1;
			if (is_paging(vcpu) && !load_pdptrs(vcpu, cr3))
				return 1;
519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535
		}
		/*
		 * We don't check reserved bits in nonpae mode, because
		 * this isn't enforced, and VMware depends on this.
		 */
	}

	/*
	 * Does the new cr3 value map to physical memory? (Note, we
	 * catch an invalid cr3 even in real-mode, because it would
	 * cause trouble later on when we turn on paging anyway.)
	 *
	 * A real CPU would silently accept an invalid cr3 and would
	 * attempt to use it - with largely undefined (and often hard
	 * to debug) behavior on the guest side.
	 */
	if (unlikely(!gfn_to_memslot(vcpu->kvm, cr3 >> PAGE_SHIFT)))
536 537 538 539 540 541 542 543 544
		return 1;
	vcpu->arch.cr3 = cr3;
	vcpu->arch.mmu.new_cr3(vcpu);
	return 0;
}

void kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
{
	if (__kvm_set_cr3(vcpu, cr3))
545
		kvm_inject_gp(vcpu, 0);
546
}
547
EXPORT_SYMBOL_GPL(kvm_set_cr3);
548

549
int __kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
550
{
551 552
	if (cr8 & CR8_RESERVED_BITS)
		return 1;
553 554 555
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_lapic_set_tpr(vcpu, cr8);
	else
556
		vcpu->arch.cr8 = cr8;
557 558 559 560 561 562 563
	return 0;
}

void kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
{
	if (__kvm_set_cr8(vcpu, cr8))
		kvm_inject_gp(vcpu, 0);
564
}
565
EXPORT_SYMBOL_GPL(kvm_set_cr8);
566

567
unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu)
568 569 570 571
{
	if (irqchip_in_kernel(vcpu->kvm))
		return kvm_lapic_get_cr8(vcpu);
	else
572
		return vcpu->arch.cr8;
573
}
574
EXPORT_SYMBOL_GPL(kvm_get_cr8);
575

576
static int __kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val)
577 578 579 580 581 582 583 584
{
	switch (dr) {
	case 0 ... 3:
		vcpu->arch.db[dr] = val;
		if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
			vcpu->arch.eff_db[dr] = val;
		break;
	case 4:
585 586
		if (kvm_read_cr4_bits(vcpu, X86_CR4_DE))
			return 1; /* #UD */
587 588
		/* fall through */
	case 6:
589 590
		if (val & 0xffffffff00000000ULL)
			return -1; /* #GP */
591 592 593
		vcpu->arch.dr6 = (val & DR6_VOLATILE) | DR6_FIXED_1;
		break;
	case 5:
594 595
		if (kvm_read_cr4_bits(vcpu, X86_CR4_DE))
			return 1; /* #UD */
596 597
		/* fall through */
	default: /* 7 */
598 599
		if (val & 0xffffffff00000000ULL)
			return -1; /* #GP */
600 601 602 603 604 605 606 607 608 609
		vcpu->arch.dr7 = (val & DR7_VOLATILE) | DR7_FIXED_1;
		if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)) {
			kvm_x86_ops->set_dr7(vcpu, vcpu->arch.dr7);
			vcpu->arch.switch_db_regs = (val & DR7_BP_EN_MASK);
		}
		break;
	}

	return 0;
}
610 611 612 613 614 615 616 617 618 619 620 621 622

int kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val)
{
	int res;

	res = __kvm_set_dr(vcpu, dr, val);
	if (res > 0)
		kvm_queue_exception(vcpu, UD_VECTOR);
	else if (res < 0)
		kvm_inject_gp(vcpu, 0);

	return res;
}
623 624
EXPORT_SYMBOL_GPL(kvm_set_dr);

625
static int _kvm_get_dr(struct kvm_vcpu *vcpu, int dr, unsigned long *val)
626 627 628 629 630 631
{
	switch (dr) {
	case 0 ... 3:
		*val = vcpu->arch.db[dr];
		break;
	case 4:
632
		if (kvm_read_cr4_bits(vcpu, X86_CR4_DE))
633 634 635 636 637 638
			return 1;
		/* fall through */
	case 6:
		*val = vcpu->arch.dr6;
		break;
	case 5:
639
		if (kvm_read_cr4_bits(vcpu, X86_CR4_DE))
640 641 642 643 644 645 646 647 648
			return 1;
		/* fall through */
	default: /* 7 */
		*val = vcpu->arch.dr7;
		break;
	}

	return 0;
}
649 650 651 652 653 654 655 656 657

int kvm_get_dr(struct kvm_vcpu *vcpu, int dr, unsigned long *val)
{
	if (_kvm_get_dr(vcpu, dr, val)) {
		kvm_queue_exception(vcpu, UD_VECTOR);
		return 1;
	}
	return 0;
}
658 659
EXPORT_SYMBOL_GPL(kvm_get_dr);

660 661 662 663 664
static inline u32 bit(int bitno)
{
	return 1 << (bitno & 31);
}

665 666 667 668 669
/*
 * List of msr numbers which we expose to userspace through KVM_GET_MSRS
 * and KVM_SET_MSRS, and KVM_GET_MSR_INDEX_LIST.
 *
 * This list is modified at module load time to reflect the
670 671
 * capabilities of the host cpu. This capabilities test skips MSRs that are
 * kvm-specific. Those are put in the beginning of the list.
672
 */
673

674
#define KVM_SAVE_MSRS_BEGIN	7
675
static u32 msrs_to_save[] = {
676
	MSR_KVM_SYSTEM_TIME, MSR_KVM_WALL_CLOCK,
677
	MSR_KVM_SYSTEM_TIME_NEW, MSR_KVM_WALL_CLOCK_NEW,
678
	HV_X64_MSR_GUEST_OS_ID, HV_X64_MSR_HYPERCALL,
G
Gleb Natapov 已提交
679
	HV_X64_MSR_APIC_ASSIST_PAGE,
680 681 682 683 684
	MSR_IA32_SYSENTER_CS, MSR_IA32_SYSENTER_ESP, MSR_IA32_SYSENTER_EIP,
	MSR_K6_STAR,
#ifdef CONFIG_X86_64
	MSR_CSTAR, MSR_KERNEL_GS_BASE, MSR_SYSCALL_MASK, MSR_LSTAR,
#endif
685
	MSR_IA32_TSC, MSR_IA32_PERF_STATUS, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA
686 687 688 689 690 691 692 693
};

static unsigned num_msrs_to_save;

static u32 emulated_msrs[] = {
	MSR_IA32_MISC_ENABLE,
};

694
static int set_efer(struct kvm_vcpu *vcpu, u64 efer)
695
{
696 697
	if (efer & efer_reserved_bits)
		return 1;
698 699

	if (is_paging(vcpu)
700 701
	    && (vcpu->arch.efer & EFER_LME) != (efer & EFER_LME))
		return 1;
702

A
Alexander Graf 已提交
703 704 705 706
	if (efer & EFER_FFXSR) {
		struct kvm_cpuid_entry2 *feat;

		feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
707 708
		if (!feat || !(feat->edx & bit(X86_FEATURE_FXSR_OPT)))
			return 1;
A
Alexander Graf 已提交
709 710
	}

711 712 713 714
	if (efer & EFER_SVME) {
		struct kvm_cpuid_entry2 *feat;

		feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
715 716
		if (!feat || !(feat->ecx & bit(X86_FEATURE_SVM)))
			return 1;
717 718
	}

719
	efer &= ~EFER_LMA;
720
	efer |= vcpu->arch.efer & EFER_LMA;
721

722 723
	kvm_x86_ops->set_efer(vcpu, efer);

724
	vcpu->arch.efer = efer;
725 726 727

	vcpu->arch.mmu.base_role.nxe = (efer & EFER_NX) && !tdp_enabled;
	kvm_mmu_reset_context(vcpu);
728 729

	return 0;
730 731
}

732 733 734 735 736 737 738
void kvm_enable_efer_bits(u64 mask)
{
       efer_reserved_bits &= ~mask;
}
EXPORT_SYMBOL_GPL(kvm_enable_efer_bits);


739 740 741 742 743 744 745 746 747 748
/*
 * Writes msr value into into the appropriate "register".
 * Returns 0 on success, non-0 otherwise.
 * Assumes vcpu_load() was already called.
 */
int kvm_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data)
{
	return kvm_x86_ops->set_msr(vcpu, msr_index, data);
}

749 750 751 752 753 754 755 756
/*
 * Adapt set_msr() to msr_io()'s calling convention
 */
static int do_set_msr(struct kvm_vcpu *vcpu, unsigned index, u64 *data)
{
	return kvm_set_msr(vcpu, index, *data);
}

757 758
static void kvm_write_wall_clock(struct kvm *kvm, gpa_t wall_clock)
{
759 760
	int version;
	int r;
761
	struct pvclock_wall_clock wc;
762
	struct timespec boot;
763 764 765 766

	if (!wall_clock)
		return;

767 768 769 770 771 772 773 774
	r = kvm_read_guest(kvm, wall_clock, &version, sizeof(version));
	if (r)
		return;

	if (version & 1)
		++version;  /* first time write, random junk */

	++version;
775 776 777

	kvm_write_guest(kvm, wall_clock, &version, sizeof(version));

778 779 780 781 782 783
	/*
	 * The guest calculates current wall clock time by adding
	 * system time (updated by kvm_write_guest_time below) to the
	 * wall clock specified here.  guest system time equals host
	 * system time for us, thus we must fill in host boot time here.
	 */
784
	getboottime(&boot);
785 786 787 788

	wc.sec = boot.tv_sec;
	wc.nsec = boot.tv_nsec;
	wc.version = version;
789 790 791 792 793 794 795

	kvm_write_guest(kvm, wall_clock, &wc, sizeof(wc));

	version++;
	kvm_write_guest(kvm, wall_clock, &version, sizeof(version));
}

796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830
static uint32_t div_frac(uint32_t dividend, uint32_t divisor)
{
	uint32_t quotient, remainder;

	/* Don't try to replace with do_div(), this one calculates
	 * "(dividend << 32) / divisor" */
	__asm__ ( "divl %4"
		  : "=a" (quotient), "=d" (remainder)
		  : "0" (0), "1" (dividend), "r" (divisor) );
	return quotient;
}

static void kvm_set_time_scale(uint32_t tsc_khz, struct pvclock_vcpu_time_info *hv_clock)
{
	uint64_t nsecs = 1000000000LL;
	int32_t  shift = 0;
	uint64_t tps64;
	uint32_t tps32;

	tps64 = tsc_khz * 1000LL;
	while (tps64 > nsecs*2) {
		tps64 >>= 1;
		shift--;
	}

	tps32 = (uint32_t)tps64;
	while (tps32 <= (uint32_t)nsecs) {
		tps32 <<= 1;
		shift++;
	}

	hv_clock->tsc_shift = shift;
	hv_clock->tsc_to_system_mul = div_frac(nsecs, tps32);

	pr_debug("%s: tsc_khz %u, tsc_shift %d, tsc_mul %u\n",
831
		 __func__, tsc_khz, hv_clock->tsc_shift,
832 833 834
		 hv_clock->tsc_to_system_mul);
}

835 836
static DEFINE_PER_CPU(unsigned long, cpu_tsc_khz);

837 838 839 840 841 842
static void kvm_write_guest_time(struct kvm_vcpu *v)
{
	struct timespec ts;
	unsigned long flags;
	struct kvm_vcpu_arch *vcpu = &v->arch;
	void *shared_kaddr;
843
	unsigned long this_tsc_khz;
844 845 846 847

	if ((!vcpu->time_page))
		return;

848 849 850 851
	this_tsc_khz = get_cpu_var(cpu_tsc_khz);
	if (unlikely(vcpu->hv_clock_tsc_khz != this_tsc_khz)) {
		kvm_set_time_scale(this_tsc_khz, &vcpu->hv_clock);
		vcpu->hv_clock_tsc_khz = this_tsc_khz;
852
	}
853
	put_cpu_var(cpu_tsc_khz);
854

855 856
	/* Keep irq disabled to prevent changes to the clock */
	local_irq_save(flags);
857
	kvm_get_msr(v, MSR_IA32_TSC, &vcpu->hv_clock.tsc_timestamp);
858
	ktime_get_ts(&ts);
859
	monotonic_to_bootbased(&ts);
860 861 862 863 864
	local_irq_restore(flags);

	/* With all the info we got, fill in the values */

	vcpu->hv_clock.system_time = ts.tv_nsec +
865 866
				     (NSEC_PER_SEC * (u64)ts.tv_sec) + v->kvm->arch.kvmclock_offset;

867 868
	vcpu->hv_clock.flags = 0;

869 870 871
	/*
	 * The interface expects us to write an even number signaling that the
	 * update is finished. Since the guest won't see the intermediate
872
	 * state, we just increase by 2 at the end.
873
	 */
874
	vcpu->hv_clock.version += 2;
875 876 877 878

	shared_kaddr = kmap_atomic(vcpu->time_page, KM_USER0);

	memcpy(shared_kaddr + vcpu->time_offset, &vcpu->hv_clock,
879
	       sizeof(vcpu->hv_clock));
880 881 882 883 884 885

	kunmap_atomic(shared_kaddr, KM_USER0);

	mark_page_dirty(v->kvm, vcpu->time >> PAGE_SHIFT);
}

886 887 888 889 890 891 892 893 894 895
static int kvm_request_guest_time_update(struct kvm_vcpu *v)
{
	struct kvm_vcpu_arch *vcpu = &v->arch;

	if (!vcpu->time_page)
		return 0;
	set_bit(KVM_REQ_KVMCLOCK_UPDATE, &v->requests);
	return 1;
}

A
Avi Kivity 已提交
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919
static bool msr_mtrr_valid(unsigned msr)
{
	switch (msr) {
	case 0x200 ... 0x200 + 2 * KVM_NR_VAR_MTRR - 1:
	case MSR_MTRRfix64K_00000:
	case MSR_MTRRfix16K_80000:
	case MSR_MTRRfix16K_A0000:
	case MSR_MTRRfix4K_C0000:
	case MSR_MTRRfix4K_C8000:
	case MSR_MTRRfix4K_D0000:
	case MSR_MTRRfix4K_D8000:
	case MSR_MTRRfix4K_E0000:
	case MSR_MTRRfix4K_E8000:
	case MSR_MTRRfix4K_F0000:
	case MSR_MTRRfix4K_F8000:
	case MSR_MTRRdefType:
	case MSR_IA32_CR_PAT:
		return true;
	case 0x2f8:
		return true;
	}
	return false;
}

920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956
static bool valid_pat_type(unsigned t)
{
	return t < 8 && (1 << t) & 0xf3; /* 0, 1, 4, 5, 6, 7 */
}

static bool valid_mtrr_type(unsigned t)
{
	return t < 8 && (1 << t) & 0x73; /* 0, 1, 4, 5, 6 */
}

static bool mtrr_valid(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
	int i;

	if (!msr_mtrr_valid(msr))
		return false;

	if (msr == MSR_IA32_CR_PAT) {
		for (i = 0; i < 8; i++)
			if (!valid_pat_type((data >> (i * 8)) & 0xff))
				return false;
		return true;
	} else if (msr == MSR_MTRRdefType) {
		if (data & ~0xcff)
			return false;
		return valid_mtrr_type(data & 0xff);
	} else if (msr >= MSR_MTRRfix64K_00000 && msr <= MSR_MTRRfix4K_F8000) {
		for (i = 0; i < 8 ; i++)
			if (!valid_mtrr_type((data >> (i * 8)) & 0xff))
				return false;
		return true;
	}

	/* variable MTRRs */
	return valid_mtrr_type(data & 0xff);
}

A
Avi Kivity 已提交
957 958
static int set_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
S
Sheng Yang 已提交
959 960
	u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges;

961
	if (!mtrr_valid(vcpu, msr, data))
A
Avi Kivity 已提交
962 963
		return 1;

S
Sheng Yang 已提交
964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990
	if (msr == MSR_MTRRdefType) {
		vcpu->arch.mtrr_state.def_type = data;
		vcpu->arch.mtrr_state.enabled = (data & 0xc00) >> 10;
	} else if (msr == MSR_MTRRfix64K_00000)
		p[0] = data;
	else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000)
		p[1 + msr - MSR_MTRRfix16K_80000] = data;
	else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000)
		p[3 + msr - MSR_MTRRfix4K_C0000] = data;
	else if (msr == MSR_IA32_CR_PAT)
		vcpu->arch.pat = data;
	else {	/* Variable MTRRs */
		int idx, is_mtrr_mask;
		u64 *pt;

		idx = (msr - 0x200) / 2;
		is_mtrr_mask = msr - 0x200 - 2 * idx;
		if (!is_mtrr_mask)
			pt =
			  (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].base_lo;
		else
			pt =
			  (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].mask_lo;
		*pt = data;
	}

	kvm_mmu_reset_context(vcpu);
A
Avi Kivity 已提交
991 992
	return 0;
}
993

H
Huang Ying 已提交
994
static int set_msr_mce(struct kvm_vcpu *vcpu, u32 msr, u64 data)
995
{
H
Huang Ying 已提交
996 997 998
	u64 mcg_cap = vcpu->arch.mcg_cap;
	unsigned bank_num = mcg_cap & 0xff;

999 1000
	switch (msr) {
	case MSR_IA32_MCG_STATUS:
H
Huang Ying 已提交
1001
		vcpu->arch.mcg_status = data;
1002
		break;
1003
	case MSR_IA32_MCG_CTL:
H
Huang Ying 已提交
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
		if (!(mcg_cap & MCG_CTL_P))
			return 1;
		if (data != 0 && data != ~(u64)0)
			return -1;
		vcpu->arch.mcg_ctl = data;
		break;
	default:
		if (msr >= MSR_IA32_MC0_CTL &&
		    msr < MSR_IA32_MC0_CTL + 4 * bank_num) {
			u32 offset = msr - MSR_IA32_MC0_CTL;
1014 1015 1016 1017 1018
			/* only 0 or all 1s can be written to IA32_MCi_CTL
			 * some Linux kernels though clear bit 10 in bank 4 to
			 * workaround a BIOS/GART TBL issue on AMD K8s, ignore
			 * this to avoid an uncatched #GP in the guest
			 */
H
Huang Ying 已提交
1019
			if ((offset & 0x3) == 0 &&
1020
			    data != 0 && (data | (1 << 10)) != ~(u64)0)
H
Huang Ying 已提交
1021 1022 1023 1024 1025 1026 1027 1028 1029
				return -1;
			vcpu->arch.mce_banks[offset] = data;
			break;
		}
		return 1;
	}
	return 0;
}

E
Ed Swierk 已提交
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061
static int xen_hvm_config(struct kvm_vcpu *vcpu, u64 data)
{
	struct kvm *kvm = vcpu->kvm;
	int lm = is_long_mode(vcpu);
	u8 *blob_addr = lm ? (u8 *)(long)kvm->arch.xen_hvm_config.blob_addr_64
		: (u8 *)(long)kvm->arch.xen_hvm_config.blob_addr_32;
	u8 blob_size = lm ? kvm->arch.xen_hvm_config.blob_size_64
		: kvm->arch.xen_hvm_config.blob_size_32;
	u32 page_num = data & ~PAGE_MASK;
	u64 page_addr = data & PAGE_MASK;
	u8 *page;
	int r;

	r = -E2BIG;
	if (page_num >= blob_size)
		goto out;
	r = -ENOMEM;
	page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	if (!page)
		goto out;
	r = -EFAULT;
	if (copy_from_user(page, blob_addr + (page_num * PAGE_SIZE), PAGE_SIZE))
		goto out_free;
	if (kvm_write_guest(kvm, page_addr, page, PAGE_SIZE))
		goto out_free;
	r = 0;
out_free:
	kfree(page);
out:
	return r;
}

1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
static bool kvm_hv_hypercall_enabled(struct kvm *kvm)
{
	return kvm->arch.hv_hypercall & HV_X64_MSR_HYPERCALL_ENABLE;
}

static bool kvm_hv_msr_partition_wide(u32 msr)
{
	bool r = false;
	switch (msr) {
	case HV_X64_MSR_GUEST_OS_ID:
	case HV_X64_MSR_HYPERCALL:
		r = true;
		break;
	}

	return r;
}

static int set_msr_hyperv_pw(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
	struct kvm *kvm = vcpu->kvm;

	switch (msr) {
	case HV_X64_MSR_GUEST_OS_ID:
		kvm->arch.hv_guest_os_id = data;
		/* setting guest os id to zero disables hypercall page */
		if (!kvm->arch.hv_guest_os_id)
			kvm->arch.hv_hypercall &= ~HV_X64_MSR_HYPERCALL_ENABLE;
		break;
	case HV_X64_MSR_HYPERCALL: {
		u64 gfn;
		unsigned long addr;
		u8 instructions[4];

		/* if guest os id is not set hypercall should remain disabled */
		if (!kvm->arch.hv_guest_os_id)
			break;
		if (!(data & HV_X64_MSR_HYPERCALL_ENABLE)) {
			kvm->arch.hv_hypercall = data;
			break;
		}
		gfn = data >> HV_X64_MSR_HYPERCALL_PAGE_ADDRESS_SHIFT;
		addr = gfn_to_hva(kvm, gfn);
		if (kvm_is_error_hva(addr))
			return 1;
		kvm_x86_ops->patch_hypercall(vcpu, instructions);
		((unsigned char *)instructions)[3] = 0xc3; /* ret */
		if (copy_to_user((void __user *)addr, instructions, 4))
			return 1;
		kvm->arch.hv_hypercall = data;
		break;
	}
	default:
		pr_unimpl(vcpu, "HYPER-V unimplemented wrmsr: 0x%x "
			  "data 0x%llx\n", msr, data);
		return 1;
	}
	return 0;
}

static int set_msr_hyperv(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
G
Gleb Natapov 已提交
1124 1125 1126
	switch (msr) {
	case HV_X64_MSR_APIC_ASSIST_PAGE: {
		unsigned long addr;
1127

G
Gleb Natapov 已提交
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153
		if (!(data & HV_X64_MSR_APIC_ASSIST_PAGE_ENABLE)) {
			vcpu->arch.hv_vapic = data;
			break;
		}
		addr = gfn_to_hva(vcpu->kvm, data >>
				  HV_X64_MSR_APIC_ASSIST_PAGE_ADDRESS_SHIFT);
		if (kvm_is_error_hva(addr))
			return 1;
		if (clear_user((void __user *)addr, PAGE_SIZE))
			return 1;
		vcpu->arch.hv_vapic = data;
		break;
	}
	case HV_X64_MSR_EOI:
		return kvm_hv_vapic_msr_write(vcpu, APIC_EOI, data);
	case HV_X64_MSR_ICR:
		return kvm_hv_vapic_msr_write(vcpu, APIC_ICR, data);
	case HV_X64_MSR_TPR:
		return kvm_hv_vapic_msr_write(vcpu, APIC_TASKPRI, data);
	default:
		pr_unimpl(vcpu, "HYPER-V unimplemented wrmsr: 0x%x "
			  "data 0x%llx\n", msr, data);
		return 1;
	}

	return 0;
1154 1155
}

1156 1157 1158 1159
int kvm_set_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
	switch (msr) {
	case MSR_EFER:
1160
		return set_efer(vcpu, data);
1161 1162
	case MSR_K7_HWCR:
		data &= ~(u64)0x40;	/* ignore flush filter disable */
1163
		data &= ~(u64)0x100;	/* ignore ignne emulation enable */
1164 1165 1166 1167 1168
		if (data != 0) {
			pr_unimpl(vcpu, "unimplemented HWCR wrmsr: 0x%llx\n",
				data);
			return 1;
		}
1169
		break;
1170 1171 1172 1173 1174 1175
	case MSR_FAM10H_MMIO_CONF_BASE:
		if (data != 0) {
			pr_unimpl(vcpu, "unimplemented MMIO_CONF_BASE wrmsr: "
				"0x%llx\n", data);
			return 1;
		}
1176
		break;
1177
	case MSR_AMD64_NB_CFG:
1178
		break;
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
	case MSR_IA32_DEBUGCTLMSR:
		if (!data) {
			/* We support the non-activated case already */
			break;
		} else if (data & ~(DEBUGCTLMSR_LBR | DEBUGCTLMSR_BTF)) {
			/* Values other than LBR and BTF are vendor-specific,
			   thus reserved and should throw a #GP */
			return 1;
		}
		pr_unimpl(vcpu, "%s: MSR_IA32_DEBUGCTLMSR 0x%llx, nop\n",
			__func__, data);
		break;
1191 1192
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_UCODE_WRITE:
1193
	case MSR_VM_HSAVE_PA:
A
Andre Przywara 已提交
1194
	case MSR_AMD64_PATCH_LOADER:
1195
		break;
A
Avi Kivity 已提交
1196 1197
	case 0x200 ... 0x2ff:
		return set_msr_mtrr(vcpu, msr, data);
1198 1199 1200
	case MSR_IA32_APICBASE:
		kvm_set_apic_base(vcpu, data);
		break;
G
Gleb Natapov 已提交
1201 1202
	case APIC_BASE_MSR ... APIC_BASE_MSR + 0x3ff:
		return kvm_x2apic_msr_write(vcpu, msr, data);
1203
	case MSR_IA32_MISC_ENABLE:
1204
		vcpu->arch.ia32_misc_enable_msr = data;
1205
		break;
1206
	case MSR_KVM_WALL_CLOCK_NEW:
1207 1208 1209 1210
	case MSR_KVM_WALL_CLOCK:
		vcpu->kvm->arch.wall_clock = data;
		kvm_write_wall_clock(vcpu->kvm, data);
		break;
1211
	case MSR_KVM_SYSTEM_TIME_NEW:
1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
	case MSR_KVM_SYSTEM_TIME: {
		if (vcpu->arch.time_page) {
			kvm_release_page_dirty(vcpu->arch.time_page);
			vcpu->arch.time_page = NULL;
		}

		vcpu->arch.time = data;

		/* we verify if the enable bit is set... */
		if (!(data & 1))
			break;

		/* ...but clean it before doing the actual write */
		vcpu->arch.time_offset = data & ~(PAGE_MASK | 1);

		vcpu->arch.time_page =
				gfn_to_page(vcpu->kvm, data >> PAGE_SHIFT);

		if (is_error_page(vcpu->arch.time_page)) {
			kvm_release_page_clean(vcpu->arch.time_page);
			vcpu->arch.time_page = NULL;
		}

1235
		kvm_request_guest_time_update(vcpu);
1236 1237
		break;
	}
H
Huang Ying 已提交
1238 1239 1240 1241
	case MSR_IA32_MCG_CTL:
	case MSR_IA32_MCG_STATUS:
	case MSR_IA32_MC0_CTL ... MSR_IA32_MC0_CTL + 4 * KVM_MAX_MCE_BANKS - 1:
		return set_msr_mce(vcpu, msr, data);
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271

	/* Performance counters are not protected by a CPUID bit,
	 * so we should check all of them in the generic path for the sake of
	 * cross vendor migration.
	 * Writing a zero into the event select MSRs disables them,
	 * which we perfectly emulate ;-). Any other value should be at least
	 * reported, some guests depend on them.
	 */
	case MSR_P6_EVNTSEL0:
	case MSR_P6_EVNTSEL1:
	case MSR_K7_EVNTSEL0:
	case MSR_K7_EVNTSEL1:
	case MSR_K7_EVNTSEL2:
	case MSR_K7_EVNTSEL3:
		if (data != 0)
			pr_unimpl(vcpu, "unimplemented perfctr wrmsr: "
				"0x%x data 0x%llx\n", msr, data);
		break;
	/* at least RHEL 4 unconditionally writes to the perfctr registers,
	 * so we ignore writes to make it happy.
	 */
	case MSR_P6_PERFCTR0:
	case MSR_P6_PERFCTR1:
	case MSR_K7_PERFCTR0:
	case MSR_K7_PERFCTR1:
	case MSR_K7_PERFCTR2:
	case MSR_K7_PERFCTR3:
		pr_unimpl(vcpu, "unimplemented perfctr wrmsr: "
			"0x%x data 0x%llx\n", msr, data);
		break;
1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
	case HV_X64_MSR_GUEST_OS_ID ... HV_X64_MSR_SINT15:
		if (kvm_hv_msr_partition_wide(msr)) {
			int r;
			mutex_lock(&vcpu->kvm->lock);
			r = set_msr_hyperv_pw(vcpu, msr, data);
			mutex_unlock(&vcpu->kvm->lock);
			return r;
		} else
			return set_msr_hyperv(vcpu, msr, data);
		break;
1282
	default:
E
Ed Swierk 已提交
1283 1284
		if (msr && (msr == vcpu->kvm->arch.xen_hvm_config.msr))
			return xen_hvm_config(vcpu, data);
1285 1286 1287 1288 1289 1290 1291 1292 1293
		if (!ignore_msrs) {
			pr_unimpl(vcpu, "unhandled wrmsr: 0x%x data %llx\n",
				msr, data);
			return 1;
		} else {
			pr_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n",
				msr, data);
			break;
		}
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
	}
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_set_msr_common);


/*
 * Reads an msr value (of 'msr_index') into 'pdata'.
 * Returns 0 on success, non-0 otherwise.
 * Assumes vcpu_load() was already called.
 */
int kvm_get_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 *pdata)
{
	return kvm_x86_ops->get_msr(vcpu, msr_index, pdata);
}

A
Avi Kivity 已提交
1310 1311
static int get_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
S
Sheng Yang 已提交
1312 1313
	u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges;

A
Avi Kivity 已提交
1314 1315 1316
	if (!msr_mtrr_valid(msr))
		return 1;

S
Sheng Yang 已提交
1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342
	if (msr == MSR_MTRRdefType)
		*pdata = vcpu->arch.mtrr_state.def_type +
			 (vcpu->arch.mtrr_state.enabled << 10);
	else if (msr == MSR_MTRRfix64K_00000)
		*pdata = p[0];
	else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000)
		*pdata = p[1 + msr - MSR_MTRRfix16K_80000];
	else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000)
		*pdata = p[3 + msr - MSR_MTRRfix4K_C0000];
	else if (msr == MSR_IA32_CR_PAT)
		*pdata = vcpu->arch.pat;
	else {	/* Variable MTRRs */
		int idx, is_mtrr_mask;
		u64 *pt;

		idx = (msr - 0x200) / 2;
		is_mtrr_mask = msr - 0x200 - 2 * idx;
		if (!is_mtrr_mask)
			pt =
			  (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].base_lo;
		else
			pt =
			  (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].mask_lo;
		*pdata = *pt;
	}

A
Avi Kivity 已提交
1343 1344 1345
	return 0;
}

H
Huang Ying 已提交
1346
static int get_msr_mce(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
1347 1348
{
	u64 data;
H
Huang Ying 已提交
1349 1350
	u64 mcg_cap = vcpu->arch.mcg_cap;
	unsigned bank_num = mcg_cap & 0xff;
1351 1352 1353 1354

	switch (msr) {
	case MSR_IA32_P5_MC_ADDR:
	case MSR_IA32_P5_MC_TYPE:
H
Huang Ying 已提交
1355 1356
		data = 0;
		break;
1357
	case MSR_IA32_MCG_CAP:
H
Huang Ying 已提交
1358 1359
		data = vcpu->arch.mcg_cap;
		break;
1360
	case MSR_IA32_MCG_CTL:
H
Huang Ying 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
		if (!(mcg_cap & MCG_CTL_P))
			return 1;
		data = vcpu->arch.mcg_ctl;
		break;
	case MSR_IA32_MCG_STATUS:
		data = vcpu->arch.mcg_status;
		break;
	default:
		if (msr >= MSR_IA32_MC0_CTL &&
		    msr < MSR_IA32_MC0_CTL + 4 * bank_num) {
			u32 offset = msr - MSR_IA32_MC0_CTL;
			data = vcpu->arch.mce_banks[offset];
			break;
		}
		return 1;
	}
	*pdata = data;
	return 0;
}

1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
static int get_msr_hyperv_pw(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
	u64 data = 0;
	struct kvm *kvm = vcpu->kvm;

	switch (msr) {
	case HV_X64_MSR_GUEST_OS_ID:
		data = kvm->arch.hv_guest_os_id;
		break;
	case HV_X64_MSR_HYPERCALL:
		data = kvm->arch.hv_hypercall;
		break;
	default:
		pr_unimpl(vcpu, "Hyper-V unhandled rdmsr: 0x%x\n", msr);
		return 1;
	}

	*pdata = data;
	return 0;
}

static int get_msr_hyperv(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
	u64 data = 0;

	switch (msr) {
	case HV_X64_MSR_VP_INDEX: {
		int r;
		struct kvm_vcpu *v;
		kvm_for_each_vcpu(r, v, vcpu->kvm)
			if (v == vcpu)
				data = r;
		break;
	}
G
Gleb Natapov 已提交
1415 1416 1417 1418 1419 1420
	case HV_X64_MSR_EOI:
		return kvm_hv_vapic_msr_read(vcpu, APIC_EOI, pdata);
	case HV_X64_MSR_ICR:
		return kvm_hv_vapic_msr_read(vcpu, APIC_ICR, pdata);
	case HV_X64_MSR_TPR:
		return kvm_hv_vapic_msr_read(vcpu, APIC_TASKPRI, pdata);
1421 1422 1423 1424 1425 1426 1427 1428
	default:
		pr_unimpl(vcpu, "Hyper-V unhandled rdmsr: 0x%x\n", msr);
		return 1;
	}
	*pdata = data;
	return 0;
}

H
Huang Ying 已提交
1429 1430 1431 1432 1433 1434
int kvm_get_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
	u64 data;

	switch (msr) {
	case MSR_IA32_PLATFORM_ID:
1435 1436
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_EBL_CR_POWERON:
1437 1438 1439 1440 1441
	case MSR_IA32_DEBUGCTLMSR:
	case MSR_IA32_LASTBRANCHFROMIP:
	case MSR_IA32_LASTBRANCHTOIP:
	case MSR_IA32_LASTINTFROMIP:
	case MSR_IA32_LASTINTTOIP:
1442 1443
	case MSR_K8_SYSCFG:
	case MSR_K7_HWCR:
1444
	case MSR_VM_HSAVE_PA:
A
Amit Shah 已提交
1445 1446
	case MSR_P6_PERFCTR0:
	case MSR_P6_PERFCTR1:
1447 1448
	case MSR_P6_EVNTSEL0:
	case MSR_P6_EVNTSEL1:
A
Amit Shah 已提交
1449
	case MSR_K7_EVNTSEL0:
A
Amit Shah 已提交
1450
	case MSR_K7_PERFCTR0:
1451
	case MSR_K8_INT_PENDING_MSG:
1452
	case MSR_AMD64_NB_CFG:
1453
	case MSR_FAM10H_MMIO_CONF_BASE:
1454 1455
		data = 0;
		break;
A
Avi Kivity 已提交
1456 1457 1458 1459 1460
	case MSR_MTRRcap:
		data = 0x500 | KVM_NR_VAR_MTRR;
		break;
	case 0x200 ... 0x2ff:
		return get_msr_mtrr(vcpu, msr, pdata);
1461 1462 1463 1464 1465 1466
	case 0xcd: /* fsb frequency */
		data = 3;
		break;
	case MSR_IA32_APICBASE:
		data = kvm_get_apic_base(vcpu);
		break;
G
Gleb Natapov 已提交
1467 1468 1469
	case APIC_BASE_MSR ... APIC_BASE_MSR + 0x3ff:
		return kvm_x2apic_msr_read(vcpu, msr, pdata);
		break;
1470
	case MSR_IA32_MISC_ENABLE:
1471
		data = vcpu->arch.ia32_misc_enable_msr;
1472
		break;
1473 1474 1475 1476 1477 1478
	case MSR_IA32_PERF_STATUS:
		/* TSC increment by tick */
		data = 1000ULL;
		/* CPU multiplier */
		data |= (((uint64_t)4ULL) << 40);
		break;
1479
	case MSR_EFER:
1480
		data = vcpu->arch.efer;
1481
		break;
1482
	case MSR_KVM_WALL_CLOCK:
1483
	case MSR_KVM_WALL_CLOCK_NEW:
1484 1485 1486
		data = vcpu->kvm->arch.wall_clock;
		break;
	case MSR_KVM_SYSTEM_TIME:
1487
	case MSR_KVM_SYSTEM_TIME_NEW:
1488 1489
		data = vcpu->arch.time;
		break;
H
Huang Ying 已提交
1490 1491 1492 1493 1494 1495 1496
	case MSR_IA32_P5_MC_ADDR:
	case MSR_IA32_P5_MC_TYPE:
	case MSR_IA32_MCG_CAP:
	case MSR_IA32_MCG_CTL:
	case MSR_IA32_MCG_STATUS:
	case MSR_IA32_MC0_CTL ... MSR_IA32_MC0_CTL + 4 * KVM_MAX_MCE_BANKS - 1:
		return get_msr_mce(vcpu, msr, pdata);
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506
	case HV_X64_MSR_GUEST_OS_ID ... HV_X64_MSR_SINT15:
		if (kvm_hv_msr_partition_wide(msr)) {
			int r;
			mutex_lock(&vcpu->kvm->lock);
			r = get_msr_hyperv_pw(vcpu, msr, pdata);
			mutex_unlock(&vcpu->kvm->lock);
			return r;
		} else
			return get_msr_hyperv(vcpu, msr, pdata);
		break;
1507
	default:
1508 1509 1510 1511 1512 1513 1514 1515
		if (!ignore_msrs) {
			pr_unimpl(vcpu, "unhandled rdmsr: 0x%x\n", msr);
			return 1;
		} else {
			pr_unimpl(vcpu, "ignored rdmsr: 0x%x\n", msr);
			data = 0;
		}
		break;
1516 1517 1518 1519 1520 1521
	}
	*pdata = data;
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_get_msr_common);

1522 1523 1524 1525 1526 1527 1528 1529 1530 1531
/*
 * Read or write a bunch of msrs. All parameters are kernel addresses.
 *
 * @return number of msrs set successfully.
 */
static int __msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs *msrs,
		    struct kvm_msr_entry *entries,
		    int (*do_msr)(struct kvm_vcpu *vcpu,
				  unsigned index, u64 *data))
{
1532
	int i, idx;
1533 1534 1535

	vcpu_load(vcpu);

1536
	idx = srcu_read_lock(&vcpu->kvm->srcu);
1537 1538 1539
	for (i = 0; i < msrs->nmsrs; ++i)
		if (do_msr(vcpu, entries[i].index, &entries[i].data))
			break;
1540
	srcu_read_unlock(&vcpu->kvm->srcu, idx);
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571

	vcpu_put(vcpu);

	return i;
}

/*
 * Read or write a bunch of msrs. Parameters are user addresses.
 *
 * @return number of msrs set successfully.
 */
static int msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs __user *user_msrs,
		  int (*do_msr)(struct kvm_vcpu *vcpu,
				unsigned index, u64 *data),
		  int writeback)
{
	struct kvm_msrs msrs;
	struct kvm_msr_entry *entries;
	int r, n;
	unsigned size;

	r = -EFAULT;
	if (copy_from_user(&msrs, user_msrs, sizeof msrs))
		goto out;

	r = -E2BIG;
	if (msrs.nmsrs >= MAX_IO_MSRS)
		goto out;

	r = -ENOMEM;
	size = sizeof(struct kvm_msr_entry) * msrs.nmsrs;
1572
	entries = kmalloc(size, GFP_KERNEL);
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
	if (!entries)
		goto out;

	r = -EFAULT;
	if (copy_from_user(entries, user_msrs->entries, size))
		goto out_free;

	r = n = __msr_io(vcpu, &msrs, entries, do_msr);
	if (r < 0)
		goto out_free;

	r = -EFAULT;
	if (writeback && copy_to_user(user_msrs->entries, entries, size))
		goto out_free;

	r = n;

out_free:
1591
	kfree(entries);
1592 1593 1594 1595
out:
	return r;
}

1596 1597 1598 1599 1600 1601 1602 1603 1604
int kvm_dev_ioctl_check_extension(long ext)
{
	int r;

	switch (ext) {
	case KVM_CAP_IRQCHIP:
	case KVM_CAP_HLT:
	case KVM_CAP_MMU_SHADOW_CACHE_CONTROL:
	case KVM_CAP_SET_TSS_ADDR:
1605
	case KVM_CAP_EXT_CPUID:
1606
	case KVM_CAP_CLOCKSOURCE:
S
Sheng Yang 已提交
1607
	case KVM_CAP_PIT:
1608
	case KVM_CAP_NOP_IO_DELAY:
1609
	case KVM_CAP_MP_STATE:
1610
	case KVM_CAP_SYNC_MMU:
1611
	case KVM_CAP_REINJECT_CONTROL:
1612
	case KVM_CAP_IRQ_INJECT_STATUS:
1613
	case KVM_CAP_ASSIGN_DEV_IRQ:
G
Gregory Haskins 已提交
1614
	case KVM_CAP_IRQFD:
G
Gregory Haskins 已提交
1615
	case KVM_CAP_IOEVENTFD:
1616
	case KVM_CAP_PIT2:
B
Beth Kon 已提交
1617
	case KVM_CAP_PIT_STATE2:
1618
	case KVM_CAP_SET_IDENTITY_MAP_ADDR:
E
Ed Swierk 已提交
1619
	case KVM_CAP_XEN_HVM:
1620
	case KVM_CAP_ADJUST_CLOCK:
J
Jan Kiszka 已提交
1621
	case KVM_CAP_VCPU_EVENTS:
1622
	case KVM_CAP_HYPERV:
G
Gleb Natapov 已提交
1623
	case KVM_CAP_HYPERV_VAPIC:
1624
	case KVM_CAP_HYPERV_SPIN:
1625
	case KVM_CAP_PCI_SEGMENT:
1626
	case KVM_CAP_DEBUGREGS:
1627
	case KVM_CAP_X86_ROBUST_SINGLESTEP:
1628 1629
		r = 1;
		break;
1630 1631 1632
	case KVM_CAP_COALESCED_MMIO:
		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
		break;
1633 1634 1635
	case KVM_CAP_VAPIC:
		r = !kvm_x86_ops->cpu_has_accelerated_tpr();
		break;
1636 1637 1638
	case KVM_CAP_NR_VCPUS:
		r = KVM_MAX_VCPUS;
		break;
1639 1640 1641
	case KVM_CAP_NR_MEMSLOTS:
		r = KVM_MEMORY_SLOTS;
		break;
1642 1643
	case KVM_CAP_PV_MMU:	/* obsolete */
		r = 0;
1644
		break;
B
Ben-Ami Yassour 已提交
1645
	case KVM_CAP_IOMMU:
J
Joerg Roedel 已提交
1646
		r = iommu_found();
B
Ben-Ami Yassour 已提交
1647
		break;
H
Huang Ying 已提交
1648 1649 1650
	case KVM_CAP_MCE:
		r = KVM_MAX_MCE_BANKS;
		break;
1651 1652 1653 1654 1655 1656 1657 1658
	default:
		r = 0;
		break;
	}
	return r;

}

1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678
long kvm_arch_dev_ioctl(struct file *filp,
			unsigned int ioctl, unsigned long arg)
{
	void __user *argp = (void __user *)arg;
	long r;

	switch (ioctl) {
	case KVM_GET_MSR_INDEX_LIST: {
		struct kvm_msr_list __user *user_msr_list = argp;
		struct kvm_msr_list msr_list;
		unsigned n;

		r = -EFAULT;
		if (copy_from_user(&msr_list, user_msr_list, sizeof msr_list))
			goto out;
		n = msr_list.nmsrs;
		msr_list.nmsrs = num_msrs_to_save + ARRAY_SIZE(emulated_msrs);
		if (copy_to_user(user_msr_list, &msr_list, sizeof msr_list))
			goto out;
		r = -E2BIG;
J
Jan Kiszka 已提交
1679
		if (n < msr_list.nmsrs)
1680 1681 1682 1683 1684
			goto out;
		r = -EFAULT;
		if (copy_to_user(user_msr_list->indices, &msrs_to_save,
				 num_msrs_to_save * sizeof(u32)))
			goto out;
J
Jan Kiszka 已提交
1685
		if (copy_to_user(user_msr_list->indices + num_msrs_to_save,
1686 1687 1688 1689 1690 1691
				 &emulated_msrs,
				 ARRAY_SIZE(emulated_msrs) * sizeof(u32)))
			goto out;
		r = 0;
		break;
	}
1692 1693 1694 1695 1696 1697 1698 1699
	case KVM_GET_SUPPORTED_CPUID: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_dev_ioctl_get_supported_cpuid(&cpuid,
1700
						      cpuid_arg->entries);
1701 1702 1703 1704 1705 1706 1707 1708 1709
		if (r)
			goto out;

		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
H
Huang Ying 已提交
1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
	case KVM_X86_GET_MCE_CAP_SUPPORTED: {
		u64 mce_cap;

		mce_cap = KVM_MCE_CAP_SUPPORTED;
		r = -EFAULT;
		if (copy_to_user(argp, &mce_cap, sizeof mce_cap))
			goto out;
		r = 0;
		break;
	}
1720 1721 1722 1723 1724 1725 1726
	default:
		r = -EINVAL;
	}
out:
	return r;
}

1727 1728 1729
void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
{
	kvm_x86_ops->vcpu_load(vcpu, cpu);
Z
Zachary Amsden 已提交
1730 1731 1732 1733 1734 1735
	if (unlikely(per_cpu(cpu_tsc_khz, cpu) == 0)) {
		unsigned long khz = cpufreq_quick_get(cpu);
		if (!khz)
			khz = tsc_khz;
		per_cpu(cpu_tsc_khz, cpu) = khz;
	}
1736
	kvm_request_guest_time_update(vcpu);
1737 1738 1739 1740
}

void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
1741
	kvm_x86_ops->vcpu_put(vcpu);
1742
	kvm_put_guest_fpu(vcpu);
1743 1744
}

1745
static int is_efer_nx(void)
1746
{
1747
	unsigned long long efer = 0;
1748

1749
	rdmsrl_safe(MSR_EFER, &efer);
1750 1751 1752 1753 1754 1755 1756 1757
	return efer & EFER_NX;
}

static void cpuid_fix_nx_cap(struct kvm_vcpu *vcpu)
{
	int i;
	struct kvm_cpuid_entry2 *e, *entry;

1758
	entry = NULL;
1759 1760
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		e = &vcpu->arch.cpuid_entries[i];
1761 1762 1763 1764 1765
		if (e->function == 0x80000001) {
			entry = e;
			break;
		}
	}
1766
	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
1767 1768 1769 1770 1771
		entry->edx &= ~(1 << 20);
		printk(KERN_INFO "kvm: guest NX capability removed\n");
	}
}

1772
/* when an old userspace process fills a new kernel module */
1773 1774 1775
static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
				    struct kvm_cpuid *cpuid,
				    struct kvm_cpuid_entry __user *entries)
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790
{
	int r, i;
	struct kvm_cpuid_entry *cpuid_entries;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry) * cpuid->nent);
	if (!cpuid_entries)
		goto out;
	r = -EFAULT;
	if (copy_from_user(cpuid_entries, entries,
			   cpuid->nent * sizeof(struct kvm_cpuid_entry)))
		goto out_free;
1791
	vcpu_load(vcpu);
1792
	for (i = 0; i < cpuid->nent; i++) {
1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804
		vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
		vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
		vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
		vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
		vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
		vcpu->arch.cpuid_entries[i].index = 0;
		vcpu->arch.cpuid_entries[i].flags = 0;
		vcpu->arch.cpuid_entries[i].padding[0] = 0;
		vcpu->arch.cpuid_entries[i].padding[1] = 0;
		vcpu->arch.cpuid_entries[i].padding[2] = 0;
	}
	vcpu->arch.cpuid_nent = cpuid->nent;
1805 1806
	cpuid_fix_nx_cap(vcpu);
	r = 0;
1807
	kvm_apic_set_version(vcpu);
1808
	kvm_x86_ops->cpuid_update(vcpu);
1809
	vcpu_put(vcpu);
1810 1811 1812 1813 1814 1815 1816 1817

out_free:
	vfree(cpuid_entries);
out:
	return r;
}

static int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
1818 1819
				     struct kvm_cpuid2 *cpuid,
				     struct kvm_cpuid_entry2 __user *entries)
1820 1821 1822 1823 1824 1825 1826
{
	int r;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -EFAULT;
1827
	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
1828
			   cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
1829
		goto out;
1830
	vcpu_load(vcpu);
1831
	vcpu->arch.cpuid_nent = cpuid->nent;
1832
	kvm_apic_set_version(vcpu);
1833
	kvm_x86_ops->cpuid_update(vcpu);
1834
	vcpu_put(vcpu);
1835 1836 1837 1838 1839 1840
	return 0;

out:
	return r;
}

1841
static int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
1842 1843
				     struct kvm_cpuid2 *cpuid,
				     struct kvm_cpuid_entry2 __user *entries)
1844 1845 1846
{
	int r;

1847
	vcpu_load(vcpu);
1848
	r = -E2BIG;
1849
	if (cpuid->nent < vcpu->arch.cpuid_nent)
1850 1851
		goto out;
	r = -EFAULT;
1852
	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
1853
			 vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
1854 1855 1856 1857
		goto out;
	return 0;

out:
1858
	cpuid->nent = vcpu->arch.cpuid_nent;
1859
	vcpu_put(vcpu);
1860 1861 1862 1863
	return r;
}

static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
1864
			   u32 index)
1865 1866 1867 1868
{
	entry->function = function;
	entry->index = index;
	cpuid_count(entry->function, entry->index,
1869
		    &entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
1870 1871 1872
	entry->flags = 0;
}

1873 1874
#define F(x) bit(X86_FEATURE_##x)

1875 1876 1877
static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			 u32 index, int *nent, int maxnent)
{
1878
	unsigned f_nx = is_efer_nx() ? F(NX) : 0;
1879
#ifdef CONFIG_X86_64
1880 1881
	unsigned f_gbpages = (kvm_x86_ops->get_lpage_level() == PT_PDPE_LEVEL)
				? F(GBPAGES) : 0;
1882 1883
	unsigned f_lm = F(LM);
#else
1884
	unsigned f_gbpages = 0;
1885
	unsigned f_lm = 0;
1886
#endif
1887
	unsigned f_rdtscp = kvm_x86_ops->rdtscp_supported() ? F(RDTSCP) : 0;
1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906

	/* cpuid 1.edx */
	const u32 kvm_supported_word0_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SEP) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* PSN */ | F(CLFLSH) |
		0 /* Reserved, DS, ACPI */ | F(MMX) |
		F(FXSR) | F(XMM) | F(XMM2) | F(SELFSNOOP) |
		0 /* HTT, TM, Reserved, PBE */;
	/* cpuid 0x80000001.edx */
	const u32 kvm_supported_word1_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SYSCALL) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* Reserved */ |
		f_nx | 0 /* Reserved */ | F(MMXEXT) | F(MMX) |
1907
		F(FXSR) | F(FXSR_OPT) | f_gbpages | f_rdtscp |
1908 1909 1910
		0 /* Reserved */ | f_lm | F(3DNOWEXT) | F(3DNOW);
	/* cpuid 1.ecx */
	const u32 kvm_supported_word4_x86_features =
A
Avi Kivity 已提交
1911 1912 1913 1914 1915
		F(XMM3) | 0 /* Reserved, DTES64, MONITOR */ |
		0 /* DS-CPL, VMX, SMX, EST */ |
		0 /* TM2 */ | F(SSSE3) | 0 /* CNXT-ID */ | 0 /* Reserved */ |
		0 /* Reserved */ | F(CX16) | 0 /* xTPR Update, PDCM */ |
		0 /* Reserved, DCA */ | F(XMM4_1) |
G
Gleb Natapov 已提交
1916
		F(XMM4_2) | F(X2APIC) | F(MOVBE) | F(POPCNT) |
A
Avi Kivity 已提交
1917
		0 /* Reserved, XSAVE, OSXSAVE */;
1918
	/* cpuid 0x80000001.ecx */
1919
	const u32 kvm_supported_word6_x86_features =
1920 1921 1922 1923
		F(LAHF_LM) | F(CMP_LEGACY) | F(SVM) | 0 /* ExtApicSpace */ |
		F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) |
		F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(SSE5) |
		0 /* SKINIT */ | 0 /* WDT */;
1924

1925
	/* all calls to cpuid_count() should be made on the same cpu */
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
	get_cpu();
	do_cpuid_1_ent(entry, function, index);
	++*nent;

	switch (function) {
	case 0:
		entry->eax = min(entry->eax, (u32)0xb);
		break;
	case 1:
		entry->edx &= kvm_supported_word0_x86_features;
1936
		entry->ecx &= kvm_supported_word4_x86_features;
1937 1938 1939
		/* we support x2apic emulation even if host does not support
		 * it since we emulate x2apic in software */
		entry->ecx |= F(X2APIC);
1940 1941 1942 1943 1944 1945 1946 1947 1948
		break;
	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	 * may return different values. This forces us to get_cpu() before
	 * issuing the first command, and also to emulate this annoying behavior
	 * in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	case 2: {
		int t, times = entry->eax & 0xff;

		entry->flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
1949
		entry->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
1950 1951 1952 1953 1954 1955 1956 1957 1958
		for (t = 1; t < times && *nent < maxnent; ++t) {
			do_cpuid_1_ent(&entry[t], function, 0);
			entry[t].flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
			++*nent;
		}
		break;
	}
	/* function 4 and 0xb have additional index. */
	case 4: {
1959
		int i, cache_type;
1960 1961 1962

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until cache_type is zero */
1963 1964
		for (i = 1; *nent < maxnent; ++i) {
			cache_type = entry[i - 1].eax & 0x1f;
1965 1966
			if (!cache_type)
				break;
1967 1968
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1969 1970 1971 1972 1973 1974
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 0xb: {
1975
		int i, level_type;
1976 1977 1978

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until level_type is zero */
1979
		for (i = 1; *nent < maxnent; ++i) {
1980
			level_type = entry[i - 1].ecx & 0xff00;
1981 1982
			if (!level_type)
				break;
1983 1984
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1985 1986 1987 1988 1989
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
	case KVM_CPUID_SIGNATURE: {
		char signature[12] = "KVMKVMKVM\0\0";
		u32 *sigptr = (u32 *)signature;
		entry->eax = 0;
		entry->ebx = sigptr[0];
		entry->ecx = sigptr[1];
		entry->edx = sigptr[2];
		break;
	}
	case KVM_CPUID_FEATURES:
		entry->eax = (1 << KVM_FEATURE_CLOCKSOURCE) |
			     (1 << KVM_FEATURE_NOP_IO_DELAY) |
2002 2003
			     (1 << KVM_FEATURE_CLOCKSOURCE2) |
			     (1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT);
2004 2005 2006 2007
		entry->ebx = 0;
		entry->ecx = 0;
		entry->edx = 0;
		break;
2008 2009 2010 2011 2012 2013 2014 2015
	case 0x80000000:
		entry->eax = min(entry->eax, 0x8000001a);
		break;
	case 0x80000001:
		entry->edx &= kvm_supported_word1_x86_features;
		entry->ecx &= kvm_supported_word6_x86_features;
		break;
	}
2016 2017 2018

	kvm_x86_ops->set_supported_cpuid(function, entry);

2019 2020 2021
	put_cpu();
}

2022 2023
#undef F

2024
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
2025
				     struct kvm_cpuid_entry2 __user *entries)
2026 2027 2028 2029 2030 2031 2032
{
	struct kvm_cpuid_entry2 *cpuid_entries;
	int limit, nent = 0, r = -E2BIG;
	u32 func;

	if (cpuid->nent < 1)
		goto out;
2033 2034
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		cpuid->nent = KVM_MAX_CPUID_ENTRIES;
2035 2036 2037 2038 2039 2040 2041 2042 2043
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	if (!cpuid_entries)
		goto out;

	do_cpuid_ent(&cpuid_entries[0], 0, 0, &nent, cpuid->nent);
	limit = cpuid_entries[0].eax;
	for (func = 1; func <= limit && nent < cpuid->nent; ++func)
		do_cpuid_ent(&cpuid_entries[nent], func, 0,
2044
			     &nent, cpuid->nent);
2045 2046 2047 2048 2049 2050 2051 2052
	r = -E2BIG;
	if (nent >= cpuid->nent)
		goto out_free;

	do_cpuid_ent(&cpuid_entries[nent], 0x80000000, 0, &nent, cpuid->nent);
	limit = cpuid_entries[nent - 1].eax;
	for (func = 0x80000001; func <= limit && nent < cpuid->nent; ++func)
		do_cpuid_ent(&cpuid_entries[nent], func, 0,
2053
			     &nent, cpuid->nent);
2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070



	r = -E2BIG;
	if (nent >= cpuid->nent)
		goto out_free;

	do_cpuid_ent(&cpuid_entries[nent], KVM_CPUID_SIGNATURE, 0, &nent,
		     cpuid->nent);

	r = -E2BIG;
	if (nent >= cpuid->nent)
		goto out_free;

	do_cpuid_ent(&cpuid_entries[nent], KVM_CPUID_FEATURES, 0, &nent,
		     cpuid->nent);

2071 2072 2073 2074
	r = -E2BIG;
	if (nent >= cpuid->nent)
		goto out_free;

2075 2076
	r = -EFAULT;
	if (copy_to_user(entries, cpuid_entries,
2077
			 nent * sizeof(struct kvm_cpuid_entry2)))
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
		goto out_free;
	cpuid->nent = nent;
	r = 0;

out_free:
	vfree(cpuid_entries);
out:
	return r;
}

2088 2089 2090 2091
static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
2092
	memcpy(s->regs, vcpu->arch.apic->regs, sizeof *s);
2093 2094 2095 2096 2097 2098 2099 2100 2101
	vcpu_put(vcpu);

	return 0;
}

static int kvm_vcpu_ioctl_set_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
2102
	memcpy(vcpu->arch.apic->regs, s->regs, sizeof *s);
2103
	kvm_apic_post_state_restore(vcpu);
2104
	update_cr8_intercept(vcpu);
2105 2106 2107 2108 2109
	vcpu_put(vcpu);

	return 0;
}

2110 2111 2112 2113 2114 2115 2116 2117 2118
static int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu,
				    struct kvm_interrupt *irq)
{
	if (irq->irq < 0 || irq->irq >= 256)
		return -EINVAL;
	if (irqchip_in_kernel(vcpu->kvm))
		return -ENXIO;
	vcpu_load(vcpu);

2119
	kvm_queue_interrupt(vcpu, irq->irq, false);
2120 2121 2122 2123 2124 2125

	vcpu_put(vcpu);

	return 0;
}

2126 2127 2128 2129 2130 2131 2132 2133 2134
static int kvm_vcpu_ioctl_nmi(struct kvm_vcpu *vcpu)
{
	vcpu_load(vcpu);
	kvm_inject_nmi(vcpu);
	vcpu_put(vcpu);

	return 0;
}

2135 2136 2137 2138 2139 2140 2141 2142 2143
static int vcpu_ioctl_tpr_access_reporting(struct kvm_vcpu *vcpu,
					   struct kvm_tpr_access_ctl *tac)
{
	if (tac->flags)
		return -EINVAL;
	vcpu->arch.tpr_access_reporting = !!tac->enabled;
	return 0;
}

H
Huang Ying 已提交
2144 2145 2146 2147 2148 2149
static int kvm_vcpu_ioctl_x86_setup_mce(struct kvm_vcpu *vcpu,
					u64 mcg_cap)
{
	int r;
	unsigned bank_num = mcg_cap & 0xff, bank;

2150
	vcpu_load(vcpu);
H
Huang Ying 已提交
2151
	r = -EINVAL;
2152
	if (!bank_num || bank_num >= KVM_MAX_MCE_BANKS)
H
Huang Ying 已提交
2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164
		goto out;
	if (mcg_cap & ~(KVM_MCE_CAP_SUPPORTED | 0xff | 0xff0000))
		goto out;
	r = 0;
	vcpu->arch.mcg_cap = mcg_cap;
	/* Init IA32_MCG_CTL to all 1s */
	if (mcg_cap & MCG_CTL_P)
		vcpu->arch.mcg_ctl = ~(u64)0;
	/* Init IA32_MCi_CTL to all 1s */
	for (bank = 0; bank < bank_num; bank++)
		vcpu->arch.mce_banks[bank*4] = ~(u64)0;
out:
2165
	vcpu_put(vcpu);
H
Huang Ying 已提交
2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193
	return r;
}

static int kvm_vcpu_ioctl_x86_set_mce(struct kvm_vcpu *vcpu,
				      struct kvm_x86_mce *mce)
{
	u64 mcg_cap = vcpu->arch.mcg_cap;
	unsigned bank_num = mcg_cap & 0xff;
	u64 *banks = vcpu->arch.mce_banks;

	if (mce->bank >= bank_num || !(mce->status & MCI_STATUS_VAL))
		return -EINVAL;
	/*
	 * if IA32_MCG_CTL is not all 1s, the uncorrected error
	 * reporting is disabled
	 */
	if ((mce->status & MCI_STATUS_UC) && (mcg_cap & MCG_CTL_P) &&
	    vcpu->arch.mcg_ctl != ~(u64)0)
		return 0;
	banks += 4 * mce->bank;
	/*
	 * if IA32_MCi_CTL is not all 1s, the uncorrected error
	 * reporting is disabled for the bank
	 */
	if ((mce->status & MCI_STATUS_UC) && banks[0] != ~(u64)0)
		return 0;
	if (mce->status & MCI_STATUS_UC) {
		if ((vcpu->arch.mcg_status & MCG_STATUS_MCIP) ||
2194
		    !kvm_read_cr4_bits(vcpu, X86_CR4_MCE)) {
H
Huang Ying 已提交
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219
			printk(KERN_DEBUG "kvm: set_mce: "
			       "injects mce exception while "
			       "previous one is in progress!\n");
			set_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests);
			return 0;
		}
		if (banks[1] & MCI_STATUS_VAL)
			mce->status |= MCI_STATUS_OVER;
		banks[2] = mce->addr;
		banks[3] = mce->misc;
		vcpu->arch.mcg_status = mce->mcg_status;
		banks[1] = mce->status;
		kvm_queue_exception(vcpu, MC_VECTOR);
	} else if (!(banks[1] & MCI_STATUS_VAL)
		   || !(banks[1] & MCI_STATUS_UC)) {
		if (banks[1] & MCI_STATUS_VAL)
			mce->status |= MCI_STATUS_OVER;
		banks[2] = mce->addr;
		banks[3] = mce->misc;
		banks[1] = mce->status;
	} else
		banks[1] |= MCI_STATUS_OVER;
	return 0;
}

J
Jan Kiszka 已提交
2220 2221 2222 2223 2224
static void kvm_vcpu_ioctl_x86_get_vcpu_events(struct kvm_vcpu *vcpu,
					       struct kvm_vcpu_events *events)
{
	vcpu_load(vcpu);

2225 2226 2227
	events->exception.injected =
		vcpu->arch.exception.pending &&
		!kvm_exception_is_soft(vcpu->arch.exception.nr);
J
Jan Kiszka 已提交
2228 2229 2230 2231
	events->exception.nr = vcpu->arch.exception.nr;
	events->exception.has_error_code = vcpu->arch.exception.has_error_code;
	events->exception.error_code = vcpu->arch.exception.error_code;

2232 2233
	events->interrupt.injected =
		vcpu->arch.interrupt.pending && !vcpu->arch.interrupt.soft;
J
Jan Kiszka 已提交
2234
	events->interrupt.nr = vcpu->arch.interrupt.nr;
2235
	events->interrupt.soft = 0;
2236 2237 2238
	events->interrupt.shadow =
		kvm_x86_ops->get_interrupt_shadow(vcpu,
			KVM_X86_SHADOW_INT_MOV_SS | KVM_X86_SHADOW_INT_STI);
J
Jan Kiszka 已提交
2239 2240 2241 2242 2243 2244 2245

	events->nmi.injected = vcpu->arch.nmi_injected;
	events->nmi.pending = vcpu->arch.nmi_pending;
	events->nmi.masked = kvm_x86_ops->get_nmi_mask(vcpu);

	events->sipi_vector = vcpu->arch.sipi_vector;

2246
	events->flags = (KVM_VCPUEVENT_VALID_NMI_PENDING
2247 2248
			 | KVM_VCPUEVENT_VALID_SIPI_VECTOR
			 | KVM_VCPUEVENT_VALID_SHADOW);
J
Jan Kiszka 已提交
2249 2250 2251 2252 2253 2254 2255

	vcpu_put(vcpu);
}

static int kvm_vcpu_ioctl_x86_set_vcpu_events(struct kvm_vcpu *vcpu,
					      struct kvm_vcpu_events *events)
{
2256
	if (events->flags & ~(KVM_VCPUEVENT_VALID_NMI_PENDING
2257 2258
			      | KVM_VCPUEVENT_VALID_SIPI_VECTOR
			      | KVM_VCPUEVENT_VALID_SHADOW))
J
Jan Kiszka 已提交
2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272
		return -EINVAL;

	vcpu_load(vcpu);

	vcpu->arch.exception.pending = events->exception.injected;
	vcpu->arch.exception.nr = events->exception.nr;
	vcpu->arch.exception.has_error_code = events->exception.has_error_code;
	vcpu->arch.exception.error_code = events->exception.error_code;

	vcpu->arch.interrupt.pending = events->interrupt.injected;
	vcpu->arch.interrupt.nr = events->interrupt.nr;
	vcpu->arch.interrupt.soft = events->interrupt.soft;
	if (vcpu->arch.interrupt.pending && irqchip_in_kernel(vcpu->kvm))
		kvm_pic_clear_isr_ack(vcpu->kvm);
2273 2274 2275
	if (events->flags & KVM_VCPUEVENT_VALID_SHADOW)
		kvm_x86_ops->set_interrupt_shadow(vcpu,
						  events->interrupt.shadow);
J
Jan Kiszka 已提交
2276 2277

	vcpu->arch.nmi_injected = events->nmi.injected;
2278 2279
	if (events->flags & KVM_VCPUEVENT_VALID_NMI_PENDING)
		vcpu->arch.nmi_pending = events->nmi.pending;
J
Jan Kiszka 已提交
2280 2281
	kvm_x86_ops->set_nmi_mask(vcpu, events->nmi.masked);

2282 2283
	if (events->flags & KVM_VCPUEVENT_VALID_SIPI_VECTOR)
		vcpu->arch.sipi_vector = events->sipi_vector;
J
Jan Kiszka 已提交
2284 2285 2286 2287 2288 2289

	vcpu_put(vcpu);

	return 0;
}

2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319
static void kvm_vcpu_ioctl_x86_get_debugregs(struct kvm_vcpu *vcpu,
					     struct kvm_debugregs *dbgregs)
{
	vcpu_load(vcpu);

	memcpy(dbgregs->db, vcpu->arch.db, sizeof(vcpu->arch.db));
	dbgregs->dr6 = vcpu->arch.dr6;
	dbgregs->dr7 = vcpu->arch.dr7;
	dbgregs->flags = 0;

	vcpu_put(vcpu);
}

static int kvm_vcpu_ioctl_x86_set_debugregs(struct kvm_vcpu *vcpu,
					    struct kvm_debugregs *dbgregs)
{
	if (dbgregs->flags)
		return -EINVAL;

	vcpu_load(vcpu);

	memcpy(vcpu->arch.db, dbgregs->db, sizeof(vcpu->arch.db));
	vcpu->arch.dr6 = dbgregs->dr6;
	vcpu->arch.dr7 = dbgregs->dr7;

	vcpu_put(vcpu);

	return 0;
}

2320 2321 2322 2323 2324 2325
long kvm_arch_vcpu_ioctl(struct file *filp,
			 unsigned int ioctl, unsigned long arg)
{
	struct kvm_vcpu *vcpu = filp->private_data;
	void __user *argp = (void __user *)arg;
	int r;
2326
	struct kvm_lapic_state *lapic = NULL;
2327 2328 2329

	switch (ioctl) {
	case KVM_GET_LAPIC: {
2330 2331 2332
		r = -EINVAL;
		if (!vcpu->arch.apic)
			goto out;
2333
		lapic = kzalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
2334

2335 2336 2337 2338
		r = -ENOMEM;
		if (!lapic)
			goto out;
		r = kvm_vcpu_ioctl_get_lapic(vcpu, lapic);
2339 2340 2341
		if (r)
			goto out;
		r = -EFAULT;
2342
		if (copy_to_user(argp, lapic, sizeof(struct kvm_lapic_state)))
2343 2344 2345 2346 2347
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_LAPIC: {
2348 2349 2350
		r = -EINVAL;
		if (!vcpu->arch.apic)
			goto out;
2351 2352 2353 2354
		lapic = kmalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
		r = -ENOMEM;
		if (!lapic)
			goto out;
2355
		r = -EFAULT;
2356
		if (copy_from_user(lapic, argp, sizeof(struct kvm_lapic_state)))
2357
			goto out;
2358
		r = kvm_vcpu_ioctl_set_lapic(vcpu, lapic);
2359 2360 2361 2362 2363
		if (r)
			goto out;
		r = 0;
		break;
	}
2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375
	case KVM_INTERRUPT: {
		struct kvm_interrupt irq;

		r = -EFAULT;
		if (copy_from_user(&irq, argp, sizeof irq))
			goto out;
		r = kvm_vcpu_ioctl_interrupt(vcpu, &irq);
		if (r)
			goto out;
		r = 0;
		break;
	}
2376 2377 2378 2379 2380 2381 2382
	case KVM_NMI: {
		r = kvm_vcpu_ioctl_nmi(vcpu);
		if (r)
			goto out;
		r = 0;
		break;
	}
2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
	case KVM_SET_CPUID: {
		struct kvm_cpuid __user *cpuid_arg = argp;
		struct kvm_cpuid cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_set_cpuid(vcpu, &cpuid, cpuid_arg->entries);
		if (r)
			goto out;
		break;
	}
2395 2396 2397 2398 2399 2400 2401 2402
	case KVM_SET_CPUID2: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_set_cpuid2(vcpu, &cpuid,
2403
					      cpuid_arg->entries);
2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415
		if (r)
			goto out;
		break;
	}
	case KVM_GET_CPUID2: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_get_cpuid2(vcpu, &cpuid,
2416
					      cpuid_arg->entries);
2417 2418 2419 2420 2421 2422 2423 2424
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
2425 2426 2427 2428 2429 2430
	case KVM_GET_MSRS:
		r = msr_io(vcpu, argp, kvm_get_msr, 1);
		break;
	case KVM_SET_MSRS:
		r = msr_io(vcpu, argp, do_set_msr, 0);
		break;
2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445
	case KVM_TPR_ACCESS_REPORTING: {
		struct kvm_tpr_access_ctl tac;

		r = -EFAULT;
		if (copy_from_user(&tac, argp, sizeof tac))
			goto out;
		r = vcpu_ioctl_tpr_access_reporting(vcpu, &tac);
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(argp, &tac, sizeof tac))
			goto out;
		r = 0;
		break;
	};
A
Avi Kivity 已提交
2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458
	case KVM_SET_VAPIC_ADDR: {
		struct kvm_vapic_addr va;

		r = -EINVAL;
		if (!irqchip_in_kernel(vcpu->kvm))
			goto out;
		r = -EFAULT;
		if (copy_from_user(&va, argp, sizeof va))
			goto out;
		r = 0;
		kvm_lapic_set_vapic_addr(vcpu, va.vapic_addr);
		break;
	}
H
Huang Ying 已提交
2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473
	case KVM_X86_SETUP_MCE: {
		u64 mcg_cap;

		r = -EFAULT;
		if (copy_from_user(&mcg_cap, argp, sizeof mcg_cap))
			goto out;
		r = kvm_vcpu_ioctl_x86_setup_mce(vcpu, mcg_cap);
		break;
	}
	case KVM_X86_SET_MCE: {
		struct kvm_x86_mce mce;

		r = -EFAULT;
		if (copy_from_user(&mce, argp, sizeof mce))
			goto out;
2474
		vcpu_load(vcpu);
H
Huang Ying 已提交
2475
		r = kvm_vcpu_ioctl_x86_set_mce(vcpu, &mce);
2476
		vcpu_put(vcpu);
H
Huang Ying 已提交
2477 2478
		break;
	}
J
Jan Kiszka 已提交
2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499
	case KVM_GET_VCPU_EVENTS: {
		struct kvm_vcpu_events events;

		kvm_vcpu_ioctl_x86_get_vcpu_events(vcpu, &events);

		r = -EFAULT;
		if (copy_to_user(argp, &events, sizeof(struct kvm_vcpu_events)))
			break;
		r = 0;
		break;
	}
	case KVM_SET_VCPU_EVENTS: {
		struct kvm_vcpu_events events;

		r = -EFAULT;
		if (copy_from_user(&events, argp, sizeof(struct kvm_vcpu_events)))
			break;

		r = kvm_vcpu_ioctl_x86_set_vcpu_events(vcpu, &events);
		break;
	}
2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522
	case KVM_GET_DEBUGREGS: {
		struct kvm_debugregs dbgregs;

		kvm_vcpu_ioctl_x86_get_debugregs(vcpu, &dbgregs);

		r = -EFAULT;
		if (copy_to_user(argp, &dbgregs,
				 sizeof(struct kvm_debugregs)))
			break;
		r = 0;
		break;
	}
	case KVM_SET_DEBUGREGS: {
		struct kvm_debugregs dbgregs;

		r = -EFAULT;
		if (copy_from_user(&dbgregs, argp,
				   sizeof(struct kvm_debugregs)))
			break;

		r = kvm_vcpu_ioctl_x86_set_debugregs(vcpu, &dbgregs);
		break;
	}
2523 2524 2525 2526
	default:
		r = -EINVAL;
	}
out:
2527
	kfree(lapic);
2528 2529 2530
	return r;
}

2531 2532 2533 2534 2535 2536 2537 2538 2539 2540
static int kvm_vm_ioctl_set_tss_addr(struct kvm *kvm, unsigned long addr)
{
	int ret;

	if (addr > (unsigned int)(-3 * PAGE_SIZE))
		return -1;
	ret = kvm_x86_ops->set_tss_addr(kvm, addr);
	return ret;
}

2541 2542 2543 2544 2545 2546 2547
static int kvm_vm_ioctl_set_identity_map_addr(struct kvm *kvm,
					      u64 ident_addr)
{
	kvm->arch.ept_identity_map_addr = ident_addr;
	return 0;
}

2548 2549 2550 2551 2552 2553
static int kvm_vm_ioctl_set_nr_mmu_pages(struct kvm *kvm,
					  u32 kvm_nr_mmu_pages)
{
	if (kvm_nr_mmu_pages < KVM_MIN_ALLOC_MMU_PAGES)
		return -EINVAL;

2554
	mutex_lock(&kvm->slots_lock);
2555
	spin_lock(&kvm->mmu_lock);
2556 2557

	kvm_mmu_change_mmu_pages(kvm, kvm_nr_mmu_pages);
2558
	kvm->arch.n_requested_mmu_pages = kvm_nr_mmu_pages;
2559

2560
	spin_unlock(&kvm->mmu_lock);
2561
	mutex_unlock(&kvm->slots_lock);
2562 2563 2564 2565 2566
	return 0;
}

static int kvm_vm_ioctl_get_nr_mmu_pages(struct kvm *kvm)
{
2567
	return kvm->arch.n_alloc_mmu_pages;
2568 2569
}

2570 2571 2572 2573 2574 2575
gfn_t unalias_gfn_instantiation(struct kvm *kvm, gfn_t gfn)
{
	int i;
	struct kvm_mem_alias *alias;
	struct kvm_mem_aliases *aliases;

2576
	aliases = kvm_aliases(kvm);
2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588

	for (i = 0; i < aliases->naliases; ++i) {
		alias = &aliases->aliases[i];
		if (alias->flags & KVM_ALIAS_INVALID)
			continue;
		if (gfn >= alias->base_gfn
		    && gfn < alias->base_gfn + alias->npages)
			return alias->target_gfn + gfn - alias->base_gfn;
	}
	return gfn;
}

2589 2590 2591 2592
gfn_t unalias_gfn(struct kvm *kvm, gfn_t gfn)
{
	int i;
	struct kvm_mem_alias *alias;
2593 2594
	struct kvm_mem_aliases *aliases;

2595
	aliases = kvm_aliases(kvm);
2596

2597 2598
	for (i = 0; i < aliases->naliases; ++i) {
		alias = &aliases->aliases[i];
2599 2600 2601 2602 2603 2604 2605
		if (gfn >= alias->base_gfn
		    && gfn < alias->base_gfn + alias->npages)
			return alias->target_gfn + gfn - alias->base_gfn;
	}
	return gfn;
}

2606 2607 2608 2609 2610 2611 2612 2613 2614 2615
/*
 * Set a new alias region.  Aliases map a portion of physical memory into
 * another portion.  This is useful for memory windows, for example the PC
 * VGA region.
 */
static int kvm_vm_ioctl_set_memory_alias(struct kvm *kvm,
					 struct kvm_memory_alias *alias)
{
	int r, n;
	struct kvm_mem_alias *p;
2616
	struct kvm_mem_aliases *aliases, *old_aliases;
2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632

	r = -EINVAL;
	/* General sanity checks */
	if (alias->memory_size & (PAGE_SIZE - 1))
		goto out;
	if (alias->guest_phys_addr & (PAGE_SIZE - 1))
		goto out;
	if (alias->slot >= KVM_ALIAS_SLOTS)
		goto out;
	if (alias->guest_phys_addr + alias->memory_size
	    < alias->guest_phys_addr)
		goto out;
	if (alias->target_phys_addr + alias->memory_size
	    < alias->target_phys_addr)
		goto out;

2633 2634 2635 2636 2637
	r = -ENOMEM;
	aliases = kzalloc(sizeof(struct kvm_mem_aliases), GFP_KERNEL);
	if (!aliases)
		goto out;

2638
	mutex_lock(&kvm->slots_lock);
2639

2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654
	/* invalidate any gfn reference in case of deletion/shrinking */
	memcpy(aliases, kvm->arch.aliases, sizeof(struct kvm_mem_aliases));
	aliases->aliases[alias->slot].flags |= KVM_ALIAS_INVALID;
	old_aliases = kvm->arch.aliases;
	rcu_assign_pointer(kvm->arch.aliases, aliases);
	synchronize_srcu_expedited(&kvm->srcu);
	kvm_mmu_zap_all(kvm);
	kfree(old_aliases);

	r = -ENOMEM;
	aliases = kzalloc(sizeof(struct kvm_mem_aliases), GFP_KERNEL);
	if (!aliases)
		goto out_unlock;

	memcpy(aliases, kvm->arch.aliases, sizeof(struct kvm_mem_aliases));
2655 2656

	p = &aliases->aliases[alias->slot];
2657 2658 2659
	p->base_gfn = alias->guest_phys_addr >> PAGE_SHIFT;
	p->npages = alias->memory_size >> PAGE_SHIFT;
	p->target_gfn = alias->target_phys_addr >> PAGE_SHIFT;
2660
	p->flags &= ~(KVM_ALIAS_INVALID);
2661 2662

	for (n = KVM_ALIAS_SLOTS; n > 0; --n)
2663
		if (aliases->aliases[n - 1].npages)
2664
			break;
2665
	aliases->naliases = n;
2666

2667 2668 2669 2670 2671
	old_aliases = kvm->arch.aliases;
	rcu_assign_pointer(kvm->arch.aliases, aliases);
	synchronize_srcu_expedited(&kvm->srcu);
	kfree(old_aliases);
	r = 0;
2672

2673
out_unlock:
2674
	mutex_unlock(&kvm->slots_lock);
2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695
out:
	return r;
}

static int kvm_vm_ioctl_get_irqchip(struct kvm *kvm, struct kvm_irqchip *chip)
{
	int r;

	r = 0;
	switch (chip->chip_id) {
	case KVM_IRQCHIP_PIC_MASTER:
		memcpy(&chip->chip.pic,
			&pic_irqchip(kvm)->pics[0],
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_PIC_SLAVE:
		memcpy(&chip->chip.pic,
			&pic_irqchip(kvm)->pics[1],
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_IOAPIC:
G
Gleb Natapov 已提交
2696
		r = kvm_get_ioapic(kvm, &chip->chip.ioapic);
2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711
		break;
	default:
		r = -EINVAL;
		break;
	}
	return r;
}

static int kvm_vm_ioctl_set_irqchip(struct kvm *kvm, struct kvm_irqchip *chip)
{
	int r;

	r = 0;
	switch (chip->chip_id) {
	case KVM_IRQCHIP_PIC_MASTER:
2712
		raw_spin_lock(&pic_irqchip(kvm)->lock);
2713 2714 2715
		memcpy(&pic_irqchip(kvm)->pics[0],
			&chip->chip.pic,
			sizeof(struct kvm_pic_state));
2716
		raw_spin_unlock(&pic_irqchip(kvm)->lock);
2717 2718
		break;
	case KVM_IRQCHIP_PIC_SLAVE:
2719
		raw_spin_lock(&pic_irqchip(kvm)->lock);
2720 2721 2722
		memcpy(&pic_irqchip(kvm)->pics[1],
			&chip->chip.pic,
			sizeof(struct kvm_pic_state));
2723
		raw_spin_unlock(&pic_irqchip(kvm)->lock);
2724 2725
		break;
	case KVM_IRQCHIP_IOAPIC:
G
Gleb Natapov 已提交
2726
		r = kvm_set_ioapic(kvm, &chip->chip.ioapic);
2727 2728 2729 2730 2731 2732 2733 2734 2735
		break;
	default:
		r = -EINVAL;
		break;
	}
	kvm_pic_update_irq(pic_irqchip(kvm));
	return r;
}

2736 2737 2738 2739
static int kvm_vm_ioctl_get_pit(struct kvm *kvm, struct kvm_pit_state *ps)
{
	int r = 0;

2740
	mutex_lock(&kvm->arch.vpit->pit_state.lock);
2741
	memcpy(ps, &kvm->arch.vpit->pit_state, sizeof(struct kvm_pit_state));
2742
	mutex_unlock(&kvm->arch.vpit->pit_state.lock);
2743 2744 2745 2746 2747 2748 2749
	return r;
}

static int kvm_vm_ioctl_set_pit(struct kvm *kvm, struct kvm_pit_state *ps)
{
	int r = 0;

2750
	mutex_lock(&kvm->arch.vpit->pit_state.lock);
2751
	memcpy(&kvm->arch.vpit->pit_state, ps, sizeof(struct kvm_pit_state));
B
Beth Kon 已提交
2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781
	kvm_pit_load_count(kvm, 0, ps->channels[0].count, 0);
	mutex_unlock(&kvm->arch.vpit->pit_state.lock);
	return r;
}

static int kvm_vm_ioctl_get_pit2(struct kvm *kvm, struct kvm_pit_state2 *ps)
{
	int r = 0;

	mutex_lock(&kvm->arch.vpit->pit_state.lock);
	memcpy(ps->channels, &kvm->arch.vpit->pit_state.channels,
		sizeof(ps->channels));
	ps->flags = kvm->arch.vpit->pit_state.flags;
	mutex_unlock(&kvm->arch.vpit->pit_state.lock);
	return r;
}

static int kvm_vm_ioctl_set_pit2(struct kvm *kvm, struct kvm_pit_state2 *ps)
{
	int r = 0, start = 0;
	u32 prev_legacy, cur_legacy;
	mutex_lock(&kvm->arch.vpit->pit_state.lock);
	prev_legacy = kvm->arch.vpit->pit_state.flags & KVM_PIT_FLAGS_HPET_LEGACY;
	cur_legacy = ps->flags & KVM_PIT_FLAGS_HPET_LEGACY;
	if (!prev_legacy && cur_legacy)
		start = 1;
	memcpy(&kvm->arch.vpit->pit_state.channels, &ps->channels,
	       sizeof(kvm->arch.vpit->pit_state.channels));
	kvm->arch.vpit->pit_state.flags = ps->flags;
	kvm_pit_load_count(kvm, 0, kvm->arch.vpit->pit_state.channels[0].count, start);
2782
	mutex_unlock(&kvm->arch.vpit->pit_state.lock);
2783 2784 2785
	return r;
}

2786 2787 2788 2789 2790
static int kvm_vm_ioctl_reinject(struct kvm *kvm,
				 struct kvm_reinject_control *control)
{
	if (!kvm->arch.vpit)
		return -ENXIO;
2791
	mutex_lock(&kvm->arch.vpit->pit_state.lock);
2792
	kvm->arch.vpit->pit_state.pit_timer.reinject = control->pit_reinject;
2793
	mutex_unlock(&kvm->arch.vpit->pit_state.lock);
2794 2795 2796
	return 0;
}

2797 2798 2799 2800 2801 2802
/*
 * Get (and clear) the dirty memory log for a memory slot.
 */
int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
				      struct kvm_dirty_log *log)
{
2803
	int r, i;
2804
	struct kvm_memory_slot *memslot;
2805
	unsigned long n;
M
Marcelo Tosatti 已提交
2806
	unsigned long is_dirty = 0;
2807

2808
	mutex_lock(&kvm->slots_lock);
2809

M
Marcelo Tosatti 已提交
2810 2811 2812 2813 2814 2815 2816 2817 2818
	r = -EINVAL;
	if (log->slot >= KVM_MEMORY_SLOTS)
		goto out;

	memslot = &kvm->memslots->memslots[log->slot];
	r = -ENOENT;
	if (!memslot->dirty_bitmap)
		goto out;

2819
	n = kvm_dirty_bitmap_bytes(memslot);
M
Marcelo Tosatti 已提交
2820 2821 2822

	for (i = 0; !is_dirty && i < n/sizeof(long); i++)
		is_dirty = memslot->dirty_bitmap[i];
2823 2824 2825

	/* If nothing is dirty, don't bother messing with page tables. */
	if (is_dirty) {
M
Marcelo Tosatti 已提交
2826
		struct kvm_memslots *slots, *old_slots;
2827
		unsigned long *dirty_bitmap;
M
Marcelo Tosatti 已提交
2828

2829
		spin_lock(&kvm->mmu_lock);
2830
		kvm_mmu_slot_remove_write_access(kvm, log->slot);
2831
		spin_unlock(&kvm->mmu_lock);
M
Marcelo Tosatti 已提交
2832

2833 2834 2835 2836 2837
		r = -ENOMEM;
		dirty_bitmap = vmalloc(n);
		if (!dirty_bitmap)
			goto out;
		memset(dirty_bitmap, 0, n);
M
Marcelo Tosatti 已提交
2838

2839 2840 2841 2842 2843 2844
		r = -ENOMEM;
		slots = kzalloc(sizeof(struct kvm_memslots), GFP_KERNEL);
		if (!slots) {
			vfree(dirty_bitmap);
			goto out;
		}
M
Marcelo Tosatti 已提交
2845 2846 2847 2848 2849 2850 2851 2852
		memcpy(slots, kvm->memslots, sizeof(struct kvm_memslots));
		slots->memslots[log->slot].dirty_bitmap = dirty_bitmap;

		old_slots = kvm->memslots;
		rcu_assign_pointer(kvm->memslots, slots);
		synchronize_srcu_expedited(&kvm->srcu);
		dirty_bitmap = old_slots->memslots[log->slot].dirty_bitmap;
		kfree(old_slots);
2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863

		r = -EFAULT;
		if (copy_to_user(log->dirty_bitmap, dirty_bitmap, n)) {
			vfree(dirty_bitmap);
			goto out;
		}
		vfree(dirty_bitmap);
	} else {
		r = -EFAULT;
		if (clear_user(log->dirty_bitmap, n))
			goto out;
2864
	}
M
Marcelo Tosatti 已提交
2865

2866 2867
	r = 0;
out:
2868
	mutex_unlock(&kvm->slots_lock);
2869 2870 2871
	return r;
}

2872 2873 2874 2875 2876
long kvm_arch_vm_ioctl(struct file *filp,
		       unsigned int ioctl, unsigned long arg)
{
	struct kvm *kvm = filp->private_data;
	void __user *argp = (void __user *)arg;
2877
	int r = -ENOTTY;
2878 2879 2880 2881 2882 2883 2884
	/*
	 * This union makes it completely explicit to gcc-3.x
	 * that these two variables' stack usage should be
	 * combined, not added together.
	 */
	union {
		struct kvm_pit_state ps;
B
Beth Kon 已提交
2885
		struct kvm_pit_state2 ps2;
2886
		struct kvm_memory_alias alias;
2887
		struct kvm_pit_config pit_config;
2888
	} u;
2889 2890 2891 2892 2893 2894 2895

	switch (ioctl) {
	case KVM_SET_TSS_ADDR:
		r = kvm_vm_ioctl_set_tss_addr(kvm, arg);
		if (r < 0)
			goto out;
		break;
2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906
	case KVM_SET_IDENTITY_MAP_ADDR: {
		u64 ident_addr;

		r = -EFAULT;
		if (copy_from_user(&ident_addr, argp, sizeof ident_addr))
			goto out;
		r = kvm_vm_ioctl_set_identity_map_addr(kvm, ident_addr);
		if (r < 0)
			goto out;
		break;
	}
2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930
	case KVM_SET_MEMORY_REGION: {
		struct kvm_memory_region kvm_mem;
		struct kvm_userspace_memory_region kvm_userspace_mem;

		r = -EFAULT;
		if (copy_from_user(&kvm_mem, argp, sizeof kvm_mem))
			goto out;
		kvm_userspace_mem.slot = kvm_mem.slot;
		kvm_userspace_mem.flags = kvm_mem.flags;
		kvm_userspace_mem.guest_phys_addr = kvm_mem.guest_phys_addr;
		kvm_userspace_mem.memory_size = kvm_mem.memory_size;
		r = kvm_vm_ioctl_set_memory_region(kvm, &kvm_userspace_mem, 0);
		if (r)
			goto out;
		break;
	}
	case KVM_SET_NR_MMU_PAGES:
		r = kvm_vm_ioctl_set_nr_mmu_pages(kvm, arg);
		if (r)
			goto out;
		break;
	case KVM_GET_NR_MMU_PAGES:
		r = kvm_vm_ioctl_get_nr_mmu_pages(kvm);
		break;
2931
	case KVM_SET_MEMORY_ALIAS:
2932
		r = -EFAULT;
2933
		if (copy_from_user(&u.alias, argp, sizeof(struct kvm_memory_alias)))
2934
			goto out;
2935
		r = kvm_vm_ioctl_set_memory_alias(kvm, &u.alias);
2936 2937 2938
		if (r)
			goto out;
		break;
2939 2940 2941 2942 2943 2944 2945
	case KVM_CREATE_IRQCHIP: {
		struct kvm_pic *vpic;

		mutex_lock(&kvm->lock);
		r = -EEXIST;
		if (kvm->arch.vpic)
			goto create_irqchip_unlock;
2946
		r = -ENOMEM;
2947 2948
		vpic = kvm_create_pic(kvm);
		if (vpic) {
2949 2950
			r = kvm_ioapic_init(kvm);
			if (r) {
2951 2952
				kvm_io_bus_unregister_dev(kvm, KVM_PIO_BUS,
							  &vpic->dev);
2953 2954
				kfree(vpic);
				goto create_irqchip_unlock;
2955 2956
			}
		} else
2957 2958 2959 2960
			goto create_irqchip_unlock;
		smp_wmb();
		kvm->arch.vpic = vpic;
		smp_wmb();
2961 2962
		r = kvm_setup_default_irq_routing(kvm);
		if (r) {
2963
			mutex_lock(&kvm->irq_lock);
2964 2965
			kvm_ioapic_destroy(kvm);
			kvm_destroy_pic(kvm);
2966
			mutex_unlock(&kvm->irq_lock);
2967
		}
2968 2969
	create_irqchip_unlock:
		mutex_unlock(&kvm->lock);
2970
		break;
2971
	}
S
Sheng Yang 已提交
2972
	case KVM_CREATE_PIT:
2973 2974 2975 2976 2977 2978 2979 2980
		u.pit_config.flags = KVM_PIT_SPEAKER_DUMMY;
		goto create_pit;
	case KVM_CREATE_PIT2:
		r = -EFAULT;
		if (copy_from_user(&u.pit_config, argp,
				   sizeof(struct kvm_pit_config)))
			goto out;
	create_pit:
2981
		mutex_lock(&kvm->slots_lock);
A
Avi Kivity 已提交
2982 2983 2984
		r = -EEXIST;
		if (kvm->arch.vpit)
			goto create_pit_unlock;
S
Sheng Yang 已提交
2985
		r = -ENOMEM;
2986
		kvm->arch.vpit = kvm_create_pit(kvm, u.pit_config.flags);
S
Sheng Yang 已提交
2987 2988
		if (kvm->arch.vpit)
			r = 0;
A
Avi Kivity 已提交
2989
	create_pit_unlock:
2990
		mutex_unlock(&kvm->slots_lock);
S
Sheng Yang 已提交
2991
		break;
2992
	case KVM_IRQ_LINE_STATUS:
2993 2994 2995 2996 2997 2998
	case KVM_IRQ_LINE: {
		struct kvm_irq_level irq_event;

		r = -EFAULT;
		if (copy_from_user(&irq_event, argp, sizeof irq_event))
			goto out;
2999
		r = -ENXIO;
3000
		if (irqchip_in_kernel(kvm)) {
3001 3002 3003 3004
			__s32 status;
			status = kvm_set_irq(kvm, KVM_USERSPACE_IRQ_SOURCE_ID,
					irq_event.irq, irq_event.level);
			if (ioctl == KVM_IRQ_LINE_STATUS) {
3005
				r = -EFAULT;
3006 3007 3008 3009 3010
				irq_event.status = status;
				if (copy_to_user(argp, &irq_event,
							sizeof irq_event))
					goto out;
			}
3011 3012 3013 3014 3015 3016
			r = 0;
		}
		break;
	}
	case KVM_GET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
3017
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
3018

3019 3020
		r = -ENOMEM;
		if (!chip)
3021
			goto out;
3022 3023 3024
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto get_irqchip_out;
3025 3026
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
3027 3028
			goto get_irqchip_out;
		r = kvm_vm_ioctl_get_irqchip(kvm, chip);
3029
		if (r)
3030
			goto get_irqchip_out;
3031
		r = -EFAULT;
3032 3033
		if (copy_to_user(argp, chip, sizeof *chip))
			goto get_irqchip_out;
3034
		r = 0;
3035 3036 3037 3038
	get_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
3039 3040 3041 3042
		break;
	}
	case KVM_SET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
3043
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
3044

3045 3046
		r = -ENOMEM;
		if (!chip)
3047
			goto out;
3048 3049 3050
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto set_irqchip_out;
3051 3052
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
3053 3054
			goto set_irqchip_out;
		r = kvm_vm_ioctl_set_irqchip(kvm, chip);
3055
		if (r)
3056
			goto set_irqchip_out;
3057
		r = 0;
3058 3059 3060 3061
	set_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
3062 3063
		break;
	}
3064 3065
	case KVM_GET_PIT: {
		r = -EFAULT;
3066
		if (copy_from_user(&u.ps, argp, sizeof(struct kvm_pit_state)))
3067 3068 3069 3070
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
3071
		r = kvm_vm_ioctl_get_pit(kvm, &u.ps);
3072 3073 3074
		if (r)
			goto out;
		r = -EFAULT;
3075
		if (copy_to_user(argp, &u.ps, sizeof(struct kvm_pit_state)))
3076 3077 3078 3079 3080 3081
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_PIT: {
		r = -EFAULT;
3082
		if (copy_from_user(&u.ps, argp, sizeof u.ps))
3083 3084 3085 3086
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
3087
		r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
3088 3089 3090 3091 3092
		if (r)
			goto out;
		r = 0;
		break;
	}
B
Beth Kon 已提交
3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118
	case KVM_GET_PIT2: {
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
		r = kvm_vm_ioctl_get_pit2(kvm, &u.ps2);
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(argp, &u.ps2, sizeof(u.ps2)))
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_PIT2: {
		r = -EFAULT;
		if (copy_from_user(&u.ps2, argp, sizeof(u.ps2)))
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
		r = kvm_vm_ioctl_set_pit2(kvm, &u.ps2);
		if (r)
			goto out;
		r = 0;
		break;
	}
3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129
	case KVM_REINJECT_CONTROL: {
		struct kvm_reinject_control control;
		r =  -EFAULT;
		if (copy_from_user(&control, argp, sizeof(control)))
			goto out;
		r = kvm_vm_ioctl_reinject(kvm, &control);
		if (r)
			goto out;
		r = 0;
		break;
	}
E
Ed Swierk 已提交
3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
	case KVM_XEN_HVM_CONFIG: {
		r = -EFAULT;
		if (copy_from_user(&kvm->arch.xen_hvm_config, argp,
				   sizeof(struct kvm_xen_hvm_config)))
			goto out;
		r = -EINVAL;
		if (kvm->arch.xen_hvm_config.flags)
			goto out;
		r = 0;
		break;
	}
3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178
	case KVM_SET_CLOCK: {
		struct timespec now;
		struct kvm_clock_data user_ns;
		u64 now_ns;
		s64 delta;

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

		r = -EINVAL;
		if (user_ns.flags)
			goto out;

		r = 0;
		ktime_get_ts(&now);
		now_ns = timespec_to_ns(&now);
		delta = user_ns.clock - now_ns;
		kvm->arch.kvmclock_offset = delta;
		break;
	}
	case KVM_GET_CLOCK: {
		struct timespec now;
		struct kvm_clock_data user_ns;
		u64 now_ns;

		ktime_get_ts(&now);
		now_ns = timespec_to_ns(&now);
		user_ns.clock = kvm->arch.kvmclock_offset + now_ns;
		user_ns.flags = 0;

		r = -EFAULT;
		if (copy_to_user(argp, &user_ns, sizeof(user_ns)))
			goto out;
		r = 0;
		break;
	}

3179 3180 3181 3182 3183 3184 3185
	default:
		;
	}
out:
	return r;
}

3186
static void kvm_init_msr_list(void)
3187 3188 3189 3190
{
	u32 dummy[2];
	unsigned i, j;

3191 3192
	/* skip the first msrs in the list. KVM-specific */
	for (i = j = KVM_SAVE_MSRS_BEGIN; i < ARRAY_SIZE(msrs_to_save); i++) {
3193 3194 3195 3196 3197 3198 3199 3200 3201
		if (rdmsr_safe(msrs_to_save[i], &dummy[0], &dummy[1]) < 0)
			continue;
		if (j < i)
			msrs_to_save[j] = msrs_to_save[i];
		j++;
	}
	num_msrs_to_save = j;
}

3202 3203
static int vcpu_mmio_write(struct kvm_vcpu *vcpu, gpa_t addr, int len,
			   const void *v)
3204
{
3205 3206 3207
	if (vcpu->arch.apic &&
	    !kvm_iodevice_write(&vcpu->arch.apic->dev, addr, len, v))
		return 0;
3208

M
Marcelo Tosatti 已提交
3209
	return kvm_io_bus_write(vcpu->kvm, KVM_MMIO_BUS, addr, len, v);
3210 3211
}

3212
static int vcpu_mmio_read(struct kvm_vcpu *vcpu, gpa_t addr, int len, void *v)
3213
{
3214 3215 3216
	if (vcpu->arch.apic &&
	    !kvm_iodevice_read(&vcpu->arch.apic->dev, addr, len, v))
		return 0;
3217

M
Marcelo Tosatti 已提交
3218
	return kvm_io_bus_read(vcpu->kvm, KVM_MMIO_BUS, addr, len, v);
3219 3220
}

3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232
static void kvm_set_segment(struct kvm_vcpu *vcpu,
			struct kvm_segment *var, int seg)
{
	kvm_x86_ops->set_segment(vcpu, var, seg);
}

void kvm_get_segment(struct kvm_vcpu *vcpu,
		     struct kvm_segment *var, int seg)
{
	kvm_x86_ops->get_segment(vcpu, var, seg);
}

3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261
gpa_t kvm_mmu_gva_to_gpa_read(struct kvm_vcpu *vcpu, gva_t gva, u32 *error)
{
	u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
	return vcpu->arch.mmu.gva_to_gpa(vcpu, gva, access, error);
}

 gpa_t kvm_mmu_gva_to_gpa_fetch(struct kvm_vcpu *vcpu, gva_t gva, u32 *error)
{
	u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
	access |= PFERR_FETCH_MASK;
	return vcpu->arch.mmu.gva_to_gpa(vcpu, gva, access, error);
}

gpa_t kvm_mmu_gva_to_gpa_write(struct kvm_vcpu *vcpu, gva_t gva, u32 *error)
{
	u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
	access |= PFERR_WRITE_MASK;
	return vcpu->arch.mmu.gva_to_gpa(vcpu, gva, access, error);
}

/* uses this to access any guest's mapped memory without checking CPL */
gpa_t kvm_mmu_gva_to_gpa_system(struct kvm_vcpu *vcpu, gva_t gva, u32 *error)
{
	return vcpu->arch.mmu.gva_to_gpa(vcpu, gva, 0, error);
}

static int kvm_read_guest_virt_helper(gva_t addr, void *val, unsigned int bytes,
				      struct kvm_vcpu *vcpu, u32 access,
				      u32 *error)
3262 3263
{
	void *data = val;
3264
	int r = X86EMUL_CONTINUE;
3265 3266

	while (bytes) {
3267
		gpa_t gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr, access, error);
3268
		unsigned offset = addr & (PAGE_SIZE-1);
3269
		unsigned toread = min(bytes, (unsigned)PAGE_SIZE - offset);
3270 3271
		int ret;

3272 3273 3274 3275
		if (gpa == UNMAPPED_GVA) {
			r = X86EMUL_PROPAGATE_FAULT;
			goto out;
		}
3276
		ret = kvm_read_guest(vcpu->kvm, gpa, data, toread);
3277
		if (ret < 0) {
3278
			r = X86EMUL_IO_NEEDED;
3279 3280
			goto out;
		}
3281

3282 3283 3284
		bytes -= toread;
		data += toread;
		addr += toread;
3285
	}
3286 3287
out:
	return r;
3288
}
3289

3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312
/* used for instruction fetching */
static int kvm_fetch_guest_virt(gva_t addr, void *val, unsigned int bytes,
				struct kvm_vcpu *vcpu, u32 *error)
{
	u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
	return kvm_read_guest_virt_helper(addr, val, bytes, vcpu,
					  access | PFERR_FETCH_MASK, error);
}

static int kvm_read_guest_virt(gva_t addr, void *val, unsigned int bytes,
			       struct kvm_vcpu *vcpu, u32 *error)
{
	u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
	return kvm_read_guest_virt_helper(addr, val, bytes, vcpu, access,
					  error);
}

static int kvm_read_guest_virt_system(gva_t addr, void *val, unsigned int bytes,
			       struct kvm_vcpu *vcpu, u32 *error)
{
	return kvm_read_guest_virt_helper(addr, val, bytes, vcpu, 0, error);
}

3313
static int kvm_write_guest_virt_system(gva_t addr, void *val,
3314
				       unsigned int bytes,
3315
				       struct kvm_vcpu *vcpu,
3316
				       u32 *error)
3317 3318 3319 3320 3321
{
	void *data = val;
	int r = X86EMUL_CONTINUE;

	while (bytes) {
3322 3323
		gpa_t gpa =  vcpu->arch.mmu.gva_to_gpa(vcpu, addr,
						       PFERR_WRITE_MASK, error);
3324 3325 3326 3327 3328 3329 3330 3331 3332 3333
		unsigned offset = addr & (PAGE_SIZE-1);
		unsigned towrite = min(bytes, (unsigned)PAGE_SIZE - offset);
		int ret;

		if (gpa == UNMAPPED_GVA) {
			r = X86EMUL_PROPAGATE_FAULT;
			goto out;
		}
		ret = kvm_write_guest(vcpu->kvm, gpa, data, towrite);
		if (ret < 0) {
3334
			r = X86EMUL_IO_NEEDED;
3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345
			goto out;
		}

		bytes -= towrite;
		data += towrite;
		addr += towrite;
	}
out:
	return r;
}

3346 3347 3348
static int emulator_read_emulated(unsigned long addr,
				  void *val,
				  unsigned int bytes,
3349
				  unsigned int *error_code,
3350 3351 3352 3353 3354 3355
				  struct kvm_vcpu *vcpu)
{
	gpa_t                 gpa;

	if (vcpu->mmio_read_completed) {
		memcpy(val, vcpu->mmio_data, bytes);
A
Avi Kivity 已提交
3356 3357
		trace_kvm_mmio(KVM_TRACE_MMIO_READ, bytes,
			       vcpu->mmio_phys_addr, *(u64 *)val);
3358 3359 3360 3361
		vcpu->mmio_read_completed = 0;
		return X86EMUL_CONTINUE;
	}

3362
	gpa = kvm_mmu_gva_to_gpa_read(vcpu, addr, error_code);
3363

3364
	if (gpa == UNMAPPED_GVA)
3365
		return X86EMUL_PROPAGATE_FAULT;
3366 3367 3368 3369 3370

	/* For APIC access vmexit */
	if ((gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
		goto mmio;

3371
	if (kvm_read_guest_virt(addr, val, bytes, vcpu, NULL)
3372
				== X86EMUL_CONTINUE)
3373 3374 3375 3376 3377 3378
		return X86EMUL_CONTINUE;

mmio:
	/*
	 * Is this MMIO handled locally?
	 */
A
Avi Kivity 已提交
3379 3380
	if (!vcpu_mmio_read(vcpu, gpa, bytes, val)) {
		trace_kvm_mmio(KVM_TRACE_MMIO_READ, bytes, gpa, *(u64 *)val);
3381 3382
		return X86EMUL_CONTINUE;
	}
A
Avi Kivity 已提交
3383 3384

	trace_kvm_mmio(KVM_TRACE_MMIO_READ_UNSATISFIED, bytes, gpa, 0);
3385 3386

	vcpu->mmio_needed = 1;
3387 3388 3389 3390
	vcpu->run->exit_reason = KVM_EXIT_MMIO;
	vcpu->run->mmio.phys_addr = vcpu->mmio_phys_addr = gpa;
	vcpu->run->mmio.len = vcpu->mmio_size = bytes;
	vcpu->run->mmio.is_write = vcpu->mmio_is_write = 0;
3391

3392
	return X86EMUL_IO_NEEDED;
3393 3394
}

3395
int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
3396
			  const void *val, int bytes)
3397 3398 3399 3400
{
	int ret;

	ret = kvm_write_guest(vcpu->kvm, gpa, val, bytes);
3401
	if (ret < 0)
3402
		return 0;
3403
	kvm_mmu_pte_write(vcpu, gpa, val, bytes, 1);
3404 3405 3406 3407 3408 3409
	return 1;
}

static int emulator_write_emulated_onepage(unsigned long addr,
					   const void *val,
					   unsigned int bytes,
3410
					   unsigned int *error_code,
3411 3412
					   struct kvm_vcpu *vcpu)
{
3413 3414
	gpa_t                 gpa;

3415
	gpa = kvm_mmu_gva_to_gpa_write(vcpu, addr, error_code);
3416

3417
	if (gpa == UNMAPPED_GVA)
3418 3419 3420 3421 3422 3423 3424 3425 3426 3427
		return X86EMUL_PROPAGATE_FAULT;

	/* For APIC access vmexit */
	if ((gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
		goto mmio;

	if (emulator_write_phys(vcpu, gpa, val, bytes))
		return X86EMUL_CONTINUE;

mmio:
A
Avi Kivity 已提交
3428
	trace_kvm_mmio(KVM_TRACE_MMIO_WRITE, bytes, gpa, *(u64 *)val);
3429 3430 3431
	/*
	 * Is this MMIO handled locally?
	 */
3432
	if (!vcpu_mmio_write(vcpu, gpa, bytes, val))
3433 3434 3435
		return X86EMUL_CONTINUE;

	vcpu->mmio_needed = 1;
3436 3437 3438 3439 3440
	vcpu->run->exit_reason = KVM_EXIT_MMIO;
	vcpu->run->mmio.phys_addr = vcpu->mmio_phys_addr = gpa;
	vcpu->run->mmio.len = vcpu->mmio_size = bytes;
	vcpu->run->mmio.is_write = vcpu->mmio_is_write = 1;
	memcpy(vcpu->run->mmio.data, val, bytes);
3441 3442 3443 3444 3445

	return X86EMUL_CONTINUE;
}

int emulator_write_emulated(unsigned long addr,
3446 3447
			    const void *val,
			    unsigned int bytes,
3448
			    unsigned int *error_code,
3449
			    struct kvm_vcpu *vcpu)
3450 3451 3452 3453 3454 3455
{
	/* Crossing a page boundary? */
	if (((addr + bytes - 1) ^ addr) & PAGE_MASK) {
		int rc, now;

		now = -addr & ~PAGE_MASK;
3456 3457
		rc = emulator_write_emulated_onepage(addr, val, now, error_code,
						     vcpu);
3458 3459 3460 3461 3462 3463
		if (rc != X86EMUL_CONTINUE)
			return rc;
		addr += now;
		val += now;
		bytes -= now;
	}
3464 3465
	return emulator_write_emulated_onepage(addr, val, bytes, error_code,
					       vcpu);
3466 3467
}

3468 3469 3470 3471 3472 3473 3474
#define CMPXCHG_TYPE(t, ptr, old, new) \
	(cmpxchg((t *)(ptr), *(t *)(old), *(t *)(new)) == *(t *)(old))

#ifdef CONFIG_X86_64
#  define CMPXCHG64(ptr, old, new) CMPXCHG_TYPE(u64, ptr, old, new)
#else
#  define CMPXCHG64(ptr, old, new) \
3475
	(cmpxchg64((u64 *)(ptr), *(u64 *)(old), *(u64 *)(new)) == *(u64 *)(old))
3476 3477
#endif

3478 3479 3480 3481
static int emulator_cmpxchg_emulated(unsigned long addr,
				     const void *old,
				     const void *new,
				     unsigned int bytes,
3482
				     unsigned int *error_code,
3483 3484
				     struct kvm_vcpu *vcpu)
{
3485 3486 3487 3488
	gpa_t gpa;
	struct page *page;
	char *kaddr;
	bool exchanged;
3489

3490 3491 3492
	/* guests cmpxchg8b have to be emulated atomically */
	if (bytes > 8 || (bytes & (bytes - 1)))
		goto emul_write;
3493

3494
	gpa = kvm_mmu_gva_to_gpa_write(vcpu, addr, NULL);
3495

3496 3497 3498
	if (gpa == UNMAPPED_GVA ||
	    (gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
		goto emul_write;
3499

3500 3501
	if (((gpa + bytes - 1) & PAGE_MASK) != (gpa & PAGE_MASK))
		goto emul_write;
3502

3503
	page = gfn_to_page(vcpu->kvm, gpa >> PAGE_SHIFT);
3504

3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521
	kaddr = kmap_atomic(page, KM_USER0);
	kaddr += offset_in_page(gpa);
	switch (bytes) {
	case 1:
		exchanged = CMPXCHG_TYPE(u8, kaddr, old, new);
		break;
	case 2:
		exchanged = CMPXCHG_TYPE(u16, kaddr, old, new);
		break;
	case 4:
		exchanged = CMPXCHG_TYPE(u32, kaddr, old, new);
		break;
	case 8:
		exchanged = CMPXCHG64(kaddr, old, new);
		break;
	default:
		BUG();
3522
	}
3523 3524 3525 3526 3527 3528
	kunmap_atomic(kaddr, KM_USER0);
	kvm_release_page_dirty(page);

	if (!exchanged)
		return X86EMUL_CMPXCHG_FAILED;

3529 3530 3531
	kvm_mmu_pte_write(vcpu, gpa, new, bytes, 1);

	return X86EMUL_CONTINUE;
3532

3533
emul_write:
3534
	printk_once(KERN_WARNING "kvm: emulating exchange as write\n");
3535

3536
	return emulator_write_emulated(addr, new, bytes, error_code, vcpu);
3537 3538
}

3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557
static int kernel_pio(struct kvm_vcpu *vcpu, void *pd)
{
	/* TODO: String I/O for in kernel device */
	int r;

	if (vcpu->arch.pio.in)
		r = kvm_io_bus_read(vcpu->kvm, KVM_PIO_BUS, vcpu->arch.pio.port,
				    vcpu->arch.pio.size, pd);
	else
		r = kvm_io_bus_write(vcpu->kvm, KVM_PIO_BUS,
				     vcpu->arch.pio.port, vcpu->arch.pio.size,
				     pd);
	return r;
}


static int emulator_pio_in_emulated(int size, unsigned short port, void *val,
			     unsigned int count, struct kvm_vcpu *vcpu)
{
3558
	if (vcpu->arch.pio.count)
3559 3560 3561 3562 3563 3564
		goto data_avail;

	trace_kvm_pio(1, port, size, 1);

	vcpu->arch.pio.port = port;
	vcpu->arch.pio.in = 1;
3565
	vcpu->arch.pio.count  = count;
3566 3567 3568 3569 3570
	vcpu->arch.pio.size = size;

	if (!kernel_pio(vcpu, vcpu->arch.pio_data)) {
	data_avail:
		memcpy(val, vcpu->arch.pio_data, size * count);
3571
		vcpu->arch.pio.count = 0;
3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592
		return 1;
	}

	vcpu->run->exit_reason = KVM_EXIT_IO;
	vcpu->run->io.direction = KVM_EXIT_IO_IN;
	vcpu->run->io.size = size;
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
	vcpu->run->io.count = count;
	vcpu->run->io.port = port;

	return 0;
}

static int emulator_pio_out_emulated(int size, unsigned short port,
			      const void *val, unsigned int count,
			      struct kvm_vcpu *vcpu)
{
	trace_kvm_pio(0, port, size, 1);

	vcpu->arch.pio.port = port;
	vcpu->arch.pio.in = 0;
3593
	vcpu->arch.pio.count = count;
3594 3595 3596 3597 3598
	vcpu->arch.pio.size = size;

	memcpy(vcpu->arch.pio_data, val, size * count);

	if (!kernel_pio(vcpu, vcpu->arch.pio_data)) {
3599
		vcpu->arch.pio.count = 0;
3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612
		return 1;
	}

	vcpu->run->exit_reason = KVM_EXIT_IO;
	vcpu->run->io.direction = KVM_EXIT_IO_OUT;
	vcpu->run->io.size = size;
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
	vcpu->run->io.count = count;
	vcpu->run->io.port = port;

	return 0;
}

3613 3614 3615 3616 3617 3618 3619
static unsigned long get_segment_base(struct kvm_vcpu *vcpu, int seg)
{
	return kvm_x86_ops->get_segment_base(vcpu, seg);
}

int emulate_invlpg(struct kvm_vcpu *vcpu, gva_t address)
{
M
Marcelo Tosatti 已提交
3620
	kvm_mmu_invlpg(vcpu, address);
3621 3622 3623 3624 3625
	return X86EMUL_CONTINUE;
}

int emulate_clts(struct kvm_vcpu *vcpu)
{
3626
	kvm_x86_ops->set_cr0(vcpu, kvm_read_cr0_bits(vcpu, ~X86_CR0_TS));
A
Avi Kivity 已提交
3627
	kvm_x86_ops->fpu_activate(vcpu);
3628 3629 3630
	return X86EMUL_CONTINUE;
}

3631
int emulator_get_dr(int dr, unsigned long *dest, struct kvm_vcpu *vcpu)
3632
{
3633
	return _kvm_get_dr(vcpu, dr, dest);
3634 3635
}

3636
int emulator_set_dr(int dr, unsigned long value, struct kvm_vcpu *vcpu)
3637
{
3638 3639

	return __kvm_set_dr(vcpu, dr, value);
3640 3641 3642 3643 3644
}

void kvm_report_emulation_failure(struct kvm_vcpu *vcpu, const char *context)
{
	u8 opcodes[4];
3645
	unsigned long rip = kvm_rip_read(vcpu);
3646 3647
	unsigned long rip_linear;

3648
	if (!printk_ratelimit())
3649 3650
		return;

3651 3652
	rip_linear = rip + get_segment_base(vcpu, VCPU_SREG_CS);

3653
	kvm_read_guest_virt(rip_linear, (void *)opcodes, 4, vcpu, NULL);
3654 3655 3656 3657 3658 3659

	printk(KERN_ERR "emulation failed (%s) rip %lx %02x %02x %02x %02x\n",
	       context, rip, opcodes[0], opcodes[1], opcodes[2], opcodes[3]);
}
EXPORT_SYMBOL_GPL(kvm_report_emulation_failure);

3660
static u64 mk_cr_64(u64 curr_cr, u32 new_val)
3661
{
3662
	return (curr_cr & ~((1ULL << 32) - 1)) | new_val;
3663 3664
}

3665
static unsigned long emulator_get_cr(int cr, struct kvm_vcpu *vcpu)
3666
{
3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692
	unsigned long value;

	switch (cr) {
	case 0:
		value = kvm_read_cr0(vcpu);
		break;
	case 2:
		value = vcpu->arch.cr2;
		break;
	case 3:
		value = vcpu->arch.cr3;
		break;
	case 4:
		value = kvm_read_cr4(vcpu);
		break;
	case 8:
		value = kvm_get_cr8(vcpu);
		break;
	default:
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
		return 0;
	}

	return value;
}

3693
static int emulator_set_cr(int cr, unsigned long val, struct kvm_vcpu *vcpu)
3694
{
3695 3696
	int res = 0;

3697 3698
	switch (cr) {
	case 0:
3699
		res = __kvm_set_cr0(vcpu, mk_cr_64(kvm_read_cr0(vcpu), val));
3700 3701 3702 3703 3704
		break;
	case 2:
		vcpu->arch.cr2 = val;
		break;
	case 3:
3705
		res = __kvm_set_cr3(vcpu, val);
3706 3707
		break;
	case 4:
3708
		res = __kvm_set_cr4(vcpu, mk_cr_64(kvm_read_cr4(vcpu), val));
3709 3710
		break;
	case 8:
3711
		res = __kvm_set_cr8(vcpu, val & 0xfUL);
3712 3713 3714
		break;
	default:
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
3715
		res = -1;
3716
	}
3717 3718

	return res;
3719 3720
}

3721 3722 3723 3724 3725
static int emulator_get_cpl(struct kvm_vcpu *vcpu)
{
	return kvm_x86_ops->get_cpl(vcpu);
}

3726 3727 3728 3729 3730
static void emulator_get_gdt(struct desc_ptr *dt, struct kvm_vcpu *vcpu)
{
	kvm_x86_ops->get_gdt(vcpu, dt);
}

3731 3732 3733 3734 3735 3736
static unsigned long emulator_get_cached_segment_base(int seg,
						      struct kvm_vcpu *vcpu)
{
	return get_segment_base(vcpu, seg);
}

3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808
static bool emulator_get_cached_descriptor(struct desc_struct *desc, int seg,
					   struct kvm_vcpu *vcpu)
{
	struct kvm_segment var;

	kvm_get_segment(vcpu, &var, seg);

	if (var.unusable)
		return false;

	if (var.g)
		var.limit >>= 12;
	set_desc_limit(desc, var.limit);
	set_desc_base(desc, (unsigned long)var.base);
	desc->type = var.type;
	desc->s = var.s;
	desc->dpl = var.dpl;
	desc->p = var.present;
	desc->avl = var.avl;
	desc->l = var.l;
	desc->d = var.db;
	desc->g = var.g;

	return true;
}

static void emulator_set_cached_descriptor(struct desc_struct *desc, int seg,
					   struct kvm_vcpu *vcpu)
{
	struct kvm_segment var;

	/* needed to preserve selector */
	kvm_get_segment(vcpu, &var, seg);

	var.base = get_desc_base(desc);
	var.limit = get_desc_limit(desc);
	if (desc->g)
		var.limit = (var.limit << 12) | 0xfff;
	var.type = desc->type;
	var.present = desc->p;
	var.dpl = desc->dpl;
	var.db = desc->d;
	var.s = desc->s;
	var.l = desc->l;
	var.g = desc->g;
	var.avl = desc->avl;
	var.present = desc->p;
	var.unusable = !var.present;
	var.padding = 0;

	kvm_set_segment(vcpu, &var, seg);
	return;
}

static u16 emulator_get_segment_selector(int seg, struct kvm_vcpu *vcpu)
{
	struct kvm_segment kvm_seg;

	kvm_get_segment(vcpu, &kvm_seg, seg);
	return kvm_seg.selector;
}

static void emulator_set_segment_selector(u16 sel, int seg,
					  struct kvm_vcpu *vcpu)
{
	struct kvm_segment kvm_seg;

	kvm_get_segment(vcpu, &kvm_seg, seg);
	kvm_seg.selector = sel;
	kvm_set_segment(vcpu, &kvm_seg, seg);
}

3809
static struct x86_emulate_ops emulate_ops = {
3810
	.read_std            = kvm_read_guest_virt_system,
3811
	.write_std           = kvm_write_guest_virt_system,
3812
	.fetch               = kvm_fetch_guest_virt,
3813 3814 3815
	.read_emulated       = emulator_read_emulated,
	.write_emulated      = emulator_write_emulated,
	.cmpxchg_emulated    = emulator_cmpxchg_emulated,
3816 3817
	.pio_in_emulated     = emulator_pio_in_emulated,
	.pio_out_emulated    = emulator_pio_out_emulated,
3818 3819 3820 3821
	.get_cached_descriptor = emulator_get_cached_descriptor,
	.set_cached_descriptor = emulator_set_cached_descriptor,
	.get_segment_selector = emulator_get_segment_selector,
	.set_segment_selector = emulator_set_segment_selector,
3822
	.get_cached_segment_base = emulator_get_cached_segment_base,
3823
	.get_gdt             = emulator_get_gdt,
3824 3825
	.get_cr              = emulator_get_cr,
	.set_cr              = emulator_set_cr,
3826
	.cpl                 = emulator_get_cpl,
3827 3828
	.get_dr              = emulator_get_dr,
	.set_dr              = emulator_set_dr,
3829 3830
	.set_msr             = kvm_set_msr,
	.get_msr             = kvm_get_msr,
3831 3832
};

3833 3834 3835 3836 3837 3838 3839 3840
static void cache_all_regs(struct kvm_vcpu *vcpu)
{
	kvm_register_read(vcpu, VCPU_REGS_RAX);
	kvm_register_read(vcpu, VCPU_REGS_RSP);
	kvm_register_read(vcpu, VCPU_REGS_RIP);
	vcpu->arch.regs_dirty = ~0;
}

3841 3842 3843
int emulate_instruction(struct kvm_vcpu *vcpu,
			unsigned long cr2,
			u16 error_code,
3844
			int emulation_type)
3845
{
3846
	int r, shadow_mask;
3847
	struct decode_cache *c;
3848

3849
	kvm_clear_exception_queue(vcpu);
3850
	vcpu->arch.mmio_fault_cr2 = cr2;
3851
	/*
3852
	 * TODO: fix emulate.c to use guest_read/write_register
3853 3854 3855 3856 3857
	 * instead of direct ->regs accesses, can save hundred cycles
	 * on Intel for instructions that don't read/change RSP, for
	 * for example.
	 */
	cache_all_regs(vcpu);
3858

3859
	if (!(emulation_type & EMULTYPE_NO_DECODE)) {
3860 3861 3862
		int cs_db, cs_l;
		kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);

3863
		vcpu->arch.emulate_ctxt.vcpu = vcpu;
3864
		vcpu->arch.emulate_ctxt.eflags = kvm_x86_ops->get_rflags(vcpu);
3865
		vcpu->arch.emulate_ctxt.eip = kvm_rip_read(vcpu);
3866
		vcpu->arch.emulate_ctxt.mode =
3867
			(!is_protmode(vcpu)) ? X86EMUL_MODE_REAL :
3868
			(vcpu->arch.emulate_ctxt.eflags & X86_EFLAGS_VM)
3869
			? X86EMUL_MODE_VM86 : cs_l
3870 3871 3872
			? X86EMUL_MODE_PROT64 :	cs_db
			? X86EMUL_MODE_PROT32 : X86EMUL_MODE_PROT16;

3873
		r = x86_decode_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
A
Avi Kivity 已提交
3874
		trace_kvm_emulate_insn_start(vcpu);
3875

3876 3877
		/* Only allow emulation of specific instructions on #UD
		 * (namely VMMCALL, sysenter, sysexit, syscall)*/
3878
		c = &vcpu->arch.emulate_ctxt.decode;
3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902
		if (emulation_type & EMULTYPE_TRAP_UD) {
			if (!c->twobyte)
				return EMULATE_FAIL;
			switch (c->b) {
			case 0x01: /* VMMCALL */
				if (c->modrm_mod != 3 || c->modrm_rm != 1)
					return EMULATE_FAIL;
				break;
			case 0x34: /* sysenter */
			case 0x35: /* sysexit */
				if (c->modrm_mod != 0 || c->modrm_rm != 0)
					return EMULATE_FAIL;
				break;
			case 0x05: /* syscall */
				if (c->modrm_mod != 0 || c->modrm_rm != 0)
					return EMULATE_FAIL;
				break;
			default:
				return EMULATE_FAIL;
			}

			if (!(c->modrm_reg == 0 || c->modrm_reg == 3))
				return EMULATE_FAIL;
		}
3903

3904
		++vcpu->stat.insn_emulation;
3905
		if (r)  {
3906
			++vcpu->stat.insn_emulation_fail;
A
Avi Kivity 已提交
3907
			trace_kvm_emulate_insn_failed(vcpu);
3908 3909 3910 3911 3912 3913
			if (kvm_mmu_unprotect_page_virt(vcpu, cr2))
				return EMULATE_DONE;
			return EMULATE_FAIL;
		}
	}

3914 3915 3916 3917 3918
	if (emulation_type & EMULTYPE_SKIP) {
		kvm_rip_write(vcpu, vcpu->arch.emulate_ctxt.decode.eip);
		return EMULATE_DONE;
	}

3919
restart:
3920
	r = x86_emulate_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
3921

3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933
	if (r) { /* emulation failed */
		/*
		 * if emulation was due to access to shadowed page table
		 * and it failed try to unshadow page and re-entetr the
		 * guest to let CPU execute the instruction.
		 */
		if (kvm_mmu_unprotect_page_virt(vcpu, cr2))
			return EMULATE_DONE;

		trace_kvm_emulate_insn_failed(vcpu);
		kvm_report_emulation_failure(vcpu, "mmio");
		return EMULATE_FAIL;
3934 3935
	}

3936 3937
	shadow_mask = vcpu->arch.emulate_ctxt.interruptibility;
	kvm_x86_ops->set_interrupt_shadow(vcpu, shadow_mask);
3938
	kvm_x86_ops->set_rflags(vcpu, vcpu->arch.emulate_ctxt.eflags);
3939
	kvm_rip_write(vcpu, vcpu->arch.emulate_ctxt.eip);
3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952

	if (vcpu->arch.pio.count) {
		if (!vcpu->arch.pio.in)
			vcpu->arch.pio.count = 0;
		return EMULATE_DO_MMIO;
	}

	if (vcpu->mmio_needed) {
		if (vcpu->mmio_is_write)
			vcpu->mmio_needed = 0;
		return EMULATE_DO_MMIO;
	}

3953 3954
	if (vcpu->arch.exception.pending)
		vcpu->arch.emulate_ctxt.restart = false;
3955

3956 3957
	if (vcpu->arch.emulate_ctxt.restart)
		goto restart;
3958

3959
	return EMULATE_DONE;
3960
}
3961
EXPORT_SYMBOL_GPL(emulate_instruction);
3962

3963
int kvm_fast_pio_out(struct kvm_vcpu *vcpu, int size, unsigned short port)
3964
{
3965 3966 3967
	unsigned long val = kvm_register_read(vcpu, VCPU_REGS_RAX);
	int ret = emulator_pio_out_emulated(size, port, &val, 1, vcpu);
	/* do not return to emulator after return from userspace */
3968
	vcpu->arch.pio.count = 0;
3969 3970
	return ret;
}
3971
EXPORT_SYMBOL_GPL(kvm_fast_pio_out);
3972

3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989
static void bounce_off(void *info)
{
	/* nothing */
}

static int kvmclock_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
				     void *data)
{
	struct cpufreq_freqs *freq = data;
	struct kvm *kvm;
	struct kvm_vcpu *vcpu;
	int i, send_ipi = 0;

	if (val == CPUFREQ_PRECHANGE && freq->old > freq->new)
		return 0;
	if (val == CPUFREQ_POSTCHANGE && freq->old < freq->new)
		return 0;
3990
	per_cpu(cpu_tsc_khz, freq->cpu) = freq->new;
3991 3992 3993

	spin_lock(&kvm_lock);
	list_for_each_entry(kvm, &vm_list, vm_list) {
3994
		kvm_for_each_vcpu(i, vcpu, kvm) {
3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026
			if (vcpu->cpu != freq->cpu)
				continue;
			if (!kvm_request_guest_time_update(vcpu))
				continue;
			if (vcpu->cpu != smp_processor_id())
				send_ipi++;
		}
	}
	spin_unlock(&kvm_lock);

	if (freq->old < freq->new && send_ipi) {
		/*
		 * We upscale the frequency.  Must make the guest
		 * doesn't see old kvmclock values while running with
		 * the new frequency, otherwise we risk the guest sees
		 * time go backwards.
		 *
		 * In case we update the frequency for another cpu
		 * (which might be in guest context) send an interrupt
		 * to kick the cpu out of guest context.  Next time
		 * guest context is entered kvmclock will be updated,
		 * so the guest will not see stale values.
		 */
		smp_call_function_single(freq->cpu, bounce_off, NULL, 1);
	}
	return 0;
}

static struct notifier_block kvmclock_cpufreq_notifier_block = {
        .notifier_call  = kvmclock_cpufreq_notifier
};

4027 4028 4029 4030 4031 4032 4033
static void kvm_timer_init(void)
{
	int cpu;

	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) {
		cpufreq_register_notifier(&kvmclock_cpufreq_notifier_block,
					  CPUFREQ_TRANSITION_NOTIFIER);
Z
Zachary Amsden 已提交
4034 4035 4036 4037 4038 4039
		for_each_online_cpu(cpu) {
			unsigned long khz = cpufreq_get(cpu);
			if (!khz)
				khz = tsc_khz;
			per_cpu(cpu_tsc_khz, cpu) = khz;
		}
4040 4041 4042
	} else {
		for_each_possible_cpu(cpu)
			per_cpu(cpu_tsc_khz, cpu) = tsc_khz;
4043 4044 4045
	}
}

4046 4047 4048 4049 4050 4051 4052 4053 4054 4055
static DEFINE_PER_CPU(struct kvm_vcpu *, current_vcpu);

static int kvm_is_in_guest(void)
{
	return percpu_read(current_vcpu) != NULL;
}

static int kvm_is_user_mode(void)
{
	int user_mode = 3;
4056

4057 4058
	if (percpu_read(current_vcpu))
		user_mode = kvm_x86_ops->get_cpl(percpu_read(current_vcpu));
4059

4060 4061 4062 4063 4064 4065
	return user_mode != 0;
}

static unsigned long kvm_get_guest_ip(void)
{
	unsigned long ip = 0;
4066

4067 4068
	if (percpu_read(current_vcpu))
		ip = kvm_rip_read(percpu_read(current_vcpu));
4069

4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090
	return ip;
}

static struct perf_guest_info_callbacks kvm_guest_cbs = {
	.is_in_guest		= kvm_is_in_guest,
	.is_user_mode		= kvm_is_user_mode,
	.get_guest_ip		= kvm_get_guest_ip,
};

void kvm_before_handle_nmi(struct kvm_vcpu *vcpu)
{
	percpu_write(current_vcpu, vcpu);
}
EXPORT_SYMBOL_GPL(kvm_before_handle_nmi);

void kvm_after_handle_nmi(struct kvm_vcpu *vcpu)
{
	percpu_write(current_vcpu, NULL);
}
EXPORT_SYMBOL_GPL(kvm_after_handle_nmi);

4091
int kvm_arch_init(void *opaque)
4092
{
4093
	int r;
4094 4095 4096 4097
	struct kvm_x86_ops *ops = (struct kvm_x86_ops *)opaque;

	if (kvm_x86_ops) {
		printk(KERN_ERR "kvm: already loaded the other module\n");
4098 4099
		r = -EEXIST;
		goto out;
4100 4101 4102 4103
	}

	if (!ops->cpu_has_kvm_support()) {
		printk(KERN_ERR "kvm: no hardware support\n");
4104 4105
		r = -EOPNOTSUPP;
		goto out;
4106 4107 4108
	}
	if (ops->disabled_by_bios()) {
		printk(KERN_ERR "kvm: disabled by bios\n");
4109 4110
		r = -EOPNOTSUPP;
		goto out;
4111 4112
	}

4113 4114 4115 4116 4117 4118
	r = kvm_mmu_module_init();
	if (r)
		goto out;

	kvm_init_msr_list();

4119
	kvm_x86_ops = ops;
4120
	kvm_mmu_set_nonpresent_ptes(0ull, 0ull);
S
Sheng Yang 已提交
4121 4122
	kvm_mmu_set_base_ptes(PT_PRESENT_MASK);
	kvm_mmu_set_mask_ptes(PT_USER_MASK, PT_ACCESSED_MASK,
4123
			PT_DIRTY_MASK, PT64_NX_MASK, 0);
4124

4125
	kvm_timer_init();
4126

4127 4128
	perf_register_guest_info_callbacks(&kvm_guest_cbs);

4129
	return 0;
4130 4131 4132

out:
	return r;
4133
}
4134

4135 4136
void kvm_arch_exit(void)
{
4137 4138
	perf_unregister_guest_info_callbacks(&kvm_guest_cbs);

4139 4140 4141
	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
		cpufreq_unregister_notifier(&kvmclock_cpufreq_notifier_block,
					    CPUFREQ_TRANSITION_NOTIFIER);
4142
	kvm_x86_ops = NULL;
4143 4144
	kvm_mmu_module_exit();
}
4145

4146 4147 4148 4149
int kvm_emulate_halt(struct kvm_vcpu *vcpu)
{
	++vcpu->stat.halt_exits;
	if (irqchip_in_kernel(vcpu->kvm)) {
4150
		vcpu->arch.mp_state = KVM_MP_STATE_HALTED;
4151 4152 4153 4154 4155 4156 4157 4158
		return 1;
	} else {
		vcpu->run->exit_reason = KVM_EXIT_HLT;
		return 0;
	}
}
EXPORT_SYMBOL_GPL(kvm_emulate_halt);

4159 4160 4161 4162 4163 4164 4165 4166 4167
static inline gpa_t hc_gpa(struct kvm_vcpu *vcpu, unsigned long a0,
			   unsigned long a1)
{
	if (is_long_mode(vcpu))
		return a0;
	else
		return a0 | ((gpa_t)a1 << 32);
}

4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178
int kvm_hv_hypercall(struct kvm_vcpu *vcpu)
{
	u64 param, ingpa, outgpa, ret;
	uint16_t code, rep_idx, rep_cnt, res = HV_STATUS_SUCCESS, rep_done = 0;
	bool fast, longmode;
	int cs_db, cs_l;

	/*
	 * hypercall generates UD from non zero cpl and real mode
	 * per HYPER-V spec
	 */
4179
	if (kvm_x86_ops->get_cpl(vcpu) != 0 || !is_protmode(vcpu)) {
4180 4181 4182 4183 4184 4185 4186 4187
		kvm_queue_exception(vcpu, UD_VECTOR);
		return 0;
	}

	kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
	longmode = is_long_mode(vcpu) && cs_l == 1;

	if (!longmode) {
4188 4189 4190 4191 4192 4193
		param = ((u64)kvm_register_read(vcpu, VCPU_REGS_RDX) << 32) |
			(kvm_register_read(vcpu, VCPU_REGS_RAX) & 0xffffffff);
		ingpa = ((u64)kvm_register_read(vcpu, VCPU_REGS_RBX) << 32) |
			(kvm_register_read(vcpu, VCPU_REGS_RCX) & 0xffffffff);
		outgpa = ((u64)kvm_register_read(vcpu, VCPU_REGS_RDI) << 32) |
			(kvm_register_read(vcpu, VCPU_REGS_RSI) & 0xffffffff);
4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209
	}
#ifdef CONFIG_X86_64
	else {
		param = kvm_register_read(vcpu, VCPU_REGS_RCX);
		ingpa = kvm_register_read(vcpu, VCPU_REGS_RDX);
		outgpa = kvm_register_read(vcpu, VCPU_REGS_R8);
	}
#endif

	code = param & 0xffff;
	fast = (param >> 16) & 0x1;
	rep_cnt = (param >> 32) & 0xfff;
	rep_idx = (param >> 48) & 0xfff;

	trace_kvm_hv_hypercall(code, fast, rep_cnt, rep_idx, ingpa, outgpa);

4210 4211 4212 4213 4214 4215 4216 4217
	switch (code) {
	case HV_X64_HV_NOTIFY_LONG_SPIN_WAIT:
		kvm_vcpu_on_spin(vcpu);
		break;
	default:
		res = HV_STATUS_INVALID_HYPERCALL_CODE;
		break;
	}
4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229

	ret = res | (((u64)rep_done & 0xfff) << 32);
	if (longmode) {
		kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
	} else {
		kvm_register_write(vcpu, VCPU_REGS_RDX, ret >> 32);
		kvm_register_write(vcpu, VCPU_REGS_RAX, ret & 0xffffffff);
	}

	return 1;
}

4230 4231 4232
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
{
	unsigned long nr, a0, a1, a2, a3, ret;
4233
	int r = 1;
4234

4235 4236 4237
	if (kvm_hv_hypercall_enabled(vcpu->kvm))
		return kvm_hv_hypercall(vcpu);

4238 4239 4240 4241 4242
	nr = kvm_register_read(vcpu, VCPU_REGS_RAX);
	a0 = kvm_register_read(vcpu, VCPU_REGS_RBX);
	a1 = kvm_register_read(vcpu, VCPU_REGS_RCX);
	a2 = kvm_register_read(vcpu, VCPU_REGS_RDX);
	a3 = kvm_register_read(vcpu, VCPU_REGS_RSI);
4243

4244
	trace_kvm_hypercall(nr, a0, a1, a2, a3);
F
Feng (Eric) Liu 已提交
4245

4246 4247 4248 4249 4250 4251 4252 4253
	if (!is_long_mode(vcpu)) {
		nr &= 0xFFFFFFFF;
		a0 &= 0xFFFFFFFF;
		a1 &= 0xFFFFFFFF;
		a2 &= 0xFFFFFFFF;
		a3 &= 0xFFFFFFFF;
	}

4254 4255 4256 4257 4258
	if (kvm_x86_ops->get_cpl(vcpu) != 0) {
		ret = -KVM_EPERM;
		goto out;
	}

4259
	switch (nr) {
A
Avi Kivity 已提交
4260 4261 4262
	case KVM_HC_VAPIC_POLL_IRQ:
		ret = 0;
		break;
4263 4264 4265
	case KVM_HC_MMU_OP:
		r = kvm_pv_mmu_op(vcpu, a0, hc_gpa(vcpu, a1, a2), &ret);
		break;
4266 4267 4268 4269
	default:
		ret = -KVM_ENOSYS;
		break;
	}
4270
out:
4271
	kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
A
Amit Shah 已提交
4272
	++vcpu->stat.hypercalls;
4273
	return r;
4274 4275 4276 4277 4278 4279
}
EXPORT_SYMBOL_GPL(kvm_emulate_hypercall);

int kvm_fix_hypercall(struct kvm_vcpu *vcpu)
{
	char instruction[3];
4280
	unsigned long rip = kvm_rip_read(vcpu);
4281 4282 4283 4284 4285 4286 4287 4288 4289 4290

	/*
	 * Blow out the MMU to ensure that no other VCPU has an active mapping
	 * to ensure that the updated hypercall appears atomically across all
	 * VCPUs.
	 */
	kvm_mmu_zap_all(vcpu->kvm);

	kvm_x86_ops->patch_hypercall(vcpu, instruction);

4291
	return emulator_write_emulated(rip, instruction, 3, NULL, vcpu);
4292 4293 4294 4295
}

void realmode_lgdt(struct kvm_vcpu *vcpu, u16 limit, unsigned long base)
{
4296
	struct desc_ptr dt = { limit, base };
4297 4298 4299 4300 4301 4302

	kvm_x86_ops->set_gdt(vcpu, &dt);
}

void realmode_lidt(struct kvm_vcpu *vcpu, u16 limit, unsigned long base)
{
4303
	struct desc_ptr dt = { limit, base };
4304 4305 4306 4307

	kvm_x86_ops->set_idt(vcpu, &dt);
}

4308 4309
static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
{
4310 4311
	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	int j, nent = vcpu->arch.cpuid_nent;
4312 4313 4314

	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	/* when no next entry is found, the current entry[i] is reselected */
4315
	for (j = i + 1; ; j = (j + 1) % nent) {
4316
		struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334
		if (ej->function == e->function) {
			ej->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
			return j;
		}
	}
	return 0; /* silence gcc, even though control never reaches here */
}

/* find an entry with matching function, matching index (if needed), and that
 * should be read next (if it's stateful) */
static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	u32 function, u32 index)
{
	if (e->function != function)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
4335
	    !(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
4336 4337 4338 4339
		return 0;
	return 1;
}

4340 4341
struct kvm_cpuid_entry2 *kvm_find_cpuid_entry(struct kvm_vcpu *vcpu,
					      u32 function, u32 index)
4342 4343
{
	int i;
4344
	struct kvm_cpuid_entry2 *best = NULL;
4345

4346
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
4347 4348
		struct kvm_cpuid_entry2 *e;

4349
		e = &vcpu->arch.cpuid_entries[i];
4350 4351 4352
		if (is_matching_cpuid_entry(e, function, index)) {
			if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
				move_to_next_stateful_cpuid_entry(vcpu, i);
4353 4354 4355 4356 4357 4358 4359 4360 4361 4362
			best = e;
			break;
		}
		/*
		 * Both basic or both extended?
		 */
		if (((e->function ^ function) & 0x80000000) == 0)
			if (!best || e->function > best->function)
				best = e;
	}
4363 4364
	return best;
}
4365
EXPORT_SYMBOL_GPL(kvm_find_cpuid_entry);
4366

4367 4368 4369 4370
int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
{
	struct kvm_cpuid_entry2 *best;

4371 4372 4373
	best = kvm_find_cpuid_entry(vcpu, 0x80000000, 0);
	if (!best || best->eax < 0x80000008)
		goto not_found;
4374 4375 4376
	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
	if (best)
		return best->eax & 0xff;
4377
not_found:
4378 4379 4380
	return 36;
}

4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392
void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
{
	u32 function, index;
	struct kvm_cpuid_entry2 *best;

	function = kvm_register_read(vcpu, VCPU_REGS_RAX);
	index = kvm_register_read(vcpu, VCPU_REGS_RCX);
	kvm_register_write(vcpu, VCPU_REGS_RAX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RBX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RCX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RDX, 0);
	best = kvm_find_cpuid_entry(vcpu, function, index);
4393
	if (best) {
4394 4395 4396 4397
		kvm_register_write(vcpu, VCPU_REGS_RAX, best->eax);
		kvm_register_write(vcpu, VCPU_REGS_RBX, best->ebx);
		kvm_register_write(vcpu, VCPU_REGS_RCX, best->ecx);
		kvm_register_write(vcpu, VCPU_REGS_RDX, best->edx);
4398 4399
	}
	kvm_x86_ops->skip_emulated_instruction(vcpu);
4400 4401 4402 4403 4404
	trace_kvm_cpuid(function,
			kvm_register_read(vcpu, VCPU_REGS_RAX),
			kvm_register_read(vcpu, VCPU_REGS_RBX),
			kvm_register_read(vcpu, VCPU_REGS_RCX),
			kvm_register_read(vcpu, VCPU_REGS_RDX));
4405 4406
}
EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);
4407

4408 4409 4410 4411 4412 4413
/*
 * Check if userspace requested an interrupt window, and that the
 * interrupt window is open.
 *
 * No need to exit to userspace if we already have an interrupt queued.
 */
A
Avi Kivity 已提交
4414
static int dm_request_for_irq_injection(struct kvm_vcpu *vcpu)
4415
{
4416
	return (!irqchip_in_kernel(vcpu->kvm) && !kvm_cpu_has_interrupt(vcpu) &&
A
Avi Kivity 已提交
4417
		vcpu->run->request_interrupt_window &&
4418
		kvm_arch_interrupt_allowed(vcpu));
4419 4420
}

A
Avi Kivity 已提交
4421
static void post_kvm_run_save(struct kvm_vcpu *vcpu)
4422
{
A
Avi Kivity 已提交
4423 4424
	struct kvm_run *kvm_run = vcpu->run;

4425
	kvm_run->if_flag = (kvm_get_rflags(vcpu) & X86_EFLAGS_IF) != 0;
4426
	kvm_run->cr8 = kvm_get_cr8(vcpu);
4427
	kvm_run->apic_base = kvm_get_apic_base(vcpu);
4428
	if (irqchip_in_kernel(vcpu->kvm))
4429
		kvm_run->ready_for_interrupt_injection = 1;
4430
	else
4431
		kvm_run->ready_for_interrupt_injection =
4432 4433 4434
			kvm_arch_interrupt_allowed(vcpu) &&
			!kvm_cpu_has_interrupt(vcpu) &&
			!kvm_event_needs_reinjection(vcpu);
4435 4436
}

A
Avi Kivity 已提交
4437 4438 4439 4440 4441 4442 4443 4444 4445
static void vapic_enter(struct kvm_vcpu *vcpu)
{
	struct kvm_lapic *apic = vcpu->arch.apic;
	struct page *page;

	if (!apic || !apic->vapic_addr)
		return;

	page = gfn_to_page(vcpu->kvm, apic->vapic_addr >> PAGE_SHIFT);
4446 4447

	vcpu->arch.apic->vapic_page = page;
A
Avi Kivity 已提交
4448 4449 4450 4451 4452
}

static void vapic_exit(struct kvm_vcpu *vcpu)
{
	struct kvm_lapic *apic = vcpu->arch.apic;
4453
	int idx;
A
Avi Kivity 已提交
4454 4455 4456 4457

	if (!apic || !apic->vapic_addr)
		return;

4458
	idx = srcu_read_lock(&vcpu->kvm->srcu);
A
Avi Kivity 已提交
4459 4460
	kvm_release_page_dirty(apic->vapic_page);
	mark_page_dirty(vcpu->kvm, apic->vapic_addr >> PAGE_SHIFT);
4461
	srcu_read_unlock(&vcpu->kvm->srcu, idx);
A
Avi Kivity 已提交
4462 4463
}

4464 4465 4466 4467 4468 4469 4470
static void update_cr8_intercept(struct kvm_vcpu *vcpu)
{
	int max_irr, tpr;

	if (!kvm_x86_ops->update_cr8_intercept)
		return;

4471 4472 4473
	if (!vcpu->arch.apic)
		return;

4474 4475 4476 4477
	if (!vcpu->arch.apic->vapic_addr)
		max_irr = kvm_lapic_find_highest_irr(vcpu);
	else
		max_irr = -1;
4478 4479 4480 4481 4482 4483 4484 4485 4486

	if (max_irr != -1)
		max_irr >>= 4;

	tpr = kvm_lapic_get_cr8(vcpu);

	kvm_x86_ops->update_cr8_intercept(vcpu, tpr, max_irr);
}

A
Avi Kivity 已提交
4487
static void inject_pending_event(struct kvm_vcpu *vcpu)
4488 4489
{
	/* try to reinject previous events if any */
4490
	if (vcpu->arch.exception.pending) {
A
Avi Kivity 已提交
4491 4492 4493
		trace_kvm_inj_exception(vcpu->arch.exception.nr,
					vcpu->arch.exception.has_error_code,
					vcpu->arch.exception.error_code);
4494 4495
		kvm_x86_ops->queue_exception(vcpu, vcpu->arch.exception.nr,
					  vcpu->arch.exception.has_error_code,
4496 4497
					  vcpu->arch.exception.error_code,
					  vcpu->arch.exception.reinject);
4498 4499 4500
		return;
	}

4501 4502 4503 4504 4505 4506
	if (vcpu->arch.nmi_injected) {
		kvm_x86_ops->set_nmi(vcpu);
		return;
	}

	if (vcpu->arch.interrupt.pending) {
4507
		kvm_x86_ops->set_irq(vcpu);
4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519
		return;
	}

	/* try to inject new event if pending */
	if (vcpu->arch.nmi_pending) {
		if (kvm_x86_ops->nmi_allowed(vcpu)) {
			vcpu->arch.nmi_pending = false;
			vcpu->arch.nmi_injected = true;
			kvm_x86_ops->set_nmi(vcpu);
		}
	} else if (kvm_cpu_has_interrupt(vcpu)) {
		if (kvm_x86_ops->interrupt_allowed(vcpu)) {
4520 4521 4522
			kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu),
					    false);
			kvm_x86_ops->set_irq(vcpu);
4523 4524 4525 4526
		}
	}
}

A
Avi Kivity 已提交
4527
static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
4528 4529
{
	int r;
4530
	bool req_int_win = !irqchip_in_kernel(vcpu->kvm) &&
A
Avi Kivity 已提交
4531
		vcpu->run->request_interrupt_window;
4532

4533 4534 4535 4536
	if (vcpu->requests)
		if (test_and_clear_bit(KVM_REQ_MMU_RELOAD, &vcpu->requests))
			kvm_mmu_unload(vcpu);

4537 4538 4539 4540
	r = kvm_mmu_reload(vcpu);
	if (unlikely(r))
		goto out;

4541 4542
	if (vcpu->requests) {
		if (test_and_clear_bit(KVM_REQ_MIGRATE_TIMER, &vcpu->requests))
M
Marcelo Tosatti 已提交
4543
			__kvm_migrate_timers(vcpu);
4544 4545
		if (test_and_clear_bit(KVM_REQ_KVMCLOCK_UPDATE, &vcpu->requests))
			kvm_write_guest_time(vcpu);
4546 4547
		if (test_and_clear_bit(KVM_REQ_MMU_SYNC, &vcpu->requests))
			kvm_mmu_sync_roots(vcpu);
4548 4549
		if (test_and_clear_bit(KVM_REQ_TLB_FLUSH, &vcpu->requests))
			kvm_x86_ops->tlb_flush(vcpu);
A
Avi Kivity 已提交
4550 4551
		if (test_and_clear_bit(KVM_REQ_REPORT_TPR_ACCESS,
				       &vcpu->requests)) {
A
Avi Kivity 已提交
4552
			vcpu->run->exit_reason = KVM_EXIT_TPR_ACCESS;
A
Avi Kivity 已提交
4553 4554 4555
			r = 0;
			goto out;
		}
J
Joerg Roedel 已提交
4556
		if (test_and_clear_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests)) {
A
Avi Kivity 已提交
4557
			vcpu->run->exit_reason = KVM_EXIT_SHUTDOWN;
J
Joerg Roedel 已提交
4558 4559 4560
			r = 0;
			goto out;
		}
4561 4562 4563 4564
		if (test_and_clear_bit(KVM_REQ_DEACTIVATE_FPU, &vcpu->requests)) {
			vcpu->fpu_active = 0;
			kvm_x86_ops->fpu_deactivate(vcpu);
		}
4565
	}
A
Avi Kivity 已提交
4566

4567 4568 4569
	preempt_disable();

	kvm_x86_ops->prepare_guest_switch(vcpu);
4570 4571
	if (vcpu->fpu_active)
		kvm_load_guest_fpu(vcpu);
4572 4573 4574

	local_irq_disable();

4575 4576 4577
	clear_bit(KVM_REQ_KICK, &vcpu->requests);
	smp_mb__after_clear_bit();

4578
	if (vcpu->requests || need_resched() || signal_pending(current)) {
4579
		set_bit(KVM_REQ_KICK, &vcpu->requests);
4580 4581 4582 4583 4584 4585
		local_irq_enable();
		preempt_enable();
		r = 1;
		goto out;
	}

A
Avi Kivity 已提交
4586
	inject_pending_event(vcpu);
4587

4588 4589 4590 4591 4592 4593
	/* enable NMI/IRQ window open exits if needed */
	if (vcpu->arch.nmi_pending)
		kvm_x86_ops->enable_nmi_window(vcpu);
	else if (kvm_cpu_has_interrupt(vcpu) || req_int_win)
		kvm_x86_ops->enable_irq_window(vcpu);

4594
	if (kvm_lapic_enabled(vcpu)) {
4595 4596
		update_cr8_intercept(vcpu);
		kvm_lapic_sync_to_vapic(vcpu);
4597
	}
A
Avi Kivity 已提交
4598

4599
	srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
4600

4601 4602
	kvm_guest_enter();

4603 4604 4605 4606 4607 4608 4609
	if (unlikely(vcpu->arch.switch_db_regs)) {
		set_debugreg(0, 7);
		set_debugreg(vcpu->arch.eff_db[0], 0);
		set_debugreg(vcpu->arch.eff_db[1], 1);
		set_debugreg(vcpu->arch.eff_db[2], 2);
		set_debugreg(vcpu->arch.eff_db[3], 3);
	}
4610

4611
	trace_kvm_entry(vcpu->vcpu_id);
A
Avi Kivity 已提交
4612
	kvm_x86_ops->run(vcpu);
4613

4614 4615 4616 4617 4618 4619 4620
	/*
	 * If the guest has used debug registers, at least dr7
	 * will be disabled while returning to the host.
	 * If we don't have active breakpoints in the host, we don't
	 * care about the messed up debug address registers. But if
	 * we have some of them active, restore the old state.
	 */
4621
	if (hw_breakpoint_active())
4622
		hw_breakpoint_restore();
4623

4624
	set_bit(KVM_REQ_KICK, &vcpu->requests);
4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640
	local_irq_enable();

	++vcpu->stat.exits;

	/*
	 * We must have an instruction between local_irq_enable() and
	 * kvm_guest_exit(), so the timer interrupt isn't delayed by
	 * the interrupt shadow.  The stat.exits increment will do nicely.
	 * But we need to prevent reordering, hence this barrier():
	 */
	barrier();

	kvm_guest_exit();

	preempt_enable();

4641
	vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
4642

4643 4644 4645 4646
	/*
	 * Profile KVM exit RIPs:
	 */
	if (unlikely(prof_on == KVM_PROFILING)) {
4647 4648
		unsigned long rip = kvm_rip_read(vcpu);
		profile_hit(KVM_PROFILING, (void *)rip);
4649 4650
	}

4651

A
Avi Kivity 已提交
4652 4653
	kvm_lapic_sync_from_vapic(vcpu);

A
Avi Kivity 已提交
4654
	r = kvm_x86_ops->handle_exit(vcpu);
4655 4656 4657
out:
	return r;
}
4658

4659

A
Avi Kivity 已提交
4660
static int __vcpu_run(struct kvm_vcpu *vcpu)
4661 4662
{
	int r;
4663
	struct kvm *kvm = vcpu->kvm;
4664 4665

	if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED)) {
4666 4667
		pr_debug("vcpu %d received sipi with vector # %x\n",
			 vcpu->vcpu_id, vcpu->arch.sipi_vector);
4668
		kvm_lapic_reset(vcpu);
4669
		r = kvm_arch_vcpu_reset(vcpu);
4670 4671 4672
		if (r)
			return r;
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
4673 4674
	}

4675
	vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
4676 4677 4678 4679
	vapic_enter(vcpu);

	r = 1;
	while (r > 0) {
4680
		if (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE)
A
Avi Kivity 已提交
4681
			r = vcpu_enter_guest(vcpu);
4682
		else {
4683
			srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
4684
			kvm_vcpu_block(vcpu);
4685
			vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
4686
			if (test_and_clear_bit(KVM_REQ_UNHALT, &vcpu->requests))
4687 4688 4689
			{
				switch(vcpu->arch.mp_state) {
				case KVM_MP_STATE_HALTED:
4690
					vcpu->arch.mp_state =
4691 4692 4693 4694 4695 4696 4697 4698 4699
						KVM_MP_STATE_RUNNABLE;
				case KVM_MP_STATE_RUNNABLE:
					break;
				case KVM_MP_STATE_SIPI_RECEIVED:
				default:
					r = -EINTR;
					break;
				}
			}
4700 4701
		}

4702 4703 4704 4705 4706 4707 4708
		if (r <= 0)
			break;

		clear_bit(KVM_REQ_PENDING_TIMER, &vcpu->requests);
		if (kvm_cpu_has_pending_timer(vcpu))
			kvm_inject_pending_timer_irqs(vcpu);

A
Avi Kivity 已提交
4709
		if (dm_request_for_irq_injection(vcpu)) {
4710
			r = -EINTR;
A
Avi Kivity 已提交
4711
			vcpu->run->exit_reason = KVM_EXIT_INTR;
4712 4713 4714 4715
			++vcpu->stat.request_irq_exits;
		}
		if (signal_pending(current)) {
			r = -EINTR;
A
Avi Kivity 已提交
4716
			vcpu->run->exit_reason = KVM_EXIT_INTR;
4717 4718 4719
			++vcpu->stat.signal_exits;
		}
		if (need_resched()) {
4720
			srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
4721
			kvm_resched(vcpu);
4722
			vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
4723
		}
4724 4725
	}

4726
	srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
4727

A
Avi Kivity 已提交
4728 4729
	vapic_exit(vcpu);

4730 4731 4732 4733 4734 4735 4736 4737 4738 4739
	return r;
}

int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	int r;
	sigset_t sigsaved;

	vcpu_load(vcpu);

4740 4741 4742
	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &vcpu->sigset, &sigsaved);

4743
	if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
4744
		kvm_vcpu_block(vcpu);
4745
		clear_bit(KVM_REQ_UNHALT, &vcpu->requests);
4746 4747
		r = -EAGAIN;
		goto out;
4748 4749 4750 4751
	}

	/* re-sync apic's tpr */
	if (!irqchip_in_kernel(vcpu->kvm))
4752
		kvm_set_cr8(vcpu, kvm_run->cr8);
4753

4754 4755 4756 4757 4758 4759
	if (vcpu->arch.pio.count || vcpu->mmio_needed ||
	    vcpu->arch.emulate_ctxt.restart) {
		if (vcpu->mmio_needed) {
			memcpy(vcpu->mmio_data, kvm_run->mmio.data, 8);
			vcpu->mmio_read_completed = 1;
			vcpu->mmio_needed = 0;
4760
		}
4761
		vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
4762
		r = emulate_instruction(vcpu, 0, 0, EMULTYPE_NO_DECODE);
4763
		srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
4764 4765 4766 4767 4768
		if (r == EMULATE_DO_MMIO) {
			r = 0;
			goto out;
		}
	}
4769 4770 4771
	if (kvm_run->exit_reason == KVM_EXIT_HYPERCALL)
		kvm_register_write(vcpu, VCPU_REGS_RAX,
				     kvm_run->hypercall.ret);
4772

A
Avi Kivity 已提交
4773
	r = __vcpu_run(vcpu);
4774 4775

out:
4776
	post_kvm_run_save(vcpu);
4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787
	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &sigsaved, NULL);

	vcpu_put(vcpu);
	return r;
}

int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
{
	vcpu_load(vcpu);

4788 4789 4790 4791 4792 4793 4794 4795
	regs->rax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	regs->rbx = kvm_register_read(vcpu, VCPU_REGS_RBX);
	regs->rcx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	regs->rdx = kvm_register_read(vcpu, VCPU_REGS_RDX);
	regs->rsi = kvm_register_read(vcpu, VCPU_REGS_RSI);
	regs->rdi = kvm_register_read(vcpu, VCPU_REGS_RDI);
	regs->rsp = kvm_register_read(vcpu, VCPU_REGS_RSP);
	regs->rbp = kvm_register_read(vcpu, VCPU_REGS_RBP);
4796
#ifdef CONFIG_X86_64
4797 4798 4799 4800 4801 4802 4803 4804
	regs->r8 = kvm_register_read(vcpu, VCPU_REGS_R8);
	regs->r9 = kvm_register_read(vcpu, VCPU_REGS_R9);
	regs->r10 = kvm_register_read(vcpu, VCPU_REGS_R10);
	regs->r11 = kvm_register_read(vcpu, VCPU_REGS_R11);
	regs->r12 = kvm_register_read(vcpu, VCPU_REGS_R12);
	regs->r13 = kvm_register_read(vcpu, VCPU_REGS_R13);
	regs->r14 = kvm_register_read(vcpu, VCPU_REGS_R14);
	regs->r15 = kvm_register_read(vcpu, VCPU_REGS_R15);
4805 4806
#endif

4807
	regs->rip = kvm_rip_read(vcpu);
4808
	regs->rflags = kvm_get_rflags(vcpu);
4809 4810 4811 4812 4813 4814 4815 4816 4817 4818

	vcpu_put(vcpu);

	return 0;
}

int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
{
	vcpu_load(vcpu);

4819 4820 4821 4822 4823 4824 4825 4826
	kvm_register_write(vcpu, VCPU_REGS_RAX, regs->rax);
	kvm_register_write(vcpu, VCPU_REGS_RBX, regs->rbx);
	kvm_register_write(vcpu, VCPU_REGS_RCX, regs->rcx);
	kvm_register_write(vcpu, VCPU_REGS_RDX, regs->rdx);
	kvm_register_write(vcpu, VCPU_REGS_RSI, regs->rsi);
	kvm_register_write(vcpu, VCPU_REGS_RDI, regs->rdi);
	kvm_register_write(vcpu, VCPU_REGS_RSP, regs->rsp);
	kvm_register_write(vcpu, VCPU_REGS_RBP, regs->rbp);
4827
#ifdef CONFIG_X86_64
4828 4829 4830 4831 4832 4833 4834 4835
	kvm_register_write(vcpu, VCPU_REGS_R8, regs->r8);
	kvm_register_write(vcpu, VCPU_REGS_R9, regs->r9);
	kvm_register_write(vcpu, VCPU_REGS_R10, regs->r10);
	kvm_register_write(vcpu, VCPU_REGS_R11, regs->r11);
	kvm_register_write(vcpu, VCPU_REGS_R12, regs->r12);
	kvm_register_write(vcpu, VCPU_REGS_R13, regs->r13);
	kvm_register_write(vcpu, VCPU_REGS_R14, regs->r14);
	kvm_register_write(vcpu, VCPU_REGS_R15, regs->r15);
4836 4837
#endif

4838
	kvm_rip_write(vcpu, regs->rip);
4839
	kvm_set_rflags(vcpu, regs->rflags);
4840

4841 4842
	vcpu->arch.exception.pending = false;

4843 4844 4845 4846 4847 4848 4849 4850 4851
	vcpu_put(vcpu);

	return 0;
}

void kvm_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l)
{
	struct kvm_segment cs;

4852
	kvm_get_segment(vcpu, &cs, VCPU_SREG_CS);
4853 4854 4855 4856 4857 4858 4859 4860
	*db = cs.db;
	*l = cs.l;
}
EXPORT_SYMBOL_GPL(kvm_get_cs_db_l_bits);

int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
				  struct kvm_sregs *sregs)
{
4861
	struct desc_ptr dt;
4862 4863 4864

	vcpu_load(vcpu);

4865 4866 4867 4868 4869 4870
	kvm_get_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
	kvm_get_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
	kvm_get_segment(vcpu, &sregs->es, VCPU_SREG_ES);
	kvm_get_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
	kvm_get_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
	kvm_get_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
4871

4872 4873
	kvm_get_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_get_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
4874 4875

	kvm_x86_ops->get_idt(vcpu, &dt);
4876 4877
	sregs->idt.limit = dt.size;
	sregs->idt.base = dt.address;
4878
	kvm_x86_ops->get_gdt(vcpu, &dt);
4879 4880
	sregs->gdt.limit = dt.size;
	sregs->gdt.base = dt.address;
4881

4882
	sregs->cr0 = kvm_read_cr0(vcpu);
4883 4884
	sregs->cr2 = vcpu->arch.cr2;
	sregs->cr3 = vcpu->arch.cr3;
4885
	sregs->cr4 = kvm_read_cr4(vcpu);
4886
	sregs->cr8 = kvm_get_cr8(vcpu);
4887
	sregs->efer = vcpu->arch.efer;
4888 4889
	sregs->apic_base = kvm_get_apic_base(vcpu);

G
Gleb Natapov 已提交
4890
	memset(sregs->interrupt_bitmap, 0, sizeof sregs->interrupt_bitmap);
4891

4892
	if (vcpu->arch.interrupt.pending && !vcpu->arch.interrupt.soft)
4893 4894
		set_bit(vcpu->arch.interrupt.nr,
			(unsigned long *)sregs->interrupt_bitmap);
4895

4896 4897 4898 4899 4900
	vcpu_put(vcpu);

	return 0;
}

4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918
int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
				    struct kvm_mp_state *mp_state)
{
	vcpu_load(vcpu);
	mp_state->mp_state = vcpu->arch.mp_state;
	vcpu_put(vcpu);
	return 0;
}

int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
				    struct kvm_mp_state *mp_state)
{
	vcpu_load(vcpu);
	vcpu->arch.mp_state = mp_state->mp_state;
	vcpu_put(vcpu);
	return 0;
}

4919 4920
int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int reason,
		    bool has_error_code, u32 error_code)
4921
{
4922 4923
	int cs_db, cs_l, ret;
	cache_all_regs(vcpu);
4924

4925
	kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
4926

4927 4928 4929 4930 4931 4932 4933 4934 4935
	vcpu->arch.emulate_ctxt.vcpu = vcpu;
	vcpu->arch.emulate_ctxt.eflags = kvm_x86_ops->get_rflags(vcpu);
	vcpu->arch.emulate_ctxt.eip = kvm_rip_read(vcpu);
	vcpu->arch.emulate_ctxt.mode =
		(!is_protmode(vcpu)) ? X86EMUL_MODE_REAL :
		(vcpu->arch.emulate_ctxt.eflags & X86_EFLAGS_VM)
		? X86EMUL_MODE_VM86 : cs_l
		? X86EMUL_MODE_PROT64 :	cs_db
		? X86EMUL_MODE_PROT32 : X86EMUL_MODE_PROT16;
4936

4937
	ret = emulator_task_switch(&vcpu->arch.emulate_ctxt, &emulate_ops,
4938 4939
				   tss_selector, reason, has_error_code,
				   error_code);
4940 4941

	if (ret)
4942
		return EMULATE_FAIL;
4943

4944
	kvm_rip_write(vcpu, vcpu->arch.emulate_ctxt.eip);
4945 4946
	kvm_x86_ops->set_rflags(vcpu, vcpu->arch.emulate_ctxt.eflags);
	return EMULATE_DONE;
4947 4948 4949
}
EXPORT_SYMBOL_GPL(kvm_task_switch);

4950 4951 4952 4953
int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
				  struct kvm_sregs *sregs)
{
	int mmu_reset_needed = 0;
G
Gleb Natapov 已提交
4954
	int pending_vec, max_bits;
4955
	struct desc_ptr dt;
4956 4957 4958

	vcpu_load(vcpu);

4959 4960
	dt.size = sregs->idt.limit;
	dt.address = sregs->idt.base;
4961
	kvm_x86_ops->set_idt(vcpu, &dt);
4962 4963
	dt.size = sregs->gdt.limit;
	dt.address = sregs->gdt.base;
4964 4965
	kvm_x86_ops->set_gdt(vcpu, &dt);

4966 4967
	vcpu->arch.cr2 = sregs->cr2;
	mmu_reset_needed |= vcpu->arch.cr3 != sregs->cr3;
4968
	vcpu->arch.cr3 = sregs->cr3;
4969

4970
	kvm_set_cr8(vcpu, sregs->cr8);
4971

4972
	mmu_reset_needed |= vcpu->arch.efer != sregs->efer;
4973 4974 4975
	kvm_x86_ops->set_efer(vcpu, sregs->efer);
	kvm_set_apic_base(vcpu, sregs->apic_base);

4976
	mmu_reset_needed |= kvm_read_cr0(vcpu) != sregs->cr0;
4977
	kvm_x86_ops->set_cr0(vcpu, sregs->cr0);
4978
	vcpu->arch.cr0 = sregs->cr0;
4979

4980
	mmu_reset_needed |= kvm_read_cr4(vcpu) != sregs->cr4;
4981
	kvm_x86_ops->set_cr4(vcpu, sregs->cr4);
4982
	if (!is_long_mode(vcpu) && is_pae(vcpu)) {
4983
		load_pdptrs(vcpu, vcpu->arch.cr3);
4984 4985
		mmu_reset_needed = 1;
	}
4986 4987 4988 4989

	if (mmu_reset_needed)
		kvm_mmu_reset_context(vcpu);

G
Gleb Natapov 已提交
4990 4991 4992 4993
	max_bits = (sizeof sregs->interrupt_bitmap) << 3;
	pending_vec = find_first_bit(
		(const unsigned long *)sregs->interrupt_bitmap, max_bits);
	if (pending_vec < max_bits) {
4994
		kvm_queue_interrupt(vcpu, pending_vec, false);
G
Gleb Natapov 已提交
4995 4996 4997
		pr_debug("Set back pending irq %d\n", pending_vec);
		if (irqchip_in_kernel(vcpu->kvm))
			kvm_pic_clear_isr_ack(vcpu->kvm);
4998 4999
	}

5000 5001 5002 5003 5004 5005
	kvm_set_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
	kvm_set_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
	kvm_set_segment(vcpu, &sregs->es, VCPU_SREG_ES);
	kvm_set_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
	kvm_set_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
	kvm_set_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
5006

5007 5008
	kvm_set_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_set_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
5009

5010 5011
	update_cr8_intercept(vcpu);

M
Marcelo Tosatti 已提交
5012
	/* Older userspace won't unhalt the vcpu on reset. */
5013
	if (kvm_vcpu_is_bsp(vcpu) && kvm_rip_read(vcpu) == 0xfff0 &&
M
Marcelo Tosatti 已提交
5014
	    sregs->cs.selector == 0xf000 && sregs->cs.base == 0xffff0000 &&
5015
	    !is_protmode(vcpu))
M
Marcelo Tosatti 已提交
5016 5017
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;

5018 5019 5020 5021 5022
	vcpu_put(vcpu);

	return 0;
}

J
Jan Kiszka 已提交
5023 5024
int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
					struct kvm_guest_debug *dbg)
5025
{
5026
	unsigned long rflags;
5027
	int i, r;
5028 5029 5030

	vcpu_load(vcpu);

5031 5032 5033 5034 5035 5036 5037 5038 5039 5040
	if (dbg->control & (KVM_GUESTDBG_INJECT_DB | KVM_GUESTDBG_INJECT_BP)) {
		r = -EBUSY;
		if (vcpu->arch.exception.pending)
			goto unlock_out;
		if (dbg->control & KVM_GUESTDBG_INJECT_DB)
			kvm_queue_exception(vcpu, DB_VECTOR);
		else
			kvm_queue_exception(vcpu, BP_VECTOR);
	}

5041 5042 5043 5044 5045
	/*
	 * Read rflags as long as potentially injected trace flags are still
	 * filtered out.
	 */
	rflags = kvm_get_rflags(vcpu);
5046 5047 5048 5049 5050 5051

	vcpu->guest_debug = dbg->control;
	if (!(vcpu->guest_debug & KVM_GUESTDBG_ENABLE))
		vcpu->guest_debug = 0;

	if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP) {
5052 5053 5054 5055 5056 5057 5058 5059 5060 5061
		for (i = 0; i < KVM_NR_DB_REGS; ++i)
			vcpu->arch.eff_db[i] = dbg->arch.debugreg[i];
		vcpu->arch.switch_db_regs =
			(dbg->arch.debugreg[7] & DR7_BP_EN_MASK);
	} else {
		for (i = 0; i < KVM_NR_DB_REGS; i++)
			vcpu->arch.eff_db[i] = vcpu->arch.db[i];
		vcpu->arch.switch_db_regs = (vcpu->arch.dr7 & DR7_BP_EN_MASK);
	}

J
Jan Kiszka 已提交
5062 5063 5064
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
		vcpu->arch.singlestep_rip = kvm_rip_read(vcpu) +
			get_segment_base(vcpu, VCPU_SREG_CS);
5065

5066 5067 5068 5069 5070
	/*
	 * Trigger an rflags update that will inject or remove the trace
	 * flags.
	 */
	kvm_set_rflags(vcpu, rflags);
5071

5072
	kvm_x86_ops->set_guest_debug(vcpu, dbg);
5073

5074
	r = 0;
J
Jan Kiszka 已提交
5075

5076
unlock_out:
5077 5078 5079 5080 5081
	vcpu_put(vcpu);

	return r;
}

5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102
/*
 * fxsave fpu state.  Taken from x86_64/processor.h.  To be killed when
 * we have asm/x86/processor.h
 */
struct fxsave {
	u16	cwd;
	u16	swd;
	u16	twd;
	u16	fop;
	u64	rip;
	u64	rdp;
	u32	mxcsr;
	u32	mxcsr_mask;
	u32	st_space[32];	/* 8*16 bytes for each FP-reg = 128 bytes */
#ifdef CONFIG_X86_64
	u32	xmm_space[64];	/* 16*16 bytes for each XMM-reg = 256 bytes */
#else
	u32	xmm_space[32];	/* 8*16 bytes for each XMM-reg = 128 bytes */
#endif
};

5103 5104 5105 5106 5107 5108 5109 5110
/*
 * Translate a guest virtual address to a guest physical address.
 */
int kvm_arch_vcpu_ioctl_translate(struct kvm_vcpu *vcpu,
				    struct kvm_translation *tr)
{
	unsigned long vaddr = tr->linear_address;
	gpa_t gpa;
5111
	int idx;
5112 5113

	vcpu_load(vcpu);
5114
	idx = srcu_read_lock(&vcpu->kvm->srcu);
5115
	gpa = kvm_mmu_gva_to_gpa_system(vcpu, vaddr, NULL);
5116
	srcu_read_unlock(&vcpu->kvm->srcu, idx);
5117 5118 5119 5120 5121 5122 5123 5124 5125
	tr->physical_address = gpa;
	tr->valid = gpa != UNMAPPED_GVA;
	tr->writeable = 1;
	tr->usermode = 0;
	vcpu_put(vcpu);

	return 0;
}

5126 5127
int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
5128
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147

	vcpu_load(vcpu);

	memcpy(fpu->fpr, fxsave->st_space, 128);
	fpu->fcw = fxsave->cwd;
	fpu->fsw = fxsave->swd;
	fpu->ftwx = fxsave->twd;
	fpu->last_opcode = fxsave->fop;
	fpu->last_ip = fxsave->rip;
	fpu->last_dp = fxsave->rdp;
	memcpy(fpu->xmm, fxsave->xmm_space, sizeof fxsave->xmm_space);

	vcpu_put(vcpu);

	return 0;
}

int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
5148
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169

	vcpu_load(vcpu);

	memcpy(fxsave->st_space, fpu->fpr, 128);
	fxsave->cwd = fpu->fcw;
	fxsave->swd = fpu->fsw;
	fxsave->twd = fpu->ftwx;
	fxsave->fop = fpu->last_opcode;
	fxsave->rip = fpu->last_ip;
	fxsave->rdp = fpu->last_dp;
	memcpy(fxsave->xmm_space, fpu->xmm, sizeof fxsave->xmm_space);

	vcpu_put(vcpu);

	return 0;
}

void fx_init(struct kvm_vcpu *vcpu)
{
	unsigned after_mxcsr_mask;

5170 5171 5172 5173 5174 5175 5176
	/*
	 * Touch the fpu the first time in non atomic context as if
	 * this is the first fpu instruction the exception handler
	 * will fire before the instruction returns and it'll have to
	 * allocate ram with GFP_KERNEL.
	 */
	if (!used_math())
5177
		kvm_fx_save(&vcpu->arch.host_fx_image);
5178

5179 5180
	/* Initialize guest FPU by resetting ours and saving into guest's */
	preempt_disable();
5181 5182 5183 5184
	kvm_fx_save(&vcpu->arch.host_fx_image);
	kvm_fx_finit();
	kvm_fx_save(&vcpu->arch.guest_fx_image);
	kvm_fx_restore(&vcpu->arch.host_fx_image);
5185 5186
	preempt_enable();

5187
	vcpu->arch.cr0 |= X86_CR0_ET;
5188
	after_mxcsr_mask = offsetof(struct i387_fxsave_struct, st_space);
5189 5190
	vcpu->arch.guest_fx_image.mxcsr = 0x1f80;
	memset((void *)&vcpu->arch.guest_fx_image + after_mxcsr_mask,
5191 5192 5193 5194 5195 5196
	       0, sizeof(struct i387_fxsave_struct) - after_mxcsr_mask);
}
EXPORT_SYMBOL_GPL(fx_init);

void kvm_load_guest_fpu(struct kvm_vcpu *vcpu)
{
5197
	if (vcpu->guest_fpu_loaded)
5198 5199 5200
		return;

	vcpu->guest_fpu_loaded = 1;
5201 5202
	kvm_fx_save(&vcpu->arch.host_fx_image);
	kvm_fx_restore(&vcpu->arch.guest_fx_image);
5203
	trace_kvm_fpu(1);
5204 5205 5206 5207 5208 5209 5210 5211
}

void kvm_put_guest_fpu(struct kvm_vcpu *vcpu)
{
	if (!vcpu->guest_fpu_loaded)
		return;

	vcpu->guest_fpu_loaded = 0;
5212 5213
	kvm_fx_save(&vcpu->arch.guest_fx_image);
	kvm_fx_restore(&vcpu->arch.host_fx_image);
A
Avi Kivity 已提交
5214
	++vcpu->stat.fpu_reload;
5215
	set_bit(KVM_REQ_DEACTIVATE_FPU, &vcpu->requests);
5216
	trace_kvm_fpu(0);
5217
}
5218 5219 5220

void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
5221 5222 5223 5224 5225
	if (vcpu->arch.time_page) {
		kvm_release_page_dirty(vcpu->arch.time_page);
		vcpu->arch.time_page = NULL;
	}

5226 5227 5228 5229 5230 5231
	kvm_x86_ops->vcpu_free(vcpu);
}

struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm,
						unsigned int id)
{
5232 5233
	return kvm_x86_ops->vcpu_create(kvm, id);
}
5234

5235 5236 5237
int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
{
	int r;
5238 5239

	/* We do fxsave: this must be aligned. */
5240
	BUG_ON((unsigned long)&vcpu->arch.host_fx_image & 0xF);
5241

S
Sheng Yang 已提交
5242
	vcpu->arch.mtrr_state.have_fixed = 1;
5243 5244 5245 5246 5247 5248 5249 5250
	vcpu_load(vcpu);
	r = kvm_arch_vcpu_reset(vcpu);
	if (r == 0)
		r = kvm_mmu_setup(vcpu);
	vcpu_put(vcpu);
	if (r < 0)
		goto free_vcpu;

5251
	return 0;
5252 5253
free_vcpu:
	kvm_x86_ops->vcpu_free(vcpu);
5254
	return r;
5255 5256
}

5257
void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
5258 5259 5260 5261 5262 5263 5264 5265 5266 5267
{
	vcpu_load(vcpu);
	kvm_mmu_unload(vcpu);
	vcpu_put(vcpu);

	kvm_x86_ops->vcpu_free(vcpu);
}

int kvm_arch_vcpu_reset(struct kvm_vcpu *vcpu)
{
5268 5269 5270
	vcpu->arch.nmi_pending = false;
	vcpu->arch.nmi_injected = false;

5271 5272 5273 5274 5275
	vcpu->arch.switch_db_regs = 0;
	memset(vcpu->arch.db, 0, sizeof(vcpu->arch.db));
	vcpu->arch.dr6 = DR6_FIXED_1;
	vcpu->arch.dr7 = DR7_FIXED_1;

5276 5277 5278
	return kvm_x86_ops->vcpu_reset(vcpu);
}

5279
int kvm_arch_hardware_enable(void *garbage)
5280
{
5281 5282 5283 5284 5285 5286 5287 5288
	/*
	 * Since this may be called from a hotplug notifcation,
	 * we can't get the CPU frequency directly.
	 */
	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) {
		int cpu = raw_smp_processor_id();
		per_cpu(cpu_tsc_khz, cpu) = 0;
	}
A
Avi Kivity 已提交
5289 5290 5291

	kvm_shared_msr_cpu_online();

5292
	return kvm_x86_ops->hardware_enable(garbage);
5293 5294 5295 5296 5297
}

void kvm_arch_hardware_disable(void *garbage)
{
	kvm_x86_ops->hardware_disable(garbage);
5298
	drop_user_return_notifiers(garbage);
5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324
}

int kvm_arch_hardware_setup(void)
{
	return kvm_x86_ops->hardware_setup();
}

void kvm_arch_hardware_unsetup(void)
{
	kvm_x86_ops->hardware_unsetup();
}

void kvm_arch_check_processor_compat(void *rtn)
{
	kvm_x86_ops->check_processor_compatibility(rtn);
}

int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
{
	struct page *page;
	struct kvm *kvm;
	int r;

	BUG_ON(vcpu->kvm == NULL);
	kvm = vcpu->kvm;

5325
	vcpu->arch.mmu.root_hpa = INVALID_PAGE;
5326
	if (!irqchip_in_kernel(kvm) || kvm_vcpu_is_bsp(vcpu))
5327
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
5328
	else
5329
		vcpu->arch.mp_state = KVM_MP_STATE_UNINITIALIZED;
5330 5331 5332 5333 5334 5335

	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
	if (!page) {
		r = -ENOMEM;
		goto fail;
	}
5336
	vcpu->arch.pio_data = page_address(page);
5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347

	r = kvm_mmu_create(vcpu);
	if (r < 0)
		goto fail_free_pio_data;

	if (irqchip_in_kernel(kvm)) {
		r = kvm_create_lapic(vcpu);
		if (r < 0)
			goto fail_mmu_destroy;
	}

H
Huang Ying 已提交
5348 5349 5350 5351
	vcpu->arch.mce_banks = kzalloc(KVM_MAX_MCE_BANKS * sizeof(u64) * 4,
				       GFP_KERNEL);
	if (!vcpu->arch.mce_banks) {
		r = -ENOMEM;
5352
		goto fail_free_lapic;
H
Huang Ying 已提交
5353 5354 5355
	}
	vcpu->arch.mcg_cap = KVM_MAX_MCE_BANKS;

5356
	return 0;
5357 5358
fail_free_lapic:
	kvm_free_lapic(vcpu);
5359 5360 5361
fail_mmu_destroy:
	kvm_mmu_destroy(vcpu);
fail_free_pio_data:
5362
	free_page((unsigned long)vcpu->arch.pio_data);
5363 5364 5365 5366 5367 5368
fail:
	return r;
}

void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
{
5369 5370
	int idx;

5371
	kfree(vcpu->arch.mce_banks);
5372
	kvm_free_lapic(vcpu);
5373
	idx = srcu_read_lock(&vcpu->kvm->srcu);
5374
	kvm_mmu_destroy(vcpu);
5375
	srcu_read_unlock(&vcpu->kvm->srcu, idx);
5376
	free_page((unsigned long)vcpu->arch.pio_data);
5377
}
5378 5379 5380 5381 5382 5383 5384 5385

struct  kvm *kvm_arch_create_vm(void)
{
	struct kvm *kvm = kzalloc(sizeof(struct kvm), GFP_KERNEL);

	if (!kvm)
		return ERR_PTR(-ENOMEM);

5386 5387 5388 5389 5390 5391
	kvm->arch.aliases = kzalloc(sizeof(struct kvm_mem_aliases), GFP_KERNEL);
	if (!kvm->arch.aliases) {
		kfree(kvm);
		return ERR_PTR(-ENOMEM);
	}

5392
	INIT_LIST_HEAD(&kvm->arch.active_mmu_pages);
B
Ben-Ami Yassour 已提交
5393
	INIT_LIST_HEAD(&kvm->arch.assigned_dev_head);
5394

5395 5396 5397
	/* Reserve bit 0 of irq_sources_bitmap for userspace irq source */
	set_bit(KVM_USERSPACE_IRQ_SOURCE_ID, &kvm->arch.irq_sources_bitmap);

5398 5399
	rdtscll(kvm->arch.vm_init_tsc);

5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412
	return kvm;
}

static void kvm_unload_vcpu_mmu(struct kvm_vcpu *vcpu)
{
	vcpu_load(vcpu);
	kvm_mmu_unload(vcpu);
	vcpu_put(vcpu);
}

static void kvm_free_vcpus(struct kvm *kvm)
{
	unsigned int i;
5413
	struct kvm_vcpu *vcpu;
5414 5415 5416 5417

	/*
	 * Unpin any mmu pages first.
	 */
5418 5419 5420 5421 5422 5423 5424 5425
	kvm_for_each_vcpu(i, vcpu, kvm)
		kvm_unload_vcpu_mmu(vcpu);
	kvm_for_each_vcpu(i, vcpu, kvm)
		kvm_arch_vcpu_free(vcpu);

	mutex_lock(&kvm->lock);
	for (i = 0; i < atomic_read(&kvm->online_vcpus); i++)
		kvm->vcpus[i] = NULL;
5426

5427 5428
	atomic_set(&kvm->online_vcpus, 0);
	mutex_unlock(&kvm->lock);
5429 5430
}

5431 5432
void kvm_arch_sync_events(struct kvm *kvm)
{
5433
	kvm_free_all_assigned_devices(kvm);
5434 5435
}

5436 5437
void kvm_arch_destroy_vm(struct kvm *kvm)
{
5438
	kvm_iommu_unmap_guest(kvm);
S
Sheng Yang 已提交
5439
	kvm_free_pit(kvm);
5440 5441
	kfree(kvm->arch.vpic);
	kfree(kvm->arch.vioapic);
5442 5443
	kvm_free_vcpus(kvm);
	kvm_free_physmem(kvm);
5444 5445
	if (kvm->arch.apic_access_page)
		put_page(kvm->arch.apic_access_page);
5446 5447
	if (kvm->arch.ept_identity_pagetable)
		put_page(kvm->arch.ept_identity_pagetable);
5448
	cleanup_srcu_struct(&kvm->srcu);
5449
	kfree(kvm->arch.aliases);
5450 5451
	kfree(kvm);
}
5452

5453 5454
int kvm_arch_prepare_memory_region(struct kvm *kvm,
				struct kvm_memory_slot *memslot,
5455
				struct kvm_memory_slot old,
5456
				struct kvm_userspace_memory_region *mem,
5457 5458
				int user_alloc)
{
5459
	int npages = memslot->npages;
5460 5461 5462 5463 5464 5465

	/*To keep backward compatibility with older userspace,
	 *x86 needs to hanlde !user_alloc case.
	 */
	if (!user_alloc) {
		if (npages && !old.rmap) {
5466 5467
			unsigned long userspace_addr;

5468
			down_write(&current->mm->mmap_sem);
5469 5470 5471
			userspace_addr = do_mmap(NULL, 0,
						 npages * PAGE_SIZE,
						 PROT_READ | PROT_WRITE,
5472
						 MAP_PRIVATE | MAP_ANONYMOUS,
5473
						 0);
5474
			up_write(&current->mm->mmap_sem);
5475

5476 5477 5478 5479
			if (IS_ERR((void *)userspace_addr))
				return PTR_ERR((void *)userspace_addr);

			memslot->userspace_addr = userspace_addr;
5480 5481 5482
		}
	}

5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507

	return 0;
}

void kvm_arch_commit_memory_region(struct kvm *kvm,
				struct kvm_userspace_memory_region *mem,
				struct kvm_memory_slot old,
				int user_alloc)
{

	int npages = mem->memory_size >> PAGE_SHIFT;

	if (!user_alloc && !old.user_alloc && old.rmap && !npages) {
		int ret;

		down_write(&current->mm->mmap_sem);
		ret = do_munmap(current->mm, old.userspace_addr,
				old.npages * PAGE_SIZE);
		up_write(&current->mm->mmap_sem);
		if (ret < 0)
			printk(KERN_WARNING
			       "kvm_vm_ioctl_set_memory_region: "
			       "failed to munmap memory\n");
	}

5508
	spin_lock(&kvm->mmu_lock);
5509
	if (!kvm->arch.n_requested_mmu_pages) {
5510 5511 5512 5513 5514
		unsigned int nr_mmu_pages = kvm_mmu_calculate_mmu_pages(kvm);
		kvm_mmu_change_mmu_pages(kvm, nr_mmu_pages);
	}

	kvm_mmu_slot_remove_write_access(kvm, mem->slot);
5515
	spin_unlock(&kvm->mmu_lock);
5516
}
5517

5518 5519 5520
void kvm_arch_flush_shadow(struct kvm *kvm)
{
	kvm_mmu_zap_all(kvm);
5521
	kvm_reload_remote_mmus(kvm);
5522 5523
}

5524 5525
int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
{
5526
	return vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE
5527 5528 5529 5530
		|| vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED
		|| vcpu->arch.nmi_pending ||
		(kvm_arch_interrupt_allowed(vcpu) &&
		 kvm_cpu_has_interrupt(vcpu));
5531
}
5532 5533 5534

void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
{
5535 5536
	int me;
	int cpu = vcpu->cpu;
5537 5538 5539 5540 5541

	if (waitqueue_active(&vcpu->wq)) {
		wake_up_interruptible(&vcpu->wq);
		++vcpu->stat.halt_wakeup;
	}
5542 5543 5544 5545 5546

	me = get_cpu();
	if (cpu != me && (unsigned)cpu < nr_cpu_ids && cpu_online(cpu))
		if (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests))
			smp_send_reschedule(cpu);
5547
	put_cpu();
5548
}
5549 5550 5551 5552 5553

int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu)
{
	return kvm_x86_ops->interrupt_allowed(vcpu);
}
5554

J
Jan Kiszka 已提交
5555 5556 5557 5558 5559 5560 5561 5562 5563
bool kvm_is_linear_rip(struct kvm_vcpu *vcpu, unsigned long linear_rip)
{
	unsigned long current_rip = kvm_rip_read(vcpu) +
		get_segment_base(vcpu, VCPU_SREG_CS);

	return current_rip == linear_rip;
}
EXPORT_SYMBOL_GPL(kvm_is_linear_rip);

5564 5565 5566 5567 5568 5569
unsigned long kvm_get_rflags(struct kvm_vcpu *vcpu)
{
	unsigned long rflags;

	rflags = kvm_x86_ops->get_rflags(vcpu);
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
5570
		rflags &= ~X86_EFLAGS_TF;
5571 5572 5573 5574 5575 5576 5577
	return rflags;
}
EXPORT_SYMBOL_GPL(kvm_get_rflags);

void kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
{
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP &&
J
Jan Kiszka 已提交
5578
	    kvm_is_linear_rip(vcpu, vcpu->arch.singlestep_rip))
5579
		rflags |= X86_EFLAGS_TF;
5580 5581 5582 5583
	kvm_x86_ops->set_rflags(vcpu, rflags);
}
EXPORT_SYMBOL_GPL(kvm_set_rflags);

5584 5585 5586 5587 5588
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_exit);
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_inj_virq);
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_page_fault);
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_msr);
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_cr);
5589
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmrun);
5590
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmexit);
5591
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmexit_inject);
5592
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_intr_vmexit);
5593
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_invlpga);
5594
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_skinit);
5595
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_intercepts);