vmx.c 98.4 KB
Newer Older
A
Avi Kivity 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * This module enables machines with Intel VT-x extensions to run virtual
 * machines without emulation or binary translation.
 *
 * Copyright (C) 2006 Qumranet, Inc.
 *
 * Authors:
 *   Avi Kivity   <avi@qumranet.com>
 *   Yaniv Kamay  <yaniv@qumranet.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */

18
#include "irq.h"
19
#include "mmu.h"
A
Avi Kivity 已提交
20

21
#include <linux/kvm_host.h>
A
Avi Kivity 已提交
22
#include <linux/module.h>
23
#include <linux/kernel.h>
A
Avi Kivity 已提交
24 25
#include <linux/mm.h>
#include <linux/highmem.h>
A
Alexey Dobriyan 已提交
26
#include <linux/sched.h>
27
#include <linux/moduleparam.h>
28
#include "kvm_cache_regs.h"
29
#include "x86.h"
A
Avi Kivity 已提交
30

A
Avi Kivity 已提交
31
#include <asm/io.h>
A
Anthony Liguori 已提交
32
#include <asm/desc.h>
33
#include <asm/vmx.h>
34
#include <asm/virtext.h>
A
Andi Kleen 已提交
35
#include <asm/mce.h>
A
Avi Kivity 已提交
36

37 38
#define __ex(x) __kvm_handle_fault_on_reboot(x)

A
Avi Kivity 已提交
39 40 41
MODULE_AUTHOR("Qumranet");
MODULE_LICENSE("GPL");

42
static int __read_mostly bypass_guest_pf = 1;
43
module_param(bypass_guest_pf, bool, S_IRUGO);
44

45
static int __read_mostly enable_vpid = 1;
46
module_param_named(vpid, enable_vpid, bool, 0444);
47

48
static int __read_mostly flexpriority_enabled = 1;
49
module_param_named(flexpriority, flexpriority_enabled, bool, S_IRUGO);
50

51
static int __read_mostly enable_ept = 1;
52
module_param_named(ept, enable_ept, bool, S_IRUGO);
S
Sheng Yang 已提交
53

54
static int __read_mostly emulate_invalid_guest_state = 0;
55
module_param(emulate_invalid_guest_state, bool, S_IRUGO);
56

57 58 59 60 61 62 63
struct vmcs {
	u32 revision_id;
	u32 abort;
	char data[0];
};

struct vcpu_vmx {
R
Rusty Russell 已提交
64
	struct kvm_vcpu       vcpu;
65
	struct list_head      local_vcpus_link;
66
	unsigned long         host_rsp;
67
	int                   launched;
68
	u8                    fail;
69
	u32                   idt_vectoring_info;
70 71 72 73 74 75 76 77 78 79 80 81
	struct kvm_msr_entry *guest_msrs;
	struct kvm_msr_entry *host_msrs;
	int                   nmsrs;
	int                   save_nmsrs;
	int                   msr_offset_efer;
#ifdef CONFIG_X86_64
	int                   msr_offset_kernel_gs_base;
#endif
	struct vmcs          *vmcs;
	struct {
		int           loaded;
		u16           fs_sel, gs_sel, ldt_sel;
82 83
		int           gs_ldt_reload_needed;
		int           fs_reload_needed;
84
		int           guest_efer_loaded;
M
Mike Day 已提交
85
	} host_state;
86 87 88 89 90 91 92
	struct {
		struct {
			bool pending;
			u8 vector;
			unsigned rip;
		} irq;
	} rmode;
93
	int vpid;
94
	bool emulation_required;
95
	enum emulation_result invalid_state_emulation_result;
96 97 98 99 100

	/* Support for vnmi-less CPUs */
	int soft_vnmi_blocked;
	ktime_t entry_time;
	s64 vnmi_blocked_time;
A
Andi Kleen 已提交
101
	u32 exit_reason;
102 103 104 105
};

static inline struct vcpu_vmx *to_vmx(struct kvm_vcpu *vcpu)
{
R
Rusty Russell 已提交
106
	return container_of(vcpu, struct vcpu_vmx, vcpu);
107 108
}

109
static int init_rmode(struct kvm *kvm);
110
static u64 construct_eptp(unsigned long root_hpa);
111

A
Avi Kivity 已提交
112 113
static DEFINE_PER_CPU(struct vmcs *, vmxarea);
static DEFINE_PER_CPU(struct vmcs *, current_vmcs);
114
static DEFINE_PER_CPU(struct list_head, vcpus_on_cpu);
A
Avi Kivity 已提交
115

116 117
static unsigned long *vmx_io_bitmap_a;
static unsigned long *vmx_io_bitmap_b;
118 119
static unsigned long *vmx_msr_bitmap_legacy;
static unsigned long *vmx_msr_bitmap_longmode;
120

121 122 123
static DECLARE_BITMAP(vmx_vpid_bitmap, VMX_NR_VPIDS);
static DEFINE_SPINLOCK(vmx_vpid_lock);

124
static struct vmcs_config {
A
Avi Kivity 已提交
125 126 127
	int size;
	int order;
	u32 revision_id;
128 129
	u32 pin_based_exec_ctrl;
	u32 cpu_based_exec_ctrl;
130
	u32 cpu_based_2nd_exec_ctrl;
131 132 133
	u32 vmexit_ctrl;
	u32 vmentry_ctrl;
} vmcs_config;
A
Avi Kivity 已提交
134

H
Hannes Eder 已提交
135
static struct vmx_capability {
S
Sheng Yang 已提交
136 137 138 139
	u32 ept;
	u32 vpid;
} vmx_capability;

A
Avi Kivity 已提交
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163
#define VMX_SEGMENT_FIELD(seg)					\
	[VCPU_SREG_##seg] = {                                   \
		.selector = GUEST_##seg##_SELECTOR,		\
		.base = GUEST_##seg##_BASE,		   	\
		.limit = GUEST_##seg##_LIMIT,		   	\
		.ar_bytes = GUEST_##seg##_AR_BYTES,	   	\
	}

static struct kvm_vmx_segment_field {
	unsigned selector;
	unsigned base;
	unsigned limit;
	unsigned ar_bytes;
} kvm_vmx_segment_fields[] = {
	VMX_SEGMENT_FIELD(CS),
	VMX_SEGMENT_FIELD(DS),
	VMX_SEGMENT_FIELD(ES),
	VMX_SEGMENT_FIELD(FS),
	VMX_SEGMENT_FIELD(GS),
	VMX_SEGMENT_FIELD(SS),
	VMX_SEGMENT_FIELD(TR),
	VMX_SEGMENT_FIELD(LDTR),
};

164 165 166 167
/*
 * Keep MSR_K6_STAR at the end, as setup_msrs() will try to optimize it
 * away by decrementing the array size.
 */
A
Avi Kivity 已提交
168
static const u32 vmx_msr_index[] = {
169
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
170 171 172 173
	MSR_SYSCALL_MASK, MSR_LSTAR, MSR_CSTAR, MSR_KERNEL_GS_BASE,
#endif
	MSR_EFER, MSR_K6_STAR,
};
174
#define NR_VMX_MSR ARRAY_SIZE(vmx_msr_index)
A
Avi Kivity 已提交
175

176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
static void load_msrs(struct kvm_msr_entry *e, int n)
{
	int i;

	for (i = 0; i < n; ++i)
		wrmsrl(e[i].index, e[i].data);
}

static void save_msrs(struct kvm_msr_entry *e, int n)
{
	int i;

	for (i = 0; i < n; ++i)
		rdmsrl(e[i].index, e[i].data);
}

A
Avi Kivity 已提交
192 193 194 195
static inline int is_page_fault(u32 intr_info)
{
	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VECTOR_MASK |
			     INTR_INFO_VALID_MASK)) ==
196
		(INTR_TYPE_HARD_EXCEPTION | PF_VECTOR | INTR_INFO_VALID_MASK);
A
Avi Kivity 已提交
197 198
}

199 200 201 202
static inline int is_no_device(u32 intr_info)
{
	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VECTOR_MASK |
			     INTR_INFO_VALID_MASK)) ==
203
		(INTR_TYPE_HARD_EXCEPTION | NM_VECTOR | INTR_INFO_VALID_MASK);
204 205
}

206 207 208 209
static inline int is_invalid_opcode(u32 intr_info)
{
	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VECTOR_MASK |
			     INTR_INFO_VALID_MASK)) ==
210
		(INTR_TYPE_HARD_EXCEPTION | UD_VECTOR | INTR_INFO_VALID_MASK);
211 212
}

A
Avi Kivity 已提交
213 214 215 216 217 218
static inline int is_external_interrupt(u32 intr_info)
{
	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VALID_MASK))
		== (INTR_TYPE_EXT_INTR | INTR_INFO_VALID_MASK);
}

A
Andi Kleen 已提交
219 220 221 222 223 224 225
static inline int is_machine_check(u32 intr_info)
{
	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VECTOR_MASK |
			     INTR_INFO_VALID_MASK)) ==
		(INTR_TYPE_HARD_EXCEPTION | MC_VECTOR | INTR_INFO_VALID_MASK);
}

S
Sheng Yang 已提交
226 227
static inline int cpu_has_vmx_msr_bitmap(void)
{
228
	return vmcs_config.cpu_based_exec_ctrl & CPU_BASED_USE_MSR_BITMAPS;
S
Sheng Yang 已提交
229 230
}

231 232
static inline int cpu_has_vmx_tpr_shadow(void)
{
233
	return vmcs_config.cpu_based_exec_ctrl & CPU_BASED_TPR_SHADOW;
234 235 236 237
}

static inline int vm_need_tpr_shadow(struct kvm *kvm)
{
238
	return (cpu_has_vmx_tpr_shadow()) && (irqchip_in_kernel(kvm));
239 240
}

241 242
static inline int cpu_has_secondary_exec_ctrls(void)
{
243 244
	return vmcs_config.cpu_based_exec_ctrl &
		CPU_BASED_ACTIVATE_SECONDARY_CONTROLS;
245 246
}

247
static inline bool cpu_has_vmx_virtualize_apic_accesses(void)
248
{
249 250 251 252 253 254 255 256
	return vmcs_config.cpu_based_2nd_exec_ctrl &
		SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES;
}

static inline bool cpu_has_vmx_flexpriority(void)
{
	return cpu_has_vmx_tpr_shadow() &&
		cpu_has_vmx_virtualize_apic_accesses();
257 258
}

S
Sheng Yang 已提交
259 260
static inline int cpu_has_vmx_invept_individual_addr(void)
{
261
	return !!(vmx_capability.ept & VMX_EPT_EXTENT_INDIVIDUAL_BIT);
S
Sheng Yang 已提交
262 263 264 265
}

static inline int cpu_has_vmx_invept_context(void)
{
266
	return !!(vmx_capability.ept & VMX_EPT_EXTENT_CONTEXT_BIT);
S
Sheng Yang 已提交
267 268 269 270
}

static inline int cpu_has_vmx_invept_global(void)
{
271
	return !!(vmx_capability.ept & VMX_EPT_EXTENT_GLOBAL_BIT);
S
Sheng Yang 已提交
272 273 274 275
}

static inline int cpu_has_vmx_ept(void)
{
276 277
	return vmcs_config.cpu_based_2nd_exec_ctrl &
		SECONDARY_EXEC_ENABLE_EPT;
S
Sheng Yang 已提交
278 279
}

280 281
static inline int vm_need_virtualize_apic_accesses(struct kvm *kvm)
{
282 283 284
	return flexpriority_enabled &&
		(cpu_has_vmx_virtualize_apic_accesses()) &&
		(irqchip_in_kernel(kvm));
285 286
}

287 288
static inline int cpu_has_vmx_vpid(void)
{
289 290
	return vmcs_config.cpu_based_2nd_exec_ctrl &
		SECONDARY_EXEC_ENABLE_VPID;
291 292
}

293 294 295 296 297
static inline int cpu_has_virtual_nmis(void)
{
	return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS;
}

298 299 300 301 302
static inline bool report_flexpriority(void)
{
	return flexpriority_enabled;
}

R
Rusty Russell 已提交
303
static int __find_msr_index(struct vcpu_vmx *vmx, u32 msr)
304 305 306
{
	int i;

307 308
	for (i = 0; i < vmx->nmsrs; ++i)
		if (vmx->guest_msrs[i].index == msr)
309 310 311 312
			return i;
	return -1;
}

313 314 315 316 317 318 319 320
static inline void __invvpid(int ext, u16 vpid, gva_t gva)
{
    struct {
	u64 vpid : 16;
	u64 rsvd : 48;
	u64 gva;
    } operand = { vpid, 0, gva };

321
    asm volatile (__ex(ASM_VMX_INVVPID)
322 323 324 325 326
		  /* CF==1 or ZF==1 --> rc = -1 */
		  "; ja 1f ; ud2 ; 1:"
		  : : "a"(&operand), "c"(ext) : "cc", "memory");
}

327 328 329 330 331 332
static inline void __invept(int ext, u64 eptp, gpa_t gpa)
{
	struct {
		u64 eptp, gpa;
	} operand = {eptp, gpa};

333
	asm volatile (__ex(ASM_VMX_INVEPT)
334 335 336 337 338
			/* CF==1 or ZF==1 --> rc = -1 */
			"; ja 1f ; ud2 ; 1:\n"
			: : "a" (&operand), "c" (ext) : "cc", "memory");
}

R
Rusty Russell 已提交
339
static struct kvm_msr_entry *find_msr_entry(struct vcpu_vmx *vmx, u32 msr)
340 341 342
{
	int i;

R
Rusty Russell 已提交
343
	i = __find_msr_index(vmx, msr);
344
	if (i >= 0)
345
		return &vmx->guest_msrs[i];
A
Al Viro 已提交
346
	return NULL;
347 348
}

A
Avi Kivity 已提交
349 350 351 352 353
static void vmcs_clear(struct vmcs *vmcs)
{
	u64 phys_addr = __pa(vmcs);
	u8 error;

354
	asm volatile (__ex(ASM_VMX_VMCLEAR_RAX) "; setna %0"
A
Avi Kivity 已提交
355 356 357 358 359 360 361 362 363
		      : "=g"(error) : "a"(&phys_addr), "m"(phys_addr)
		      : "cc", "memory");
	if (error)
		printk(KERN_ERR "kvm: vmclear fail: %p/%llx\n",
		       vmcs, phys_addr);
}

static void __vcpu_clear(void *arg)
{
R
Rusty Russell 已提交
364
	struct vcpu_vmx *vmx = arg;
365
	int cpu = raw_smp_processor_id();
A
Avi Kivity 已提交
366

R
Rusty Russell 已提交
367
	if (vmx->vcpu.cpu == cpu)
368 369
		vmcs_clear(vmx->vmcs);
	if (per_cpu(current_vmcs, cpu) == vmx->vmcs)
A
Avi Kivity 已提交
370
		per_cpu(current_vmcs, cpu) = NULL;
371
	rdtscll(vmx->vcpu.arch.host_tsc);
372 373 374
	list_del(&vmx->local_vcpus_link);
	vmx->vcpu.cpu = -1;
	vmx->launched = 0;
A
Avi Kivity 已提交
375 376
}

R
Rusty Russell 已提交
377
static void vcpu_clear(struct vcpu_vmx *vmx)
A
Avi Kivity 已提交
378
{
379 380
	if (vmx->vcpu.cpu == -1)
		return;
381
	smp_call_function_single(vmx->vcpu.cpu, __vcpu_clear, vmx, 1);
A
Avi Kivity 已提交
382 383
}

384 385 386 387 388 389 390 391
static inline void vpid_sync_vcpu_all(struct vcpu_vmx *vmx)
{
	if (vmx->vpid == 0)
		return;

	__invvpid(VMX_VPID_EXTENT_SINGLE_CONTEXT, vmx->vpid, 0);
}

392 393 394 395 396 397 398 399
static inline void ept_sync_global(void)
{
	if (cpu_has_vmx_invept_global())
		__invept(VMX_EPT_EXTENT_GLOBAL, 0, 0);
}

static inline void ept_sync_context(u64 eptp)
{
400
	if (enable_ept) {
401 402 403 404 405 406 407 408 409
		if (cpu_has_vmx_invept_context())
			__invept(VMX_EPT_EXTENT_CONTEXT, eptp, 0);
		else
			ept_sync_global();
	}
}

static inline void ept_sync_individual_addr(u64 eptp, gpa_t gpa)
{
410
	if (enable_ept) {
411 412 413 414 415 416 417 418
		if (cpu_has_vmx_invept_individual_addr())
			__invept(VMX_EPT_EXTENT_INDIVIDUAL_ADDR,
					eptp, gpa);
		else
			ept_sync_context(eptp);
	}
}

A
Avi Kivity 已提交
419 420 421 422
static unsigned long vmcs_readl(unsigned long field)
{
	unsigned long value;

423
	asm volatile (__ex(ASM_VMX_VMREAD_RDX_RAX)
A
Avi Kivity 已提交
424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
		      : "=a"(value) : "d"(field) : "cc");
	return value;
}

static u16 vmcs_read16(unsigned long field)
{
	return vmcs_readl(field);
}

static u32 vmcs_read32(unsigned long field)
{
	return vmcs_readl(field);
}

static u64 vmcs_read64(unsigned long field)
{
440
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
441 442 443 444 445 446
	return vmcs_readl(field);
#else
	return vmcs_readl(field) | ((u64)vmcs_readl(field+1) << 32);
#endif
}

447 448 449 450 451 452 453
static noinline void vmwrite_error(unsigned long field, unsigned long value)
{
	printk(KERN_ERR "vmwrite error: reg %lx value %lx (err %d)\n",
	       field, value, vmcs_read32(VM_INSTRUCTION_ERROR));
	dump_stack();
}

A
Avi Kivity 已提交
454 455 456 457
static void vmcs_writel(unsigned long field, unsigned long value)
{
	u8 error;

458
	asm volatile (__ex(ASM_VMX_VMWRITE_RAX_RDX) "; setna %0"
M
Mike Day 已提交
459
		       : "=q"(error) : "a"(value), "d"(field) : "cc");
460 461
	if (unlikely(error))
		vmwrite_error(field, value);
A
Avi Kivity 已提交
462 463 464 465 466 467 468 469 470 471 472 473 474 475 476
}

static void vmcs_write16(unsigned long field, u16 value)
{
	vmcs_writel(field, value);
}

static void vmcs_write32(unsigned long field, u32 value)
{
	vmcs_writel(field, value);
}

static void vmcs_write64(unsigned long field, u64 value)
{
	vmcs_writel(field, value);
477
#ifndef CONFIG_X86_64
A
Avi Kivity 已提交
478 479 480 481 482
	asm volatile ("");
	vmcs_writel(field+1, value >> 32);
#endif
}

483 484 485 486 487 488 489 490 491 492
static void vmcs_clear_bits(unsigned long field, u32 mask)
{
	vmcs_writel(field, vmcs_readl(field) & ~mask);
}

static void vmcs_set_bits(unsigned long field, u32 mask)
{
	vmcs_writel(field, vmcs_readl(field) | mask);
}

493 494 495 496
static void update_exception_bitmap(struct kvm_vcpu *vcpu)
{
	u32 eb;

A
Andi Kleen 已提交
497
	eb = (1u << PF_VECTOR) | (1u << UD_VECTOR) | (1u << MC_VECTOR);
498 499
	if (!vcpu->fpu_active)
		eb |= 1u << NM_VECTOR;
J
Jan Kiszka 已提交
500 501 502 503 504 505 506
	if (vcpu->guest_debug & KVM_GUESTDBG_ENABLE) {
		if (vcpu->guest_debug &
		    (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))
			eb |= 1u << DB_VECTOR;
		if (vcpu->guest_debug & KVM_GUESTDBG_USE_SW_BP)
			eb |= 1u << BP_VECTOR;
	}
507
	if (vcpu->arch.rmode.vm86_active)
508
		eb = ~0;
509
	if (enable_ept)
510
		eb &= ~(1u << PF_VECTOR); /* bypass_guest_pf = 0 */
511 512 513
	vmcs_write32(EXCEPTION_BITMAP, eb);
}

514 515 516 517 518 519
static void reload_tss(void)
{
	/*
	 * VT restores TR but not its size.  Useless.
	 */
	struct descriptor_table gdt;
520
	struct desc_struct *descs;
521

522
	kvm_get_gdt(&gdt);
523 524 525 526 527
	descs = (void *)gdt.base;
	descs[GDT_ENTRY_TSS].type = 9; /* available TSS */
	load_TR_desc();
}

R
Rusty Russell 已提交
528
static void load_transition_efer(struct vcpu_vmx *vmx)
529
{
530
	int efer_offset = vmx->msr_offset_efer;
531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
	u64 host_efer = vmx->host_msrs[efer_offset].data;
	u64 guest_efer = vmx->guest_msrs[efer_offset].data;
	u64 ignore_bits;

	if (efer_offset < 0)
		return;
	/*
	 * NX is emulated; LMA and LME handled by hardware; SCE meaninless
	 * outside long mode
	 */
	ignore_bits = EFER_NX | EFER_SCE;
#ifdef CONFIG_X86_64
	ignore_bits |= EFER_LMA | EFER_LME;
	/* SCE is meaningful only in long mode on Intel */
	if (guest_efer & EFER_LMA)
		ignore_bits &= ~(u64)EFER_SCE;
#endif
	if ((guest_efer & ~ignore_bits) == (host_efer & ~ignore_bits))
		return;
550

551 552 553 554
	vmx->host_state.guest_efer_loaded = 1;
	guest_efer &= ~ignore_bits;
	guest_efer |= host_efer & ignore_bits;
	wrmsrl(MSR_EFER, guest_efer);
R
Rusty Russell 已提交
555
	vmx->vcpu.stat.efer_reload++;
556 557
}

558 559 560 561 562 563 564 565
static void reload_host_efer(struct vcpu_vmx *vmx)
{
	if (vmx->host_state.guest_efer_loaded) {
		vmx->host_state.guest_efer_loaded = 0;
		load_msrs(vmx->host_msrs + vmx->msr_offset_efer, 1);
	}
}

566
static void vmx_save_host_state(struct kvm_vcpu *vcpu)
567
{
568 569
	struct vcpu_vmx *vmx = to_vmx(vcpu);

570
	if (vmx->host_state.loaded)
571 572
		return;

573
	vmx->host_state.loaded = 1;
574 575 576 577
	/*
	 * Set host fs and gs selectors.  Unfortunately, 22.2.3 does not
	 * allow segment selectors with cpl > 0 or ti == 1.
	 */
578
	vmx->host_state.ldt_sel = kvm_read_ldt();
579
	vmx->host_state.gs_ldt_reload_needed = vmx->host_state.ldt_sel;
580
	vmx->host_state.fs_sel = kvm_read_fs();
581
	if (!(vmx->host_state.fs_sel & 7)) {
582
		vmcs_write16(HOST_FS_SELECTOR, vmx->host_state.fs_sel);
583 584
		vmx->host_state.fs_reload_needed = 0;
	} else {
585
		vmcs_write16(HOST_FS_SELECTOR, 0);
586
		vmx->host_state.fs_reload_needed = 1;
587
	}
588
	vmx->host_state.gs_sel = kvm_read_gs();
589 590
	if (!(vmx->host_state.gs_sel & 7))
		vmcs_write16(HOST_GS_SELECTOR, vmx->host_state.gs_sel);
591 592
	else {
		vmcs_write16(HOST_GS_SELECTOR, 0);
593
		vmx->host_state.gs_ldt_reload_needed = 1;
594 595 596 597 598 599
	}

#ifdef CONFIG_X86_64
	vmcs_writel(HOST_FS_BASE, read_msr(MSR_FS_BASE));
	vmcs_writel(HOST_GS_BASE, read_msr(MSR_GS_BASE));
#else
600 601
	vmcs_writel(HOST_FS_BASE, segment_base(vmx->host_state.fs_sel));
	vmcs_writel(HOST_GS_BASE, segment_base(vmx->host_state.gs_sel));
602
#endif
603 604

#ifdef CONFIG_X86_64
M
Mike Day 已提交
605
	if (is_long_mode(&vmx->vcpu))
606 607
		save_msrs(vmx->host_msrs +
			  vmx->msr_offset_kernel_gs_base, 1);
M
Mike Day 已提交
608

609
#endif
610
	load_msrs(vmx->guest_msrs, vmx->save_nmsrs);
611
	load_transition_efer(vmx);
612 613
}

614
static void __vmx_load_host_state(struct vcpu_vmx *vmx)
615
{
616
	unsigned long flags;
617

618
	if (!vmx->host_state.loaded)
619 620
		return;

621
	++vmx->vcpu.stat.host_state_reload;
622
	vmx->host_state.loaded = 0;
623
	if (vmx->host_state.fs_reload_needed)
624
		kvm_load_fs(vmx->host_state.fs_sel);
625
	if (vmx->host_state.gs_ldt_reload_needed) {
626
		kvm_load_ldt(vmx->host_state.ldt_sel);
627 628 629 630
		/*
		 * If we have to reload gs, we must take care to
		 * preserve our gs base.
		 */
631
		local_irq_save(flags);
632
		kvm_load_gs(vmx->host_state.gs_sel);
633 634 635
#ifdef CONFIG_X86_64
		wrmsrl(MSR_GS_BASE, vmcs_readl(HOST_GS_BASE));
#endif
636
		local_irq_restore(flags);
637
	}
638
	reload_tss();
639 640
	save_msrs(vmx->guest_msrs, vmx->save_nmsrs);
	load_msrs(vmx->host_msrs, vmx->save_nmsrs);
641
	reload_host_efer(vmx);
642 643
}

644 645 646 647 648 649 650
static void vmx_load_host_state(struct vcpu_vmx *vmx)
{
	preempt_disable();
	__vmx_load_host_state(vmx);
	preempt_enable();
}

A
Avi Kivity 已提交
651 652 653 654
/*
 * Switches to specified vcpu, until a matching vcpu_put(), but assumes
 * vcpu mutex is already taken.
 */
655
static void vmx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
A
Avi Kivity 已提交
656
{
657 658
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	u64 phys_addr = __pa(vmx->vmcs);
659
	u64 tsc_this, delta, new_offset;
A
Avi Kivity 已提交
660

661
	if (vcpu->cpu != cpu) {
R
Rusty Russell 已提交
662
		vcpu_clear(vmx);
M
Marcelo Tosatti 已提交
663
		kvm_migrate_timers(vcpu);
664
		vpid_sync_vcpu_all(vmx);
665 666 667 668
		local_irq_disable();
		list_add(&vmx->local_vcpus_link,
			 &per_cpu(vcpus_on_cpu, cpu));
		local_irq_enable();
669
	}
A
Avi Kivity 已提交
670

671
	if (per_cpu(current_vmcs, cpu) != vmx->vmcs) {
A
Avi Kivity 已提交
672 673
		u8 error;

674
		per_cpu(current_vmcs, cpu) = vmx->vmcs;
675
		asm volatile (__ex(ASM_VMX_VMPTRLD_RAX) "; setna %0"
A
Avi Kivity 已提交
676 677 678 679
			      : "=g"(error) : "a"(&phys_addr), "m"(phys_addr)
			      : "cc");
		if (error)
			printk(KERN_ERR "kvm: vmptrld %p/%llx fail\n",
680
			       vmx->vmcs, phys_addr);
A
Avi Kivity 已提交
681 682 683 684 685 686 687 688 689 690 691
	}

	if (vcpu->cpu != cpu) {
		struct descriptor_table dt;
		unsigned long sysenter_esp;

		vcpu->cpu = cpu;
		/*
		 * Linux uses per-cpu TSS and GDT, so set these when switching
		 * processors.
		 */
692 693
		vmcs_writel(HOST_TR_BASE, kvm_read_tr_base()); /* 22.2.4 */
		kvm_get_gdt(&dt);
A
Avi Kivity 已提交
694 695 696 697
		vmcs_writel(HOST_GDTR_BASE, dt.base);   /* 22.2.4 */

		rdmsrl(MSR_IA32_SYSENTER_ESP, sysenter_esp);
		vmcs_writel(HOST_IA32_SYSENTER_ESP, sysenter_esp); /* 22.2.3 */
698 699 700 701 702

		/*
		 * Make sure the time stamp counter is monotonous.
		 */
		rdtscll(tsc_this);
703 704 705 706 707
		if (tsc_this < vcpu->arch.host_tsc) {
			delta = vcpu->arch.host_tsc - tsc_this;
			new_offset = vmcs_read64(TSC_OFFSET) + delta;
			vmcs_write64(TSC_OFFSET, new_offset);
		}
A
Avi Kivity 已提交
708 709 710 711 712
	}
}

static void vmx_vcpu_put(struct kvm_vcpu *vcpu)
{
713
	__vmx_load_host_state(to_vmx(vcpu));
A
Avi Kivity 已提交
714 715
}

716 717 718 719 720
static void vmx_fpu_activate(struct kvm_vcpu *vcpu)
{
	if (vcpu->fpu_active)
		return;
	vcpu->fpu_active = 1;
721
	vmcs_clear_bits(GUEST_CR0, X86_CR0_TS);
722
	if (vcpu->arch.cr0 & X86_CR0_TS)
723
		vmcs_set_bits(GUEST_CR0, X86_CR0_TS);
724 725 726 727 728 729 730 731
	update_exception_bitmap(vcpu);
}

static void vmx_fpu_deactivate(struct kvm_vcpu *vcpu)
{
	if (!vcpu->fpu_active)
		return;
	vcpu->fpu_active = 0;
732
	vmcs_set_bits(GUEST_CR0, X86_CR0_TS);
733 734 735
	update_exception_bitmap(vcpu);
}

A
Avi Kivity 已提交
736 737 738 739 740 741 742
static unsigned long vmx_get_rflags(struct kvm_vcpu *vcpu)
{
	return vmcs_readl(GUEST_RFLAGS);
}

static void vmx_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
{
743
	if (vcpu->arch.rmode.vm86_active)
744
		rflags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
A
Avi Kivity 已提交
745 746 747
	vmcs_writel(GUEST_RFLAGS, rflags);
}

748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776
static u32 vmx_get_interrupt_shadow(struct kvm_vcpu *vcpu, int mask)
{
	u32 interruptibility = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO);
	int ret = 0;

	if (interruptibility & GUEST_INTR_STATE_STI)
		ret |= X86_SHADOW_INT_STI;
	if (interruptibility & GUEST_INTR_STATE_MOV_SS)
		ret |= X86_SHADOW_INT_MOV_SS;

	return ret & mask;
}

static void vmx_set_interrupt_shadow(struct kvm_vcpu *vcpu, int mask)
{
	u32 interruptibility_old = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO);
	u32 interruptibility = interruptibility_old;

	interruptibility &= ~(GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS);

	if (mask & X86_SHADOW_INT_MOV_SS)
		interruptibility |= GUEST_INTR_STATE_MOV_SS;
	if (mask & X86_SHADOW_INT_STI)
		interruptibility |= GUEST_INTR_STATE_STI;

	if ((interruptibility != interruptibility_old))
		vmcs_write32(GUEST_INTERRUPTIBILITY_INFO, interruptibility);
}

A
Avi Kivity 已提交
777 778 779 780
static void skip_emulated_instruction(struct kvm_vcpu *vcpu)
{
	unsigned long rip;

781
	rip = kvm_rip_read(vcpu);
A
Avi Kivity 已提交
782
	rip += vmcs_read32(VM_EXIT_INSTRUCTION_LEN);
783
	kvm_rip_write(vcpu, rip);
A
Avi Kivity 已提交
784

785 786
	/* skipping an emulated instruction also counts */
	vmx_set_interrupt_shadow(vcpu, 0);
A
Avi Kivity 已提交
787 788
}

789 790 791
static void vmx_queue_exception(struct kvm_vcpu *vcpu, unsigned nr,
				bool has_error_code, u32 error_code)
{
792
	struct vcpu_vmx *vmx = to_vmx(vcpu);
793
	u32 intr_info = nr | INTR_INFO_VALID_MASK;
794

795
	if (has_error_code) {
796
		vmcs_write32(VM_ENTRY_EXCEPTION_ERROR_CODE, error_code);
797 798
		intr_info |= INTR_INFO_DELIVER_CODE_MASK;
	}
799

800
	if (vcpu->arch.rmode.vm86_active) {
801 802 803
		vmx->rmode.irq.pending = true;
		vmx->rmode.irq.vector = nr;
		vmx->rmode.irq.rip = kvm_rip_read(vcpu);
804
		if (nr == BP_VECTOR || nr == OF_VECTOR)
805
			vmx->rmode.irq.rip++;
806 807
		intr_info |= INTR_TYPE_SOFT_INTR;
		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, intr_info);
808 809 810 811 812
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
		kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1);
		return;
	}

813 814 815
	if (kvm_exception_is_soft(nr)) {
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN,
			     vmx->vcpu.arch.event_exit_inst_len);
816 817 818 819 820
		intr_info |= INTR_TYPE_SOFT_EXCEPTION;
	} else
		intr_info |= INTR_TYPE_HARD_EXCEPTION;

	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, intr_info);
821 822
}

823 824 825
/*
 * Swap MSR entry in host/guest MSR entry array.
 */
826
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
827
static void move_msr_up(struct vcpu_vmx *vmx, int from, int to)
828
{
829 830 831 832 833 834 835 836
	struct kvm_msr_entry tmp;

	tmp = vmx->guest_msrs[to];
	vmx->guest_msrs[to] = vmx->guest_msrs[from];
	vmx->guest_msrs[from] = tmp;
	tmp = vmx->host_msrs[to];
	vmx->host_msrs[to] = vmx->host_msrs[from];
	vmx->host_msrs[from] = tmp;
837
}
838
#endif
839

840 841 842 843 844
/*
 * Set up the vmcs to automatically save and restore system
 * msrs.  Don't touch the 64-bit msrs if the guest is in legacy
 * mode, as fiddling with msrs is very expensive.
 */
R
Rusty Russell 已提交
845
static void setup_msrs(struct vcpu_vmx *vmx)
846
{
847
	int save_nmsrs;
848
	unsigned long *msr_bitmap;
849

850
	vmx_load_host_state(vmx);
851 852
	save_nmsrs = 0;
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
853
	if (is_long_mode(&vmx->vcpu)) {
854 855
		int index;

R
Rusty Russell 已提交
856
		index = __find_msr_index(vmx, MSR_SYSCALL_MASK);
857
		if (index >= 0)
R
Rusty Russell 已提交
858 859
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_LSTAR);
860
		if (index >= 0)
R
Rusty Russell 已提交
861 862
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_CSTAR);
863
		if (index >= 0)
R
Rusty Russell 已提交
864 865
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_KERNEL_GS_BASE);
866
		if (index >= 0)
R
Rusty Russell 已提交
867
			move_msr_up(vmx, index, save_nmsrs++);
868 869 870 871
		/*
		 * MSR_K6_STAR is only needed on long mode guests, and only
		 * if efer.sce is enabled.
		 */
R
Rusty Russell 已提交
872
		index = __find_msr_index(vmx, MSR_K6_STAR);
873
		if ((index >= 0) && (vmx->vcpu.arch.shadow_efer & EFER_SCE))
R
Rusty Russell 已提交
874
			move_msr_up(vmx, index, save_nmsrs++);
875 876
	}
#endif
877
	vmx->save_nmsrs = save_nmsrs;
878

879
#ifdef CONFIG_X86_64
880
	vmx->msr_offset_kernel_gs_base =
R
Rusty Russell 已提交
881
		__find_msr_index(vmx, MSR_KERNEL_GS_BASE);
882
#endif
R
Rusty Russell 已提交
883
	vmx->msr_offset_efer = __find_msr_index(vmx, MSR_EFER);
884 885 886 887 888 889 890 891 892

	if (cpu_has_vmx_msr_bitmap()) {
		if (is_long_mode(&vmx->vcpu))
			msr_bitmap = vmx_msr_bitmap_longmode;
		else
			msr_bitmap = vmx_msr_bitmap_legacy;

		vmcs_write64(MSR_BITMAP, __pa(msr_bitmap));
	}
893 894
}

A
Avi Kivity 已提交
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911
/*
 * reads and returns guest's timestamp counter "register"
 * guest_tsc = host_tsc + tsc_offset    -- 21.3
 */
static u64 guest_read_tsc(void)
{
	u64 host_tsc, tsc_offset;

	rdtscll(host_tsc);
	tsc_offset = vmcs_read64(TSC_OFFSET);
	return host_tsc + tsc_offset;
}

/*
 * writes 'guest_tsc' into guest's timestamp counter "register"
 * guest_tsc = host_tsc + tsc_offset ==> tsc_offset = guest_tsc - host_tsc
 */
912
static void guest_write_tsc(u64 guest_tsc, u64 host_tsc)
A
Avi Kivity 已提交
913 914 915 916 917 918 919 920 921 922 923 924
{
	vmcs_write64(TSC_OFFSET, guest_tsc - host_tsc);
}

/*
 * Reads an msr value (of 'msr_index') into 'pdata'.
 * Returns 0 on success, non-0 otherwise.
 * Assumes vcpu_load() was already called.
 */
static int vmx_get_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 *pdata)
{
	u64 data;
925
	struct kvm_msr_entry *msr;
A
Avi Kivity 已提交
926 927 928 929 930 931 932

	if (!pdata) {
		printk(KERN_ERR "BUG: get_msr called with NULL pdata\n");
		return -EINVAL;
	}

	switch (msr_index) {
933
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
934 935 936 937 938 939 940
	case MSR_FS_BASE:
		data = vmcs_readl(GUEST_FS_BASE);
		break;
	case MSR_GS_BASE:
		data = vmcs_readl(GUEST_GS_BASE);
		break;
	case MSR_EFER:
941
		return kvm_get_msr_common(vcpu, msr_index, pdata);
A
Avi Kivity 已提交
942 943 944 945 946 947 948 949
#endif
	case MSR_IA32_TIME_STAMP_COUNTER:
		data = guest_read_tsc();
		break;
	case MSR_IA32_SYSENTER_CS:
		data = vmcs_read32(GUEST_SYSENTER_CS);
		break;
	case MSR_IA32_SYSENTER_EIP:
A
Avi Kivity 已提交
950
		data = vmcs_readl(GUEST_SYSENTER_EIP);
A
Avi Kivity 已提交
951 952
		break;
	case MSR_IA32_SYSENTER_ESP:
A
Avi Kivity 已提交
953
		data = vmcs_readl(GUEST_SYSENTER_ESP);
A
Avi Kivity 已提交
954 955
		break;
	default:
956
		vmx_load_host_state(to_vmx(vcpu));
R
Rusty Russell 已提交
957
		msr = find_msr_entry(to_vmx(vcpu), msr_index);
958 959 960
		if (msr) {
			data = msr->data;
			break;
A
Avi Kivity 已提交
961
		}
962
		return kvm_get_msr_common(vcpu, msr_index, pdata);
A
Avi Kivity 已提交
963 964 965 966 967 968 969 970 971 972 973 974 975
	}

	*pdata = data;
	return 0;
}

/*
 * Writes msr value into into the appropriate "register".
 * Returns 0 on success, non-0 otherwise.
 * Assumes vcpu_load() was already called.
 */
static int vmx_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data)
{
976 977
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	struct kvm_msr_entry *msr;
978
	u64 host_tsc;
979 980
	int ret = 0;

A
Avi Kivity 已提交
981
	switch (msr_index) {
982
	case MSR_EFER:
983
		vmx_load_host_state(vmx);
984 985
		ret = kvm_set_msr_common(vcpu, msr_index, data);
		break;
986
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
987 988 989 990 991 992 993 994 995 996 997
	case MSR_FS_BASE:
		vmcs_writel(GUEST_FS_BASE, data);
		break;
	case MSR_GS_BASE:
		vmcs_writel(GUEST_GS_BASE, data);
		break;
#endif
	case MSR_IA32_SYSENTER_CS:
		vmcs_write32(GUEST_SYSENTER_CS, data);
		break;
	case MSR_IA32_SYSENTER_EIP:
A
Avi Kivity 已提交
998
		vmcs_writel(GUEST_SYSENTER_EIP, data);
A
Avi Kivity 已提交
999 1000
		break;
	case MSR_IA32_SYSENTER_ESP:
A
Avi Kivity 已提交
1001
		vmcs_writel(GUEST_SYSENTER_ESP, data);
A
Avi Kivity 已提交
1002
		break;
A
Avi Kivity 已提交
1003
	case MSR_IA32_TIME_STAMP_COUNTER:
1004 1005
		rdtscll(host_tsc);
		guest_write_tsc(data, host_tsc);
1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
		break;
	case MSR_P6_PERFCTR0:
	case MSR_P6_PERFCTR1:
	case MSR_P6_EVNTSEL0:
	case MSR_P6_EVNTSEL1:
		/*
		 * Just discard all writes to the performance counters; this
		 * should keep both older linux and windows 64-bit guests
		 * happy
		 */
		pr_unimpl(vcpu, "unimplemented perfctr wrmsr: 0x%x data 0x%llx\n", msr_index, data);

A
Avi Kivity 已提交
1018
		break;
S
Sheng Yang 已提交
1019 1020 1021 1022 1023 1024 1025
	case MSR_IA32_CR_PAT:
		if (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_PAT) {
			vmcs_write64(GUEST_IA32_PAT, data);
			vcpu->arch.pat = data;
			break;
		}
		/* Otherwise falls through to kvm_set_msr_common */
A
Avi Kivity 已提交
1026
	default:
1027
		vmx_load_host_state(vmx);
R
Rusty Russell 已提交
1028
		msr = find_msr_entry(vmx, msr_index);
1029 1030 1031
		if (msr) {
			msr->data = data;
			break;
A
Avi Kivity 已提交
1032
		}
1033
		ret = kvm_set_msr_common(vcpu, msr_index, data);
A
Avi Kivity 已提交
1034 1035
	}

1036
	return ret;
A
Avi Kivity 已提交
1037 1038
}

1039
static void vmx_cache_reg(struct kvm_vcpu *vcpu, enum kvm_reg reg)
A
Avi Kivity 已提交
1040
{
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051
	__set_bit(reg, (unsigned long *)&vcpu->arch.regs_avail);
	switch (reg) {
	case VCPU_REGS_RSP:
		vcpu->arch.regs[VCPU_REGS_RSP] = vmcs_readl(GUEST_RSP);
		break;
	case VCPU_REGS_RIP:
		vcpu->arch.regs[VCPU_REGS_RIP] = vmcs_readl(GUEST_RIP);
		break;
	default:
		break;
	}
A
Avi Kivity 已提交
1052 1053
}

J
Jan Kiszka 已提交
1054
static int set_guest_debug(struct kvm_vcpu *vcpu, struct kvm_guest_debug *dbg)
A
Avi Kivity 已提交
1055
{
J
Jan Kiszka 已提交
1056 1057
	int old_debug = vcpu->guest_debug;
	unsigned long flags;
A
Avi Kivity 已提交
1058

J
Jan Kiszka 已提交
1059 1060 1061
	vcpu->guest_debug = dbg->control;
	if (!(vcpu->guest_debug & KVM_GUESTDBG_ENABLE))
		vcpu->guest_debug = 0;
A
Avi Kivity 已提交
1062

1063 1064 1065 1066 1067
	if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)
		vmcs_writel(GUEST_DR7, dbg->arch.debugreg[7]);
	else
		vmcs_writel(GUEST_DR7, vcpu->arch.dr7);

J
Jan Kiszka 已提交
1068 1069 1070 1071
	flags = vmcs_readl(GUEST_RFLAGS);
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
		flags |= X86_EFLAGS_TF | X86_EFLAGS_RF;
	else if (old_debug & KVM_GUESTDBG_SINGLESTEP)
A
Avi Kivity 已提交
1072
		flags &= ~(X86_EFLAGS_TF | X86_EFLAGS_RF);
J
Jan Kiszka 已提交
1073
	vmcs_writel(GUEST_RFLAGS, flags);
A
Avi Kivity 已提交
1074

1075
	update_exception_bitmap(vcpu);
A
Avi Kivity 已提交
1076 1077 1078 1079 1080 1081

	return 0;
}

static __init int cpu_has_kvm_support(void)
{
1082
	return cpu_has_vmx();
A
Avi Kivity 已提交
1083 1084 1085 1086 1087 1088 1089
}

static __init int vmx_disabled_by_bios(void)
{
	u64 msr;

	rdmsrl(MSR_IA32_FEATURE_CONTROL, msr);
1090 1091 1092
	return (msr & (FEATURE_CONTROL_LOCKED |
		       FEATURE_CONTROL_VMXON_ENABLED))
	    == FEATURE_CONTROL_LOCKED;
1093
	/* locked but not enabled */
A
Avi Kivity 已提交
1094 1095
}

A
Avi Kivity 已提交
1096
static void hardware_enable(void *garbage)
A
Avi Kivity 已提交
1097 1098 1099 1100 1101
{
	int cpu = raw_smp_processor_id();
	u64 phys_addr = __pa(per_cpu(vmxarea, cpu));
	u64 old;

1102
	INIT_LIST_HEAD(&per_cpu(vcpus_on_cpu, cpu));
A
Avi Kivity 已提交
1103
	rdmsrl(MSR_IA32_FEATURE_CONTROL, old);
1104 1105 1106 1107
	if ((old & (FEATURE_CONTROL_LOCKED |
		    FEATURE_CONTROL_VMXON_ENABLED))
	    != (FEATURE_CONTROL_LOCKED |
		FEATURE_CONTROL_VMXON_ENABLED))
A
Avi Kivity 已提交
1108
		/* enable and lock */
1109
		wrmsrl(MSR_IA32_FEATURE_CONTROL, old |
1110 1111
		       FEATURE_CONTROL_LOCKED |
		       FEATURE_CONTROL_VMXON_ENABLED);
1112
	write_cr4(read_cr4() | X86_CR4_VMXE); /* FIXME: not cpu hotplug safe */
1113 1114
	asm volatile (ASM_VMX_VMXON_RAX
		      : : "a"(&phys_addr), "m"(phys_addr)
A
Avi Kivity 已提交
1115 1116 1117
		      : "memory", "cc");
}

1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
static void vmclear_local_vcpus(void)
{
	int cpu = raw_smp_processor_id();
	struct vcpu_vmx *vmx, *n;

	list_for_each_entry_safe(vmx, n, &per_cpu(vcpus_on_cpu, cpu),
				 local_vcpus_link)
		__vcpu_clear(vmx);
}

1128 1129 1130 1131 1132

/* Just like cpu_vmxoff(), but with the __kvm_handle_fault_on_reboot()
 * tricks.
 */
static void kvm_cpu_vmxoff(void)
A
Avi Kivity 已提交
1133
{
1134
	asm volatile (__ex(ASM_VMX_VMXOFF) : : : "cc");
1135
	write_cr4(read_cr4() & ~X86_CR4_VMXE);
A
Avi Kivity 已提交
1136 1137
}

1138 1139 1140 1141 1142 1143
static void hardware_disable(void *garbage)
{
	vmclear_local_vcpus();
	kvm_cpu_vmxoff();
}

1144
static __init int adjust_vmx_controls(u32 ctl_min, u32 ctl_opt,
M
Mike Day 已提交
1145
				      u32 msr, u32 *result)
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156
{
	u32 vmx_msr_low, vmx_msr_high;
	u32 ctl = ctl_min | ctl_opt;

	rdmsr(msr, vmx_msr_low, vmx_msr_high);

	ctl &= vmx_msr_high; /* bit == 0 in high word ==> must be zero */
	ctl |= vmx_msr_low;  /* bit == 1 in low word  ==> must be one  */

	/* Ensure minimum (required) set of control bits are supported. */
	if (ctl_min & ~ctl)
Y
Yang, Sheng 已提交
1157
		return -EIO;
1158 1159 1160 1161 1162

	*result = ctl;
	return 0;
}

Y
Yang, Sheng 已提交
1163
static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
A
Avi Kivity 已提交
1164 1165
{
	u32 vmx_msr_low, vmx_msr_high;
S
Sheng Yang 已提交
1166
	u32 min, opt, min2, opt2;
1167 1168
	u32 _pin_based_exec_control = 0;
	u32 _cpu_based_exec_control = 0;
1169
	u32 _cpu_based_2nd_exec_control = 0;
1170 1171 1172 1173
	u32 _vmexit_control = 0;
	u32 _vmentry_control = 0;

	min = PIN_BASED_EXT_INTR_MASK | PIN_BASED_NMI_EXITING;
1174
	opt = PIN_BASED_VIRTUAL_NMIS;
1175 1176
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PINBASED_CTLS,
				&_pin_based_exec_control) < 0)
Y
Yang, Sheng 已提交
1177
		return -EIO;
1178 1179 1180 1181 1182 1183

	min = CPU_BASED_HLT_EXITING |
#ifdef CONFIG_X86_64
	      CPU_BASED_CR8_LOAD_EXITING |
	      CPU_BASED_CR8_STORE_EXITING |
#endif
S
Sheng Yang 已提交
1184 1185
	      CPU_BASED_CR3_LOAD_EXITING |
	      CPU_BASED_CR3_STORE_EXITING |
1186 1187
	      CPU_BASED_USE_IO_BITMAPS |
	      CPU_BASED_MOV_DR_EXITING |
M
Marcelo Tosatti 已提交
1188 1189
	      CPU_BASED_USE_TSC_OFFSETING |
	      CPU_BASED_INVLPG_EXITING;
1190
	opt = CPU_BASED_TPR_SHADOW |
S
Sheng Yang 已提交
1191
	      CPU_BASED_USE_MSR_BITMAPS |
1192
	      CPU_BASED_ACTIVATE_SECONDARY_CONTROLS;
1193 1194
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PROCBASED_CTLS,
				&_cpu_based_exec_control) < 0)
Y
Yang, Sheng 已提交
1195
		return -EIO;
1196 1197 1198 1199 1200
#ifdef CONFIG_X86_64
	if ((_cpu_based_exec_control & CPU_BASED_TPR_SHADOW))
		_cpu_based_exec_control &= ~CPU_BASED_CR8_LOAD_EXITING &
					   ~CPU_BASED_CR8_STORE_EXITING;
#endif
1201
	if (_cpu_based_exec_control & CPU_BASED_ACTIVATE_SECONDARY_CONTROLS) {
S
Sheng Yang 已提交
1202 1203
		min2 = 0;
		opt2 = SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |
1204
			SECONDARY_EXEC_WBINVD_EXITING |
S
Sheng Yang 已提交
1205 1206 1207 1208
			SECONDARY_EXEC_ENABLE_VPID |
			SECONDARY_EXEC_ENABLE_EPT;
		if (adjust_vmx_controls(min2, opt2,
					MSR_IA32_VMX_PROCBASED_CTLS2,
1209 1210 1211 1212 1213 1214 1215 1216
					&_cpu_based_2nd_exec_control) < 0)
			return -EIO;
	}
#ifndef CONFIG_X86_64
	if (!(_cpu_based_2nd_exec_control &
				SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES))
		_cpu_based_exec_control &= ~CPU_BASED_TPR_SHADOW;
#endif
S
Sheng Yang 已提交
1217
	if (_cpu_based_2nd_exec_control & SECONDARY_EXEC_ENABLE_EPT) {
M
Marcelo Tosatti 已提交
1218 1219
		/* CR3 accesses and invlpg don't need to cause VM Exits when EPT
		   enabled */
S
Sheng Yang 已提交
1220
		min &= ~(CPU_BASED_CR3_LOAD_EXITING |
M
Marcelo Tosatti 已提交
1221 1222
			 CPU_BASED_CR3_STORE_EXITING |
			 CPU_BASED_INVLPG_EXITING);
S
Sheng Yang 已提交
1223 1224 1225 1226 1227 1228
		if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PROCBASED_CTLS,
					&_cpu_based_exec_control) < 0)
			return -EIO;
		rdmsr(MSR_IA32_VMX_EPT_VPID_CAP,
		      vmx_capability.ept, vmx_capability.vpid);
	}
1229 1230 1231 1232 1233

	min = 0;
#ifdef CONFIG_X86_64
	min |= VM_EXIT_HOST_ADDR_SPACE_SIZE;
#endif
S
Sheng Yang 已提交
1234
	opt = VM_EXIT_SAVE_IA32_PAT | VM_EXIT_LOAD_IA32_PAT;
1235 1236
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_EXIT_CTLS,
				&_vmexit_control) < 0)
Y
Yang, Sheng 已提交
1237
		return -EIO;
1238

S
Sheng Yang 已提交
1239 1240
	min = 0;
	opt = VM_ENTRY_LOAD_IA32_PAT;
1241 1242
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_ENTRY_CTLS,
				&_vmentry_control) < 0)
Y
Yang, Sheng 已提交
1243
		return -EIO;
A
Avi Kivity 已提交
1244

N
Nguyen Anh Quynh 已提交
1245
	rdmsr(MSR_IA32_VMX_BASIC, vmx_msr_low, vmx_msr_high);
1246 1247 1248

	/* IA-32 SDM Vol 3B: VMCS size is never greater than 4kB. */
	if ((vmx_msr_high & 0x1fff) > PAGE_SIZE)
Y
Yang, Sheng 已提交
1249
		return -EIO;
1250 1251 1252 1253

#ifdef CONFIG_X86_64
	/* IA-32 SDM Vol 3B: 64-bit CPUs always have VMX_BASIC_MSR[48]==0. */
	if (vmx_msr_high & (1u<<16))
Y
Yang, Sheng 已提交
1254
		return -EIO;
1255 1256 1257 1258
#endif

	/* Require Write-Back (WB) memory type for VMCS accesses. */
	if (((vmx_msr_high >> 18) & 15) != 6)
Y
Yang, Sheng 已提交
1259
		return -EIO;
1260

Y
Yang, Sheng 已提交
1261 1262 1263
	vmcs_conf->size = vmx_msr_high & 0x1fff;
	vmcs_conf->order = get_order(vmcs_config.size);
	vmcs_conf->revision_id = vmx_msr_low;
1264

Y
Yang, Sheng 已提交
1265 1266
	vmcs_conf->pin_based_exec_ctrl = _pin_based_exec_control;
	vmcs_conf->cpu_based_exec_ctrl = _cpu_based_exec_control;
1267
	vmcs_conf->cpu_based_2nd_exec_ctrl = _cpu_based_2nd_exec_control;
Y
Yang, Sheng 已提交
1268 1269
	vmcs_conf->vmexit_ctrl         = _vmexit_control;
	vmcs_conf->vmentry_ctrl        = _vmentry_control;
1270 1271

	return 0;
N
Nguyen Anh Quynh 已提交
1272
}
A
Avi Kivity 已提交
1273 1274 1275 1276 1277 1278 1279

static struct vmcs *alloc_vmcs_cpu(int cpu)
{
	int node = cpu_to_node(cpu);
	struct page *pages;
	struct vmcs *vmcs;

1280
	pages = alloc_pages_node(node, GFP_KERNEL, vmcs_config.order);
A
Avi Kivity 已提交
1281 1282 1283
	if (!pages)
		return NULL;
	vmcs = page_address(pages);
1284 1285
	memset(vmcs, 0, vmcs_config.size);
	vmcs->revision_id = vmcs_config.revision_id; /* vmcs revision id */
A
Avi Kivity 已提交
1286 1287 1288 1289 1290
	return vmcs;
}

static struct vmcs *alloc_vmcs(void)
{
1291
	return alloc_vmcs_cpu(raw_smp_processor_id());
A
Avi Kivity 已提交
1292 1293 1294 1295
}

static void free_vmcs(struct vmcs *vmcs)
{
1296
	free_pages((unsigned long)vmcs, vmcs_config.order);
A
Avi Kivity 已提交
1297 1298
}

1299
static void free_kvm_area(void)
A
Avi Kivity 已提交
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
{
	int cpu;

	for_each_online_cpu(cpu)
		free_vmcs(per_cpu(vmxarea, cpu));
}

static __init int alloc_kvm_area(void)
{
	int cpu;

	for_each_online_cpu(cpu) {
		struct vmcs *vmcs;

		vmcs = alloc_vmcs_cpu(cpu);
		if (!vmcs) {
			free_kvm_area();
			return -ENOMEM;
		}

		per_cpu(vmxarea, cpu) = vmcs;
	}
	return 0;
}

static __init int hardware_setup(void)
{
Y
Yang, Sheng 已提交
1327 1328
	if (setup_vmcs_config(&vmcs_config) < 0)
		return -EIO;
1329 1330 1331 1332

	if (boot_cpu_has(X86_FEATURE_NX))
		kvm_enable_efer_bits(EFER_NX);

S
Sheng Yang 已提交
1333 1334 1335 1336 1337 1338 1339 1340 1341
	if (!cpu_has_vmx_vpid())
		enable_vpid = 0;

	if (!cpu_has_vmx_ept())
		enable_ept = 0;

	if (!cpu_has_vmx_flexpriority())
		flexpriority_enabled = 0;

1342 1343 1344
	if (!cpu_has_vmx_tpr_shadow())
		kvm_x86_ops->update_cr8_intercept = NULL;

A
Avi Kivity 已提交
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
	return alloc_kvm_area();
}

static __exit void hardware_unsetup(void)
{
	free_kvm_area();
}

static void fix_pmode_dataseg(int seg, struct kvm_save_segment *save)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];

1357
	if (vmcs_readl(sf->base) == save->base && (save->base & AR_S_MASK)) {
A
Avi Kivity 已提交
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
		vmcs_write16(sf->selector, save->selector);
		vmcs_writel(sf->base, save->base);
		vmcs_write32(sf->limit, save->limit);
		vmcs_write32(sf->ar_bytes, save->ar);
	} else {
		u32 dpl = (vmcs_read16(sf->selector) & SELECTOR_RPL_MASK)
			<< AR_DPL_SHIFT;
		vmcs_write32(sf->ar_bytes, 0x93 | dpl);
	}
}

static void enter_pmode(struct kvm_vcpu *vcpu)
{
	unsigned long flags;
1372
	struct vcpu_vmx *vmx = to_vmx(vcpu);
A
Avi Kivity 已提交
1373

1374
	vmx->emulation_required = 1;
1375
	vcpu->arch.rmode.vm86_active = 0;
A
Avi Kivity 已提交
1376

1377 1378 1379
	vmcs_writel(GUEST_TR_BASE, vcpu->arch.rmode.tr.base);
	vmcs_write32(GUEST_TR_LIMIT, vcpu->arch.rmode.tr.limit);
	vmcs_write32(GUEST_TR_AR_BYTES, vcpu->arch.rmode.tr.ar);
A
Avi Kivity 已提交
1380 1381

	flags = vmcs_readl(GUEST_RFLAGS);
1382
	flags &= ~(X86_EFLAGS_IOPL | X86_EFLAGS_VM);
1383
	flags |= (vcpu->arch.rmode.save_iopl << IOPL_SHIFT);
A
Avi Kivity 已提交
1384 1385
	vmcs_writel(GUEST_RFLAGS, flags);

1386 1387
	vmcs_writel(GUEST_CR4, (vmcs_readl(GUEST_CR4) & ~X86_CR4_VME) |
			(vmcs_readl(CR4_READ_SHADOW) & X86_CR4_VME));
A
Avi Kivity 已提交
1388 1389 1390

	update_exception_bitmap(vcpu);

1391 1392 1393
	if (emulate_invalid_guest_state)
		return;

1394 1395 1396 1397
	fix_pmode_dataseg(VCPU_SREG_ES, &vcpu->arch.rmode.es);
	fix_pmode_dataseg(VCPU_SREG_DS, &vcpu->arch.rmode.ds);
	fix_pmode_dataseg(VCPU_SREG_GS, &vcpu->arch.rmode.gs);
	fix_pmode_dataseg(VCPU_SREG_FS, &vcpu->arch.rmode.fs);
A
Avi Kivity 已提交
1398 1399 1400 1401 1402 1403 1404 1405 1406

	vmcs_write16(GUEST_SS_SELECTOR, 0);
	vmcs_write32(GUEST_SS_AR_BYTES, 0x93);

	vmcs_write16(GUEST_CS_SELECTOR,
		     vmcs_read16(GUEST_CS_SELECTOR) & ~SELECTOR_RPL_MASK);
	vmcs_write32(GUEST_CS_AR_BYTES, 0x9b);
}

M
Mike Day 已提交
1407
static gva_t rmode_tss_base(struct kvm *kvm)
A
Avi Kivity 已提交
1408
{
1409
	if (!kvm->arch.tss_addr) {
1410 1411 1412 1413
		gfn_t base_gfn = kvm->memslots[0].base_gfn +
				 kvm->memslots[0].npages - 3;
		return base_gfn << PAGE_SHIFT;
	}
1414
	return kvm->arch.tss_addr;
A
Avi Kivity 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
}

static void fix_rmode_seg(int seg, struct kvm_save_segment *save)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];

	save->selector = vmcs_read16(sf->selector);
	save->base = vmcs_readl(sf->base);
	save->limit = vmcs_read32(sf->limit);
	save->ar = vmcs_read32(sf->ar_bytes);
1425 1426
	vmcs_write16(sf->selector, save->base >> 4);
	vmcs_write32(sf->base, save->base & 0xfffff);
A
Avi Kivity 已提交
1427 1428 1429 1430 1431 1432 1433
	vmcs_write32(sf->limit, 0xffff);
	vmcs_write32(sf->ar_bytes, 0xf3);
}

static void enter_rmode(struct kvm_vcpu *vcpu)
{
	unsigned long flags;
1434
	struct vcpu_vmx *vmx = to_vmx(vcpu);
A
Avi Kivity 已提交
1435

1436
	vmx->emulation_required = 1;
1437
	vcpu->arch.rmode.vm86_active = 1;
A
Avi Kivity 已提交
1438

1439
	vcpu->arch.rmode.tr.base = vmcs_readl(GUEST_TR_BASE);
A
Avi Kivity 已提交
1440 1441
	vmcs_writel(GUEST_TR_BASE, rmode_tss_base(vcpu->kvm));

1442
	vcpu->arch.rmode.tr.limit = vmcs_read32(GUEST_TR_LIMIT);
A
Avi Kivity 已提交
1443 1444
	vmcs_write32(GUEST_TR_LIMIT, RMODE_TSS_SIZE - 1);

1445
	vcpu->arch.rmode.tr.ar = vmcs_read32(GUEST_TR_AR_BYTES);
A
Avi Kivity 已提交
1446 1447 1448
	vmcs_write32(GUEST_TR_AR_BYTES, 0x008b);

	flags = vmcs_readl(GUEST_RFLAGS);
1449 1450
	vcpu->arch.rmode.save_iopl
		= (flags & X86_EFLAGS_IOPL) >> IOPL_SHIFT;
A
Avi Kivity 已提交
1451

1452
	flags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
A
Avi Kivity 已提交
1453 1454

	vmcs_writel(GUEST_RFLAGS, flags);
1455
	vmcs_writel(GUEST_CR4, vmcs_readl(GUEST_CR4) | X86_CR4_VME);
A
Avi Kivity 已提交
1456 1457
	update_exception_bitmap(vcpu);

1458 1459 1460
	if (emulate_invalid_guest_state)
		goto continue_rmode;

A
Avi Kivity 已提交
1461 1462 1463 1464 1465
	vmcs_write16(GUEST_SS_SELECTOR, vmcs_readl(GUEST_SS_BASE) >> 4);
	vmcs_write32(GUEST_SS_LIMIT, 0xffff);
	vmcs_write32(GUEST_SS_AR_BYTES, 0xf3);

	vmcs_write32(GUEST_CS_AR_BYTES, 0xf3);
1466
	vmcs_write32(GUEST_CS_LIMIT, 0xffff);
1467 1468
	if (vmcs_readl(GUEST_CS_BASE) == 0xffff0000)
		vmcs_writel(GUEST_CS_BASE, 0xf0000);
A
Avi Kivity 已提交
1469 1470
	vmcs_write16(GUEST_CS_SELECTOR, vmcs_readl(GUEST_CS_BASE) >> 4);

1471 1472 1473 1474
	fix_rmode_seg(VCPU_SREG_ES, &vcpu->arch.rmode.es);
	fix_rmode_seg(VCPU_SREG_DS, &vcpu->arch.rmode.ds);
	fix_rmode_seg(VCPU_SREG_GS, &vcpu->arch.rmode.gs);
	fix_rmode_seg(VCPU_SREG_FS, &vcpu->arch.rmode.fs);
1475

1476
continue_rmode:
1477
	kvm_mmu_reset_context(vcpu);
1478
	init_rmode(vcpu->kvm);
A
Avi Kivity 已提交
1479 1480
}

1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503
static void vmx_set_efer(struct kvm_vcpu *vcpu, u64 efer)
{
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	struct kvm_msr_entry *msr = find_msr_entry(vmx, MSR_EFER);

	vcpu->arch.shadow_efer = efer;
	if (!msr)
		return;
	if (efer & EFER_LMA) {
		vmcs_write32(VM_ENTRY_CONTROLS,
			     vmcs_read32(VM_ENTRY_CONTROLS) |
			     VM_ENTRY_IA32E_MODE);
		msr->data = efer;
	} else {
		vmcs_write32(VM_ENTRY_CONTROLS,
			     vmcs_read32(VM_ENTRY_CONTROLS) &
			     ~VM_ENTRY_IA32E_MODE);

		msr->data = efer & ~EFER_LME;
	}
	setup_msrs(vmx);
}

1504
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
1505 1506 1507 1508 1509 1510 1511 1512

static void enter_lmode(struct kvm_vcpu *vcpu)
{
	u32 guest_tr_ar;

	guest_tr_ar = vmcs_read32(GUEST_TR_AR_BYTES);
	if ((guest_tr_ar & AR_TYPE_MASK) != AR_TYPE_BUSY_64_TSS) {
		printk(KERN_DEBUG "%s: tss fixup for long mode. \n",
1513
		       __func__);
A
Avi Kivity 已提交
1514 1515 1516 1517
		vmcs_write32(GUEST_TR_AR_BYTES,
			     (guest_tr_ar & ~AR_TYPE_MASK)
			     | AR_TYPE_BUSY_64_TSS);
	}
1518
	vcpu->arch.shadow_efer |= EFER_LMA;
1519
	vmx_set_efer(vcpu, vcpu->arch.shadow_efer);
A
Avi Kivity 已提交
1520 1521 1522 1523
}

static void exit_lmode(struct kvm_vcpu *vcpu)
{
1524
	vcpu->arch.shadow_efer &= ~EFER_LMA;
A
Avi Kivity 已提交
1525 1526 1527

	vmcs_write32(VM_ENTRY_CONTROLS,
		     vmcs_read32(VM_ENTRY_CONTROLS)
1528
		     & ~VM_ENTRY_IA32E_MODE);
A
Avi Kivity 已提交
1529 1530 1531 1532
}

#endif

1533 1534 1535
static void vmx_flush_tlb(struct kvm_vcpu *vcpu)
{
	vpid_sync_vcpu_all(to_vmx(vcpu));
1536
	if (enable_ept)
1537
		ept_sync_context(construct_eptp(vcpu->arch.mmu.root_hpa));
1538 1539
}

1540
static void vmx_decache_cr4_guest_bits(struct kvm_vcpu *vcpu)
1541
{
1542 1543
	vcpu->arch.cr4 &= KVM_GUEST_CR4_MASK;
	vcpu->arch.cr4 |= vmcs_readl(GUEST_CR4) & ~KVM_GUEST_CR4_MASK;
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
static void ept_load_pdptrs(struct kvm_vcpu *vcpu)
{
	if (is_paging(vcpu) && is_pae(vcpu) && !is_long_mode(vcpu)) {
		if (!load_pdptrs(vcpu, vcpu->arch.cr3)) {
			printk(KERN_ERR "EPT: Fail to load pdptrs!\n");
			return;
		}
		vmcs_write64(GUEST_PDPTR0, vcpu->arch.pdptrs[0]);
		vmcs_write64(GUEST_PDPTR1, vcpu->arch.pdptrs[1]);
		vmcs_write64(GUEST_PDPTR2, vcpu->arch.pdptrs[2]);
		vmcs_write64(GUEST_PDPTR3, vcpu->arch.pdptrs[3]);
	}
}

static void vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4);

static void ept_update_paging_mode_cr0(unsigned long *hw_cr0,
					unsigned long cr0,
					struct kvm_vcpu *vcpu)
{
	if (!(cr0 & X86_CR0_PG)) {
		/* From paging/starting to nonpaging */
		vmcs_write32(CPU_BASED_VM_EXEC_CONTROL,
1569
			     vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) |
1570 1571 1572 1573 1574 1575 1576 1577 1578
			     (CPU_BASED_CR3_LOAD_EXITING |
			      CPU_BASED_CR3_STORE_EXITING));
		vcpu->arch.cr0 = cr0;
		vmx_set_cr4(vcpu, vcpu->arch.cr4);
		*hw_cr0 |= X86_CR0_PE | X86_CR0_PG;
		*hw_cr0 &= ~X86_CR0_WP;
	} else if (!is_paging(vcpu)) {
		/* From nonpaging to paging */
		vmcs_write32(CPU_BASED_VM_EXEC_CONTROL,
1579
			     vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) &
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598
			     ~(CPU_BASED_CR3_LOAD_EXITING |
			       CPU_BASED_CR3_STORE_EXITING));
		vcpu->arch.cr0 = cr0;
		vmx_set_cr4(vcpu, vcpu->arch.cr4);
		if (!(vcpu->arch.cr0 & X86_CR0_WP))
			*hw_cr0 &= ~X86_CR0_WP;
	}
}

static void ept_update_paging_mode_cr4(unsigned long *hw_cr4,
					struct kvm_vcpu *vcpu)
{
	if (!is_paging(vcpu)) {
		*hw_cr4 &= ~X86_CR4_PAE;
		*hw_cr4 |= X86_CR4_PSE;
	} else if (!(vcpu->arch.cr4 & X86_CR4_PAE))
		*hw_cr4 &= ~X86_CR4_PAE;
}

A
Avi Kivity 已提交
1599 1600
static void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
{
1601 1602 1603
	unsigned long hw_cr0 = (cr0 & ~KVM_GUEST_CR0_MASK) |
				KVM_VM_CR0_ALWAYS_ON;

1604 1605
	vmx_fpu_deactivate(vcpu);

1606
	if (vcpu->arch.rmode.vm86_active && (cr0 & X86_CR0_PE))
A
Avi Kivity 已提交
1607 1608
		enter_pmode(vcpu);

1609
	if (!vcpu->arch.rmode.vm86_active && !(cr0 & X86_CR0_PE))
A
Avi Kivity 已提交
1610 1611
		enter_rmode(vcpu);

1612
#ifdef CONFIG_X86_64
1613
	if (vcpu->arch.shadow_efer & EFER_LME) {
1614
		if (!is_paging(vcpu) && (cr0 & X86_CR0_PG))
A
Avi Kivity 已提交
1615
			enter_lmode(vcpu);
1616
		if (is_paging(vcpu) && !(cr0 & X86_CR0_PG))
A
Avi Kivity 已提交
1617 1618 1619 1620
			exit_lmode(vcpu);
	}
#endif

1621
	if (enable_ept)
1622 1623
		ept_update_paging_mode_cr0(&hw_cr0, cr0, vcpu);

A
Avi Kivity 已提交
1624
	vmcs_writel(CR0_READ_SHADOW, cr0);
1625
	vmcs_writel(GUEST_CR0, hw_cr0);
1626
	vcpu->arch.cr0 = cr0;
1627

1628
	if (!(cr0 & X86_CR0_TS) || !(cr0 & X86_CR0_PE))
1629
		vmx_fpu_activate(vcpu);
A
Avi Kivity 已提交
1630 1631
}

1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
static u64 construct_eptp(unsigned long root_hpa)
{
	u64 eptp;

	/* TODO write the value reading from MSR */
	eptp = VMX_EPT_DEFAULT_MT |
		VMX_EPT_DEFAULT_GAW << VMX_EPT_GAW_EPTP_SHIFT;
	eptp |= (root_hpa & PAGE_MASK);

	return eptp;
}

A
Avi Kivity 已提交
1644 1645
static void vmx_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
{
1646 1647 1648 1649
	unsigned long guest_cr3;
	u64 eptp;

	guest_cr3 = cr3;
1650
	if (enable_ept) {
1651 1652 1653 1654 1655 1656 1657 1658
		eptp = construct_eptp(cr3);
		vmcs_write64(EPT_POINTER, eptp);
		ept_sync_context(eptp);
		ept_load_pdptrs(vcpu);
		guest_cr3 = is_paging(vcpu) ? vcpu->arch.cr3 :
			VMX_EPT_IDENTITY_PAGETABLE_ADDR;
	}

1659
	vmx_flush_tlb(vcpu);
1660
	vmcs_writel(GUEST_CR3, guest_cr3);
1661
	if (vcpu->arch.cr0 & X86_CR0_PE)
1662
		vmx_fpu_deactivate(vcpu);
A
Avi Kivity 已提交
1663 1664 1665 1666
}

static void vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
{
1667
	unsigned long hw_cr4 = cr4 | (vcpu->arch.rmode.vm86_active ?
1668 1669
		    KVM_RMODE_VM_CR4_ALWAYS_ON : KVM_PMODE_VM_CR4_ALWAYS_ON);

1670
	vcpu->arch.cr4 = cr4;
1671
	if (enable_ept)
1672 1673 1674 1675
		ept_update_paging_mode_cr4(&hw_cr4, vcpu);

	vmcs_writel(CR4_READ_SHADOW, cr4);
	vmcs_writel(GUEST_CR4, hw_cr4);
A
Avi Kivity 已提交
1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694
}

static u64 vmx_get_segment_base(struct kvm_vcpu *vcpu, int seg)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];

	return vmcs_readl(sf->base);
}

static void vmx_get_segment(struct kvm_vcpu *vcpu,
			    struct kvm_segment *var, int seg)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];
	u32 ar;

	var->base = vmcs_readl(sf->base);
	var->limit = vmcs_read32(sf->limit);
	var->selector = vmcs_read16(sf->selector);
	ar = vmcs_read32(sf->ar_bytes);
1695
	if ((ar & AR_UNUSABLE_MASK) && !emulate_invalid_guest_state)
A
Avi Kivity 已提交
1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707
		ar = 0;
	var->type = ar & 15;
	var->s = (ar >> 4) & 1;
	var->dpl = (ar >> 5) & 3;
	var->present = (ar >> 7) & 1;
	var->avl = (ar >> 12) & 1;
	var->l = (ar >> 13) & 1;
	var->db = (ar >> 14) & 1;
	var->g = (ar >> 15) & 1;
	var->unusable = (ar >> 16) & 1;
}

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721
static int vmx_get_cpl(struct kvm_vcpu *vcpu)
{
	struct kvm_segment kvm_seg;

	if (!(vcpu->arch.cr0 & X86_CR0_PE)) /* if real mode */
		return 0;

	if (vmx_get_rflags(vcpu) & X86_EFLAGS_VM) /* if virtual 8086 */
		return 3;

	vmx_get_segment(vcpu, &kvm_seg, VCPU_SREG_CS);
	return kvm_seg.selector & 3;
}

1722
static u32 vmx_segment_access_rights(struct kvm_segment *var)
A
Avi Kivity 已提交
1723 1724 1725
{
	u32 ar;

1726
	if (var->unusable)
A
Avi Kivity 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
		ar = 1 << 16;
	else {
		ar = var->type & 15;
		ar |= (var->s & 1) << 4;
		ar |= (var->dpl & 3) << 5;
		ar |= (var->present & 1) << 7;
		ar |= (var->avl & 1) << 12;
		ar |= (var->l & 1) << 13;
		ar |= (var->db & 1) << 14;
		ar |= (var->g & 1) << 15;
	}
1738 1739
	if (ar == 0) /* a 0 value means unusable */
		ar = AR_UNUSABLE_MASK;
1740 1741 1742 1743 1744 1745 1746 1747 1748 1749

	return ar;
}

static void vmx_set_segment(struct kvm_vcpu *vcpu,
			    struct kvm_segment *var, int seg)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];
	u32 ar;

1750
	if (vcpu->arch.rmode.vm86_active && seg == VCPU_SREG_TR) {
1751 1752 1753 1754
		vcpu->arch.rmode.tr.selector = var->selector;
		vcpu->arch.rmode.tr.base = var->base;
		vcpu->arch.rmode.tr.limit = var->limit;
		vcpu->arch.rmode.tr.ar = vmx_segment_access_rights(var);
1755 1756 1757 1758 1759
		return;
	}
	vmcs_writel(sf->base, var->base);
	vmcs_write32(sf->limit, var->limit);
	vmcs_write16(sf->selector, var->selector);
1760
	if (vcpu->arch.rmode.vm86_active && var->s) {
1761 1762 1763 1764 1765 1766 1767 1768
		/*
		 * Hack real-mode segments into vm86 compatibility.
		 */
		if (var->base == 0xffff0000 && var->selector == 0xf000)
			vmcs_writel(sf->base, 0xf0000);
		ar = 0xf3;
	} else
		ar = vmx_segment_access_rights(var);
A
Avi Kivity 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
	vmcs_write32(sf->ar_bytes, ar);
}

static void vmx_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l)
{
	u32 ar = vmcs_read32(GUEST_CS_AR_BYTES);

	*db = (ar >> 14) & 1;
	*l = (ar >> 13) & 1;
}

static void vmx_get_idt(struct kvm_vcpu *vcpu, struct descriptor_table *dt)
{
	dt->limit = vmcs_read32(GUEST_IDTR_LIMIT);
	dt->base = vmcs_readl(GUEST_IDTR_BASE);
}

static void vmx_set_idt(struct kvm_vcpu *vcpu, struct descriptor_table *dt)
{
	vmcs_write32(GUEST_IDTR_LIMIT, dt->limit);
	vmcs_writel(GUEST_IDTR_BASE, dt->base);
}

static void vmx_get_gdt(struct kvm_vcpu *vcpu, struct descriptor_table *dt)
{
	dt->limit = vmcs_read32(GUEST_GDTR_LIMIT);
	dt->base = vmcs_readl(GUEST_GDTR_BASE);
}

static void vmx_set_gdt(struct kvm_vcpu *vcpu, struct descriptor_table *dt)
{
	vmcs_write32(GUEST_GDTR_LIMIT, dt->limit);
	vmcs_writel(GUEST_GDTR_BASE, dt->base);
}

1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829
static bool rmode_segment_valid(struct kvm_vcpu *vcpu, int seg)
{
	struct kvm_segment var;
	u32 ar;

	vmx_get_segment(vcpu, &var, seg);
	ar = vmx_segment_access_rights(&var);

	if (var.base != (var.selector << 4))
		return false;
	if (var.limit != 0xffff)
		return false;
	if (ar != 0xf3)
		return false;

	return true;
}

static bool code_segment_valid(struct kvm_vcpu *vcpu)
{
	struct kvm_segment cs;
	unsigned int cs_rpl;

	vmx_get_segment(vcpu, &cs, VCPU_SREG_CS);
	cs_rpl = cs.selector & SELECTOR_RPL_MASK;

1830 1831
	if (cs.unusable)
		return false;
1832 1833 1834 1835
	if (~cs.type & (AR_TYPE_CODE_MASK|AR_TYPE_ACCESSES_MASK))
		return false;
	if (!cs.s)
		return false;
1836
	if (cs.type & AR_TYPE_WRITEABLE_MASK) {
1837 1838
		if (cs.dpl > cs_rpl)
			return false;
1839
	} else {
1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857
		if (cs.dpl != cs_rpl)
			return false;
	}
	if (!cs.present)
		return false;

	/* TODO: Add Reserved field check, this'll require a new member in the kvm_segment_field structure */
	return true;
}

static bool stack_segment_valid(struct kvm_vcpu *vcpu)
{
	struct kvm_segment ss;
	unsigned int ss_rpl;

	vmx_get_segment(vcpu, &ss, VCPU_SREG_SS);
	ss_rpl = ss.selector & SELECTOR_RPL_MASK;

1858 1859 1860
	if (ss.unusable)
		return true;
	if (ss.type != 3 && ss.type != 7)
1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
		return false;
	if (!ss.s)
		return false;
	if (ss.dpl != ss_rpl) /* DPL != RPL */
		return false;
	if (!ss.present)
		return false;

	return true;
}

static bool data_segment_valid(struct kvm_vcpu *vcpu, int seg)
{
	struct kvm_segment var;
	unsigned int rpl;

	vmx_get_segment(vcpu, &var, seg);
	rpl = var.selector & SELECTOR_RPL_MASK;

1880 1881
	if (var.unusable)
		return true;
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902
	if (!var.s)
		return false;
	if (!var.present)
		return false;
	if (~var.type & (AR_TYPE_CODE_MASK|AR_TYPE_WRITEABLE_MASK)) {
		if (var.dpl < rpl) /* DPL < RPL */
			return false;
	}

	/* TODO: Add other members to kvm_segment_field to allow checking for other access
	 * rights flags
	 */
	return true;
}

static bool tr_valid(struct kvm_vcpu *vcpu)
{
	struct kvm_segment tr;

	vmx_get_segment(vcpu, &tr, VCPU_SREG_TR);

1903 1904
	if (tr.unusable)
		return false;
1905 1906
	if (tr.selector & SELECTOR_TI_MASK)	/* TI = 1 */
		return false;
1907
	if (tr.type != 3 && tr.type != 11) /* TODO: Check if guest is in IA32e mode */
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920
		return false;
	if (!tr.present)
		return false;

	return true;
}

static bool ldtr_valid(struct kvm_vcpu *vcpu)
{
	struct kvm_segment ldtr;

	vmx_get_segment(vcpu, &ldtr, VCPU_SREG_LDTR);

1921 1922
	if (ldtr.unusable)
		return true;
1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993
	if (ldtr.selector & SELECTOR_TI_MASK)	/* TI = 1 */
		return false;
	if (ldtr.type != 2)
		return false;
	if (!ldtr.present)
		return false;

	return true;
}

static bool cs_ss_rpl_check(struct kvm_vcpu *vcpu)
{
	struct kvm_segment cs, ss;

	vmx_get_segment(vcpu, &cs, VCPU_SREG_CS);
	vmx_get_segment(vcpu, &ss, VCPU_SREG_SS);

	return ((cs.selector & SELECTOR_RPL_MASK) ==
		 (ss.selector & SELECTOR_RPL_MASK));
}

/*
 * Check if guest state is valid. Returns true if valid, false if
 * not.
 * We assume that registers are always usable
 */
static bool guest_state_valid(struct kvm_vcpu *vcpu)
{
	/* real mode guest state checks */
	if (!(vcpu->arch.cr0 & X86_CR0_PE)) {
		if (!rmode_segment_valid(vcpu, VCPU_SREG_CS))
			return false;
		if (!rmode_segment_valid(vcpu, VCPU_SREG_SS))
			return false;
		if (!rmode_segment_valid(vcpu, VCPU_SREG_DS))
			return false;
		if (!rmode_segment_valid(vcpu, VCPU_SREG_ES))
			return false;
		if (!rmode_segment_valid(vcpu, VCPU_SREG_FS))
			return false;
		if (!rmode_segment_valid(vcpu, VCPU_SREG_GS))
			return false;
	} else {
	/* protected mode guest state checks */
		if (!cs_ss_rpl_check(vcpu))
			return false;
		if (!code_segment_valid(vcpu))
			return false;
		if (!stack_segment_valid(vcpu))
			return false;
		if (!data_segment_valid(vcpu, VCPU_SREG_DS))
			return false;
		if (!data_segment_valid(vcpu, VCPU_SREG_ES))
			return false;
		if (!data_segment_valid(vcpu, VCPU_SREG_FS))
			return false;
		if (!data_segment_valid(vcpu, VCPU_SREG_GS))
			return false;
		if (!tr_valid(vcpu))
			return false;
		if (!ldtr_valid(vcpu))
			return false;
	}
	/* TODO:
	 * - Add checks on RIP
	 * - Add checks on RFLAGS
	 */

	return true;
}

M
Mike Day 已提交
1994
static int init_rmode_tss(struct kvm *kvm)
A
Avi Kivity 已提交
1995 1996
{
	gfn_t fn = rmode_tss_base(kvm) >> PAGE_SHIFT;
1997
	u16 data = 0;
1998
	int ret = 0;
1999
	int r;
A
Avi Kivity 已提交
2000

2001 2002
	r = kvm_clear_guest_page(kvm, fn, 0, PAGE_SIZE);
	if (r < 0)
2003
		goto out;
2004
	data = TSS_BASE_SIZE + TSS_REDIRECTION_SIZE;
2005 2006
	r = kvm_write_guest_page(kvm, fn++, &data,
			TSS_IOPB_BASE_OFFSET, sizeof(u16));
2007
	if (r < 0)
2008
		goto out;
2009 2010
	r = kvm_clear_guest_page(kvm, fn++, 0, PAGE_SIZE);
	if (r < 0)
2011
		goto out;
2012 2013
	r = kvm_clear_guest_page(kvm, fn, 0, PAGE_SIZE);
	if (r < 0)
2014
		goto out;
2015
	data = ~0;
2016 2017 2018
	r = kvm_write_guest_page(kvm, fn, &data,
				 RMODE_TSS_SIZE - 2 * PAGE_SIZE - 1,
				 sizeof(u8));
2019
	if (r < 0)
2020 2021 2022 2023 2024
		goto out;

	ret = 1;
out:
	return ret;
A
Avi Kivity 已提交
2025 2026
}

2027 2028 2029 2030 2031 2032
static int init_rmode_identity_map(struct kvm *kvm)
{
	int i, r, ret;
	pfn_t identity_map_pfn;
	u32 tmp;

2033
	if (!enable_ept)
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
		return 1;
	if (unlikely(!kvm->arch.ept_identity_pagetable)) {
		printk(KERN_ERR "EPT: identity-mapping pagetable "
			"haven't been allocated!\n");
		return 0;
	}
	if (likely(kvm->arch.ept_identity_pagetable_done))
		return 1;
	ret = 0;
	identity_map_pfn = VMX_EPT_IDENTITY_PAGETABLE_ADDR >> PAGE_SHIFT;
	r = kvm_clear_guest_page(kvm, identity_map_pfn, 0, PAGE_SIZE);
	if (r < 0)
		goto out;
	/* Set up identity-mapping pagetable for EPT in real mode */
	for (i = 0; i < PT32_ENT_PER_PAGE; i++) {
		tmp = (i << 22) + (_PAGE_PRESENT | _PAGE_RW | _PAGE_USER |
			_PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_PSE);
		r = kvm_write_guest_page(kvm, identity_map_pfn,
				&tmp, i * sizeof(tmp), sizeof(tmp));
		if (r < 0)
			goto out;
	}
	kvm->arch.ept_identity_pagetable_done = true;
	ret = 1;
out:
	return ret;
}

A
Avi Kivity 已提交
2062 2063 2064 2065 2066 2067 2068
static void seg_setup(int seg)
{
	struct kvm_vmx_segment_field *sf = &kvm_vmx_segment_fields[seg];

	vmcs_write16(sf->selector, 0);
	vmcs_writel(sf->base, 0);
	vmcs_write32(sf->limit, 0xffff);
2069
	vmcs_write32(sf->ar_bytes, 0xf3);
A
Avi Kivity 已提交
2070 2071
}

2072 2073 2074 2075 2076
static int alloc_apic_access_page(struct kvm *kvm)
{
	struct kvm_userspace_memory_region kvm_userspace_mem;
	int r = 0;

2077
	down_write(&kvm->slots_lock);
2078
	if (kvm->arch.apic_access_page)
2079 2080 2081 2082 2083 2084 2085 2086
		goto out;
	kvm_userspace_mem.slot = APIC_ACCESS_PAGE_PRIVATE_MEMSLOT;
	kvm_userspace_mem.flags = 0;
	kvm_userspace_mem.guest_phys_addr = 0xfee00000ULL;
	kvm_userspace_mem.memory_size = PAGE_SIZE;
	r = __kvm_set_memory_region(kvm, &kvm_userspace_mem, 0);
	if (r)
		goto out;
2087

2088
	kvm->arch.apic_access_page = gfn_to_page(kvm, 0xfee00);
2089
out:
2090
	up_write(&kvm->slots_lock);
2091 2092 2093
	return r;
}

2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
static int alloc_identity_pagetable(struct kvm *kvm)
{
	struct kvm_userspace_memory_region kvm_userspace_mem;
	int r = 0;

	down_write(&kvm->slots_lock);
	if (kvm->arch.ept_identity_pagetable)
		goto out;
	kvm_userspace_mem.slot = IDENTITY_PAGETABLE_PRIVATE_MEMSLOT;
	kvm_userspace_mem.flags = 0;
	kvm_userspace_mem.guest_phys_addr = VMX_EPT_IDENTITY_PAGETABLE_ADDR;
	kvm_userspace_mem.memory_size = PAGE_SIZE;
	r = __kvm_set_memory_region(kvm, &kvm_userspace_mem, 0);
	if (r)
		goto out;

	kvm->arch.ept_identity_pagetable = gfn_to_page(kvm,
			VMX_EPT_IDENTITY_PAGETABLE_ADDR >> PAGE_SHIFT);
out:
	up_write(&kvm->slots_lock);
	return r;
}

2117 2118 2119 2120 2121
static void allocate_vpid(struct vcpu_vmx *vmx)
{
	int vpid;

	vmx->vpid = 0;
2122
	if (!enable_vpid)
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132
		return;
	spin_lock(&vmx_vpid_lock);
	vpid = find_first_zero_bit(vmx_vpid_bitmap, VMX_NR_VPIDS);
	if (vpid < VMX_NR_VPIDS) {
		vmx->vpid = vpid;
		__set_bit(vpid, vmx_vpid_bitmap);
	}
	spin_unlock(&vmx_vpid_lock);
}

2133
static void __vmx_disable_intercept_for_msr(unsigned long *msr_bitmap, u32 msr)
S
Sheng Yang 已提交
2134
{
2135
	int f = sizeof(unsigned long);
S
Sheng Yang 已提交
2136 2137 2138 2139 2140 2141 2142 2143 2144 2145

	if (!cpu_has_vmx_msr_bitmap())
		return;

	/*
	 * See Intel PRM Vol. 3, 20.6.9 (MSR-Bitmap Address). Early manuals
	 * have the write-low and read-high bitmap offsets the wrong way round.
	 * We can control MSRs 0x00000000-0x00001fff and 0xc0000000-0xc0001fff.
	 */
	if (msr <= 0x1fff) {
2146 2147
		__clear_bit(msr, msr_bitmap + 0x000 / f); /* read-low */
		__clear_bit(msr, msr_bitmap + 0x800 / f); /* write-low */
S
Sheng Yang 已提交
2148 2149
	} else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff)) {
		msr &= 0x1fff;
2150 2151
		__clear_bit(msr, msr_bitmap + 0x400 / f); /* read-high */
		__clear_bit(msr, msr_bitmap + 0xc00 / f); /* write-high */
S
Sheng Yang 已提交
2152 2153 2154
	}
}

2155 2156 2157 2158 2159 2160 2161
static void vmx_disable_intercept_for_msr(u32 msr, bool longmode_only)
{
	if (!longmode_only)
		__vmx_disable_intercept_for_msr(vmx_msr_bitmap_legacy, msr);
	__vmx_disable_intercept_for_msr(vmx_msr_bitmap_longmode, msr);
}

A
Avi Kivity 已提交
2162 2163 2164
/*
 * Sets up the vmcs for emulated real mode.
 */
R
Rusty Russell 已提交
2165
static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
A
Avi Kivity 已提交
2166
{
S
Sheng Yang 已提交
2167
	u32 host_sysenter_cs, msr_low, msr_high;
A
Avi Kivity 已提交
2168
	u32 junk;
2169
	u64 host_pat, tsc_this, tsc_base;
A
Avi Kivity 已提交
2170 2171 2172
	unsigned long a;
	struct descriptor_table dt;
	int i;
2173
	unsigned long kvm_vmx_return;
2174
	u32 exec_control;
A
Avi Kivity 已提交
2175 2176

	/* I/O */
2177 2178
	vmcs_write64(IO_BITMAP_A, __pa(vmx_io_bitmap_a));
	vmcs_write64(IO_BITMAP_B, __pa(vmx_io_bitmap_b));
A
Avi Kivity 已提交
2179

S
Sheng Yang 已提交
2180
	if (cpu_has_vmx_msr_bitmap())
2181
		vmcs_write64(MSR_BITMAP, __pa(vmx_msr_bitmap_legacy));
S
Sheng Yang 已提交
2182

A
Avi Kivity 已提交
2183 2184 2185
	vmcs_write64(VMCS_LINK_POINTER, -1ull); /* 22.3.1.5 */

	/* Control */
2186 2187
	vmcs_write32(PIN_BASED_VM_EXEC_CONTROL,
		vmcs_config.pin_based_exec_ctrl);
2188 2189 2190 2191 2192 2193 2194 2195 2196

	exec_control = vmcs_config.cpu_based_exec_ctrl;
	if (!vm_need_tpr_shadow(vmx->vcpu.kvm)) {
		exec_control &= ~CPU_BASED_TPR_SHADOW;
#ifdef CONFIG_X86_64
		exec_control |= CPU_BASED_CR8_STORE_EXITING |
				CPU_BASED_CR8_LOAD_EXITING;
#endif
	}
2197
	if (!enable_ept)
S
Sheng Yang 已提交
2198
		exec_control |= CPU_BASED_CR3_STORE_EXITING |
2199 2200
				CPU_BASED_CR3_LOAD_EXITING  |
				CPU_BASED_INVLPG_EXITING;
2201
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, exec_control);
A
Avi Kivity 已提交
2202

2203 2204 2205 2206 2207
	if (cpu_has_secondary_exec_ctrls()) {
		exec_control = vmcs_config.cpu_based_2nd_exec_ctrl;
		if (!vm_need_virtualize_apic_accesses(vmx->vcpu.kvm))
			exec_control &=
				~SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES;
2208 2209
		if (vmx->vpid == 0)
			exec_control &= ~SECONDARY_EXEC_ENABLE_VPID;
2210
		if (!enable_ept)
S
Sheng Yang 已提交
2211
			exec_control &= ~SECONDARY_EXEC_ENABLE_EPT;
2212 2213
		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
	}
2214

2215 2216
	vmcs_write32(PAGE_FAULT_ERROR_CODE_MASK, !!bypass_guest_pf);
	vmcs_write32(PAGE_FAULT_ERROR_CODE_MATCH, !!bypass_guest_pf);
A
Avi Kivity 已提交
2217 2218 2219 2220 2221 2222 2223 2224 2225
	vmcs_write32(CR3_TARGET_COUNT, 0);           /* 22.2.1 */

	vmcs_writel(HOST_CR0, read_cr0());  /* 22.2.3 */
	vmcs_writel(HOST_CR4, read_cr4());  /* 22.2.3, 22.2.5 */
	vmcs_writel(HOST_CR3, read_cr3());  /* 22.2.3  FIXME: shadow tables */

	vmcs_write16(HOST_CS_SELECTOR, __KERNEL_CS);  /* 22.2.4 */
	vmcs_write16(HOST_DS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
	vmcs_write16(HOST_ES_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
2226 2227
	vmcs_write16(HOST_FS_SELECTOR, kvm_read_fs());    /* 22.2.4 */
	vmcs_write16(HOST_GS_SELECTOR, kvm_read_gs());    /* 22.2.4 */
A
Avi Kivity 已提交
2228
	vmcs_write16(HOST_SS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
2229
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
	rdmsrl(MSR_FS_BASE, a);
	vmcs_writel(HOST_FS_BASE, a); /* 22.2.4 */
	rdmsrl(MSR_GS_BASE, a);
	vmcs_writel(HOST_GS_BASE, a); /* 22.2.4 */
#else
	vmcs_writel(HOST_FS_BASE, 0); /* 22.2.4 */
	vmcs_writel(HOST_GS_BASE, 0); /* 22.2.4 */
#endif

	vmcs_write16(HOST_TR_SELECTOR, GDT_ENTRY_TSS*8);  /* 22.2.4 */

2241
	kvm_get_idt(&dt);
A
Avi Kivity 已提交
2242 2243
	vmcs_writel(HOST_IDTR_BASE, dt.base);   /* 22.2.4 */

M
Mike Day 已提交
2244
	asm("mov $.Lkvm_vmx_return, %0" : "=r"(kvm_vmx_return));
2245
	vmcs_writel(HOST_RIP, kvm_vmx_return); /* 22.2.5 */
2246 2247 2248
	vmcs_write32(VM_EXIT_MSR_STORE_COUNT, 0);
	vmcs_write32(VM_EXIT_MSR_LOAD_COUNT, 0);
	vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, 0);
A
Avi Kivity 已提交
2249 2250 2251 2252 2253 2254 2255 2256

	rdmsr(MSR_IA32_SYSENTER_CS, host_sysenter_cs, junk);
	vmcs_write32(HOST_IA32_SYSENTER_CS, host_sysenter_cs);
	rdmsrl(MSR_IA32_SYSENTER_ESP, a);
	vmcs_writel(HOST_IA32_SYSENTER_ESP, a);   /* 22.2.3 */
	rdmsrl(MSR_IA32_SYSENTER_EIP, a);
	vmcs_writel(HOST_IA32_SYSENTER_EIP, a);   /* 22.2.3 */

S
Sheng Yang 已提交
2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270
	if (vmcs_config.vmexit_ctrl & VM_EXIT_LOAD_IA32_PAT) {
		rdmsr(MSR_IA32_CR_PAT, msr_low, msr_high);
		host_pat = msr_low | ((u64) msr_high << 32);
		vmcs_write64(HOST_IA32_PAT, host_pat);
	}
	if (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_PAT) {
		rdmsr(MSR_IA32_CR_PAT, msr_low, msr_high);
		host_pat = msr_low | ((u64) msr_high << 32);
		/* Write the default value follow host pat */
		vmcs_write64(GUEST_IA32_PAT, host_pat);
		/* Keep arch.pat sync with GUEST_IA32_PAT */
		vmx->vcpu.arch.pat = host_pat;
	}

A
Avi Kivity 已提交
2271 2272 2273 2274
	for (i = 0; i < NR_VMX_MSR; ++i) {
		u32 index = vmx_msr_index[i];
		u32 data_low, data_high;
		u64 data;
2275
		int j = vmx->nmsrs;
A
Avi Kivity 已提交
2276 2277 2278

		if (rdmsr_safe(index, &data_low, &data_high) < 0)
			continue;
2279 2280
		if (wrmsr_safe(index, data_low, data_high) < 0)
			continue;
A
Avi Kivity 已提交
2281
		data = data_low | ((u64)data_high << 32);
2282 2283 2284 2285 2286
		vmx->host_msrs[j].index = index;
		vmx->host_msrs[j].reserved = 0;
		vmx->host_msrs[j].data = data;
		vmx->guest_msrs[j] = vmx->host_msrs[j];
		++vmx->nmsrs;
A
Avi Kivity 已提交
2287 2288
	}

2289
	vmcs_write32(VM_EXIT_CONTROLS, vmcs_config.vmexit_ctrl);
A
Avi Kivity 已提交
2290 2291

	/* 22.2.1, 20.8.1 */
2292 2293
	vmcs_write32(VM_ENTRY_CONTROLS, vmcs_config.vmentry_ctrl);

2294 2295 2296
	vmcs_writel(CR0_GUEST_HOST_MASK, ~0UL);
	vmcs_writel(CR4_GUEST_HOST_MASK, KVM_GUEST_CR4_MASK);

2297 2298 2299 2300 2301 2302
	tsc_base = vmx->vcpu.kvm->arch.vm_init_tsc;
	rdtscll(tsc_this);
	if (tsc_this < vmx->vcpu.kvm->arch.vm_init_tsc)
		tsc_base = tsc_this;

	guest_write_tsc(0, tsc_base);
2303

2304 2305 2306
	return 0;
}

2307 2308 2309 2310 2311 2312 2313 2314 2315
static int init_rmode(struct kvm *kvm)
{
	if (!init_rmode_tss(kvm))
		return 0;
	if (!init_rmode_identity_map(kvm))
		return 0;
	return 1;
}

2316 2317 2318 2319 2320 2321
static int vmx_vcpu_reset(struct kvm_vcpu *vcpu)
{
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	u64 msr;
	int ret;

2322
	vcpu->arch.regs_avail = ~((1 << VCPU_REGS_RIP) | (1 << VCPU_REGS_RSP));
2323
	down_read(&vcpu->kvm->slots_lock);
2324
	if (!init_rmode(vmx->vcpu.kvm)) {
2325 2326 2327 2328
		ret = -ENOMEM;
		goto out;
	}

2329
	vmx->vcpu.arch.rmode.vm86_active = 0;
2330

2331 2332
	vmx->soft_vnmi_blocked = 0;

2333
	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
2334
	kvm_set_cr8(&vmx->vcpu, 0);
2335 2336 2337 2338 2339 2340 2341
	msr = 0xfee00000 | MSR_IA32_APICBASE_ENABLE;
	if (vmx->vcpu.vcpu_id == 0)
		msr |= MSR_IA32_APICBASE_BSP;
	kvm_set_apic_base(&vmx->vcpu, msr);

	fx_init(&vmx->vcpu);

2342
	seg_setup(VCPU_SREG_CS);
2343 2344 2345 2346 2347 2348 2349 2350
	/*
	 * GUEST_CS_BASE should really be 0xffff0000, but VT vm86 mode
	 * insists on having GUEST_CS_BASE == GUEST_CS_SELECTOR << 4.  Sigh.
	 */
	if (vmx->vcpu.vcpu_id == 0) {
		vmcs_write16(GUEST_CS_SELECTOR, 0xf000);
		vmcs_writel(GUEST_CS_BASE, 0x000f0000);
	} else {
2351 2352
		vmcs_write16(GUEST_CS_SELECTOR, vmx->vcpu.arch.sipi_vector << 8);
		vmcs_writel(GUEST_CS_BASE, vmx->vcpu.arch.sipi_vector << 12);
2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376
	}

	seg_setup(VCPU_SREG_DS);
	seg_setup(VCPU_SREG_ES);
	seg_setup(VCPU_SREG_FS);
	seg_setup(VCPU_SREG_GS);
	seg_setup(VCPU_SREG_SS);

	vmcs_write16(GUEST_TR_SELECTOR, 0);
	vmcs_writel(GUEST_TR_BASE, 0);
	vmcs_write32(GUEST_TR_LIMIT, 0xffff);
	vmcs_write32(GUEST_TR_AR_BYTES, 0x008b);

	vmcs_write16(GUEST_LDTR_SELECTOR, 0);
	vmcs_writel(GUEST_LDTR_BASE, 0);
	vmcs_write32(GUEST_LDTR_LIMIT, 0xffff);
	vmcs_write32(GUEST_LDTR_AR_BYTES, 0x00082);

	vmcs_write32(GUEST_SYSENTER_CS, 0);
	vmcs_writel(GUEST_SYSENTER_ESP, 0);
	vmcs_writel(GUEST_SYSENTER_EIP, 0);

	vmcs_writel(GUEST_RFLAGS, 0x02);
	if (vmx->vcpu.vcpu_id == 0)
2377
		kvm_rip_write(vcpu, 0xfff0);
2378
	else
2379 2380
		kvm_rip_write(vcpu, 0);
	kvm_register_write(vcpu, VCPU_REGS_RSP, 0);
2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398

	vmcs_writel(GUEST_DR7, 0x400);

	vmcs_writel(GUEST_GDTR_BASE, 0);
	vmcs_write32(GUEST_GDTR_LIMIT, 0xffff);

	vmcs_writel(GUEST_IDTR_BASE, 0);
	vmcs_write32(GUEST_IDTR_LIMIT, 0xffff);

	vmcs_write32(GUEST_ACTIVITY_STATE, 0);
	vmcs_write32(GUEST_INTERRUPTIBILITY_INFO, 0);
	vmcs_write32(GUEST_PENDING_DBG_EXCEPTIONS, 0);

	/* Special registers */
	vmcs_write64(GUEST_IA32_DEBUGCTL, 0);

	setup_msrs(vmx);

A
Avi Kivity 已提交
2399 2400
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, 0);  /* 22.2.1 */

2401 2402 2403 2404
	if (cpu_has_vmx_tpr_shadow()) {
		vmcs_write64(VIRTUAL_APIC_PAGE_ADDR, 0);
		if (vm_need_tpr_shadow(vmx->vcpu.kvm))
			vmcs_write64(VIRTUAL_APIC_PAGE_ADDR,
2405
				page_to_phys(vmx->vcpu.arch.apic->regs_page));
2406 2407 2408 2409 2410
		vmcs_write32(TPR_THRESHOLD, 0);
	}

	if (vm_need_virtualize_apic_accesses(vmx->vcpu.kvm))
		vmcs_write64(APIC_ACCESS_ADDR,
2411
			     page_to_phys(vmx->vcpu.kvm->arch.apic_access_page));
A
Avi Kivity 已提交
2412

2413 2414 2415
	if (vmx->vpid != 0)
		vmcs_write16(VIRTUAL_PROCESSOR_ID, vmx->vpid);

2416 2417
	vmx->vcpu.arch.cr0 = 0x60000010;
	vmx_set_cr0(&vmx->vcpu, vmx->vcpu.arch.cr0); /* enter rmode */
R
Rusty Russell 已提交
2418 2419 2420 2421
	vmx_set_cr4(&vmx->vcpu, 0);
	vmx_set_efer(&vmx->vcpu, 0);
	vmx_fpu_activate(&vmx->vcpu);
	update_exception_bitmap(&vmx->vcpu);
A
Avi Kivity 已提交
2422

2423 2424
	vpid_sync_vcpu_all(vmx);

2425
	ret = 0;
A
Avi Kivity 已提交
2426

2427 2428 2429
	/* HACK: Don't enable emulation on guest boot/reset */
	vmx->emulation_required = 0;

A
Avi Kivity 已提交
2430
out:
2431
	up_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
2432 2433 2434
	return ret;
}

2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457
static void enable_irq_window(struct kvm_vcpu *vcpu)
{
	u32 cpu_based_vm_exec_control;

	cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL);
	cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING;
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control);
}

static void enable_nmi_window(struct kvm_vcpu *vcpu)
{
	u32 cpu_based_vm_exec_control;

	if (!cpu_has_virtual_nmis()) {
		enable_irq_window(vcpu);
		return;
	}

	cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL);
	cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_NMI_PENDING;
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control);
}

2458
static void vmx_inject_irq(struct kvm_vcpu *vcpu)
2459
{
2460
	struct vcpu_vmx *vmx = to_vmx(vcpu);
2461 2462
	uint32_t intr;
	int irq = vcpu->arch.interrupt.nr;
2463

F
Feng (Eric) Liu 已提交
2464 2465
	KVMTRACE_1D(INJ_VIRQ, vcpu, (u32)irq, handler);

2466
	++vcpu->stat.irq_injections;
2467
	if (vcpu->arch.rmode.vm86_active) {
2468 2469
		vmx->rmode.irq.pending = true;
		vmx->rmode.irq.vector = irq;
2470
		vmx->rmode.irq.rip = kvm_rip_read(vcpu);
2471 2472 2473
		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			     irq | INTR_TYPE_SOFT_INTR | INTR_INFO_VALID_MASK);
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
2474
		kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1);
2475 2476
		return;
	}
2477 2478 2479 2480 2481 2482 2483 2484
	intr = irq | INTR_INFO_VALID_MASK;
	if (vcpu->arch.interrupt.soft) {
		intr |= INTR_TYPE_SOFT_INTR;
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN,
			     vmx->vcpu.arch.event_exit_inst_len);
	} else
		intr |= INTR_TYPE_EXT_INTR;
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, intr);
2485 2486
}

2487 2488
static void vmx_inject_nmi(struct kvm_vcpu *vcpu)
{
J
Jan Kiszka 已提交
2489 2490
	struct vcpu_vmx *vmx = to_vmx(vcpu);

2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503
	if (!cpu_has_virtual_nmis()) {
		/*
		 * Tracking the NMI-blocked state in software is built upon
		 * finding the next open IRQ window. This, in turn, depends on
		 * well-behaving guests: They have to keep IRQs disabled at
		 * least as long as the NMI handler runs. Otherwise we may
		 * cause NMI nesting, maybe breaking the guest. But as this is
		 * highly unlikely, we can live with the residual risk.
		 */
		vmx->soft_vnmi_blocked = 1;
		vmx->vnmi_blocked_time = 0;
	}

2504
	++vcpu->stat.nmi_injections;
2505
	if (vcpu->arch.rmode.vm86_active) {
J
Jan Kiszka 已提交
2506 2507 2508 2509 2510 2511 2512 2513 2514 2515
		vmx->rmode.irq.pending = true;
		vmx->rmode.irq.vector = NMI_VECTOR;
		vmx->rmode.irq.rip = kvm_rip_read(vcpu);
		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			     NMI_VECTOR | INTR_TYPE_SOFT_INTR |
			     INTR_INFO_VALID_MASK);
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
		kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1);
		return;
	}
2516 2517 2518 2519
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			INTR_TYPE_NMI_INTR | INTR_INFO_VALID_MASK | NMI_VECTOR);
}

2520
static int vmx_nmi_allowed(struct kvm_vcpu *vcpu)
2521
{
2522
	if (!cpu_has_virtual_nmis() && to_vmx(vcpu)->soft_vnmi_blocked)
2523
		return 0;
2524

2525 2526 2527
	return	!(vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) &
			(GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS |
				GUEST_INTR_STATE_NMI));
2528 2529
}

2530 2531
static int vmx_interrupt_allowed(struct kvm_vcpu *vcpu)
{
2532 2533 2534
	return (vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF) &&
		!(vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) &
			(GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS));
2535 2536
}

2537 2538 2539 2540
static int vmx_set_tss_addr(struct kvm *kvm, unsigned int addr)
{
	int ret;
	struct kvm_userspace_memory_region tss_mem = {
2541
		.slot = TSS_PRIVATE_MEMSLOT,
2542 2543 2544 2545 2546 2547 2548 2549
		.guest_phys_addr = addr,
		.memory_size = PAGE_SIZE * 3,
		.flags = 0,
	};

	ret = kvm_set_memory_region(kvm, &tss_mem, 0);
	if (ret)
		return ret;
2550
	kvm->arch.tss_addr = addr;
2551 2552 2553
	return 0;
}

A
Avi Kivity 已提交
2554 2555 2556
static int handle_rmode_exception(struct kvm_vcpu *vcpu,
				  int vec, u32 err_code)
{
2557 2558 2559 2560 2561
	/*
	 * Instruction with address size override prefix opcode 0x67
	 * Cause the #SS fault with 0 error code in VM86 mode.
	 */
	if (((vec == GP_VECTOR) || (vec == SS_VECTOR)) && err_code == 0)
2562
		if (emulate_instruction(vcpu, NULL, 0, 0, 0) == EMULATE_DONE)
A
Avi Kivity 已提交
2563
			return 1;
2564 2565 2566 2567 2568 2569 2570
	/*
	 * Forward all other exceptions that are valid in real mode.
	 * FIXME: Breaks guest debugging in real mode, needs to be fixed with
	 *        the required debugging infrastructure rework.
	 */
	switch (vec) {
	case DB_VECTOR:
J
Jan Kiszka 已提交
2571 2572 2573 2574 2575
		if (vcpu->guest_debug &
		    (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))
			return 0;
		kvm_queue_exception(vcpu, vec);
		return 1;
2576
	case BP_VECTOR:
J
Jan Kiszka 已提交
2577 2578 2579 2580
		if (vcpu->guest_debug & KVM_GUESTDBG_USE_SW_BP)
			return 0;
		/* fall through */
	case DE_VECTOR:
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590
	case OF_VECTOR:
	case BR_VECTOR:
	case UD_VECTOR:
	case DF_VECTOR:
	case SS_VECTOR:
	case GP_VECTOR:
	case MF_VECTOR:
		kvm_queue_exception(vcpu, vec);
		return 1;
	}
A
Avi Kivity 已提交
2591 2592 2593
	return 0;
}

A
Andi Kleen 已提交
2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618
/*
 * Trigger machine check on the host. We assume all the MSRs are already set up
 * by the CPU and that we still run on the same CPU as the MCE occurred on.
 * We pass a fake environment to the machine check handler because we want
 * the guest to be always treated like user space, no matter what context
 * it used internally.
 */
static void kvm_machine_check(void)
{
#if defined(CONFIG_X86_MCE) && defined(CONFIG_X86_64)
	struct pt_regs regs = {
		.cs = 3, /* Fake ring 3 no matter what the guest ran on */
		.flags = X86_EFLAGS_IF,
	};

	do_machine_check(&regs, 0);
#endif
}

static int handle_machine_check(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	/* already handled by vcpu_run */
	return 1;
}

A
Avi Kivity 已提交
2619 2620
static int handle_exception(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2621
	struct vcpu_vmx *vmx = to_vmx(vcpu);
J
Jan Kiszka 已提交
2622
	u32 intr_info, ex_no, error_code;
2623
	unsigned long cr2, rip, dr6;
A
Avi Kivity 已提交
2624 2625 2626
	u32 vect_info;
	enum emulation_result er;

2627
	vect_info = vmx->idt_vectoring_info;
A
Avi Kivity 已提交
2628 2629
	intr_info = vmcs_read32(VM_EXIT_INTR_INFO);

A
Andi Kleen 已提交
2630 2631 2632
	if (is_machine_check(intr_info))
		return handle_machine_check(vcpu, kvm_run);

A
Avi Kivity 已提交
2633
	if ((vect_info & VECTORING_INFO_VALID_MASK) &&
M
Mike Day 已提交
2634
						!is_page_fault(intr_info))
A
Avi Kivity 已提交
2635
		printk(KERN_ERR "%s: unexpected, vectoring info 0x%x "
2636
		       "intr info 0x%x\n", __func__, vect_info, intr_info);
A
Avi Kivity 已提交
2637

2638
	if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR)
2639
		return 1;  /* already handled by vmx_vcpu_run() */
2640 2641

	if (is_no_device(intr_info)) {
2642
		vmx_fpu_activate(vcpu);
2643 2644 2645
		return 1;
	}

2646
	if (is_invalid_opcode(intr_info)) {
2647
		er = emulate_instruction(vcpu, kvm_run, 0, 0, EMULTYPE_TRAP_UD);
2648
		if (er != EMULATE_DONE)
2649
			kvm_queue_exception(vcpu, UD_VECTOR);
2650 2651 2652
		return 1;
	}

A
Avi Kivity 已提交
2653
	error_code = 0;
2654
	rip = kvm_rip_read(vcpu);
2655
	if (intr_info & INTR_INFO_DELIVER_CODE_MASK)
A
Avi Kivity 已提交
2656 2657
		error_code = vmcs_read32(VM_EXIT_INTR_ERROR_CODE);
	if (is_page_fault(intr_info)) {
2658
		/* EPT won't cause page fault directly */
2659
		if (enable_ept)
2660
			BUG();
A
Avi Kivity 已提交
2661
		cr2 = vmcs_readl(EXIT_QUALIFICATION);
F
Feng (Eric) Liu 已提交
2662 2663
		KVMTRACE_3D(PAGE_FAULT, vcpu, error_code, (u32)cr2,
			    (u32)((u64)cr2 >> 32), handler);
2664
		if (kvm_event_needs_reinjection(vcpu))
2665
			kvm_mmu_unprotect_page_virt(vcpu, cr2);
2666
		return kvm_mmu_page_fault(vcpu, cr2, error_code);
A
Avi Kivity 已提交
2667 2668
	}

2669
	if (vcpu->arch.rmode.vm86_active &&
A
Avi Kivity 已提交
2670
	    handle_rmode_exception(vcpu, intr_info & INTR_INFO_VECTOR_MASK,
2671
								error_code)) {
2672 2673
		if (vcpu->arch.halt_request) {
			vcpu->arch.halt_request = 0;
2674 2675
			return kvm_emulate_halt(vcpu);
		}
A
Avi Kivity 已提交
2676
		return 1;
2677
	}
A
Avi Kivity 已提交
2678

J
Jan Kiszka 已提交
2679
	ex_no = intr_info & INTR_INFO_VECTOR_MASK;
2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692
	switch (ex_no) {
	case DB_VECTOR:
		dr6 = vmcs_readl(EXIT_QUALIFICATION);
		if (!(vcpu->guest_debug &
		      (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))) {
			vcpu->arch.dr6 = dr6 | DR6_FIXED_1;
			kvm_queue_exception(vcpu, DB_VECTOR);
			return 1;
		}
		kvm_run->debug.arch.dr6 = dr6 | DR6_FIXED_1;
		kvm_run->debug.arch.dr7 = vmcs_readl(GUEST_DR7);
		/* fall through */
	case BP_VECTOR:
A
Avi Kivity 已提交
2693
		kvm_run->exit_reason = KVM_EXIT_DEBUG;
J
Jan Kiszka 已提交
2694 2695
		kvm_run->debug.arch.pc = vmcs_readl(GUEST_CS_BASE) + rip;
		kvm_run->debug.arch.exception = ex_no;
2696 2697
		break;
	default:
J
Jan Kiszka 已提交
2698 2699 2700
		kvm_run->exit_reason = KVM_EXIT_EXCEPTION;
		kvm_run->ex.exception = ex_no;
		kvm_run->ex.error_code = error_code;
2701
		break;
A
Avi Kivity 已提交
2702 2703 2704 2705 2706 2707 2708
	}
	return 0;
}

static int handle_external_interrupt(struct kvm_vcpu *vcpu,
				     struct kvm_run *kvm_run)
{
A
Avi Kivity 已提交
2709
	++vcpu->stat.irq_exits;
F
Feng (Eric) Liu 已提交
2710
	KVMTRACE_1D(INTR, vcpu, vmcs_read32(VM_EXIT_INTR_INFO), handler);
A
Avi Kivity 已提交
2711 2712 2713
	return 1;
}

2714 2715 2716 2717 2718
static int handle_triple_fault(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	kvm_run->exit_reason = KVM_EXIT_SHUTDOWN;
	return 0;
}
A
Avi Kivity 已提交
2719 2720 2721

static int handle_io(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2722
	unsigned long exit_qualification;
2723
	int size, in, string;
2724
	unsigned port;
A
Avi Kivity 已提交
2725

A
Avi Kivity 已提交
2726
	++vcpu->stat.io_exits;
2727
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
2728
	string = (exit_qualification & 16) != 0;
2729 2730

	if (string) {
2731 2732
		if (emulate_instruction(vcpu,
					kvm_run, 0, 0, 0) == EMULATE_DO_MMIO)
2733 2734 2735 2736 2737 2738
			return 0;
		return 1;
	}

	size = (exit_qualification & 7) + 1;
	in = (exit_qualification & 8) != 0;
2739
	port = exit_qualification >> 16;
2740

2741
	skip_emulated_instruction(vcpu);
L
Laurent Vivier 已提交
2742
	return kvm_emulate_pio(vcpu, kvm_run, in, size, port);
A
Avi Kivity 已提交
2743 2744
}

I
Ingo Molnar 已提交
2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755
static void
vmx_patch_hypercall(struct kvm_vcpu *vcpu, unsigned char *hypercall)
{
	/*
	 * Patch in the VMCALL instruction:
	 */
	hypercall[0] = 0x0f;
	hypercall[1] = 0x01;
	hypercall[2] = 0xc1;
}

A
Avi Kivity 已提交
2756 2757
static int handle_cr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2758
	unsigned long exit_qualification;
A
Avi Kivity 已提交
2759 2760 2761
	int cr;
	int reg;

2762
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
A
Avi Kivity 已提交
2763 2764 2765 2766
	cr = exit_qualification & 15;
	reg = (exit_qualification >> 8) & 15;
	switch ((exit_qualification >> 4) & 3) {
	case 0: /* mov to cr */
2767 2768 2769 2770
		KVMTRACE_3D(CR_WRITE, vcpu, (u32)cr,
			    (u32)kvm_register_read(vcpu, reg),
			    (u32)((u64)kvm_register_read(vcpu, reg) >> 32),
			    handler);
A
Avi Kivity 已提交
2771 2772
		switch (cr) {
		case 0:
2773
			kvm_set_cr0(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2774 2775 2776
			skip_emulated_instruction(vcpu);
			return 1;
		case 3:
2777
			kvm_set_cr3(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2778 2779 2780
			skip_emulated_instruction(vcpu);
			return 1;
		case 4:
2781
			kvm_set_cr4(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2782 2783
			skip_emulated_instruction(vcpu);
			return 1;
2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795
		case 8: {
				u8 cr8_prev = kvm_get_cr8(vcpu);
				u8 cr8 = kvm_register_read(vcpu, reg);
				kvm_set_cr8(vcpu, cr8);
				skip_emulated_instruction(vcpu);
				if (irqchip_in_kernel(vcpu->kvm))
					return 1;
				if (cr8_prev <= cr8)
					return 1;
				kvm_run->exit_reason = KVM_EXIT_SET_TPR;
				return 0;
			}
A
Avi Kivity 已提交
2796 2797
		};
		break;
2798
	case 2: /* clts */
2799
		vmx_fpu_deactivate(vcpu);
2800 2801
		vcpu->arch.cr0 &= ~X86_CR0_TS;
		vmcs_writel(CR0_READ_SHADOW, vcpu->arch.cr0);
2802
		vmx_fpu_activate(vcpu);
F
Feng (Eric) Liu 已提交
2803
		KVMTRACE_0D(CLTS, vcpu, handler);
2804 2805
		skip_emulated_instruction(vcpu);
		return 1;
A
Avi Kivity 已提交
2806 2807 2808
	case 1: /*mov from cr*/
		switch (cr) {
		case 3:
2809
			kvm_register_write(vcpu, reg, vcpu->arch.cr3);
F
Feng (Eric) Liu 已提交
2810
			KVMTRACE_3D(CR_READ, vcpu, (u32)cr,
2811 2812
				    (u32)kvm_register_read(vcpu, reg),
				    (u32)((u64)kvm_register_read(vcpu, reg) >> 32),
F
Feng (Eric) Liu 已提交
2813
				    handler);
A
Avi Kivity 已提交
2814 2815 2816
			skip_emulated_instruction(vcpu);
			return 1;
		case 8:
2817
			kvm_register_write(vcpu, reg, kvm_get_cr8(vcpu));
F
Feng (Eric) Liu 已提交
2818
			KVMTRACE_2D(CR_READ, vcpu, (u32)cr,
2819
				    (u32)kvm_register_read(vcpu, reg), handler);
A
Avi Kivity 已提交
2820 2821 2822 2823 2824
			skip_emulated_instruction(vcpu);
			return 1;
		}
		break;
	case 3: /* lmsw */
2825
		kvm_lmsw(vcpu, (exit_qualification >> LMSW_SOURCE_DATA_SHIFT) & 0x0f);
A
Avi Kivity 已提交
2826 2827 2828 2829 2830 2831 2832

		skip_emulated_instruction(vcpu);
		return 1;
	default:
		break;
	}
	kvm_run->exit_reason = 0;
2833
	pr_unimpl(vcpu, "unhandled control register: op %d cr %d\n",
A
Avi Kivity 已提交
2834 2835 2836 2837 2838 2839
	       (int)(exit_qualification >> 4) & 3, cr);
	return 0;
}

static int handle_dr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2840
	unsigned long exit_qualification;
A
Avi Kivity 已提交
2841 2842 2843
	unsigned long val;
	int dr, reg;

2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868
	dr = vmcs_readl(GUEST_DR7);
	if (dr & DR7_GD) {
		/*
		 * As the vm-exit takes precedence over the debug trap, we
		 * need to emulate the latter, either for the host or the
		 * guest debugging itself.
		 */
		if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP) {
			kvm_run->debug.arch.dr6 = vcpu->arch.dr6;
			kvm_run->debug.arch.dr7 = dr;
			kvm_run->debug.arch.pc =
				vmcs_readl(GUEST_CS_BASE) +
				vmcs_readl(GUEST_RIP);
			kvm_run->debug.arch.exception = DB_VECTOR;
			kvm_run->exit_reason = KVM_EXIT_DEBUG;
			return 0;
		} else {
			vcpu->arch.dr7 &= ~DR7_GD;
			vcpu->arch.dr6 |= DR6_BD;
			vmcs_writel(GUEST_DR7, vcpu->arch.dr7);
			kvm_queue_exception(vcpu, DB_VECTOR);
			return 1;
		}
	}

2869
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
2870 2871 2872
	dr = exit_qualification & DEBUG_REG_ACCESS_NUM;
	reg = DEBUG_REG_ACCESS_REG(exit_qualification);
	if (exit_qualification & TYPE_MOV_FROM_DR) {
A
Avi Kivity 已提交
2873
		switch (dr) {
2874 2875 2876
		case 0 ... 3:
			val = vcpu->arch.db[dr];
			break;
A
Avi Kivity 已提交
2877
		case 6:
2878
			val = vcpu->arch.dr6;
A
Avi Kivity 已提交
2879 2880
			break;
		case 7:
2881
			val = vcpu->arch.dr7;
A
Avi Kivity 已提交
2882 2883 2884 2885
			break;
		default:
			val = 0;
		}
2886
		kvm_register_write(vcpu, reg, val);
F
Feng (Eric) Liu 已提交
2887
		KVMTRACE_2D(DR_READ, vcpu, (u32)dr, (u32)val, handler);
A
Avi Kivity 已提交
2888
	} else {
2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920
		val = vcpu->arch.regs[reg];
		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 ... 5:
			if (vcpu->arch.cr4 & X86_CR4_DE)
				kvm_queue_exception(vcpu, UD_VECTOR);
			break;
		case 6:
			if (val & 0xffffffff00000000ULL) {
				kvm_queue_exception(vcpu, GP_VECTOR);
				break;
			}
			vcpu->arch.dr6 = (val & DR6_VOLATILE) | DR6_FIXED_1;
			break;
		case 7:
			if (val & 0xffffffff00000000ULL) {
				kvm_queue_exception(vcpu, GP_VECTOR);
				break;
			}
			vcpu->arch.dr7 = (val & DR7_VOLATILE) | DR7_FIXED_1;
			if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)) {
				vmcs_writel(GUEST_DR7, vcpu->arch.dr7);
				vcpu->arch.switch_db_regs =
					(val & DR7_BP_EN_MASK);
			}
			break;
		}
		KVMTRACE_2D(DR_WRITE, vcpu, (u32)dr, (u32)val, handler);
A
Avi Kivity 已提交
2921 2922 2923 2924 2925 2926 2927
	}
	skip_emulated_instruction(vcpu);
	return 1;
}

static int handle_cpuid(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2928 2929
	kvm_emulate_cpuid(vcpu);
	return 1;
A
Avi Kivity 已提交
2930 2931 2932 2933
}

static int handle_rdmsr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2934
	u32 ecx = vcpu->arch.regs[VCPU_REGS_RCX];
A
Avi Kivity 已提交
2935 2936 2937
	u64 data;

	if (vmx_get_msr(vcpu, ecx, &data)) {
2938
		kvm_inject_gp(vcpu, 0);
A
Avi Kivity 已提交
2939 2940 2941
		return 1;
	}

F
Feng (Eric) Liu 已提交
2942 2943 2944
	KVMTRACE_3D(MSR_READ, vcpu, ecx, (u32)data, (u32)(data >> 32),
		    handler);

A
Avi Kivity 已提交
2945
	/* FIXME: handling of bits 32:63 of rax, rdx */
2946 2947
	vcpu->arch.regs[VCPU_REGS_RAX] = data & -1u;
	vcpu->arch.regs[VCPU_REGS_RDX] = (data >> 32) & -1u;
A
Avi Kivity 已提交
2948 2949 2950 2951 2952 2953
	skip_emulated_instruction(vcpu);
	return 1;
}

static int handle_wrmsr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2954 2955 2956
	u32 ecx = vcpu->arch.regs[VCPU_REGS_RCX];
	u64 data = (vcpu->arch.regs[VCPU_REGS_RAX] & -1u)
		| ((u64)(vcpu->arch.regs[VCPU_REGS_RDX] & -1u) << 32);
A
Avi Kivity 已提交
2957

F
Feng (Eric) Liu 已提交
2958 2959 2960
	KVMTRACE_3D(MSR_WRITE, vcpu, ecx, (u32)data, (u32)(data >> 32),
		    handler);

A
Avi Kivity 已提交
2961
	if (vmx_set_msr(vcpu, ecx, data) != 0) {
2962
		kvm_inject_gp(vcpu, 0);
A
Avi Kivity 已提交
2963 2964 2965 2966 2967 2968 2969
		return 1;
	}

	skip_emulated_instruction(vcpu);
	return 1;
}

2970 2971 2972 2973 2974 2975
static int handle_tpr_below_threshold(struct kvm_vcpu *vcpu,
				      struct kvm_run *kvm_run)
{
	return 1;
}

A
Avi Kivity 已提交
2976 2977 2978
static int handle_interrupt_window(struct kvm_vcpu *vcpu,
				   struct kvm_run *kvm_run)
{
2979 2980 2981 2982 2983 2984
	u32 cpu_based_vm_exec_control;

	/* clear pending irq */
	cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL);
	cpu_based_vm_exec_control &= ~CPU_BASED_VIRTUAL_INTR_PENDING;
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control);
F
Feng (Eric) Liu 已提交
2985 2986

	KVMTRACE_0D(PEND_INTR, vcpu, handler);
2987
	++vcpu->stat.irq_window_exits;
F
Feng (Eric) Liu 已提交
2988

2989 2990 2991 2992
	/*
	 * If the user space waits to inject interrupts, exit as soon as
	 * possible
	 */
2993 2994 2995
	if (!irqchip_in_kernel(vcpu->kvm) &&
	    kvm_run->request_interrupt_window &&
	    !kvm_cpu_has_interrupt(vcpu)) {
2996 2997 2998
		kvm_run->exit_reason = KVM_EXIT_IRQ_WINDOW_OPEN;
		return 0;
	}
A
Avi Kivity 已提交
2999 3000 3001 3002 3003 3004
	return 1;
}

static int handle_halt(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	skip_emulated_instruction(vcpu);
3005
	return kvm_emulate_halt(vcpu);
A
Avi Kivity 已提交
3006 3007
}

3008 3009
static int handle_vmcall(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
3010
	skip_emulated_instruction(vcpu);
3011 3012
	kvm_emulate_hypercall(vcpu);
	return 1;
3013 3014
}

M
Marcelo Tosatti 已提交
3015 3016
static int handle_invlpg(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
3017
	unsigned long exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
M
Marcelo Tosatti 已提交
3018 3019 3020 3021 3022 3023

	kvm_mmu_invlpg(vcpu, exit_qualification);
	skip_emulated_instruction(vcpu);
	return 1;
}

E
Eddie Dong 已提交
3024 3025 3026 3027 3028 3029 3030
static int handle_wbinvd(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	skip_emulated_instruction(vcpu);
	/* TODO: Add support for VT-d/pass-through device */
	return 1;
}

3031 3032
static int handle_apic_access(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
3033
	unsigned long exit_qualification;
3034 3035 3036
	enum emulation_result er;
	unsigned long offset;

3037
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050
	offset = exit_qualification & 0xffful;

	er = emulate_instruction(vcpu, kvm_run, 0, 0, 0);

	if (er !=  EMULATE_DONE) {
		printk(KERN_ERR
		       "Fail to handle apic access vmexit! Offset is 0x%lx\n",
		       offset);
		return -ENOTSUPP;
	}
	return 1;
}

3051 3052
static int handle_task_switch(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
J
Jan Kiszka 已提交
3053
	struct vcpu_vmx *vmx = to_vmx(vcpu);
3054 3055
	unsigned long exit_qualification;
	u16 tss_selector;
3056 3057 3058 3059
	int reason, type, idt_v;

	idt_v = (vmx->idt_vectoring_info & VECTORING_INFO_VALID_MASK);
	type = (vmx->idt_vectoring_info & VECTORING_INFO_TYPE_MASK);
3060 3061 3062 3063

	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);

	reason = (u32)exit_qualification >> 30;
3064 3065 3066 3067 3068 3069 3070 3071 3072
	if (reason == TASK_SWITCH_GATE && idt_v) {
		switch (type) {
		case INTR_TYPE_NMI_INTR:
			vcpu->arch.nmi_injected = false;
			if (cpu_has_virtual_nmis())
				vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO,
					      GUEST_INTR_STATE_NMI);
			break;
		case INTR_TYPE_EXT_INTR:
3073
		case INTR_TYPE_SOFT_INTR:
3074 3075 3076 3077 3078 3079 3080 3081 3082
			kvm_clear_interrupt_queue(vcpu);
			break;
		case INTR_TYPE_HARD_EXCEPTION:
		case INTR_TYPE_SOFT_EXCEPTION:
			kvm_clear_exception_queue(vcpu);
			break;
		default:
			break;
		}
J
Jan Kiszka 已提交
3083
	}
3084 3085
	tss_selector = exit_qualification;

3086 3087 3088 3089 3090
	if (!idt_v || (type != INTR_TYPE_HARD_EXCEPTION &&
		       type != INTR_TYPE_EXT_INTR &&
		       type != INTR_TYPE_NMI_INTR))
		skip_emulated_instruction(vcpu);

3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102
	if (!kvm_task_switch(vcpu, tss_selector, reason))
		return 0;

	/* clear all local breakpoint enable flags */
	vmcs_writel(GUEST_DR7, vmcs_readl(GUEST_DR7) & ~55);

	/*
	 * TODO: What about debug traps on tss switch?
	 *       Are we supposed to inject them and update dr6?
	 */

	return 1;
3103 3104
}

3105 3106
static int handle_ept_violation(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
3107
	unsigned long exit_qualification;
3108 3109 3110
	gpa_t gpa;
	int gla_validity;

3111
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122

	if (exit_qualification & (1 << 6)) {
		printk(KERN_ERR "EPT: GPA exceeds GAW!\n");
		return -ENOTSUPP;
	}

	gla_validity = (exit_qualification >> 7) & 0x3;
	if (gla_validity != 0x3 && gla_validity != 0x1 && gla_validity != 0) {
		printk(KERN_ERR "EPT: Handling EPT violation failed!\n");
		printk(KERN_ERR "EPT: GPA: 0x%lx, GVA: 0x%lx\n",
			(long unsigned int)vmcs_read64(GUEST_PHYSICAL_ADDRESS),
3123
			vmcs_readl(GUEST_LINEAR_ADDRESS));
3124 3125 3126 3127 3128 3129 3130 3131
		printk(KERN_ERR "EPT: Exit qualification is 0x%lx\n",
			(long unsigned int)exit_qualification);
		kvm_run->exit_reason = KVM_EXIT_UNKNOWN;
		kvm_run->hw.hardware_exit_reason = 0;
		return -ENOTSUPP;
	}

	gpa = vmcs_read64(GUEST_PHYSICAL_ADDRESS);
3132
	return kvm_mmu_page_fault(vcpu, gpa & PAGE_MASK, 0);
3133 3134
}

3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147
static int handle_nmi_window(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	u32 cpu_based_vm_exec_control;

	/* clear pending NMI */
	cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL);
	cpu_based_vm_exec_control &= ~CPU_BASED_VIRTUAL_NMI_PENDING;
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control);
	++vcpu->stat.nmi_window_exits;

	return 1;
}

3148 3149 3150
static void handle_invalid_guest_state(struct kvm_vcpu *vcpu,
				struct kvm_run *kvm_run)
{
3151 3152
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	enum emulation_result err = EMULATE_DONE;
3153 3154 3155 3156 3157 3158 3159

	preempt_enable();
	local_irq_enable();

	while (!guest_state_valid(vcpu)) {
		err = emulate_instruction(vcpu, kvm_run, 0, 0, 0);

3160 3161 3162 3163 3164 3165
		if (err == EMULATE_DO_MMIO)
			break;

		if (err != EMULATE_DONE) {
			kvm_report_emulation_failure(vcpu, "emulation failure");
			return;
3166 3167 3168 3169 3170 3171 3172 3173 3174 3175
		}

		if (signal_pending(current))
			break;
		if (need_resched())
			schedule();
	}

	local_irq_disable();
	preempt_disable();
3176 3177

	vmx->invalid_state_emulation_result = err;
3178 3179
}

A
Avi Kivity 已提交
3180 3181 3182 3183 3184 3185 3186 3187 3188
/*
 * The exit handlers return 1 if the exit was handled fully and guest execution
 * may resume.  Otherwise they set the kvm_run parameter to indicate what needs
 * to be done to userspace and return 0.
 */
static int (*kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu,
				      struct kvm_run *kvm_run) = {
	[EXIT_REASON_EXCEPTION_NMI]           = handle_exception,
	[EXIT_REASON_EXTERNAL_INTERRUPT]      = handle_external_interrupt,
3189
	[EXIT_REASON_TRIPLE_FAULT]            = handle_triple_fault,
3190
	[EXIT_REASON_NMI_WINDOW]	      = handle_nmi_window,
A
Avi Kivity 已提交
3191 3192 3193 3194 3195 3196 3197 3198
	[EXIT_REASON_IO_INSTRUCTION]          = handle_io,
	[EXIT_REASON_CR_ACCESS]               = handle_cr,
	[EXIT_REASON_DR_ACCESS]               = handle_dr,
	[EXIT_REASON_CPUID]                   = handle_cpuid,
	[EXIT_REASON_MSR_READ]                = handle_rdmsr,
	[EXIT_REASON_MSR_WRITE]               = handle_wrmsr,
	[EXIT_REASON_PENDING_INTERRUPT]       = handle_interrupt_window,
	[EXIT_REASON_HLT]                     = handle_halt,
M
Marcelo Tosatti 已提交
3199
	[EXIT_REASON_INVLPG]		      = handle_invlpg,
3200
	[EXIT_REASON_VMCALL]                  = handle_vmcall,
3201 3202
	[EXIT_REASON_TPR_BELOW_THRESHOLD]     = handle_tpr_below_threshold,
	[EXIT_REASON_APIC_ACCESS]             = handle_apic_access,
E
Eddie Dong 已提交
3203
	[EXIT_REASON_WBINVD]                  = handle_wbinvd,
3204
	[EXIT_REASON_TASK_SWITCH]             = handle_task_switch,
3205
	[EXIT_REASON_EPT_VIOLATION]	      = handle_ept_violation,
A
Andi Kleen 已提交
3206
	[EXIT_REASON_MCE_DURING_VMENTRY]      = handle_machine_check,
A
Avi Kivity 已提交
3207 3208 3209
};

static const int kvm_vmx_max_exit_handlers =
3210
	ARRAY_SIZE(kvm_vmx_exit_handlers);
A
Avi Kivity 已提交
3211 3212 3213 3214 3215

/*
 * The guest has exited.  See if we can fix it or if we need userspace
 * assistance.
 */
3216
static int vmx_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
3217
{
3218
	struct vcpu_vmx *vmx = to_vmx(vcpu);
A
Andi Kleen 已提交
3219
	u32 exit_reason = vmx->exit_reason;
3220
	u32 vectoring_info = vmx->idt_vectoring_info;
3221

3222 3223
	KVMTRACE_3D(VMEXIT, vcpu, exit_reason, (u32)kvm_rip_read(vcpu),
		    (u32)((u64)kvm_rip_read(vcpu) >> 32), entryexit);
F
Feng (Eric) Liu 已提交
3224

3225 3226
	/* If we need to emulate an MMIO from handle_invalid_guest_state
	 * we just return 0 */
3227 3228 3229
	if (vmx->emulation_required && emulate_invalid_guest_state) {
		if (guest_state_valid(vcpu))
			vmx->emulation_required = 0;
3230
		return vmx->invalid_state_emulation_result != EMULATE_DO_MMIO;
3231
	}
3232

3233 3234
	/* Access CR3 don't cause VMExit in paging mode, so we need
	 * to sync with guest real CR3. */
3235
	if (enable_ept && is_paging(vcpu)) {
3236 3237 3238 3239
		vcpu->arch.cr3 = vmcs_readl(GUEST_CR3);
		ept_load_pdptrs(vcpu);
	}

3240 3241 3242 3243 3244 3245
	if (unlikely(vmx->fail)) {
		kvm_run->exit_reason = KVM_EXIT_FAIL_ENTRY;
		kvm_run->fail_entry.hardware_entry_failure_reason
			= vmcs_read32(VM_INSTRUCTION_ERROR);
		return 0;
	}
A
Avi Kivity 已提交
3246

M
Mike Day 已提交
3247
	if ((vectoring_info & VECTORING_INFO_VALID_MASK) &&
3248
			(exit_reason != EXIT_REASON_EXCEPTION_NMI &&
J
Jan Kiszka 已提交
3249 3250 3251 3252 3253
			exit_reason != EXIT_REASON_EPT_VIOLATION &&
			exit_reason != EXIT_REASON_TASK_SWITCH))
		printk(KERN_WARNING "%s: unexpected, valid vectoring info "
		       "(0x%x) and exit reason is 0x%x\n",
		       __func__, vectoring_info, exit_reason);
3254 3255

	if (unlikely(!cpu_has_virtual_nmis() && vmx->soft_vnmi_blocked)) {
3256
		if (vmx_interrupt_allowed(vcpu)) {
3257 3258
			vmx->soft_vnmi_blocked = 0;
		} else if (vmx->vnmi_blocked_time > 1000000000LL &&
3259
			   vcpu->arch.nmi_pending) {
3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272
			/*
			 * This CPU don't support us in finding the end of an
			 * NMI-blocked window if the guest runs with IRQs
			 * disabled. So we pull the trigger after 1 s of
			 * futile waiting, but inform the user about this.
			 */
			printk(KERN_WARNING "%s: Breaking out of NMI-blocked "
			       "state on VCPU %d after 1 s timeout\n",
			       __func__, vcpu->vcpu_id);
			vmx->soft_vnmi_blocked = 0;
		}
	}

A
Avi Kivity 已提交
3273 3274 3275 3276 3277 3278 3279 3280 3281 3282
	if (exit_reason < kvm_vmx_max_exit_handlers
	    && kvm_vmx_exit_handlers[exit_reason])
		return kvm_vmx_exit_handlers[exit_reason](vcpu, kvm_run);
	else {
		kvm_run->exit_reason = KVM_EXIT_UNKNOWN;
		kvm_run->hw.hardware_exit_reason = exit_reason;
	}
	return 0;
}

3283
static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
3284
{
3285
	if (irr == -1 || tpr < irr) {
3286 3287 3288 3289
		vmcs_write32(TPR_THRESHOLD, 0);
		return;
	}

3290
	vmcs_write32(TPR_THRESHOLD, irr);
3291 3292
}

3293 3294 3295
static void vmx_complete_interrupts(struct vcpu_vmx *vmx)
{
	u32 exit_intr_info;
3296
	u32 idt_vectoring_info = vmx->idt_vectoring_info;
3297 3298
	bool unblock_nmi;
	u8 vector;
3299 3300
	int type;
	bool idtv_info_valid;
3301 3302

	exit_intr_info = vmcs_read32(VM_EXIT_INTR_INFO);
3303

A
Andi Kleen 已提交
3304 3305 3306 3307 3308 3309 3310 3311
	vmx->exit_reason = vmcs_read32(VM_EXIT_REASON);

	/* Handle machine checks before interrupts are enabled */
	if ((vmx->exit_reason == EXIT_REASON_MCE_DURING_VMENTRY)
	    || (vmx->exit_reason == EXIT_REASON_EXCEPTION_NMI
		&& is_machine_check(exit_intr_info)))
		kvm_machine_check();

3312 3313 3314 3315 3316 3317 3318 3319 3320
	/* We need to handle NMIs before interrupts are enabled */
	if ((exit_intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR &&
	    (exit_intr_info & INTR_INFO_VALID_MASK)) {
		KVMTRACE_0D(NMI, &vmx->vcpu, handler);
		asm("int $2");
	}

	idtv_info_valid = idt_vectoring_info & VECTORING_INFO_VALID_MASK;

3321 3322 3323 3324
	if (cpu_has_virtual_nmis()) {
		unblock_nmi = (exit_intr_info & INTR_INFO_UNBLOCK_NMI) != 0;
		vector = exit_intr_info & INTR_INFO_VECTOR_MASK;
		/*
3325
		 * SDM 3: 27.7.1.2 (September 2008)
3326 3327
		 * Re-set bit "block by NMI" before VM entry if vmexit caused by
		 * a guest IRET fault.
3328 3329 3330 3331 3332
		 * SDM 3: 23.2.2 (September 2008)
		 * Bit 12 is undefined in any of the following cases:
		 *  If the VM exit sets the valid bit in the IDT-vectoring
		 *   information field.
		 *  If the VM exit is due to a double fault.
3333
		 */
3334 3335
		if ((exit_intr_info & INTR_INFO_VALID_MASK) && unblock_nmi &&
		    vector != DF_VECTOR && !idtv_info_valid)
3336 3337
			vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO,
				      GUEST_INTR_STATE_NMI);
3338 3339 3340
	} else if (unlikely(vmx->soft_vnmi_blocked))
		vmx->vnmi_blocked_time +=
			ktime_to_ns(ktime_sub(ktime_get(), vmx->entry_time));
3341

3342 3343 3344 3345 3346 3347 3348
	vmx->vcpu.arch.nmi_injected = false;
	kvm_clear_exception_queue(&vmx->vcpu);
	kvm_clear_interrupt_queue(&vmx->vcpu);

	if (!idtv_info_valid)
		return;

3349 3350
	vector = idt_vectoring_info & VECTORING_INFO_VECTOR_MASK;
	type = idt_vectoring_info & VECTORING_INFO_TYPE_MASK;
3351

3352
	switch (type) {
3353 3354
	case INTR_TYPE_NMI_INTR:
		vmx->vcpu.arch.nmi_injected = true;
3355
		/*
3356
		 * SDM 3: 27.7.1.2 (September 2008)
3357 3358
		 * Clear bit "block by NMI" before VM entry if a NMI
		 * delivery faulted.
3359
		 */
3360 3361 3362 3363
		vmcs_clear_bits(GUEST_INTERRUPTIBILITY_INFO,
				GUEST_INTR_STATE_NMI);
		break;
	case INTR_TYPE_SOFT_EXCEPTION:
3364 3365 3366 3367
		vmx->vcpu.arch.event_exit_inst_len =
			vmcs_read32(VM_EXIT_INSTRUCTION_LEN);
		/* fall through */
	case INTR_TYPE_HARD_EXCEPTION:
3368
		if (idt_vectoring_info & VECTORING_INFO_DELIVER_CODE_MASK) {
3369 3370
			u32 err = vmcs_read32(IDT_VECTORING_ERROR_CODE);
			kvm_queue_exception_e(&vmx->vcpu, vector, err);
3371 3372
		} else
			kvm_queue_exception(&vmx->vcpu, vector);
3373
		break;
3374 3375 3376 3377
	case INTR_TYPE_SOFT_INTR:
		vmx->vcpu.arch.event_exit_inst_len =
			vmcs_read32(VM_EXIT_INSTRUCTION_LEN);
		/* fall through */
3378
	case INTR_TYPE_EXT_INTR:
3379 3380
		kvm_queue_interrupt(&vmx->vcpu, vector,
			type == INTR_TYPE_SOFT_INTR);
3381 3382 3383
		break;
	default:
		break;
3384
	}
3385 3386
}

3387 3388 3389 3390 3391 3392 3393 3394 3395
/*
 * Failure to inject an interrupt should give us the information
 * in IDT_VECTORING_INFO_FIELD.  However, if the failure occurs
 * when fetching the interrupt redirection bitmap in the real-mode
 * tss, this doesn't happen.  So we do it ourselves.
 */
static void fixup_rmode_irq(struct vcpu_vmx *vmx)
{
	vmx->rmode.irq.pending = 0;
3396
	if (kvm_rip_read(&vmx->vcpu) + 1 != vmx->rmode.irq.rip)
3397
		return;
3398
	kvm_rip_write(&vmx->vcpu, vmx->rmode.irq.rip);
3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409
	if (vmx->idt_vectoring_info & VECTORING_INFO_VALID_MASK) {
		vmx->idt_vectoring_info &= ~VECTORING_INFO_TYPE_MASK;
		vmx->idt_vectoring_info |= INTR_TYPE_EXT_INTR;
		return;
	}
	vmx->idt_vectoring_info =
		VECTORING_INFO_VALID_MASK
		| INTR_TYPE_EXT_INTR
		| vmx->rmode.irq.vector;
}

3410 3411 3412 3413 3414 3415 3416 3417
#ifdef CONFIG_X86_64
#define R "r"
#define Q "q"
#else
#define R "e"
#define Q "l"
#endif

3418
static void vmx_vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
A
Avi Kivity 已提交
3419
{
3420
	struct vcpu_vmx *vmx = to_vmx(vcpu);
3421

3422 3423 3424 3425
	/* Record the guest's net vcpu time for enforced NMI injections. */
	if (unlikely(!cpu_has_virtual_nmis() && vmx->soft_vnmi_blocked))
		vmx->entry_time = ktime_get();

3426 3427 3428 3429 3430 3431
	/* Handle invalid guest state instead of entering VMX */
	if (vmx->emulation_required && emulate_invalid_guest_state) {
		handle_invalid_guest_state(vcpu, kvm_run);
		return;
	}

3432 3433 3434 3435 3436
	if (test_bit(VCPU_REGS_RSP, (unsigned long *)&vcpu->arch.regs_dirty))
		vmcs_writel(GUEST_RSP, vcpu->arch.regs[VCPU_REGS_RSP]);
	if (test_bit(VCPU_REGS_RIP, (unsigned long *)&vcpu->arch.regs_dirty))
		vmcs_writel(GUEST_RIP, vcpu->arch.regs[VCPU_REGS_RIP]);

3437 3438 3439 3440 3441
	/*
	 * Loading guest fpu may have cleared host cr0.ts
	 */
	vmcs_writel(HOST_CR0, read_cr0());

3442 3443
	set_debugreg(vcpu->arch.dr6, 6);

M
Mike Day 已提交
3444
	asm(
A
Avi Kivity 已提交
3445
		/* Store host registers */
3446 3447
		"push %%"R"dx; push %%"R"bp;"
		"push %%"R"cx \n\t"
3448 3449 3450
		"cmp %%"R"sp, %c[host_rsp](%0) \n\t"
		"je 1f \n\t"
		"mov %%"R"sp, %c[host_rsp](%0) \n\t"
3451
		__ex(ASM_VMX_VMWRITE_RSP_RDX) "\n\t"
3452
		"1: \n\t"
A
Avi Kivity 已提交
3453
		/* Check if vmlaunch of vmresume is needed */
3454
		"cmpl $0, %c[launched](%0) \n\t"
A
Avi Kivity 已提交
3455
		/* Load guest registers.  Don't clobber flags. */
3456 3457 3458 3459 3460 3461 3462 3463
		"mov %c[cr2](%0), %%"R"ax \n\t"
		"mov %%"R"ax, %%cr2 \n\t"
		"mov %c[rax](%0), %%"R"ax \n\t"
		"mov %c[rbx](%0), %%"R"bx \n\t"
		"mov %c[rdx](%0), %%"R"dx \n\t"
		"mov %c[rsi](%0), %%"R"si \n\t"
		"mov %c[rdi](%0), %%"R"di \n\t"
		"mov %c[rbp](%0), %%"R"bp \n\t"
3464
#ifdef CONFIG_X86_64
3465 3466 3467 3468 3469 3470 3471 3472
		"mov %c[r8](%0),  %%r8  \n\t"
		"mov %c[r9](%0),  %%r9  \n\t"
		"mov %c[r10](%0), %%r10 \n\t"
		"mov %c[r11](%0), %%r11 \n\t"
		"mov %c[r12](%0), %%r12 \n\t"
		"mov %c[r13](%0), %%r13 \n\t"
		"mov %c[r14](%0), %%r14 \n\t"
		"mov %c[r15](%0), %%r15 \n\t"
A
Avi Kivity 已提交
3473
#endif
3474 3475
		"mov %c[rcx](%0), %%"R"cx \n\t" /* kills %0 (ecx) */

A
Avi Kivity 已提交
3476
		/* Enter guest mode */
3477
		"jne .Llaunched \n\t"
3478
		__ex(ASM_VMX_VMLAUNCH) "\n\t"
3479
		"jmp .Lkvm_vmx_return \n\t"
3480
		".Llaunched: " __ex(ASM_VMX_VMRESUME) "\n\t"
3481
		".Lkvm_vmx_return: "
A
Avi Kivity 已提交
3482
		/* Save guest registers, load host registers, keep flags */
3483 3484 3485 3486 3487 3488 3489 3490
		"xchg %0,     (%%"R"sp) \n\t"
		"mov %%"R"ax, %c[rax](%0) \n\t"
		"mov %%"R"bx, %c[rbx](%0) \n\t"
		"push"Q" (%%"R"sp); pop"Q" %c[rcx](%0) \n\t"
		"mov %%"R"dx, %c[rdx](%0) \n\t"
		"mov %%"R"si, %c[rsi](%0) \n\t"
		"mov %%"R"di, %c[rdi](%0) \n\t"
		"mov %%"R"bp, %c[rbp](%0) \n\t"
3491
#ifdef CONFIG_X86_64
3492 3493 3494 3495 3496 3497 3498 3499
		"mov %%r8,  %c[r8](%0) \n\t"
		"mov %%r9,  %c[r9](%0) \n\t"
		"mov %%r10, %c[r10](%0) \n\t"
		"mov %%r11, %c[r11](%0) \n\t"
		"mov %%r12, %c[r12](%0) \n\t"
		"mov %%r13, %c[r13](%0) \n\t"
		"mov %%r14, %c[r14](%0) \n\t"
		"mov %%r15, %c[r15](%0) \n\t"
A
Avi Kivity 已提交
3500
#endif
3501 3502 3503 3504
		"mov %%cr2, %%"R"ax   \n\t"
		"mov %%"R"ax, %c[cr2](%0) \n\t"

		"pop  %%"R"bp; pop  %%"R"bp; pop  %%"R"dx \n\t"
3505 3506 3507 3508
		"setbe %c[fail](%0) \n\t"
	      : : "c"(vmx), "d"((unsigned long)HOST_RSP),
		[launched]"i"(offsetof(struct vcpu_vmx, launched)),
		[fail]"i"(offsetof(struct vcpu_vmx, fail)),
3509
		[host_rsp]"i"(offsetof(struct vcpu_vmx, host_rsp)),
3510 3511 3512 3513 3514 3515 3516
		[rax]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RAX])),
		[rbx]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RBX])),
		[rcx]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RCX])),
		[rdx]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RDX])),
		[rsi]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RSI])),
		[rdi]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RDI])),
		[rbp]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_RBP])),
3517
#ifdef CONFIG_X86_64
3518 3519 3520 3521 3522 3523 3524 3525
		[r8]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R8])),
		[r9]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R9])),
		[r10]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R10])),
		[r11]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R11])),
		[r12]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R12])),
		[r13]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R13])),
		[r14]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R14])),
		[r15]"i"(offsetof(struct vcpu_vmx, vcpu.arch.regs[VCPU_REGS_R15])),
A
Avi Kivity 已提交
3526
#endif
3527
		[cr2]"i"(offsetof(struct vcpu_vmx, vcpu.arch.cr2))
3528
	      : "cc", "memory"
3529
		, R"bx", R"di", R"si"
3530 3531 3532 3533
#ifdef CONFIG_X86_64
		, "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15"
#endif
	      );
A
Avi Kivity 已提交
3534

3535 3536 3537
	vcpu->arch.regs_avail = ~((1 << VCPU_REGS_RIP) | (1 << VCPU_REGS_RSP));
	vcpu->arch.regs_dirty = 0;

3538 3539
	get_debugreg(vcpu->arch.dr6, 6);

3540
	vmx->idt_vectoring_info = vmcs_read32(IDT_VECTORING_INFO_FIELD);
3541 3542
	if (vmx->rmode.irq.pending)
		fixup_rmode_irq(vmx);
3543

M
Mike Day 已提交
3544
	asm("mov %0, %%ds; mov %0, %%es" : : "r"(__USER_DS));
3545
	vmx->launched = 1;
3546

3547
	vmx_complete_interrupts(vmx);
A
Avi Kivity 已提交
3548 3549
}

3550 3551 3552
#undef R
#undef Q

A
Avi Kivity 已提交
3553 3554
static void vmx_free_vmcs(struct kvm_vcpu *vcpu)
{
3555 3556 3557
	struct vcpu_vmx *vmx = to_vmx(vcpu);

	if (vmx->vmcs) {
3558
		vcpu_clear(vmx);
3559 3560
		free_vmcs(vmx->vmcs);
		vmx->vmcs = NULL;
A
Avi Kivity 已提交
3561 3562 3563 3564 3565
	}
}

static void vmx_free_vcpu(struct kvm_vcpu *vcpu)
{
R
Rusty Russell 已提交
3566 3567
	struct vcpu_vmx *vmx = to_vmx(vcpu);

3568 3569 3570 3571
	spin_lock(&vmx_vpid_lock);
	if (vmx->vpid != 0)
		__clear_bit(vmx->vpid, vmx_vpid_bitmap);
	spin_unlock(&vmx_vpid_lock);
A
Avi Kivity 已提交
3572
	vmx_free_vmcs(vcpu);
R
Rusty Russell 已提交
3573 3574 3575
	kfree(vmx->host_msrs);
	kfree(vmx->guest_msrs);
	kvm_vcpu_uninit(vcpu);
3576
	kmem_cache_free(kvm_vcpu_cache, vmx);
A
Avi Kivity 已提交
3577 3578
}

R
Rusty Russell 已提交
3579
static struct kvm_vcpu *vmx_create_vcpu(struct kvm *kvm, unsigned int id)
A
Avi Kivity 已提交
3580
{
R
Rusty Russell 已提交
3581
	int err;
3582
	struct vcpu_vmx *vmx = kmem_cache_zalloc(kvm_vcpu_cache, GFP_KERNEL);
3583
	int cpu;
A
Avi Kivity 已提交
3584

3585
	if (!vmx)
R
Rusty Russell 已提交
3586 3587
		return ERR_PTR(-ENOMEM);

3588 3589
	allocate_vpid(vmx);

R
Rusty Russell 已提交
3590 3591 3592
	err = kvm_vcpu_init(&vmx->vcpu, kvm, id);
	if (err)
		goto free_vcpu;
3593

3594
	vmx->guest_msrs = kmalloc(PAGE_SIZE, GFP_KERNEL);
R
Rusty Russell 已提交
3595 3596 3597 3598
	if (!vmx->guest_msrs) {
		err = -ENOMEM;
		goto uninit_vcpu;
	}
3599

3600 3601
	vmx->host_msrs = kmalloc(PAGE_SIZE, GFP_KERNEL);
	if (!vmx->host_msrs)
R
Rusty Russell 已提交
3602
		goto free_guest_msrs;
3603

3604 3605
	vmx->vmcs = alloc_vmcs();
	if (!vmx->vmcs)
R
Rusty Russell 已提交
3606
		goto free_msrs;
3607 3608 3609

	vmcs_clear(vmx->vmcs);

3610 3611
	cpu = get_cpu();
	vmx_vcpu_load(&vmx->vcpu, cpu);
R
Rusty Russell 已提交
3612
	err = vmx_vcpu_setup(vmx);
R
Rusty Russell 已提交
3613
	vmx_vcpu_put(&vmx->vcpu);
3614
	put_cpu();
R
Rusty Russell 已提交
3615 3616
	if (err)
		goto free_vmcs;
3617 3618 3619
	if (vm_need_virtualize_apic_accesses(kvm))
		if (alloc_apic_access_page(kvm) != 0)
			goto free_vmcs;
R
Rusty Russell 已提交
3620

3621
	if (enable_ept)
3622 3623 3624
		if (alloc_identity_pagetable(kvm) != 0)
			goto free_vmcs;

R
Rusty Russell 已提交
3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635
	return &vmx->vcpu;

free_vmcs:
	free_vmcs(vmx->vmcs);
free_msrs:
	kfree(vmx->host_msrs);
free_guest_msrs:
	kfree(vmx->guest_msrs);
uninit_vcpu:
	kvm_vcpu_uninit(&vmx->vcpu);
free_vcpu:
3636
	kmem_cache_free(kvm_vcpu_cache, vmx);
R
Rusty Russell 已提交
3637
	return ERR_PTR(err);
A
Avi Kivity 已提交
3638 3639
}

Y
Yang, Sheng 已提交
3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653
static void __init vmx_check_processor_compat(void *rtn)
{
	struct vmcs_config vmcs_conf;

	*(int *)rtn = 0;
	if (setup_vmcs_config(&vmcs_conf) < 0)
		*(int *)rtn = -EIO;
	if (memcmp(&vmcs_config, &vmcs_conf, sizeof(struct vmcs_config)) != 0) {
		printk(KERN_ERR "kvm: CPU %d feature inconsistency!\n",
				smp_processor_id());
		*(int *)rtn = -EIO;
	}
}

3654 3655 3656 3657 3658
static int get_ept_level(void)
{
	return VMX_EPT_DEFAULT_GAW + 1;
}

3659
static u64 vmx_get_mt_mask(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio)
S
Sheng Yang 已提交
3660
{
3661 3662
	u64 ret;

3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673
	/* For VT-d and EPT combination
	 * 1. MMIO: always map as UC
	 * 2. EPT with VT-d:
	 *   a. VT-d without snooping control feature: can't guarantee the
	 *	result, try to trust guest.
	 *   b. VT-d with snooping control feature: snooping control feature of
	 *	VT-d engine can guarantee the cache correctness. Just set it
	 *	to WB to keep consistent with host. So the same as item 3.
	 * 3. EPT without VT-d: always map as WB and set IGMT=1 to keep
	 *    consistent with host MTRR
	 */
3674 3675
	if (is_mmio)
		ret = MTRR_TYPE_UNCACHABLE << VMX_EPT_MT_EPTE_SHIFT;
3676 3677 3678 3679
	else if (vcpu->kvm->arch.iommu_domain &&
		!(vcpu->kvm->arch.iommu_flags & KVM_IOMMU_CACHE_COHERENCY))
		ret = kvm_get_guest_memory_type(vcpu, gfn) <<
		      VMX_EPT_MT_EPTE_SHIFT;
3680
	else
3681 3682
		ret = (MTRR_TYPE_WRBACK << VMX_EPT_MT_EPTE_SHIFT)
			| VMX_EPT_IGMT_BIT;
3683 3684

	return ret;
S
Sheng Yang 已提交
3685 3686
}

3687
static struct kvm_x86_ops vmx_x86_ops = {
A
Avi Kivity 已提交
3688 3689 3690 3691
	.cpu_has_kvm_support = cpu_has_kvm_support,
	.disabled_by_bios = vmx_disabled_by_bios,
	.hardware_setup = hardware_setup,
	.hardware_unsetup = hardware_unsetup,
Y
Yang, Sheng 已提交
3692
	.check_processor_compatibility = vmx_check_processor_compat,
A
Avi Kivity 已提交
3693 3694
	.hardware_enable = hardware_enable,
	.hardware_disable = hardware_disable,
3695
	.cpu_has_accelerated_tpr = report_flexpriority,
A
Avi Kivity 已提交
3696 3697 3698

	.vcpu_create = vmx_create_vcpu,
	.vcpu_free = vmx_free_vcpu,
3699
	.vcpu_reset = vmx_vcpu_reset,
A
Avi Kivity 已提交
3700

3701
	.prepare_guest_switch = vmx_save_host_state,
A
Avi Kivity 已提交
3702 3703 3704 3705 3706 3707 3708 3709 3710
	.vcpu_load = vmx_vcpu_load,
	.vcpu_put = vmx_vcpu_put,

	.set_guest_debug = set_guest_debug,
	.get_msr = vmx_get_msr,
	.set_msr = vmx_set_msr,
	.get_segment_base = vmx_get_segment_base,
	.get_segment = vmx_get_segment,
	.set_segment = vmx_set_segment,
3711
	.get_cpl = vmx_get_cpl,
A
Avi Kivity 已提交
3712
	.get_cs_db_l_bits = vmx_get_cs_db_l_bits,
3713
	.decache_cr4_guest_bits = vmx_decache_cr4_guest_bits,
A
Avi Kivity 已提交
3714 3715 3716 3717 3718 3719 3720 3721
	.set_cr0 = vmx_set_cr0,
	.set_cr3 = vmx_set_cr3,
	.set_cr4 = vmx_set_cr4,
	.set_efer = vmx_set_efer,
	.get_idt = vmx_get_idt,
	.set_idt = vmx_set_idt,
	.get_gdt = vmx_get_gdt,
	.set_gdt = vmx_set_gdt,
3722
	.cache_reg = vmx_cache_reg,
A
Avi Kivity 已提交
3723 3724 3725 3726 3727 3728
	.get_rflags = vmx_get_rflags,
	.set_rflags = vmx_set_rflags,

	.tlb_flush = vmx_flush_tlb,

	.run = vmx_vcpu_run,
3729
	.handle_exit = vmx_handle_exit,
A
Avi Kivity 已提交
3730
	.skip_emulated_instruction = skip_emulated_instruction,
3731 3732
	.set_interrupt_shadow = vmx_set_interrupt_shadow,
	.get_interrupt_shadow = vmx_get_interrupt_shadow,
I
Ingo Molnar 已提交
3733
	.patch_hypercall = vmx_patch_hypercall,
E
Eddie Dong 已提交
3734
	.set_irq = vmx_inject_irq,
3735
	.set_nmi = vmx_inject_nmi,
3736
	.queue_exception = vmx_queue_exception,
3737
	.interrupt_allowed = vmx_interrupt_allowed,
3738 3739 3740 3741 3742
	.nmi_allowed = vmx_nmi_allowed,
	.enable_nmi_window = enable_nmi_window,
	.enable_irq_window = enable_irq_window,
	.update_cr8_intercept = update_cr8_intercept,

3743
	.set_tss_addr = vmx_set_tss_addr,
3744
	.get_tdp_level = get_ept_level,
3745
	.get_mt_mask = vmx_get_mt_mask,
A
Avi Kivity 已提交
3746 3747 3748 3749
};

static int __init vmx_init(void)
{
3750 3751
	int r;

3752
	vmx_io_bitmap_a = (unsigned long *)__get_free_page(GFP_KERNEL);
3753 3754 3755
	if (!vmx_io_bitmap_a)
		return -ENOMEM;

3756
	vmx_io_bitmap_b = (unsigned long *)__get_free_page(GFP_KERNEL);
3757 3758 3759 3760 3761
	if (!vmx_io_bitmap_b) {
		r = -ENOMEM;
		goto out;
	}

3762 3763
	vmx_msr_bitmap_legacy = (unsigned long *)__get_free_page(GFP_KERNEL);
	if (!vmx_msr_bitmap_legacy) {
S
Sheng Yang 已提交
3764 3765 3766 3767
		r = -ENOMEM;
		goto out1;
	}

3768 3769 3770 3771 3772 3773
	vmx_msr_bitmap_longmode = (unsigned long *)__get_free_page(GFP_KERNEL);
	if (!vmx_msr_bitmap_longmode) {
		r = -ENOMEM;
		goto out2;
	}

3774 3775 3776 3777
	/*
	 * Allow direct access to the PC debug port (it is often used for I/O
	 * delays, but the vmexits simply slow things down).
	 */
3778 3779
	memset(vmx_io_bitmap_a, 0xff, PAGE_SIZE);
	clear_bit(0x80, vmx_io_bitmap_a);
3780

3781
	memset(vmx_io_bitmap_b, 0xff, PAGE_SIZE);
3782

3783 3784
	memset(vmx_msr_bitmap_legacy, 0xff, PAGE_SIZE);
	memset(vmx_msr_bitmap_longmode, 0xff, PAGE_SIZE);
S
Sheng Yang 已提交
3785

3786 3787
	set_bit(0, vmx_vpid_bitmap); /* 0 is reserved for host */

3788
	r = kvm_init(&vmx_x86_ops, sizeof(struct vcpu_vmx), THIS_MODULE);
3789
	if (r)
3790
		goto out3;
S
Sheng Yang 已提交
3791

3792 3793 3794 3795 3796 3797
	vmx_disable_intercept_for_msr(MSR_FS_BASE, false);
	vmx_disable_intercept_for_msr(MSR_GS_BASE, false);
	vmx_disable_intercept_for_msr(MSR_KERNEL_GS_BASE, true);
	vmx_disable_intercept_for_msr(MSR_IA32_SYSENTER_CS, false);
	vmx_disable_intercept_for_msr(MSR_IA32_SYSENTER_ESP, false);
	vmx_disable_intercept_for_msr(MSR_IA32_SYSENTER_EIP, false);
3798

3799
	if (enable_ept) {
3800
		bypass_guest_pf = 0;
3801
		kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK |
3802
			VMX_EPT_WRITABLE_MASK);
3803
		kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull,
3804
				VMX_EPT_EXECUTABLE_MASK);
3805 3806 3807
		kvm_enable_tdp();
	} else
		kvm_disable_tdp();
3808

3809 3810 3811
	if (bypass_guest_pf)
		kvm_mmu_set_nonpresent_ptes(~0xffeull, 0ull);

3812 3813
	ept_sync_global();

3814 3815
	return 0;

3816 3817
out3:
	free_page((unsigned long)vmx_msr_bitmap_longmode);
S
Sheng Yang 已提交
3818
out2:
3819
	free_page((unsigned long)vmx_msr_bitmap_legacy);
3820
out1:
3821
	free_page((unsigned long)vmx_io_bitmap_b);
3822
out:
3823
	free_page((unsigned long)vmx_io_bitmap_a);
3824
	return r;
A
Avi Kivity 已提交
3825 3826 3827 3828
}

static void __exit vmx_exit(void)
{
3829 3830
	free_page((unsigned long)vmx_msr_bitmap_legacy);
	free_page((unsigned long)vmx_msr_bitmap_longmode);
3831 3832
	free_page((unsigned long)vmx_io_bitmap_b);
	free_page((unsigned long)vmx_io_bitmap_a);
3833

3834
	kvm_exit();
A
Avi Kivity 已提交
3835 3836 3837 3838
}

module_init(vmx_init)
module_exit(vmx_exit)