vmx.c 96.8 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
Avi Kivity 已提交
35

36 37
#define __ex(x) __kvm_handle_fault_on_reboot(x)

A
Avi Kivity 已提交
38 39 40
MODULE_AUTHOR("Qumranet");
MODULE_LICENSE("GPL");

41 42 43
static int bypass_guest_pf = 1;
module_param(bypass_guest_pf, bool, 0);

44 45 46
static int enable_vpid = 1;
module_param(enable_vpid, bool, 0);

47 48 49
static int flexpriority_enabled = 1;
module_param(flexpriority_enabled, bool, 0);

50
static int enable_ept = 1;
S
Sheng Yang 已提交
51 52
module_param(enable_ept, bool, 0);

53 54 55
static int emulate_invalid_guest_state = 0;
module_param(emulate_invalid_guest_state, bool, 0);

56 57 58 59 60 61 62
struct vmcs {
	u32 revision_id;
	u32 abort;
	char data[0];
};

struct vcpu_vmx {
R
Rusty Russell 已提交
63
	struct kvm_vcpu       vcpu;
64
	struct list_head      local_vcpus_link;
65
	unsigned long         host_rsp;
66
	int                   launched;
67
	u8                    fail;
68
	u32                   idt_vectoring_info;
69 70 71 72 73 74 75 76 77 78 79 80
	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;
81 82
		int           gs_ldt_reload_needed;
		int           fs_reload_needed;
83
		int           guest_efer_loaded;
M
Mike Day 已提交
84
	} host_state;
85 86 87 88 89 90 91
	struct {
		struct {
			bool pending;
			u8 vector;
			unsigned rip;
		} irq;
	} rmode;
92
	int vpid;
93
	bool emulation_required;
94 95 96 97 98

	/* Support for vnmi-less CPUs */
	int soft_vnmi_blocked;
	ktime_t entry_time;
	s64 vnmi_blocked_time;
99 100 101 102
};

static inline struct vcpu_vmx *to_vmx(struct kvm_vcpu *vcpu)
{
R
Rusty Russell 已提交
103
	return container_of(vcpu, struct vcpu_vmx, vcpu);
104 105
}

106
static int init_rmode(struct kvm *kvm);
107
static u64 construct_eptp(unsigned long root_hpa);
108

A
Avi Kivity 已提交
109 110
static DEFINE_PER_CPU(struct vmcs *, vmxarea);
static DEFINE_PER_CPU(struct vmcs *, current_vmcs);
111
static DEFINE_PER_CPU(struct list_head, vcpus_on_cpu);
A
Avi Kivity 已提交
112

113 114
static struct page *vmx_io_bitmap_a;
static struct page *vmx_io_bitmap_b;
S
Sheng Yang 已提交
115
static struct page *vmx_msr_bitmap;
116

117 118 119
static DECLARE_BITMAP(vmx_vpid_bitmap, VMX_NR_VPIDS);
static DEFINE_SPINLOCK(vmx_vpid_lock);

120
static struct vmcs_config {
A
Avi Kivity 已提交
121 122 123
	int size;
	int order;
	u32 revision_id;
124 125
	u32 pin_based_exec_ctrl;
	u32 cpu_based_exec_ctrl;
126
	u32 cpu_based_2nd_exec_ctrl;
127 128 129
	u32 vmexit_ctrl;
	u32 vmentry_ctrl;
} vmcs_config;
A
Avi Kivity 已提交
130

H
Hannes Eder 已提交
131
static struct vmx_capability {
S
Sheng Yang 已提交
132 133 134 135
	u32 ept;
	u32 vpid;
} vmx_capability;

A
Avi Kivity 已提交
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159
#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),
};

160 161 162 163
/*
 * 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
164
static const u32 vmx_msr_index[] = {
165
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
166 167 168 169
	MSR_SYSCALL_MASK, MSR_LSTAR, MSR_CSTAR, MSR_KERNEL_GS_BASE,
#endif
	MSR_EFER, MSR_K6_STAR,
};
170
#define NR_VMX_MSR ARRAY_SIZE(vmx_msr_index)
A
Avi Kivity 已提交
171

172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187
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
188 189 190 191
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)) ==
192
		(INTR_TYPE_HARD_EXCEPTION | PF_VECTOR | INTR_INFO_VALID_MASK);
A
Avi Kivity 已提交
193 194
}

195 196 197 198
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)) ==
199
		(INTR_TYPE_HARD_EXCEPTION | NM_VECTOR | INTR_INFO_VALID_MASK);
200 201
}

202 203 204 205
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)) ==
206
		(INTR_TYPE_HARD_EXCEPTION | UD_VECTOR | INTR_INFO_VALID_MASK);
207 208
}

A
Avi Kivity 已提交
209 210 211 212 213 214
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);
}

S
Sheng Yang 已提交
215 216 217 218 219
static inline int cpu_has_vmx_msr_bitmap(void)
{
	return (vmcs_config.cpu_based_exec_ctrl & CPU_BASED_USE_MSR_BITMAPS);
}

220 221 222 223 224 225 226 227 228 229
static inline int cpu_has_vmx_tpr_shadow(void)
{
	return (vmcs_config.cpu_based_exec_ctrl & CPU_BASED_TPR_SHADOW);
}

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

230 231 232 233 234 235
static inline int cpu_has_secondary_exec_ctrls(void)
{
	return (vmcs_config.cpu_based_exec_ctrl &
		CPU_BASED_ACTIVATE_SECONDARY_CONTROLS);
}

236
static inline bool cpu_has_vmx_virtualize_apic_accesses(void)
237
{
238 239 240
	return flexpriority_enabled
		&& (vmcs_config.cpu_based_2nd_exec_ctrl &
		    SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES);
241 242
}

S
Sheng Yang 已提交
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
static inline int cpu_has_vmx_invept_individual_addr(void)
{
	return (!!(vmx_capability.ept & VMX_EPT_EXTENT_INDIVIDUAL_BIT));
}

static inline int cpu_has_vmx_invept_context(void)
{
	return (!!(vmx_capability.ept & VMX_EPT_EXTENT_CONTEXT_BIT));
}

static inline int cpu_has_vmx_invept_global(void)
{
	return (!!(vmx_capability.ept & VMX_EPT_EXTENT_GLOBAL_BIT));
}

static inline int cpu_has_vmx_ept(void)
{
	return (vmcs_config.cpu_based_2nd_exec_ctrl &
		SECONDARY_EXEC_ENABLE_EPT);
}

static inline int vm_need_ept(void)
{
	return (cpu_has_vmx_ept() && enable_ept);
}

269 270 271 272 273 274
static inline int vm_need_virtualize_apic_accesses(struct kvm *kvm)
{
	return ((cpu_has_vmx_virtualize_apic_accesses()) &&
		(irqchip_in_kernel(kvm)));
}

275 276 277 278 279 280
static inline int cpu_has_vmx_vpid(void)
{
	return (vmcs_config.cpu_based_2nd_exec_ctrl &
		SECONDARY_EXEC_ENABLE_VPID);
}

281 282 283 284 285
static inline int cpu_has_virtual_nmis(void)
{
	return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS;
}

R
Rusty Russell 已提交
286
static int __find_msr_index(struct vcpu_vmx *vmx, u32 msr)
287 288 289
{
	int i;

290 291
	for (i = 0; i < vmx->nmsrs; ++i)
		if (vmx->guest_msrs[i].index == msr)
292 293 294 295
			return i;
	return -1;
}

296 297 298 299 300 301 302 303
static inline void __invvpid(int ext, u16 vpid, gva_t gva)
{
    struct {
	u64 vpid : 16;
	u64 rsvd : 48;
	u64 gva;
    } operand = { vpid, 0, gva };

304
    asm volatile (__ex(ASM_VMX_INVVPID)
305 306 307 308 309
		  /* CF==1 or ZF==1 --> rc = -1 */
		  "; ja 1f ; ud2 ; 1:"
		  : : "a"(&operand), "c"(ext) : "cc", "memory");
}

310 311 312 313 314 315
static inline void __invept(int ext, u64 eptp, gpa_t gpa)
{
	struct {
		u64 eptp, gpa;
	} operand = {eptp, gpa};

316
	asm volatile (__ex(ASM_VMX_INVEPT)
317 318 319 320 321
			/* CF==1 or ZF==1 --> rc = -1 */
			"; ja 1f ; ud2 ; 1:\n"
			: : "a" (&operand), "c" (ext) : "cc", "memory");
}

R
Rusty Russell 已提交
322
static struct kvm_msr_entry *find_msr_entry(struct vcpu_vmx *vmx, u32 msr)
323 324 325
{
	int i;

R
Rusty Russell 已提交
326
	i = __find_msr_index(vmx, msr);
327
	if (i >= 0)
328
		return &vmx->guest_msrs[i];
A
Al Viro 已提交
329
	return NULL;
330 331
}

A
Avi Kivity 已提交
332 333 334 335 336
static void vmcs_clear(struct vmcs *vmcs)
{
	u64 phys_addr = __pa(vmcs);
	u8 error;

337
	asm volatile (__ex(ASM_VMX_VMCLEAR_RAX) "; setna %0"
A
Avi Kivity 已提交
338 339 340 341 342 343 344 345 346
		      : "=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
347
	struct vcpu_vmx *vmx = arg;
348
	int cpu = raw_smp_processor_id();
A
Avi Kivity 已提交
349

R
Rusty Russell 已提交
350
	if (vmx->vcpu.cpu == cpu)
351 352
		vmcs_clear(vmx->vmcs);
	if (per_cpu(current_vmcs, cpu) == vmx->vmcs)
A
Avi Kivity 已提交
353
		per_cpu(current_vmcs, cpu) = NULL;
354
	rdtscll(vmx->vcpu.arch.host_tsc);
355 356 357
	list_del(&vmx->local_vcpus_link);
	vmx->vcpu.cpu = -1;
	vmx->launched = 0;
A
Avi Kivity 已提交
358 359
}

R
Rusty Russell 已提交
360
static void vcpu_clear(struct vcpu_vmx *vmx)
A
Avi Kivity 已提交
361
{
362 363
	if (vmx->vcpu.cpu == -1)
		return;
364
	smp_call_function_single(vmx->vcpu.cpu, __vcpu_clear, vmx, 1);
A
Avi Kivity 已提交
365 366
}

367 368 369 370 371 372 373 374
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);
}

375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
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)
{
	if (vm_need_ept()) {
		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)
{
	if (vm_need_ept()) {
		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
402 403 404 405
static unsigned long vmcs_readl(unsigned long field)
{
	unsigned long value;

406
	asm volatile (__ex(ASM_VMX_VMREAD_RDX_RAX)
A
Avi Kivity 已提交
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
		      : "=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)
{
423
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
424 425 426 427 428 429
	return vmcs_readl(field);
#else
	return vmcs_readl(field) | ((u64)vmcs_readl(field+1) << 32);
#endif
}

430 431 432 433 434 435 436
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
437 438 439 440
static void vmcs_writel(unsigned long field, unsigned long value)
{
	u8 error;

441
	asm volatile (__ex(ASM_VMX_VMWRITE_RAX_RDX) "; setna %0"
M
Mike Day 已提交
442
		       : "=q"(error) : "a"(value), "d"(field) : "cc");
443 444
	if (unlikely(error))
		vmwrite_error(field, value);
A
Avi Kivity 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
}

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);
460
#ifndef CONFIG_X86_64
A
Avi Kivity 已提交
461 462 463 464 465
	asm volatile ("");
	vmcs_writel(field+1, value >> 32);
#endif
}

466 467 468 469 470 471 472 473 474 475
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);
}

476 477 478 479
static void update_exception_bitmap(struct kvm_vcpu *vcpu)
{
	u32 eb;

480
	eb = (1u << PF_VECTOR) | (1u << UD_VECTOR);
481 482
	if (!vcpu->fpu_active)
		eb |= 1u << NM_VECTOR;
J
Jan Kiszka 已提交
483 484 485 486 487 488 489
	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;
	}
490
	if (vcpu->arch.rmode.active)
491
		eb = ~0;
492 493
	if (vm_need_ept())
		eb &= ~(1u << PF_VECTOR); /* bypass_guest_pf = 0 */
494 495 496
	vmcs_write32(EXCEPTION_BITMAP, eb);
}

497 498 499 500 501 502
static void reload_tss(void)
{
	/*
	 * VT restores TR but not its size.  Useless.
	 */
	struct descriptor_table gdt;
503
	struct desc_struct *descs;
504

505
	kvm_get_gdt(&gdt);
506 507 508 509 510
	descs = (void *)gdt.base;
	descs[GDT_ENTRY_TSS].type = 9; /* available TSS */
	load_TR_desc();
}

R
Rusty Russell 已提交
511
static void load_transition_efer(struct vcpu_vmx *vmx)
512
{
513
	int efer_offset = vmx->msr_offset_efer;
514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
	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;
533

534 535 536 537
	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
538
	vmx->vcpu.stat.efer_reload++;
539 540
}

541 542 543 544 545 546 547 548
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);
	}
}

549
static void vmx_save_host_state(struct kvm_vcpu *vcpu)
550
{
551 552
	struct vcpu_vmx *vmx = to_vmx(vcpu);

553
	if (vmx->host_state.loaded)
554 555
		return;

556
	vmx->host_state.loaded = 1;
557 558 559 560
	/*
	 * Set host fs and gs selectors.  Unfortunately, 22.2.3 does not
	 * allow segment selectors with cpl > 0 or ti == 1.
	 */
561
	vmx->host_state.ldt_sel = kvm_read_ldt();
562
	vmx->host_state.gs_ldt_reload_needed = vmx->host_state.ldt_sel;
563
	vmx->host_state.fs_sel = kvm_read_fs();
564
	if (!(vmx->host_state.fs_sel & 7)) {
565
		vmcs_write16(HOST_FS_SELECTOR, vmx->host_state.fs_sel);
566 567
		vmx->host_state.fs_reload_needed = 0;
	} else {
568
		vmcs_write16(HOST_FS_SELECTOR, 0);
569
		vmx->host_state.fs_reload_needed = 1;
570
	}
571
	vmx->host_state.gs_sel = kvm_read_gs();
572 573
	if (!(vmx->host_state.gs_sel & 7))
		vmcs_write16(HOST_GS_SELECTOR, vmx->host_state.gs_sel);
574 575
	else {
		vmcs_write16(HOST_GS_SELECTOR, 0);
576
		vmx->host_state.gs_ldt_reload_needed = 1;
577 578 579 580 581 582
	}

#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
583 584
	vmcs_writel(HOST_FS_BASE, segment_base(vmx->host_state.fs_sel));
	vmcs_writel(HOST_GS_BASE, segment_base(vmx->host_state.gs_sel));
585
#endif
586 587

#ifdef CONFIG_X86_64
M
Mike Day 已提交
588
	if (is_long_mode(&vmx->vcpu))
589 590
		save_msrs(vmx->host_msrs +
			  vmx->msr_offset_kernel_gs_base, 1);
M
Mike Day 已提交
591

592
#endif
593
	load_msrs(vmx->guest_msrs, vmx->save_nmsrs);
594
	load_transition_efer(vmx);
595 596
}

597
static void __vmx_load_host_state(struct vcpu_vmx *vmx)
598
{
599
	unsigned long flags;
600

601
	if (!vmx->host_state.loaded)
602 603
		return;

604
	++vmx->vcpu.stat.host_state_reload;
605
	vmx->host_state.loaded = 0;
606
	if (vmx->host_state.fs_reload_needed)
607
		kvm_load_fs(vmx->host_state.fs_sel);
608
	if (vmx->host_state.gs_ldt_reload_needed) {
609
		kvm_load_ldt(vmx->host_state.ldt_sel);
610 611 612 613
		/*
		 * If we have to reload gs, we must take care to
		 * preserve our gs base.
		 */
614
		local_irq_save(flags);
615
		kvm_load_gs(vmx->host_state.gs_sel);
616 617 618
#ifdef CONFIG_X86_64
		wrmsrl(MSR_GS_BASE, vmcs_readl(HOST_GS_BASE));
#endif
619
		local_irq_restore(flags);
620
	}
621
	reload_tss();
622 623
	save_msrs(vmx->guest_msrs, vmx->save_nmsrs);
	load_msrs(vmx->host_msrs, vmx->save_nmsrs);
624
	reload_host_efer(vmx);
625 626
}

627 628 629 630 631 632 633
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
634 635 636 637
/*
 * Switches to specified vcpu, until a matching vcpu_put(), but assumes
 * vcpu mutex is already taken.
 */
638
static void vmx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
A
Avi Kivity 已提交
639
{
640 641
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	u64 phys_addr = __pa(vmx->vmcs);
642
	u64 tsc_this, delta, new_offset;
A
Avi Kivity 已提交
643

644
	if (vcpu->cpu != cpu) {
R
Rusty Russell 已提交
645
		vcpu_clear(vmx);
M
Marcelo Tosatti 已提交
646
		kvm_migrate_timers(vcpu);
647
		vpid_sync_vcpu_all(vmx);
648 649 650 651
		local_irq_disable();
		list_add(&vmx->local_vcpus_link,
			 &per_cpu(vcpus_on_cpu, cpu));
		local_irq_enable();
652
	}
A
Avi Kivity 已提交
653

654
	if (per_cpu(current_vmcs, cpu) != vmx->vmcs) {
A
Avi Kivity 已提交
655 656
		u8 error;

657
		per_cpu(current_vmcs, cpu) = vmx->vmcs;
658
		asm volatile (__ex(ASM_VMX_VMPTRLD_RAX) "; setna %0"
A
Avi Kivity 已提交
659 660 661 662
			      : "=g"(error) : "a"(&phys_addr), "m"(phys_addr)
			      : "cc");
		if (error)
			printk(KERN_ERR "kvm: vmptrld %p/%llx fail\n",
663
			       vmx->vmcs, phys_addr);
A
Avi Kivity 已提交
664 665 666 667 668 669 670 671 672 673 674
	}

	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.
		 */
675 676
		vmcs_writel(HOST_TR_BASE, kvm_read_tr_base()); /* 22.2.4 */
		kvm_get_gdt(&dt);
A
Avi Kivity 已提交
677 678 679 680
		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 */
681 682 683 684 685

		/*
		 * Make sure the time stamp counter is monotonous.
		 */
		rdtscll(tsc_this);
686 687 688 689 690
		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
691 692 693 694 695
	}
}

static void vmx_vcpu_put(struct kvm_vcpu *vcpu)
{
696
	__vmx_load_host_state(to_vmx(vcpu));
A
Avi Kivity 已提交
697 698
}

699 700 701 702 703
static void vmx_fpu_activate(struct kvm_vcpu *vcpu)
{
	if (vcpu->fpu_active)
		return;
	vcpu->fpu_active = 1;
704
	vmcs_clear_bits(GUEST_CR0, X86_CR0_TS);
705
	if (vcpu->arch.cr0 & X86_CR0_TS)
706
		vmcs_set_bits(GUEST_CR0, X86_CR0_TS);
707 708 709 710 711 712 713 714
	update_exception_bitmap(vcpu);
}

static void vmx_fpu_deactivate(struct kvm_vcpu *vcpu)
{
	if (!vcpu->fpu_active)
		return;
	vcpu->fpu_active = 0;
715
	vmcs_set_bits(GUEST_CR0, X86_CR0_TS);
716 717 718
	update_exception_bitmap(vcpu);
}

A
Avi Kivity 已提交
719 720 721 722 723 724 725
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)
{
726
	if (vcpu->arch.rmode.active)
727
		rflags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
A
Avi Kivity 已提交
728 729 730 731 732 733 734 735
	vmcs_writel(GUEST_RFLAGS, rflags);
}

static void skip_emulated_instruction(struct kvm_vcpu *vcpu)
{
	unsigned long rip;
	u32 interruptibility;

736
	rip = kvm_rip_read(vcpu);
A
Avi Kivity 已提交
737
	rip += vmcs_read32(VM_EXIT_INSTRUCTION_LEN);
738
	kvm_rip_write(vcpu, rip);
A
Avi Kivity 已提交
739 740 741 742 743 744 745 746 747

	/*
	 * We emulated an instruction, so temporary interrupt blocking
	 * should be removed, if set.
	 */
	interruptibility = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO);
	if (interruptibility & 3)
		vmcs_write32(GUEST_INTERRUPTIBILITY_INFO,
			     interruptibility & ~3);
748
	vcpu->arch.interrupt_window_open = 1;
A
Avi Kivity 已提交
749 750
}

751 752 753
static void vmx_queue_exception(struct kvm_vcpu *vcpu, unsigned nr,
				bool has_error_code, u32 error_code)
{
754
	struct vcpu_vmx *vmx = to_vmx(vcpu);
755
	u32 intr_info = nr | INTR_INFO_VALID_MASK;
756

757
	if (has_error_code) {
758
		vmcs_write32(VM_ENTRY_EXCEPTION_ERROR_CODE, error_code);
759 760
		intr_info |= INTR_INFO_DELIVER_CODE_MASK;
	}
761 762 763 764 765

	if (vcpu->arch.rmode.active) {
		vmx->rmode.irq.pending = true;
		vmx->rmode.irq.vector = nr;
		vmx->rmode.irq.rip = kvm_rip_read(vcpu);
766
		if (nr == BP_VECTOR || nr == OF_VECTOR)
767
			vmx->rmode.irq.rip++;
768 769
		intr_info |= INTR_TYPE_SOFT_INTR;
		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, intr_info);
770 771 772 773 774
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
		kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1);
		return;
	}

775 776 777 778 779 780 781
	if (nr == BP_VECTOR || nr == OF_VECTOR) {
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
		intr_info |= INTR_TYPE_SOFT_EXCEPTION;
	} else
		intr_info |= INTR_TYPE_HARD_EXCEPTION;

	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, intr_info);
782 783 784 785
}

static bool vmx_exception_injected(struct kvm_vcpu *vcpu)
{
786
	return false;
787 788
}

789 790 791
/*
 * Swap MSR entry in host/guest MSR entry array.
 */
792
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
793
static void move_msr_up(struct vcpu_vmx *vmx, int from, int to)
794
{
795 796 797 798 799 800 801 802
	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;
803
}
804
#endif
805

806 807 808 809 810
/*
 * 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
811
static void setup_msrs(struct vcpu_vmx *vmx)
812
{
813
	int save_nmsrs;
814

815
	vmx_load_host_state(vmx);
816 817
	save_nmsrs = 0;
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
818
	if (is_long_mode(&vmx->vcpu)) {
819 820
		int index;

R
Rusty Russell 已提交
821
		index = __find_msr_index(vmx, MSR_SYSCALL_MASK);
822
		if (index >= 0)
R
Rusty Russell 已提交
823 824
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_LSTAR);
825
		if (index >= 0)
R
Rusty Russell 已提交
826 827
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_CSTAR);
828
		if (index >= 0)
R
Rusty Russell 已提交
829 830
			move_msr_up(vmx, index, save_nmsrs++);
		index = __find_msr_index(vmx, MSR_KERNEL_GS_BASE);
831
		if (index >= 0)
R
Rusty Russell 已提交
832
			move_msr_up(vmx, index, save_nmsrs++);
833 834 835 836
		/*
		 * MSR_K6_STAR is only needed on long mode guests, and only
		 * if efer.sce is enabled.
		 */
R
Rusty Russell 已提交
837
		index = __find_msr_index(vmx, MSR_K6_STAR);
838
		if ((index >= 0) && (vmx->vcpu.arch.shadow_efer & EFER_SCE))
R
Rusty Russell 已提交
839
			move_msr_up(vmx, index, save_nmsrs++);
840 841
	}
#endif
842
	vmx->save_nmsrs = save_nmsrs;
843

844
#ifdef CONFIG_X86_64
845
	vmx->msr_offset_kernel_gs_base =
R
Rusty Russell 已提交
846
		__find_msr_index(vmx, MSR_KERNEL_GS_BASE);
847
#endif
R
Rusty Russell 已提交
848
	vmx->msr_offset_efer = __find_msr_index(vmx, MSR_EFER);
849 850
}

A
Avi Kivity 已提交
851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883
/*
 * 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
 */
static void guest_write_tsc(u64 guest_tsc)
{
	u64 host_tsc;

	rdtscll(host_tsc);
	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;
884
	struct kvm_msr_entry *msr;
A
Avi Kivity 已提交
885 886 887 888 889 890 891

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

	switch (msr_index) {
892
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
893 894 895 896 897 898 899
	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:
900
		return kvm_get_msr_common(vcpu, msr_index, pdata);
A
Avi Kivity 已提交
901 902 903 904 905 906 907 908
#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
909
		data = vmcs_readl(GUEST_SYSENTER_EIP);
A
Avi Kivity 已提交
910 911
		break;
	case MSR_IA32_SYSENTER_ESP:
A
Avi Kivity 已提交
912
		data = vmcs_readl(GUEST_SYSENTER_ESP);
A
Avi Kivity 已提交
913 914
		break;
	default:
915
		vmx_load_host_state(to_vmx(vcpu));
R
Rusty Russell 已提交
916
		msr = find_msr_entry(to_vmx(vcpu), msr_index);
917 918 919
		if (msr) {
			data = msr->data;
			break;
A
Avi Kivity 已提交
920
		}
921
		return kvm_get_msr_common(vcpu, msr_index, pdata);
A
Avi Kivity 已提交
922 923 924 925 926 927 928 929 930 931 932 933 934
	}

	*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)
{
935 936
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	struct kvm_msr_entry *msr;
937 938
	int ret = 0;

A
Avi Kivity 已提交
939
	switch (msr_index) {
940
#ifdef CONFIG_X86_64
941
	case MSR_EFER:
942
		vmx_load_host_state(vmx);
943 944
		ret = kvm_set_msr_common(vcpu, msr_index, data);
		break;
A
Avi Kivity 已提交
945 946 947 948 949 950 951 952 953 954 955
	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
956
		vmcs_writel(GUEST_SYSENTER_EIP, data);
A
Avi Kivity 已提交
957 958
		break;
	case MSR_IA32_SYSENTER_ESP:
A
Avi Kivity 已提交
959
		vmcs_writel(GUEST_SYSENTER_ESP, data);
A
Avi Kivity 已提交
960
		break;
A
Avi Kivity 已提交
961
	case MSR_IA32_TIME_STAMP_COUNTER:
A
Avi Kivity 已提交
962
		guest_write_tsc(data);
963 964 965 966 967 968 969 970 971 972 973 974
		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
975
		break;
S
Sheng Yang 已提交
976 977 978 979 980 981 982
	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
983
	default:
984
		vmx_load_host_state(vmx);
R
Rusty Russell 已提交
985
		msr = find_msr_entry(vmx, msr_index);
986 987 988
		if (msr) {
			msr->data = data;
			break;
A
Avi Kivity 已提交
989
		}
990
		ret = kvm_set_msr_common(vcpu, msr_index, data);
A
Avi Kivity 已提交
991 992
	}

993
	return ret;
A
Avi Kivity 已提交
994 995
}

996
static void vmx_cache_reg(struct kvm_vcpu *vcpu, enum kvm_reg reg)
A
Avi Kivity 已提交
997
{
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
	__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
1009 1010
}

J
Jan Kiszka 已提交
1011
static int set_guest_debug(struct kvm_vcpu *vcpu, struct kvm_guest_debug *dbg)
A
Avi Kivity 已提交
1012
{
J
Jan Kiszka 已提交
1013 1014
	int old_debug = vcpu->guest_debug;
	unsigned long flags;
A
Avi Kivity 已提交
1015

J
Jan Kiszka 已提交
1016 1017 1018
	vcpu->guest_debug = dbg->control;
	if (!(vcpu->guest_debug & KVM_GUESTDBG_ENABLE))
		vcpu->guest_debug = 0;
A
Avi Kivity 已提交
1019

J
Jan Kiszka 已提交
1020 1021 1022 1023
	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
1024
		flags &= ~(X86_EFLAGS_TF | X86_EFLAGS_RF);
J
Jan Kiszka 已提交
1025
	vmcs_writel(GUEST_RFLAGS, flags);
A
Avi Kivity 已提交
1026

1027
	update_exception_bitmap(vcpu);
A
Avi Kivity 已提交
1028 1029 1030 1031

	return 0;
}

E
Eddie Dong 已提交
1032 1033
static int vmx_get_irq(struct kvm_vcpu *vcpu)
{
1034 1035 1036
	if (!vcpu->arch.interrupt.pending)
		return -1;
	return vcpu->arch.interrupt.nr;
E
Eddie Dong 已提交
1037 1038
}

A
Avi Kivity 已提交
1039 1040
static __init int cpu_has_kvm_support(void)
{
1041
	return cpu_has_vmx();
A
Avi Kivity 已提交
1042 1043 1044 1045 1046 1047 1048
}

static __init int vmx_disabled_by_bios(void)
{
	u64 msr;

	rdmsrl(MSR_IA32_FEATURE_CONTROL, msr);
1049 1050 1051
	return (msr & (FEATURE_CONTROL_LOCKED |
		       FEATURE_CONTROL_VMXON_ENABLED))
	    == FEATURE_CONTROL_LOCKED;
1052
	/* locked but not enabled */
A
Avi Kivity 已提交
1053 1054
}

A
Avi Kivity 已提交
1055
static void hardware_enable(void *garbage)
A
Avi Kivity 已提交
1056 1057 1058 1059 1060
{
	int cpu = raw_smp_processor_id();
	u64 phys_addr = __pa(per_cpu(vmxarea, cpu));
	u64 old;

1061
	INIT_LIST_HEAD(&per_cpu(vcpus_on_cpu, cpu));
A
Avi Kivity 已提交
1062
	rdmsrl(MSR_IA32_FEATURE_CONTROL, old);
1063 1064 1065 1066
	if ((old & (FEATURE_CONTROL_LOCKED |
		    FEATURE_CONTROL_VMXON_ENABLED))
	    != (FEATURE_CONTROL_LOCKED |
		FEATURE_CONTROL_VMXON_ENABLED))
A
Avi Kivity 已提交
1067
		/* enable and lock */
1068
		wrmsrl(MSR_IA32_FEATURE_CONTROL, old |
1069 1070
		       FEATURE_CONTROL_LOCKED |
		       FEATURE_CONTROL_VMXON_ENABLED);
1071
	write_cr4(read_cr4() | X86_CR4_VMXE); /* FIXME: not cpu hotplug safe */
1072 1073
	asm volatile (ASM_VMX_VMXON_RAX
		      : : "a"(&phys_addr), "m"(phys_addr)
A
Avi Kivity 已提交
1074 1075 1076
		      : "memory", "cc");
}

1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
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);
}

1087 1088 1089 1090 1091

/* 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
1092
{
1093
	asm volatile (__ex(ASM_VMX_VMXOFF) : : : "cc");
1094
	write_cr4(read_cr4() & ~X86_CR4_VMXE);
A
Avi Kivity 已提交
1095 1096
}

1097 1098 1099 1100 1101 1102
static void hardware_disable(void *garbage)
{
	vmclear_local_vcpus();
	kvm_cpu_vmxoff();
}

1103
static __init int adjust_vmx_controls(u32 ctl_min, u32 ctl_opt,
M
Mike Day 已提交
1104
				      u32 msr, u32 *result)
1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
{
	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
1116
		return -EIO;
1117 1118 1119 1120 1121

	*result = ctl;
	return 0;
}

Y
Yang, Sheng 已提交
1122
static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
A
Avi Kivity 已提交
1123 1124
{
	u32 vmx_msr_low, vmx_msr_high;
S
Sheng Yang 已提交
1125
	u32 min, opt, min2, opt2;
1126 1127
	u32 _pin_based_exec_control = 0;
	u32 _cpu_based_exec_control = 0;
1128
	u32 _cpu_based_2nd_exec_control = 0;
1129 1130 1131 1132
	u32 _vmexit_control = 0;
	u32 _vmentry_control = 0;

	min = PIN_BASED_EXT_INTR_MASK | PIN_BASED_NMI_EXITING;
1133
	opt = PIN_BASED_VIRTUAL_NMIS;
1134 1135
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PINBASED_CTLS,
				&_pin_based_exec_control) < 0)
Y
Yang, Sheng 已提交
1136
		return -EIO;
1137 1138 1139 1140 1141 1142

	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
1143 1144
	      CPU_BASED_CR3_LOAD_EXITING |
	      CPU_BASED_CR3_STORE_EXITING |
1145 1146
	      CPU_BASED_USE_IO_BITMAPS |
	      CPU_BASED_MOV_DR_EXITING |
M
Marcelo Tosatti 已提交
1147 1148
	      CPU_BASED_USE_TSC_OFFSETING |
	      CPU_BASED_INVLPG_EXITING;
1149
	opt = CPU_BASED_TPR_SHADOW |
S
Sheng Yang 已提交
1150
	      CPU_BASED_USE_MSR_BITMAPS |
1151
	      CPU_BASED_ACTIVATE_SECONDARY_CONTROLS;
1152 1153
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PROCBASED_CTLS,
				&_cpu_based_exec_control) < 0)
Y
Yang, Sheng 已提交
1154
		return -EIO;
1155 1156 1157 1158 1159
#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
1160
	if (_cpu_based_exec_control & CPU_BASED_ACTIVATE_SECONDARY_CONTROLS) {
S
Sheng Yang 已提交
1161 1162
		min2 = 0;
		opt2 = SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |
1163
			SECONDARY_EXEC_WBINVD_EXITING |
S
Sheng Yang 已提交
1164 1165 1166 1167
			SECONDARY_EXEC_ENABLE_VPID |
			SECONDARY_EXEC_ENABLE_EPT;
		if (adjust_vmx_controls(min2, opt2,
					MSR_IA32_VMX_PROCBASED_CTLS2,
1168 1169 1170 1171 1172 1173 1174 1175
					&_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
1176
	if (_cpu_based_2nd_exec_control & SECONDARY_EXEC_ENABLE_EPT) {
M
Marcelo Tosatti 已提交
1177 1178
		/* CR3 accesses and invlpg don't need to cause VM Exits when EPT
		   enabled */
S
Sheng Yang 已提交
1179
		min &= ~(CPU_BASED_CR3_LOAD_EXITING |
M
Marcelo Tosatti 已提交
1180 1181
			 CPU_BASED_CR3_STORE_EXITING |
			 CPU_BASED_INVLPG_EXITING);
S
Sheng Yang 已提交
1182 1183 1184 1185 1186 1187
		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);
	}
1188 1189 1190 1191 1192

	min = 0;
#ifdef CONFIG_X86_64
	min |= VM_EXIT_HOST_ADDR_SPACE_SIZE;
#endif
S
Sheng Yang 已提交
1193
	opt = VM_EXIT_SAVE_IA32_PAT | VM_EXIT_LOAD_IA32_PAT;
1194 1195
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_EXIT_CTLS,
				&_vmexit_control) < 0)
Y
Yang, Sheng 已提交
1196
		return -EIO;
1197

S
Sheng Yang 已提交
1198 1199
	min = 0;
	opt = VM_ENTRY_LOAD_IA32_PAT;
1200 1201
	if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_ENTRY_CTLS,
				&_vmentry_control) < 0)
Y
Yang, Sheng 已提交
1202
		return -EIO;
A
Avi Kivity 已提交
1203

N
Nguyen Anh Quynh 已提交
1204
	rdmsr(MSR_IA32_VMX_BASIC, vmx_msr_low, vmx_msr_high);
1205 1206 1207

	/* 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
1208
		return -EIO;
1209 1210 1211 1212

#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
1213
		return -EIO;
1214 1215 1216 1217
#endif

	/* Require Write-Back (WB) memory type for VMCS accesses. */
	if (((vmx_msr_high >> 18) & 15) != 6)
Y
Yang, Sheng 已提交
1218
		return -EIO;
1219

Y
Yang, Sheng 已提交
1220 1221 1222
	vmcs_conf->size = vmx_msr_high & 0x1fff;
	vmcs_conf->order = get_order(vmcs_config.size);
	vmcs_conf->revision_id = vmx_msr_low;
1223

Y
Yang, Sheng 已提交
1224 1225
	vmcs_conf->pin_based_exec_ctrl = _pin_based_exec_control;
	vmcs_conf->cpu_based_exec_ctrl = _cpu_based_exec_control;
1226
	vmcs_conf->cpu_based_2nd_exec_ctrl = _cpu_based_2nd_exec_control;
Y
Yang, Sheng 已提交
1227 1228
	vmcs_conf->vmexit_ctrl         = _vmexit_control;
	vmcs_conf->vmentry_ctrl        = _vmentry_control;
1229 1230

	return 0;
N
Nguyen Anh Quynh 已提交
1231
}
A
Avi Kivity 已提交
1232 1233 1234 1235 1236 1237 1238

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

1239
	pages = alloc_pages_node(node, GFP_KERNEL, vmcs_config.order);
A
Avi Kivity 已提交
1240 1241 1242
	if (!pages)
		return NULL;
	vmcs = page_address(pages);
1243 1244
	memset(vmcs, 0, vmcs_config.size);
	vmcs->revision_id = vmcs_config.revision_id; /* vmcs revision id */
A
Avi Kivity 已提交
1245 1246 1247 1248 1249
	return vmcs;
}

static struct vmcs *alloc_vmcs(void)
{
1250
	return alloc_vmcs_cpu(raw_smp_processor_id());
A
Avi Kivity 已提交
1251 1252 1253 1254
}

static void free_vmcs(struct vmcs *vmcs)
{
1255
	free_pages((unsigned long)vmcs, vmcs_config.order);
A
Avi Kivity 已提交
1256 1257
}

1258
static void free_kvm_area(void)
A
Avi Kivity 已提交
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
{
	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
1286 1287
	if (setup_vmcs_config(&vmcs_config) < 0)
		return -EIO;
1288 1289 1290 1291

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

A
Avi Kivity 已提交
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
	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];

1304
	if (vmcs_readl(sf->base) == save->base && (save->base & AR_S_MASK)) {
A
Avi Kivity 已提交
1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
		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;
1319
	struct vcpu_vmx *vmx = to_vmx(vcpu);
A
Avi Kivity 已提交
1320

1321
	vmx->emulation_required = 1;
1322
	vcpu->arch.rmode.active = 0;
A
Avi Kivity 已提交
1323

1324 1325 1326
	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
1327 1328

	flags = vmcs_readl(GUEST_RFLAGS);
1329
	flags &= ~(X86_EFLAGS_IOPL | X86_EFLAGS_VM);
1330
	flags |= (vcpu->arch.rmode.save_iopl << IOPL_SHIFT);
A
Avi Kivity 已提交
1331 1332
	vmcs_writel(GUEST_RFLAGS, flags);

1333 1334
	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
1335 1336 1337

	update_exception_bitmap(vcpu);

1338 1339 1340
	if (emulate_invalid_guest_state)
		return;

1341 1342 1343 1344
	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
1345 1346 1347 1348 1349 1350 1351 1352 1353

	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
1354
static gva_t rmode_tss_base(struct kvm *kvm)
A
Avi Kivity 已提交
1355
{
1356
	if (!kvm->arch.tss_addr) {
1357 1358 1359 1360
		gfn_t base_gfn = kvm->memslots[0].base_gfn +
				 kvm->memslots[0].npages - 3;
		return base_gfn << PAGE_SHIFT;
	}
1361
	return kvm->arch.tss_addr;
A
Avi Kivity 已提交
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
}

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);
1372 1373
	vmcs_write16(sf->selector, save->base >> 4);
	vmcs_write32(sf->base, save->base & 0xfffff);
A
Avi Kivity 已提交
1374 1375 1376 1377 1378 1379 1380
	vmcs_write32(sf->limit, 0xffff);
	vmcs_write32(sf->ar_bytes, 0xf3);
}

static void enter_rmode(struct kvm_vcpu *vcpu)
{
	unsigned long flags;
1381
	struct vcpu_vmx *vmx = to_vmx(vcpu);
A
Avi Kivity 已提交
1382

1383
	vmx->emulation_required = 1;
1384
	vcpu->arch.rmode.active = 1;
A
Avi Kivity 已提交
1385

1386
	vcpu->arch.rmode.tr.base = vmcs_readl(GUEST_TR_BASE);
A
Avi Kivity 已提交
1387 1388
	vmcs_writel(GUEST_TR_BASE, rmode_tss_base(vcpu->kvm));

1389
	vcpu->arch.rmode.tr.limit = vmcs_read32(GUEST_TR_LIMIT);
A
Avi Kivity 已提交
1390 1391
	vmcs_write32(GUEST_TR_LIMIT, RMODE_TSS_SIZE - 1);

1392
	vcpu->arch.rmode.tr.ar = vmcs_read32(GUEST_TR_AR_BYTES);
A
Avi Kivity 已提交
1393 1394 1395
	vmcs_write32(GUEST_TR_AR_BYTES, 0x008b);

	flags = vmcs_readl(GUEST_RFLAGS);
1396 1397
	vcpu->arch.rmode.save_iopl
		= (flags & X86_EFLAGS_IOPL) >> IOPL_SHIFT;
A
Avi Kivity 已提交
1398

1399
	flags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
A
Avi Kivity 已提交
1400 1401

	vmcs_writel(GUEST_RFLAGS, flags);
1402
	vmcs_writel(GUEST_CR4, vmcs_readl(GUEST_CR4) | X86_CR4_VME);
A
Avi Kivity 已提交
1403 1404
	update_exception_bitmap(vcpu);

1405 1406 1407
	if (emulate_invalid_guest_state)
		goto continue_rmode;

A
Avi Kivity 已提交
1408 1409 1410 1411 1412
	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);
1413
	vmcs_write32(GUEST_CS_LIMIT, 0xffff);
1414 1415
	if (vmcs_readl(GUEST_CS_BASE) == 0xffff0000)
		vmcs_writel(GUEST_CS_BASE, 0xf0000);
A
Avi Kivity 已提交
1416 1417
	vmcs_write16(GUEST_CS_SELECTOR, vmcs_readl(GUEST_CS_BASE) >> 4);

1418 1419 1420 1421
	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);
1422

1423
continue_rmode:
1424
	kvm_mmu_reset_context(vcpu);
1425
	init_rmode(vcpu->kvm);
A
Avi Kivity 已提交
1426 1427
}

1428
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
1429 1430 1431 1432 1433 1434 1435 1436

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",
1437
		       __func__);
A
Avi Kivity 已提交
1438 1439 1440 1441 1442
		vmcs_write32(GUEST_TR_AR_BYTES,
			     (guest_tr_ar & ~AR_TYPE_MASK)
			     | AR_TYPE_BUSY_64_TSS);
	}

1443
	vcpu->arch.shadow_efer |= EFER_LMA;
A
Avi Kivity 已提交
1444

R
Rusty Russell 已提交
1445
	find_msr_entry(to_vmx(vcpu), MSR_EFER)->data |= EFER_LMA | EFER_LME;
A
Avi Kivity 已提交
1446 1447
	vmcs_write32(VM_ENTRY_CONTROLS,
		     vmcs_read32(VM_ENTRY_CONTROLS)
1448
		     | VM_ENTRY_IA32E_MODE);
A
Avi Kivity 已提交
1449 1450 1451 1452
}

static void exit_lmode(struct kvm_vcpu *vcpu)
{
1453
	vcpu->arch.shadow_efer &= ~EFER_LMA;
A
Avi Kivity 已提交
1454 1455 1456

	vmcs_write32(VM_ENTRY_CONTROLS,
		     vmcs_read32(VM_ENTRY_CONTROLS)
1457
		     & ~VM_ENTRY_IA32E_MODE);
A
Avi Kivity 已提交
1458 1459 1460 1461
}

#endif

1462 1463 1464
static void vmx_flush_tlb(struct kvm_vcpu *vcpu)
{
	vpid_sync_vcpu_all(to_vmx(vcpu));
1465 1466
	if (vm_need_ept())
		ept_sync_context(construct_eptp(vcpu->arch.mmu.root_hpa));
1467 1468
}

1469
static void vmx_decache_cr4_guest_bits(struct kvm_vcpu *vcpu)
1470
{
1471 1472
	vcpu->arch.cr4 &= KVM_GUEST_CR4_MASK;
	vcpu->arch.cr4 |= vmcs_readl(GUEST_CR4) & ~KVM_GUEST_CR4_MASK;
1473 1474
}

1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
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,
1498
			     vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) |
1499 1500 1501 1502 1503 1504 1505 1506 1507
			     (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,
1508
			     vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) &
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527
			     ~(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
1528 1529
static void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
{
1530 1531 1532
	unsigned long hw_cr0 = (cr0 & ~KVM_GUEST_CR0_MASK) |
				KVM_VM_CR0_ALWAYS_ON;

1533 1534
	vmx_fpu_deactivate(vcpu);

1535
	if (vcpu->arch.rmode.active && (cr0 & X86_CR0_PE))
A
Avi Kivity 已提交
1536 1537
		enter_pmode(vcpu);

1538
	if (!vcpu->arch.rmode.active && !(cr0 & X86_CR0_PE))
A
Avi Kivity 已提交
1539 1540
		enter_rmode(vcpu);

1541
#ifdef CONFIG_X86_64
1542
	if (vcpu->arch.shadow_efer & EFER_LME) {
1543
		if (!is_paging(vcpu) && (cr0 & X86_CR0_PG))
A
Avi Kivity 已提交
1544
			enter_lmode(vcpu);
1545
		if (is_paging(vcpu) && !(cr0 & X86_CR0_PG))
A
Avi Kivity 已提交
1546 1547 1548 1549
			exit_lmode(vcpu);
	}
#endif

1550 1551 1552
	if (vm_need_ept())
		ept_update_paging_mode_cr0(&hw_cr0, cr0, vcpu);

A
Avi Kivity 已提交
1553
	vmcs_writel(CR0_READ_SHADOW, cr0);
1554
	vmcs_writel(GUEST_CR0, hw_cr0);
1555
	vcpu->arch.cr0 = cr0;
1556

1557
	if (!(cr0 & X86_CR0_TS) || !(cr0 & X86_CR0_PE))
1558
		vmx_fpu_activate(vcpu);
A
Avi Kivity 已提交
1559 1560
}

1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572
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
1573 1574
static void vmx_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
{
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587
	unsigned long guest_cr3;
	u64 eptp;

	guest_cr3 = cr3;
	if (vm_need_ept()) {
		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;
	}

1588
	vmx_flush_tlb(vcpu);
1589
	vmcs_writel(GUEST_CR3, guest_cr3);
1590
	if (vcpu->arch.cr0 & X86_CR0_PE)
1591
		vmx_fpu_deactivate(vcpu);
A
Avi Kivity 已提交
1592 1593 1594 1595
}

static void vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
{
1596 1597 1598
	unsigned long hw_cr4 = cr4 | (vcpu->arch.rmode.active ?
		    KVM_RMODE_VM_CR4_ALWAYS_ON : KVM_PMODE_VM_CR4_ALWAYS_ON);

1599
	vcpu->arch.cr4 = cr4;
1600 1601 1602 1603 1604
	if (vm_need_ept())
		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
1605 1606 1607 1608
}

static void vmx_set_efer(struct kvm_vcpu *vcpu, u64 efer)
{
R
Rusty Russell 已提交
1609 1610
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	struct kvm_msr_entry *msr = find_msr_entry(vmx, MSR_EFER);
A
Avi Kivity 已提交
1611

1612
	vcpu->arch.shadow_efer = efer;
1613 1614
	if (!msr)
		return;
A
Avi Kivity 已提交
1615 1616 1617
	if (efer & EFER_LMA) {
		vmcs_write32(VM_ENTRY_CONTROLS,
				     vmcs_read32(VM_ENTRY_CONTROLS) |
1618
				     VM_ENTRY_IA32E_MODE);
A
Avi Kivity 已提交
1619 1620 1621 1622 1623
		msr->data = efer;

	} else {
		vmcs_write32(VM_ENTRY_CONTROLS,
				     vmcs_read32(VM_ENTRY_CONTROLS) &
1624
				     ~VM_ENTRY_IA32E_MODE);
A
Avi Kivity 已提交
1625 1626 1627

		msr->data = efer & ~EFER_LME;
	}
R
Rusty Russell 已提交
1628
	setup_msrs(vmx);
A
Avi Kivity 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660
}

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);
	if (ar & AR_UNUSABLE_MASK)
		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;
}

1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674
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;
}

1675
static u32 vmx_segment_access_rights(struct kvm_segment *var)
A
Avi Kivity 已提交
1676 1677 1678
{
	u32 ar;

1679
	if (var->unusable)
A
Avi Kivity 已提交
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690
		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;
	}
1691 1692
	if (ar == 0) /* a 0 value means unusable */
		ar = AR_UNUSABLE_MASK;
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702

	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;

1703 1704 1705 1706 1707
	if (vcpu->arch.rmode.active && seg == VCPU_SREG_TR) {
		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);
1708 1709 1710 1711 1712
		return;
	}
	vmcs_writel(sf->base, var->base);
	vmcs_write32(sf->limit, var->limit);
	vmcs_write16(sf->selector, var->selector);
1713
	if (vcpu->arch.rmode.active && var->s) {
1714 1715 1716 1717 1718 1719 1720 1721
		/*
		 * 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
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756
	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);
}

1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 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 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 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936
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;

	if (~cs.type & (AR_TYPE_CODE_MASK|AR_TYPE_ACCESSES_MASK))
		return false;
	if (!cs.s)
		return false;
	if (!(~cs.type & (AR_TYPE_CODE_MASK|AR_TYPE_WRITEABLE_MASK))) {
		if (cs.dpl > cs_rpl)
			return false;
	} else if (cs.type & AR_TYPE_CODE_MASK) {
		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;

	if ((ss.type != 3) || (ss.type != 7))
		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;

	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);

	if (tr.selector & SELECTOR_TI_MASK)	/* TI = 1 */
		return false;
	if ((tr.type != 3) || (tr.type != 11)) /* TODO: Check if guest is in IA32e mode */
		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);

	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
1937
static int init_rmode_tss(struct kvm *kvm)
A
Avi Kivity 已提交
1938 1939
{
	gfn_t fn = rmode_tss_base(kvm) >> PAGE_SHIFT;
1940
	u16 data = 0;
1941
	int ret = 0;
1942
	int r;
A
Avi Kivity 已提交
1943

1944 1945
	r = kvm_clear_guest_page(kvm, fn, 0, PAGE_SIZE);
	if (r < 0)
1946
		goto out;
1947
	data = TSS_BASE_SIZE + TSS_REDIRECTION_SIZE;
1948 1949
	r = kvm_write_guest_page(kvm, fn++, &data,
			TSS_IOPB_BASE_OFFSET, sizeof(u16));
1950
	if (r < 0)
1951
		goto out;
1952 1953
	r = kvm_clear_guest_page(kvm, fn++, 0, PAGE_SIZE);
	if (r < 0)
1954
		goto out;
1955 1956
	r = kvm_clear_guest_page(kvm, fn, 0, PAGE_SIZE);
	if (r < 0)
1957
		goto out;
1958
	data = ~0;
1959 1960 1961
	r = kvm_write_guest_page(kvm, fn, &data,
				 RMODE_TSS_SIZE - 2 * PAGE_SIZE - 1,
				 sizeof(u8));
1962
	if (r < 0)
1963 1964 1965 1966 1967
		goto out;

	ret = 1;
out:
	return ret;
A
Avi Kivity 已提交
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 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
static int init_rmode_identity_map(struct kvm *kvm)
{
	int i, r, ret;
	pfn_t identity_map_pfn;
	u32 tmp;

	if (!vm_need_ept())
		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
2005 2006 2007 2008 2009 2010 2011
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);
2012
	vmcs_write32(sf->ar_bytes, 0xf3);
A
Avi Kivity 已提交
2013 2014
}

2015 2016 2017 2018 2019
static int alloc_apic_access_page(struct kvm *kvm)
{
	struct kvm_userspace_memory_region kvm_userspace_mem;
	int r = 0;

2020
	down_write(&kvm->slots_lock);
2021
	if (kvm->arch.apic_access_page)
2022 2023 2024 2025 2026 2027 2028 2029
		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;
2030

2031
	kvm->arch.apic_access_page = gfn_to_page(kvm, 0xfee00);
2032
out:
2033
	up_write(&kvm->slots_lock);
2034 2035 2036
	return r;
}

2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059
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;
}

2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075
static void allocate_vpid(struct vcpu_vmx *vmx)
{
	int vpid;

	vmx->vpid = 0;
	if (!enable_vpid || !cpu_has_vmx_vpid())
		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);
}

2076
static void vmx_disable_intercept_for_msr(struct page *msr_bitmap, u32 msr)
S
Sheng Yang 已提交
2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
{
	void *va;

	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.
	 */
	va = kmap(msr_bitmap);
	if (msr <= 0x1fff) {
		__clear_bit(msr, va + 0x000); /* read-low */
		__clear_bit(msr, va + 0x800); /* write-low */
	} else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff)) {
		msr &= 0x1fff;
		__clear_bit(msr, va + 0x400); /* read-high */
		__clear_bit(msr, va + 0xc00); /* write-high */
	}
	kunmap(msr_bitmap);
}

A
Avi Kivity 已提交
2100 2101 2102
/*
 * Sets up the vmcs for emulated real mode.
 */
R
Rusty Russell 已提交
2103
static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
A
Avi Kivity 已提交
2104
{
S
Sheng Yang 已提交
2105
	u32 host_sysenter_cs, msr_low, msr_high;
A
Avi Kivity 已提交
2106
	u32 junk;
S
Sheng Yang 已提交
2107
	u64 host_pat;
A
Avi Kivity 已提交
2108 2109 2110
	unsigned long a;
	struct descriptor_table dt;
	int i;
2111
	unsigned long kvm_vmx_return;
2112
	u32 exec_control;
A
Avi Kivity 已提交
2113 2114

	/* I/O */
2115 2116
	vmcs_write64(IO_BITMAP_A, page_to_phys(vmx_io_bitmap_a));
	vmcs_write64(IO_BITMAP_B, page_to_phys(vmx_io_bitmap_b));
A
Avi Kivity 已提交
2117

S
Sheng Yang 已提交
2118 2119 2120
	if (cpu_has_vmx_msr_bitmap())
		vmcs_write64(MSR_BITMAP, page_to_phys(vmx_msr_bitmap));

A
Avi Kivity 已提交
2121 2122 2123
	vmcs_write64(VMCS_LINK_POINTER, -1ull); /* 22.3.1.5 */

	/* Control */
2124 2125
	vmcs_write32(PIN_BASED_VM_EXEC_CONTROL,
		vmcs_config.pin_based_exec_ctrl);
2126 2127 2128 2129 2130 2131 2132 2133 2134

	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
	}
S
Sheng Yang 已提交
2135 2136
	if (!vm_need_ept())
		exec_control |= CPU_BASED_CR3_STORE_EXITING |
2137 2138
				CPU_BASED_CR3_LOAD_EXITING  |
				CPU_BASED_INVLPG_EXITING;
2139
	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, exec_control);
A
Avi Kivity 已提交
2140

2141 2142 2143 2144 2145
	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;
2146 2147
		if (vmx->vpid == 0)
			exec_control &= ~SECONDARY_EXEC_ENABLE_VPID;
S
Sheng Yang 已提交
2148 2149
		if (!vm_need_ept())
			exec_control &= ~SECONDARY_EXEC_ENABLE_EPT;
2150 2151
		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
	}
2152

2153 2154
	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
2155 2156 2157 2158 2159 2160 2161 2162 2163
	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 */
2164 2165
	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
2166
	vmcs_write16(HOST_SS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
2167
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178
	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 */

2179
	kvm_get_idt(&dt);
A
Avi Kivity 已提交
2180 2181
	vmcs_writel(HOST_IDTR_BASE, dt.base);   /* 22.2.4 */

M
Mike Day 已提交
2182
	asm("mov $.Lkvm_vmx_return, %0" : "=r"(kvm_vmx_return));
2183
	vmcs_writel(HOST_RIP, kvm_vmx_return); /* 22.2.5 */
2184 2185 2186
	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
2187 2188 2189 2190 2191 2192 2193 2194

	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
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
	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
2209 2210 2211 2212
	for (i = 0; i < NR_VMX_MSR; ++i) {
		u32 index = vmx_msr_index[i];
		u32 data_low, data_high;
		u64 data;
2213
		int j = vmx->nmsrs;
A
Avi Kivity 已提交
2214 2215 2216

		if (rdmsr_safe(index, &data_low, &data_high) < 0)
			continue;
2217 2218
		if (wrmsr_safe(index, data_low, data_high) < 0)
			continue;
A
Avi Kivity 已提交
2219
		data = data_low | ((u64)data_high << 32);
2220 2221 2222 2223 2224
		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
2225 2226
	}

2227
	vmcs_write32(VM_EXIT_CONTROLS, vmcs_config.vmexit_ctrl);
A
Avi Kivity 已提交
2228 2229

	/* 22.2.1, 20.8.1 */
2230 2231
	vmcs_write32(VM_ENTRY_CONTROLS, vmcs_config.vmentry_ctrl);

2232 2233 2234
	vmcs_writel(CR0_GUEST_HOST_MASK, ~0UL);
	vmcs_writel(CR4_GUEST_HOST_MASK, KVM_GUEST_CR4_MASK);

2235

2236 2237 2238
	return 0;
}

2239 2240 2241 2242 2243 2244 2245 2246 2247
static int init_rmode(struct kvm *kvm)
{
	if (!init_rmode_tss(kvm))
		return 0;
	if (!init_rmode_identity_map(kvm))
		return 0;
	return 1;
}

2248 2249 2250 2251 2252 2253
static int vmx_vcpu_reset(struct kvm_vcpu *vcpu)
{
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	u64 msr;
	int ret;

2254
	vcpu->arch.regs_avail = ~((1 << VCPU_REGS_RIP) | (1 << VCPU_REGS_RSP));
2255
	down_read(&vcpu->kvm->slots_lock);
2256
	if (!init_rmode(vmx->vcpu.kvm)) {
2257 2258 2259 2260
		ret = -ENOMEM;
		goto out;
	}

2261
	vmx->vcpu.arch.rmode.active = 0;
2262

2263 2264
	vmx->soft_vnmi_blocked = 0;

2265
	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
2266
	kvm_set_cr8(&vmx->vcpu, 0);
2267 2268 2269 2270 2271 2272 2273
	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);

2274
	seg_setup(VCPU_SREG_CS);
2275 2276 2277 2278 2279 2280 2281 2282
	/*
	 * 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 {
2283 2284
		vmcs_write16(GUEST_CS_SELECTOR, vmx->vcpu.arch.sipi_vector << 8);
		vmcs_writel(GUEST_CS_BASE, vmx->vcpu.arch.sipi_vector << 12);
2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
	}

	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)
2309
		kvm_rip_write(vcpu, 0xfff0);
2310
	else
2311 2312
		kvm_rip_write(vcpu, 0);
	kvm_register_write(vcpu, VCPU_REGS_RSP, 0);
2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332

	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);

	guest_write_tsc(0);

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

	setup_msrs(vmx);

A
Avi Kivity 已提交
2333 2334
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, 0);  /* 22.2.1 */

2335 2336 2337 2338
	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,
2339
				page_to_phys(vmx->vcpu.arch.apic->regs_page));
2340 2341 2342 2343 2344
		vmcs_write32(TPR_THRESHOLD, 0);
	}

	if (vm_need_virtualize_apic_accesses(vmx->vcpu.kvm))
		vmcs_write64(APIC_ACCESS_ADDR,
2345
			     page_to_phys(vmx->vcpu.kvm->arch.apic_access_page));
A
Avi Kivity 已提交
2346

2347 2348 2349
	if (vmx->vpid != 0)
		vmcs_write16(VIRTUAL_PROCESSOR_ID, vmx->vpid);

2350 2351
	vmx->vcpu.arch.cr0 = 0x60000010;
	vmx_set_cr0(&vmx->vcpu, vmx->vcpu.arch.cr0); /* enter rmode */
R
Rusty Russell 已提交
2352 2353 2354 2355
	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
2356

2357 2358
	vpid_sync_vcpu_all(vmx);

2359
	ret = 0;
A
Avi Kivity 已提交
2360

2361 2362 2363
	/* HACK: Don't enable emulation on guest boot/reset */
	vmx->emulation_required = 0;

A
Avi Kivity 已提交
2364
out:
2365
	up_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
2366 2367 2368
	return ret;
}

2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391
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);
}

2392 2393
static void vmx_inject_irq(struct kvm_vcpu *vcpu, int irq)
{
2394 2395
	struct vcpu_vmx *vmx = to_vmx(vcpu);

F
Feng (Eric) Liu 已提交
2396 2397
	KVMTRACE_1D(INJ_VIRQ, vcpu, (u32)irq, handler);

2398
	++vcpu->stat.irq_injections;
2399
	if (vcpu->arch.rmode.active) {
2400 2401
		vmx->rmode.irq.pending = true;
		vmx->rmode.irq.vector = irq;
2402
		vmx->rmode.irq.rip = kvm_rip_read(vcpu);
2403 2404 2405
		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			     irq | INTR_TYPE_SOFT_INTR | INTR_INFO_VALID_MASK);
		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1);
2406
		kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1);
2407 2408 2409 2410 2411 2412
		return;
	}
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			irq | INTR_TYPE_EXT_INTR | INTR_INFO_VALID_MASK);
}

2413 2414
static void vmx_inject_nmi(struct kvm_vcpu *vcpu)
{
J
Jan Kiszka 已提交
2415 2416
	struct vcpu_vmx *vmx = to_vmx(vcpu);

2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
	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;
	}

2430
	++vcpu->stat.nmi_injections;
J
Jan Kiszka 已提交
2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
	if (vcpu->arch.rmode.active) {
		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;
	}
2442 2443 2444 2445
	vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
			INTR_TYPE_NMI_INTR | INTR_INFO_VALID_MASK | NMI_VECTOR);
}

2446 2447 2448 2449 2450 2451 2452 2453
static void vmx_update_window_states(struct kvm_vcpu *vcpu)
{
	u32 guest_intr = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO);

	vcpu->arch.nmi_window_open =
		!(guest_intr & (GUEST_INTR_STATE_STI |
				GUEST_INTR_STATE_MOV_SS |
				GUEST_INTR_STATE_NMI));
2454 2455
	if (!cpu_has_virtual_nmis() && to_vmx(vcpu)->soft_vnmi_blocked)
		vcpu->arch.nmi_window_open = 0;
2456 2457 2458 2459 2460 2461 2462

	vcpu->arch.interrupt_window_open =
		((vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF) &&
		 !(guest_intr & (GUEST_INTR_STATE_STI |
				 GUEST_INTR_STATE_MOV_SS)));
}

A
Avi Kivity 已提交
2463 2464
static void kvm_do_inject_irq(struct kvm_vcpu *vcpu)
{
2465 2466
	int word_index = __ffs(vcpu->arch.irq_summary);
	int bit_index = __ffs(vcpu->arch.irq_pending[word_index]);
A
Avi Kivity 已提交
2467 2468
	int irq = word_index * BITS_PER_LONG + bit_index;

2469 2470 2471
	clear_bit(bit_index, &vcpu->arch.irq_pending[word_index]);
	if (!vcpu->arch.irq_pending[word_index])
		clear_bit(word_index, &vcpu->arch.irq_summary);
2472
	kvm_queue_interrupt(vcpu, irq);
A
Avi Kivity 已提交
2473 2474
}

2475 2476 2477 2478 2479
static void do_interrupt_requests(struct kvm_vcpu *vcpu,
				       struct kvm_run *kvm_run)
{
	vmx_update_window_states(vcpu);

2480 2481 2482 2483 2484
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
		vmcs_clear_bits(GUEST_INTERRUPTIBILITY_INFO,
				GUEST_INTR_STATE_STI |
				GUEST_INTR_STATE_MOV_SS);

2485
	if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) {
2486 2487 2488
		if (vcpu->arch.interrupt.pending) {
			enable_nmi_window(vcpu);
		} else if (vcpu->arch.nmi_window_open) {
2489 2490 2491 2492
			vcpu->arch.nmi_pending = false;
			vcpu->arch.nmi_injected = true;
		} else {
			enable_nmi_window(vcpu);
2493 2494
			return;
		}
2495 2496 2497
	}
	if (vcpu->arch.nmi_injected) {
		vmx_inject_nmi(vcpu);
2498
		if (vcpu->arch.nmi_pending)
2499
			enable_nmi_window(vcpu);
2500 2501 2502 2503
		else if (vcpu->arch.irq_summary
			 || kvm_run->request_interrupt_window)
			enable_irq_window(vcpu);
		return;
2504 2505
	}

2506 2507 2508 2509 2510 2511 2512
	if (vcpu->arch.interrupt_window_open) {
		if (vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending)
			kvm_do_inject_irq(vcpu);

		if (vcpu->arch.interrupt.pending)
			vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr);
	}
2513 2514
	if (!vcpu->arch.interrupt_window_open &&
	    (vcpu->arch.irq_summary || kvm_run->request_interrupt_window))
2515
		enable_irq_window(vcpu);
A
Avi Kivity 已提交
2516 2517
}

2518 2519 2520 2521
static int vmx_set_tss_addr(struct kvm *kvm, unsigned int addr)
{
	int ret;
	struct kvm_userspace_memory_region tss_mem = {
2522
		.slot = TSS_PRIVATE_MEMSLOT,
2523 2524 2525 2526 2527 2528 2529 2530
		.guest_phys_addr = addr,
		.memory_size = PAGE_SIZE * 3,
		.flags = 0,
	};

	ret = kvm_set_memory_region(kvm, &tss_mem, 0);
	if (ret)
		return ret;
2531
	kvm->arch.tss_addr = addr;
2532 2533 2534
	return 0;
}

A
Avi Kivity 已提交
2535 2536 2537
static int handle_rmode_exception(struct kvm_vcpu *vcpu,
				  int vec, u32 err_code)
{
2538 2539 2540 2541 2542
	/*
	 * 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)
2543
		if (emulate_instruction(vcpu, NULL, 0, 0, 0) == EMULATE_DONE)
A
Avi Kivity 已提交
2544
			return 1;
2545 2546 2547 2548 2549 2550 2551
	/*
	 * 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
2552 2553 2554 2555 2556
		if (vcpu->guest_debug &
		    (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))
			return 0;
		kvm_queue_exception(vcpu, vec);
		return 1;
2557
	case BP_VECTOR:
J
Jan Kiszka 已提交
2558 2559 2560 2561
		if (vcpu->guest_debug & KVM_GUESTDBG_USE_SW_BP)
			return 0;
		/* fall through */
	case DE_VECTOR:
2562 2563 2564 2565 2566 2567 2568 2569 2570 2571
	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
2572 2573 2574 2575 2576
	return 0;
}

static int handle_exception(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2577
	struct vcpu_vmx *vmx = to_vmx(vcpu);
J
Jan Kiszka 已提交
2578
	u32 intr_info, ex_no, error_code;
2579
	unsigned long cr2, rip, dr6;
A
Avi Kivity 已提交
2580 2581 2582
	u32 vect_info;
	enum emulation_result er;

2583
	vect_info = vmx->idt_vectoring_info;
A
Avi Kivity 已提交
2584 2585 2586
	intr_info = vmcs_read32(VM_EXIT_INTR_INFO);

	if ((vect_info & VECTORING_INFO_VALID_MASK) &&
M
Mike Day 已提交
2587
						!is_page_fault(intr_info))
A
Avi Kivity 已提交
2588
		printk(KERN_ERR "%s: unexpected, vectoring info 0x%x "
2589
		       "intr info 0x%x\n", __func__, vect_info, intr_info);
A
Avi Kivity 已提交
2590

2591
	if (!irqchip_in_kernel(vcpu->kvm) && is_external_interrupt(vect_info)) {
A
Avi Kivity 已提交
2592
		int irq = vect_info & VECTORING_INFO_VECTOR_MASK;
2593 2594
		set_bit(irq, vcpu->arch.irq_pending);
		set_bit(irq / BITS_PER_LONG, &vcpu->arch.irq_summary);
A
Avi Kivity 已提交
2595 2596
	}

2597
	if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR)
2598
		return 1;  /* already handled by vmx_vcpu_run() */
2599 2600

	if (is_no_device(intr_info)) {
2601
		vmx_fpu_activate(vcpu);
2602 2603 2604
		return 1;
	}

2605
	if (is_invalid_opcode(intr_info)) {
2606
		er = emulate_instruction(vcpu, kvm_run, 0, 0, EMULTYPE_TRAP_UD);
2607
		if (er != EMULATE_DONE)
2608
			kvm_queue_exception(vcpu, UD_VECTOR);
2609 2610 2611
		return 1;
	}

A
Avi Kivity 已提交
2612
	error_code = 0;
2613
	rip = kvm_rip_read(vcpu);
2614
	if (intr_info & INTR_INFO_DELIVER_CODE_MASK)
A
Avi Kivity 已提交
2615 2616
		error_code = vmcs_read32(VM_EXIT_INTR_ERROR_CODE);
	if (is_page_fault(intr_info)) {
2617 2618 2619
		/* EPT won't cause page fault directly */
		if (vm_need_ept())
			BUG();
A
Avi Kivity 已提交
2620
		cr2 = vmcs_readl(EXIT_QUALIFICATION);
F
Feng (Eric) Liu 已提交
2621 2622
		KVMTRACE_3D(PAGE_FAULT, vcpu, error_code, (u32)cr2,
			    (u32)((u64)cr2 >> 32), handler);
2623
		if (vcpu->arch.interrupt.pending || vcpu->arch.exception.pending)
2624
			kvm_mmu_unprotect_page_virt(vcpu, cr2);
2625
		return kvm_mmu_page_fault(vcpu, cr2, error_code);
A
Avi Kivity 已提交
2626 2627
	}

2628
	if (vcpu->arch.rmode.active &&
A
Avi Kivity 已提交
2629
	    handle_rmode_exception(vcpu, intr_info & INTR_INFO_VECTOR_MASK,
2630
								error_code)) {
2631 2632
		if (vcpu->arch.halt_request) {
			vcpu->arch.halt_request = 0;
2633 2634
			return kvm_emulate_halt(vcpu);
		}
A
Avi Kivity 已提交
2635
		return 1;
2636
	}
A
Avi Kivity 已提交
2637

J
Jan Kiszka 已提交
2638
	ex_no = intr_info & INTR_INFO_VECTOR_MASK;
2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651
	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
2652
		kvm_run->exit_reason = KVM_EXIT_DEBUG;
J
Jan Kiszka 已提交
2653 2654
		kvm_run->debug.arch.pc = vmcs_readl(GUEST_CS_BASE) + rip;
		kvm_run->debug.arch.exception = ex_no;
2655 2656
		break;
	default:
J
Jan Kiszka 已提交
2657 2658 2659
		kvm_run->exit_reason = KVM_EXIT_EXCEPTION;
		kvm_run->ex.exception = ex_no;
		kvm_run->ex.error_code = error_code;
2660
		break;
A
Avi Kivity 已提交
2661 2662 2663 2664 2665 2666 2667
	}
	return 0;
}

static int handle_external_interrupt(struct kvm_vcpu *vcpu,
				     struct kvm_run *kvm_run)
{
A
Avi Kivity 已提交
2668
	++vcpu->stat.irq_exits;
F
Feng (Eric) Liu 已提交
2669
	KVMTRACE_1D(INTR, vcpu, vmcs_read32(VM_EXIT_INTR_INFO), handler);
A
Avi Kivity 已提交
2670 2671 2672
	return 1;
}

2673 2674 2675 2676 2677
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
2678 2679 2680

static int handle_io(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2681
	unsigned long exit_qualification;
2682 2683
	int size, down, in, string, rep;
	unsigned port;
A
Avi Kivity 已提交
2684

A
Avi Kivity 已提交
2685
	++vcpu->stat.io_exits;
2686
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
2687
	string = (exit_qualification & 16) != 0;
2688 2689

	if (string) {
2690 2691
		if (emulate_instruction(vcpu,
					kvm_run, 0, 0, 0) == EMULATE_DO_MMIO)
2692 2693 2694 2695 2696 2697
			return 0;
		return 1;
	}

	size = (exit_qualification & 7) + 1;
	in = (exit_qualification & 8) != 0;
2698 2699 2700
	down = (vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_DF) != 0;
	rep = (exit_qualification & 32) != 0;
	port = exit_qualification >> 16;
2701

2702
	skip_emulated_instruction(vcpu);
L
Laurent Vivier 已提交
2703
	return kvm_emulate_pio(vcpu, kvm_run, in, size, port);
A
Avi Kivity 已提交
2704 2705
}

I
Ingo Molnar 已提交
2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716
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
2717 2718
static int handle_cr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2719
	unsigned long exit_qualification;
A
Avi Kivity 已提交
2720 2721 2722
	int cr;
	int reg;

2723
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
A
Avi Kivity 已提交
2724 2725 2726 2727
	cr = exit_qualification & 15;
	reg = (exit_qualification >> 8) & 15;
	switch ((exit_qualification >> 4) & 3) {
	case 0: /* mov to cr */
2728 2729 2730 2731
		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
2732 2733
		switch (cr) {
		case 0:
2734
			kvm_set_cr0(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2735 2736 2737
			skip_emulated_instruction(vcpu);
			return 1;
		case 3:
2738
			kvm_set_cr3(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2739 2740 2741
			skip_emulated_instruction(vcpu);
			return 1;
		case 4:
2742
			kvm_set_cr4(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2743 2744 2745
			skip_emulated_instruction(vcpu);
			return 1;
		case 8:
2746
			kvm_set_cr8(vcpu, kvm_register_read(vcpu, reg));
A
Avi Kivity 已提交
2747
			skip_emulated_instruction(vcpu);
2748 2749
			if (irqchip_in_kernel(vcpu->kvm))
				return 1;
2750 2751
			kvm_run->exit_reason = KVM_EXIT_SET_TPR;
			return 0;
A
Avi Kivity 已提交
2752 2753
		};
		break;
2754
	case 2: /* clts */
2755
		vmx_fpu_deactivate(vcpu);
2756 2757
		vcpu->arch.cr0 &= ~X86_CR0_TS;
		vmcs_writel(CR0_READ_SHADOW, vcpu->arch.cr0);
2758
		vmx_fpu_activate(vcpu);
F
Feng (Eric) Liu 已提交
2759
		KVMTRACE_0D(CLTS, vcpu, handler);
2760 2761
		skip_emulated_instruction(vcpu);
		return 1;
A
Avi Kivity 已提交
2762 2763 2764
	case 1: /*mov from cr*/
		switch (cr) {
		case 3:
2765
			kvm_register_write(vcpu, reg, vcpu->arch.cr3);
F
Feng (Eric) Liu 已提交
2766
			KVMTRACE_3D(CR_READ, vcpu, (u32)cr,
2767 2768
				    (u32)kvm_register_read(vcpu, reg),
				    (u32)((u64)kvm_register_read(vcpu, reg) >> 32),
F
Feng (Eric) Liu 已提交
2769
				    handler);
A
Avi Kivity 已提交
2770 2771 2772
			skip_emulated_instruction(vcpu);
			return 1;
		case 8:
2773
			kvm_register_write(vcpu, reg, kvm_get_cr8(vcpu));
F
Feng (Eric) Liu 已提交
2774
			KVMTRACE_2D(CR_READ, vcpu, (u32)cr,
2775
				    (u32)kvm_register_read(vcpu, reg), handler);
A
Avi Kivity 已提交
2776 2777 2778 2779 2780
			skip_emulated_instruction(vcpu);
			return 1;
		}
		break;
	case 3: /* lmsw */
2781
		kvm_lmsw(vcpu, (exit_qualification >> LMSW_SOURCE_DATA_SHIFT) & 0x0f);
A
Avi Kivity 已提交
2782 2783 2784 2785 2786 2787 2788

		skip_emulated_instruction(vcpu);
		return 1;
	default:
		break;
	}
	kvm_run->exit_reason = 0;
2789
	pr_unimpl(vcpu, "unhandled control register: op %d cr %d\n",
A
Avi Kivity 已提交
2790 2791 2792 2793 2794 2795
	       (int)(exit_qualification >> 4) & 3, cr);
	return 0;
}

static int handle_dr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2796
	unsigned long exit_qualification;
A
Avi Kivity 已提交
2797 2798 2799
	unsigned long val;
	int dr, reg;

2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824
	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;
		}
	}

2825
	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
2826 2827 2828
	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
2829
		switch (dr) {
2830 2831 2832
		case 0 ... 3:
			val = vcpu->arch.db[dr];
			break;
A
Avi Kivity 已提交
2833
		case 6:
2834
			val = vcpu->arch.dr6;
A
Avi Kivity 已提交
2835 2836
			break;
		case 7:
2837
			val = vcpu->arch.dr7;
A
Avi Kivity 已提交
2838 2839 2840 2841
			break;
		default:
			val = 0;
		}
2842
		kvm_register_write(vcpu, reg, val);
F
Feng (Eric) Liu 已提交
2843
		KVMTRACE_2D(DR_READ, vcpu, (u32)dr, (u32)val, handler);
A
Avi Kivity 已提交
2844
	} else {
2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876
		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
2877 2878 2879 2880 2881 2882 2883
	}
	skip_emulated_instruction(vcpu);
	return 1;
}

static int handle_cpuid(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2884 2885
	kvm_emulate_cpuid(vcpu);
	return 1;
A
Avi Kivity 已提交
2886 2887 2888 2889
}

static int handle_rdmsr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2890
	u32 ecx = vcpu->arch.regs[VCPU_REGS_RCX];
A
Avi Kivity 已提交
2891 2892 2893
	u64 data;

	if (vmx_get_msr(vcpu, ecx, &data)) {
2894
		kvm_inject_gp(vcpu, 0);
A
Avi Kivity 已提交
2895 2896 2897
		return 1;
	}

F
Feng (Eric) Liu 已提交
2898 2899 2900
	KVMTRACE_3D(MSR_READ, vcpu, ecx, (u32)data, (u32)(data >> 32),
		    handler);

A
Avi Kivity 已提交
2901
	/* FIXME: handling of bits 32:63 of rax, rdx */
2902 2903
	vcpu->arch.regs[VCPU_REGS_RAX] = data & -1u;
	vcpu->arch.regs[VCPU_REGS_RDX] = (data >> 32) & -1u;
A
Avi Kivity 已提交
2904 2905 2906 2907 2908 2909
	skip_emulated_instruction(vcpu);
	return 1;
}

static int handle_wrmsr(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2910 2911 2912
	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
2913

F
Feng (Eric) Liu 已提交
2914 2915 2916
	KVMTRACE_3D(MSR_WRITE, vcpu, ecx, (u32)data, (u32)(data >> 32),
		    handler);

A
Avi Kivity 已提交
2917
	if (vmx_set_msr(vcpu, ecx, data) != 0) {
2918
		kvm_inject_gp(vcpu, 0);
A
Avi Kivity 已提交
2919 2920 2921 2922 2923 2924 2925
		return 1;
	}

	skip_emulated_instruction(vcpu);
	return 1;
}

2926 2927 2928 2929 2930 2931
static int handle_tpr_below_threshold(struct kvm_vcpu *vcpu,
				      struct kvm_run *kvm_run)
{
	return 1;
}

A
Avi Kivity 已提交
2932 2933 2934
static int handle_interrupt_window(struct kvm_vcpu *vcpu,
				   struct kvm_run *kvm_run)
{
2935 2936 2937 2938 2939 2940
	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
2941 2942

	KVMTRACE_0D(PEND_INTR, vcpu, handler);
2943
	++vcpu->stat.irq_window_exits;
F
Feng (Eric) Liu 已提交
2944

2945 2946 2947 2948 2949
	/*
	 * If the user space waits to inject interrupts, exit as soon as
	 * possible
	 */
	if (kvm_run->request_interrupt_window &&
2950
	    !vcpu->arch.irq_summary) {
2951 2952 2953
		kvm_run->exit_reason = KVM_EXIT_IRQ_WINDOW_OPEN;
		return 0;
	}
A
Avi Kivity 已提交
2954 2955 2956 2957 2958 2959
	return 1;
}

static int handle_halt(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	skip_emulated_instruction(vcpu);
2960
	return kvm_emulate_halt(vcpu);
A
Avi Kivity 已提交
2961 2962
}

2963 2964
static int handle_vmcall(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
2965
	skip_emulated_instruction(vcpu);
2966 2967
	kvm_emulate_hypercall(vcpu);
	return 1;
2968 2969
}

M
Marcelo Tosatti 已提交
2970 2971 2972 2973 2974 2975 2976 2977 2978
static int handle_invlpg(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	u64 exit_qualification = vmcs_read64(EXIT_QUALIFICATION);

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

E
Eddie Dong 已提交
2979 2980 2981 2982 2983 2984 2985
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;
}

2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005
static int handle_apic_access(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	u64 exit_qualification;
	enum emulation_result er;
	unsigned long offset;

	exit_qualification = vmcs_read64(EXIT_QUALIFICATION);
	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;
}

3006 3007
static int handle_task_switch(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
J
Jan Kiszka 已提交
3008
	struct vcpu_vmx *vmx = to_vmx(vcpu);
3009 3010 3011 3012 3013 3014 3015
	unsigned long exit_qualification;
	u16 tss_selector;
	int reason;

	exit_qualification = vmcs_readl(EXIT_QUALIFICATION);

	reason = (u32)exit_qualification >> 30;
J
Jan Kiszka 已提交
3016 3017 3018 3019 3020 3021 3022 3023 3024
	if (reason == TASK_SWITCH_GATE && vmx->vcpu.arch.nmi_injected &&
	    (vmx->idt_vectoring_info & VECTORING_INFO_VALID_MASK) &&
	    (vmx->idt_vectoring_info & VECTORING_INFO_TYPE_MASK)
	    == INTR_TYPE_NMI_INTR) {
		vcpu->arch.nmi_injected = false;
		if (cpu_has_virtual_nmis())
			vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO,
				      GUEST_INTR_STATE_NMI);
	}
3025 3026
	tss_selector = exit_qualification;

3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038
	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;
3039 3040
}

3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098
static int handle_ept_violation(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	u64 exit_qualification;
	enum emulation_result er;
	gpa_t gpa;
	unsigned long hva;
	int gla_validity;
	int r;

	exit_qualification = vmcs_read64(EXIT_QUALIFICATION);

	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),
			(long unsigned int)vmcs_read64(GUEST_LINEAR_ADDRESS));
		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);
	hva = gfn_to_hva(vcpu->kvm, gpa >> PAGE_SHIFT);
	if (!kvm_is_error_hva(hva)) {
		r = kvm_mmu_page_fault(vcpu, gpa & PAGE_MASK, 0);
		if (r < 0) {
			printk(KERN_ERR "EPT: Not enough memory!\n");
			return -ENOMEM;
		}
		return 1;
	} else {
		/* must be MMIO */
		er = emulate_instruction(vcpu, kvm_run, 0, 0, 0);

		if (er == EMULATE_FAIL) {
			printk(KERN_ERR
			 "EPT: Fail to handle EPT violation vmexit!er is %d\n",
			 er);
			printk(KERN_ERR "EPT: GPA: 0x%lx, GVA: 0x%lx\n",
			 (long unsigned int)vmcs_read64(GUEST_PHYSICAL_ADDRESS),
			 (long unsigned int)vmcs_read64(GUEST_LINEAR_ADDRESS));
			printk(KERN_ERR "EPT: Exit qualification is 0x%lx\n",
				(long unsigned int)exit_qualification);
			return -ENOTSUPP;
		} else if (er == EMULATE_DO_MMIO)
			return 0;
	}
	return 1;
}

3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
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;
}

3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123
static void handle_invalid_guest_state(struct kvm_vcpu *vcpu,
				struct kvm_run *kvm_run)
{
	struct vcpu_vmx *vmx = to_vmx(vcpu);
	int err;

	preempt_enable();
	local_irq_enable();

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

3124 3125 3126 3127 3128 3129
		if (err == EMULATE_DO_MMIO)
			break;

		if (err != EMULATE_DONE) {
			kvm_report_emulation_failure(vcpu, "emulation failure");
			return;
3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
		}

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

	local_irq_disable();
	preempt_disable();

3141 3142 3143 3144
	/* Guest state should be valid now except if we need to
	 * emulate an MMIO */
	if (guest_state_valid(vcpu))
		vmx->emulation_required = 0;
3145 3146
}

A
Avi Kivity 已提交
3147 3148 3149 3150 3151 3152 3153 3154 3155
/*
 * 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,
3156
	[EXIT_REASON_TRIPLE_FAULT]            = handle_triple_fault,
3157
	[EXIT_REASON_NMI_WINDOW]	      = handle_nmi_window,
A
Avi Kivity 已提交
3158 3159 3160 3161 3162 3163 3164 3165
	[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
3166
	[EXIT_REASON_INVLPG]		      = handle_invlpg,
3167
	[EXIT_REASON_VMCALL]                  = handle_vmcall,
3168 3169
	[EXIT_REASON_TPR_BELOW_THRESHOLD]     = handle_tpr_below_threshold,
	[EXIT_REASON_APIC_ACCESS]             = handle_apic_access,
E
Eddie Dong 已提交
3170
	[EXIT_REASON_WBINVD]                  = handle_wbinvd,
3171
	[EXIT_REASON_TASK_SWITCH]             = handle_task_switch,
3172
	[EXIT_REASON_EPT_VIOLATION]	      = handle_ept_violation,
A
Avi Kivity 已提交
3173 3174 3175
};

static const int kvm_vmx_max_exit_handlers =
3176
	ARRAY_SIZE(kvm_vmx_exit_handlers);
A
Avi Kivity 已提交
3177 3178 3179 3180 3181 3182 3183 3184

/*
 * The guest has exited.  See if we can fix it or if we need userspace
 * assistance.
 */
static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
{
	u32 exit_reason = vmcs_read32(VM_EXIT_REASON);
3185
	struct vcpu_vmx *vmx = to_vmx(vcpu);
3186
	u32 vectoring_info = vmx->idt_vectoring_info;
3187

3188 3189
	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
3190

3191 3192 3193 3194 3195
	/* If we need to emulate an MMIO from handle_invalid_guest_state
	 * we just return 0 */
	if (vmx->emulation_required && emulate_invalid_guest_state)
		return 0;

3196 3197 3198 3199 3200 3201 3202
	/* Access CR3 don't cause VMExit in paging mode, so we need
	 * to sync with guest real CR3. */
	if (vm_need_ept() && is_paging(vcpu)) {
		vcpu->arch.cr3 = vmcs_readl(GUEST_CR3);
		ept_load_pdptrs(vcpu);
	}

3203 3204 3205 3206 3207 3208
	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
3209

M
Mike Day 已提交
3210
	if ((vectoring_info & VECTORING_INFO_VALID_MASK) &&
3211
			(exit_reason != EXIT_REASON_EXCEPTION_NMI &&
J
Jan Kiszka 已提交
3212 3213 3214 3215 3216
			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);
3217 3218 3219 3220 3221 3222

	if (unlikely(!cpu_has_virtual_nmis() && vmx->soft_vnmi_blocked)) {
		if (vcpu->arch.interrupt_window_open) {
			vmx->soft_vnmi_blocked = 0;
			vcpu->arch.nmi_window_open = 1;
		} else if (vmx->vnmi_blocked_time > 1000000000LL &&
3223
			   vcpu->arch.nmi_pending) {
3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237
			/*
			 * 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;
			vmx->vcpu.arch.nmi_window_open = 1;
		}
	}

A
Avi Kivity 已提交
3238 3239 3240 3241 3242 3243 3244 3245 3246 3247
	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;
}

3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264
static void update_tpr_threshold(struct kvm_vcpu *vcpu)
{
	int max_irr, tpr;

	if (!vm_need_tpr_shadow(vcpu->kvm))
		return;

	if (!kvm_lapic_enabled(vcpu) ||
	    ((max_irr = kvm_lapic_find_highest_irr(vcpu)) == -1)) {
		vmcs_write32(TPR_THRESHOLD, 0);
		return;
	}

	tpr = (kvm_lapic_get_cr8(vcpu) & 0x0f) << 4;
	vmcs_write32(TPR_THRESHOLD, (max_irr > tpr) ? tpr >> 4 : max_irr >> 4);
}

3265 3266 3267
static void vmx_complete_interrupts(struct vcpu_vmx *vmx)
{
	u32 exit_intr_info;
3268
	u32 idt_vectoring_info;
3269 3270
	bool unblock_nmi;
	u8 vector;
3271 3272
	int type;
	bool idtv_info_valid;
3273
	u32 error;
3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286

	exit_intr_info = vmcs_read32(VM_EXIT_INTR_INFO);
	if (cpu_has_virtual_nmis()) {
		unblock_nmi = (exit_intr_info & INTR_INFO_UNBLOCK_NMI) != 0;
		vector = exit_intr_info & INTR_INFO_VECTOR_MASK;
		/*
		 * SDM 3: 25.7.1.2
		 * Re-set bit "block by NMI" before VM entry if vmexit caused by
		 * a guest IRET fault.
		 */
		if (unblock_nmi && vector != DF_VECTOR)
			vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO,
				      GUEST_INTR_STATE_NMI);
3287 3288 3289
	} else if (unlikely(vmx->soft_vnmi_blocked))
		vmx->vnmi_blocked_time +=
			ktime_to_ns(ktime_sub(ktime_get(), vmx->entry_time));
3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306

	idt_vectoring_info = vmx->idt_vectoring_info;
	idtv_info_valid = idt_vectoring_info & VECTORING_INFO_VALID_MASK;
	vector = idt_vectoring_info & VECTORING_INFO_VECTOR_MASK;
	type = idt_vectoring_info & VECTORING_INFO_TYPE_MASK;
	if (vmx->vcpu.arch.nmi_injected) {
		/*
		 * SDM 3: 25.7.1.2
		 * Clear bit "block by NMI" before VM entry if a NMI delivery
		 * faulted.
		 */
		if (idtv_info_valid && type == INTR_TYPE_NMI_INTR)
			vmcs_clear_bits(GUEST_INTERRUPTIBILITY_INFO,
					GUEST_INTR_STATE_NMI);
		else
			vmx->vcpu.arch.nmi_injected = false;
	}
3307
	kvm_clear_exception_queue(&vmx->vcpu);
3308 3309
	if (idtv_info_valid && (type == INTR_TYPE_HARD_EXCEPTION ||
				type == INTR_TYPE_SOFT_EXCEPTION)) {
3310 3311 3312 3313 3314 3315 3316
		if (idt_vectoring_info & VECTORING_INFO_DELIVER_CODE_MASK) {
			error = vmcs_read32(IDT_VECTORING_ERROR_CODE);
			kvm_queue_exception_e(&vmx->vcpu, vector, error);
		} else
			kvm_queue_exception(&vmx->vcpu, vector);
		vmx->idt_vectoring_info = 0;
	}
3317 3318 3319 3320 3321
	kvm_clear_interrupt_queue(&vmx->vcpu);
	if (idtv_info_valid && type == INTR_TYPE_EXT_INTR) {
		kvm_queue_interrupt(&vmx->vcpu, vector);
		vmx->idt_vectoring_info = 0;
	}
3322 3323
}

3324 3325
static void vmx_intr_assist(struct kvm_vcpu *vcpu)
{
3326 3327
	update_tpr_threshold(vcpu);

3328 3329
	vmx_update_window_states(vcpu);

3330 3331 3332 3333 3334
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
		vmcs_clear_bits(GUEST_INTERRUPTIBILITY_INFO,
				GUEST_INTR_STATE_STI |
				GUEST_INTR_STATE_MOV_SS);

3335 3336 3337 3338 3339 3340 3341 3342
	if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) {
		if (vcpu->arch.interrupt.pending) {
			enable_nmi_window(vcpu);
		} else if (vcpu->arch.nmi_window_open) {
			vcpu->arch.nmi_pending = false;
			vcpu->arch.nmi_injected = true;
		} else {
			enable_nmi_window(vcpu);
3343 3344 3345
			return;
		}
	}
3346 3347 3348 3349 3350 3351 3352 3353
	if (vcpu->arch.nmi_injected) {
		vmx_inject_nmi(vcpu);
		if (vcpu->arch.nmi_pending)
			enable_nmi_window(vcpu);
		else if (kvm_cpu_has_interrupt(vcpu))
			enable_irq_window(vcpu);
		return;
	}
3354
	if (!vcpu->arch.interrupt.pending && kvm_cpu_has_interrupt(vcpu)) {
3355
		if (vcpu->arch.interrupt_window_open)
3356 3357 3358 3359 3360 3361
			kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu));
		else
			enable_irq_window(vcpu);
	}
	if (vcpu->arch.interrupt.pending) {
		vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr);
3362 3363
		if (kvm_cpu_has_interrupt(vcpu))
			enable_irq_window(vcpu);
3364
	}
3365 3366
}

3367 3368 3369 3370 3371 3372 3373 3374 3375
/*
 * 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;
3376
	if (kvm_rip_read(&vmx->vcpu) + 1 != vmx->rmode.irq.rip)
3377
		return;
3378
	kvm_rip_write(&vmx->vcpu, vmx->rmode.irq.rip);
3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389
	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;
}

3390 3391 3392 3393 3394 3395 3396 3397
#ifdef CONFIG_X86_64
#define R "r"
#define Q "q"
#else
#define R "e"
#define Q "l"
#endif

3398
static void vmx_vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
A
Avi Kivity 已提交
3399
{
3400
	struct vcpu_vmx *vmx = to_vmx(vcpu);
3401
	u32 intr_info;
3402

3403 3404 3405 3406
	/* 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();

3407 3408 3409 3410 3411 3412
	/* Handle invalid guest state instead of entering VMX */
	if (vmx->emulation_required && emulate_invalid_guest_state) {
		handle_invalid_guest_state(vcpu, kvm_run);
		return;
	}

3413 3414 3415 3416 3417
	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]);

3418 3419 3420 3421 3422
	/*
	 * Loading guest fpu may have cleared host cr0.ts
	 */
	vmcs_writel(HOST_CR0, read_cr0());

3423 3424
	set_debugreg(vcpu->arch.dr6, 6);

M
Mike Day 已提交
3425
	asm(
A
Avi Kivity 已提交
3426
		/* Store host registers */
3427 3428
		"push %%"R"dx; push %%"R"bp;"
		"push %%"R"cx \n\t"
3429 3430 3431
		"cmp %%"R"sp, %c[host_rsp](%0) \n\t"
		"je 1f \n\t"
		"mov %%"R"sp, %c[host_rsp](%0) \n\t"
3432
		__ex(ASM_VMX_VMWRITE_RSP_RDX) "\n\t"
3433
		"1: \n\t"
A
Avi Kivity 已提交
3434
		/* Check if vmlaunch of vmresume is needed */
3435
		"cmpl $0, %c[launched](%0) \n\t"
A
Avi Kivity 已提交
3436
		/* Load guest registers.  Don't clobber flags. */
3437 3438 3439 3440 3441 3442 3443 3444
		"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"
3445
#ifdef CONFIG_X86_64
3446 3447 3448 3449 3450 3451 3452 3453
		"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
3454
#endif
3455 3456
		"mov %c[rcx](%0), %%"R"cx \n\t" /* kills %0 (ecx) */

A
Avi Kivity 已提交
3457
		/* Enter guest mode */
3458
		"jne .Llaunched \n\t"
3459
		__ex(ASM_VMX_VMLAUNCH) "\n\t"
3460
		"jmp .Lkvm_vmx_return \n\t"
3461
		".Llaunched: " __ex(ASM_VMX_VMRESUME) "\n\t"
3462
		".Lkvm_vmx_return: "
A
Avi Kivity 已提交
3463
		/* Save guest registers, load host registers, keep flags */
3464 3465 3466 3467 3468 3469 3470 3471
		"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"
3472
#ifdef CONFIG_X86_64
3473 3474 3475 3476 3477 3478 3479 3480
		"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
3481
#endif
3482 3483 3484 3485
		"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"
3486 3487 3488 3489
		"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)),
3490
		[host_rsp]"i"(offsetof(struct vcpu_vmx, host_rsp)),
3491 3492 3493 3494 3495 3496 3497
		[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])),
3498
#ifdef CONFIG_X86_64
3499 3500 3501 3502 3503 3504 3505 3506
		[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
3507
#endif
3508
		[cr2]"i"(offsetof(struct vcpu_vmx, vcpu.arch.cr2))
3509
	      : "cc", "memory"
3510
		, R"bx", R"di", R"si"
3511 3512 3513 3514
#ifdef CONFIG_X86_64
		, "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15"
#endif
	      );
A
Avi Kivity 已提交
3515

3516 3517 3518
	vcpu->arch.regs_avail = ~((1 << VCPU_REGS_RIP) | (1 << VCPU_REGS_RSP));
	vcpu->arch.regs_dirty = 0;

3519 3520
	get_debugreg(vcpu->arch.dr6, 6);

3521
	vmx->idt_vectoring_info = vmcs_read32(IDT_VECTORING_INFO_FIELD);
3522 3523
	if (vmx->rmode.irq.pending)
		fixup_rmode_irq(vmx);
3524

3525
	vmx_update_window_states(vcpu);
A
Avi Kivity 已提交
3526

M
Mike Day 已提交
3527
	asm("mov %0, %%ds; mov %0, %%es" : : "r"(__USER_DS));
3528
	vmx->launched = 1;
3529 3530 3531 3532

	intr_info = vmcs_read32(VM_EXIT_INTR_INFO);

	/* We need to handle NMIs before interrupts are enabled */
3533
	if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR &&
3534
	    (intr_info & INTR_INFO_VALID_MASK)) {
F
Feng (Eric) Liu 已提交
3535
		KVMTRACE_0D(NMI, vcpu, handler);
3536
		asm("int $2");
F
Feng (Eric) Liu 已提交
3537
	}
3538 3539

	vmx_complete_interrupts(vmx);
A
Avi Kivity 已提交
3540 3541
}

3542 3543 3544
#undef R
#undef Q

A
Avi Kivity 已提交
3545 3546
static void vmx_free_vmcs(struct kvm_vcpu *vcpu)
{
3547 3548 3549
	struct vcpu_vmx *vmx = to_vmx(vcpu);

	if (vmx->vmcs) {
3550
		vcpu_clear(vmx);
3551 3552
		free_vmcs(vmx->vmcs);
		vmx->vmcs = NULL;
A
Avi Kivity 已提交
3553 3554 3555 3556 3557
	}
}

static void vmx_free_vcpu(struct kvm_vcpu *vcpu)
{
R
Rusty Russell 已提交
3558 3559
	struct vcpu_vmx *vmx = to_vmx(vcpu);

3560 3561 3562 3563
	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
3564
	vmx_free_vmcs(vcpu);
R
Rusty Russell 已提交
3565 3566 3567
	kfree(vmx->host_msrs);
	kfree(vmx->guest_msrs);
	kvm_vcpu_uninit(vcpu);
3568
	kmem_cache_free(kvm_vcpu_cache, vmx);
A
Avi Kivity 已提交
3569 3570
}

R
Rusty Russell 已提交
3571
static struct kvm_vcpu *vmx_create_vcpu(struct kvm *kvm, unsigned int id)
A
Avi Kivity 已提交
3572
{
R
Rusty Russell 已提交
3573
	int err;
3574
	struct vcpu_vmx *vmx = kmem_cache_zalloc(kvm_vcpu_cache, GFP_KERNEL);
3575
	int cpu;
A
Avi Kivity 已提交
3576

3577
	if (!vmx)
R
Rusty Russell 已提交
3578 3579
		return ERR_PTR(-ENOMEM);

3580 3581
	allocate_vpid(vmx);

R
Rusty Russell 已提交
3582 3583 3584
	err = kvm_vcpu_init(&vmx->vcpu, kvm, id);
	if (err)
		goto free_vcpu;
3585

3586
	vmx->guest_msrs = kmalloc(PAGE_SIZE, GFP_KERNEL);
R
Rusty Russell 已提交
3587 3588 3589 3590
	if (!vmx->guest_msrs) {
		err = -ENOMEM;
		goto uninit_vcpu;
	}
3591

3592 3593
	vmx->host_msrs = kmalloc(PAGE_SIZE, GFP_KERNEL);
	if (!vmx->host_msrs)
R
Rusty Russell 已提交
3594
		goto free_guest_msrs;
3595

3596 3597
	vmx->vmcs = alloc_vmcs();
	if (!vmx->vmcs)
R
Rusty Russell 已提交
3598
		goto free_msrs;
3599 3600 3601

	vmcs_clear(vmx->vmcs);

3602 3603
	cpu = get_cpu();
	vmx_vcpu_load(&vmx->vcpu, cpu);
R
Rusty Russell 已提交
3604
	err = vmx_vcpu_setup(vmx);
R
Rusty Russell 已提交
3605
	vmx_vcpu_put(&vmx->vcpu);
3606
	put_cpu();
R
Rusty Russell 已提交
3607 3608
	if (err)
		goto free_vmcs;
3609 3610 3611
	if (vm_need_virtualize_apic_accesses(kvm))
		if (alloc_apic_access_page(kvm) != 0)
			goto free_vmcs;
R
Rusty Russell 已提交
3612

3613 3614 3615 3616
	if (vm_need_ept())
		if (alloc_identity_pagetable(kvm) != 0)
			goto free_vmcs;

R
Rusty Russell 已提交
3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627
	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:
3628
	kmem_cache_free(kvm_vcpu_cache, vmx);
R
Rusty Russell 已提交
3629
	return ERR_PTR(err);
A
Avi Kivity 已提交
3630 3631
}

Y
Yang, Sheng 已提交
3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645
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;
	}
}

3646 3647 3648 3649 3650
static int get_ept_level(void)
{
	return VMX_EPT_DEFAULT_GAW + 1;
}

S
Sheng Yang 已提交
3651 3652 3653 3654 3655
static int vmx_get_mt_mask_shift(void)
{
	return VMX_EPT_MT_EPTE_SHIFT;
}

3656
static struct kvm_x86_ops vmx_x86_ops = {
A
Avi Kivity 已提交
3657 3658 3659 3660
	.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
3661
	.check_processor_compatibility = vmx_check_processor_compat,
A
Avi Kivity 已提交
3662 3663
	.hardware_enable = hardware_enable,
	.hardware_disable = hardware_disable,
3664
	.cpu_has_accelerated_tpr = cpu_has_vmx_virtualize_apic_accesses,
A
Avi Kivity 已提交
3665 3666 3667

	.vcpu_create = vmx_create_vcpu,
	.vcpu_free = vmx_free_vcpu,
3668
	.vcpu_reset = vmx_vcpu_reset,
A
Avi Kivity 已提交
3669

3670
	.prepare_guest_switch = vmx_save_host_state,
A
Avi Kivity 已提交
3671 3672 3673 3674 3675 3676 3677 3678 3679
	.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,
3680
	.get_cpl = vmx_get_cpl,
A
Avi Kivity 已提交
3681
	.get_cs_db_l_bits = vmx_get_cs_db_l_bits,
3682
	.decache_cr4_guest_bits = vmx_decache_cr4_guest_bits,
A
Avi Kivity 已提交
3683 3684 3685 3686 3687 3688 3689 3690
	.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,
3691
	.cache_reg = vmx_cache_reg,
A
Avi Kivity 已提交
3692 3693 3694 3695 3696 3697
	.get_rflags = vmx_get_rflags,
	.set_rflags = vmx_set_rflags,

	.tlb_flush = vmx_flush_tlb,

	.run = vmx_vcpu_run,
3698
	.handle_exit = kvm_handle_exit,
A
Avi Kivity 已提交
3699
	.skip_emulated_instruction = skip_emulated_instruction,
I
Ingo Molnar 已提交
3700
	.patch_hypercall = vmx_patch_hypercall,
E
Eddie Dong 已提交
3701 3702
	.get_irq = vmx_get_irq,
	.set_irq = vmx_inject_irq,
3703 3704
	.queue_exception = vmx_queue_exception,
	.exception_injected = vmx_exception_injected,
3705 3706
	.inject_pending_irq = vmx_intr_assist,
	.inject_pending_vectors = do_interrupt_requests,
3707 3708

	.set_tss_addr = vmx_set_tss_addr,
3709
	.get_tdp_level = get_ept_level,
S
Sheng Yang 已提交
3710
	.get_mt_mask_shift = vmx_get_mt_mask_shift,
A
Avi Kivity 已提交
3711 3712 3713 3714
};

static int __init vmx_init(void)
{
S
Sheng Yang 已提交
3715
	void *va;
3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727
	int r;

	vmx_io_bitmap_a = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
	if (!vmx_io_bitmap_a)
		return -ENOMEM;

	vmx_io_bitmap_b = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
	if (!vmx_io_bitmap_b) {
		r = -ENOMEM;
		goto out;
	}

S
Sheng Yang 已提交
3728 3729 3730 3731 3732 3733
	vmx_msr_bitmap = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
	if (!vmx_msr_bitmap) {
		r = -ENOMEM;
		goto out1;
	}

3734 3735 3736 3737
	/*
	 * Allow direct access to the PC debug port (it is often used for I/O
	 * delays, but the vmexits simply slow things down).
	 */
S
Sheng Yang 已提交
3738 3739 3740
	va = kmap(vmx_io_bitmap_a);
	memset(va, 0xff, PAGE_SIZE);
	clear_bit(0x80, va);
3741
	kunmap(vmx_io_bitmap_a);
3742

S
Sheng Yang 已提交
3743 3744
	va = kmap(vmx_io_bitmap_b);
	memset(va, 0xff, PAGE_SIZE);
3745
	kunmap(vmx_io_bitmap_b);
3746

S
Sheng Yang 已提交
3747 3748 3749 3750
	va = kmap(vmx_msr_bitmap);
	memset(va, 0xff, PAGE_SIZE);
	kunmap(vmx_msr_bitmap);

3751 3752
	set_bit(0, vmx_vpid_bitmap); /* 0 is reserved for host */

3753
	r = kvm_init(&vmx_x86_ops, sizeof(struct vcpu_vmx), THIS_MODULE);
3754
	if (r)
S
Sheng Yang 已提交
3755 3756 3757 3758 3759 3760 3761
		goto out2;

	vmx_disable_intercept_for_msr(vmx_msr_bitmap, MSR_FS_BASE);
	vmx_disable_intercept_for_msr(vmx_msr_bitmap, MSR_GS_BASE);
	vmx_disable_intercept_for_msr(vmx_msr_bitmap, MSR_IA32_SYSENTER_CS);
	vmx_disable_intercept_for_msr(vmx_msr_bitmap, MSR_IA32_SYSENTER_ESP);
	vmx_disable_intercept_for_msr(vmx_msr_bitmap, MSR_IA32_SYSENTER_EIP);
3762

3763
	if (vm_need_ept()) {
3764
		bypass_guest_pf = 0;
3765
		kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK |
3766
			VMX_EPT_WRITABLE_MASK);
3767
		kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull,
S
Sheng Yang 已提交
3768 3769
				VMX_EPT_EXECUTABLE_MASK,
				VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT);
3770 3771 3772
		kvm_enable_tdp();
	} else
		kvm_disable_tdp();
3773

3774 3775 3776
	if (bypass_guest_pf)
		kvm_mmu_set_nonpresent_ptes(~0xffeull, 0ull);

3777 3778
	ept_sync_global();

3779 3780
	return 0;

S
Sheng Yang 已提交
3781 3782
out2:
	__free_page(vmx_msr_bitmap);
3783 3784 3785 3786 3787
out1:
	__free_page(vmx_io_bitmap_b);
out:
	__free_page(vmx_io_bitmap_a);
	return r;
A
Avi Kivity 已提交
3788 3789 3790 3791
}

static void __exit vmx_exit(void)
{
S
Sheng Yang 已提交
3792
	__free_page(vmx_msr_bitmap);
3793 3794 3795
	__free_page(vmx_io_bitmap_b);
	__free_page(vmx_io_bitmap_a);

3796
	kvm_exit();
A
Avi Kivity 已提交
3797 3798 3799 3800
}

module_init(vmx_init)
module_exit(vmx_exit)