svm.c 110.0 KB
Newer Older
A
Avi Kivity 已提交
1 2 3 4 5 6
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * AMD SVM support
 *
 * Copyright (C) 2006 Qumranet, Inc.
N
Nicolas Kaiser 已提交
7
 * Copyright 2010 Red Hat, Inc. and/or its affiliates.
A
Avi Kivity 已提交
8 9 10 11 12 13 14 15 16
 *
 * Authors:
 *   Yaniv Kamay  <yaniv@qumranet.com>
 *   Avi Kivity   <avi@qumranet.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */
17 18
#include <linux/kvm_host.h>

19
#include "irq.h"
20
#include "mmu.h"
21
#include "kvm_cache_regs.h"
22
#include "x86.h"
A
Avi Kivity 已提交
23

A
Avi Kivity 已提交
24
#include <linux/module.h>
25
#include <linux/mod_devicetable.h>
26
#include <linux/kernel.h>
A
Avi Kivity 已提交
27 28
#include <linux/vmalloc.h>
#include <linux/highmem.h>
A
Alexey Dobriyan 已提交
29
#include <linux/sched.h>
30
#include <linux/ftrace_event.h>
31
#include <linux/slab.h>
A
Avi Kivity 已提交
32

33
#include <asm/perf_event.h>
34
#include <asm/tlbflush.h>
A
Avi Kivity 已提交
35
#include <asm/desc.h>
G
Gleb Natapov 已提交
36
#include <asm/kvm_para.h>
A
Avi Kivity 已提交
37

38
#include <asm/virtext.h>
39
#include "trace.h"
40

41 42
#define __ex(x) __kvm_handle_fault_on_reboot(x)

A
Avi Kivity 已提交
43 44 45
MODULE_AUTHOR("Qumranet");
MODULE_LICENSE("GPL");

46 47 48 49 50 51
static const struct x86_cpu_id svm_cpu_id[] = {
	X86_FEATURE_MATCH(X86_FEATURE_SVM),
	{}
};
MODULE_DEVICE_TABLE(x86cpu, svm_cpu_id);

A
Avi Kivity 已提交
52 53 54 55 56 57
#define IOPM_ALLOC_ORDER 2
#define MSRPM_ALLOC_ORDER 1

#define SEG_TYPE_LDT 2
#define SEG_TYPE_BUSY_TSS16 3

58 59 60 61
#define SVM_FEATURE_NPT            (1 <<  0)
#define SVM_FEATURE_LBRV           (1 <<  1)
#define SVM_FEATURE_SVML           (1 <<  2)
#define SVM_FEATURE_NRIP           (1 <<  3)
62 63 64 65
#define SVM_FEATURE_TSC_RATE       (1 <<  4)
#define SVM_FEATURE_VMCB_CLEAN     (1 <<  5)
#define SVM_FEATURE_FLUSH_ASID     (1 <<  6)
#define SVM_FEATURE_DECODE_ASSIST  (1 <<  7)
66
#define SVM_FEATURE_PAUSE_FILTER   (1 << 10)
67

68 69 70 71
#define NESTED_EXIT_HOST	0	/* Exit handled on host level */
#define NESTED_EXIT_DONE	1	/* Exit caused nested vmexit  */
#define NESTED_EXIT_CONTINUE	2	/* Further checks needed      */

72 73
#define DEBUGCTL_RESERVED_BITS (~(0x3fULL))

74
#define TSC_RATIO_RSVD          0xffffff0000000000ULL
75 76
#define TSC_RATIO_MIN		0x0000000000000001ULL
#define TSC_RATIO_MAX		0x000000ffffffffffULL
77

78 79
static bool erratum_383_found __read_mostly;

A
Avi Kivity 已提交
80 81 82 83 84 85 86 87 88 89 90 91
static const u32 host_save_user_msrs[] = {
#ifdef CONFIG_X86_64
	MSR_STAR, MSR_LSTAR, MSR_CSTAR, MSR_SYSCALL_MASK, MSR_KERNEL_GS_BASE,
	MSR_FS_BASE,
#endif
	MSR_IA32_SYSENTER_CS, MSR_IA32_SYSENTER_ESP, MSR_IA32_SYSENTER_EIP,
};

#define NR_HOST_SAVE_USER_MSRS ARRAY_SIZE(host_save_user_msrs)

struct kvm_vcpu;

92 93 94
struct nested_state {
	struct vmcb *hsave;
	u64 hsave_msr;
95
	u64 vm_cr_msr;
96 97 98 99 100 101 102
	u64 vmcb;

	/* These are the merged vectors */
	u32 *msrpm;

	/* gpa pointers to the real vectors */
	u64 vmcb_msrpm;
103
	u64 vmcb_iopm;
J
Joerg Roedel 已提交
104

105 106 107
	/* A VMEXIT is required but not yet emulated */
	bool exit_required;

J
Joerg Roedel 已提交
108
	/* cache for intercepts of the guest */
109
	u32 intercept_cr;
110
	u32 intercept_dr;
J
Joerg Roedel 已提交
111 112 113
	u32 intercept_exceptions;
	u64 intercept;

114 115
	/* Nested Paging related state */
	u64 nested_cr3;
116 117
};

118 119 120
#define MSRPM_OFFSETS	16
static u32 msrpm_offsets[MSRPM_OFFSETS] __read_mostly;

121 122 123 124 125 126
/*
 * Set osvw_len to higher value when updated Revision Guides
 * are published and we know what the new status bits are
 */
static uint64_t osvw_len = 4, osvw_status;

A
Avi Kivity 已提交
127 128 129 130 131 132 133 134 135 136 137 138
struct vcpu_svm {
	struct kvm_vcpu vcpu;
	struct vmcb *vmcb;
	unsigned long vmcb_pa;
	struct svm_cpu_data *svm_data;
	uint64_t asid_generation;
	uint64_t sysenter_esp;
	uint64_t sysenter_eip;

	u64 next_rip;

	u64 host_user_msrs[NR_HOST_SAVE_USER_MSRS];
139
	struct {
140 141 142
		u16 fs;
		u16 gs;
		u16 ldt;
143 144
		u64 gs_base;
	} host;
A
Avi Kivity 已提交
145 146 147

	u32 *msrpm;

148 149
	ulong nmi_iret_rip;

150
	struct nested_state nested;
J
Jan Kiszka 已提交
151 152

	bool nmi_singlestep;
153 154 155

	unsigned int3_injected;
	unsigned long int3_rip;
G
Gleb Natapov 已提交
156
	u32 apf_reason;
157 158

	u64  tsc_ratio;
A
Avi Kivity 已提交
159 160
};

161 162 163
static DEFINE_PER_CPU(u64, current_tsc_ratio);
#define TSC_RATIO_DEFAULT	0x0100000000ULL

164 165
#define MSR_INVALID			0xffffffffU

166
static const struct svm_direct_access_msrs {
167 168 169
	u32 index;   /* Index of the MSR */
	bool always; /* True if intercept is always on */
} direct_access_msrs[] = {
B
Brian Gerst 已提交
170
	{ .index = MSR_STAR,				.always = true  },
171 172 173 174 175 176 177 178 179 180 181 182 183 184
	{ .index = MSR_IA32_SYSENTER_CS,		.always = true  },
#ifdef CONFIG_X86_64
	{ .index = MSR_GS_BASE,				.always = true  },
	{ .index = MSR_FS_BASE,				.always = true  },
	{ .index = MSR_KERNEL_GS_BASE,			.always = true  },
	{ .index = MSR_LSTAR,				.always = true  },
	{ .index = MSR_CSTAR,				.always = true  },
	{ .index = MSR_SYSCALL_MASK,			.always = true  },
#endif
	{ .index = MSR_IA32_LASTBRANCHFROMIP,		.always = false },
	{ .index = MSR_IA32_LASTBRANCHTOIP,		.always = false },
	{ .index = MSR_IA32_LASTINTFROMIP,		.always = false },
	{ .index = MSR_IA32_LASTINTTOIP,		.always = false },
	{ .index = MSR_INVALID,				.always = false },
A
Avi Kivity 已提交
185 186
};

187 188 189 190
/* enable NPT for AMD64 and X86 with PAE */
#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
static bool npt_enabled = true;
#else
J
Joerg Roedel 已提交
191
static bool npt_enabled;
192
#endif
193

194 195
/* allow nested paging (virtualized MMU) for all guests */
static int npt = true;
196
module_param(npt, int, S_IRUGO);
197

198 199
/* allow nested virtualization in KVM/SVM */
static int nested = true;
200 201
module_param(nested, int, S_IRUGO);

202
static void svm_flush_tlb(struct kvm_vcpu *vcpu);
203
static void svm_complete_interrupts(struct vcpu_svm *svm);
204

205
static int nested_svm_exit_handled(struct vcpu_svm *svm);
206
static int nested_svm_intercept(struct vcpu_svm *svm);
207 208 209
static int nested_svm_vmexit(struct vcpu_svm *svm);
static int nested_svm_check_exception(struct vcpu_svm *svm, unsigned nr,
				      bool has_error_code, u32 error_code);
210
static u64 __scale_tsc(u64 ratio, u64 tsc);
211

212
enum {
213 214
	VMCB_INTERCEPTS, /* Intercept vectors, TSC offset,
			    pause filter count */
215
	VMCB_PERM_MAP,   /* IOPM Base and MSRPM Base */
216
	VMCB_ASID,	 /* ASID */
217
	VMCB_INTR,	 /* int_ctl, int_vector */
218
	VMCB_NPT,        /* npt_en, nCR3, gPAT */
219
	VMCB_CR,	 /* CR0, CR3, CR4, EFER */
220
	VMCB_DR,         /* DR6, DR7 */
221
	VMCB_DT,         /* GDT, IDT */
222
	VMCB_SEG,        /* CS, DS, SS, ES, CPL */
223
	VMCB_CR2,        /* CR2 only */
224
	VMCB_LBR,        /* DBGCTL, BR_FROM, BR_TO, LAST_EX_FROM, LAST_EX_TO */
225 226 227
	VMCB_DIRTY_MAX,
};

228 229
/* TPR and CR2 are always written before VMRUN */
#define VMCB_ALWAYS_DIRTY_MASK	((1U << VMCB_INTR) | (1U << VMCB_CR2))
230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246

static inline void mark_all_dirty(struct vmcb *vmcb)
{
	vmcb->control.clean = 0;
}

static inline void mark_all_clean(struct vmcb *vmcb)
{
	vmcb->control.clean = ((1 << VMCB_DIRTY_MAX) - 1)
			       & ~VMCB_ALWAYS_DIRTY_MASK;
}

static inline void mark_dirty(struct vmcb *vmcb, int bit)
{
	vmcb->control.clean &= ~(1 << bit);
}

247 248
static inline struct vcpu_svm *to_svm(struct kvm_vcpu *vcpu)
{
R
Rusty Russell 已提交
249
	return container_of(vcpu, struct vcpu_svm, vcpu);
250 251
}

252 253 254 255 256
static void recalc_intercepts(struct vcpu_svm *svm)
{
	struct vmcb_control_area *c, *h;
	struct nested_state *g;

257 258
	mark_dirty(svm->vmcb, VMCB_INTERCEPTS);

259 260 261 262 263 264 265
	if (!is_guest_mode(&svm->vcpu))
		return;

	c = &svm->vmcb->control;
	h = &svm->nested.hsave->control;
	g = &svm->nested;

266
	c->intercept_cr = h->intercept_cr | g->intercept_cr;
267
	c->intercept_dr = h->intercept_dr | g->intercept_dr;
268 269 270 271
	c->intercept_exceptions = h->intercept_exceptions | g->intercept_exceptions;
	c->intercept = h->intercept | g->intercept;
}

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
static inline struct vmcb *get_host_vmcb(struct vcpu_svm *svm)
{
	if (is_guest_mode(&svm->vcpu))
		return svm->nested.hsave;
	else
		return svm->vmcb;
}

static inline void set_cr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_cr |= (1U << bit);

	recalc_intercepts(svm);
}

static inline void clr_cr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_cr &= ~(1U << bit);

	recalc_intercepts(svm);
}

static inline bool is_cr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	return vmcb->control.intercept_cr & (1U << bit);
}

305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
static inline void set_dr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_dr |= (1U << bit);

	recalc_intercepts(svm);
}

static inline void clr_dr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_dr &= ~(1U << bit);

	recalc_intercepts(svm);
}

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
static inline void set_exception_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_exceptions |= (1U << bit);

	recalc_intercepts(svm);
}

static inline void clr_exception_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept_exceptions &= ~(1U << bit);

	recalc_intercepts(svm);
}

341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358
static inline void set_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept |= (1ULL << bit);

	recalc_intercepts(svm);
}

static inline void clr_intercept(struct vcpu_svm *svm, int bit)
{
	struct vmcb *vmcb = get_host_vmcb(svm);

	vmcb->control.intercept &= ~(1ULL << bit);

	recalc_intercepts(svm);
}

359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
static inline void enable_gif(struct vcpu_svm *svm)
{
	svm->vcpu.arch.hflags |= HF_GIF_MASK;
}

static inline void disable_gif(struct vcpu_svm *svm)
{
	svm->vcpu.arch.hflags &= ~HF_GIF_MASK;
}

static inline bool gif_set(struct vcpu_svm *svm)
{
	return !!(svm->vcpu.arch.hflags & HF_GIF_MASK);
}

374
static unsigned long iopm_base;
A
Avi Kivity 已提交
375 376 377 378

struct kvm_ldttss_desc {
	u16 limit0;
	u16 base0;
J
Joerg Roedel 已提交
379 380
	unsigned base1:8, type:5, dpl:2, p:1;
	unsigned limit1:4, zero0:3, g:1, base2:8;
A
Avi Kivity 已提交
381 382 383 384 385 386 387
	u32 base3;
	u32 zero1;
} __attribute__((packed));

struct svm_cpu_data {
	int cpu;

A
Avi Kivity 已提交
388 389 390
	u64 asid_generation;
	u32 max_asid;
	u32 next_asid;
A
Avi Kivity 已提交
391 392 393 394 395 396 397 398 399 400 401 402
	struct kvm_ldttss_desc *tss_desc;

	struct page *save_area;
};

static DEFINE_PER_CPU(struct svm_cpu_data *, svm_data);

struct svm_init_data {
	int cpu;
	int r;
};

403
static const u32 msrpm_ranges[] = {0, 0xc0000000, 0xc0010000};
A
Avi Kivity 已提交
404

405
#define NUM_MSR_MAPS ARRAY_SIZE(msrpm_ranges)
A
Avi Kivity 已提交
406 407 408
#define MSRS_RANGE_SIZE 2048
#define MSRS_IN_RANGE (MSRS_RANGE_SIZE * 8 / 2)

409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
static u32 svm_msrpm_offset(u32 msr)
{
	u32 offset;
	int i;

	for (i = 0; i < NUM_MSR_MAPS; i++) {
		if (msr < msrpm_ranges[i] ||
		    msr >= msrpm_ranges[i] + MSRS_IN_RANGE)
			continue;

		offset  = (msr - msrpm_ranges[i]) / 4; /* 4 msrs per u8 */
		offset += (i * MSRS_RANGE_SIZE);       /* add range offset */

		/* Now we have the u8 offset - but need the u32 offset */
		return offset / 4;
	}

	/* MSR not in any range */
	return MSR_INVALID;
}

A
Avi Kivity 已提交
430 431 432 433
#define MAX_INST_SIZE 15

static inline void clgi(void)
{
434
	asm volatile (__ex(SVM_CLGI));
A
Avi Kivity 已提交
435 436 437 438
}

static inline void stgi(void)
{
439
	asm volatile (__ex(SVM_STGI));
A
Avi Kivity 已提交
440 441 442 443
}

static inline void invlpga(unsigned long addr, u32 asid)
{
J
Joerg Roedel 已提交
444
	asm volatile (__ex(SVM_INVLPGA) : : "a"(addr), "c"(asid));
A
Avi Kivity 已提交
445 446
}

447 448 449 450 451 452 453 454 455
static int get_npt_level(void)
{
#ifdef CONFIG_X86_64
	return PT64_ROOT_LEVEL;
#else
	return PT32E_ROOT_LEVEL;
#endif
}

A
Avi Kivity 已提交
456 457
static void svm_set_efer(struct kvm_vcpu *vcpu, u64 efer)
{
458
	vcpu->arch.efer = efer;
459
	if (!npt_enabled && !(efer & EFER_LMA))
460
		efer &= ~EFER_LME;
A
Avi Kivity 已提交
461

462
	to_svm(vcpu)->vmcb->save.efer = efer | EFER_SVME;
463
	mark_dirty(to_svm(vcpu)->vmcb, VMCB_CR);
A
Avi Kivity 已提交
464 465 466 467 468 469 470 471
}

static int is_external_interrupt(u32 info)
{
	info &= SVM_EVTINJ_TYPE_MASK | SVM_EVTINJ_VALID;
	return info == (SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_INTR);
}

472 473 474 475 476 477
static u32 svm_get_interrupt_shadow(struct kvm_vcpu *vcpu, int mask)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u32 ret = 0;

	if (svm->vmcb->control.int_state & SVM_INTERRUPT_SHADOW_MASK)
478
		ret |= KVM_X86_SHADOW_INT_STI | KVM_X86_SHADOW_INT_MOV_SS;
479 480 481 482 483 484 485 486 487 488 489 490 491 492
	return ret & mask;
}

static void svm_set_interrupt_shadow(struct kvm_vcpu *vcpu, int mask)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	if (mask == 0)
		svm->vmcb->control.int_state &= ~SVM_INTERRUPT_SHADOW_MASK;
	else
		svm->vmcb->control.int_state |= SVM_INTERRUPT_SHADOW_MASK;

}

A
Avi Kivity 已提交
493 494
static void skip_emulated_instruction(struct kvm_vcpu *vcpu)
{
495 496
	struct vcpu_svm *svm = to_svm(vcpu);

497 498 499
	if (svm->vmcb->control.next_rip != 0)
		svm->next_rip = svm->vmcb->control.next_rip;

500
	if (!svm->next_rip) {
501
		if (emulate_instruction(vcpu, EMULTYPE_SKIP) !=
502 503
				EMULATE_DONE)
			printk(KERN_DEBUG "%s: NOP\n", __func__);
A
Avi Kivity 已提交
504 505
		return;
	}
506 507 508
	if (svm->next_rip - kvm_rip_read(vcpu) > MAX_INST_SIZE)
		printk(KERN_ERR "%s: ip 0x%lx next 0x%llx\n",
		       __func__, kvm_rip_read(vcpu), svm->next_rip);
A
Avi Kivity 已提交
509

510
	kvm_rip_write(vcpu, svm->next_rip);
511
	svm_set_interrupt_shadow(vcpu, 0);
A
Avi Kivity 已提交
512 513
}

J
Jan Kiszka 已提交
514
static void svm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr,
515 516
				bool has_error_code, u32 error_code,
				bool reinject)
J
Jan Kiszka 已提交
517 518 519
{
	struct vcpu_svm *svm = to_svm(vcpu);

J
Joerg Roedel 已提交
520 521 522 523
	/*
	 * If we are within a nested VM we'd better #VMEXIT and let the guest
	 * handle the exception
	 */
524 525
	if (!reinject &&
	    nested_svm_check_exception(svm, nr, has_error_code, error_code))
J
Jan Kiszka 已提交
526 527
		return;

528
	if (nr == BP_VECTOR && !static_cpu_has(X86_FEATURE_NRIPS)) {
529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
		unsigned long rip, old_rip = kvm_rip_read(&svm->vcpu);

		/*
		 * For guest debugging where we have to reinject #BP if some
		 * INT3 is guest-owned:
		 * Emulate nRIP by moving RIP forward. Will fail if injection
		 * raises a fault that is not intercepted. Still better than
		 * failing in all cases.
		 */
		skip_emulated_instruction(&svm->vcpu);
		rip = kvm_rip_read(&svm->vcpu);
		svm->int3_rip = rip + svm->vmcb->save.cs.base;
		svm->int3_injected = rip - old_rip;
	}

J
Jan Kiszka 已提交
544 545 546 547 548 549 550
	svm->vmcb->control.event_inj = nr
		| SVM_EVTINJ_VALID
		| (has_error_code ? SVM_EVTINJ_VALID_ERR : 0)
		| SVM_EVTINJ_TYPE_EXEPT;
	svm->vmcb->control.event_inj_err = error_code;
}

551 552 553 554 555 556
static void svm_init_erratum_383(void)
{
	u32 low, high;
	int err;
	u64 val;

557
	if (!cpu_has_amd_erratum(amd_erratum_383))
558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
		return;

	/* Use _safe variants to not break nested virtualization */
	val = native_read_msr_safe(MSR_AMD64_DC_CFG, &err);
	if (err)
		return;

	val |= (1ULL << 47);

	low  = lower_32_bits(val);
	high = upper_32_bits(val);

	native_write_msr_safe(MSR_AMD64_DC_CFG, low, high);

	erratum_383_found = true;
}

575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
static void svm_init_osvw(struct kvm_vcpu *vcpu)
{
	/*
	 * Guests should see errata 400 and 415 as fixed (assuming that
	 * HLT and IO instructions are intercepted).
	 */
	vcpu->arch.osvw.length = (osvw_len >= 3) ? (osvw_len) : 3;
	vcpu->arch.osvw.status = osvw_status & ~(6ULL);

	/*
	 * By increasing VCPU's osvw.length to 3 we are telling the guest that
	 * all osvw.status bits inside that length, including bit 0 (which is
	 * reserved for erratum 298), are valid. However, if host processor's
	 * osvw_len is 0 then osvw_status[0] carries no information. We need to
	 * be conservative here and therefore we tell the guest that erratum 298
	 * is present (because we really don't know).
	 */
	if (osvw_len == 0 && boot_cpu_data.x86 == 0x10)
		vcpu->arch.osvw.status |= 1;
}

A
Avi Kivity 已提交
596 597
static int has_svm(void)
{
598
	const char *msg;
A
Avi Kivity 已提交
599

600
	if (!cpu_has_svm(&msg)) {
J
Joe Perches 已提交
601
		printk(KERN_INFO "has_svm: %s\n", msg);
A
Avi Kivity 已提交
602 603 604 605 606 607 608 609
		return 0;
	}

	return 1;
}

static void svm_hardware_disable(void *garbage)
{
610 611 612 613
	/* Make sure we clean up behind us */
	if (static_cpu_has(X86_FEATURE_TSCRATEMSR))
		wrmsrl(MSR_AMD64_TSC_RATIO, TSC_RATIO_DEFAULT);

614
	cpu_svm_disable();
615 616

	amd_pmu_disable_virt();
A
Avi Kivity 已提交
617 618
}

619
static int svm_hardware_enable(void *garbage)
A
Avi Kivity 已提交
620 621
{

622
	struct svm_cpu_data *sd;
A
Avi Kivity 已提交
623
	uint64_t efer;
624
	struct desc_ptr gdt_descr;
A
Avi Kivity 已提交
625 626 627
	struct desc_struct *gdt;
	int me = raw_smp_processor_id();

628 629 630 631
	rdmsrl(MSR_EFER, efer);
	if (efer & EFER_SVME)
		return -EBUSY;

A
Avi Kivity 已提交
632
	if (!has_svm()) {
633
		pr_err("%s: err EOPNOTSUPP on %d\n", __func__, me);
634
		return -EINVAL;
A
Avi Kivity 已提交
635
	}
636 637
	sd = per_cpu(svm_data, me);
	if (!sd) {
638
		pr_err("%s: svm_data is NULL on %d\n", __func__, me);
639
		return -EINVAL;
A
Avi Kivity 已提交
640 641
	}

642 643 644
	sd->asid_generation = 1;
	sd->max_asid = cpuid_ebx(SVM_CPUID_FUNC) - 1;
	sd->next_asid = sd->max_asid + 1;
A
Avi Kivity 已提交
645

646
	native_store_gdt(&gdt_descr);
647
	gdt = (struct desc_struct *)gdt_descr.address;
648
	sd->tss_desc = (struct kvm_ldttss_desc *)(gdt + GDT_ENTRY_TSS);
A
Avi Kivity 已提交
649

650
	wrmsrl(MSR_EFER, efer | EFER_SVME);
A
Avi Kivity 已提交
651

652
	wrmsrl(MSR_VM_HSAVE_PA, page_to_pfn(sd->save_area) << PAGE_SHIFT);
653

654 655 656 657 658
	if (static_cpu_has(X86_FEATURE_TSCRATEMSR)) {
		wrmsrl(MSR_AMD64_TSC_RATIO, TSC_RATIO_DEFAULT);
		__get_cpu_var(current_tsc_ratio) = TSC_RATIO_DEFAULT;
	}

659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688

	/*
	 * Get OSVW bits.
	 *
	 * Note that it is possible to have a system with mixed processor
	 * revisions and therefore different OSVW bits. If bits are not the same
	 * on different processors then choose the worst case (i.e. if erratum
	 * is present on one processor and not on another then assume that the
	 * erratum is present everywhere).
	 */
	if (cpu_has(&boot_cpu_data, X86_FEATURE_OSVW)) {
		uint64_t len, status = 0;
		int err;

		len = native_read_msr_safe(MSR_AMD64_OSVW_ID_LENGTH, &err);
		if (!err)
			status = native_read_msr_safe(MSR_AMD64_OSVW_STATUS,
						      &err);

		if (err)
			osvw_status = osvw_len = 0;
		else {
			if (len < osvw_len)
				osvw_len = len;
			osvw_status |= status;
			osvw_status &= (1ULL << osvw_len) - 1;
		}
	} else
		osvw_status = osvw_len = 0;

689 690
	svm_init_erratum_383();

691 692
	amd_pmu_enable_virt();

693
	return 0;
A
Avi Kivity 已提交
694 695
}

696 697
static void svm_cpu_uninit(int cpu)
{
698
	struct svm_cpu_data *sd = per_cpu(svm_data, raw_smp_processor_id());
699

700
	if (!sd)
701 702 703
		return;

	per_cpu(svm_data, raw_smp_processor_id()) = NULL;
704 705
	__free_page(sd->save_area);
	kfree(sd);
706 707
}

A
Avi Kivity 已提交
708 709
static int svm_cpu_init(int cpu)
{
710
	struct svm_cpu_data *sd;
A
Avi Kivity 已提交
711 712
	int r;

713 714
	sd = kzalloc(sizeof(struct svm_cpu_data), GFP_KERNEL);
	if (!sd)
A
Avi Kivity 已提交
715
		return -ENOMEM;
716 717
	sd->cpu = cpu;
	sd->save_area = alloc_page(GFP_KERNEL);
A
Avi Kivity 已提交
718
	r = -ENOMEM;
719
	if (!sd->save_area)
A
Avi Kivity 已提交
720 721
		goto err_1;

722
	per_cpu(svm_data, cpu) = sd;
A
Avi Kivity 已提交
723 724 725 726

	return 0;

err_1:
727
	kfree(sd);
A
Avi Kivity 已提交
728 729 730 731
	return r;

}

732 733 734 735 736 737 738 739 740 741 742
static bool valid_msr_intercept(u32 index)
{
	int i;

	for (i = 0; direct_access_msrs[i].index != MSR_INVALID; i++)
		if (direct_access_msrs[i].index == index)
			return true;

	return false;
}

743 744
static void set_msr_interception(u32 *msrpm, unsigned msr,
				 int read, int write)
A
Avi Kivity 已提交
745
{
746 747 748
	u8 bit_read, bit_write;
	unsigned long tmp;
	u32 offset;
A
Avi Kivity 已提交
749

750 751 752 753 754 755
	/*
	 * If this warning triggers extend the direct_access_msrs list at the
	 * beginning of the file
	 */
	WARN_ON(!valid_msr_intercept(msr));

756 757 758 759 760 761 762 763 764 765 766
	offset    = svm_msrpm_offset(msr);
	bit_read  = 2 * (msr & 0x0f);
	bit_write = 2 * (msr & 0x0f) + 1;
	tmp       = msrpm[offset];

	BUG_ON(offset == MSR_INVALID);

	read  ? clear_bit(bit_read,  &tmp) : set_bit(bit_read,  &tmp);
	write ? clear_bit(bit_write, &tmp) : set_bit(bit_write, &tmp);

	msrpm[offset] = tmp;
A
Avi Kivity 已提交
767 768
}

769
static void svm_vcpu_init_msrpm(u32 *msrpm)
A
Avi Kivity 已提交
770 771 772
{
	int i;

773 774
	memset(msrpm, 0xff, PAGE_SIZE * (1 << MSRPM_ALLOC_ORDER));

775 776 777 778 779 780
	for (i = 0; direct_access_msrs[i].index != MSR_INVALID; i++) {
		if (!direct_access_msrs[i].always)
			continue;

		set_msr_interception(msrpm, direct_access_msrs[i].index, 1, 1);
	}
781 782
}

783 784 785 786 787 788 789 790
static void add_msr_offset(u32 offset)
{
	int i;

	for (i = 0; i < MSRPM_OFFSETS; ++i) {

		/* Offset already in list? */
		if (msrpm_offsets[i] == offset)
791
			return;
792 793 794 795 796 797 798 799 800

		/* Slot used by another offset? */
		if (msrpm_offsets[i] != MSR_INVALID)
			continue;

		/* Add offset to list */
		msrpm_offsets[i] = offset;

		return;
A
Avi Kivity 已提交
801
	}
802 803 804 805 806

	/*
	 * If this BUG triggers the msrpm_offsets table has an overflow. Just
	 * increase MSRPM_OFFSETS in this case.
	 */
807
	BUG();
A
Avi Kivity 已提交
808 809
}

810
static void init_msrpm_offsets(void)
811
{
812
	int i;
813

814 815 816 817 818 819 820 821 822 823
	memset(msrpm_offsets, 0xff, sizeof(msrpm_offsets));

	for (i = 0; direct_access_msrs[i].index != MSR_INVALID; i++) {
		u32 offset;

		offset = svm_msrpm_offset(direct_access_msrs[i].index);
		BUG_ON(offset == MSR_INVALID);

		add_msr_offset(offset);
	}
824 825
}

826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
static void svm_enable_lbrv(struct vcpu_svm *svm)
{
	u32 *msrpm = svm->msrpm;

	svm->vmcb->control.lbr_ctl = 1;
	set_msr_interception(msrpm, MSR_IA32_LASTBRANCHFROMIP, 1, 1);
	set_msr_interception(msrpm, MSR_IA32_LASTBRANCHTOIP, 1, 1);
	set_msr_interception(msrpm, MSR_IA32_LASTINTFROMIP, 1, 1);
	set_msr_interception(msrpm, MSR_IA32_LASTINTTOIP, 1, 1);
}

static void svm_disable_lbrv(struct vcpu_svm *svm)
{
	u32 *msrpm = svm->msrpm;

	svm->vmcb->control.lbr_ctl = 0;
	set_msr_interception(msrpm, MSR_IA32_LASTBRANCHFROMIP, 0, 0);
	set_msr_interception(msrpm, MSR_IA32_LASTBRANCHTOIP, 0, 0);
	set_msr_interception(msrpm, MSR_IA32_LASTINTFROMIP, 0, 0);
	set_msr_interception(msrpm, MSR_IA32_LASTINTTOIP, 0, 0);
}

A
Avi Kivity 已提交
848 849 850 851
static __init int svm_hardware_setup(void)
{
	int cpu;
	struct page *iopm_pages;
852
	void *iopm_va;
A
Avi Kivity 已提交
853 854 855 856 857 858
	int r;

	iopm_pages = alloc_pages(GFP_KERNEL, IOPM_ALLOC_ORDER);

	if (!iopm_pages)
		return -ENOMEM;
859 860 861

	iopm_va = page_address(iopm_pages);
	memset(iopm_va, 0xff, PAGE_SIZE * (1 << IOPM_ALLOC_ORDER));
A
Avi Kivity 已提交
862 863
	iopm_base = page_to_pfn(iopm_pages) << PAGE_SHIFT;

864 865
	init_msrpm_offsets();

866 867 868
	if (boot_cpu_has(X86_FEATURE_NX))
		kvm_enable_efer_bits(EFER_NX);

A
Alexander Graf 已提交
869 870 871
	if (boot_cpu_has(X86_FEATURE_FXSR_OPT))
		kvm_enable_efer_bits(EFER_FFXSR);

872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888
	if (boot_cpu_has(X86_FEATURE_TSCRATEMSR)) {
		u64 max;

		kvm_has_tsc_control = true;

		/*
		 * Make sure the user can only configure tsc_khz values that
		 * fit into a signed integer.
		 * A min value is not calculated needed because it will always
		 * be 1 on all machines and a value of 0 is used to disable
		 * tsc-scaling for the vcpu.
		 */
		max = min(0x7fffffffULL, __scale_tsc(tsc_khz, TSC_RATIO_MAX));

		kvm_max_guest_tsc_khz = max;
	}

889 890
	if (nested) {
		printk(KERN_INFO "kvm: Nested Virtualization enabled\n");
891
		kvm_enable_efer_bits(EFER_SVME | EFER_LMSLE);
892 893
	}

Z
Zachary Amsden 已提交
894
	for_each_possible_cpu(cpu) {
A
Avi Kivity 已提交
895 896
		r = svm_cpu_init(cpu);
		if (r)
897
			goto err;
A
Avi Kivity 已提交
898
	}
899

900
	if (!boot_cpu_has(X86_FEATURE_NPT))
901 902
		npt_enabled = false;

903 904 905 906 907
	if (npt_enabled && !npt) {
		printk(KERN_INFO "kvm: Nested Paging disabled\n");
		npt_enabled = false;
	}

908
	if (npt_enabled) {
909
		printk(KERN_INFO "kvm: Nested Paging enabled\n");
910
		kvm_enable_tdp();
911 912
	} else
		kvm_disable_tdp();
913

A
Avi Kivity 已提交
914 915
	return 0;

916
err:
A
Avi Kivity 已提交
917 918 919 920 921 922 923
	__free_pages(iopm_pages, IOPM_ALLOC_ORDER);
	iopm_base = 0;
	return r;
}

static __exit void svm_hardware_unsetup(void)
{
924 925
	int cpu;

Z
Zachary Amsden 已提交
926
	for_each_possible_cpu(cpu)
927 928
		svm_cpu_uninit(cpu);

A
Avi Kivity 已提交
929
	__free_pages(pfn_to_page(iopm_base >> PAGE_SHIFT), IOPM_ALLOC_ORDER);
930
	iopm_base = 0;
A
Avi Kivity 已提交
931 932 933 934 935 936
}

static void init_seg(struct vmcb_seg *seg)
{
	seg->selector = 0;
	seg->attrib = SVM_SELECTOR_P_MASK | SVM_SELECTOR_S_MASK |
J
Joerg Roedel 已提交
937
		      SVM_SELECTOR_WRITE_MASK; /* Read/Write Data Segment */
A
Avi Kivity 已提交
938 939 940 941 942 943 944 945 946 947 948 949
	seg->limit = 0xffff;
	seg->base = 0;
}

static void init_sys_seg(struct vmcb_seg *seg, uint32_t type)
{
	seg->selector = 0;
	seg->attrib = SVM_SELECTOR_P_MASK | type;
	seg->limit = 0xffff;
	seg->base = 0;
}

950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975
static u64 __scale_tsc(u64 ratio, u64 tsc)
{
	u64 mult, frac, _tsc;

	mult  = ratio >> 32;
	frac  = ratio & ((1ULL << 32) - 1);

	_tsc  = tsc;
	_tsc *= mult;
	_tsc += (tsc >> 32) * frac;
	_tsc += ((tsc & ((1ULL << 32) - 1)) * frac) >> 32;

	return _tsc;
}

static u64 svm_scale_tsc(struct kvm_vcpu *vcpu, u64 tsc)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u64 _tsc = tsc;

	if (svm->tsc_ratio != TSC_RATIO_DEFAULT)
		_tsc = __scale_tsc(svm->tsc_ratio, tsc);

	return _tsc;
}

976
static void svm_set_tsc_khz(struct kvm_vcpu *vcpu, u32 user_tsc_khz, bool scale)
977 978 979 980 981
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u64 ratio;
	u64 khz;

982 983 984
	/* Guest TSC same frequency as host TSC? */
	if (!scale) {
		svm->tsc_ratio = TSC_RATIO_DEFAULT;
985
		return;
986
	}
987

988 989 990 991 992 993 994
	/* TSC scaling supported? */
	if (!boot_cpu_has(X86_FEATURE_TSCRATEMSR)) {
		if (user_tsc_khz > tsc_khz) {
			vcpu->arch.tsc_catchup = 1;
			vcpu->arch.tsc_always_catchup = 1;
		} else
			WARN(1, "user requested TSC rate below hardware speed\n");
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
		return;
	}

	khz = user_tsc_khz;

	/* TSC scaling required  - calculate ratio */
	ratio = khz << 32;
	do_div(ratio, tsc_khz);

	if (ratio == 0 || ratio & TSC_RATIO_RSVD) {
		WARN_ONCE(1, "Invalid TSC ratio - virtual-tsc-khz=%u\n",
				user_tsc_khz);
		return;
	}
	svm->tsc_ratio             = ratio;
}

1012 1013 1014 1015 1016
static void svm_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u64 g_tsc_offset = 0;

1017
	if (is_guest_mode(vcpu)) {
1018 1019 1020 1021 1022 1023
		g_tsc_offset = svm->vmcb->control.tsc_offset -
			       svm->nested.hsave->control.tsc_offset;
		svm->nested.hsave->control.tsc_offset = offset;
	}

	svm->vmcb->control.tsc_offset = offset + g_tsc_offset;
1024 1025

	mark_dirty(svm->vmcb, VMCB_INTERCEPTS);
1026 1027
}

1028
static void svm_adjust_tsc_offset(struct kvm_vcpu *vcpu, s64 adjustment, bool host)
Z
Zachary Amsden 已提交
1029 1030 1031
{
	struct vcpu_svm *svm = to_svm(vcpu);

1032 1033 1034 1035
	WARN_ON(adjustment < 0);
	if (host)
		adjustment = svm_scale_tsc(vcpu, adjustment);

Z
Zachary Amsden 已提交
1036
	svm->vmcb->control.tsc_offset += adjustment;
1037
	if (is_guest_mode(vcpu))
Z
Zachary Amsden 已提交
1038
		svm->nested.hsave->control.tsc_offset += adjustment;
1039
	mark_dirty(svm->vmcb, VMCB_INTERCEPTS);
Z
Zachary Amsden 已提交
1040 1041
}

1042 1043 1044 1045 1046 1047 1048 1049 1050
static u64 svm_compute_tsc_offset(struct kvm_vcpu *vcpu, u64 target_tsc)
{
	u64 tsc;

	tsc = svm_scale_tsc(vcpu, native_read_tsc());

	return target_tsc - tsc;
}

1051
static void init_vmcb(struct vcpu_svm *svm)
A
Avi Kivity 已提交
1052
{
1053 1054
	struct vmcb_control_area *control = &svm->vmcb->control;
	struct vmcb_save_area *save = &svm->vmcb->save;
A
Avi Kivity 已提交
1055

1056
	svm->vcpu.fpu_active = 1;
1057
	svm->vcpu.arch.hflags = 0;
1058

1059 1060 1061 1062 1063 1064 1065
	set_cr_intercept(svm, INTERCEPT_CR0_READ);
	set_cr_intercept(svm, INTERCEPT_CR3_READ);
	set_cr_intercept(svm, INTERCEPT_CR4_READ);
	set_cr_intercept(svm, INTERCEPT_CR0_WRITE);
	set_cr_intercept(svm, INTERCEPT_CR3_WRITE);
	set_cr_intercept(svm, INTERCEPT_CR4_WRITE);
	set_cr_intercept(svm, INTERCEPT_CR8_WRITE);
A
Avi Kivity 已提交
1066

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	set_dr_intercept(svm, INTERCEPT_DR0_READ);
	set_dr_intercept(svm, INTERCEPT_DR1_READ);
	set_dr_intercept(svm, INTERCEPT_DR2_READ);
	set_dr_intercept(svm, INTERCEPT_DR3_READ);
	set_dr_intercept(svm, INTERCEPT_DR4_READ);
	set_dr_intercept(svm, INTERCEPT_DR5_READ);
	set_dr_intercept(svm, INTERCEPT_DR6_READ);
	set_dr_intercept(svm, INTERCEPT_DR7_READ);

	set_dr_intercept(svm, INTERCEPT_DR0_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR1_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR2_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR3_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR4_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR5_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR6_WRITE);
	set_dr_intercept(svm, INTERCEPT_DR7_WRITE);
A
Avi Kivity 已提交
1084

1085 1086 1087
	set_exception_intercept(svm, PF_VECTOR);
	set_exception_intercept(svm, UD_VECTOR);
	set_exception_intercept(svm, MC_VECTOR);
A
Avi Kivity 已提交
1088

1089 1090 1091 1092
	set_intercept(svm, INTERCEPT_INTR);
	set_intercept(svm, INTERCEPT_NMI);
	set_intercept(svm, INTERCEPT_SMI);
	set_intercept(svm, INTERCEPT_SELECTIVE_CR0);
A
Avi Kivity 已提交
1093
	set_intercept(svm, INTERCEPT_RDPMC);
1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
	set_intercept(svm, INTERCEPT_CPUID);
	set_intercept(svm, INTERCEPT_INVD);
	set_intercept(svm, INTERCEPT_HLT);
	set_intercept(svm, INTERCEPT_INVLPG);
	set_intercept(svm, INTERCEPT_INVLPGA);
	set_intercept(svm, INTERCEPT_IOIO_PROT);
	set_intercept(svm, INTERCEPT_MSR_PROT);
	set_intercept(svm, INTERCEPT_TASK_SWITCH);
	set_intercept(svm, INTERCEPT_SHUTDOWN);
	set_intercept(svm, INTERCEPT_VMRUN);
	set_intercept(svm, INTERCEPT_VMMCALL);
	set_intercept(svm, INTERCEPT_VMLOAD);
	set_intercept(svm, INTERCEPT_VMSAVE);
	set_intercept(svm, INTERCEPT_STGI);
	set_intercept(svm, INTERCEPT_CLGI);
	set_intercept(svm, INTERCEPT_SKINIT);
	set_intercept(svm, INTERCEPT_WBINVD);
	set_intercept(svm, INTERCEPT_MONITOR);
	set_intercept(svm, INTERCEPT_MWAIT);
J
Joerg Roedel 已提交
1113
	set_intercept(svm, INTERCEPT_XSETBV);
A
Avi Kivity 已提交
1114 1115

	control->iopm_base_pa = iopm_base;
1116
	control->msrpm_base_pa = __pa(svm->msrpm);
A
Avi Kivity 已提交
1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129
	control->int_ctl = V_INTR_MASKING_MASK;

	init_seg(&save->es);
	init_seg(&save->ss);
	init_seg(&save->ds);
	init_seg(&save->fs);
	init_seg(&save->gs);

	save->cs.selector = 0xf000;
	/* Executable/Readable Code Segment */
	save->cs.attrib = SVM_SELECTOR_READ_MASK | SVM_SELECTOR_P_MASK |
		SVM_SELECTOR_S_MASK | SVM_SELECTOR_CODE_MASK;
	save->cs.limit = 0xffff;
1130 1131 1132 1133 1134 1135 1136
	/*
	 * cs.base should really be 0xffff0000, but vmx can't handle that, so
	 * be consistent with it.
	 *
	 * Replace when we have real mode working for vmx.
	 */
	save->cs.base = 0xf0000;
A
Avi Kivity 已提交
1137 1138 1139 1140 1141 1142 1143

	save->gdtr.limit = 0xffff;
	save->idtr.limit = 0xffff;

	init_sys_seg(&save->ldtr, SEG_TYPE_LDT);
	init_sys_seg(&save->tr, SEG_TYPE_BUSY_TSS16);

1144
	svm_set_efer(&svm->vcpu, 0);
M
Mike Day 已提交
1145
	save->dr6 = 0xffff0ff0;
1146
	kvm_set_rflags(&svm->vcpu, 2);
A
Avi Kivity 已提交
1147
	save->rip = 0x0000fff0;
1148
	svm->vcpu.arch.regs[VCPU_REGS_RIP] = save->rip;
A
Avi Kivity 已提交
1149

J
Joerg Roedel 已提交
1150 1151
	/*
	 * This is the guest-visible cr0 value.
1152
	 * svm_set_cr0() sets PG and WP and clears NW and CD on save->cr0.
A
Avi Kivity 已提交
1153
	 */
1154 1155
	svm->vcpu.arch.cr0 = 0;
	(void)kvm_set_cr0(&svm->vcpu, X86_CR0_NW | X86_CR0_CD | X86_CR0_ET);
1156

1157
	save->cr4 = X86_CR4_PAE;
A
Avi Kivity 已提交
1158
	/* rdx = ?? */
1159 1160 1161 1162

	if (npt_enabled) {
		/* Setup VMCB for Nested Paging */
		control->nested_ctl = 1;
1163
		clr_intercept(svm, INTERCEPT_INVLPG);
1164
		clr_exception_intercept(svm, PF_VECTOR);
1165 1166
		clr_cr_intercept(svm, INTERCEPT_CR3_READ);
		clr_cr_intercept(svm, INTERCEPT_CR3_WRITE);
1167 1168 1169 1170
		save->g_pat = 0x0007040600070406ULL;
		save->cr3 = 0;
		save->cr4 = 0;
	}
1171
	svm->asid_generation = 0;
1172

1173
	svm->nested.vmcb = 0;
1174 1175
	svm->vcpu.arch.hflags = 0;

1176
	if (boot_cpu_has(X86_FEATURE_PAUSEFILTER)) {
1177
		control->pause_filter_count = 3000;
1178
		set_intercept(svm, INTERCEPT_PAUSE);
1179 1180
	}

1181 1182
	mark_all_dirty(svm->vmcb);

1183
	enable_gif(svm);
A
Avi Kivity 已提交
1184 1185
}

1186
static int svm_vcpu_reset(struct kvm_vcpu *vcpu)
1187 1188 1189
{
	struct vcpu_svm *svm = to_svm(vcpu);

1190
	init_vmcb(svm);
A
Avi Kivity 已提交
1191

1192
	if (!kvm_vcpu_is_bsp(vcpu)) {
1193
		kvm_rip_write(vcpu, 0);
1194 1195
		svm->vmcb->save.cs.base = svm->vcpu.arch.sipi_vector << 12;
		svm->vmcb->save.cs.selector = svm->vcpu.arch.sipi_vector << 8;
A
Avi Kivity 已提交
1196
	}
1197 1198
	vcpu->arch.regs_avail = ~0;
	vcpu->arch.regs_dirty = ~0;
1199 1200

	return 0;
1201 1202
}

R
Rusty Russell 已提交
1203
static struct kvm_vcpu *svm_create_vcpu(struct kvm *kvm, unsigned int id)
A
Avi Kivity 已提交
1204
{
1205
	struct vcpu_svm *svm;
A
Avi Kivity 已提交
1206
	struct page *page;
1207
	struct page *msrpm_pages;
A
Alexander Graf 已提交
1208
	struct page *hsave_page;
A
Alexander Graf 已提交
1209
	struct page *nested_msrpm_pages;
R
Rusty Russell 已提交
1210
	int err;
A
Avi Kivity 已提交
1211

1212
	svm = kmem_cache_zalloc(kvm_vcpu_cache, GFP_KERNEL);
R
Rusty Russell 已提交
1213 1214 1215 1216 1217
	if (!svm) {
		err = -ENOMEM;
		goto out;
	}

1218 1219
	svm->tsc_ratio = TSC_RATIO_DEFAULT;

R
Rusty Russell 已提交
1220 1221 1222 1223
	err = kvm_vcpu_init(&svm->vcpu, kvm, id);
	if (err)
		goto free_svm;

1224
	err = -ENOMEM;
A
Avi Kivity 已提交
1225
	page = alloc_page(GFP_KERNEL);
1226
	if (!page)
R
Rusty Russell 已提交
1227
		goto uninit;
A
Avi Kivity 已提交
1228

1229 1230
	msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER);
	if (!msrpm_pages)
1231
		goto free_page1;
A
Alexander Graf 已提交
1232 1233 1234

	nested_msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER);
	if (!nested_msrpm_pages)
1235
		goto free_page2;
1236

A
Alexander Graf 已提交
1237 1238
	hsave_page = alloc_page(GFP_KERNEL);
	if (!hsave_page)
1239 1240
		goto free_page3;

1241
	svm->nested.hsave = page_address(hsave_page);
A
Alexander Graf 已提交
1242

1243 1244 1245
	svm->msrpm = page_address(msrpm_pages);
	svm_vcpu_init_msrpm(svm->msrpm);

1246
	svm->nested.msrpm = page_address(nested_msrpm_pages);
1247
	svm_vcpu_init_msrpm(svm->nested.msrpm);
A
Alexander Graf 已提交
1248

1249 1250 1251 1252
	svm->vmcb = page_address(page);
	clear_page(svm->vmcb);
	svm->vmcb_pa = page_to_pfn(page) << PAGE_SHIFT;
	svm->asid_generation = 0;
1253
	init_vmcb(svm);
1254
	kvm_write_tsc(&svm->vcpu, 0);
1255

1256 1257 1258 1259
	err = fx_init(&svm->vcpu);
	if (err)
		goto free_page4;

1260
	svm->vcpu.arch.apic_base = 0xfee00000 | MSR_IA32_APICBASE_ENABLE;
1261
	if (kvm_vcpu_is_bsp(&svm->vcpu))
1262
		svm->vcpu.arch.apic_base |= MSR_IA32_APICBASE_BSP;
A
Avi Kivity 已提交
1263

1264 1265
	svm_init_osvw(&svm->vcpu);

R
Rusty Russell 已提交
1266
	return &svm->vcpu;
1267

1268 1269
free_page4:
	__free_page(hsave_page);
1270 1271 1272 1273 1274 1275
free_page3:
	__free_pages(nested_msrpm_pages, MSRPM_ALLOC_ORDER);
free_page2:
	__free_pages(msrpm_pages, MSRPM_ALLOC_ORDER);
free_page1:
	__free_page(page);
R
Rusty Russell 已提交
1276 1277 1278
uninit:
	kvm_vcpu_uninit(&svm->vcpu);
free_svm:
1279
	kmem_cache_free(kvm_vcpu_cache, svm);
R
Rusty Russell 已提交
1280 1281
out:
	return ERR_PTR(err);
A
Avi Kivity 已提交
1282 1283 1284 1285
}

static void svm_free_vcpu(struct kvm_vcpu *vcpu)
{
1286 1287
	struct vcpu_svm *svm = to_svm(vcpu);

R
Rusty Russell 已提交
1288
	__free_page(pfn_to_page(svm->vmcb_pa >> PAGE_SHIFT));
1289
	__free_pages(virt_to_page(svm->msrpm), MSRPM_ALLOC_ORDER);
1290 1291
	__free_page(virt_to_page(svm->nested.hsave));
	__free_pages(virt_to_page(svm->nested.msrpm), MSRPM_ALLOC_ORDER);
R
Rusty Russell 已提交
1292
	kvm_vcpu_uninit(vcpu);
1293
	kmem_cache_free(kvm_vcpu_cache, svm);
A
Avi Kivity 已提交
1294 1295
}

1296
static void svm_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
A
Avi Kivity 已提交
1297
{
1298
	struct vcpu_svm *svm = to_svm(vcpu);
1299
	int i;
1300 1301

	if (unlikely(cpu != vcpu->cpu)) {
1302
		svm->asid_generation = 0;
1303
		mark_all_dirty(svm->vmcb);
1304
	}
1305

1306 1307 1308
#ifdef CONFIG_X86_64
	rdmsrl(MSR_GS_BASE, to_svm(vcpu)->host.gs_base);
#endif
1309 1310 1311 1312
	savesegment(fs, svm->host.fs);
	savesegment(gs, svm->host.gs);
	svm->host.ldt = kvm_read_ldt();

1313
	for (i = 0; i < NR_HOST_SAVE_USER_MSRS; i++)
1314
		rdmsrl(host_save_user_msrs[i], svm->host_user_msrs[i]);
1315 1316 1317 1318 1319 1320

	if (static_cpu_has(X86_FEATURE_TSCRATEMSR) &&
	    svm->tsc_ratio != __get_cpu_var(current_tsc_ratio)) {
		__get_cpu_var(current_tsc_ratio) = svm->tsc_ratio;
		wrmsrl(MSR_AMD64_TSC_RATIO, svm->tsc_ratio);
	}
A
Avi Kivity 已提交
1321 1322 1323 1324
}

static void svm_vcpu_put(struct kvm_vcpu *vcpu)
{
1325
	struct vcpu_svm *svm = to_svm(vcpu);
1326 1327
	int i;

1328
	++vcpu->stat.host_state_reload;
1329 1330 1331 1332
	kvm_load_ldt(svm->host.ldt);
#ifdef CONFIG_X86_64
	loadsegment(fs, svm->host.fs);
	wrmsrl(MSR_KERNEL_GS_BASE, current->thread.gs);
1333
	load_gs_index(svm->host.gs);
1334
#else
1335
#ifdef CONFIG_X86_32_LAZY_GS
1336
	loadsegment(gs, svm->host.gs);
1337
#endif
1338
#endif
1339
	for (i = 0; i < NR_HOST_SAVE_USER_MSRS; i++)
1340
		wrmsrl(host_save_user_msrs[i], svm->host_user_msrs[i]);
A
Avi Kivity 已提交
1341 1342
}

K
Kevin Wolf 已提交
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
static void svm_update_cpl(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	int cpl;

	if (!is_protmode(vcpu))
		cpl = 0;
	else if (svm->vmcb->save.rflags & X86_EFLAGS_VM)
		cpl = 3;
	else
		cpl = svm->vmcb->save.cs.selector & 0x3;

	svm->vmcb->save.cpl = cpl;
}

A
Avi Kivity 已提交
1358 1359
static unsigned long svm_get_rflags(struct kvm_vcpu *vcpu)
{
1360
	return to_svm(vcpu)->vmcb->save.rflags;
A
Avi Kivity 已提交
1361 1362 1363 1364
}

static void svm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
{
1365 1366
	unsigned long old_rflags = to_svm(vcpu)->vmcb->save.rflags;

1367
	to_svm(vcpu)->vmcb->save.rflags = rflags;
1368 1369
	if ((old_rflags ^ rflags) & X86_EFLAGS_VM)
		svm_update_cpl(vcpu);
A
Avi Kivity 已提交
1370 1371
}

A
Avi Kivity 已提交
1372 1373 1374 1375 1376
static void svm_cache_reg(struct kvm_vcpu *vcpu, enum kvm_reg reg)
{
	switch (reg) {
	case VCPU_EXREG_PDPTR:
		BUG_ON(!npt_enabled);
1377
		load_pdptrs(vcpu, vcpu->arch.walk_mmu, kvm_read_cr3(vcpu));
A
Avi Kivity 已提交
1378 1379 1380 1381 1382 1383
		break;
	default:
		BUG();
	}
}

1384 1385
static void svm_set_vintr(struct vcpu_svm *svm)
{
1386
	set_intercept(svm, INTERCEPT_VINTR);
1387 1388 1389 1390
}

static void svm_clear_vintr(struct vcpu_svm *svm)
{
1391
	clr_intercept(svm, INTERCEPT_VINTR);
1392 1393
}

A
Avi Kivity 已提交
1394 1395
static struct vmcb_seg *svm_seg(struct kvm_vcpu *vcpu, int seg)
{
1396
	struct vmcb_save_area *save = &to_svm(vcpu)->vmcb->save;
A
Avi Kivity 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408

	switch (seg) {
	case VCPU_SREG_CS: return &save->cs;
	case VCPU_SREG_DS: return &save->ds;
	case VCPU_SREG_ES: return &save->es;
	case VCPU_SREG_FS: return &save->fs;
	case VCPU_SREG_GS: return &save->gs;
	case VCPU_SREG_SS: return &save->ss;
	case VCPU_SREG_TR: return &save->tr;
	case VCPU_SREG_LDTR: return &save->ldtr;
	}
	BUG();
A
Al Viro 已提交
1409
	return NULL;
A
Avi Kivity 已提交
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
}

static u64 svm_get_segment_base(struct kvm_vcpu *vcpu, int seg)
{
	struct vmcb_seg *s = svm_seg(vcpu, seg);

	return s->base;
}

static void svm_get_segment(struct kvm_vcpu *vcpu,
			    struct kvm_segment *var, int seg)
{
	struct vmcb_seg *s = svm_seg(vcpu, seg);

	var->base = s->base;
	var->limit = s->limit;
	var->selector = s->selector;
	var->type = s->attrib & SVM_SELECTOR_TYPE_MASK;
	var->s = (s->attrib >> SVM_SELECTOR_S_SHIFT) & 1;
	var->dpl = (s->attrib >> SVM_SELECTOR_DPL_SHIFT) & 3;
	var->present = (s->attrib >> SVM_SELECTOR_P_SHIFT) & 1;
	var->avl = (s->attrib >> SVM_SELECTOR_AVL_SHIFT) & 1;
	var->l = (s->attrib >> SVM_SELECTOR_L_SHIFT) & 1;
	var->db = (s->attrib >> SVM_SELECTOR_DB_SHIFT) & 1;
	var->g = (s->attrib >> SVM_SELECTOR_G_SHIFT) & 1;
1435

J
Joerg Roedel 已提交
1436 1437
	/*
	 * AMD's VMCB does not have an explicit unusable field, so emulate it
1438 1439 1440 1441
	 * for cross vendor migration purposes by "not present"
	 */
	var->unusable = !var->present || (var->type == 0);

1442 1443 1444 1445 1446 1447 1448
	switch (seg) {
	case VCPU_SREG_CS:
		/*
		 * SVM always stores 0 for the 'G' bit in the CS selector in
		 * the VMCB on a VMEXIT. This hurts cross-vendor migration:
		 * Intel's VMENTRY has a check on the 'G' bit.
		 */
1449
		var->g = s->limit > 0xfffff;
1450 1451 1452 1453 1454 1455
		break;
	case VCPU_SREG_TR:
		/*
		 * Work around a bug where the busy flag in the tr selector
		 * isn't exposed
		 */
1456
		var->type |= 0x2;
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471
		break;
	case VCPU_SREG_DS:
	case VCPU_SREG_ES:
	case VCPU_SREG_FS:
	case VCPU_SREG_GS:
		/*
		 * The accessed bit must always be set in the segment
		 * descriptor cache, although it can be cleared in the
		 * descriptor, the cached bit always remains at 1. Since
		 * Intel has a check on this, set it here to support
		 * cross-vendor migration.
		 */
		if (!var->unusable)
			var->type |= 0x1;
		break;
1472
	case VCPU_SREG_SS:
J
Joerg Roedel 已提交
1473 1474
		/*
		 * On AMD CPUs sometimes the DB bit in the segment
1475 1476 1477 1478 1479 1480 1481
		 * descriptor is left as 1, although the whole segment has
		 * been made unusable. Clear it here to pass an Intel VMX
		 * entry check when cross vendor migrating.
		 */
		if (var->unusable)
			var->db = 0;
		break;
1482
	}
A
Avi Kivity 已提交
1483 1484
}

1485 1486 1487 1488 1489 1490 1491
static int svm_get_cpl(struct kvm_vcpu *vcpu)
{
	struct vmcb_save_area *save = &to_svm(vcpu)->vmcb->save;

	return save->cpl;
}

1492
static void svm_get_idt(struct kvm_vcpu *vcpu, struct desc_ptr *dt)
A
Avi Kivity 已提交
1493
{
1494 1495
	struct vcpu_svm *svm = to_svm(vcpu);

1496 1497
	dt->size = svm->vmcb->save.idtr.limit;
	dt->address = svm->vmcb->save.idtr.base;
A
Avi Kivity 已提交
1498 1499
}

1500
static void svm_set_idt(struct kvm_vcpu *vcpu, struct desc_ptr *dt)
A
Avi Kivity 已提交
1501
{
1502 1503
	struct vcpu_svm *svm = to_svm(vcpu);

1504 1505
	svm->vmcb->save.idtr.limit = dt->size;
	svm->vmcb->save.idtr.base = dt->address ;
1506
	mark_dirty(svm->vmcb, VMCB_DT);
A
Avi Kivity 已提交
1507 1508
}

1509
static void svm_get_gdt(struct kvm_vcpu *vcpu, struct desc_ptr *dt)
A
Avi Kivity 已提交
1510
{
1511 1512
	struct vcpu_svm *svm = to_svm(vcpu);

1513 1514
	dt->size = svm->vmcb->save.gdtr.limit;
	dt->address = svm->vmcb->save.gdtr.base;
A
Avi Kivity 已提交
1515 1516
}

1517
static void svm_set_gdt(struct kvm_vcpu *vcpu, struct desc_ptr *dt)
A
Avi Kivity 已提交
1518
{
1519 1520
	struct vcpu_svm *svm = to_svm(vcpu);

1521 1522
	svm->vmcb->save.gdtr.limit = dt->size;
	svm->vmcb->save.gdtr.base = dt->address ;
1523
	mark_dirty(svm->vmcb, VMCB_DT);
A
Avi Kivity 已提交
1524 1525
}

1526 1527 1528 1529
static void svm_decache_cr0_guest_bits(struct kvm_vcpu *vcpu)
{
}

1530 1531 1532 1533
static void svm_decache_cr3(struct kvm_vcpu *vcpu)
{
}

1534
static void svm_decache_cr4_guest_bits(struct kvm_vcpu *vcpu)
1535 1536 1537
{
}

A
Avi Kivity 已提交
1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
static void update_cr0_intercept(struct vcpu_svm *svm)
{
	ulong gcr0 = svm->vcpu.arch.cr0;
	u64 *hcr0 = &svm->vmcb->save.cr0;

	if (!svm->vcpu.fpu_active)
		*hcr0 |= SVM_CR0_SELECTIVE_MASK;
	else
		*hcr0 = (*hcr0 & ~SVM_CR0_SELECTIVE_MASK)
			| (gcr0 & SVM_CR0_SELECTIVE_MASK);

1549
	mark_dirty(svm->vmcb, VMCB_CR);
A
Avi Kivity 已提交
1550 1551

	if (gcr0 == *hcr0 && svm->vcpu.fpu_active) {
1552 1553
		clr_cr_intercept(svm, INTERCEPT_CR0_READ);
		clr_cr_intercept(svm, INTERCEPT_CR0_WRITE);
A
Avi Kivity 已提交
1554
	} else {
1555 1556
		set_cr_intercept(svm, INTERCEPT_CR0_READ);
		set_cr_intercept(svm, INTERCEPT_CR0_WRITE);
A
Avi Kivity 已提交
1557 1558 1559
	}
}

A
Avi Kivity 已提交
1560 1561
static void svm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
{
1562 1563
	struct vcpu_svm *svm = to_svm(vcpu);

1564
#ifdef CONFIG_X86_64
1565
	if (vcpu->arch.efer & EFER_LME) {
1566
		if (!is_paging(vcpu) && (cr0 & X86_CR0_PG)) {
1567
			vcpu->arch.efer |= EFER_LMA;
1568
			svm->vmcb->save.efer |= EFER_LMA | EFER_LME;
A
Avi Kivity 已提交
1569 1570
		}

M
Mike Day 已提交
1571
		if (is_paging(vcpu) && !(cr0 & X86_CR0_PG)) {
1572
			vcpu->arch.efer &= ~EFER_LMA;
1573
			svm->vmcb->save.efer &= ~(EFER_LMA | EFER_LME);
A
Avi Kivity 已提交
1574 1575 1576
		}
	}
#endif
1577
	vcpu->arch.cr0 = cr0;
1578 1579 1580

	if (!npt_enabled)
		cr0 |= X86_CR0_PG | X86_CR0_WP;
1581 1582

	if (!vcpu->fpu_active)
J
Joerg Roedel 已提交
1583
		cr0 |= X86_CR0_TS;
1584 1585 1586 1587 1588 1589
	/*
	 * re-enable caching here because the QEMU bios
	 * does not do it - this results in some delay at
	 * reboot
	 */
	cr0 &= ~(X86_CR0_CD | X86_CR0_NW);
1590
	svm->vmcb->save.cr0 = cr0;
1591
	mark_dirty(svm->vmcb, VMCB_CR);
A
Avi Kivity 已提交
1592
	update_cr0_intercept(svm);
A
Avi Kivity 已提交
1593 1594
}

1595
static int svm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
A
Avi Kivity 已提交
1596
{
1597
	unsigned long host_cr4_mce = read_cr4() & X86_CR4_MCE;
1598 1599
	unsigned long old_cr4 = to_svm(vcpu)->vmcb->save.cr4;

1600 1601 1602
	if (cr4 & X86_CR4_VMXE)
		return 1;

1603
	if (npt_enabled && ((old_cr4 ^ cr4) & X86_CR4_PGE))
1604
		svm_flush_tlb(vcpu);
1605

1606 1607 1608
	vcpu->arch.cr4 = cr4;
	if (!npt_enabled)
		cr4 |= X86_CR4_PAE;
1609
	cr4 |= host_cr4_mce;
1610
	to_svm(vcpu)->vmcb->save.cr4 = cr4;
1611
	mark_dirty(to_svm(vcpu)->vmcb, VMCB_CR);
1612
	return 0;
A
Avi Kivity 已提交
1613 1614 1615 1616 1617
}

static void svm_set_segment(struct kvm_vcpu *vcpu,
			    struct kvm_segment *var, int seg)
{
1618
	struct vcpu_svm *svm = to_svm(vcpu);
A
Avi Kivity 已提交
1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
	struct vmcb_seg *s = svm_seg(vcpu, seg);

	s->base = var->base;
	s->limit = var->limit;
	s->selector = var->selector;
	if (var->unusable)
		s->attrib = 0;
	else {
		s->attrib = (var->type & SVM_SELECTOR_TYPE_MASK);
		s->attrib |= (var->s & 1) << SVM_SELECTOR_S_SHIFT;
		s->attrib |= (var->dpl & 3) << SVM_SELECTOR_DPL_SHIFT;
		s->attrib |= (var->present & 1) << SVM_SELECTOR_P_SHIFT;
		s->attrib |= (var->avl & 1) << SVM_SELECTOR_AVL_SHIFT;
		s->attrib |= (var->l & 1) << SVM_SELECTOR_L_SHIFT;
		s->attrib |= (var->db & 1) << SVM_SELECTOR_DB_SHIFT;
		s->attrib |= (var->g & 1) << SVM_SELECTOR_G_SHIFT;
	}
	if (seg == VCPU_SREG_CS)
K
Kevin Wolf 已提交
1637
		svm_update_cpl(vcpu);
A
Avi Kivity 已提交
1638

1639
	mark_dirty(svm->vmcb, VMCB_SEG);
A
Avi Kivity 已提交
1640 1641
}

1642
static void update_db_bp_intercept(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
1643
{
J
Jan Kiszka 已提交
1644 1645
	struct vcpu_svm *svm = to_svm(vcpu);

1646 1647
	clr_exception_intercept(svm, DB_VECTOR);
	clr_exception_intercept(svm, BP_VECTOR);
1648

J
Jan Kiszka 已提交
1649
	if (svm->nmi_singlestep)
1650
		set_exception_intercept(svm, DB_VECTOR);
1651

J
Jan Kiszka 已提交
1652 1653 1654
	if (vcpu->guest_debug & KVM_GUESTDBG_ENABLE) {
		if (vcpu->guest_debug &
		    (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))
1655
			set_exception_intercept(svm, DB_VECTOR);
J
Jan Kiszka 已提交
1656
		if (vcpu->guest_debug & KVM_GUESTDBG_USE_SW_BP)
1657
			set_exception_intercept(svm, BP_VECTOR);
J
Jan Kiszka 已提交
1658 1659
	} else
		vcpu->guest_debug = 0;
1660 1661
}

1662
static void new_asid(struct vcpu_svm *svm, struct svm_cpu_data *sd)
A
Avi Kivity 已提交
1663
{
1664 1665 1666
	if (sd->next_asid > sd->max_asid) {
		++sd->asid_generation;
		sd->next_asid = 1;
1667
		svm->vmcb->control.tlb_ctl = TLB_CONTROL_FLUSH_ALL_ASID;
A
Avi Kivity 已提交
1668 1669
	}

1670 1671
	svm->asid_generation = sd->asid_generation;
	svm->vmcb->control.asid = sd->next_asid++;
1672 1673

	mark_dirty(svm->vmcb, VMCB_ASID);
A
Avi Kivity 已提交
1674 1675
}

1676
static void svm_set_dr7(struct kvm_vcpu *vcpu, unsigned long value)
A
Avi Kivity 已提交
1677
{
1678 1679
	struct vcpu_svm *svm = to_svm(vcpu);

1680
	svm->vmcb->save.dr7 = value;
1681
	mark_dirty(svm->vmcb, VMCB_DR);
A
Avi Kivity 已提交
1682 1683
}

A
Avi Kivity 已提交
1684
static int pf_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
1685
{
G
Gleb Natapov 已提交
1686
	u64 fault_address = svm->vmcb->control.exit_info_2;
A
Avi Kivity 已提交
1687
	u32 error_code;
G
Gleb Natapov 已提交
1688
	int r = 1;
A
Avi Kivity 已提交
1689

G
Gleb Natapov 已提交
1690 1691 1692
	switch (svm->apf_reason) {
	default:
		error_code = svm->vmcb->control.exit_info_1;
1693

G
Gleb Natapov 已提交
1694 1695 1696
		trace_kvm_page_fault(fault_address, error_code);
		if (!npt_enabled && kvm_event_needs_reinjection(&svm->vcpu))
			kvm_mmu_unprotect_page_virt(&svm->vcpu, fault_address);
1697 1698 1699
		r = kvm_mmu_page_fault(&svm->vcpu, fault_address, error_code,
			svm->vmcb->control.insn_bytes,
			svm->vmcb->control.insn_len);
G
Gleb Natapov 已提交
1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
		break;
	case KVM_PV_REASON_PAGE_NOT_PRESENT:
		svm->apf_reason = 0;
		local_irq_disable();
		kvm_async_pf_task_wait(fault_address);
		local_irq_enable();
		break;
	case KVM_PV_REASON_PAGE_READY:
		svm->apf_reason = 0;
		local_irq_disable();
		kvm_async_pf_task_wake(fault_address);
		local_irq_enable();
		break;
	}
	return r;
A
Avi Kivity 已提交
1715 1716
}

A
Avi Kivity 已提交
1717
static int db_interception(struct vcpu_svm *svm)
J
Jan Kiszka 已提交
1718
{
A
Avi Kivity 已提交
1719 1720
	struct kvm_run *kvm_run = svm->vcpu.run;

J
Jan Kiszka 已提交
1721
	if (!(svm->vcpu.guest_debug &
1722
	      (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP)) &&
J
Jan Kiszka 已提交
1723
		!svm->nmi_singlestep) {
J
Jan Kiszka 已提交
1724 1725 1726
		kvm_queue_exception(&svm->vcpu, DB_VECTOR);
		return 1;
	}
1727

J
Jan Kiszka 已提交
1728 1729
	if (svm->nmi_singlestep) {
		svm->nmi_singlestep = false;
1730 1731 1732
		if (!(svm->vcpu.guest_debug & KVM_GUESTDBG_SINGLESTEP))
			svm->vmcb->save.rflags &=
				~(X86_EFLAGS_TF | X86_EFLAGS_RF);
1733
		update_db_bp_intercept(&svm->vcpu);
1734 1735 1736
	}

	if (svm->vcpu.guest_debug &
J
Joerg Roedel 已提交
1737
	    (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP)) {
1738 1739 1740 1741 1742 1743 1744 1745
		kvm_run->exit_reason = KVM_EXIT_DEBUG;
		kvm_run->debug.arch.pc =
			svm->vmcb->save.cs.base + svm->vmcb->save.rip;
		kvm_run->debug.arch.exception = DB_VECTOR;
		return 0;
	}

	return 1;
J
Jan Kiszka 已提交
1746 1747
}

A
Avi Kivity 已提交
1748
static int bp_interception(struct vcpu_svm *svm)
J
Jan Kiszka 已提交
1749
{
A
Avi Kivity 已提交
1750 1751
	struct kvm_run *kvm_run = svm->vcpu.run;

J
Jan Kiszka 已提交
1752 1753 1754 1755 1756 1757
	kvm_run->exit_reason = KVM_EXIT_DEBUG;
	kvm_run->debug.arch.pc = svm->vmcb->save.cs.base + svm->vmcb->save.rip;
	kvm_run->debug.arch.exception = BP_VECTOR;
	return 0;
}

A
Avi Kivity 已提交
1758
static int ud_interception(struct vcpu_svm *svm)
1759 1760 1761
{
	int er;

1762
	er = emulate_instruction(&svm->vcpu, EMULTYPE_TRAP_UD);
1763
	if (er != EMULATE_DONE)
1764
		kvm_queue_exception(&svm->vcpu, UD_VECTOR);
1765 1766 1767
	return 1;
}

A
Avi Kivity 已提交
1768
static void svm_fpu_activate(struct kvm_vcpu *vcpu)
A
Anthony Liguori 已提交
1769
{
A
Avi Kivity 已提交
1770
	struct vcpu_svm *svm = to_svm(vcpu);
1771

1772
	clr_exception_intercept(svm, NM_VECTOR);
1773

R
Rusty Russell 已提交
1774
	svm->vcpu.fpu_active = 1;
A
Avi Kivity 已提交
1775
	update_cr0_intercept(svm);
A
Avi Kivity 已提交
1776
}
1777

A
Avi Kivity 已提交
1778 1779 1780
static int nm_interception(struct vcpu_svm *svm)
{
	svm_fpu_activate(&svm->vcpu);
1781
	return 1;
A
Anthony Liguori 已提交
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
static bool is_erratum_383(void)
{
	int err, i;
	u64 value;

	if (!erratum_383_found)
		return false;

	value = native_read_msr_safe(MSR_IA32_MC0_STATUS, &err);
	if (err)
		return false;

	/* Bit 62 may or may not be set for this mce */
	value &= ~(1ULL << 62);

	if (value != 0xb600000000010015ULL)
		return false;

	/* Clear MCi_STATUS registers */
	for (i = 0; i < 6; ++i)
		native_write_msr_safe(MSR_IA32_MCx_STATUS(i), 0, 0);

	value = native_read_msr_safe(MSR_IA32_MCG_STATUS, &err);
	if (!err) {
		u32 low, high;

		value &= ~(1ULL << 2);
		low    = lower_32_bits(value);
		high   = upper_32_bits(value);

		native_write_msr_safe(MSR_IA32_MCG_STATUS, low, high);
	}

	/* Flush tlb to evict multi-match entries */
	__flush_tlb_all();

	return true;
}

1823
static void svm_handle_mce(struct vcpu_svm *svm)
1824
{
1825 1826 1827 1828 1829 1830 1831
	if (is_erratum_383()) {
		/*
		 * Erratum 383 triggered. Guest state is corrupt so kill the
		 * guest.
		 */
		pr_err("KVM: Guest triggered AMD Erratum 383\n");

1832
		kvm_make_request(KVM_REQ_TRIPLE_FAULT, &svm->vcpu);
1833 1834 1835 1836

		return;
	}

1837 1838 1839 1840 1841 1842 1843 1844
	/*
	 * On an #MC intercept the MCE handler is not called automatically in
	 * the host. So do it by hand here.
	 */
	asm volatile (
		"int $0x12\n");
	/* not sure if we ever come back to this point */

1845 1846 1847 1848 1849
	return;
}

static int mc_interception(struct vcpu_svm *svm)
{
1850 1851 1852
	return 1;
}

A
Avi Kivity 已提交
1853
static int shutdown_interception(struct vcpu_svm *svm)
1854
{
A
Avi Kivity 已提交
1855 1856
	struct kvm_run *kvm_run = svm->vcpu.run;

1857 1858 1859 1860
	/*
	 * VMCB is undefined after a SHUTDOWN intercept
	 * so reinitialize it.
	 */
1861
	clear_page(svm->vmcb);
1862
	init_vmcb(svm);
1863 1864 1865 1866 1867

	kvm_run->exit_reason = KVM_EXIT_SHUTDOWN;
	return 0;
}

A
Avi Kivity 已提交
1868
static int io_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
1869
{
1870
	struct kvm_vcpu *vcpu = &svm->vcpu;
M
Mike Day 已提交
1871
	u32 io_info = svm->vmcb->control.exit_info_1; /* address size bug? */
1872
	int size, in, string;
1873
	unsigned port;
A
Avi Kivity 已提交
1874

R
Rusty Russell 已提交
1875
	++svm->vcpu.stat.io_exits;
1876
	string = (io_info & SVM_IOIO_STR_MASK) != 0;
1877
	in = (io_info & SVM_IOIO_TYPE_MASK) != 0;
1878
	if (string || in)
1879
		return emulate_instruction(vcpu, 0) == EMULATE_DONE;
1880

1881 1882
	port = io_info >> 16;
	size = (io_info & SVM_IOIO_SIZE_MASK) >> SVM_IOIO_SIZE_SHIFT;
1883
	svm->next_rip = svm->vmcb->control.exit_info_2;
1884
	skip_emulated_instruction(&svm->vcpu);
1885 1886

	return kvm_fast_pio_out(vcpu, size, port);
A
Avi Kivity 已提交
1887 1888
}

A
Avi Kivity 已提交
1889
static int nmi_interception(struct vcpu_svm *svm)
1890 1891 1892 1893
{
	return 1;
}

A
Avi Kivity 已提交
1894
static int intr_interception(struct vcpu_svm *svm)
1895 1896 1897 1898 1899
{
	++svm->vcpu.stat.irq_exits;
	return 1;
}

A
Avi Kivity 已提交
1900
static int nop_on_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
1901 1902 1903 1904
{
	return 1;
}

A
Avi Kivity 已提交
1905
static int halt_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
1906
{
1907
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 1;
R
Rusty Russell 已提交
1908 1909
	skip_emulated_instruction(&svm->vcpu);
	return kvm_emulate_halt(&svm->vcpu);
A
Avi Kivity 已提交
1910 1911
}

A
Avi Kivity 已提交
1912
static int vmmcall_interception(struct vcpu_svm *svm)
1913
{
1914
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
R
Rusty Russell 已提交
1915
	skip_emulated_instruction(&svm->vcpu);
1916 1917
	kvm_emulate_hypercall(&svm->vcpu);
	return 1;
1918 1919
}

1920 1921 1922 1923 1924 1925 1926
static unsigned long nested_svm_get_tdp_cr3(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	return svm->nested.nested_cr3;
}

1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940
static u64 nested_svm_get_tdp_pdptr(struct kvm_vcpu *vcpu, int index)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u64 cr3 = svm->nested.nested_cr3;
	u64 pdpte;
	int ret;

	ret = kvm_read_guest_page(vcpu->kvm, gpa_to_gfn(cr3), &pdpte,
				  offset_in_page(cr3) + index * 8, 8);
	if (ret)
		return 0;
	return pdpte;
}

1941 1942 1943 1944 1945 1946
static void nested_svm_set_tdp_cr3(struct kvm_vcpu *vcpu,
				   unsigned long root)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	svm->vmcb->control.nested_cr3 = root;
1947
	mark_dirty(svm->vmcb, VMCB_NPT);
1948
	svm_flush_tlb(vcpu);
1949 1950
}

1951 1952
static void nested_svm_inject_npf_exit(struct kvm_vcpu *vcpu,
				       struct x86_exception *fault)
1953 1954 1955 1956 1957
{
	struct vcpu_svm *svm = to_svm(vcpu);

	svm->vmcb->control.exit_code = SVM_EXIT_NPF;
	svm->vmcb->control.exit_code_hi = 0;
1958 1959
	svm->vmcb->control.exit_info_1 = fault->error_code;
	svm->vmcb->control.exit_info_2 = fault->address;
1960 1961 1962 1963

	nested_svm_vmexit(svm);
}

1964 1965 1966 1967 1968 1969 1970 1971
static int nested_svm_init_mmu_context(struct kvm_vcpu *vcpu)
{
	int r;

	r = kvm_init_shadow_mmu(vcpu, &vcpu->arch.mmu);

	vcpu->arch.mmu.set_cr3           = nested_svm_set_tdp_cr3;
	vcpu->arch.mmu.get_cr3           = nested_svm_get_tdp_cr3;
1972
	vcpu->arch.mmu.get_pdptr         = nested_svm_get_tdp_pdptr;
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
	vcpu->arch.mmu.inject_page_fault = nested_svm_inject_npf_exit;
	vcpu->arch.mmu.shadow_root_level = get_npt_level();
	vcpu->arch.walk_mmu              = &vcpu->arch.nested_mmu;

	return r;
}

static void nested_svm_uninit_mmu_context(struct kvm_vcpu *vcpu)
{
	vcpu->arch.walk_mmu = &vcpu->arch.mmu;
}

1985 1986
static int nested_svm_check_permissions(struct vcpu_svm *svm)
{
1987
	if (!(svm->vcpu.arch.efer & EFER_SVME)
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
	    || !is_paging(&svm->vcpu)) {
		kvm_queue_exception(&svm->vcpu, UD_VECTOR);
		return 1;
	}

	if (svm->vmcb->save.cpl) {
		kvm_inject_gp(&svm->vcpu, 0);
		return 1;
	}

       return 0;
}

2001 2002 2003
static int nested_svm_check_exception(struct vcpu_svm *svm, unsigned nr,
				      bool has_error_code, u32 error_code)
{
2004 2005
	int vmexit;

2006
	if (!is_guest_mode(&svm->vcpu))
2007
		return 0;
2008

2009 2010 2011 2012 2013
	svm->vmcb->control.exit_code = SVM_EXIT_EXCP_BASE + nr;
	svm->vmcb->control.exit_code_hi = 0;
	svm->vmcb->control.exit_info_1 = error_code;
	svm->vmcb->control.exit_info_2 = svm->vcpu.arch.cr2;

2014 2015 2016 2017 2018
	vmexit = nested_svm_intercept(svm);
	if (vmexit == NESTED_EXIT_DONE)
		svm->nested.exit_required = true;

	return vmexit;
2019 2020
}

2021 2022
/* This function returns true if it is save to enable the irq window */
static inline bool nested_svm_intr(struct vcpu_svm *svm)
2023
{
2024
	if (!is_guest_mode(&svm->vcpu))
2025
		return true;
2026

2027
	if (!(svm->vcpu.arch.hflags & HF_VINTR_MASK))
2028
		return true;
2029

2030
	if (!(svm->vcpu.arch.hflags & HF_HIF_MASK))
2031
		return false;
2032

2033 2034 2035 2036 2037 2038 2039 2040
	/*
	 * if vmexit was already requested (by intercepted exception
	 * for instance) do not overwrite it with "external interrupt"
	 * vmexit.
	 */
	if (svm->nested.exit_required)
		return false;

2041 2042 2043
	svm->vmcb->control.exit_code   = SVM_EXIT_INTR;
	svm->vmcb->control.exit_info_1 = 0;
	svm->vmcb->control.exit_info_2 = 0;
2044

2045 2046 2047
	if (svm->nested.intercept & 1ULL) {
		/*
		 * The #vmexit can't be emulated here directly because this
G
Guo Chao 已提交
2048
		 * code path runs with irqs and preemption disabled. A
2049 2050 2051 2052
		 * #vmexit emulation might sleep. Only signal request for
		 * the #vmexit here.
		 */
		svm->nested.exit_required = true;
2053
		trace_kvm_nested_intr_vmexit(svm->vmcb->save.rip);
2054
		return false;
2055 2056
	}

2057
	return true;
2058 2059
}

2060 2061 2062
/* This function returns true if it is save to enable the nmi window */
static inline bool nested_svm_nmi(struct vcpu_svm *svm)
{
2063
	if (!is_guest_mode(&svm->vcpu))
2064 2065 2066 2067 2068 2069 2070 2071 2072
		return true;

	if (!(svm->nested.intercept & (1ULL << INTERCEPT_NMI)))
		return true;

	svm->vmcb->control.exit_code = SVM_EXIT_NMI;
	svm->nested.exit_required = true;

	return false;
2073 2074
}

2075
static void *nested_svm_map(struct vcpu_svm *svm, u64 gpa, struct page **_page)
2076 2077 2078
{
	struct page *page;

2079 2080
	might_sleep();

2081 2082 2083 2084
	page = gfn_to_page(svm->vcpu.kvm, gpa >> PAGE_SHIFT);
	if (is_error_page(page))
		goto error;

2085 2086 2087
	*_page = page;

	return kmap(page);
2088 2089 2090 2091 2092 2093 2094

error:
	kvm_inject_gp(&svm->vcpu, 0);

	return NULL;
}

2095
static void nested_svm_unmap(struct page *page)
2096
{
2097
	kunmap(page);
2098 2099 2100
	kvm_release_page_dirty(page);
}

2101 2102 2103 2104 2105
static int nested_svm_intercept_ioio(struct vcpu_svm *svm)
{
	unsigned port;
	u8 val, bit;
	u64 gpa;
2106

2107 2108
	if (!(svm->nested.intercept & (1ULL << INTERCEPT_IOIO_PROT)))
		return NESTED_EXIT_HOST;
2109

2110 2111 2112 2113 2114 2115 2116 2117 2118
	port = svm->vmcb->control.exit_info_1 >> 16;
	gpa  = svm->nested.vmcb_iopm + (port / 8);
	bit  = port % 8;
	val  = 0;

	if (kvm_read_guest(svm->vcpu.kvm, gpa, &val, 1))
		val &= (1 << bit);

	return val ? NESTED_EXIT_DONE : NESTED_EXIT_HOST;
2119 2120
}

2121
static int nested_svm_exit_handled_msr(struct vcpu_svm *svm)
2122
{
2123 2124
	u32 offset, msr, value;
	int write, mask;
2125

2126
	if (!(svm->nested.intercept & (1ULL << INTERCEPT_MSR_PROT)))
2127
		return NESTED_EXIT_HOST;
2128

2129 2130 2131 2132
	msr    = svm->vcpu.arch.regs[VCPU_REGS_RCX];
	offset = svm_msrpm_offset(msr);
	write  = svm->vmcb->control.exit_info_1 & 1;
	mask   = 1 << ((2 * (msr & 0xf)) + write);
2133

2134 2135
	if (offset == MSR_INVALID)
		return NESTED_EXIT_DONE;
2136

2137 2138
	/* Offset is in 32 bit units but need in 8 bit units */
	offset *= 4;
2139

2140 2141
	if (kvm_read_guest(svm->vcpu.kvm, svm->nested.vmcb_msrpm + offset, &value, 4))
		return NESTED_EXIT_DONE;
2142

2143
	return (value & mask) ? NESTED_EXIT_DONE : NESTED_EXIT_HOST;
2144 2145
}

2146
static int nested_svm_exit_special(struct vcpu_svm *svm)
2147 2148
{
	u32 exit_code = svm->vmcb->control.exit_code;
2149

2150 2151 2152
	switch (exit_code) {
	case SVM_EXIT_INTR:
	case SVM_EXIT_NMI:
2153
	case SVM_EXIT_EXCP_BASE + MC_VECTOR:
2154 2155
		return NESTED_EXIT_HOST;
	case SVM_EXIT_NPF:
J
Joerg Roedel 已提交
2156
		/* For now we are always handling NPFs when using them */
2157 2158 2159 2160
		if (npt_enabled)
			return NESTED_EXIT_HOST;
		break;
	case SVM_EXIT_EXCP_BASE + PF_VECTOR:
G
Gleb Natapov 已提交
2161 2162
		/* When we're shadowing, trap PFs, but not async PF */
		if (!npt_enabled && svm->apf_reason == 0)
2163 2164
			return NESTED_EXIT_HOST;
		break;
2165 2166 2167
	case SVM_EXIT_EXCP_BASE + NM_VECTOR:
		nm_interception(svm);
		break;
2168 2169
	default:
		break;
2170 2171
	}

2172 2173 2174 2175 2176 2177
	return NESTED_EXIT_CONTINUE;
}

/*
 * If this function returns true, this #vmexit was already handled
 */
2178
static int nested_svm_intercept(struct vcpu_svm *svm)
2179 2180 2181 2182
{
	u32 exit_code = svm->vmcb->control.exit_code;
	int vmexit = NESTED_EXIT_HOST;

2183
	switch (exit_code) {
2184
	case SVM_EXIT_MSR:
2185
		vmexit = nested_svm_exit_handled_msr(svm);
2186
		break;
2187 2188 2189
	case SVM_EXIT_IOIO:
		vmexit = nested_svm_intercept_ioio(svm);
		break;
2190 2191 2192
	case SVM_EXIT_READ_CR0 ... SVM_EXIT_WRITE_CR8: {
		u32 bit = 1U << (exit_code - SVM_EXIT_READ_CR0);
		if (svm->nested.intercept_cr & bit)
2193
			vmexit = NESTED_EXIT_DONE;
2194 2195
		break;
	}
2196 2197 2198
	case SVM_EXIT_READ_DR0 ... SVM_EXIT_WRITE_DR7: {
		u32 bit = 1U << (exit_code - SVM_EXIT_READ_DR0);
		if (svm->nested.intercept_dr & bit)
2199
			vmexit = NESTED_EXIT_DONE;
2200 2201 2202 2203
		break;
	}
	case SVM_EXIT_EXCP_BASE ... SVM_EXIT_EXCP_BASE + 0x1f: {
		u32 excp_bits = 1 << (exit_code - SVM_EXIT_EXCP_BASE);
J
Joerg Roedel 已提交
2204
		if (svm->nested.intercept_exceptions & excp_bits)
2205
			vmexit = NESTED_EXIT_DONE;
G
Gleb Natapov 已提交
2206 2207 2208 2209
		/* async page fault always cause vmexit */
		else if ((exit_code == SVM_EXIT_EXCP_BASE + PF_VECTOR) &&
			 svm->apf_reason != 0)
			vmexit = NESTED_EXIT_DONE;
2210 2211
		break;
	}
2212 2213 2214 2215
	case SVM_EXIT_ERR: {
		vmexit = NESTED_EXIT_DONE;
		break;
	}
2216 2217
	default: {
		u64 exit_bits = 1ULL << (exit_code - SVM_EXIT_INTR);
J
Joerg Roedel 已提交
2218
		if (svm->nested.intercept & exit_bits)
2219
			vmexit = NESTED_EXIT_DONE;
2220 2221 2222
	}
	}

2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
	return vmexit;
}

static int nested_svm_exit_handled(struct vcpu_svm *svm)
{
	int vmexit;

	vmexit = nested_svm_intercept(svm);

	if (vmexit == NESTED_EXIT_DONE)
2233 2234 2235
		nested_svm_vmexit(svm);

	return vmexit;
2236 2237
}

2238 2239 2240 2241 2242
static inline void copy_vmcb_control_area(struct vmcb *dst_vmcb, struct vmcb *from_vmcb)
{
	struct vmcb_control_area *dst  = &dst_vmcb->control;
	struct vmcb_control_area *from = &from_vmcb->control;

2243
	dst->intercept_cr         = from->intercept_cr;
2244
	dst->intercept_dr         = from->intercept_dr;
2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267
	dst->intercept_exceptions = from->intercept_exceptions;
	dst->intercept            = from->intercept;
	dst->iopm_base_pa         = from->iopm_base_pa;
	dst->msrpm_base_pa        = from->msrpm_base_pa;
	dst->tsc_offset           = from->tsc_offset;
	dst->asid                 = from->asid;
	dst->tlb_ctl              = from->tlb_ctl;
	dst->int_ctl              = from->int_ctl;
	dst->int_vector           = from->int_vector;
	dst->int_state            = from->int_state;
	dst->exit_code            = from->exit_code;
	dst->exit_code_hi         = from->exit_code_hi;
	dst->exit_info_1          = from->exit_info_1;
	dst->exit_info_2          = from->exit_info_2;
	dst->exit_int_info        = from->exit_int_info;
	dst->exit_int_info_err    = from->exit_int_info_err;
	dst->nested_ctl           = from->nested_ctl;
	dst->event_inj            = from->event_inj;
	dst->event_inj_err        = from->event_inj_err;
	dst->nested_cr3           = from->nested_cr3;
	dst->lbr_ctl              = from->lbr_ctl;
}

2268
static int nested_svm_vmexit(struct vcpu_svm *svm)
2269
{
2270
	struct vmcb *nested_vmcb;
2271
	struct vmcb *hsave = svm->nested.hsave;
J
Joerg Roedel 已提交
2272
	struct vmcb *vmcb = svm->vmcb;
2273
	struct page *page;
2274

2275 2276 2277 2278
	trace_kvm_nested_vmexit_inject(vmcb->control.exit_code,
				       vmcb->control.exit_info_1,
				       vmcb->control.exit_info_2,
				       vmcb->control.exit_int_info,
2279 2280
				       vmcb->control.exit_int_info_err,
				       KVM_ISA_SVM);
2281

2282
	nested_vmcb = nested_svm_map(svm, svm->nested.vmcb, &page);
2283 2284 2285
	if (!nested_vmcb)
		return 1;

2286 2287
	/* Exit Guest-Mode */
	leave_guest_mode(&svm->vcpu);
2288 2289
	svm->nested.vmcb = 0;

2290
	/* Give the current vmcb to the guest */
J
Joerg Roedel 已提交
2291 2292 2293 2294 2295 2296 2297 2298
	disable_gif(svm);

	nested_vmcb->save.es     = vmcb->save.es;
	nested_vmcb->save.cs     = vmcb->save.cs;
	nested_vmcb->save.ss     = vmcb->save.ss;
	nested_vmcb->save.ds     = vmcb->save.ds;
	nested_vmcb->save.gdtr   = vmcb->save.gdtr;
	nested_vmcb->save.idtr   = vmcb->save.idtr;
2299
	nested_vmcb->save.efer   = svm->vcpu.arch.efer;
2300
	nested_vmcb->save.cr0    = kvm_read_cr0(&svm->vcpu);
2301
	nested_vmcb->save.cr3    = kvm_read_cr3(&svm->vcpu);
J
Joerg Roedel 已提交
2302
	nested_vmcb->save.cr2    = vmcb->save.cr2;
2303
	nested_vmcb->save.cr4    = svm->vcpu.arch.cr4;
2304
	nested_vmcb->save.rflags = kvm_get_rflags(&svm->vcpu);
J
Joerg Roedel 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320
	nested_vmcb->save.rip    = vmcb->save.rip;
	nested_vmcb->save.rsp    = vmcb->save.rsp;
	nested_vmcb->save.rax    = vmcb->save.rax;
	nested_vmcb->save.dr7    = vmcb->save.dr7;
	nested_vmcb->save.dr6    = vmcb->save.dr6;
	nested_vmcb->save.cpl    = vmcb->save.cpl;

	nested_vmcb->control.int_ctl           = vmcb->control.int_ctl;
	nested_vmcb->control.int_vector        = vmcb->control.int_vector;
	nested_vmcb->control.int_state         = vmcb->control.int_state;
	nested_vmcb->control.exit_code         = vmcb->control.exit_code;
	nested_vmcb->control.exit_code_hi      = vmcb->control.exit_code_hi;
	nested_vmcb->control.exit_info_1       = vmcb->control.exit_info_1;
	nested_vmcb->control.exit_info_2       = vmcb->control.exit_info_2;
	nested_vmcb->control.exit_int_info     = vmcb->control.exit_int_info;
	nested_vmcb->control.exit_int_info_err = vmcb->control.exit_int_info_err;
2321
	nested_vmcb->control.next_rip          = vmcb->control.next_rip;
2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337

	/*
	 * If we emulate a VMRUN/#VMEXIT in the same host #vmexit cycle we have
	 * to make sure that we do not lose injected events. So check event_inj
	 * here and copy it to exit_int_info if it is valid.
	 * Exit_int_info and event_inj can't be both valid because the case
	 * below only happens on a VMRUN instruction intercept which has
	 * no valid exit_int_info set.
	 */
	if (vmcb->control.event_inj & SVM_EVTINJ_VALID) {
		struct vmcb_control_area *nc = &nested_vmcb->control;

		nc->exit_int_info     = vmcb->control.event_inj;
		nc->exit_int_info_err = vmcb->control.event_inj_err;
	}

J
Joerg Roedel 已提交
2338 2339 2340
	nested_vmcb->control.tlb_ctl           = 0;
	nested_vmcb->control.event_inj         = 0;
	nested_vmcb->control.event_inj_err     = 0;
2341 2342 2343 2344 2345 2346

	/* We always set V_INTR_MASKING and remember the old value in hflags */
	if (!(svm->vcpu.arch.hflags & HF_VINTR_MASK))
		nested_vmcb->control.int_ctl &= ~V_INTR_MASKING_MASK;

	/* Restore the original control entries */
2347
	copy_vmcb_control_area(vmcb, hsave);
2348

2349 2350
	kvm_clear_exception_queue(&svm->vcpu);
	kvm_clear_interrupt_queue(&svm->vcpu);
2351

2352 2353
	svm->nested.nested_cr3 = 0;

2354 2355 2356 2357 2358 2359 2360
	/* Restore selected save entries */
	svm->vmcb->save.es = hsave->save.es;
	svm->vmcb->save.cs = hsave->save.cs;
	svm->vmcb->save.ss = hsave->save.ss;
	svm->vmcb->save.ds = hsave->save.ds;
	svm->vmcb->save.gdtr = hsave->save.gdtr;
	svm->vmcb->save.idtr = hsave->save.idtr;
2361
	kvm_set_rflags(&svm->vcpu, hsave->save.rflags);
2362 2363 2364 2365 2366 2367 2368
	svm_set_efer(&svm->vcpu, hsave->save.efer);
	svm_set_cr0(&svm->vcpu, hsave->save.cr0 | X86_CR0_PE);
	svm_set_cr4(&svm->vcpu, hsave->save.cr4);
	if (npt_enabled) {
		svm->vmcb->save.cr3 = hsave->save.cr3;
		svm->vcpu.arch.cr3 = hsave->save.cr3;
	} else {
2369
		(void)kvm_set_cr3(&svm->vcpu, hsave->save.cr3);
2370 2371 2372 2373 2374 2375 2376 2377
	}
	kvm_register_write(&svm->vcpu, VCPU_REGS_RAX, hsave->save.rax);
	kvm_register_write(&svm->vcpu, VCPU_REGS_RSP, hsave->save.rsp);
	kvm_register_write(&svm->vcpu, VCPU_REGS_RIP, hsave->save.rip);
	svm->vmcb->save.dr7 = 0;
	svm->vmcb->save.cpl = 0;
	svm->vmcb->control.exit_int_info = 0;

2378 2379
	mark_all_dirty(svm->vmcb);

2380
	nested_svm_unmap(page);
2381

2382
	nested_svm_uninit_mmu_context(&svm->vcpu);
2383 2384 2385 2386 2387
	kvm_mmu_reset_context(&svm->vcpu);
	kvm_mmu_load(&svm->vcpu);

	return 0;
}
A
Alexander Graf 已提交
2388

2389
static bool nested_svm_vmrun_msrpm(struct vcpu_svm *svm)
A
Alexander Graf 已提交
2390
{
2391 2392
	/*
	 * This function merges the msr permission bitmaps of kvm and the
G
Guo Chao 已提交
2393
	 * nested vmcb. It is optimized in that it only merges the parts where
2394 2395
	 * the kvm msr permission bitmap may contain zero bits
	 */
A
Alexander Graf 已提交
2396
	int i;
2397

2398 2399
	if (!(svm->nested.intercept & (1ULL << INTERCEPT_MSR_PROT)))
		return true;
2400

2401 2402 2403
	for (i = 0; i < MSRPM_OFFSETS; i++) {
		u32 value, p;
		u64 offset;
2404

2405 2406
		if (msrpm_offsets[i] == 0xffffffff)
			break;
A
Alexander Graf 已提交
2407

2408 2409
		p      = msrpm_offsets[i];
		offset = svm->nested.vmcb_msrpm + (p * 4);
2410 2411 2412 2413 2414 2415

		if (kvm_read_guest(svm->vcpu.kvm, offset, &value, 4))
			return false;

		svm->nested.msrpm[p] = svm->msrpm[p] | value;
	}
A
Alexander Graf 已提交
2416

2417
	svm->vmcb->control.msrpm_base_pa = __pa(svm->nested.msrpm);
2418 2419

	return true;
A
Alexander Graf 已提交
2420 2421
}

2422 2423 2424 2425 2426
static bool nested_vmcb_checks(struct vmcb *vmcb)
{
	if ((vmcb->control.intercept & (1ULL << INTERCEPT_VMRUN)) == 0)
		return false;

2427 2428 2429
	if (vmcb->control.asid == 0)
		return false;

2430 2431 2432
	if (vmcb->control.nested_ctl && !npt_enabled)
		return false;

2433 2434 2435
	return true;
}

2436
static bool nested_svm_vmrun(struct vcpu_svm *svm)
A
Alexander Graf 已提交
2437
{
2438
	struct vmcb *nested_vmcb;
2439
	struct vmcb *hsave = svm->nested.hsave;
J
Joerg Roedel 已提交
2440
	struct vmcb *vmcb = svm->vmcb;
2441
	struct page *page;
2442
	u64 vmcb_gpa;
A
Alexander Graf 已提交
2443

2444
	vmcb_gpa = svm->vmcb->save.rax;
A
Alexander Graf 已提交
2445

2446
	nested_vmcb = nested_svm_map(svm, svm->vmcb->save.rax, &page);
2447 2448 2449
	if (!nested_vmcb)
		return false;

2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460
	if (!nested_vmcb_checks(nested_vmcb)) {
		nested_vmcb->control.exit_code    = SVM_EXIT_ERR;
		nested_vmcb->control.exit_code_hi = 0;
		nested_vmcb->control.exit_info_1  = 0;
		nested_vmcb->control.exit_info_2  = 0;

		nested_svm_unmap(page);

		return false;
	}

2461
	trace_kvm_nested_vmrun(svm->vmcb->save.rip, vmcb_gpa,
2462 2463 2464 2465 2466
			       nested_vmcb->save.rip,
			       nested_vmcb->control.int_ctl,
			       nested_vmcb->control.event_inj,
			       nested_vmcb->control.nested_ctl);

2467 2468
	trace_kvm_nested_intercepts(nested_vmcb->control.intercept_cr & 0xffff,
				    nested_vmcb->control.intercept_cr >> 16,
2469 2470 2471
				    nested_vmcb->control.intercept_exceptions,
				    nested_vmcb->control.intercept);

A
Alexander Graf 已提交
2472
	/* Clear internal status */
2473 2474
	kvm_clear_exception_queue(&svm->vcpu);
	kvm_clear_interrupt_queue(&svm->vcpu);
A
Alexander Graf 已提交
2475

J
Joerg Roedel 已提交
2476 2477 2478 2479
	/*
	 * Save the old vmcb, so we don't need to pick what we save, but can
	 * restore everything when a VMEXIT occurs
	 */
J
Joerg Roedel 已提交
2480 2481 2482 2483 2484 2485
	hsave->save.es     = vmcb->save.es;
	hsave->save.cs     = vmcb->save.cs;
	hsave->save.ss     = vmcb->save.ss;
	hsave->save.ds     = vmcb->save.ds;
	hsave->save.gdtr   = vmcb->save.gdtr;
	hsave->save.idtr   = vmcb->save.idtr;
2486
	hsave->save.efer   = svm->vcpu.arch.efer;
2487
	hsave->save.cr0    = kvm_read_cr0(&svm->vcpu);
J
Joerg Roedel 已提交
2488
	hsave->save.cr4    = svm->vcpu.arch.cr4;
2489
	hsave->save.rflags = kvm_get_rflags(&svm->vcpu);
2490
	hsave->save.rip    = kvm_rip_read(&svm->vcpu);
J
Joerg Roedel 已提交
2491 2492 2493 2494 2495
	hsave->save.rsp    = vmcb->save.rsp;
	hsave->save.rax    = vmcb->save.rax;
	if (npt_enabled)
		hsave->save.cr3    = vmcb->save.cr3;
	else
2496
		hsave->save.cr3    = kvm_read_cr3(&svm->vcpu);
J
Joerg Roedel 已提交
2497

2498
	copy_vmcb_control_area(hsave, vmcb);
A
Alexander Graf 已提交
2499

2500
	if (kvm_get_rflags(&svm->vcpu) & X86_EFLAGS_IF)
A
Alexander Graf 已提交
2501 2502 2503 2504
		svm->vcpu.arch.hflags |= HF_HIF_MASK;
	else
		svm->vcpu.arch.hflags &= ~HF_HIF_MASK;

2505 2506 2507 2508 2509 2510
	if (nested_vmcb->control.nested_ctl) {
		kvm_mmu_unload(&svm->vcpu);
		svm->nested.nested_cr3 = nested_vmcb->control.nested_cr3;
		nested_svm_init_mmu_context(&svm->vcpu);
	}

A
Alexander Graf 已提交
2511 2512 2513 2514 2515 2516 2517
	/* Load the nested guest state */
	svm->vmcb->save.es = nested_vmcb->save.es;
	svm->vmcb->save.cs = nested_vmcb->save.cs;
	svm->vmcb->save.ss = nested_vmcb->save.ss;
	svm->vmcb->save.ds = nested_vmcb->save.ds;
	svm->vmcb->save.gdtr = nested_vmcb->save.gdtr;
	svm->vmcb->save.idtr = nested_vmcb->save.idtr;
2518
	kvm_set_rflags(&svm->vcpu, nested_vmcb->save.rflags);
A
Alexander Graf 已提交
2519 2520 2521 2522 2523 2524
	svm_set_efer(&svm->vcpu, nested_vmcb->save.efer);
	svm_set_cr0(&svm->vcpu, nested_vmcb->save.cr0);
	svm_set_cr4(&svm->vcpu, nested_vmcb->save.cr4);
	if (npt_enabled) {
		svm->vmcb->save.cr3 = nested_vmcb->save.cr3;
		svm->vcpu.arch.cr3 = nested_vmcb->save.cr3;
2525
	} else
2526
		(void)kvm_set_cr3(&svm->vcpu, nested_vmcb->save.cr3);
2527 2528 2529 2530

	/* Guest paging mode is active - reset mmu */
	kvm_mmu_reset_context(&svm->vcpu);

J
Joerg Roedel 已提交
2531
	svm->vmcb->save.cr2 = svm->vcpu.arch.cr2 = nested_vmcb->save.cr2;
A
Alexander Graf 已提交
2532 2533 2534
	kvm_register_write(&svm->vcpu, VCPU_REGS_RAX, nested_vmcb->save.rax);
	kvm_register_write(&svm->vcpu, VCPU_REGS_RSP, nested_vmcb->save.rsp);
	kvm_register_write(&svm->vcpu, VCPU_REGS_RIP, nested_vmcb->save.rip);
J
Joerg Roedel 已提交
2535

A
Alexander Graf 已提交
2536 2537 2538 2539 2540 2541 2542 2543
	/* In case we don't even reach vcpu_run, the fields are not updated */
	svm->vmcb->save.rax = nested_vmcb->save.rax;
	svm->vmcb->save.rsp = nested_vmcb->save.rsp;
	svm->vmcb->save.rip = nested_vmcb->save.rip;
	svm->vmcb->save.dr7 = nested_vmcb->save.dr7;
	svm->vmcb->save.dr6 = nested_vmcb->save.dr6;
	svm->vmcb->save.cpl = nested_vmcb->save.cpl;

2544
	svm->nested.vmcb_msrpm = nested_vmcb->control.msrpm_base_pa & ~0x0fffULL;
2545
	svm->nested.vmcb_iopm  = nested_vmcb->control.iopm_base_pa  & ~0x0fffULL;
A
Alexander Graf 已提交
2546

J
Joerg Roedel 已提交
2547
	/* cache intercepts */
2548
	svm->nested.intercept_cr         = nested_vmcb->control.intercept_cr;
2549
	svm->nested.intercept_dr         = nested_vmcb->control.intercept_dr;
J
Joerg Roedel 已提交
2550 2551 2552
	svm->nested.intercept_exceptions = nested_vmcb->control.intercept_exceptions;
	svm->nested.intercept            = nested_vmcb->control.intercept;

2553
	svm_flush_tlb(&svm->vcpu);
A
Alexander Graf 已提交
2554 2555 2556 2557 2558 2559
	svm->vmcb->control.int_ctl = nested_vmcb->control.int_ctl | V_INTR_MASKING_MASK;
	if (nested_vmcb->control.int_ctl & V_INTR_MASKING_MASK)
		svm->vcpu.arch.hflags |= HF_VINTR_MASK;
	else
		svm->vcpu.arch.hflags &= ~HF_VINTR_MASK;

2560 2561
	if (svm->vcpu.arch.hflags & HF_VINTR_MASK) {
		/* We only want the cr8 intercept bits of the guest */
2562 2563
		clr_cr_intercept(svm, INTERCEPT_CR8_READ);
		clr_cr_intercept(svm, INTERCEPT_CR8_WRITE);
2564 2565
	}

2566
	/* We don't want to see VMMCALLs from a nested guest */
2567
	clr_intercept(svm, INTERCEPT_VMMCALL);
2568

2569
	svm->vmcb->control.lbr_ctl = nested_vmcb->control.lbr_ctl;
A
Alexander Graf 已提交
2570 2571 2572 2573 2574 2575
	svm->vmcb->control.int_vector = nested_vmcb->control.int_vector;
	svm->vmcb->control.int_state = nested_vmcb->control.int_state;
	svm->vmcb->control.tsc_offset += nested_vmcb->control.tsc_offset;
	svm->vmcb->control.event_inj = nested_vmcb->control.event_inj;
	svm->vmcb->control.event_inj_err = nested_vmcb->control.event_inj_err;

2576
	nested_svm_unmap(page);
2577

2578 2579 2580
	/* Enter Guest-Mode */
	enter_guest_mode(&svm->vcpu);

2581 2582 2583 2584 2585 2586
	/*
	 * Merge guest and host intercepts - must be called  with vcpu in
	 * guest-mode to take affect here
	 */
	recalc_intercepts(svm);

2587
	svm->nested.vmcb = vmcb_gpa;
2588

2589
	enable_gif(svm);
A
Alexander Graf 已提交
2590

2591 2592
	mark_all_dirty(svm->vmcb);

2593
	return true;
A
Alexander Graf 已提交
2594 2595
}

2596
static void nested_svm_vmloadsave(struct vmcb *from_vmcb, struct vmcb *to_vmcb)
2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611
{
	to_vmcb->save.fs = from_vmcb->save.fs;
	to_vmcb->save.gs = from_vmcb->save.gs;
	to_vmcb->save.tr = from_vmcb->save.tr;
	to_vmcb->save.ldtr = from_vmcb->save.ldtr;
	to_vmcb->save.kernel_gs_base = from_vmcb->save.kernel_gs_base;
	to_vmcb->save.star = from_vmcb->save.star;
	to_vmcb->save.lstar = from_vmcb->save.lstar;
	to_vmcb->save.cstar = from_vmcb->save.cstar;
	to_vmcb->save.sfmask = from_vmcb->save.sfmask;
	to_vmcb->save.sysenter_cs = from_vmcb->save.sysenter_cs;
	to_vmcb->save.sysenter_esp = from_vmcb->save.sysenter_esp;
	to_vmcb->save.sysenter_eip = from_vmcb->save.sysenter_eip;
}

A
Avi Kivity 已提交
2612
static int vmload_interception(struct vcpu_svm *svm)
2613
{
2614
	struct vmcb *nested_vmcb;
2615
	struct page *page;
2616

2617 2618 2619
	if (nested_svm_check_permissions(svm))
		return 1;

2620
	nested_vmcb = nested_svm_map(svm, svm->vmcb->save.rax, &page);
2621 2622 2623
	if (!nested_vmcb)
		return 1;

2624 2625 2626
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
	skip_emulated_instruction(&svm->vcpu);

2627
	nested_svm_vmloadsave(nested_vmcb, svm->vmcb);
2628
	nested_svm_unmap(page);
2629 2630 2631 2632

	return 1;
}

A
Avi Kivity 已提交
2633
static int vmsave_interception(struct vcpu_svm *svm)
2634
{
2635
	struct vmcb *nested_vmcb;
2636
	struct page *page;
2637

2638 2639 2640
	if (nested_svm_check_permissions(svm))
		return 1;

2641
	nested_vmcb = nested_svm_map(svm, svm->vmcb->save.rax, &page);
2642 2643 2644
	if (!nested_vmcb)
		return 1;

2645 2646 2647
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
	skip_emulated_instruction(&svm->vcpu);

2648
	nested_svm_vmloadsave(svm->vmcb, nested_vmcb);
2649
	nested_svm_unmap(page);
2650 2651 2652 2653

	return 1;
}

A
Avi Kivity 已提交
2654
static int vmrun_interception(struct vcpu_svm *svm)
A
Alexander Graf 已提交
2655 2656 2657 2658
{
	if (nested_svm_check_permissions(svm))
		return 1;

2659 2660
	/* Save rip after vmrun instruction */
	kvm_rip_write(&svm->vcpu, kvm_rip_read(&svm->vcpu) + 3);
A
Alexander Graf 已提交
2661

2662
	if (!nested_svm_vmrun(svm))
A
Alexander Graf 已提交
2663 2664
		return 1;

2665
	if (!nested_svm_vmrun_msrpm(svm))
2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677
		goto failed;

	return 1;

failed:

	svm->vmcb->control.exit_code    = SVM_EXIT_ERR;
	svm->vmcb->control.exit_code_hi = 0;
	svm->vmcb->control.exit_info_1  = 0;
	svm->vmcb->control.exit_info_2  = 0;

	nested_svm_vmexit(svm);
A
Alexander Graf 已提交
2678 2679 2680 2681

	return 1;
}

A
Avi Kivity 已提交
2682
static int stgi_interception(struct vcpu_svm *svm)
2683 2684 2685 2686 2687 2688
{
	if (nested_svm_check_permissions(svm))
		return 1;

	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
	skip_emulated_instruction(&svm->vcpu);
2689
	kvm_make_request(KVM_REQ_EVENT, &svm->vcpu);
2690

2691
	enable_gif(svm);
2692 2693 2694 2695

	return 1;
}

A
Avi Kivity 已提交
2696
static int clgi_interception(struct vcpu_svm *svm)
2697 2698 2699 2700 2701 2702 2703
{
	if (nested_svm_check_permissions(svm))
		return 1;

	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
	skip_emulated_instruction(&svm->vcpu);

2704
	disable_gif(svm);
2705 2706 2707 2708 2709

	/* After a CLGI no interrupts should come */
	svm_clear_vintr(svm);
	svm->vmcb->control.int_ctl &= ~V_IRQ_MASK;

2710 2711
	mark_dirty(svm->vmcb, VMCB_INTR);

2712 2713 2714
	return 1;
}

A
Avi Kivity 已提交
2715
static int invlpga_interception(struct vcpu_svm *svm)
A
Alexander Graf 已提交
2716 2717 2718
{
	struct kvm_vcpu *vcpu = &svm->vcpu;

2719 2720 2721
	trace_kvm_invlpga(svm->vmcb->save.rip, vcpu->arch.regs[VCPU_REGS_RCX],
			  vcpu->arch.regs[VCPU_REGS_RAX]);

A
Alexander Graf 已提交
2722 2723 2724 2725 2726 2727 2728 2729
	/* Let's treat INVLPGA the same as INVLPG (can be optimized!) */
	kvm_mmu_invlpg(vcpu, vcpu->arch.regs[VCPU_REGS_RAX]);

	svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
	skip_emulated_instruction(&svm->vcpu);
	return 1;
}

2730 2731 2732 2733 2734 2735 2736 2737
static int skinit_interception(struct vcpu_svm *svm)
{
	trace_kvm_skinit(svm->vmcb->save.rip, svm->vcpu.arch.regs[VCPU_REGS_RAX]);

	kvm_queue_exception(&svm->vcpu, UD_VECTOR);
	return 1;
}

J
Joerg Roedel 已提交
2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750
static int xsetbv_interception(struct vcpu_svm *svm)
{
	u64 new_bv = kvm_read_edx_eax(&svm->vcpu);
	u32 index = kvm_register_read(&svm->vcpu, VCPU_REGS_RCX);

	if (kvm_set_xcr(&svm->vcpu, index, new_bv) == 0) {
		svm->next_rip = kvm_rip_read(&svm->vcpu) + 3;
		skip_emulated_instruction(&svm->vcpu);
	}

	return 1;
}

A
Avi Kivity 已提交
2751
static int invalid_op_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
2752
{
2753
	kvm_queue_exception(&svm->vcpu, UD_VECTOR);
A
Avi Kivity 已提交
2754 2755 2756
	return 1;
}

A
Avi Kivity 已提交
2757
static int task_switch_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
2758
{
2759
	u16 tss_selector;
2760 2761 2762
	int reason;
	int int_type = svm->vmcb->control.exit_int_info &
		SVM_EXITINTINFO_TYPE_MASK;
2763
	int int_vec = svm->vmcb->control.exit_int_info & SVM_EVTINJ_VEC_MASK;
2764 2765 2766 2767
	uint32_t type =
		svm->vmcb->control.exit_int_info & SVM_EXITINTINFO_TYPE_MASK;
	uint32_t idt_v =
		svm->vmcb->control.exit_int_info & SVM_EXITINTINFO_VALID;
2768 2769
	bool has_error_code = false;
	u32 error_code = 0;
2770 2771

	tss_selector = (u16)svm->vmcb->control.exit_info_1;
2772

2773 2774
	if (svm->vmcb->control.exit_info_2 &
	    (1ULL << SVM_EXITINFOSHIFT_TS_REASON_IRET))
2775 2776 2777 2778
		reason = TASK_SWITCH_IRET;
	else if (svm->vmcb->control.exit_info_2 &
		 (1ULL << SVM_EXITINFOSHIFT_TS_REASON_JMP))
		reason = TASK_SWITCH_JMP;
2779
	else if (idt_v)
2780 2781 2782 2783
		reason = TASK_SWITCH_GATE;
	else
		reason = TASK_SWITCH_CALL;

2784 2785 2786 2787 2788 2789
	if (reason == TASK_SWITCH_GATE) {
		switch (type) {
		case SVM_EXITINTINFO_TYPE_NMI:
			svm->vcpu.arch.nmi_injected = false;
			break;
		case SVM_EXITINTINFO_TYPE_EXEPT:
2790 2791 2792 2793 2794 2795
			if (svm->vmcb->control.exit_info_2 &
			    (1ULL << SVM_EXITINFOSHIFT_TS_HAS_ERROR_CODE)) {
				has_error_code = true;
				error_code =
					(u32)svm->vmcb->control.exit_info_2;
			}
2796 2797 2798 2799 2800 2801 2802 2803 2804
			kvm_clear_exception_queue(&svm->vcpu);
			break;
		case SVM_EXITINTINFO_TYPE_INTR:
			kvm_clear_interrupt_queue(&svm->vcpu);
			break;
		default:
			break;
		}
	}
2805

2806 2807 2808
	if (reason != TASK_SWITCH_GATE ||
	    int_type == SVM_EXITINTINFO_TYPE_SOFT ||
	    (int_type == SVM_EXITINTINFO_TYPE_EXEPT &&
2809 2810
	     (int_vec == OF_VECTOR || int_vec == BP_VECTOR)))
		skip_emulated_instruction(&svm->vcpu);
2811

2812 2813 2814 2815
	if (int_type != SVM_EXITINTINFO_TYPE_SOFT)
		int_vec = -1;

	if (kvm_task_switch(&svm->vcpu, tss_selector, int_vec, reason,
2816 2817 2818 2819 2820 2821 2822
				has_error_code, error_code) == EMULATE_FAIL) {
		svm->vcpu.run->exit_reason = KVM_EXIT_INTERNAL_ERROR;
		svm->vcpu.run->internal.suberror = KVM_INTERNAL_ERROR_EMULATION;
		svm->vcpu.run->internal.ndata = 0;
		return 0;
	}
	return 1;
A
Avi Kivity 已提交
2823 2824
}

A
Avi Kivity 已提交
2825
static int cpuid_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
2826
{
2827
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 2;
R
Rusty Russell 已提交
2828
	kvm_emulate_cpuid(&svm->vcpu);
2829
	return 1;
A
Avi Kivity 已提交
2830 2831
}

A
Avi Kivity 已提交
2832
static int iret_interception(struct vcpu_svm *svm)
2833 2834
{
	++svm->vcpu.stat.nmi_window_exits;
2835
	clr_intercept(svm, INTERCEPT_IRET);
2836
	svm->vcpu.arch.hflags |= HF_IRET_MASK;
2837
	svm->nmi_iret_rip = kvm_rip_read(&svm->vcpu);
2838 2839 2840
	return 1;
}

A
Avi Kivity 已提交
2841
static int invlpg_interception(struct vcpu_svm *svm)
M
Marcelo Tosatti 已提交
2842
{
2843 2844 2845 2846 2847 2848
	if (!static_cpu_has(X86_FEATURE_DECODEASSISTS))
		return emulate_instruction(&svm->vcpu, 0) == EMULATE_DONE;

	kvm_mmu_invlpg(&svm->vcpu, svm->vmcb->control.exit_info_1);
	skip_emulated_instruction(&svm->vcpu);
	return 1;
M
Marcelo Tosatti 已提交
2849 2850
}

A
Avi Kivity 已提交
2851
static int emulate_on_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
2852
{
2853
	return emulate_instruction(&svm->vcpu, 0) == EMULATE_DONE;
A
Avi Kivity 已提交
2854 2855
}

A
Avi Kivity 已提交
2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868
static int rdpmc_interception(struct vcpu_svm *svm)
{
	int err;

	if (!static_cpu_has(X86_FEATURE_NRIPS))
		return emulate_on_interception(svm);

	err = kvm_rdpmc(&svm->vcpu);
	kvm_complete_insn_gp(&svm->vcpu, err);

	return 1;
}

2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891
bool check_selective_cr0_intercepted(struct vcpu_svm *svm, unsigned long val)
{
	unsigned long cr0 = svm->vcpu.arch.cr0;
	bool ret = false;
	u64 intercept;

	intercept = svm->nested.intercept;

	if (!is_guest_mode(&svm->vcpu) ||
	    (!(intercept & (1ULL << INTERCEPT_SELECTIVE_CR0))))
		return false;

	cr0 &= ~SVM_CR0_SELECTIVE_MASK;
	val &= ~SVM_CR0_SELECTIVE_MASK;

	if (cr0 ^ val) {
		svm->vmcb->control.exit_code = SVM_EXIT_CR0_SEL_WRITE;
		ret = (nested_svm_exit_handled(svm) == NESTED_EXIT_DONE);
	}

	return ret;
}

2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914
#define CR_VALID (1ULL << 63)

static int cr_interception(struct vcpu_svm *svm)
{
	int reg, cr;
	unsigned long val;
	int err;

	if (!static_cpu_has(X86_FEATURE_DECODEASSISTS))
		return emulate_on_interception(svm);

	if (unlikely((svm->vmcb->control.exit_info_1 & CR_VALID) == 0))
		return emulate_on_interception(svm);

	reg = svm->vmcb->control.exit_info_1 & SVM_EXITINFO_REG_MASK;
	cr = svm->vmcb->control.exit_code - SVM_EXIT_READ_CR0;

	err = 0;
	if (cr >= 16) { /* mov to cr */
		cr -= 16;
		val = kvm_register_read(&svm->vcpu, reg);
		switch (cr) {
		case 0:
2915 2916
			if (!check_selective_cr0_intercepted(svm, val))
				err = kvm_set_cr0(&svm->vcpu, val);
2917 2918 2919
			else
				return 1;

2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943
			break;
		case 3:
			err = kvm_set_cr3(&svm->vcpu, val);
			break;
		case 4:
			err = kvm_set_cr4(&svm->vcpu, val);
			break;
		case 8:
			err = kvm_set_cr8(&svm->vcpu, val);
			break;
		default:
			WARN(1, "unhandled write to CR%d", cr);
			kvm_queue_exception(&svm->vcpu, UD_VECTOR);
			return 1;
		}
	} else { /* mov from cr */
		switch (cr) {
		case 0:
			val = kvm_read_cr0(&svm->vcpu);
			break;
		case 2:
			val = svm->vcpu.arch.cr2;
			break;
		case 3:
2944
			val = kvm_read_cr3(&svm->vcpu);
2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963
			break;
		case 4:
			val = kvm_read_cr4(&svm->vcpu);
			break;
		case 8:
			val = kvm_get_cr8(&svm->vcpu);
			break;
		default:
			WARN(1, "unhandled read from CR%d", cr);
			kvm_queue_exception(&svm->vcpu, UD_VECTOR);
			return 1;
		}
		kvm_register_write(&svm->vcpu, reg, val);
	}
	kvm_complete_insn_gp(&svm->vcpu, err);

	return 1;
}

2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984
static int dr_interception(struct vcpu_svm *svm)
{
	int reg, dr;
	unsigned long val;
	int err;

	if (!boot_cpu_has(X86_FEATURE_DECODEASSISTS))
		return emulate_on_interception(svm);

	reg = svm->vmcb->control.exit_info_1 & SVM_EXITINFO_REG_MASK;
	dr = svm->vmcb->control.exit_code - SVM_EXIT_READ_DR0;

	if (dr >= 16) { /* mov to DRn */
		val = kvm_register_read(&svm->vcpu, reg);
		kvm_set_dr(&svm->vcpu, dr - 16, val);
	} else {
		err = kvm_get_dr(&svm->vcpu, dr, &val);
		if (!err)
			kvm_register_write(&svm->vcpu, reg, val);
	}

2985 2986
	skip_emulated_instruction(&svm->vcpu);

2987 2988 2989
	return 1;
}

A
Avi Kivity 已提交
2990
static int cr8_write_interception(struct vcpu_svm *svm)
2991
{
A
Avi Kivity 已提交
2992
	struct kvm_run *kvm_run = svm->vcpu.run;
A
Andre Przywara 已提交
2993
	int r;
A
Avi Kivity 已提交
2994

2995 2996
	u8 cr8_prev = kvm_get_cr8(&svm->vcpu);
	/* instruction emulation calls kvm_set_cr8() */
2997
	r = cr_interception(svm);
2998
	if (irqchip_in_kernel(svm->vcpu.kvm)) {
2999
		clr_cr_intercept(svm, INTERCEPT_CR8_WRITE);
3000
		return r;
3001
	}
3002
	if (cr8_prev <= kvm_get_cr8(&svm->vcpu))
3003
		return r;
3004 3005 3006 3007
	kvm_run->exit_reason = KVM_EXIT_SET_TPR;
	return 0;
}

N
Nadav Har'El 已提交
3008 3009 3010 3011 3012 3013 3014
u64 svm_read_l1_tsc(struct kvm_vcpu *vcpu)
{
	struct vmcb *vmcb = get_host_vmcb(to_svm(vcpu));
	return vmcb->control.tsc_offset +
		svm_scale_tsc(vcpu, native_read_tsc());
}

A
Avi Kivity 已提交
3015 3016
static int svm_get_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 *data)
{
3017 3018
	struct vcpu_svm *svm = to_svm(vcpu);

A
Avi Kivity 已提交
3019
	switch (ecx) {
3020
	case MSR_IA32_TSC: {
3021
		*data = svm->vmcb->control.tsc_offset +
3022 3023
			svm_scale_tsc(vcpu, native_read_tsc());

A
Avi Kivity 已提交
3024 3025
		break;
	}
B
Brian Gerst 已提交
3026
	case MSR_STAR:
3027
		*data = svm->vmcb->save.star;
A
Avi Kivity 已提交
3028
		break;
3029
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
3030
	case MSR_LSTAR:
3031
		*data = svm->vmcb->save.lstar;
A
Avi Kivity 已提交
3032 3033
		break;
	case MSR_CSTAR:
3034
		*data = svm->vmcb->save.cstar;
A
Avi Kivity 已提交
3035 3036
		break;
	case MSR_KERNEL_GS_BASE:
3037
		*data = svm->vmcb->save.kernel_gs_base;
A
Avi Kivity 已提交
3038 3039
		break;
	case MSR_SYSCALL_MASK:
3040
		*data = svm->vmcb->save.sfmask;
A
Avi Kivity 已提交
3041 3042 3043
		break;
#endif
	case MSR_IA32_SYSENTER_CS:
3044
		*data = svm->vmcb->save.sysenter_cs;
A
Avi Kivity 已提交
3045 3046
		break;
	case MSR_IA32_SYSENTER_EIP:
3047
		*data = svm->sysenter_eip;
A
Avi Kivity 已提交
3048 3049
		break;
	case MSR_IA32_SYSENTER_ESP:
3050
		*data = svm->sysenter_esp;
A
Avi Kivity 已提交
3051
		break;
J
Joerg Roedel 已提交
3052 3053 3054 3055 3056
	/*
	 * Nobody will change the following 5 values in the VMCB so we can
	 * safely return them on rdmsr. They will always be 0 until LBRV is
	 * implemented.
	 */
3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071
	case MSR_IA32_DEBUGCTLMSR:
		*data = svm->vmcb->save.dbgctl;
		break;
	case MSR_IA32_LASTBRANCHFROMIP:
		*data = svm->vmcb->save.br_from;
		break;
	case MSR_IA32_LASTBRANCHTOIP:
		*data = svm->vmcb->save.br_to;
		break;
	case MSR_IA32_LASTINTFROMIP:
		*data = svm->vmcb->save.last_excp_from;
		break;
	case MSR_IA32_LASTINTTOIP:
		*data = svm->vmcb->save.last_excp_to;
		break;
A
Alexander Graf 已提交
3072
	case MSR_VM_HSAVE_PA:
3073
		*data = svm->nested.hsave_msr;
A
Alexander Graf 已提交
3074
		break;
3075
	case MSR_VM_CR:
3076
		*data = svm->nested.vm_cr_msr;
3077
		break;
3078 3079 3080
	case MSR_IA32_UCODE_REV:
		*data = 0x01000065;
		break;
A
Avi Kivity 已提交
3081
	default:
3082
		return kvm_get_msr_common(vcpu, ecx, data);
A
Avi Kivity 已提交
3083 3084 3085 3086
	}
	return 0;
}

A
Avi Kivity 已提交
3087
static int rdmsr_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
3088
{
3089
	u32 ecx = svm->vcpu.arch.regs[VCPU_REGS_RCX];
A
Avi Kivity 已提交
3090 3091
	u64 data;

3092 3093
	if (svm_get_msr(&svm->vcpu, ecx, &data)) {
		trace_kvm_msr_read_ex(ecx);
3094
		kvm_inject_gp(&svm->vcpu, 0);
3095
	} else {
3096
		trace_kvm_msr_read(ecx, data);
3097

3098
		svm->vcpu.arch.regs[VCPU_REGS_RAX] = data & 0xffffffff;
3099
		svm->vcpu.arch.regs[VCPU_REGS_RDX] = data >> 32;
3100
		svm->next_rip = kvm_rip_read(&svm->vcpu) + 2;
R
Rusty Russell 已提交
3101
		skip_emulated_instruction(&svm->vcpu);
A
Avi Kivity 已提交
3102 3103 3104 3105
	}
	return 1;
}

3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130
static int svm_set_vm_cr(struct kvm_vcpu *vcpu, u64 data)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	int svm_dis, chg_mask;

	if (data & ~SVM_VM_CR_VALID_MASK)
		return 1;

	chg_mask = SVM_VM_CR_VALID_MASK;

	if (svm->nested.vm_cr_msr & SVM_VM_CR_SVM_DIS_MASK)
		chg_mask &= ~(SVM_VM_CR_SVM_LOCK_MASK | SVM_VM_CR_SVM_DIS_MASK);

	svm->nested.vm_cr_msr &= ~chg_mask;
	svm->nested.vm_cr_msr |= (data & chg_mask);

	svm_dis = svm->nested.vm_cr_msr & SVM_VM_CR_SVM_DIS_MASK;

	/* check for svm_disable while efer.svme is set */
	if (svm_dis && (vcpu->arch.efer & EFER_SVME))
		return 1;

	return 0;
}

A
Avi Kivity 已提交
3131 3132
static int svm_set_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 data)
{
3133 3134
	struct vcpu_svm *svm = to_svm(vcpu);

A
Avi Kivity 已提交
3135
	switch (ecx) {
3136
	case MSR_IA32_TSC:
3137
		kvm_write_tsc(vcpu, data);
A
Avi Kivity 已提交
3138
		break;
B
Brian Gerst 已提交
3139
	case MSR_STAR:
3140
		svm->vmcb->save.star = data;
A
Avi Kivity 已提交
3141
		break;
3142
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
3143
	case MSR_LSTAR:
3144
		svm->vmcb->save.lstar = data;
A
Avi Kivity 已提交
3145 3146
		break;
	case MSR_CSTAR:
3147
		svm->vmcb->save.cstar = data;
A
Avi Kivity 已提交
3148 3149
		break;
	case MSR_KERNEL_GS_BASE:
3150
		svm->vmcb->save.kernel_gs_base = data;
A
Avi Kivity 已提交
3151 3152
		break;
	case MSR_SYSCALL_MASK:
3153
		svm->vmcb->save.sfmask = data;
A
Avi Kivity 已提交
3154 3155 3156
		break;
#endif
	case MSR_IA32_SYSENTER_CS:
3157
		svm->vmcb->save.sysenter_cs = data;
A
Avi Kivity 已提交
3158 3159
		break;
	case MSR_IA32_SYSENTER_EIP:
3160
		svm->sysenter_eip = data;
3161
		svm->vmcb->save.sysenter_eip = data;
A
Avi Kivity 已提交
3162 3163
		break;
	case MSR_IA32_SYSENTER_ESP:
3164
		svm->sysenter_esp = data;
3165
		svm->vmcb->save.sysenter_esp = data;
A
Avi Kivity 已提交
3166
		break;
3167
	case MSR_IA32_DEBUGCTLMSR:
3168
		if (!boot_cpu_has(X86_FEATURE_LBRV)) {
3169 3170
			vcpu_unimpl(vcpu, "%s: MSR_IA32_DEBUGCTL 0x%llx, nop\n",
				    __func__, data);
3171 3172 3173 3174 3175 3176
			break;
		}
		if (data & DEBUGCTL_RESERVED_BITS)
			return 1;

		svm->vmcb->save.dbgctl = data;
3177
		mark_dirty(svm->vmcb, VMCB_LBR);
3178 3179 3180 3181
		if (data & (1ULL<<0))
			svm_enable_lbrv(svm);
		else
			svm_disable_lbrv(svm);
3182
		break;
A
Alexander Graf 已提交
3183
	case MSR_VM_HSAVE_PA:
3184
		svm->nested.hsave_msr = data;
3185
		break;
3186
	case MSR_VM_CR:
3187
		return svm_set_vm_cr(vcpu, data);
3188
	case MSR_VM_IGNNE:
3189
		vcpu_unimpl(vcpu, "unimplemented wrmsr: 0x%x data 0x%llx\n", ecx, data);
3190
		break;
A
Avi Kivity 已提交
3191
	default:
3192
		return kvm_set_msr_common(vcpu, ecx, data);
A
Avi Kivity 已提交
3193 3194 3195 3196
	}
	return 0;
}

A
Avi Kivity 已提交
3197
static int wrmsr_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
3198
{
3199
	u32 ecx = svm->vcpu.arch.regs[VCPU_REGS_RCX];
3200
	u64 data = (svm->vcpu.arch.regs[VCPU_REGS_RAX] & -1u)
3201
		| ((u64)(svm->vcpu.arch.regs[VCPU_REGS_RDX] & -1u) << 32);
3202 3203


3204
	svm->next_rip = kvm_rip_read(&svm->vcpu) + 2;
3205 3206
	if (svm_set_msr(&svm->vcpu, ecx, data)) {
		trace_kvm_msr_write_ex(ecx, data);
3207
		kvm_inject_gp(&svm->vcpu, 0);
3208 3209
	} else {
		trace_kvm_msr_write(ecx, data);
R
Rusty Russell 已提交
3210
		skip_emulated_instruction(&svm->vcpu);
3211
	}
A
Avi Kivity 已提交
3212 3213 3214
	return 1;
}

A
Avi Kivity 已提交
3215
static int msr_interception(struct vcpu_svm *svm)
A
Avi Kivity 已提交
3216
{
R
Rusty Russell 已提交
3217
	if (svm->vmcb->control.exit_info_1)
A
Avi Kivity 已提交
3218
		return wrmsr_interception(svm);
A
Avi Kivity 已提交
3219
	else
A
Avi Kivity 已提交
3220
		return rdmsr_interception(svm);
A
Avi Kivity 已提交
3221 3222
}

A
Avi Kivity 已提交
3223
static int interrupt_window_interception(struct vcpu_svm *svm)
3224
{
A
Avi Kivity 已提交
3225 3226
	struct kvm_run *kvm_run = svm->vcpu.run;

3227
	kvm_make_request(KVM_REQ_EVENT, &svm->vcpu);
3228
	svm_clear_vintr(svm);
3229
	svm->vmcb->control.int_ctl &= ~V_IRQ_MASK;
3230
	mark_dirty(svm->vmcb, VMCB_INTR);
3231
	++svm->vcpu.stat.irq_window_exits;
3232 3233 3234 3235
	/*
	 * If the user space waits to inject interrupts, exit as soon as
	 * possible
	 */
3236 3237 3238
	if (!irqchip_in_kernel(svm->vcpu.kvm) &&
	    kvm_run->request_interrupt_window &&
	    !kvm_cpu_has_interrupt(&svm->vcpu)) {
3239 3240 3241 3242 3243 3244 3245
		kvm_run->exit_reason = KVM_EXIT_IRQ_WINDOW_OPEN;
		return 0;
	}

	return 1;
}

3246 3247 3248 3249 3250 3251
static int pause_interception(struct vcpu_svm *svm)
{
	kvm_vcpu_on_spin(&(svm->vcpu));
	return 1;
}

3252
static int (*const svm_exit_handlers[])(struct vcpu_svm *svm) = {
3253 3254 3255 3256
	[SVM_EXIT_READ_CR0]			= cr_interception,
	[SVM_EXIT_READ_CR3]			= cr_interception,
	[SVM_EXIT_READ_CR4]			= cr_interception,
	[SVM_EXIT_READ_CR8]			= cr_interception,
A
Avi Kivity 已提交
3257
	[SVM_EXIT_CR0_SEL_WRITE]		= emulate_on_interception,
3258
	[SVM_EXIT_WRITE_CR0]			= cr_interception,
3259 3260
	[SVM_EXIT_WRITE_CR3]			= cr_interception,
	[SVM_EXIT_WRITE_CR4]			= cr_interception,
J
Joerg Roedel 已提交
3261
	[SVM_EXIT_WRITE_CR8]			= cr8_write_interception,
3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277
	[SVM_EXIT_READ_DR0]			= dr_interception,
	[SVM_EXIT_READ_DR1]			= dr_interception,
	[SVM_EXIT_READ_DR2]			= dr_interception,
	[SVM_EXIT_READ_DR3]			= dr_interception,
	[SVM_EXIT_READ_DR4]			= dr_interception,
	[SVM_EXIT_READ_DR5]			= dr_interception,
	[SVM_EXIT_READ_DR6]			= dr_interception,
	[SVM_EXIT_READ_DR7]			= dr_interception,
	[SVM_EXIT_WRITE_DR0]			= dr_interception,
	[SVM_EXIT_WRITE_DR1]			= dr_interception,
	[SVM_EXIT_WRITE_DR2]			= dr_interception,
	[SVM_EXIT_WRITE_DR3]			= dr_interception,
	[SVM_EXIT_WRITE_DR4]			= dr_interception,
	[SVM_EXIT_WRITE_DR5]			= dr_interception,
	[SVM_EXIT_WRITE_DR6]			= dr_interception,
	[SVM_EXIT_WRITE_DR7]			= dr_interception,
J
Jan Kiszka 已提交
3278 3279
	[SVM_EXIT_EXCP_BASE + DB_VECTOR]	= db_interception,
	[SVM_EXIT_EXCP_BASE + BP_VECTOR]	= bp_interception,
3280
	[SVM_EXIT_EXCP_BASE + UD_VECTOR]	= ud_interception,
J
Joerg Roedel 已提交
3281 3282 3283 3284
	[SVM_EXIT_EXCP_BASE + PF_VECTOR]	= pf_interception,
	[SVM_EXIT_EXCP_BASE + NM_VECTOR]	= nm_interception,
	[SVM_EXIT_EXCP_BASE + MC_VECTOR]	= mc_interception,
	[SVM_EXIT_INTR]				= intr_interception,
3285
	[SVM_EXIT_NMI]				= nmi_interception,
A
Avi Kivity 已提交
3286 3287
	[SVM_EXIT_SMI]				= nop_on_interception,
	[SVM_EXIT_INIT]				= nop_on_interception,
3288
	[SVM_EXIT_VINTR]			= interrupt_window_interception,
A
Avi Kivity 已提交
3289
	[SVM_EXIT_RDPMC]			= rdpmc_interception,
A
Avi Kivity 已提交
3290
	[SVM_EXIT_CPUID]			= cpuid_interception,
3291
	[SVM_EXIT_IRET]                         = iret_interception,
3292
	[SVM_EXIT_INVD]                         = emulate_on_interception,
3293
	[SVM_EXIT_PAUSE]			= pause_interception,
A
Avi Kivity 已提交
3294
	[SVM_EXIT_HLT]				= halt_interception,
M
Marcelo Tosatti 已提交
3295
	[SVM_EXIT_INVLPG]			= invlpg_interception,
A
Alexander Graf 已提交
3296
	[SVM_EXIT_INVLPGA]			= invlpga_interception,
J
Joerg Roedel 已提交
3297
	[SVM_EXIT_IOIO]				= io_interception,
A
Avi Kivity 已提交
3298 3299
	[SVM_EXIT_MSR]				= msr_interception,
	[SVM_EXIT_TASK_SWITCH]			= task_switch_interception,
3300
	[SVM_EXIT_SHUTDOWN]			= shutdown_interception,
A
Alexander Graf 已提交
3301
	[SVM_EXIT_VMRUN]			= vmrun_interception,
3302
	[SVM_EXIT_VMMCALL]			= vmmcall_interception,
3303 3304
	[SVM_EXIT_VMLOAD]			= vmload_interception,
	[SVM_EXIT_VMSAVE]			= vmsave_interception,
3305 3306
	[SVM_EXIT_STGI]				= stgi_interception,
	[SVM_EXIT_CLGI]				= clgi_interception,
3307
	[SVM_EXIT_SKINIT]			= skinit_interception,
3308
	[SVM_EXIT_WBINVD]                       = emulate_on_interception,
3309 3310
	[SVM_EXIT_MONITOR]			= invalid_op_interception,
	[SVM_EXIT_MWAIT]			= invalid_op_interception,
J
Joerg Roedel 已提交
3311
	[SVM_EXIT_XSETBV]			= xsetbv_interception,
3312
	[SVM_EXIT_NPF]				= pf_interception,
A
Avi Kivity 已提交
3313 3314
};

3315
static void dump_vmcb(struct kvm_vcpu *vcpu)
3316 3317 3318 3319 3320 3321
{
	struct vcpu_svm *svm = to_svm(vcpu);
	struct vmcb_control_area *control = &svm->vmcb->control;
	struct vmcb_save_area *save = &svm->vmcb->save;

	pr_err("VMCB Control Area:\n");
3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347
	pr_err("%-20s%04x\n", "cr_read:", control->intercept_cr & 0xffff);
	pr_err("%-20s%04x\n", "cr_write:", control->intercept_cr >> 16);
	pr_err("%-20s%04x\n", "dr_read:", control->intercept_dr & 0xffff);
	pr_err("%-20s%04x\n", "dr_write:", control->intercept_dr >> 16);
	pr_err("%-20s%08x\n", "exceptions:", control->intercept_exceptions);
	pr_err("%-20s%016llx\n", "intercepts:", control->intercept);
	pr_err("%-20s%d\n", "pause filter count:", control->pause_filter_count);
	pr_err("%-20s%016llx\n", "iopm_base_pa:", control->iopm_base_pa);
	pr_err("%-20s%016llx\n", "msrpm_base_pa:", control->msrpm_base_pa);
	pr_err("%-20s%016llx\n", "tsc_offset:", control->tsc_offset);
	pr_err("%-20s%d\n", "asid:", control->asid);
	pr_err("%-20s%d\n", "tlb_ctl:", control->tlb_ctl);
	pr_err("%-20s%08x\n", "int_ctl:", control->int_ctl);
	pr_err("%-20s%08x\n", "int_vector:", control->int_vector);
	pr_err("%-20s%08x\n", "int_state:", control->int_state);
	pr_err("%-20s%08x\n", "exit_code:", control->exit_code);
	pr_err("%-20s%016llx\n", "exit_info1:", control->exit_info_1);
	pr_err("%-20s%016llx\n", "exit_info2:", control->exit_info_2);
	pr_err("%-20s%08x\n", "exit_int_info:", control->exit_int_info);
	pr_err("%-20s%08x\n", "exit_int_info_err:", control->exit_int_info_err);
	pr_err("%-20s%lld\n", "nested_ctl:", control->nested_ctl);
	pr_err("%-20s%016llx\n", "nested_cr3:", control->nested_cr3);
	pr_err("%-20s%08x\n", "event_inj:", control->event_inj);
	pr_err("%-20s%08x\n", "event_inj_err:", control->event_inj_err);
	pr_err("%-20s%lld\n", "lbr_ctl:", control->lbr_ctl);
	pr_err("%-20s%016llx\n", "next_rip:", control->next_rip);
3348
	pr_err("VMCB State Save Area:\n");
3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "es:",
	       save->es.selector, save->es.attrib,
	       save->es.limit, save->es.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "cs:",
	       save->cs.selector, save->cs.attrib,
	       save->cs.limit, save->cs.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "ss:",
	       save->ss.selector, save->ss.attrib,
	       save->ss.limit, save->ss.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "ds:",
	       save->ds.selector, save->ds.attrib,
	       save->ds.limit, save->ds.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "fs:",
	       save->fs.selector, save->fs.attrib,
	       save->fs.limit, save->fs.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "gs:",
	       save->gs.selector, save->gs.attrib,
	       save->gs.limit, save->gs.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "gdtr:",
	       save->gdtr.selector, save->gdtr.attrib,
	       save->gdtr.limit, save->gdtr.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "ldtr:",
	       save->ldtr.selector, save->ldtr.attrib,
	       save->ldtr.limit, save->ldtr.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "idtr:",
	       save->idtr.selector, save->idtr.attrib,
	       save->idtr.limit, save->idtr.base);
	pr_err("%-5s s: %04x a: %04x l: %08x b: %016llx\n",
	       "tr:",
	       save->tr.selector, save->tr.attrib,
	       save->tr.limit, save->tr.base);
3389 3390
	pr_err("cpl:            %d                efer:         %016llx\n",
		save->cpl, save->efer);
3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "cr0:", save->cr0, "cr2:", save->cr2);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "cr3:", save->cr3, "cr4:", save->cr4);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "dr6:", save->dr6, "dr7:", save->dr7);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "rip:", save->rip, "rflags:", save->rflags);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "rsp:", save->rsp, "rax:", save->rax);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "star:", save->star, "lstar:", save->lstar);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "cstar:", save->cstar, "sfmask:", save->sfmask);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "kernel_gs_base:", save->kernel_gs_base,
	       "sysenter_cs:", save->sysenter_cs);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "sysenter_esp:", save->sysenter_esp,
	       "sysenter_eip:", save->sysenter_eip);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "gpat:", save->g_pat, "dbgctl:", save->dbgctl);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "br_from:", save->br_from, "br_to:", save->br_to);
	pr_err("%-15s %016llx %-13s %016llx\n",
	       "excp_from:", save->last_excp_from,
	       "excp_to:", save->last_excp_to);
3418 3419
}

3420 3421 3422 3423 3424 3425 3426 3427
static void svm_get_exit_info(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2)
{
	struct vmcb_control_area *control = &to_svm(vcpu)->vmcb->control;

	*info1 = control->exit_info_1;
	*info2 = control->exit_info_2;
}

A
Avi Kivity 已提交
3428
static int handle_exit(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
3429
{
3430
	struct vcpu_svm *svm = to_svm(vcpu);
A
Avi Kivity 已提交
3431
	struct kvm_run *kvm_run = vcpu->run;
3432
	u32 exit_code = svm->vmcb->control.exit_code;
A
Avi Kivity 已提交
3433

3434
	if (!is_cr_intercept(svm, INTERCEPT_CR0_WRITE))
3435 3436 3437
		vcpu->arch.cr0 = svm->vmcb->save.cr0;
	if (npt_enabled)
		vcpu->arch.cr3 = svm->vmcb->save.cr3;
3438

3439 3440 3441 3442 3443 3444 3445
	if (unlikely(svm->nested.exit_required)) {
		nested_svm_vmexit(svm);
		svm->nested.exit_required = false;

		return 1;
	}

3446
	if (is_guest_mode(vcpu)) {
3447 3448
		int vmexit;

3449 3450 3451 3452
		trace_kvm_nested_vmexit(svm->vmcb->save.rip, exit_code,
					svm->vmcb->control.exit_info_1,
					svm->vmcb->control.exit_info_2,
					svm->vmcb->control.exit_int_info,
3453 3454
					svm->vmcb->control.exit_int_info_err,
					KVM_ISA_SVM);
3455

3456 3457 3458 3459 3460 3461
		vmexit = nested_svm_exit_special(svm);

		if (vmexit == NESTED_EXIT_CONTINUE)
			vmexit = nested_svm_exit_handled(svm);

		if (vmexit == NESTED_EXIT_DONE)
3462 3463 3464
			return 1;
	}

3465 3466
	svm_complete_interrupts(svm);

3467 3468 3469 3470
	if (svm->vmcb->control.exit_code == SVM_EXIT_ERR) {
		kvm_run->exit_reason = KVM_EXIT_FAIL_ENTRY;
		kvm_run->fail_entry.hardware_entry_failure_reason
			= svm->vmcb->control.exit_code;
3471 3472
		pr_err("KVM: FAILED VMRUN WITH VMCB:\n");
		dump_vmcb(vcpu);
3473 3474 3475
		return 0;
	}

3476
	if (is_external_interrupt(svm->vmcb->control.exit_int_info) &&
3477
	    exit_code != SVM_EXIT_EXCP_BASE + PF_VECTOR &&
3478 3479
	    exit_code != SVM_EXIT_NPF && exit_code != SVM_EXIT_TASK_SWITCH &&
	    exit_code != SVM_EXIT_INTR && exit_code != SVM_EXIT_NMI)
A
Avi Kivity 已提交
3480 3481
		printk(KERN_ERR "%s: unexpected exit_ini_info 0x%x "
		       "exit_code 0x%x\n",
3482
		       __func__, svm->vmcb->control.exit_int_info,
A
Avi Kivity 已提交
3483 3484
		       exit_code);

3485
	if (exit_code >= ARRAY_SIZE(svm_exit_handlers)
J
Joe Perches 已提交
3486
	    || !svm_exit_handlers[exit_code]) {
A
Avi Kivity 已提交
3487
		kvm_run->exit_reason = KVM_EXIT_UNKNOWN;
3488
		kvm_run->hw.hardware_exit_reason = exit_code;
A
Avi Kivity 已提交
3489 3490 3491
		return 0;
	}

A
Avi Kivity 已提交
3492
	return svm_exit_handlers[exit_code](svm);
A
Avi Kivity 已提交
3493 3494 3495 3496 3497 3498
}

static void reload_tss(struct kvm_vcpu *vcpu)
{
	int cpu = raw_smp_processor_id();

3499 3500
	struct svm_cpu_data *sd = per_cpu(svm_data, cpu);
	sd->tss_desc->type = 9; /* available 32/64-bit TSS */
A
Avi Kivity 已提交
3501 3502 3503
	load_TR_desc();
}

R
Rusty Russell 已提交
3504
static void pre_svm_run(struct vcpu_svm *svm)
A
Avi Kivity 已提交
3505 3506 3507
{
	int cpu = raw_smp_processor_id();

3508
	struct svm_cpu_data *sd = per_cpu(svm_data, cpu);
A
Avi Kivity 已提交
3509

3510
	/* FIXME: handle wraparound of asid_generation */
3511 3512
	if (svm->asid_generation != sd->asid_generation)
		new_asid(svm, sd);
A
Avi Kivity 已提交
3513 3514
}

3515 3516 3517 3518 3519 3520
static void svm_inject_nmi(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	svm->vmcb->control.event_inj = SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_NMI;
	vcpu->arch.hflags |= HF_NMI_MASK;
3521
	set_intercept(svm, INTERCEPT_IRET);
3522 3523
	++vcpu->stat.nmi_injections;
}
A
Avi Kivity 已提交
3524

3525
static inline void svm_inject_irq(struct vcpu_svm *svm, int irq)
A
Avi Kivity 已提交
3526 3527 3528
{
	struct vmcb_control_area *control;

R
Rusty Russell 已提交
3529
	control = &svm->vmcb->control;
3530
	control->int_vector = irq;
A
Avi Kivity 已提交
3531 3532 3533
	control->int_ctl &= ~V_INTR_PRIO_MASK;
	control->int_ctl |= V_IRQ_MASK |
		((/*control->int_vector >> 4*/ 0xf) << V_INTR_PRIO_SHIFT);
3534
	mark_dirty(svm->vmcb, VMCB_INTR);
A
Avi Kivity 已提交
3535 3536
}

3537
static void svm_set_irq(struct kvm_vcpu *vcpu)
E
Eddie Dong 已提交
3538 3539 3540
{
	struct vcpu_svm *svm = to_svm(vcpu);

3541
	BUG_ON(!(gif_set(svm)));
3542

3543 3544 3545
	trace_kvm_inj_virq(vcpu->arch.interrupt.nr);
	++vcpu->stat.irq_injections;

3546 3547
	svm->vmcb->control.event_inj = vcpu->arch.interrupt.nr |
		SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_INTR;
E
Eddie Dong 已提交
3548 3549
}

3550
static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
3551 3552 3553
{
	struct vcpu_svm *svm = to_svm(vcpu);

3554
	if (is_guest_mode(vcpu) && (vcpu->arch.hflags & HF_VINTR_MASK))
3555 3556
		return;

3557
	if (irr == -1)
3558 3559
		return;

3560
	if (tpr >= irr)
3561
		set_cr_intercept(svm, INTERCEPT_CR8_WRITE);
3562
}
3563

3564 3565 3566 3567
static int svm_nmi_allowed(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	struct vmcb *vmcb = svm->vmcb;
J
Joerg Roedel 已提交
3568 3569 3570 3571 3572 3573
	int ret;
	ret = !(vmcb->control.int_state & SVM_INTERRUPT_SHADOW_MASK) &&
	      !(svm->vcpu.arch.hflags & HF_NMI_MASK);
	ret = ret && gif_set(svm) && nested_svm_nmi(svm);

	return ret;
3574 3575
}

J
Jan Kiszka 已提交
3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588
static bool svm_get_nmi_mask(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	return !!(svm->vcpu.arch.hflags & HF_NMI_MASK);
}

static void svm_set_nmi_mask(struct kvm_vcpu *vcpu, bool masked)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	if (masked) {
		svm->vcpu.arch.hflags |= HF_NMI_MASK;
3589
		set_intercept(svm, INTERCEPT_IRET);
J
Jan Kiszka 已提交
3590 3591
	} else {
		svm->vcpu.arch.hflags &= ~HF_NMI_MASK;
3592
		clr_intercept(svm, INTERCEPT_IRET);
J
Jan Kiszka 已提交
3593 3594 3595
	}
}

3596 3597 3598 3599
static int svm_interrupt_allowed(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	struct vmcb *vmcb = svm->vmcb;
3600 3601 3602 3603 3604 3605
	int ret;

	if (!gif_set(svm) ||
	     (vmcb->control.int_state & SVM_INTERRUPT_SHADOW_MASK))
		return 0;

3606
	ret = !!(kvm_get_rflags(vcpu) & X86_EFLAGS_IF);
3607

3608
	if (is_guest_mode(vcpu))
3609 3610 3611
		return ret && !(svm->vcpu.arch.hflags & HF_VINTR_MASK);

	return ret;
3612 3613
}

3614
static void enable_irq_window(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
3615
{
3616 3617
	struct vcpu_svm *svm = to_svm(vcpu);

J
Joerg Roedel 已提交
3618 3619 3620 3621 3622 3623
	/*
	 * In case GIF=0 we can't rely on the CPU to tell us when GIF becomes
	 * 1, because that's a separate STGI/VMRUN intercept.  The next time we
	 * get that intercept, this function will be called again though and
	 * we'll get the vintr intercept.
	 */
3624
	if (gif_set(svm) && nested_svm_intr(svm)) {
3625 3626 3627
		svm_set_vintr(svm);
		svm_inject_irq(svm, 0x0);
	}
3628 3629
}

3630
static void enable_nmi_window(struct kvm_vcpu *vcpu)
3631
{
3632
	struct vcpu_svm *svm = to_svm(vcpu);
3633

3634 3635 3636 3637
	if ((svm->vcpu.arch.hflags & (HF_NMI_MASK | HF_IRET_MASK))
	    == HF_NMI_MASK)
		return; /* IRET will cause a vm exit */

J
Joerg Roedel 已提交
3638 3639 3640 3641
	/*
	 * Something prevents NMI from been injected. Single step over possible
	 * problem (IRET or exception injection or interrupt shadow)
	 */
J
Jan Kiszka 已提交
3642
	svm->nmi_singlestep = true;
3643
	svm->vmcb->save.rflags |= (X86_EFLAGS_TF | X86_EFLAGS_RF);
3644
	update_db_bp_intercept(vcpu);
3645 3646
}

3647 3648 3649 3650 3651
static int svm_set_tss_addr(struct kvm *kvm, unsigned int addr)
{
	return 0;
}

3652 3653
static void svm_flush_tlb(struct kvm_vcpu *vcpu)
{
3654 3655 3656 3657 3658 3659
	struct vcpu_svm *svm = to_svm(vcpu);

	if (static_cpu_has(X86_FEATURE_FLUSHBYASID))
		svm->vmcb->control.tlb_ctl = TLB_CONTROL_FLUSH_ASID;
	else
		svm->asid_generation--;
3660 3661
}

3662 3663 3664 3665
static void svm_prepare_guest_switch(struct kvm_vcpu *vcpu)
{
}

3666 3667 3668 3669
static inline void sync_cr8_to_lapic(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);

3670
	if (is_guest_mode(vcpu) && (vcpu->arch.hflags & HF_VINTR_MASK))
3671 3672
		return;

3673
	if (!is_cr_intercept(svm, INTERCEPT_CR8_WRITE)) {
3674
		int cr8 = svm->vmcb->control.int_ctl & V_TPR_MASK;
3675
		kvm_set_cr8(vcpu, cr8);
3676 3677 3678
	}
}

3679 3680 3681 3682 3683
static inline void sync_lapic_to_cr8(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	u64 cr8;

3684
	if (is_guest_mode(vcpu) && (vcpu->arch.hflags & HF_VINTR_MASK))
3685 3686
		return;

3687 3688 3689 3690 3691
	cr8 = kvm_get_cr8(vcpu);
	svm->vmcb->control.int_ctl &= ~V_TPR_MASK;
	svm->vmcb->control.int_ctl |= cr8 & V_TPR_MASK;
}

3692 3693 3694 3695 3696
static void svm_complete_interrupts(struct vcpu_svm *svm)
{
	u8 vector;
	int type;
	u32 exitintinfo = svm->vmcb->control.exit_int_info;
3697 3698 3699
	unsigned int3_injected = svm->int3_injected;

	svm->int3_injected = 0;
3700

3701 3702 3703 3704 3705 3706
	/*
	 * If we've made progress since setting HF_IRET_MASK, we've
	 * executed an IRET and can allow NMI injection.
	 */
	if ((svm->vcpu.arch.hflags & HF_IRET_MASK)
	    && kvm_rip_read(&svm->vcpu) != svm->nmi_iret_rip) {
3707
		svm->vcpu.arch.hflags &= ~(HF_NMI_MASK | HF_IRET_MASK);
3708 3709
		kvm_make_request(KVM_REQ_EVENT, &svm->vcpu);
	}
3710

3711 3712 3713 3714 3715 3716 3717
	svm->vcpu.arch.nmi_injected = false;
	kvm_clear_exception_queue(&svm->vcpu);
	kvm_clear_interrupt_queue(&svm->vcpu);

	if (!(exitintinfo & SVM_EXITINTINFO_VALID))
		return;

3718 3719
	kvm_make_request(KVM_REQ_EVENT, &svm->vcpu);

3720 3721 3722 3723 3724 3725 3726 3727
	vector = exitintinfo & SVM_EXITINTINFO_VEC_MASK;
	type = exitintinfo & SVM_EXITINTINFO_TYPE_MASK;

	switch (type) {
	case SVM_EXITINTINFO_TYPE_NMI:
		svm->vcpu.arch.nmi_injected = true;
		break;
	case SVM_EXITINTINFO_TYPE_EXEPT:
3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738
		/*
		 * In case of software exceptions, do not reinject the vector,
		 * but re-execute the instruction instead. Rewind RIP first
		 * if we emulated INT3 before.
		 */
		if (kvm_exception_is_soft(vector)) {
			if (vector == BP_VECTOR && int3_injected &&
			    kvm_is_linear_rip(&svm->vcpu, svm->int3_rip))
				kvm_rip_write(&svm->vcpu,
					      kvm_rip_read(&svm->vcpu) -
					      int3_injected);
3739
			break;
3740
		}
3741 3742
		if (exitintinfo & SVM_EXITINTINFO_VALID_ERR) {
			u32 err = svm->vmcb->control.exit_int_info_err;
3743
			kvm_requeue_exception_e(&svm->vcpu, vector, err);
3744 3745

		} else
3746
			kvm_requeue_exception(&svm->vcpu, vector);
3747 3748
		break;
	case SVM_EXITINTINFO_TYPE_INTR:
3749
		kvm_queue_interrupt(&svm->vcpu, vector, false);
3750 3751 3752 3753 3754 3755
		break;
	default:
		break;
	}
}

A
Avi Kivity 已提交
3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766
static void svm_cancel_injection(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);
	struct vmcb_control_area *control = &svm->vmcb->control;

	control->exit_int_info = control->event_inj;
	control->exit_int_info_err = control->event_inj_err;
	control->event_inj = 0;
	svm_complete_interrupts(svm);
}

A
Avi Kivity 已提交
3767
static void svm_vcpu_run(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
3768
{
3769
	struct vcpu_svm *svm = to_svm(vcpu);
3770

3771 3772 3773 3774
	svm->vmcb->save.rax = vcpu->arch.regs[VCPU_REGS_RAX];
	svm->vmcb->save.rsp = vcpu->arch.regs[VCPU_REGS_RSP];
	svm->vmcb->save.rip = vcpu->arch.regs[VCPU_REGS_RIP];

3775 3776 3777 3778 3779 3780 3781
	/*
	 * A vmexit emulation is required before the vcpu can be executed
	 * again.
	 */
	if (unlikely(svm->nested.exit_required))
		return;

R
Rusty Russell 已提交
3782
	pre_svm_run(svm);
A
Avi Kivity 已提交
3783

3784 3785
	sync_lapic_to_cr8(vcpu);

3786
	svm->vmcb->save.cr2 = vcpu->arch.cr2;
A
Avi Kivity 已提交
3787

3788 3789 3790
	clgi();

	local_irq_enable();
3791

A
Avi Kivity 已提交
3792
	asm volatile (
A
Avi Kivity 已提交
3793 3794 3795 3796 3797 3798 3799
		"push %%" _ASM_BP "; \n\t"
		"mov %c[rbx](%[svm]), %%" _ASM_BX " \n\t"
		"mov %c[rcx](%[svm]), %%" _ASM_CX " \n\t"
		"mov %c[rdx](%[svm]), %%" _ASM_DX " \n\t"
		"mov %c[rsi](%[svm]), %%" _ASM_SI " \n\t"
		"mov %c[rdi](%[svm]), %%" _ASM_DI " \n\t"
		"mov %c[rbp](%[svm]), %%" _ASM_BP " \n\t"
3800
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
3801 3802 3803 3804 3805 3806 3807 3808
		"mov %c[r8](%[svm]),  %%r8  \n\t"
		"mov %c[r9](%[svm]),  %%r9  \n\t"
		"mov %c[r10](%[svm]), %%r10 \n\t"
		"mov %c[r11](%[svm]), %%r11 \n\t"
		"mov %c[r12](%[svm]), %%r12 \n\t"
		"mov %c[r13](%[svm]), %%r13 \n\t"
		"mov %c[r14](%[svm]), %%r14 \n\t"
		"mov %c[r15](%[svm]), %%r15 \n\t"
A
Avi Kivity 已提交
3809 3810 3811
#endif

		/* Enter guest mode */
A
Avi Kivity 已提交
3812 3813
		"push %%" _ASM_AX " \n\t"
		"mov %c[vmcb](%[svm]), %%" _ASM_AX " \n\t"
3814 3815 3816
		__ex(SVM_VMLOAD) "\n\t"
		__ex(SVM_VMRUN) "\n\t"
		__ex(SVM_VMSAVE) "\n\t"
A
Avi Kivity 已提交
3817
		"pop %%" _ASM_AX " \n\t"
A
Avi Kivity 已提交
3818 3819

		/* Save guest registers, load host registers */
A
Avi Kivity 已提交
3820 3821 3822 3823 3824 3825
		"mov %%" _ASM_BX ", %c[rbx](%[svm]) \n\t"
		"mov %%" _ASM_CX ", %c[rcx](%[svm]) \n\t"
		"mov %%" _ASM_DX ", %c[rdx](%[svm]) \n\t"
		"mov %%" _ASM_SI ", %c[rsi](%[svm]) \n\t"
		"mov %%" _ASM_DI ", %c[rdi](%[svm]) \n\t"
		"mov %%" _ASM_BP ", %c[rbp](%[svm]) \n\t"
3826
#ifdef CONFIG_X86_64
R
Rusty Russell 已提交
3827 3828 3829 3830 3831 3832 3833 3834
		"mov %%r8,  %c[r8](%[svm]) \n\t"
		"mov %%r9,  %c[r9](%[svm]) \n\t"
		"mov %%r10, %c[r10](%[svm]) \n\t"
		"mov %%r11, %c[r11](%[svm]) \n\t"
		"mov %%r12, %c[r12](%[svm]) \n\t"
		"mov %%r13, %c[r13](%[svm]) \n\t"
		"mov %%r14, %c[r14](%[svm]) \n\t"
		"mov %%r15, %c[r15](%[svm]) \n\t"
A
Avi Kivity 已提交
3835
#endif
A
Avi Kivity 已提交
3836
		"pop %%" _ASM_BP
A
Avi Kivity 已提交
3837
		:
R
Rusty Russell 已提交
3838
		: [svm]"a"(svm),
A
Avi Kivity 已提交
3839
		  [vmcb]"i"(offsetof(struct vcpu_svm, vmcb_pa)),
3840 3841 3842 3843 3844 3845
		  [rbx]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RBX])),
		  [rcx]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RCX])),
		  [rdx]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RDX])),
		  [rsi]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RSI])),
		  [rdi]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RDI])),
		  [rbp]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_RBP]))
3846
#ifdef CONFIG_X86_64
3847 3848 3849 3850 3851 3852 3853 3854
		  , [r8]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R8])),
		  [r9]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R9])),
		  [r10]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R10])),
		  [r11]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R11])),
		  [r12]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R12])),
		  [r13]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R13])),
		  [r14]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R14])),
		  [r15]"i"(offsetof(struct vcpu_svm, vcpu.arch.regs[VCPU_REGS_R15]))
A
Avi Kivity 已提交
3855
#endif
3856 3857
		: "cc", "memory"
#ifdef CONFIG_X86_64
A
Avi Kivity 已提交
3858
		, "rbx", "rcx", "rdx", "rsi", "rdi"
3859
		, "r8", "r9", "r10", "r11" , "r12", "r13", "r14", "r15"
A
Avi Kivity 已提交
3860 3861
#else
		, "ebx", "ecx", "edx", "esi", "edi"
3862 3863
#endif
		);
A
Avi Kivity 已提交
3864

3865 3866 3867
#ifdef CONFIG_X86_64
	wrmsrl(MSR_GS_BASE, svm->host.gs_base);
#else
3868
	loadsegment(fs, svm->host.fs);
3869 3870 3871
#ifndef CONFIG_X86_32_LAZY_GS
	loadsegment(gs, svm->host.gs);
#endif
3872
#endif
A
Avi Kivity 已提交
3873 3874 3875

	reload_tss(vcpu);

3876 3877
	local_irq_disable();

3878 3879 3880 3881 3882
	vcpu->arch.cr2 = svm->vmcb->save.cr2;
	vcpu->arch.regs[VCPU_REGS_RAX] = svm->vmcb->save.rax;
	vcpu->arch.regs[VCPU_REGS_RSP] = svm->vmcb->save.rsp;
	vcpu->arch.regs[VCPU_REGS_RIP] = svm->vmcb->save.rip;

3883 3884
	trace_kvm_exit(svm->vmcb->control.exit_code, vcpu, KVM_ISA_SVM);

3885 3886 3887 3888 3889 3890 3891 3892 3893 3894
	if (unlikely(svm->vmcb->control.exit_code == SVM_EXIT_NMI))
		kvm_before_handle_nmi(&svm->vcpu);

	stgi();

	/* Any pending NMI will happen here */

	if (unlikely(svm->vmcb->control.exit_code == SVM_EXIT_NMI))
		kvm_after_handle_nmi(&svm->vcpu);

3895 3896
	sync_cr8_to_lapic(vcpu);

3897
	svm->next_rip = 0;
3898

3899 3900
	svm->vmcb->control.tlb_ctl = TLB_CONTROL_DO_NOTHING;

G
Gleb Natapov 已提交
3901 3902 3903 3904
	/* if exit due to PF check for async PF */
	if (svm->vmcb->control.exit_code == SVM_EXIT_EXCP_BASE + PF_VECTOR)
		svm->apf_reason = kvm_read_and_reset_pf_reason();

A
Avi Kivity 已提交
3905 3906 3907 3908
	if (npt_enabled) {
		vcpu->arch.regs_avail &= ~(1 << VCPU_EXREG_PDPTR);
		vcpu->arch.regs_dirty &= ~(1 << VCPU_EXREG_PDPTR);
	}
3909 3910 3911 3912 3913 3914 3915 3916

	/*
	 * We need to handle MC intercepts here before the vcpu has a chance to
	 * change the physical cpu
	 */
	if (unlikely(svm->vmcb->control.exit_code ==
		     SVM_EXIT_EXCP_BASE + MC_VECTOR))
		svm_handle_mce(svm);
3917 3918

	mark_all_clean(svm->vmcb);
A
Avi Kivity 已提交
3919 3920 3921 3922
}

static void svm_set_cr3(struct kvm_vcpu *vcpu, unsigned long root)
{
3923 3924 3925
	struct vcpu_svm *svm = to_svm(vcpu);

	svm->vmcb->save.cr3 = root;
3926
	mark_dirty(svm->vmcb, VMCB_CR);
3927
	svm_flush_tlb(vcpu);
A
Avi Kivity 已提交
3928 3929
}

3930 3931 3932 3933 3934
static void set_tdp_cr3(struct kvm_vcpu *vcpu, unsigned long root)
{
	struct vcpu_svm *svm = to_svm(vcpu);

	svm->vmcb->control.nested_cr3 = root;
3935
	mark_dirty(svm->vmcb, VMCB_NPT);
3936 3937

	/* Also sync guest cr3 here in case we live migrate */
3938
	svm->vmcb->save.cr3 = kvm_read_cr3(vcpu);
3939
	mark_dirty(svm->vmcb, VMCB_CR);
3940

3941
	svm_flush_tlb(vcpu);
3942 3943
}

A
Avi Kivity 已提交
3944 3945
static int is_disabled(void)
{
3946 3947 3948 3949 3950 3951
	u64 vm_cr;

	rdmsrl(MSR_VM_CR, vm_cr);
	if (vm_cr & (1 << SVM_VM_CR_SVM_DISABLE))
		return 1;

A
Avi Kivity 已提交
3952 3953 3954
	return 0;
}

I
Ingo Molnar 已提交
3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965
static void
svm_patch_hypercall(struct kvm_vcpu *vcpu, unsigned char *hypercall)
{
	/*
	 * Patch in the VMMCALL instruction:
	 */
	hypercall[0] = 0x0f;
	hypercall[1] = 0x01;
	hypercall[2] = 0xd9;
}

Y
Yang, Sheng 已提交
3966 3967 3968 3969 3970
static void svm_check_processor_compat(void *rtn)
{
	*(int *)rtn = 0;
}

3971 3972 3973 3974 3975
static bool svm_cpu_has_accelerated_tpr(void)
{
	return false;
}

3976
static u64 svm_get_mt_mask(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio)
S
Sheng Yang 已提交
3977 3978 3979 3980
{
	return 0;
}

3981 3982 3983 3984
static void svm_cpuid_update(struct kvm_vcpu *vcpu)
{
}

3985 3986
static void svm_set_supported_cpuid(u32 func, struct kvm_cpuid_entry2 *entry)
{
3987
	switch (func) {
3988 3989 3990 3991
	case 0x80000001:
		if (nested)
			entry->ecx |= (1 << 2); /* Set SVM bit */
		break;
3992 3993 3994 3995 3996
	case 0x8000000A:
		entry->eax = 1; /* SVM revision 1 */
		entry->ebx = 8; /* Lets support 8 ASIDs in case we add proper
				   ASID emulation to nested SVM */
		entry->ecx = 0; /* Reserved */
3997 3998 3999 4000
		entry->edx = 0; /* Per default do not support any
				   additional features */

		/* Support next_rip if host supports it */
4001
		if (boot_cpu_has(X86_FEATURE_NRIPS))
4002
			entry->edx |= SVM_FEATURE_NRIP;
4003

4004 4005 4006 4007
		/* Support NPT for the guest if enabled */
		if (npt_enabled)
			entry->edx |= SVM_FEATURE_NPT;

4008 4009
		break;
	}
4010 4011
}

4012
static int svm_get_lpage_level(void)
4013
{
4014
	return PT_PDPE_LEVEL;
4015 4016
}

4017 4018 4019 4020 4021
static bool svm_rdtscp_supported(void)
{
	return false;
}

4022 4023 4024 4025 4026
static bool svm_invpcid_supported(void)
{
	return false;
}

4027 4028 4029 4030 4031
static bool svm_has_wbinvd_exit(void)
{
	return true;
}

4032 4033 4034 4035
static void svm_fpu_deactivate(struct kvm_vcpu *vcpu)
{
	struct vcpu_svm *svm = to_svm(vcpu);

4036
	set_exception_intercept(svm, NM_VECTOR);
4037
	update_cr0_intercept(svm);
4038 4039
}

4040
#define PRE_EX(exit)  { .exit_code = (exit), \
4041
			.stage = X86_ICPT_PRE_EXCEPT, }
4042
#define POST_EX(exit) { .exit_code = (exit), \
4043
			.stage = X86_ICPT_POST_EXCEPT, }
4044
#define POST_MEM(exit) { .exit_code = (exit), \
4045
			.stage = X86_ICPT_POST_MEMACCESS, }
4046

4047
static const struct __x86_intercept {
4048 4049 4050 4051 4052 4053 4054 4055
	u32 exit_code;
	enum x86_intercept_stage stage;
} x86_intercept_map[] = {
	[x86_intercept_cr_read]		= POST_EX(SVM_EXIT_READ_CR0),
	[x86_intercept_cr_write]	= POST_EX(SVM_EXIT_WRITE_CR0),
	[x86_intercept_clts]		= POST_EX(SVM_EXIT_WRITE_CR0),
	[x86_intercept_lmsw]		= POST_EX(SVM_EXIT_WRITE_CR0),
	[x86_intercept_smsw]		= POST_EX(SVM_EXIT_READ_CR0),
4056 4057
	[x86_intercept_dr_read]		= POST_EX(SVM_EXIT_READ_DR0),
	[x86_intercept_dr_write]	= POST_EX(SVM_EXIT_WRITE_DR0),
4058 4059 4060 4061 4062 4063 4064 4065
	[x86_intercept_sldt]		= POST_EX(SVM_EXIT_LDTR_READ),
	[x86_intercept_str]		= POST_EX(SVM_EXIT_TR_READ),
	[x86_intercept_lldt]		= POST_EX(SVM_EXIT_LDTR_WRITE),
	[x86_intercept_ltr]		= POST_EX(SVM_EXIT_TR_WRITE),
	[x86_intercept_sgdt]		= POST_EX(SVM_EXIT_GDTR_READ),
	[x86_intercept_sidt]		= POST_EX(SVM_EXIT_IDTR_READ),
	[x86_intercept_lgdt]		= POST_EX(SVM_EXIT_GDTR_WRITE),
	[x86_intercept_lidt]		= POST_EX(SVM_EXIT_IDTR_WRITE),
4066 4067 4068 4069 4070 4071 4072 4073
	[x86_intercept_vmrun]		= POST_EX(SVM_EXIT_VMRUN),
	[x86_intercept_vmmcall]		= POST_EX(SVM_EXIT_VMMCALL),
	[x86_intercept_vmload]		= POST_EX(SVM_EXIT_VMLOAD),
	[x86_intercept_vmsave]		= POST_EX(SVM_EXIT_VMSAVE),
	[x86_intercept_stgi]		= POST_EX(SVM_EXIT_STGI),
	[x86_intercept_clgi]		= POST_EX(SVM_EXIT_CLGI),
	[x86_intercept_skinit]		= POST_EX(SVM_EXIT_SKINIT),
	[x86_intercept_invlpga]		= POST_EX(SVM_EXIT_INVLPGA),
4074 4075 4076
	[x86_intercept_rdtscp]		= POST_EX(SVM_EXIT_RDTSCP),
	[x86_intercept_monitor]		= POST_MEM(SVM_EXIT_MONITOR),
	[x86_intercept_mwait]		= POST_EX(SVM_EXIT_MWAIT),
4077 4078 4079 4080 4081 4082 4083 4084 4085
	[x86_intercept_invlpg]		= POST_EX(SVM_EXIT_INVLPG),
	[x86_intercept_invd]		= POST_EX(SVM_EXIT_INVD),
	[x86_intercept_wbinvd]		= POST_EX(SVM_EXIT_WBINVD),
	[x86_intercept_wrmsr]		= POST_EX(SVM_EXIT_MSR),
	[x86_intercept_rdtsc]		= POST_EX(SVM_EXIT_RDTSC),
	[x86_intercept_rdmsr]		= POST_EX(SVM_EXIT_MSR),
	[x86_intercept_rdpmc]		= POST_EX(SVM_EXIT_RDPMC),
	[x86_intercept_cpuid]		= PRE_EX(SVM_EXIT_CPUID),
	[x86_intercept_rsm]		= PRE_EX(SVM_EXIT_RSM),
4086 4087 4088 4089 4090 4091 4092
	[x86_intercept_pause]		= PRE_EX(SVM_EXIT_PAUSE),
	[x86_intercept_pushf]		= PRE_EX(SVM_EXIT_PUSHF),
	[x86_intercept_popf]		= PRE_EX(SVM_EXIT_POPF),
	[x86_intercept_intn]		= PRE_EX(SVM_EXIT_SWINT),
	[x86_intercept_iret]		= PRE_EX(SVM_EXIT_IRET),
	[x86_intercept_icebp]		= PRE_EX(SVM_EXIT_ICEBP),
	[x86_intercept_hlt]		= POST_EX(SVM_EXIT_HLT),
4093 4094 4095 4096
	[x86_intercept_in]		= POST_EX(SVM_EXIT_IOIO),
	[x86_intercept_ins]		= POST_EX(SVM_EXIT_IOIO),
	[x86_intercept_out]		= POST_EX(SVM_EXIT_IOIO),
	[x86_intercept_outs]		= POST_EX(SVM_EXIT_IOIO),
4097 4098
};

4099
#undef PRE_EX
4100
#undef POST_EX
4101
#undef POST_MEM
4102

4103 4104 4105 4106
static int svm_check_intercept(struct kvm_vcpu *vcpu,
			       struct x86_instruction_info *info,
			       enum x86_intercept_stage stage)
{
4107 4108 4109 4110 4111 4112 4113 4114 4115 4116
	struct vcpu_svm *svm = to_svm(vcpu);
	int vmexit, ret = X86EMUL_CONTINUE;
	struct __x86_intercept icpt_info;
	struct vmcb *vmcb = svm->vmcb;

	if (info->intercept >= ARRAY_SIZE(x86_intercept_map))
		goto out;

	icpt_info = x86_intercept_map[info->intercept];

4117
	if (stage != icpt_info.stage)
4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155
		goto out;

	switch (icpt_info.exit_code) {
	case SVM_EXIT_READ_CR0:
		if (info->intercept == x86_intercept_cr_read)
			icpt_info.exit_code += info->modrm_reg;
		break;
	case SVM_EXIT_WRITE_CR0: {
		unsigned long cr0, val;
		u64 intercept;

		if (info->intercept == x86_intercept_cr_write)
			icpt_info.exit_code += info->modrm_reg;

		if (icpt_info.exit_code != SVM_EXIT_WRITE_CR0)
			break;

		intercept = svm->nested.intercept;

		if (!(intercept & (1ULL << INTERCEPT_SELECTIVE_CR0)))
			break;

		cr0 = vcpu->arch.cr0 & ~SVM_CR0_SELECTIVE_MASK;
		val = info->src_val  & ~SVM_CR0_SELECTIVE_MASK;

		if (info->intercept == x86_intercept_lmsw) {
			cr0 &= 0xfUL;
			val &= 0xfUL;
			/* lmsw can't clear PE - catch this here */
			if (cr0 & X86_CR0_PE)
				val |= X86_CR0_PE;
		}

		if (cr0 ^ val)
			icpt_info.exit_code = SVM_EXIT_CR0_SEL_WRITE;

		break;
	}
4156 4157 4158 4159
	case SVM_EXIT_READ_DR0:
	case SVM_EXIT_WRITE_DR0:
		icpt_info.exit_code += info->modrm_reg;
		break;
4160 4161 4162 4163 4164 4165
	case SVM_EXIT_MSR:
		if (info->intercept == x86_intercept_wrmsr)
			vmcb->control.exit_info_1 = 1;
		else
			vmcb->control.exit_info_1 = 0;
		break;
4166 4167 4168 4169 4170 4171 4172
	case SVM_EXIT_PAUSE:
		/*
		 * We get this for NOP only, but pause
		 * is rep not, check this here
		 */
		if (info->rep_prefix != REPE_PREFIX)
			goto out;
4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204
	case SVM_EXIT_IOIO: {
		u64 exit_info;
		u32 bytes;

		exit_info = (vcpu->arch.regs[VCPU_REGS_RDX] & 0xffff) << 16;

		if (info->intercept == x86_intercept_in ||
		    info->intercept == x86_intercept_ins) {
			exit_info |= SVM_IOIO_TYPE_MASK;
			bytes = info->src_bytes;
		} else {
			bytes = info->dst_bytes;
		}

		if (info->intercept == x86_intercept_outs ||
		    info->intercept == x86_intercept_ins)
			exit_info |= SVM_IOIO_STR_MASK;

		if (info->rep_prefix)
			exit_info |= SVM_IOIO_REP_MASK;

		bytes = min(bytes, 4u);

		exit_info |= bytes << SVM_IOIO_SIZE_SHIFT;

		exit_info |= (u32)info->ad_bytes << (SVM_IOIO_ASIZE_SHIFT - 1);

		vmcb->control.exit_info_1 = exit_info;
		vmcb->control.exit_info_2 = info->next_rip;

		break;
	}
4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217
	default:
		break;
	}

	vmcb->control.next_rip  = info->next_rip;
	vmcb->control.exit_code = icpt_info.exit_code;
	vmexit = nested_svm_exit_handled(svm);

	ret = (vmexit == NESTED_EXIT_DONE) ? X86EMUL_INTERCEPTED
					   : X86EMUL_CONTINUE;

out:
	return ret;
4218 4219
}

4220
static struct kvm_x86_ops svm_x86_ops = {
A
Avi Kivity 已提交
4221 4222 4223 4224
	.cpu_has_kvm_support = has_svm,
	.disabled_by_bios = is_disabled,
	.hardware_setup = svm_hardware_setup,
	.hardware_unsetup = svm_hardware_unsetup,
Y
Yang, Sheng 已提交
4225
	.check_processor_compatibility = svm_check_processor_compat,
A
Avi Kivity 已提交
4226 4227
	.hardware_enable = svm_hardware_enable,
	.hardware_disable = svm_hardware_disable,
4228
	.cpu_has_accelerated_tpr = svm_cpu_has_accelerated_tpr,
A
Avi Kivity 已提交
4229 4230 4231

	.vcpu_create = svm_create_vcpu,
	.vcpu_free = svm_free_vcpu,
4232
	.vcpu_reset = svm_vcpu_reset,
A
Avi Kivity 已提交
4233

4234
	.prepare_guest_switch = svm_prepare_guest_switch,
A
Avi Kivity 已提交
4235 4236 4237
	.vcpu_load = svm_vcpu_load,
	.vcpu_put = svm_vcpu_put,

4238
	.update_db_bp_intercept = update_db_bp_intercept,
A
Avi Kivity 已提交
4239 4240 4241 4242 4243
	.get_msr = svm_get_msr,
	.set_msr = svm_set_msr,
	.get_segment_base = svm_get_segment_base,
	.get_segment = svm_get_segment,
	.set_segment = svm_set_segment,
4244
	.get_cpl = svm_get_cpl,
4245
	.get_cs_db_l_bits = kvm_get_cs_db_l_bits,
4246
	.decache_cr0_guest_bits = svm_decache_cr0_guest_bits,
4247
	.decache_cr3 = svm_decache_cr3,
4248
	.decache_cr4_guest_bits = svm_decache_cr4_guest_bits,
A
Avi Kivity 已提交
4249 4250 4251 4252 4253 4254 4255 4256
	.set_cr0 = svm_set_cr0,
	.set_cr3 = svm_set_cr3,
	.set_cr4 = svm_set_cr4,
	.set_efer = svm_set_efer,
	.get_idt = svm_get_idt,
	.set_idt = svm_set_idt,
	.get_gdt = svm_get_gdt,
	.set_gdt = svm_set_gdt,
4257
	.set_dr7 = svm_set_dr7,
A
Avi Kivity 已提交
4258
	.cache_reg = svm_cache_reg,
A
Avi Kivity 已提交
4259 4260
	.get_rflags = svm_get_rflags,
	.set_rflags = svm_set_rflags,
A
Avi Kivity 已提交
4261
	.fpu_activate = svm_fpu_activate,
4262
	.fpu_deactivate = svm_fpu_deactivate,
A
Avi Kivity 已提交
4263 4264 4265 4266

	.tlb_flush = svm_flush_tlb,

	.run = svm_vcpu_run,
4267
	.handle_exit = handle_exit,
A
Avi Kivity 已提交
4268
	.skip_emulated_instruction = skip_emulated_instruction,
4269 4270
	.set_interrupt_shadow = svm_set_interrupt_shadow,
	.get_interrupt_shadow = svm_get_interrupt_shadow,
I
Ingo Molnar 已提交
4271
	.patch_hypercall = svm_patch_hypercall,
E
Eddie Dong 已提交
4272
	.set_irq = svm_set_irq,
4273
	.set_nmi = svm_inject_nmi,
4274
	.queue_exception = svm_queue_exception,
A
Avi Kivity 已提交
4275
	.cancel_injection = svm_cancel_injection,
4276
	.interrupt_allowed = svm_interrupt_allowed,
4277
	.nmi_allowed = svm_nmi_allowed,
J
Jan Kiszka 已提交
4278 4279
	.get_nmi_mask = svm_get_nmi_mask,
	.set_nmi_mask = svm_set_nmi_mask,
4280 4281 4282
	.enable_nmi_window = enable_nmi_window,
	.enable_irq_window = enable_irq_window,
	.update_cr8_intercept = update_cr8_intercept,
4283 4284

	.set_tss_addr = svm_set_tss_addr,
4285
	.get_tdp_level = get_npt_level,
4286
	.get_mt_mask = svm_get_mt_mask,
4287

4288 4289
	.get_exit_info = svm_get_exit_info,

4290
	.get_lpage_level = svm_get_lpage_level,
4291 4292

	.cpuid_update = svm_cpuid_update,
4293 4294

	.rdtscp_supported = svm_rdtscp_supported,
4295
	.invpcid_supported = svm_invpcid_supported,
4296 4297

	.set_supported_cpuid = svm_set_supported_cpuid,
4298 4299

	.has_wbinvd_exit = svm_has_wbinvd_exit,
4300

4301
	.set_tsc_khz = svm_set_tsc_khz,
4302
	.write_tsc_offset = svm_write_tsc_offset,
Z
Zachary Amsden 已提交
4303
	.adjust_tsc_offset = svm_adjust_tsc_offset,
4304
	.compute_tsc_offset = svm_compute_tsc_offset,
N
Nadav Har'El 已提交
4305
	.read_l1_tsc = svm_read_l1_tsc,
4306 4307

	.set_tdp_cr3 = set_tdp_cr3,
4308 4309

	.check_intercept = svm_check_intercept,
A
Avi Kivity 已提交
4310 4311 4312 4313
};

static int __init svm_init(void)
{
4314
	return kvm_init(&svm_x86_ops, sizeof(struct vcpu_svm),
4315
			__alignof__(struct vcpu_svm), THIS_MODULE);
A
Avi Kivity 已提交
4316 4317 4318 4319
}

static void __exit svm_exit(void)
{
4320
	kvm_exit();
A
Avi Kivity 已提交
4321 4322 4323 4324
}

module_init(svm_init)
module_exit(svm_exit)