x86.c 110.8 KB
Newer Older
1 2 3 4 5 6
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * derived from drivers/kvm/kvm_main.c
 *
 * Copyright (C) 2006 Qumranet, Inc.
B
Ben-Ami Yassour 已提交
7 8
 * Copyright (C) 2008 Qumranet, Inc.
 * Copyright IBM Corporation, 2008
9 10 11 12
 *
 * Authors:
 *   Avi Kivity   <avi@qumranet.com>
 *   Yaniv Kamay  <yaniv@qumranet.com>
B
Ben-Ami Yassour 已提交
13 14
 *   Amit Shah    <amit.shah@qumranet.com>
 *   Ben-Ami Yassour <benami@il.ibm.com>
15 16 17 18 19 20
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */

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

29
#include <linux/clocksource.h>
B
Ben-Ami Yassour 已提交
30
#include <linux/interrupt.h>
31 32 33
#include <linux/kvm.h>
#include <linux/fs.h>
#include <linux/vmalloc.h>
34
#include <linux/module.h>
35
#include <linux/mman.h>
36
#include <linux/highmem.h>
J
Joerg Roedel 已提交
37
#include <linux/iommu.h>
B
Ben-Ami Yassour 已提交
38
#include <linux/intel-iommu.h>
39
#include <linux/cpufreq.h>
40 41

#include <asm/uaccess.h>
42
#include <asm/msr.h>
43
#include <asm/desc.h>
S
Sheng Yang 已提交
44
#include <asm/mtrr.h>
45

46
#define MAX_IO_MSRS 256
47 48 49 50 51 52 53 54 55 56 57
#define CR0_RESERVED_BITS						\
	(~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
			  | X86_CR0_ET | X86_CR0_NE | X86_CR0_WP | X86_CR0_AM \
			  | X86_CR0_NW | X86_CR0_CD | X86_CR0_PG))
#define CR4_RESERVED_BITS						\
	(~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\
			  | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE	\
			  | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR	\
			  | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE))

#define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
58 59 60 61 62 63 64 65 66
/* EFER defaults:
 * - enable syscall per default because its emulated by KVM
 * - enable LME and LMA per default on 64 bit KVM
 */
#ifdef CONFIG_X86_64
static u64 __read_mostly efer_reserved_bits = 0xfffffffffffffafeULL;
#else
static u64 __read_mostly efer_reserved_bits = 0xfffffffffffffffeULL;
#endif
67

68 69
#define VM_STAT(x) offsetof(struct kvm, stat.x), KVM_STAT_VM
#define VCPU_STAT(x) offsetof(struct kvm_vcpu, stat.x), KVM_STAT_VCPU
70

71 72
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
				    struct kvm_cpuid_entry2 __user *entries);
73 74
struct kvm_cpuid_entry2 *kvm_find_cpuid_entry(struct kvm_vcpu *vcpu,
					      u32 function, u32 index);
75

76
struct kvm_x86_ops *kvm_x86_ops;
77
EXPORT_SYMBOL_GPL(kvm_x86_ops);
78

79
struct kvm_stats_debugfs_item debugfs_entries[] = {
80 81 82 83 84 85 86 87 88
	{ "pf_fixed", VCPU_STAT(pf_fixed) },
	{ "pf_guest", VCPU_STAT(pf_guest) },
	{ "tlb_flush", VCPU_STAT(tlb_flush) },
	{ "invlpg", VCPU_STAT(invlpg) },
	{ "exits", VCPU_STAT(exits) },
	{ "io_exits", VCPU_STAT(io_exits) },
	{ "mmio_exits", VCPU_STAT(mmio_exits) },
	{ "signal_exits", VCPU_STAT(signal_exits) },
	{ "irq_window", VCPU_STAT(irq_window_exits) },
89
	{ "nmi_window", VCPU_STAT(nmi_window_exits) },
90 91
	{ "halt_exits", VCPU_STAT(halt_exits) },
	{ "halt_wakeup", VCPU_STAT(halt_wakeup) },
A
Amit Shah 已提交
92
	{ "hypercalls", VCPU_STAT(hypercalls) },
93 94 95 96 97 98 99
	{ "request_irq", VCPU_STAT(request_irq_exits) },
	{ "irq_exits", VCPU_STAT(irq_exits) },
	{ "host_state_reload", VCPU_STAT(host_state_reload) },
	{ "efer_reload", VCPU_STAT(efer_reload) },
	{ "fpu_reload", VCPU_STAT(fpu_reload) },
	{ "insn_emulation", VCPU_STAT(insn_emulation) },
	{ "insn_emulation_fail", VCPU_STAT(insn_emulation_fail) },
100
	{ "irq_injections", VCPU_STAT(irq_injections) },
101
	{ "nmi_injections", VCPU_STAT(nmi_injections) },
A
Avi Kivity 已提交
102 103 104 105 106 107
	{ "mmu_shadow_zapped", VM_STAT(mmu_shadow_zapped) },
	{ "mmu_pte_write", VM_STAT(mmu_pte_write) },
	{ "mmu_pte_updated", VM_STAT(mmu_pte_updated) },
	{ "mmu_pde_zapped", VM_STAT(mmu_pde_zapped) },
	{ "mmu_flooded", VM_STAT(mmu_flooded) },
	{ "mmu_recycled", VM_STAT(mmu_recycled) },
A
Avi Kivity 已提交
108
	{ "mmu_cache_miss", VM_STAT(mmu_cache_miss) },
109
	{ "mmu_unsync", VM_STAT(mmu_unsync) },
110
	{ "remote_tlb_flush", VM_STAT(remote_tlb_flush) },
M
Marcelo Tosatti 已提交
111
	{ "largepages", VM_STAT(lpages) },
112 113 114
	{ NULL }
};

115 116 117
unsigned long segment_base(u16 selector)
{
	struct descriptor_table gdt;
118
	struct desc_struct *d;
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
	unsigned long table_base;
	unsigned long v;

	if (selector == 0)
		return 0;

	asm("sgdt %0" : "=m"(gdt));
	table_base = gdt.base;

	if (selector & 4) {           /* from ldt */
		u16 ldt_selector;

		asm("sldt %0" : "=g"(ldt_selector));
		table_base = segment_base(ldt_selector);
	}
134 135 136
	d = (struct desc_struct *)(table_base + (selector & ~7));
	v = d->base0 | ((unsigned long)d->base1 << 16) |
		((unsigned long)d->base2 << 24);
137
#ifdef CONFIG_X86_64
138 139
	if (d->s == 0 && (d->type == 2 || d->type == 9 || d->type == 11))
		v |= ((unsigned long)((struct ldttss_desc64 *)d)->base3) << 32;
140 141 142 143 144
#endif
	return v;
}
EXPORT_SYMBOL_GPL(segment_base);

145 146 147
u64 kvm_get_apic_base(struct kvm_vcpu *vcpu)
{
	if (irqchip_in_kernel(vcpu->kvm))
148
		return vcpu->arch.apic_base;
149
	else
150
		return vcpu->arch.apic_base;
151 152 153 154 155 156 157 158 159
}
EXPORT_SYMBOL_GPL(kvm_get_apic_base);

void kvm_set_apic_base(struct kvm_vcpu *vcpu, u64 data)
{
	/* TODO: reserve bits check */
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_lapic_set_base(vcpu, data);
	else
160
		vcpu->arch.apic_base = data;
161 162 163
}
EXPORT_SYMBOL_GPL(kvm_set_apic_base);

164 165
void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr)
{
166 167 168 169
	WARN_ON(vcpu->arch.exception.pending);
	vcpu->arch.exception.pending = true;
	vcpu->arch.exception.has_error_code = false;
	vcpu->arch.exception.nr = nr;
170 171 172
}
EXPORT_SYMBOL_GPL(kvm_queue_exception);

173 174 175 176
void kvm_inject_page_fault(struct kvm_vcpu *vcpu, unsigned long addr,
			   u32 error_code)
{
	++vcpu->stat.pf_guest;
177

J
Joerg Roedel 已提交
178 179 180 181 182 183 184 185 186 187
	if (vcpu->arch.exception.pending) {
		if (vcpu->arch.exception.nr == PF_VECTOR) {
			printk(KERN_DEBUG "kvm: inject_page_fault:"
					" double fault 0x%lx\n", addr);
			vcpu->arch.exception.nr = DF_VECTOR;
			vcpu->arch.exception.error_code = 0;
		} else if (vcpu->arch.exception.nr == DF_VECTOR) {
			/* triple fault -> shutdown */
			set_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests);
		}
188 189
		return;
	}
190
	vcpu->arch.cr2 = addr;
191 192 193
	kvm_queue_exception_e(vcpu, PF_VECTOR, error_code);
}

194 195 196 197 198 199
void kvm_inject_nmi(struct kvm_vcpu *vcpu)
{
	vcpu->arch.nmi_pending = 1;
}
EXPORT_SYMBOL_GPL(kvm_inject_nmi);

200 201
void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
{
202 203 204 205 206
	WARN_ON(vcpu->arch.exception.pending);
	vcpu->arch.exception.pending = true;
	vcpu->arch.exception.has_error_code = true;
	vcpu->arch.exception.nr = nr;
	vcpu->arch.exception.error_code = error_code;
207 208 209 210 211
}
EXPORT_SYMBOL_GPL(kvm_queue_exception_e);

static void __queue_exception(struct kvm_vcpu *vcpu)
{
212 213 214
	kvm_x86_ops->queue_exception(vcpu, vcpu->arch.exception.nr,
				     vcpu->arch.exception.has_error_code,
				     vcpu->arch.exception.error_code);
215 216
}

217 218 219 220 221 222 223 224 225
/*
 * Load the pae pdptrs.  Return true is they are all valid.
 */
int load_pdptrs(struct kvm_vcpu *vcpu, unsigned long cr3)
{
	gfn_t pdpt_gfn = cr3 >> PAGE_SHIFT;
	unsigned offset = ((cr3 & (PAGE_SIZE-1)) >> 5) << 2;
	int i;
	int ret;
226
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
227 228 229 230 231 232 233 234

	ret = kvm_read_guest_page(vcpu->kvm, pdpt_gfn, pdpte,
				  offset * sizeof(u64), sizeof(pdpte));
	if (ret < 0) {
		ret = 0;
		goto out;
	}
	for (i = 0; i < ARRAY_SIZE(pdpte); ++i) {
235 236
		if (is_present_pte(pdpte[i]) &&
		    (pdpte[i] & vcpu->arch.mmu.rsvd_bits_mask[0][2])) {
237 238 239 240 241 242
			ret = 0;
			goto out;
		}
	}
	ret = 1;

243
	memcpy(vcpu->arch.pdptrs, pdpte, sizeof(vcpu->arch.pdptrs));
244 245 246 247
out:

	return ret;
}
248
EXPORT_SYMBOL_GPL(load_pdptrs);
249

250 251
static bool pdptrs_changed(struct kvm_vcpu *vcpu)
{
252
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
253 254 255 256 257 258
	bool changed = true;
	int r;

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

259
	r = kvm_read_guest(vcpu->kvm, vcpu->arch.cr3 & ~31u, pdpte, sizeof(pdpte));
260 261
	if (r < 0)
		goto out;
262
	changed = memcmp(pdpte, vcpu->arch.pdptrs, sizeof(pdpte)) != 0;
263 264 265 266 267
out:

	return changed;
}

268
void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
269 270 271
{
	if (cr0 & CR0_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr0: 0x%lx #GP, reserved bits 0x%lx\n",
272
		       cr0, vcpu->arch.cr0);
273
		kvm_inject_gp(vcpu, 0);
274 275 276 277 278
		return;
	}

	if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD)) {
		printk(KERN_DEBUG "set_cr0: #GP, CD == 0 && NW == 1\n");
279
		kvm_inject_gp(vcpu, 0);
280 281 282 283 284 285
		return;
	}

	if ((cr0 & X86_CR0_PG) && !(cr0 & X86_CR0_PE)) {
		printk(KERN_DEBUG "set_cr0: #GP, set PG flag "
		       "and a clear PE flag\n");
286
		kvm_inject_gp(vcpu, 0);
287 288 289 290 291
		return;
	}

	if (!is_paging(vcpu) && (cr0 & X86_CR0_PG)) {
#ifdef CONFIG_X86_64
292
		if ((vcpu->arch.shadow_efer & EFER_LME)) {
293 294 295 296 297
			int cs_db, cs_l;

			if (!is_pae(vcpu)) {
				printk(KERN_DEBUG "set_cr0: #GP, start paging "
				       "in long mode while PAE is disabled\n");
298
				kvm_inject_gp(vcpu, 0);
299 300 301 302 303 304
				return;
			}
			kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
			if (cs_l) {
				printk(KERN_DEBUG "set_cr0: #GP, start paging "
				       "in long mode while CS.L == 1\n");
305
				kvm_inject_gp(vcpu, 0);
306 307 308 309 310
				return;

			}
		} else
#endif
311
		if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
312 313
			printk(KERN_DEBUG "set_cr0: #GP, pdptrs "
			       "reserved bits\n");
314
			kvm_inject_gp(vcpu, 0);
315 316 317 318 319 320
			return;
		}

	}

	kvm_x86_ops->set_cr0(vcpu, cr0);
321
	vcpu->arch.cr0 = cr0;
322 323 324 325

	kvm_mmu_reset_context(vcpu);
	return;
}
326
EXPORT_SYMBOL_GPL(kvm_set_cr0);
327

328
void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw)
329
{
330
	kvm_set_cr0(vcpu, (vcpu->arch.cr0 & ~0x0ful) | (msw & 0x0f));
F
Feng (Eric) Liu 已提交
331 332 333
	KVMTRACE_1D(LMSW, vcpu,
		    (u32)((vcpu->arch.cr0 & ~0x0ful) | (msw & 0x0f)),
		    handler);
334
}
335
EXPORT_SYMBOL_GPL(kvm_lmsw);
336

337
void kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
338
{
339 340 341
	unsigned long old_cr4 = vcpu->arch.cr4;
	unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE;

342 343
	if (cr4 & CR4_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr4: #GP, reserved bits\n");
344
		kvm_inject_gp(vcpu, 0);
345 346 347 348 349 350 351
		return;
	}

	if (is_long_mode(vcpu)) {
		if (!(cr4 & X86_CR4_PAE)) {
			printk(KERN_DEBUG "set_cr4: #GP, clearing PAE while "
			       "in long mode\n");
352
			kvm_inject_gp(vcpu, 0);
353 354
			return;
		}
355 356
	} else if (is_paging(vcpu) && (cr4 & X86_CR4_PAE)
		   && ((cr4 ^ old_cr4) & pdptr_bits)
357
		   && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
358
		printk(KERN_DEBUG "set_cr4: #GP, pdptrs reserved bits\n");
359
		kvm_inject_gp(vcpu, 0);
360 361 362 363 364
		return;
	}

	if (cr4 & X86_CR4_VMXE) {
		printk(KERN_DEBUG "set_cr4: #GP, setting VMXE\n");
365
		kvm_inject_gp(vcpu, 0);
366 367 368
		return;
	}
	kvm_x86_ops->set_cr4(vcpu, cr4);
369
	vcpu->arch.cr4 = cr4;
370
	vcpu->arch.mmu.base_role.cr4_pge = (cr4 & X86_CR4_PGE) && !tdp_enabled;
371 372
	kvm_mmu_reset_context(vcpu);
}
373
EXPORT_SYMBOL_GPL(kvm_set_cr4);
374

375
void kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
376
{
377
	if (cr3 == vcpu->arch.cr3 && !pdptrs_changed(vcpu)) {
378
		kvm_mmu_sync_roots(vcpu);
379 380 381 382
		kvm_mmu_flush_tlb(vcpu);
		return;
	}

383 384 385
	if (is_long_mode(vcpu)) {
		if (cr3 & CR3_L_MODE_RESERVED_BITS) {
			printk(KERN_DEBUG "set_cr3: #GP, reserved bits\n");
386
			kvm_inject_gp(vcpu, 0);
387 388 389 390 391 392 393
			return;
		}
	} else {
		if (is_pae(vcpu)) {
			if (cr3 & CR3_PAE_RESERVED_BITS) {
				printk(KERN_DEBUG
				       "set_cr3: #GP, reserved bits\n");
394
				kvm_inject_gp(vcpu, 0);
395 396 397 398 399
				return;
			}
			if (is_paging(vcpu) && !load_pdptrs(vcpu, cr3)) {
				printk(KERN_DEBUG "set_cr3: #GP, pdptrs "
				       "reserved bits\n");
400
				kvm_inject_gp(vcpu, 0);
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
				return;
			}
		}
		/*
		 * We don't check reserved bits in nonpae mode, because
		 * this isn't enforced, and VMware depends on this.
		 */
	}

	/*
	 * Does the new cr3 value map to physical memory? (Note, we
	 * catch an invalid cr3 even in real-mode, because it would
	 * cause trouble later on when we turn on paging anyway.)
	 *
	 * A real CPU would silently accept an invalid cr3 and would
	 * attempt to use it - with largely undefined (and often hard
	 * to debug) behavior on the guest side.
	 */
	if (unlikely(!gfn_to_memslot(vcpu->kvm, cr3 >> PAGE_SHIFT)))
420
		kvm_inject_gp(vcpu, 0);
421
	else {
422 423
		vcpu->arch.cr3 = cr3;
		vcpu->arch.mmu.new_cr3(vcpu);
424 425
	}
}
426
EXPORT_SYMBOL_GPL(kvm_set_cr3);
427

428
void kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
429 430 431
{
	if (cr8 & CR8_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr8: #GP, reserved bits 0x%lx\n", cr8);
432
		kvm_inject_gp(vcpu, 0);
433 434 435 436 437
		return;
	}
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_lapic_set_tpr(vcpu, cr8);
	else
438
		vcpu->arch.cr8 = cr8;
439
}
440
EXPORT_SYMBOL_GPL(kvm_set_cr8);
441

442
unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu)
443 444 445 446
{
	if (irqchip_in_kernel(vcpu->kvm))
		return kvm_lapic_get_cr8(vcpu);
	else
447
		return vcpu->arch.cr8;
448
}
449
EXPORT_SYMBOL_GPL(kvm_get_cr8);
450

451 452 453 454 455
static inline u32 bit(int bitno)
{
	return 1 << (bitno & 31);
}

456 457 458 459 460 461 462 463 464 465 466 467 468
/*
 * List of msr numbers which we expose to userspace through KVM_GET_MSRS
 * and KVM_SET_MSRS, and KVM_GET_MSR_INDEX_LIST.
 *
 * This list is modified at module load time to reflect the
 * capabilities of the host cpu.
 */
static u32 msrs_to_save[] = {
	MSR_IA32_SYSENTER_CS, MSR_IA32_SYSENTER_ESP, MSR_IA32_SYSENTER_EIP,
	MSR_K6_STAR,
#ifdef CONFIG_X86_64
	MSR_CSTAR, MSR_KERNEL_GS_BASE, MSR_SYSCALL_MASK, MSR_LSTAR,
#endif
469
	MSR_IA32_TIME_STAMP_COUNTER, MSR_KVM_SYSTEM_TIME, MSR_KVM_WALL_CLOCK,
A
Alexander Graf 已提交
470
	MSR_IA32_PERF_STATUS, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA
471 472 473 474 475 476 477 478
};

static unsigned num_msrs_to_save;

static u32 emulated_msrs[] = {
	MSR_IA32_MISC_ENABLE,
};

479 480
static void set_efer(struct kvm_vcpu *vcpu, u64 efer)
{
481
	if (efer & efer_reserved_bits) {
482 483
		printk(KERN_DEBUG "set_efer: 0x%llx #GP, reserved bits\n",
		       efer);
484
		kvm_inject_gp(vcpu, 0);
485 486 487 488
		return;
	}

	if (is_paging(vcpu)
489
	    && (vcpu->arch.shadow_efer & EFER_LME) != (efer & EFER_LME)) {
490
		printk(KERN_DEBUG "set_efer: #GP, change LME while paging\n");
491
		kvm_inject_gp(vcpu, 0);
492 493 494
		return;
	}

A
Alexander Graf 已提交
495 496 497 498 499 500 501 502 503 504 505
	if (efer & EFER_FFXSR) {
		struct kvm_cpuid_entry2 *feat;

		feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
		if (!feat || !(feat->edx & bit(X86_FEATURE_FXSR_OPT))) {
			printk(KERN_DEBUG "set_efer: #GP, enable FFXSR w/o CPUID capability\n");
			kvm_inject_gp(vcpu, 0);
			return;
		}
	}

506 507 508 509 510 511 512 513 514 515 516
	if (efer & EFER_SVME) {
		struct kvm_cpuid_entry2 *feat;

		feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
		if (!feat || !(feat->ecx & bit(X86_FEATURE_SVM))) {
			printk(KERN_DEBUG "set_efer: #GP, enable SVM w/o SVM\n");
			kvm_inject_gp(vcpu, 0);
			return;
		}
	}

517 518 519
	kvm_x86_ops->set_efer(vcpu, efer);

	efer &= ~EFER_LMA;
520
	efer |= vcpu->arch.shadow_efer & EFER_LMA;
521

522
	vcpu->arch.shadow_efer = efer;
523 524 525

	vcpu->arch.mmu.base_role.nxe = (efer & EFER_NX) && !tdp_enabled;
	kvm_mmu_reset_context(vcpu);
526 527
}

528 529 530 531 532 533 534
void kvm_enable_efer_bits(u64 mask)
{
       efer_reserved_bits &= ~mask;
}
EXPORT_SYMBOL_GPL(kvm_enable_efer_bits);


535 536 537 538 539 540 541 542 543 544
/*
 * Writes msr value into into the appropriate "register".
 * Returns 0 on success, non-0 otherwise.
 * Assumes vcpu_load() was already called.
 */
int kvm_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data)
{
	return kvm_x86_ops->set_msr(vcpu, msr_index, data);
}

545 546 547 548 549 550 551 552
/*
 * Adapt set_msr() to msr_io()'s calling convention
 */
static int do_set_msr(struct kvm_vcpu *vcpu, unsigned index, u64 *data)
{
	return kvm_set_msr(vcpu, index, *data);
}

553 554 555
static void kvm_write_wall_clock(struct kvm *kvm, gpa_t wall_clock)
{
	static int version;
556 557
	struct pvclock_wall_clock wc;
	struct timespec now, sys, boot;
558 559 560 561 562 563 564 565

	if (!wall_clock)
		return;

	version++;

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

566 567 568 569 570 571 572 573 574 575 576 577 578
	/*
	 * The guest calculates current wall clock time by adding
	 * system time (updated by kvm_write_guest_time below) to the
	 * wall clock specified here.  guest system time equals host
	 * system time for us, thus we must fill in host boot time here.
	 */
	now = current_kernel_time();
	ktime_get_ts(&sys);
	boot = ns_to_timespec(timespec_to_ns(&now) - timespec_to_ns(&sys));

	wc.sec = boot.tv_sec;
	wc.nsec = boot.tv_nsec;
	wc.version = version;
579 580 581 582 583 584 585

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

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

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
static uint32_t div_frac(uint32_t dividend, uint32_t divisor)
{
	uint32_t quotient, remainder;

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

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

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

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

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

	pr_debug("%s: tsc_khz %u, tsc_shift %d, tsc_mul %u\n",
621
		 __func__, tsc_khz, hv_clock->tsc_shift,
622 623 624
		 hv_clock->tsc_to_system_mul);
}

625 626
static DEFINE_PER_CPU(unsigned long, cpu_tsc_khz);

627 628 629 630 631 632
static void kvm_write_guest_time(struct kvm_vcpu *v)
{
	struct timespec ts;
	unsigned long flags;
	struct kvm_vcpu_arch *vcpu = &v->arch;
	void *shared_kaddr;
633
	unsigned long this_tsc_khz;
634 635 636 637

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

638 639 640 641
	this_tsc_khz = get_cpu_var(cpu_tsc_khz);
	if (unlikely(vcpu->hv_clock_tsc_khz != this_tsc_khz)) {
		kvm_set_time_scale(this_tsc_khz, &vcpu->hv_clock);
		vcpu->hv_clock_tsc_khz = this_tsc_khz;
642
	}
643
	put_cpu_var(cpu_tsc_khz);
644

645 646 647 648 649 650 651 652 653 654 655 656 657 658
	/* Keep irq disabled to prevent changes to the clock */
	local_irq_save(flags);
	kvm_get_msr(v, MSR_IA32_TIME_STAMP_COUNTER,
			  &vcpu->hv_clock.tsc_timestamp);
	ktime_get_ts(&ts);
	local_irq_restore(flags);

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

	vcpu->hv_clock.system_time = ts.tv_nsec +
				     (NSEC_PER_SEC * (u64)ts.tv_sec);
	/*
	 * The interface expects us to write an even number signaling that the
	 * update is finished. Since the guest won't see the intermediate
659
	 * state, we just increase by 2 at the end.
660
	 */
661
	vcpu->hv_clock.version += 2;
662 663 664 665

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

	memcpy(shared_kaddr + vcpu->time_offset, &vcpu->hv_clock,
666
	       sizeof(vcpu->hv_clock));
667 668 669 670 671 672

	kunmap_atomic(shared_kaddr, KM_USER0);

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

673 674 675 676 677 678 679 680 681 682
static int kvm_request_guest_time_update(struct kvm_vcpu *v)
{
	struct kvm_vcpu_arch *vcpu = &v->arch;

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

A
Avi Kivity 已提交
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708
static bool msr_mtrr_valid(unsigned msr)
{
	switch (msr) {
	case 0x200 ... 0x200 + 2 * KVM_NR_VAR_MTRR - 1:
	case MSR_MTRRfix64K_00000:
	case MSR_MTRRfix16K_80000:
	case MSR_MTRRfix16K_A0000:
	case MSR_MTRRfix4K_C0000:
	case MSR_MTRRfix4K_C8000:
	case MSR_MTRRfix4K_D0000:
	case MSR_MTRRfix4K_D8000:
	case MSR_MTRRfix4K_E0000:
	case MSR_MTRRfix4K_E8000:
	case MSR_MTRRfix4K_F0000:
	case MSR_MTRRfix4K_F8000:
	case MSR_MTRRdefType:
	case MSR_IA32_CR_PAT:
		return true;
	case 0x2f8:
		return true;
	}
	return false;
}

static int set_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
S
Sheng Yang 已提交
709 710
	u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges;

A
Avi Kivity 已提交
711 712 713
	if (!msr_mtrr_valid(msr))
		return 1;

S
Sheng Yang 已提交
714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
	if (msr == MSR_MTRRdefType) {
		vcpu->arch.mtrr_state.def_type = data;
		vcpu->arch.mtrr_state.enabled = (data & 0xc00) >> 10;
	} else if (msr == MSR_MTRRfix64K_00000)
		p[0] = data;
	else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000)
		p[1 + msr - MSR_MTRRfix16K_80000] = data;
	else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000)
		p[3 + msr - MSR_MTRRfix4K_C0000] = data;
	else if (msr == MSR_IA32_CR_PAT)
		vcpu->arch.pat = data;
	else {	/* Variable MTRRs */
		int idx, is_mtrr_mask;
		u64 *pt;

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

	kvm_mmu_reset_context(vcpu);
A
Avi Kivity 已提交
741 742
	return 0;
}
743 744 745 746 747 748 749 750 751

int kvm_set_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 data)
{
	switch (msr) {
	case MSR_EFER:
		set_efer(vcpu, data);
		break;
	case MSR_IA32_MC0_STATUS:
		pr_unimpl(vcpu, "%s: MSR_IA32_MC0_STATUS 0x%llx, nop\n",
752
		       __func__, data);
753 754 755
		break;
	case MSR_IA32_MCG_STATUS:
		pr_unimpl(vcpu, "%s: MSR_IA32_MCG_STATUS 0x%llx, nop\n",
756
			__func__, data);
757
		break;
758 759
	case MSR_IA32_MCG_CTL:
		pr_unimpl(vcpu, "%s: MSR_IA32_MCG_CTL 0x%llx, nop\n",
760
			__func__, data);
761
		break;
762 763 764 765 766 767 768 769 770 771 772 773
	case MSR_IA32_DEBUGCTLMSR:
		if (!data) {
			/* We support the non-activated case already */
			break;
		} else if (data & ~(DEBUGCTLMSR_LBR | DEBUGCTLMSR_BTF)) {
			/* Values other than LBR and BTF are vendor-specific,
			   thus reserved and should throw a #GP */
			return 1;
		}
		pr_unimpl(vcpu, "%s: MSR_IA32_DEBUGCTLMSR 0x%llx, nop\n",
			__func__, data);
		break;
774 775
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_UCODE_WRITE:
776
	case MSR_VM_HSAVE_PA:
777
		break;
A
Avi Kivity 已提交
778 779
	case 0x200 ... 0x2ff:
		return set_msr_mtrr(vcpu, msr, data);
780 781 782 783
	case MSR_IA32_APICBASE:
		kvm_set_apic_base(vcpu, data);
		break;
	case MSR_IA32_MISC_ENABLE:
784
		vcpu->arch.ia32_misc_enable_msr = data;
785
		break;
786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
	case MSR_KVM_WALL_CLOCK:
		vcpu->kvm->arch.wall_clock = data;
		kvm_write_wall_clock(vcpu->kvm, data);
		break;
	case MSR_KVM_SYSTEM_TIME: {
		if (vcpu->arch.time_page) {
			kvm_release_page_dirty(vcpu->arch.time_page);
			vcpu->arch.time_page = NULL;
		}

		vcpu->arch.time = data;

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

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

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

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

813
		kvm_request_guest_time_update(vcpu);
814 815
		break;
	}
816
	default:
817
		pr_unimpl(vcpu, "unhandled wrmsr: 0x%x data %llx\n", msr, data);
818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
		return 1;
	}
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_set_msr_common);


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

A
Avi Kivity 已提交
835 836
static int get_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
S
Sheng Yang 已提交
837 838
	u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges;

A
Avi Kivity 已提交
839 840 841
	if (!msr_mtrr_valid(msr))
		return 1;

S
Sheng Yang 已提交
842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867
	if (msr == MSR_MTRRdefType)
		*pdata = vcpu->arch.mtrr_state.def_type +
			 (vcpu->arch.mtrr_state.enabled << 10);
	else if (msr == MSR_MTRRfix64K_00000)
		*pdata = p[0];
	else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000)
		*pdata = p[1 + msr - MSR_MTRRfix16K_80000];
	else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000)
		*pdata = p[3 + msr - MSR_MTRRfix4K_C0000];
	else if (msr == MSR_IA32_CR_PAT)
		*pdata = vcpu->arch.pat;
	else {	/* Variable MTRRs */
		int idx, is_mtrr_mask;
		u64 *pt;

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

A
Avi Kivity 已提交
868 869 870
	return 0;
}

871 872 873 874 875 876 877 878 879 880 881 882 883
int kvm_get_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
	u64 data;

	switch (msr) {
	case 0xc0010010: /* SYSCFG */
	case 0xc0010015: /* HWCR */
	case MSR_IA32_PLATFORM_ID:
	case MSR_IA32_P5_MC_ADDR:
	case MSR_IA32_P5_MC_TYPE:
	case MSR_IA32_MC0_CTL:
	case MSR_IA32_MCG_STATUS:
	case MSR_IA32_MCG_CAP:
884
	case MSR_IA32_MCG_CTL:
885 886 887 888 889
	case MSR_IA32_MC0_MISC:
	case MSR_IA32_MC0_MISC+4:
	case MSR_IA32_MC0_MISC+8:
	case MSR_IA32_MC0_MISC+12:
	case MSR_IA32_MC0_MISC+16:
J
Joerg Roedel 已提交
890
	case MSR_IA32_MC0_MISC+20:
891 892
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_EBL_CR_POWERON:
893 894 895 896 897
	case MSR_IA32_DEBUGCTLMSR:
	case MSR_IA32_LASTBRANCHFROMIP:
	case MSR_IA32_LASTBRANCHTOIP:
	case MSR_IA32_LASTINTFROMIP:
	case MSR_IA32_LASTINTTOIP:
898
	case MSR_VM_HSAVE_PA:
899 900
	case MSR_P6_EVNTSEL0:
	case MSR_P6_EVNTSEL1:
901 902
		data = 0;
		break;
A
Avi Kivity 已提交
903 904 905 906 907
	case MSR_MTRRcap:
		data = 0x500 | KVM_NR_VAR_MTRR;
		break;
	case 0x200 ... 0x2ff:
		return get_msr_mtrr(vcpu, msr, pdata);
908 909 910 911 912 913 914
	case 0xcd: /* fsb frequency */
		data = 3;
		break;
	case MSR_IA32_APICBASE:
		data = kvm_get_apic_base(vcpu);
		break;
	case MSR_IA32_MISC_ENABLE:
915
		data = vcpu->arch.ia32_misc_enable_msr;
916
		break;
917 918 919 920 921 922
	case MSR_IA32_PERF_STATUS:
		/* TSC increment by tick */
		data = 1000ULL;
		/* CPU multiplier */
		data |= (((uint64_t)4ULL) << 40);
		break;
923
	case MSR_EFER:
924
		data = vcpu->arch.shadow_efer;
925
		break;
926 927 928 929 930 931
	case MSR_KVM_WALL_CLOCK:
		data = vcpu->kvm->arch.wall_clock;
		break;
	case MSR_KVM_SYSTEM_TIME:
		data = vcpu->arch.time;
		break;
932 933 934 935 936 937 938 939 940
	default:
		pr_unimpl(vcpu, "unhandled rdmsr: 0x%x\n", msr);
		return 1;
	}
	*pdata = data;
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_get_msr_common);

941 942 943 944 945 946 947 948 949 950 951 952 953 954
/*
 * Read or write a bunch of msrs. All parameters are kernel addresses.
 *
 * @return number of msrs set successfully.
 */
static int __msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs *msrs,
		    struct kvm_msr_entry *entries,
		    int (*do_msr)(struct kvm_vcpu *vcpu,
				  unsigned index, u64 *data))
{
	int i;

	vcpu_load(vcpu);

955
	down_read(&vcpu->kvm->slots_lock);
956 957 958
	for (i = 0; i < msrs->nmsrs; ++i)
		if (do_msr(vcpu, entries[i].index, &entries[i].data))
			break;
959
	up_read(&vcpu->kvm->slots_lock);
960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014

	vcpu_put(vcpu);

	return i;
}

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

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

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

	r = -ENOMEM;
	size = sizeof(struct kvm_msr_entry) * msrs.nmsrs;
	entries = vmalloc(size);
	if (!entries)
		goto out;

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

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

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

	r = n;

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

1015 1016 1017 1018 1019 1020 1021 1022 1023
int kvm_dev_ioctl_check_extension(long ext)
{
	int r;

	switch (ext) {
	case KVM_CAP_IRQCHIP:
	case KVM_CAP_HLT:
	case KVM_CAP_MMU_SHADOW_CACHE_CONTROL:
	case KVM_CAP_SET_TSS_ADDR:
1024
	case KVM_CAP_EXT_CPUID:
1025
	case KVM_CAP_CLOCKSOURCE:
S
Sheng Yang 已提交
1026
	case KVM_CAP_PIT:
1027
	case KVM_CAP_NOP_IO_DELAY:
1028
	case KVM_CAP_MP_STATE:
1029
	case KVM_CAP_SYNC_MMU:
1030
	case KVM_CAP_REINJECT_CONTROL:
1031
	case KVM_CAP_IRQ_INJECT_STATUS:
1032
	case KVM_CAP_ASSIGN_DEV_IRQ:
1033 1034
		r = 1;
		break;
1035 1036 1037
	case KVM_CAP_COALESCED_MMIO:
		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
		break;
1038 1039 1040
	case KVM_CAP_VAPIC:
		r = !kvm_x86_ops->cpu_has_accelerated_tpr();
		break;
1041 1042 1043
	case KVM_CAP_NR_VCPUS:
		r = KVM_MAX_VCPUS;
		break;
1044 1045 1046
	case KVM_CAP_NR_MEMSLOTS:
		r = KVM_MEMORY_SLOTS;
		break;
1047 1048 1049
	case KVM_CAP_PV_MMU:
		r = !tdp_enabled;
		break;
B
Ben-Ami Yassour 已提交
1050
	case KVM_CAP_IOMMU:
J
Joerg Roedel 已提交
1051
		r = iommu_found();
B
Ben-Ami Yassour 已提交
1052
		break;
1053 1054 1055 1056 1057 1058 1059 1060
	default:
		r = 0;
		break;
	}
	return r;

}

1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
long kvm_arch_dev_ioctl(struct file *filp,
			unsigned int ioctl, unsigned long arg)
{
	void __user *argp = (void __user *)arg;
	long r;

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

		r = -EFAULT;
		if (copy_from_user(&msr_list, user_msr_list, sizeof msr_list))
			goto out;
		n = msr_list.nmsrs;
		msr_list.nmsrs = num_msrs_to_save + ARRAY_SIZE(emulated_msrs);
		if (copy_to_user(user_msr_list, &msr_list, sizeof msr_list))
			goto out;
		r = -E2BIG;
		if (n < num_msrs_to_save)
			goto out;
		r = -EFAULT;
		if (copy_to_user(user_msr_list->indices, &msrs_to_save,
				 num_msrs_to_save * sizeof(u32)))
			goto out;
		if (copy_to_user(user_msr_list->indices
				 + num_msrs_to_save * sizeof(u32),
				 &emulated_msrs,
				 ARRAY_SIZE(emulated_msrs) * sizeof(u32)))
			goto out;
		r = 0;
		break;
	}
1095 1096 1097 1098 1099 1100 1101 1102
	case KVM_GET_SUPPORTED_CPUID: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_dev_ioctl_get_supported_cpuid(&cpuid,
1103
						      cpuid_arg->entries);
1104 1105 1106 1107 1108 1109 1110 1111 1112
		if (r)
			goto out;

		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
1113 1114 1115 1116 1117 1118 1119
	default:
		r = -EINVAL;
	}
out:
	return r;
}

1120 1121 1122
void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
{
	kvm_x86_ops->vcpu_load(vcpu, cpu);
1123
	kvm_request_guest_time_update(vcpu);
1124 1125 1126 1127 1128
}

void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
	kvm_x86_ops->vcpu_put(vcpu);
1129
	kvm_put_guest_fpu(vcpu);
1130 1131
}

1132
static int is_efer_nx(void)
1133
{
1134
	unsigned long long efer = 0;
1135

1136
	rdmsrl_safe(MSR_EFER, &efer);
1137 1138 1139 1140 1141 1142 1143 1144
	return efer & EFER_NX;
}

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

1145
	entry = NULL;
1146 1147
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		e = &vcpu->arch.cpuid_entries[i];
1148 1149 1150 1151 1152
		if (e->function == 0x80000001) {
			entry = e;
			break;
		}
	}
1153
	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
1154 1155 1156 1157 1158
		entry->edx &= ~(1 << 20);
		printk(KERN_INFO "kvm: guest NX capability removed\n");
	}
}

1159
/* when an old userspace process fills a new kernel module */
1160 1161 1162
static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
				    struct kvm_cpuid *cpuid,
				    struct kvm_cpuid_entry __user *entries)
1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
{
	int r, i;
	struct kvm_cpuid_entry *cpuid_entries;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry) * cpuid->nent);
	if (!cpuid_entries)
		goto out;
	r = -EFAULT;
	if (copy_from_user(cpuid_entries, entries,
			   cpuid->nent * sizeof(struct kvm_cpuid_entry)))
		goto out_free;
	for (i = 0; i < cpuid->nent; i++) {
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
		vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
		vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
		vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
		vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
		vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
		vcpu->arch.cpuid_entries[i].index = 0;
		vcpu->arch.cpuid_entries[i].flags = 0;
		vcpu->arch.cpuid_entries[i].padding[0] = 0;
		vcpu->arch.cpuid_entries[i].padding[1] = 0;
		vcpu->arch.cpuid_entries[i].padding[2] = 0;
	}
	vcpu->arch.cpuid_nent = cpuid->nent;
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
	cpuid_fix_nx_cap(vcpu);
	r = 0;

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

static int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
1201 1202
				     struct kvm_cpuid2 *cpuid,
				     struct kvm_cpuid_entry2 __user *entries)
1203 1204 1205 1206 1207 1208 1209
{
	int r;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -EFAULT;
1210
	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
1211
			   cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
1212
		goto out;
1213
	vcpu->arch.cpuid_nent = cpuid->nent;
1214 1215 1216 1217 1218 1219
	return 0;

out:
	return r;
}

1220
static int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
1221 1222
				     struct kvm_cpuid2 *cpuid,
				     struct kvm_cpuid_entry2 __user *entries)
1223 1224 1225 1226
{
	int r;

	r = -E2BIG;
1227
	if (cpuid->nent < vcpu->arch.cpuid_nent)
1228 1229
		goto out;
	r = -EFAULT;
1230
	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
1231
			 vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
1232 1233 1234 1235
		goto out;
	return 0;

out:
1236
	cpuid->nent = vcpu->arch.cpuid_nent;
1237 1238 1239 1240
	return r;
}

static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
1241
			   u32 index)
1242 1243 1244 1245
{
	entry->function = function;
	entry->index = index;
	cpuid_count(entry->function, entry->index,
1246
		    &entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
1247 1248 1249
	entry->flags = 0;
}

1250 1251
#define F(x) bit(X86_FEATURE_##x)

1252 1253 1254
static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			 u32 index, int *nent, int maxnent)
{
1255
	unsigned f_nx = is_efer_nx() ? F(NX) : 0;
1256
#ifdef CONFIG_X86_64
1257 1258 1259
	unsigned f_lm = F(LM);
#else
	unsigned f_lm = 0;
1260
#endif
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283

	/* cpuid 1.edx */
	const u32 kvm_supported_word0_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SEP) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* PSN */ | F(CLFLSH) |
		0 /* Reserved, DS, ACPI */ | F(MMX) |
		F(FXSR) | F(XMM) | F(XMM2) | F(SELFSNOOP) |
		0 /* HTT, TM, Reserved, PBE */;
	/* cpuid 0x80000001.edx */
	const u32 kvm_supported_word1_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SYSCALL) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* Reserved */ |
		f_nx | 0 /* Reserved */ | F(MMXEXT) | F(MMX) |
		F(FXSR) | F(FXSR_OPT) | 0 /* GBPAGES */ | 0 /* RDTSCP */ |
		0 /* Reserved */ | f_lm | F(3DNOWEXT) | F(3DNOW);
	/* cpuid 1.ecx */
	const u32 kvm_supported_word4_x86_features =
A
Avi Kivity 已提交
1284 1285 1286 1287 1288 1289 1290
		F(XMM3) | 0 /* Reserved, DTES64, MONITOR */ |
		0 /* DS-CPL, VMX, SMX, EST */ |
		0 /* TM2 */ | F(SSSE3) | 0 /* CNXT-ID */ | 0 /* Reserved */ |
		0 /* Reserved */ | F(CX16) | 0 /* xTPR Update, PDCM */ |
		0 /* Reserved, DCA */ | F(XMM4_1) |
		F(XMM4_2) | 0 /* x2APIC */ | F(MOVBE) | F(POPCNT) |
		0 /* Reserved, XSAVE, OSXSAVE */;
1291
	/* cpuid 0x80000001.ecx */
1292
	const u32 kvm_supported_word6_x86_features =
1293 1294 1295 1296
		F(LAHF_LM) | F(CMP_LEGACY) | F(SVM) | 0 /* ExtApicSpace */ |
		F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) |
		F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(SSE5) |
		0 /* SKINIT */ | 0 /* WDT */;
1297

1298
	/* all calls to cpuid_count() should be made on the same cpu */
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
	get_cpu();
	do_cpuid_1_ent(entry, function, index);
	++*nent;

	switch (function) {
	case 0:
		entry->eax = min(entry->eax, (u32)0xb);
		break;
	case 1:
		entry->edx &= kvm_supported_word0_x86_features;
1309
		entry->ecx &= kvm_supported_word4_x86_features;
1310 1311 1312 1313 1314 1315 1316 1317 1318
		break;
	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	 * may return different values. This forces us to get_cpu() before
	 * issuing the first command, and also to emulate this annoying behavior
	 * in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	case 2: {
		int t, times = entry->eax & 0xff;

		entry->flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
1319
		entry->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
1320 1321 1322 1323 1324 1325 1326 1327 1328
		for (t = 1; t < times && *nent < maxnent; ++t) {
			do_cpuid_1_ent(&entry[t], function, 0);
			entry[t].flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
			++*nent;
		}
		break;
	}
	/* function 4 and 0xb have additional index. */
	case 4: {
1329
		int i, cache_type;
1330 1331 1332

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until cache_type is zero */
1333 1334
		for (i = 1; *nent < maxnent; ++i) {
			cache_type = entry[i - 1].eax & 0x1f;
1335 1336
			if (!cache_type)
				break;
1337 1338
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1339 1340 1341 1342 1343 1344
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 0xb: {
1345
		int i, level_type;
1346 1347 1348

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until level_type is zero */
1349
		for (i = 1; *nent < maxnent; ++i) {
1350
			level_type = entry[i - 1].ecx & 0xff00;
1351 1352
			if (!level_type)
				break;
1353 1354
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 0x80000000:
		entry->eax = min(entry->eax, 0x8000001a);
		break;
	case 0x80000001:
		entry->edx &= kvm_supported_word1_x86_features;
		entry->ecx &= kvm_supported_word6_x86_features;
		break;
	}
	put_cpu();
}

1371 1372
#undef F

1373
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
1374
				     struct kvm_cpuid_entry2 __user *entries)
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390
{
	struct kvm_cpuid_entry2 *cpuid_entries;
	int limit, nent = 0, r = -E2BIG;
	u32 func;

	if (cpuid->nent < 1)
		goto out;
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	if (!cpuid_entries)
		goto out;

	do_cpuid_ent(&cpuid_entries[0], 0, 0, &nent, cpuid->nent);
	limit = cpuid_entries[0].eax;
	for (func = 1; func <= limit && nent < cpuid->nent; ++func)
		do_cpuid_ent(&cpuid_entries[nent], func, 0,
1391
			     &nent, cpuid->nent);
1392 1393 1394 1395 1396 1397 1398 1399
	r = -E2BIG;
	if (nent >= cpuid->nent)
		goto out_free;

	do_cpuid_ent(&cpuid_entries[nent], 0x80000000, 0, &nent, cpuid->nent);
	limit = cpuid_entries[nent - 1].eax;
	for (func = 0x80000001; func <= limit && nent < cpuid->nent; ++func)
		do_cpuid_ent(&cpuid_entries[nent], func, 0,
1400
			     &nent, cpuid->nent);
1401 1402
	r = -EFAULT;
	if (copy_to_user(entries, cpuid_entries,
1403
			 nent * sizeof(struct kvm_cpuid_entry2)))
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
		goto out_free;
	cpuid->nent = nent;
	r = 0;

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

1414 1415 1416 1417
static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
1418
	memcpy(s->regs, vcpu->arch.apic->regs, sizeof *s);
1419 1420 1421 1422 1423 1424 1425 1426 1427
	vcpu_put(vcpu);

	return 0;
}

static int kvm_vcpu_ioctl_set_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
1428
	memcpy(vcpu->arch.apic->regs, s->regs, sizeof *s);
1429 1430 1431 1432 1433 1434
	kvm_apic_post_state_restore(vcpu);
	vcpu_put(vcpu);

	return 0;
}

1435 1436 1437 1438 1439 1440 1441 1442 1443
static int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu,
				    struct kvm_interrupt *irq)
{
	if (irq->irq < 0 || irq->irq >= 256)
		return -EINVAL;
	if (irqchip_in_kernel(vcpu->kvm))
		return -ENXIO;
	vcpu_load(vcpu);

1444
	kvm_queue_interrupt(vcpu, irq->irq, false);
1445 1446 1447 1448 1449 1450

	vcpu_put(vcpu);

	return 0;
}

1451 1452 1453 1454 1455 1456 1457 1458 1459
static int kvm_vcpu_ioctl_nmi(struct kvm_vcpu *vcpu)
{
	vcpu_load(vcpu);
	kvm_inject_nmi(vcpu);
	vcpu_put(vcpu);

	return 0;
}

1460 1461 1462 1463 1464 1465 1466 1467 1468
static int vcpu_ioctl_tpr_access_reporting(struct kvm_vcpu *vcpu,
					   struct kvm_tpr_access_ctl *tac)
{
	if (tac->flags)
		return -EINVAL;
	vcpu->arch.tpr_access_reporting = !!tac->enabled;
	return 0;
}

1469 1470 1471 1472 1473 1474
long kvm_arch_vcpu_ioctl(struct file *filp,
			 unsigned int ioctl, unsigned long arg)
{
	struct kvm_vcpu *vcpu = filp->private_data;
	void __user *argp = (void __user *)arg;
	int r;
1475
	struct kvm_lapic_state *lapic = NULL;
1476 1477 1478

	switch (ioctl) {
	case KVM_GET_LAPIC: {
1479
		lapic = kzalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
1480

1481 1482 1483 1484
		r = -ENOMEM;
		if (!lapic)
			goto out;
		r = kvm_vcpu_ioctl_get_lapic(vcpu, lapic);
1485 1486 1487
		if (r)
			goto out;
		r = -EFAULT;
1488
		if (copy_to_user(argp, lapic, sizeof(struct kvm_lapic_state)))
1489 1490 1491 1492 1493
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_LAPIC: {
1494 1495 1496 1497
		lapic = kmalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
		r = -ENOMEM;
		if (!lapic)
			goto out;
1498
		r = -EFAULT;
1499
		if (copy_from_user(lapic, argp, sizeof(struct kvm_lapic_state)))
1500
			goto out;
1501
		r = kvm_vcpu_ioctl_set_lapic(vcpu, lapic);
1502 1503 1504 1505 1506
		if (r)
			goto out;
		r = 0;
		break;
	}
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
	case KVM_INTERRUPT: {
		struct kvm_interrupt irq;

		r = -EFAULT;
		if (copy_from_user(&irq, argp, sizeof irq))
			goto out;
		r = kvm_vcpu_ioctl_interrupt(vcpu, &irq);
		if (r)
			goto out;
		r = 0;
		break;
	}
1519 1520 1521 1522 1523 1524 1525
	case KVM_NMI: {
		r = kvm_vcpu_ioctl_nmi(vcpu);
		if (r)
			goto out;
		r = 0;
		break;
	}
1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
	case KVM_SET_CPUID: {
		struct kvm_cpuid __user *cpuid_arg = argp;
		struct kvm_cpuid cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_set_cpuid(vcpu, &cpuid, cpuid_arg->entries);
		if (r)
			goto out;
		break;
	}
1538 1539 1540 1541 1542 1543 1544 1545
	case KVM_SET_CPUID2: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_set_cpuid2(vcpu, &cpuid,
1546
					      cpuid_arg->entries);
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
		if (r)
			goto out;
		break;
	}
	case KVM_GET_CPUID2: {
		struct kvm_cpuid2 __user *cpuid_arg = argp;
		struct kvm_cpuid2 cpuid;

		r = -EFAULT;
		if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
			goto out;
		r = kvm_vcpu_ioctl_get_cpuid2(vcpu, &cpuid,
1559
					      cpuid_arg->entries);
1560 1561 1562 1563 1564 1565 1566 1567
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
1568 1569 1570 1571 1572 1573
	case KVM_GET_MSRS:
		r = msr_io(vcpu, argp, kvm_get_msr, 1);
		break;
	case KVM_SET_MSRS:
		r = msr_io(vcpu, argp, do_set_msr, 0);
		break;
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
	case KVM_TPR_ACCESS_REPORTING: {
		struct kvm_tpr_access_ctl tac;

		r = -EFAULT;
		if (copy_from_user(&tac, argp, sizeof tac))
			goto out;
		r = vcpu_ioctl_tpr_access_reporting(vcpu, &tac);
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(argp, &tac, sizeof tac))
			goto out;
		r = 0;
		break;
	};
A
Avi Kivity 已提交
1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601
	case KVM_SET_VAPIC_ADDR: {
		struct kvm_vapic_addr va;

		r = -EINVAL;
		if (!irqchip_in_kernel(vcpu->kvm))
			goto out;
		r = -EFAULT;
		if (copy_from_user(&va, argp, sizeof va))
			goto out;
		r = 0;
		kvm_lapic_set_vapic_addr(vcpu, va.vapic_addr);
		break;
	}
1602 1603 1604 1605
	default:
		r = -EINVAL;
	}
out:
1606
	kfree(lapic);
1607 1608 1609
	return r;
}

1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
static int kvm_vm_ioctl_set_tss_addr(struct kvm *kvm, unsigned long addr)
{
	int ret;

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

static int kvm_vm_ioctl_set_nr_mmu_pages(struct kvm *kvm,
					  u32 kvm_nr_mmu_pages)
{
	if (kvm_nr_mmu_pages < KVM_MIN_ALLOC_MMU_PAGES)
		return -EINVAL;

1626
	down_write(&kvm->slots_lock);
1627
	spin_lock(&kvm->mmu_lock);
1628 1629

	kvm_mmu_change_mmu_pages(kvm, kvm_nr_mmu_pages);
1630
	kvm->arch.n_requested_mmu_pages = kvm_nr_mmu_pages;
1631

1632
	spin_unlock(&kvm->mmu_lock);
1633
	up_write(&kvm->slots_lock);
1634 1635 1636 1637 1638
	return 0;
}

static int kvm_vm_ioctl_get_nr_mmu_pages(struct kvm *kvm)
{
1639
	return kvm->arch.n_alloc_mmu_pages;
1640 1641
}

1642 1643 1644 1645 1646
gfn_t unalias_gfn(struct kvm *kvm, gfn_t gfn)
{
	int i;
	struct kvm_mem_alias *alias;

1647 1648
	for (i = 0; i < kvm->arch.naliases; ++i) {
		alias = &kvm->arch.aliases[i];
1649 1650 1651 1652 1653 1654 1655
		if (gfn >= alias->base_gfn
		    && gfn < alias->base_gfn + alias->npages)
			return alias->target_gfn + gfn - alias->base_gfn;
	}
	return gfn;
}

1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
/*
 * Set a new alias region.  Aliases map a portion of physical memory into
 * another portion.  This is useful for memory windows, for example the PC
 * VGA region.
 */
static int kvm_vm_ioctl_set_memory_alias(struct kvm *kvm,
					 struct kvm_memory_alias *alias)
{
	int r, n;
	struct kvm_mem_alias *p;

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

1682
	down_write(&kvm->slots_lock);
1683
	spin_lock(&kvm->mmu_lock);
1684

1685
	p = &kvm->arch.aliases[alias->slot];
1686 1687 1688 1689 1690
	p->base_gfn = alias->guest_phys_addr >> PAGE_SHIFT;
	p->npages = alias->memory_size >> PAGE_SHIFT;
	p->target_gfn = alias->target_phys_addr >> PAGE_SHIFT;

	for (n = KVM_ALIAS_SLOTS; n > 0; --n)
1691
		if (kvm->arch.aliases[n - 1].npages)
1692
			break;
1693
	kvm->arch.naliases = n;
1694

1695
	spin_unlock(&kvm->mmu_lock);
1696 1697
	kvm_mmu_zap_all(kvm);

1698
	up_write(&kvm->slots_lock);
1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762

	return 0;

out:
	return r;
}

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

	r = 0;
	switch (chip->chip_id) {
	case KVM_IRQCHIP_PIC_MASTER:
		memcpy(&chip->chip.pic,
			&pic_irqchip(kvm)->pics[0],
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_PIC_SLAVE:
		memcpy(&chip->chip.pic,
			&pic_irqchip(kvm)->pics[1],
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_IOAPIC:
		memcpy(&chip->chip.ioapic,
			ioapic_irqchip(kvm),
			sizeof(struct kvm_ioapic_state));
		break;
	default:
		r = -EINVAL;
		break;
	}
	return r;
}

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

	r = 0;
	switch (chip->chip_id) {
	case KVM_IRQCHIP_PIC_MASTER:
		memcpy(&pic_irqchip(kvm)->pics[0],
			&chip->chip.pic,
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_PIC_SLAVE:
		memcpy(&pic_irqchip(kvm)->pics[1],
			&chip->chip.pic,
			sizeof(struct kvm_pic_state));
		break;
	case KVM_IRQCHIP_IOAPIC:
		memcpy(ioapic_irqchip(kvm),
			&chip->chip.ioapic,
			sizeof(struct kvm_ioapic_state));
		break;
	default:
		r = -EINVAL;
		break;
	}
	kvm_pic_update_irq(pic_irqchip(kvm));
	return r;
}

1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
static int kvm_vm_ioctl_get_pit(struct kvm *kvm, struct kvm_pit_state *ps)
{
	int r = 0;

	memcpy(ps, &kvm->arch.vpit->pit_state, sizeof(struct kvm_pit_state));
	return r;
}

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

	memcpy(&kvm->arch.vpit->pit_state, ps, sizeof(struct kvm_pit_state));
	kvm_pit_load_count(kvm, 0, ps->channels[0].count);
	return r;
}

1780 1781 1782 1783 1784 1785 1786 1787 1788
static int kvm_vm_ioctl_reinject(struct kvm *kvm,
				 struct kvm_reinject_control *control)
{
	if (!kvm->arch.vpit)
		return -ENXIO;
	kvm->arch.vpit->pit_state.pit_timer.reinject = control->pit_reinject;
	return 0;
}

1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
/*
 * Get (and clear) the dirty memory log for a memory slot.
 */
int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
				      struct kvm_dirty_log *log)
{
	int r;
	int n;
	struct kvm_memory_slot *memslot;
	int is_dirty = 0;

1800
	down_write(&kvm->slots_lock);
1801 1802 1803 1804 1805 1806 1807

	r = kvm_get_dirty_log(kvm, log, &is_dirty);
	if (r)
		goto out;

	/* If nothing is dirty, don't bother messing with page tables. */
	if (is_dirty) {
1808
		spin_lock(&kvm->mmu_lock);
1809
		kvm_mmu_slot_remove_write_access(kvm, log->slot);
1810
		spin_unlock(&kvm->mmu_lock);
1811 1812 1813 1814 1815 1816 1817
		kvm_flush_remote_tlbs(kvm);
		memslot = &kvm->memslots[log->slot];
		n = ALIGN(memslot->npages, BITS_PER_LONG) / 8;
		memset(memslot->dirty_bitmap, 0, n);
	}
	r = 0;
out:
1818
	up_write(&kvm->slots_lock);
1819 1820 1821
	return r;
}

1822 1823 1824 1825 1826 1827
long kvm_arch_vm_ioctl(struct file *filp,
		       unsigned int ioctl, unsigned long arg)
{
	struct kvm *kvm = filp->private_data;
	void __user *argp = (void __user *)arg;
	int r = -EINVAL;
1828 1829 1830 1831 1832 1833 1834 1835 1836
	/*
	 * This union makes it completely explicit to gcc-3.x
	 * that these two variables' stack usage should be
	 * combined, not added together.
	 */
	union {
		struct kvm_pit_state ps;
		struct kvm_memory_alias alias;
	} u;
1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867

	switch (ioctl) {
	case KVM_SET_TSS_ADDR:
		r = kvm_vm_ioctl_set_tss_addr(kvm, arg);
		if (r < 0)
			goto out;
		break;
	case KVM_SET_MEMORY_REGION: {
		struct kvm_memory_region kvm_mem;
		struct kvm_userspace_memory_region kvm_userspace_mem;

		r = -EFAULT;
		if (copy_from_user(&kvm_mem, argp, sizeof kvm_mem))
			goto out;
		kvm_userspace_mem.slot = kvm_mem.slot;
		kvm_userspace_mem.flags = kvm_mem.flags;
		kvm_userspace_mem.guest_phys_addr = kvm_mem.guest_phys_addr;
		kvm_userspace_mem.memory_size = kvm_mem.memory_size;
		r = kvm_vm_ioctl_set_memory_region(kvm, &kvm_userspace_mem, 0);
		if (r)
			goto out;
		break;
	}
	case KVM_SET_NR_MMU_PAGES:
		r = kvm_vm_ioctl_set_nr_mmu_pages(kvm, arg);
		if (r)
			goto out;
		break;
	case KVM_GET_NR_MMU_PAGES:
		r = kvm_vm_ioctl_get_nr_mmu_pages(kvm);
		break;
1868
	case KVM_SET_MEMORY_ALIAS:
1869
		r = -EFAULT;
1870
		if (copy_from_user(&u.alias, argp, sizeof(struct kvm_memory_alias)))
1871
			goto out;
1872
		r = kvm_vm_ioctl_set_memory_alias(kvm, &u.alias);
1873 1874 1875 1876 1877
		if (r)
			goto out;
		break;
	case KVM_CREATE_IRQCHIP:
		r = -ENOMEM;
1878 1879
		kvm->arch.vpic = kvm_create_pic(kvm);
		if (kvm->arch.vpic) {
1880 1881
			r = kvm_ioapic_init(kvm);
			if (r) {
1882 1883
				kfree(kvm->arch.vpic);
				kvm->arch.vpic = NULL;
1884 1885 1886 1887
				goto out;
			}
		} else
			goto out;
1888 1889 1890 1891 1892 1893
		r = kvm_setup_default_irq_routing(kvm);
		if (r) {
			kfree(kvm->arch.vpic);
			kfree(kvm->arch.vioapic);
			goto out;
		}
1894
		break;
S
Sheng Yang 已提交
1895
	case KVM_CREATE_PIT:
A
Avi Kivity 已提交
1896 1897 1898 1899
		mutex_lock(&kvm->lock);
		r = -EEXIST;
		if (kvm->arch.vpit)
			goto create_pit_unlock;
S
Sheng Yang 已提交
1900 1901 1902 1903
		r = -ENOMEM;
		kvm->arch.vpit = kvm_create_pit(kvm);
		if (kvm->arch.vpit)
			r = 0;
A
Avi Kivity 已提交
1904 1905
	create_pit_unlock:
		mutex_unlock(&kvm->lock);
S
Sheng Yang 已提交
1906
		break;
1907
	case KVM_IRQ_LINE_STATUS:
1908 1909 1910 1911 1912 1913 1914
	case KVM_IRQ_LINE: {
		struct kvm_irq_level irq_event;

		r = -EFAULT;
		if (copy_from_user(&irq_event, argp, sizeof irq_event))
			goto out;
		if (irqchip_in_kernel(kvm)) {
1915
			__s32 status;
1916
			mutex_lock(&kvm->lock);
1917 1918
			status = kvm_set_irq(kvm, KVM_USERSPACE_IRQ_SOURCE_ID,
					irq_event.irq, irq_event.level);
1919
			mutex_unlock(&kvm->lock);
1920 1921 1922 1923 1924 1925
			if (ioctl == KVM_IRQ_LINE_STATUS) {
				irq_event.status = status;
				if (copy_to_user(argp, &irq_event,
							sizeof irq_event))
					goto out;
			}
1926 1927 1928 1929 1930 1931
			r = 0;
		}
		break;
	}
	case KVM_GET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
1932
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
1933

1934 1935
		r = -ENOMEM;
		if (!chip)
1936
			goto out;
1937 1938 1939
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto get_irqchip_out;
1940 1941
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
1942 1943
			goto get_irqchip_out;
		r = kvm_vm_ioctl_get_irqchip(kvm, chip);
1944
		if (r)
1945
			goto get_irqchip_out;
1946
		r = -EFAULT;
1947 1948
		if (copy_to_user(argp, chip, sizeof *chip))
			goto get_irqchip_out;
1949
		r = 0;
1950 1951 1952 1953
	get_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
1954 1955 1956 1957
		break;
	}
	case KVM_SET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
1958
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
1959

1960 1961
		r = -ENOMEM;
		if (!chip)
1962
			goto out;
1963 1964 1965
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto set_irqchip_out;
1966 1967
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
1968 1969
			goto set_irqchip_out;
		r = kvm_vm_ioctl_set_irqchip(kvm, chip);
1970
		if (r)
1971
			goto set_irqchip_out;
1972
		r = 0;
1973 1974 1975 1976
	set_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
1977 1978
		break;
	}
1979 1980
	case KVM_GET_PIT: {
		r = -EFAULT;
1981
		if (copy_from_user(&u.ps, argp, sizeof(struct kvm_pit_state)))
1982 1983 1984 1985
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
1986
		r = kvm_vm_ioctl_get_pit(kvm, &u.ps);
1987 1988 1989
		if (r)
			goto out;
		r = -EFAULT;
1990
		if (copy_to_user(argp, &u.ps, sizeof(struct kvm_pit_state)))
1991 1992 1993 1994 1995 1996
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_PIT: {
		r = -EFAULT;
1997
		if (copy_from_user(&u.ps, argp, sizeof u.ps))
1998 1999 2000 2001
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
2002
		r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
2003 2004 2005 2006 2007
		if (r)
			goto out;
		r = 0;
		break;
	}
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
	case KVM_REINJECT_CONTROL: {
		struct kvm_reinject_control control;
		r =  -EFAULT;
		if (copy_from_user(&control, argp, sizeof(control)))
			goto out;
		r = kvm_vm_ioctl_reinject(kvm, &control);
		if (r)
			goto out;
		r = 0;
		break;
	}
2019 2020 2021 2022 2023 2024 2025
	default:
		;
	}
out:
	return r;
}

2026
static void kvm_init_msr_list(void)
2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040
{
	u32 dummy[2];
	unsigned i, j;

	for (i = j = 0; i < ARRAY_SIZE(msrs_to_save); i++) {
		if (rdmsr_safe(msrs_to_save[i], &dummy[0], &dummy[1]) < 0)
			continue;
		if (j < i)
			msrs_to_save[j] = msrs_to_save[i];
		j++;
	}
	num_msrs_to_save = j;
}

2041 2042 2043 2044
/*
 * Only apic need an MMIO device hook, so shortcut now..
 */
static struct kvm_io_device *vcpu_find_pervcpu_dev(struct kvm_vcpu *vcpu,
2045 2046
						gpa_t addr, int len,
						int is_write)
2047 2048 2049
{
	struct kvm_io_device *dev;

2050 2051
	if (vcpu->arch.apic) {
		dev = &vcpu->arch.apic->dev;
2052
		if (dev->in_range(dev, addr, len, is_write))
2053 2054 2055 2056 2057 2058 2059
			return dev;
	}
	return NULL;
}


static struct kvm_io_device *vcpu_find_mmio_dev(struct kvm_vcpu *vcpu,
2060 2061
						gpa_t addr, int len,
						int is_write)
2062 2063 2064
{
	struct kvm_io_device *dev;

2065
	dev = vcpu_find_pervcpu_dev(vcpu, addr, len, is_write);
2066
	if (dev == NULL)
2067 2068
		dev = kvm_io_bus_find_dev(&vcpu->kvm->mmio_bus, addr, len,
					  is_write);
2069 2070 2071
	return dev;
}

2072 2073
static int kvm_read_guest_virt(gva_t addr, void *val, unsigned int bytes,
			       struct kvm_vcpu *vcpu)
2074 2075
{
	void *data = val;
2076
	int r = X86EMUL_CONTINUE;
2077 2078

	while (bytes) {
2079
		gpa_t gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2080
		unsigned offset = addr & (PAGE_SIZE-1);
2081
		unsigned toread = min(bytes, (unsigned)PAGE_SIZE - offset);
2082 2083
		int ret;

2084 2085 2086 2087
		if (gpa == UNMAPPED_GVA) {
			r = X86EMUL_PROPAGATE_FAULT;
			goto out;
		}
2088
		ret = kvm_read_guest(vcpu->kvm, gpa, data, toread);
2089 2090 2091 2092
		if (ret < 0) {
			r = X86EMUL_UNHANDLEABLE;
			goto out;
		}
2093

2094 2095 2096
		bytes -= toread;
		data += toread;
		addr += toread;
2097
	}
2098 2099
out:
	return r;
2100
}
2101

2102 2103
static int kvm_write_guest_virt(gva_t addr, void *val, unsigned int bytes,
				struct kvm_vcpu *vcpu)
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131
{
	void *data = val;
	int r = X86EMUL_CONTINUE;

	while (bytes) {
		gpa_t gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
		unsigned offset = addr & (PAGE_SIZE-1);
		unsigned towrite = min(bytes, (unsigned)PAGE_SIZE - offset);
		int ret;

		if (gpa == UNMAPPED_GVA) {
			r = X86EMUL_PROPAGATE_FAULT;
			goto out;
		}
		ret = kvm_write_guest(vcpu->kvm, gpa, data, towrite);
		if (ret < 0) {
			r = X86EMUL_UNHANDLEABLE;
			goto out;
		}

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

2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146

static int emulator_read_emulated(unsigned long addr,
				  void *val,
				  unsigned int bytes,
				  struct kvm_vcpu *vcpu)
{
	struct kvm_io_device *mmio_dev;
	gpa_t                 gpa;

	if (vcpu->mmio_read_completed) {
		memcpy(val, vcpu->mmio_data, bytes);
		vcpu->mmio_read_completed = 0;
		return X86EMUL_CONTINUE;
	}

2147
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2148 2149 2150 2151 2152

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

2153 2154
	if (kvm_read_guest_virt(addr, val, bytes, vcpu)
				== X86EMUL_CONTINUE)
2155 2156 2157 2158 2159 2160 2161 2162
		return X86EMUL_CONTINUE;
	if (gpa == UNMAPPED_GVA)
		return X86EMUL_PROPAGATE_FAULT;

mmio:
	/*
	 * Is this MMIO handled locally?
	 */
2163
	mutex_lock(&vcpu->kvm->lock);
2164
	mmio_dev = vcpu_find_mmio_dev(vcpu, gpa, bytes, 0);
2165 2166
	if (mmio_dev) {
		kvm_iodevice_read(mmio_dev, gpa, bytes, val);
2167
		mutex_unlock(&vcpu->kvm->lock);
2168 2169
		return X86EMUL_CONTINUE;
	}
2170
	mutex_unlock(&vcpu->kvm->lock);
2171 2172 2173 2174 2175 2176 2177 2178 2179

	vcpu->mmio_needed = 1;
	vcpu->mmio_phys_addr = gpa;
	vcpu->mmio_size = bytes;
	vcpu->mmio_is_write = 0;

	return X86EMUL_UNHANDLEABLE;
}

2180
int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
2181
			  const void *val, int bytes)
2182 2183 2184 2185
{
	int ret;

	ret = kvm_write_guest(vcpu->kvm, gpa, val, bytes);
2186
	if (ret < 0)
2187
		return 0;
2188
	kvm_mmu_pte_write(vcpu, gpa, val, bytes, 1);
2189 2190 2191 2192 2193 2194 2195 2196 2197
	return 1;
}

static int emulator_write_emulated_onepage(unsigned long addr,
					   const void *val,
					   unsigned int bytes,
					   struct kvm_vcpu *vcpu)
{
	struct kvm_io_device *mmio_dev;
2198 2199 2200
	gpa_t                 gpa;

	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2201 2202

	if (gpa == UNMAPPED_GVA) {
2203
		kvm_inject_page_fault(vcpu, addr, 2);
2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217
		return X86EMUL_PROPAGATE_FAULT;
	}

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

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

mmio:
	/*
	 * Is this MMIO handled locally?
	 */
2218
	mutex_lock(&vcpu->kvm->lock);
2219
	mmio_dev = vcpu_find_mmio_dev(vcpu, gpa, bytes, 1);
2220 2221
	if (mmio_dev) {
		kvm_iodevice_write(mmio_dev, gpa, bytes, val);
2222
		mutex_unlock(&vcpu->kvm->lock);
2223 2224
		return X86EMUL_CONTINUE;
	}
2225
	mutex_unlock(&vcpu->kvm->lock);
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268

	vcpu->mmio_needed = 1;
	vcpu->mmio_phys_addr = gpa;
	vcpu->mmio_size = bytes;
	vcpu->mmio_is_write = 1;
	memcpy(vcpu->mmio_data, val, bytes);

	return X86EMUL_CONTINUE;
}

int emulator_write_emulated(unsigned long addr,
				   const void *val,
				   unsigned int bytes,
				   struct kvm_vcpu *vcpu)
{
	/* Crossing a page boundary? */
	if (((addr + bytes - 1) ^ addr) & PAGE_MASK) {
		int rc, now;

		now = -addr & ~PAGE_MASK;
		rc = emulator_write_emulated_onepage(addr, val, now, vcpu);
		if (rc != X86EMUL_CONTINUE)
			return rc;
		addr += now;
		val += now;
		bytes -= now;
	}
	return emulator_write_emulated_onepage(addr, val, bytes, vcpu);
}
EXPORT_SYMBOL_GPL(emulator_write_emulated);

static int emulator_cmpxchg_emulated(unsigned long addr,
				     const void *old,
				     const void *new,
				     unsigned int bytes,
				     struct kvm_vcpu *vcpu)
{
	static int reported;

	if (!reported) {
		reported = 1;
		printk(KERN_WARNING "kvm: emulating exchange as write\n");
	}
2269 2270 2271
#ifndef CONFIG_X86_64
	/* guests cmpxchg8b have to be emulated atomically */
	if (bytes == 8) {
2272
		gpa_t gpa;
2273
		struct page *page;
A
Andrew Morton 已提交
2274
		char *kaddr;
2275 2276
		u64 val;

2277 2278
		gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);

2279 2280 2281 2282 2283 2284 2285 2286
		if (gpa == UNMAPPED_GVA ||
		   (gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
			goto emul_write;

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

		val = *(u64 *)new;
2287

2288
		page = gfn_to_page(vcpu->kvm, gpa >> PAGE_SHIFT);
2289

A
Andrew Morton 已提交
2290 2291 2292
		kaddr = kmap_atomic(page, KM_USER0);
		set_64bit((u64 *)(kaddr + offset_in_page(gpa)), val);
		kunmap_atomic(kaddr, KM_USER0);
2293 2294
		kvm_release_page_dirty(page);
	}
2295
emul_write:
2296 2297
#endif

2298 2299 2300 2301 2302 2303 2304 2305 2306 2307
	return emulator_write_emulated(addr, new, bytes, vcpu);
}

static unsigned long get_segment_base(struct kvm_vcpu *vcpu, int seg)
{
	return kvm_x86_ops->get_segment_base(vcpu, seg);
}

int emulate_invlpg(struct kvm_vcpu *vcpu, gva_t address)
{
M
Marcelo Tosatti 已提交
2308
	kvm_mmu_invlpg(vcpu, address);
2309 2310 2311 2312 2313
	return X86EMUL_CONTINUE;
}

int emulate_clts(struct kvm_vcpu *vcpu)
{
J
Joerg Roedel 已提交
2314
	KVMTRACE_0D(CLTS, vcpu, handler);
2315
	kvm_x86_ops->set_cr0(vcpu, vcpu->arch.cr0 & ~X86_CR0_TS);
2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327
	return X86EMUL_CONTINUE;
}

int emulator_get_dr(struct x86_emulate_ctxt *ctxt, int dr, unsigned long *dest)
{
	struct kvm_vcpu *vcpu = ctxt->vcpu;

	switch (dr) {
	case 0 ... 3:
		*dest = kvm_x86_ops->get_dr(vcpu, dr);
		return X86EMUL_CONTINUE;
	default:
2328
		pr_unimpl(vcpu, "%s: unexpected dr %u\n", __func__, dr);
2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348
		return X86EMUL_UNHANDLEABLE;
	}
}

int emulator_set_dr(struct x86_emulate_ctxt *ctxt, int dr, unsigned long value)
{
	unsigned long mask = (ctxt->mode == X86EMUL_MODE_PROT64) ? ~0ULL : ~0U;
	int exception;

	kvm_x86_ops->set_dr(ctxt->vcpu, dr, value & mask, &exception);
	if (exception) {
		/* FIXME: better handling */
		return X86EMUL_UNHANDLEABLE;
	}
	return X86EMUL_CONTINUE;
}

void kvm_report_emulation_failure(struct kvm_vcpu *vcpu, const char *context)
{
	u8 opcodes[4];
2349
	unsigned long rip = kvm_rip_read(vcpu);
2350 2351
	unsigned long rip_linear;

2352
	if (!printk_ratelimit())
2353 2354
		return;

2355 2356
	rip_linear = rip + get_segment_base(vcpu, VCPU_SREG_CS);

2357
	kvm_read_guest_virt(rip_linear, (void *)opcodes, 4, vcpu);
2358 2359 2360 2361 2362 2363

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

2364
static struct x86_emulate_ops emulate_ops = {
2365
	.read_std            = kvm_read_guest_virt,
2366 2367 2368 2369 2370
	.read_emulated       = emulator_read_emulated,
	.write_emulated      = emulator_write_emulated,
	.cmpxchg_emulated    = emulator_cmpxchg_emulated,
};

2371 2372 2373 2374 2375 2376 2377 2378
static void cache_all_regs(struct kvm_vcpu *vcpu)
{
	kvm_register_read(vcpu, VCPU_REGS_RAX);
	kvm_register_read(vcpu, VCPU_REGS_RSP);
	kvm_register_read(vcpu, VCPU_REGS_RIP);
	vcpu->arch.regs_dirty = ~0;
}

2379 2380 2381 2382
int emulate_instruction(struct kvm_vcpu *vcpu,
			struct kvm_run *run,
			unsigned long cr2,
			u16 error_code,
2383
			int emulation_type)
2384
{
2385
	int r, shadow_mask;
2386
	struct decode_cache *c;
2387

2388
	kvm_clear_exception_queue(vcpu);
2389
	vcpu->arch.mmio_fault_cr2 = cr2;
2390 2391 2392 2393 2394 2395 2396
	/*
	 * TODO: fix x86_emulate.c to use guest_read/write_register
	 * instead of direct ->regs accesses, can save hundred cycles
	 * on Intel for instructions that don't read/change RSP, for
	 * for example.
	 */
	cache_all_regs(vcpu);
2397 2398

	vcpu->mmio_is_write = 0;
2399
	vcpu->arch.pio.string = 0;
2400

2401
	if (!(emulation_type & EMULTYPE_NO_DECODE)) {
2402 2403 2404
		int cs_db, cs_l;
		kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);

2405 2406 2407 2408
		vcpu->arch.emulate_ctxt.vcpu = vcpu;
		vcpu->arch.emulate_ctxt.eflags = kvm_x86_ops->get_rflags(vcpu);
		vcpu->arch.emulate_ctxt.mode =
			(vcpu->arch.emulate_ctxt.eflags & X86_EFLAGS_VM)
2409 2410 2411 2412
			? X86EMUL_MODE_REAL : cs_l
			? X86EMUL_MODE_PROT64 :	cs_db
			? X86EMUL_MODE_PROT32 : X86EMUL_MODE_PROT16;

2413
		r = x86_decode_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
2414 2415 2416 2417 2418 2419 2420 2421 2422 2423

		/* Reject the instructions other than VMCALL/VMMCALL when
		 * try to emulate invalid opcode */
		c = &vcpu->arch.emulate_ctxt.decode;
		if ((emulation_type & EMULTYPE_TRAP_UD) &&
		    (!(c->twobyte && c->b == 0x01 &&
		      (c->modrm_reg == 0 || c->modrm_reg == 3) &&
		       c->modrm_mod == 3 && c->modrm_rm == 1)))
			return EMULATE_FAIL;

2424
		++vcpu->stat.insn_emulation;
2425
		if (r)  {
2426
			++vcpu->stat.insn_emulation_fail;
2427 2428 2429 2430 2431 2432
			if (kvm_mmu_unprotect_page_virt(vcpu, cr2))
				return EMULATE_DONE;
			return EMULATE_FAIL;
		}
	}

2433 2434 2435 2436 2437
	if (emulation_type & EMULTYPE_SKIP) {
		kvm_rip_write(vcpu, vcpu->arch.emulate_ctxt.decode.eip);
		return EMULATE_DONE;
	}

2438
	r = x86_emulate_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
2439 2440 2441 2442
	shadow_mask = vcpu->arch.emulate_ctxt.interruptibility;

	if (r == 0)
		kvm_x86_ops->set_interrupt_shadow(vcpu, shadow_mask);
2443

2444
	if (vcpu->arch.pio.string)
2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464
		return EMULATE_DO_MMIO;

	if ((r || vcpu->mmio_is_write) && run) {
		run->exit_reason = KVM_EXIT_MMIO;
		run->mmio.phys_addr = vcpu->mmio_phys_addr;
		memcpy(run->mmio.data, vcpu->mmio_data, 8);
		run->mmio.len = vcpu->mmio_size;
		run->mmio.is_write = vcpu->mmio_is_write;
	}

	if (r) {
		if (kvm_mmu_unprotect_page_virt(vcpu, cr2))
			return EMULATE_DONE;
		if (!vcpu->mmio_needed) {
			kvm_report_emulation_failure(vcpu, "mmio");
			return EMULATE_FAIL;
		}
		return EMULATE_DO_MMIO;
	}

2465
	kvm_x86_ops->set_rflags(vcpu, vcpu->arch.emulate_ctxt.eflags);
2466 2467 2468 2469 2470 2471 2472 2473 2474 2475

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

	return EMULATE_DONE;
}
EXPORT_SYMBOL_GPL(emulate_instruction);

2476 2477
static int pio_copy_data(struct kvm_vcpu *vcpu)
{
2478
	void *p = vcpu->arch.pio_data;
I
Izik Eidus 已提交
2479
	gva_t q = vcpu->arch.pio.guest_gva;
2480
	unsigned bytes;
I
Izik Eidus 已提交
2481
	int ret;
2482

2483 2484
	bytes = vcpu->arch.pio.size * vcpu->arch.pio.cur_count;
	if (vcpu->arch.pio.in)
I
Izik Eidus 已提交
2485
		ret = kvm_write_guest_virt(q, p, bytes, vcpu);
2486
	else
I
Izik Eidus 已提交
2487 2488
		ret = kvm_read_guest_virt(q, p, bytes, vcpu);
	return ret;
2489 2490 2491 2492
}

int complete_pio(struct kvm_vcpu *vcpu)
{
2493
	struct kvm_pio_request *io = &vcpu->arch.pio;
2494 2495
	long delta;
	int r;
2496
	unsigned long val;
2497 2498

	if (!io->string) {
2499 2500 2501 2502 2503
		if (io->in) {
			val = kvm_register_read(vcpu, VCPU_REGS_RAX);
			memcpy(&val, vcpu->arch.pio_data, io->size);
			kvm_register_write(vcpu, VCPU_REGS_RAX, val);
		}
2504 2505 2506
	} else {
		if (io->in) {
			r = pio_copy_data(vcpu);
2507
			if (r)
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517
				return r;
		}

		delta = 1;
		if (io->rep) {
			delta *= io->cur_count;
			/*
			 * The size of the register should really depend on
			 * current address size.
			 */
2518 2519 2520
			val = kvm_register_read(vcpu, VCPU_REGS_RCX);
			val -= delta;
			kvm_register_write(vcpu, VCPU_REGS_RCX, val);
2521 2522 2523 2524
		}
		if (io->down)
			delta = -delta;
		delta *= io->size;
2525 2526 2527 2528 2529 2530 2531 2532 2533
		if (io->in) {
			val = kvm_register_read(vcpu, VCPU_REGS_RDI);
			val += delta;
			kvm_register_write(vcpu, VCPU_REGS_RDI, val);
		} else {
			val = kvm_register_read(vcpu, VCPU_REGS_RSI);
			val += delta;
			kvm_register_write(vcpu, VCPU_REGS_RSI, val);
		}
2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
	}

	io->count -= io->cur_count;
	io->cur_count = 0;

	return 0;
}

static void kernel_pio(struct kvm_io_device *pio_dev,
		       struct kvm_vcpu *vcpu,
		       void *pd)
{
	/* TODO: String I/O for in kernel device */

	mutex_lock(&vcpu->kvm->lock);
2549 2550 2551
	if (vcpu->arch.pio.in)
		kvm_iodevice_read(pio_dev, vcpu->arch.pio.port,
				  vcpu->arch.pio.size,
2552 2553
				  pd);
	else
2554 2555
		kvm_iodevice_write(pio_dev, vcpu->arch.pio.port,
				   vcpu->arch.pio.size,
2556 2557 2558 2559 2560 2561 2562
				   pd);
	mutex_unlock(&vcpu->kvm->lock);
}

static void pio_string_write(struct kvm_io_device *pio_dev,
			     struct kvm_vcpu *vcpu)
{
2563 2564
	struct kvm_pio_request *io = &vcpu->arch.pio;
	void *pd = vcpu->arch.pio_data;
2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577
	int i;

	mutex_lock(&vcpu->kvm->lock);
	for (i = 0; i < io->cur_count; i++) {
		kvm_iodevice_write(pio_dev, io->port,
				   io->size,
				   pd);
		pd += io->size;
	}
	mutex_unlock(&vcpu->kvm->lock);
}

static struct kvm_io_device *vcpu_find_pio_dev(struct kvm_vcpu *vcpu,
2578 2579
					       gpa_t addr, int len,
					       int is_write)
2580
{
2581
	return kvm_io_bus_find_dev(&vcpu->kvm->pio_bus, addr, len, is_write);
2582 2583 2584 2585 2586 2587
}

int kvm_emulate_pio(struct kvm_vcpu *vcpu, struct kvm_run *run, int in,
		  int size, unsigned port)
{
	struct kvm_io_device *pio_dev;
2588
	unsigned long val;
2589 2590 2591

	vcpu->run->exit_reason = KVM_EXIT_IO;
	vcpu->run->io.direction = in ? KVM_EXIT_IO_IN : KVM_EXIT_IO_OUT;
2592
	vcpu->run->io.size = vcpu->arch.pio.size = size;
2593
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
2594 2595 2596 2597 2598 2599
	vcpu->run->io.count = vcpu->arch.pio.count = vcpu->arch.pio.cur_count = 1;
	vcpu->run->io.port = vcpu->arch.pio.port = port;
	vcpu->arch.pio.in = in;
	vcpu->arch.pio.string = 0;
	vcpu->arch.pio.down = 0;
	vcpu->arch.pio.rep = 0;
2600

F
Feng (Eric) Liu 已提交
2601 2602 2603 2604 2605 2606 2607
	if (vcpu->run->io.direction == KVM_EXIT_IO_IN)
		KVMTRACE_2D(IO_READ, vcpu, vcpu->run->io.port, (u32)size,
			    handler);
	else
		KVMTRACE_2D(IO_WRITE, vcpu, vcpu->run->io.port, (u32)size,
			    handler);

2608 2609
	val = kvm_register_read(vcpu, VCPU_REGS_RAX);
	memcpy(vcpu->arch.pio_data, &val, 4);
2610

2611
	pio_dev = vcpu_find_pio_dev(vcpu, port, size, !in);
2612
	if (pio_dev) {
2613
		kernel_pio(pio_dev, vcpu, vcpu->arch.pio_data);
2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625
		complete_pio(vcpu);
		return 1;
	}
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_emulate_pio);

int kvm_emulate_pio_string(struct kvm_vcpu *vcpu, struct kvm_run *run, int in,
		  int size, unsigned long count, int down,
		  gva_t address, int rep, unsigned port)
{
	unsigned now, in_page;
I
Izik Eidus 已提交
2626
	int ret = 0;
2627 2628 2629 2630
	struct kvm_io_device *pio_dev;

	vcpu->run->exit_reason = KVM_EXIT_IO;
	vcpu->run->io.direction = in ? KVM_EXIT_IO_IN : KVM_EXIT_IO_OUT;
2631
	vcpu->run->io.size = vcpu->arch.pio.size = size;
2632
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
2633 2634 2635 2636 2637 2638
	vcpu->run->io.count = vcpu->arch.pio.count = vcpu->arch.pio.cur_count = count;
	vcpu->run->io.port = vcpu->arch.pio.port = port;
	vcpu->arch.pio.in = in;
	vcpu->arch.pio.string = 1;
	vcpu->arch.pio.down = down;
	vcpu->arch.pio.rep = rep;
2639

F
Feng (Eric) Liu 已提交
2640 2641 2642 2643 2644 2645 2646
	if (vcpu->run->io.direction == KVM_EXIT_IO_IN)
		KVMTRACE_2D(IO_READ, vcpu, vcpu->run->io.port, (u32)size,
			    handler);
	else
		KVMTRACE_2D(IO_WRITE, vcpu, vcpu->run->io.port, (u32)size,
			    handler);

2647 2648 2649 2650 2651 2652 2653 2654 2655 2656
	if (!count) {
		kvm_x86_ops->skip_emulated_instruction(vcpu);
		return 1;
	}

	if (!down)
		in_page = PAGE_SIZE - offset_in_page(address);
	else
		in_page = offset_in_page(address) + size;
	now = min(count, (unsigned long)in_page / size);
I
Izik Eidus 已提交
2657
	if (!now)
2658 2659 2660 2661 2662 2663
		now = 1;
	if (down) {
		/*
		 * String I/O in reverse.  Yuck.  Kill the guest, fix later.
		 */
		pr_unimpl(vcpu, "guest string pio down\n");
2664
		kvm_inject_gp(vcpu, 0);
2665 2666 2667
		return 1;
	}
	vcpu->run->io.count = now;
2668
	vcpu->arch.pio.cur_count = now;
2669

2670
	if (vcpu->arch.pio.cur_count == vcpu->arch.pio.count)
2671 2672
		kvm_x86_ops->skip_emulated_instruction(vcpu);

I
Izik Eidus 已提交
2673
	vcpu->arch.pio.guest_gva = address;
2674

2675 2676 2677
	pio_dev = vcpu_find_pio_dev(vcpu, port,
				    vcpu->arch.pio.cur_count,
				    !vcpu->arch.pio.in);
2678
	if (!vcpu->arch.pio.in) {
2679 2680
		/* string PIO write */
		ret = pio_copy_data(vcpu);
I
Izik Eidus 已提交
2681 2682 2683 2684 2685
		if (ret == X86EMUL_PROPAGATE_FAULT) {
			kvm_inject_gp(vcpu, 0);
			return 1;
		}
		if (ret == 0 && pio_dev) {
2686 2687
			pio_string_write(pio_dev, vcpu);
			complete_pio(vcpu);
2688
			if (vcpu->arch.pio.count == 0)
2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699
				ret = 1;
		}
	} else if (pio_dev)
		pr_unimpl(vcpu, "no string pio read support yet, "
		       "port %x size %d count %ld\n",
			port, size, count);

	return ret;
}
EXPORT_SYMBOL_GPL(kvm_emulate_pio_string);

2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762
static void bounce_off(void *info)
{
	/* nothing */
}

static unsigned int  ref_freq;
static unsigned long tsc_khz_ref;

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

	if (!ref_freq)
		ref_freq = freq->old;

	if (val == CPUFREQ_PRECHANGE && freq->old > freq->new)
		return 0;
	if (val == CPUFREQ_POSTCHANGE && freq->old < freq->new)
		return 0;
	per_cpu(cpu_tsc_khz, freq->cpu) = cpufreq_scale(tsc_khz_ref, ref_freq, freq->new);

	spin_lock(&kvm_lock);
	list_for_each_entry(kvm, &vm_list, vm_list) {
		for (i = 0; i < KVM_MAX_VCPUS; ++i) {
			vcpu = kvm->vcpus[i];
			if (!vcpu)
				continue;
			if (vcpu->cpu != freq->cpu)
				continue;
			if (!kvm_request_guest_time_update(vcpu))
				continue;
			if (vcpu->cpu != smp_processor_id())
				send_ipi++;
		}
	}
	spin_unlock(&kvm_lock);

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

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

2763
int kvm_arch_init(void *opaque)
2764
{
2765
	int r, cpu;
2766 2767 2768 2769
	struct kvm_x86_ops *ops = (struct kvm_x86_ops *)opaque;

	if (kvm_x86_ops) {
		printk(KERN_ERR "kvm: already loaded the other module\n");
2770 2771
		r = -EEXIST;
		goto out;
2772 2773 2774 2775
	}

	if (!ops->cpu_has_kvm_support()) {
		printk(KERN_ERR "kvm: no hardware support\n");
2776 2777
		r = -EOPNOTSUPP;
		goto out;
2778 2779 2780
	}
	if (ops->disabled_by_bios()) {
		printk(KERN_ERR "kvm: disabled by bios\n");
2781 2782
		r = -EOPNOTSUPP;
		goto out;
2783 2784
	}

2785 2786 2787 2788 2789 2790
	r = kvm_mmu_module_init();
	if (r)
		goto out;

	kvm_init_msr_list();

2791
	kvm_x86_ops = ops;
2792
	kvm_mmu_set_nonpresent_ptes(0ull, 0ull);
S
Sheng Yang 已提交
2793 2794
	kvm_mmu_set_base_ptes(PT_PRESENT_MASK);
	kvm_mmu_set_mask_ptes(PT_USER_MASK, PT_ACCESSED_MASK,
2795
			PT_DIRTY_MASK, PT64_NX_MASK, 0);
2796 2797 2798 2799 2800 2801 2802 2803 2804

	for_each_possible_cpu(cpu)
		per_cpu(cpu_tsc_khz, cpu) = tsc_khz;
	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) {
		tsc_khz_ref = tsc_khz;
		cpufreq_register_notifier(&kvmclock_cpufreq_notifier_block,
					  CPUFREQ_TRANSITION_NOTIFIER);
	}

2805
	return 0;
2806 2807 2808

out:
	return r;
2809
}
2810

2811 2812
void kvm_arch_exit(void)
{
2813 2814 2815
	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
		cpufreq_unregister_notifier(&kvmclock_cpufreq_notifier_block,
					    CPUFREQ_TRANSITION_NOTIFIER);
2816
	kvm_x86_ops = NULL;
2817 2818
	kvm_mmu_module_exit();
}
2819

2820 2821 2822
int kvm_emulate_halt(struct kvm_vcpu *vcpu)
{
	++vcpu->stat.halt_exits;
F
Feng (Eric) Liu 已提交
2823
	KVMTRACE_0D(HLT, vcpu, handler);
2824
	if (irqchip_in_kernel(vcpu->kvm)) {
2825
		vcpu->arch.mp_state = KVM_MP_STATE_HALTED;
2826 2827 2828 2829 2830 2831 2832 2833
		return 1;
	} else {
		vcpu->run->exit_reason = KVM_EXIT_HLT;
		return 0;
	}
}
EXPORT_SYMBOL_GPL(kvm_emulate_halt);

2834 2835 2836 2837 2838 2839 2840 2841 2842
static inline gpa_t hc_gpa(struct kvm_vcpu *vcpu, unsigned long a0,
			   unsigned long a1)
{
	if (is_long_mode(vcpu))
		return a0;
	else
		return a0 | ((gpa_t)a1 << 32);
}

2843 2844 2845
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
{
	unsigned long nr, a0, a1, a2, a3, ret;
2846
	int r = 1;
2847

2848 2849 2850 2851 2852
	nr = kvm_register_read(vcpu, VCPU_REGS_RAX);
	a0 = kvm_register_read(vcpu, VCPU_REGS_RBX);
	a1 = kvm_register_read(vcpu, VCPU_REGS_RCX);
	a2 = kvm_register_read(vcpu, VCPU_REGS_RDX);
	a3 = kvm_register_read(vcpu, VCPU_REGS_RSI);
2853

F
Feng (Eric) Liu 已提交
2854 2855
	KVMTRACE_1D(VMMCALL, vcpu, (u32)nr, handler);

2856 2857 2858 2859 2860 2861 2862 2863 2864
	if (!is_long_mode(vcpu)) {
		nr &= 0xFFFFFFFF;
		a0 &= 0xFFFFFFFF;
		a1 &= 0xFFFFFFFF;
		a2 &= 0xFFFFFFFF;
		a3 &= 0xFFFFFFFF;
	}

	switch (nr) {
A
Avi Kivity 已提交
2865 2866 2867
	case KVM_HC_VAPIC_POLL_IRQ:
		ret = 0;
		break;
2868 2869 2870
	case KVM_HC_MMU_OP:
		r = kvm_pv_mmu_op(vcpu, a0, hc_gpa(vcpu, a1, a2), &ret);
		break;
2871 2872 2873 2874
	default:
		ret = -KVM_ENOSYS;
		break;
	}
2875
	kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
A
Amit Shah 已提交
2876
	++vcpu->stat.hypercalls;
2877
	return r;
2878 2879 2880 2881 2882 2883 2884
}
EXPORT_SYMBOL_GPL(kvm_emulate_hypercall);

int kvm_fix_hypercall(struct kvm_vcpu *vcpu)
{
	char instruction[3];
	int ret = 0;
2885
	unsigned long rip = kvm_rip_read(vcpu);
2886 2887 2888 2889 2890 2891 2892 2893 2894 2895


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

	kvm_x86_ops->patch_hypercall(vcpu, instruction);
2896
	if (emulator_write_emulated(rip, instruction, 3, vcpu)
2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924
	    != X86EMUL_CONTINUE)
		ret = -EFAULT;

	return ret;
}

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

void realmode_lgdt(struct kvm_vcpu *vcpu, u16 limit, unsigned long base)
{
	struct descriptor_table dt = { limit, base };

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

void realmode_lidt(struct kvm_vcpu *vcpu, u16 limit, unsigned long base)
{
	struct descriptor_table dt = { limit, base };

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

void realmode_lmsw(struct kvm_vcpu *vcpu, unsigned long msw,
		   unsigned long *rflags)
{
2925
	kvm_lmsw(vcpu, msw);
2926 2927 2928 2929 2930
	*rflags = kvm_x86_ops->get_rflags(vcpu);
}

unsigned long realmode_get_cr(struct kvm_vcpu *vcpu, int cr)
{
J
Joerg Roedel 已提交
2931 2932
	unsigned long value;

2933 2934 2935
	kvm_x86_ops->decache_cr4_guest_bits(vcpu);
	switch (cr) {
	case 0:
J
Joerg Roedel 已提交
2936 2937
		value = vcpu->arch.cr0;
		break;
2938
	case 2:
J
Joerg Roedel 已提交
2939 2940
		value = vcpu->arch.cr2;
		break;
2941
	case 3:
J
Joerg Roedel 已提交
2942 2943
		value = vcpu->arch.cr3;
		break;
2944
	case 4:
J
Joerg Roedel 已提交
2945 2946
		value = vcpu->arch.cr4;
		break;
2947
	case 8:
J
Joerg Roedel 已提交
2948 2949
		value = kvm_get_cr8(vcpu);
		break;
2950
	default:
2951
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
2952 2953
		return 0;
	}
J
Joerg Roedel 已提交
2954 2955 2956 2957
	KVMTRACE_3D(CR_READ, vcpu, (u32)cr, (u32)value,
		    (u32)((u64)value >> 32), handler);

	return value;
2958 2959 2960 2961 2962
}

void realmode_set_cr(struct kvm_vcpu *vcpu, int cr, unsigned long val,
		     unsigned long *rflags)
{
J
Joerg Roedel 已提交
2963 2964 2965
	KVMTRACE_3D(CR_WRITE, vcpu, (u32)cr, (u32)val,
		    (u32)((u64)val >> 32), handler);

2966 2967
	switch (cr) {
	case 0:
2968
		kvm_set_cr0(vcpu, mk_cr_64(vcpu->arch.cr0, val));
2969 2970 2971
		*rflags = kvm_x86_ops->get_rflags(vcpu);
		break;
	case 2:
2972
		vcpu->arch.cr2 = val;
2973 2974
		break;
	case 3:
2975
		kvm_set_cr3(vcpu, val);
2976 2977
		break;
	case 4:
2978
		kvm_set_cr4(vcpu, mk_cr_64(vcpu->arch.cr4, val));
2979
		break;
2980
	case 8:
2981
		kvm_set_cr8(vcpu, val & 0xfUL);
2982
		break;
2983
	default:
2984
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
2985 2986 2987
	}
}

2988 2989
static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
{
2990 2991
	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	int j, nent = vcpu->arch.cpuid_nent;
2992 2993 2994

	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	/* when no next entry is found, the current entry[i] is reselected */
2995
	for (j = i + 1; ; j = (j + 1) % nent) {
2996
		struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014
		if (ej->function == e->function) {
			ej->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
			return j;
		}
	}
	return 0; /* silence gcc, even though control never reaches here */
}

/* find an entry with matching function, matching index (if needed), and that
 * should be read next (if it's stateful) */
static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	u32 function, u32 index)
{
	if (e->function != function)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
3015
	    !(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
3016 3017 3018 3019
		return 0;
	return 1;
}

3020 3021
struct kvm_cpuid_entry2 *kvm_find_cpuid_entry(struct kvm_vcpu *vcpu,
					      u32 function, u32 index)
3022 3023
{
	int i;
3024
	struct kvm_cpuid_entry2 *best = NULL;
3025

3026
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
3027 3028
		struct kvm_cpuid_entry2 *e;

3029
		e = &vcpu->arch.cpuid_entries[i];
3030 3031 3032
		if (is_matching_cpuid_entry(e, function, index)) {
			if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
				move_to_next_stateful_cpuid_entry(vcpu, i);
3033 3034 3035 3036 3037 3038 3039 3040 3041 3042
			best = e;
			break;
		}
		/*
		 * Both basic or both extended?
		 */
		if (((e->function ^ function) & 0x80000000) == 0)
			if (!best || e->function > best->function)
				best = e;
	}
3043 3044 3045
	return best;
}

3046 3047 3048 3049 3050 3051 3052 3053 3054 3055
int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
{
	struct kvm_cpuid_entry2 *best;

	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
	if (best)
		return best->eax & 0xff;
	return 36;
}

3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067
void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
{
	u32 function, index;
	struct kvm_cpuid_entry2 *best;

	function = kvm_register_read(vcpu, VCPU_REGS_RAX);
	index = kvm_register_read(vcpu, VCPU_REGS_RCX);
	kvm_register_write(vcpu, VCPU_REGS_RAX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RBX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RCX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RDX, 0);
	best = kvm_find_cpuid_entry(vcpu, function, index);
3068
	if (best) {
3069 3070 3071 3072
		kvm_register_write(vcpu, VCPU_REGS_RAX, best->eax);
		kvm_register_write(vcpu, VCPU_REGS_RBX, best->ebx);
		kvm_register_write(vcpu, VCPU_REGS_RCX, best->ecx);
		kvm_register_write(vcpu, VCPU_REGS_RDX, best->edx);
3073 3074
	}
	kvm_x86_ops->skip_emulated_instruction(vcpu);
F
Feng (Eric) Liu 已提交
3075
	KVMTRACE_5D(CPUID, vcpu, function,
3076 3077 3078 3079
		    (u32)kvm_register_read(vcpu, VCPU_REGS_RAX),
		    (u32)kvm_register_read(vcpu, VCPU_REGS_RBX),
		    (u32)kvm_register_read(vcpu, VCPU_REGS_RCX),
		    (u32)kvm_register_read(vcpu, VCPU_REGS_RDX), handler);
3080 3081
}
EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);
3082

3083 3084 3085 3086 3087 3088 3089 3090 3091
/*
 * Check if userspace requested an interrupt window, and that the
 * interrupt window is open.
 *
 * No need to exit to userspace if we already have an interrupt queued.
 */
static int dm_request_for_irq_injection(struct kvm_vcpu *vcpu,
					  struct kvm_run *kvm_run)
{
3092
	return (!irqchip_in_kernel(vcpu->kvm) && !kvm_cpu_has_interrupt(vcpu) &&
3093
		kvm_run->request_interrupt_window &&
3094
		kvm_arch_interrupt_allowed(vcpu));
3095 3096 3097 3098 3099 3100
}

static void post_kvm_run_save(struct kvm_vcpu *vcpu,
			      struct kvm_run *kvm_run)
{
	kvm_run->if_flag = (kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF) != 0;
3101
	kvm_run->cr8 = kvm_get_cr8(vcpu);
3102
	kvm_run->apic_base = kvm_get_apic_base(vcpu);
3103
	if (irqchip_in_kernel(vcpu->kvm))
3104
		kvm_run->ready_for_interrupt_injection = 1;
3105
	else
3106
		kvm_run->ready_for_interrupt_injection =
3107 3108 3109
			kvm_arch_interrupt_allowed(vcpu) &&
			!kvm_cpu_has_interrupt(vcpu) &&
			!kvm_event_needs_reinjection(vcpu);
3110 3111
}

A
Avi Kivity 已提交
3112 3113 3114 3115 3116 3117 3118 3119 3120
static void vapic_enter(struct kvm_vcpu *vcpu)
{
	struct kvm_lapic *apic = vcpu->arch.apic;
	struct page *page;

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

	page = gfn_to_page(vcpu->kvm, apic->vapic_addr >> PAGE_SHIFT);
3121 3122

	vcpu->arch.apic->vapic_page = page;
A
Avi Kivity 已提交
3123 3124 3125 3126 3127 3128 3129 3130 3131
}

static void vapic_exit(struct kvm_vcpu *vcpu)
{
	struct kvm_lapic *apic = vcpu->arch.apic;

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

3132
	down_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
3133 3134
	kvm_release_page_dirty(apic->vapic_page);
	mark_page_dirty(vcpu->kvm, apic->vapic_addr >> PAGE_SHIFT);
3135
	up_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
3136 3137
}

3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154
static void update_cr8_intercept(struct kvm_vcpu *vcpu)
{
	int max_irr, tpr;

	if (!kvm_x86_ops->update_cr8_intercept)
		return;

	max_irr = kvm_lapic_find_highest_irr(vcpu);

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

	tpr = kvm_lapic_get_cr8(vcpu);

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

3155
static void inject_pending_irq(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
3156
{
3157 3158 3159
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
		kvm_x86_ops->set_interrupt_shadow(vcpu, 0);

3160 3161 3162 3163 3164 3165 3166
	/* try to reinject previous events if any */
	if (vcpu->arch.nmi_injected) {
		kvm_x86_ops->set_nmi(vcpu);
		return;
	}

	if (vcpu->arch.interrupt.pending) {
3167
		kvm_x86_ops->set_irq(vcpu);
3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179
		return;
	}

	/* try to inject new event if pending */
	if (vcpu->arch.nmi_pending) {
		if (kvm_x86_ops->nmi_allowed(vcpu)) {
			vcpu->arch.nmi_pending = false;
			vcpu->arch.nmi_injected = true;
			kvm_x86_ops->set_nmi(vcpu);
		}
	} else if (kvm_cpu_has_interrupt(vcpu)) {
		if (kvm_x86_ops->interrupt_allowed(vcpu)) {
3180 3181 3182
			kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu),
					    false);
			kvm_x86_ops->set_irq(vcpu);
3183 3184 3185 3186
		}
	}
}

3187
static int vcpu_enter_guest(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
3188 3189
{
	int r;
3190 3191
	bool req_int_win = !irqchip_in_kernel(vcpu->kvm) &&
		kvm_run->request_interrupt_window;
3192

3193 3194 3195 3196
	if (vcpu->requests)
		if (test_and_clear_bit(KVM_REQ_MMU_RELOAD, &vcpu->requests))
			kvm_mmu_unload(vcpu);

3197 3198 3199 3200
	r = kvm_mmu_reload(vcpu);
	if (unlikely(r))
		goto out;

3201 3202
	if (vcpu->requests) {
		if (test_and_clear_bit(KVM_REQ_MIGRATE_TIMER, &vcpu->requests))
M
Marcelo Tosatti 已提交
3203
			__kvm_migrate_timers(vcpu);
3204 3205
		if (test_and_clear_bit(KVM_REQ_KVMCLOCK_UPDATE, &vcpu->requests))
			kvm_write_guest_time(vcpu);
3206 3207
		if (test_and_clear_bit(KVM_REQ_MMU_SYNC, &vcpu->requests))
			kvm_mmu_sync_roots(vcpu);
3208 3209
		if (test_and_clear_bit(KVM_REQ_TLB_FLUSH, &vcpu->requests))
			kvm_x86_ops->tlb_flush(vcpu);
A
Avi Kivity 已提交
3210 3211 3212 3213 3214 3215
		if (test_and_clear_bit(KVM_REQ_REPORT_TPR_ACCESS,
				       &vcpu->requests)) {
			kvm_run->exit_reason = KVM_EXIT_TPR_ACCESS;
			r = 0;
			goto out;
		}
J
Joerg Roedel 已提交
3216 3217 3218 3219 3220
		if (test_and_clear_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests)) {
			kvm_run->exit_reason = KVM_EXIT_SHUTDOWN;
			r = 0;
			goto out;
		}
3221
	}
A
Avi Kivity 已提交
3222

3223 3224 3225 3226 3227 3228 3229
	preempt_disable();

	kvm_x86_ops->prepare_guest_switch(vcpu);
	kvm_load_guest_fpu(vcpu);

	local_irq_disable();

3230 3231 3232
	clear_bit(KVM_REQ_KICK, &vcpu->requests);
	smp_mb__after_clear_bit();

3233
	if (vcpu->requests || need_resched() || signal_pending(current)) {
3234 3235 3236 3237 3238 3239
		local_irq_enable();
		preempt_enable();
		r = 1;
		goto out;
	}

3240
	if (vcpu->arch.exception.pending)
3241
		__queue_exception(vcpu);
3242
	else
3243
		inject_pending_irq(vcpu, kvm_run);
3244

3245 3246 3247 3248 3249 3250
	/* enable NMI/IRQ window open exits if needed */
	if (vcpu->arch.nmi_pending)
		kvm_x86_ops->enable_nmi_window(vcpu);
	else if (kvm_cpu_has_interrupt(vcpu) || req_int_win)
		kvm_x86_ops->enable_irq_window(vcpu);

3251 3252 3253 3254 3255 3256
	if (kvm_lapic_enabled(vcpu)) {
		if (!vcpu->arch.apic->vapic_addr)
			update_cr8_intercept(vcpu);
		else
			kvm_lapic_sync_to_vapic(vcpu);
	}
A
Avi Kivity 已提交
3257

3258 3259
	up_read(&vcpu->kvm->slots_lock);

3260 3261
	kvm_guest_enter();

3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275
	get_debugreg(vcpu->arch.host_dr6, 6);
	get_debugreg(vcpu->arch.host_dr7, 7);
	if (unlikely(vcpu->arch.switch_db_regs)) {
		get_debugreg(vcpu->arch.host_db[0], 0);
		get_debugreg(vcpu->arch.host_db[1], 1);
		get_debugreg(vcpu->arch.host_db[2], 2);
		get_debugreg(vcpu->arch.host_db[3], 3);

		set_debugreg(0, 7);
		set_debugreg(vcpu->arch.eff_db[0], 0);
		set_debugreg(vcpu->arch.eff_db[1], 1);
		set_debugreg(vcpu->arch.eff_db[2], 2);
		set_debugreg(vcpu->arch.eff_db[3], 3);
	}
3276

F
Feng (Eric) Liu 已提交
3277
	KVMTRACE_0D(VMENTRY, vcpu, entryexit);
3278 3279
	kvm_x86_ops->run(vcpu, kvm_run);

3280 3281 3282 3283 3284 3285 3286 3287 3288 3289
	if (unlikely(vcpu->arch.switch_db_regs)) {
		set_debugreg(0, 7);
		set_debugreg(vcpu->arch.host_db[0], 0);
		set_debugreg(vcpu->arch.host_db[1], 1);
		set_debugreg(vcpu->arch.host_db[2], 2);
		set_debugreg(vcpu->arch.host_db[3], 3);
	}
	set_debugreg(vcpu->arch.host_dr6, 6);
	set_debugreg(vcpu->arch.host_dr7, 7);

3290
	set_bit(KVM_REQ_KICK, &vcpu->requests);
3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306
	local_irq_enable();

	++vcpu->stat.exits;

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

	kvm_guest_exit();

	preempt_enable();

3307 3308
	down_read(&vcpu->kvm->slots_lock);

3309 3310 3311 3312
	/*
	 * Profile KVM exit RIPs:
	 */
	if (unlikely(prof_on == KVM_PROFILING)) {
3313 3314
		unsigned long rip = kvm_rip_read(vcpu);
		profile_hit(KVM_PROFILING, (void *)rip);
3315 3316
	}

3317

A
Avi Kivity 已提交
3318 3319
	kvm_lapic_sync_from_vapic(vcpu);

3320
	r = kvm_x86_ops->handle_exit(kvm_run, vcpu);
3321 3322 3323
out:
	return r;
}
3324

3325

3326 3327 3328 3329 3330
static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
	int r;

	if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED)) {
3331 3332
		pr_debug("vcpu %d received sipi with vector # %x\n",
			 vcpu->vcpu_id, vcpu->arch.sipi_vector);
3333
		kvm_lapic_reset(vcpu);
3334
		r = kvm_arch_vcpu_reset(vcpu);
3335 3336 3337
		if (r)
			return r;
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
3338 3339
	}

3340 3341 3342 3343 3344
	down_read(&vcpu->kvm->slots_lock);
	vapic_enter(vcpu);

	r = 1;
	while (r > 0) {
3345
		if (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE)
3346 3347 3348 3349 3350 3351
			r = vcpu_enter_guest(vcpu, kvm_run);
		else {
			up_read(&vcpu->kvm->slots_lock);
			kvm_vcpu_block(vcpu);
			down_read(&vcpu->kvm->slots_lock);
			if (test_and_clear_bit(KVM_REQ_UNHALT, &vcpu->requests))
3352 3353 3354
			{
				switch(vcpu->arch.mp_state) {
				case KVM_MP_STATE_HALTED:
3355
					vcpu->arch.mp_state =
3356 3357 3358 3359 3360 3361 3362 3363 3364
						KVM_MP_STATE_RUNNABLE;
				case KVM_MP_STATE_RUNNABLE:
					break;
				case KVM_MP_STATE_SIPI_RECEIVED:
				default:
					r = -EINTR;
					break;
				}
			}
3365 3366
		}

3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387
		if (r <= 0)
			break;

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

		if (dm_request_for_irq_injection(vcpu, kvm_run)) {
			r = -EINTR;
			kvm_run->exit_reason = KVM_EXIT_INTR;
			++vcpu->stat.request_irq_exits;
		}
		if (signal_pending(current)) {
			r = -EINTR;
			kvm_run->exit_reason = KVM_EXIT_INTR;
			++vcpu->stat.signal_exits;
		}
		if (need_resched()) {
			up_read(&vcpu->kvm->slots_lock);
			kvm_resched(vcpu);
			down_read(&vcpu->kvm->slots_lock);
3388
		}
3389 3390
	}

3391
	up_read(&vcpu->kvm->slots_lock);
3392 3393
	post_kvm_run_save(vcpu, kvm_run);

A
Avi Kivity 已提交
3394 3395
	vapic_exit(vcpu);

3396 3397 3398 3399 3400 3401 3402 3403 3404 3405
	return r;
}

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

	vcpu_load(vcpu);

3406 3407 3408
	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &vcpu->sigset, &sigsaved);

3409
	if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
3410
		kvm_vcpu_block(vcpu);
3411
		clear_bit(KVM_REQ_UNHALT, &vcpu->requests);
3412 3413
		r = -EAGAIN;
		goto out;
3414 3415 3416 3417
	}

	/* re-sync apic's tpr */
	if (!irqchip_in_kernel(vcpu->kvm))
3418
		kvm_set_cr8(vcpu, kvm_run->cr8);
3419

3420
	if (vcpu->arch.pio.cur_count) {
3421 3422 3423 3424 3425 3426 3427 3428 3429
		r = complete_pio(vcpu);
		if (r)
			goto out;
	}
#if CONFIG_HAS_IOMEM
	if (vcpu->mmio_needed) {
		memcpy(vcpu->mmio_data, kvm_run->mmio.data, 8);
		vcpu->mmio_read_completed = 1;
		vcpu->mmio_needed = 0;
3430 3431

		down_read(&vcpu->kvm->slots_lock);
3432
		r = emulate_instruction(vcpu, kvm_run,
3433 3434
					vcpu->arch.mmio_fault_cr2, 0,
					EMULTYPE_NO_DECODE);
3435
		up_read(&vcpu->kvm->slots_lock);
3436 3437 3438 3439 3440 3441 3442 3443 3444
		if (r == EMULATE_DO_MMIO) {
			/*
			 * Read-modify-write.  Back to userspace.
			 */
			r = 0;
			goto out;
		}
	}
#endif
3445 3446 3447
	if (kvm_run->exit_reason == KVM_EXIT_HYPERCALL)
		kvm_register_write(vcpu, VCPU_REGS_RAX,
				     kvm_run->hypercall.ret);
3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462

	r = __vcpu_run(vcpu, kvm_run);

out:
	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &sigsaved, NULL);

	vcpu_put(vcpu);
	return r;
}

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

3463 3464 3465 3466 3467 3468 3469 3470
	regs->rax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	regs->rbx = kvm_register_read(vcpu, VCPU_REGS_RBX);
	regs->rcx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	regs->rdx = kvm_register_read(vcpu, VCPU_REGS_RDX);
	regs->rsi = kvm_register_read(vcpu, VCPU_REGS_RSI);
	regs->rdi = kvm_register_read(vcpu, VCPU_REGS_RDI);
	regs->rsp = kvm_register_read(vcpu, VCPU_REGS_RSP);
	regs->rbp = kvm_register_read(vcpu, VCPU_REGS_RBP);
3471
#ifdef CONFIG_X86_64
3472 3473 3474 3475 3476 3477 3478 3479
	regs->r8 = kvm_register_read(vcpu, VCPU_REGS_R8);
	regs->r9 = kvm_register_read(vcpu, VCPU_REGS_R9);
	regs->r10 = kvm_register_read(vcpu, VCPU_REGS_R10);
	regs->r11 = kvm_register_read(vcpu, VCPU_REGS_R11);
	regs->r12 = kvm_register_read(vcpu, VCPU_REGS_R12);
	regs->r13 = kvm_register_read(vcpu, VCPU_REGS_R13);
	regs->r14 = kvm_register_read(vcpu, VCPU_REGS_R14);
	regs->r15 = kvm_register_read(vcpu, VCPU_REGS_R15);
3480 3481
#endif

3482
	regs->rip = kvm_rip_read(vcpu);
3483 3484 3485 3486 3487
	regs->rflags = kvm_x86_ops->get_rflags(vcpu);

	/*
	 * Don't leak debug flags in case they were set for guest debugging
	 */
J
Jan Kiszka 已提交
3488
	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499
		regs->rflags &= ~(X86_EFLAGS_TF | X86_EFLAGS_RF);

	vcpu_put(vcpu);

	return 0;
}

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

3500 3501 3502 3503 3504 3505 3506 3507
	kvm_register_write(vcpu, VCPU_REGS_RAX, regs->rax);
	kvm_register_write(vcpu, VCPU_REGS_RBX, regs->rbx);
	kvm_register_write(vcpu, VCPU_REGS_RCX, regs->rcx);
	kvm_register_write(vcpu, VCPU_REGS_RDX, regs->rdx);
	kvm_register_write(vcpu, VCPU_REGS_RSI, regs->rsi);
	kvm_register_write(vcpu, VCPU_REGS_RDI, regs->rdi);
	kvm_register_write(vcpu, VCPU_REGS_RSP, regs->rsp);
	kvm_register_write(vcpu, VCPU_REGS_RBP, regs->rbp);
3508
#ifdef CONFIG_X86_64
3509 3510 3511 3512 3513 3514 3515 3516 3517
	kvm_register_write(vcpu, VCPU_REGS_R8, regs->r8);
	kvm_register_write(vcpu, VCPU_REGS_R9, regs->r9);
	kvm_register_write(vcpu, VCPU_REGS_R10, regs->r10);
	kvm_register_write(vcpu, VCPU_REGS_R11, regs->r11);
	kvm_register_write(vcpu, VCPU_REGS_R12, regs->r12);
	kvm_register_write(vcpu, VCPU_REGS_R13, regs->r13);
	kvm_register_write(vcpu, VCPU_REGS_R14, regs->r14);
	kvm_register_write(vcpu, VCPU_REGS_R15, regs->r15);

3518 3519
#endif

3520
	kvm_rip_write(vcpu, regs->rip);
3521 3522 3523
	kvm_x86_ops->set_rflags(vcpu, regs->rflags);


3524 3525
	vcpu->arch.exception.pending = false;

3526 3527 3528 3529 3530
	vcpu_put(vcpu);

	return 0;
}

3531 3532
void kvm_get_segment(struct kvm_vcpu *vcpu,
		     struct kvm_segment *var, int seg)
3533
{
3534
	kvm_x86_ops->get_segment(vcpu, var, seg);
3535 3536 3537 3538 3539 3540
}

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

3541
	kvm_get_segment(vcpu, &cs, VCPU_SREG_CS);
3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553
	*db = cs.db;
	*l = cs.l;
}
EXPORT_SYMBOL_GPL(kvm_get_cs_db_l_bits);

int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
				  struct kvm_sregs *sregs)
{
	struct descriptor_table dt;

	vcpu_load(vcpu);

3554 3555 3556 3557 3558 3559
	kvm_get_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
	kvm_get_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
	kvm_get_segment(vcpu, &sregs->es, VCPU_SREG_ES);
	kvm_get_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
	kvm_get_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
	kvm_get_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
3560

3561 3562
	kvm_get_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_get_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
3563 3564 3565 3566 3567 3568 3569 3570 3571

	kvm_x86_ops->get_idt(vcpu, &dt);
	sregs->idt.limit = dt.limit;
	sregs->idt.base = dt.base;
	kvm_x86_ops->get_gdt(vcpu, &dt);
	sregs->gdt.limit = dt.limit;
	sregs->gdt.base = dt.base;

	kvm_x86_ops->decache_cr4_guest_bits(vcpu);
3572 3573 3574 3575
	sregs->cr0 = vcpu->arch.cr0;
	sregs->cr2 = vcpu->arch.cr2;
	sregs->cr3 = vcpu->arch.cr3;
	sregs->cr4 = vcpu->arch.cr4;
3576
	sregs->cr8 = kvm_get_cr8(vcpu);
3577
	sregs->efer = vcpu->arch.shadow_efer;
3578 3579
	sregs->apic_base = kvm_get_apic_base(vcpu);

G
Gleb Natapov 已提交
3580
	memset(sregs->interrupt_bitmap, 0, sizeof sregs->interrupt_bitmap);
3581

3582 3583 3584
	if (vcpu->arch.interrupt.pending)
		set_bit(vcpu->arch.interrupt.nr,
			(unsigned long *)sregs->interrupt_bitmap);
3585

3586 3587 3588 3589 3590
	vcpu_put(vcpu);

	return 0;
}

3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608
int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
				    struct kvm_mp_state *mp_state)
{
	vcpu_load(vcpu);
	mp_state->mp_state = vcpu->arch.mp_state;
	vcpu_put(vcpu);
	return 0;
}

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

3609
static void kvm_set_segment(struct kvm_vcpu *vcpu,
3610 3611
			struct kvm_segment *var, int seg)
{
3612
	kvm_x86_ops->set_segment(vcpu, var, seg);
3613 3614
}

3615 3616 3617 3618 3619 3620 3621 3622
static void seg_desct_to_kvm_desct(struct desc_struct *seg_desc, u16 selector,
				   struct kvm_segment *kvm_desct)
{
	kvm_desct->base = seg_desc->base0;
	kvm_desct->base |= seg_desc->base1 << 16;
	kvm_desct->base |= seg_desc->base2 << 24;
	kvm_desct->limit = seg_desc->limit0;
	kvm_desct->limit |= seg_desc->limit << 16;
3623 3624 3625 3626
	if (seg_desc->g) {
		kvm_desct->limit <<= 12;
		kvm_desct->limit |= 0xfff;
	}
3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642
	kvm_desct->selector = selector;
	kvm_desct->type = seg_desc->type;
	kvm_desct->present = seg_desc->p;
	kvm_desct->dpl = seg_desc->dpl;
	kvm_desct->db = seg_desc->d;
	kvm_desct->s = seg_desc->s;
	kvm_desct->l = seg_desc->l;
	kvm_desct->g = seg_desc->g;
	kvm_desct->avl = seg_desc->avl;
	if (!selector)
		kvm_desct->unusable = 1;
	else
		kvm_desct->unusable = 0;
	kvm_desct->padding = 0;
}

A
Amit Shah 已提交
3643 3644 3645
static void get_segment_descriptor_dtable(struct kvm_vcpu *vcpu,
					  u16 selector,
					  struct descriptor_table *dtable)
3646 3647 3648 3649
{
	if (selector & 1 << 2) {
		struct kvm_segment kvm_seg;

3650
		kvm_get_segment(vcpu, &kvm_seg, VCPU_SREG_LDTR);
3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665

		if (kvm_seg.unusable)
			dtable->limit = 0;
		else
			dtable->limit = kvm_seg.limit;
		dtable->base = kvm_seg.base;
	}
	else
		kvm_x86_ops->get_gdt(vcpu, dtable);
}

/* allowed just for 8 bytes segments */
static int load_guest_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector,
					 struct desc_struct *seg_desc)
{
3666
	gpa_t gpa;
3667 3668 3669
	struct descriptor_table dtable;
	u16 index = selector >> 3;

A
Amit Shah 已提交
3670
	get_segment_descriptor_dtable(vcpu, selector, &dtable);
3671 3672 3673 3674 3675

	if (dtable.limit < index * 8 + 7) {
		kvm_queue_exception_e(vcpu, GP_VECTOR, selector & 0xfffc);
		return 1;
	}
3676 3677 3678
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, dtable.base);
	gpa += index * 8;
	return kvm_read_guest(vcpu->kvm, gpa, seg_desc, 8);
3679 3680 3681 3682 3683 3684
}

/* allowed just for 8 bytes segments */
static int save_guest_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector,
					 struct desc_struct *seg_desc)
{
3685
	gpa_t gpa;
3686 3687 3688
	struct descriptor_table dtable;
	u16 index = selector >> 3;

A
Amit Shah 已提交
3689
	get_segment_descriptor_dtable(vcpu, selector, &dtable);
3690 3691 3692

	if (dtable.limit < index * 8 + 7)
		return 1;
3693 3694 3695
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, dtable.base);
	gpa += index * 8;
	return kvm_write_guest(vcpu->kvm, gpa, seg_desc, 8);
3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706
}

static u32 get_tss_base_addr(struct kvm_vcpu *vcpu,
			     struct desc_struct *seg_desc)
{
	u32 base_addr;

	base_addr = seg_desc->base0;
	base_addr |= (seg_desc->base1 << 16);
	base_addr |= (seg_desc->base2 << 24);

3707
	return vcpu->arch.mmu.gva_to_gpa(vcpu, base_addr);
3708 3709 3710 3711 3712 3713
}

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

3714
	kvm_get_segment(vcpu, &kvm_seg, seg);
3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
	return kvm_seg.selector;
}

static int load_segment_descriptor_to_kvm_desct(struct kvm_vcpu *vcpu,
						u16 selector,
						struct kvm_segment *kvm_seg)
{
	struct desc_struct seg_desc;

	if (load_guest_segment_descriptor(vcpu, selector, &seg_desc))
		return 1;
	seg_desct_to_kvm_desct(&seg_desc, selector, kvm_seg);
	return 0;
}

3730
static int kvm_load_realmode_segment(struct kvm_vcpu *vcpu, u16 selector, int seg)
3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749
{
	struct kvm_segment segvar = {
		.base = selector << 4,
		.limit = 0xffff,
		.selector = selector,
		.type = 3,
		.present = 1,
		.dpl = 3,
		.db = 0,
		.s = 1,
		.l = 0,
		.g = 0,
		.avl = 0,
		.unusable = 0,
	};
	kvm_x86_ops->set_segment(vcpu, &segvar, seg);
	return 0;
}

3750 3751
int kvm_load_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector,
				int type_bits, int seg)
3752 3753 3754
{
	struct kvm_segment kvm_seg;

3755 3756
	if (!(vcpu->arch.cr0 & X86_CR0_PE))
		return kvm_load_realmode_segment(vcpu, selector, seg);
3757 3758 3759 3760 3761 3762 3763 3764 3765
	if (load_segment_descriptor_to_kvm_desct(vcpu, selector, &kvm_seg))
		return 1;
	kvm_seg.type |= type_bits;

	if (seg != VCPU_SREG_SS && seg != VCPU_SREG_CS &&
	    seg != VCPU_SREG_LDTR)
		if (!kvm_seg.s)
			kvm_seg.unusable = 1;

3766
	kvm_set_segment(vcpu, &kvm_seg, seg);
3767 3768 3769 3770 3771 3772 3773
	return 0;
}

static void save_state_to_tss32(struct kvm_vcpu *vcpu,
				struct tss_segment_32 *tss)
{
	tss->cr3 = vcpu->arch.cr3;
3774
	tss->eip = kvm_rip_read(vcpu);
3775
	tss->eflags = kvm_x86_ops->get_rflags(vcpu);
3776 3777 3778 3779 3780 3781 3782 3783
	tss->eax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	tss->ecx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	tss->edx = kvm_register_read(vcpu, VCPU_REGS_RDX);
	tss->ebx = kvm_register_read(vcpu, VCPU_REGS_RBX);
	tss->esp = kvm_register_read(vcpu, VCPU_REGS_RSP);
	tss->ebp = kvm_register_read(vcpu, VCPU_REGS_RBP);
	tss->esi = kvm_register_read(vcpu, VCPU_REGS_RSI);
	tss->edi = kvm_register_read(vcpu, VCPU_REGS_RDI);
3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797
	tss->es = get_segment_selector(vcpu, VCPU_SREG_ES);
	tss->cs = get_segment_selector(vcpu, VCPU_SREG_CS);
	tss->ss = get_segment_selector(vcpu, VCPU_SREG_SS);
	tss->ds = get_segment_selector(vcpu, VCPU_SREG_DS);
	tss->fs = get_segment_selector(vcpu, VCPU_SREG_FS);
	tss->gs = get_segment_selector(vcpu, VCPU_SREG_GS);
	tss->ldt_selector = get_segment_selector(vcpu, VCPU_SREG_LDTR);
}

static int load_state_from_tss32(struct kvm_vcpu *vcpu,
				  struct tss_segment_32 *tss)
{
	kvm_set_cr3(vcpu, tss->cr3);

3798
	kvm_rip_write(vcpu, tss->eip);
3799 3800
	kvm_x86_ops->set_rflags(vcpu, tss->eflags | 2);

3801 3802 3803 3804 3805 3806 3807 3808
	kvm_register_write(vcpu, VCPU_REGS_RAX, tss->eax);
	kvm_register_write(vcpu, VCPU_REGS_RCX, tss->ecx);
	kvm_register_write(vcpu, VCPU_REGS_RDX, tss->edx);
	kvm_register_write(vcpu, VCPU_REGS_RBX, tss->ebx);
	kvm_register_write(vcpu, VCPU_REGS_RSP, tss->esp);
	kvm_register_write(vcpu, VCPU_REGS_RBP, tss->ebp);
	kvm_register_write(vcpu, VCPU_REGS_RSI, tss->esi);
	kvm_register_write(vcpu, VCPU_REGS_RDI, tss->edi);
3809

3810
	if (kvm_load_segment_descriptor(vcpu, tss->ldt_selector, 0, VCPU_SREG_LDTR))
3811 3812
		return 1;

3813
	if (kvm_load_segment_descriptor(vcpu, tss->es, 1, VCPU_SREG_ES))
3814 3815
		return 1;

3816
	if (kvm_load_segment_descriptor(vcpu, tss->cs, 9, VCPU_SREG_CS))
3817 3818
		return 1;

3819
	if (kvm_load_segment_descriptor(vcpu, tss->ss, 1, VCPU_SREG_SS))
3820 3821
		return 1;

3822
	if (kvm_load_segment_descriptor(vcpu, tss->ds, 1, VCPU_SREG_DS))
3823 3824
		return 1;

3825
	if (kvm_load_segment_descriptor(vcpu, tss->fs, 1, VCPU_SREG_FS))
3826 3827
		return 1;

3828
	if (kvm_load_segment_descriptor(vcpu, tss->gs, 1, VCPU_SREG_GS))
3829 3830 3831 3832 3833 3834 3835
		return 1;
	return 0;
}

static void save_state_to_tss16(struct kvm_vcpu *vcpu,
				struct tss_segment_16 *tss)
{
3836
	tss->ip = kvm_rip_read(vcpu);
3837
	tss->flag = kvm_x86_ops->get_rflags(vcpu);
3838 3839 3840 3841 3842 3843 3844 3845
	tss->ax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	tss->cx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	tss->dx = kvm_register_read(vcpu, VCPU_REGS_RDX);
	tss->bx = kvm_register_read(vcpu, VCPU_REGS_RBX);
	tss->sp = kvm_register_read(vcpu, VCPU_REGS_RSP);
	tss->bp = kvm_register_read(vcpu, VCPU_REGS_RBP);
	tss->si = kvm_register_read(vcpu, VCPU_REGS_RSI);
	tss->di = kvm_register_read(vcpu, VCPU_REGS_RDI);
3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857

	tss->es = get_segment_selector(vcpu, VCPU_SREG_ES);
	tss->cs = get_segment_selector(vcpu, VCPU_SREG_CS);
	tss->ss = get_segment_selector(vcpu, VCPU_SREG_SS);
	tss->ds = get_segment_selector(vcpu, VCPU_SREG_DS);
	tss->ldt = get_segment_selector(vcpu, VCPU_SREG_LDTR);
	tss->prev_task_link = get_segment_selector(vcpu, VCPU_SREG_TR);
}

static int load_state_from_tss16(struct kvm_vcpu *vcpu,
				 struct tss_segment_16 *tss)
{
3858
	kvm_rip_write(vcpu, tss->ip);
3859
	kvm_x86_ops->set_rflags(vcpu, tss->flag | 2);
3860 3861 3862 3863 3864 3865 3866 3867
	kvm_register_write(vcpu, VCPU_REGS_RAX, tss->ax);
	kvm_register_write(vcpu, VCPU_REGS_RCX, tss->cx);
	kvm_register_write(vcpu, VCPU_REGS_RDX, tss->dx);
	kvm_register_write(vcpu, VCPU_REGS_RBX, tss->bx);
	kvm_register_write(vcpu, VCPU_REGS_RSP, tss->sp);
	kvm_register_write(vcpu, VCPU_REGS_RBP, tss->bp);
	kvm_register_write(vcpu, VCPU_REGS_RSI, tss->si);
	kvm_register_write(vcpu, VCPU_REGS_RDI, tss->di);
3868

3869
	if (kvm_load_segment_descriptor(vcpu, tss->ldt, 0, VCPU_SREG_LDTR))
3870 3871
		return 1;

3872
	if (kvm_load_segment_descriptor(vcpu, tss->es, 1, VCPU_SREG_ES))
3873 3874
		return 1;

3875
	if (kvm_load_segment_descriptor(vcpu, tss->cs, 9, VCPU_SREG_CS))
3876 3877
		return 1;

3878
	if (kvm_load_segment_descriptor(vcpu, tss->ss, 1, VCPU_SREG_SS))
3879 3880
		return 1;

3881
	if (kvm_load_segment_descriptor(vcpu, tss->ds, 1, VCPU_SREG_DS))
3882 3883 3884 3885
		return 1;
	return 0;
}

3886
static int kvm_task_switch_16(struct kvm_vcpu *vcpu, u16 tss_selector,
3887 3888
			      u16 old_tss_sel, u32 old_tss_base,
			      struct desc_struct *nseg_desc)
3889 3890 3891 3892
{
	struct tss_segment_16 tss_segment_16;
	int ret = 0;

3893 3894
	if (kvm_read_guest(vcpu->kvm, old_tss_base, &tss_segment_16,
			   sizeof tss_segment_16))
3895 3896 3897 3898
		goto out;

	save_state_to_tss16(vcpu, &tss_segment_16);

3899 3900
	if (kvm_write_guest(vcpu->kvm, old_tss_base, &tss_segment_16,
			    sizeof tss_segment_16))
3901
		goto out;
3902 3903 3904 3905 3906

	if (kvm_read_guest(vcpu->kvm, get_tss_base_addr(vcpu, nseg_desc),
			   &tss_segment_16, sizeof tss_segment_16))
		goto out;

3907 3908 3909 3910 3911 3912 3913 3914 3915 3916
	if (old_tss_sel != 0xffff) {
		tss_segment_16.prev_task_link = old_tss_sel;

		if (kvm_write_guest(vcpu->kvm,
				    get_tss_base_addr(vcpu, nseg_desc),
				    &tss_segment_16.prev_task_link,
				    sizeof tss_segment_16.prev_task_link))
			goto out;
	}

3917 3918 3919 3920 3921 3922 3923 3924
	if (load_state_from_tss16(vcpu, &tss_segment_16))
		goto out;

	ret = 1;
out:
	return ret;
}

3925
static int kvm_task_switch_32(struct kvm_vcpu *vcpu, u16 tss_selector,
3926
		       u16 old_tss_sel, u32 old_tss_base,
3927 3928 3929 3930 3931
		       struct desc_struct *nseg_desc)
{
	struct tss_segment_32 tss_segment_32;
	int ret = 0;

3932 3933
	if (kvm_read_guest(vcpu->kvm, old_tss_base, &tss_segment_32,
			   sizeof tss_segment_32))
3934 3935 3936 3937
		goto out;

	save_state_to_tss32(vcpu, &tss_segment_32);

3938 3939 3940 3941 3942 3943
	if (kvm_write_guest(vcpu->kvm, old_tss_base, &tss_segment_32,
			    sizeof tss_segment_32))
		goto out;

	if (kvm_read_guest(vcpu->kvm, get_tss_base_addr(vcpu, nseg_desc),
			   &tss_segment_32, sizeof tss_segment_32))
3944
		goto out;
3945

3946 3947 3948 3949 3950 3951 3952 3953 3954 3955
	if (old_tss_sel != 0xffff) {
		tss_segment_32.prev_task_link = old_tss_sel;

		if (kvm_write_guest(vcpu->kvm,
				    get_tss_base_addr(vcpu, nseg_desc),
				    &tss_segment_32.prev_task_link,
				    sizeof tss_segment_32.prev_task_link))
			goto out;
	}

3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969
	if (load_state_from_tss32(vcpu, &tss_segment_32))
		goto out;

	ret = 1;
out:
	return ret;
}

int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int reason)
{
	struct kvm_segment tr_seg;
	struct desc_struct cseg_desc;
	struct desc_struct nseg_desc;
	int ret = 0;
3970 3971
	u32 old_tss_base = get_segment_base(vcpu, VCPU_SREG_TR);
	u16 old_tss_sel = get_segment_selector(vcpu, VCPU_SREG_TR);
3972

3973
	old_tss_base = vcpu->arch.mmu.gva_to_gpa(vcpu, old_tss_base);
3974

3975 3976 3977
	/* FIXME: Handle errors. Failure to read either TSS or their
	 * descriptors should generate a pagefault.
	 */
3978 3979 3980
	if (load_guest_segment_descriptor(vcpu, tss_selector, &nseg_desc))
		goto out;

3981
	if (load_guest_segment_descriptor(vcpu, old_tss_sel, &cseg_desc))
3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999
		goto out;

	if (reason != TASK_SWITCH_IRET) {
		int cpl;

		cpl = kvm_x86_ops->get_cpl(vcpu);
		if ((tss_selector & 3) > nseg_desc.dpl || cpl > nseg_desc.dpl) {
			kvm_queue_exception_e(vcpu, GP_VECTOR, 0);
			return 1;
		}
	}

	if (!nseg_desc.p || (nseg_desc.limit0 | nseg_desc.limit << 16) < 0x67) {
		kvm_queue_exception_e(vcpu, TS_VECTOR, tss_selector & 0xfffc);
		return 1;
	}

	if (reason == TASK_SWITCH_IRET || reason == TASK_SWITCH_JMP) {
4000
		cseg_desc.type &= ~(1 << 1); //clear the B flag
4001
		save_guest_segment_descriptor(vcpu, old_tss_sel, &cseg_desc);
4002 4003 4004 4005 4006 4007 4008
	}

	if (reason == TASK_SWITCH_IRET) {
		u32 eflags = kvm_x86_ops->get_rflags(vcpu);
		kvm_x86_ops->set_rflags(vcpu, eflags & ~X86_EFLAGS_NT);
	}

4009 4010 4011 4012
	/* set back link to prev task only if NT bit is set in eflags
	   note that old_tss_sel is not used afetr this point */
	if (reason != TASK_SWITCH_CALL && reason != TASK_SWITCH_GATE)
		old_tss_sel = 0xffff;
4013

4014 4015 4016 4017 4018
	/* set back link to prev task only if NT bit is set in eflags
	   note that old_tss_sel is not used afetr this point */
	if (reason != TASK_SWITCH_CALL && reason != TASK_SWITCH_GATE)
		old_tss_sel = 0xffff;

4019
	if (nseg_desc.type & 8)
4020 4021
		ret = kvm_task_switch_32(vcpu, tss_selector, old_tss_sel,
					 old_tss_base, &nseg_desc);
4022
	else
4023 4024
		ret = kvm_task_switch_16(vcpu, tss_selector, old_tss_sel,
					 old_tss_base, &nseg_desc);
4025 4026 4027 4028 4029 4030 4031

	if (reason == TASK_SWITCH_CALL || reason == TASK_SWITCH_GATE) {
		u32 eflags = kvm_x86_ops->get_rflags(vcpu);
		kvm_x86_ops->set_rflags(vcpu, eflags | X86_EFLAGS_NT);
	}

	if (reason != TASK_SWITCH_IRET) {
4032
		nseg_desc.type |= (1 << 1);
4033 4034 4035 4036 4037 4038 4039
		save_guest_segment_descriptor(vcpu, tss_selector,
					      &nseg_desc);
	}

	kvm_x86_ops->set_cr0(vcpu, vcpu->arch.cr0 | X86_CR0_TS);
	seg_desct_to_kvm_desct(&nseg_desc, tss_selector, &tr_seg);
	tr_seg.type = 11;
4040
	kvm_set_segment(vcpu, &tr_seg, VCPU_SREG_TR);
4041 4042 4043 4044 4045
out:
	return ret;
}
EXPORT_SYMBOL_GPL(kvm_task_switch);

4046 4047 4048 4049
int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
				  struct kvm_sregs *sregs)
{
	int mmu_reset_needed = 0;
G
Gleb Natapov 已提交
4050
	int pending_vec, max_bits;
4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061
	struct descriptor_table dt;

	vcpu_load(vcpu);

	dt.limit = sregs->idt.limit;
	dt.base = sregs->idt.base;
	kvm_x86_ops->set_idt(vcpu, &dt);
	dt.limit = sregs->gdt.limit;
	dt.base = sregs->gdt.base;
	kvm_x86_ops->set_gdt(vcpu, &dt);

4062 4063
	vcpu->arch.cr2 = sregs->cr2;
	mmu_reset_needed |= vcpu->arch.cr3 != sregs->cr3;
4064 4065 4066 4067 4068 4069 4070

	down_read(&vcpu->kvm->slots_lock);
	if (gfn_to_memslot(vcpu->kvm, sregs->cr3 >> PAGE_SHIFT))
		vcpu->arch.cr3 = sregs->cr3;
	else
		set_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests);
	up_read(&vcpu->kvm->slots_lock);
4071

4072
	kvm_set_cr8(vcpu, sregs->cr8);
4073

4074
	mmu_reset_needed |= vcpu->arch.shadow_efer != sregs->efer;
4075 4076 4077 4078 4079
	kvm_x86_ops->set_efer(vcpu, sregs->efer);
	kvm_set_apic_base(vcpu, sregs->apic_base);

	kvm_x86_ops->decache_cr4_guest_bits(vcpu);

4080
	mmu_reset_needed |= vcpu->arch.cr0 != sregs->cr0;
4081
	kvm_x86_ops->set_cr0(vcpu, sregs->cr0);
4082
	vcpu->arch.cr0 = sregs->cr0;
4083

4084
	mmu_reset_needed |= vcpu->arch.cr4 != sregs->cr4;
4085 4086
	kvm_x86_ops->set_cr4(vcpu, sregs->cr4);
	if (!is_long_mode(vcpu) && is_pae(vcpu))
4087
		load_pdptrs(vcpu, vcpu->arch.cr3);
4088 4089 4090 4091

	if (mmu_reset_needed)
		kvm_mmu_reset_context(vcpu);

G
Gleb Natapov 已提交
4092 4093 4094 4095
	max_bits = (sizeof sregs->interrupt_bitmap) << 3;
	pending_vec = find_first_bit(
		(const unsigned long *)sregs->interrupt_bitmap, max_bits);
	if (pending_vec < max_bits) {
4096
		kvm_queue_interrupt(vcpu, pending_vec, false);
G
Gleb Natapov 已提交
4097 4098 4099
		pr_debug("Set back pending irq %d\n", pending_vec);
		if (irqchip_in_kernel(vcpu->kvm))
			kvm_pic_clear_isr_ack(vcpu->kvm);
4100 4101
	}

4102 4103 4104 4105 4106 4107
	kvm_set_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
	kvm_set_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
	kvm_set_segment(vcpu, &sregs->es, VCPU_SREG_ES);
	kvm_set_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
	kvm_set_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
	kvm_set_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
4108

4109 4110
	kvm_set_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_set_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
4111

M
Marcelo Tosatti 已提交
4112 4113 4114 4115 4116 4117
	/* Older userspace won't unhalt the vcpu on reset. */
	if (vcpu->vcpu_id == 0 && kvm_rip_read(vcpu) == 0xfff0 &&
	    sregs->cs.selector == 0xf000 && sregs->cs.base == 0xffff0000 &&
	    !(vcpu->arch.cr0 & X86_CR0_PE))
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;

4118 4119 4120 4121 4122
	vcpu_put(vcpu);

	return 0;
}

J
Jan Kiszka 已提交
4123 4124
int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
					struct kvm_guest_debug *dbg)
4125
{
4126
	int i, r;
4127 4128 4129

	vcpu_load(vcpu);

4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141
	if ((dbg->control & (KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_HW_BP)) ==
	    (KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_HW_BP)) {
		for (i = 0; i < KVM_NR_DB_REGS; ++i)
			vcpu->arch.eff_db[i] = dbg->arch.debugreg[i];
		vcpu->arch.switch_db_regs =
			(dbg->arch.debugreg[7] & DR7_BP_EN_MASK);
	} else {
		for (i = 0; i < KVM_NR_DB_REGS; i++)
			vcpu->arch.eff_db[i] = vcpu->arch.db[i];
		vcpu->arch.switch_db_regs = (vcpu->arch.dr7 & DR7_BP_EN_MASK);
	}

4142 4143
	r = kvm_x86_ops->set_guest_debug(vcpu, dbg);

J
Jan Kiszka 已提交
4144 4145 4146 4147 4148
	if (dbg->control & KVM_GUESTDBG_INJECT_DB)
		kvm_queue_exception(vcpu, DB_VECTOR);
	else if (dbg->control & KVM_GUESTDBG_INJECT_BP)
		kvm_queue_exception(vcpu, BP_VECTOR);

4149 4150 4151 4152 4153
	vcpu_put(vcpu);

	return r;
}

4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174
/*
 * fxsave fpu state.  Taken from x86_64/processor.h.  To be killed when
 * we have asm/x86/processor.h
 */
struct fxsave {
	u16	cwd;
	u16	swd;
	u16	twd;
	u16	fop;
	u64	rip;
	u64	rdp;
	u32	mxcsr;
	u32	mxcsr_mask;
	u32	st_space[32];	/* 8*16 bytes for each FP-reg = 128 bytes */
#ifdef CONFIG_X86_64
	u32	xmm_space[64];	/* 16*16 bytes for each XMM-reg = 256 bytes */
#else
	u32	xmm_space[32];	/* 8*16 bytes for each XMM-reg = 128 bytes */
#endif
};

4175 4176 4177 4178 4179 4180 4181 4182 4183 4184
/*
 * Translate a guest virtual address to a guest physical address.
 */
int kvm_arch_vcpu_ioctl_translate(struct kvm_vcpu *vcpu,
				    struct kvm_translation *tr)
{
	unsigned long vaddr = tr->linear_address;
	gpa_t gpa;

	vcpu_load(vcpu);
4185
	down_read(&vcpu->kvm->slots_lock);
4186
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, vaddr);
4187
	up_read(&vcpu->kvm->slots_lock);
4188 4189 4190 4191 4192 4193 4194 4195 4196
	tr->physical_address = gpa;
	tr->valid = gpa != UNMAPPED_GVA;
	tr->writeable = 1;
	tr->usermode = 0;
	vcpu_put(vcpu);

	return 0;
}

4197 4198
int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
4199
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218

	vcpu_load(vcpu);

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

	vcpu_put(vcpu);

	return 0;
}

int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
4219
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240

	vcpu_load(vcpu);

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

	vcpu_put(vcpu);

	return 0;
}

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

4241 4242 4243 4244 4245 4246 4247
	/*
	 * Touch the fpu the first time in non atomic context as if
	 * this is the first fpu instruction the exception handler
	 * will fire before the instruction returns and it'll have to
	 * allocate ram with GFP_KERNEL.
	 */
	if (!used_math())
4248
		kvm_fx_save(&vcpu->arch.host_fx_image);
4249

4250 4251
	/* Initialize guest FPU by resetting ours and saving into guest's */
	preempt_disable();
4252 4253 4254 4255
	kvm_fx_save(&vcpu->arch.host_fx_image);
	kvm_fx_finit();
	kvm_fx_save(&vcpu->arch.guest_fx_image);
	kvm_fx_restore(&vcpu->arch.host_fx_image);
4256 4257
	preempt_enable();

4258
	vcpu->arch.cr0 |= X86_CR0_ET;
4259
	after_mxcsr_mask = offsetof(struct i387_fxsave_struct, st_space);
4260 4261
	vcpu->arch.guest_fx_image.mxcsr = 0x1f80;
	memset((void *)&vcpu->arch.guest_fx_image + after_mxcsr_mask,
4262 4263 4264 4265 4266 4267 4268 4269 4270 4271
	       0, sizeof(struct i387_fxsave_struct) - after_mxcsr_mask);
}
EXPORT_SYMBOL_GPL(fx_init);

void kvm_load_guest_fpu(struct kvm_vcpu *vcpu)
{
	if (!vcpu->fpu_active || vcpu->guest_fpu_loaded)
		return;

	vcpu->guest_fpu_loaded = 1;
4272 4273
	kvm_fx_save(&vcpu->arch.host_fx_image);
	kvm_fx_restore(&vcpu->arch.guest_fx_image);
4274 4275 4276 4277 4278 4279 4280 4281 4282
}
EXPORT_SYMBOL_GPL(kvm_load_guest_fpu);

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

	vcpu->guest_fpu_loaded = 0;
4283 4284
	kvm_fx_save(&vcpu->arch.guest_fx_image);
	kvm_fx_restore(&vcpu->arch.host_fx_image);
A
Avi Kivity 已提交
4285
	++vcpu->stat.fpu_reload;
4286 4287
}
EXPORT_SYMBOL_GPL(kvm_put_guest_fpu);
4288 4289 4290

void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
4291 4292 4293 4294 4295
	if (vcpu->arch.time_page) {
		kvm_release_page_dirty(vcpu->arch.time_page);
		vcpu->arch.time_page = NULL;
	}

4296 4297 4298 4299 4300 4301
	kvm_x86_ops->vcpu_free(vcpu);
}

struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm,
						unsigned int id)
{
4302 4303
	return kvm_x86_ops->vcpu_create(kvm, id);
}
4304

4305 4306 4307
int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
{
	int r;
4308 4309

	/* We do fxsave: this must be aligned. */
4310
	BUG_ON((unsigned long)&vcpu->arch.host_fx_image & 0xF);
4311

S
Sheng Yang 已提交
4312
	vcpu->arch.mtrr_state.have_fixed = 1;
4313 4314 4315 4316 4317 4318 4319 4320
	vcpu_load(vcpu);
	r = kvm_arch_vcpu_reset(vcpu);
	if (r == 0)
		r = kvm_mmu_setup(vcpu);
	vcpu_put(vcpu);
	if (r < 0)
		goto free_vcpu;

4321
	return 0;
4322 4323
free_vcpu:
	kvm_x86_ops->vcpu_free(vcpu);
4324
	return r;
4325 4326
}

4327
void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
4328 4329 4330 4331 4332 4333 4334 4335 4336 4337
{
	vcpu_load(vcpu);
	kvm_mmu_unload(vcpu);
	vcpu_put(vcpu);

	kvm_x86_ops->vcpu_free(vcpu);
}

int kvm_arch_vcpu_reset(struct kvm_vcpu *vcpu)
{
4338 4339 4340
	vcpu->arch.nmi_pending = false;
	vcpu->arch.nmi_injected = false;

4341 4342 4343 4344 4345
	vcpu->arch.switch_db_regs = 0;
	memset(vcpu->arch.db, 0, sizeof(vcpu->arch.db));
	vcpu->arch.dr6 = DR6_FIXED_1;
	vcpu->arch.dr7 = DR7_FIXED_1;

4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382
	return kvm_x86_ops->vcpu_reset(vcpu);
}

void kvm_arch_hardware_enable(void *garbage)
{
	kvm_x86_ops->hardware_enable(garbage);
}

void kvm_arch_hardware_disable(void *garbage)
{
	kvm_x86_ops->hardware_disable(garbage);
}

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

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

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

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

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

4383
	vcpu->arch.mmu.root_hpa = INVALID_PAGE;
4384
	if (!irqchip_in_kernel(kvm) || vcpu->vcpu_id == 0)
4385
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
4386
	else
4387
		vcpu->arch.mp_state = KVM_MP_STATE_UNINITIALIZED;
4388 4389 4390 4391 4392 4393

	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
	if (!page) {
		r = -ENOMEM;
		goto fail;
	}
4394
	vcpu->arch.pio_data = page_address(page);
4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410

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

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

	return 0;

fail_mmu_destroy:
	kvm_mmu_destroy(vcpu);
fail_free_pio_data:
4411
	free_page((unsigned long)vcpu->arch.pio_data);
4412 4413 4414 4415 4416 4417 4418
fail:
	return r;
}

void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
{
	kvm_free_lapic(vcpu);
4419
	down_read(&vcpu->kvm->slots_lock);
4420
	kvm_mmu_destroy(vcpu);
4421
	up_read(&vcpu->kvm->slots_lock);
4422
	free_page((unsigned long)vcpu->arch.pio_data);
4423
}
4424 4425 4426 4427 4428 4429 4430 4431

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

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

4432
	INIT_LIST_HEAD(&kvm->arch.active_mmu_pages);
B
Ben-Ami Yassour 已提交
4433
	INIT_LIST_HEAD(&kvm->arch.assigned_dev_head);
4434

4435 4436 4437
	/* Reserve bit 0 of irq_sources_bitmap for userspace irq source */
	set_bit(KVM_USERSPACE_IRQ_SOURCE_ID, &kvm->arch.irq_sources_bitmap);

4438 4439
	rdtscll(kvm->arch.vm_init_tsc);

4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468
	return kvm;
}

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

static void kvm_free_vcpus(struct kvm *kvm)
{
	unsigned int i;

	/*
	 * Unpin any mmu pages first.
	 */
	for (i = 0; i < KVM_MAX_VCPUS; ++i)
		if (kvm->vcpus[i])
			kvm_unload_vcpu_mmu(kvm->vcpus[i]);
	for (i = 0; i < KVM_MAX_VCPUS; ++i) {
		if (kvm->vcpus[i]) {
			kvm_arch_vcpu_free(kvm->vcpus[i]);
			kvm->vcpus[i] = NULL;
		}
	}

}

4469 4470
void kvm_arch_sync_events(struct kvm *kvm)
{
4471
	kvm_free_all_assigned_devices(kvm);
4472 4473
}

4474 4475
void kvm_arch_destroy_vm(struct kvm *kvm)
{
4476
	kvm_iommu_unmap_guest(kvm);
S
Sheng Yang 已提交
4477
	kvm_free_pit(kvm);
4478 4479
	kfree(kvm->arch.vpic);
	kfree(kvm->arch.vioapic);
4480 4481
	kvm_free_vcpus(kvm);
	kvm_free_physmem(kvm);
4482 4483
	if (kvm->arch.apic_access_page)
		put_page(kvm->arch.apic_access_page);
4484 4485
	if (kvm->arch.ept_identity_pagetable)
		put_page(kvm->arch.ept_identity_pagetable);
4486 4487
	kfree(kvm);
}
4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501

int kvm_arch_set_memory_region(struct kvm *kvm,
				struct kvm_userspace_memory_region *mem,
				struct kvm_memory_slot old,
				int user_alloc)
{
	int npages = mem->memory_size >> PAGE_SHIFT;
	struct kvm_memory_slot *memslot = &kvm->memslots[mem->slot];

	/*To keep backward compatibility with older userspace,
	 *x86 needs to hanlde !user_alloc case.
	 */
	if (!user_alloc) {
		if (npages && !old.rmap) {
4502 4503
			unsigned long userspace_addr;

4504
			down_write(&current->mm->mmap_sem);
4505 4506 4507
			userspace_addr = do_mmap(NULL, 0,
						 npages * PAGE_SIZE,
						 PROT_READ | PROT_WRITE,
4508
						 MAP_PRIVATE | MAP_ANONYMOUS,
4509
						 0);
4510
			up_write(&current->mm->mmap_sem);
4511

4512 4513 4514 4515 4516 4517 4518
			if (IS_ERR((void *)userspace_addr))
				return PTR_ERR((void *)userspace_addr);

			/* set userspace_addr atomically for kvm_hva_to_rmapp */
			spin_lock(&kvm->mmu_lock);
			memslot->userspace_addr = userspace_addr;
			spin_unlock(&kvm->mmu_lock);
4519 4520 4521 4522
		} else {
			if (!old.user_alloc && old.rmap) {
				int ret;

4523
				down_write(&current->mm->mmap_sem);
4524 4525
				ret = do_munmap(current->mm, old.userspace_addr,
						old.npages * PAGE_SIZE);
4526
				up_write(&current->mm->mmap_sem);
4527 4528 4529 4530 4531 4532 4533 4534
				if (ret < 0)
					printk(KERN_WARNING
				       "kvm_vm_ioctl_set_memory_region: "
				       "failed to munmap memory\n");
			}
		}
	}

4535
	spin_lock(&kvm->mmu_lock);
4536
	if (!kvm->arch.n_requested_mmu_pages) {
4537 4538 4539 4540 4541
		unsigned int nr_mmu_pages = kvm_mmu_calculate_mmu_pages(kvm);
		kvm_mmu_change_mmu_pages(kvm, nr_mmu_pages);
	}

	kvm_mmu_slot_remove_write_access(kvm, mem->slot);
4542
	spin_unlock(&kvm->mmu_lock);
4543 4544 4545 4546
	kvm_flush_remote_tlbs(kvm);

	return 0;
}
4547

4548 4549 4550
void kvm_arch_flush_shadow(struct kvm *kvm)
{
	kvm_mmu_zap_all(kvm);
4551
	kvm_reload_remote_mmus(kvm);
4552 4553
}

4554 4555
int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
{
4556
	return vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE
4557 4558
	       || vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED
	       || vcpu->arch.nmi_pending;
4559
}
4560 4561 4562

void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
{
4563 4564
	int me;
	int cpu = vcpu->cpu;
4565 4566 4567 4568 4569

	if (waitqueue_active(&vcpu->wq)) {
		wake_up_interruptible(&vcpu->wq);
		++vcpu->stat.halt_wakeup;
	}
4570 4571 4572 4573 4574

	me = get_cpu();
	if (cpu != me && (unsigned)cpu < nr_cpu_ids && cpu_online(cpu))
		if (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests))
			smp_send_reschedule(cpu);
4575
	put_cpu();
4576
}
4577 4578 4579 4580 4581

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