x86.c 105.9 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
#include <linux/kvm.h>
#include <linux/fs.h>
B
Ben-Ami Yassour 已提交
33
#include <linux/pci.h>
34
#include <linux/vmalloc.h>
35
#include <linux/module.h>
36
#include <linux/mman.h>
37
#include <linux/highmem.h>
B
Ben-Ami Yassour 已提交
38
#include <linux/intel-iommu.h>
39 40

#include <asm/uaccess.h>
41
#include <asm/msr.h>
42
#include <asm/desc.h>
43

44
#define MAX_IO_MSRS 256
45 46 47 48 49 50 51 52 53 54 55
#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)
56 57 58 59 60 61 62 63 64
/* 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
65

66 67
#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
68

69 70 71
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
				    struct kvm_cpuid_entry2 __user *entries);

72
struct kvm_x86_ops *kvm_x86_ops;
73
EXPORT_SYMBOL_GPL(kvm_x86_ops);
74

75
struct kvm_stats_debugfs_item debugfs_entries[] = {
76 77 78 79 80 81 82 83 84
	{ "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) },
85
	{ "nmi_window", VCPU_STAT(nmi_window_exits) },
86 87
	{ "halt_exits", VCPU_STAT(halt_exits) },
	{ "halt_wakeup", VCPU_STAT(halt_wakeup) },
A
Amit Shah 已提交
88
	{ "hypercalls", VCPU_STAT(hypercalls) },
89 90 91 92 93 94 95
	{ "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) },
96
	{ "irq_injections", VCPU_STAT(irq_injections) },
A
Avi Kivity 已提交
97 98 99 100 101 102
	{ "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
103
	{ "mmu_cache_miss", VM_STAT(mmu_cache_miss) },
104
	{ "remote_tlb_flush", VM_STAT(remote_tlb_flush) },
M
Marcelo Tosatti 已提交
105
	{ "largepages", VM_STAT(lpages) },
106 107 108
	{ NULL }
};

109
static struct kvm_assigned_dev_kernel *kvm_find_assigned_dev(struct list_head *head,
B
Ben-Ami Yassour 已提交
110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
						      int assigned_dev_id)
{
	struct list_head *ptr;
	struct kvm_assigned_dev_kernel *match;

	list_for_each(ptr, head) {
		match = list_entry(ptr, struct kvm_assigned_dev_kernel, list);
		if (match->assigned_dev_id == assigned_dev_id)
			return match;
	}
	return NULL;
}

static void kvm_assigned_dev_interrupt_work_handler(struct work_struct *work)
{
	struct kvm_assigned_dev_kernel *assigned_dev;

	assigned_dev = container_of(work, struct kvm_assigned_dev_kernel,
				    interrupt_work);

	/* This is taken to safely inject irq inside the guest. When
	 * the interrupt injection (or the ioapic code) uses a
	 * finer-grained lock, update this
	 */
	mutex_lock(&assigned_dev->kvm->lock);
	kvm_set_irq(assigned_dev->kvm,
		    assigned_dev->guest_irq, 1);
	mutex_unlock(&assigned_dev->kvm->lock);
	kvm_put_kvm(assigned_dev->kvm);
}

/* FIXME: Implement the OR logic needed to make shared interrupts on
 * this line behave properly
 */
static irqreturn_t kvm_assigned_dev_intr(int irq, void *dev_id)
{
	struct kvm_assigned_dev_kernel *assigned_dev =
		(struct kvm_assigned_dev_kernel *) dev_id;

	kvm_get_kvm(assigned_dev->kvm);
	schedule_work(&assigned_dev->interrupt_work);
	disable_irq_nosync(irq);
	return IRQ_HANDLED;
}

/* Ack the irq line for an assigned device */
static void kvm_assigned_dev_ack_irq(struct kvm_irq_ack_notifier *kian)
{
	struct kvm_assigned_dev_kernel *dev;

	if (kian->gsi == -1)
		return;

	dev = container_of(kian, struct kvm_assigned_dev_kernel,
			   ack_notifier);
	kvm_set_irq(dev->kvm, dev->guest_irq, 0);
	enable_irq(dev->host_irq);
}

169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205
static void kvm_free_assigned_device(struct kvm *kvm,
				     struct kvm_assigned_dev_kernel
				     *assigned_dev)
{
	if (irqchip_in_kernel(kvm) && assigned_dev->irq_requested)
		free_irq(assigned_dev->host_irq, (void *)assigned_dev);

	kvm_unregister_irq_ack_notifier(kvm, &assigned_dev->ack_notifier);

	if (cancel_work_sync(&assigned_dev->interrupt_work))
		/* We had pending work. That means we will have to take
		 * care of kvm_put_kvm.
		 */
		kvm_put_kvm(kvm);

	pci_release_regions(assigned_dev->dev);
	pci_disable_device(assigned_dev->dev);
	pci_dev_put(assigned_dev->dev);

	list_del(&assigned_dev->list);
	kfree(assigned_dev);
}

static void kvm_free_all_assigned_devices(struct kvm *kvm)
{
	struct list_head *ptr, *ptr2;
	struct kvm_assigned_dev_kernel *assigned_dev;

	list_for_each_safe(ptr, ptr2, &kvm->arch.assigned_dev_head) {
		assigned_dev = list_entry(ptr,
					  struct kvm_assigned_dev_kernel,
					  list);

		kvm_free_assigned_device(kvm, assigned_dev);
	}
}

B
Ben-Ami Yassour 已提交
206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
static int kvm_vm_ioctl_assign_irq(struct kvm *kvm,
				   struct kvm_assigned_irq
				   *assigned_irq)
{
	int r = 0;
	struct kvm_assigned_dev_kernel *match;

	mutex_lock(&kvm->lock);

	match = kvm_find_assigned_dev(&kvm->arch.assigned_dev_head,
				      assigned_irq->assigned_dev_id);
	if (!match) {
		mutex_unlock(&kvm->lock);
		return -EINVAL;
	}

	if (match->irq_requested) {
		match->guest_irq = assigned_irq->guest_irq;
		match->ack_notifier.gsi = assigned_irq->guest_irq;
		mutex_unlock(&kvm->lock);
		return 0;
	}

	INIT_WORK(&match->interrupt_work,
		  kvm_assigned_dev_interrupt_work_handler);

	if (irqchip_in_kernel(kvm)) {
233
		if (!capable(CAP_SYS_RAWIO)) {
234 235
			r = -EPERM;
			goto out_release;
236 237
		}

B
Ben-Ami Yassour 已提交
238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
		if (assigned_irq->host_irq)
			match->host_irq = assigned_irq->host_irq;
		else
			match->host_irq = match->dev->irq;
		match->guest_irq = assigned_irq->guest_irq;
		match->ack_notifier.gsi = assigned_irq->guest_irq;
		match->ack_notifier.irq_acked = kvm_assigned_dev_ack_irq;
		kvm_register_irq_ack_notifier(kvm, &match->ack_notifier);

		/* Even though this is PCI, we don't want to use shared
		 * interrupts. Sharing host devices with guest-assigned devices
		 * on the same interrupt line is not a happy situation: there
		 * are going to be long delays in accepting, acking, etc.
		 */
		if (request_irq(match->host_irq, kvm_assigned_dev_intr, 0,
				"kvm_assigned_device", (void *)match)) {
			r = -EIO;
255
			goto out_release;
B
Ben-Ami Yassour 已提交
256 257 258 259 260 261
		}
	}

	match->irq_requested = true;
	mutex_unlock(&kvm->lock);
	return r;
262 263 264 265
out_release:
	mutex_unlock(&kvm->lock);
	kvm_free_assigned_device(kvm, match);
	return r;
B
Ben-Ami Yassour 已提交
266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318
}

static int kvm_vm_ioctl_assign_device(struct kvm *kvm,
				      struct kvm_assigned_pci_dev *assigned_dev)
{
	int r = 0;
	struct kvm_assigned_dev_kernel *match;
	struct pci_dev *dev;

	mutex_lock(&kvm->lock);

	match = kvm_find_assigned_dev(&kvm->arch.assigned_dev_head,
				      assigned_dev->assigned_dev_id);
	if (match) {
		/* device already assigned */
		r = -EINVAL;
		goto out;
	}

	match = kzalloc(sizeof(struct kvm_assigned_dev_kernel), GFP_KERNEL);
	if (match == NULL) {
		printk(KERN_INFO "%s: Couldn't allocate memory\n",
		       __func__);
		r = -ENOMEM;
		goto out;
	}
	dev = pci_get_bus_and_slot(assigned_dev->busnr,
				   assigned_dev->devfn);
	if (!dev) {
		printk(KERN_INFO "%s: host device not found\n", __func__);
		r = -EINVAL;
		goto out_free;
	}
	if (pci_enable_device(dev)) {
		printk(KERN_INFO "%s: Could not enable PCI device\n", __func__);
		r = -EBUSY;
		goto out_put;
	}
	r = pci_request_regions(dev, "kvm_assigned_device");
	if (r) {
		printk(KERN_INFO "%s: Could not get access to device regions\n",
		       __func__);
		goto out_disable;
	}
	match->assigned_dev_id = assigned_dev->assigned_dev_id;
	match->host_busnr = assigned_dev->busnr;
	match->host_devfn = assigned_dev->devfn;
	match->dev = dev;

	match->kvm = kvm;

	list_add(&match->list, &kvm->arch.assigned_dev_head);

B
Ben-Ami Yassour 已提交
319 320 321 322 323 324
	if (assigned_dev->flags & KVM_DEV_ASSIGN_ENABLE_IOMMU) {
		r = kvm_iommu_map_guest(kvm, match);
		if (r)
			goto out_list_del;
	}

B
Ben-Ami Yassour 已提交
325 326 327
out:
	mutex_unlock(&kvm->lock);
	return r;
B
Ben-Ami Yassour 已提交
328 329 330
out_list_del:
	list_del(&match->list);
	pci_release_regions(dev);
B
Ben-Ami Yassour 已提交
331 332 333 334 335 336 337 338 339 340
out_disable:
	pci_disable_device(dev);
out_put:
	pci_dev_put(dev);
out_free:
	kfree(match);
	mutex_unlock(&kvm->lock);
	return r;
}

341 342 343
unsigned long segment_base(u16 selector)
{
	struct descriptor_table gdt;
344
	struct desc_struct *d;
345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
	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);
	}
360 361 362
	d = (struct desc_struct *)(table_base + (selector & ~7));
	v = d->base0 | ((unsigned long)d->base1 << 16) |
		((unsigned long)d->base2 << 24);
363
#ifdef CONFIG_X86_64
364 365
	if (d->s == 0 && (d->type == 2 || d->type == 9 || d->type == 11))
		v |= ((unsigned long)((struct ldttss_desc64 *)d)->base3) << 32;
366 367 368 369 370
#endif
	return v;
}
EXPORT_SYMBOL_GPL(segment_base);

371 372 373
u64 kvm_get_apic_base(struct kvm_vcpu *vcpu)
{
	if (irqchip_in_kernel(vcpu->kvm))
374
		return vcpu->arch.apic_base;
375
	else
376
		return vcpu->arch.apic_base;
377 378 379 380 381 382 383 384 385
}
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
386
		vcpu->arch.apic_base = data;
387 388 389
}
EXPORT_SYMBOL_GPL(kvm_set_apic_base);

390 391
void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr)
{
392 393 394 395
	WARN_ON(vcpu->arch.exception.pending);
	vcpu->arch.exception.pending = true;
	vcpu->arch.exception.has_error_code = false;
	vcpu->arch.exception.nr = nr;
396 397 398
}
EXPORT_SYMBOL_GPL(kvm_queue_exception);

399 400 401 402
void kvm_inject_page_fault(struct kvm_vcpu *vcpu, unsigned long addr,
			   u32 error_code)
{
	++vcpu->stat.pf_guest;
J
Joerg Roedel 已提交
403 404 405 406 407 408 409 410 411 412
	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);
		}
413 414
		return;
	}
415
	vcpu->arch.cr2 = addr;
416 417 418
	kvm_queue_exception_e(vcpu, PF_VECTOR, error_code);
}

419 420 421 422 423 424
void kvm_inject_nmi(struct kvm_vcpu *vcpu)
{
	vcpu->arch.nmi_pending = 1;
}
EXPORT_SYMBOL_GPL(kvm_inject_nmi);

425 426
void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
{
427 428 429 430 431
	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;
432 433 434 435 436
}
EXPORT_SYMBOL_GPL(kvm_queue_exception_e);

static void __queue_exception(struct kvm_vcpu *vcpu)
{
437 438 439
	kvm_x86_ops->queue_exception(vcpu, vcpu->arch.exception.nr,
				     vcpu->arch.exception.has_error_code,
				     vcpu->arch.exception.error_code);
440 441
}

442 443 444 445 446 447 448 449 450
/*
 * 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;
451
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
452 453 454 455 456 457 458 459 460 461 462 463 464 465 466

	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) {
		if ((pdpte[i] & 1) && (pdpte[i] & 0xfffffff0000001e6ull)) {
			ret = 0;
			goto out;
		}
	}
	ret = 1;

467
	memcpy(vcpu->arch.pdptrs, pdpte, sizeof(vcpu->arch.pdptrs));
468 469 470 471
out:

	return ret;
}
472
EXPORT_SYMBOL_GPL(load_pdptrs);
473

474 475
static bool pdptrs_changed(struct kvm_vcpu *vcpu)
{
476
	u64 pdpte[ARRAY_SIZE(vcpu->arch.pdptrs)];
477 478 479 480 481 482
	bool changed = true;
	int r;

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

483
	r = kvm_read_guest(vcpu->kvm, vcpu->arch.cr3 & ~31u, pdpte, sizeof(pdpte));
484 485
	if (r < 0)
		goto out;
486
	changed = memcmp(pdpte, vcpu->arch.pdptrs, sizeof(pdpte)) != 0;
487 488 489 490 491
out:

	return changed;
}

492
void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
493 494 495
{
	if (cr0 & CR0_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr0: 0x%lx #GP, reserved bits 0x%lx\n",
496
		       cr0, vcpu->arch.cr0);
497
		kvm_inject_gp(vcpu, 0);
498 499 500 501 502
		return;
	}

	if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD)) {
		printk(KERN_DEBUG "set_cr0: #GP, CD == 0 && NW == 1\n");
503
		kvm_inject_gp(vcpu, 0);
504 505 506 507 508 509
		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");
510
		kvm_inject_gp(vcpu, 0);
511 512 513 514 515
		return;
	}

	if (!is_paging(vcpu) && (cr0 & X86_CR0_PG)) {
#ifdef CONFIG_X86_64
516
		if ((vcpu->arch.shadow_efer & EFER_LME)) {
517 518 519 520 521
			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");
522
				kvm_inject_gp(vcpu, 0);
523 524 525 526 527 528
				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");
529
				kvm_inject_gp(vcpu, 0);
530 531 532 533 534
				return;

			}
		} else
#endif
535
		if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
536 537
			printk(KERN_DEBUG "set_cr0: #GP, pdptrs "
			       "reserved bits\n");
538
			kvm_inject_gp(vcpu, 0);
539 540 541 542 543 544
			return;
		}

	}

	kvm_x86_ops->set_cr0(vcpu, cr0);
545
	vcpu->arch.cr0 = cr0;
546 547 548 549

	kvm_mmu_reset_context(vcpu);
	return;
}
550
EXPORT_SYMBOL_GPL(kvm_set_cr0);
551

552
void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw)
553
{
554
	kvm_set_cr0(vcpu, (vcpu->arch.cr0 & ~0x0ful) | (msw & 0x0f));
F
Feng (Eric) Liu 已提交
555 556 557
	KVMTRACE_1D(LMSW, vcpu,
		    (u32)((vcpu->arch.cr0 & ~0x0ful) | (msw & 0x0f)),
		    handler);
558
}
559
EXPORT_SYMBOL_GPL(kvm_lmsw);
560

561
void kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
562 563 564
{
	if (cr4 & CR4_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr4: #GP, reserved bits\n");
565
		kvm_inject_gp(vcpu, 0);
566 567 568 569 570 571 572
		return;
	}

	if (is_long_mode(vcpu)) {
		if (!(cr4 & X86_CR4_PAE)) {
			printk(KERN_DEBUG "set_cr4: #GP, clearing PAE while "
			       "in long mode\n");
573
			kvm_inject_gp(vcpu, 0);
574 575 576
			return;
		}
	} else if (is_paging(vcpu) && !is_pae(vcpu) && (cr4 & X86_CR4_PAE)
577
		   && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
578
		printk(KERN_DEBUG "set_cr4: #GP, pdptrs reserved bits\n");
579
		kvm_inject_gp(vcpu, 0);
580 581 582 583 584
		return;
	}

	if (cr4 & X86_CR4_VMXE) {
		printk(KERN_DEBUG "set_cr4: #GP, setting VMXE\n");
585
		kvm_inject_gp(vcpu, 0);
586 587 588
		return;
	}
	kvm_x86_ops->set_cr4(vcpu, cr4);
589
	vcpu->arch.cr4 = cr4;
590 591
	kvm_mmu_reset_context(vcpu);
}
592
EXPORT_SYMBOL_GPL(kvm_set_cr4);
593

594
void kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
595
{
596
	if (cr3 == vcpu->arch.cr3 && !pdptrs_changed(vcpu)) {
597 598 599 600
		kvm_mmu_flush_tlb(vcpu);
		return;
	}

601 602 603
	if (is_long_mode(vcpu)) {
		if (cr3 & CR3_L_MODE_RESERVED_BITS) {
			printk(KERN_DEBUG "set_cr3: #GP, reserved bits\n");
604
			kvm_inject_gp(vcpu, 0);
605 606 607 608 609 610 611
			return;
		}
	} else {
		if (is_pae(vcpu)) {
			if (cr3 & CR3_PAE_RESERVED_BITS) {
				printk(KERN_DEBUG
				       "set_cr3: #GP, reserved bits\n");
612
				kvm_inject_gp(vcpu, 0);
613 614 615 616 617
				return;
			}
			if (is_paging(vcpu) && !load_pdptrs(vcpu, cr3)) {
				printk(KERN_DEBUG "set_cr3: #GP, pdptrs "
				       "reserved bits\n");
618
				kvm_inject_gp(vcpu, 0);
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637
				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)))
638
		kvm_inject_gp(vcpu, 0);
639
	else {
640 641
		vcpu->arch.cr3 = cr3;
		vcpu->arch.mmu.new_cr3(vcpu);
642 643
	}
}
644
EXPORT_SYMBOL_GPL(kvm_set_cr3);
645

646
void kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
647 648 649
{
	if (cr8 & CR8_RESERVED_BITS) {
		printk(KERN_DEBUG "set_cr8: #GP, reserved bits 0x%lx\n", cr8);
650
		kvm_inject_gp(vcpu, 0);
651 652 653 654 655
		return;
	}
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_lapic_set_tpr(vcpu, cr8);
	else
656
		vcpu->arch.cr8 = cr8;
657
}
658
EXPORT_SYMBOL_GPL(kvm_set_cr8);
659

660
unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu)
661 662 663 664
{
	if (irqchip_in_kernel(vcpu->kvm))
		return kvm_lapic_get_cr8(vcpu);
	else
665
		return vcpu->arch.cr8;
666
}
667
EXPORT_SYMBOL_GPL(kvm_get_cr8);
668

669 670 671 672 673 674 675 676 677 678 679 680 681
/*
 * 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
682
	MSR_IA32_TIME_STAMP_COUNTER, MSR_KVM_SYSTEM_TIME, MSR_KVM_WALL_CLOCK,
683
	MSR_IA32_PERF_STATUS,
684 685 686 687 688 689 690 691
};

static unsigned num_msrs_to_save;

static u32 emulated_msrs[] = {
	MSR_IA32_MISC_ENABLE,
};

692 693
static void set_efer(struct kvm_vcpu *vcpu, u64 efer)
{
694
	if (efer & efer_reserved_bits) {
695 696
		printk(KERN_DEBUG "set_efer: 0x%llx #GP, reserved bits\n",
		       efer);
697
		kvm_inject_gp(vcpu, 0);
698 699 700 701
		return;
	}

	if (is_paging(vcpu)
702
	    && (vcpu->arch.shadow_efer & EFER_LME) != (efer & EFER_LME)) {
703
		printk(KERN_DEBUG "set_efer: #GP, change LME while paging\n");
704
		kvm_inject_gp(vcpu, 0);
705 706 707 708 709 710
		return;
	}

	kvm_x86_ops->set_efer(vcpu, efer);

	efer &= ~EFER_LMA;
711
	efer |= vcpu->arch.shadow_efer & EFER_LMA;
712

713
	vcpu->arch.shadow_efer = efer;
714 715
}

716 717 718 719 720 721 722
void kvm_enable_efer_bits(u64 mask)
{
       efer_reserved_bits &= ~mask;
}
EXPORT_SYMBOL_GPL(kvm_enable_efer_bits);


723 724 725 726 727 728 729 730 731 732
/*
 * 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);
}

733 734 735 736 737 738 739 740
/*
 * 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);
}

741 742 743
static void kvm_write_wall_clock(struct kvm *kvm, gpa_t wall_clock)
{
	static int version;
744 745
	struct pvclock_wall_clock wc;
	struct timespec now, sys, boot;
746 747 748 749 750 751 752 753

	if (!wall_clock)
		return;

	version++;

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

754 755 756 757 758 759 760 761 762 763 764 765 766
	/*
	 * 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;
767 768 769 770 771 772 773

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

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

774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
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",
		 __FUNCTION__, tsc_khz, hv_clock->tsc_shift,
		 hv_clock->tsc_to_system_mul);
}

813 814 815 816 817 818 819 820 821 822
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;

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

823 824 825 826 827
	if (unlikely(vcpu->hv_clock_tsc_khz != tsc_khz)) {
		kvm_set_time_scale(tsc_khz, &vcpu->hv_clock);
		vcpu->hv_clock_tsc_khz = tsc_khz;
	}

828 829 830 831 832 833 834 835 836 837 838 839 840 841
	/* 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
842
	 * state, we just increase by 2 at the end.
843
	 */
844
	vcpu->hv_clock.version += 2;
845 846 847 848

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

	memcpy(shared_kaddr + vcpu->time_offset, &vcpu->hv_clock,
849
	       sizeof(vcpu->hv_clock));
850 851 852 853 854 855

	kunmap_atomic(shared_kaddr, KM_USER0);

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

A
Avi Kivity 已提交
856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887
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)
{
	if (!msr_mtrr_valid(msr))
		return 1;

	vcpu->arch.mtrr[msr - 0x200] = data;
	return 0;
}
888 889 890 891 892 893 894 895 896

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",
897
		       __func__, data);
898 899 900
		break;
	case MSR_IA32_MCG_STATUS:
		pr_unimpl(vcpu, "%s: MSR_IA32_MCG_STATUS 0x%llx, nop\n",
901
			__func__, data);
902
		break;
903 904
	case MSR_IA32_MCG_CTL:
		pr_unimpl(vcpu, "%s: MSR_IA32_MCG_CTL 0x%llx, nop\n",
905
			__func__, data);
906
		break;
907 908 909 910 911 912 913 914 915 916 917 918
	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;
919 920 921
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_UCODE_WRITE:
		break;
A
Avi Kivity 已提交
922 923
	case 0x200 ... 0x2ff:
		return set_msr_mtrr(vcpu, msr, data);
924 925 926 927
	case MSR_IA32_APICBASE:
		kvm_set_apic_base(vcpu, data);
		break;
	case MSR_IA32_MISC_ENABLE:
928
		vcpu->arch.ia32_misc_enable_msr = data;
929
		break;
930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959
	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;
		}

		kvm_write_guest_time(vcpu);
		break;
	}
960
	default:
961
		pr_unimpl(vcpu, "unhandled wrmsr: 0x%x data %llx\n", msr, data);
962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978
		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
979 980 981 982 983 984 985 986 987
static int get_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
{
	if (!msr_mtrr_valid(msr))
		return 1;

	*pdata = vcpu->arch.mtrr[msr - 0x200];
	return 0;
}

988 989 990 991 992 993 994 995 996 997 998 999 1000
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:
1001
	case MSR_IA32_MCG_CTL:
1002 1003 1004 1005 1006
	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
1007
	case MSR_IA32_MC0_MISC+20:
1008 1009
	case MSR_IA32_UCODE_REV:
	case MSR_IA32_EBL_CR_POWERON:
1010 1011 1012 1013 1014
	case MSR_IA32_DEBUGCTLMSR:
	case MSR_IA32_LASTBRANCHFROMIP:
	case MSR_IA32_LASTBRANCHTOIP:
	case MSR_IA32_LASTINTFROMIP:
	case MSR_IA32_LASTINTTOIP:
1015 1016
		data = 0;
		break;
A
Avi Kivity 已提交
1017 1018 1019 1020 1021
	case MSR_MTRRcap:
		data = 0x500 | KVM_NR_VAR_MTRR;
		break;
	case 0x200 ... 0x2ff:
		return get_msr_mtrr(vcpu, msr, pdata);
1022 1023 1024 1025 1026 1027 1028
	case 0xcd: /* fsb frequency */
		data = 3;
		break;
	case MSR_IA32_APICBASE:
		data = kvm_get_apic_base(vcpu);
		break;
	case MSR_IA32_MISC_ENABLE:
1029
		data = vcpu->arch.ia32_misc_enable_msr;
1030
		break;
1031 1032 1033 1034 1035 1036
	case MSR_IA32_PERF_STATUS:
		/* TSC increment by tick */
		data = 1000ULL;
		/* CPU multiplier */
		data |= (((uint64_t)4ULL) << 40);
		break;
1037
	case MSR_EFER:
1038
		data = vcpu->arch.shadow_efer;
1039
		break;
1040 1041 1042 1043 1044 1045
	case MSR_KVM_WALL_CLOCK:
		data = vcpu->kvm->arch.wall_clock;
		break;
	case MSR_KVM_SYSTEM_TIME:
		data = vcpu->arch.time;
		break;
1046 1047 1048 1049 1050 1051 1052 1053 1054
	default:
		pr_unimpl(vcpu, "unhandled rdmsr: 0x%x\n", msr);
		return 1;
	}
	*pdata = data;
	return 0;
}
EXPORT_SYMBOL_GPL(kvm_get_msr_common);

1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
/*
 * 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);

1069
	down_read(&vcpu->kvm->slots_lock);
1070 1071 1072
	for (i = 0; i < msrs->nmsrs; ++i)
		if (do_msr(vcpu, entries[i].index, &entries[i].data))
			break;
1073
	up_read(&vcpu->kvm->slots_lock);
1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128

	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;
}

1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
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_USER_MEMORY:
	case KVM_CAP_SET_TSS_ADDR:
1139
	case KVM_CAP_EXT_CPUID:
1140
	case KVM_CAP_CLOCKSOURCE:
S
Sheng Yang 已提交
1141
	case KVM_CAP_PIT:
1142
	case KVM_CAP_NOP_IO_DELAY:
1143
	case KVM_CAP_MP_STATE:
1144
	case KVM_CAP_SYNC_MMU:
1145 1146
		r = 1;
		break;
1147 1148 1149
	case KVM_CAP_COALESCED_MMIO:
		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
		break;
1150 1151 1152
	case KVM_CAP_VAPIC:
		r = !kvm_x86_ops->cpu_has_accelerated_tpr();
		break;
1153 1154 1155
	case KVM_CAP_NR_VCPUS:
		r = KVM_MAX_VCPUS;
		break;
1156 1157 1158
	case KVM_CAP_NR_MEMSLOTS:
		r = KVM_MEMORY_SLOTS;
		break;
1159 1160 1161
	case KVM_CAP_PV_MMU:
		r = !tdp_enabled;
		break;
B
Ben-Ami Yassour 已提交
1162 1163 1164
	case KVM_CAP_IOMMU:
		r = intel_iommu_found();
		break;
1165 1166 1167 1168 1169 1170 1171 1172
	default:
		r = 0;
		break;
	}
	return r;

}

1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206
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;
	}
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
	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,
			cpuid_arg->entries);
		if (r)
			goto out;

		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
1225 1226 1227 1228 1229 1230 1231
	default:
		r = -EINVAL;
	}
out:
	return r;
}

1232 1233 1234
void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
{
	kvm_x86_ops->vcpu_load(vcpu, cpu);
1235
	kvm_write_guest_time(vcpu);
1236 1237 1238 1239 1240
}

void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
	kvm_x86_ops->vcpu_put(vcpu);
1241
	kvm_put_guest_fpu(vcpu);
1242 1243
}

1244
static int is_efer_nx(void)
1245 1246 1247 1248
{
	u64 efer;

	rdmsrl(MSR_EFER, efer);
1249 1250 1251 1252 1253 1254 1255 1256
	return efer & EFER_NX;
}

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

1257
	entry = NULL;
1258 1259
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		e = &vcpu->arch.cpuid_entries[i];
1260 1261 1262 1263 1264
		if (e->function == 0x80000001) {
			entry = e;
			break;
		}
	}
1265
	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
1266 1267 1268 1269 1270
		entry->edx &= ~(1 << 20);
		printk(KERN_INFO "kvm: guest NX capability removed\n");
	}
}

1271
/* when an old userspace process fills a new kernel module */
1272 1273 1274
static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
				    struct kvm_cpuid *cpuid,
				    struct kvm_cpuid_entry __user *entries)
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
{
	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++) {
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
		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;
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
	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,
				    struct kvm_cpuid2 *cpuid,
				    struct kvm_cpuid_entry2 __user *entries)
1315 1316 1317 1318 1319 1320 1321
{
	int r;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -EFAULT;
1322
	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
1323
			   cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
1324
		goto out;
1325
	vcpu->arch.cpuid_nent = cpuid->nent;
1326 1327 1328 1329 1330 1331
	return 0;

out:
	return r;
}

1332 1333 1334 1335 1336 1337 1338
static int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
				    struct kvm_cpuid2 *cpuid,
				    struct kvm_cpuid_entry2 __user *entries)
{
	int r;

	r = -E2BIG;
1339
	if (cpuid->nent < vcpu->arch.cpuid_nent)
1340 1341
		goto out;
	r = -EFAULT;
1342 1343
	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
			   vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
1344 1345 1346 1347
		goto out;
	return 0;

out:
1348
	cpuid->nent = vcpu->arch.cpuid_nent;
1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
	return r;
}

static inline u32 bit(int bitno)
{
	return 1 << (bitno & 31);
}

static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			  u32 index)
{
	entry->function = function;
	entry->index = index;
	cpuid_count(entry->function, entry->index,
		&entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
	entry->flags = 0;
}

static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			 u32 index, int *nent, int maxnent)
{
	const u32 kvm_supported_word0_x86_features = bit(X86_FEATURE_FPU) |
		bit(X86_FEATURE_VME) | bit(X86_FEATURE_DE) |
		bit(X86_FEATURE_PSE) | bit(X86_FEATURE_TSC) |
		bit(X86_FEATURE_MSR) | bit(X86_FEATURE_PAE) |
		bit(X86_FEATURE_CX8) | bit(X86_FEATURE_APIC) |
		bit(X86_FEATURE_SEP) | bit(X86_FEATURE_PGE) |
		bit(X86_FEATURE_CMOV) | bit(X86_FEATURE_PSE36) |
		bit(X86_FEATURE_CLFLSH) | bit(X86_FEATURE_MMX) |
		bit(X86_FEATURE_FXSR) | bit(X86_FEATURE_XMM) |
		bit(X86_FEATURE_XMM2) | bit(X86_FEATURE_SELFSNOOP);
	const u32 kvm_supported_word1_x86_features = bit(X86_FEATURE_FPU) |
		bit(X86_FEATURE_VME) | bit(X86_FEATURE_DE) |
		bit(X86_FEATURE_PSE) | bit(X86_FEATURE_TSC) |
		bit(X86_FEATURE_MSR) | bit(X86_FEATURE_PAE) |
		bit(X86_FEATURE_CX8) | bit(X86_FEATURE_APIC) |
		bit(X86_FEATURE_PGE) |
		bit(X86_FEATURE_CMOV) | bit(X86_FEATURE_PSE36) |
		bit(X86_FEATURE_MMX) | bit(X86_FEATURE_FXSR) |
		bit(X86_FEATURE_SYSCALL) |
		(bit(X86_FEATURE_NX) && is_efer_nx()) |
#ifdef CONFIG_X86_64
		bit(X86_FEATURE_LM) |
#endif
		bit(X86_FEATURE_MMXEXT) |
		bit(X86_FEATURE_3DNOWEXT) |
		bit(X86_FEATURE_3DNOW);
	const u32 kvm_supported_word3_x86_features =
		bit(X86_FEATURE_XMM3) | bit(X86_FEATURE_CX16);
	const u32 kvm_supported_word6_x86_features =
		bit(X86_FEATURE_LAHF_LM) | bit(X86_FEATURE_CMP_LEGACY);

	/* all func 2 cpuid_count() should be called on the same cpu */
	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;
		entry->ecx &= kvm_supported_word3_x86_features;
		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;
		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: {
1431
		int i, cache_type;
1432 1433 1434

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until cache_type is zero */
1435 1436
		for (i = 1; *nent < maxnent; ++i) {
			cache_type = entry[i - 1].eax & 0x1f;
1437 1438
			if (!cache_type)
				break;
1439 1440
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1441 1442 1443 1444 1445 1446
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 0xb: {
1447
		int i, level_type;
1448 1449 1450

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until level_type is zero */
1451 1452
		for (i = 1; *nent < maxnent; ++i) {
			level_type = entry[i - 1].ecx & 0xff;
1453 1454
			if (!level_type)
				break;
1455 1456
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472
			       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();
}

1473
static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
				    struct kvm_cpuid_entry2 __user *entries)
{
	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,
				&nent, cpuid->nent);
	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,
			       &nent, cpuid->nent);
	r = -EFAULT;
	if (copy_to_user(entries, cpuid_entries,
			nent * sizeof(struct kvm_cpuid_entry2)))
		goto out_free;
	cpuid->nent = nent;
	r = 0;

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

1514 1515 1516 1517
static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
1518
	memcpy(s->regs, vcpu->arch.apic->regs, sizeof *s);
1519 1520 1521 1522 1523 1524 1525 1526 1527
	vcpu_put(vcpu);

	return 0;
}

static int kvm_vcpu_ioctl_set_lapic(struct kvm_vcpu *vcpu,
				    struct kvm_lapic_state *s)
{
	vcpu_load(vcpu);
1528
	memcpy(vcpu->arch.apic->regs, s->regs, sizeof *s);
1529 1530 1531 1532 1533 1534
	kvm_apic_post_state_restore(vcpu);
	vcpu_put(vcpu);

	return 0;
}

1535 1536 1537 1538 1539 1540 1541 1542 1543
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);

1544 1545
	set_bit(irq->irq, vcpu->arch.irq_pending);
	set_bit(irq->irq / BITS_PER_LONG, &vcpu->arch.irq_summary);
1546 1547 1548 1549 1550 1551

	vcpu_put(vcpu);

	return 0;
}

1552 1553 1554 1555 1556 1557 1558 1559 1560
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;
}

1561 1562 1563 1564 1565 1566
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;
1567
	struct kvm_lapic_state *lapic = NULL;
1568 1569 1570

	switch (ioctl) {
	case KVM_GET_LAPIC: {
1571
		lapic = kzalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
1572

1573 1574 1575 1576
		r = -ENOMEM;
		if (!lapic)
			goto out;
		r = kvm_vcpu_ioctl_get_lapic(vcpu, lapic);
1577 1578 1579
		if (r)
			goto out;
		r = -EFAULT;
1580
		if (copy_to_user(argp, lapic, sizeof(struct kvm_lapic_state)))
1581 1582 1583 1584 1585
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_LAPIC: {
1586 1587 1588 1589
		lapic = kmalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
		r = -ENOMEM;
		if (!lapic)
			goto out;
1590
		r = -EFAULT;
1591
		if (copy_from_user(lapic, argp, sizeof(struct kvm_lapic_state)))
1592
			goto out;
1593
		r = kvm_vcpu_ioctl_set_lapic(vcpu, lapic);
1594 1595 1596 1597 1598
		if (r)
			goto out;
		r = 0;
		break;
	}
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610
	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;
	}
1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622
	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;
	}
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652
	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,
				cpuid_arg->entries);
		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,
				cpuid_arg->entries);
		if (r)
			goto out;
		r = -EFAULT;
		if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
			goto out;
		r = 0;
		break;
	}
1653 1654 1655 1656 1657 1658
	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;
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
	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
1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686
	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;
	}
1687 1688 1689 1690
	default:
		r = -EINVAL;
	}
out:
1691 1692
	if (lapic)
		kfree(lapic);
1693 1694 1695
	return r;
}

1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711
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;

1712
	down_write(&kvm->slots_lock);
1713 1714

	kvm_mmu_change_mmu_pages(kvm, kvm_nr_mmu_pages);
1715
	kvm->arch.n_requested_mmu_pages = kvm_nr_mmu_pages;
1716

1717
	up_write(&kvm->slots_lock);
1718 1719 1720 1721 1722
	return 0;
}

static int kvm_vm_ioctl_get_nr_mmu_pages(struct kvm *kvm)
{
1723
	return kvm->arch.n_alloc_mmu_pages;
1724 1725
}

1726 1727 1728 1729 1730
gfn_t unalias_gfn(struct kvm *kvm, gfn_t gfn)
{
	int i;
	struct kvm_mem_alias *alias;

1731 1732
	for (i = 0; i < kvm->arch.naliases; ++i) {
		alias = &kvm->arch.aliases[i];
1733 1734 1735 1736 1737 1738 1739
		if (gfn >= alias->base_gfn
		    && gfn < alias->base_gfn + alias->npages)
			return alias->target_gfn + gfn - alias->base_gfn;
	}
	return gfn;
}

1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765
/*
 * 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;

1766
	down_write(&kvm->slots_lock);
1767
	spin_lock(&kvm->mmu_lock);
1768

1769
	p = &kvm->arch.aliases[alias->slot];
1770 1771 1772 1773 1774
	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)
1775
		if (kvm->arch.aliases[n - 1].npages)
1776
			break;
1777
	kvm->arch.naliases = n;
1778

1779
	spin_unlock(&kvm->mmu_lock);
1780 1781
	kvm_mmu_zap_all(kvm);

1782
	up_write(&kvm->slots_lock);
1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846

	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;
}

1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863
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;
}

1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874
/*
 * 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;

1875
	down_write(&kvm->slots_lock);
1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890

	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) {
		kvm_mmu_slot_remove_write_access(kvm, log->slot);
		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:
1891
	up_write(&kvm->slots_lock);
1892 1893 1894
	return r;
}

1895 1896 1897 1898 1899 1900
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;
1901 1902 1903 1904 1905 1906 1907 1908 1909
	/*
	 * 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;
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940

	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;
1941
	case KVM_SET_MEMORY_ALIAS:
1942
		r = -EFAULT;
1943
		if (copy_from_user(&u.alias, argp, sizeof(struct kvm_memory_alias)))
1944
			goto out;
1945
		r = kvm_vm_ioctl_set_memory_alias(kvm, &u.alias);
1946 1947 1948 1949 1950
		if (r)
			goto out;
		break;
	case KVM_CREATE_IRQCHIP:
		r = -ENOMEM;
1951 1952
		kvm->arch.vpic = kvm_create_pic(kvm);
		if (kvm->arch.vpic) {
1953 1954
			r = kvm_ioapic_init(kvm);
			if (r) {
1955 1956
				kfree(kvm->arch.vpic);
				kvm->arch.vpic = NULL;
1957 1958 1959 1960 1961
				goto out;
			}
		} else
			goto out;
		break;
S
Sheng Yang 已提交
1962 1963 1964 1965 1966 1967
	case KVM_CREATE_PIT:
		r = -ENOMEM;
		kvm->arch.vpit = kvm_create_pit(kvm);
		if (kvm->arch.vpit)
			r = 0;
		break;
1968 1969 1970 1971 1972 1973 1974 1975
	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)) {
			mutex_lock(&kvm->lock);
1976
			kvm_set_irq(kvm, irq_event.irq, irq_event.level);
1977 1978 1979 1980 1981 1982 1983
			mutex_unlock(&kvm->lock);
			r = 0;
		}
		break;
	}
	case KVM_GET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
1984
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
1985

1986 1987
		r = -ENOMEM;
		if (!chip)
1988
			goto out;
1989 1990 1991
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto get_irqchip_out;
1992 1993
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
1994 1995
			goto get_irqchip_out;
		r = kvm_vm_ioctl_get_irqchip(kvm, chip);
1996
		if (r)
1997
			goto get_irqchip_out;
1998
		r = -EFAULT;
1999 2000
		if (copy_to_user(argp, chip, sizeof *chip))
			goto get_irqchip_out;
2001
		r = 0;
2002 2003 2004 2005
	get_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
2006 2007 2008 2009
		break;
	}
	case KVM_SET_IRQCHIP: {
		/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
2010
		struct kvm_irqchip *chip = kmalloc(sizeof(*chip), GFP_KERNEL);
2011

2012 2013
		r = -ENOMEM;
		if (!chip)
2014
			goto out;
2015 2016 2017
		r = -EFAULT;
		if (copy_from_user(chip, argp, sizeof *chip))
			goto set_irqchip_out;
2018 2019
		r = -ENXIO;
		if (!irqchip_in_kernel(kvm))
2020 2021
			goto set_irqchip_out;
		r = kvm_vm_ioctl_set_irqchip(kvm, chip);
2022
		if (r)
2023
			goto set_irqchip_out;
2024
		r = 0;
2025 2026 2027 2028
	set_irqchip_out:
		kfree(chip);
		if (r)
			goto out;
2029 2030
		break;
	}
B
Ben-Ami Yassour 已提交
2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052
	case KVM_ASSIGN_PCI_DEVICE: {
		struct kvm_assigned_pci_dev assigned_dev;

		r = -EFAULT;
		if (copy_from_user(&assigned_dev, argp, sizeof assigned_dev))
			goto out;
		r = kvm_vm_ioctl_assign_device(kvm, &assigned_dev);
		if (r)
			goto out;
		break;
	}
	case KVM_ASSIGN_IRQ: {
		struct kvm_assigned_irq assigned_irq;

		r = -EFAULT;
		if (copy_from_user(&assigned_irq, argp, sizeof assigned_irq))
			goto out;
		r = kvm_vm_ioctl_assign_irq(kvm, &assigned_irq);
		if (r)
			goto out;
		break;
	}
2053 2054
	case KVM_GET_PIT: {
		r = -EFAULT;
2055
		if (copy_from_user(&u.ps, argp, sizeof(struct kvm_pit_state)))
2056 2057 2058 2059
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
2060
		r = kvm_vm_ioctl_get_pit(kvm, &u.ps);
2061 2062 2063
		if (r)
			goto out;
		r = -EFAULT;
2064
		if (copy_to_user(argp, &u.ps, sizeof(struct kvm_pit_state)))
2065 2066 2067 2068 2069 2070
			goto out;
		r = 0;
		break;
	}
	case KVM_SET_PIT: {
		r = -EFAULT;
2071
		if (copy_from_user(&u.ps, argp, sizeof u.ps))
2072 2073 2074 2075
			goto out;
		r = -ENXIO;
		if (!kvm->arch.vpit)
			goto out;
2076
		r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
2077 2078 2079 2080 2081
		if (r)
			goto out;
		r = 0;
		break;
	}
2082 2083 2084 2085 2086 2087 2088
	default:
		;
	}
out:
	return r;
}

2089
static void kvm_init_msr_list(void)
2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103
{
	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;
}

2104 2105 2106 2107
/*
 * Only apic need an MMIO device hook, so shortcut now..
 */
static struct kvm_io_device *vcpu_find_pervcpu_dev(struct kvm_vcpu *vcpu,
2108 2109
						gpa_t addr, int len,
						int is_write)
2110 2111 2112
{
	struct kvm_io_device *dev;

2113 2114
	if (vcpu->arch.apic) {
		dev = &vcpu->arch.apic->dev;
2115
		if (dev->in_range(dev, addr, len, is_write))
2116 2117 2118 2119 2120 2121 2122
			return dev;
	}
	return NULL;
}


static struct kvm_io_device *vcpu_find_mmio_dev(struct kvm_vcpu *vcpu,
2123 2124
						gpa_t addr, int len,
						int is_write)
2125 2126 2127
{
	struct kvm_io_device *dev;

2128
	dev = vcpu_find_pervcpu_dev(vcpu, addr, len, is_write);
2129
	if (dev == NULL)
2130 2131
		dev = kvm_io_bus_find_dev(&vcpu->kvm->mmio_bus, addr, len,
					  is_write);
2132 2133 2134 2135 2136 2137 2138 2139 2140
	return dev;
}

int emulator_read_std(unsigned long addr,
			     void *val,
			     unsigned int bytes,
			     struct kvm_vcpu *vcpu)
{
	void *data = val;
2141
	int r = X86EMUL_CONTINUE;
2142 2143

	while (bytes) {
2144
		gpa_t gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2145 2146 2147 2148
		unsigned offset = addr & (PAGE_SIZE-1);
		unsigned tocopy = min(bytes, (unsigned)PAGE_SIZE - offset);
		int ret;

2149 2150 2151 2152
		if (gpa == UNMAPPED_GVA) {
			r = X86EMUL_PROPAGATE_FAULT;
			goto out;
		}
2153
		ret = kvm_read_guest(vcpu->kvm, gpa, data, tocopy);
2154 2155 2156 2157
		if (ret < 0) {
			r = X86EMUL_UNHANDLEABLE;
			goto out;
		}
2158 2159 2160 2161 2162

		bytes -= tocopy;
		data += tocopy;
		addr += tocopy;
	}
2163 2164
out:
	return r;
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
}
EXPORT_SYMBOL_GPL(emulator_read_std);

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;
	}

2182
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197

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

	if (emulator_read_std(addr, val, bytes, vcpu)
			== X86EMUL_CONTINUE)
		return X86EMUL_CONTINUE;
	if (gpa == UNMAPPED_GVA)
		return X86EMUL_PROPAGATE_FAULT;

mmio:
	/*
	 * Is this MMIO handled locally?
	 */
2198
	mutex_lock(&vcpu->kvm->lock);
2199
	mmio_dev = vcpu_find_mmio_dev(vcpu, gpa, bytes, 0);
2200 2201
	if (mmio_dev) {
		kvm_iodevice_read(mmio_dev, gpa, bytes, val);
2202
		mutex_unlock(&vcpu->kvm->lock);
2203 2204
		return X86EMUL_CONTINUE;
	}
2205
	mutex_unlock(&vcpu->kvm->lock);
2206 2207 2208 2209 2210 2211 2212 2213 2214

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

	return X86EMUL_UNHANDLEABLE;
}

2215
int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
2216
			  const void *val, int bytes)
2217 2218 2219 2220
{
	int ret;

	ret = kvm_write_guest(vcpu->kvm, gpa, val, bytes);
2221
	if (ret < 0)
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
		return 0;
	kvm_mmu_pte_write(vcpu, gpa, val, bytes);
	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;
2233 2234 2235
	gpa_t                 gpa;

	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);
2236 2237

	if (gpa == UNMAPPED_GVA) {
2238
		kvm_inject_page_fault(vcpu, addr, 2);
2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252
		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?
	 */
2253
	mutex_lock(&vcpu->kvm->lock);
2254
	mmio_dev = vcpu_find_mmio_dev(vcpu, gpa, bytes, 1);
2255 2256
	if (mmio_dev) {
		kvm_iodevice_write(mmio_dev, gpa, bytes, val);
2257
		mutex_unlock(&vcpu->kvm->lock);
2258 2259
		return X86EMUL_CONTINUE;
	}
2260
	mutex_unlock(&vcpu->kvm->lock);
2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303

	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");
	}
2304 2305 2306
#ifndef CONFIG_X86_64
	/* guests cmpxchg8b have to be emulated atomically */
	if (bytes == 8) {
2307
		gpa_t gpa;
2308
		struct page *page;
A
Andrew Morton 已提交
2309
		char *kaddr;
2310 2311
		u64 val;

2312 2313
		gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, addr);

2314 2315 2316 2317 2318 2319 2320 2321
		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;
2322

2323
		page = gfn_to_page(vcpu->kvm, gpa >> PAGE_SHIFT);
2324

A
Andrew Morton 已提交
2325 2326 2327
		kaddr = kmap_atomic(page, KM_USER0);
		set_64bit((u64 *)(kaddr + offset_in_page(gpa)), val);
		kunmap_atomic(kaddr, KM_USER0);
2328 2329
		kvm_release_page_dirty(page);
	}
2330
emul_write:
2331 2332
#endif

2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347
	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)
{
	return X86EMUL_CONTINUE;
}

int emulate_clts(struct kvm_vcpu *vcpu)
{
J
Joerg Roedel 已提交
2348
	KVMTRACE_0D(CLTS, vcpu, handler);
2349
	kvm_x86_ops->set_cr0(vcpu, vcpu->arch.cr0 & ~X86_CR0_TS);
2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361
	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:
2362
		pr_unimpl(vcpu, "%s: unexpected dr %u\n", __func__, dr);
2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382
		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];
2383
	unsigned long rip = kvm_rip_read(vcpu);
2384 2385
	unsigned long rip_linear;

2386
	if (!printk_ratelimit())
2387 2388
		return;

2389 2390
	rip_linear = rip + get_segment_base(vcpu, VCPU_SREG_CS);

2391 2392 2393 2394 2395 2396 2397
	emulator_read_std(rip_linear, (void *)opcodes, 4, vcpu);

	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);

2398
static struct x86_emulate_ops emulate_ops = {
2399 2400 2401 2402 2403 2404
	.read_std            = emulator_read_std,
	.read_emulated       = emulator_read_emulated,
	.write_emulated      = emulator_write_emulated,
	.cmpxchg_emulated    = emulator_cmpxchg_emulated,
};

2405 2406 2407 2408 2409 2410 2411 2412
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;
}

2413 2414 2415 2416
int emulate_instruction(struct kvm_vcpu *vcpu,
			struct kvm_run *run,
			unsigned long cr2,
			u16 error_code,
2417
			int emulation_type)
2418 2419
{
	int r;
2420
	struct decode_cache *c;
2421

2422
	kvm_clear_exception_queue(vcpu);
2423
	vcpu->arch.mmio_fault_cr2 = cr2;
2424 2425 2426 2427 2428 2429 2430
	/*
	 * 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);
2431 2432

	vcpu->mmio_is_write = 0;
2433
	vcpu->arch.pio.string = 0;
2434

2435
	if (!(emulation_type & EMULTYPE_NO_DECODE)) {
2436 2437 2438
		int cs_db, cs_l;
		kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);

2439 2440 2441 2442
		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)
2443 2444 2445 2446
			? X86EMUL_MODE_REAL : cs_l
			? X86EMUL_MODE_PROT64 :	cs_db
			? X86EMUL_MODE_PROT32 : X86EMUL_MODE_PROT16;

2447
		r = x86_decode_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
2448 2449 2450 2451 2452 2453 2454 2455 2456 2457

		/* 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;

2458
		++vcpu->stat.insn_emulation;
2459
		if (r)  {
2460
			++vcpu->stat.insn_emulation_fail;
2461 2462 2463 2464 2465 2466
			if (kvm_mmu_unprotect_page_virt(vcpu, cr2))
				return EMULATE_DONE;
			return EMULATE_FAIL;
		}
	}

2467
	r = x86_emulate_insn(&vcpu->arch.emulate_ctxt, &emulate_ops);
2468

2469
	if (vcpu->arch.pio.string)
2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489
		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;
	}

2490
	kvm_x86_ops->set_rflags(vcpu, vcpu->arch.emulate_ctxt.eflags);
2491 2492 2493 2494 2495 2496 2497 2498 2499 2500

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

	return EMULATE_DONE;
}
EXPORT_SYMBOL_GPL(emulate_instruction);

2501 2502 2503 2504
static void free_pio_guest_pages(struct kvm_vcpu *vcpu)
{
	int i;

2505 2506 2507 2508
	for (i = 0; i < ARRAY_SIZE(vcpu->arch.pio.guest_pages); ++i)
		if (vcpu->arch.pio.guest_pages[i]) {
			kvm_release_page_dirty(vcpu->arch.pio.guest_pages[i]);
			vcpu->arch.pio.guest_pages[i] = NULL;
2509 2510 2511 2512 2513
		}
}

static int pio_copy_data(struct kvm_vcpu *vcpu)
{
2514
	void *p = vcpu->arch.pio_data;
2515 2516
	void *q;
	unsigned bytes;
2517
	int nr_pages = vcpu->arch.pio.guest_pages[1] ? 2 : 1;
2518

2519
	q = vmap(vcpu->arch.pio.guest_pages, nr_pages, VM_READ|VM_WRITE,
2520 2521 2522 2523 2524
		 PAGE_KERNEL);
	if (!q) {
		free_pio_guest_pages(vcpu);
		return -ENOMEM;
	}
2525 2526 2527
	q += vcpu->arch.pio.guest_page_offset;
	bytes = vcpu->arch.pio.size * vcpu->arch.pio.cur_count;
	if (vcpu->arch.pio.in)
2528 2529 2530
		memcpy(q, p, bytes);
	else
		memcpy(p, q, bytes);
2531
	q -= vcpu->arch.pio.guest_page_offset;
2532 2533 2534 2535 2536 2537 2538
	vunmap(q);
	free_pio_guest_pages(vcpu);
	return 0;
}

int complete_pio(struct kvm_vcpu *vcpu)
{
2539
	struct kvm_pio_request *io = &vcpu->arch.pio;
2540 2541
	long delta;
	int r;
2542
	unsigned long val;
2543 2544

	if (!io->string) {
2545 2546 2547 2548 2549
		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);
		}
2550 2551 2552
	} else {
		if (io->in) {
			r = pio_copy_data(vcpu);
2553
			if (r)
2554 2555 2556 2557 2558 2559 2560 2561 2562 2563
				return r;
		}

		delta = 1;
		if (io->rep) {
			delta *= io->cur_count;
			/*
			 * The size of the register should really depend on
			 * current address size.
			 */
2564 2565 2566
			val = kvm_register_read(vcpu, VCPU_REGS_RCX);
			val -= delta;
			kvm_register_write(vcpu, VCPU_REGS_RCX, val);
2567 2568 2569 2570
		}
		if (io->down)
			delta = -delta;
		delta *= io->size;
2571 2572 2573 2574 2575 2576 2577 2578 2579
		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);
		}
2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594
	}

	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);
2595 2596 2597
	if (vcpu->arch.pio.in)
		kvm_iodevice_read(pio_dev, vcpu->arch.pio.port,
				  vcpu->arch.pio.size,
2598 2599
				  pd);
	else
2600 2601
		kvm_iodevice_write(pio_dev, vcpu->arch.pio.port,
				   vcpu->arch.pio.size,
2602 2603 2604 2605 2606 2607 2608
				   pd);
	mutex_unlock(&vcpu->kvm->lock);
}

static void pio_string_write(struct kvm_io_device *pio_dev,
			     struct kvm_vcpu *vcpu)
{
2609 2610
	struct kvm_pio_request *io = &vcpu->arch.pio;
	void *pd = vcpu->arch.pio_data;
2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623
	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,
2624 2625
					       gpa_t addr, int len,
					       int is_write)
2626
{
2627
	return kvm_io_bus_find_dev(&vcpu->kvm->pio_bus, addr, len, is_write);
2628 2629 2630 2631 2632 2633
}

int kvm_emulate_pio(struct kvm_vcpu *vcpu, struct kvm_run *run, int in,
		  int size, unsigned port)
{
	struct kvm_io_device *pio_dev;
2634
	unsigned long val;
2635 2636 2637

	vcpu->run->exit_reason = KVM_EXIT_IO;
	vcpu->run->io.direction = in ? KVM_EXIT_IO_IN : KVM_EXIT_IO_OUT;
2638
	vcpu->run->io.size = vcpu->arch.pio.size = size;
2639
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
2640 2641 2642 2643 2644 2645 2646
	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.guest_page_offset = 0;
	vcpu->arch.pio.rep = 0;
2647

F
Feng (Eric) Liu 已提交
2648 2649 2650 2651 2652 2653 2654
	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);

2655 2656
	val = kvm_register_read(vcpu, VCPU_REGS_RAX);
	memcpy(vcpu->arch.pio_data, &val, 4);
2657 2658 2659

	kvm_x86_ops->skip_emulated_instruction(vcpu);

2660
	pio_dev = vcpu_find_pio_dev(vcpu, port, size, !in);
2661
	if (pio_dev) {
2662
		kernel_pio(pio_dev, vcpu, vcpu->arch.pio_data);
2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681
		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;
	int i, ret = 0;
	int nr_pages = 1;
	struct page *page;
	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;
2682
	vcpu->run->io.size = vcpu->arch.pio.size = size;
2683
	vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
2684 2685 2686 2687 2688 2689 2690
	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.guest_page_offset = offset_in_page(address);
	vcpu->arch.pio.rep = rep;
2691

F
Feng (Eric) Liu 已提交
2692 2693 2694 2695 2696 2697 2698
	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);

2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722
	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);
	if (!now) {
		/*
		 * String I/O straddles page boundary.  Pin two guest pages
		 * so that we satisfy atomicity constraints.  Do just one
		 * transaction to avoid complexity.
		 */
		nr_pages = 2;
		now = 1;
	}
	if (down) {
		/*
		 * String I/O in reverse.  Yuck.  Kill the guest, fix later.
		 */
		pr_unimpl(vcpu, "guest string pio down\n");
2723
		kvm_inject_gp(vcpu, 0);
2724 2725 2726
		return 1;
	}
	vcpu->run->io.count = now;
2727
	vcpu->arch.pio.cur_count = now;
2728

2729
	if (vcpu->arch.pio.cur_count == vcpu->arch.pio.count)
2730 2731 2732 2733
		kvm_x86_ops->skip_emulated_instruction(vcpu);

	for (i = 0; i < nr_pages; ++i) {
		page = gva_to_page(vcpu, address + i * PAGE_SIZE);
2734
		vcpu->arch.pio.guest_pages[i] = page;
2735
		if (!page) {
2736
			kvm_inject_gp(vcpu, 0);
2737 2738 2739 2740 2741
			free_pio_guest_pages(vcpu);
			return 1;
		}
	}

2742 2743 2744
	pio_dev = vcpu_find_pio_dev(vcpu, port,
				    vcpu->arch.pio.cur_count,
				    !vcpu->arch.pio.in);
2745
	if (!vcpu->arch.pio.in) {
2746 2747 2748 2749 2750
		/* string PIO write */
		ret = pio_copy_data(vcpu);
		if (ret >= 0 && pio_dev) {
			pio_string_write(pio_dev, vcpu);
			complete_pio(vcpu);
2751
			if (vcpu->arch.pio.count == 0)
2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762
				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);

2763
int kvm_arch_init(void *opaque)
2764
{
2765
	int r;
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 2795
	kvm_mmu_set_base_ptes(PT_PRESENT_MASK);
	kvm_mmu_set_mask_ptes(PT_USER_MASK, PT_ACCESSED_MASK,
			PT_DIRTY_MASK, PT64_NX_MASK, 0);
2796
	return 0;
2797 2798 2799

out:
	return r;
2800
}
2801

2802 2803 2804
void kvm_arch_exit(void)
{
	kvm_x86_ops = NULL;
2805 2806
	kvm_mmu_module_exit();
}
2807

2808 2809 2810
int kvm_emulate_halt(struct kvm_vcpu *vcpu)
{
	++vcpu->stat.halt_exits;
F
Feng (Eric) Liu 已提交
2811
	KVMTRACE_0D(HLT, vcpu, handler);
2812
	if (irqchip_in_kernel(vcpu->kvm)) {
2813
		vcpu->arch.mp_state = KVM_MP_STATE_HALTED;
2814 2815 2816 2817 2818 2819 2820 2821
		return 1;
	} else {
		vcpu->run->exit_reason = KVM_EXIT_HLT;
		return 0;
	}
}
EXPORT_SYMBOL_GPL(kvm_emulate_halt);

2822 2823 2824 2825 2826 2827 2828 2829 2830
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);
}

2831 2832 2833
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
{
	unsigned long nr, a0, a1, a2, a3, ret;
2834
	int r = 1;
2835

2836 2837 2838 2839 2840
	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);
2841

F
Feng (Eric) Liu 已提交
2842 2843
	KVMTRACE_1D(VMMCALL, vcpu, (u32)nr, handler);

2844 2845 2846 2847 2848 2849 2850 2851 2852
	if (!is_long_mode(vcpu)) {
		nr &= 0xFFFFFFFF;
		a0 &= 0xFFFFFFFF;
		a1 &= 0xFFFFFFFF;
		a2 &= 0xFFFFFFFF;
		a3 &= 0xFFFFFFFF;
	}

	switch (nr) {
A
Avi Kivity 已提交
2853 2854 2855
	case KVM_HC_VAPIC_POLL_IRQ:
		ret = 0;
		break;
2856 2857 2858
	case KVM_HC_MMU_OP:
		r = kvm_pv_mmu_op(vcpu, a0, hc_gpa(vcpu, a1, a2), &ret);
		break;
2859 2860 2861 2862
	default:
		ret = -KVM_ENOSYS;
		break;
	}
2863
	kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
A
Amit Shah 已提交
2864
	++vcpu->stat.hypercalls;
2865
	return r;
2866 2867 2868 2869 2870 2871 2872
}
EXPORT_SYMBOL_GPL(kvm_emulate_hypercall);

int kvm_fix_hypercall(struct kvm_vcpu *vcpu)
{
	char instruction[3];
	int ret = 0;
2873
	unsigned long rip = kvm_rip_read(vcpu);
2874 2875 2876 2877 2878 2879 2880 2881 2882 2883


	/*
	 * 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);
2884
	if (emulator_write_emulated(rip, instruction, 3, vcpu)
2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912
	    != 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)
{
2913
	kvm_lmsw(vcpu, msw);
2914 2915 2916 2917 2918
	*rflags = kvm_x86_ops->get_rflags(vcpu);
}

unsigned long realmode_get_cr(struct kvm_vcpu *vcpu, int cr)
{
J
Joerg Roedel 已提交
2919 2920
	unsigned long value;

2921 2922 2923
	kvm_x86_ops->decache_cr4_guest_bits(vcpu);
	switch (cr) {
	case 0:
J
Joerg Roedel 已提交
2924 2925
		value = vcpu->arch.cr0;
		break;
2926
	case 2:
J
Joerg Roedel 已提交
2927 2928
		value = vcpu->arch.cr2;
		break;
2929
	case 3:
J
Joerg Roedel 已提交
2930 2931
		value = vcpu->arch.cr3;
		break;
2932
	case 4:
J
Joerg Roedel 已提交
2933 2934
		value = vcpu->arch.cr4;
		break;
2935
	case 8:
J
Joerg Roedel 已提交
2936 2937
		value = kvm_get_cr8(vcpu);
		break;
2938
	default:
2939
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
2940 2941
		return 0;
	}
J
Joerg Roedel 已提交
2942 2943 2944 2945
	KVMTRACE_3D(CR_READ, vcpu, (u32)cr, (u32)value,
		    (u32)((u64)value >> 32), handler);

	return value;
2946 2947 2948 2949 2950
}

void realmode_set_cr(struct kvm_vcpu *vcpu, int cr, unsigned long val,
		     unsigned long *rflags)
{
J
Joerg Roedel 已提交
2951 2952 2953
	KVMTRACE_3D(CR_WRITE, vcpu, (u32)cr, (u32)val,
		    (u32)((u64)val >> 32), handler);

2954 2955
	switch (cr) {
	case 0:
2956
		kvm_set_cr0(vcpu, mk_cr_64(vcpu->arch.cr0, val));
2957 2958 2959
		*rflags = kvm_x86_ops->get_rflags(vcpu);
		break;
	case 2:
2960
		vcpu->arch.cr2 = val;
2961 2962
		break;
	case 3:
2963
		kvm_set_cr3(vcpu, val);
2964 2965
		break;
	case 4:
2966
		kvm_set_cr4(vcpu, mk_cr_64(vcpu->arch.cr4, val));
2967
		break;
2968
	case 8:
2969
		kvm_set_cr8(vcpu, val & 0xfUL);
2970
		break;
2971
	default:
2972
		vcpu_printf(vcpu, "%s: unexpected cr %u\n", __func__, cr);
2973 2974 2975
	}
}

2976 2977
static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
{
2978 2979
	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	int j, nent = vcpu->arch.cpuid_nent;
2980 2981 2982 2983

	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	/* when no next entry is found, the current entry[i] is reselected */
	for (j = i + 1; j == i; j = (j + 1) % nent) {
2984
		struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007
		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) &&
		!(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
		return 0;
	return 1;
}

3008 3009 3010
void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
{
	int i;
3011 3012
	u32 function, index;
	struct kvm_cpuid_entry2 *e, *best;
3013

3014 3015 3016 3017 3018 3019
	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);
3020
	best = NULL;
3021 3022
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		e = &vcpu->arch.cpuid_entries[i];
3023 3024 3025
		if (is_matching_cpuid_entry(e, function, index)) {
			if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
				move_to_next_stateful_cpuid_entry(vcpu, i);
3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036
			best = e;
			break;
		}
		/*
		 * Both basic or both extended?
		 */
		if (((e->function ^ function) & 0x80000000) == 0)
			if (!best || e->function > best->function)
				best = e;
	}
	if (best) {
3037 3038 3039 3040
		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);
3041 3042
	}
	kvm_x86_ops->skip_emulated_instruction(vcpu);
F
Feng (Eric) Liu 已提交
3043
	KVMTRACE_5D(CPUID, vcpu, function,
3044 3045 3046 3047
		    (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);
3048 3049
}
EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);
3050

3051 3052 3053 3054 3055 3056 3057 3058 3059
/*
 * 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)
{
3060
	return (!vcpu->arch.irq_summary &&
3061
		kvm_run->request_interrupt_window &&
3062
		vcpu->arch.interrupt_window_open &&
3063 3064 3065 3066 3067 3068 3069
		(kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF));
}

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;
3070
	kvm_run->cr8 = kvm_get_cr8(vcpu);
3071 3072 3073 3074 3075
	kvm_run->apic_base = kvm_get_apic_base(vcpu);
	if (irqchip_in_kernel(vcpu->kvm))
		kvm_run->ready_for_interrupt_injection = 1;
	else
		kvm_run->ready_for_interrupt_injection =
3076 3077
					(vcpu->arch.interrupt_window_open &&
					 vcpu->arch.irq_summary == 0);
3078 3079
}

A
Avi Kivity 已提交
3080 3081 3082 3083 3084 3085 3086 3087 3088
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);
3089 3090

	vcpu->arch.apic->vapic_page = page;
A
Avi Kivity 已提交
3091 3092 3093 3094 3095 3096 3097 3098 3099
}

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

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

3100
	down_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
3101 3102
	kvm_release_page_dirty(apic->vapic_page);
	mark_page_dirty(vcpu->kvm, apic->vapic_addr >> PAGE_SHIFT);
3103
	up_read(&vcpu->kvm->slots_lock);
A
Avi Kivity 已提交
3104 3105
}

3106
static int vcpu_enter_guest(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
3107 3108 3109
{
	int r;

3110 3111 3112 3113
	if (vcpu->requests)
		if (test_and_clear_bit(KVM_REQ_MMU_RELOAD, &vcpu->requests))
			kvm_mmu_unload(vcpu);

3114 3115 3116 3117
	r = kvm_mmu_reload(vcpu);
	if (unlikely(r))
		goto out;

3118 3119
	if (vcpu->requests) {
		if (test_and_clear_bit(KVM_REQ_MIGRATE_TIMER, &vcpu->requests))
M
Marcelo Tosatti 已提交
3120
			__kvm_migrate_timers(vcpu);
3121 3122
		if (test_and_clear_bit(KVM_REQ_TLB_FLUSH, &vcpu->requests))
			kvm_x86_ops->tlb_flush(vcpu);
A
Avi Kivity 已提交
3123 3124 3125 3126 3127 3128
		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
3129 3130 3131 3132 3133
		if (test_and_clear_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests)) {
			kvm_run->exit_reason = KVM_EXIT_SHUTDOWN;
			r = 0;
			goto out;
		}
3134
	}
A
Avi Kivity 已提交
3135

3136
	clear_bit(KVM_REQ_PENDING_TIMER, &vcpu->requests);
3137 3138 3139 3140 3141 3142 3143 3144 3145
	kvm_inject_pending_timer_irqs(vcpu);

	preempt_disable();

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

	local_irq_disable();

3146
	if (vcpu->requests || need_resched() || signal_pending(current)) {
3147 3148 3149 3150 3151 3152
		local_irq_enable();
		preempt_enable();
		r = 1;
		goto out;
	}

3153 3154 3155
	if (vcpu->guest_debug.enabled)
		kvm_x86_ops->guest_debug_pre(vcpu);

3156 3157 3158 3159 3160 3161 3162
	vcpu->guest_mode = 1;
	/*
	 * Make sure that guest_mode assignment won't happen after
	 * testing the pending IRQ vector bitmap.
	 */
	smp_wmb();

3163
	if (vcpu->arch.exception.pending)
3164 3165
		__queue_exception(vcpu);
	else if (irqchip_in_kernel(vcpu->kvm))
3166
		kvm_x86_ops->inject_pending_irq(vcpu);
3167
	else
3168 3169
		kvm_x86_ops->inject_pending_vectors(vcpu, kvm_run);

A
Avi Kivity 已提交
3170 3171
	kvm_lapic_sync_to_vapic(vcpu);

3172 3173
	up_read(&vcpu->kvm->slots_lock);

3174 3175 3176
	kvm_guest_enter();


F
Feng (Eric) Liu 已提交
3177
	KVMTRACE_0D(VMENTRY, vcpu, entryexit);
3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196
	kvm_x86_ops->run(vcpu, kvm_run);

	vcpu->guest_mode = 0;
	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();

3197 3198
	down_read(&vcpu->kvm->slots_lock);

3199 3200 3201 3202
	/*
	 * Profile KVM exit RIPs:
	 */
	if (unlikely(prof_on == KVM_PROFILING)) {
3203 3204
		unsigned long rip = kvm_rip_read(vcpu);
		profile_hit(KVM_PROFILING, (void *)rip);
3205 3206
	}

3207 3208
	if (vcpu->arch.exception.pending && kvm_x86_ops->exception_injected(vcpu))
		vcpu->arch.exception.pending = false;
3209

A
Avi Kivity 已提交
3210 3211
	kvm_lapic_sync_from_vapic(vcpu);

3212
	r = kvm_x86_ops->handle_exit(kvm_run, vcpu);
3213 3214 3215
out:
	return r;
}
3216

3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228
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)) {
		printk("vcpu %d received sipi with vector # %x\n",
		       vcpu->vcpu_id, vcpu->arch.sipi_vector);
		kvm_lapic_reset(vcpu);
		r = kvm_x86_ops->vcpu_reset(vcpu);
		if (r)
			return r;
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
3229 3230
	}

3231 3232 3233 3234 3235
	down_read(&vcpu->kvm->slots_lock);
	vapic_enter(vcpu);

	r = 1;
	while (r > 0) {
3236
		if (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE)
3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266
			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))
				if (vcpu->arch.mp_state == KVM_MP_STATE_HALTED)
					vcpu->arch.mp_state =
							KVM_MP_STATE_RUNNABLE;
			if (vcpu->arch.mp_state != KVM_MP_STATE_RUNNABLE)
				r = -EINTR;
		}

		if (r > 0) {
			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);
			}
		}
3267 3268
	}

3269
	up_read(&vcpu->kvm->slots_lock);
3270 3271
	post_kvm_run_save(vcpu, kvm_run);

A
Avi Kivity 已提交
3272 3273
	vapic_exit(vcpu);

3274 3275 3276 3277 3278 3279 3280 3281 3282 3283
	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);

3284 3285 3286
	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &vcpu->sigset, &sigsaved);

3287
	if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
3288
		kvm_vcpu_block(vcpu);
3289
		clear_bit(KVM_REQ_UNHALT, &vcpu->requests);
3290 3291
		r = -EAGAIN;
		goto out;
3292 3293 3294 3295
	}

	/* re-sync apic's tpr */
	if (!irqchip_in_kernel(vcpu->kvm))
3296
		kvm_set_cr8(vcpu, kvm_run->cr8);
3297

3298
	if (vcpu->arch.pio.cur_count) {
3299 3300 3301 3302 3303 3304 3305 3306 3307
		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;
3308 3309

		down_read(&vcpu->kvm->slots_lock);
3310
		r = emulate_instruction(vcpu, kvm_run,
3311 3312
					vcpu->arch.mmio_fault_cr2, 0,
					EMULTYPE_NO_DECODE);
3313
		up_read(&vcpu->kvm->slots_lock);
3314 3315 3316 3317 3318 3319 3320 3321 3322
		if (r == EMULATE_DO_MMIO) {
			/*
			 * Read-modify-write.  Back to userspace.
			 */
			r = 0;
			goto out;
		}
	}
#endif
3323 3324 3325
	if (kvm_run->exit_reason == KVM_EXIT_HYPERCALL)
		kvm_register_write(vcpu, VCPU_REGS_RAX,
				     kvm_run->hypercall.ret);
3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340

	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);

3341 3342 3343 3344 3345 3346 3347 3348
	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);
3349
#ifdef CONFIG_X86_64
3350 3351 3352 3353 3354 3355 3356 3357
	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);
3358 3359
#endif

3360
	regs->rip = kvm_rip_read(vcpu);
3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377
	regs->rflags = kvm_x86_ops->get_rflags(vcpu);

	/*
	 * Don't leak debug flags in case they were set for guest debugging
	 */
	if (vcpu->guest_debug.enabled && vcpu->guest_debug.singlestep)
		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);

3378 3379 3380 3381 3382 3383 3384 3385
	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);
3386
#ifdef CONFIG_X86_64
3387 3388 3389 3390 3391 3392 3393 3394 3395
	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);

3396 3397
#endif

3398
	kvm_rip_write(vcpu, regs->rip);
3399 3400 3401
	kvm_x86_ops->set_rflags(vcpu, regs->rflags);


3402 3403
	vcpu->arch.exception.pending = false;

3404 3405 3406 3407 3408
	vcpu_put(vcpu);

	return 0;
}

3409 3410
void kvm_get_segment(struct kvm_vcpu *vcpu,
		     struct kvm_segment *var, int seg)
3411
{
3412
	kvm_x86_ops->get_segment(vcpu, var, seg);
3413 3414 3415 3416 3417 3418
}

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

3419
	kvm_get_segment(vcpu, &cs, VCPU_SREG_CS);
3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432
	*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;
	int pending_vec;

	vcpu_load(vcpu);

3433 3434 3435 3436 3437 3438
	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);
3439

3440 3441
	kvm_get_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_get_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
3442 3443 3444 3445 3446 3447 3448 3449 3450

	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);
3451 3452 3453 3454
	sregs->cr0 = vcpu->arch.cr0;
	sregs->cr2 = vcpu->arch.cr2;
	sregs->cr3 = vcpu->arch.cr3;
	sregs->cr4 = vcpu->arch.cr4;
3455
	sregs->cr8 = kvm_get_cr8(vcpu);
3456
	sregs->efer = vcpu->arch.shadow_efer;
3457 3458 3459 3460 3461 3462 3463 3464 3465 3466
	sregs->apic_base = kvm_get_apic_base(vcpu);

	if (irqchip_in_kernel(vcpu->kvm)) {
		memset(sregs->interrupt_bitmap, 0,
		       sizeof sregs->interrupt_bitmap);
		pending_vec = kvm_x86_ops->get_irq(vcpu);
		if (pending_vec >= 0)
			set_bit(pending_vec,
				(unsigned long *)sregs->interrupt_bitmap);
	} else
3467
		memcpy(sregs->interrupt_bitmap, vcpu->arch.irq_pending,
3468 3469 3470 3471 3472 3473 3474
		       sizeof sregs->interrupt_bitmap);

	vcpu_put(vcpu);

	return 0;
}

3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492
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;
}

3493
static void kvm_set_segment(struct kvm_vcpu *vcpu,
3494 3495
			struct kvm_segment *var, int seg)
{
3496
	kvm_x86_ops->set_segment(vcpu, var, seg);
3497 3498
}

3499 3500 3501 3502 3503 3504 3505 3506
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;
3507 3508 3509 3510
	if (seg_desc->g) {
		kvm_desct->limit <<= 12;
		kvm_desct->limit |= 0xfff;
	}
3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533
	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;
}

static void get_segment_descritptor_dtable(struct kvm_vcpu *vcpu,
					   u16 selector,
					   struct descriptor_table *dtable)
{
	if (selector & 1 << 2) {
		struct kvm_segment kvm_seg;

3534
		kvm_get_segment(vcpu, &kvm_seg, VCPU_SREG_LDTR);
3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549

		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)
{
3550
	gpa_t gpa;
3551 3552 3553 3554 3555 3556 3557 3558 3559
	struct descriptor_table dtable;
	u16 index = selector >> 3;

	get_segment_descritptor_dtable(vcpu, selector, &dtable);

	if (dtable.limit < index * 8 + 7) {
		kvm_queue_exception_e(vcpu, GP_VECTOR, selector & 0xfffc);
		return 1;
	}
3560 3561 3562
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, dtable.base);
	gpa += index * 8;
	return kvm_read_guest(vcpu->kvm, gpa, seg_desc, 8);
3563 3564 3565 3566 3567 3568
}

/* allowed just for 8 bytes segments */
static int save_guest_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector,
					 struct desc_struct *seg_desc)
{
3569
	gpa_t gpa;
3570 3571 3572 3573 3574 3575 3576
	struct descriptor_table dtable;
	u16 index = selector >> 3;

	get_segment_descritptor_dtable(vcpu, selector, &dtable);

	if (dtable.limit < index * 8 + 7)
		return 1;
3577 3578 3579
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, dtable.base);
	gpa += index * 8;
	return kvm_write_guest(vcpu->kvm, gpa, seg_desc, 8);
3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590
}

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);

3591
	return vcpu->arch.mmu.gva_to_gpa(vcpu, base_addr);
3592 3593 3594 3595 3596 3597
}

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

3598
	kvm_get_segment(vcpu, &kvm_seg, seg);
3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613
	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;
}

3614
static int kvm_load_realmode_segment(struct kvm_vcpu *vcpu, u16 selector, int seg)
3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633
{
	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;
}

3634 3635
int kvm_load_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector,
				int type_bits, int seg)
3636 3637 3638
{
	struct kvm_segment kvm_seg;

3639 3640
	if (!(vcpu->arch.cr0 & X86_CR0_PE))
		return kvm_load_realmode_segment(vcpu, selector, seg);
3641 3642 3643 3644 3645 3646 3647 3648 3649
	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;

3650
	kvm_set_segment(vcpu, &kvm_seg, seg);
3651 3652 3653 3654 3655 3656 3657
	return 0;
}

static void save_state_to_tss32(struct kvm_vcpu *vcpu,
				struct tss_segment_32 *tss)
{
	tss->cr3 = vcpu->arch.cr3;
3658
	tss->eip = kvm_rip_read(vcpu);
3659
	tss->eflags = kvm_x86_ops->get_rflags(vcpu);
3660 3661 3662 3663 3664 3665 3666 3667
	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);
3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682
	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);
	tss->prev_task_link = get_segment_selector(vcpu, VCPU_SREG_TR);
}

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

3683
	kvm_rip_write(vcpu, tss->eip);
3684 3685
	kvm_x86_ops->set_rflags(vcpu, tss->eflags | 2);

3686 3687 3688 3689 3690 3691 3692 3693
	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);
3694

3695
	if (kvm_load_segment_descriptor(vcpu, tss->ldt_selector, 0, VCPU_SREG_LDTR))
3696 3697
		return 1;

3698
	if (kvm_load_segment_descriptor(vcpu, tss->es, 1, VCPU_SREG_ES))
3699 3700
		return 1;

3701
	if (kvm_load_segment_descriptor(vcpu, tss->cs, 9, VCPU_SREG_CS))
3702 3703
		return 1;

3704
	if (kvm_load_segment_descriptor(vcpu, tss->ss, 1, VCPU_SREG_SS))
3705 3706
		return 1;

3707
	if (kvm_load_segment_descriptor(vcpu, tss->ds, 1, VCPU_SREG_DS))
3708 3709
		return 1;

3710
	if (kvm_load_segment_descriptor(vcpu, tss->fs, 1, VCPU_SREG_FS))
3711 3712
		return 1;

3713
	if (kvm_load_segment_descriptor(vcpu, tss->gs, 1, VCPU_SREG_GS))
3714 3715 3716 3717 3718 3719 3720
		return 1;
	return 0;
}

static void save_state_to_tss16(struct kvm_vcpu *vcpu,
				struct tss_segment_16 *tss)
{
3721
	tss->ip = kvm_rip_read(vcpu);
3722
	tss->flag = kvm_x86_ops->get_rflags(vcpu);
3723 3724 3725 3726 3727 3728 3729 3730
	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);
3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742

	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)
{
3743
	kvm_rip_write(vcpu, tss->ip);
3744
	kvm_x86_ops->set_rflags(vcpu, tss->flag | 2);
3745 3746 3747 3748 3749 3750 3751 3752
	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);
3753

3754
	if (kvm_load_segment_descriptor(vcpu, tss->ldt, 0, VCPU_SREG_LDTR))
3755 3756
		return 1;

3757
	if (kvm_load_segment_descriptor(vcpu, tss->es, 1, VCPU_SREG_ES))
3758 3759
		return 1;

3760
	if (kvm_load_segment_descriptor(vcpu, tss->cs, 9, VCPU_SREG_CS))
3761 3762
		return 1;

3763
	if (kvm_load_segment_descriptor(vcpu, tss->ss, 1, VCPU_SREG_SS))
3764 3765
		return 1;

3766
	if (kvm_load_segment_descriptor(vcpu, tss->ds, 1, VCPU_SREG_DS))
3767 3768 3769 3770
		return 1;
	return 0;
}

3771
static int kvm_task_switch_16(struct kvm_vcpu *vcpu, u16 tss_selector,
3772
		       u32 old_tss_base,
3773 3774 3775 3776 3777
		       struct desc_struct *nseg_desc)
{
	struct tss_segment_16 tss_segment_16;
	int ret = 0;

3778 3779
	if (kvm_read_guest(vcpu->kvm, old_tss_base, &tss_segment_16,
			   sizeof tss_segment_16))
3780 3781 3782 3783
		goto out;

	save_state_to_tss16(vcpu, &tss_segment_16);

3784 3785
	if (kvm_write_guest(vcpu->kvm, old_tss_base, &tss_segment_16,
			    sizeof tss_segment_16))
3786
		goto out;
3787 3788 3789 3790 3791

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

3792 3793 3794 3795 3796 3797 3798 3799
	if (load_state_from_tss16(vcpu, &tss_segment_16))
		goto out;

	ret = 1;
out:
	return ret;
}

3800
static int kvm_task_switch_32(struct kvm_vcpu *vcpu, u16 tss_selector,
3801
		       u32 old_tss_base,
3802 3803 3804 3805 3806
		       struct desc_struct *nseg_desc)
{
	struct tss_segment_32 tss_segment_32;
	int ret = 0;

3807 3808
	if (kvm_read_guest(vcpu->kvm, old_tss_base, &tss_segment_32,
			   sizeof tss_segment_32))
3809 3810 3811 3812
		goto out;

	save_state_to_tss32(vcpu, &tss_segment_32);

3813 3814 3815 3816 3817 3818
	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))
3819
		goto out;
3820

3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834
	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;
3835 3836
	u32 old_tss_base = get_segment_base(vcpu, VCPU_SREG_TR);
	u16 old_tss_sel = get_segment_selector(vcpu, VCPU_SREG_TR);
3837

3838
	old_tss_base = vcpu->arch.mmu.gva_to_gpa(vcpu, old_tss_base);
3839

3840 3841 3842
	/* FIXME: Handle errors. Failure to read either TSS or their
	 * descriptors should generate a pagefault.
	 */
3843 3844 3845
	if (load_guest_segment_descriptor(vcpu, tss_selector, &nseg_desc))
		goto out;

3846
	if (load_guest_segment_descriptor(vcpu, old_tss_sel, &cseg_desc))
3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864
		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) {
3865
		cseg_desc.type &= ~(1 << 1); //clear the B flag
3866
		save_guest_segment_descriptor(vcpu, old_tss_sel, &cseg_desc);
3867 3868 3869 3870 3871 3872 3873 3874 3875 3876
	}

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

	kvm_x86_ops->skip_emulated_instruction(vcpu);

	if (nseg_desc.type & 8)
3877
		ret = kvm_task_switch_32(vcpu, tss_selector, old_tss_base,
3878 3879
					 &nseg_desc);
	else
3880
		ret = kvm_task_switch_16(vcpu, tss_selector, old_tss_base,
3881 3882 3883 3884 3885 3886 3887 3888
					 &nseg_desc);

	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) {
3889
		nseg_desc.type |= (1 << 1);
3890 3891 3892 3893 3894 3895 3896
		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;
3897
	kvm_set_segment(vcpu, &tr_seg, VCPU_SREG_TR);
3898 3899 3900 3901 3902
out:
	return ret;
}
EXPORT_SYMBOL_GPL(kvm_task_switch);

3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918
int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
				  struct kvm_sregs *sregs)
{
	int mmu_reset_needed = 0;
	int i, pending_vec, max_bits;
	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);

3919 3920 3921
	vcpu->arch.cr2 = sregs->cr2;
	mmu_reset_needed |= vcpu->arch.cr3 != sregs->cr3;
	vcpu->arch.cr3 = sregs->cr3;
3922

3923
	kvm_set_cr8(vcpu, sregs->cr8);
3924

3925
	mmu_reset_needed |= vcpu->arch.shadow_efer != sregs->efer;
3926 3927 3928 3929 3930
	kvm_x86_ops->set_efer(vcpu, sregs->efer);
	kvm_set_apic_base(vcpu, sregs->apic_base);

	kvm_x86_ops->decache_cr4_guest_bits(vcpu);

3931
	mmu_reset_needed |= vcpu->arch.cr0 != sregs->cr0;
3932
	kvm_x86_ops->set_cr0(vcpu, sregs->cr0);
3933
	vcpu->arch.cr0 = sregs->cr0;
3934

3935
	mmu_reset_needed |= vcpu->arch.cr4 != sregs->cr4;
3936 3937
	kvm_x86_ops->set_cr4(vcpu, sregs->cr4);
	if (!is_long_mode(vcpu) && is_pae(vcpu))
3938
		load_pdptrs(vcpu, vcpu->arch.cr3);
3939 3940 3941 3942 3943

	if (mmu_reset_needed)
		kvm_mmu_reset_context(vcpu);

	if (!irqchip_in_kernel(vcpu->kvm)) {
3944 3945 3946 3947 3948 3949
		memcpy(vcpu->arch.irq_pending, sregs->interrupt_bitmap,
		       sizeof vcpu->arch.irq_pending);
		vcpu->arch.irq_summary = 0;
		for (i = 0; i < ARRAY_SIZE(vcpu->arch.irq_pending); ++i)
			if (vcpu->arch.irq_pending[i])
				__set_bit(i, &vcpu->arch.irq_summary);
3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962
	} else {
		max_bits = (sizeof sregs->interrupt_bitmap) << 3;
		pending_vec = find_first_bit(
			(const unsigned long *)sregs->interrupt_bitmap,
			max_bits);
		/* Only pending external irq is handled here */
		if (pending_vec < max_bits) {
			kvm_x86_ops->set_irq(vcpu, pending_vec);
			pr_debug("Set back pending irq %d\n",
				 pending_vec);
		}
	}

3963 3964 3965 3966 3967 3968
	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);
3969

3970 3971
	kvm_set_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
	kvm_set_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
3972

M
Marcelo Tosatti 已提交
3973 3974 3975 3976 3977 3978
	/* 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;

3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997
	vcpu_put(vcpu);

	return 0;
}

int kvm_arch_vcpu_ioctl_debug_guest(struct kvm_vcpu *vcpu,
				    struct kvm_debug_guest *dbg)
{
	int r;

	vcpu_load(vcpu);

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

	vcpu_put(vcpu);

	return r;
}

3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018
/*
 * 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
};

4019 4020 4021 4022 4023 4024 4025 4026 4027 4028
/*
 * 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);
4029
	down_read(&vcpu->kvm->slots_lock);
4030
	gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, vaddr);
4031
	up_read(&vcpu->kvm->slots_lock);
4032 4033 4034 4035 4036 4037 4038 4039 4040
	tr->physical_address = gpa;
	tr->valid = gpa != UNMAPPED_GVA;
	tr->writeable = 1;
	tr->usermode = 0;
	vcpu_put(vcpu);

	return 0;
}

4041 4042
int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
4043
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062

	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)
{
4063
	struct fxsave *fxsave = (struct fxsave *)&vcpu->arch.guest_fx_image;
4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084

	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;

4085 4086 4087 4088 4089 4090 4091
	/*
	 * 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())
4092
		kvm_fx_save(&vcpu->arch.host_fx_image);
4093

4094 4095
	/* Initialize guest FPU by resetting ours and saving into guest's */
	preempt_disable();
4096 4097 4098 4099
	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);
4100 4101
	preempt_enable();

4102
	vcpu->arch.cr0 |= X86_CR0_ET;
4103
	after_mxcsr_mask = offsetof(struct i387_fxsave_struct, st_space);
4104 4105
	vcpu->arch.guest_fx_image.mxcsr = 0x1f80;
	memset((void *)&vcpu->arch.guest_fx_image + after_mxcsr_mask,
4106 4107 4108 4109 4110 4111 4112 4113 4114 4115
	       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;
4116 4117
	kvm_fx_save(&vcpu->arch.host_fx_image);
	kvm_fx_restore(&vcpu->arch.guest_fx_image);
4118 4119 4120 4121 4122 4123 4124 4125 4126
}
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;
4127 4128
	kvm_fx_save(&vcpu->arch.guest_fx_image);
	kvm_fx_restore(&vcpu->arch.host_fx_image);
A
Avi Kivity 已提交
4129
	++vcpu->stat.fpu_reload;
4130 4131
}
EXPORT_SYMBOL_GPL(kvm_put_guest_fpu);
4132 4133 4134 4135 4136 4137 4138 4139 4140

void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
	kvm_x86_ops->vcpu_free(vcpu);
}

struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm,
						unsigned int id)
{
4141 4142
	return kvm_x86_ops->vcpu_create(kvm, id);
}
4143

4144 4145 4146
int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
{
	int r;
4147 4148

	/* We do fxsave: this must be aligned. */
4149
	BUG_ON((unsigned long)&vcpu->arch.host_fx_image & 0xF);
4150 4151 4152 4153 4154 4155 4156 4157 4158

	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;

4159
	return 0;
4160 4161
free_vcpu:
	kvm_x86_ops->vcpu_free(vcpu);
4162
	return r;
4163 4164
}

4165
void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212
{
	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)
{
	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;

4213
	vcpu->arch.mmu.root_hpa = INVALID_PAGE;
4214
	if (!irqchip_in_kernel(kvm) || vcpu->vcpu_id == 0)
4215
		vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
4216
	else
4217
		vcpu->arch.mp_state = KVM_MP_STATE_UNINITIALIZED;
4218 4219 4220 4221 4222 4223

	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
	if (!page) {
		r = -ENOMEM;
		goto fail;
	}
4224
	vcpu->arch.pio_data = page_address(page);
4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240

	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:
4241
	free_page((unsigned long)vcpu->arch.pio_data);
4242 4243 4244 4245 4246 4247 4248
fail:
	return r;
}

void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
{
	kvm_free_lapic(vcpu);
4249
	down_read(&vcpu->kvm->slots_lock);
4250
	kvm_mmu_destroy(vcpu);
4251
	up_read(&vcpu->kvm->slots_lock);
4252
	free_page((unsigned long)vcpu->arch.pio_data);
4253
}
4254 4255 4256 4257 4258 4259 4260 4261

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

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

4262
	INIT_LIST_HEAD(&kvm->arch.active_mmu_pages);
B
Ben-Ami Yassour 已提交
4263
	INIT_LIST_HEAD(&kvm->arch.assigned_dev_head);
4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295

	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;
		}
	}

}

void kvm_arch_destroy_vm(struct kvm *kvm)
{
B
Ben-Ami Yassour 已提交
4296
	kvm_iommu_unmap_guest(kvm);
4297
	kvm_free_all_assigned_devices(kvm);
S
Sheng Yang 已提交
4298
	kvm_free_pit(kvm);
4299 4300
	kfree(kvm->arch.vpic);
	kfree(kvm->arch.vioapic);
4301 4302
	kvm_free_vcpus(kvm);
	kvm_free_physmem(kvm);
4303 4304
	if (kvm->arch.apic_access_page)
		put_page(kvm->arch.apic_access_page);
4305 4306
	if (kvm->arch.ept_identity_pagetable)
		put_page(kvm->arch.ept_identity_pagetable);
4307 4308
	kfree(kvm);
}
4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322

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) {
4323 4324
			unsigned long userspace_addr;

4325
			down_write(&current->mm->mmap_sem);
4326 4327 4328
			userspace_addr = do_mmap(NULL, 0,
						 npages * PAGE_SIZE,
						 PROT_READ | PROT_WRITE,
4329
						 MAP_PRIVATE | MAP_ANONYMOUS,
4330
						 0);
4331
			up_write(&current->mm->mmap_sem);
4332

4333 4334 4335 4336 4337 4338 4339
			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);
4340 4341 4342 4343
		} else {
			if (!old.user_alloc && old.rmap) {
				int ret;

4344
				down_write(&current->mm->mmap_sem);
4345 4346
				ret = do_munmap(current->mm, old.userspace_addr,
						old.npages * PAGE_SIZE);
4347
				up_write(&current->mm->mmap_sem);
4348 4349 4350 4351 4352 4353 4354 4355
				if (ret < 0)
					printk(KERN_WARNING
				       "kvm_vm_ioctl_set_memory_region: "
				       "failed to munmap memory\n");
			}
		}
	}

4356
	if (!kvm->arch.n_requested_mmu_pages) {
4357 4358 4359 4360 4361 4362 4363 4364 4365
		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);
	kvm_flush_remote_tlbs(kvm);

	return 0;
}
4366

4367 4368 4369 4370 4371
void kvm_arch_flush_shadow(struct kvm *kvm)
{
	kvm_mmu_zap_all(kvm);
}

4372 4373
int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
{
4374 4375
	return vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE
	       || vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED;
4376
}
4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388

static void vcpu_kick_intr(void *info)
{
#ifdef DEBUG
	struct kvm_vcpu *vcpu = (struct kvm_vcpu *)info;
	printk(KERN_DEBUG "vcpu_kick_intr %p \n", vcpu);
#endif
}

void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
{
	int ipi_pcpu = vcpu->cpu;
4389
	int cpu = get_cpu();
4390 4391 4392 4393 4394

	if (waitqueue_active(&vcpu->wq)) {
		wake_up_interruptible(&vcpu->wq);
		++vcpu->stat.halt_wakeup;
	}
4395 4396 4397 4398 4399
	/*
	 * We may be called synchronously with irqs disabled in guest mode,
	 * So need not to call smp_call_function_single() in that case.
	 */
	if (vcpu->guest_mode && vcpu->cpu != cpu)
4400
		smp_call_function_single(ipi_pcpu, vcpu_kick_intr, vcpu, 0);
4401
	put_cpu();
4402
}