mmu.c 34.9 KB
Newer Older
A
Avi Kivity 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
/*
 * Kernel-based Virtual Machine driver for Linux
 *
 * This module enables machines with Intel VT-x extensions to run virtual
 * machines without emulation or binary translation.
 *
 * MMU support
 *
 * Copyright (C) 2006 Qumranet, Inc.
 *
 * Authors:
 *   Yaniv Kamay  <yaniv@qumranet.com>
 *   Avi Kivity   <avi@qumranet.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */
#include <linux/types.h>
#include <linux/string.h>
#include <asm/page.h>
#include <linux/mm.h>
#include <linux/highmem.h>
#include <linux/module.h>

#include "vmx.h"
#include "kvm.h"

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
#undef MMU_DEBUG

#undef AUDIT

#ifdef AUDIT
static void kvm_mmu_audit(struct kvm_vcpu *vcpu, const char *msg);
#else
static void kvm_mmu_audit(struct kvm_vcpu *vcpu, const char *msg) {}
#endif

#ifdef MMU_DEBUG

#define pgprintk(x...) do { if (dbg) printk(x); } while (0)
#define rmap_printk(x...) do { if (dbg) printk(x); } while (0)

#else

#define pgprintk(x...) do { } while (0)
#define rmap_printk(x...) do { } while (0)

#endif

#if defined(MMU_DEBUG) || defined(AUDIT)
static int dbg = 1;
#endif
A
Avi Kivity 已提交
54 55 56 57 58 59 60

#define ASSERT(x)							\
	if (!(x)) {							\
		printk(KERN_WARNING "assertion failed %s:%d: %s\n",	\
		       __FILE__, __LINE__, #x);				\
	}

61 62 63 64
#define PT64_PT_BITS 9
#define PT64_ENT_PER_PAGE (1 << PT64_PT_BITS)
#define PT32_PT_BITS 10
#define PT32_ENT_PER_PAGE (1 << PT32_PT_BITS)
A
Avi Kivity 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89

#define PT_WRITABLE_SHIFT 1

#define PT_PRESENT_MASK (1ULL << 0)
#define PT_WRITABLE_MASK (1ULL << PT_WRITABLE_SHIFT)
#define PT_USER_MASK (1ULL << 2)
#define PT_PWT_MASK (1ULL << 3)
#define PT_PCD_MASK (1ULL << 4)
#define PT_ACCESSED_MASK (1ULL << 5)
#define PT_DIRTY_MASK (1ULL << 6)
#define PT_PAGE_SIZE_MASK (1ULL << 7)
#define PT_PAT_MASK (1ULL << 7)
#define PT_GLOBAL_MASK (1ULL << 8)
#define PT64_NX_MASK (1ULL << 63)

#define PT_PAT_SHIFT 7
#define PT_DIR_PAT_SHIFT 12
#define PT_DIR_PAT_MASK (1ULL << PT_DIR_PAT_SHIFT)

#define PT32_DIR_PSE36_SIZE 4
#define PT32_DIR_PSE36_SHIFT 13
#define PT32_DIR_PSE36_MASK (((1ULL << PT32_DIR_PSE36_SIZE) - 1) << PT32_DIR_PSE36_SHIFT)


#define PT32_PTE_COPY_MASK \
90
	(PT_PRESENT_MASK | PT_ACCESSED_MASK | PT_DIRTY_MASK | PT_GLOBAL_MASK)
A
Avi Kivity 已提交
91

92
#define PT64_PTE_COPY_MASK (PT64_NX_MASK | PT32_PTE_COPY_MASK)
A
Avi Kivity 已提交
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145

#define PT_FIRST_AVAIL_BITS_SHIFT 9
#define PT64_SECOND_AVAIL_BITS_SHIFT 52

#define PT_SHADOW_PS_MARK (1ULL << PT_FIRST_AVAIL_BITS_SHIFT)
#define PT_SHADOW_IO_MARK (1ULL << PT_FIRST_AVAIL_BITS_SHIFT)

#define PT_SHADOW_WRITABLE_SHIFT (PT_FIRST_AVAIL_BITS_SHIFT + 1)
#define PT_SHADOW_WRITABLE_MASK (1ULL << PT_SHADOW_WRITABLE_SHIFT)

#define PT_SHADOW_USER_SHIFT (PT_SHADOW_WRITABLE_SHIFT + 1)
#define PT_SHADOW_USER_MASK (1ULL << (PT_SHADOW_USER_SHIFT))

#define PT_SHADOW_BITS_OFFSET (PT_SHADOW_WRITABLE_SHIFT - PT_WRITABLE_SHIFT)

#define VALID_PAGE(x) ((x) != INVALID_PAGE)

#define PT64_LEVEL_BITS 9

#define PT64_LEVEL_SHIFT(level) \
		( PAGE_SHIFT + (level - 1) * PT64_LEVEL_BITS )

#define PT64_LEVEL_MASK(level) \
		(((1ULL << PT64_LEVEL_BITS) - 1) << PT64_LEVEL_SHIFT(level))

#define PT64_INDEX(address, level)\
	(((address) >> PT64_LEVEL_SHIFT(level)) & ((1 << PT64_LEVEL_BITS) - 1))


#define PT32_LEVEL_BITS 10

#define PT32_LEVEL_SHIFT(level) \
		( PAGE_SHIFT + (level - 1) * PT32_LEVEL_BITS )

#define PT32_LEVEL_MASK(level) \
		(((1ULL << PT32_LEVEL_BITS) - 1) << PT32_LEVEL_SHIFT(level))

#define PT32_INDEX(address, level)\
	(((address) >> PT32_LEVEL_SHIFT(level)) & ((1 << PT32_LEVEL_BITS) - 1))


#define PT64_BASE_ADDR_MASK (((1ULL << 52) - 1) & PAGE_MASK)
#define PT64_DIR_BASE_ADDR_MASK \
	(PT64_BASE_ADDR_MASK & ~((1ULL << (PAGE_SHIFT + PT64_LEVEL_BITS)) - 1))

#define PT32_BASE_ADDR_MASK PAGE_MASK
#define PT32_DIR_BASE_ADDR_MASK \
	(PAGE_MASK & ~((1ULL << (PAGE_SHIFT + PT32_LEVEL_BITS)) - 1))


#define PFERR_PRESENT_MASK (1U << 0)
#define PFERR_WRITE_MASK (1U << 1)
#define PFERR_USER_MASK (1U << 2)
146
#define PFERR_FETCH_MASK (1U << 4)
A
Avi Kivity 已提交
147 148 149 150 151 152 153 154

#define PT64_ROOT_LEVEL 4
#define PT32_ROOT_LEVEL 2
#define PT32E_ROOT_LEVEL 3

#define PT_DIRECTORY_LEVEL 2
#define PT_PAGE_TABLE_LEVEL 1

155 156 157 158 159 160 161
#define RMAP_EXT 4

struct kvm_rmap_desc {
	u64 *shadow_ptes[RMAP_EXT];
	struct kvm_rmap_desc *more;
};

A
Avi Kivity 已提交
162 163 164 165 166 167 168 169 170 171
static int is_write_protection(struct kvm_vcpu *vcpu)
{
	return vcpu->cr0 & CR0_WP_MASK;
}

static int is_cpuid_PSE36(void)
{
	return 1;
}

172 173 174 175 176
static int is_nx(struct kvm_vcpu *vcpu)
{
	return vcpu->shadow_efer & EFER_NX;
}

A
Avi Kivity 已提交
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
static int is_present_pte(unsigned long pte)
{
	return pte & PT_PRESENT_MASK;
}

static int is_writeble_pte(unsigned long pte)
{
	return pte & PT_WRITABLE_MASK;
}

static int is_io_pte(unsigned long pte)
{
	return pte & PT_SHADOW_IO_MARK;
}

192 193 194 195 196 197
static int is_rmap_pte(u64 pte)
{
	return (pte & (PT_WRITABLE_MASK | PT_PRESENT_MASK))
		== (PT_WRITABLE_MASK | PT_PRESENT_MASK);
}

198 199
static int mmu_topup_memory_cache(struct kvm_mmu_memory_cache *cache,
				  size_t objsize, int min)
200 201 202 203
{
	void *obj;

	if (cache->nobjs >= min)
204
		return 0;
205 206 207
	while (cache->nobjs < ARRAY_SIZE(cache->objects)) {
		obj = kzalloc(objsize, GFP_NOWAIT);
		if (!obj)
208
			return -ENOMEM;
209 210
		cache->objects[cache->nobjs++] = obj;
	}
211
	return 0;
212 213 214 215 216 217 218 219
}

static void mmu_free_memory_cache(struct kvm_mmu_memory_cache *mc)
{
	while (mc->nobjs)
		kfree(mc->objects[--mc->nobjs]);
}

220
static int mmu_topup_memory_caches(struct kvm_vcpu *vcpu)
221
{
222 223 224 225 226 227 228 229 230 231
	int r;

	r = mmu_topup_memory_cache(&vcpu->mmu_pte_chain_cache,
				   sizeof(struct kvm_pte_chain), 4);
	if (r)
		goto out;
	r = mmu_topup_memory_cache(&vcpu->mmu_rmap_desc_cache,
				   sizeof(struct kvm_rmap_desc), 1);
out:
	return r;
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
}

static void mmu_free_memory_caches(struct kvm_vcpu *vcpu)
{
	mmu_free_memory_cache(&vcpu->mmu_pte_chain_cache);
	mmu_free_memory_cache(&vcpu->mmu_rmap_desc_cache);
}

static void *mmu_memory_cache_alloc(struct kvm_mmu_memory_cache *mc,
				    size_t size)
{
	void *p;

	BUG_ON(!mc->nobjs);
	p = mc->objects[--mc->nobjs];
	memset(p, 0, size);
	return p;
}

static void mmu_memory_cache_free(struct kvm_mmu_memory_cache *mc, void *obj)
{
	if (mc->nobjs < KVM_NR_MEM_OBJS)
		mc->objects[mc->nobjs++] = obj;
	else
		kfree(obj);
}

static struct kvm_pte_chain *mmu_alloc_pte_chain(struct kvm_vcpu *vcpu)
{
	return mmu_memory_cache_alloc(&vcpu->mmu_pte_chain_cache,
				      sizeof(struct kvm_pte_chain));
}

static void mmu_free_pte_chain(struct kvm_vcpu *vcpu,
			       struct kvm_pte_chain *pc)
{
	mmu_memory_cache_free(&vcpu->mmu_pte_chain_cache, pc);
}

static struct kvm_rmap_desc *mmu_alloc_rmap_desc(struct kvm_vcpu *vcpu)
{
	return mmu_memory_cache_alloc(&vcpu->mmu_rmap_desc_cache,
				      sizeof(struct kvm_rmap_desc));
}

static void mmu_free_rmap_desc(struct kvm_vcpu *vcpu,
			       struct kvm_rmap_desc *rd)
{
	mmu_memory_cache_free(&vcpu->mmu_rmap_desc_cache, rd);
}

283 284 285 286 287 288 289 290 291
/*
 * Reverse mapping data structures:
 *
 * If page->private bit zero is zero, then page->private points to the
 * shadow page table entry that points to page_address(page).
 *
 * If page->private bit zero is one, (then page->private & ~1) points
 * to a struct kvm_rmap_desc containing more mappings.
 */
292
static void rmap_add(struct kvm_vcpu *vcpu, u64 *spte)
293 294 295 296 297 298 299 300
{
	struct page *page;
	struct kvm_rmap_desc *desc;
	int i;

	if (!is_rmap_pte(*spte))
		return;
	page = pfn_to_page((*spte & PT64_BASE_ADDR_MASK) >> PAGE_SHIFT);
301
	if (!page_private(page)) {
302
		rmap_printk("rmap_add: %p %llx 0->1\n", spte, *spte);
303 304
		set_page_private(page,(unsigned long)spte);
	} else if (!(page_private(page) & 1)) {
305
		rmap_printk("rmap_add: %p %llx 1->many\n", spte, *spte);
306
		desc = mmu_alloc_rmap_desc(vcpu);
307
		desc->shadow_ptes[0] = (u64 *)page_private(page);
308
		desc->shadow_ptes[1] = spte;
309
		set_page_private(page,(unsigned long)desc | 1);
310 311
	} else {
		rmap_printk("rmap_add: %p %llx many->many\n", spte, *spte);
312
		desc = (struct kvm_rmap_desc *)(page_private(page) & ~1ul);
313 314 315
		while (desc->shadow_ptes[RMAP_EXT-1] && desc->more)
			desc = desc->more;
		if (desc->shadow_ptes[RMAP_EXT-1]) {
316
			desc->more = mmu_alloc_rmap_desc(vcpu);
317 318 319 320 321 322 323 324
			desc = desc->more;
		}
		for (i = 0; desc->shadow_ptes[i]; ++i)
			;
		desc->shadow_ptes[i] = spte;
	}
}

325 326
static void rmap_desc_remove_entry(struct kvm_vcpu *vcpu,
				   struct page *page,
327 328 329 330 331 332 333 334 335
				   struct kvm_rmap_desc *desc,
				   int i,
				   struct kvm_rmap_desc *prev_desc)
{
	int j;

	for (j = RMAP_EXT - 1; !desc->shadow_ptes[j] && j > i; --j)
		;
	desc->shadow_ptes[i] = desc->shadow_ptes[j];
A
Al Viro 已提交
336
	desc->shadow_ptes[j] = NULL;
337 338 339
	if (j != 0)
		return;
	if (!prev_desc && !desc->more)
340
		set_page_private(page,(unsigned long)desc->shadow_ptes[0]);
341 342 343 344
	else
		if (prev_desc)
			prev_desc->more = desc->more;
		else
345
			set_page_private(page,(unsigned long)desc->more | 1);
346
	mmu_free_rmap_desc(vcpu, desc);
347 348
}

349
static void rmap_remove(struct kvm_vcpu *vcpu, u64 *spte)
350 351 352 353 354 355 356 357 358
{
	struct page *page;
	struct kvm_rmap_desc *desc;
	struct kvm_rmap_desc *prev_desc;
	int i;

	if (!is_rmap_pte(*spte))
		return;
	page = pfn_to_page((*spte & PT64_BASE_ADDR_MASK) >> PAGE_SHIFT);
359
	if (!page_private(page)) {
360 361
		printk(KERN_ERR "rmap_remove: %p %llx 0->BUG\n", spte, *spte);
		BUG();
362
	} else if (!(page_private(page) & 1)) {
363
		rmap_printk("rmap_remove:  %p %llx 1->0\n", spte, *spte);
364
		if ((u64 *)page_private(page) != spte) {
365 366 367 368
			printk(KERN_ERR "rmap_remove:  %p %llx 1->BUG\n",
			       spte, *spte);
			BUG();
		}
369
		set_page_private(page,0);
370 371
	} else {
		rmap_printk("rmap_remove:  %p %llx many->many\n", spte, *spte);
372
		desc = (struct kvm_rmap_desc *)(page_private(page) & ~1ul);
373 374 375 376
		prev_desc = NULL;
		while (desc) {
			for (i = 0; i < RMAP_EXT && desc->shadow_ptes[i]; ++i)
				if (desc->shadow_ptes[i] == spte) {
377 378
					rmap_desc_remove_entry(vcpu, page,
							       desc, i,
379 380 381 382 383 384 385 386 387 388
							       prev_desc);
					return;
				}
			prev_desc = desc;
			desc = desc->more;
		}
		BUG();
	}
}

389
static void rmap_write_protect(struct kvm_vcpu *vcpu, u64 gfn)
390
{
391
	struct kvm *kvm = vcpu->kvm;
392 393 394 395 396 397 398 399 400
	struct page *page;
	struct kvm_memory_slot *slot;
	struct kvm_rmap_desc *desc;
	u64 *spte;

	slot = gfn_to_memslot(kvm, gfn);
	BUG_ON(!slot);
	page = gfn_to_page(slot, gfn);

401 402 403
	while (page_private(page)) {
		if (!(page_private(page) & 1))
			spte = (u64 *)page_private(page);
404
		else {
405
			desc = (struct kvm_rmap_desc *)(page_private(page) & ~1ul);
406 407 408 409 410 411 412 413
			spte = desc->shadow_ptes[0];
		}
		BUG_ON(!spte);
		BUG_ON((*spte & PT64_BASE_ADDR_MASK) !=
		       page_to_pfn(page) << PAGE_SHIFT);
		BUG_ON(!(*spte & PT_PRESENT_MASK));
		BUG_ON(!(*spte & PT_WRITABLE_MASK));
		rmap_printk("rmap_write_protect: spte %p %llx\n", spte, *spte);
414
		rmap_remove(vcpu, spte);
415
		kvm_arch_ops->tlb_flush(vcpu);
416 417 418 419
		*spte &= ~(u64)PT_WRITABLE_MASK;
	}
}

A
Avi Kivity 已提交
420 421
static int is_empty_shadow_page(hpa_t page_hpa)
{
422 423 424 425
	u64 *pos;
	u64 *end;

	for (pos = __va(page_hpa), end = pos + PAGE_SIZE / sizeof(u64);
A
Avi Kivity 已提交
426
		      pos != end; pos++)
427 428 429
		if (*pos != 0) {
			printk(KERN_ERR "%s: %p %llx\n", __FUNCTION__,
			       pos, *pos);
A
Avi Kivity 已提交
430
			return 0;
431
		}
A
Avi Kivity 已提交
432 433 434
	return 1;
}

435 436 437 438
static void kvm_mmu_free_page(struct kvm_vcpu *vcpu, hpa_t page_hpa)
{
	struct kvm_mmu_page *page_head = page_header(page_hpa);

439
	ASSERT(is_empty_shadow_page(page_hpa));
440 441 442 443 444 445
	list_del(&page_head->link);
	page_head->page_hpa = page_hpa;
	list_add(&page_head->link, &vcpu->free_pages);
	++vcpu->kvm->n_free_mmu_pages;
}

446 447 448 449 450
static unsigned kvm_page_table_hashfn(gfn_t gfn)
{
	return gfn;
}

451 452
static struct kvm_mmu_page *kvm_mmu_alloc_page(struct kvm_vcpu *vcpu,
					       u64 *parent_pte)
A
Avi Kivity 已提交
453 454 455 456
{
	struct kvm_mmu_page *page;

	if (list_empty(&vcpu->free_pages))
457
		return NULL;
A
Avi Kivity 已提交
458 459 460 461 462 463 464

	page = list_entry(vcpu->free_pages.next, struct kvm_mmu_page, link);
	list_del(&page->link);
	list_add(&page->link, &vcpu->kvm->active_mmu_pages);
	ASSERT(is_empty_shadow_page(page->page_hpa));
	page->slot_bitmap = 0;
	page->global = 1;
465
	page->multimapped = 0;
A
Avi Kivity 已提交
466
	page->parent_pte = parent_pte;
A
Avi Kivity 已提交
467
	--vcpu->kvm->n_free_mmu_pages;
468
	return page;
A
Avi Kivity 已提交
469 470
}

471 472
static void mmu_page_add_parent_pte(struct kvm_vcpu *vcpu,
				    struct kvm_mmu_page *page, u64 *parent_pte)
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487
{
	struct kvm_pte_chain *pte_chain;
	struct hlist_node *node;
	int i;

	if (!parent_pte)
		return;
	if (!page->multimapped) {
		u64 *old = page->parent_pte;

		if (!old) {
			page->parent_pte = parent_pte;
			return;
		}
		page->multimapped = 1;
488
		pte_chain = mmu_alloc_pte_chain(vcpu);
489 490 491 492 493 494 495 496 497 498 499 500 501
		INIT_HLIST_HEAD(&page->parent_ptes);
		hlist_add_head(&pte_chain->link, &page->parent_ptes);
		pte_chain->parent_ptes[0] = old;
	}
	hlist_for_each_entry(pte_chain, node, &page->parent_ptes, link) {
		if (pte_chain->parent_ptes[NR_PTE_CHAIN_ENTRIES-1])
			continue;
		for (i = 0; i < NR_PTE_CHAIN_ENTRIES; ++i)
			if (!pte_chain->parent_ptes[i]) {
				pte_chain->parent_ptes[i] = parent_pte;
				return;
			}
	}
502
	pte_chain = mmu_alloc_pte_chain(vcpu);
503 504 505 506 507
	BUG_ON(!pte_chain);
	hlist_add_head(&pte_chain->link, &page->parent_ptes);
	pte_chain->parent_ptes[0] = parent_pte;
}

508 509
static void mmu_page_remove_parent_pte(struct kvm_vcpu *vcpu,
				       struct kvm_mmu_page *page,
510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526
				       u64 *parent_pte)
{
	struct kvm_pte_chain *pte_chain;
	struct hlist_node *node;
	int i;

	if (!page->multimapped) {
		BUG_ON(page->parent_pte != parent_pte);
		page->parent_pte = NULL;
		return;
	}
	hlist_for_each_entry(pte_chain, node, &page->parent_ptes, link)
		for (i = 0; i < NR_PTE_CHAIN_ENTRIES; ++i) {
			if (!pte_chain->parent_ptes[i])
				break;
			if (pte_chain->parent_ptes[i] != parent_pte)
				continue;
527 528
			while (i + 1 < NR_PTE_CHAIN_ENTRIES
				&& pte_chain->parent_ptes[i + 1]) {
529 530 531 532 533
				pte_chain->parent_ptes[i]
					= pte_chain->parent_ptes[i + 1];
				++i;
			}
			pte_chain->parent_ptes[i] = NULL;
534 535
			if (i == 0) {
				hlist_del(&pte_chain->link);
536
				mmu_free_pte_chain(vcpu, pte_chain);
537 538 539 540 541
				if (hlist_empty(&page->parent_ptes)) {
					page->multimapped = 0;
					page->parent_pte = NULL;
				}
			}
542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
			return;
		}
	BUG();
}

static struct kvm_mmu_page *kvm_mmu_lookup_page(struct kvm_vcpu *vcpu,
						gfn_t gfn)
{
	unsigned index;
	struct hlist_head *bucket;
	struct kvm_mmu_page *page;
	struct hlist_node *node;

	pgprintk("%s: looking for gfn %lx\n", __FUNCTION__, gfn);
	index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
	bucket = &vcpu->kvm->mmu_page_hash[index];
	hlist_for_each_entry(page, node, bucket, hash_link)
		if (page->gfn == gfn && !page->role.metaphysical) {
			pgprintk("%s: found role %x\n",
				 __FUNCTION__, page->role.word);
			return page;
		}
	return NULL;
}

static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
					     gfn_t gfn,
					     gva_t gaddr,
					     unsigned level,
					     int metaphysical,
					     u64 *parent_pte)
{
	union kvm_mmu_page_role role;
	unsigned index;
	unsigned quadrant;
	struct hlist_head *bucket;
	struct kvm_mmu_page *page;
	struct hlist_node *node;

	role.word = 0;
	role.glevels = vcpu->mmu.root_level;
	role.level = level;
	role.metaphysical = metaphysical;
	if (vcpu->mmu.root_level <= PT32_ROOT_LEVEL) {
		quadrant = gaddr >> (PAGE_SHIFT + (PT64_PT_BITS * level));
		quadrant &= (1 << ((PT32_PT_BITS - PT64_PT_BITS) * level)) - 1;
		role.quadrant = quadrant;
	}
	pgprintk("%s: looking gfn %lx role %x\n", __FUNCTION__,
		 gfn, role.word);
	index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
	bucket = &vcpu->kvm->mmu_page_hash[index];
	hlist_for_each_entry(page, node, bucket, hash_link)
		if (page->gfn == gfn && page->role.word == role.word) {
596
			mmu_page_add_parent_pte(vcpu, page, parent_pte);
597 598 599 600 601 602 603 604 605 606
			pgprintk("%s: found\n", __FUNCTION__);
			return page;
		}
	page = kvm_mmu_alloc_page(vcpu, parent_pte);
	if (!page)
		return page;
	pgprintk("%s: adding gfn %lx role %x\n", __FUNCTION__, gfn, role.word);
	page->gfn = gfn;
	page->role = role;
	hlist_add_head(&page->hash_link, bucket);
607
	if (!metaphysical)
608
		rmap_write_protect(vcpu, gfn);
609 610 611
	return page;
}

612 613 614
static void kvm_mmu_page_unlink_children(struct kvm_vcpu *vcpu,
					 struct kvm_mmu_page *page)
{
615 616 617 618 619 620 621 622 623
	unsigned i;
	u64 *pt;
	u64 ent;

	pt = __va(page->page_hpa);

	if (page->role.level == PT_PAGE_TABLE_LEVEL) {
		for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
			if (pt[i] & PT_PRESENT_MASK)
624
				rmap_remove(vcpu, &pt[i]);
625 626
			pt[i] = 0;
		}
627
		kvm_arch_ops->tlb_flush(vcpu);
628 629 630 631 632 633 634 635 636 637
		return;
	}

	for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
		ent = pt[i];

		pt[i] = 0;
		if (!(ent & PT_PRESENT_MASK))
			continue;
		ent &= PT64_BASE_ADDR_MASK;
638
		mmu_page_remove_parent_pte(vcpu, page_header(ent), &pt[i]);
639
	}
640 641
}

642 643 644 645
static void kvm_mmu_put_page(struct kvm_vcpu *vcpu,
			     struct kvm_mmu_page *page,
			     u64 *parent_pte)
{
646
	mmu_page_remove_parent_pte(vcpu, page, parent_pte);
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
}

static void kvm_mmu_zap_page(struct kvm_vcpu *vcpu,
			     struct kvm_mmu_page *page)
{
	u64 *parent_pte;

	while (page->multimapped || page->parent_pte) {
		if (!page->multimapped)
			parent_pte = page->parent_pte;
		else {
			struct kvm_pte_chain *chain;

			chain = container_of(page->parent_ptes.first,
					     struct kvm_pte_chain, link);
			parent_pte = chain->parent_ptes[0];
		}
664
		BUG_ON(!parent_pte);
665 666 667
		kvm_mmu_put_page(vcpu, page, parent_pte);
		*parent_pte = 0;
	}
668
	kvm_mmu_page_unlink_children(vcpu, page);
669 670 671 672 673 674 675
	if (!page->root_count) {
		hlist_del(&page->hash_link);
		kvm_mmu_free_page(vcpu, page->page_hpa);
	} else {
		list_del(&page->link);
		list_add(&page->link, &vcpu->kvm->active_mmu_pages);
	}
676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
}

static int kvm_mmu_unprotect_page(struct kvm_vcpu *vcpu, gfn_t gfn)
{
	unsigned index;
	struct hlist_head *bucket;
	struct kvm_mmu_page *page;
	struct hlist_node *node, *n;
	int r;

	pgprintk("%s: looking for gfn %lx\n", __FUNCTION__, gfn);
	r = 0;
	index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
	bucket = &vcpu->kvm->mmu_page_hash[index];
	hlist_for_each_entry_safe(page, node, n, bucket, hash_link)
		if (page->gfn == gfn && !page->role.metaphysical) {
692 693
			pgprintk("%s: gfn %lx role %x\n", __FUNCTION__, gfn,
				 page->role.word);
694 695 696 697
			kvm_mmu_zap_page(vcpu, page);
			r = 1;
		}
	return r;
698 699
}

A
Avi Kivity 已提交
700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
static void page_header_update_slot(struct kvm *kvm, void *pte, gpa_t gpa)
{
	int slot = memslot_id(kvm, gfn_to_memslot(kvm, gpa >> PAGE_SHIFT));
	struct kvm_mmu_page *page_head = page_header(__pa(pte));

	__set_bit(slot, &page_head->slot_bitmap);
}

hpa_t safe_gpa_to_hpa(struct kvm_vcpu *vcpu, gpa_t gpa)
{
	hpa_t hpa = gpa_to_hpa(vcpu, gpa);

	return is_error_hpa(hpa) ? bad_page_address | (gpa & ~PAGE_MASK): hpa;
}

hpa_t gpa_to_hpa(struct kvm_vcpu *vcpu, gpa_t gpa)
{
	struct kvm_memory_slot *slot;
	struct page *page;

	ASSERT((gpa & HPA_ERR_MASK) == 0);
	slot = gfn_to_memslot(vcpu->kvm, gpa >> PAGE_SHIFT);
	if (!slot)
		return gpa | HPA_ERR_MASK;
	page = gfn_to_page(slot, gpa >> PAGE_SHIFT);
	return ((hpa_t)page_to_pfn(page) << PAGE_SHIFT)
		| (gpa & (PAGE_SIZE-1));
}

hpa_t gva_to_hpa(struct kvm_vcpu *vcpu, gva_t gva)
{
	gpa_t gpa = vcpu->mmu.gva_to_gpa(vcpu, gva);

	if (gpa == UNMAPPED_GVA)
		return UNMAPPED_GVA;
	return gpa_to_hpa(vcpu, gpa);
}

static void nonpaging_new_cr3(struct kvm_vcpu *vcpu)
{
}

static int nonpaging_map(struct kvm_vcpu *vcpu, gva_t v, hpa_t p)
{
	int level = PT32E_ROOT_LEVEL;
	hpa_t table_addr = vcpu->mmu.root_hpa;

	for (; ; level--) {
		u32 index = PT64_INDEX(v, level);
		u64 *table;
750
		u64 pte;
A
Avi Kivity 已提交
751 752 753 754 755

		ASSERT(VALID_PAGE(table_addr));
		table = __va(table_addr);

		if (level == 1) {
756 757 758
			pte = table[index];
			if (is_present_pte(pte) && is_writeble_pte(pte))
				return 0;
A
Avi Kivity 已提交
759 760 761 762
			mark_page_dirty(vcpu->kvm, v >> PAGE_SHIFT);
			page_header_update_slot(vcpu->kvm, table, v);
			table[index] = p | PT_PRESENT_MASK | PT_WRITABLE_MASK |
								PT_USER_MASK;
763
			rmap_add(vcpu, &table[index]);
A
Avi Kivity 已提交
764 765 766 767
			return 0;
		}

		if (table[index] == 0) {
768
			struct kvm_mmu_page *new_table;
769
			gfn_t pseudo_gfn;
A
Avi Kivity 已提交
770

771 772 773 774 775
			pseudo_gfn = (v & PT64_DIR_BASE_ADDR_MASK)
				>> PAGE_SHIFT;
			new_table = kvm_mmu_get_page(vcpu, pseudo_gfn,
						     v, level - 1,
						     1, &table[index]);
776
			if (!new_table) {
A
Avi Kivity 已提交
777 778 779 780
				pgprintk("nonpaging_map: ENOMEM\n");
				return -ENOMEM;
			}

781 782
			table[index] = new_table->page_hpa | PT_PRESENT_MASK
				| PT_WRITABLE_MASK | PT_USER_MASK;
A
Avi Kivity 已提交
783 784 785 786 787
		}
		table_addr = table[index] & PT64_BASE_ADDR_MASK;
	}
}

788 789 790
static void mmu_free_roots(struct kvm_vcpu *vcpu)
{
	int i;
791
	struct kvm_mmu_page *page;
792 793 794 795 796 797

#ifdef CONFIG_X86_64
	if (vcpu->mmu.shadow_root_level == PT64_ROOT_LEVEL) {
		hpa_t root = vcpu->mmu.root_hpa;

		ASSERT(VALID_PAGE(root));
798 799
		page = page_header(root);
		--page->root_count;
800 801 802 803 804 805 806 807 808
		vcpu->mmu.root_hpa = INVALID_PAGE;
		return;
	}
#endif
	for (i = 0; i < 4; ++i) {
		hpa_t root = vcpu->mmu.pae_root[i];

		ASSERT(VALID_PAGE(root));
		root &= PT64_BASE_ADDR_MASK;
809 810
		page = page_header(root);
		--page->root_count;
811 812 813 814 815 816 817 818
		vcpu->mmu.pae_root[i] = INVALID_PAGE;
	}
	vcpu->mmu.root_hpa = INVALID_PAGE;
}

static void mmu_alloc_roots(struct kvm_vcpu *vcpu)
{
	int i;
819
	gfn_t root_gfn;
820 821
	struct kvm_mmu_page *page;

822
	root_gfn = vcpu->cr3 >> PAGE_SHIFT;
823 824 825 826 827 828

#ifdef CONFIG_X86_64
	if (vcpu->mmu.shadow_root_level == PT64_ROOT_LEVEL) {
		hpa_t root = vcpu->mmu.root_hpa;

		ASSERT(!VALID_PAGE(root));
829 830 831
		page = kvm_mmu_get_page(vcpu, root_gfn, 0,
					PT64_ROOT_LEVEL, 0, NULL);
		root = page->page_hpa;
832
		++page->root_count;
833 834 835 836 837 838 839 840
		vcpu->mmu.root_hpa = root;
		return;
	}
#endif
	for (i = 0; i < 4; ++i) {
		hpa_t root = vcpu->mmu.pae_root[i];

		ASSERT(!VALID_PAGE(root));
841 842 843 844
		if (vcpu->mmu.root_level == PT32E_ROOT_LEVEL)
			root_gfn = vcpu->pdptrs[i] >> PAGE_SHIFT;
		else if (vcpu->mmu.root_level == 0)
			root_gfn = 0;
845
		page = kvm_mmu_get_page(vcpu, root_gfn, i << 30,
846
					PT32_ROOT_LEVEL, !is_paging(vcpu),
847 848
					NULL);
		root = page->page_hpa;
849
		++page->root_count;
850 851 852 853 854
		vcpu->mmu.pae_root[i] = root | PT_PRESENT_MASK;
	}
	vcpu->mmu.root_hpa = __pa(vcpu->mmu.pae_root);
}

A
Avi Kivity 已提交
855 856 857 858 859 860 861 862 863
static gpa_t nonpaging_gva_to_gpa(struct kvm_vcpu *vcpu, gva_t vaddr)
{
	return vaddr;
}

static int nonpaging_page_fault(struct kvm_vcpu *vcpu, gva_t gva,
			       u32 error_code)
{
	gpa_t addr = gva;
A
Avi Kivity 已提交
864
	hpa_t paddr;
865
	int r;
A
Avi Kivity 已提交
866

867 868 869
	r = mmu_topup_memory_caches(vcpu);
	if (r)
		return r;
870

A
Avi Kivity 已提交
871 872 873 874
	ASSERT(vcpu);
	ASSERT(VALID_PAGE(vcpu->mmu.root_hpa));


A
Avi Kivity 已提交
875
	paddr = gpa_to_hpa(vcpu , addr & PT64_BASE_ADDR_MASK);
A
Avi Kivity 已提交
876

A
Avi Kivity 已提交
877 878
	if (is_error_hpa(paddr))
		return 1;
A
Avi Kivity 已提交
879

A
Avi Kivity 已提交
880
	return nonpaging_map(vcpu, addr & PAGE_MASK, paddr);
A
Avi Kivity 已提交
881 882 883 884
}

static void nonpaging_free(struct kvm_vcpu *vcpu)
{
885
	mmu_free_roots(vcpu);
A
Avi Kivity 已提交
886 887 888 889 890 891 892 893 894 895
}

static int nonpaging_init_context(struct kvm_vcpu *vcpu)
{
	struct kvm_mmu *context = &vcpu->mmu;

	context->new_cr3 = nonpaging_new_cr3;
	context->page_fault = nonpaging_page_fault;
	context->gva_to_gpa = nonpaging_gva_to_gpa;
	context->free = nonpaging_free;
896
	context->root_level = 0;
A
Avi Kivity 已提交
897
	context->shadow_root_level = PT32E_ROOT_LEVEL;
898
	mmu_alloc_roots(vcpu);
A
Avi Kivity 已提交
899 900 901 902 903 904 905 906 907 908 909 910 911
	ASSERT(VALID_PAGE(context->root_hpa));
	kvm_arch_ops->set_cr3(vcpu, context->root_hpa);
	return 0;
}

static void kvm_mmu_flush_tlb(struct kvm_vcpu *vcpu)
{
	++kvm_stat.tlb_flush;
	kvm_arch_ops->tlb_flush(vcpu);
}

static void paging_new_cr3(struct kvm_vcpu *vcpu)
{
912
	pgprintk("%s: cr3 %lx\n", __FUNCTION__, vcpu->cr3);
913
	mmu_free_roots(vcpu);
914 915
	if (unlikely(vcpu->kvm->n_free_mmu_pages < KVM_MIN_FREE_MMU_PAGES))
		kvm_mmu_free_some_pages(vcpu);
916
	mmu_alloc_roots(vcpu);
A
Avi Kivity 已提交
917
	kvm_mmu_flush_tlb(vcpu);
918
	kvm_arch_ops->set_cr3(vcpu, vcpu->mmu.root_hpa);
A
Avi Kivity 已提交
919 920 921 922 923 924 925 926 927 928 929
}

static void mark_pagetable_nonglobal(void *shadow_pte)
{
	page_header(__pa(shadow_pte))->global = 0;
}

static inline void set_pte_common(struct kvm_vcpu *vcpu,
			     u64 *shadow_pte,
			     gpa_t gaddr,
			     int dirty,
930 931
			     u64 access_bits,
			     gfn_t gfn)
A
Avi Kivity 已提交
932 933 934 935 936 937
{
	hpa_t paddr;

	*shadow_pte |= access_bits << PT_SHADOW_BITS_OFFSET;
	if (!dirty)
		access_bits &= ~PT_WRITABLE_MASK;
938

939
	paddr = gpa_to_hpa(vcpu, gaddr & PT64_BASE_ADDR_MASK);
A
Avi Kivity 已提交
940 941 942 943 944 945 946 947 948 949

	*shadow_pte |= access_bits;

	if (!(*shadow_pte & PT_GLOBAL_MASK))
		mark_pagetable_nonglobal(shadow_pte);

	if (is_error_hpa(paddr)) {
		*shadow_pte |= gaddr;
		*shadow_pte |= PT_SHADOW_IO_MARK;
		*shadow_pte &= ~PT_PRESENT_MASK;
950
		return;
A
Avi Kivity 已提交
951
	}
952 953 954 955 956 957

	*shadow_pte |= paddr;

	if (access_bits & PT_WRITABLE_MASK) {
		struct kvm_mmu_page *shadow;

958
		shadow = kvm_mmu_lookup_page(vcpu, gfn);
959 960
		if (shadow) {
			pgprintk("%s: found shadow page for %lx, marking ro\n",
961
				 __FUNCTION__, gfn);
962
			access_bits &= ~PT_WRITABLE_MASK;
963 964 965 966
			if (is_writeble_pte(*shadow_pte)) {
				    *shadow_pte &= ~PT_WRITABLE_MASK;
				    kvm_arch_ops->tlb_flush(vcpu);
			}
967 968 969 970 971 972 973
		}
	}

	if (access_bits & PT_WRITABLE_MASK)
		mark_page_dirty(vcpu->kvm, gaddr >> PAGE_SHIFT);

	page_header_update_slot(vcpu->kvm, shadow_pte, gaddr);
974
	rmap_add(vcpu, shadow_pte);
A
Avi Kivity 已提交
975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
}

static void inject_page_fault(struct kvm_vcpu *vcpu,
			      u64 addr,
			      u32 err_code)
{
	kvm_arch_ops->inject_page_fault(vcpu, addr, err_code);
}

static inline int fix_read_pf(u64 *shadow_ent)
{
	if ((*shadow_ent & PT_SHADOW_USER_MASK) &&
	    !(*shadow_ent & PT_USER_MASK)) {
		/*
		 * If supervisor write protect is disabled, we shadow kernel
		 * pages as user pages so we can trap the write access.
		 */
		*shadow_ent |= PT_USER_MASK;
		*shadow_ent &= ~PT_WRITABLE_MASK;

		return 1;

	}
	return 0;
}

static void paging_free(struct kvm_vcpu *vcpu)
{
	nonpaging_free(vcpu);
}

#define PTTYPE 64
#include "paging_tmpl.h"
#undef PTTYPE

#define PTTYPE 32
#include "paging_tmpl.h"
#undef PTTYPE

1014
static int paging64_init_context_common(struct kvm_vcpu *vcpu, int level)
A
Avi Kivity 已提交
1015 1016 1017 1018 1019 1020 1021 1022
{
	struct kvm_mmu *context = &vcpu->mmu;

	ASSERT(is_pae(vcpu));
	context->new_cr3 = paging_new_cr3;
	context->page_fault = paging64_page_fault;
	context->gva_to_gpa = paging64_gva_to_gpa;
	context->free = paging_free;
1023 1024 1025
	context->root_level = level;
	context->shadow_root_level = level;
	mmu_alloc_roots(vcpu);
A
Avi Kivity 已提交
1026 1027 1028 1029 1030 1031
	ASSERT(VALID_PAGE(context->root_hpa));
	kvm_arch_ops->set_cr3(vcpu, context->root_hpa |
		    (vcpu->cr3 & (CR3_PCD_MASK | CR3_WPT_MASK)));
	return 0;
}

1032 1033 1034 1035 1036
static int paging64_init_context(struct kvm_vcpu *vcpu)
{
	return paging64_init_context_common(vcpu, PT64_ROOT_LEVEL);
}

A
Avi Kivity 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
static int paging32_init_context(struct kvm_vcpu *vcpu)
{
	struct kvm_mmu *context = &vcpu->mmu;

	context->new_cr3 = paging_new_cr3;
	context->page_fault = paging32_page_fault;
	context->gva_to_gpa = paging32_gva_to_gpa;
	context->free = paging_free;
	context->root_level = PT32_ROOT_LEVEL;
	context->shadow_root_level = PT32E_ROOT_LEVEL;
1047
	mmu_alloc_roots(vcpu);
A
Avi Kivity 已提交
1048 1049 1050 1051 1052 1053 1054 1055
	ASSERT(VALID_PAGE(context->root_hpa));
	kvm_arch_ops->set_cr3(vcpu, context->root_hpa |
		    (vcpu->cr3 & (CR3_PCD_MASK | CR3_WPT_MASK)));
	return 0;
}

static int paging32E_init_context(struct kvm_vcpu *vcpu)
{
1056
	return paging64_init_context_common(vcpu, PT32E_ROOT_LEVEL);
A
Avi Kivity 已提交
1057 1058 1059 1060 1061 1062 1063 1064 1065
}

static int init_kvm_mmu(struct kvm_vcpu *vcpu)
{
	ASSERT(vcpu);
	ASSERT(!VALID_PAGE(vcpu->mmu.root_hpa));

	if (!is_paging(vcpu))
		return nonpaging_init_context(vcpu);
A
Avi Kivity 已提交
1066
	else if (is_long_mode(vcpu))
A
Avi Kivity 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084
		return paging64_init_context(vcpu);
	else if (is_pae(vcpu))
		return paging32E_init_context(vcpu);
	else
		return paging32_init_context(vcpu);
}

static void destroy_kvm_mmu(struct kvm_vcpu *vcpu)
{
	ASSERT(vcpu);
	if (VALID_PAGE(vcpu->mmu.root_hpa)) {
		vcpu->mmu.free(vcpu);
		vcpu->mmu.root_hpa = INVALID_PAGE;
	}
}

int kvm_mmu_reset_context(struct kvm_vcpu *vcpu)
{
1085 1086
	int r;

A
Avi Kivity 已提交
1087
	destroy_kvm_mmu(vcpu);
1088 1089 1090
	r = init_kvm_mmu(vcpu);
	if (r < 0)
		goto out;
1091
	r = mmu_topup_memory_caches(vcpu);
1092 1093
out:
	return r;
A
Avi Kivity 已提交
1094 1095
}

1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
static void mmu_pre_write_zap_pte(struct kvm_vcpu *vcpu,
				  struct kvm_mmu_page *page,
				  u64 *spte)
{
	u64 pte;
	struct kvm_mmu_page *child;

	pte = *spte;
	if (is_present_pte(pte)) {
		if (page->role.level == PT_PAGE_TABLE_LEVEL)
			rmap_remove(vcpu, spte);
		else {
			child = page_header(pte & PT64_BASE_ADDR_MASK);
			mmu_page_remove_parent_pte(vcpu, child, spte);
		}
	}
	*spte = 0;
}

1115 1116
void kvm_mmu_pre_write(struct kvm_vcpu *vcpu, gpa_t gpa, int bytes)
{
1117 1118
	gfn_t gfn = gpa >> PAGE_SHIFT;
	struct kvm_mmu_page *page;
1119
	struct hlist_node *node, *n;
1120 1121 1122 1123
	struct hlist_head *bucket;
	unsigned index;
	u64 *spte;
	unsigned offset = offset_in_page(gpa);
1124
	unsigned pte_size;
1125
	unsigned page_offset;
1126
	unsigned misaligned;
1127
	int level;
1128
	int flooded = 0;
1129
	int npte;
1130

1131
	pgprintk("%s: gpa %llx bytes %d\n", __FUNCTION__, gpa, bytes);
1132 1133 1134 1135 1136 1137 1138 1139
	if (gfn == vcpu->last_pt_write_gfn) {
		++vcpu->last_pt_write_count;
		if (vcpu->last_pt_write_count >= 3)
			flooded = 1;
	} else {
		vcpu->last_pt_write_gfn = gfn;
		vcpu->last_pt_write_count = 1;
	}
1140 1141
	index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
	bucket = &vcpu->kvm->mmu_page_hash[index];
1142
	hlist_for_each_entry_safe(page, node, n, bucket, hash_link) {
1143 1144
		if (page->gfn != gfn || page->role.metaphysical)
			continue;
1145 1146
		pte_size = page->role.glevels == PT32_ROOT_LEVEL ? 4 : 8;
		misaligned = (offset ^ (offset + bytes - 1)) & ~(pte_size - 1);
1147
		if (misaligned || flooded) {
1148 1149 1150 1151
			/*
			 * Misaligned accesses are too much trouble to fix
			 * up; also, they usually indicate a page is not used
			 * as a page table.
1152 1153 1154 1155 1156
			 *
			 * If we're seeing too many writes to a page,
			 * it may no longer be a page table, or we may be
			 * forking, in which case it is better to unmap the
			 * page.
1157 1158 1159 1160 1161 1162
			 */
			pgprintk("misaligned: gpa %llx bytes %d role %x\n",
				 gpa, bytes, page->role.word);
			kvm_mmu_zap_page(vcpu, page);
			continue;
		}
1163 1164
		page_offset = offset;
		level = page->role.level;
1165
		npte = 1;
1166
		if (page->role.glevels == PT32_ROOT_LEVEL) {
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
			page_offset <<= 1;	/* 32->64 */
			/*
			 * A 32-bit pde maps 4MB while the shadow pdes map
			 * only 2MB.  So we need to double the offset again
			 * and zap two pdes instead of one.
			 */
			if (level == PT32_ROOT_LEVEL) {
				page_offset <<= 1;
				npte = 2;
			}
1177 1178 1179 1180
			page_offset &= ~PAGE_MASK;
		}
		spte = __va(page->page_hpa);
		spte += page_offset / sizeof(*spte);
1181 1182 1183
		while (npte--) {
			mmu_pre_write_zap_pte(vcpu, page, spte);
			++spte;
1184 1185
		}
	}
1186 1187 1188 1189 1190 1191
}

void kvm_mmu_post_write(struct kvm_vcpu *vcpu, gpa_t gpa, int bytes)
{
}

1192 1193 1194 1195 1196 1197 1198
int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva)
{
	gpa_t gpa = vcpu->mmu.gva_to_gpa(vcpu, gva);

	return kvm_mmu_unprotect_page(vcpu, gpa >> PAGE_SHIFT);
}

A
Avi Kivity 已提交
1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
void kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu)
{
	while (vcpu->kvm->n_free_mmu_pages < KVM_REFILL_PAGES) {
		struct kvm_mmu_page *page;

		page = container_of(vcpu->kvm->active_mmu_pages.prev,
				    struct kvm_mmu_page, link);
		kvm_mmu_zap_page(vcpu, page);
	}
}
EXPORT_SYMBOL_GPL(kvm_mmu_free_some_pages);

A
Avi Kivity 已提交
1211 1212
static void free_mmu_pages(struct kvm_vcpu *vcpu)
{
1213
	struct kvm_mmu_page *page;
A
Avi Kivity 已提交
1214

1215 1216 1217 1218 1219 1220
	while (!list_empty(&vcpu->kvm->active_mmu_pages)) {
		page = container_of(vcpu->kvm->active_mmu_pages.next,
				    struct kvm_mmu_page, link);
		kvm_mmu_zap_page(vcpu, page);
	}
	while (!list_empty(&vcpu->free_pages)) {
A
Avi Kivity 已提交
1221 1222 1223 1224 1225 1226
		page = list_entry(vcpu->free_pages.next,
				  struct kvm_mmu_page, link);
		list_del(&page->link);
		__free_page(pfn_to_page(page->page_hpa >> PAGE_SHIFT));
		page->page_hpa = INVALID_PAGE;
	}
1227
	free_page((unsigned long)vcpu->mmu.pae_root);
A
Avi Kivity 已提交
1228 1229 1230 1231
}

static int alloc_mmu_pages(struct kvm_vcpu *vcpu)
{
1232
	struct page *page;
A
Avi Kivity 已提交
1233 1234 1235 1236 1237 1238 1239 1240
	int i;

	ASSERT(vcpu);

	for (i = 0; i < KVM_NUM_MMU_PAGES; i++) {
		struct kvm_mmu_page *page_header = &vcpu->page_header_buf[i];

		INIT_LIST_HEAD(&page_header->link);
1241
		if ((page = alloc_page(GFP_KERNEL)) == NULL)
A
Avi Kivity 已提交
1242
			goto error_1;
1243
		set_page_private(page, (unsigned long)page_header);
A
Avi Kivity 已提交
1244 1245 1246
		page_header->page_hpa = (hpa_t)page_to_pfn(page) << PAGE_SHIFT;
		memset(__va(page_header->page_hpa), 0, PAGE_SIZE);
		list_add(&page_header->link, &vcpu->free_pages);
A
Avi Kivity 已提交
1247
		++vcpu->kvm->n_free_mmu_pages;
A
Avi Kivity 已提交
1248
	}
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261

	/*
	 * When emulating 32-bit mode, cr3 is only 32 bits even on x86_64.
	 * Therefore we need to allocate shadow page tables in the first
	 * 4GB of memory, which happens to fit the DMA32 zone.
	 */
	page = alloc_page(GFP_KERNEL | __GFP_DMA32);
	if (!page)
		goto error_1;
	vcpu->mmu.pae_root = page_address(page);
	for (i = 0; i < 4; ++i)
		vcpu->mmu.pae_root[i] = INVALID_PAGE;

A
Avi Kivity 已提交
1262 1263 1264 1265 1266 1267 1268
	return 0;

error_1:
	free_mmu_pages(vcpu);
	return -ENOMEM;
}

1269
int kvm_mmu_create(struct kvm_vcpu *vcpu)
A
Avi Kivity 已提交
1270 1271 1272 1273 1274
{
	ASSERT(vcpu);
	ASSERT(!VALID_PAGE(vcpu->mmu.root_hpa));
	ASSERT(list_empty(&vcpu->free_pages));

1275 1276
	return alloc_mmu_pages(vcpu);
}
A
Avi Kivity 已提交
1277

1278 1279 1280 1281 1282
int kvm_mmu_setup(struct kvm_vcpu *vcpu)
{
	ASSERT(vcpu);
	ASSERT(!VALID_PAGE(vcpu->mmu.root_hpa));
	ASSERT(!list_empty(&vcpu->free_pages));
1283

1284
	return init_kvm_mmu(vcpu);
A
Avi Kivity 已提交
1285 1286 1287 1288 1289 1290 1291 1292
}

void kvm_mmu_destroy(struct kvm_vcpu *vcpu)
{
	ASSERT(vcpu);

	destroy_kvm_mmu(vcpu);
	free_mmu_pages(vcpu);
1293
	mmu_free_memory_caches(vcpu);
A
Avi Kivity 已提交
1294 1295
}

1296
void kvm_mmu_slot_remove_write_access(struct kvm_vcpu *vcpu, int slot)
A
Avi Kivity 已提交
1297
{
1298
	struct kvm *kvm = vcpu->kvm;
A
Avi Kivity 已提交
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
	struct kvm_mmu_page *page;

	list_for_each_entry(page, &kvm->active_mmu_pages, link) {
		int i;
		u64 *pt;

		if (!test_bit(slot, &page->slot_bitmap))
			continue;

		pt = __va(page->page_hpa);
		for (i = 0; i < PT64_ENT_PER_PAGE; ++i)
			/* avoid RMW */
1311
			if (pt[i] & PT_WRITABLE_MASK) {
1312
				rmap_remove(vcpu, &pt[i]);
A
Avi Kivity 已提交
1313
				pt[i] &= ~PT_WRITABLE_MASK;
1314
			}
A
Avi Kivity 已提交
1315 1316
	}
}
1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 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 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476

#ifdef AUDIT

static const char *audit_msg;

static gva_t canonicalize(gva_t gva)
{
#ifdef CONFIG_X86_64
	gva = (long long)(gva << 16) >> 16;
#endif
	return gva;
}

static void audit_mappings_page(struct kvm_vcpu *vcpu, u64 page_pte,
				gva_t va, int level)
{
	u64 *pt = __va(page_pte & PT64_BASE_ADDR_MASK);
	int i;
	gva_t va_delta = 1ul << (PAGE_SHIFT + 9 * (level - 1));

	for (i = 0; i < PT64_ENT_PER_PAGE; ++i, va += va_delta) {
		u64 ent = pt[i];

		if (!ent & PT_PRESENT_MASK)
			continue;

		va = canonicalize(va);
		if (level > 1)
			audit_mappings_page(vcpu, ent, va, level - 1);
		else {
			gpa_t gpa = vcpu->mmu.gva_to_gpa(vcpu, va);
			hpa_t hpa = gpa_to_hpa(vcpu, gpa);

			if ((ent & PT_PRESENT_MASK)
			    && (ent & PT64_BASE_ADDR_MASK) != hpa)
				printk(KERN_ERR "audit error: (%s) levels %d"
				       " gva %lx gpa %llx hpa %llx ent %llx\n",
				       audit_msg, vcpu->mmu.root_level,
				       va, gpa, hpa, ent);
		}
	}
}

static void audit_mappings(struct kvm_vcpu *vcpu)
{
	int i;

	if (vcpu->mmu.root_level == 4)
		audit_mappings_page(vcpu, vcpu->mmu.root_hpa, 0, 4);
	else
		for (i = 0; i < 4; ++i)
			if (vcpu->mmu.pae_root[i] & PT_PRESENT_MASK)
				audit_mappings_page(vcpu,
						    vcpu->mmu.pae_root[i],
						    i << 30,
						    2);
}

static int count_rmaps(struct kvm_vcpu *vcpu)
{
	int nmaps = 0;
	int i, j, k;

	for (i = 0; i < KVM_MEMORY_SLOTS; ++i) {
		struct kvm_memory_slot *m = &vcpu->kvm->memslots[i];
		struct kvm_rmap_desc *d;

		for (j = 0; j < m->npages; ++j) {
			struct page *page = m->phys_mem[j];

			if (!page->private)
				continue;
			if (!(page->private & 1)) {
				++nmaps;
				continue;
			}
			d = (struct kvm_rmap_desc *)(page->private & ~1ul);
			while (d) {
				for (k = 0; k < RMAP_EXT; ++k)
					if (d->shadow_ptes[k])
						++nmaps;
					else
						break;
				d = d->more;
			}
		}
	}
	return nmaps;
}

static int count_writable_mappings(struct kvm_vcpu *vcpu)
{
	int nmaps = 0;
	struct kvm_mmu_page *page;
	int i;

	list_for_each_entry(page, &vcpu->kvm->active_mmu_pages, link) {
		u64 *pt = __va(page->page_hpa);

		if (page->role.level != PT_PAGE_TABLE_LEVEL)
			continue;

		for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
			u64 ent = pt[i];

			if (!(ent & PT_PRESENT_MASK))
				continue;
			if (!(ent & PT_WRITABLE_MASK))
				continue;
			++nmaps;
		}
	}
	return nmaps;
}

static void audit_rmap(struct kvm_vcpu *vcpu)
{
	int n_rmap = count_rmaps(vcpu);
	int n_actual = count_writable_mappings(vcpu);

	if (n_rmap != n_actual)
		printk(KERN_ERR "%s: (%s) rmap %d actual %d\n",
		       __FUNCTION__, audit_msg, n_rmap, n_actual);
}

static void audit_write_protection(struct kvm_vcpu *vcpu)
{
	struct kvm_mmu_page *page;

	list_for_each_entry(page, &vcpu->kvm->active_mmu_pages, link) {
		hfn_t hfn;
		struct page *pg;

		if (page->role.metaphysical)
			continue;

		hfn = gpa_to_hpa(vcpu, (gpa_t)page->gfn << PAGE_SHIFT)
			>> PAGE_SHIFT;
		pg = pfn_to_page(hfn);
		if (pg->private)
			printk(KERN_ERR "%s: (%s) shadow page has writable"
			       " mappings: gfn %lx role %x\n",
			       __FUNCTION__, audit_msg, page->gfn,
			       page->role.word);
	}
}

static void kvm_mmu_audit(struct kvm_vcpu *vcpu, const char *msg)
{
	int olddbg = dbg;

	dbg = 0;
	audit_msg = msg;
	audit_rmap(vcpu);
	audit_write_protection(vcpu);
	audit_mappings(vcpu);
	dbg = olddbg;
}

#endif