gmap.c 66.0 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
/*
 *  KVM guest address space mapping code
 *
 *    Copyright IBM Corp. 2007, 2016
 *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
 */

#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/smp.h>
#include <linux/spinlock.h>
#include <linux/slab.h>
#include <linux/swapops.h>
#include <linux/ksm.h>
#include <linux/mman.h>

#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/gmap.h>
#include <asm/tlb.h>

24 25
#define GMAP_SHADOW_FAKE_TABLE 1ULL

26
/**
27
 * gmap_alloc - allocate and initialize a guest address space
28
 * @mm: pointer to the parent mm_struct
29
 * @limit: maximum address of the gmap address space
30 31 32
 *
 * Returns a guest address space structure.
 */
33
static struct gmap *gmap_alloc(unsigned long limit)
34 35 36 37 38 39
{
	struct gmap *gmap;
	struct page *page;
	unsigned long *table;
	unsigned long etype, atype;

40 41
	if (limit < _REGION3_SIZE) {
		limit = _REGION3_SIZE - 1;
42 43
		atype = _ASCE_TYPE_SEGMENT;
		etype = _SEGMENT_ENTRY_EMPTY;
44 45
	} else if (limit < _REGION2_SIZE) {
		limit = _REGION2_SIZE - 1;
46 47
		atype = _ASCE_TYPE_REGION3;
		etype = _REGION3_ENTRY_EMPTY;
48 49
	} else if (limit < _REGION1_SIZE) {
		limit = _REGION1_SIZE - 1;
50 51 52 53 54 55 56 57 58 59 60
		atype = _ASCE_TYPE_REGION2;
		etype = _REGION2_ENTRY_EMPTY;
	} else {
		limit = -1UL;
		atype = _ASCE_TYPE_REGION1;
		etype = _REGION1_ENTRY_EMPTY;
	}
	gmap = kzalloc(sizeof(struct gmap), GFP_KERNEL);
	if (!gmap)
		goto out;
	INIT_LIST_HEAD(&gmap->crst_list);
61 62
	INIT_LIST_HEAD(&gmap->children);
	INIT_LIST_HEAD(&gmap->pt_list);
63 64
	INIT_RADIX_TREE(&gmap->guest_to_host, GFP_KERNEL);
	INIT_RADIX_TREE(&gmap->host_to_guest, GFP_ATOMIC);
65
	INIT_RADIX_TREE(&gmap->host_to_rmap, GFP_ATOMIC);
66
	spin_lock_init(&gmap->guest_table_lock);
67
	spin_lock_init(&gmap->shadow_lock);
68
	atomic_set(&gmap->ref_count, 1);
69
	page = alloc_pages(GFP_KERNEL, CRST_ALLOC_ORDER);
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
	if (!page)
		goto out_free;
	page->index = 0;
	list_add(&page->lru, &gmap->crst_list);
	table = (unsigned long *) page_to_phys(page);
	crst_table_init(table, etype);
	gmap->table = table;
	gmap->asce = atype | _ASCE_TABLE_LENGTH |
		_ASCE_USER_BITS | __pa(table);
	gmap->asce_end = limit;
	return gmap;

out_free:
	kfree(gmap);
out:
	return NULL;
}
87 88 89 90 91 92 93 94 95 96 97

/**
 * gmap_create - create a guest address space
 * @mm: pointer to the parent mm_struct
 * @limit: maximum size of the gmap address space
 *
 * Returns a guest address space structure.
 */
struct gmap *gmap_create(struct mm_struct *mm, unsigned long limit)
{
	struct gmap *gmap;
98
	unsigned long gmap_asce;
99 100 101 102 103

	gmap = gmap_alloc(limit);
	if (!gmap)
		return NULL;
	gmap->mm = mm;
104
	spin_lock(&mm->context.lock);
105
	list_add_rcu(&gmap->list, &mm->context.gmap_list);
106 107 108 109 110
	if (list_is_singular(&mm->context.gmap_list))
		gmap_asce = gmap->asce;
	else
		gmap_asce = -1UL;
	WRITE_ONCE(mm->context.gmap_asce, gmap_asce);
111
	spin_unlock(&mm->context.lock);
112 113 114
	return gmap;
}
EXPORT_SYMBOL_GPL(gmap_create);
115 116 117 118

static void gmap_flush_tlb(struct gmap *gmap)
{
	if (MACHINE_HAS_IDTE)
119
		__tlb_flush_idte(gmap->asce);
120 121 122 123 124 125 126 127 128
	else
		__tlb_flush_global();
}

static void gmap_radix_tree_free(struct radix_tree_root *root)
{
	struct radix_tree_iter iter;
	unsigned long indices[16];
	unsigned long index;
H
Heiko Carstens 已提交
129
	void __rcu **slot;
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
	int i, nr;

	/* A radix tree is freed by deleting all of its entries */
	index = 0;
	do {
		nr = 0;
		radix_tree_for_each_slot(slot, root, &iter, index) {
			indices[nr] = iter.index;
			if (++nr == 16)
				break;
		}
		for (i = 0; i < nr; i++) {
			index = indices[i];
			radix_tree_delete(root, index);
		}
	} while (nr > 0);
}

148 149 150 151 152 153
static void gmap_rmap_radix_tree_free(struct radix_tree_root *root)
{
	struct gmap_rmap *rmap, *rnext, *head;
	struct radix_tree_iter iter;
	unsigned long indices[16];
	unsigned long index;
H
Heiko Carstens 已提交
154
	void __rcu **slot;
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
	int i, nr;

	/* A radix tree is freed by deleting all of its entries */
	index = 0;
	do {
		nr = 0;
		radix_tree_for_each_slot(slot, root, &iter, index) {
			indices[nr] = iter.index;
			if (++nr == 16)
				break;
		}
		for (i = 0; i < nr; i++) {
			index = indices[i];
			head = radix_tree_delete(root, index);
			gmap_for_each_rmap_safe(rmap, rnext, head)
				kfree(rmap);
		}
	} while (nr > 0);
}

175 176 177
/**
 * gmap_free - free a guest address space
 * @gmap: pointer to the guest address space structure
178 179
 *
 * No locks required. There are no references to this gmap anymore.
180
 */
181
static void gmap_free(struct gmap *gmap)
182 183 184
{
	struct page *page, *next;

185 186 187
	/* Flush tlb of all gmaps (if not already done for shadows) */
	if (!(gmap_is_shadow(gmap) && gmap->removed))
		gmap_flush_tlb(gmap);
188 189
	/* Free all segment & region tables. */
	list_for_each_entry_safe(page, next, &gmap->crst_list, lru)
190
		__free_pages(page, CRST_ALLOC_ORDER);
191 192
	gmap_radix_tree_free(&gmap->guest_to_host);
	gmap_radix_tree_free(&gmap->host_to_guest);
193 194 195 196 197 198 199 200 201 202 203

	/* Free additional data for a shadow gmap */
	if (gmap_is_shadow(gmap)) {
		/* Free all page tables. */
		list_for_each_entry_safe(page, next, &gmap->pt_list, lru)
			page_table_free_pgste(page);
		gmap_rmap_radix_tree_free(&gmap->host_to_rmap);
		/* Release reference to the parent */
		gmap_put(gmap->parent);
	}

204 205
	kfree(gmap);
}
206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238

/**
 * gmap_get - increase reference counter for guest address space
 * @gmap: pointer to the guest address space structure
 *
 * Returns the gmap pointer
 */
struct gmap *gmap_get(struct gmap *gmap)
{
	atomic_inc(&gmap->ref_count);
	return gmap;
}
EXPORT_SYMBOL_GPL(gmap_get);

/**
 * gmap_put - decrease reference counter for guest address space
 * @gmap: pointer to the guest address space structure
 *
 * If the reference counter reaches zero the guest address space is freed.
 */
void gmap_put(struct gmap *gmap)
{
	if (atomic_dec_return(&gmap->ref_count) == 0)
		gmap_free(gmap);
}
EXPORT_SYMBOL_GPL(gmap_put);

/**
 * gmap_remove - remove a guest address space but do not free it yet
 * @gmap: pointer to the guest address space structure
 */
void gmap_remove(struct gmap *gmap)
{
239
	struct gmap *sg, *next;
240
	unsigned long gmap_asce;
241 242 243 244 245 246 247 248 249 250

	/* Remove all shadow gmaps linked to this gmap */
	if (!list_empty(&gmap->children)) {
		spin_lock(&gmap->shadow_lock);
		list_for_each_entry_safe(sg, next, &gmap->children, list) {
			list_del(&sg->list);
			gmap_put(sg);
		}
		spin_unlock(&gmap->shadow_lock);
	}
251
	/* Remove gmap from the pre-mm list */
252
	spin_lock(&gmap->mm->context.lock);
253
	list_del_rcu(&gmap->list);
254 255 256 257 258 259 260 261
	if (list_empty(&gmap->mm->context.gmap_list))
		gmap_asce = 0;
	else if (list_is_singular(&gmap->mm->context.gmap_list))
		gmap_asce = list_first_entry(&gmap->mm->context.gmap_list,
					     struct gmap, list)->asce;
	else
		gmap_asce = -1UL;
	WRITE_ONCE(gmap->mm->context.gmap_asce, gmap_asce);
262
	spin_unlock(&gmap->mm->context.lock);
263 264 265 266 267
	synchronize_rcu();
	/* Put reference */
	gmap_put(gmap);
}
EXPORT_SYMBOL_GPL(gmap_remove);
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288

/**
 * gmap_enable - switch primary space to the guest address space
 * @gmap: pointer to the guest address space structure
 */
void gmap_enable(struct gmap *gmap)
{
	S390_lowcore.gmap = (unsigned long) gmap;
}
EXPORT_SYMBOL_GPL(gmap_enable);

/**
 * gmap_disable - switch back to the standard primary address space
 * @gmap: pointer to the guest address space structure
 */
void gmap_disable(struct gmap *gmap)
{
	S390_lowcore.gmap = 0UL;
}
EXPORT_SYMBOL_GPL(gmap_disable);

289 290 291 292 293 294 295 296 297 298 299
/**
 * gmap_get_enabled - get a pointer to the currently enabled gmap
 *
 * Returns a pointer to the currently enabled gmap. 0 if none is enabled.
 */
struct gmap *gmap_get_enabled(void)
{
	return (struct gmap *) S390_lowcore.gmap;
}
EXPORT_SYMBOL_GPL(gmap_get_enabled);

300 301 302 303 304 305 306 307 308 309
/*
 * gmap_alloc_table is assumed to be called with mmap_sem held
 */
static int gmap_alloc_table(struct gmap *gmap, unsigned long *table,
			    unsigned long init, unsigned long gaddr)
{
	struct page *page;
	unsigned long *new;

	/* since we dont free the gmap table until gmap_free we can unlock */
310
	page = alloc_pages(GFP_KERNEL, CRST_ALLOC_ORDER);
311 312 313 314
	if (!page)
		return -ENOMEM;
	new = (unsigned long *) page_to_phys(page);
	crst_table_init(new, init);
315
	spin_lock(&gmap->guest_table_lock);
316 317 318 319 320 321 322
	if (*table & _REGION_ENTRY_INVALID) {
		list_add(&page->lru, &gmap->crst_list);
		*table = (unsigned long) new | _REGION_ENTRY_LENGTH |
			(*table & _REGION_ENTRY_TYPE_MASK);
		page->index = gaddr;
		page = NULL;
	}
323
	spin_unlock(&gmap->guest_table_lock);
324
	if (page)
325
		__free_pages(page, CRST_ALLOC_ORDER);
326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358
	return 0;
}

/**
 * __gmap_segment_gaddr - find virtual address from segment pointer
 * @entry: pointer to a segment table entry in the guest address space
 *
 * Returns the virtual address in the guest address space for the segment
 */
static unsigned long __gmap_segment_gaddr(unsigned long *entry)
{
	struct page *page;
	unsigned long offset, mask;

	offset = (unsigned long) entry / sizeof(unsigned long);
	offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE;
	mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1);
	page = virt_to_page((void *)((unsigned long) entry & mask));
	return page->index + offset;
}

/**
 * __gmap_unlink_by_vmaddr - unlink a single segment via a host address
 * @gmap: pointer to the guest address space structure
 * @vmaddr: address in the host process address space
 *
 * Returns 1 if a TLB flush is required
 */
static int __gmap_unlink_by_vmaddr(struct gmap *gmap, unsigned long vmaddr)
{
	unsigned long *entry;
	int flush = 0;

359
	BUG_ON(gmap_is_shadow(gmap));
360 361 362
	spin_lock(&gmap->guest_table_lock);
	entry = radix_tree_delete(&gmap->host_to_guest, vmaddr >> PMD_SHIFT);
	if (entry) {
363 364
		flush = (*entry != _SEGMENT_ENTRY_EMPTY);
		*entry = _SEGMENT_ENTRY_EMPTY;
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398
	}
	spin_unlock(&gmap->guest_table_lock);
	return flush;
}

/**
 * __gmap_unmap_by_gaddr - unmap a single segment via a guest address
 * @gmap: pointer to the guest address space structure
 * @gaddr: address in the guest address space
 *
 * Returns 1 if a TLB flush is required
 */
static int __gmap_unmap_by_gaddr(struct gmap *gmap, unsigned long gaddr)
{
	unsigned long vmaddr;

	vmaddr = (unsigned long) radix_tree_delete(&gmap->guest_to_host,
						   gaddr >> PMD_SHIFT);
	return vmaddr ? __gmap_unlink_by_vmaddr(gmap, vmaddr) : 0;
}

/**
 * gmap_unmap_segment - unmap segment from the guest address space
 * @gmap: pointer to the guest address space structure
 * @to: address in the guest address space
 * @len: length of the memory area to unmap
 *
 * Returns 0 if the unmap succeeded, -EINVAL if not.
 */
int gmap_unmap_segment(struct gmap *gmap, unsigned long to, unsigned long len)
{
	unsigned long off;
	int flush;

399
	BUG_ON(gmap_is_shadow(gmap));
400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430
	if ((to | len) & (PMD_SIZE - 1))
		return -EINVAL;
	if (len == 0 || to + len < to)
		return -EINVAL;

	flush = 0;
	down_write(&gmap->mm->mmap_sem);
	for (off = 0; off < len; off += PMD_SIZE)
		flush |= __gmap_unmap_by_gaddr(gmap, to + off);
	up_write(&gmap->mm->mmap_sem);
	if (flush)
		gmap_flush_tlb(gmap);
	return 0;
}
EXPORT_SYMBOL_GPL(gmap_unmap_segment);

/**
 * gmap_map_segment - map a segment to the guest address space
 * @gmap: pointer to the guest address space structure
 * @from: source address in the parent address space
 * @to: target address in the guest address space
 * @len: length of the memory area to map
 *
 * Returns 0 if the mmap succeeded, -EINVAL or -ENOMEM if not.
 */
int gmap_map_segment(struct gmap *gmap, unsigned long from,
		     unsigned long to, unsigned long len)
{
	unsigned long off;
	int flush;

431
	BUG_ON(gmap_is_shadow(gmap));
432 433 434
	if ((from | to | len) & (PMD_SIZE - 1))
		return -EINVAL;
	if (len == 0 || from + len < from || to + len < to ||
435
	    from + len - 1 > TASK_SIZE_MAX || to + len - 1 > gmap->asce_end)
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468
		return -EINVAL;

	flush = 0;
	down_write(&gmap->mm->mmap_sem);
	for (off = 0; off < len; off += PMD_SIZE) {
		/* Remove old translation */
		flush |= __gmap_unmap_by_gaddr(gmap, to + off);
		/* Store new translation */
		if (radix_tree_insert(&gmap->guest_to_host,
				      (to + off) >> PMD_SHIFT,
				      (void *) from + off))
			break;
	}
	up_write(&gmap->mm->mmap_sem);
	if (flush)
		gmap_flush_tlb(gmap);
	if (off >= len)
		return 0;
	gmap_unmap_segment(gmap, to, len);
	return -ENOMEM;
}
EXPORT_SYMBOL_GPL(gmap_map_segment);

/**
 * __gmap_translate - translate a guest address to a user space address
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: guest address
 *
 * Returns user space address which corresponds to the guest address or
 * -EFAULT if no such mapping exists.
 * This function does not establish potentially missing page table entries.
 * The mmap_sem of the mm that belongs to the address space must be held
 * when this function gets called.
469 470
 *
 * Note: Can also be called for shadow gmaps.
471 472 473 474 475 476 477
 */
unsigned long __gmap_translate(struct gmap *gmap, unsigned long gaddr)
{
	unsigned long vmaddr;

	vmaddr = (unsigned long)
		radix_tree_lookup(&gmap->guest_to_host, gaddr >> PMD_SHIFT);
478
	/* Note: guest_to_host is empty for a shadow gmap */
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514
	return vmaddr ? (vmaddr | (gaddr & ~PMD_MASK)) : -EFAULT;
}
EXPORT_SYMBOL_GPL(__gmap_translate);

/**
 * gmap_translate - translate a guest address to a user space address
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: guest address
 *
 * Returns user space address which corresponds to the guest address or
 * -EFAULT if no such mapping exists.
 * This function does not establish potentially missing page table entries.
 */
unsigned long gmap_translate(struct gmap *gmap, unsigned long gaddr)
{
	unsigned long rc;

	down_read(&gmap->mm->mmap_sem);
	rc = __gmap_translate(gmap, gaddr);
	up_read(&gmap->mm->mmap_sem);
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_translate);

/**
 * gmap_unlink - disconnect a page table from the gmap shadow tables
 * @gmap: pointer to guest mapping meta data structure
 * @table: pointer to the host page table
 * @vmaddr: vm address associated with the host page table
 */
void gmap_unlink(struct mm_struct *mm, unsigned long *table,
		 unsigned long vmaddr)
{
	struct gmap *gmap;
	int flush;

515 516
	rcu_read_lock();
	list_for_each_entry_rcu(gmap, &mm->context.gmap_list, list) {
517 518 519 520
		flush = __gmap_unlink_by_vmaddr(gmap, vmaddr);
		if (flush)
			gmap_flush_tlb(gmap);
	}
521
	rcu_read_unlock();
522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
}

/**
 * gmap_link - set up shadow page tables to connect a host to a guest address
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: guest address
 * @vmaddr: vm address
 *
 * Returns 0 on success, -ENOMEM for out of memory conditions, and -EFAULT
 * if the vm address is already mapped to a different guest segment.
 * The mmap_sem of the mm that belongs to the address space must be held
 * when this function gets called.
 */
int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr)
{
	struct mm_struct *mm;
	unsigned long *table;
	spinlock_t *ptl;
	pgd_t *pgd;
541
	p4d_t *p4d;
542 543 544 545
	pud_t *pud;
	pmd_t *pmd;
	int rc;

546
	BUG_ON(gmap_is_shadow(gmap));
547 548 549
	/* Create higher level tables in the gmap page table */
	table = gmap->table;
	if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION1) {
550
		table += (gaddr & _REGION1_INDEX) >> _REGION1_SHIFT;
551 552
		if ((*table & _REGION_ENTRY_INVALID) &&
		    gmap_alloc_table(gmap, table, _REGION2_ENTRY_EMPTY,
553
				     gaddr & _REGION1_MASK))
554 555 556 557
			return -ENOMEM;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
	}
	if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION2) {
558
		table += (gaddr & _REGION2_INDEX) >> _REGION2_SHIFT;
559 560
		if ((*table & _REGION_ENTRY_INVALID) &&
		    gmap_alloc_table(gmap, table, _REGION3_ENTRY_EMPTY,
561
				     gaddr & _REGION2_MASK))
562 563 564 565
			return -ENOMEM;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
	}
	if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION3) {
566
		table += (gaddr & _REGION3_INDEX) >> _REGION3_SHIFT;
567 568
		if ((*table & _REGION_ENTRY_INVALID) &&
		    gmap_alloc_table(gmap, table, _SEGMENT_ENTRY_EMPTY,
569
				     gaddr & _REGION3_MASK))
570 571 572
			return -ENOMEM;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
	}
573
	table += (gaddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;
574 575 576 577
	/* Walk the parent mm page table */
	mm = gmap->mm;
	pgd = pgd_offset(mm, vmaddr);
	VM_BUG_ON(pgd_none(*pgd));
578 579 580
	p4d = p4d_offset(pgd, vmaddr);
	VM_BUG_ON(p4d_none(*p4d));
	pud = pud_offset(p4d, vmaddr);
581
	VM_BUG_ON(pud_none(*pud));
582 583 584
	/* large puds cannot yet be handled */
	if (pud_large(*pud))
		return -EFAULT;
585 586 587 588 589 590 591 592 593 594 595
	pmd = pmd_offset(pud, vmaddr);
	VM_BUG_ON(pmd_none(*pmd));
	/* large pmds cannot yet be handled */
	if (pmd_large(*pmd))
		return -EFAULT;
	/* Link gmap segment table entry location to page table. */
	rc = radix_tree_preload(GFP_KERNEL);
	if (rc)
		return rc;
	ptl = pmd_lock(mm, pmd);
	spin_lock(&gmap->guest_table_lock);
596
	if (*table == _SEGMENT_ENTRY_EMPTY) {
597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693
		rc = radix_tree_insert(&gmap->host_to_guest,
				       vmaddr >> PMD_SHIFT, table);
		if (!rc)
			*table = pmd_val(*pmd);
	} else
		rc = 0;
	spin_unlock(&gmap->guest_table_lock);
	spin_unlock(ptl);
	radix_tree_preload_end();
	return rc;
}

/**
 * gmap_fault - resolve a fault on a guest address
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: guest address
 * @fault_flags: flags to pass down to handle_mm_fault()
 *
 * Returns 0 on success, -ENOMEM for out of memory conditions, and -EFAULT
 * if the vm address is already mapped to a different guest segment.
 */
int gmap_fault(struct gmap *gmap, unsigned long gaddr,
	       unsigned int fault_flags)
{
	unsigned long vmaddr;
	int rc;
	bool unlocked;

	down_read(&gmap->mm->mmap_sem);

retry:
	unlocked = false;
	vmaddr = __gmap_translate(gmap, gaddr);
	if (IS_ERR_VALUE(vmaddr)) {
		rc = vmaddr;
		goto out_up;
	}
	if (fixup_user_fault(current, gmap->mm, vmaddr, fault_flags,
			     &unlocked)) {
		rc = -EFAULT;
		goto out_up;
	}
	/*
	 * In the case that fixup_user_fault unlocked the mmap_sem during
	 * faultin redo __gmap_translate to not race with a map/unmap_segment.
	 */
	if (unlocked)
		goto retry;

	rc = __gmap_link(gmap, gaddr, vmaddr);
out_up:
	up_read(&gmap->mm->mmap_sem);
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_fault);

/*
 * this function is assumed to be called with mmap_sem held
 */
void __gmap_zap(struct gmap *gmap, unsigned long gaddr)
{
	unsigned long vmaddr;
	spinlock_t *ptl;
	pte_t *ptep;

	/* Find the vm address for the guest address */
	vmaddr = (unsigned long) radix_tree_lookup(&gmap->guest_to_host,
						   gaddr >> PMD_SHIFT);
	if (vmaddr) {
		vmaddr |= gaddr & ~PMD_MASK;
		/* Get pointer to the page table entry */
		ptep = get_locked_pte(gmap->mm, vmaddr, &ptl);
		if (likely(ptep))
			ptep_zap_unused(gmap->mm, vmaddr, ptep, 0);
		pte_unmap_unlock(ptep, ptl);
	}
}
EXPORT_SYMBOL_GPL(__gmap_zap);

void gmap_discard(struct gmap *gmap, unsigned long from, unsigned long to)
{
	unsigned long gaddr, vmaddr, size;
	struct vm_area_struct *vma;

	down_read(&gmap->mm->mmap_sem);
	for (gaddr = from; gaddr < to;
	     gaddr = (gaddr + PMD_SIZE) & PMD_MASK) {
		/* Find the vm address for the guest address */
		vmaddr = (unsigned long)
			radix_tree_lookup(&gmap->guest_to_host,
					  gaddr >> PMD_SHIFT);
		if (!vmaddr)
			continue;
		vmaddr |= gaddr & ~PMD_MASK;
		/* Find vma in the parent mm */
		vma = find_vma(gmap->mm, vmaddr);
		size = min(to - gaddr, PMD_SIZE - (gaddr & ~PMD_MASK));
694
		zap_page_range(vma, vmaddr, size);
695 696 697 698 699 700 701 702 703
	}
	up_read(&gmap->mm->mmap_sem);
}
EXPORT_SYMBOL_GPL(gmap_discard);

static LIST_HEAD(gmap_notifier_list);
static DEFINE_SPINLOCK(gmap_notifier_lock);

/**
704
 * gmap_register_pte_notifier - register a pte invalidation callback
705 706
 * @nb: pointer to the gmap notifier block
 */
707
void gmap_register_pte_notifier(struct gmap_notifier *nb)
708 709
{
	spin_lock(&gmap_notifier_lock);
710
	list_add_rcu(&nb->list, &gmap_notifier_list);
711 712
	spin_unlock(&gmap_notifier_lock);
}
713
EXPORT_SYMBOL_GPL(gmap_register_pte_notifier);
714 715

/**
716
 * gmap_unregister_pte_notifier - remove a pte invalidation callback
717 718
 * @nb: pointer to the gmap notifier block
 */
719
void gmap_unregister_pte_notifier(struct gmap_notifier *nb)
720 721
{
	spin_lock(&gmap_notifier_lock);
722
	list_del_rcu(&nb->list);
723
	spin_unlock(&gmap_notifier_lock);
724
	synchronize_rcu();
725
}
726
EXPORT_SYMBOL_GPL(gmap_unregister_pte_notifier);
727

728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
/**
 * gmap_call_notifier - call all registered invalidation callbacks
 * @gmap: pointer to guest mapping meta data structure
 * @start: start virtual address in the guest address space
 * @end: end virtual address in the guest address space
 */
static void gmap_call_notifier(struct gmap *gmap, unsigned long start,
			       unsigned long end)
{
	struct gmap_notifier *nb;

	list_for_each_entry(nb, &gmap_notifier_list, list)
		nb->notifier_call(gmap, start, end);
}

743
/**
744 745 746
 * gmap_table_walk - walk the gmap page tables
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
747 748 749 750 751 752 753 754 755 756 757
 * @level: page table level to stop at
 *
 * Returns a table entry pointer for the given guest address and @level
 * @level=0 : returns a pointer to a page table table entry (or NULL)
 * @level=1 : returns a pointer to a segment table entry (or NULL)
 * @level=2 : returns a pointer to a region-3 table entry (or NULL)
 * @level=3 : returns a pointer to a region-2 table entry (or NULL)
 * @level=4 : returns a pointer to a region-1 table entry (or NULL)
 *
 * Returns NULL if the gmap page tables could not be walked to the
 * requested level.
758
 *
759
 * Note: Can also be called for shadow gmaps.
760 761
 */
static inline unsigned long *gmap_table_walk(struct gmap *gmap,
762
					     unsigned long gaddr, int level)
763 764 765
{
	unsigned long *table;

766 767 768 769 770 771
	if ((gmap->asce & _ASCE_TYPE_MASK) + 4 < (level * 4))
		return NULL;
	if (gmap_is_shadow(gmap) && gmap->removed)
		return NULL;
	if (gaddr & (-1UL << (31 + ((gmap->asce & _ASCE_TYPE_MASK) >> 2)*11)))
		return NULL;
772 773 774
	table = gmap->table;
	switch (gmap->asce & _ASCE_TYPE_MASK) {
	case _ASCE_TYPE_REGION1:
775
		table += (gaddr & _REGION1_INDEX) >> _REGION1_SHIFT;
776 777
		if (level == 4)
			break;
778 779 780 781 782
		if (*table & _REGION_ENTRY_INVALID)
			return NULL;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
		/* Fallthrough */
	case _ASCE_TYPE_REGION2:
783
		table += (gaddr & _REGION2_INDEX) >> _REGION2_SHIFT;
784 785
		if (level == 3)
			break;
786 787 788 789 790
		if (*table & _REGION_ENTRY_INVALID)
			return NULL;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
		/* Fallthrough */
	case _ASCE_TYPE_REGION3:
791
		table += (gaddr & _REGION3_INDEX) >> _REGION3_SHIFT;
792 793
		if (level == 2)
			break;
794 795 796 797 798
		if (*table & _REGION_ENTRY_INVALID)
			return NULL;
		table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
		/* Fallthrough */
	case _ASCE_TYPE_SEGMENT:
799
		table += (gaddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;
800 801 802 803 804
		if (level == 1)
			break;
		if (*table & _REGION_ENTRY_INVALID)
			return NULL;
		table = (unsigned long *)(*table & _SEGMENT_ENTRY_ORIGIN);
805
		table += (gaddr & _PAGE_INDEX) >> _PAGE_SHIFT;
806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
	}
	return table;
}

/**
 * gmap_pte_op_walk - walk the gmap page table, get the page table lock
 *		      and return the pte pointer
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @ptl: pointer to the spinlock pointer
 *
 * Returns a pointer to the locked pte for a guest address, or NULL
 */
static pte_t *gmap_pte_op_walk(struct gmap *gmap, unsigned long gaddr,
			       spinlock_t **ptl)
{
	unsigned long *table;

824
	BUG_ON(gmap_is_shadow(gmap));
825
	/* Walk the gmap page table, lock and get pte pointer */
826
	table = gmap_table_walk(gmap, gaddr, 1); /* get segment pointer */
827
	if (!table || *table & _SEGMENT_ENTRY_INVALID)
828 829 830 831 832 833 834 835 836
		return NULL;
	return pte_alloc_map_lock(gmap->mm, (pmd_t *) table, gaddr, ptl);
}

/**
 * gmap_pte_op_fixup - force a page in and connect the gmap page table
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @vmaddr: address in the host process address space
837
 * @prot: indicates access rights: PROT_NONE, PROT_READ or PROT_WRITE
838 839 840 841 842 843
 *
 * Returns 0 if the caller can retry __gmap_translate (might fail again),
 * -ENOMEM if out of memory and -EFAULT if anything goes wrong while fixing
 * up or connecting the gmap page table.
 */
static int gmap_pte_op_fixup(struct gmap *gmap, unsigned long gaddr,
844
			     unsigned long vmaddr, int prot)
845 846
{
	struct mm_struct *mm = gmap->mm;
847
	unsigned int fault_flags;
848 849
	bool unlocked = false;

850
	BUG_ON(gmap_is_shadow(gmap));
851 852
	fault_flags = (prot == PROT_WRITE) ? FAULT_FLAG_WRITE : 0;
	if (fixup_user_fault(current, mm, vmaddr, fault_flags, &unlocked))
853 854 855 856 857 858
		return -EFAULT;
	if (unlocked)
		/* lost mmap_sem, caller has to retry __gmap_translate */
		return 0;
	/* Connect the page tables */
	return __gmap_link(gmap, gaddr, vmaddr);
859 860 861
}

/**
862 863 864 865 866
 * gmap_pte_op_end - release the page table lock
 * @ptl: pointer to the spinlock pointer
 */
static void gmap_pte_op_end(spinlock_t *ptl)
{
867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914
	if (ptl)
		spin_unlock(ptl);
}

/**
 * gmap_pmd_op_walk - walk the gmap tables, get the guest table lock
 *		      and return the pmd pointer
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 *
 * Returns a pointer to the pmd for a guest address, or NULL
 */
static inline pmd_t *gmap_pmd_op_walk(struct gmap *gmap, unsigned long gaddr)
{
	pmd_t *pmdp;

	BUG_ON(gmap_is_shadow(gmap));
	spin_lock(&gmap->guest_table_lock);
	pmdp = (pmd_t *) gmap_table_walk(gmap, gaddr, 1);

	if (!pmdp || pmd_none(*pmdp)) {
		spin_unlock(&gmap->guest_table_lock);
		return NULL;
	}

	/* 4k page table entries are locked via the pte (pte_alloc_map_lock). */
	if (!pmd_large(*pmdp))
		spin_unlock(&gmap->guest_table_lock);
	return pmdp;
}

/**
 * gmap_pmd_op_end - release the guest_table_lock if needed
 * @gmap: pointer to the guest mapping meta data structure
 * @pmdp: pointer to the pmd
 */
static inline void gmap_pmd_op_end(struct gmap *gmap, pmd_t *pmdp)
{
	if (pmd_large(*pmdp))
		spin_unlock(&gmap->guest_table_lock);
}

/*
 * gmap_protect_pte - remove access rights to memory and set pgste bits
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @pmdp: pointer to the pmd associated with the pte
 * @prot: indicates access rights: PROT_NONE, PROT_READ or PROT_WRITE
915
 * @bits: notification bits to set
916 917 918 919 920 921 922 923 924 925 926 927
 *
 * Returns 0 if successfully protected, -ENOMEM if out of memory and
 * -EAGAIN if a fixup is needed.
 *
 * Expected to be called with sg->mm->mmap_sem in read
 */
static int gmap_protect_pte(struct gmap *gmap, unsigned long gaddr,
			    pmd_t *pmdp, int prot, unsigned long bits)
{
	int rc;
	pte_t *ptep;
	spinlock_t *ptl = NULL;
928
	unsigned long pbits = 0;
929 930 931 932 933 934 935 936

	if (pmd_val(*pmdp) & _SEGMENT_ENTRY_INVALID)
		return -EAGAIN;

	ptep = pte_alloc_map_lock(gmap->mm, pmdp, gaddr, &ptl);
	if (!ptep)
		return -ENOMEM;

937 938
	pbits |= (bits & GMAP_NOTIFY_MPROT) ? PGSTE_IN_BIT : 0;
	pbits |= (bits & GMAP_NOTIFY_SHADOW) ? PGSTE_VSIE_BIT : 0;
939
	/* Protect and unlock. */
940
	rc = ptep_force_prot(gmap->mm, gaddr, ptep, prot, pbits);
941 942
	gmap_pte_op_end(ptl);
	return rc;
943 944
}

945 946
/*
 * gmap_protect_range - remove access rights to memory and set pgste bits
947 948 949
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @len: size of area
950 951 952 953 954 955 956
 * @prot: indicates access rights: PROT_NONE, PROT_READ or PROT_WRITE
 * @bits: pgste notification bits to set
 *
 * Returns 0 if successfully protected, -ENOMEM if out of memory and
 * -EFAULT if gaddr is invalid (or mapping for shadows is missing).
 *
 * Called with sg->mm->mmap_sem in read.
957
 */
958 959
static int gmap_protect_range(struct gmap *gmap, unsigned long gaddr,
			      unsigned long len, int prot, unsigned long bits)
960
{
961
	unsigned long vmaddr;
962
	pmd_t *pmdp;
963 964
	int rc;

965
	BUG_ON(gmap_is_shadow(gmap));
966 967
	while (len) {
		rc = -EAGAIN;
968 969 970 971 972 973 974 975 976
		pmdp = gmap_pmd_op_walk(gmap, gaddr);
		if (pmdp) {
			rc = gmap_protect_pte(gmap, gaddr, pmdp, prot,
					      bits);
			if (!rc) {
				len -= PAGE_SIZE;
				gaddr += PAGE_SIZE;
			}
			gmap_pmd_op_end(gmap, pmdp);
977 978 979 980 981
		}
		if (rc) {
			vmaddr = __gmap_translate(gmap, gaddr);
			if (IS_ERR_VALUE(vmaddr))
				return vmaddr;
982
			rc = gmap_pte_op_fixup(gmap, gaddr, vmaddr, prot);
983 984 985 986 987 988
			if (rc)
				return rc;
		}
	}
	return 0;
}
989

990 991 992
/**
 * gmap_mprotect_notify - change access rights for a range of ptes and
 *                        call the notifier if any pte changes again
993 994 995
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @len: size of area
996
 * @prot: indicates access rights: PROT_NONE, PROT_READ or PROT_WRITE
997
 *
998 999 1000 1001 1002
 * Returns 0 if for each page in the given range a gmap mapping exists,
 * the new access rights could be set and the notifier could be armed.
 * If the gmap mapping is missing for one or more pages -EFAULT is
 * returned. If no memory could be allocated -ENOMEM is returned.
 * This function establishes missing page table entries.
1003
 */
1004 1005
int gmap_mprotect_notify(struct gmap *gmap, unsigned long gaddr,
			 unsigned long len, int prot)
1006
{
1007
	int rc;
1008

1009
	if ((gaddr & ~PAGE_MASK) || (len & ~PAGE_MASK) || gmap_is_shadow(gmap))
1010
		return -EINVAL;
1011
	if (!MACHINE_HAS_ESOP && prot == PROT_READ)
1012 1013
		return -EINVAL;
	down_read(&gmap->mm->mmap_sem);
1014
	rc = gmap_protect_range(gmap, gaddr, len, prot, GMAP_NOTIFY_MPROT);
1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027
	up_read(&gmap->mm->mmap_sem);
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_mprotect_notify);

/**
 * gmap_read_table - get an unsigned long value from a guest page table using
 *                   absolute addressing, without marking the page referenced.
 * @gmap: pointer to guest mapping meta data structure
 * @gaddr: virtual address in the guest address space
 * @val: pointer to the unsigned long value to return
 *
 * Returns 0 if the value was read, -ENOMEM if out of memory and -EFAULT
1028 1029
 * if reading using the virtual address failed. -EINVAL if called on a gmap
 * shadow.
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
 *
 * Called with gmap->mm->mmap_sem in read.
 */
int gmap_read_table(struct gmap *gmap, unsigned long gaddr, unsigned long *val)
{
	unsigned long address, vmaddr;
	spinlock_t *ptl;
	pte_t *ptep, pte;
	int rc;

1040 1041 1042
	if (gmap_is_shadow(gmap))
		return -EINVAL;

1043
	while (1) {
1044 1045 1046
		rc = -EAGAIN;
		ptep = gmap_pte_op_walk(gmap, gaddr, &ptl);
		if (ptep) {
1047 1048 1049 1050 1051 1052 1053 1054 1055
			pte = *ptep;
			if (pte_present(pte) && (pte_val(pte) & _PAGE_READ)) {
				address = pte_val(pte) & PAGE_MASK;
				address += gaddr & ~PAGE_MASK;
				*val = *(unsigned long *) address;
				pte_val(*ptep) |= _PAGE_YOUNG;
				/* Do *NOT* clear the _PAGE_INVALID bit! */
				rc = 0;
			}
1056
			gmap_pte_op_end(ptl);
1057
		}
1058 1059 1060 1061 1062
		if (!rc)
			break;
		vmaddr = __gmap_translate(gmap, gaddr);
		if (IS_ERR_VALUE(vmaddr)) {
			rc = vmaddr;
1063 1064
			break;
		}
1065
		rc = gmap_pte_op_fixup(gmap, gaddr, vmaddr, PROT_READ);
1066 1067 1068 1069 1070
		if (rc)
			break;
	}
	return rc;
}
1071
EXPORT_SYMBOL_GPL(gmap_read_table);
1072 1073

/**
1074 1075 1076 1077
 * gmap_insert_rmap - add a rmap to the host_to_rmap radix tree
 * @sg: pointer to the shadow guest address space structure
 * @vmaddr: vm address associated with the rmap
 * @rmap: pointer to the rmap structure
1078
 *
1079
 * Called with the sg->guest_table_lock
1080
 */
1081 1082
static inline void gmap_insert_rmap(struct gmap *sg, unsigned long vmaddr,
				    struct gmap_rmap *rmap)
1083
{
H
Heiko Carstens 已提交
1084
	void __rcu **slot;
1085

1086 1087 1088 1089 1090
	BUG_ON(!gmap_is_shadow(sg));
	slot = radix_tree_lookup_slot(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT);
	if (slot) {
		rmap->next = radix_tree_deref_slot_protected(slot,
							&sg->guest_table_lock);
1091
		radix_tree_replace_slot(&sg->host_to_rmap, slot, rmap);
1092 1093 1094 1095 1096 1097 1098 1099
	} else {
		rmap->next = NULL;
		radix_tree_insert(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT,
				  rmap);
	}
}

/**
1100
 * gmap_protect_rmap - restrict access rights to memory (RO) and create an rmap
1101 1102 1103 1104 1105 1106 1107 1108 1109
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow gmap
 * @paddr: address in the parent guest address space
 * @len: length of the memory area to protect
 *
 * Returns 0 if successfully protected and the rmap was created, -ENOMEM
 * if out of memory and -EFAULT if paddr is invalid.
 */
static int gmap_protect_rmap(struct gmap *sg, unsigned long raddr,
1110
			     unsigned long paddr, unsigned long len)
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129
{
	struct gmap *parent;
	struct gmap_rmap *rmap;
	unsigned long vmaddr;
	spinlock_t *ptl;
	pte_t *ptep;
	int rc;

	BUG_ON(!gmap_is_shadow(sg));
	parent = sg->parent;
	while (len) {
		vmaddr = __gmap_translate(parent, paddr);
		if (IS_ERR_VALUE(vmaddr))
			return vmaddr;
		rmap = kzalloc(sizeof(*rmap), GFP_KERNEL);
		if (!rmap)
			return -ENOMEM;
		rmap->raddr = raddr;
		rc = radix_tree_preload(GFP_KERNEL);
1130
		if (rc) {
1131 1132 1133 1134 1135 1136 1137
			kfree(rmap);
			return rc;
		}
		rc = -EAGAIN;
		ptep = gmap_pte_op_walk(parent, paddr, &ptl);
		if (ptep) {
			spin_lock(&sg->guest_table_lock);
1138
			rc = ptep_force_prot(parent->mm, paddr, ptep, PROT_READ,
1139 1140 1141 1142 1143 1144 1145 1146 1147
					     PGSTE_VSIE_BIT);
			if (!rc)
				gmap_insert_rmap(sg, vmaddr, rmap);
			spin_unlock(&sg->guest_table_lock);
			gmap_pte_op_end(ptl);
		}
		radix_tree_preload_end();
		if (rc) {
			kfree(rmap);
1148
			rc = gmap_pte_op_fixup(parent, paddr, vmaddr, PROT_READ);
1149
			if (rc)
1150
				return rc;
1151 1152
			continue;
		}
1153
		paddr += PAGE_SIZE;
1154
		len -= PAGE_SIZE;
1155
	}
1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 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
	return 0;
}

#define _SHADOW_RMAP_MASK	0x7
#define _SHADOW_RMAP_REGION1	0x5
#define _SHADOW_RMAP_REGION2	0x4
#define _SHADOW_RMAP_REGION3	0x3
#define _SHADOW_RMAP_SEGMENT	0x2
#define _SHADOW_RMAP_PGTABLE	0x1

/**
 * gmap_idte_one - invalidate a single region or segment table entry
 * @asce: region or segment table *origin* + table-type bits
 * @vaddr: virtual address to identify the table entry to flush
 *
 * The invalid bit of a single region or segment table entry is set
 * and the associated TLB entries depending on the entry are flushed.
 * The table-type of the @asce identifies the portion of the @vaddr
 * that is used as the invalidation index.
 */
static inline void gmap_idte_one(unsigned long asce, unsigned long vaddr)
{
	asm volatile(
		"	.insn	rrf,0xb98e0000,%0,%1,0,0"
		: : "a" (asce), "a" (vaddr) : "cc", "memory");
}

/**
 * gmap_unshadow_page - remove a page from a shadow page table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 *
 * Called with the sg->guest_table_lock
 */
static void gmap_unshadow_page(struct gmap *sg, unsigned long raddr)
{
	unsigned long *table;

	BUG_ON(!gmap_is_shadow(sg));
	table = gmap_table_walk(sg, raddr, 0); /* get page table pointer */
	if (!table || *table & _PAGE_INVALID)
		return;
1198
	gmap_call_notifier(sg, raddr, raddr + _PAGE_SIZE - 1);
1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
	ptep_unshadow_pte(sg->mm, raddr, (pte_t *) table);
}

/**
 * __gmap_unshadow_pgt - remove all entries from a shadow page table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 * @pgt: pointer to the start of a shadow page table
 *
 * Called with the sg->guest_table_lock
 */
static void __gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr,
				unsigned long *pgt)
{
	int i;

	BUG_ON(!gmap_is_shadow(sg));
1216
	for (i = 0; i < _PAGE_ENTRIES; i++, raddr += _PAGE_SIZE)
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
		pgt[i] = _PAGE_INVALID;
}

/**
 * gmap_unshadow_pgt - remove a shadow page table from a segment entry
 * @sg: pointer to the shadow guest address space structure
 * @raddr: address in the shadow guest address space
 *
 * Called with the sg->guest_table_lock
 */
static void gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr)
{
	unsigned long sto, *ste, *pgt;
	struct page *page;

	BUG_ON(!gmap_is_shadow(sg));
	ste = gmap_table_walk(sg, raddr, 1); /* get segment pointer */
1234
	if (!ste || !(*ste & _SEGMENT_ENTRY_ORIGIN))
1235
		return;
1236 1237
	gmap_call_notifier(sg, raddr, raddr + _SEGMENT_SIZE - 1);
	sto = (unsigned long) (ste - ((raddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT));
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
	gmap_idte_one(sto | _ASCE_TYPE_SEGMENT, raddr);
	pgt = (unsigned long *)(*ste & _SEGMENT_ENTRY_ORIGIN);
	*ste = _SEGMENT_ENTRY_EMPTY;
	__gmap_unshadow_pgt(sg, raddr, pgt);
	/* Free page table */
	page = pfn_to_page(__pa(pgt) >> PAGE_SHIFT);
	list_del(&page->lru);
	page_table_free_pgste(page);
}

/**
 * __gmap_unshadow_sgt - remove all entries from a shadow segment table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 * @sgt: pointer to the start of a shadow segment table
 *
 * Called with the sg->guest_table_lock
 */
static void __gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr,
				unsigned long *sgt)
{
H
Heiko Carstens 已提交
1259
	unsigned long *pgt;
1260 1261 1262 1263
	struct page *page;
	int i;

	BUG_ON(!gmap_is_shadow(sg));
1264
	for (i = 0; i < _CRST_ENTRIES; i++, raddr += _SEGMENT_SIZE) {
1265
		if (!(sgt[i] & _SEGMENT_ENTRY_ORIGIN))
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
			continue;
		pgt = (unsigned long *)(sgt[i] & _REGION_ENTRY_ORIGIN);
		sgt[i] = _SEGMENT_ENTRY_EMPTY;
		__gmap_unshadow_pgt(sg, raddr, pgt);
		/* Free page table */
		page = pfn_to_page(__pa(pgt) >> PAGE_SHIFT);
		list_del(&page->lru);
		page_table_free_pgste(page);
	}
}

/**
 * gmap_unshadow_sgt - remove a shadow segment table from a region-3 entry
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 *
 * Called with the shadow->guest_table_lock
 */
static void gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr)
{
	unsigned long r3o, *r3e, *sgt;
	struct page *page;

	BUG_ON(!gmap_is_shadow(sg));
	r3e = gmap_table_walk(sg, raddr, 2); /* get region-3 pointer */
1291
	if (!r3e || !(*r3e & _REGION_ENTRY_ORIGIN))
1292
		return;
1293 1294
	gmap_call_notifier(sg, raddr, raddr + _REGION3_SIZE - 1);
	r3o = (unsigned long) (r3e - ((raddr & _REGION3_INDEX) >> _REGION3_SHIFT));
1295 1296 1297 1298 1299 1300 1301
	gmap_idte_one(r3o | _ASCE_TYPE_REGION3, raddr);
	sgt = (unsigned long *)(*r3e & _REGION_ENTRY_ORIGIN);
	*r3e = _REGION3_ENTRY_EMPTY;
	__gmap_unshadow_sgt(sg, raddr, sgt);
	/* Free segment table */
	page = pfn_to_page(__pa(sgt) >> PAGE_SHIFT);
	list_del(&page->lru);
1302
	__free_pages(page, CRST_ALLOC_ORDER);
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
}

/**
 * __gmap_unshadow_r3t - remove all entries from a shadow region-3 table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: address in the shadow guest address space
 * @r3t: pointer to the start of a shadow region-3 table
 *
 * Called with the sg->guest_table_lock
 */
static void __gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr,
				unsigned long *r3t)
{
H
Heiko Carstens 已提交
1316
	unsigned long *sgt;
1317 1318 1319 1320
	struct page *page;
	int i;

	BUG_ON(!gmap_is_shadow(sg));
1321
	for (i = 0; i < _CRST_ENTRIES; i++, raddr += _REGION3_SIZE) {
1322
		if (!(r3t[i] & _REGION_ENTRY_ORIGIN))
1323 1324 1325 1326 1327 1328 1329
			continue;
		sgt = (unsigned long *)(r3t[i] & _REGION_ENTRY_ORIGIN);
		r3t[i] = _REGION3_ENTRY_EMPTY;
		__gmap_unshadow_sgt(sg, raddr, sgt);
		/* Free segment table */
		page = pfn_to_page(__pa(sgt) >> PAGE_SHIFT);
		list_del(&page->lru);
1330
		__free_pages(page, CRST_ALLOC_ORDER);
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
	}
}

/**
 * gmap_unshadow_r3t - remove a shadow region-3 table from a region-2 entry
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 *
 * Called with the sg->guest_table_lock
 */
static void gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr)
{
	unsigned long r2o, *r2e, *r3t;
	struct page *page;

	BUG_ON(!gmap_is_shadow(sg));
	r2e = gmap_table_walk(sg, raddr, 3); /* get region-2 pointer */
1348
	if (!r2e || !(*r2e & _REGION_ENTRY_ORIGIN))
1349
		return;
1350 1351
	gmap_call_notifier(sg, raddr, raddr + _REGION2_SIZE - 1);
	r2o = (unsigned long) (r2e - ((raddr & _REGION2_INDEX) >> _REGION2_SHIFT));
1352 1353 1354 1355 1356 1357 1358
	gmap_idte_one(r2o | _ASCE_TYPE_REGION2, raddr);
	r3t = (unsigned long *)(*r2e & _REGION_ENTRY_ORIGIN);
	*r2e = _REGION2_ENTRY_EMPTY;
	__gmap_unshadow_r3t(sg, raddr, r3t);
	/* Free region 3 table */
	page = pfn_to_page(__pa(r3t) >> PAGE_SHIFT);
	list_del(&page->lru);
1359
	__free_pages(page, CRST_ALLOC_ORDER);
1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
}

/**
 * __gmap_unshadow_r2t - remove all entries from a shadow region-2 table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 * @r2t: pointer to the start of a shadow region-2 table
 *
 * Called with the sg->guest_table_lock
 */
static void __gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr,
				unsigned long *r2t)
{
H
Heiko Carstens 已提交
1373
	unsigned long *r3t;
1374 1375 1376 1377
	struct page *page;
	int i;

	BUG_ON(!gmap_is_shadow(sg));
1378
	for (i = 0; i < _CRST_ENTRIES; i++, raddr += _REGION2_SIZE) {
1379
		if (!(r2t[i] & _REGION_ENTRY_ORIGIN))
1380 1381 1382 1383 1384 1385 1386
			continue;
		r3t = (unsigned long *)(r2t[i] & _REGION_ENTRY_ORIGIN);
		r2t[i] = _REGION2_ENTRY_EMPTY;
		__gmap_unshadow_r3t(sg, raddr, r3t);
		/* Free region 3 table */
		page = pfn_to_page(__pa(r3t) >> PAGE_SHIFT);
		list_del(&page->lru);
1387
		__free_pages(page, CRST_ALLOC_ORDER);
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
	}
}

/**
 * gmap_unshadow_r2t - remove a shadow region-2 table from a region-1 entry
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 *
 * Called with the sg->guest_table_lock
 */
static void gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr)
{
	unsigned long r1o, *r1e, *r2t;
	struct page *page;

	BUG_ON(!gmap_is_shadow(sg));
	r1e = gmap_table_walk(sg, raddr, 4); /* get region-1 pointer */
1405
	if (!r1e || !(*r1e & _REGION_ENTRY_ORIGIN))
1406
		return;
1407 1408
	gmap_call_notifier(sg, raddr, raddr + _REGION1_SIZE - 1);
	r1o = (unsigned long) (r1e - ((raddr & _REGION1_INDEX) >> _REGION1_SHIFT));
1409 1410 1411 1412 1413 1414 1415
	gmap_idte_one(r1o | _ASCE_TYPE_REGION1, raddr);
	r2t = (unsigned long *)(*r1e & _REGION_ENTRY_ORIGIN);
	*r1e = _REGION1_ENTRY_EMPTY;
	__gmap_unshadow_r2t(sg, raddr, r2t);
	/* Free region 2 table */
	page = pfn_to_page(__pa(r2t) >> PAGE_SHIFT);
	list_del(&page->lru);
1416
	__free_pages(page, CRST_ALLOC_ORDER);
1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
}

/**
 * __gmap_unshadow_r1t - remove all entries from a shadow region-1 table
 * @sg: pointer to the shadow guest address space structure
 * @raddr: rmap address in the shadow guest address space
 * @r1t: pointer to the start of a shadow region-1 table
 *
 * Called with the shadow->guest_table_lock
 */
static void __gmap_unshadow_r1t(struct gmap *sg, unsigned long raddr,
				unsigned long *r1t)
{
	unsigned long asce, *r2t;
	struct page *page;
	int i;

	BUG_ON(!gmap_is_shadow(sg));
	asce = (unsigned long) r1t | _ASCE_TYPE_REGION1;
1436
	for (i = 0; i < _CRST_ENTRIES; i++, raddr += _REGION1_SIZE) {
1437
		if (!(r1t[i] & _REGION_ENTRY_ORIGIN))
1438 1439 1440 1441 1442 1443 1444 1445 1446
			continue;
		r2t = (unsigned long *)(r1t[i] & _REGION_ENTRY_ORIGIN);
		__gmap_unshadow_r2t(sg, raddr, r2t);
		/* Clear entry and flush translation r1t -> r2t */
		gmap_idte_one(asce, raddr);
		r1t[i] = _REGION1_ENTRY_EMPTY;
		/* Free region 2 table */
		page = pfn_to_page(__pa(r2t) >> PAGE_SHIFT);
		list_del(&page->lru);
1447
		__free_pages(page, CRST_ALLOC_ORDER);
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465
	}
}

/**
 * gmap_unshadow - remove a shadow page table completely
 * @sg: pointer to the shadow guest address space structure
 *
 * Called with sg->guest_table_lock
 */
static void gmap_unshadow(struct gmap *sg)
{
	unsigned long *table;

	BUG_ON(!gmap_is_shadow(sg));
	if (sg->removed)
		return;
	sg->removed = 1;
	gmap_call_notifier(sg, 0, -1UL);
1466
	gmap_flush_tlb(sg);
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
	table = (unsigned long *)(sg->asce & _ASCE_ORIGIN);
	switch (sg->asce & _ASCE_TYPE_MASK) {
	case _ASCE_TYPE_REGION1:
		__gmap_unshadow_r1t(sg, 0, table);
		break;
	case _ASCE_TYPE_REGION2:
		__gmap_unshadow_r2t(sg, 0, table);
		break;
	case _ASCE_TYPE_REGION3:
		__gmap_unshadow_r3t(sg, 0, table);
		break;
	case _ASCE_TYPE_SEGMENT:
		__gmap_unshadow_sgt(sg, 0, table);
		break;
	}
}

/**
 * gmap_find_shadow - find a specific asce in the list of shadow tables
 * @parent: pointer to the parent gmap
 * @asce: ASCE for which the shadow table is created
1488
 * @edat_level: edat level to be used for the shadow translation
1489 1490
 *
 * Returns the pointer to a gmap if a shadow table with the given asce is
1491 1492
 * already available, ERR_PTR(-EAGAIN) if another one is just being created,
 * otherwise NULL
1493
 */
1494 1495
static struct gmap *gmap_find_shadow(struct gmap *parent, unsigned long asce,
				     int edat_level)
1496 1497 1498 1499
{
	struct gmap *sg;

	list_for_each_entry(sg, &parent->children, list) {
1500 1501
		if (sg->orig_asce != asce || sg->edat_level != edat_level ||
		    sg->removed)
1502
			continue;
1503 1504
		if (!sg->initialized)
			return ERR_PTR(-EAGAIN);
1505 1506 1507 1508 1509 1510
		atomic_inc(&sg->ref_count);
		return sg;
	}
	return NULL;
}

1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
/**
 * gmap_shadow_valid - check if a shadow guest address space matches the
 *                     given properties and is still valid
 * @sg: pointer to the shadow guest address space structure
 * @asce: ASCE for which the shadow table is requested
 * @edat_level: edat level to be used for the shadow translation
 *
 * Returns 1 if the gmap shadow is still valid and matches the given
 * properties, the caller can continue using it. Returns 0 otherwise, the
 * caller has to request a new shadow gmap in this case.
 *
 */
int gmap_shadow_valid(struct gmap *sg, unsigned long asce, int edat_level)
{
	if (sg->removed)
		return 0;
	return sg->orig_asce == asce && sg->edat_level == edat_level;
}
EXPORT_SYMBOL_GPL(gmap_shadow_valid);

1531 1532 1533 1534
/**
 * gmap_shadow - create/find a shadow guest address space
 * @parent: pointer to the parent gmap
 * @asce: ASCE for which the shadow table is created
1535
 * @edat_level: edat level to be used for the shadow translation
1536 1537 1538 1539 1540 1541
 *
 * The pages of the top level page table referred by the asce parameter
 * will be set to read-only and marked in the PGSTEs of the kvm process.
 * The shadow table will be removed automatically on any change to the
 * PTE mapping for the source table.
 *
1542 1543 1544
 * Returns a guest address space structure, ERR_PTR(-ENOMEM) if out of memory,
 * ERR_PTR(-EAGAIN) if the caller has to retry and ERR_PTR(-EFAULT) if the
 * parent gmap table could not be protected.
1545
 */
1546 1547
struct gmap *gmap_shadow(struct gmap *parent, unsigned long asce,
			 int edat_level)
1548 1549 1550 1551 1552 1553 1554
{
	struct gmap *sg, *new;
	unsigned long limit;
	int rc;

	BUG_ON(gmap_is_shadow(parent));
	spin_lock(&parent->shadow_lock);
1555
	sg = gmap_find_shadow(parent, asce, edat_level);
1556 1557 1558 1559 1560
	spin_unlock(&parent->shadow_lock);
	if (sg)
		return sg;
	/* Create a new shadow gmap */
	limit = -1UL >> (33 - (((asce & _ASCE_TYPE_MASK) >> 2) * 11));
1561 1562
	if (asce & _ASCE_REAL_SPACE)
		limit = -1UL;
1563 1564
	new = gmap_alloc(limit);
	if (!new)
1565
		return ERR_PTR(-ENOMEM);
1566 1567 1568
	new->mm = parent->mm;
	new->parent = gmap_get(parent);
	new->orig_asce = asce;
1569
	new->edat_level = edat_level;
1570 1571 1572
	new->initialized = false;
	spin_lock(&parent->shadow_lock);
	/* Recheck if another CPU created the same shadow */
1573
	sg = gmap_find_shadow(parent, asce, edat_level);
1574 1575 1576 1577 1578
	if (sg) {
		spin_unlock(&parent->shadow_lock);
		gmap_free(new);
		return sg;
	}
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
	if (asce & _ASCE_REAL_SPACE) {
		/* only allow one real-space gmap shadow */
		list_for_each_entry(sg, &parent->children, list) {
			if (sg->orig_asce & _ASCE_REAL_SPACE) {
				spin_lock(&sg->guest_table_lock);
				gmap_unshadow(sg);
				spin_unlock(&sg->guest_table_lock);
				list_del(&sg->list);
				gmap_put(sg);
				break;
			}
		}
	}
1592 1593
	atomic_set(&new->ref_count, 2);
	list_add(&new->list, &parent->children);
1594 1595 1596 1597 1598 1599
	if (asce & _ASCE_REAL_SPACE) {
		/* nothing to protect, return right away */
		new->initialized = true;
		spin_unlock(&parent->shadow_lock);
		return new;
	}
1600 1601
	spin_unlock(&parent->shadow_lock);
	/* protect after insertion, so it will get properly invalidated */
1602 1603
	down_read(&parent->mm->mmap_sem);
	rc = gmap_protect_range(parent, asce & _ASCE_ORIGIN,
1604
				((asce & _ASCE_TABLE_LENGTH) + 1) * PAGE_SIZE,
1605
				PROT_READ, GMAP_NOTIFY_SHADOW);
1606
	up_read(&parent->mm->mmap_sem);
1607 1608
	spin_lock(&parent->shadow_lock);
	new->initialized = true;
1609
	if (rc) {
1610
		list_del(&new->list);
1611
		gmap_free(new);
1612 1613 1614 1615
		new = ERR_PTR(rc);
	}
	spin_unlock(&parent->shadow_lock);
	return new;
1616 1617 1618 1619 1620 1621 1622 1623
}
EXPORT_SYMBOL_GPL(gmap_shadow);

/**
 * gmap_shadow_r2t - create an empty shadow region 2 table
 * @sg: pointer to the shadow guest address space structure
 * @saddr: faulting address in the shadow gmap
 * @r2t: parent gmap address of the region 2 table to get shadowed
1624
 * @fake: r2t references contiguous guest memory block, not a r2t
1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
 *
 * The r2t parameter specifies the address of the source table. The
 * four pages of the source table are made read-only in the parent gmap
 * address space. A write to the source table area @r2t will automatically
 * remove the shadow r2 table and all of its decendents.
 *
 * Returns 0 if successfully shadowed or already shadowed, -EAGAIN if the
 * shadow table structure is incomplete, -ENOMEM if out of memory and
 * -EFAULT if an address in the parent gmap could not be resolved.
 *
 * Called with sg->mm->mmap_sem in read.
 */
1637 1638
int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t,
		    int fake)
1639 1640 1641 1642 1643 1644 1645 1646
{
	unsigned long raddr, origin, offset, len;
	unsigned long *s_r2t, *table;
	struct page *page;
	int rc;

	BUG_ON(!gmap_is_shadow(sg));
	/* Allocate a shadow region second table */
1647
	page = alloc_pages(GFP_KERNEL, CRST_ALLOC_ORDER);
1648 1649 1650
	if (!page)
		return -ENOMEM;
	page->index = r2t & _REGION_ENTRY_ORIGIN;
1651 1652
	if (fake)
		page->index |= GMAP_SHADOW_FAKE_TABLE;
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
	s_r2t = (unsigned long *) page_to_phys(page);
	/* Install shadow region second table */
	spin_lock(&sg->guest_table_lock);
	table = gmap_table_walk(sg, saddr, 4); /* get region-1 pointer */
	if (!table) {
		rc = -EAGAIN;		/* Race with unshadow */
		goto out_free;
	}
	if (!(*table & _REGION_ENTRY_INVALID)) {
		rc = 0;			/* Already established */
		goto out_free;
1664 1665 1666
	} else if (*table & _REGION_ENTRY_ORIGIN) {
		rc = -EAGAIN;		/* Race with shadow */
		goto out_free;
1667 1668
	}
	crst_table_init(s_r2t, _REGION2_ENTRY_EMPTY);
1669 1670 1671
	/* mark as invalid as long as the parent table is not protected */
	*table = (unsigned long) s_r2t | _REGION_ENTRY_LENGTH |
		 _REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_INVALID;
1672 1673
	if (sg->edat_level >= 1)
		*table |= (r2t & _REGION_ENTRY_PROTECT);
1674
	list_add(&page->lru, &sg->crst_list);
1675 1676 1677 1678 1679 1680
	if (fake) {
		/* nothing to protect for fake tables */
		*table &= ~_REGION_ENTRY_INVALID;
		spin_unlock(&sg->guest_table_lock);
		return 0;
	}
1681 1682
	spin_unlock(&sg->guest_table_lock);
	/* Make r2t read-only in parent gmap page table */
1683
	raddr = (saddr & _REGION1_MASK) | _SHADOW_RMAP_REGION1;
1684
	origin = r2t & _REGION_ENTRY_ORIGIN;
1685 1686
	offset = ((r2t & _REGION_ENTRY_OFFSET) >> 6) * PAGE_SIZE;
	len = ((r2t & _REGION_ENTRY_LENGTH) + 1) * PAGE_SIZE - offset;
1687
	rc = gmap_protect_rmap(sg, raddr, origin + offset, len);
1688 1689 1690 1691 1692 1693 1694 1695 1696
	spin_lock(&sg->guest_table_lock);
	if (!rc) {
		table = gmap_table_walk(sg, saddr, 4);
		if (!table || (*table & _REGION_ENTRY_ORIGIN) !=
			      (unsigned long) s_r2t)
			rc = -EAGAIN;		/* Race with unshadow */
		else
			*table &= ~_REGION_ENTRY_INVALID;
	} else {
1697 1698
		gmap_unshadow_r2t(sg, raddr);
	}
1699
	spin_unlock(&sg->guest_table_lock);
1700 1701 1702
	return rc;
out_free:
	spin_unlock(&sg->guest_table_lock);
1703
	__free_pages(page, CRST_ALLOC_ORDER);
1704 1705
	return rc;
}
1706 1707 1708 1709 1710 1711 1712
EXPORT_SYMBOL_GPL(gmap_shadow_r2t);

/**
 * gmap_shadow_r3t - create a shadow region 3 table
 * @sg: pointer to the shadow guest address space structure
 * @saddr: faulting address in the shadow gmap
 * @r3t: parent gmap address of the region 3 table to get shadowed
1713
 * @fake: r3t references contiguous guest memory block, not a r3t
1714 1715 1716 1717 1718 1719 1720
 *
 * Returns 0 if successfully shadowed or already shadowed, -EAGAIN if the
 * shadow table structure is incomplete, -ENOMEM if out of memory and
 * -EFAULT if an address in the parent gmap could not be resolved.
 *
 * Called with sg->mm->mmap_sem in read.
 */
1721 1722
int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t,
		    int fake)
1723 1724 1725 1726 1727 1728 1729 1730
{
	unsigned long raddr, origin, offset, len;
	unsigned long *s_r3t, *table;
	struct page *page;
	int rc;

	BUG_ON(!gmap_is_shadow(sg));
	/* Allocate a shadow region second table */
1731
	page = alloc_pages(GFP_KERNEL, CRST_ALLOC_ORDER);
1732 1733 1734
	if (!page)
		return -ENOMEM;
	page->index = r3t & _REGION_ENTRY_ORIGIN;
1735 1736
	if (fake)
		page->index |= GMAP_SHADOW_FAKE_TABLE;
1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747
	s_r3t = (unsigned long *) page_to_phys(page);
	/* Install shadow region second table */
	spin_lock(&sg->guest_table_lock);
	table = gmap_table_walk(sg, saddr, 3); /* get region-2 pointer */
	if (!table) {
		rc = -EAGAIN;		/* Race with unshadow */
		goto out_free;
	}
	if (!(*table & _REGION_ENTRY_INVALID)) {
		rc = 0;			/* Already established */
		goto out_free;
1748 1749
	} else if (*table & _REGION_ENTRY_ORIGIN) {
		rc = -EAGAIN;		/* Race with shadow */
1750 1751
	}
	crst_table_init(s_r3t, _REGION3_ENTRY_EMPTY);
1752 1753 1754
	/* mark as invalid as long as the parent table is not protected */
	*table = (unsigned long) s_r3t | _REGION_ENTRY_LENGTH |
		 _REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_INVALID;
1755 1756
	if (sg->edat_level >= 1)
		*table |= (r3t & _REGION_ENTRY_PROTECT);
1757
	list_add(&page->lru, &sg->crst_list);
1758 1759 1760 1761 1762 1763
	if (fake) {
		/* nothing to protect for fake tables */
		*table &= ~_REGION_ENTRY_INVALID;
		spin_unlock(&sg->guest_table_lock);
		return 0;
	}
1764 1765
	spin_unlock(&sg->guest_table_lock);
	/* Make r3t read-only in parent gmap page table */
1766
	raddr = (saddr & _REGION2_MASK) | _SHADOW_RMAP_REGION2;
1767
	origin = r3t & _REGION_ENTRY_ORIGIN;
1768 1769
	offset = ((r3t & _REGION_ENTRY_OFFSET) >> 6) * PAGE_SIZE;
	len = ((r3t & _REGION_ENTRY_LENGTH) + 1) * PAGE_SIZE - offset;
1770
	rc = gmap_protect_rmap(sg, raddr, origin + offset, len);
1771 1772 1773 1774 1775 1776 1777 1778 1779
	spin_lock(&sg->guest_table_lock);
	if (!rc) {
		table = gmap_table_walk(sg, saddr, 3);
		if (!table || (*table & _REGION_ENTRY_ORIGIN) !=
			      (unsigned long) s_r3t)
			rc = -EAGAIN;		/* Race with unshadow */
		else
			*table &= ~_REGION_ENTRY_INVALID;
	} else {
1780 1781
		gmap_unshadow_r3t(sg, raddr);
	}
1782
	spin_unlock(&sg->guest_table_lock);
1783 1784 1785
	return rc;
out_free:
	spin_unlock(&sg->guest_table_lock);
1786
	__free_pages(page, CRST_ALLOC_ORDER);
1787 1788 1789 1790 1791 1792 1793 1794 1795
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_shadow_r3t);

/**
 * gmap_shadow_sgt - create a shadow segment table
 * @sg: pointer to the shadow guest address space structure
 * @saddr: faulting address in the shadow gmap
 * @sgt: parent gmap address of the segment table to get shadowed
1796
 * @fake: sgt references contiguous guest memory block, not a sgt
1797 1798 1799 1800 1801 1802 1803
 *
 * Returns: 0 if successfully shadowed or already shadowed, -EAGAIN if the
 * shadow table structure is incomplete, -ENOMEM if out of memory and
 * -EFAULT if an address in the parent gmap could not be resolved.
 *
 * Called with sg->mm->mmap_sem in read.
 */
1804 1805
int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt,
		    int fake)
1806 1807 1808 1809 1810 1811
{
	unsigned long raddr, origin, offset, len;
	unsigned long *s_sgt, *table;
	struct page *page;
	int rc;

1812
	BUG_ON(!gmap_is_shadow(sg) || (sgt & _REGION3_ENTRY_LARGE));
1813
	/* Allocate a shadow segment table */
1814
	page = alloc_pages(GFP_KERNEL, CRST_ALLOC_ORDER);
1815 1816 1817
	if (!page)
		return -ENOMEM;
	page->index = sgt & _REGION_ENTRY_ORIGIN;
1818 1819
	if (fake)
		page->index |= GMAP_SHADOW_FAKE_TABLE;
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
	s_sgt = (unsigned long *) page_to_phys(page);
	/* Install shadow region second table */
	spin_lock(&sg->guest_table_lock);
	table = gmap_table_walk(sg, saddr, 2); /* get region-3 pointer */
	if (!table) {
		rc = -EAGAIN;		/* Race with unshadow */
		goto out_free;
	}
	if (!(*table & _REGION_ENTRY_INVALID)) {
		rc = 0;			/* Already established */
		goto out_free;
1831 1832 1833
	} else if (*table & _REGION_ENTRY_ORIGIN) {
		rc = -EAGAIN;		/* Race with shadow */
		goto out_free;
1834 1835
	}
	crst_table_init(s_sgt, _SEGMENT_ENTRY_EMPTY);
1836 1837 1838
	/* mark as invalid as long as the parent table is not protected */
	*table = (unsigned long) s_sgt | _REGION_ENTRY_LENGTH |
		 _REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_INVALID;
1839 1840
	if (sg->edat_level >= 1)
		*table |= sgt & _REGION_ENTRY_PROTECT;
1841
	list_add(&page->lru, &sg->crst_list);
1842 1843 1844 1845 1846 1847
	if (fake) {
		/* nothing to protect for fake tables */
		*table &= ~_REGION_ENTRY_INVALID;
		spin_unlock(&sg->guest_table_lock);
		return 0;
	}
1848 1849
	spin_unlock(&sg->guest_table_lock);
	/* Make sgt read-only in parent gmap page table */
1850
	raddr = (saddr & _REGION3_MASK) | _SHADOW_RMAP_REGION3;
1851
	origin = sgt & _REGION_ENTRY_ORIGIN;
1852 1853
	offset = ((sgt & _REGION_ENTRY_OFFSET) >> 6) * PAGE_SIZE;
	len = ((sgt & _REGION_ENTRY_LENGTH) + 1) * PAGE_SIZE - offset;
1854
	rc = gmap_protect_rmap(sg, raddr, origin + offset, len);
1855 1856 1857 1858 1859 1860 1861 1862 1863
	spin_lock(&sg->guest_table_lock);
	if (!rc) {
		table = gmap_table_walk(sg, saddr, 2);
		if (!table || (*table & _REGION_ENTRY_ORIGIN) !=
			      (unsigned long) s_sgt)
			rc = -EAGAIN;		/* Race with unshadow */
		else
			*table &= ~_REGION_ENTRY_INVALID;
	} else {
1864 1865
		gmap_unshadow_sgt(sg, raddr);
	}
1866
	spin_unlock(&sg->guest_table_lock);
1867 1868 1869
	return rc;
out_free:
	spin_unlock(&sg->guest_table_lock);
1870
	__free_pages(page, CRST_ALLOC_ORDER);
1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_shadow_sgt);

/**
 * gmap_shadow_lookup_pgtable - find a shadow page table
 * @sg: pointer to the shadow guest address space structure
 * @saddr: the address in the shadow aguest address space
 * @pgt: parent gmap address of the page table to get shadowed
 * @dat_protection: if the pgtable is marked as protected by dat
1881
 * @fake: pgt references contiguous guest memory block, not a pgtable
1882 1883 1884 1885 1886 1887 1888
 *
 * Returns 0 if the shadow page table was found and -EAGAIN if the page
 * table was not found.
 *
 * Called with sg->mm->mmap_sem in read.
 */
int gmap_shadow_pgt_lookup(struct gmap *sg, unsigned long saddr,
1889 1890
			   unsigned long *pgt, int *dat_protection,
			   int *fake)
1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
{
	unsigned long *table;
	struct page *page;
	int rc;

	BUG_ON(!gmap_is_shadow(sg));
	spin_lock(&sg->guest_table_lock);
	table = gmap_table_walk(sg, saddr, 1); /* get segment pointer */
	if (table && !(*table & _SEGMENT_ENTRY_INVALID)) {
		/* Shadow page tables are full pages (pte+pgste) */
		page = pfn_to_page(*table >> PAGE_SHIFT);
1902
		*pgt = page->index & ~GMAP_SHADOW_FAKE_TABLE;
1903
		*dat_protection = !!(*table & _SEGMENT_ENTRY_PROTECT);
1904
		*fake = !!(page->index & GMAP_SHADOW_FAKE_TABLE);
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919
		rc = 0;
	} else  {
		rc = -EAGAIN;
	}
	spin_unlock(&sg->guest_table_lock);
	return rc;

}
EXPORT_SYMBOL_GPL(gmap_shadow_pgt_lookup);

/**
 * gmap_shadow_pgt - instantiate a shadow page table
 * @sg: pointer to the shadow guest address space structure
 * @saddr: faulting address in the shadow gmap
 * @pgt: parent gmap address of the page table to get shadowed
1920
 * @fake: pgt references contiguous guest memory block, not a pgtable
1921 1922 1923 1924 1925 1926 1927
 *
 * Returns 0 if successfully shadowed or already shadowed, -EAGAIN if the
 * shadow table structure is incomplete, -ENOMEM if out of memory,
 * -EFAULT if an address in the parent gmap could not be resolved and
 *
 * Called with gmap->mm->mmap_sem in read
 */
1928 1929
int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt,
		    int fake)
1930 1931 1932 1933 1934 1935
{
	unsigned long raddr, origin;
	unsigned long *s_pgt, *table;
	struct page *page;
	int rc;

1936
	BUG_ON(!gmap_is_shadow(sg) || (pgt & _SEGMENT_ENTRY_LARGE));
1937 1938 1939 1940 1941
	/* Allocate a shadow page table */
	page = page_table_alloc_pgste(sg->mm);
	if (!page)
		return -ENOMEM;
	page->index = pgt & _SEGMENT_ENTRY_ORIGIN;
1942 1943
	if (fake)
		page->index |= GMAP_SHADOW_FAKE_TABLE;
1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954
	s_pgt = (unsigned long *) page_to_phys(page);
	/* Install shadow page table */
	spin_lock(&sg->guest_table_lock);
	table = gmap_table_walk(sg, saddr, 1); /* get segment pointer */
	if (!table) {
		rc = -EAGAIN;		/* Race with unshadow */
		goto out_free;
	}
	if (!(*table & _SEGMENT_ENTRY_INVALID)) {
		rc = 0;			/* Already established */
		goto out_free;
1955 1956 1957
	} else if (*table & _SEGMENT_ENTRY_ORIGIN) {
		rc = -EAGAIN;		/* Race with shadow */
		goto out_free;
1958
	}
1959
	/* mark as invalid as long as the parent table is not protected */
1960
	*table = (unsigned long) s_pgt | _SEGMENT_ENTRY |
1961
		 (pgt & _SEGMENT_ENTRY_PROTECT) | _SEGMENT_ENTRY_INVALID;
1962
	list_add(&page->lru, &sg->pt_list);
1963 1964 1965 1966 1967 1968
	if (fake) {
		/* nothing to protect for fake tables */
		*table &= ~_SEGMENT_ENTRY_INVALID;
		spin_unlock(&sg->guest_table_lock);
		return 0;
	}
1969 1970
	spin_unlock(&sg->guest_table_lock);
	/* Make pgt read-only in parent gmap page table (not the pgste) */
1971
	raddr = (saddr & _SEGMENT_MASK) | _SHADOW_RMAP_SEGMENT;
1972
	origin = pgt & _SEGMENT_ENTRY_ORIGIN & PAGE_MASK;
1973
	rc = gmap_protect_rmap(sg, raddr, origin, PAGE_SIZE);
1974 1975 1976 1977 1978 1979 1980 1981 1982
	spin_lock(&sg->guest_table_lock);
	if (!rc) {
		table = gmap_table_walk(sg, saddr, 1);
		if (!table || (*table & _SEGMENT_ENTRY_ORIGIN) !=
			      (unsigned long) s_pgt)
			rc = -EAGAIN;		/* Race with unshadow */
		else
			*table &= ~_SEGMENT_ENTRY_INVALID;
	} else {
1983 1984
		gmap_unshadow_pgt(sg, raddr);
	}
1985
	spin_unlock(&sg->guest_table_lock);
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
	return rc;
out_free:
	spin_unlock(&sg->guest_table_lock);
	page_table_free_pgste(page);
	return rc;

}
EXPORT_SYMBOL_GPL(gmap_shadow_pgt);

/**
 * gmap_shadow_page - create a shadow page mapping
 * @sg: pointer to the shadow guest address space structure
 * @saddr: faulting address in the shadow gmap
1999
 * @pte: pte in parent gmap address space to get shadowed
2000 2001 2002 2003 2004 2005 2006
 *
 * Returns 0 if successfully shadowed or already shadowed, -EAGAIN if the
 * shadow table structure is incomplete, -ENOMEM if out of memory and
 * -EFAULT if an address in the parent gmap could not be resolved.
 *
 * Called with sg->mm->mmap_sem in read.
 */
2007
int gmap_shadow_page(struct gmap *sg, unsigned long saddr, pte_t pte)
2008 2009 2010
{
	struct gmap *parent;
	struct gmap_rmap *rmap;
2011
	unsigned long vmaddr, paddr;
2012 2013
	spinlock_t *ptl;
	pte_t *sptep, *tptep;
2014
	int prot;
2015 2016 2017 2018
	int rc;

	BUG_ON(!gmap_is_shadow(sg));
	parent = sg->parent;
2019
	prot = (pte_val(pte) & _PAGE_PROTECT) ? PROT_READ : PROT_WRITE;
2020 2021 2022 2023 2024 2025 2026

	rmap = kzalloc(sizeof(*rmap), GFP_KERNEL);
	if (!rmap)
		return -ENOMEM;
	rmap->raddr = (saddr & PAGE_MASK) | _SHADOW_RMAP_PGTABLE;

	while (1) {
2027
		paddr = pte_val(pte) & PAGE_MASK;
2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047
		vmaddr = __gmap_translate(parent, paddr);
		if (IS_ERR_VALUE(vmaddr)) {
			rc = vmaddr;
			break;
		}
		rc = radix_tree_preload(GFP_KERNEL);
		if (rc)
			break;
		rc = -EAGAIN;
		sptep = gmap_pte_op_walk(parent, paddr, &ptl);
		if (sptep) {
			spin_lock(&sg->guest_table_lock);
			/* Get page table pointer */
			tptep = (pte_t *) gmap_table_walk(sg, saddr, 0);
			if (!tptep) {
				spin_unlock(&sg->guest_table_lock);
				gmap_pte_op_end(ptl);
				radix_tree_preload_end();
				break;
			}
2048
			rc = ptep_shadow_pte(sg->mm, saddr, sptep, tptep, pte);
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060
			if (rc > 0) {
				/* Success and a new mapping */
				gmap_insert_rmap(sg, vmaddr, rmap);
				rmap = NULL;
				rc = 0;
			}
			gmap_pte_op_end(ptl);
			spin_unlock(&sg->guest_table_lock);
		}
		radix_tree_preload_end();
		if (!rc)
			break;
2061
		rc = gmap_pte_op_fixup(parent, paddr, vmaddr, prot);
2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075
		if (rc)
			break;
	}
	kfree(rmap);
	return rc;
}
EXPORT_SYMBOL_GPL(gmap_shadow_page);

/**
 * gmap_shadow_notify - handle notifications for shadow gmap
 *
 * Called with sg->parent->shadow_lock.
 */
static void gmap_shadow_notify(struct gmap *sg, unsigned long vmaddr,
2076
			       unsigned long gaddr)
2077 2078
{
	struct gmap_rmap *rmap, *rnext, *head;
2079
	unsigned long start, end, bits, raddr;
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089

	BUG_ON(!gmap_is_shadow(sg));

	spin_lock(&sg->guest_table_lock);
	if (sg->removed) {
		spin_unlock(&sg->guest_table_lock);
		return;
	}
	/* Check for top level table */
	start = sg->orig_asce & _ASCE_ORIGIN;
2090
	end = start + ((sg->orig_asce & _ASCE_TABLE_LENGTH) + 1) * PAGE_SIZE;
2091 2092
	if (!(sg->orig_asce & _ASCE_REAL_SPACE) && gaddr >= start &&
	    gaddr < end) {
2093 2094 2095 2096 2097 2098 2099 2100
		/* The complete shadow table has to go */
		gmap_unshadow(sg);
		spin_unlock(&sg->guest_table_lock);
		list_del(&sg->list);
		gmap_put(sg);
		return;
	}
	/* Remove the page table tree from on specific entry */
2101
	head = radix_tree_delete(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT);
2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125
	gmap_for_each_rmap_safe(rmap, rnext, head) {
		bits = rmap->raddr & _SHADOW_RMAP_MASK;
		raddr = rmap->raddr ^ bits;
		switch (bits) {
		case _SHADOW_RMAP_REGION1:
			gmap_unshadow_r2t(sg, raddr);
			break;
		case _SHADOW_RMAP_REGION2:
			gmap_unshadow_r3t(sg, raddr);
			break;
		case _SHADOW_RMAP_REGION3:
			gmap_unshadow_sgt(sg, raddr);
			break;
		case _SHADOW_RMAP_SEGMENT:
			gmap_unshadow_pgt(sg, raddr);
			break;
		case _SHADOW_RMAP_PGTABLE:
			gmap_unshadow_page(sg, raddr);
			break;
		}
		kfree(rmap);
	}
	spin_unlock(&sg->guest_table_lock);
}
2126 2127 2128 2129 2130 2131

/**
 * ptep_notify - call all invalidation callbacks for a specific pte.
 * @mm: pointer to the process mm_struct
 * @addr: virtual address in the process address space
 * @pte: pointer to the page table entry
2132
 * @bits: bits from the pgste that caused the notify call
2133 2134 2135 2136
 *
 * This function is assumed to be called with the page table lock held
 * for the pte to notify.
 */
2137 2138
void ptep_notify(struct mm_struct *mm, unsigned long vmaddr,
		 pte_t *pte, unsigned long bits)
2139
{
2140
	unsigned long offset, gaddr = 0;
2141
	unsigned long *table;
2142
	struct gmap *gmap, *sg, *next;
2143 2144

	offset = ((unsigned long) pte) & (255 * sizeof(pte_t));
2145
	offset = offset * (PAGE_SIZE / sizeof(pte_t));
2146 2147 2148
	rcu_read_lock();
	list_for_each_entry_rcu(gmap, &mm->context.gmap_list, list) {
		spin_lock(&gmap->guest_table_lock);
2149 2150
		table = radix_tree_lookup(&gmap->host_to_guest,
					  vmaddr >> PMD_SHIFT);
2151 2152 2153
		if (table)
			gaddr = __gmap_segment_gaddr(table) + offset;
		spin_unlock(&gmap->guest_table_lock);
2154 2155 2156 2157 2158 2159 2160
		if (!table)
			continue;

		if (!list_empty(&gmap->children) && (bits & PGSTE_VSIE_BIT)) {
			spin_lock(&gmap->shadow_lock);
			list_for_each_entry_safe(sg, next,
						 &gmap->children, list)
2161
				gmap_shadow_notify(sg, vmaddr, gaddr);
2162 2163 2164
			spin_unlock(&gmap->shadow_lock);
		}
		if (bits & PGSTE_IN_BIT)
2165
			gmap_call_notifier(gmap, gaddr, gaddr + PAGE_SIZE - 1);
2166
	}
2167
	rcu_read_unlock();
2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188
}
EXPORT_SYMBOL_GPL(ptep_notify);

static inline void thp_split_mm(struct mm_struct *mm)
{
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
	struct vm_area_struct *vma;
	unsigned long addr;

	for (vma = mm->mmap; vma != NULL; vma = vma->vm_next) {
		for (addr = vma->vm_start;
		     addr < vma->vm_end;
		     addr += PAGE_SIZE)
			follow_page(vma, addr, FOLL_SPLIT);
		vma->vm_flags &= ~VM_HUGEPAGE;
		vma->vm_flags |= VM_NOHUGEPAGE;
	}
	mm->def_flags |= VM_NOHUGEPAGE;
#endif
}

2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219
/*
 * Remove all empty zero pages from the mapping for lazy refaulting
 * - This must be called after mm->context.has_pgste is set, to avoid
 *   future creation of zero pages
 * - This must be called after THP was enabled
 */
static int __zap_zero_pages(pmd_t *pmd, unsigned long start,
			   unsigned long end, struct mm_walk *walk)
{
	unsigned long addr;

	for (addr = start; addr != end; addr += PAGE_SIZE) {
		pte_t *ptep;
		spinlock_t *ptl;

		ptep = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
		if (is_zero_pfn(pte_pfn(*ptep)))
			ptep_xchg_direct(walk->mm, addr, ptep, __pte(_PAGE_INVALID));
		pte_unmap_unlock(ptep, ptl);
	}
	return 0;
}

static inline void zap_zero_pages(struct mm_struct *mm)
{
	struct mm_walk walk = { .pmd_entry = __zap_zero_pages };

	walk.mm = mm;
	walk_page_range(0, TASK_SIZE, &walk);
}

2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236
/*
 * switch on pgstes for its userspace process (for kvm)
 */
int s390_enable_sie(void)
{
	struct mm_struct *mm = current->mm;

	/* Do we have pgstes? if yes, we are done */
	if (mm_has_pgste(mm))
		return 0;
	/* Fail if the page tables are 2K */
	if (!mm_alloc_pgste(mm))
		return -EINVAL;
	down_write(&mm->mmap_sem);
	mm->context.has_pgste = 1;
	/* split thp mappings and disable thp for future mappings */
	thp_split_mm(mm);
2237
	zap_zero_pages(mm);
2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262
	up_write(&mm->mmap_sem);
	return 0;
}
EXPORT_SYMBOL_GPL(s390_enable_sie);

/*
 * Enable storage key handling from now on and initialize the storage
 * keys with the default key.
 */
static int __s390_enable_skey(pte_t *pte, unsigned long addr,
			      unsigned long next, struct mm_walk *walk)
{
	/* Clear storage key */
	ptep_zap_key(walk->mm, addr, pte);
	return 0;
}

int s390_enable_skey(void)
{
	struct mm_walk walk = { .pte_entry = __s390_enable_skey };
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	int rc = 0;

	down_write(&mm->mmap_sem);
2263
	if (mm_uses_skeys(mm))
2264 2265
		goto out_up;

2266
	mm->context.uses_skeys = 1;
2267 2268 2269
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
		if (ksm_madvise(vma, vma->vm_start, vma->vm_end,
				MADV_UNMERGEABLE, &vma->vm_flags)) {
2270
			mm->context.uses_skeys = 0;
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 2304 2305
			rc = -ENOMEM;
			goto out_up;
		}
	}
	mm->def_flags &= ~VM_MERGEABLE;

	walk.mm = mm;
	walk_page_range(0, TASK_SIZE, &walk);

out_up:
	up_write(&mm->mmap_sem);
	return rc;
}
EXPORT_SYMBOL_GPL(s390_enable_skey);

/*
 * Reset CMMA state, make all pages stable again.
 */
static int __s390_reset_cmma(pte_t *pte, unsigned long addr,
			     unsigned long next, struct mm_walk *walk)
{
	ptep_zap_unused(walk->mm, addr, pte, 1);
	return 0;
}

void s390_reset_cmma(struct mm_struct *mm)
{
	struct mm_walk walk = { .pte_entry = __s390_reset_cmma };

	down_write(&mm->mmap_sem);
	walk.mm = mm;
	walk_page_range(0, TASK_SIZE, &walk);
	up_write(&mm->mmap_sem);
}
EXPORT_SYMBOL_GPL(s390_reset_cmma);