hugetlbpage.c 26.6 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
B
Becky Bruce 已提交
2
 * PPC Huge TLB Page Support for Kernel.
L
Linus Torvalds 已提交
3 4
 *
 * Copyright (C) 2003 David Gibson, IBM Corporation.
B
Becky Bruce 已提交
5
 * Copyright (C) 2011 Becky Bruce, Freescale Semiconductor
L
Linus Torvalds 已提交
6 7 8 9 10 11
 *
 * Based on the IA-32 version:
 * Copyright (C) 2002, Rohit Seth <rohit.seth@intel.com>
 */

#include <linux/mm.h>
12
#include <linux/io.h>
13
#include <linux/slab.h>
L
Linus Torvalds 已提交
14
#include <linux/hugetlb.h>
15
#include <linux/export.h>
B
Becky Bruce 已提交
16 17 18
#include <linux/of_fdt.h>
#include <linux/memblock.h>
#include <linux/bootmem.h>
19
#include <linux/moduleparam.h>
20
#include <asm/pgtable.h>
L
Linus Torvalds 已提交
21 22
#include <asm/pgalloc.h>
#include <asm/tlb.h>
B
Becky Bruce 已提交
23
#include <asm/setup.h>
24 25 26
#include <asm/hugetlb.h>

#ifdef CONFIG_HUGETLB_PAGE
L
Linus Torvalds 已提交
27

28 29 30
#define PAGE_SHIFT_64K	16
#define PAGE_SHIFT_16M	24
#define PAGE_SHIFT_16G	34
31

B
Becky Bruce 已提交
32
unsigned int HPAGE_SHIFT;
33

B
Becky Bruce 已提交
34 35
/*
 * Tracks gpages after the device tree is scanned and before the
36 37 38 39
 * huge_boot_pages list is ready.  On non-Freescale implementations, this is
 * just used to track 16G pages and so is a single array.  FSL-based
 * implementations may have more than one gpage size, so we need multiple
 * arrays
B
Becky Bruce 已提交
40
 */
41
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
42 43 44 45 46 47
#define MAX_NUMBER_GPAGES	128
struct psize_gpages {
	u64 gpage_list[MAX_NUMBER_GPAGES];
	unsigned int nr_gpages;
};
static struct psize_gpages gpage_freearray[MMU_PAGE_COUNT];
48 49 50 51
#else
#define MAX_NUMBER_GPAGES	1024
static u64 gpage_freearray[MAX_NUMBER_GPAGES];
static unsigned nr_gpages;
B
Becky Bruce 已提交
52
#endif
53

54 55
#define hugepd_none(hpd)	((hpd).pd == 0)

56 57 58 59 60 61 62 63 64
#ifdef CONFIG_PPC_BOOK3S_64
/*
 * At this point we do the placement change only for BOOK3S 64. This would
 * possibly work on other subarchs.
 */

/*
 * We have PGD_INDEX_SIZ = 12 and PTE_INDEX_SIZE = 8, so that we can have
 * 16GB hugepage pte in PGD and 16MB hugepage pte at PMD;
65 66 67
 *
 * Defined in such a way that we can optimize away code block at build time
 * if CONFIG_HUGETLB_PAGE=n.
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108
 */
int pmd_huge(pmd_t pmd)
{
	/*
	 * leaf pte for huge page, bottom two bits != 00
	 */
	return ((pmd_val(pmd) & 0x3) != 0x0);
}

int pud_huge(pud_t pud)
{
	/*
	 * leaf pte for huge page, bottom two bits != 00
	 */
	return ((pud_val(pud) & 0x3) != 0x0);
}

int pgd_huge(pgd_t pgd)
{
	/*
	 * leaf pte for huge page, bottom two bits != 00
	 */
	return ((pgd_val(pgd) & 0x3) != 0x0);
}
#else
int pmd_huge(pmd_t pmd)
{
	return 0;
}

int pud_huge(pud_t pud)
{
	return 0;
}

int pgd_huge(pgd_t pgd)
{
	return 0;
}
#endif

109 110
pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
{
111
	/* Only called for hugetlbfs pages, hence can ignore THP */
112
	return __find_linux_pte_or_hugepte(mm->pgd, addr, NULL);
113 114
}

115
static int __hugepte_alloc(struct mm_struct *mm, hugepd_t *hpdp,
116
			   unsigned long address, unsigned pdshift, unsigned pshift)
117
{
B
Becky Bruce 已提交
118 119 120
	struct kmem_cache *cachep;
	pte_t *new;

121
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
122 123 124
	int i;
	int num_hugepd = 1 << (pshift - pdshift);
	cachep = hugepte_cache;
125 126
#else
	cachep = PGT_CACHE(pdshift - pshift);
B
Becky Bruce 已提交
127 128 129
#endif

	new = kmem_cache_zalloc(cachep, GFP_KERNEL|__GFP_REPEAT);
130

131 132 133
	BUG_ON(pshift > HUGEPD_SHIFT_MASK);
	BUG_ON((unsigned long)new & HUGEPD_SHIFT_MASK);

134 135 136 137
	if (! new)
		return -ENOMEM;

	spin_lock(&mm->page_table_lock);
138
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
139 140 141 142 143 144 145 146 147 148
	/*
	 * We have multiple higher-level entries that point to the same
	 * actual pte location.  Fill in each as we go and backtrack on error.
	 * We need all of these so the DTLB pgtable walk code can find the
	 * right higher-level entry without knowing if it's a hugepage or not.
	 */
	for (i = 0; i < num_hugepd; i++, hpdp++) {
		if (unlikely(!hugepd_none(*hpdp)))
			break;
		else
149
			/* We use the old format for PPC_FSL_BOOK3E */
B
Becky Bruce 已提交
150 151 152 153 154 155 156 157
			hpdp->pd = ((unsigned long)new & ~PD_HUGE) | pshift;
	}
	/* If we bailed from the for loop early, an error occurred, clean up */
	if (i < num_hugepd) {
		for (i = i - 1 ; i >= 0; i--, hpdp--)
			hpdp->pd = 0;
		kmem_cache_free(cachep, new);
	}
158 159 160
#else
	if (!hugepd_none(*hpdp))
		kmem_cache_free(cachep, new);
161 162 163 164 165
	else {
#ifdef CONFIG_PPC_BOOK3S_64
		hpdp->pd = (unsigned long)new |
			    (shift_to_mmu_psize(pshift) << 2);
#else
166
		hpdp->pd = ((unsigned long)new & ~PD_HUGE) | pshift;
167 168
#endif
	}
B
Becky Bruce 已提交
169
#endif
170 171 172 173
	spin_unlock(&mm->page_table_lock);
	return 0;
}

174 175 176 177 178 179 180 181 182 183 184 185
/*
 * These macros define how to determine which level of the page table holds
 * the hpdp.
 */
#ifdef CONFIG_PPC_FSL_BOOK3E
#define HUGEPD_PGD_SHIFT PGDIR_SHIFT
#define HUGEPD_PUD_SHIFT PUD_SHIFT
#else
#define HUGEPD_PGD_SHIFT PUD_SHIFT
#define HUGEPD_PUD_SHIFT PMD_SHIFT
#endif

186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
#ifdef CONFIG_PPC_BOOK3S_64
/*
 * At this point we do the placement change only for BOOK3S 64. This would
 * possibly work on other subarchs.
 */
pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr, unsigned long sz)
{
	pgd_t *pg;
	pud_t *pu;
	pmd_t *pm;
	hugepd_t *hpdp = NULL;
	unsigned pshift = __ffs(sz);
	unsigned pdshift = PGDIR_SHIFT;

	addr &= ~(sz-1);
	pg = pgd_offset(mm, addr);

	if (pshift == PGDIR_SHIFT)
		/* 16GB huge page */
		return (pte_t *) pg;
	else if (pshift > PUD_SHIFT)
		/*
		 * We need to use hugepd table
		 */
		hpdp = (hugepd_t *)pg;
	else {
		pdshift = PUD_SHIFT;
		pu = pud_alloc(mm, pg, addr);
		if (pshift == PUD_SHIFT)
			return (pte_t *)pu;
		else if (pshift > PMD_SHIFT)
			hpdp = (hugepd_t *)pu;
		else {
			pdshift = PMD_SHIFT;
			pm = pmd_alloc(mm, pu, addr);
			if (pshift == PMD_SHIFT)
				/* 16MB hugepage */
				return (pte_t *)pm;
			else
				hpdp = (hugepd_t *)pm;
		}
	}
	if (!hpdp)
		return NULL;

	BUG_ON(!hugepd_none(*hpdp) && !hugepd_ok(*hpdp));

	if (hugepd_none(*hpdp) && __hugepte_alloc(mm, hpdp, addr, pdshift, pshift))
		return NULL;

236
	return hugepte_offset(*hpdp, addr, pdshift);
237 238 239 240
}

#else

241
pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr, unsigned long sz)
242
{
243 244 245 246 247 248 249 250 251 252
	pgd_t *pg;
	pud_t *pu;
	pmd_t *pm;
	hugepd_t *hpdp = NULL;
	unsigned pshift = __ffs(sz);
	unsigned pdshift = PGDIR_SHIFT;

	addr &= ~(sz-1);

	pg = pgd_offset(mm, addr);
253 254

	if (pshift >= HUGEPD_PGD_SHIFT) {
255 256 257 258
		hpdp = (hugepd_t *)pg;
	} else {
		pdshift = PUD_SHIFT;
		pu = pud_alloc(mm, pg, addr);
259
		if (pshift >= HUGEPD_PUD_SHIFT) {
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
			hpdp = (hugepd_t *)pu;
		} else {
			pdshift = PMD_SHIFT;
			pm = pmd_alloc(mm, pu, addr);
			hpdp = (hugepd_t *)pm;
		}
	}

	if (!hpdp)
		return NULL;

	BUG_ON(!hugepd_none(*hpdp) && !hugepd_ok(*hpdp));

	if (hugepd_none(*hpdp) && __hugepte_alloc(mm, hpdp, addr, pdshift, pshift))
		return NULL;

276
	return hugepte_offset(*hpdp, addr, pdshift);
277
}
278
#endif
279

280
#ifdef CONFIG_PPC_FSL_BOOK3E
281
/* Build list of addresses of gigantic pages.  This function is used in early
282
 * boot before the buddy allocator is setup.
283
 */
B
Becky Bruce 已提交
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
void add_gpage(u64 addr, u64 page_size, unsigned long number_of_pages)
{
	unsigned int idx = shift_to_mmu_psize(__ffs(page_size));
	int i;

	if (addr == 0)
		return;

	gpage_freearray[idx].nr_gpages = number_of_pages;

	for (i = 0; i < number_of_pages; i++) {
		gpage_freearray[idx].gpage_list[i] = addr;
		addr += page_size;
	}
}

/*
 * Moves the gigantic page addresses from the temporary list to the
 * huge_boot_pages list.
 */
int alloc_bootmem_huge_page(struct hstate *hstate)
{
	struct huge_bootmem_page *m;
307
	int idx = shift_to_mmu_psize(huge_page_shift(hstate));
B
Becky Bruce 已提交
308 309 310 311 312 313 314 315 316 317
	int nr_gpages = gpage_freearray[idx].nr_gpages;

	if (nr_gpages == 0)
		return 0;

#ifdef CONFIG_HIGHMEM
	/*
	 * If gpages can be in highmem we can't use the trick of storing the
	 * data structure in the page; allocate space for this
	 */
318
	m = memblock_virt_alloc(sizeof(struct huge_bootmem_page), 0);
B
Becky Bruce 已提交
319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
	m->phys = gpage_freearray[idx].gpage_list[--nr_gpages];
#else
	m = phys_to_virt(gpage_freearray[idx].gpage_list[--nr_gpages]);
#endif

	list_add(&m->list, &huge_boot_pages);
	gpage_freearray[idx].nr_gpages = nr_gpages;
	gpage_freearray[idx].gpage_list[nr_gpages] = 0;
	m->hstate = hstate;

	return 1;
}
/*
 * Scan the command line hugepagesz= options for gigantic pages; store those in
 * a list that we use to allocate the memory once all options are parsed.
 */

unsigned long gpage_npages[MMU_PAGE_COUNT];

338 339
static int __init do_gpage_early_setup(char *param, char *val,
				       const char *unused)
B
Becky Bruce 已提交
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357
{
	static phys_addr_t size;
	unsigned long npages;

	/*
	 * The hugepagesz and hugepages cmdline options are interleaved.  We
	 * use the size variable to keep track of whether or not this was done
	 * properly and skip over instances where it is incorrect.  Other
	 * command-line parsing code will issue warnings, so we don't need to.
	 *
	 */
	if ((strcmp(param, "default_hugepagesz") == 0) ||
	    (strcmp(param, "hugepagesz") == 0)) {
		size = memparse(val, NULL);
	} else if (strcmp(param, "hugepages") == 0) {
		if (size != 0) {
			if (sscanf(val, "%lu", &npages) <= 0)
				npages = 0;
358 359 360 361 362 363 364
			if (npages > MAX_NUMBER_GPAGES) {
				pr_warn("MMU: %lu pages requested for page "
					"size %llu KB, limiting to "
					__stringify(MAX_NUMBER_GPAGES) "\n",
					npages, size / 1024);
				npages = MAX_NUMBER_GPAGES;
			}
B
Becky Bruce 已提交
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
			gpage_npages[shift_to_mmu_psize(__ffs(size))] = npages;
			size = 0;
		}
	}
	return 0;
}


/*
 * This function allocates physical space for pages that are larger than the
 * buddy allocator can handle.  We want to allocate these in highmem because
 * the amount of lowmem is limited.  This means that this function MUST be
 * called before lowmem_end_addr is set up in MMU_init() in order for the lmb
 * allocate to grab highmem.
 */
void __init reserve_hugetlb_gpages(void)
{
	static __initdata char cmdline[COMMAND_LINE_SIZE];
	phys_addr_t size, base;
	int i;

	strlcpy(cmdline, boot_command_line, COMMAND_LINE_SIZE);
387 388
	parse_args("hugetlb gpages", cmdline, NULL, 0, 0, 0,
			&do_gpage_early_setup);
B
Becky Bruce 已提交
389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408

	/*
	 * Walk gpage list in reverse, allocating larger page sizes first.
	 * Skip over unsupported sizes, or sizes that have 0 gpages allocated.
	 * When we reach the point in the list where pages are no longer
	 * considered gpages, we're done.
	 */
	for (i = MMU_PAGE_COUNT-1; i >= 0; i--) {
		if (mmu_psize_defs[i].shift == 0 || gpage_npages[i] == 0)
			continue;
		else if (mmu_psize_to_shift(i) < (MAX_ORDER + PAGE_SHIFT))
			break;

		size = (phys_addr_t)(1ULL << mmu_psize_to_shift(i));
		base = memblock_alloc_base(size * gpage_npages[i], size,
					   MEMBLOCK_ALLOC_ANYWHERE);
		add_gpage(base, size, gpage_npages[i]);
	}
}

409
#else /* !PPC_FSL_BOOK3E */
B
Becky Bruce 已提交
410 411

/* Build list of addresses of gigantic pages.  This function is used in early
412
 * boot before the buddy allocator is setup.
B
Becky Bruce 已提交
413 414
 */
void add_gpage(u64 addr, u64 page_size, unsigned long number_of_pages)
415 416 417 418 419 420 421 422 423 424 425
{
	if (!addr)
		return;
	while (number_of_pages > 0) {
		gpage_freearray[nr_gpages] = addr;
		nr_gpages++;
		number_of_pages--;
		addr += page_size;
	}
}

426
/* Moves the gigantic page addresses from the temporary list to the
427 428 429
 * huge_boot_pages list.
 */
int alloc_bootmem_huge_page(struct hstate *hstate)
430 431 432 433 434 435 436
{
	struct huge_bootmem_page *m;
	if (nr_gpages == 0)
		return 0;
	m = phys_to_virt(gpage_freearray[--nr_gpages]);
	gpage_freearray[nr_gpages] = 0;
	list_add(&m->list, &huge_boot_pages);
437
	m->hstate = hstate;
438 439
	return 1;
}
B
Becky Bruce 已提交
440
#endif
441

442 443 444 445 446
int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep)
{
	return 0;
}

447
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
#define HUGEPD_FREELIST_SIZE \
	((PAGE_SIZE - sizeof(struct hugepd_freelist)) / sizeof(pte_t))

struct hugepd_freelist {
	struct rcu_head	rcu;
	unsigned int index;
	void *ptes[0];
};

static DEFINE_PER_CPU(struct hugepd_freelist *, hugepd_freelist_cur);

static void hugepd_free_rcu_callback(struct rcu_head *head)
{
	struct hugepd_freelist *batch =
		container_of(head, struct hugepd_freelist, rcu);
	unsigned int i;

	for (i = 0; i < batch->index; i++)
		kmem_cache_free(hugepte_cache, batch->ptes[i]);

	free_page((unsigned long)batch);
}

static void hugepd_free(struct mmu_gather *tlb, void *hugepte)
{
	struct hugepd_freelist **batchp;

475
	batchp = this_cpu_ptr(&hugepd_freelist_cur);
B
Becky Bruce 已提交
476 477 478 479 480

	if (atomic_read(&tlb->mm->mm_users) < 2 ||
	    cpumask_equal(mm_cpumask(tlb->mm),
			  cpumask_of(smp_processor_id()))) {
		kmem_cache_free(hugepte_cache, hugepte);
481
        put_cpu_var(hugepd_freelist_cur);
B
Becky Bruce 已提交
482 483 484 485 486 487 488 489 490 491 492 493 494
		return;
	}

	if (*batchp == NULL) {
		*batchp = (struct hugepd_freelist *)__get_free_page(GFP_ATOMIC);
		(*batchp)->index = 0;
	}

	(*batchp)->ptes[(*batchp)->index++] = hugepte;
	if ((*batchp)->index == HUGEPD_FREELIST_SIZE) {
		call_rcu_sched(&(*batchp)->rcu, hugepd_free_rcu_callback);
		*batchp = NULL;
	}
495
	put_cpu_var(hugepd_freelist_cur);
B
Becky Bruce 已提交
496 497 498
}
#endif

499 500 501
static void free_hugepd_range(struct mmu_gather *tlb, hugepd_t *hpdp, int pdshift,
			      unsigned long start, unsigned long end,
			      unsigned long floor, unsigned long ceiling)
502 503
{
	pte_t *hugepte = hugepd_page(*hpdp);
B
Becky Bruce 已提交
504 505
	int i;

506
	unsigned long pdmask = ~((1UL << pdshift) - 1);
B
Becky Bruce 已提交
507 508
	unsigned int num_hugepd = 1;

509 510
#ifdef CONFIG_PPC_FSL_BOOK3E
	/* Note: On fsl the hpdp may be the first of several */
B
Becky Bruce 已提交
511
	num_hugepd = (1 << (hugepd_shift(*hpdp) - pdshift));
512 513
#else
	unsigned int shift = hugepd_shift(*hpdp);
B
Becky Bruce 已提交
514
#endif
515 516 517 518 519 520 521 522 523 524 525

	start &= pdmask;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= pdmask;
		if (! ceiling)
			return;
	}
	if (end - 1 > ceiling - 1)
		return;
526

B
Becky Bruce 已提交
527 528 529
	for (i = 0; i < num_hugepd; i++, hpdp++)
		hpdp->pd = 0;

530
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
531
	hugepd_free(tlb, hugepte);
532 533
#else
	pgtable_free_tlb(tlb, hugepte, pdshift - shift);
B
Becky Bruce 已提交
534
#endif
535 536 537 538
}

static void hugetlb_free_pmd_range(struct mmu_gather *tlb, pud_t *pud,
				   unsigned long addr, unsigned long end,
539
				   unsigned long floor, unsigned long ceiling)
540 541 542 543 544 545 546
{
	pmd_t *pmd;
	unsigned long next;
	unsigned long start;

	start = addr;
	do {
547
		pmd = pmd_offset(pud, addr);
548
		next = pmd_addr_end(addr, end);
549
		if (!is_hugepd(__hugepd(pmd_val(*pmd)))) {
550 551 552 553 554
			/*
			 * if it is not hugepd pointer, we should already find
			 * it cleared.
			 */
			WARN_ON(!pmd_none_or_clear_bad(pmd));
555
			continue;
556
		}
557 558 559 560 561 562 563 564 565
#ifdef CONFIG_PPC_FSL_BOOK3E
		/*
		 * Increment next by the size of the huge mapping since
		 * there may be more than one entry at this level for a
		 * single hugepage, but all of them point to
		 * the same kmem cache that holds the hugepte.
		 */
		next = addr + (1 << hugepd_shift(*(hugepd_t *)pmd));
#endif
566 567
		free_hugepd_range(tlb, (hugepd_t *)pmd, PMD_SHIFT,
				  addr, next, floor, ceiling);
568
	} while (addr = next, addr != end);
569 570 571 572 573 574 575 576

	start &= PUD_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PUD_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
577
	}
578 579
	if (end - 1 > ceiling - 1)
		return;
L
Linus Torvalds 已提交
580

581 582
	pmd = pmd_offset(pud, start);
	pud_clear(pud);
583
	pmd_free_tlb(tlb, pmd, start);
584
	mm_dec_nr_pmds(tlb->mm);
585 586 587 588 589 590 591 592 593 594 595 596
}

static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd,
				   unsigned long addr, unsigned long end,
				   unsigned long floor, unsigned long ceiling)
{
	pud_t *pud;
	unsigned long next;
	unsigned long start;

	start = addr;
	do {
597
		pud = pud_offset(pgd, addr);
598
		next = pud_addr_end(addr, end);
599
		if (!is_hugepd(__hugepd(pud_val(*pud)))) {
600 601
			if (pud_none_or_clear_bad(pud))
				continue;
602
			hugetlb_free_pmd_range(tlb, pud, addr, next, floor,
603
					       ceiling);
604
		} else {
605 606 607 608 609 610 611 612 613
#ifdef CONFIG_PPC_FSL_BOOK3E
			/*
			 * Increment next by the size of the huge mapping since
			 * there may be more than one entry at this level for a
			 * single hugepage, but all of them point to
			 * the same kmem cache that holds the hugepte.
			 */
			next = addr + (1 << hugepd_shift(*(hugepd_t *)pud));
#endif
614 615
			free_hugepd_range(tlb, (hugepd_t *)pud, PUD_SHIFT,
					  addr, next, floor, ceiling);
616
		}
617
	} while (addr = next, addr != end);
618 619 620 621 622 623 624 625 626 627 628 629 630 631

	start &= PGDIR_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PGDIR_MASK;
		if (!ceiling)
			return;
	}
	if (end - 1 > ceiling - 1)
		return;

	pud = pud_offset(pgd, start);
	pgd_clear(pgd);
632
	pud_free_tlb(tlb, pud, start);
633 634 635 636 637
}

/*
 * This function frees user-level page tables of a process.
 */
638
void hugetlb_free_pgd_range(struct mmu_gather *tlb,
639 640 641 642 643 644 645
			    unsigned long addr, unsigned long end,
			    unsigned long floor, unsigned long ceiling)
{
	pgd_t *pgd;
	unsigned long next;

	/*
646 647 648 649 650 651 652 653 654 655
	 * Because there are a number of different possible pagetable
	 * layouts for hugepage ranges, we limit knowledge of how
	 * things should be laid out to the allocation path
	 * (huge_pte_alloc(), above).  Everything else works out the
	 * structure as it goes from information in the hugepd
	 * pointers.  That means that we can't here use the
	 * optimization used in the normal page free_pgd_range(), of
	 * checking whether we're actually covering a large enough
	 * range to have to do anything at the top level of the walk
	 * instead of at the bottom.
656
	 *
657 658 659
	 * To make sense of this, you should probably go read the big
	 * block comment at the top of the normal free_pgd_range(),
	 * too.
660 661 662 663
	 */

	do {
		next = pgd_addr_end(addr, end);
B
Becky Bruce 已提交
664
		pgd = pgd_offset(tlb->mm, addr);
665
		if (!is_hugepd(__hugepd(pgd_val(*pgd)))) {
666 667 668 669
			if (pgd_none_or_clear_bad(pgd))
				continue;
			hugetlb_free_pud_range(tlb, pgd, addr, next, floor, ceiling);
		} else {
670
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
671 672
			/*
			 * Increment next by the size of the huge mapping since
673 674 675
			 * there may be more than one entry at the pgd level
			 * for a single hugepage, but all of them point to the
			 * same kmem cache that holds the hugepte.
B
Becky Bruce 已提交
676 677 678
			 */
			next = addr + (1 << hugepd_shift(*(hugepd_t *)pgd));
#endif
679 680
			free_hugepd_range(tlb, (hugepd_t *)pgd, PGDIR_SHIFT,
					  addr, next, floor, ceiling);
681
		}
B
Becky Bruce 已提交
682
	} while (addr = next, addr != end);
L
Linus Torvalds 已提交
683 684
}

685 686 687 688
/*
 * We are holding mmap_sem, so a parallel huge page collapse cannot run.
 * To prevent hugepage split, disable irq.
 */
L
Linus Torvalds 已提交
689 690 691
struct page *
follow_huge_addr(struct mm_struct *mm, unsigned long address, int write)
{
692
	pte_t *ptep, pte;
693
	unsigned shift;
694
	unsigned long mask, flags;
695 696 697 698 699 700 701
	struct page *page = ERR_PTR(-EINVAL);

	local_irq_save(flags);
	ptep = find_linux_pte_or_hugepte(mm->pgd, address, &shift);
	if (!ptep)
		goto no_page;
	pte = READ_ONCE(*ptep);
702
	/*
703
	 * Verify it is a huge page else bail.
704 705 706
	 * Transparent hugepages are handled by generic code. We can skip them
	 * here.
	 */
707 708
	if (!shift || pmd_trans_huge(__pmd(pte_val(pte))))
		goto no_page;
L
Linus Torvalds 已提交
709

710 711 712
	if (!pte_present(pte)) {
		page = NULL;
		goto no_page;
713
	}
714
	mask = (1UL << shift) - 1;
715
	page = pte_page(pte);
716 717
	if (page)
		page += (address & mask) / PAGE_SIZE;
L
Linus Torvalds 已提交
718

719
no_page:
720
	local_irq_restore(flags);
L
Linus Torvalds 已提交
721 722 723 724 725 726 727 728 729 730 731
	return page;
}

struct page *
follow_huge_pmd(struct mm_struct *mm, unsigned long address,
		pmd_t *pmd, int write)
{
	BUG();
	return NULL;
}

732 733 734 735 736 737 738 739
struct page *
follow_huge_pud(struct mm_struct *mm, unsigned long address,
		pud_t *pud, int write)
{
	BUG();
	return NULL;
}

D
David Gibson 已提交
740 741 742 743 744 745 746
static unsigned long hugepte_addr_end(unsigned long addr, unsigned long end,
				      unsigned long sz)
{
	unsigned long __boundary = (addr + sz) & ~(sz-1);
	return (__boundary - 1 < end - 1) ? __boundary : end;
}

747 748
int gup_huge_pd(hugepd_t hugepd, unsigned long addr, unsigned pdshift,
		unsigned long end, int write, struct page **pages, int *nr)
749 750
{
	pte_t *ptep;
751
	unsigned long sz = 1UL << hugepd_shift(hugepd);
D
David Gibson 已提交
752
	unsigned long next;
753 754 755

	ptep = hugepte_offset(hugepd, addr, pdshift);
	do {
D
David Gibson 已提交
756
		next = hugepte_addr_end(addr, end, sz);
757 758
		if (!gup_hugepte(ptep, sz, addr, end, write, pages, nr))
			return 0;
D
David Gibson 已提交
759
	} while (ptep++, addr = next, addr != end);
760 761 762

	return 1;
}
L
Linus Torvalds 已提交
763

764
#ifdef CONFIG_PPC_MM_SLICES
L
Linus Torvalds 已提交
765 766 767 768
unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
					unsigned long len, unsigned long pgoff,
					unsigned long flags)
{
769 770
	struct hstate *hstate = hstate_file(file);
	int mmu_psize = shift_to_mmu_psize(huge_page_shift(hstate));
771

772
	return slice_get_unmapped_area(addr, len, flags, mmu_psize, 1);
L
Linus Torvalds 已提交
773
}
774
#endif
L
Linus Torvalds 已提交
775

776 777
unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
{
778
#ifdef CONFIG_PPC_MM_SLICES
779 780 781
	unsigned int psize = get_slice_psize(vma->vm_mm, vma->vm_start);

	return 1UL << mmu_psize_to_shift(psize);
B
Becky Bruce 已提交
782 783 784 785 786 787 788 789 790 791 792 793 794
#else
	if (!is_vm_hugetlb_page(vma))
		return PAGE_SIZE;

	return huge_page_size(hstate_vma(vma));
#endif
}

static inline bool is_power_of_4(unsigned long x)
{
	if (is_power_of_2(x))
		return (__ilog2(x) % 2) ? false : true;
	return false;
795 796
}

797
static int __init add_huge_page_size(unsigned long long size)
798
{
799 800
	int shift = __ffs(size);
	int mmu_psize;
801

802
	/* Check that it is a page size supported by the hardware and
803
	 * that it fits within pagetable and slice limits. */
B
Becky Bruce 已提交
804 805 806 807
#ifdef CONFIG_PPC_FSL_BOOK3E
	if ((size < PAGE_SIZE) || !is_power_of_4(size))
		return -EINVAL;
#else
808 809 810
	if (!is_power_of_2(size)
	    || (shift > SLICE_HIGH_SHIFT) || (shift <= PAGE_SHIFT))
		return -EINVAL;
B
Becky Bruce 已提交
811
#endif
812

813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832
	if ((mmu_psize = shift_to_mmu_psize(shift)) < 0)
		return -EINVAL;

#ifdef CONFIG_SPU_FS_64K_LS
	/* Disable support for 64K huge pages when 64K SPU local store
	 * support is enabled as the current implementation conflicts.
	 */
	if (shift == PAGE_SHIFT_64K)
		return -EINVAL;
#endif /* CONFIG_SPU_FS_64K_LS */

	BUG_ON(mmu_psize_defs[mmu_psize].shift != shift);

	/* Return if huge page size has already been setup */
	if (size_to_hstate(size))
		return 0;

	hugetlb_add_hstate(shift - PAGE_SHIFT);

	return 0;
833 834 835 836 837 838 839 840
}

static int __init hugepage_setup_sz(char *str)
{
	unsigned long long size;

	size = memparse(str, &str);

841
	if (add_huge_page_size(size) != 0)
842 843 844 845 846 847
		printk(KERN_WARNING "Invalid huge page size specified(%llu)\n", size);

	return 1;
}
__setup("hugepagesz=", hugepage_setup_sz);

848
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887
struct kmem_cache *hugepte_cache;
static int __init hugetlbpage_init(void)
{
	int psize;

	for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
		unsigned shift;

		if (!mmu_psize_defs[psize].shift)
			continue;

		shift = mmu_psize_to_shift(psize);

		/* Don't treat normal page sizes as huge... */
		if (shift != PAGE_SHIFT)
			if (add_huge_page_size(1ULL << shift) < 0)
				continue;
	}

	/*
	 * Create a kmem cache for hugeptes.  The bottom bits in the pte have
	 * size information encoded in them, so align them to allow this
	 */
	hugepte_cache =  kmem_cache_create("hugepte-cache", sizeof(pte_t),
					   HUGEPD_SHIFT_MASK + 1, 0, NULL);
	if (hugepte_cache == NULL)
		panic("%s: Unable to create kmem cache for hugeptes\n",
		      __func__);

	/* Default hpage size = 4M */
	if (mmu_psize_defs[MMU_PAGE_4M].shift)
		HPAGE_SHIFT = mmu_psize_defs[MMU_PAGE_4M].shift;
	else
		panic("%s: Unable to set default huge page size\n", __func__);


	return 0;
}
#else
888 889
static int __init hugetlbpage_init(void)
{
890
	int psize;
891

892
	if (!mmu_has_feature(MMU_FTR_16M_PAGE))
893
		return -ENODEV;
894

895 896 897
	for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
		unsigned shift;
		unsigned pdshift;
898

899 900
		if (!mmu_psize_defs[psize].shift)
			continue;
901

902 903 904 905 906 907 908 909 910 911 912
		shift = mmu_psize_to_shift(psize);

		if (add_huge_page_size(1ULL << shift) < 0)
			continue;

		if (shift < PMD_SHIFT)
			pdshift = PMD_SHIFT;
		else if (shift < PUD_SHIFT)
			pdshift = PUD_SHIFT;
		else
			pdshift = PGDIR_SHIFT;
913 914 915 916 917 918 919 920 921 922
		/*
		 * if we have pdshift and shift value same, we don't
		 * use pgt cache for hugepd.
		 */
		if (pdshift != shift) {
			pgtable_cache_add(pdshift - shift, NULL);
			if (!PGT_CACHE(pdshift - shift))
				panic("hugetlbpage_init(): could not create "
				      "pgtable cache for %d bit pagesize\n", shift);
		}
923
	}
924

925 926 927 928 929 930 931 932
	/* Set default large page size. Currently, we pick 16M or 1M
	 * depending on what is available
	 */
	if (mmu_psize_defs[MMU_PAGE_16M].shift)
		HPAGE_SHIFT = mmu_psize_defs[MMU_PAGE_16M].shift;
	else if (mmu_psize_defs[MMU_PAGE_1M].shift)
		HPAGE_SHIFT = mmu_psize_defs[MMU_PAGE_1M].shift;

933 934
	return 0;
}
B
Becky Bruce 已提交
935
#endif
936
module_init(hugetlbpage_init);
937 938 939 940

void flush_dcache_icache_hugepage(struct page *page)
{
	int i;
B
Becky Bruce 已提交
941
	void *start;
942 943 944

	BUG_ON(!PageCompound(page));

B
Becky Bruce 已提交
945 946 947 948
	for (i = 0; i < (1UL << compound_order(page)); i++) {
		if (!PageHighMem(page)) {
			__flush_dcache_icache(page_address(page+i));
		} else {
949
			start = kmap_atomic(page+i);
B
Becky Bruce 已提交
950
			__flush_dcache_icache(start);
951
			kunmap_atomic(start);
B
Becky Bruce 已提交
952 953
		}
	}
954
}
955 956 957 958 959 960 961 962 963

#endif /* CONFIG_HUGETLB_PAGE */

/*
 * We have 4 cases for pgds and pmds:
 * (1) invalid (all zeroes)
 * (2) pointer to next table, as normal; bottom 6 bits == 0
 * (3) leaf pte for huge page, bottom two bits != 00
 * (4) hugepd pointer, bottom two bits == 00, next 4 bits indicate size of table
964 965 966
 *
 * So long as we atomically load page table pointers we are safe against teardown,
 * we can follow the address down to the the page and take a ref on it.
967 968
 * This function need to be called with interrupts disabled. We use this variant
 * when we have MSR[EE] = 0 but the paca->soft_enabled = 1
969
 */
970

971 972
pte_t *__find_linux_pte_or_hugepte(pgd_t *pgdir, unsigned long ea,
				   unsigned *shift)
973
{
974 975 976
	pgd_t pgd, *pgdp;
	pud_t pud, *pudp;
	pmd_t pmd, *pmdp;
977 978 979 980 981 982 983
	pte_t *ret_pte;
	hugepd_t *hpdp = NULL;
	unsigned pdshift = PGDIR_SHIFT;

	if (shift)
		*shift = 0;

984
	pgdp = pgdir + pgd_index(ea);
985
	pgd  = READ_ONCE(*pgdp);
986
	/*
987 988 989 990
	 * Always operate on the local stack value. This make sure the
	 * value don't get updated by a parallel THP split/collapse,
	 * page fault or a page unmap. The return pte_t * is still not
	 * stable. So should be checked there for above conditions.
991
	 */
992
	if (pgd_none(pgd))
993
		return NULL;
994 995
	else if (pgd_huge(pgd)) {
		ret_pte = (pte_t *) pgdp;
996
		goto out;
997
	} else if (is_hugepd(__hugepd(pgd_val(pgd))))
998
		hpdp = (hugepd_t *)&pgd;
999
	else {
1000 1001 1002 1003 1004
		/*
		 * Even if we end up with an unmap, the pgtable will not
		 * be freed, because we do an rcu free and here we are
		 * irq disabled
		 */
1005
		pdshift = PUD_SHIFT;
1006
		pudp = pud_offset(&pgd, ea);
1007
		pud  = READ_ONCE(*pudp);
1008

1009
		if (pud_none(pud))
1010
			return NULL;
1011 1012
		else if (pud_huge(pud)) {
			ret_pte = (pte_t *) pudp;
1013
			goto out;
1014
		} else if (is_hugepd(__hugepd(pud_val(pud))))
1015
			hpdp = (hugepd_t *)&pud;
1016
		else {
1017
			pdshift = PMD_SHIFT;
1018
			pmdp = pmd_offset(&pud, ea);
1019
			pmd  = READ_ONCE(*pmdp);
1020 1021 1022 1023
			/*
			 * A hugepage collapse is captured by pmd_none, because
			 * it mark the pmd none and do a hpte invalidate.
			 *
1024 1025 1026
			 * We don't worry about pmd_trans_splitting here, The
			 * caller if it needs to handle the splitting case
			 * should check for that.
1027
			 */
1028
			if (pmd_none(pmd))
1029
				return NULL;
1030

1031 1032
			if (pmd_huge(pmd) || pmd_large(pmd)) {
				ret_pte = (pte_t *) pmdp;
1033
				goto out;
1034
			} else if (is_hugepd(__hugepd(pmd_val(pmd))))
1035
				hpdp = (hugepd_t *)&pmd;
1036
			else
1037
				return pte_offset_kernel(&pmd, ea);
1038 1039 1040 1041 1042
		}
	}
	if (!hpdp)
		return NULL;

1043
	ret_pte = hugepte_offset(*hpdp, ea, pdshift);
1044 1045 1046 1047 1048 1049
	pdshift = hugepd_shift(*hpdp);
out:
	if (shift)
		*shift = pdshift;
	return ret_pte;
}
1050
EXPORT_SYMBOL_GPL(__find_linux_pte_or_hugepte);
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064

int gup_hugepte(pte_t *ptep, unsigned long sz, unsigned long addr,
		unsigned long end, int write, struct page **pages, int *nr)
{
	unsigned long mask;
	unsigned long pte_end;
	struct page *head, *page, *tail;
	pte_t pte;
	int refs;

	pte_end = (addr + sz) & ~(sz-1);
	if (pte_end < end)
		end = pte_end;

1065
	pte = READ_ONCE(*ptep);
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
	mask = _PAGE_PRESENT | _PAGE_USER;
	if (write)
		mask |= _PAGE_RW;

	if ((pte_val(pte) & mask) != mask)
		return 0;

	/* hugepages are never "special" */
	VM_BUG_ON(!pfn_valid(pte_pfn(pte)));

	refs = 0;
	head = pte_page(pte);

	page = head + ((addr & (sz-1)) >> PAGE_SHIFT);
	tail = page;
	do {
		VM_BUG_ON(compound_head(page) != head);
		pages[*nr] = page;
		(*nr)++;
		page++;
		refs++;
	} while (addr += PAGE_SIZE, addr != end);

	if (!page_cache_add_speculative(head, refs)) {
		*nr -= refs;
		return 0;
	}

	if (unlikely(pte_val(pte) != pte_val(*ptep))) {
		/* Could be optimized better */
		*nr -= refs;
		while (refs--)
			put_page(head);
		return 0;
	}

	/*
	 * Any tail page need their mapcount reference taken before we
	 * return.
	 */
	while (refs--) {
		if (PageTail(tail))
			get_huge_page_tail(tail);
		tail++;
	}

	return 1;
}