gup.c 81.1 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-only
2 3 4 5 6 7
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/spinlock.h>

#include <linux/mm.h>
8
#include <linux/memremap.h>
9 10 11 12 13
#include <linux/pagemap.h>
#include <linux/rmap.h>
#include <linux/swap.h>
#include <linux/swapops.h>

14
#include <linux/sched/signal.h>
15
#include <linux/rwsem.h>
16
#include <linux/hugetlb.h>
17 18 19
#include <linux/migrate.h>
#include <linux/mm_inline.h>
#include <linux/sched/mm.h>
20

21
#include <asm/mmu_context.h>
22
#include <asm/pgtable.h>
23
#include <asm/tlbflush.h>
24

25 26
#include "internal.h"

27 28 29 30 31
struct follow_page_context {
	struct dev_pagemap *pgmap;
	unsigned int page_mask;
};

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
static void hpage_pincount_add(struct page *page, int refs)
{
	VM_BUG_ON_PAGE(!hpage_pincount_available(page), page);
	VM_BUG_ON_PAGE(page != compound_head(page), page);

	atomic_add(refs, compound_pincount_ptr(page));
}

static void hpage_pincount_sub(struct page *page, int refs)
{
	VM_BUG_ON_PAGE(!hpage_pincount_available(page), page);
	VM_BUG_ON_PAGE(page != compound_head(page), page);

	atomic_sub(refs, compound_pincount_ptr(page));
}

48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
/*
 * Return the compound head page with ref appropriately incremented,
 * or NULL if that failed.
 */
static inline struct page *try_get_compound_head(struct page *page, int refs)
{
	struct page *head = compound_head(page);

	if (WARN_ON_ONCE(page_ref_count(head) < 0))
		return NULL;
	if (unlikely(!page_cache_add_speculative(head, refs)))
		return NULL;
	return head;
}

John Hubbard's avatar
John Hubbard committed
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
/*
 * try_grab_compound_head() - attempt to elevate a page's refcount, by a
 * flags-dependent amount.
 *
 * "grab" names in this file mean, "look at flags to decide whether to use
 * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount.
 *
 * Either FOLL_PIN or FOLL_GET (or neither) must be set, but not both at the
 * same time. (That's true throughout the get_user_pages*() and
 * pin_user_pages*() APIs.) Cases:
 *
 *    FOLL_GET: page's refcount will be incremented by 1.
 *    FOLL_PIN: page's refcount will be incremented by GUP_PIN_COUNTING_BIAS.
 *
 * Return: head page (with refcount appropriately incremented) for success, or
 * NULL upon failure. If neither FOLL_GET nor FOLL_PIN was set, that's
 * considered failure, and furthermore, a likely bug in the caller, so a warning
 * is also emitted.
 */
static __maybe_unused struct page *try_grab_compound_head(struct page *page,
							  int refs,
							  unsigned int flags)
{
	if (flags & FOLL_GET)
		return try_get_compound_head(page, refs);
	else if (flags & FOLL_PIN) {
89 90
		int orig_refs = refs;

91 92 93 94 95 96 97 98
		/*
		 * Can't do FOLL_LONGTERM + FOLL_PIN with CMA in the gup fast
		 * path, so fail and let the caller fall back to the slow path.
		 */
		if (unlikely(flags & FOLL_LONGTERM) &&
				is_migrate_cma_page(page))
			return NULL;

99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
		/*
		 * When pinning a compound page of order > 1 (which is what
		 * hpage_pincount_available() checks for), use an exact count to
		 * track it, via hpage_pincount_add/_sub().
		 *
		 * However, be sure to *also* increment the normal page refcount
		 * field at least once, so that the page really is pinned.
		 */
		if (!hpage_pincount_available(page))
			refs *= GUP_PIN_COUNTING_BIAS;

		page = try_get_compound_head(page, refs);
		if (!page)
			return NULL;

		if (hpage_pincount_available(page))
			hpage_pincount_add(page, refs);

117 118 119
		mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_ACQUIRED,
				    orig_refs);

120
		return page;
John Hubbard's avatar
John Hubbard committed
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
	}

	WARN_ON_ONCE(1);
	return NULL;
}

/**
 * try_grab_page() - elevate a page's refcount by a flag-dependent amount
 *
 * This might not do anything at all, depending on the flags argument.
 *
 * "grab" names in this file mean, "look at flags to decide whether to use
 * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount.
 *
 * @page:    pointer to page to be grabbed
 * @flags:   gup flags: these are the FOLL_* flag values.
 *
 * Either FOLL_PIN or FOLL_GET (or neither) may be set, but not both at the same
 * time. Cases:
 *
 *    FOLL_GET: page's refcount will be incremented by 1.
 *    FOLL_PIN: page's refcount will be incremented by GUP_PIN_COUNTING_BIAS.
 *
 * Return: true for success, or if no action was required (if neither FOLL_PIN
 * nor FOLL_GET was set, nothing is done). False for failure: FOLL_GET or
 * FOLL_PIN was set, but the page could not be grabbed.
 */
bool __must_check try_grab_page(struct page *page, unsigned int flags)
{
	WARN_ON_ONCE((flags & (FOLL_GET | FOLL_PIN)) == (FOLL_GET | FOLL_PIN));

	if (flags & FOLL_GET)
		return try_get_page(page);
	else if (flags & FOLL_PIN) {
155 156
		int refs = 1;

John Hubbard's avatar
John Hubbard committed
157 158 159 160 161
		page = compound_head(page);

		if (WARN_ON_ONCE(page_ref_count(page) <= 0))
			return false;

162 163 164 165 166 167 168 169 170 171 172 173
		if (hpage_pincount_available(page))
			hpage_pincount_add(page, 1);
		else
			refs = GUP_PIN_COUNTING_BIAS;

		/*
		 * Similar to try_grab_compound_head(): even if using the
		 * hpage_pincount_add/_sub() routines, be sure to
		 * *also* increment the normal page refcount field at least
		 * once, so that the page really is pinned.
		 */
		page_ref_add(page, refs);
174 175

		mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_ACQUIRED, 1);
John Hubbard's avatar
John Hubbard committed
176 177 178 179 180 181 182 183
	}

	return true;
}

#ifdef CONFIG_DEV_PAGEMAP_OPS
static bool __unpin_devmap_managed_user_page(struct page *page)
{
184
	int count, refs = 1;
John Hubbard's avatar
John Hubbard committed
185 186 187 188

	if (!page_is_devmap_managed(page))
		return false;

189 190 191 192 193 194
	if (hpage_pincount_available(page))
		hpage_pincount_sub(page, 1);
	else
		refs = GUP_PIN_COUNTING_BIAS;

	count = page_ref_sub_return(page, refs);
John Hubbard's avatar
John Hubbard committed
195

196
	mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_RELEASED, 1);
John Hubbard's avatar
John Hubbard committed
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
	/*
	 * devmap page refcounts are 1-based, rather than 0-based: if
	 * refcount is 1, then the page is free and the refcount is
	 * stable because nobody holds a reference on the page.
	 */
	if (count == 1)
		free_devmap_managed_page(page);
	else if (!count)
		__put_page(page);

	return true;
}
#else
static bool __unpin_devmap_managed_user_page(struct page *page)
{
	return false;
}
#endif /* CONFIG_DEV_PAGEMAP_OPS */

/**
 * unpin_user_page() - release a dma-pinned page
 * @page:            pointer to page to be released
 *
 * Pages that were pinned via pin_user_pages*() must be released via either
 * unpin_user_page(), or one of the unpin_user_pages*() routines. This is so
 * that such pages can be separately tracked and uniquely handled. In
 * particular, interactions with RDMA and filesystems need special handling.
 */
void unpin_user_page(struct page *page)
{
227 228
	int refs = 1;

John Hubbard's avatar
John Hubbard committed
229 230 231 232 233 234 235 236 237 238 239
	page = compound_head(page);

	/*
	 * For devmap managed pages we need to catch refcount transition from
	 * GUP_PIN_COUNTING_BIAS to 1, when refcount reach one it means the
	 * page is free and we need to inform the device driver through
	 * callback. See include/linux/memremap.h and HMM for details.
	 */
	if (__unpin_devmap_managed_user_page(page))
		return;

240 241 242 243 244 245
	if (hpage_pincount_available(page))
		hpage_pincount_sub(page, 1);
	else
		refs = GUP_PIN_COUNTING_BIAS;

	if (page_ref_sub_and_test(page, refs))
John Hubbard's avatar
John Hubbard committed
246
		__put_page(page);
247 248

	mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_RELEASED, 1);
John Hubbard's avatar
John Hubbard committed
249 250 251
}
EXPORT_SYMBOL(unpin_user_page);

252
/**
253
 * unpin_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages
254
 * @pages:  array of pages to be maybe marked dirty, and definitely released.
255
 * @npages: number of pages in the @pages array.
256
 * @make_dirty: whether to mark the pages dirty
257 258 259 260 261
 *
 * "gup-pinned page" refers to a page that has had one of the get_user_pages()
 * variants called on that page.
 *
 * For each page in the @pages array, make that page (or its head page, if a
262
 * compound page) dirty, if @make_dirty is true, and if the page was previously
263 264
 * listed as clean. In any case, releases all pages using unpin_user_page(),
 * possibly via unpin_user_pages(), for the non-dirty case.
265
 *
266
 * Please see the unpin_user_page() documentation for details.
267
 *
268 269 270
 * set_page_dirty_lock() is used internally. If instead, set_page_dirty() is
 * required, then the caller should a) verify that this is really correct,
 * because _lock() is usually required, and b) hand code it:
271
 * set_page_dirty_lock(), unpin_user_page().
272 273
 *
 */
274 275
void unpin_user_pages_dirty_lock(struct page **pages, unsigned long npages,
				 bool make_dirty)
276
{
277
	unsigned long index;
278

279 280 281 282 283 284 285
	/*
	 * TODO: this can be optimized for huge pages: if a series of pages is
	 * physically contiguous and part of the same compound page, then a
	 * single operation to the head page should suffice.
	 */

	if (!make_dirty) {
286
		unpin_user_pages(pages, npages);
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
		return;
	}

	for (index = 0; index < npages; index++) {
		struct page *page = compound_head(pages[index]);
		/*
		 * Checking PageDirty at this point may race with
		 * clear_page_dirty_for_io(), but that's OK. Two key
		 * cases:
		 *
		 * 1) This code sees the page as already dirty, so it
		 * skips the call to set_page_dirty(). That could happen
		 * because clear_page_dirty_for_io() called
		 * page_mkclean(), followed by set_page_dirty().
		 * However, now the page is going to get written back,
		 * which meets the original intention of setting it
		 * dirty, so all is well: clear_page_dirty_for_io() goes
		 * on to call TestClearPageDirty(), and write the page
		 * back.
		 *
		 * 2) This code sees the page as clean, so it calls
		 * set_page_dirty(). The page stays dirty, despite being
		 * written back, so it gets written back again in the
		 * next writeback cycle. This is harmless.
		 */
		if (!PageDirty(page))
			set_page_dirty_lock(page);
314
		unpin_user_page(page);
315
	}
316
}
317
EXPORT_SYMBOL(unpin_user_pages_dirty_lock);
318 319

/**
320
 * unpin_user_pages() - release an array of gup-pinned pages.
321 322 323
 * @pages:  array of pages to be marked dirty and released.
 * @npages: number of pages in the @pages array.
 *
324
 * For each page in the @pages array, release the page using unpin_user_page().
325
 *
326
 * Please see the unpin_user_page() documentation for details.
327
 */
328
void unpin_user_pages(struct page **pages, unsigned long npages)
329 330 331 332 333 334 335 336 337
{
	unsigned long index;

	/*
	 * TODO: this can be optimized for huge pages: if a series of pages is
	 * physically contiguous and part of the same compound page, then a
	 * single operation to the head page should suffice.
	 */
	for (index = 0; index < npages; index++)
338
		unpin_user_page(pages[index]);
339
}
340
EXPORT_SYMBOL(unpin_user_pages);
341

342
#ifdef CONFIG_MMU
343 344
static struct page *no_page_table(struct vm_area_struct *vma,
		unsigned int flags)
345
{
346 347 348 349 350 351 352 353 354 355 356 357
	/*
	 * When core dumping an enormous anonymous area that nobody
	 * has touched so far, we don't want to allocate unnecessary pages or
	 * page tables.  Return error instead of NULL to skip handle_mm_fault,
	 * then get_dump_page() will return NULL to leave a hole in the dump.
	 * But we can only make this optimization where a hole would surely
	 * be zero-filled if handle_mm_fault() actually did handle it.
	 */
	if ((flags & FOLL_DUMP) && (!vma->vm_ops || !vma->vm_ops->fault))
		return ERR_PTR(-EFAULT);
	return NULL;
}
358

359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
static int follow_pfn_pte(struct vm_area_struct *vma, unsigned long address,
		pte_t *pte, unsigned int flags)
{
	/* No page to get reference */
	if (flags & FOLL_GET)
		return -EFAULT;

	if (flags & FOLL_TOUCH) {
		pte_t entry = *pte;

		if (flags & FOLL_WRITE)
			entry = pte_mkdirty(entry);
		entry = pte_mkyoung(entry);

		if (!pte_same(*pte, entry)) {
			set_pte_at(vma->vm_mm, address, pte, entry);
			update_mmu_cache(vma, address, pte);
		}
	}

	/* Proper page table entry exists, but no corresponding struct page */
	return -EEXIST;
}

383 384 385 386 387 388
/*
 * FOLL_FORCE can write to even unwritable pte's, but only
 * after we've gone through a COW cycle and they are dirty.
 */
static inline bool can_follow_write_pte(pte_t pte, unsigned int flags)
{
389
	return pte_write(pte) ||
390 391 392
		((flags & FOLL_FORCE) && (flags & FOLL_COW) && pte_dirty(pte));
}

393
static struct page *follow_page_pte(struct vm_area_struct *vma,
394 395
		unsigned long address, pmd_t *pmd, unsigned int flags,
		struct dev_pagemap **pgmap)
396 397 398 399 400
{
	struct mm_struct *mm = vma->vm_mm;
	struct page *page;
	spinlock_t *ptl;
	pte_t *ptep, pte;
401
	int ret;
402

403 404 405 406
	/* FOLL_GET and FOLL_PIN are mutually exclusive. */
	if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) ==
			 (FOLL_PIN | FOLL_GET)))
		return ERR_PTR(-EINVAL);
407
retry:
408
	if (unlikely(pmd_bad(*pmd)))
409
		return no_page_table(vma, flags);
410 411 412 413 414 415 416 417 418 419 420 421

	ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
	pte = *ptep;
	if (!pte_present(pte)) {
		swp_entry_t entry;
		/*
		 * KSM's break_ksm() relies upon recognizing a ksm page
		 * even while it is being migrated, so for that case we
		 * need migration_entry_wait().
		 */
		if (likely(!(flags & FOLL_MIGRATION)))
			goto no_page;
422
		if (pte_none(pte))
423 424 425 426 427 428
			goto no_page;
		entry = pte_to_swp_entry(pte);
		if (!is_migration_entry(entry))
			goto no_page;
		pte_unmap_unlock(ptep, ptl);
		migration_entry_wait(mm, pmd, address);
429
		goto retry;
430
	}
431
	if ((flags & FOLL_NUMA) && pte_protnone(pte))
432
		goto no_page;
433
	if ((flags & FOLL_WRITE) && !can_follow_write_pte(pte, flags)) {
434 435 436
		pte_unmap_unlock(ptep, ptl);
		return NULL;
	}
437 438

	page = vm_normal_page(vma, address, pte);
John Hubbard's avatar
John Hubbard committed
439
	if (!page && pte_devmap(pte) && (flags & (FOLL_GET | FOLL_PIN))) {
440
		/*
John Hubbard's avatar
John Hubbard committed
441 442 443
		 * Only return device mapping pages in the FOLL_GET or FOLL_PIN
		 * case since they are only valid while holding the pgmap
		 * reference.
444
		 */
445 446
		*pgmap = get_dev_pagemap(pte_pfn(pte), *pgmap);
		if (*pgmap)
447 448 449 450
			page = pte_page(pte);
		else
			goto no_page;
	} else if (unlikely(!page)) {
451 452 453 454 455 456 457 458 459 460 461 462 463
		if (flags & FOLL_DUMP) {
			/* Avoid special (like zero) pages in core dumps */
			page = ERR_PTR(-EFAULT);
			goto out;
		}

		if (is_zero_pfn(pte_pfn(pte))) {
			page = pte_page(pte);
		} else {
			ret = follow_pfn_pte(vma, address, ptep, flags);
			page = ERR_PTR(ret);
			goto out;
		}
464 465
	}

466 467 468 469 470 471 472 473 474 475 476 477
	if (flags & FOLL_SPLIT && PageTransCompound(page)) {
		get_page(page);
		pte_unmap_unlock(ptep, ptl);
		lock_page(page);
		ret = split_huge_page(page);
		unlock_page(page);
		put_page(page);
		if (ret)
			return ERR_PTR(ret);
		goto retry;
	}

John Hubbard's avatar
John Hubbard committed
478 479 480 481
	/* try_grab_page() does nothing unless FOLL_GET or FOLL_PIN is set. */
	if (unlikely(!try_grab_page(page, flags))) {
		page = ERR_PTR(-ENOMEM);
		goto out;
482
	}
483 484 485 486 487 488 489 490 491 492 493 494 495
	/*
	 * We need to make the page accessible if and only if we are going
	 * to access its content (the FOLL_PIN case).  Please see
	 * Documentation/core-api/pin_user_pages.rst for details.
	 */
	if (flags & FOLL_PIN) {
		ret = arch_make_page_accessible(page);
		if (ret) {
			unpin_user_page(page);
			page = ERR_PTR(ret);
			goto out;
		}
	}
496 497 498 499 500 501 502 503 504 505 506
	if (flags & FOLL_TOUCH) {
		if ((flags & FOLL_WRITE) &&
		    !pte_dirty(pte) && !PageDirty(page))
			set_page_dirty(page);
		/*
		 * pte_mkyoung() would be more correct here, but atomic care
		 * is needed to avoid losing the dirty bit: it is easier to use
		 * mark_page_accessed().
		 */
		mark_page_accessed(page);
	}
Eric B Munson's avatar
Eric B Munson committed
507
	if ((flags & FOLL_MLOCK) && (vma->vm_flags & VM_LOCKED)) {
508 509 510 511
		/* Do not mlock pte-mapped THP */
		if (PageTransCompound(page))
			goto out;

512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
		/*
		 * The preliminary mapping check is mainly to avoid the
		 * pointless overhead of lock_page on the ZERO_PAGE
		 * which might bounce very badly if there is contention.
		 *
		 * If the page is already locked, we don't need to
		 * handle it now - vmscan will handle it later if and
		 * when it attempts to reclaim the page.
		 */
		if (page->mapping && trylock_page(page)) {
			lru_add_drain();  /* push cached pages to LRU */
			/*
			 * Because we lock page here, and migration is
			 * blocked by the pte's page reference, and we
			 * know the page is still mapped, we don't even
			 * need to check for file-cache page truncation.
			 */
			mlock_vma_page(page);
			unlock_page(page);
		}
	}
533
out:
534 535 536 537 538
	pte_unmap_unlock(ptep, ptl);
	return page;
no_page:
	pte_unmap_unlock(ptep, ptl);
	if (!pte_none(pte))
539 540 541 542
		return NULL;
	return no_page_table(vma, flags);
}

543 544
static struct page *follow_pmd_mask(struct vm_area_struct *vma,
				    unsigned long address, pud_t *pudp,
545 546
				    unsigned int flags,
				    struct follow_page_context *ctx)
547
{
548
	pmd_t *pmd, pmdval;
549 550 551 552
	spinlock_t *ptl;
	struct page *page;
	struct mm_struct *mm = vma->vm_mm;

553
	pmd = pmd_offset(pudp, address);
554 555 556 557 558 559
	/*
	 * The READ_ONCE() will stabilize the pmdval in a register or
	 * on the stack so that it will stop changing under the code.
	 */
	pmdval = READ_ONCE(*pmd);
	if (pmd_none(pmdval))
560
		return no_page_table(vma, flags);
561
	if (pmd_huge(pmdval) && is_vm_hugetlb_page(vma)) {
562 563 564 565
		page = follow_huge_pmd(mm, address, pmd, flags);
		if (page)
			return page;
		return no_page_table(vma, flags);
566
	}
567
	if (is_hugepd(__hugepd(pmd_val(pmdval)))) {
568
		page = follow_huge_pd(vma, address,
569
				      __hugepd(pmd_val(pmdval)), flags,
570 571 572 573 574
				      PMD_SHIFT);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
575
retry:
576
	if (!pmd_present(pmdval)) {
577 578 579
		if (likely(!(flags & FOLL_MIGRATION)))
			return no_page_table(vma, flags);
		VM_BUG_ON(thp_migration_supported() &&
580 581
				  !is_pmd_migration_entry(pmdval));
		if (is_pmd_migration_entry(pmdval))
582
			pmd_migration_entry_wait(mm, pmd);
583 584 585 586 587 588 589
		pmdval = READ_ONCE(*pmd);
		/*
		 * MADV_DONTNEED may convert the pmd to null because
		 * mmap_sem is held in read mode
		 */
		if (pmd_none(pmdval))
			return no_page_table(vma, flags);
590 591
		goto retry;
	}
592
	if (pmd_devmap(pmdval)) {
593
		ptl = pmd_lock(mm, pmd);
594
		page = follow_devmap_pmd(vma, address, pmd, flags, &ctx->pgmap);
595 596 597 598
		spin_unlock(ptl);
		if (page)
			return page;
	}
599
	if (likely(!pmd_trans_huge(pmdval)))
600
		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
601

602
	if ((flags & FOLL_NUMA) && pmd_protnone(pmdval))
603 604
		return no_page_table(vma, flags);

605
retry_locked:
606
	ptl = pmd_lock(mm, pmd);
607 608 609 610
	if (unlikely(pmd_none(*pmd))) {
		spin_unlock(ptl);
		return no_page_table(vma, flags);
	}
611 612 613 614 615 616 617
	if (unlikely(!pmd_present(*pmd))) {
		spin_unlock(ptl);
		if (likely(!(flags & FOLL_MIGRATION)))
			return no_page_table(vma, flags);
		pmd_migration_entry_wait(mm, pmd);
		goto retry_locked;
	}
618 619
	if (unlikely(!pmd_trans_huge(*pmd))) {
		spin_unlock(ptl);
620
		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
621
	}
Song Liu's avatar
Song Liu committed
622
	if (flags & (FOLL_SPLIT | FOLL_SPLIT_PMD)) {
623 624 625 626 627
		int ret;
		page = pmd_page(*pmd);
		if (is_huge_zero_page(page)) {
			spin_unlock(ptl);
			ret = 0;
628
			split_huge_pmd(vma, pmd, address);
629 630
			if (pmd_trans_unstable(pmd))
				ret = -EBUSY;
Song Liu's avatar
Song Liu committed
631
		} else if (flags & FOLL_SPLIT) {
632 633 634 635
			if (unlikely(!try_get_page(page))) {
				spin_unlock(ptl);
				return ERR_PTR(-ENOMEM);
			}
636
			spin_unlock(ptl);
637 638 639 640
			lock_page(page);
			ret = split_huge_page(page);
			unlock_page(page);
			put_page(page);
641 642
			if (pmd_none(*pmd))
				return no_page_table(vma, flags);
Song Liu's avatar
Song Liu committed
643 644 645 646
		} else {  /* flags & FOLL_SPLIT_PMD */
			spin_unlock(ptl);
			split_huge_pmd(vma, pmd, address);
			ret = pte_alloc(mm, pmd) ? -ENOMEM : 0;
647 648 649
		}

		return ret ? ERR_PTR(ret) :
650
			follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
651
	}
652 653
	page = follow_trans_huge_pmd(vma, address, pmd, flags);
	spin_unlock(ptl);
654
	ctx->page_mask = HPAGE_PMD_NR - 1;
655
	return page;
656 657
}

658 659
static struct page *follow_pud_mask(struct vm_area_struct *vma,
				    unsigned long address, p4d_t *p4dp,
660 661
				    unsigned int flags,
				    struct follow_page_context *ctx)
662 663 664 665 666 667 668 669 670
{
	pud_t *pud;
	spinlock_t *ptl;
	struct page *page;
	struct mm_struct *mm = vma->vm_mm;

	pud = pud_offset(p4dp, address);
	if (pud_none(*pud))
		return no_page_table(vma, flags);
671
	if (pud_huge(*pud) && is_vm_hugetlb_page(vma)) {
672 673 674 675 676
		page = follow_huge_pud(mm, address, pud, flags);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
677 678 679 680 681 682 683 684
	if (is_hugepd(__hugepd(pud_val(*pud)))) {
		page = follow_huge_pd(vma, address,
				      __hugepd(pud_val(*pud)), flags,
				      PUD_SHIFT);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
685 686
	if (pud_devmap(*pud)) {
		ptl = pud_lock(mm, pud);
687
		page = follow_devmap_pud(vma, address, pud, flags, &ctx->pgmap);
688 689 690 691 692 693 694
		spin_unlock(ptl);
		if (page)
			return page;
	}
	if (unlikely(pud_bad(*pud)))
		return no_page_table(vma, flags);

695
	return follow_pmd_mask(vma, address, pud, flags, ctx);
696 697 698 699
}

static struct page *follow_p4d_mask(struct vm_area_struct *vma,
				    unsigned long address, pgd_t *pgdp,
700 701
				    unsigned int flags,
				    struct follow_page_context *ctx)
702 703
{
	p4d_t *p4d;
704
	struct page *page;
705 706 707 708 709 710 711 712

	p4d = p4d_offset(pgdp, address);
	if (p4d_none(*p4d))
		return no_page_table(vma, flags);
	BUILD_BUG_ON(p4d_huge(*p4d));
	if (unlikely(p4d_bad(*p4d)))
		return no_page_table(vma, flags);

713 714 715 716 717 718 719 720
	if (is_hugepd(__hugepd(p4d_val(*p4d)))) {
		page = follow_huge_pd(vma, address,
				      __hugepd(p4d_val(*p4d)), flags,
				      P4D_SHIFT);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
721
	return follow_pud_mask(vma, address, p4d, flags, ctx);
722 723 724 725 726 727 728
}

/**
 * follow_page_mask - look up a page descriptor from a user-virtual address
 * @vma: vm_area_struct mapping @address
 * @address: virtual address to look up
 * @flags: flags modifying lookup behaviour
729 730
 * @ctx: contains dev_pagemap for %ZONE_DEVICE memory pinning and a
 *       pointer to output page_mask
731 732 733
 *
 * @flags can have FOLL_ flags set, defined in <linux/mm.h>
 *
734 735 736 737 738 739
 * When getting pages from ZONE_DEVICE memory, the @ctx->pgmap caches
 * the device's dev_pagemap metadata to avoid repeating expensive lookups.
 *
 * On output, the @ctx->page_mask is set according to the size of the page.
 *
 * Return: the mapped (struct page *), %NULL if no mapping exists, or
740 741 742
 * an error pointer if there is a mapping to something not represented
 * by a page descriptor (see also vm_normal_page()).
 */
743
static struct page *follow_page_mask(struct vm_area_struct *vma,
744
			      unsigned long address, unsigned int flags,
745
			      struct follow_page_context *ctx)
746 747 748 749 750
{
	pgd_t *pgd;
	struct page *page;
	struct mm_struct *mm = vma->vm_mm;

751
	ctx->page_mask = 0;
752 753 754 755

	/* make this handle hugepd */
	page = follow_huge_addr(mm, address, flags & FOLL_WRITE);
	if (!IS_ERR(page)) {
John Hubbard's avatar
John Hubbard committed
756
		WARN_ON_ONCE(flags & (FOLL_GET | FOLL_PIN));
757 758 759 760 761 762 763 764
		return page;
	}

	pgd = pgd_offset(mm, address);

	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
		return no_page_table(vma, flags);

765 766 767 768 769 770
	if (pgd_huge(*pgd)) {
		page = follow_huge_pgd(mm, address, pgd, flags);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
771 772 773 774 775 776 777 778
	if (is_hugepd(__hugepd(pgd_val(*pgd)))) {
		page = follow_huge_pd(vma, address,
				      __hugepd(pgd_val(*pgd)), flags,
				      PGDIR_SHIFT);
		if (page)
			return page;
		return no_page_table(vma, flags);
	}
779

780 781 782 783 784 785 786 787 788 789 790 791 792
	return follow_p4d_mask(vma, address, pgd, flags, ctx);
}

struct page *follow_page(struct vm_area_struct *vma, unsigned long address,
			 unsigned int foll_flags)
{
	struct follow_page_context ctx = { NULL };
	struct page *page;

	page = follow_page_mask(vma, address, foll_flags, &ctx);
	if (ctx.pgmap)
		put_dev_pagemap(ctx.pgmap);
	return page;
793 794
}

795 796 797 798 799
static int get_gate_page(struct mm_struct *mm, unsigned long address,
		unsigned int gup_flags, struct vm_area_struct **vma,
		struct page **page)
{
	pgd_t *pgd;
800
	p4d_t *p4d;
801 802 803 804 805 806 807 808 809 810 811 812
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
	int ret = -EFAULT;

	/* user gate pages are read-only */
	if (gup_flags & FOLL_WRITE)
		return -EFAULT;
	if (address > TASK_SIZE)
		pgd = pgd_offset_k(address);
	else
		pgd = pgd_offset_gate(mm, address);
813 814
	if (pgd_none(*pgd))
		return -EFAULT;
815
	p4d = p4d_offset(pgd, address);
816 817
	if (p4d_none(*p4d))
		return -EFAULT;
818
	pud = pud_offset(p4d, address);
819 820
	if (pud_none(*pud))
		return -EFAULT;
821
	pmd = pmd_offset(pud, address);
822
	if (!pmd_present(*pmd))
823 824 825 826 827 828 829 830 831 832 833 834 835 836
		return -EFAULT;
	VM_BUG_ON(pmd_trans_huge(*pmd));
	pte = pte_offset_map(pmd, address);
	if (pte_none(*pte))
		goto unmap;
	*vma = get_gate_vma(mm);
	if (!page)
		goto out;
	*page = vm_normal_page(*vma, address, *pte);
	if (!*page) {
		if ((gup_flags & FOLL_DUMP) || !is_zero_pfn(pte_pfn(*pte)))
			goto unmap;
		*page = pte_page(*pte);
	}
837 838 839 840
	if (unlikely(!try_get_page(*page))) {
		ret = -ENOMEM;
		goto unmap;
	}
841 842 843 844 845 846 847
out:
	ret = 0;
unmap:
	pte_unmap(pte);
	return ret;
}

848
/*
849 850 851
 * mmap_sem must be held on entry.  If @locked != NULL and *@flags
 * does not include FOLL_NOWAIT, the mmap_sem may be released.  If it
 * is, *@locked will be set to 0 and -EBUSY returned.
852
 */
853
static int faultin_page(struct task_struct *tsk, struct vm_area_struct *vma,
854
		unsigned long address, unsigned int *flags, int *locked)
855 856
{
	unsigned int fault_flags = 0;
857
	vm_fault_t ret;
858

Eric B Munson's avatar
Eric B Munson committed
859 860 861
	/* mlock all present pages, but do not fault in new pages */
	if ((*flags & (FOLL_POPULATE | FOLL_MLOCK)) == FOLL_MLOCK)
		return -ENOENT;
862 863
	if (*flags & FOLL_WRITE)
		fault_flags |= FAULT_FLAG_WRITE;
864 865
	if (*flags & FOLL_REMOTE)
		fault_flags |= FAULT_FLAG_REMOTE;
866
	if (locked)
867 868 869
		fault_flags |= FAULT_FLAG_ALLOW_RETRY;
	if (*flags & FOLL_NOWAIT)
		fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_RETRY_NOWAIT;
870 871 872 873
	if (*flags & FOLL_TRIED) {
		VM_WARN_ON_ONCE(fault_flags & FAULT_FLAG_ALLOW_RETRY);
		fault_flags |= FAULT_FLAG_TRIED;
	}
874

875
	ret = handle_mm_fault(vma, address, fault_flags);
876
	if (ret & VM_FAULT_ERROR) {
877 878 879 880
		int err = vm_fault_to_errno(ret, *flags);

		if (err)
			return err;
881 882 883 884 885 886 887 888 889 890 891
		BUG();
	}

	if (tsk) {
		if (ret & VM_FAULT_MAJOR)
			tsk->maj_flt++;
		else
			tsk->min_flt++;
	}

	if (ret & VM_FAULT_RETRY) {
892 893
		if (locked && !(fault_flags & FAULT_FLAG_RETRY_NOWAIT))
			*locked = 0;
894 895 896 897 898 899 900 901 902 903 904 905 906
		return -EBUSY;
	}

	/*
	 * The VM_FAULT_WRITE bit tells us that do_wp_page has broken COW when
	 * necessary, even if maybe_mkwrite decided not to set pte_write. We
	 * can thus safely do subsequent page lookups as if they were reads.
	 * But only do so when looping for pte_write is futile: in some cases
	 * userspace may also be wanting to write to the gotten user page,
	 * which a read fault here might prevent (a readonly page might get
	 * reCOWed by userspace write).
	 */
	if ((ret & VM_FAULT_WRITE) && !(vma->vm_flags & VM_WRITE))
907
		*flags |= FOLL_COW;
908 909 910
	return 0;
}

911 912 913
static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags)
{
	vm_flags_t vm_flags = vma->vm_flags;
914 915
	int write = (gup_flags & FOLL_WRITE);
	int foreign = (gup_flags & FOLL_REMOTE);
916 917 918 919

	if (vm_flags & (VM_IO | VM_PFNMAP))
		return -EFAULT;

920 921 922
	if (gup_flags & FOLL_ANON && !vma_is_anonymous(vma))
		return -EFAULT;

923
	if (write) {
924 925 926 927 928 929 930 931 932 933 934 935
		if (!(vm_flags & VM_WRITE)) {
			if (!(gup_flags & FOLL_FORCE))
				return -EFAULT;
			/*
			 * We used to let the write,force case do COW in a
			 * VM_MAYWRITE VM_SHARED !VM_WRITE vma, so ptrace could
			 * set a breakpoint in a read-only mapping of an
			 * executable, without corrupting the file (yet only
			 * when that file had been opened for writing!).
			 * Anon pages in shared mappings are surprising: now
			 * just reject it.
			 */
936
			if (!is_cow_mapping(vm_flags))
937 938 939 940 941 942 943 944 945 946 947 948
				return -EFAULT;
		}
	} else if (!(vm_flags & VM_READ)) {
		if (!(gup_flags & FOLL_FORCE))
			return -EFAULT;
		/*
		 * Is there actually any vma we can reach here which does not
		 * have VM_MAYREAD set?
		 */
		if (!(vm_flags & VM_MAYREAD))
			return -EFAULT;
	}
949 950 951 952 953
	/*
	 * gups are always data accesses, not instruction
	 * fetches, so execute=false here
	 */
	if (!arch_vma_access_permitted(vma, write, false, foreign))
954
		return -EFAULT;
955 956 957
	return 0;
}

958 959 960 961 962 963 964 965 966 967 968 969
/**
 * __get_user_pages() - pin user pages in memory
 * @tsk:	task_struct of target task
 * @mm:		mm_struct of target mm
 * @start:	starting user address
 * @nr_pages:	number of pages from start to pin
 * @gup_flags:	flags modifying pin behaviour
 * @pages:	array that receives pointers to the pages pinned.
 *		Should be at least nr_pages long. Or NULL, if caller
 *		only intends to ensure the pages are faulted in.
 * @vmas:	array of pointers to vmas corresponding to each page.
 *		Or NULL if the caller does not require them.
970
 * @locked:     whether we're still with the mmap_sem held
971
 *
972 973 974 975 976 977 978 979 980 981 982
 * Returns either number of pages pinned (which may be less than the
 * number requested), or an error. Details about the return value:
 *
 * -- If nr_pages is 0, returns 0.
 * -- If nr_pages is >0, but no pages were pinned, returns -errno.
 * -- If nr_pages is >0, and some pages were pinned, returns the number of
 *    pages pinned. Again, this may be less than nr_pages.
 *
 * The caller is responsible for releasing returned @pages, via put_page().
 *
 * @vmas are valid only as long as mmap_sem is held.
983
 *
984
 * Must be called with mmap_sem held.  It may be released.  See below.
985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004
 *
 * __get_user_pages walks a process's page tables and takes a reference to
 * each struct page that each user address corresponds to at a given
 * instant. That is, it takes the page that would be accessed if a user
 * thread accesses the given user virtual address at that instant.
 *
 * This does not guarantee that the page exists in the user mappings when
 * __get_user_pages returns, and there may even be a completely different
 * page there in some cases (eg. if mmapped pagecache has been invalidated
 * and subsequently re faulted). However it does guarantee that the page
 * won't be freed completely. And mostly callers simply care that the page
 * contains data that was valid *at some point in time*. Typically, an IO
 * or similar operation cannot guarantee anything stronger anyway because
 * locks can't be held over the syscall boundary.
 *
 * If @gup_flags & FOLL_WRITE == 0, the page must not be written to. If
 * the page is written to, set_page_dirty (or set_page_dirty_lock, as
 * appropriate) must be called after the page is finished with, and
 * before put_page is called.
 *
1005 1006 1007
 * If @locked != NULL, *@locked will be set to 0 when mmap_sem is
 * released by an up_read().  That can happen if @gup_flags does not
 * have FOLL_NOWAIT.
1008
 *
1009
 * A caller using such a combination of @locked and @gup_flags
1010 1011 1012
 * must therefore hold the mmap_sem for reading only, and recognize
 * when it's been released.  Otherwise, it must be held for either
 * reading or writing and will not be released.
1013 1014 1015 1016 1017
 *
 * In most cases, get_user_pages or get_user_pages_fast should be used
 * instead of __get_user_pages. __get_user_pages should be used only if
 * you need some special @gup_flags.
 */
1018
static long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
1019 1020
		unsigned long start, unsigned long nr_pages,
		unsigned int gup_flags, struct page **pages,
1021
		struct vm_area_struct **vmas, int *locked)
1022
{
1023
	long ret = 0, i = 0;
1024
	struct vm_area_struct *vma = NULL;
1025
	struct follow_page_context ctx = { NULL };
1026 1027 1028 1029

	if (!nr_pages)
		return 0;

1030 1031
	start = untagged_addr(start);

1032
	VM_BUG_ON(!!pages != !!(gup_flags & (FOLL_GET | FOLL_PIN)));
1033 1034 1035 1036 1037 1038 1039 1040 1041 1042

	/*
	 * If FOLL_FORCE is set then do not force a full fault as the hinting
	 * fault information is unrelated to the reference behaviour of a task
	 * using the address space
	 */
	if (!(gup_flags & FOLL_FORCE))
		gup_flags |= FOLL_NUMA;

	do {
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
		struct page *page;
		unsigned int foll_flags = gup_flags;
		unsigned int page_increm;

		/* first iteration or cross vma bound */
		if (!vma || start >= vma->vm_end) {
			vma = find_extend_vma(mm, start);
			if (!vma && in_gate_area(mm, start)) {
				ret = get_gate_page(mm, start & PAGE_MASK,
						gup_flags, &vma,
						pages ? &pages[i] : NULL);
				if (ret)
1055
					goto out;
1056
				ctx.page_mask = 0;
1057 1058
				goto next_page;
			}
1059

1060 1061 1062 1063
			if (!vma || check_vma_flags(vma, gup_flags)) {
				ret = -EFAULT;
				goto out;
			}
1064 1065 1066
			if (is_vm_hugetlb_page(vma)) {
				i = follow_hugetlb_page(mm, vma, pages, vmas,
						&start, &nr_pages, i,
1067
						gup_flags, locked);
1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
				if (locked && *locked == 0) {
					/*
					 * We've got a VM_FAULT_RETRY
					 * and we've lost mmap_sem.
					 * We must stop here.
					 */
					BUG_ON(gup_flags & FOLL_NOWAIT);
					BUG_ON(ret != 0);
					goto out;
				}
1078
				continue;
1079
			}
1080 1081 1082 1083 1084 1085
		}
retry:
		/*
		 * If we have a pending SIGKILL, don't keep faulting pages and
		 * potentially allocating memory.
		 */
1086
		if (fatal_signal_pending(current)) {
1087 1088 1089
			ret = -ERESTARTSYS;
			goto out;
		}
1090
		cond_resched();
1091 1092

		page = follow_page_mask(vma, start, foll_flags, &ctx);
1093 1094
		if (!page) {
			ret = faultin_page(tsk, vma, start, &foll_flags,
1095
					   locked);
1096 1097 1098
			switch (ret) {
			case 0:
				goto retry;
1099 1100 1101
			case -EBUSY:
				ret = 0;
				/* FALLTHRU */
1102 1103 1104
			case -EFAULT:
			case -ENOMEM:
			case -EHWPOISON:
1105
				goto out;
1106 1107
			case -ENOENT:
				goto next_page;
1108
			}
1109
			BUG();
1110 1111 1112 1113 1114 1115 1116
		} else if (PTR_ERR(page) == -EEXIST) {
			/*
			 * Proper page table entry exists, but no corresponding
			 * struct page.
			 */
			goto next_page;
		} else if (IS_ERR(page)) {
1117 1118
			ret = PTR_ERR(page);
			goto out;
1119
		}
1120 1121 1122 1123
		if (pages) {
			pages[i] = page;
			flush_anon_page(vma, page, start);
			flush_dcache_page(page);
1124
			ctx.page_mask = 0;
1125 1126
		}
next_page:
1127 1128
		if (vmas) {
			vmas[i] = vma;
1129
			ctx.page_mask = 0;
1130
		}