filemap.c 67.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11
/*
 *	linux/mm/filemap.c
 *
 * Copyright (C) 1994-1999  Linus Torvalds
 */

/*
 * This file handles the generic file mmap semantics used by
 * most "normal" filesystems (but you don't /have/ to use this:
 * the NFS filesystem used to do this differently, for example)
 */
12
#include <linux/export.h>
L
Linus Torvalds 已提交
13 14
#include <linux/compiler.h>
#include <linux/fs.h>
15
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
16
#include <linux/aio.h>
17
#include <linux/capability.h>
L
Linus Torvalds 已提交
18
#include <linux/kernel_stat.h>
19
#include <linux/gfp.h>
L
Linus Torvalds 已提交
20 21 22 23 24 25 26 27
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/file.h>
#include <linux/uio.h>
#include <linux/hash.h>
#include <linux/writeback.h>
28
#include <linux/backing-dev.h>
L
Linus Torvalds 已提交
29 30 31
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/security.h>
32
#include <linux/cpuset.h>
N
Nick Piggin 已提交
33
#include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
34
#include <linux/memcontrol.h>
35
#include <linux/cleancache.h>
36 37
#include "internal.h"

R
Robert Jarzmik 已提交
38 39 40
#define CREATE_TRACE_POINTS
#include <trace/events/filemap.h>

L
Linus Torvalds 已提交
41 42 43
/*
 * FIXME: remove all knowledge of the buffer layer from the core VM
 */
44
#include <linux/buffer_head.h> /* for try_to_free_buffers */
L
Linus Torvalds 已提交
45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

#include <asm/mman.h>

/*
 * Shared mappings implemented 30.11.1994. It's not fully working yet,
 * though.
 *
 * Shared mappings now work. 15.8.1995  Bruno.
 *
 * finished 'unifying' the page and buffer cache and SMP-threaded the
 * page-cache, 21.05.1999, Ingo Molnar <mingo@redhat.com>
 *
 * SMP-threaded pagemap-LRU 1999, Andrea Arcangeli <andrea@suse.de>
 */

/*
 * Lock ordering:
 *
63
 *  ->i_mmap_mutex		(truncate_pagecache)
L
Linus Torvalds 已提交
64
 *    ->private_lock		(__free_pte->__set_page_dirty_buffers)
65 66
 *      ->swap_lock		(exclusive_swap_page, others)
 *        ->mapping->tree_lock
L
Linus Torvalds 已提交
67
 *
68
 *  ->i_mutex
69
 *    ->i_mmap_mutex		(truncate->unmap_mapping_range)
L
Linus Torvalds 已提交
70 71
 *
 *  ->mmap_sem
72
 *    ->i_mmap_mutex
73
 *      ->page_table_lock or pte_lock	(various, mainly in memory.c)
L
Linus Torvalds 已提交
74 75 76 77 78
 *        ->mapping->tree_lock	(arch-dependent flush_dcache_mmap_lock)
 *
 *  ->mmap_sem
 *    ->lock_page		(access_process_vm)
 *
79 80
 *  ->i_mutex			(generic_file_buffered_write)
 *    ->mmap_sem		(fault_in_pages_readable->do_page_fault)
L
Linus Torvalds 已提交
81
 *
82
 *  bdi->wb.list_lock
83
 *    sb_lock			(fs/fs-writeback.c)
L
Linus Torvalds 已提交
84 85
 *    ->mapping->tree_lock	(__sync_single_inode)
 *
86
 *  ->i_mmap_mutex
L
Linus Torvalds 已提交
87 88 89
 *    ->anon_vma.lock		(vma_adjust)
 *
 *  ->anon_vma.lock
90
 *    ->page_table_lock or pte_lock	(anon_vma_prepare and various)
L
Linus Torvalds 已提交
91
 *
92
 *  ->page_table_lock or pte_lock
93
 *    ->swap_lock		(try_to_unmap_one)
L
Linus Torvalds 已提交
94 95 96
 *    ->private_lock		(try_to_unmap_one)
 *    ->tree_lock		(try_to_unmap_one)
 *    ->zone.lru_lock		(follow_page->mark_page_accessed)
97
 *    ->zone.lru_lock		(check_pte_range->isolate_lru_page)
L
Linus Torvalds 已提交
98 99
 *    ->private_lock		(page_remove_rmap->set_page_dirty)
 *    ->tree_lock		(page_remove_rmap->set_page_dirty)
100
 *    bdi.wb->list_lock		(page_remove_rmap->set_page_dirty)
101
 *    ->inode->i_lock		(page_remove_rmap->set_page_dirty)
102
 *    bdi.wb->list_lock		(zap_pte_range->set_page_dirty)
103
 *    ->inode->i_lock		(zap_pte_range->set_page_dirty)
L
Linus Torvalds 已提交
104 105
 *    ->private_lock		(zap_pte_range->__set_page_dirty_buffers)
 *
106 107
 * ->i_mmap_mutex
 *   ->tasklist_lock            (memory_failure, collect_procs_ao)
L
Linus Torvalds 已提交
108 109 110
 */

/*
111
 * Delete a page from the page cache and free it. Caller has to make
L
Linus Torvalds 已提交
112
 * sure the page is locked and that nobody else uses it - or that usage
N
Nick Piggin 已提交
113
 * is safe.  The caller must hold the mapping's tree_lock.
L
Linus Torvalds 已提交
114
 */
115
void __delete_from_page_cache(struct page *page)
L
Linus Torvalds 已提交
116 117 118
{
	struct address_space *mapping = page->mapping;

R
Robert Jarzmik 已提交
119
	trace_mm_filemap_delete_from_page_cache(page);
120 121 122 123 124 125 126 127
	/*
	 * if we're uptodate, flush out into the cleancache, otherwise
	 * invalidate any existing cleancache entries.  We can't leave
	 * stale data around in the cleancache once our page is gone
	 */
	if (PageUptodate(page) && PageMappedToDisk(page))
		cleancache_put_page(page);
	else
128
		cleancache_invalidate_page(mapping, page);
129

L
Linus Torvalds 已提交
130 131
	radix_tree_delete(&mapping->page_tree, page->index);
	page->mapping = NULL;
132
	/* Leave page->index set: truncation lookup relies upon it */
L
Linus Torvalds 已提交
133
	mapping->nrpages--;
134
	__dec_zone_page_state(page, NR_FILE_PAGES);
135 136
	if (PageSwapBacked(page))
		__dec_zone_page_state(page, NR_SHMEM);
137
	BUG_ON(page_mapped(page));
138 139 140 141 142 143 144 145 146 147 148 149

	/*
	 * Some filesystems seem to re-dirty the page even after
	 * the VM has canceled the dirty bit (eg ext3 journaling).
	 *
	 * Fix it up by doing a final dirty accounting check after
	 * having removed the page entirely.
	 */
	if (PageDirty(page) && mapping_cap_account_dirty(mapping)) {
		dec_zone_page_state(page, NR_FILE_DIRTY);
		dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE);
	}
L
Linus Torvalds 已提交
150 151
}

152 153 154 155 156 157 158 159 160
/**
 * delete_from_page_cache - delete page from page cache
 * @page: the page which the kernel is trying to remove from page cache
 *
 * This must be called only on pages that have been verified to be in the page
 * cache and locked.  It will never put the page into the free list, the caller
 * has a reference on the page.
 */
void delete_from_page_cache(struct page *page)
L
Linus Torvalds 已提交
161 162
{
	struct address_space *mapping = page->mapping;
163
	void (*freepage)(struct page *);
L
Linus Torvalds 已提交
164

M
Matt Mackall 已提交
165
	BUG_ON(!PageLocked(page));
L
Linus Torvalds 已提交
166

167
	freepage = mapping->a_ops->freepage;
N
Nick Piggin 已提交
168
	spin_lock_irq(&mapping->tree_lock);
169
	__delete_from_page_cache(page);
N
Nick Piggin 已提交
170
	spin_unlock_irq(&mapping->tree_lock);
171
	mem_cgroup_uncharge_cache_page(page);
172 173 174

	if (freepage)
		freepage(page);
175 176 177 178
	page_cache_release(page);
}
EXPORT_SYMBOL(delete_from_page_cache);

J
Jens Axboe 已提交
179
static int sleep_on_page(void *word)
L
Linus Torvalds 已提交
180 181 182 183 184
{
	io_schedule();
	return 0;
}

J
Jens Axboe 已提交
185
static int sleep_on_page_killable(void *word)
M
Matthew Wilcox 已提交
186
{
J
Jens Axboe 已提交
187
	sleep_on_page(word);
M
Matthew Wilcox 已提交
188 189 190
	return fatal_signal_pending(current) ? -EINTR : 0;
}

191 192 193 194 195 196 197 198 199 200 201
static int filemap_check_errors(struct address_space *mapping)
{
	int ret = 0;
	/* Check for outstanding write errors */
	if (test_and_clear_bit(AS_ENOSPC, &mapping->flags))
		ret = -ENOSPC;
	if (test_and_clear_bit(AS_EIO, &mapping->flags))
		ret = -EIO;
	return ret;
}

L
Linus Torvalds 已提交
202
/**
203
 * __filemap_fdatawrite_range - start writeback on mapping dirty pages in range
204 205
 * @mapping:	address space structure to write
 * @start:	offset in bytes where the range starts
206
 * @end:	offset in bytes where the range ends (inclusive)
207
 * @sync_mode:	enable synchronous operation
L
Linus Torvalds 已提交
208
 *
209 210 211
 * Start writeback against all of a mapping's dirty pages that lie
 * within the byte offsets <start, end> inclusive.
 *
L
Linus Torvalds 已提交
212
 * If sync_mode is WB_SYNC_ALL then this is a "data integrity" operation, as
213
 * opposed to a regular memory cleansing writeback.  The difference between
L
Linus Torvalds 已提交
214 215 216
 * these two operations is that if a dirty page/buffer is encountered, it must
 * be waited upon, and not just skipped over.
 */
217 218
int __filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
				loff_t end, int sync_mode)
L
Linus Torvalds 已提交
219 220 221 222
{
	int ret;
	struct writeback_control wbc = {
		.sync_mode = sync_mode,
223
		.nr_to_write = LONG_MAX,
224 225
		.range_start = start,
		.range_end = end,
L
Linus Torvalds 已提交
226 227 228 229 230 231 232 233 234 235 236 237
	};

	if (!mapping_cap_writeback_dirty(mapping))
		return 0;

	ret = do_writepages(mapping, &wbc);
	return ret;
}

static inline int __filemap_fdatawrite(struct address_space *mapping,
	int sync_mode)
{
238
	return __filemap_fdatawrite_range(mapping, 0, LLONG_MAX, sync_mode);
L
Linus Torvalds 已提交
239 240 241 242 243 244 245 246
}

int filemap_fdatawrite(struct address_space *mapping)
{
	return __filemap_fdatawrite(mapping, WB_SYNC_ALL);
}
EXPORT_SYMBOL(filemap_fdatawrite);

247
int filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
248
				loff_t end)
L
Linus Torvalds 已提交
249 250 251
{
	return __filemap_fdatawrite_range(mapping, start, end, WB_SYNC_ALL);
}
252
EXPORT_SYMBOL(filemap_fdatawrite_range);
L
Linus Torvalds 已提交
253

254 255 256 257
/**
 * filemap_flush - mostly a non-blocking flush
 * @mapping:	target address_space
 *
L
Linus Torvalds 已提交
258 259 260 261 262 263 264 265 266
 * This is a mostly non-blocking flush.  Not suitable for data-integrity
 * purposes - I/O may not be started against all dirty pages.
 */
int filemap_flush(struct address_space *mapping)
{
	return __filemap_fdatawrite(mapping, WB_SYNC_NONE);
}
EXPORT_SYMBOL(filemap_flush);

267
/**
268 269 270 271
 * filemap_fdatawait_range - wait for writeback to complete
 * @mapping:		address space structure to wait for
 * @start_byte:		offset in bytes where the range starts
 * @end_byte:		offset in bytes where the range ends (inclusive)
272
 *
273 274
 * Walk the list of under-writeback pages of the given address space
 * in the given range and wait for all of them.
L
Linus Torvalds 已提交
275
 */
276 277
int filemap_fdatawait_range(struct address_space *mapping, loff_t start_byte,
			    loff_t end_byte)
L
Linus Torvalds 已提交
278
{
279 280
	pgoff_t index = start_byte >> PAGE_CACHE_SHIFT;
	pgoff_t end = end_byte >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
281 282
	struct pagevec pvec;
	int nr_pages;
283
	int ret2, ret = 0;
L
Linus Torvalds 已提交
284

285
	if (end_byte < start_byte)
286
		goto out;
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302

	pagevec_init(&pvec, 0);
	while ((index <= end) &&
			(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
			PAGECACHE_TAG_WRITEBACK,
			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1)) != 0) {
		unsigned i;

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];

			/* until radix tree lookup accepts end_index */
			if (page->index > end)
				continue;

			wait_on_page_writeback(page);
303
			if (TestClearPageError(page))
L
Linus Torvalds 已提交
304 305 306 307 308
				ret = -EIO;
		}
		pagevec_release(&pvec);
		cond_resched();
	}
309 310 311 312
out:
	ret2 = filemap_check_errors(mapping);
	if (!ret)
		ret = ret2;
L
Linus Torvalds 已提交
313 314 315

	return ret;
}
316 317
EXPORT_SYMBOL(filemap_fdatawait_range);

L
Linus Torvalds 已提交
318
/**
319
 * filemap_fdatawait - wait for all under-writeback pages to complete
L
Linus Torvalds 已提交
320
 * @mapping: address space structure to wait for
321 322 323
 *
 * Walk the list of under-writeback pages of the given address space
 * and wait for all of them.
L
Linus Torvalds 已提交
324 325 326 327 328 329 330 331
 */
int filemap_fdatawait(struct address_space *mapping)
{
	loff_t i_size = i_size_read(mapping->host);

	if (i_size == 0)
		return 0;

332
	return filemap_fdatawait_range(mapping, 0, i_size - 1);
L
Linus Torvalds 已提交
333 334 335 336 337
}
EXPORT_SYMBOL(filemap_fdatawait);

int filemap_write_and_wait(struct address_space *mapping)
{
338
	int err = 0;
L
Linus Torvalds 已提交
339 340

	if (mapping->nrpages) {
341 342 343 344 345 346 347 348 349 350 351 352
		err = filemap_fdatawrite(mapping);
		/*
		 * Even if the above returned error, the pages may be
		 * written partially (e.g. -ENOSPC), so we wait for it.
		 * But the -EIO is special case, it may indicate the worst
		 * thing (e.g. bug) happened, so we avoid waiting for it.
		 */
		if (err != -EIO) {
			int err2 = filemap_fdatawait(mapping);
			if (!err)
				err = err2;
		}
353 354
	} else {
		err = filemap_check_errors(mapping);
L
Linus Torvalds 已提交
355
	}
356
	return err;
L
Linus Torvalds 已提交
357
}
358
EXPORT_SYMBOL(filemap_write_and_wait);
L
Linus Torvalds 已提交
359

360 361 362 363 364 365
/**
 * filemap_write_and_wait_range - write out & wait on a file range
 * @mapping:	the address_space for the pages
 * @lstart:	offset in bytes where the range starts
 * @lend:	offset in bytes where the range ends (inclusive)
 *
366 367 368 369 370
 * Write out and wait upon file offsets lstart->lend, inclusive.
 *
 * Note that `lend' is inclusive (describes the last byte to be written) so
 * that this function can be used to write to the very end-of-file (end = -1).
 */
L
Linus Torvalds 已提交
371 372 373
int filemap_write_and_wait_range(struct address_space *mapping,
				 loff_t lstart, loff_t lend)
{
374
	int err = 0;
L
Linus Torvalds 已提交
375 376

	if (mapping->nrpages) {
377 378 379 380
		err = __filemap_fdatawrite_range(mapping, lstart, lend,
						 WB_SYNC_ALL);
		/* See comment of filemap_write_and_wait() */
		if (err != -EIO) {
381 382
			int err2 = filemap_fdatawait_range(mapping,
						lstart, lend);
383 384 385
			if (!err)
				err = err2;
		}
386 387
	} else {
		err = filemap_check_errors(mapping);
L
Linus Torvalds 已提交
388
	}
389
	return err;
L
Linus Torvalds 已提交
390
}
391
EXPORT_SYMBOL(filemap_write_and_wait_range);
L
Linus Torvalds 已提交
392

393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
/**
 * replace_page_cache_page - replace a pagecache page with a new one
 * @old:	page to be replaced
 * @new:	page to replace with
 * @gfp_mask:	allocation mode
 *
 * This function replaces a page in the pagecache with a new one.  On
 * success it acquires the pagecache reference for the new page and
 * drops it for the old page.  Both the old and new pages must be
 * locked.  This function does not add the new page to the LRU, the
 * caller must do that.
 *
 * The remove + add is atomic.  The only way this function can fail is
 * memory allocation failure.
 */
int replace_page_cache_page(struct page *old, struct page *new, gfp_t gfp_mask)
{
	int error;

412 413 414
	VM_BUG_ON_PAGE(!PageLocked(old), old);
	VM_BUG_ON_PAGE(!PageLocked(new), new);
	VM_BUG_ON_PAGE(new->mapping, new);
415 416 417 418 419 420 421 422 423 424 425 426 427 428

	error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
	if (!error) {
		struct address_space *mapping = old->mapping;
		void (*freepage)(struct page *);

		pgoff_t offset = old->index;
		freepage = mapping->a_ops->freepage;

		page_cache_get(new);
		new->mapping = mapping;
		new->index = offset;

		spin_lock_irq(&mapping->tree_lock);
429
		__delete_from_page_cache(old);
430 431 432 433 434 435 436
		error = radix_tree_insert(&mapping->page_tree, offset, new);
		BUG_ON(error);
		mapping->nrpages++;
		__inc_zone_page_state(new, NR_FILE_PAGES);
		if (PageSwapBacked(new))
			__inc_zone_page_state(new, NR_SHMEM);
		spin_unlock_irq(&mapping->tree_lock);
437 438
		/* mem_cgroup codes must not be called under tree_lock */
		mem_cgroup_replace_page_cache(old, new);
439 440 441 442 443 444 445 446 447 448
		radix_tree_preload_end();
		if (freepage)
			freepage(old);
		page_cache_release(old);
	}

	return error;
}
EXPORT_SYMBOL_GPL(replace_page_cache_page);

449
/**
N
Nick Piggin 已提交
450
 * add_to_page_cache_locked - add a locked page to the pagecache
451 452 453 454 455
 * @page:	page to add
 * @mapping:	the page's address_space
 * @offset:	page index
 * @gfp_mask:	page allocation mode
 *
N
Nick Piggin 已提交
456
 * This function is used to add a page to the pagecache. It must be locked.
L
Linus Torvalds 已提交
457 458
 * This function does not add the page to the LRU.  The caller must do that.
 */
N
Nick Piggin 已提交
459
int add_to_page_cache_locked(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
460
		pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
461
{
N
Nick Piggin 已提交
462 463
	int error;

464 465
	VM_BUG_ON_PAGE(!PageLocked(page), page);
	VM_BUG_ON_PAGE(PageSwapBacked(page), page);
N
Nick Piggin 已提交
466 467

	error = mem_cgroup_cache_charge(page, current->mm,
K
KAMEZAWA Hiroyuki 已提交
468
					gfp_mask & GFP_RECLAIM_MASK);
B
Balbir Singh 已提交
469
	if (error)
470
		return error;
L
Linus Torvalds 已提交
471

472
	error = radix_tree_maybe_preload(gfp_mask & ~__GFP_HIGHMEM);
473
	if (error) {
474
		mem_cgroup_uncharge_cache_page(page);
475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
		return error;
	}

	page_cache_get(page);
	page->mapping = mapping;
	page->index = offset;

	spin_lock_irq(&mapping->tree_lock);
	error = radix_tree_insert(&mapping->page_tree, offset, page);
	radix_tree_preload_end();
	if (unlikely(error))
		goto err_insert;
	mapping->nrpages++;
	__inc_zone_page_state(page, NR_FILE_PAGES);
	spin_unlock_irq(&mapping->tree_lock);
	trace_mm_filemap_add_to_page_cache(page);
	return 0;
err_insert:
	page->mapping = NULL;
	/* Leave page->index set: truncation relies upon it */
	spin_unlock_irq(&mapping->tree_lock);
	mem_cgroup_uncharge_cache_page(page);
	page_cache_release(page);
L
Linus Torvalds 已提交
498 499
	return error;
}
N
Nick Piggin 已提交
500
EXPORT_SYMBOL(add_to_page_cache_locked);
L
Linus Torvalds 已提交
501 502

int add_to_page_cache_lru(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
503
				pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
504
{
505 506 507
	int ret;

	ret = add_to_page_cache(page, mapping, offset, gfp_mask);
508 509
	if (ret == 0)
		lru_cache_add_file(page);
L
Linus Torvalds 已提交
510 511
	return ret;
}
512
EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
L
Linus Torvalds 已提交
513

514
#ifdef CONFIG_NUMA
515
struct page *__page_cache_alloc(gfp_t gfp)
516
{
517 518 519
	int n;
	struct page *page;

520
	if (cpuset_do_page_mem_spread()) {
521 522 523 524 525 526 527
		unsigned int cpuset_mems_cookie;
		do {
			cpuset_mems_cookie = get_mems_allowed();
			n = cpuset_mem_spread_node();
			page = alloc_pages_exact_node(n, gfp, 0);
		} while (!put_mems_allowed(cpuset_mems_cookie) && !page);

528
		return page;
529
	}
530
	return alloc_pages(gfp, 0);
531
}
532
EXPORT_SYMBOL(__page_cache_alloc);
533 534
#endif

L
Linus Torvalds 已提交
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556
/*
 * In order to wait for pages to become available there must be
 * waitqueues associated with pages. By using a hash table of
 * waitqueues where the bucket discipline is to maintain all
 * waiters on the same queue and wake all when any of the pages
 * become available, and for the woken contexts to check to be
 * sure the appropriate page became available, this saves space
 * at a cost of "thundering herd" phenomena during rare hash
 * collisions.
 */
static wait_queue_head_t *page_waitqueue(struct page *page)
{
	const struct zone *zone = page_zone(page);

	return &zone->wait_table[hash_ptr(page, zone->wait_table_bits)];
}

static inline void wake_up_page(struct page *page, int bit)
{
	__wake_up_bit(page_waitqueue(page), &page->flags, bit);
}

H
Harvey Harrison 已提交
557
void wait_on_page_bit(struct page *page, int bit_nr)
L
Linus Torvalds 已提交
558 559 560 561
{
	DEFINE_WAIT_BIT(wait, &page->flags, bit_nr);

	if (test_bit(bit_nr, &page->flags))
J
Jens Axboe 已提交
562
		__wait_on_bit(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
563 564 565 566
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(wait_on_page_bit);

567 568 569 570 571 572 573 574 575 576 577
int wait_on_page_bit_killable(struct page *page, int bit_nr)
{
	DEFINE_WAIT_BIT(wait, &page->flags, bit_nr);

	if (!test_bit(bit_nr, &page->flags))
		return 0;

	return __wait_on_bit(page_waitqueue(page), &wait,
			     sleep_on_page_killable, TASK_KILLABLE);
}

578 579
/**
 * add_page_wait_queue - Add an arbitrary waiter to a page's wait queue
R
Randy Dunlap 已提交
580 581
 * @page: Page defining the wait queue of interest
 * @waiter: Waiter to add to the queue
582 583 584 585 586 587 588 589 590 591 592 593 594 595
 *
 * Add an arbitrary @waiter to the wait queue for the nominated @page.
 */
void add_page_wait_queue(struct page *page, wait_queue_t *waiter)
{
	wait_queue_head_t *q = page_waitqueue(page);
	unsigned long flags;

	spin_lock_irqsave(&q->lock, flags);
	__add_wait_queue(q, waiter);
	spin_unlock_irqrestore(&q->lock, flags);
}
EXPORT_SYMBOL_GPL(add_page_wait_queue);

L
Linus Torvalds 已提交
596
/**
597
 * unlock_page - unlock a locked page
L
Linus Torvalds 已提交
598 599 600 601 602 603 604
 * @page: the page
 *
 * Unlocks the page and wakes up sleepers in ___wait_on_page_locked().
 * Also wakes sleepers in wait_on_page_writeback() because the wakeup
 * mechananism between PageLocked pages and PageWriteback pages is shared.
 * But that's OK - sleepers in wait_on_page_writeback() just go back to sleep.
 *
N
Nick Piggin 已提交
605 606
 * The mb is necessary to enforce ordering between the clear_bit and the read
 * of the waitqueue (to avoid SMP races with a parallel wait_on_page_locked()).
L
Linus Torvalds 已提交
607
 */
H
Harvey Harrison 已提交
608
void unlock_page(struct page *page)
L
Linus Torvalds 已提交
609
{
610
	VM_BUG_ON_PAGE(!PageLocked(page), page);
N
Nick Piggin 已提交
611 612
	clear_bit_unlock(PG_locked, &page->flags);
	smp_mb__after_clear_bit();
L
Linus Torvalds 已提交
613 614 615 616
	wake_up_page(page, PG_locked);
}
EXPORT_SYMBOL(unlock_page);

617 618 619
/**
 * end_page_writeback - end writeback against a page
 * @page: the page
L
Linus Torvalds 已提交
620 621 622
 */
void end_page_writeback(struct page *page)
{
623 624 625 626 627 628
	if (TestClearPageReclaim(page))
		rotate_reclaimable_page(page);

	if (!test_clear_page_writeback(page))
		BUG();

L
Linus Torvalds 已提交
629 630 631 632 633
	smp_mb__after_clear_bit();
	wake_up_page(page, PG_writeback);
}
EXPORT_SYMBOL(end_page_writeback);

634 635 636
/**
 * __lock_page - get a lock on the page, assuming we need to sleep to get it
 * @page: the page to lock
L
Linus Torvalds 已提交
637
 */
H
Harvey Harrison 已提交
638
void __lock_page(struct page *page)
L
Linus Torvalds 已提交
639 640 641
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

J
Jens Axboe 已提交
642
	__wait_on_bit_lock(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
643 644 645 646
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(__lock_page);

H
Harvey Harrison 已提交
647
int __lock_page_killable(struct page *page)
M
Matthew Wilcox 已提交
648 649 650 651
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

	return __wait_on_bit_lock(page_waitqueue(page), &wait,
J
Jens Axboe 已提交
652
					sleep_on_page_killable, TASK_KILLABLE);
M
Matthew Wilcox 已提交
653
}
654
EXPORT_SYMBOL_GPL(__lock_page_killable);
M
Matthew Wilcox 已提交
655

656 657 658
int __lock_page_or_retry(struct page *page, struct mm_struct *mm,
			 unsigned int flags)
{
659 660 661 662 663 664 665 666 667 668 669 670
	if (flags & FAULT_FLAG_ALLOW_RETRY) {
		/*
		 * CAUTION! In this case, mmap_sem is not released
		 * even though return 0.
		 */
		if (flags & FAULT_FLAG_RETRY_NOWAIT)
			return 0;

		up_read(&mm->mmap_sem);
		if (flags & FAULT_FLAG_KILLABLE)
			wait_on_page_locked_killable(page);
		else
671
			wait_on_page_locked(page);
672
		return 0;
673 674 675 676 677 678 679 680 681 682 683 684
	} else {
		if (flags & FAULT_FLAG_KILLABLE) {
			int ret;

			ret = __lock_page_killable(page);
			if (ret) {
				up_read(&mm->mmap_sem);
				return 0;
			}
		} else
			__lock_page(page);
		return 1;
685 686 687
	}
}

688 689 690 691 692
/**
 * find_get_page - find and get a page reference
 * @mapping: the address_space to search
 * @offset: the page index
 *
N
Nick Piggin 已提交
693 694
 * Is there a pagecache struct page at the given (mapping, offset) tuple?
 * If yes, increment its refcount and return it; if no, return NULL.
L
Linus Torvalds 已提交
695
 */
N
Nick Piggin 已提交
696
struct page *find_get_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
697
{
N
Nick Piggin 已提交
698
	void **pagep;
L
Linus Torvalds 已提交
699 700
	struct page *page;

N
Nick Piggin 已提交
701 702 703 704 705 706
	rcu_read_lock();
repeat:
	page = NULL;
	pagep = radix_tree_lookup_slot(&mapping->page_tree, offset);
	if (pagep) {
		page = radix_tree_deref_slot(pagep);
N
Nick Piggin 已提交
707 708
		if (unlikely(!page))
			goto out;
709
		if (radix_tree_exception(page)) {
710 711 712 713 714 715 716 717
			if (radix_tree_deref_retry(page))
				goto repeat;
			/*
			 * Otherwise, shmem/tmpfs must be storing a swap entry
			 * here as an exceptional entry: so return it without
			 * attempting to raise page count.
			 */
			goto out;
718
		}
N
Nick Piggin 已提交
719 720 721 722 723 724 725 726 727 728 729 730 731
		if (!page_cache_get_speculative(page))
			goto repeat;

		/*
		 * Has the page moved?
		 * This is part of the lockless pagecache protocol. See
		 * include/linux/pagemap.h for details.
		 */
		if (unlikely(page != *pagep)) {
			page_cache_release(page);
			goto repeat;
		}
	}
N
Nick Piggin 已提交
732
out:
N
Nick Piggin 已提交
733 734
	rcu_read_unlock();

L
Linus Torvalds 已提交
735 736 737 738 739 740
	return page;
}
EXPORT_SYMBOL(find_get_page);

/**
 * find_lock_page - locate, pin and lock a pagecache page
741 742
 * @mapping: the address_space to search
 * @offset: the page index
L
Linus Torvalds 已提交
743 744 745 746 747 748
 *
 * Locates the desired pagecache page, locks it, increments its reference
 * count and returns its address.
 *
 * Returns zero if the page was not present. find_lock_page() may sleep.
 */
N
Nick Piggin 已提交
749
struct page *find_lock_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
750 751 752 753
{
	struct page *page;

repeat:
N
Nick Piggin 已提交
754
	page = find_get_page(mapping, offset);
755
	if (page && !radix_tree_exception(page)) {
N
Nick Piggin 已提交
756 757 758 759 760 761
		lock_page(page);
		/* Has the page been truncated? */
		if (unlikely(page->mapping != mapping)) {
			unlock_page(page);
			page_cache_release(page);
			goto repeat;
L
Linus Torvalds 已提交
762
		}
763
		VM_BUG_ON_PAGE(page->index != offset, page);
L
Linus Torvalds 已提交
764 765 766 767 768 769 770
	}
	return page;
}
EXPORT_SYMBOL(find_lock_page);

/**
 * find_or_create_page - locate or add a pagecache page
771 772 773
 * @mapping: the page's address_space
 * @index: the page's index into the mapping
 * @gfp_mask: page allocation mode
L
Linus Torvalds 已提交
774 775 776 777 778 779 780 781 782 783 784 785 786
 *
 * Locates a page in the pagecache.  If the page is not present, a new page
 * is allocated using @gfp_mask and is added to the pagecache and to the VM's
 * LRU list.  The returned page is locked and has its reference count
 * incremented.
 *
 * find_or_create_page() may sleep, even if @gfp_flags specifies an atomic
 * allocation!
 *
 * find_or_create_page() returns the desired page's address, or zero on
 * memory exhaustion.
 */
struct page *find_or_create_page(struct address_space *mapping,
787
		pgoff_t index, gfp_t gfp_mask)
L
Linus Torvalds 已提交
788
{
N
Nick Piggin 已提交
789
	struct page *page;
L
Linus Torvalds 已提交
790 791 792 793
	int err;
repeat:
	page = find_lock_page(mapping, index);
	if (!page) {
N
Nick Piggin 已提交
794 795 796
		page = __page_cache_alloc(gfp_mask);
		if (!page)
			return NULL;
N
Nick Piggin 已提交
797 798 799 800 801 802 803 804
		/*
		 * We want a regular kernel memory (not highmem or DMA etc)
		 * allocation for the radix tree nodes, but we need to honour
		 * the context-specific requirements the caller has asked for.
		 * GFP_RECLAIM_MASK collects those requirements.
		 */
		err = add_to_page_cache_lru(page, mapping, index,
			(gfp_mask & GFP_RECLAIM_MASK));
N
Nick Piggin 已提交
805 806 807 808 809
		if (unlikely(err)) {
			page_cache_release(page);
			page = NULL;
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
		}
	}
	return page;
}
EXPORT_SYMBOL(find_or_create_page);

/**
 * find_get_pages - gang pagecache lookup
 * @mapping:	The address_space to search
 * @start:	The starting page index
 * @nr_pages:	The maximum number of pages
 * @pages:	Where the resulting pages are placed
 *
 * find_get_pages() will search for and return a group of up to
 * @nr_pages pages in the mapping.  The pages are placed at @pages.
 * find_get_pages() takes a reference against the returned pages.
 *
 * The search returns a group of mapping-contiguous pages with ascending
 * indexes.  There may be holes in the indices due to not-present pages.
 *
 * find_get_pages() returns the number of pages which were found.
 */
unsigned find_get_pages(struct address_space *mapping, pgoff_t start,
			    unsigned int nr_pages, struct page **pages)
{
835 836 837 838 839 840
	struct radix_tree_iter iter;
	void **slot;
	unsigned ret = 0;

	if (unlikely(!nr_pages))
		return 0;
N
Nick Piggin 已提交
841 842 843

	rcu_read_lock();
restart:
844
	radix_tree_for_each_slot(slot, &mapping->page_tree, &iter, start) {
N
Nick Piggin 已提交
845 846
		struct page *page;
repeat:
847
		page = radix_tree_deref_slot(slot);
N
Nick Piggin 已提交
848 849
		if (unlikely(!page))
			continue;
850

851
		if (radix_tree_exception(page)) {
852 853 854 855 856 857
			if (radix_tree_deref_retry(page)) {
				/*
				 * Transient condition which can only trigger
				 * when entry at index 0 moves out of or back
				 * to root: none yet gotten, safe to restart.
				 */
858
				WARN_ON(iter.index);
859 860
				goto restart;
			}
861
			/*
862 863 864
			 * Otherwise, shmem/tmpfs must be storing a swap entry
			 * here as an exceptional entry: so skip over it -
			 * we only reach this from invalidate_mapping_pages().
865
			 */
866
			continue;
N
Nick Piggin 已提交
867
		}
N
Nick Piggin 已提交
868 869 870 871 872

		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
873
		if (unlikely(page != *slot)) {
N
Nick Piggin 已提交
874 875 876
			page_cache_release(page);
			goto repeat;
		}
L
Linus Torvalds 已提交
877

N
Nick Piggin 已提交
878
		pages[ret] = page;
879 880
		if (++ret == nr_pages)
			break;
N
Nick Piggin 已提交
881
	}
882

N
Nick Piggin 已提交
883
	rcu_read_unlock();
L
Linus Torvalds 已提交
884 885 886
	return ret;
}

887 888 889 890 891 892 893 894 895 896 897 898 899 900 901
/**
 * find_get_pages_contig - gang contiguous pagecache lookup
 * @mapping:	The address_space to search
 * @index:	The starting page index
 * @nr_pages:	The maximum number of pages
 * @pages:	Where the resulting pages are placed
 *
 * find_get_pages_contig() works exactly like find_get_pages(), except
 * that the returned number of pages are guaranteed to be contiguous.
 *
 * find_get_pages_contig() returns the number of pages which were found.
 */
unsigned find_get_pages_contig(struct address_space *mapping, pgoff_t index,
			       unsigned int nr_pages, struct page **pages)
{
902 903 904 905 906 907
	struct radix_tree_iter iter;
	void **slot;
	unsigned int ret = 0;

	if (unlikely(!nr_pages))
		return 0;
N
Nick Piggin 已提交
908 909 910

	rcu_read_lock();
restart:
911
	radix_tree_for_each_contig(slot, &mapping->page_tree, &iter, index) {
N
Nick Piggin 已提交
912 913
		struct page *page;
repeat:
914 915
		page = radix_tree_deref_slot(slot);
		/* The hole, there no reason to continue */
N
Nick Piggin 已提交
916
		if (unlikely(!page))
917
			break;
918

919
		if (radix_tree_exception(page)) {
920 921 922 923 924 925 926 927
			if (radix_tree_deref_retry(page)) {
				/*
				 * Transient condition which can only trigger
				 * when entry at index 0 moves out of or back
				 * to root: none yet gotten, safe to restart.
				 */
				goto restart;
			}
928
			/*
929 930 931
			 * Otherwise, shmem/tmpfs must be storing a swap entry
			 * here as an exceptional entry: so stop looking for
			 * contiguous pages.
932
			 */
933
			break;
934
		}
935

N
Nick Piggin 已提交
936 937 938 939
		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
940
		if (unlikely(page != *slot)) {
N
Nick Piggin 已提交
941 942 943 944
			page_cache_release(page);
			goto repeat;
		}

N
Nick Piggin 已提交
945 946 947 948 949
		/*
		 * must check mapping and index after taking the ref.
		 * otherwise we can get both false positives and false
		 * negatives, which is just confusing to the caller.
		 */
950
		if (page->mapping == NULL || page->index != iter.index) {
N
Nick Piggin 已提交
951 952 953 954
			page_cache_release(page);
			break;
		}

N
Nick Piggin 已提交
955
		pages[ret] = page;
956 957
		if (++ret == nr_pages)
			break;
958
	}
N
Nick Piggin 已提交
959 960
	rcu_read_unlock();
	return ret;
961
}
962
EXPORT_SYMBOL(find_get_pages_contig);
963

964 965 966 967 968 969 970 971
/**
 * find_get_pages_tag - find and return pages that match @tag
 * @mapping:	the address_space to search
 * @index:	the starting page index
 * @tag:	the tag index
 * @nr_pages:	the maximum number of pages
 * @pages:	where the resulting pages are placed
 *
L
Linus Torvalds 已提交
972
 * Like find_get_pages, except we only return pages which are tagged with
973
 * @tag.   We update @index to index the next page for the traversal.
L
Linus Torvalds 已提交
974 975 976 977
 */
unsigned find_get_pages_tag(struct address_space *mapping, pgoff_t *index,
			int tag, unsigned int nr_pages, struct page **pages)
{
978 979 980 981 982 983
	struct radix_tree_iter iter;
	void **slot;
	unsigned ret = 0;

	if (unlikely(!nr_pages))
		return 0;
N
Nick Piggin 已提交
984 985 986

	rcu_read_lock();
restart:
987 988
	radix_tree_for_each_tagged(slot, &mapping->page_tree,
				   &iter, *index, tag) {
N
Nick Piggin 已提交
989 990
		struct page *page;
repeat:
991
		page = radix_tree_deref_slot(slot);
N
Nick Piggin 已提交
992 993
		if (unlikely(!page))
			continue;
994

995
		if (radix_tree_exception(page)) {
996 997 998 999 1000 1001 1002 1003
			if (radix_tree_deref_retry(page)) {
				/*
				 * Transient condition which can only trigger
				 * when entry at index 0 moves out of or back
				 * to root: none yet gotten, safe to restart.
				 */
				goto restart;
			}
1004
			/*
1005 1006
			 * This function is never used on a shmem/tmpfs
			 * mapping, so a swap entry won't be found here.
1007
			 */
1008
			BUG();
1009
		}
N
Nick Piggin 已提交
1010 1011 1012 1013 1014

		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
1015
		if (unlikely(page != *slot)) {
N
Nick Piggin 已提交
1016 1017 1018 1019 1020
			page_cache_release(page);
			goto repeat;
		}

		pages[ret] = page;
1021 1022
		if (++ret == nr_pages)
			break;
N
Nick Piggin 已提交
1023
	}
1024

N
Nick Piggin 已提交
1025
	rcu_read_unlock();
L
Linus Torvalds 已提交
1026 1027 1028

	if (ret)
		*index = pages[ret - 1]->index + 1;
N
Nick Piggin 已提交
1029

L
Linus Torvalds 已提交
1030 1031
	return ret;
}
1032
EXPORT_SYMBOL(find_get_pages_tag);
L
Linus Torvalds 已提交
1033

1034 1035 1036 1037 1038
/**
 * grab_cache_page_nowait - returns locked page at given index in given cache
 * @mapping: target address_space
 * @index: the page index
 *
1039
 * Same as grab_cache_page(), but do not wait if the page is unavailable.
L
Linus Torvalds 已提交
1040 1041 1042 1043 1044 1045 1046 1047
 * This is intended for speculative data generators, where the data can
 * be regenerated if the page couldn't be grabbed.  This routine should
 * be safe to call while holding the lock for another page.
 *
 * Clear __GFP_FS when allocating the page to avoid recursion into the fs
 * and deadlock against the caller's locked page.
 */
struct page *
1048
grab_cache_page_nowait(struct address_space *mapping, pgoff_t index)
L
Linus Torvalds 已提交
1049 1050 1051 1052
{
	struct page *page = find_get_page(mapping, index);

	if (page) {
N
Nick Piggin 已提交
1053
		if (trylock_page(page))
L
Linus Torvalds 已提交
1054 1055 1056 1057
			return page;
		page_cache_release(page);
		return NULL;
	}
1058
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~__GFP_FS);
N
Nick Piggin 已提交
1059
	if (page && add_to_page_cache_lru(page, mapping, index, GFP_NOFS)) {
L
Linus Torvalds 已提交
1060 1061 1062 1063 1064 1065 1066
		page_cache_release(page);
		page = NULL;
	}
	return page;
}
EXPORT_SYMBOL(grab_cache_page_nowait);

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
/*
 * CD/DVDs are error prone. When a medium error occurs, the driver may fail
 * a _large_ part of the i/o request. Imagine the worst scenario:
 *
 *      ---R__________________________________________B__________
 *         ^ reading here                             ^ bad block(assume 4k)
 *
 * read(R) => miss => readahead(R...B) => media error => frustrating retries
 * => failing the whole request => read(R) => read(R+1) =>
 * readahead(R+1...B+1) => bang => read(R+2) => read(R+3) =>
 * readahead(R+3...B+2) => bang => read(R+3) => read(R+4) =>
 * readahead(R+4...B+3) => bang => read(R+4) => read(R+5) => ......
 *
 * It is going insane. Fix it by quickly scaling down the readahead size.
 */
static void shrink_readahead_size_eio(struct file *filp,
					struct file_ra_state *ra)
{
	ra->ra_pages /= 4;
}

1088
/**
C
Christoph Hellwig 已提交
1089
 * do_generic_file_read - generic file read routine
1090 1091 1092 1093
 * @filp:	the file to read
 * @ppos:	current file position
 * @desc:	read_descriptor
 *
L
Linus Torvalds 已提交
1094
 * This is a generic file read routine, and uses the
1095
 * mapping->a_ops->readpage() function for the actual low-level stuff.
L
Linus Torvalds 已提交
1096 1097 1098 1099
 *
 * This is really ugly. But the goto's actually try to clarify some
 * of the logic when it comes to error handling etc.
 */
C
Christoph Hellwig 已提交
1100
static void do_generic_file_read(struct file *filp, loff_t *ppos,
1101
		read_descriptor_t *desc)
L
Linus Torvalds 已提交
1102
{
C
Christoph Hellwig 已提交
1103
	struct address_space *mapping = filp->f_mapping;
L
Linus Torvalds 已提交
1104
	struct inode *inode = mapping->host;
C
Christoph Hellwig 已提交
1105
	struct file_ra_state *ra = &filp->f_ra;
1106 1107 1108 1109
	pgoff_t index;
	pgoff_t last_index;
	pgoff_t prev_index;
	unsigned long offset;      /* offset into pagecache page */
1110
	unsigned int prev_offset;
L
Linus Torvalds 已提交
1111 1112 1113
	int error;

	index = *ppos >> PAGE_CACHE_SHIFT;
1114 1115
	prev_index = ra->prev_pos >> PAGE_CACHE_SHIFT;
	prev_offset = ra->prev_pos & (PAGE_CACHE_SIZE-1);
L
Linus Torvalds 已提交
1116 1117 1118 1119 1120
	last_index = (*ppos + desc->count + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	for (;;) {
		struct page *page;
1121
		pgoff_t end_index;
N
NeilBrown 已提交
1122
		loff_t isize;
L
Linus Torvalds 已提交
1123 1124 1125 1126 1127
		unsigned long nr, ret;

		cond_resched();
find_page:
		page = find_get_page(mapping, index);
1128
		if (!page) {
1129
			page_cache_sync_readahead(mapping,
1130
					ra, filp,
1131 1132 1133 1134 1135 1136
					index, last_index - index);
			page = find_get_page(mapping, index);
			if (unlikely(page == NULL))
				goto no_cached_page;
		}
		if (PageReadahead(page)) {
1137
			page_cache_async_readahead(mapping,
1138
					ra, filp, page,
1139
					index, last_index - index);
L
Linus Torvalds 已提交
1140
		}
1141 1142 1143 1144
		if (!PageUptodate(page)) {
			if (inode->i_blkbits == PAGE_CACHE_SHIFT ||
					!mapping->a_ops->is_partially_uptodate)
				goto page_not_up_to_date;
N
Nick Piggin 已提交
1145
			if (!trylock_page(page))
1146
				goto page_not_up_to_date;
1147 1148 1149
			/* Did it get truncated before we got the lock? */
			if (!page->mapping)
				goto page_not_up_to_date_locked;
1150 1151 1152 1153 1154
			if (!mapping->a_ops->is_partially_uptodate(page,
								desc, offset))
				goto page_not_up_to_date_locked;
			unlock_page(page);
		}
L
Linus Torvalds 已提交
1155
page_ok:
N
NeilBrown 已提交
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
		/*
		 * i_size must be checked after we know the page is Uptodate.
		 *
		 * Checking i_size after the check allows us to calculate
		 * the correct value for "nr", which means the zero-filled
		 * part of the page is not copied back to userspace (unless
		 * another truncate extends the file - this is desired though).
		 */

		isize = i_size_read(inode);
		end_index = (isize - 1) >> PAGE_CACHE_SHIFT;
		if (unlikely(!isize || index > end_index)) {
			page_cache_release(page);
			goto out;
		}

		/* nr is the maximum number of bytes to copy from this page */
		nr = PAGE_CACHE_SIZE;
		if (index == end_index) {
			nr = ((isize - 1) & ~PAGE_CACHE_MASK) + 1;
			if (nr <= offset) {
				page_cache_release(page);
				goto out;
			}
		}
		nr = nr - offset;
L
Linus Torvalds 已提交
1182 1183 1184 1185 1186 1187 1188 1189 1190

		/* If users can be writing to this page using arbitrary
		 * virtual addresses, take care about potential aliasing
		 * before reading the page on the kernel side.
		 */
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

		/*
1191 1192
		 * When a sequential read accesses a page several times,
		 * only mark it as accessed the first time.
L
Linus Torvalds 已提交
1193
		 */
1194
		if (prev_index != index || offset != prev_offset)
L
Linus Torvalds 已提交
1195 1196 1197 1198 1199 1200 1201
			mark_page_accessed(page);
		prev_index = index;

		/*
		 * Ok, we have the page, and it's up-to-date, so
		 * now we can copy it to user space...
		 *
1202 1203
		 * The file_read_actor routine returns how many bytes were
		 * actually used..
L
Linus Torvalds 已提交
1204 1205 1206 1207 1208
		 * NOTE! This may not be the same as how much of a user buffer
		 * we filled up (we may be padding etc), so we can only update
		 * "pos" here (the actor routine has to update the user buffer
		 * pointers and the remaining count).
		 */
1209
		ret = file_read_actor(desc, page, offset, nr);
L
Linus Torvalds 已提交
1210 1211 1212
		offset += ret;
		index += offset >> PAGE_CACHE_SHIFT;
		offset &= ~PAGE_CACHE_MASK;
J
Jan Kara 已提交
1213
		prev_offset = offset;
L
Linus Torvalds 已提交
1214 1215 1216 1217 1218 1219 1220 1221

		page_cache_release(page);
		if (ret == nr && desc->count)
			continue;
		goto out;

page_not_up_to_date:
		/* Get exclusive access to the page ... */
1222 1223 1224
		error = lock_page_killable(page);
		if (unlikely(error))
			goto readpage_error;
L
Linus Torvalds 已提交
1225

1226
page_not_up_to_date_locked:
N
Nick Piggin 已提交
1227
		/* Did it get truncated before we got the lock? */
L
Linus Torvalds 已提交
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
		if (!page->mapping) {
			unlock_page(page);
			page_cache_release(page);
			continue;
		}

		/* Did somebody else fill it already? */
		if (PageUptodate(page)) {
			unlock_page(page);
			goto page_ok;
		}

readpage:
1241 1242 1243 1244 1245 1246
		/*
		 * A previous I/O error may have been due to temporary
		 * failures, eg. multipath errors.
		 * PG_error will be set again if readpage fails.
		 */
		ClearPageError(page);
L
Linus Torvalds 已提交
1247 1248 1249
		/* Start the actual read. The read will unlock the page. */
		error = mapping->a_ops->readpage(filp, page);

1250 1251 1252 1253 1254
		if (unlikely(error)) {
			if (error == AOP_TRUNCATED_PAGE) {
				page_cache_release(page);
				goto find_page;
			}
L
Linus Torvalds 已提交
1255
			goto readpage_error;
1256
		}
L
Linus Torvalds 已提交
1257 1258

		if (!PageUptodate(page)) {
1259 1260 1261
			error = lock_page_killable(page);
			if (unlikely(error))
				goto readpage_error;
L
Linus Torvalds 已提交
1262 1263 1264
			if (!PageUptodate(page)) {
				if (page->mapping == NULL) {
					/*
1265
					 * invalidate_mapping_pages got it
L
Linus Torvalds 已提交
1266 1267 1268 1269 1270 1271
					 */
					unlock_page(page);
					page_cache_release(page);
					goto find_page;
				}
				unlock_page(page);
1272
				shrink_readahead_size_eio(filp, ra);
1273 1274
				error = -EIO;
				goto readpage_error;
L
Linus Torvalds 已提交
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
			}
			unlock_page(page);
		}

		goto page_ok;

readpage_error:
		/* UHHUH! A synchronous read error occurred. Report it */
		desc->error = error;
		page_cache_release(page);
		goto out;

no_cached_page:
		/*
		 * Ok, it wasn't cached, so we need to create a new
		 * page..
		 */
N
Nick Piggin 已提交
1292 1293 1294 1295
		page = page_cache_alloc_cold(mapping);
		if (!page) {
			desc->error = -ENOMEM;
			goto out;
L
Linus Torvalds 已提交
1296
		}
N
Nick Piggin 已提交
1297
		error = add_to_page_cache_lru(page, mapping,
L
Linus Torvalds 已提交
1298 1299
						index, GFP_KERNEL);
		if (error) {
N
Nick Piggin 已提交
1300
			page_cache_release(page);
L
Linus Torvalds 已提交
1301 1302 1303 1304 1305 1306 1307 1308 1309
			if (error == -EEXIST)
				goto find_page;
			desc->error = error;
			goto out;
		}
		goto readpage;
	}

out:
1310 1311 1312
	ra->prev_pos = prev_index;
	ra->prev_pos <<= PAGE_CACHE_SHIFT;
	ra->prev_pos |= prev_offset;
L
Linus Torvalds 已提交
1313

1314
	*ppos = ((loff_t)index << PAGE_CACHE_SHIFT) + offset;
1315
	file_accessed(filp);
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
}

int file_read_actor(read_descriptor_t *desc, struct page *page,
			unsigned long offset, unsigned long size)
{
	char *kaddr;
	unsigned long left, count = desc->count;

	if (size > count)
		size = count;

	/*
	 * Faults on the destination of a read are common, so do it before
	 * taking the kmap.
	 */
	if (!fault_in_pages_writeable(desc->arg.buf, size)) {
1332
		kaddr = kmap_atomic(page);
L
Linus Torvalds 已提交
1333 1334
		left = __copy_to_user_inatomic(desc->arg.buf,
						kaddr + offset, size);
1335
		kunmap_atomic(kaddr);
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355
		if (left == 0)
			goto success;
	}

	/* Do it the slow way */
	kaddr = kmap(page);
	left = __copy_to_user(desc->arg.buf, kaddr + offset, size);
	kunmap(page);

	if (left) {
		size -= left;
		desc->error = -EFAULT;
	}
success:
	desc->count = count - size;
	desc->written += size;
	desc->arg.buf += size;
	return size;
}

1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
/*
 * Performs necessary checks before doing a write
 * @iov:	io vector request
 * @nr_segs:	number of segments in the iovec
 * @count:	number of bytes to write
 * @access_flags: type of access: %VERIFY_READ or %VERIFY_WRITE
 *
 * Adjust number of segments and amount of bytes to write (nr_segs should be
 * properly initialized first). Returns appropriate error code that caller
 * should return or zero in case that write should be allowed.
 */
int generic_segment_checks(const struct iovec *iov,
			unsigned long *nr_segs, size_t *count, int access_flags)
{
	unsigned long   seg;
	size_t cnt = 0;
	for (seg = 0; seg < *nr_segs; seg++) {
		const struct iovec *iv = &iov[seg];

		/*
		 * If any segment has a negative length, or the cumulative
		 * length ever wraps negative then return -EINVAL.
		 */
		cnt += iv->iov_len;
		if (unlikely((ssize_t)(cnt|iv->iov_len) < 0))
			return -EINVAL;
		if (access_ok(access_flags, iv->iov_base, iv->iov_len))
			continue;
		if (seg == 0)
			return -EFAULT;
		*nr_segs = seg;
		cnt -= iv->iov_len;	/* This segment is no good */
		break;
	}
	*count = cnt;
	return 0;
}
EXPORT_SYMBOL(generic_segment_checks);

1395
/**
H
Henrik Kretzschmar 已提交
1396
 * generic_file_aio_read - generic filesystem read routine
1397 1398 1399
 * @iocb:	kernel I/O control block
 * @iov:	io vector request
 * @nr_segs:	number of segments in the iovec
H
Henrik Kretzschmar 已提交
1400
 * @pos:	current file position
1401
 *
L
Linus Torvalds 已提交
1402 1403 1404 1405
 * This is the "read()" routine for all filesystems
 * that can use the page cache directly.
 */
ssize_t
1406 1407
generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
1408 1409 1410
{
	struct file *filp = iocb->ki_filp;
	ssize_t retval;
1411
	unsigned long seg = 0;
L
Linus Torvalds 已提交
1412
	size_t count;
1413
	loff_t *ppos = &iocb->ki_pos;
L
Linus Torvalds 已提交
1414 1415

	count = 0;
1416 1417 1418
	retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
	if (retval)
		return retval;
L
Linus Torvalds 已提交
1419 1420 1421

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (filp->f_flags & O_DIRECT) {
1422
		loff_t size;
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430
		struct address_space *mapping;
		struct inode *inode;

		mapping = filp->f_mapping;
		inode = mapping->host;
		if (!count)
			goto out; /* skip atime */
		size = i_size_read(inode);
1431
		retval = filemap_write_and_wait_range(mapping, pos,
1432
					pos + iov_length(iov, nr_segs) - 1);
1433 1434 1435 1436 1437 1438 1439 1440
		if (!retval) {
			retval = mapping->a_ops->direct_IO(READ, iocb,
							   iov, pos, nr_segs);
		}
		if (retval > 0) {
			*ppos = pos + retval;
			count -= retval;
		}
1441

1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
		/*
		 * Btrfs can have a short DIO read if we encounter
		 * compressed extents, so if there was an error, or if
		 * we've already read everything we wanted to, or if
		 * there was a short read because we hit EOF, go ahead
		 * and return.  Otherwise fallthrough to buffered io for
		 * the rest of the read.
		 */
		if (retval < 0 || !count || *ppos >= size) {
			file_accessed(filp);
			goto out;
1453
		}
L
Linus Torvalds 已提交
1454 1455
	}

1456
	count = retval;
H
Hugh Dickins 已提交
1457 1458
	for (seg = 0; seg < nr_segs; seg++) {
		read_descriptor_t desc;
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472
		loff_t offset = 0;

		/*
		 * If we did a short DIO read we need to skip the section of the
		 * iov that we've already read data into.
		 */
		if (count) {
			if (count > iov[seg].iov_len) {
				count -= iov[seg].iov_len;
				continue;
			}
			offset = count;
			count = 0;
		}
L
Linus Torvalds 已提交
1473

H
Hugh Dickins 已提交
1474
		desc.written = 0;
1475 1476
		desc.arg.buf = iov[seg].iov_base + offset;
		desc.count = iov[seg].iov_len - offset;
H
Hugh Dickins 已提交
1477 1478 1479
		if (desc.count == 0)
			continue;
		desc.error = 0;
1480
		do_generic_file_read(filp, ppos, &desc);
H
Hugh Dickins 已提交
1481 1482 1483 1484
		retval += desc.written;
		if (desc.error) {
			retval = retval ?: desc.error;
			break;
L
Linus Torvalds 已提交
1485
		}
H
Hugh Dickins 已提交
1486 1487
		if (desc.count > 0)
			break;
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492 1493 1494
	}
out:
	return retval;
}
EXPORT_SYMBOL(generic_file_aio_read);

#ifdef CONFIG_MMU
1495 1496 1497 1498 1499
/**
 * page_cache_read - adds requested page to the page cache if not already there
 * @file:	file to read
 * @offset:	page index
 *
L
Linus Torvalds 已提交
1500 1501 1502
 * This adds the requested page to the page cache if it isn't already there,
 * and schedules an I/O to read in its contents from disk.
 */
H
Harvey Harrison 已提交
1503
static int page_cache_read(struct file *file, pgoff_t offset)
L
Linus Torvalds 已提交
1504 1505 1506
{
	struct address_space *mapping = file->f_mapping;
	struct page *page; 
1507
	int ret;
L
Linus Torvalds 已提交
1508

1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
	do {
		page = page_cache_alloc_cold(mapping);
		if (!page)
			return -ENOMEM;

		ret = add_to_page_cache_lru(page, mapping, offset, GFP_KERNEL);
		if (ret == 0)
			ret = mapping->a_ops->readpage(file, page);
		else if (ret == -EEXIST)
			ret = 0; /* losing race to add is OK */
L
Linus Torvalds 已提交
1519 1520 1521

		page_cache_release(page);

1522 1523 1524
	} while (ret == AOP_TRUNCATED_PAGE);
		
	return ret;
L
Linus Torvalds 已提交
1525 1526 1527 1528
}

#define MMAP_LOTSAMISS  (100)

1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541
/*
 * Synchronous readahead happens when we don't even find
 * a page in the page cache at all.
 */
static void do_sync_mmap_readahead(struct vm_area_struct *vma,
				   struct file_ra_state *ra,
				   struct file *file,
				   pgoff_t offset)
{
	unsigned long ra_pages;
	struct address_space *mapping = file->f_mapping;

	/* If we don't want any read-ahead, don't bother */
1542
	if (vma->vm_flags & VM_RAND_READ)
1543
		return;
1544 1545
	if (!ra->ra_pages)
		return;
1546

1547
	if (vma->vm_flags & VM_SEQ_READ) {
1548 1549
		page_cache_sync_readahead(mapping, ra, file, offset,
					  ra->ra_pages);
1550 1551 1552
		return;
	}

1553 1554
	/* Avoid banging the cache line if not needed */
	if (ra->mmap_miss < MMAP_LOTSAMISS * 10)
1555 1556 1557 1558 1559 1560 1561 1562 1563
		ra->mmap_miss++;

	/*
	 * Do we miss much more than hit in this file? If so,
	 * stop bothering with read-ahead. It will only hurt.
	 */
	if (ra->mmap_miss > MMAP_LOTSAMISS)
		return;

1564 1565 1566
	/*
	 * mmap read-around
	 */
1567
	ra_pages = max_sane_readahead(ra->ra_pages);
1568 1569
	ra->start = max_t(long, 0, offset - ra_pages / 2);
	ra->size = ra_pages;
1570
	ra->async_size = ra_pages / 4;
1571
	ra_submit(ra, mapping, file);
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
}

/*
 * Asynchronous readahead happens when we find the page and PG_readahead,
 * so we want to possibly extend the readahead further..
 */
static void do_async_mmap_readahead(struct vm_area_struct *vma,
				    struct file_ra_state *ra,
				    struct file *file,
				    struct page *page,
				    pgoff_t offset)
{
	struct address_space *mapping = file->f_mapping;

	/* If we don't want any read-ahead, don't bother */
1587
	if (vma->vm_flags & VM_RAND_READ)
1588 1589 1590 1591
		return;
	if (ra->mmap_miss > 0)
		ra->mmap_miss--;
	if (PageReadahead(page))
1592 1593
		page_cache_async_readahead(mapping, ra, file,
					   page, offset, ra->ra_pages);
1594 1595
}

1596
/**
1597
 * filemap_fault - read in file data for page fault handling
N
Nick Piggin 已提交
1598 1599
 * @vma:	vma in which the fault was taken
 * @vmf:	struct vm_fault containing details of the fault
1600
 *
1601
 * filemap_fault() is invoked via the vma operations vector for a
L
Linus Torvalds 已提交
1602 1603 1604 1605 1606 1607
 * mapped memory region to read in file data during a page fault.
 *
 * The goto's are kind of ugly, but this streamlines the normal case of having
 * it in the page cache, and handles the special cases reasonably without
 * having a lot of duplicated code.
 */
N
Nick Piggin 已提交
1608
int filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1609 1610
{
	int error;
1611
	struct file *file = vma->vm_file;
L
Linus Torvalds 已提交
1612 1613 1614
	struct address_space *mapping = file->f_mapping;
	struct file_ra_state *ra = &file->f_ra;
	struct inode *inode = mapping->host;
1615
	pgoff_t offset = vmf->pgoff;
L
Linus Torvalds 已提交
1616
	struct page *page;
J
Jan Kara 已提交
1617
	pgoff_t size;
N
Nick Piggin 已提交
1618
	int ret = 0;
L
Linus Torvalds 已提交
1619 1620

	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1621
	if (offset >= size)
1622
		return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1623 1624

	/*
1625
	 * Do we have something in the page cache already?
L
Linus Torvalds 已提交
1626
	 */
1627
	page = find_get_page(mapping, offset);
1628
	if (likely(page) && !(vmf->flags & FAULT_FLAG_TRIED)) {
L
Linus Torvalds 已提交
1629
		/*
1630 1631
		 * We found the page, so try async readahead before
		 * waiting for the lock.
L
Linus Torvalds 已提交
1632
		 */
1633
		do_async_mmap_readahead(vma, ra, file, page, offset);
1634
	} else if (!page) {
1635 1636 1637
		/* No page in the page cache at all */
		do_sync_mmap_readahead(vma, ra, file, offset);
		count_vm_event(PGMAJFAULT);
1638
		mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT);
1639 1640
		ret = VM_FAULT_MAJOR;
retry_find:
1641
		page = find_get_page(mapping, offset);
L
Linus Torvalds 已提交
1642 1643 1644 1645
		if (!page)
			goto no_cached_page;
	}

1646 1647
	if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags)) {
		page_cache_release(page);
1648
		return ret | VM_FAULT_RETRY;
1649
	}
1650 1651 1652 1653 1654 1655 1656

	/* Did it get truncated? */
	if (unlikely(page->mapping != mapping)) {
		unlock_page(page);
		put_page(page);
		goto retry_find;
	}
1657
	VM_BUG_ON_PAGE(page->index != offset, page);
1658

L
Linus Torvalds 已提交
1659
	/*
1660 1661
	 * We have a locked page in the page cache, now we need to check
	 * that it's up-to-date. If not, it is going to be due to an error.
L
Linus Torvalds 已提交
1662
	 */
1663
	if (unlikely(!PageUptodate(page)))
L
Linus Torvalds 已提交
1664 1665
		goto page_not_uptodate;

1666 1667 1668 1669
	/*
	 * Found the page and have a reference on it.
	 * We must recheck i_size under page lock.
	 */
1670
	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1671
	if (unlikely(offset >= size)) {
1672
		unlock_page(page);
1673
		page_cache_release(page);
1674
		return VM_FAULT_SIGBUS;
1675 1676
	}

N
Nick Piggin 已提交
1677
	vmf->page = page;
N
Nick Piggin 已提交
1678
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1679 1680 1681 1682 1683 1684

no_cached_page:
	/*
	 * We're only likely to ever get here if MADV_RANDOM is in
	 * effect.
	 */
1685
	error = page_cache_read(file, offset);
L
Linus Torvalds 已提交
1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700

	/*
	 * The page we want has now been added to the page cache.
	 * In the unlikely event that someone removed it in the
	 * meantime, we'll just come back here and read it again.
	 */
	if (error >= 0)
		goto retry_find;

	/*
	 * An error return from page_cache_read can result if the
	 * system is low on memory, or a problem occurs while trying
	 * to schedule I/O.
	 */
	if (error == -ENOMEM)
N
Nick Piggin 已提交
1701 1702
		return VM_FAULT_OOM;
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1703 1704 1705 1706 1707 1708 1709 1710 1711

page_not_uptodate:
	/*
	 * Umm, take care of errors if the page isn't up-to-date.
	 * Try to re-read it _once_. We do this synchronously,
	 * because there really aren't any performance issues here
	 * and we need to check for errors.
	 */
	ClearPageError(page);
1712
	error = mapping->a_ops->readpage(file, page);
1713 1714 1715 1716 1717
	if (!error) {
		wait_on_page_locked(page);
		if (!PageUptodate(page))
			error = -EIO;
	}
1718 1719 1720
	page_cache_release(page);

	if (!error || error == AOP_TRUNCATED_PAGE)
1721
		goto retry_find;
L
Linus Torvalds 已提交
1722

1723
	/* Things didn't work out. Return zero to tell the mm layer so. */
1724
	shrink_readahead_size_eio(file, ra);
N
Nick Piggin 已提交
1725
	return VM_FAULT_SIGBUS;
1726 1727 1728
}
EXPORT_SYMBOL(filemap_fault);

1729 1730 1731
int filemap_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
	struct page *page = vmf->page;
A
Al Viro 已提交
1732
	struct inode *inode = file_inode(vma->vm_file);
1733 1734
	int ret = VM_FAULT_LOCKED;

1735
	sb_start_pagefault(inode->i_sb);
1736 1737 1738 1739 1740 1741 1742
	file_update_time(vma->vm_file);
	lock_page(page);
	if (page->mapping != inode->i_mapping) {
		unlock_page(page);
		ret = VM_FAULT_NOPAGE;
		goto out;
	}
1743 1744 1745 1746 1747 1748
	/*
	 * We mark the page dirty already here so that when freeze is in
	 * progress, we are guaranteed that writeback during freezing will
	 * see the dirty page and writeprotect it again.
	 */
	set_page_dirty(page);
1749
	wait_for_stable_page(page);
1750
out:
1751
	sb_end_pagefault(inode->i_sb);
1752 1753 1754 1755
	return ret;
}
EXPORT_SYMBOL(filemap_page_mkwrite);

1756
const struct vm_operations_struct generic_file_vm_ops = {
1757
	.fault		= filemap_fault,
1758
	.page_mkwrite	= filemap_page_mkwrite,
1759
	.remap_pages	= generic_file_remap_pages,
L
Linus Torvalds 已提交
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797
};

/* This is used for a general mmap of a disk file */

int generic_file_mmap(struct file * file, struct vm_area_struct * vma)
{
	struct address_space *mapping = file->f_mapping;

	if (!mapping->a_ops->readpage)
		return -ENOEXEC;
	file_accessed(file);
	vma->vm_ops = &generic_file_vm_ops;
	return 0;
}

/*
 * This is for filesystems which do not implement ->writepage.
 */
int generic_file_readonly_mmap(struct file *file, struct vm_area_struct *vma)
{
	if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
		return -EINVAL;
	return generic_file_mmap(file, vma);
}
#else
int generic_file_mmap(struct file * file, struct vm_area_struct * vma)
{
	return -ENOSYS;
}
int generic_file_readonly_mmap(struct file * file, struct vm_area_struct * vma)
{
	return -ENOSYS;
}
#endif /* CONFIG_MMU */

EXPORT_SYMBOL(generic_file_mmap);
EXPORT_SYMBOL(generic_file_readonly_mmap);

1798
static struct page *__read_cache_page(struct address_space *mapping,
1799
				pgoff_t index,
1800
				int (*filler)(void *, struct page *),
1801 1802
				void *data,
				gfp_t gfp)
L
Linus Torvalds 已提交
1803
{
N
Nick Piggin 已提交
1804
	struct page *page;
L
Linus Torvalds 已提交
1805 1806 1807 1808
	int err;
repeat:
	page = find_get_page(mapping, index);
	if (!page) {
1809
		page = __page_cache_alloc(gfp | __GFP_COLD);
N
Nick Piggin 已提交
1810 1811
		if (!page)
			return ERR_PTR(-ENOMEM);
1812
		err = add_to_page_cache_lru(page, mapping, index, gfp);
N
Nick Piggin 已提交
1813 1814 1815 1816
		if (unlikely(err)) {
			page_cache_release(page);
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
			/* Presumably ENOMEM for radix tree node */
			return ERR_PTR(err);
		}
		err = filler(data, page);
		if (err < 0) {
			page_cache_release(page);
			page = ERR_PTR(err);
		}
	}
	return page;
}

1829
static struct page *do_read_cache_page(struct address_space *mapping,
1830
				pgoff_t index,
1831
				int (*filler)(void *, struct page *),
1832 1833 1834
				void *data,
				gfp_t gfp)

L
Linus Torvalds 已提交
1835 1836 1837 1838 1839
{
	struct page *page;
	int err;

retry:
1840
	page = __read_cache_page(mapping, index, filler, data, gfp);
L
Linus Torvalds 已提交
1841
	if (IS_ERR(page))
1842
		return page;
L
Linus Torvalds 已提交
1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858
	if (PageUptodate(page))
		goto out;

	lock_page(page);
	if (!page->mapping) {
		unlock_page(page);
		page_cache_release(page);
		goto retry;
	}
	if (PageUptodate(page)) {
		unlock_page(page);
		goto out;
	}
	err = filler(data, page);
	if (err < 0) {
		page_cache_release(page);
1859
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1860
	}
1861
out:
1862 1863 1864
	mark_page_accessed(page);
	return page;
}
1865 1866 1867 1868 1869 1870

/**
 * read_cache_page_async - read into page cache, fill it if needed
 * @mapping:	the page's address_space
 * @index:	the page index
 * @filler:	function to perform the read
1871
 * @data:	first arg to filler(data, page) function, often left as NULL
1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
 *
 * Same as read_cache_page, but don't wait for page to become unlocked
 * after submitting it to the filler.
 *
 * Read into the page cache. If a page already exists, and PageUptodate() is
 * not set, try to fill the page but don't wait for it to become unlocked.
 *
 * If the page does not get brought uptodate, return -EIO.
 */
struct page *read_cache_page_async(struct address_space *mapping,
				pgoff_t index,
1883
				int (*filler)(void *, struct page *),
1884 1885 1886 1887
				void *data)
{
	return do_read_cache_page(mapping, index, filler, data, mapping_gfp_mask(mapping));
}
1888 1889
EXPORT_SYMBOL(read_cache_page_async);

1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908
static struct page *wait_on_page_read(struct page *page)
{
	if (!IS_ERR(page)) {
		wait_on_page_locked(page);
		if (!PageUptodate(page)) {
			page_cache_release(page);
			page = ERR_PTR(-EIO);
		}
	}
	return page;
}

/**
 * read_cache_page_gfp - read into page cache, using specified page allocation flags.
 * @mapping:	the page's address_space
 * @index:	the page index
 * @gfp:	the page allocator flags to use if allocating
 *
 * This is the same as "read_mapping_page(mapping, index, NULL)", but with
1909
 * any new page allocations done using the specified allocation flags.
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
 *
 * If the page does not get brought uptodate, return -EIO.
 */
struct page *read_cache_page_gfp(struct address_space *mapping,
				pgoff_t index,
				gfp_t gfp)
{
	filler_t *filler = (filler_t *)mapping->a_ops->readpage;

	return wait_on_page_read(do_read_cache_page(mapping, index, filler, NULL, gfp));
}
EXPORT_SYMBOL(read_cache_page_gfp);

1923 1924 1925 1926 1927
/**
 * read_cache_page - read into page cache, fill it if needed
 * @mapping:	the page's address_space
 * @index:	the page index
 * @filler:	function to perform the read
1928
 * @data:	first arg to filler(data, page) function, often left as NULL
1929 1930 1931 1932 1933 1934 1935
 *
 * Read into the page cache. If a page already exists, and PageUptodate() is
 * not set, try to fill the page then wait for it to become unlocked.
 *
 * If the page does not get brought uptodate, return -EIO.
 */
struct page *read_cache_page(struct address_space *mapping,
1936
				pgoff_t index,
1937
				int (*filler)(void *, struct page *),
1938 1939
				void *data)
{
1940
	return wait_on_page_read(read_cache_page_async(mapping, index, filler, data));
L
Linus Torvalds 已提交
1941 1942 1943
}
EXPORT_SYMBOL(read_cache_page);

N
Nick Piggin 已提交
1944
static size_t __iovec_copy_from_user_inatomic(char *vaddr,
L
Linus Torvalds 已提交
1945 1946
			const struct iovec *iov, size_t base, size_t bytes)
{
1947
	size_t copied = 0, left = 0;
L
Linus Torvalds 已提交
1948 1949 1950 1951 1952 1953

	while (bytes) {
		char __user *buf = iov->iov_base + base;
		int copy = min(bytes, iov->iov_len - base);

		base = 0;
1954
		left = __copy_from_user_inatomic(vaddr, buf, copy);
L
Linus Torvalds 已提交
1955 1956 1957 1958 1959
		copied += copy;
		bytes -= copy;
		vaddr += copy;
		iov++;

1960
		if (unlikely(left))
L
Linus Torvalds 已提交
1961 1962 1963 1964 1965
			break;
	}
	return copied - left;
}

N
Nick Piggin 已提交
1966 1967
/*
 * Copy as much as we can into the page and return the number of bytes which
1968
 * were successfully copied.  If a fault is encountered then return the number of
N
Nick Piggin 已提交
1969 1970 1971 1972 1973 1974 1975 1976 1977
 * bytes which were copied.
 */
size_t iov_iter_copy_from_user_atomic(struct page *page,
		struct iov_iter *i, unsigned long offset, size_t bytes)
{
	char *kaddr;
	size_t copied;

	BUG_ON(!in_atomic());
1978
	kaddr = kmap_atomic(page);
N
Nick Piggin 已提交
1979 1980 1981
	if (likely(i->nr_segs == 1)) {
		int left;
		char __user *buf = i->iov->iov_base + i->iov_offset;
1982
		left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
1983 1984 1985 1986 1987
		copied = bytes - left;
	} else {
		copied = __iovec_copy_from_user_inatomic(kaddr + offset,
						i->iov, i->iov_offset, bytes);
	}
1988
	kunmap_atomic(kaddr);
N
Nick Piggin 已提交
1989 1990 1991

	return copied;
}
N
Nick Piggin 已提交
1992
EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
N
Nick Piggin 已提交
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

/*
 * This has the same sideeffects and return value as
 * iov_iter_copy_from_user_atomic().
 * The difference is that it attempts to resolve faults.
 * Page must not be locked.
 */
size_t iov_iter_copy_from_user(struct page *page,
		struct iov_iter *i, unsigned long offset, size_t bytes)
{
	char *kaddr;
	size_t copied;

	kaddr = kmap(page);
	if (likely(i->nr_segs == 1)) {
		int left;
		char __user *buf = i->iov->iov_base + i->iov_offset;
2010
		left = __copy_from_user(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
2011 2012 2013 2014 2015 2016 2017 2018
		copied = bytes - left;
	} else {
		copied = __iovec_copy_from_user_inatomic(kaddr + offset,
						i->iov, i->iov_offset, bytes);
	}
	kunmap(page);
	return copied;
}
N
Nick Piggin 已提交
2019
EXPORT_SYMBOL(iov_iter_copy_from_user);
N
Nick Piggin 已提交
2020

N
Nick Piggin 已提交
2021
void iov_iter_advance(struct iov_iter *i, size_t bytes)
N
Nick Piggin 已提交
2022
{
N
Nick Piggin 已提交
2023 2024
	BUG_ON(i->count < bytes);

N
Nick Piggin 已提交
2025 2026
	if (likely(i->nr_segs == 1)) {
		i->iov_offset += bytes;
N
Nick Piggin 已提交
2027
		i->count -= bytes;
N
Nick Piggin 已提交
2028 2029 2030
	} else {
		const struct iovec *iov = i->iov;
		size_t base = i->iov_offset;
2031
		unsigned long nr_segs = i->nr_segs;
N
Nick Piggin 已提交
2032

2033 2034
		/*
		 * The !iov->iov_len check ensures we skip over unlikely
N
Nick Piggin 已提交
2035
		 * zero-length segments (without overruning the iovec).
2036
		 */
2037
		while (bytes || unlikely(i->count && !iov->iov_len)) {
N
Nick Piggin 已提交
2038
			int copy;
N
Nick Piggin 已提交
2039

N
Nick Piggin 已提交
2040 2041 2042
			copy = min(bytes, iov->iov_len - base);
			BUG_ON(!i->count || i->count < copy);
			i->count -= copy;
N
Nick Piggin 已提交
2043 2044 2045 2046
			bytes -= copy;
			base += copy;
			if (iov->iov_len == base) {
				iov++;
2047
				nr_segs--;
N
Nick Piggin 已提交
2048 2049 2050 2051 2052
				base = 0;
			}
		}
		i->iov = iov;
		i->iov_offset = base;
2053
		i->nr_segs = nr_segs;
N
Nick Piggin 已提交
2054 2055
	}
}
N
Nick Piggin 已提交
2056
EXPORT_SYMBOL(iov_iter_advance);
N
Nick Piggin 已提交
2057

2058 2059 2060 2061 2062 2063 2064 2065 2066 2067
/*
 * Fault in the first iovec of the given iov_iter, to a maximum length
 * of bytes. Returns 0 on success, or non-zero if the memory could not be
 * accessed (ie. because it is an invalid address).
 *
 * writev-intensive code may want this to prefault several iovecs -- that
 * would be possible (callers must not rely on the fact that _only_ the
 * first iovec will be faulted with the current implementation).
 */
int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes)
N
Nick Piggin 已提交
2068 2069
{
	char __user *buf = i->iov->iov_base + i->iov_offset;
2070 2071
	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	return fault_in_pages_readable(buf, bytes);
N
Nick Piggin 已提交
2072
}
N
Nick Piggin 已提交
2073
EXPORT_SYMBOL(iov_iter_fault_in_readable);
N
Nick Piggin 已提交
2074 2075 2076 2077

/*
 * Return the count of just the current iov_iter segment.
 */
2078
size_t iov_iter_single_seg_count(const struct iov_iter *i)
N
Nick Piggin 已提交
2079 2080 2081 2082 2083 2084 2085
{
	const struct iovec *iov = i->iov;
	if (i->nr_segs == 1)
		return i->count;
	else
		return min(i->count, iov->iov_len - i->iov_offset);
}
N
Nick Piggin 已提交
2086
EXPORT_SYMBOL(iov_iter_single_seg_count);
N
Nick Piggin 已提交
2087

L
Linus Torvalds 已提交
2088 2089 2090
/*
 * Performs necessary checks before doing a write
 *
2091
 * Can adjust writing position or amount of bytes to write.
L
Linus Torvalds 已提交
2092 2093 2094 2095 2096 2097
 * Returns appropriate error code that caller should return or
 * zero in case that write should be allowed.
 */
inline int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk)
{
	struct inode *inode = file->f_mapping->host;
J
Jiri Slaby 已提交
2098
	unsigned long limit = rlimit(RLIMIT_FSIZE);
L
Linus Torvalds 已提交
2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149

        if (unlikely(*pos < 0))
                return -EINVAL;

	if (!isblk) {
		/* FIXME: this is for backwards compatibility with 2.4 */
		if (file->f_flags & O_APPEND)
                        *pos = i_size_read(inode);

		if (limit != RLIM_INFINITY) {
			if (*pos >= limit) {
				send_sig(SIGXFSZ, current, 0);
				return -EFBIG;
			}
			if (*count > limit - (typeof(limit))*pos) {
				*count = limit - (typeof(limit))*pos;
			}
		}
	}

	/*
	 * LFS rule
	 */
	if (unlikely(*pos + *count > MAX_NON_LFS &&
				!(file->f_flags & O_LARGEFILE))) {
		if (*pos >= MAX_NON_LFS) {
			return -EFBIG;
		}
		if (*count > MAX_NON_LFS - (unsigned long)*pos) {
			*count = MAX_NON_LFS - (unsigned long)*pos;
		}
	}

	/*
	 * Are we about to exceed the fs block limit ?
	 *
	 * If we have written data it becomes a short write.  If we have
	 * exceeded without writing data we send a signal and return EFBIG.
	 * Linus frestrict idea will clean these up nicely..
	 */
	if (likely(!isblk)) {
		if (unlikely(*pos >= inode->i_sb->s_maxbytes)) {
			if (*count || *pos > inode->i_sb->s_maxbytes) {
				return -EFBIG;
			}
			/* zero-length writes at ->s_maxbytes are OK */
		}

		if (unlikely(*pos + *count > inode->i_sb->s_maxbytes))
			*count = inode->i_sb->s_maxbytes - *pos;
	} else {
2150
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
		loff_t isize;
		if (bdev_read_only(I_BDEV(inode)))
			return -EPERM;
		isize = i_size_read(inode);
		if (*pos >= isize) {
			if (*count || *pos > isize)
				return -ENOSPC;
		}

		if (*pos + *count > isize)
			*count = isize - *pos;
2162 2163 2164
#else
		return -EPERM;
#endif
L
Linus Torvalds 已提交
2165 2166 2167 2168 2169
	}
	return 0;
}
EXPORT_SYMBOL(generic_write_checks);

2170 2171 2172 2173 2174 2175
int pagecache_write_begin(struct file *file, struct address_space *mapping,
				loff_t pos, unsigned len, unsigned flags,
				struct page **pagep, void **fsdata)
{
	const struct address_space_operations *aops = mapping->a_ops;

2176
	return aops->write_begin(file, mapping, pos, len, flags,
2177 2178 2179 2180 2181 2182 2183 2184 2185 2186
							pagep, fsdata);
}
EXPORT_SYMBOL(pagecache_write_begin);

int pagecache_write_end(struct file *file, struct address_space *mapping,
				loff_t pos, unsigned len, unsigned copied,
				struct page *page, void *fsdata)
{
	const struct address_space_operations *aops = mapping->a_ops;

2187 2188
	mark_page_accessed(page);
	return aops->write_end(file, mapping, pos, len, copied, page, fsdata);
2189 2190 2191
}
EXPORT_SYMBOL(pagecache_write_end);

L
Linus Torvalds 已提交
2192 2193 2194 2195 2196 2197 2198 2199 2200
ssize_t
generic_file_direct_write(struct kiocb *iocb, const struct iovec *iov,
		unsigned long *nr_segs, loff_t pos, loff_t *ppos,
		size_t count, size_t ocount)
{
	struct file	*file = iocb->ki_filp;
	struct address_space *mapping = file->f_mapping;
	struct inode	*inode = mapping->host;
	ssize_t		written;
2201 2202
	size_t		write_len;
	pgoff_t		end;
L
Linus Torvalds 已提交
2203 2204 2205 2206

	if (count != ocount)
		*nr_segs = iov_shorten((struct iovec *)iov, *nr_segs, count);

2207 2208 2209
	write_len = iov_length(iov, *nr_segs);
	end = (pos + write_len - 1) >> PAGE_CACHE_SHIFT;

2210
	written = filemap_write_and_wait_range(mapping, pos, pos + write_len - 1);
2211 2212 2213 2214 2215 2216 2217
	if (written)
		goto out;

	/*
	 * After a write we want buffered reads to be sure to go to disk to get
	 * the new data.  We invalidate clean cached page from the region we're
	 * about to write.  We do this *before* the write so that we can return
2218
	 * without clobbering -EIOCBQUEUED from ->direct_IO().
2219 2220 2221 2222
	 */
	if (mapping->nrpages) {
		written = invalidate_inode_pages2_range(mapping,
					pos >> PAGE_CACHE_SHIFT, end);
2223 2224 2225 2226 2227 2228 2229
		/*
		 * If a page can not be invalidated, return 0 to fall back
		 * to buffered write.
		 */
		if (written) {
			if (written == -EBUSY)
				return 0;
2230
			goto out;
2231
		}
2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
	}

	written = mapping->a_ops->direct_IO(WRITE, iocb, iov, pos, *nr_segs);

	/*
	 * Finally, try again to invalidate clean pages which might have been
	 * cached by non-direct readahead, or faulted in by get_user_pages()
	 * if the source of the write was an mmap'ed region of the file
	 * we're writing.  Either one is a pretty crazy thing to do,
	 * so we don't support it 100%.  If this invalidation
	 * fails, tough, the write still worked...
	 */
	if (mapping->nrpages) {
		invalidate_inode_pages2_range(mapping,
					      pos >> PAGE_CACHE_SHIFT, end);
	}

L
Linus Torvalds 已提交
2249
	if (written > 0) {
2250 2251 2252
		pos += written;
		if (pos > i_size_read(inode) && !S_ISBLK(inode->i_mode)) {
			i_size_write(inode, pos);
L
Linus Torvalds 已提交
2253 2254
			mark_inode_dirty(inode);
		}
2255
		*ppos = pos;
L
Linus Torvalds 已提交
2256
	}
2257
out:
L
Linus Torvalds 已提交
2258 2259 2260 2261
	return written;
}
EXPORT_SYMBOL(generic_file_direct_write);

N
Nick Piggin 已提交
2262 2263 2264 2265
/*
 * Find or create a page at the given pagecache position. Return the locked
 * page. This function is specifically for buffered writes.
 */
2266 2267
struct page *grab_cache_page_write_begin(struct address_space *mapping,
					pgoff_t index, unsigned flags)
N
Nick Piggin 已提交
2268 2269
{
	int status;
2270
	gfp_t gfp_mask;
N
Nick Piggin 已提交
2271
	struct page *page;
2272
	gfp_t gfp_notmask = 0;
2273

2274 2275 2276
	gfp_mask = mapping_gfp_mask(mapping);
	if (mapping_cap_account_dirty(mapping))
		gfp_mask |= __GFP_WRITE;
2277 2278
	if (flags & AOP_FLAG_NOFS)
		gfp_notmask = __GFP_FS;
N
Nick Piggin 已提交
2279 2280
repeat:
	page = find_lock_page(mapping, index);
2281
	if (page)
2282
		goto found;
N
Nick Piggin 已提交
2283

2284
	page = __page_cache_alloc(gfp_mask & ~gfp_notmask);
N
Nick Piggin 已提交
2285 2286
	if (!page)
		return NULL;
2287 2288
	status = add_to_page_cache_lru(page, mapping, index,
						GFP_KERNEL & ~gfp_notmask);
N
Nick Piggin 已提交
2289 2290 2291 2292 2293 2294
	if (unlikely(status)) {
		page_cache_release(page);
		if (status == -EEXIST)
			goto repeat;
		return NULL;
	}
2295
found:
2296
	wait_for_stable_page(page);
N
Nick Piggin 已提交
2297 2298
	return page;
}
2299
EXPORT_SYMBOL(grab_cache_page_write_begin);
N
Nick Piggin 已提交
2300

2301 2302 2303 2304 2305 2306 2307
static ssize_t generic_perform_write(struct file *file,
				struct iov_iter *i, loff_t pos)
{
	struct address_space *mapping = file->f_mapping;
	const struct address_space_operations *a_ops = mapping->a_ops;
	long status = 0;
	ssize_t written = 0;
N
Nick Piggin 已提交
2308 2309 2310 2311 2312 2313 2314
	unsigned int flags = 0;

	/*
	 * Copies from kernel address space cannot fail (NFSD is a big user).
	 */
	if (segment_eq(get_fs(), KERNEL_DS))
		flags |= AOP_FLAG_UNINTERRUPTIBLE;
2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342

	do {
		struct page *page;
		unsigned long offset;	/* Offset into pagecache page */
		unsigned long bytes;	/* Bytes to write to page */
		size_t copied;		/* Bytes copied from user */
		void *fsdata;

		offset = (pos & (PAGE_CACHE_SIZE - 1));
		bytes = min_t(unsigned long, PAGE_CACHE_SIZE - offset,
						iov_iter_count(i));

again:
		/*
		 * Bring in the user page that we will copy from _first_.
		 * Otherwise there's a nasty deadlock on copying from the
		 * same page as we're writing to, without it being marked
		 * up-to-date.
		 *
		 * Not only is this an optimisation, but it is also required
		 * to check that the address is actually valid, when atomic
		 * usercopies are used, below.
		 */
		if (unlikely(iov_iter_fault_in_readable(i, bytes))) {
			status = -EFAULT;
			break;
		}

N
Nick Piggin 已提交
2343
		status = a_ops->write_begin(file, mapping, pos, bytes, flags,
2344 2345 2346 2347
						&page, &fsdata);
		if (unlikely(status))
			break;

2348 2349 2350
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

2351 2352 2353 2354 2355
		pagefault_disable();
		copied = iov_iter_copy_from_user_atomic(page, i, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

2356
		mark_page_accessed(page);
2357 2358 2359 2360 2361 2362 2363 2364
		status = a_ops->write_end(file, mapping, pos, bytes, copied,
						page, fsdata);
		if (unlikely(status < 0))
			break;
		copied = status;

		cond_resched();

2365
		iov_iter_advance(i, copied);
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382
		if (unlikely(copied == 0)) {
			/*
			 * If we were unable to copy any data at all, we must
			 * fall back to a single segment length write.
			 *
			 * If we didn't fallback here, we could livelock
			 * because not all segments in the iov can be copied at
			 * once without a pagefault.
			 */
			bytes = min_t(unsigned long, PAGE_CACHE_SIZE - offset,
						iov_iter_single_seg_count(i));
			goto again;
		}
		pos += copied;
		written += copied;

		balance_dirty_pages_ratelimited(mapping);
2383 2384 2385 2386
		if (fatal_signal_pending(current)) {
			status = -EINTR;
			break;
		}
2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401
	} while (iov_iter_count(i));

	return written ? written : status;
}

ssize_t
generic_file_buffered_write(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos, loff_t *ppos,
		size_t count, ssize_t written)
{
	struct file *file = iocb->ki_filp;
	ssize_t status;
	struct iov_iter i;

	iov_iter_init(&i, iov, nr_segs, count, written);
2402
	status = generic_perform_write(file, &i, pos);
L
Linus Torvalds 已提交
2403 2404

	if (likely(status >= 0)) {
2405 2406
		written += status;
		*ppos = pos + status;
L
Linus Torvalds 已提交
2407 2408 2409 2410 2411 2412
  	}
	
	return written ? written : status;
}
EXPORT_SYMBOL(generic_file_buffered_write);

2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433
/**
 * __generic_file_aio_write - write data to a file
 * @iocb:	IO state structure (file, offset, etc.)
 * @iov:	vector with data to write
 * @nr_segs:	number of segments in the vector
 * @ppos:	position where to write
 *
 * This function does all the work needed for actually writing data to a
 * file. It does all basic checks, removes SUID from the file, updates
 * modification times and calls proper subroutines depending on whether we
 * do direct IO or a standard buffered write.
 *
 * It expects i_mutex to be grabbed unless we work on a block device or similar
 * object which does not need locking at all.
 *
 * This function does *not* take care of syncing data in case of O_SYNC write.
 * A caller has to handle it. This is mainly due to the fact that we want to
 * avoid syncing under i_mutex.
 */
ssize_t __generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
				 unsigned long nr_segs, loff_t *ppos)
L
Linus Torvalds 已提交
2434 2435
{
	struct file *file = iocb->ki_filp;
2436
	struct address_space * mapping = file->f_mapping;
L
Linus Torvalds 已提交
2437 2438 2439 2440 2441 2442 2443 2444
	size_t ocount;		/* original count */
	size_t count;		/* after file limit checks */
	struct inode 	*inode = mapping->host;
	loff_t		pos;
	ssize_t		written;
	ssize_t		err;

	ocount = 0;
2445 2446 2447
	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
	if (err)
		return err;
L
Linus Torvalds 已提交
2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462

	count = ocount;
	pos = *ppos;

	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;
	written = 0;

	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
	if (err)
		goto out;

	if (count == 0)
		goto out;

2463
	err = file_remove_suid(file);
L
Linus Torvalds 已提交
2464 2465 2466
	if (err)
		goto out;

2467 2468 2469
	err = file_update_time(file);
	if (err)
		goto out;
L
Linus Torvalds 已提交
2470 2471 2472

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (unlikely(file->f_flags & O_DIRECT)) {
2473 2474 2475 2476 2477
		loff_t endbyte;
		ssize_t written_buffered;

		written = generic_file_direct_write(iocb, iov, &nr_segs, pos,
							ppos, count, ocount);
L
Linus Torvalds 已提交
2478 2479 2480 2481 2482 2483 2484 2485
		if (written < 0 || written == count)
			goto out;
		/*
		 * direct-io write to a hole: fall through to buffered I/O
		 * for completing the rest of the request.
		 */
		pos += written;
		count -= written;
2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499
		written_buffered = generic_file_buffered_write(iocb, iov,
						nr_segs, pos, ppos, count,
						written);
		/*
		 * If generic_file_buffered_write() retuned a synchronous error
		 * then we want to return the number of bytes which were
		 * direct-written, or the error code if that was zero.  Note
		 * that this differs from normal direct-io semantics, which
		 * will return -EFOO even if some bytes were written.
		 */
		if (written_buffered < 0) {
			err = written_buffered;
			goto out;
		}
L
Linus Torvalds 已提交
2500

2501 2502 2503 2504 2505 2506
		/*
		 * We need to ensure that the page cache pages are written to
		 * disk and invalidated to preserve the expected O_DIRECT
		 * semantics.
		 */
		endbyte = pos + written_buffered - written - 1;
2507
		err = filemap_write_and_wait_range(file->f_mapping, pos, endbyte);
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522
		if (err == 0) {
			written = written_buffered;
			invalidate_mapping_pages(mapping,
						 pos >> PAGE_CACHE_SHIFT,
						 endbyte >> PAGE_CACHE_SHIFT);
		} else {
			/*
			 * We don't know how much we wrote, so just return
			 * the number of bytes which were direct-written
			 */
		}
	} else {
		written = generic_file_buffered_write(iocb, iov, nr_segs,
				pos, ppos, count, written);
	}
L
Linus Torvalds 已提交
2523 2524 2525 2526
out:
	current->backing_dev_info = NULL;
	return written ? written : err;
}
2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539
EXPORT_SYMBOL(__generic_file_aio_write);

/**
 * generic_file_aio_write - write data to a file
 * @iocb:	IO state structure
 * @iov:	vector with data to write
 * @nr_segs:	number of segments in the vector
 * @pos:	position in file where to write
 *
 * This is a wrapper around __generic_file_aio_write() to be used by most
 * filesystems. It takes care of syncing the file in case of O_SYNC file
 * and acquires i_mutex as needed.
 */
2540 2541
ssize_t generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
2542 2543
{
	struct file *file = iocb->ki_filp;
2544
	struct inode *inode = file->f_mapping->host;
L
Linus Torvalds 已提交
2545 2546 2547 2548
	ssize_t ret;

	BUG_ON(iocb->ki_pos != pos);

2549
	mutex_lock(&inode->i_mutex);
2550
	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
2551
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
2552

2553
	if (ret > 0) {
L
Linus Torvalds 已提交
2554 2555
		ssize_t err;

2556
		err = generic_write_sync(file, pos, ret);
2557
		if (err < 0 && ret > 0)
L
Linus Torvalds 已提交
2558 2559 2560 2561 2562 2563
			ret = err;
	}
	return ret;
}
EXPORT_SYMBOL(generic_file_aio_write);

2564 2565 2566 2567 2568 2569 2570 2571 2572 2573
/**
 * try_to_release_page() - release old fs-specific metadata on a page
 *
 * @page: the page which the kernel is trying to free
 * @gfp_mask: memory allocation flags (and I/O mode)
 *
 * The address_space is to try to release any data against the page
 * (presumably at page->private).  If the release was successful, return `1'.
 * Otherwise return zero.
 *
2574 2575 2576
 * This may also be called if PG_fscache is set on a page, indicating that the
 * page is known to the local caching routines.
 *
2577
 * The @gfp_mask argument specifies whether I/O may be performed to release
2578
 * this page (__GFP_IO), and whether the call may block (__GFP_WAIT & __GFP_FS).
2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594
 *
 */
int try_to_release_page(struct page *page, gfp_t gfp_mask)
{
	struct address_space * const mapping = page->mapping;

	BUG_ON(!PageLocked(page));
	if (PageWriteback(page))
		return 0;

	if (mapping && mapping->a_ops->releasepage)
		return mapping->a_ops->releasepage(page, gfp_mask);
	return try_to_free_buffers(page);
}

EXPORT_SYMBOL(try_to_release_page);