filemap.c 69.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 *	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)
 */
#include <linux/module.h>
#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 32
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/security.h>
#include <linux/syscalls.h>
33
#include <linux/cpuset.h>
N
Nick Piggin 已提交
34
#include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
35
#include <linux/memcontrol.h>
36
#include <linux/cleancache.h>
37 38
#include "internal.h"

L
Linus Torvalds 已提交
39 40 41
/*
 * FIXME: remove all knowledge of the buffer layer from the core VM
 */
42
#include <linux/buffer_head.h> /* for try_to_free_buffers */
L
Linus Torvalds 已提交
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

#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:
 *
61
 *  ->i_mmap_mutex		(truncate_pagecache)
L
Linus Torvalds 已提交
62
 *    ->private_lock		(__free_pte->__set_page_dirty_buffers)
63 64
 *      ->swap_lock		(exclusive_swap_page, others)
 *        ->mapping->tree_lock
L
Linus Torvalds 已提交
65
 *
66
 *  ->i_mutex
67
 *    ->i_mmap_mutex		(truncate->unmap_mapping_range)
L
Linus Torvalds 已提交
68 69
 *
 *  ->mmap_sem
70
 *    ->i_mmap_mutex
71
 *      ->page_table_lock or pte_lock	(various, mainly in memory.c)
L
Linus Torvalds 已提交
72 73 74 75 76
 *        ->mapping->tree_lock	(arch-dependent flush_dcache_mmap_lock)
 *
 *  ->mmap_sem
 *    ->lock_page		(access_process_vm)
 *
77 78
 *  ->i_mutex			(generic_file_buffered_write)
 *    ->mmap_sem		(fault_in_pages_readable->do_page_fault)
L
Linus Torvalds 已提交
79
 *
80
 *  bdi->wb.list_lock
81
 *    sb_lock			(fs/fs-writeback.c)
L
Linus Torvalds 已提交
82 83
 *    ->mapping->tree_lock	(__sync_single_inode)
 *
84
 *  ->i_mmap_mutex
L
Linus Torvalds 已提交
85 86 87
 *    ->anon_vma.lock		(vma_adjust)
 *
 *  ->anon_vma.lock
88
 *    ->page_table_lock or pte_lock	(anon_vma_prepare and various)
L
Linus Torvalds 已提交
89
 *
90
 *  ->page_table_lock or pte_lock
91
 *    ->swap_lock		(try_to_unmap_one)
L
Linus Torvalds 已提交
92 93 94
 *    ->private_lock		(try_to_unmap_one)
 *    ->tree_lock		(try_to_unmap_one)
 *    ->zone.lru_lock		(follow_page->mark_page_accessed)
95
 *    ->zone.lru_lock		(check_pte_range->isolate_lru_page)
L
Linus Torvalds 已提交
96 97
 *    ->private_lock		(page_remove_rmap->set_page_dirty)
 *    ->tree_lock		(page_remove_rmap->set_page_dirty)
98
 *    bdi.wb->list_lock		(page_remove_rmap->set_page_dirty)
99
 *    ->inode->i_lock		(page_remove_rmap->set_page_dirty)
100
 *    bdi.wb->list_lock		(zap_pte_range->set_page_dirty)
101
 *    ->inode->i_lock		(zap_pte_range->set_page_dirty)
L
Linus Torvalds 已提交
102 103
 *    ->private_lock		(zap_pte_range->__set_page_dirty_buffers)
 *
104 105
 *  (code doesn't rely on that order, so you could switch it around)
 *  ->tasklist_lock             (memory_failure, collect_procs_ao)
106
 *    ->i_mmap_mutex
L
Linus Torvalds 已提交
107 108 109
 */

/*
110
 * Delete a page from the page cache and free it. Caller has to make
L
Linus Torvalds 已提交
111
 * sure the page is locked and that nobody else uses it - or that usage
N
Nick Piggin 已提交
112
 * is safe.  The caller must hold the mapping's tree_lock.
L
Linus Torvalds 已提交
113
 */
114
void __delete_from_page_cache(struct page *page)
L
Linus Torvalds 已提交
115 116 117
{
	struct address_space *mapping = page->mapping;

118 119 120 121 122 123 124 125 126 127
	/*
	 * 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
		cleancache_flush_page(mapping, page);

L
Linus Torvalds 已提交
128 129
	radix_tree_delete(&mapping->page_tree, page->index);
	page->mapping = NULL;
130
	/* Leave page->index set: truncation lookup relies upon it */
L
Linus Torvalds 已提交
131
	mapping->nrpages--;
132
	__dec_zone_page_state(page, NR_FILE_PAGES);
133 134
	if (PageSwapBacked(page))
		__dec_zone_page_state(page, NR_SHMEM);
135
	BUG_ON(page_mapped(page));
136 137 138 139 140 141 142 143 144 145 146 147

	/*
	 * 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
148 149
}

150 151 152 153 154 155 156 157 158
/**
 * 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
159 160
{
	struct address_space *mapping = page->mapping;
161
	void (*freepage)(struct page *);
L
Linus Torvalds 已提交
162

M
Matt Mackall 已提交
163
	BUG_ON(!PageLocked(page));
L
Linus Torvalds 已提交
164

165
	freepage = mapping->a_ops->freepage;
N
Nick Piggin 已提交
166
	spin_lock_irq(&mapping->tree_lock);
167
	__delete_from_page_cache(page);
N
Nick Piggin 已提交
168
	spin_unlock_irq(&mapping->tree_lock);
169
	mem_cgroup_uncharge_cache_page(page);
170 171 172

	if (freepage)
		freepage(page);
173 174 175 176
	page_cache_release(page);
}
EXPORT_SYMBOL(delete_from_page_cache);

J
Jens Axboe 已提交
177
static int sleep_on_page(void *word)
L
Linus Torvalds 已提交
178 179 180 181 182
{
	io_schedule();
	return 0;
}

J
Jens Axboe 已提交
183
static int sleep_on_page_killable(void *word)
M
Matthew Wilcox 已提交
184
{
J
Jens Axboe 已提交
185
	sleep_on_page(word);
M
Matthew Wilcox 已提交
186 187 188
	return fatal_signal_pending(current) ? -EINTR : 0;
}

L
Linus Torvalds 已提交
189
/**
190
 * __filemap_fdatawrite_range - start writeback on mapping dirty pages in range
191 192
 * @mapping:	address space structure to write
 * @start:	offset in bytes where the range starts
193
 * @end:	offset in bytes where the range ends (inclusive)
194
 * @sync_mode:	enable synchronous operation
L
Linus Torvalds 已提交
195
 *
196 197 198
 * 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
199
 * If sync_mode is WB_SYNC_ALL then this is a "data integrity" operation, as
200
 * opposed to a regular memory cleansing writeback.  The difference between
L
Linus Torvalds 已提交
201 202 203
 * these two operations is that if a dirty page/buffer is encountered, it must
 * be waited upon, and not just skipped over.
 */
204 205
int __filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
				loff_t end, int sync_mode)
L
Linus Torvalds 已提交
206 207 208 209
{
	int ret;
	struct writeback_control wbc = {
		.sync_mode = sync_mode,
210
		.nr_to_write = LONG_MAX,
211 212
		.range_start = start,
		.range_end = end,
L
Linus Torvalds 已提交
213 214 215 216 217 218 219 220 221 222 223 224
	};

	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)
{
225
	return __filemap_fdatawrite_range(mapping, 0, LLONG_MAX, sync_mode);
L
Linus Torvalds 已提交
226 227 228 229 230 231 232 233
}

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

234
int filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
235
				loff_t end)
L
Linus Torvalds 已提交
236 237 238
{
	return __filemap_fdatawrite_range(mapping, start, end, WB_SYNC_ALL);
}
239
EXPORT_SYMBOL(filemap_fdatawrite_range);
L
Linus Torvalds 已提交
240

241 242 243 244
/**
 * filemap_flush - mostly a non-blocking flush
 * @mapping:	target address_space
 *
L
Linus Torvalds 已提交
245 246 247 248 249 250 251 252 253
 * 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);

254
/**
255 256 257 258
 * 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)
259
 *
260 261
 * 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
262
 */
263 264
int filemap_fdatawait_range(struct address_space *mapping, loff_t start_byte,
			    loff_t end_byte)
L
Linus Torvalds 已提交
265
{
266 267
	pgoff_t index = start_byte >> PAGE_CACHE_SHIFT;
	pgoff_t end = end_byte >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
268 269 270 271
	struct pagevec pvec;
	int nr_pages;
	int ret = 0;

272
	if (end_byte < start_byte)
L
Linus Torvalds 已提交
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
		return 0;

	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);
290
			if (TestClearPageError(page))
L
Linus Torvalds 已提交
291 292 293 294 295 296 297 298 299 300 301 302 303 304
				ret = -EIO;
		}
		pagevec_release(&pvec);
		cond_resched();
	}

	/* 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;
}
305 306
EXPORT_SYMBOL(filemap_fdatawait_range);

L
Linus Torvalds 已提交
307
/**
308
 * filemap_fdatawait - wait for all under-writeback pages to complete
L
Linus Torvalds 已提交
309
 * @mapping: address space structure to wait for
310 311 312
 *
 * Walk the list of under-writeback pages of the given address space
 * and wait for all of them.
L
Linus Torvalds 已提交
313 314 315 316 317 318 319 320
 */
int filemap_fdatawait(struct address_space *mapping)
{
	loff_t i_size = i_size_read(mapping->host);

	if (i_size == 0)
		return 0;

321
	return filemap_fdatawait_range(mapping, 0, i_size - 1);
L
Linus Torvalds 已提交
322 323 324 325 326
}
EXPORT_SYMBOL(filemap_fdatawait);

int filemap_write_and_wait(struct address_space *mapping)
{
327
	int err = 0;
L
Linus Torvalds 已提交
328 329

	if (mapping->nrpages) {
330 331 332 333 334 335 336 337 338 339 340 341
		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;
		}
L
Linus Torvalds 已提交
342
	}
343
	return err;
L
Linus Torvalds 已提交
344
}
345
EXPORT_SYMBOL(filemap_write_and_wait);
L
Linus Torvalds 已提交
346

347 348 349 350 351 352
/**
 * 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)
 *
353 354 355 356 357
 * 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
358 359 360
int filemap_write_and_wait_range(struct address_space *mapping,
				 loff_t lstart, loff_t lend)
{
361
	int err = 0;
L
Linus Torvalds 已提交
362 363

	if (mapping->nrpages) {
364 365 366 367
		err = __filemap_fdatawrite_range(mapping, lstart, lend,
						 WB_SYNC_ALL);
		/* See comment of filemap_write_and_wait() */
		if (err != -EIO) {
368 369
			int err2 = filemap_fdatawait_range(mapping,
						lstart, lend);
370 371 372
			if (!err)
				err = err2;
		}
L
Linus Torvalds 已提交
373
	}
374
	return err;
L
Linus Torvalds 已提交
375
}
376
EXPORT_SYMBOL(filemap_write_and_wait_range);
L
Linus Torvalds 已提交
377

378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426
/**
 * 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;
	struct mem_cgroup *memcg = NULL;

	VM_BUG_ON(!PageLocked(old));
	VM_BUG_ON(!PageLocked(new));
	VM_BUG_ON(new->mapping);

	/*
	 * This is not page migration, but prepare_migration and
	 * end_migration does enough work for charge replacement.
	 *
	 * In the longer term we probably want a specialized function
	 * for moving the charge from old to new in a more efficient
	 * manner.
	 */
	error = mem_cgroup_prepare_migration(old, new, &memcg, gfp_mask);
	if (error)
		return error;

	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);
427
		__delete_from_page_cache(old);
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447
		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);
		radix_tree_preload_end();
		if (freepage)
			freepage(old);
		page_cache_release(old);
		mem_cgroup_end_migration(memcg, old, new, true);
	} else {
		mem_cgroup_end_migration(memcg, old, new, false);
	}

	return error;
}
EXPORT_SYMBOL_GPL(replace_page_cache_page);

448
/**
N
Nick Piggin 已提交
449
 * add_to_page_cache_locked - add a locked page to the pagecache
450 451 452 453 454
 * @page:	page to add
 * @mapping:	the page's address_space
 * @offset:	page index
 * @gfp_mask:	page allocation mode
 *
N
Nick Piggin 已提交
455
 * This function is used to add a page to the pagecache. It must be locked.
L
Linus Torvalds 已提交
456 457
 * This function does not add the page to the LRU.  The caller must do that.
 */
N
Nick Piggin 已提交
458
int add_to_page_cache_locked(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
459
		pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
460
{
N
Nick Piggin 已提交
461 462 463
	int error;

	VM_BUG_ON(!PageLocked(page));
464
	VM_BUG_ON(PageSwapBacked(page));
N
Nick Piggin 已提交
465 466

	error = mem_cgroup_cache_charge(page, current->mm,
K
KAMEZAWA Hiroyuki 已提交
467
					gfp_mask & GFP_RECLAIM_MASK);
B
Balbir Singh 已提交
468 469
	if (error)
		goto out;
L
Linus Torvalds 已提交
470

B
Balbir Singh 已提交
471
	error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
L
Linus Torvalds 已提交
472
	if (error == 0) {
N
Nick Piggin 已提交
473 474 475 476
		page_cache_get(page);
		page->mapping = mapping;
		page->index = offset;

N
Nick Piggin 已提交
477
		spin_lock_irq(&mapping->tree_lock);
L
Linus Torvalds 已提交
478
		error = radix_tree_insert(&mapping->page_tree, offset, page);
N
Nick Piggin 已提交
479
		if (likely(!error)) {
L
Linus Torvalds 已提交
480
			mapping->nrpages++;
481
			__inc_zone_page_state(page, NR_FILE_PAGES);
482
			spin_unlock_irq(&mapping->tree_lock);
N
Nick Piggin 已提交
483 484
		} else {
			page->mapping = NULL;
485
			/* Leave page->index set: truncation relies upon it */
486
			spin_unlock_irq(&mapping->tree_lock);
487
			mem_cgroup_uncharge_cache_page(page);
N
Nick Piggin 已提交
488 489
			page_cache_release(page);
		}
L
Linus Torvalds 已提交
490
		radix_tree_preload_end();
B
Balbir Singh 已提交
491
	} else
492
		mem_cgroup_uncharge_cache_page(page);
493
out:
L
Linus Torvalds 已提交
494 495
	return error;
}
N
Nick Piggin 已提交
496
EXPORT_SYMBOL(add_to_page_cache_locked);
L
Linus Torvalds 已提交
497 498

int add_to_page_cache_lru(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
499
				pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
500
{
501 502 503
	int ret;

	ret = add_to_page_cache(page, mapping, offset, gfp_mask);
504 505
	if (ret == 0)
		lru_cache_add_file(page);
L
Linus Torvalds 已提交
506 507
	return ret;
}
508
EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
L
Linus Torvalds 已提交
509

510
#ifdef CONFIG_NUMA
511
struct page *__page_cache_alloc(gfp_t gfp)
512
{
513 514 515
	int n;
	struct page *page;

516
	if (cpuset_do_page_mem_spread()) {
517 518 519 520 521
		get_mems_allowed();
		n = cpuset_mem_spread_node();
		page = alloc_pages_exact_node(n, gfp, 0);
		put_mems_allowed();
		return page;
522
	}
523
	return alloc_pages(gfp, 0);
524
}
525
EXPORT_SYMBOL(__page_cache_alloc);
526 527
#endif

L
Linus Torvalds 已提交
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
/*
 * 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
550
void wait_on_page_bit(struct page *page, int bit_nr)
L
Linus Torvalds 已提交
551 552 553 554
{
	DEFINE_WAIT_BIT(wait, &page->flags, bit_nr);

	if (test_bit(bit_nr, &page->flags))
J
Jens Axboe 已提交
555
		__wait_on_bit(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
556 557 558 559
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(wait_on_page_bit);

560 561 562 563 564 565 566 567 568 569 570
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);
}

571 572
/**
 * add_page_wait_queue - Add an arbitrary waiter to a page's wait queue
R
Randy Dunlap 已提交
573 574
 * @page: Page defining the wait queue of interest
 * @waiter: Waiter to add to the queue
575 576 577 578 579 580 581 582 583 584 585 586 587 588
 *
 * 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
589
/**
590
 * unlock_page - unlock a locked page
L
Linus Torvalds 已提交
591 592 593 594 595 596 597
 * @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
598 599
 * 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
600
 */
H
Harvey Harrison 已提交
601
void unlock_page(struct page *page)
L
Linus Torvalds 已提交
602
{
N
Nick Piggin 已提交
603 604 605
	VM_BUG_ON(!PageLocked(page));
	clear_bit_unlock(PG_locked, &page->flags);
	smp_mb__after_clear_bit();
L
Linus Torvalds 已提交
606 607 608 609
	wake_up_page(page, PG_locked);
}
EXPORT_SYMBOL(unlock_page);

610 611 612
/**
 * end_page_writeback - end writeback against a page
 * @page: the page
L
Linus Torvalds 已提交
613 614 615
 */
void end_page_writeback(struct page *page)
{
616 617 618 619 620 621
	if (TestClearPageReclaim(page))
		rotate_reclaimable_page(page);

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

L
Linus Torvalds 已提交
622 623 624 625 626
	smp_mb__after_clear_bit();
	wake_up_page(page, PG_writeback);
}
EXPORT_SYMBOL(end_page_writeback);

627 628 629
/**
 * __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
630
 */
H
Harvey Harrison 已提交
631
void __lock_page(struct page *page)
L
Linus Torvalds 已提交
632 633 634
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

J
Jens Axboe 已提交
635
	__wait_on_bit_lock(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
636 637 638 639
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(__lock_page);

H
Harvey Harrison 已提交
640
int __lock_page_killable(struct page *page)
M
Matthew Wilcox 已提交
641 642 643 644
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

	return __wait_on_bit_lock(page_waitqueue(page), &wait,
J
Jens Axboe 已提交
645
					sleep_on_page_killable, TASK_KILLABLE);
M
Matthew Wilcox 已提交
646
}
647
EXPORT_SYMBOL_GPL(__lock_page_killable);
M
Matthew Wilcox 已提交
648

649 650 651
int __lock_page_or_retry(struct page *page, struct mm_struct *mm,
			 unsigned int flags)
{
652 653 654 655 656 657 658 659 660 661 662 663
	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
664
			wait_on_page_locked(page);
665
		return 0;
666 667 668 669 670 671 672 673 674 675 676 677
	} 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;
678 679 680
	}
}

681 682 683 684 685
/**
 * 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
686 687
 * 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
688
 */
N
Nick Piggin 已提交
689
struct page *find_get_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
690
{
N
Nick Piggin 已提交
691
	void **pagep;
L
Linus Torvalds 已提交
692 693
	struct page *page;

N
Nick Piggin 已提交
694 695 696 697 698 699
	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
700 701
		if (unlikely(!page))
			goto out;
702
		if (radix_tree_exception(page)) {
703 704 705 706 707 708 709 710
			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;
711
		}
N
Nick Piggin 已提交
712 713 714 715 716 717 718 719 720 721 722 723 724
		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
725
out:
N
Nick Piggin 已提交
726 727
	rcu_read_unlock();

L
Linus Torvalds 已提交
728 729 730 731 732 733
	return page;
}
EXPORT_SYMBOL(find_get_page);

/**
 * find_lock_page - locate, pin and lock a pagecache page
734 735
 * @mapping: the address_space to search
 * @offset: the page index
L
Linus Torvalds 已提交
736 737 738 739 740 741
 *
 * 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
742
struct page *find_lock_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
743 744 745 746
{
	struct page *page;

repeat:
N
Nick Piggin 已提交
747
	page = find_get_page(mapping, offset);
748
	if (page && !radix_tree_exception(page)) {
N
Nick Piggin 已提交
749 750 751 752 753 754
		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
755
		}
N
Nick Piggin 已提交
756
		VM_BUG_ON(page->index != offset);
L
Linus Torvalds 已提交
757 758 759 760 761 762 763
	}
	return page;
}
EXPORT_SYMBOL(find_lock_page);

/**
 * find_or_create_page - locate or add a pagecache page
764 765 766
 * @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
767 768 769 770 771 772 773 774 775 776 777 778 779
 *
 * 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,
780
		pgoff_t index, gfp_t gfp_mask)
L
Linus Torvalds 已提交
781
{
N
Nick Piggin 已提交
782
	struct page *page;
L
Linus Torvalds 已提交
783 784 785 786
	int err;
repeat:
	page = find_lock_page(mapping, index);
	if (!page) {
N
Nick Piggin 已提交
787 788 789
		page = __page_cache_alloc(gfp_mask);
		if (!page)
			return NULL;
N
Nick Piggin 已提交
790 791 792 793 794 795 796 797
		/*
		 * 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
798 799 800 801 802
		if (unlikely(err)) {
			page_cache_release(page);
			page = NULL;
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829
		}
	}
	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)
{
	unsigned int i;
	unsigned int ret;
830
	unsigned int nr_found, nr_skip;
N
Nick Piggin 已提交
831 832 833 834

	rcu_read_lock();
restart:
	nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree,
835
				(void ***)pages, NULL, start, nr_pages);
N
Nick Piggin 已提交
836
	ret = 0;
837
	nr_skip = 0;
N
Nick Piggin 已提交
838 839 840 841 842 843
	for (i = 0; i < nr_found; i++) {
		struct page *page;
repeat:
		page = radix_tree_deref_slot((void **)pages[i]);
		if (unlikely(!page))
			continue;
844

845
		if (radix_tree_exception(page)) {
846 847 848 849 850 851 852 853 854
			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.
				 */
				WARN_ON(start | i);
				goto restart;
			}
855
			/*
856 857 858
			 * 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().
859
			 */
860
			nr_skip++;
861
			continue;
N
Nick Piggin 已提交
862
		}
N
Nick Piggin 已提交
863 864 865 866 867 868 869 870 871

		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
		if (unlikely(page != *((void **)pages[i]))) {
			page_cache_release(page);
			goto repeat;
		}
L
Linus Torvalds 已提交
872

N
Nick Piggin 已提交
873 874 875
		pages[ret] = page;
		ret++;
	}
876 877 878 879 880

	/*
	 * If all entries were removed before we could secure them,
	 * try again, because callers stop trying once 0 is returned.
	 */
881
	if (unlikely(!ret && nr_found > nr_skip))
882
		goto restart;
N
Nick Piggin 已提交
883
	rcu_read_unlock();
L
Linus Torvalds 已提交
884 885 886
	return ret;
}

887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903
/**
 * 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)
{
	unsigned int i;
	unsigned int ret;
N
Nick Piggin 已提交
904 905 906 907 908
	unsigned int nr_found;

	rcu_read_lock();
restart:
	nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree,
909
				(void ***)pages, NULL, index, nr_pages);
N
Nick Piggin 已提交
910 911 912 913 914 915 916
	ret = 0;
	for (i = 0; i < nr_found; i++) {
		struct page *page;
repeat:
		page = radix_tree_deref_slot((void **)pages[i]);
		if (unlikely(!page))
			continue;
917

918
		if (radix_tree_exception(page)) {
919 920 921 922 923 924 925 926
			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;
			}
927
			/*
928 929 930
			 * Otherwise, shmem/tmpfs must be storing a swap entry
			 * here as an exceptional entry: so stop looking for
			 * contiguous pages.
931
			 */
932
			break;
933
		}
934

N
Nick Piggin 已提交
935 936 937 938 939 940 941 942 943
		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
		if (unlikely(page != *((void **)pages[i]))) {
			page_cache_release(page);
			goto repeat;
		}

N
Nick Piggin 已提交
944 945 946 947 948 949 950 951 952 953
		/*
		 * 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.
		 */
		if (page->mapping == NULL || page->index != index) {
			page_cache_release(page);
			break;
		}

N
Nick Piggin 已提交
954 955
		pages[ret] = page;
		ret++;
956 957
		index++;
	}
N
Nick Piggin 已提交
958 959
	rcu_read_unlock();
	return ret;
960
}
961
EXPORT_SYMBOL(find_get_pages_contig);
962

963 964 965 966 967 968 969 970
/**
 * 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
971
 * Like find_get_pages, except we only return pages which are tagged with
972
 * @tag.   We update @index to index the next page for the traversal.
L
Linus Torvalds 已提交
973 974 975 976 977 978
 */
unsigned find_get_pages_tag(struct address_space *mapping, pgoff_t *index,
			int tag, unsigned int nr_pages, struct page **pages)
{
	unsigned int i;
	unsigned int ret;
N
Nick Piggin 已提交
979 980 981 982 983 984 985 986 987 988 989 990 991
	unsigned int nr_found;

	rcu_read_lock();
restart:
	nr_found = radix_tree_gang_lookup_tag_slot(&mapping->page_tree,
				(void ***)pages, *index, nr_pages, tag);
	ret = 0;
	for (i = 0; i < nr_found; i++) {
		struct page *page;
repeat:
		page = radix_tree_deref_slot((void **)pages[i]);
		if (unlikely(!page))
			continue;
992

993
		if (radix_tree_exception(page)) {
994 995 996 997 998 999 1000 1001
			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;
			}
1002
			/*
1003 1004
			 * This function is never used on a shmem/tmpfs
			 * mapping, so a swap entry won't be found here.
1005
			 */
1006
			BUG();
1007
		}
N
Nick Piggin 已提交
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020

		if (!page_cache_get_speculative(page))
			goto repeat;

		/* Has the page moved? */
		if (unlikely(page != *((void **)pages[i]))) {
			page_cache_release(page);
			goto repeat;
		}

		pages[ret] = page;
		ret++;
	}
1021 1022 1023 1024 1025 1026 1027

	/*
	 * If all entries were removed before we could secure them,
	 * try again, because callers stop trying once 0 is returned.
	 */
	if (unlikely(!ret && nr_found))
		goto restart;
N
Nick Piggin 已提交
1028
	rcu_read_unlock();
L
Linus Torvalds 已提交
1029 1030 1031

	if (ret)
		*index = pages[ret - 1]->index + 1;
N
Nick Piggin 已提交
1032

L
Linus Torvalds 已提交
1033 1034
	return ret;
}
1035
EXPORT_SYMBOL(find_get_pages_tag);
L
Linus Torvalds 已提交
1036

1037 1038 1039 1040 1041
/**
 * grab_cache_page_nowait - returns locked page at given index in given cache
 * @mapping: target address_space
 * @index: the page index
 *
1042
 * Same as grab_cache_page(), but do not wait if the page is unavailable.
L
Linus Torvalds 已提交
1043 1044 1045 1046 1047 1048 1049 1050
 * 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 *
1051
grab_cache_page_nowait(struct address_space *mapping, pgoff_t index)
L
Linus Torvalds 已提交
1052 1053 1054 1055
{
	struct page *page = find_get_page(mapping, index);

	if (page) {
N
Nick Piggin 已提交
1056
		if (trylock_page(page))
L
Linus Torvalds 已提交
1057 1058 1059 1060
			return page;
		page_cache_release(page);
		return NULL;
	}
1061
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~__GFP_FS);
N
Nick Piggin 已提交
1062
	if (page && add_to_page_cache_lru(page, mapping, index, GFP_NOFS)) {
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069
		page_cache_release(page);
		page = NULL;
	}
	return page;
}
EXPORT_SYMBOL(grab_cache_page_nowait);

1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
/*
 * 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;
}

1091
/**
C
Christoph Hellwig 已提交
1092
 * do_generic_file_read - generic file read routine
1093 1094 1095 1096 1097
 * @filp:	the file to read
 * @ppos:	current file position
 * @desc:	read_descriptor
 * @actor:	read method
 *
L
Linus Torvalds 已提交
1098
 * This is a generic file read routine, and uses the
1099
 * mapping->a_ops->readpage() function for the actual low-level stuff.
L
Linus Torvalds 已提交
1100 1101 1102 1103
 *
 * 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
1104 1105
static void do_generic_file_read(struct file *filp, loff_t *ppos,
		read_descriptor_t *desc, read_actor_t actor)
L
Linus Torvalds 已提交
1106
{
C
Christoph Hellwig 已提交
1107
	struct address_space *mapping = filp->f_mapping;
L
Linus Torvalds 已提交
1108
	struct inode *inode = mapping->host;
C
Christoph Hellwig 已提交
1109
	struct file_ra_state *ra = &filp->f_ra;
1110 1111 1112 1113
	pgoff_t index;
	pgoff_t last_index;
	pgoff_t prev_index;
	unsigned long offset;      /* offset into pagecache page */
1114
	unsigned int prev_offset;
L
Linus Torvalds 已提交
1115 1116 1117
	int error;

	index = *ppos >> PAGE_CACHE_SHIFT;
1118 1119
	prev_index = ra->prev_pos >> PAGE_CACHE_SHIFT;
	prev_offset = ra->prev_pos & (PAGE_CACHE_SIZE-1);
L
Linus Torvalds 已提交
1120 1121 1122 1123 1124
	last_index = (*ppos + desc->count + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	for (;;) {
		struct page *page;
1125
		pgoff_t end_index;
N
NeilBrown 已提交
1126
		loff_t isize;
L
Linus Torvalds 已提交
1127 1128 1129 1130 1131
		unsigned long nr, ret;

		cond_resched();
find_page:
		page = find_get_page(mapping, index);
1132
		if (!page) {
1133
			page_cache_sync_readahead(mapping,
1134
					ra, filp,
1135 1136 1137 1138 1139 1140
					index, last_index - index);
			page = find_get_page(mapping, index);
			if (unlikely(page == NULL))
				goto no_cached_page;
		}
		if (PageReadahead(page)) {
1141
			page_cache_async_readahead(mapping,
1142
					ra, filp, page,
1143
					index, last_index - index);
L
Linus Torvalds 已提交
1144
		}
1145 1146 1147 1148
		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
1149
			if (!trylock_page(page))
1150
				goto page_not_up_to_date;
1151 1152 1153
			/* Did it get truncated before we got the lock? */
			if (!page->mapping)
				goto page_not_up_to_date_locked;
1154 1155 1156 1157 1158
			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
1159
page_ok:
N
NeilBrown 已提交
1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
		/*
		 * 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
1186 1187 1188 1189 1190 1191 1192 1193 1194

		/* 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);

		/*
1195 1196
		 * When a sequential read accesses a page several times,
		 * only mark it as accessed the first time.
L
Linus Torvalds 已提交
1197
		 */
1198
		if (prev_index != index || offset != prev_offset)
L
Linus Torvalds 已提交
1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
			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...
		 *
		 * The actor routine returns how many bytes were actually used..
		 * 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).
		 */
		ret = actor(desc, page, offset, nr);
		offset += ret;
		index += offset >> PAGE_CACHE_SHIFT;
		offset &= ~PAGE_CACHE_MASK;
J
Jan Kara 已提交
1216
		prev_offset = offset;
L
Linus Torvalds 已提交
1217 1218 1219 1220 1221 1222 1223 1224

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

page_not_up_to_date:
		/* Get exclusive access to the page ... */
1225 1226 1227
		error = lock_page_killable(page);
		if (unlikely(error))
			goto readpage_error;
L
Linus Torvalds 已提交
1228

1229
page_not_up_to_date_locked:
N
Nick Piggin 已提交
1230
		/* Did it get truncated before we got the lock? */
L
Linus Torvalds 已提交
1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
		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:
1244 1245 1246 1247 1248 1249
		/*
		 * 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
1250 1251 1252
		/* Start the actual read. The read will unlock the page. */
		error = mapping->a_ops->readpage(filp, page);

1253 1254 1255 1256 1257
		if (unlikely(error)) {
			if (error == AOP_TRUNCATED_PAGE) {
				page_cache_release(page);
				goto find_page;
			}
L
Linus Torvalds 已提交
1258
			goto readpage_error;
1259
		}
L
Linus Torvalds 已提交
1260 1261

		if (!PageUptodate(page)) {
1262 1263 1264
			error = lock_page_killable(page);
			if (unlikely(error))
				goto readpage_error;
L
Linus Torvalds 已提交
1265 1266 1267
			if (!PageUptodate(page)) {
				if (page->mapping == NULL) {
					/*
1268
					 * invalidate_mapping_pages got it
L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274
					 */
					unlock_page(page);
					page_cache_release(page);
					goto find_page;
				}
				unlock_page(page);
1275
				shrink_readahead_size_eio(filp, ra);
1276 1277
				error = -EIO;
				goto readpage_error;
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
			}
			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
1295 1296 1297 1298
		page = page_cache_alloc_cold(mapping);
		if (!page) {
			desc->error = -ENOMEM;
			goto out;
L
Linus Torvalds 已提交
1299
		}
N
Nick Piggin 已提交
1300
		error = add_to_page_cache_lru(page, mapping,
L
Linus Torvalds 已提交
1301 1302
						index, GFP_KERNEL);
		if (error) {
N
Nick Piggin 已提交
1303
			page_cache_release(page);
L
Linus Torvalds 已提交
1304 1305 1306 1307 1308 1309 1310 1311 1312
			if (error == -EEXIST)
				goto find_page;
			desc->error = error;
			goto out;
		}
		goto readpage;
	}

out:
1313 1314 1315
	ra->prev_pos = prev_index;
	ra->prev_pos <<= PAGE_CACHE_SHIFT;
	ra->prev_pos |= prev_offset;
L
Linus Torvalds 已提交
1316

1317
	*ppos = ((loff_t)index << PAGE_CACHE_SHIFT) + offset;
1318
	file_accessed(filp);
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
}

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)) {
		kaddr = kmap_atomic(page, KM_USER0);
		left = __copy_to_user_inatomic(desc->arg.buf,
						kaddr + offset, size);
		kunmap_atomic(kaddr, KM_USER0);
		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;
}

1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
/*
 * 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);

1398
/**
H
Henrik Kretzschmar 已提交
1399
 * generic_file_aio_read - generic filesystem read routine
1400 1401 1402
 * @iocb:	kernel I/O control block
 * @iov:	io vector request
 * @nr_segs:	number of segments in the iovec
H
Henrik Kretzschmar 已提交
1403
 * @pos:	current file position
1404
 *
L
Linus Torvalds 已提交
1405 1406 1407 1408
 * This is the "read()" routine for all filesystems
 * that can use the page cache directly.
 */
ssize_t
1409 1410
generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
1411 1412 1413
{
	struct file *filp = iocb->ki_filp;
	ssize_t retval;
1414
	unsigned long seg = 0;
L
Linus Torvalds 已提交
1415
	size_t count;
1416
	loff_t *ppos = &iocb->ki_pos;
1417
	struct blk_plug plug;
L
Linus Torvalds 已提交
1418 1419

	count = 0;
1420 1421 1422
	retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
	if (retval)
		return retval;
L
Linus Torvalds 已提交
1423

1424 1425
	blk_start_plug(&plug);

L
Linus Torvalds 已提交
1426 1427
	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (filp->f_flags & O_DIRECT) {
1428
		loff_t size;
L
Linus Torvalds 已提交
1429 1430 1431 1432 1433 1434 1435 1436 1437
		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);
		if (pos < size) {
1438 1439
			retval = filemap_write_and_wait_range(mapping, pos,
					pos + iov_length(iov, nr_segs) - 1);
1440 1441 1442 1443
			if (!retval) {
				retval = mapping->a_ops->direct_IO(READ, iocb,
							iov, pos, nr_segs);
			}
1444
			if (retval > 0) {
L
Linus Torvalds 已提交
1445
				*ppos = pos + retval;
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
				count -= retval;
			}

			/*
			 * 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) {
H
Hugh Dickins 已提交
1458 1459 1460
				file_accessed(filp);
				goto out;
			}
1461
		}
L
Linus Torvalds 已提交
1462 1463
	}

1464
	count = retval;
H
Hugh Dickins 已提交
1465 1466
	for (seg = 0; seg < nr_segs; seg++) {
		read_descriptor_t desc;
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
		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
1481

H
Hugh Dickins 已提交
1482
		desc.written = 0;
1483 1484
		desc.arg.buf = iov[seg].iov_base + offset;
		desc.count = iov[seg].iov_len - offset;
H
Hugh Dickins 已提交
1485 1486 1487 1488 1489 1490 1491 1492
		if (desc.count == 0)
			continue;
		desc.error = 0;
		do_generic_file_read(filp, ppos, &desc, file_read_actor);
		retval += desc.written;
		if (desc.error) {
			retval = retval ?: desc.error;
			break;
L
Linus Torvalds 已提交
1493
		}
H
Hugh Dickins 已提交
1494 1495
		if (desc.count > 0)
			break;
L
Linus Torvalds 已提交
1496 1497
	}
out:
1498
	blk_finish_plug(&plug);
L
Linus Torvalds 已提交
1499 1500 1501 1502 1503 1504
	return retval;
}
EXPORT_SYMBOL(generic_file_aio_read);

static ssize_t
do_readahead(struct address_space *mapping, struct file *filp,
1505
	     pgoff_t index, unsigned long nr)
L
Linus Torvalds 已提交
1506 1507 1508 1509
{
	if (!mapping || !mapping->a_ops || !mapping->a_ops->readpage)
		return -EINVAL;

1510
	force_page_cache_readahead(mapping, filp, index, nr);
L
Linus Torvalds 已提交
1511 1512 1513
	return 0;
}

1514
SYSCALL_DEFINE(readahead)(int fd, loff_t offset, size_t count)
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523
{
	ssize_t ret;
	struct file *file;

	ret = -EBADF;
	file = fget(fd);
	if (file) {
		if (file->f_mode & FMODE_READ) {
			struct address_space *mapping = file->f_mapping;
1524 1525
			pgoff_t start = offset >> PAGE_CACHE_SHIFT;
			pgoff_t end = (offset + count - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
1526 1527 1528 1529 1530 1531 1532
			unsigned long len = end - start + 1;
			ret = do_readahead(mapping, file, start, len);
		}
		fput(file);
	}
	return ret;
}
1533 1534 1535 1536 1537 1538 1539
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_readahead(long fd, loff_t offset, long count)
{
	return SYSC_readahead((int) fd, offset, (size_t) count);
}
SYSCALL_ALIAS(sys_readahead, SyS_readahead);
#endif
L
Linus Torvalds 已提交
1540 1541

#ifdef CONFIG_MMU
1542 1543 1544 1545 1546
/**
 * 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
1547 1548 1549
 * 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
1550
static int page_cache_read(struct file *file, pgoff_t offset)
L
Linus Torvalds 已提交
1551 1552 1553
{
	struct address_space *mapping = file->f_mapping;
	struct page *page; 
1554
	int ret;
L
Linus Torvalds 已提交
1555

1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
	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
1566 1567 1568

		page_cache_release(page);

1569 1570 1571
	} while (ret == AOP_TRUNCATED_PAGE);
		
	return ret;
L
Linus Torvalds 已提交
1572 1573 1574 1575
}

#define MMAP_LOTSAMISS  (100)

1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
/*
 * 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 */
	if (VM_RandomReadHint(vma))
		return;
1591 1592
	if (!ra->ra_pages)
		return;
1593

1594
	if (VM_SequentialReadHint(vma)) {
1595 1596
		page_cache_sync_readahead(mapping, ra, file, offset,
					  ra->ra_pages);
1597 1598 1599
		return;
	}

1600 1601
	/* Avoid banging the cache line if not needed */
	if (ra->mmap_miss < MMAP_LOTSAMISS * 10)
1602 1603 1604 1605 1606 1607 1608 1609 1610
		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;

1611 1612 1613
	/*
	 * mmap read-around
	 */
1614
	ra_pages = max_sane_readahead(ra->ra_pages);
1615 1616
	ra->start = max_t(long, 0, offset - ra_pages / 2);
	ra->size = ra_pages;
1617
	ra->async_size = ra_pages / 4;
1618
	ra_submit(ra, mapping, file);
1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638
}

/*
 * 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 */
	if (VM_RandomReadHint(vma))
		return;
	if (ra->mmap_miss > 0)
		ra->mmap_miss--;
	if (PageReadahead(page))
1639 1640
		page_cache_async_readahead(mapping, ra, file,
					   page, offset, ra->ra_pages);
1641 1642
}

1643
/**
1644
 * filemap_fault - read in file data for page fault handling
N
Nick Piggin 已提交
1645 1646
 * @vma:	vma in which the fault was taken
 * @vmf:	struct vm_fault containing details of the fault
1647
 *
1648
 * filemap_fault() is invoked via the vma operations vector for a
L
Linus Torvalds 已提交
1649 1650 1651 1652 1653 1654
 * 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
1655
int filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1656 1657
{
	int error;
1658
	struct file *file = vma->vm_file;
L
Linus Torvalds 已提交
1659 1660 1661
	struct address_space *mapping = file->f_mapping;
	struct file_ra_state *ra = &file->f_ra;
	struct inode *inode = mapping->host;
1662
	pgoff_t offset = vmf->pgoff;
L
Linus Torvalds 已提交
1663
	struct page *page;
J
Jan Kara 已提交
1664
	pgoff_t size;
N
Nick Piggin 已提交
1665
	int ret = 0;
L
Linus Torvalds 已提交
1666 1667

	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1668
	if (offset >= size)
1669
		return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1670 1671 1672 1673

	/*
	 * Do we have something in the page cache already?
	 */
1674 1675
	page = find_get_page(mapping, offset);
	if (likely(page)) {
L
Linus Torvalds 已提交
1676
		/*
1677 1678
		 * We found the page, so try async readahead before
		 * waiting for the lock.
L
Linus Torvalds 已提交
1679
		 */
1680 1681 1682 1683 1684
		do_async_mmap_readahead(vma, ra, file, page, offset);
	} else {
		/* No page in the page cache at all */
		do_sync_mmap_readahead(vma, ra, file, offset);
		count_vm_event(PGMAJFAULT);
1685
		mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT);
1686 1687
		ret = VM_FAULT_MAJOR;
retry_find:
1688
		page = find_get_page(mapping, offset);
L
Linus Torvalds 已提交
1689 1690 1691 1692
		if (!page)
			goto no_cached_page;
	}

1693 1694
	if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags)) {
		page_cache_release(page);
1695
		return ret | VM_FAULT_RETRY;
1696
	}
1697 1698 1699 1700 1701 1702 1703 1704 1705

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

L
Linus Torvalds 已提交
1706
	/*
1707 1708
	 * 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
1709
	 */
1710
	if (unlikely(!PageUptodate(page)))
L
Linus Torvalds 已提交
1711 1712
		goto page_not_uptodate;

1713 1714 1715 1716
	/*
	 * Found the page and have a reference on it.
	 * We must recheck i_size under page lock.
	 */
1717
	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1718
	if (unlikely(offset >= size)) {
1719
		unlock_page(page);
1720
		page_cache_release(page);
1721
		return VM_FAULT_SIGBUS;
1722 1723
	}

N
Nick Piggin 已提交
1724
	vmf->page = page;
N
Nick Piggin 已提交
1725
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1726 1727 1728 1729 1730 1731

no_cached_page:
	/*
	 * We're only likely to ever get here if MADV_RANDOM is in
	 * effect.
	 */
1732
	error = page_cache_read(file, offset);
L
Linus Torvalds 已提交
1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747

	/*
	 * 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
1748 1749
		return VM_FAULT_OOM;
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1750 1751 1752 1753 1754 1755 1756 1757 1758

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);
1759
	error = mapping->a_ops->readpage(file, page);
1760 1761 1762 1763 1764
	if (!error) {
		wait_on_page_locked(page);
		if (!PageUptodate(page))
			error = -EIO;
	}
1765 1766 1767
	page_cache_release(page);

	if (!error || error == AOP_TRUNCATED_PAGE)
1768
		goto retry_find;
L
Linus Torvalds 已提交
1769

1770
	/* Things didn't work out. Return zero to tell the mm layer so. */
1771
	shrink_readahead_size_eio(file, ra);
N
Nick Piggin 已提交
1772
	return VM_FAULT_SIGBUS;
1773 1774 1775
}
EXPORT_SYMBOL(filemap_fault);

1776
const struct vm_operations_struct generic_file_vm_ops = {
1777
	.fault		= filemap_fault,
L
Linus Torvalds 已提交
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
};

/* 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;
N
Nick Piggin 已提交
1790
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816
	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);

1817
static struct page *__read_cache_page(struct address_space *mapping,
1818
				pgoff_t index,
1819
				int (*filler)(void *, struct page *),
1820 1821
				void *data,
				gfp_t gfp)
L
Linus Torvalds 已提交
1822
{
N
Nick Piggin 已提交
1823
	struct page *page;
L
Linus Torvalds 已提交
1824 1825 1826 1827
	int err;
repeat:
	page = find_get_page(mapping, index);
	if (!page) {
1828
		page = __page_cache_alloc(gfp | __GFP_COLD);
N
Nick Piggin 已提交
1829 1830 1831 1832 1833 1834 1835
		if (!page)
			return ERR_PTR(-ENOMEM);
		err = add_to_page_cache_lru(page, mapping, index, GFP_KERNEL);
		if (unlikely(err)) {
			page_cache_release(page);
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847
			/* 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;
}

1848
static struct page *do_read_cache_page(struct address_space *mapping,
1849
				pgoff_t index,
1850
				int (*filler)(void *, struct page *),
1851 1852 1853
				void *data,
				gfp_t gfp)

L
Linus Torvalds 已提交
1854 1855 1856 1857 1858
{
	struct page *page;
	int err;

retry:
1859
	page = __read_cache_page(mapping, index, filler, data, gfp);
L
Linus Torvalds 已提交
1860
	if (IS_ERR(page))
1861
		return page;
L
Linus Torvalds 已提交
1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
	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);
1878
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1879
	}
1880
out:
1881 1882 1883
	mark_page_accessed(page);
	return page;
}
1884 1885 1886 1887 1888 1889

/**
 * 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
1890
 * @data:	first arg to filler(data, page) function, often left as NULL
1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
 *
 * 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,
1902
				int (*filler)(void *, struct page *),
1903 1904 1905 1906
				void *data)
{
	return do_read_cache_page(mapping, index, filler, data, mapping_gfp_mask(mapping));
}
1907 1908
EXPORT_SYMBOL(read_cache_page_async);

1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944
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
 * any new page allocations done using the specified allocation flags. Note
 * that the Radix tree operations will still use GFP_KERNEL, so you can't
 * expect to do this atomically or anything like that - but you can pass in
 * other page requirements.
 *
 * 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);

1945 1946 1947 1948 1949
/**
 * 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
1950
 * @data:	first arg to filler(data, page) function, often left as NULL
1951 1952 1953 1954 1955 1956 1957
 *
 * 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,
1958
				pgoff_t index,
1959
				int (*filler)(void *, struct page *),
1960 1961
				void *data)
{
1962
	return wait_on_page_read(read_cache_page_async(mapping, index, filler, data));
L
Linus Torvalds 已提交
1963 1964 1965 1966 1967 1968 1969 1970 1971
}
EXPORT_SYMBOL(read_cache_page);

/*
 * The logic we want is
 *
 *	if suid or (sgid and xgrp)
 *		remove privs
 */
1972
int should_remove_suid(struct dentry *dentry)
L
Linus Torvalds 已提交
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
{
	mode_t mode = dentry->d_inode->i_mode;
	int kill = 0;

	/* suid always must be killed */
	if (unlikely(mode & S_ISUID))
		kill = ATTR_KILL_SUID;

	/*
	 * sgid without any exec bits is just a mandatory locking mark; leave
	 * it alone.  If some exec bits are set, it's a real sgid; kill it.
	 */
	if (unlikely((mode & S_ISGID) && (mode & S_IXGRP)))
		kill |= ATTR_KILL_SGID;

1988
	if (unlikely(kill && !capable(CAP_FSETID) && S_ISREG(mode)))
1989
		return kill;
L
Linus Torvalds 已提交
1990

1991 1992
	return 0;
}
M
Mark Fasheh 已提交
1993
EXPORT_SYMBOL(should_remove_suid);
1994

1995
static int __remove_suid(struct dentry *dentry, int kill)
1996 1997 1998 1999 2000 2001 2002
{
	struct iattr newattrs;

	newattrs.ia_valid = ATTR_FORCE | kill;
	return notify_change(dentry, &newattrs);
}

2003
int file_remove_suid(struct file *file)
2004
{
2005
	struct dentry *dentry = file->f_path.dentry;
2006 2007 2008
	struct inode *inode = dentry->d_inode;
	int killsuid;
	int killpriv;
2009
	int error = 0;
2010

2011 2012 2013 2014 2015 2016 2017
	/* Fast path for nothing security related */
	if (IS_NOSEC(inode))
		return 0;

	killsuid = should_remove_suid(dentry);
	killpriv = security_inode_need_killpriv(dentry);

2018 2019 2020 2021 2022 2023
	if (killpriv < 0)
		return killpriv;
	if (killpriv)
		error = security_inode_killpriv(dentry);
	if (!error && killsuid)
		error = __remove_suid(dentry, killsuid);
A
Al Viro 已提交
2024
	if (!error && (inode->i_sb->s_flags & MS_NOSEC))
2025
		inode->i_flags |= S_NOSEC;
2026

2027
	return error;
L
Linus Torvalds 已提交
2028
}
2029
EXPORT_SYMBOL(file_remove_suid);
L
Linus Torvalds 已提交
2030

N
Nick Piggin 已提交
2031
static size_t __iovec_copy_from_user_inatomic(char *vaddr,
L
Linus Torvalds 已提交
2032 2033
			const struct iovec *iov, size_t base, size_t bytes)
{
2034
	size_t copied = 0, left = 0;
L
Linus Torvalds 已提交
2035 2036 2037 2038 2039 2040

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

		base = 0;
2041
		left = __copy_from_user_inatomic(vaddr, buf, copy);
L
Linus Torvalds 已提交
2042 2043 2044 2045 2046
		copied += copy;
		bytes -= copy;
		vaddr += copy;
		iov++;

2047
		if (unlikely(left))
L
Linus Torvalds 已提交
2048 2049 2050 2051 2052
			break;
	}
	return copied - left;
}

N
Nick Piggin 已提交
2053 2054
/*
 * Copy as much as we can into the page and return the number of bytes which
2055
 * were successfully copied.  If a fault is encountered then return the number of
N
Nick Piggin 已提交
2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068
 * 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());
	kaddr = kmap_atomic(page, KM_USER0);
	if (likely(i->nr_segs == 1)) {
		int left;
		char __user *buf = i->iov->iov_base + i->iov_offset;
2069
		left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078
		copied = bytes - left;
	} else {
		copied = __iovec_copy_from_user_inatomic(kaddr + offset,
						i->iov, i->iov_offset, bytes);
	}
	kunmap_atomic(kaddr, KM_USER0);

	return copied;
}
N
Nick Piggin 已提交
2079
EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
N
Nick Piggin 已提交
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096

/*
 * 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;
2097
		left = __copy_from_user(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
2098 2099 2100 2101 2102 2103 2104 2105
		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
2106
EXPORT_SYMBOL(iov_iter_copy_from_user);
N
Nick Piggin 已提交
2107

N
Nick Piggin 已提交
2108
void iov_iter_advance(struct iov_iter *i, size_t bytes)
N
Nick Piggin 已提交
2109
{
N
Nick Piggin 已提交
2110 2111
	BUG_ON(i->count < bytes);

N
Nick Piggin 已提交
2112 2113
	if (likely(i->nr_segs == 1)) {
		i->iov_offset += bytes;
N
Nick Piggin 已提交
2114
		i->count -= bytes;
N
Nick Piggin 已提交
2115 2116 2117 2118
	} else {
		const struct iovec *iov = i->iov;
		size_t base = i->iov_offset;

2119 2120
		/*
		 * The !iov->iov_len check ensures we skip over unlikely
N
Nick Piggin 已提交
2121
		 * zero-length segments (without overruning the iovec).
2122
		 */
2123
		while (bytes || unlikely(i->count && !iov->iov_len)) {
N
Nick Piggin 已提交
2124
			int copy;
N
Nick Piggin 已提交
2125

N
Nick Piggin 已提交
2126 2127 2128
			copy = min(bytes, iov->iov_len - base);
			BUG_ON(!i->count || i->count < copy);
			i->count -= copy;
N
Nick Piggin 已提交
2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139
			bytes -= copy;
			base += copy;
			if (iov->iov_len == base) {
				iov++;
				base = 0;
			}
		}
		i->iov = iov;
		i->iov_offset = base;
	}
}
N
Nick Piggin 已提交
2140
EXPORT_SYMBOL(iov_iter_advance);
N
Nick Piggin 已提交
2141

2142 2143 2144 2145 2146 2147 2148 2149 2150 2151
/*
 * 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
2152 2153
{
	char __user *buf = i->iov->iov_base + i->iov_offset;
2154 2155
	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	return fault_in_pages_readable(buf, bytes);
N
Nick Piggin 已提交
2156
}
N
Nick Piggin 已提交
2157
EXPORT_SYMBOL(iov_iter_fault_in_readable);
N
Nick Piggin 已提交
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169

/*
 * Return the count of just the current iov_iter segment.
 */
size_t iov_iter_single_seg_count(struct iov_iter *i)
{
	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
2170
EXPORT_SYMBOL(iov_iter_single_seg_count);
N
Nick Piggin 已提交
2171

L
Linus Torvalds 已提交
2172 2173 2174
/*
 * Performs necessary checks before doing a write
 *
2175
 * Can adjust writing position or amount of bytes to write.
L
Linus Torvalds 已提交
2176 2177 2178 2179 2180 2181
 * 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
2182
	unsigned long limit = rlimit(RLIMIT_FSIZE);
L
Linus Torvalds 已提交
2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233

        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 {
2234
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245
		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;
2246 2247 2248
#else
		return -EPERM;
#endif
L
Linus Torvalds 已提交
2249 2250 2251 2252 2253
	}
	return 0;
}
EXPORT_SYMBOL(generic_write_checks);

2254 2255 2256 2257 2258 2259
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;

2260
	return aops->write_begin(file, mapping, pos, len, flags,
2261 2262 2263 2264 2265 2266 2267 2268 2269 2270
							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;

2271 2272
	mark_page_accessed(page);
	return aops->write_end(file, mapping, pos, len, copied, page, fsdata);
2273 2274 2275
}
EXPORT_SYMBOL(pagecache_write_end);

L
Linus Torvalds 已提交
2276 2277 2278 2279 2280 2281 2282 2283 2284
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;
2285 2286
	size_t		write_len;
	pgoff_t		end;
L
Linus Torvalds 已提交
2287 2288 2289 2290

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

2291 2292 2293
	write_len = iov_length(iov, *nr_segs);
	end = (pos + write_len - 1) >> PAGE_CACHE_SHIFT;

2294
	written = filemap_write_and_wait_range(mapping, pos, pos + write_len - 1);
2295 2296 2297 2298 2299 2300 2301
	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
2302
	 * without clobbering -EIOCBQUEUED from ->direct_IO().
2303 2304 2305 2306
	 */
	if (mapping->nrpages) {
		written = invalidate_inode_pages2_range(mapping,
					pos >> PAGE_CACHE_SHIFT, end);
2307 2308 2309 2310 2311 2312 2313
		/*
		 * If a page can not be invalidated, return 0 to fall back
		 * to buffered write.
		 */
		if (written) {
			if (written == -EBUSY)
				return 0;
2314
			goto out;
2315
		}
2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
	}

	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
2333
	if (written > 0) {
2334 2335 2336
		pos += written;
		if (pos > i_size_read(inode) && !S_ISBLK(inode->i_mode)) {
			i_size_write(inode, pos);
L
Linus Torvalds 已提交
2337 2338
			mark_inode_dirty(inode);
		}
2339
		*ppos = pos;
L
Linus Torvalds 已提交
2340
	}
2341
out:
L
Linus Torvalds 已提交
2342 2343 2344 2345
	return written;
}
EXPORT_SYMBOL(generic_file_direct_write);

N
Nick Piggin 已提交
2346 2347 2348 2349
/*
 * Find or create a page at the given pagecache position. Return the locked
 * page. This function is specifically for buffered writes.
 */
2350 2351
struct page *grab_cache_page_write_begin(struct address_space *mapping,
					pgoff_t index, unsigned flags)
N
Nick Piggin 已提交
2352 2353 2354
{
	int status;
	struct page *page;
2355 2356 2357
	gfp_t gfp_notmask = 0;
	if (flags & AOP_FLAG_NOFS)
		gfp_notmask = __GFP_FS;
N
Nick Piggin 已提交
2358 2359
repeat:
	page = find_lock_page(mapping, index);
2360
	if (page)
2361
		goto found;
N
Nick Piggin 已提交
2362

2363
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~gfp_notmask);
N
Nick Piggin 已提交
2364 2365
	if (!page)
		return NULL;
2366 2367
	status = add_to_page_cache_lru(page, mapping, index,
						GFP_KERNEL & ~gfp_notmask);
N
Nick Piggin 已提交
2368 2369 2370 2371 2372 2373
	if (unlikely(status)) {
		page_cache_release(page);
		if (status == -EEXIST)
			goto repeat;
		return NULL;
	}
2374 2375
found:
	wait_on_page_writeback(page);
N
Nick Piggin 已提交
2376 2377
	return page;
}
2378
EXPORT_SYMBOL(grab_cache_page_write_begin);
N
Nick Piggin 已提交
2379

2380 2381 2382 2383 2384 2385 2386
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
2387 2388 2389 2390 2391 2392 2393
	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;
2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422

	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
2423
		status = a_ops->write_begin(file, mapping, pos, bytes, flags,
2424 2425 2426 2427
						&page, &fsdata);
		if (unlikely(status))
			break;

2428 2429 2430
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

2431 2432 2433 2434 2435
		pagefault_disable();
		copied = iov_iter_copy_from_user_atomic(page, i, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

2436
		mark_page_accessed(page);
2437 2438 2439 2440 2441 2442 2443 2444
		status = a_ops->write_end(file, mapping, pos, bytes, copied,
						page, fsdata);
		if (unlikely(status < 0))
			break;
		copied = status;

		cond_resched();

2445
		iov_iter_advance(i, copied);
2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478
		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);

	} 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);
2479
	status = generic_perform_write(file, &i, pos);
L
Linus Torvalds 已提交
2480 2481

	if (likely(status >= 0)) {
2482 2483
		written += status;
		*ppos = pos + status;
L
Linus Torvalds 已提交
2484 2485 2486 2487 2488 2489
  	}
	
	return written ? written : status;
}
EXPORT_SYMBOL(generic_file_buffered_write);

2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510
/**
 * __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
2511 2512
{
	struct file *file = iocb->ki_filp;
2513
	struct address_space * mapping = file->f_mapping;
L
Linus Torvalds 已提交
2514 2515 2516 2517 2518 2519 2520 2521
	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;
2522 2523 2524
	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
	if (err)
		return err;
L
Linus Torvalds 已提交
2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541

	count = ocount;
	pos = *ppos;

	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);

	/* 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;

2542
	err = file_remove_suid(file);
L
Linus Torvalds 已提交
2543 2544 2545
	if (err)
		goto out;

2546
	file_update_time(file);
L
Linus Torvalds 已提交
2547 2548 2549

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (unlikely(file->f_flags & O_DIRECT)) {
2550 2551 2552 2553 2554
		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
2555 2556 2557 2558 2559 2560 2561 2562
		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;
2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576
		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
2577

2578 2579 2580 2581 2582 2583
		/*
		 * 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;
2584
		err = filemap_write_and_wait_range(file->f_mapping, pos, endbyte);
2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599
		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
2600 2601 2602 2603
out:
	current->backing_dev_info = NULL;
	return written ? written : err;
}
2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616
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.
 */
2617 2618
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
2619 2620
{
	struct file *file = iocb->ki_filp;
2621
	struct inode *inode = file->f_mapping->host;
2622
	struct blk_plug plug;
L
Linus Torvalds 已提交
2623 2624 2625 2626
	ssize_t ret;

	BUG_ON(iocb->ki_pos != pos);

2627
	mutex_lock(&inode->i_mutex);
2628
	blk_start_plug(&plug);
2629
	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
2630
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
2631

2632
	if (ret > 0 || ret == -EIOCBQUEUED) {
L
Linus Torvalds 已提交
2633 2634
		ssize_t err;

2635
		err = generic_write_sync(file, pos, ret);
2636
		if (err < 0 && ret > 0)
L
Linus Torvalds 已提交
2637 2638
			ret = err;
	}
2639
	blk_finish_plug(&plug);
L
Linus Torvalds 已提交
2640 2641 2642 2643
	return ret;
}
EXPORT_SYMBOL(generic_file_aio_write);

2644 2645 2646 2647 2648 2649 2650 2651 2652 2653
/**
 * 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.
 *
2654 2655 2656
 * This may also be called if PG_fscache is set on a page, indicating that the
 * page is known to the local caching routines.
 *
2657
 * The @gfp_mask argument specifies whether I/O may be performed to release
2658
 * this page (__GFP_IO), and whether the call may block (__GFP_WAIT & __GFP_FS).
2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674
 *
 */
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);