filemap.c 69.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 *	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/slab.h>
#include <linux/compiler.h>
#include <linux/fs.h>
16
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
17
#include <linux/aio.h>
18
#include <linux/capability.h>
L
Linus Torvalds 已提交
19 20 21 22 23 24 25 26 27
#include <linux/kernel_stat.h>
#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/mm_inline.h> /* for page_is_file_cache() */
37 38
#include "internal.h"

L
Linus Torvalds 已提交
39 40 41 42 43 44 45
/*
 * FIXME: remove all knowledge of the buffer layer from the core VM
 */
#include <linux/buffer_head.h> /* for generic_osync_inode */

#include <asm/mman.h>

46

L
Linus Torvalds 已提交
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
/*
 * 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:
 *
 *  ->i_mmap_lock		(vmtruncate)
 *    ->private_lock		(__free_pte->__set_page_dirty_buffers)
64 65
 *      ->swap_lock		(exclusive_swap_page, others)
 *        ->mapping->tree_lock
L
Linus Torvalds 已提交
66
 *
67
 *  ->i_mutex
L
Linus Torvalds 已提交
68 69 70 71
 *    ->i_mmap_lock		(truncate->unmap_mapping_range)
 *
 *  ->mmap_sem
 *    ->i_mmap_lock
72
 *      ->page_table_lock or pte_lock	(various, mainly in memory.c)
L
Linus Torvalds 已提交
73 74 75 76 77
 *        ->mapping->tree_lock	(arch-dependent flush_dcache_mmap_lock)
 *
 *  ->mmap_sem
 *    ->lock_page		(access_process_vm)
 *
78 79
 *  ->i_mutex			(generic_file_buffered_write)
 *    ->mmap_sem		(fault_in_pages_readable->do_page_fault)
L
Linus Torvalds 已提交
80
 *
81
 *  ->i_mutex
L
Linus Torvalds 已提交
82 83 84 85 86 87 88 89 90 91
 *    ->i_alloc_sem             (various)
 *
 *  ->inode_lock
 *    ->sb_lock			(fs/fs-writeback.c)
 *    ->mapping->tree_lock	(__sync_single_inode)
 *
 *  ->i_mmap_lock
 *    ->anon_vma.lock		(vma_adjust)
 *
 *  ->anon_vma.lock
92
 *    ->page_table_lock or pte_lock	(anon_vma_prepare and various)
L
Linus Torvalds 已提交
93
 *
94
 *  ->page_table_lock or pte_lock
95
 *    ->swap_lock		(try_to_unmap_one)
L
Linus Torvalds 已提交
96 97 98
 *    ->private_lock		(try_to_unmap_one)
 *    ->tree_lock		(try_to_unmap_one)
 *    ->zone.lru_lock		(follow_page->mark_page_accessed)
99
 *    ->zone.lru_lock		(check_pte_range->isolate_lru_page)
L
Linus Torvalds 已提交
100 101 102 103 104 105 106 107 108 109 110 111 112
 *    ->private_lock		(page_remove_rmap->set_page_dirty)
 *    ->tree_lock		(page_remove_rmap->set_page_dirty)
 *    ->inode_lock		(page_remove_rmap->set_page_dirty)
 *    ->inode_lock		(zap_pte_range->set_page_dirty)
 *    ->private_lock		(zap_pte_range->__set_page_dirty_buffers)
 *
 *  ->task->proc_lock
 *    ->dcache_lock		(proc_pid_lookup)
 */

/*
 * Remove a page from the page cache and free it. Caller has to make
 * 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
114 115 116 117 118 119 120 121
 */
void __remove_from_page_cache(struct page *page)
{
	struct address_space *mapping = page->mapping;

	radix_tree_delete(&mapping->page_tree, page->index);
	page->mapping = NULL;
	mapping->nrpages--;
122
	__dec_zone_page_state(page, NR_FILE_PAGES);
123
	BUG_ON(page_mapped(page));
124 125 126 127 128 129 130 131 132 133 134 135

	/*
	 * 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
136 137 138 139 140 141
}

void remove_from_page_cache(struct page *page)
{
	struct address_space *mapping = page->mapping;

M
Matt Mackall 已提交
142
	BUG_ON(!PageLocked(page));
L
Linus Torvalds 已提交
143

N
Nick Piggin 已提交
144
	spin_lock_irq(&mapping->tree_lock);
L
Linus Torvalds 已提交
145
	__remove_from_page_cache(page);
N
Nick Piggin 已提交
146
	spin_unlock_irq(&mapping->tree_lock);
147
	mem_cgroup_uncharge_cache_page(page);
L
Linus Torvalds 已提交
148 149 150 151 152 153 154
}

static int sync_page(void *word)
{
	struct address_space *mapping;
	struct page *page;

155
	page = container_of((unsigned long *)word, struct page, flags);
L
Linus Torvalds 已提交
156 157

	/*
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
	 * page_mapping() is being called without PG_locked held.
	 * Some knowledge of the state and use of the page is used to
	 * reduce the requirements down to a memory barrier.
	 * The danger here is of a stale page_mapping() return value
	 * indicating a struct address_space different from the one it's
	 * associated with when it is associated with one.
	 * After smp_mb(), it's either the correct page_mapping() for
	 * the page, or an old page_mapping() and the page's own
	 * page_mapping() has gone NULL.
	 * The ->sync_page() address_space operation must tolerate
	 * page_mapping() going NULL. By an amazing coincidence,
	 * this comes about because none of the users of the page
	 * in the ->sync_page() methods make essential use of the
	 * page_mapping(), merely passing the page down to the backing
	 * device's unplug functions when it's non-NULL, which in turn
H
Hugh Dickins 已提交
173
	 * ignore it for all cases but swap, where only page_private(page) is
174 175 176
	 * of interest. When page_mapping() does go NULL, the entire
	 * call stack gracefully ignores the page and returns.
	 * -- wli
L
Linus Torvalds 已提交
177 178 179 180 181 182 183 184 185
	 */
	smp_mb();
	mapping = page_mapping(page);
	if (mapping && mapping->a_ops && mapping->a_ops->sync_page)
		mapping->a_ops->sync_page(page);
	io_schedule();
	return 0;
}

M
Matthew Wilcox 已提交
186 187 188 189 190 191
static int sync_page_killable(void *word)
{
	sync_page(word);
	return fatal_signal_pending(current) ? -EINTR : 0;
}

L
Linus Torvalds 已提交
192
/**
193
 * __filemap_fdatawrite_range - start writeback on mapping dirty pages in range
194 195
 * @mapping:	address space structure to write
 * @start:	offset in bytes where the range starts
196
 * @end:	offset in bytes where the range ends (inclusive)
197
 * @sync_mode:	enable synchronous operation
L
Linus Torvalds 已提交
198
 *
199 200 201
 * 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
202
 * If sync_mode is WB_SYNC_ALL then this is a "data integrity" operation, as
203
 * opposed to a regular memory cleansing writeback.  The difference between
L
Linus Torvalds 已提交
204 205 206
 * these two operations is that if a dirty page/buffer is encountered, it must
 * be waited upon, and not just skipped over.
 */
207 208
int __filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
				loff_t end, int sync_mode)
L
Linus Torvalds 已提交
209 210 211 212
{
	int ret;
	struct writeback_control wbc = {
		.sync_mode = sync_mode,
213
		.nr_to_write = LONG_MAX,
214 215
		.range_start = start,
		.range_end = end,
L
Linus Torvalds 已提交
216 217 218 219 220 221 222 223 224 225 226 227
	};

	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)
{
228
	return __filemap_fdatawrite_range(mapping, 0, LLONG_MAX, sync_mode);
L
Linus Torvalds 已提交
229 230 231 232 233 234 235 236
}

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

237
int filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
238
				loff_t end)
L
Linus Torvalds 已提交
239 240 241
{
	return __filemap_fdatawrite_range(mapping, start, end, WB_SYNC_ALL);
}
242
EXPORT_SYMBOL(filemap_fdatawrite_range);
L
Linus Torvalds 已提交
243

244 245 246 247
/**
 * filemap_flush - mostly a non-blocking flush
 * @mapping:	target address_space
 *
L
Linus Torvalds 已提交
248 249 250 251 252 253 254 255 256
 * 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);

257 258 259 260 261 262
/**
 * wait_on_page_writeback_range - wait for writeback to complete
 * @mapping:	target address_space
 * @start:	beginning page index
 * @end:	ending page index
 *
L
Linus Torvalds 已提交
263 264 265
 * Wait for writeback to complete against pages indexed by start->end
 * inclusive
 */
266
int wait_on_page_writeback_range(struct address_space *mapping,
L
Linus Torvalds 已提交
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
				pgoff_t start, pgoff_t end)
{
	struct pagevec pvec;
	int nr_pages;
	int ret = 0;
	pgoff_t index;

	if (end < start)
		return 0;

	pagevec_init(&pvec, 0);
	index = start;
	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);
			if (PageError(page))
				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;
}

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
/**
 * filemap_fdatawait_range - wait for all under-writeback pages to complete in a given range
 * @mapping: address space structure to wait for
 * @start:	offset in bytes where the range starts
 * @end:	offset in bytes where the range ends (inclusive)
 *
 * Walk the list of under-writeback pages of the given address space
 * in the given range and wait for all of them.
 *
 * This is just a simple wrapper so that callers don't have to convert offsets
 * to page indexes themselves
 */
int filemap_fdatawait_range(struct address_space *mapping, loff_t start,
			    loff_t end)
{
	return wait_on_page_writeback_range(mapping, start >> PAGE_CACHE_SHIFT,
					    end >> PAGE_CACHE_SHIFT);
}
EXPORT_SYMBOL(filemap_fdatawait_range);

329 330 331 332 333 334 335
/**
 * sync_page_range - write and wait on all pages in the passed range
 * @inode:	target inode
 * @mapping:	target address_space
 * @pos:	beginning offset in pages to write
 * @count:	number of bytes to write
 *
L
Linus Torvalds 已提交
336 337 338 339
 * Write and wait upon all the pages in the passed range.  This is a "data
 * integrity" operation.  It waits upon in-flight writeout before starting and
 * waiting upon new writeout.  If there was an IO error, return it.
 *
340
 * We need to re-take i_mutex during the generic_osync_inode list walk because
L
Linus Torvalds 已提交
341 342 343
 * it is otherwise livelockable.
 */
int sync_page_range(struct inode *inode, struct address_space *mapping,
344
			loff_t pos, loff_t count)
L
Linus Torvalds 已提交
345 346 347 348 349 350 351 352 353
{
	pgoff_t start = pos >> PAGE_CACHE_SHIFT;
	pgoff_t end = (pos + count - 1) >> PAGE_CACHE_SHIFT;
	int ret;

	if (!mapping_cap_writeback_dirty(mapping) || !count)
		return 0;
	ret = filemap_fdatawrite_range(mapping, pos, pos + count - 1);
	if (ret == 0) {
354
		mutex_lock(&inode->i_mutex);
L
Linus Torvalds 已提交
355
		ret = generic_osync_inode(inode, mapping, OSYNC_METADATA);
356
		mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
357 358 359 360 361 362 363
	}
	if (ret == 0)
		ret = wait_on_page_writeback_range(mapping, start, end);
	return ret;
}
EXPORT_SYMBOL(sync_page_range);

364
/**
365
 * sync_page_range_nolock - write & wait on all pages in the passed range without locking
366 367 368 369 370
 * @inode:	target inode
 * @mapping:	target address_space
 * @pos:	beginning offset in pages to write
 * @count:	number of bytes to write
 *
371
 * Note: Holding i_mutex across sync_page_range_nolock() is not a good idea
L
Linus Torvalds 已提交
372 373 374
 * as it forces O_SYNC writers to different parts of the same file
 * to be serialised right until io completion.
 */
375 376
int sync_page_range_nolock(struct inode *inode, struct address_space *mapping,
			   loff_t pos, loff_t count)
L
Linus Torvalds 已提交
377 378 379 380 381 382 383 384 385 386 387 388 389 390
{
	pgoff_t start = pos >> PAGE_CACHE_SHIFT;
	pgoff_t end = (pos + count - 1) >> PAGE_CACHE_SHIFT;
	int ret;

	if (!mapping_cap_writeback_dirty(mapping) || !count)
		return 0;
	ret = filemap_fdatawrite_range(mapping, pos, pos + count - 1);
	if (ret == 0)
		ret = generic_osync_inode(inode, mapping, OSYNC_METADATA);
	if (ret == 0)
		ret = wait_on_page_writeback_range(mapping, start, end);
	return ret;
}
391
EXPORT_SYMBOL(sync_page_range_nolock);
L
Linus Torvalds 已提交
392 393

/**
394
 * filemap_fdatawait - wait for all under-writeback pages to complete
L
Linus Torvalds 已提交
395
 * @mapping: address space structure to wait for
396 397 398
 *
 * Walk the list of under-writeback pages of the given address space
 * and wait for all of them.
L
Linus Torvalds 已提交
399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
 */
int filemap_fdatawait(struct address_space *mapping)
{
	loff_t i_size = i_size_read(mapping->host);

	if (i_size == 0)
		return 0;

	return wait_on_page_writeback_range(mapping, 0,
				(i_size - 1) >> PAGE_CACHE_SHIFT);
}
EXPORT_SYMBOL(filemap_fdatawait);

int filemap_write_and_wait(struct address_space *mapping)
{
414
	int err = 0;
L
Linus Torvalds 已提交
415 416

	if (mapping->nrpages) {
417 418 419 420 421 422 423 424 425 426 427 428
		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
429
	}
430
	return err;
L
Linus Torvalds 已提交
431
}
432
EXPORT_SYMBOL(filemap_write_and_wait);
L
Linus Torvalds 已提交
433

434 435 436 437 438 439
/**
 * 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)
 *
440 441 442 443 444
 * 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
445 446 447
int filemap_write_and_wait_range(struct address_space *mapping,
				 loff_t lstart, loff_t lend)
{
448
	int err = 0;
L
Linus Torvalds 已提交
449 450

	if (mapping->nrpages) {
451 452 453 454 455 456 457 458 459 460
		err = __filemap_fdatawrite_range(mapping, lstart, lend,
						 WB_SYNC_ALL);
		/* See comment of filemap_write_and_wait() */
		if (err != -EIO) {
			int err2 = wait_on_page_writeback_range(mapping,
						lstart >> PAGE_CACHE_SHIFT,
						lend >> PAGE_CACHE_SHIFT);
			if (!err)
				err = err2;
		}
L
Linus Torvalds 已提交
461
	}
462
	return err;
L
Linus Torvalds 已提交
463
}
464
EXPORT_SYMBOL(filemap_write_and_wait_range);
L
Linus Torvalds 已提交
465

466
/**
N
Nick Piggin 已提交
467
 * add_to_page_cache_locked - add a locked page to the pagecache
468 469 470 471 472
 * @page:	page to add
 * @mapping:	the page's address_space
 * @offset:	page index
 * @gfp_mask:	page allocation mode
 *
N
Nick Piggin 已提交
473
 * This function is used to add a page to the pagecache. It must be locked.
L
Linus Torvalds 已提交
474 475
 * This function does not add the page to the LRU.  The caller must do that.
 */
N
Nick Piggin 已提交
476
int add_to_page_cache_locked(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
477
		pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
478
{
N
Nick Piggin 已提交
479 480 481 482 483
	int error;

	VM_BUG_ON(!PageLocked(page));

	error = mem_cgroup_cache_charge(page, current->mm,
K
KAMEZAWA Hiroyuki 已提交
484
					gfp_mask & GFP_RECLAIM_MASK);
B
Balbir Singh 已提交
485 486
	if (error)
		goto out;
L
Linus Torvalds 已提交
487

B
Balbir Singh 已提交
488
	error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
L
Linus Torvalds 已提交
489
	if (error == 0) {
N
Nick Piggin 已提交
490 491 492 493
		page_cache_get(page);
		page->mapping = mapping;
		page->index = offset;

N
Nick Piggin 已提交
494
		spin_lock_irq(&mapping->tree_lock);
L
Linus Torvalds 已提交
495
		error = radix_tree_insert(&mapping->page_tree, offset, page);
N
Nick Piggin 已提交
496
		if (likely(!error)) {
L
Linus Torvalds 已提交
497
			mapping->nrpages++;
498
			__inc_zone_page_state(page, NR_FILE_PAGES);
499
			spin_unlock_irq(&mapping->tree_lock);
N
Nick Piggin 已提交
500 501
		} else {
			page->mapping = NULL;
502
			spin_unlock_irq(&mapping->tree_lock);
503
			mem_cgroup_uncharge_cache_page(page);
N
Nick Piggin 已提交
504 505
			page_cache_release(page);
		}
L
Linus Torvalds 已提交
506
		radix_tree_preload_end();
B
Balbir Singh 已提交
507
	} else
508
		mem_cgroup_uncharge_cache_page(page);
509
out:
L
Linus Torvalds 已提交
510 511
	return error;
}
N
Nick Piggin 已提交
512
EXPORT_SYMBOL(add_to_page_cache_locked);
L
Linus Torvalds 已提交
513 514

int add_to_page_cache_lru(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
515
				pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
516
{
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
	int ret;

	/*
	 * Splice_read and readahead add shmem/tmpfs pages into the page cache
	 * before shmem_readpage has a chance to mark them as SwapBacked: they
	 * need to go on the active_anon lru below, and mem_cgroup_cache_charge
	 * (called in add_to_page_cache) needs to know where they're going too.
	 */
	if (mapping_cap_swap_backed(mapping))
		SetPageSwapBacked(page);

	ret = add_to_page_cache(page, mapping, offset, gfp_mask);
	if (ret == 0) {
		if (page_is_file_cache(page))
			lru_cache_add_file(page);
		else
			lru_cache_add_active_anon(page);
	}
L
Linus Torvalds 已提交
535 536
	return ret;
}
537
EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
L
Linus Torvalds 已提交
538

539
#ifdef CONFIG_NUMA
540
struct page *__page_cache_alloc(gfp_t gfp)
541 542 543
{
	if (cpuset_do_page_mem_spread()) {
		int n = cpuset_mem_spread_node();
544
		return alloc_pages_exact_node(n, gfp, 0);
545
	}
546
	return alloc_pages(gfp, 0);
547
}
548
EXPORT_SYMBOL(__page_cache_alloc);
549 550
#endif

551 552 553 554 555 556
static int __sleep_on_page_lock(void *word)
{
	io_schedule();
	return 0;
}

L
Linus Torvalds 已提交
557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578
/*
 * 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
579
void wait_on_page_bit(struct page *page, int bit_nr)
L
Linus Torvalds 已提交
580 581 582 583 584 585 586 587 588
{
	DEFINE_WAIT_BIT(wait, &page->flags, bit_nr);

	if (test_bit(bit_nr, &page->flags))
		__wait_on_bit(page_waitqueue(page), &wait, sync_page,
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(wait_on_page_bit);

589 590
/**
 * add_page_wait_queue - Add an arbitrary waiter to a page's wait queue
R
Randy Dunlap 已提交
591 592
 * @page: Page defining the wait queue of interest
 * @waiter: Waiter to add to the queue
593 594 595 596 597 598 599 600 601 602 603 604 605 606
 *
 * 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
607
/**
608
 * unlock_page - unlock a locked page
L
Linus Torvalds 已提交
609 610 611 612 613 614 615
 * @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
616 617
 * 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
618
 */
H
Harvey Harrison 已提交
619
void unlock_page(struct page *page)
L
Linus Torvalds 已提交
620
{
N
Nick Piggin 已提交
621 622 623
	VM_BUG_ON(!PageLocked(page));
	clear_bit_unlock(PG_locked, &page->flags);
	smp_mb__after_clear_bit();
L
Linus Torvalds 已提交
624 625 626 627
	wake_up_page(page, PG_locked);
}
EXPORT_SYMBOL(unlock_page);

628 629 630
/**
 * end_page_writeback - end writeback against a page
 * @page: the page
L
Linus Torvalds 已提交
631 632 633
 */
void end_page_writeback(struct page *page)
{
634 635 636 637 638 639
	if (TestClearPageReclaim(page))
		rotate_reclaimable_page(page);

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

L
Linus Torvalds 已提交
640 641 642 643 644
	smp_mb__after_clear_bit();
	wake_up_page(page, PG_writeback);
}
EXPORT_SYMBOL(end_page_writeback);

645 646 647
/**
 * __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
648
 *
649
 * Ugly. Running sync_page() in state TASK_UNINTERRUPTIBLE is scary.  If some
L
Linus Torvalds 已提交
650 651 652 653
 * random driver's requestfn sets TASK_RUNNING, we could busywait.  However
 * chances are that on the second loop, the block layer's plug list is empty,
 * so sync_page() will then return in state TASK_UNINTERRUPTIBLE.
 */
H
Harvey Harrison 已提交
654
void __lock_page(struct page *page)
L
Linus Torvalds 已提交
655 656 657 658 659 660 661 662
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

	__wait_on_bit_lock(page_waitqueue(page), &wait, sync_page,
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(__lock_page);

H
Harvey Harrison 已提交
663
int __lock_page_killable(struct page *page)
M
Matthew Wilcox 已提交
664 665 666 667 668 669
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

	return __wait_on_bit_lock(page_waitqueue(page), &wait,
					sync_page_killable, TASK_KILLABLE);
}
670
EXPORT_SYMBOL_GPL(__lock_page_killable);
M
Matthew Wilcox 已提交
671

672 673 674 675
/**
 * __lock_page_nosync - get a lock on the page, without calling sync_page()
 * @page: the page to lock
 *
676 677 678
 * Variant of lock_page that does not require the caller to hold a reference
 * on the page's mapping.
 */
H
Harvey Harrison 已提交
679
void __lock_page_nosync(struct page *page)
680 681 682 683 684 685
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);
	__wait_on_bit_lock(page_waitqueue(page), &wait, __sleep_on_page_lock,
							TASK_UNINTERRUPTIBLE);
}

686 687 688 689 690
/**
 * 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
691 692
 * 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
693
 */
N
Nick Piggin 已提交
694
struct page *find_get_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
695
{
N
Nick Piggin 已提交
696
	void **pagep;
L
Linus Torvalds 已提交
697 698
	struct page *page;

N
Nick Piggin 已提交
699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722
	rcu_read_lock();
repeat:
	page = NULL;
	pagep = radix_tree_lookup_slot(&mapping->page_tree, offset);
	if (pagep) {
		page = radix_tree_deref_slot(pagep);
		if (unlikely(!page || page == RADIX_TREE_RETRY))
			goto repeat;

		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;
		}
	}
	rcu_read_unlock();

L
Linus Torvalds 已提交
723 724 725 726 727 728
	return page;
}
EXPORT_SYMBOL(find_get_page);

/**
 * find_lock_page - locate, pin and lock a pagecache page
729 730
 * @mapping: the address_space to search
 * @offset: the page index
L
Linus Torvalds 已提交
731 732 733 734 735 736
 *
 * 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
737
struct page *find_lock_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
738 739 740 741
{
	struct page *page;

repeat:
N
Nick Piggin 已提交
742
	page = find_get_page(mapping, offset);
L
Linus Torvalds 已提交
743
	if (page) {
N
Nick Piggin 已提交
744 745 746 747 748 749
		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
750
		}
N
Nick Piggin 已提交
751
		VM_BUG_ON(page->index != offset);
L
Linus Torvalds 已提交
752 753 754 755 756 757 758
	}
	return page;
}
EXPORT_SYMBOL(find_lock_page);

/**
 * find_or_create_page - locate or add a pagecache page
759 760 761
 * @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
762 763 764 765 766 767 768 769 770 771 772 773 774
 *
 * 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,
775
		pgoff_t index, gfp_t gfp_mask)
L
Linus Torvalds 已提交
776
{
N
Nick Piggin 已提交
777
	struct page *page;
L
Linus Torvalds 已提交
778 779 780 781
	int err;
repeat:
	page = find_lock_page(mapping, index);
	if (!page) {
N
Nick Piggin 已提交
782 783 784
		page = __page_cache_alloc(gfp_mask);
		if (!page)
			return NULL;
N
Nick Piggin 已提交
785 786 787 788 789 790 791 792
		/*
		 * 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
793 794 795 796 797
		if (unlikely(err)) {
			page_cache_release(page);
			page = NULL;
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
		}
	}
	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;
N
Nick Piggin 已提交
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852
	unsigned int nr_found;

	rcu_read_lock();
restart:
	nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree,
				(void ***)pages, start, nr_pages);
	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;
		/*
		 * this can only trigger if nr_found == 1, making livelock
		 * a non issue.
		 */
		if (unlikely(page == RADIX_TREE_RETRY))
			goto restart;

		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
853

N
Nick Piggin 已提交
854 855 856 857
		pages[ret] = page;
		ret++;
	}
	rcu_read_unlock();
L
Linus Torvalds 已提交
858 859 860
	return ret;
}

861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877
/**
 * 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
878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
	unsigned int nr_found;

	rcu_read_lock();
restart:
	nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree,
				(void ***)pages, index, nr_pages);
	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;
		/*
		 * this can only trigger if nr_found == 1, making livelock
		 * a non issue.
		 */
		if (unlikely(page == RADIX_TREE_RETRY))
			goto restart;
897

N
Nick Piggin 已提交
898
		if (page->mapping == NULL || page->index != index)
899 900
			break;

N
Nick Piggin 已提交
901 902 903 904 905 906 907 908 909 910 911
		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++;
912 913
		index++;
	}
N
Nick Piggin 已提交
914 915
	rcu_read_unlock();
	return ret;
916
}
917
EXPORT_SYMBOL(find_get_pages_contig);
918

919 920 921 922 923 924 925 926
/**
 * 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
927
 * Like find_get_pages, except we only return pages which are tagged with
928
 * @tag.   We update @index to index the next page for the traversal.
L
Linus Torvalds 已提交
929 930 931 932 933 934
 */
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
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967
	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;
		/*
		 * this can only trigger if nr_found == 1, making livelock
		 * a non issue.
		 */
		if (unlikely(page == RADIX_TREE_RETRY))
			goto restart;

		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++;
	}
	rcu_read_unlock();
L
Linus Torvalds 已提交
968 969 970

	if (ret)
		*index = pages[ret - 1]->index + 1;
N
Nick Piggin 已提交
971

L
Linus Torvalds 已提交
972 973
	return ret;
}
974
EXPORT_SYMBOL(find_get_pages_tag);
L
Linus Torvalds 已提交
975

976 977 978 979 980
/**
 * grab_cache_page_nowait - returns locked page at given index in given cache
 * @mapping: target address_space
 * @index: the page index
 *
981
 * Same as grab_cache_page(), but do not wait if the page is unavailable.
L
Linus Torvalds 已提交
982 983 984 985 986 987 988 989
 * 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 *
990
grab_cache_page_nowait(struct address_space *mapping, pgoff_t index)
L
Linus Torvalds 已提交
991 992 993 994
{
	struct page *page = find_get_page(mapping, index);

	if (page) {
N
Nick Piggin 已提交
995
		if (trylock_page(page))
L
Linus Torvalds 已提交
996 997 998 999
			return page;
		page_cache_release(page);
		return NULL;
	}
1000
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~__GFP_FS);
N
Nick Piggin 已提交
1001
	if (page && add_to_page_cache_lru(page, mapping, index, GFP_NOFS)) {
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008
		page_cache_release(page);
		page = NULL;
	}
	return page;
}
EXPORT_SYMBOL(grab_cache_page_nowait);

1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029
/*
 * 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;
}

1030
/**
C
Christoph Hellwig 已提交
1031
 * do_generic_file_read - generic file read routine
1032 1033 1034 1035 1036
 * @filp:	the file to read
 * @ppos:	current file position
 * @desc:	read_descriptor
 * @actor:	read method
 *
L
Linus Torvalds 已提交
1037
 * This is a generic file read routine, and uses the
1038
 * mapping->a_ops->readpage() function for the actual low-level stuff.
L
Linus Torvalds 已提交
1039 1040 1041 1042
 *
 * 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
1043 1044
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
1045
{
C
Christoph Hellwig 已提交
1046
	struct address_space *mapping = filp->f_mapping;
L
Linus Torvalds 已提交
1047
	struct inode *inode = mapping->host;
C
Christoph Hellwig 已提交
1048
	struct file_ra_state *ra = &filp->f_ra;
1049 1050 1051 1052
	pgoff_t index;
	pgoff_t last_index;
	pgoff_t prev_index;
	unsigned long offset;      /* offset into pagecache page */
1053
	unsigned int prev_offset;
L
Linus Torvalds 已提交
1054 1055 1056
	int error;

	index = *ppos >> PAGE_CACHE_SHIFT;
1057 1058
	prev_index = ra->prev_pos >> PAGE_CACHE_SHIFT;
	prev_offset = ra->prev_pos & (PAGE_CACHE_SIZE-1);
L
Linus Torvalds 已提交
1059 1060 1061 1062 1063
	last_index = (*ppos + desc->count + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	for (;;) {
		struct page *page;
1064
		pgoff_t end_index;
N
NeilBrown 已提交
1065
		loff_t isize;
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070
		unsigned long nr, ret;

		cond_resched();
find_page:
		page = find_get_page(mapping, index);
1071
		if (!page) {
1072
			page_cache_sync_readahead(mapping,
1073
					ra, filp,
1074 1075 1076 1077 1078 1079
					index, last_index - index);
			page = find_get_page(mapping, index);
			if (unlikely(page == NULL))
				goto no_cached_page;
		}
		if (PageReadahead(page)) {
1080
			page_cache_async_readahead(mapping,
1081
					ra, filp, page,
1082
					index, last_index - index);
L
Linus Torvalds 已提交
1083
		}
1084 1085 1086 1087
		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
1088
			if (!trylock_page(page))
1089 1090 1091 1092 1093 1094
				goto page_not_up_to_date;
			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
1095
page_ok:
N
NeilBrown 已提交
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121
		/*
		 * 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
1122 1123 1124 1125 1126 1127 1128 1129 1130

		/* 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);

		/*
1131 1132
		 * When a sequential read accesses a page several times,
		 * only mark it as accessed the first time.
L
Linus Torvalds 已提交
1133
		 */
1134
		if (prev_index != index || offset != prev_offset)
L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
			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
1152
		prev_offset = offset;
L
Linus Torvalds 已提交
1153 1154 1155 1156 1157 1158 1159 1160

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

page_not_up_to_date:
		/* Get exclusive access to the page ... */
1161 1162 1163
		error = lock_page_killable(page);
		if (unlikely(error))
			goto readpage_error;
L
Linus Torvalds 已提交
1164

1165
page_not_up_to_date_locked:
N
Nick Piggin 已提交
1166
		/* Did it get truncated before we got the lock? */
L
Linus Torvalds 已提交
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
		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:
		/* Start the actual read. The read will unlock the page. */
		error = mapping->a_ops->readpage(filp, page);

1183 1184 1185 1186 1187
		if (unlikely(error)) {
			if (error == AOP_TRUNCATED_PAGE) {
				page_cache_release(page);
				goto find_page;
			}
L
Linus Torvalds 已提交
1188
			goto readpage_error;
1189
		}
L
Linus Torvalds 已提交
1190 1191

		if (!PageUptodate(page)) {
1192 1193 1194
			error = lock_page_killable(page);
			if (unlikely(error))
				goto readpage_error;
L
Linus Torvalds 已提交
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
			if (!PageUptodate(page)) {
				if (page->mapping == NULL) {
					/*
					 * invalidate_inode_pages got it
					 */
					unlock_page(page);
					page_cache_release(page);
					goto find_page;
				}
				unlock_page(page);
1205
				shrink_readahead_size_eio(filp, ra);
1206 1207
				error = -EIO;
				goto readpage_error;
L
Linus Torvalds 已提交
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
			}
			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
1225 1226 1227 1228
		page = page_cache_alloc_cold(mapping);
		if (!page) {
			desc->error = -ENOMEM;
			goto out;
L
Linus Torvalds 已提交
1229
		}
N
Nick Piggin 已提交
1230
		error = add_to_page_cache_lru(page, mapping,
L
Linus Torvalds 已提交
1231 1232
						index, GFP_KERNEL);
		if (error) {
N
Nick Piggin 已提交
1233
			page_cache_release(page);
L
Linus Torvalds 已提交
1234 1235 1236 1237 1238 1239 1240 1241 1242
			if (error == -EEXIST)
				goto find_page;
			desc->error = error;
			goto out;
		}
		goto readpage;
	}

out:
1243 1244 1245
	ra->prev_pos = prev_index;
	ra->prev_pos <<= PAGE_CACHE_SHIFT;
	ra->prev_pos |= prev_offset;
L
Linus Torvalds 已提交
1246

1247
	*ppos = ((loff_t)index << PAGE_CACHE_SHIFT) + offset;
1248
	file_accessed(filp);
L
Linus Torvalds 已提交
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288
}

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;
}

1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327
/*
 * 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);

1328
/**
H
Henrik Kretzschmar 已提交
1329
 * generic_file_aio_read - generic filesystem read routine
1330 1331 1332
 * @iocb:	kernel I/O control block
 * @iov:	io vector request
 * @nr_segs:	number of segments in the iovec
H
Henrik Kretzschmar 已提交
1333
 * @pos:	current file position
1334
 *
L
Linus Torvalds 已提交
1335 1336 1337 1338
 * This is the "read()" routine for all filesystems
 * that can use the page cache directly.
 */
ssize_t
1339 1340
generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
1341 1342 1343 1344 1345
{
	struct file *filp = iocb->ki_filp;
	ssize_t retval;
	unsigned long seg;
	size_t count;
1346
	loff_t *ppos = &iocb->ki_pos;
L
Linus Torvalds 已提交
1347 1348

	count = 0;
1349 1350 1351
	retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
	if (retval)
		return retval;
L
Linus Torvalds 已提交
1352 1353 1354

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (filp->f_flags & O_DIRECT) {
1355
		loff_t size;
L
Linus Torvalds 已提交
1356 1357 1358 1359 1360 1361 1362 1363 1364
		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) {
1365 1366
			retval = filemap_write_and_wait_range(mapping, pos,
					pos + iov_length(iov, nr_segs) - 1);
1367 1368 1369 1370
			if (!retval) {
				retval = mapping->a_ops->direct_IO(READ, iocb,
							iov, pos, nr_segs);
			}
L
Linus Torvalds 已提交
1371 1372
			if (retval > 0)
				*ppos = pos + retval;
H
Hugh Dickins 已提交
1373 1374 1375 1376
			if (retval) {
				file_accessed(filp);
				goto out;
			}
1377
		}
L
Linus Torvalds 已提交
1378 1379
	}

H
Hugh Dickins 已提交
1380 1381
	for (seg = 0; seg < nr_segs; seg++) {
		read_descriptor_t desc;
L
Linus Torvalds 已提交
1382

H
Hugh Dickins 已提交
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393
		desc.written = 0;
		desc.arg.buf = iov[seg].iov_base;
		desc.count = iov[seg].iov_len;
		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
1394
		}
H
Hugh Dickins 已提交
1395 1396
		if (desc.count > 0)
			break;
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403 1404
	}
out:
	return retval;
}
EXPORT_SYMBOL(generic_file_aio_read);

static ssize_t
do_readahead(struct address_space *mapping, struct file *filp,
1405
	     pgoff_t index, unsigned long nr)
L
Linus Torvalds 已提交
1406 1407 1408 1409
{
	if (!mapping || !mapping->a_ops || !mapping->a_ops->readpage)
		return -EINVAL;

1410
	force_page_cache_readahead(mapping, filp, index, nr);
L
Linus Torvalds 已提交
1411 1412 1413
	return 0;
}

1414
SYSCALL_DEFINE(readahead)(int fd, loff_t offset, size_t count)
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423
{
	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;
1424 1425
			pgoff_t start = offset >> PAGE_CACHE_SHIFT;
			pgoff_t end = (offset + count - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
1426 1427 1428 1429 1430 1431 1432
			unsigned long len = end - start + 1;
			ret = do_readahead(mapping, file, start, len);
		}
		fput(file);
	}
	return ret;
}
1433 1434 1435 1436 1437 1438 1439
#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
1440 1441

#ifdef CONFIG_MMU
1442 1443 1444 1445 1446
/**
 * 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
1447 1448 1449
 * 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
1450
static int page_cache_read(struct file *file, pgoff_t offset)
L
Linus Torvalds 已提交
1451 1452 1453
{
	struct address_space *mapping = file->f_mapping;
	struct page *page; 
1454
	int ret;
L
Linus Torvalds 已提交
1455

1456 1457 1458 1459 1460 1461 1462 1463 1464 1465
	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
1466 1467 1468

		page_cache_release(page);

1469 1470 1471
	} while (ret == AOP_TRUNCATED_PAGE);
		
	return ret;
L
Linus Torvalds 已提交
1472 1473 1474 1475
}

#define MMAP_LOTSAMISS  (100)

1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491
/*
 * 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;

1492 1493
	if (VM_SequentialReadHint(vma) ||
			offset - 1 == (ra->prev_pos >> PAGE_CACHE_SHIFT)) {
1494 1495
		page_cache_sync_readahead(mapping, ra, file, offset,
					  ra->ra_pages);
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
		return;
	}

	if (ra->mmap_miss < INT_MAX)
		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;

1509 1510 1511
	/*
	 * mmap read-around
	 */
1512 1513
	ra_pages = max_sane_readahead(ra->ra_pages);
	if (ra_pages) {
1514 1515 1516 1517
		ra->start = max_t(long, 0, offset - ra_pages/2);
		ra->size = ra_pages;
		ra->async_size = 0;
		ra_submit(ra, mapping, file);
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
	}
}

/*
 * 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))
1539 1540
		page_cache_async_readahead(mapping, ra, file,
					   page, offset, ra->ra_pages);
1541 1542
}

1543
/**
1544
 * filemap_fault - read in file data for page fault handling
N
Nick Piggin 已提交
1545 1546
 * @vma:	vma in which the fault was taken
 * @vmf:	struct vm_fault containing details of the fault
1547
 *
1548
 * filemap_fault() is invoked via the vma operations vector for a
L
Linus Torvalds 已提交
1549 1550 1551 1552 1553 1554
 * 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
1555
int filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1556 1557
{
	int error;
1558
	struct file *file = vma->vm_file;
L
Linus Torvalds 已提交
1559 1560 1561
	struct address_space *mapping = file->f_mapping;
	struct file_ra_state *ra = &file->f_ra;
	struct inode *inode = mapping->host;
1562
	pgoff_t offset = vmf->pgoff;
L
Linus Torvalds 已提交
1563
	struct page *page;
J
Jan Kara 已提交
1564
	pgoff_t size;
N
Nick Piggin 已提交
1565
	int ret = 0;
L
Linus Torvalds 已提交
1566 1567

	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1568
	if (offset >= size)
1569
		return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1570 1571 1572 1573

	/*
	 * Do we have something in the page cache already?
	 */
1574 1575
	page = find_get_page(mapping, offset);
	if (likely(page)) {
L
Linus Torvalds 已提交
1576
		/*
1577 1578
		 * We found the page, so try async readahead before
		 * waiting for the lock.
L
Linus Torvalds 已提交
1579
		 */
1580 1581
		do_async_mmap_readahead(vma, ra, file, page, offset);
		lock_page(page);
L
Linus Torvalds 已提交
1582

1583 1584 1585 1586 1587
		/* Did it get truncated? */
		if (unlikely(page->mapping != mapping)) {
			unlock_page(page);
			put_page(page);
			goto no_cached_page;
L
Linus Torvalds 已提交
1588
		}
1589 1590 1591 1592 1593 1594 1595
	} else {
		/* No page in the page cache at all */
		do_sync_mmap_readahead(vma, ra, file, offset);
		count_vm_event(PGMAJFAULT);
		ret = VM_FAULT_MAJOR;
retry_find:
		page = find_lock_page(mapping, offset);
L
Linus Torvalds 已提交
1596 1597 1598 1599 1600
		if (!page)
			goto no_cached_page;
	}

	/*
1601 1602
	 * 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
1603
	 */
1604
	if (unlikely(!PageUptodate(page)))
L
Linus Torvalds 已提交
1605 1606
		goto page_not_uptodate;

1607 1608 1609 1610
	/*
	 * Found the page and have a reference on it.
	 * We must recheck i_size under page lock.
	 */
1611
	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1612
	if (unlikely(offset >= size)) {
1613
		unlock_page(page);
1614
		page_cache_release(page);
1615
		return VM_FAULT_SIGBUS;
1616 1617
	}

1618
	ra->prev_pos = (loff_t)offset << PAGE_CACHE_SHIFT;
N
Nick Piggin 已提交
1619
	vmf->page = page;
N
Nick Piggin 已提交
1620
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1621 1622 1623 1624 1625 1626

no_cached_page:
	/*
	 * We're only likely to ever get here if MADV_RANDOM is in
	 * effect.
	 */
1627
	error = page_cache_read(file, offset);
L
Linus Torvalds 已提交
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642

	/*
	 * 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
1643 1644
		return VM_FAULT_OOM;
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653

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);
1654
	error = mapping->a_ops->readpage(file, page);
1655 1656 1657 1658 1659
	if (!error) {
		wait_on_page_locked(page);
		if (!PageUptodate(page))
			error = -EIO;
	}
1660 1661 1662
	page_cache_release(page);

	if (!error || error == AOP_TRUNCATED_PAGE)
1663
		goto retry_find;
L
Linus Torvalds 已提交
1664

1665
	/* Things didn't work out. Return zero to tell the mm layer so. */
1666
	shrink_readahead_size_eio(file, ra);
N
Nick Piggin 已提交
1667
	return VM_FAULT_SIGBUS;
1668 1669 1670
}
EXPORT_SYMBOL(filemap_fault);

L
Linus Torvalds 已提交
1671
struct vm_operations_struct generic_file_vm_ops = {
1672
	.fault		= filemap_fault,
L
Linus Torvalds 已提交
1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
};

/* 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
1685
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711
	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);

1712
static struct page *__read_cache_page(struct address_space *mapping,
1713
				pgoff_t index,
L
Linus Torvalds 已提交
1714 1715 1716
				int (*filler)(void *,struct page*),
				void *data)
{
N
Nick Piggin 已提交
1717
	struct page *page;
L
Linus Torvalds 已提交
1718 1719 1720 1721
	int err;
repeat:
	page = find_get_page(mapping, index);
	if (!page) {
N
Nick Piggin 已提交
1722 1723 1724 1725 1726 1727 1728 1729
		page = page_cache_alloc_cold(mapping);
		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
1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
			/* 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;
}

1742 1743 1744 1745 1746 1747 1748
/**
 * 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
 * @data:	destination for read data
 *
1749 1750
 * Same as read_cache_page, but don't wait for page to become unlocked
 * after submitting it to the filler.
1751 1752 1753 1754 1755
 *
 * 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.
L
Linus Torvalds 已提交
1756
 */
1757
struct page *read_cache_page_async(struct address_space *mapping,
1758
				pgoff_t index,
L
Linus Torvalds 已提交
1759 1760 1761 1762 1763 1764 1765 1766 1767
				int (*filler)(void *,struct page*),
				void *data)
{
	struct page *page;
	int err;

retry:
	page = __read_cache_page(mapping, index, filler, data);
	if (IS_ERR(page))
1768
		return page;
L
Linus Torvalds 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784
	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);
1785
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1786
	}
1787
out:
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
	mark_page_accessed(page);
	return page;
}
EXPORT_SYMBOL(read_cache_page_async);

/**
 * 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
 * @data:	destination for read data
 *
 * 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,
1806
				pgoff_t index,
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819
				int (*filler)(void *,struct page*),
				void *data)
{
	struct page *page;

	page = read_cache_page_async(mapping, index, filler, data);
	if (IS_ERR(page))
		goto out;
	wait_on_page_locked(page);
	if (!PageUptodate(page)) {
		page_cache_release(page);
		page = ERR_PTR(-EIO);
	}
L
Linus Torvalds 已提交
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
 out:
	return page;
}
EXPORT_SYMBOL(read_cache_page);

/*
 * The logic we want is
 *
 *	if suid or (sgid and xgrp)
 *		remove privs
 */
1831
int should_remove_suid(struct dentry *dentry)
L
Linus Torvalds 已提交
1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846
{
	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;

1847
	if (unlikely(kill && !capable(CAP_FSETID) && S_ISREG(mode)))
1848
		return kill;
L
Linus Torvalds 已提交
1849

1850 1851
	return 0;
}
M
Mark Fasheh 已提交
1852
EXPORT_SYMBOL(should_remove_suid);
1853

1854
static int __remove_suid(struct dentry *dentry, int kill)
1855 1856 1857 1858 1859 1860 1861
{
	struct iattr newattrs;

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

1862
int file_remove_suid(struct file *file)
1863
{
1864
	struct dentry *dentry = file->f_path.dentry;
1865 1866 1867
	int killsuid = should_remove_suid(dentry);
	int killpriv = security_inode_need_killpriv(dentry);
	int error = 0;
1868

1869 1870 1871 1872 1873 1874
	if (killpriv < 0)
		return killpriv;
	if (killpriv)
		error = security_inode_killpriv(dentry);
	if (!error && killsuid)
		error = __remove_suid(dentry, killsuid);
1875

1876
	return error;
L
Linus Torvalds 已提交
1877
}
1878
EXPORT_SYMBOL(file_remove_suid);
L
Linus Torvalds 已提交
1879

N
Nick Piggin 已提交
1880
static size_t __iovec_copy_from_user_inatomic(char *vaddr,
L
Linus Torvalds 已提交
1881 1882
			const struct iovec *iov, size_t base, size_t bytes)
{
1883
	size_t copied = 0, left = 0;
L
Linus Torvalds 已提交
1884 1885 1886 1887 1888 1889

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

		base = 0;
1890
		left = __copy_from_user_inatomic(vaddr, buf, copy);
L
Linus Torvalds 已提交
1891 1892 1893 1894 1895
		copied += copy;
		bytes -= copy;
		vaddr += copy;
		iov++;

1896
		if (unlikely(left))
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901
			break;
	}
	return copied - left;
}

N
Nick Piggin 已提交
1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917
/*
 * Copy as much as we can into the page and return the number of bytes which
 * were sucessfully copied.  If a fault is encountered then return the number of
 * 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;
1918
		left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
1919 1920 1921 1922 1923 1924 1925 1926 1927
		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
1928
EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
N
Nick Piggin 已提交
1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945

/*
 * 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;
1946
		left = __copy_from_user(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
1947 1948 1949 1950 1951 1952 1953 1954
		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
1955
EXPORT_SYMBOL(iov_iter_copy_from_user);
N
Nick Piggin 已提交
1956

N
Nick Piggin 已提交
1957
void iov_iter_advance(struct iov_iter *i, size_t bytes)
N
Nick Piggin 已提交
1958
{
N
Nick Piggin 已提交
1959 1960
	BUG_ON(i->count < bytes);

N
Nick Piggin 已提交
1961 1962
	if (likely(i->nr_segs == 1)) {
		i->iov_offset += bytes;
N
Nick Piggin 已提交
1963
		i->count -= bytes;
N
Nick Piggin 已提交
1964 1965 1966 1967
	} else {
		const struct iovec *iov = i->iov;
		size_t base = i->iov_offset;

1968 1969
		/*
		 * The !iov->iov_len check ensures we skip over unlikely
N
Nick Piggin 已提交
1970
		 * zero-length segments (without overruning the iovec).
1971
		 */
1972
		while (bytes || unlikely(i->count && !iov->iov_len)) {
N
Nick Piggin 已提交
1973
			int copy;
N
Nick Piggin 已提交
1974

N
Nick Piggin 已提交
1975 1976 1977
			copy = min(bytes, iov->iov_len - base);
			BUG_ON(!i->count || i->count < copy);
			i->count -= copy;
N
Nick Piggin 已提交
1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
			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
1989
EXPORT_SYMBOL(iov_iter_advance);
N
Nick Piggin 已提交
1990

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
/*
 * 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
2001 2002
{
	char __user *buf = i->iov->iov_base + i->iov_offset;
2003 2004
	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	return fault_in_pages_readable(buf, bytes);
N
Nick Piggin 已提交
2005
}
N
Nick Piggin 已提交
2006
EXPORT_SYMBOL(iov_iter_fault_in_readable);
N
Nick Piggin 已提交
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

/*
 * 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
2019
EXPORT_SYMBOL(iov_iter_single_seg_count);
N
Nick Piggin 已提交
2020

L
Linus Torvalds 已提交
2021 2022 2023
/*
 * Performs necessary checks before doing a write
 *
2024
 * Can adjust writing position or amount of bytes to write.
L
Linus Torvalds 已提交
2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082
 * 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;
	unsigned long limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;

        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 {
2083
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094
		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;
2095 2096 2097
#else
		return -EPERM;
#endif
L
Linus Torvalds 已提交
2098 2099 2100 2101 2102
	}
	return 0;
}
EXPORT_SYMBOL(generic_write_checks);

2103 2104 2105 2106 2107 2108
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;

2109
	return aops->write_begin(file, mapping, pos, len, flags,
2110 2111 2112 2113 2114 2115 2116 2117 2118 2119
							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;

2120 2121
	mark_page_accessed(page);
	return aops->write_end(file, mapping, pos, len, copied, page, fsdata);
2122 2123 2124
}
EXPORT_SYMBOL(pagecache_write_end);

L
Linus Torvalds 已提交
2125 2126 2127 2128 2129 2130 2131 2132 2133
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;
2134 2135
	size_t		write_len;
	pgoff_t		end;
L
Linus Torvalds 已提交
2136 2137 2138 2139

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

2140 2141 2142
	write_len = iov_length(iov, *nr_segs);
	end = (pos + write_len - 1) >> PAGE_CACHE_SHIFT;

2143
	written = filemap_write_and_wait_range(mapping, pos, pos + write_len - 1);
2144 2145 2146 2147 2148 2149 2150
	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
2151
	 * without clobbering -EIOCBQUEUED from ->direct_IO().
2152 2153 2154 2155
	 */
	if (mapping->nrpages) {
		written = invalidate_inode_pages2_range(mapping,
					pos >> PAGE_CACHE_SHIFT, end);
2156 2157 2158 2159 2160 2161 2162
		/*
		 * If a page can not be invalidated, return 0 to fall back
		 * to buffered write.
		 */
		if (written) {
			if (written == -EBUSY)
				return 0;
2163
			goto out;
2164
		}
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
	}

	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
2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193
	if (written > 0) {
		loff_t end = pos + written;
		if (end > i_size_read(inode) && !S_ISBLK(inode->i_mode)) {
			i_size_write(inode,  end);
			mark_inode_dirty(inode);
		}
		*ppos = end;
	}

	/*
	 * Sync the fs metadata but not the minor inode changes and
	 * of course not the data as we did direct DMA for the IO.
2194
	 * i_mutex is held, which protects generic_osync_inode() from
2195
	 * livelocking.  AIO O_DIRECT ops attempt to sync metadata here.
L
Linus Torvalds 已提交
2196
	 */
2197
out:
2198 2199
	if ((written >= 0 || written == -EIOCBQUEUED) &&
	    ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
2200 2201 2202 2203
		int err = generic_osync_inode(inode, mapping, OSYNC_METADATA);
		if (err < 0)
			written = err;
	}
L
Linus Torvalds 已提交
2204 2205 2206 2207
	return written;
}
EXPORT_SYMBOL(generic_file_direct_write);

N
Nick Piggin 已提交
2208 2209 2210 2211
/*
 * Find or create a page at the given pagecache position. Return the locked
 * page. This function is specifically for buffered writes.
 */
2212 2213
struct page *grab_cache_page_write_begin(struct address_space *mapping,
					pgoff_t index, unsigned flags)
N
Nick Piggin 已提交
2214 2215 2216
{
	int status;
	struct page *page;
2217 2218 2219
	gfp_t gfp_notmask = 0;
	if (flags & AOP_FLAG_NOFS)
		gfp_notmask = __GFP_FS;
N
Nick Piggin 已提交
2220 2221 2222 2223 2224
repeat:
	page = find_lock_page(mapping, index);
	if (likely(page))
		return page;

2225
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~gfp_notmask);
N
Nick Piggin 已提交
2226 2227
	if (!page)
		return NULL;
2228 2229
	status = add_to_page_cache_lru(page, mapping, index,
						GFP_KERNEL & ~gfp_notmask);
N
Nick Piggin 已提交
2230 2231 2232 2233 2234 2235 2236 2237
	if (unlikely(status)) {
		page_cache_release(page);
		if (status == -EEXIST)
			goto repeat;
		return NULL;
	}
	return page;
}
2238
EXPORT_SYMBOL(grab_cache_page_write_begin);
N
Nick Piggin 已提交
2239

2240 2241 2242 2243 2244 2245 2246
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
2247 2248 2249 2250 2251 2252 2253
	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;
2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284

	do {
		struct page *page;
		pgoff_t index;		/* Pagecache index for current 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));
		index = pos >> PAGE_CACHE_SHIFT;
		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
2285
		status = a_ops->write_begin(file, mapping, pos, bytes, flags,
2286 2287 2288 2289 2290 2291 2292 2293 2294
						&page, &fsdata);
		if (unlikely(status))
			break;

		pagefault_disable();
		copied = iov_iter_copy_from_user_atomic(page, i, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

2295
		mark_page_accessed(page);
2296 2297 2298 2299 2300 2301 2302 2303
		status = a_ops->write_end(file, mapping, pos, bytes, copied,
						page, fsdata);
		if (unlikely(status < 0))
			break;
		copied = status;

		cond_resched();

2304
		iov_iter_advance(i, copied);
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 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
		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;
	struct address_space *mapping = file->f_mapping;
	const struct address_space_operations *a_ops = mapping->a_ops;
	struct inode *inode = mapping->host;
	ssize_t status;
	struct iov_iter i;

	iov_iter_init(&i, iov, nr_segs, count, written);
2341
	status = generic_perform_write(file, &i, pos);
L
Linus Torvalds 已提交
2342 2343

	if (likely(status >= 0)) {
2344 2345 2346 2347 2348 2349 2350
		written += status;
		*ppos = pos + status;

		/*
		 * For now, when the user asks for O_SYNC, we'll actually give
		 * O_DSYNC
		 */
L
Linus Torvalds 已提交
2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363
		if (unlikely((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
			if (!a_ops->writepage || !is_sync_kiocb(iocb))
				status = generic_osync_inode(inode, mapping,
						OSYNC_METADATA|OSYNC_DATA);
		}
  	}
	
	/*
	 * If we get here for O_DIRECT writes then we must have fallen through
	 * to buffered writes (block instantiation inside i_size).  So we sync
	 * the file data here, to try to honour O_DIRECT expectations.
	 */
	if (unlikely(file->f_flags & O_DIRECT) && written)
2364 2365
		status = filemap_write_and_wait_range(mapping,
					pos, pos + written - 1);
L
Linus Torvalds 已提交
2366 2367 2368 2369 2370

	return written ? written : status;
}
EXPORT_SYMBOL(generic_file_buffered_write);

2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391
/**
 * __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
2392 2393
{
	struct file *file = iocb->ki_filp;
2394
	struct address_space * mapping = file->f_mapping;
L
Linus Torvalds 已提交
2395 2396 2397 2398 2399 2400 2401 2402
	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;
2403 2404 2405
	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
	if (err)
		return err;
L
Linus Torvalds 已提交
2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422

	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;

2423
	err = file_remove_suid(file);
L
Linus Torvalds 已提交
2424 2425 2426
	if (err)
		goto out;

2427
	file_update_time(file);
L
Linus Torvalds 已提交
2428 2429 2430

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (unlikely(file->f_flags & O_DIRECT)) {
2431 2432 2433 2434 2435
		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
2436 2437 2438 2439 2440 2441 2442 2443
		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;
2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457
		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
2458

2459 2460 2461 2462 2463 2464
		/*
		 * 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;
M
Mark Fasheh 已提交
2465 2466 2467 2468
		err = do_sync_mapping_range(file->f_mapping, pos, endbyte,
					    SYNC_FILE_RANGE_WAIT_BEFORE|
					    SYNC_FILE_RANGE_WRITE|
					    SYNC_FILE_RANGE_WAIT_AFTER);
2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483
		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
2484 2485 2486 2487
out:
	current->backing_dev_info = NULL;
	return written ? written : err;
}
2488 2489
EXPORT_SYMBOL(__generic_file_aio_write);

L
Linus Torvalds 已提交
2490

2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504
/**
 * generic_file_aio_write_nolock - write data, usually to a device
 * @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() which takes care of
 * syncing the file in case of O_SYNC file. It does not take i_mutex for the
 * write itself but may do so during syncing. It is meant for users like block
 * devices which do not need i_mutex during write. If your filesystem needs to
 * do a write but already holds i_mutex, use __generic_file_aio_write()
 * directly and then sync the file like generic_file_aio_write().
 */
2505 2506
ssize_t generic_file_aio_write_nolock(struct kiocb *iocb,
		const struct iovec *iov, unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
2507 2508 2509 2510 2511 2512
{
	struct file *file = iocb->ki_filp;
	struct address_space *mapping = file->f_mapping;
	struct inode *inode = mapping->host;
	ssize_t ret;

2513 2514
	BUG_ON(iocb->ki_pos != pos);

2515
	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
L
Linus Torvalds 已提交
2516 2517

	if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
2518
		ssize_t err;
L
Linus Torvalds 已提交
2519 2520 2521 2522 2523 2524 2525

		err = sync_page_range_nolock(inode, mapping, pos, ret);
		if (err < 0)
			ret = err;
	}
	return ret;
}
2526
EXPORT_SYMBOL(generic_file_aio_write_nolock);
L
Linus Torvalds 已提交
2527

2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538
/**
 * 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.
 */
2539 2540
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
2541 2542 2543 2544 2545 2546 2547 2548
{
	struct file *file = iocb->ki_filp;
	struct address_space *mapping = file->f_mapping;
	struct inode *inode = mapping->host;
	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 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563

	if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
		ssize_t err;

		err = sync_page_range(inode, mapping, pos, ret);
		if (err < 0)
			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);