filemap.c 69.4 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/mm_inline.h> /* for page_is_file_cache() */
37
#include <linux/cleancache.h>
38 39
#include "internal.h"

L
Linus Torvalds 已提交
40 41 42
/*
 * FIXME: remove all knowledge of the buffer layer from the core VM
 */
43
#include <linux/buffer_head.h> /* for try_to_free_buffers */
L
Linus Torvalds 已提交
44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61

#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:
 *
62
 *  ->i_mmap_mutex		(truncate_pagecache)
L
Linus Torvalds 已提交
63
 *    ->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
68
 *    ->i_mmap_mutex		(truncate->unmap_mapping_range)
L
Linus Torvalds 已提交
69 70
 *
 *  ->mmap_sem
71
 *    ->i_mmap_mutex
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
 *  bdi->wb.list_lock
82
 *    sb_lock			(fs/fs-writeback.c)
L
Linus Torvalds 已提交
83 84
 *    ->mapping->tree_lock	(__sync_single_inode)
 *
85
 *  ->i_mmap_mutex
L
Linus Torvalds 已提交
86 87 88
 *    ->anon_vma.lock		(vma_adjust)
 *
 *  ->anon_vma.lock
89
 *    ->page_table_lock or pte_lock	(anon_vma_prepare and various)
L
Linus Torvalds 已提交
90
 *
91
 *  ->page_table_lock or pte_lock
92
 *    ->swap_lock		(try_to_unmap_one)
L
Linus Torvalds 已提交
93 94 95
 *    ->private_lock		(try_to_unmap_one)
 *    ->tree_lock		(try_to_unmap_one)
 *    ->zone.lru_lock		(follow_page->mark_page_accessed)
96
 *    ->zone.lru_lock		(check_pte_range->isolate_lru_page)
L
Linus Torvalds 已提交
97 98
 *    ->private_lock		(page_remove_rmap->set_page_dirty)
 *    ->tree_lock		(page_remove_rmap->set_page_dirty)
99
 *    bdi.wb->list_lock		(page_remove_rmap->set_page_dirty)
100
 *    ->inode->i_lock		(page_remove_rmap->set_page_dirty)
101
 *    bdi.wb->list_lock		(zap_pte_range->set_page_dirty)
102
 *    ->inode->i_lock		(zap_pte_range->set_page_dirty)
L
Linus Torvalds 已提交
103 104
 *    ->private_lock		(zap_pte_range->__set_page_dirty_buffers)
 *
105 106
 *  (code doesn't rely on that order, so you could switch it around)
 *  ->tasklist_lock             (memory_failure, collect_procs_ao)
107
 *    ->i_mmap_mutex
L
Linus Torvalds 已提交
108 109 110
 */

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

119 120 121 122 123 124 125 126 127 128
	/*
	 * 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
129 130
	radix_tree_delete(&mapping->page_tree, page->index);
	page->mapping = NULL;
131
	/* Leave page->index set: truncation lookup relies upon it */
L
Linus Torvalds 已提交
132
	mapping->nrpages--;
133
	__dec_zone_page_state(page, NR_FILE_PAGES);
134 135
	if (PageSwapBacked(page))
		__dec_zone_page_state(page, NR_SHMEM);
136
	BUG_ON(page_mapped(page));
137 138 139 140 141 142 143 144 145 146 147 148

	/*
	 * 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
149 150
}

151 152 153 154 155 156 157 158 159
/**
 * 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
160 161
{
	struct address_space *mapping = page->mapping;
162
	void (*freepage)(struct page *);
L
Linus Torvalds 已提交
163

M
Matt Mackall 已提交
164
	BUG_ON(!PageLocked(page));
L
Linus Torvalds 已提交
165

166
	freepage = mapping->a_ops->freepage;
N
Nick Piggin 已提交
167
	spin_lock_irq(&mapping->tree_lock);
168
	__delete_from_page_cache(page);
N
Nick Piggin 已提交
169
	spin_unlock_irq(&mapping->tree_lock);
170
	mem_cgroup_uncharge_cache_page(page);
171 172 173

	if (freepage)
		freepage(page);
174 175 176 177
	page_cache_release(page);
}
EXPORT_SYMBOL(delete_from_page_cache);

J
Jens Axboe 已提交
178
static int sleep_on_page(void *word)
L
Linus Torvalds 已提交
179 180 181 182 183
{
	io_schedule();
	return 0;
}

J
Jens Axboe 已提交
184
static int sleep_on_page_killable(void *word)
M
Matthew Wilcox 已提交
185
{
J
Jens Axboe 已提交
186
	sleep_on_page(word);
M
Matthew Wilcox 已提交
187 188 189
	return fatal_signal_pending(current) ? -EINTR : 0;
}

L
Linus Torvalds 已提交
190
/**
191
 * __filemap_fdatawrite_range - start writeback on mapping dirty pages in range
192 193
 * @mapping:	address space structure to write
 * @start:	offset in bytes where the range starts
194
 * @end:	offset in bytes where the range ends (inclusive)
195
 * @sync_mode:	enable synchronous operation
L
Linus Torvalds 已提交
196
 *
197 198 199
 * 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
200
 * If sync_mode is WB_SYNC_ALL then this is a "data integrity" operation, as
201
 * opposed to a regular memory cleansing writeback.  The difference between
L
Linus Torvalds 已提交
202 203 204
 * these two operations is that if a dirty page/buffer is encountered, it must
 * be waited upon, and not just skipped over.
 */
205 206
int __filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
				loff_t end, int sync_mode)
L
Linus Torvalds 已提交
207 208 209 210
{
	int ret;
	struct writeback_control wbc = {
		.sync_mode = sync_mode,
211
		.nr_to_write = LONG_MAX,
212 213
		.range_start = start,
		.range_end = end,
L
Linus Torvalds 已提交
214 215 216 217 218 219 220 221 222 223 224 225
	};

	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)
{
226
	return __filemap_fdatawrite_range(mapping, 0, LLONG_MAX, sync_mode);
L
Linus Torvalds 已提交
227 228 229 230 231 232 233 234
}

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

235
int filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
236
				loff_t end)
L
Linus Torvalds 已提交
237 238 239
{
	return __filemap_fdatawrite_range(mapping, start, end, WB_SYNC_ALL);
}
240
EXPORT_SYMBOL(filemap_fdatawrite_range);
L
Linus Torvalds 已提交
241

242 243 244 245
/**
 * filemap_flush - mostly a non-blocking flush
 * @mapping:	target address_space
 *
L
Linus Torvalds 已提交
246 247 248 249 250 251 252 253 254
 * 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);

255
/**
256 257 258 259
 * 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)
260
 *
261 262
 * 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
263
 */
264 265
int filemap_fdatawait_range(struct address_space *mapping, loff_t start_byte,
			    loff_t end_byte)
L
Linus Torvalds 已提交
266
{
267 268
	pgoff_t index = start_byte >> PAGE_CACHE_SHIFT;
	pgoff_t end = end_byte >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
269 270 271 272
	struct pagevec pvec;
	int nr_pages;
	int ret = 0;

273
	if (end_byte < start_byte)
L
Linus Torvalds 已提交
274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
		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);
291
			if (TestClearPageError(page))
L
Linus Torvalds 已提交
292 293 294 295 296 297 298 299 300 301 302 303 304 305
				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;
}
306 307
EXPORT_SYMBOL(filemap_fdatawait_range);

L
Linus Torvalds 已提交
308
/**
309
 * filemap_fdatawait - wait for all under-writeback pages to complete
L
Linus Torvalds 已提交
310
 * @mapping: address space structure to wait for
311 312 313
 *
 * Walk the list of under-writeback pages of the given address space
 * and wait for all of them.
L
Linus Torvalds 已提交
314 315 316 317 318 319 320 321
 */
int filemap_fdatawait(struct address_space *mapping)
{
	loff_t i_size = i_size_read(mapping->host);

	if (i_size == 0)
		return 0;

322
	return filemap_fdatawait_range(mapping, 0, i_size - 1);
L
Linus Torvalds 已提交
323 324 325 326 327
}
EXPORT_SYMBOL(filemap_fdatawait);

int filemap_write_and_wait(struct address_space *mapping)
{
328
	int err = 0;
L
Linus Torvalds 已提交
329 330

	if (mapping->nrpages) {
331 332 333 334 335 336 337 338 339 340 341 342
		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
343
	}
344
	return err;
L
Linus Torvalds 已提交
345
}
346
EXPORT_SYMBOL(filemap_write_and_wait);
L
Linus Torvalds 已提交
347

348 349 350 351 352 353
/**
 * 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)
 *
354 355 356 357 358
 * 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
359 360 361
int filemap_write_and_wait_range(struct address_space *mapping,
				 loff_t lstart, loff_t lend)
{
362
	int err = 0;
L
Linus Torvalds 已提交
363 364

	if (mapping->nrpages) {
365 366 367 368
		err = __filemap_fdatawrite_range(mapping, lstart, lend,
						 WB_SYNC_ALL);
		/* See comment of filemap_write_and_wait() */
		if (err != -EIO) {
369 370
			int err2 = filemap_fdatawait_range(mapping,
						lstart, lend);
371 372 373
			if (!err)
				err = err2;
		}
L
Linus Torvalds 已提交
374
	}
375
	return err;
L
Linus Torvalds 已提交
376
}
377
EXPORT_SYMBOL(filemap_write_and_wait_range);
L
Linus Torvalds 已提交
378

379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
/**
 * 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);
428
		__delete_from_page_cache(old);
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
		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);

449
/**
N
Nick Piggin 已提交
450
 * add_to_page_cache_locked - add a locked page to the pagecache
451 452 453 454 455
 * @page:	page to add
 * @mapping:	the page's address_space
 * @offset:	page index
 * @gfp_mask:	page allocation mode
 *
N
Nick Piggin 已提交
456
 * This function is used to add a page to the pagecache. It must be locked.
L
Linus Torvalds 已提交
457 458
 * This function does not add the page to the LRU.  The caller must do that.
 */
N
Nick Piggin 已提交
459
int add_to_page_cache_locked(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
460
		pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
461
{
N
Nick Piggin 已提交
462 463 464 465 466
	int error;

	VM_BUG_ON(!PageLocked(page));

	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);
482 483
			if (PageSwapBacked(page))
				__inc_zone_page_state(page, NR_SHMEM);
484
			spin_unlock_irq(&mapping->tree_lock);
N
Nick Piggin 已提交
485 486
		} else {
			page->mapping = NULL;
487
			/* Leave page->index set: truncation relies upon it */
488
			spin_unlock_irq(&mapping->tree_lock);
489
			mem_cgroup_uncharge_cache_page(page);
N
Nick Piggin 已提交
490 491
			page_cache_release(page);
		}
L
Linus Torvalds 已提交
492
		radix_tree_preload_end();
B
Balbir Singh 已提交
493
	} else
494
		mem_cgroup_uncharge_cache_page(page);
495
out:
L
Linus Torvalds 已提交
496 497
	return error;
}
N
Nick Piggin 已提交
498
EXPORT_SYMBOL(add_to_page_cache_locked);
L
Linus Torvalds 已提交
499 500

int add_to_page_cache_lru(struct page *page, struct address_space *mapping,
A
Al Viro 已提交
501
				pgoff_t offset, gfp_t gfp_mask)
L
Linus Torvalds 已提交
502
{
503 504 505 506 507
	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
508
	 * need to go on the anon lru below, and mem_cgroup_cache_charge
509 510 511 512 513 514 515 516 517 518
	 * (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
519
			lru_cache_add_anon(page);
520
	}
L
Linus Torvalds 已提交
521 522
	return ret;
}
523
EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
L
Linus Torvalds 已提交
524

525
#ifdef CONFIG_NUMA
526
struct page *__page_cache_alloc(gfp_t gfp)
527
{
528 529 530
	int n;
	struct page *page;

531
	if (cpuset_do_page_mem_spread()) {
532 533 534 535 536
		get_mems_allowed();
		n = cpuset_mem_spread_node();
		page = alloc_pages_exact_node(n, gfp, 0);
		put_mems_allowed();
		return page;
537
	}
538
	return alloc_pages(gfp, 0);
539
}
540
EXPORT_SYMBOL(__page_cache_alloc);
541 542
#endif

L
Linus Torvalds 已提交
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564
/*
 * 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
565
void wait_on_page_bit(struct page *page, int bit_nr)
L
Linus Torvalds 已提交
566 567 568 569
{
	DEFINE_WAIT_BIT(wait, &page->flags, bit_nr);

	if (test_bit(bit_nr, &page->flags))
J
Jens Axboe 已提交
570
		__wait_on_bit(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
571 572 573 574
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(wait_on_page_bit);

575 576 577 578 579 580 581 582 583 584 585
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);
}

586 587
/**
 * add_page_wait_queue - Add an arbitrary waiter to a page's wait queue
R
Randy Dunlap 已提交
588 589
 * @page: Page defining the wait queue of interest
 * @waiter: Waiter to add to the queue
590 591 592 593 594 595 596 597 598 599 600 601 602 603
 *
 * 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
604
/**
605
 * unlock_page - unlock a locked page
L
Linus Torvalds 已提交
606 607 608 609 610 611 612
 * @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
613 614
 * 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
615
 */
H
Harvey Harrison 已提交
616
void unlock_page(struct page *page)
L
Linus Torvalds 已提交
617
{
N
Nick Piggin 已提交
618 619 620
	VM_BUG_ON(!PageLocked(page));
	clear_bit_unlock(PG_locked, &page->flags);
	smp_mb__after_clear_bit();
L
Linus Torvalds 已提交
621 622 623 624
	wake_up_page(page, PG_locked);
}
EXPORT_SYMBOL(unlock_page);

625 626 627
/**
 * end_page_writeback - end writeback against a page
 * @page: the page
L
Linus Torvalds 已提交
628 629 630
 */
void end_page_writeback(struct page *page)
{
631 632 633 634 635 636
	if (TestClearPageReclaim(page))
		rotate_reclaimable_page(page);

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

L
Linus Torvalds 已提交
637 638 639 640 641
	smp_mb__after_clear_bit();
	wake_up_page(page, PG_writeback);
}
EXPORT_SYMBOL(end_page_writeback);

642 643 644
/**
 * __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
645
 */
H
Harvey Harrison 已提交
646
void __lock_page(struct page *page)
L
Linus Torvalds 已提交
647 648 649
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

J
Jens Axboe 已提交
650
	__wait_on_bit_lock(page_waitqueue(page), &wait, sleep_on_page,
L
Linus Torvalds 已提交
651 652 653 654
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(__lock_page);

H
Harvey Harrison 已提交
655
int __lock_page_killable(struct page *page)
M
Matthew Wilcox 已提交
656 657 658 659
{
	DEFINE_WAIT_BIT(wait, &page->flags, PG_locked);

	return __wait_on_bit_lock(page_waitqueue(page), &wait,
J
Jens Axboe 已提交
660
					sleep_on_page_killable, TASK_KILLABLE);
M
Matthew Wilcox 已提交
661
}
662
EXPORT_SYMBOL_GPL(__lock_page_killable);
M
Matthew Wilcox 已提交
663

664 665 666
int __lock_page_or_retry(struct page *page, struct mm_struct *mm,
			 unsigned int flags)
{
667 668 669 670 671 672 673 674 675 676 677 678
	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
679
			wait_on_page_locked(page);
680
		return 0;
681 682 683 684 685 686 687 688 689 690 691 692
	} 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;
693 694 695
	}
}

696 697 698 699 700
/**
 * 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
701 702
 * 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
703
 */
N
Nick Piggin 已提交
704
struct page *find_get_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
705
{
N
Nick Piggin 已提交
706
	void **pagep;
L
Linus Torvalds 已提交
707 708
	struct page *page;

N
Nick Piggin 已提交
709 710 711 712 713 714
	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
715 716 717
		if (unlikely(!page))
			goto out;
		if (radix_tree_deref_retry(page))
N
Nick Piggin 已提交
718 719 720 721 722 723 724 725 726 727 728 729 730 731 732
			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;
		}
	}
N
Nick Piggin 已提交
733
out:
N
Nick Piggin 已提交
734 735
	rcu_read_unlock();

L
Linus Torvalds 已提交
736 737 738 739 740 741
	return page;
}
EXPORT_SYMBOL(find_get_page);

/**
 * find_lock_page - locate, pin and lock a pagecache page
742 743
 * @mapping: the address_space to search
 * @offset: the page index
L
Linus Torvalds 已提交
744 745 746 747 748 749
 *
 * 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
750
struct page *find_lock_page(struct address_space *mapping, pgoff_t offset)
L
Linus Torvalds 已提交
751 752 753 754
{
	struct page *page;

repeat:
N
Nick Piggin 已提交
755
	page = find_get_page(mapping, offset);
L
Linus Torvalds 已提交
756
	if (page) {
N
Nick Piggin 已提交
757 758 759 760 761 762
		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
763
		}
N
Nick Piggin 已提交
764
		VM_BUG_ON(page->index != offset);
L
Linus Torvalds 已提交
765 766 767 768 769 770 771
	}
	return page;
}
EXPORT_SYMBOL(find_lock_page);

/**
 * find_or_create_page - locate or add a pagecache page
772 773 774
 * @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
775 776 777 778 779 780 781 782 783 784 785 786 787
 *
 * 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,
788
		pgoff_t index, gfp_t gfp_mask)
L
Linus Torvalds 已提交
789
{
N
Nick Piggin 已提交
790
	struct page *page;
L
Linus Torvalds 已提交
791 792 793 794
	int err;
repeat:
	page = find_lock_page(mapping, index);
	if (!page) {
N
Nick Piggin 已提交
795 796 797
		page = __page_cache_alloc(gfp_mask);
		if (!page)
			return NULL;
N
Nick Piggin 已提交
798 799 800 801 802 803 804 805
		/*
		 * 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
806 807 808 809 810
		if (unlikely(err)) {
			page_cache_release(page);
			page = NULL;
			if (err == -EEXIST)
				goto repeat;
L
Linus Torvalds 已提交
811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837
		}
	}
	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
838 839 840 841 842 843 844 845 846 847 848 849 850
	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;
851 852 853 854 855

		/*
		 * This can only trigger when the entry at index 0 moves out
		 * of or back to the root: none yet gotten, safe to restart.
		 */
N
Nick Piggin 已提交
856
		if (radix_tree_deref_retry(page)) {
857
			WARN_ON(start | i);
N
Nick Piggin 已提交
858
			goto restart;
N
Nick Piggin 已提交
859
		}
N
Nick Piggin 已提交
860 861 862 863 864 865 866 867 868

		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
869

N
Nick Piggin 已提交
870 871 872
		pages[ret] = page;
		ret++;
	}
873 874 875 876 877 878 879

	/*
	 * 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
880
	rcu_read_unlock();
L
Linus Torvalds 已提交
881 882 883
	return ret;
}

884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
/**
 * 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
901 902 903 904 905 906 907 908 909 910 911 912 913
	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;
914 915 916 917 918

		/*
		 * This can only trigger when the entry at index 0 moves out
		 * of or back to the root: none yet gotten, safe to restart.
		 */
N
Nick Piggin 已提交
919
		if (radix_tree_deref_retry(page))
N
Nick Piggin 已提交
920
			goto restart;
921

N
Nick Piggin 已提交
922 923 924 925 926 927 928 929 930
		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
931 932 933 934 935 936 937 938 939 940
		/*
		 * 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
941 942
		pages[ret] = page;
		ret++;
943 944
		index++;
	}
N
Nick Piggin 已提交
945 946
	rcu_read_unlock();
	return ret;
947
}
948
EXPORT_SYMBOL(find_get_pages_contig);
949

950 951 952 953 954 955 956 957
/**
 * 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
958
 * Like find_get_pages, except we only return pages which are tagged with
959
 * @tag.   We update @index to index the next page for the traversal.
L
Linus Torvalds 已提交
960 961 962 963 964 965
 */
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
966 967 968 969 970 971 972 973 974 975 976 977 978
	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;
979 980 981 982 983

		/*
		 * This can only trigger when the entry at index 0 moves out
		 * of or back to the root: none yet gotten, safe to restart.
		 */
N
Nick Piggin 已提交
984
		if (radix_tree_deref_retry(page))
N
Nick Piggin 已提交
985 986 987 988 989 990 991 992 993 994 995 996 997 998
			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++;
	}
999 1000 1001 1002 1003 1004 1005

	/*
	 * 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
1006
	rcu_read_unlock();
L
Linus Torvalds 已提交
1007 1008 1009

	if (ret)
		*index = pages[ret - 1]->index + 1;
N
Nick Piggin 已提交
1010

L
Linus Torvalds 已提交
1011 1012
	return ret;
}
1013
EXPORT_SYMBOL(find_get_pages_tag);
L
Linus Torvalds 已提交
1014

1015 1016 1017 1018 1019
/**
 * grab_cache_page_nowait - returns locked page at given index in given cache
 * @mapping: target address_space
 * @index: the page index
 *
1020
 * Same as grab_cache_page(), but do not wait if the page is unavailable.
L
Linus Torvalds 已提交
1021 1022 1023 1024 1025 1026 1027 1028
 * 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 *
1029
grab_cache_page_nowait(struct address_space *mapping, pgoff_t index)
L
Linus Torvalds 已提交
1030 1031 1032 1033
{
	struct page *page = find_get_page(mapping, index);

	if (page) {
N
Nick Piggin 已提交
1034
		if (trylock_page(page))
L
Linus Torvalds 已提交
1035 1036 1037 1038
			return page;
		page_cache_release(page);
		return NULL;
	}
1039
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~__GFP_FS);
N
Nick Piggin 已提交
1040
	if (page && add_to_page_cache_lru(page, mapping, index, GFP_NOFS)) {
L
Linus Torvalds 已提交
1041 1042 1043 1044 1045 1046 1047
		page_cache_release(page);
		page = NULL;
	}
	return page;
}
EXPORT_SYMBOL(grab_cache_page_nowait);

1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
/*
 * 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;
}

1069
/**
C
Christoph Hellwig 已提交
1070
 * do_generic_file_read - generic file read routine
1071 1072 1073 1074 1075
 * @filp:	the file to read
 * @ppos:	current file position
 * @desc:	read_descriptor
 * @actor:	read method
 *
L
Linus Torvalds 已提交
1076
 * This is a generic file read routine, and uses the
1077
 * mapping->a_ops->readpage() function for the actual low-level stuff.
L
Linus Torvalds 已提交
1078 1079 1080 1081
 *
 * 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
1082 1083
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
1084
{
C
Christoph Hellwig 已提交
1085
	struct address_space *mapping = filp->f_mapping;
L
Linus Torvalds 已提交
1086
	struct inode *inode = mapping->host;
C
Christoph Hellwig 已提交
1087
	struct file_ra_state *ra = &filp->f_ra;
1088 1089 1090 1091
	pgoff_t index;
	pgoff_t last_index;
	pgoff_t prev_index;
	unsigned long offset;      /* offset into pagecache page */
1092
	unsigned int prev_offset;
L
Linus Torvalds 已提交
1093 1094 1095
	int error;

	index = *ppos >> PAGE_CACHE_SHIFT;
1096 1097
	prev_index = ra->prev_pos >> PAGE_CACHE_SHIFT;
	prev_offset = ra->prev_pos & (PAGE_CACHE_SIZE-1);
L
Linus Torvalds 已提交
1098 1099 1100 1101 1102
	last_index = (*ppos + desc->count + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	for (;;) {
		struct page *page;
1103
		pgoff_t end_index;
N
NeilBrown 已提交
1104
		loff_t isize;
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109
		unsigned long nr, ret;

		cond_resched();
find_page:
		page = find_get_page(mapping, index);
1110
		if (!page) {
1111
			page_cache_sync_readahead(mapping,
1112
					ra, filp,
1113 1114 1115 1116 1117 1118
					index, last_index - index);
			page = find_get_page(mapping, index);
			if (unlikely(page == NULL))
				goto no_cached_page;
		}
		if (PageReadahead(page)) {
1119
			page_cache_async_readahead(mapping,
1120
					ra, filp, page,
1121
					index, last_index - index);
L
Linus Torvalds 已提交
1122
		}
1123 1124 1125 1126
		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
1127
			if (!trylock_page(page))
1128
				goto page_not_up_to_date;
1129 1130 1131
			/* Did it get truncated before we got the lock? */
			if (!page->mapping)
				goto page_not_up_to_date_locked;
1132 1133 1134 1135 1136
			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
1137
page_ok:
N
NeilBrown 已提交
1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
		/*
		 * 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
1164 1165 1166 1167 1168 1169 1170 1171 1172

		/* 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);

		/*
1173 1174
		 * When a sequential read accesses a page several times,
		 * only mark it as accessed the first time.
L
Linus Torvalds 已提交
1175
		 */
1176
		if (prev_index != index || offset != prev_offset)
L
Linus Torvalds 已提交
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
			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
1194
		prev_offset = offset;
L
Linus Torvalds 已提交
1195 1196 1197 1198 1199 1200 1201 1202

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

page_not_up_to_date:
		/* Get exclusive access to the page ... */
1203 1204 1205
		error = lock_page_killable(page);
		if (unlikely(error))
			goto readpage_error;
L
Linus Torvalds 已提交
1206

1207
page_not_up_to_date_locked:
N
Nick Piggin 已提交
1208
		/* Did it get truncated before we got the lock? */
L
Linus Torvalds 已提交
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221
		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:
1222 1223 1224 1225 1226 1227
		/*
		 * 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
1228 1229 1230
		/* Start the actual read. The read will unlock the page. */
		error = mapping->a_ops->readpage(filp, page);

1231 1232 1233 1234 1235
		if (unlikely(error)) {
			if (error == AOP_TRUNCATED_PAGE) {
				page_cache_release(page);
				goto find_page;
			}
L
Linus Torvalds 已提交
1236
			goto readpage_error;
1237
		}
L
Linus Torvalds 已提交
1238 1239

		if (!PageUptodate(page)) {
1240 1241 1242
			error = lock_page_killable(page);
			if (unlikely(error))
				goto readpage_error;
L
Linus Torvalds 已提交
1243 1244 1245
			if (!PageUptodate(page)) {
				if (page->mapping == NULL) {
					/*
1246
					 * invalidate_mapping_pages got it
L
Linus Torvalds 已提交
1247 1248 1249 1250 1251 1252
					 */
					unlock_page(page);
					page_cache_release(page);
					goto find_page;
				}
				unlock_page(page);
1253
				shrink_readahead_size_eio(filp, ra);
1254 1255
				error = -EIO;
				goto readpage_error;
L
Linus Torvalds 已提交
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
			}
			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
1273 1274 1275 1276
		page = page_cache_alloc_cold(mapping);
		if (!page) {
			desc->error = -ENOMEM;
			goto out;
L
Linus Torvalds 已提交
1277
		}
N
Nick Piggin 已提交
1278
		error = add_to_page_cache_lru(page, mapping,
L
Linus Torvalds 已提交
1279 1280
						index, GFP_KERNEL);
		if (error) {
N
Nick Piggin 已提交
1281
			page_cache_release(page);
L
Linus Torvalds 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290
			if (error == -EEXIST)
				goto find_page;
			desc->error = error;
			goto out;
		}
		goto readpage;
	}

out:
1291 1292 1293
	ra->prev_pos = prev_index;
	ra->prev_pos <<= PAGE_CACHE_SHIFT;
	ra->prev_pos |= prev_offset;
L
Linus Torvalds 已提交
1294

1295
	*ppos = ((loff_t)index << PAGE_CACHE_SHIFT) + offset;
1296
	file_accessed(filp);
L
Linus Torvalds 已提交
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 1328 1329 1330 1331 1332 1333 1334 1335 1336
}

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;
}

1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
/*
 * 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);

1376
/**
H
Henrik Kretzschmar 已提交
1377
 * generic_file_aio_read - generic filesystem read routine
1378 1379 1380
 * @iocb:	kernel I/O control block
 * @iov:	io vector request
 * @nr_segs:	number of segments in the iovec
H
Henrik Kretzschmar 已提交
1381
 * @pos:	current file position
1382
 *
L
Linus Torvalds 已提交
1383 1384 1385 1386
 * This is the "read()" routine for all filesystems
 * that can use the page cache directly.
 */
ssize_t
1387 1388
generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
		unsigned long nr_segs, loff_t pos)
L
Linus Torvalds 已提交
1389 1390 1391
{
	struct file *filp = iocb->ki_filp;
	ssize_t retval;
1392
	unsigned long seg = 0;
L
Linus Torvalds 已提交
1393
	size_t count;
1394
	loff_t *ppos = &iocb->ki_pos;
1395
	struct blk_plug plug;
L
Linus Torvalds 已提交
1396 1397

	count = 0;
1398 1399 1400
	retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
	if (retval)
		return retval;
L
Linus Torvalds 已提交
1401

1402 1403
	blk_start_plug(&plug);

L
Linus Torvalds 已提交
1404 1405
	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (filp->f_flags & O_DIRECT) {
1406
		loff_t size;
L
Linus Torvalds 已提交
1407 1408 1409 1410 1411 1412 1413 1414 1415
		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) {
1416 1417
			retval = filemap_write_and_wait_range(mapping, pos,
					pos + iov_length(iov, nr_segs) - 1);
1418 1419 1420 1421
			if (!retval) {
				retval = mapping->a_ops->direct_IO(READ, iocb,
							iov, pos, nr_segs);
			}
1422
			if (retval > 0) {
L
Linus Torvalds 已提交
1423
				*ppos = pos + retval;
1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
				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
1436 1437 1438
				file_accessed(filp);
				goto out;
			}
1439
		}
L
Linus Torvalds 已提交
1440 1441
	}

1442
	count = retval;
H
Hugh Dickins 已提交
1443 1444
	for (seg = 0; seg < nr_segs; seg++) {
		read_descriptor_t desc;
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
		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
1459

H
Hugh Dickins 已提交
1460
		desc.written = 0;
1461 1462
		desc.arg.buf = iov[seg].iov_base + offset;
		desc.count = iov[seg].iov_len - offset;
H
Hugh Dickins 已提交
1463 1464 1465 1466 1467 1468 1469 1470
		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
1471
		}
H
Hugh Dickins 已提交
1472 1473
		if (desc.count > 0)
			break;
L
Linus Torvalds 已提交
1474 1475
	}
out:
1476
	blk_finish_plug(&plug);
L
Linus Torvalds 已提交
1477 1478 1479 1480 1481 1482
	return retval;
}
EXPORT_SYMBOL(generic_file_aio_read);

static ssize_t
do_readahead(struct address_space *mapping, struct file *filp,
1483
	     pgoff_t index, unsigned long nr)
L
Linus Torvalds 已提交
1484 1485 1486 1487
{
	if (!mapping || !mapping->a_ops || !mapping->a_ops->readpage)
		return -EINVAL;

1488
	force_page_cache_readahead(mapping, filp, index, nr);
L
Linus Torvalds 已提交
1489 1490 1491
	return 0;
}

1492
SYSCALL_DEFINE(readahead)(int fd, loff_t offset, size_t count)
L
Linus Torvalds 已提交
1493 1494 1495 1496 1497 1498 1499 1500 1501
{
	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;
1502 1503
			pgoff_t start = offset >> PAGE_CACHE_SHIFT;
			pgoff_t end = (offset + count - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510
			unsigned long len = end - start + 1;
			ret = do_readahead(mapping, file, start, len);
		}
		fput(file);
	}
	return ret;
}
1511 1512 1513 1514 1515 1516 1517
#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
1518 1519

#ifdef CONFIG_MMU
1520 1521 1522 1523 1524
/**
 * 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
1525 1526 1527
 * 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
1528
static int page_cache_read(struct file *file, pgoff_t offset)
L
Linus Torvalds 已提交
1529 1530 1531
{
	struct address_space *mapping = file->f_mapping;
	struct page *page; 
1532
	int ret;
L
Linus Torvalds 已提交
1533

1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
	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
1544 1545 1546

		page_cache_release(page);

1547 1548 1549
	} while (ret == AOP_TRUNCATED_PAGE);
		
	return ret;
L
Linus Torvalds 已提交
1550 1551 1552 1553
}

#define MMAP_LOTSAMISS  (100)

1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
/*
 * 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;
1569 1570
	if (!ra->ra_pages)
		return;
1571

1572
	if (VM_SequentialReadHint(vma)) {
1573 1574
		page_cache_sync_readahead(mapping, ra, file, offset,
					  ra->ra_pages);
1575 1576 1577
		return;
	}

1578 1579
	/* Avoid banging the cache line if not needed */
	if (ra->mmap_miss < MMAP_LOTSAMISS * 10)
1580 1581 1582 1583 1584 1585 1586 1587 1588
		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;

1589 1590 1591
	/*
	 * mmap read-around
	 */
1592
	ra_pages = max_sane_readahead(ra->ra_pages);
1593 1594
	ra->start = max_t(long, 0, offset - ra_pages / 2);
	ra->size = ra_pages;
1595
	ra->async_size = ra_pages / 4;
1596
	ra_submit(ra, mapping, file);
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616
}

/*
 * 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))
1617 1618
		page_cache_async_readahead(mapping, ra, file,
					   page, offset, ra->ra_pages);
1619 1620
}

1621
/**
1622
 * filemap_fault - read in file data for page fault handling
N
Nick Piggin 已提交
1623 1624
 * @vma:	vma in which the fault was taken
 * @vmf:	struct vm_fault containing details of the fault
1625
 *
1626
 * filemap_fault() is invoked via the vma operations vector for a
L
Linus Torvalds 已提交
1627 1628 1629 1630 1631 1632
 * 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
1633
int filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1634 1635
{
	int error;
1636
	struct file *file = vma->vm_file;
L
Linus Torvalds 已提交
1637 1638 1639
	struct address_space *mapping = file->f_mapping;
	struct file_ra_state *ra = &file->f_ra;
	struct inode *inode = mapping->host;
1640
	pgoff_t offset = vmf->pgoff;
L
Linus Torvalds 已提交
1641
	struct page *page;
J
Jan Kara 已提交
1642
	pgoff_t size;
N
Nick Piggin 已提交
1643
	int ret = 0;
L
Linus Torvalds 已提交
1644 1645

	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1646
	if (offset >= size)
1647
		return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1648 1649 1650 1651

	/*
	 * Do we have something in the page cache already?
	 */
1652 1653
	page = find_get_page(mapping, offset);
	if (likely(page)) {
L
Linus Torvalds 已提交
1654
		/*
1655 1656
		 * We found the page, so try async readahead before
		 * waiting for the lock.
L
Linus Torvalds 已提交
1657
		 */
1658 1659 1660 1661 1662
		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);
1663
		mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT);
1664 1665
		ret = VM_FAULT_MAJOR;
retry_find:
1666
		page = find_get_page(mapping, offset);
L
Linus Torvalds 已提交
1667 1668 1669 1670
		if (!page)
			goto no_cached_page;
	}

1671 1672
	if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags)) {
		page_cache_release(page);
1673
		return ret | VM_FAULT_RETRY;
1674
	}
1675 1676 1677 1678 1679 1680 1681 1682 1683

	/* 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
1684
	/*
1685 1686
	 * 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
1687
	 */
1688
	if (unlikely(!PageUptodate(page)))
L
Linus Torvalds 已提交
1689 1690
		goto page_not_uptodate;

1691 1692 1693 1694
	/*
	 * Found the page and have a reference on it.
	 * We must recheck i_size under page lock.
	 */
1695
	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
1696
	if (unlikely(offset >= size)) {
1697
		unlock_page(page);
1698
		page_cache_release(page);
1699
		return VM_FAULT_SIGBUS;
1700 1701
	}

N
Nick Piggin 已提交
1702
	vmf->page = page;
N
Nick Piggin 已提交
1703
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1704 1705 1706 1707 1708 1709

no_cached_page:
	/*
	 * We're only likely to ever get here if MADV_RANDOM is in
	 * effect.
	 */
1710
	error = page_cache_read(file, offset);
L
Linus Torvalds 已提交
1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725

	/*
	 * 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
1726 1727
		return VM_FAULT_OOM;
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732 1733 1734 1735 1736

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);
1737
	error = mapping->a_ops->readpage(file, page);
1738 1739 1740 1741 1742
	if (!error) {
		wait_on_page_locked(page);
		if (!PageUptodate(page))
			error = -EIO;
	}
1743 1744 1745
	page_cache_release(page);

	if (!error || error == AOP_TRUNCATED_PAGE)
1746
		goto retry_find;
L
Linus Torvalds 已提交
1747

1748
	/* Things didn't work out. Return zero to tell the mm layer so. */
1749
	shrink_readahead_size_eio(file, ra);
N
Nick Piggin 已提交
1750
	return VM_FAULT_SIGBUS;
1751 1752 1753
}
EXPORT_SYMBOL(filemap_fault);

1754
const struct vm_operations_struct generic_file_vm_ops = {
1755
	.fault		= filemap_fault,
L
Linus Torvalds 已提交
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767
};

/* 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
1768
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
	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);

1795
static struct page *__read_cache_page(struct address_space *mapping,
1796
				pgoff_t index,
1797
				int (*filler)(void *, struct page *),
1798 1799
				void *data,
				gfp_t gfp)
L
Linus Torvalds 已提交
1800
{
N
Nick Piggin 已提交
1801
	struct page *page;
L
Linus Torvalds 已提交
1802 1803 1804 1805
	int err;
repeat:
	page = find_get_page(mapping, index);
	if (!page) {
1806
		page = __page_cache_alloc(gfp | __GFP_COLD);
N
Nick Piggin 已提交
1807 1808 1809 1810 1811 1812 1813
		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
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
			/* 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;
}

1826
static struct page *do_read_cache_page(struct address_space *mapping,
1827
				pgoff_t index,
1828
				int (*filler)(void *, struct page *),
1829 1830 1831
				void *data,
				gfp_t gfp)

L
Linus Torvalds 已提交
1832 1833 1834 1835 1836
{
	struct page *page;
	int err;

retry:
1837
	page = __read_cache_page(mapping, index, filler, data, gfp);
L
Linus Torvalds 已提交
1838
	if (IS_ERR(page))
1839
		return page;
L
Linus Torvalds 已提交
1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855
	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);
1856
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1857
	}
1858
out:
1859 1860 1861
	mark_page_accessed(page);
	return page;
}
1862 1863 1864 1865 1866 1867

/**
 * 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
1868
 * @data:	first arg to filler(data, page) function, often left as NULL
1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
 *
 * 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,
1880
				int (*filler)(void *, struct page *),
1881 1882 1883 1884
				void *data)
{
	return do_read_cache_page(mapping, index, filler, data, mapping_gfp_mask(mapping));
}
1885 1886
EXPORT_SYMBOL(read_cache_page_async);

1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
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);

1923 1924 1925 1926 1927
/**
 * read_cache_page - read into page cache, fill it if needed
 * @mapping:	the page's address_space
 * @index:	the page index
 * @filler:	function to perform the read
1928
 * @data:	first arg to filler(data, page) function, often left as NULL
1929 1930 1931 1932 1933 1934 1935
 *
 * Read into the page cache. If a page already exists, and PageUptodate() is
 * not set, try to fill the page then wait for it to become unlocked.
 *
 * If the page does not get brought uptodate, return -EIO.
 */
struct page *read_cache_page(struct address_space *mapping,
1936
				pgoff_t index,
1937
				int (*filler)(void *, struct page *),
1938 1939
				void *data)
{
1940
	return wait_on_page_read(read_cache_page_async(mapping, index, filler, data));
L
Linus Torvalds 已提交
1941 1942 1943 1944 1945 1946 1947 1948 1949
}
EXPORT_SYMBOL(read_cache_page);

/*
 * The logic we want is
 *
 *	if suid or (sgid and xgrp)
 *		remove privs
 */
1950
int should_remove_suid(struct dentry *dentry)
L
Linus Torvalds 已提交
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
{
	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;

1966
	if (unlikely(kill && !capable(CAP_FSETID) && S_ISREG(mode)))
1967
		return kill;
L
Linus Torvalds 已提交
1968

1969 1970
	return 0;
}
M
Mark Fasheh 已提交
1971
EXPORT_SYMBOL(should_remove_suid);
1972

1973
static int __remove_suid(struct dentry *dentry, int kill)
1974 1975 1976 1977 1978 1979 1980
{
	struct iattr newattrs;

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

1981
int file_remove_suid(struct file *file)
1982
{
1983
	struct dentry *dentry = file->f_path.dentry;
1984 1985 1986
	struct inode *inode = dentry->d_inode;
	int killsuid;
	int killpriv;
1987
	int error = 0;
1988

1989 1990 1991 1992 1993 1994 1995
	/* Fast path for nothing security related */
	if (IS_NOSEC(inode))
		return 0;

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

1996 1997 1998 1999 2000 2001
	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
2002
	if (!error && (inode->i_sb->s_flags & MS_NOSEC))
2003
		inode->i_flags |= S_NOSEC;
2004

2005
	return error;
L
Linus Torvalds 已提交
2006
}
2007
EXPORT_SYMBOL(file_remove_suid);
L
Linus Torvalds 已提交
2008

N
Nick Piggin 已提交
2009
static size_t __iovec_copy_from_user_inatomic(char *vaddr,
L
Linus Torvalds 已提交
2010 2011
			const struct iovec *iov, size_t base, size_t bytes)
{
2012
	size_t copied = 0, left = 0;
L
Linus Torvalds 已提交
2013 2014 2015 2016 2017 2018

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

		base = 0;
2019
		left = __copy_from_user_inatomic(vaddr, buf, copy);
L
Linus Torvalds 已提交
2020 2021 2022 2023 2024
		copied += copy;
		bytes -= copy;
		vaddr += copy;
		iov++;

2025
		if (unlikely(left))
L
Linus Torvalds 已提交
2026 2027 2028 2029 2030
			break;
	}
	return copied - left;
}

N
Nick Piggin 已提交
2031 2032
/*
 * Copy as much as we can into the page and return the number of bytes which
2033
 * were successfully copied.  If a fault is encountered then return the number of
N
Nick Piggin 已提交
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046
 * 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;
2047
		left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
2048 2049 2050 2051 2052 2053 2054 2055 2056
		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
2057
EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
N
Nick Piggin 已提交
2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074

/*
 * 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;
2075
		left = __copy_from_user(kaddr + offset, buf, bytes);
N
Nick Piggin 已提交
2076 2077 2078 2079 2080 2081 2082 2083
		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
2084
EXPORT_SYMBOL(iov_iter_copy_from_user);
N
Nick Piggin 已提交
2085

N
Nick Piggin 已提交
2086
void iov_iter_advance(struct iov_iter *i, size_t bytes)
N
Nick Piggin 已提交
2087
{
N
Nick Piggin 已提交
2088 2089
	BUG_ON(i->count < bytes);

N
Nick Piggin 已提交
2090 2091
	if (likely(i->nr_segs == 1)) {
		i->iov_offset += bytes;
N
Nick Piggin 已提交
2092
		i->count -= bytes;
N
Nick Piggin 已提交
2093 2094 2095 2096
	} else {
		const struct iovec *iov = i->iov;
		size_t base = i->iov_offset;

2097 2098
		/*
		 * The !iov->iov_len check ensures we skip over unlikely
N
Nick Piggin 已提交
2099
		 * zero-length segments (without overruning the iovec).
2100
		 */
2101
		while (bytes || unlikely(i->count && !iov->iov_len)) {
N
Nick Piggin 已提交
2102
			int copy;
N
Nick Piggin 已提交
2103

N
Nick Piggin 已提交
2104 2105 2106
			copy = min(bytes, iov->iov_len - base);
			BUG_ON(!i->count || i->count < copy);
			i->count -= copy;
N
Nick Piggin 已提交
2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117
			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
2118
EXPORT_SYMBOL(iov_iter_advance);
N
Nick Piggin 已提交
2119

2120 2121 2122 2123 2124 2125 2126 2127 2128 2129
/*
 * 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
2130 2131
{
	char __user *buf = i->iov->iov_base + i->iov_offset;
2132 2133
	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	return fault_in_pages_readable(buf, bytes);
N
Nick Piggin 已提交
2134
}
N
Nick Piggin 已提交
2135
EXPORT_SYMBOL(iov_iter_fault_in_readable);
N
Nick Piggin 已提交
2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147

/*
 * 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
2148
EXPORT_SYMBOL(iov_iter_single_seg_count);
N
Nick Piggin 已提交
2149

L
Linus Torvalds 已提交
2150 2151 2152
/*
 * Performs necessary checks before doing a write
 *
2153
 * Can adjust writing position or amount of bytes to write.
L
Linus Torvalds 已提交
2154 2155 2156 2157 2158 2159
 * 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
2160
	unsigned long limit = rlimit(RLIMIT_FSIZE);
L
Linus Torvalds 已提交
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 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

        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 {
2212
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223
		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;
2224 2225 2226
#else
		return -EPERM;
#endif
L
Linus Torvalds 已提交
2227 2228 2229 2230 2231
	}
	return 0;
}
EXPORT_SYMBOL(generic_write_checks);

2232 2233 2234 2235 2236 2237
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;

2238
	return aops->write_begin(file, mapping, pos, len, flags,
2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
							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;

2249 2250
	mark_page_accessed(page);
	return aops->write_end(file, mapping, pos, len, copied, page, fsdata);
2251 2252 2253
}
EXPORT_SYMBOL(pagecache_write_end);

L
Linus Torvalds 已提交
2254 2255 2256 2257 2258 2259 2260 2261 2262
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;
2263 2264
	size_t		write_len;
	pgoff_t		end;
L
Linus Torvalds 已提交
2265 2266 2267 2268

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

2269 2270 2271
	write_len = iov_length(iov, *nr_segs);
	end = (pos + write_len - 1) >> PAGE_CACHE_SHIFT;

2272
	written = filemap_write_and_wait_range(mapping, pos, pos + write_len - 1);
2273 2274 2275 2276 2277 2278 2279
	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
2280
	 * without clobbering -EIOCBQUEUED from ->direct_IO().
2281 2282 2283 2284
	 */
	if (mapping->nrpages) {
		written = invalidate_inode_pages2_range(mapping,
					pos >> PAGE_CACHE_SHIFT, end);
2285 2286 2287 2288 2289 2290 2291
		/*
		 * If a page can not be invalidated, return 0 to fall back
		 * to buffered write.
		 */
		if (written) {
			if (written == -EBUSY)
				return 0;
2292
			goto out;
2293
		}
2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310
	}

	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
2311
	if (written > 0) {
2312 2313 2314
		pos += written;
		if (pos > i_size_read(inode) && !S_ISBLK(inode->i_mode)) {
			i_size_write(inode, pos);
L
Linus Torvalds 已提交
2315 2316
			mark_inode_dirty(inode);
		}
2317
		*ppos = pos;
L
Linus Torvalds 已提交
2318
	}
2319
out:
L
Linus Torvalds 已提交
2320 2321 2322 2323
	return written;
}
EXPORT_SYMBOL(generic_file_direct_write);

N
Nick Piggin 已提交
2324 2325 2326 2327
/*
 * Find or create a page at the given pagecache position. Return the locked
 * page. This function is specifically for buffered writes.
 */
2328 2329
struct page *grab_cache_page_write_begin(struct address_space *mapping,
					pgoff_t index, unsigned flags)
N
Nick Piggin 已提交
2330 2331 2332
{
	int status;
	struct page *page;
2333 2334 2335
	gfp_t gfp_notmask = 0;
	if (flags & AOP_FLAG_NOFS)
		gfp_notmask = __GFP_FS;
N
Nick Piggin 已提交
2336 2337
repeat:
	page = find_lock_page(mapping, index);
2338
	if (page)
2339
		goto found;
N
Nick Piggin 已提交
2340

2341
	page = __page_cache_alloc(mapping_gfp_mask(mapping) & ~gfp_notmask);
N
Nick Piggin 已提交
2342 2343
	if (!page)
		return NULL;
2344 2345
	status = add_to_page_cache_lru(page, mapping, index,
						GFP_KERNEL & ~gfp_notmask);
N
Nick Piggin 已提交
2346 2347 2348 2349 2350 2351
	if (unlikely(status)) {
		page_cache_release(page);
		if (status == -EEXIST)
			goto repeat;
		return NULL;
	}
2352 2353
found:
	wait_on_page_writeback(page);
N
Nick Piggin 已提交
2354 2355
	return page;
}
2356
EXPORT_SYMBOL(grab_cache_page_write_begin);
N
Nick Piggin 已提交
2357

2358 2359 2360 2361 2362 2363 2364
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
2365 2366 2367 2368 2369 2370 2371
	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;
2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400

	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
2401
		status = a_ops->write_begin(file, mapping, pos, bytes, flags,
2402 2403 2404 2405
						&page, &fsdata);
		if (unlikely(status))
			break;

2406 2407 2408
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

2409 2410 2411 2412 2413
		pagefault_disable();
		copied = iov_iter_copy_from_user_atomic(page, i, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

2414
		mark_page_accessed(page);
2415 2416 2417 2418 2419 2420 2421 2422
		status = a_ops->write_end(file, mapping, pos, bytes, copied,
						page, fsdata);
		if (unlikely(status < 0))
			break;
		copied = status;

		cond_resched();

2423
		iov_iter_advance(i, copied);
2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456
		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);
2457
	status = generic_perform_write(file, &i, pos);
L
Linus Torvalds 已提交
2458 2459

	if (likely(status >= 0)) {
2460 2461
		written += status;
		*ppos = pos + status;
L
Linus Torvalds 已提交
2462 2463 2464 2465 2466 2467
  	}
	
	return written ? written : status;
}
EXPORT_SYMBOL(generic_file_buffered_write);

2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488
/**
 * __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
2489 2490
{
	struct file *file = iocb->ki_filp;
2491
	struct address_space * mapping = file->f_mapping;
L
Linus Torvalds 已提交
2492 2493 2494 2495 2496 2497 2498 2499
	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;
2500 2501 2502
	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
	if (err)
		return err;
L
Linus Torvalds 已提交
2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519

	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;

2520
	err = file_remove_suid(file);
L
Linus Torvalds 已提交
2521 2522 2523
	if (err)
		goto out;

2524
	file_update_time(file);
L
Linus Torvalds 已提交
2525 2526 2527

	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
	if (unlikely(file->f_flags & O_DIRECT)) {
2528 2529 2530 2531 2532
		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
2533 2534 2535 2536 2537 2538 2539 2540
		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;
2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554
		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
2555

2556 2557 2558 2559 2560 2561
		/*
		 * 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;
2562
		err = filemap_write_and_wait_range(file->f_mapping, pos, endbyte);
2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577
		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
2578 2579 2580 2581
out:
	current->backing_dev_info = NULL;
	return written ? written : err;
}
2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594
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.
 */
2595 2596
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
2597 2598
{
	struct file *file = iocb->ki_filp;
2599
	struct inode *inode = file->f_mapping->host;
2600
	struct blk_plug plug;
L
Linus Torvalds 已提交
2601 2602 2603 2604
	ssize_t ret;

	BUG_ON(iocb->ki_pos != pos);

2605
	mutex_lock(&inode->i_mutex);
2606
	blk_start_plug(&plug);
2607
	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
2608
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
2609

2610
	if (ret > 0 || ret == -EIOCBQUEUED) {
L
Linus Torvalds 已提交
2611 2612
		ssize_t err;

2613
		err = generic_write_sync(file, pos, ret);
2614
		if (err < 0 && ret > 0)
L
Linus Torvalds 已提交
2615 2616
			ret = err;
	}
2617
	blk_finish_plug(&plug);
L
Linus Torvalds 已提交
2618 2619 2620 2621
	return ret;
}
EXPORT_SYMBOL(generic_file_aio_write);

2622 2623 2624 2625 2626 2627 2628 2629 2630 2631
/**
 * 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.
 *
2632 2633 2634
 * This may also be called if PG_fscache is set on a page, indicating that the
 * page is known to the local caching routines.
 *
2635
 * The @gfp_mask argument specifies whether I/O may be performed to release
2636
 * this page (__GFP_IO), and whether the call may block (__GFP_WAIT & __GFP_FS).
2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652
 *
 */
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);