Blame view

fs/mpage.c 20.5 KB
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
1
2
3
4
5
6
7
8
  /*
   * fs/mpage.c
   *
   * Copyright (C) 2002, Linus Torvalds.
   *
   * Contains functions related to preparing and submitting BIOs which contain
   * multiple pagecache pages.
   *
e1f8e8744   Francois Cami   Remove Andrew Mor...
9
   * 15May2002	Andrew Morton
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
10
11
12
13
14
15
   *		Initial version
   * 27Jun2002	axboe@suse.de
   *		use bio_add_page() to build bio's just the right size
   */
  
  #include <linux/kernel.h>
630d9c472   Paul Gortmaker   fs: reduce the us...
16
  #include <linux/export.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
17
18
  #include <linux/mm.h>
  #include <linux/kdev_t.h>
5a0e3ad6a   Tejun Heo   include cleanup: ...
19
  #include <linux/gfp.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
20
21
22
23
24
25
26
  #include <linux/bio.h>
  #include <linux/fs.h>
  #include <linux/buffer_head.h>
  #include <linux/blkdev.h>
  #include <linux/highmem.h>
  #include <linux/prefetch.h>
  #include <linux/mpage.h>
02c43638e   Andrew Morton   fs/mpage.c:mpage_...
27
  #include <linux/mm_inline.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
28
29
30
  #include <linux/writeback.h>
  #include <linux/backing-dev.h>
  #include <linux/pagevec.h>
c515e1fd3   Dan Magenheimer   mm/fs: add hooks ...
31
  #include <linux/cleancache.h>
4db96b71e   Akinobu Mita   vfs: guard end of...
32
  #include "internal.h"
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
33
34
35
36
37
38
39
40
41
42
43
44
45
  
  /*
   * I/O completion handler for multipage BIOs.
   *
   * The mpage code never puts partial pages into a BIO (except for end-of-file).
   * If a page does not map to a contiguous run of blocks then it simply falls
   * back to block_read_full_page().
   *
   * Why is this?  If a page's completion depends on a number of different BIOs
   * which can complete in any order (or at the same time) then determining the
   * status of that page is hard.  See end_buffer_async_read() for the details.
   * There is no point in duplicating all that complexity.
   */
4246a0b63   Christoph Hellwig   block: add a bi_e...
46
  static void mpage_end_io(struct bio *bio)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
47
  {
2c30c71bd   Kent Overstreet   block: Convert va...
48
49
  	struct bio_vec *bv;
  	int i;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
50

2c30c71bd   Kent Overstreet   block: Convert va...
51
52
  	bio_for_each_segment_all(bv, bio, i) {
  		struct page *page = bv->bv_page;
c11f0c0b5   Jens Axboe   block/mm: make bd...
53
  		page_endio(page, op_is_write(bio_op(bio)), bio->bi_error);
2c30c71bd   Kent Overstreet   block: Convert va...
54
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
55
  	bio_put(bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
56
  }
eed25cd5b   Mike Christie   mpage: use bio op...
57
  static struct bio *mpage_bio_submit(int op, int op_flags, struct bio *bio)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
58
  {
c32b0d4b3   Hai Shan   fs/mpage.c: conso...
59
  	bio->bi_end_io = mpage_end_io;
eed25cd5b   Mike Christie   mpage: use bio op...
60
61
  	bio_set_op_attrs(bio, op, op_flags);
  	guard_bio_eod(op, bio);
4e49ea4a3   Mike Christie   block/fs/drivers:...
62
  	submit_bio(bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
63
64
65
66
67
68
  	return NULL;
  }
  
  static struct bio *
  mpage_alloc(struct block_device *bdev,
  		sector_t first_sector, int nr_vecs,
dd0fc66fb   Al Viro   [PATCH] gfp flags...
69
  		gfp_t gfp_flags)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
70
71
  {
  	struct bio *bio;
8a5c743e3   Michal Hocko   mm, memcg: use co...
72
73
  	/* Restrict the given (page cache) mask for slab allocations */
  	gfp_flags &= GFP_KERNEL;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
74
75
76
77
78
79
80
81
82
  	bio = bio_alloc(gfp_flags, nr_vecs);
  
  	if (bio == NULL && (current->flags & PF_MEMALLOC)) {
  		while (!bio && (nr_vecs /= 2))
  			bio = bio_alloc(gfp_flags, nr_vecs);
  	}
  
  	if (bio) {
  		bio->bi_bdev = bdev;
4f024f379   Kent Overstreet   block: Abstract o...
83
  		bio->bi_iter.bi_sector = first_sector;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
  	}
  	return bio;
  }
  
  /*
   * support function for mpage_readpages.  The fs supplied get_block might
   * return an up to date buffer.  This is used to map that buffer into
   * the page, which allows readpage to avoid triggering a duplicate call
   * to get_block.
   *
   * The idea is to avoid adding buffers to pages that don't already have
   * them.  So when the buffer is up to date and the page size == block size,
   * this marks the page up to date instead of adding new buffers.
   */
  static void 
  map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block) 
  {
  	struct inode *inode = page->mapping->host;
  	struct buffer_head *page_bh, *head;
  	int block = 0;
  
  	if (!page_has_buffers(page)) {
  		/*
  		 * don't make any buffers if there is only one buffer on
  		 * the page and the page just needs to be set up to date
  		 */
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
110
  		if (inode->i_blkbits == PAGE_SHIFT &&
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
  		    buffer_uptodate(bh)) {
  			SetPageUptodate(page);    
  			return;
  		}
  		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
  	}
  	head = page_buffers(page);
  	page_bh = head;
  	do {
  		if (block == page_block) {
  			page_bh->b_state = bh->b_state;
  			page_bh->b_bdev = bh->b_bdev;
  			page_bh->b_blocknr = bh->b_blocknr;
  			break;
  		}
  		page_bh = page_bh->b_this_page;
  		block++;
  	} while (page_bh != head);
  }
fa30bd058   Badari Pulavarty   [PATCH] map multi...
130
131
132
133
134
135
136
137
138
  /*
   * This is the worker routine which does all the work of mapping the disk
   * blocks and constructs largest possible bios, submits them for IO if the
   * blocks are not contiguous on the disk.
   *
   * We pass a buffer_head back and forth and use its buffer_mapped() flag to
   * represent the validity of its disk mapping and to decide when to do the next
   * get_block() call.
   */
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
139
140
  static struct bio *
  do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
fa30bd058   Badari Pulavarty   [PATCH] map multi...
141
  		sector_t *last_block_in_bio, struct buffer_head *map_bh,
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
142
143
  		unsigned long *first_logical_block, get_block_t get_block,
  		gfp_t gfp)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
144
145
146
  {
  	struct inode *inode = page->mapping->host;
  	const unsigned blkbits = inode->i_blkbits;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
147
  	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
148
149
150
  	const unsigned blocksize = 1 << blkbits;
  	sector_t block_in_file;
  	sector_t last_block;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
151
  	sector_t last_block_in_file;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
152
153
154
155
  	sector_t blocks[MAX_BUF_PER_PAGE];
  	unsigned page_block;
  	unsigned first_hole = blocks_per_page;
  	struct block_device *bdev = NULL;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
156
157
  	int length;
  	int fully_mapped = 1;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
158
159
  	unsigned nblocks;
  	unsigned relative_block;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
160
161
162
  
  	if (page_has_buffers(page))
  		goto confused;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
163
  	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
fa30bd058   Badari Pulavarty   [PATCH] map multi...
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
  	last_block = block_in_file + nr_pages * blocks_per_page;
  	last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits;
  	if (last_block > last_block_in_file)
  		last_block = last_block_in_file;
  	page_block = 0;
  
  	/*
  	 * Map blocks using the result from the previous get_blocks call first.
  	 */
  	nblocks = map_bh->b_size >> blkbits;
  	if (buffer_mapped(map_bh) && block_in_file > *first_logical_block &&
  			block_in_file < (*first_logical_block + nblocks)) {
  		unsigned map_offset = block_in_file - *first_logical_block;
  		unsigned last = nblocks - map_offset;
  
  		for (relative_block = 0; ; relative_block++) {
  			if (relative_block == last) {
  				clear_buffer_mapped(map_bh);
  				break;
  			}
  			if (page_block == blocks_per_page)
  				break;
  			blocks[page_block] = map_bh->b_blocknr + map_offset +
  						relative_block;
  			page_block++;
  			block_in_file++;
  		}
  		bdev = map_bh->b_bdev;
  	}
  
  	/*
  	 * Then do more get_blocks calls until we are done with this page.
  	 */
  	map_bh->b_page = page;
  	while (page_block < blocks_per_page) {
  		map_bh->b_state = 0;
  		map_bh->b_size = 0;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
201

1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
202
  		if (block_in_file < last_block) {
fa30bd058   Badari Pulavarty   [PATCH] map multi...
203
204
  			map_bh->b_size = (last_block-block_in_file) << blkbits;
  			if (get_block(inode, block_in_file, map_bh, 0))
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
205
  				goto confused;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
206
  			*first_logical_block = block_in_file;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
207
  		}
fa30bd058   Badari Pulavarty   [PATCH] map multi...
208
  		if (!buffer_mapped(map_bh)) {
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
209
210
211
  			fully_mapped = 0;
  			if (first_hole == blocks_per_page)
  				first_hole = page_block;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
212
213
  			page_block++;
  			block_in_file++;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
214
215
216
217
218
219
220
221
222
  			continue;
  		}
  
  		/* some filesystems will copy data into the page during
  		 * the get_block call, in which case we don't want to
  		 * read it again.  map_buffer_to_page copies the data
  		 * we just collected from get_block into the page's buffers
  		 * so readpage doesn't have to repeat the get_block call
  		 */
fa30bd058   Badari Pulavarty   [PATCH] map multi...
223
224
  		if (buffer_uptodate(map_bh)) {
  			map_buffer_to_page(page, map_bh, page_block);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
225
226
227
228
229
230
231
  			goto confused;
  		}
  	
  		if (first_hole != blocks_per_page)
  			goto confused;		/* hole -> non-hole */
  
  		/* Contiguous blocks? */
fa30bd058   Badari Pulavarty   [PATCH] map multi...
232
  		if (page_block && blocks[page_block-1] != map_bh->b_blocknr-1)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
233
  			goto confused;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
234
235
236
237
238
239
240
241
242
243
244
245
  		nblocks = map_bh->b_size >> blkbits;
  		for (relative_block = 0; ; relative_block++) {
  			if (relative_block == nblocks) {
  				clear_buffer_mapped(map_bh);
  				break;
  			} else if (page_block == blocks_per_page)
  				break;
  			blocks[page_block] = map_bh->b_blocknr+relative_block;
  			page_block++;
  			block_in_file++;
  		}
  		bdev = map_bh->b_bdev;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
246
247
248
  	}
  
  	if (first_hole != blocks_per_page) {
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
249
  		zero_user_segment(page, first_hole << blkbits, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
250
251
252
253
254
255
256
257
  		if (first_hole == 0) {
  			SetPageUptodate(page);
  			unlock_page(page);
  			goto out;
  		}
  	} else if (fully_mapped) {
  		SetPageMappedToDisk(page);
  	}
c515e1fd3   Dan Magenheimer   mm/fs: add hooks ...
258
259
260
261
262
  	if (fully_mapped && blocks_per_page == 1 && !PageUptodate(page) &&
  	    cleancache_get_page(page) == 0) {
  		SetPageUptodate(page);
  		goto confused;
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
263
264
265
266
  	/*
  	 * This page will go to BIO.  Do we need to send this BIO off first?
  	 */
  	if (bio && (*last_block_in_bio != blocks[0] - 1))
eed25cd5b   Mike Christie   mpage: use bio op...
267
  		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
268
269
270
  
  alloc_new:
  	if (bio == NULL) {
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
271
272
273
274
275
  		if (first_hole == blocks_per_page) {
  			if (!bdev_read_page(bdev, blocks[0] << (blkbits - 9),
  								page))
  				goto out;
  		}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
276
  		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
277
  				min_t(int, nr_pages, BIO_MAX_PAGES), gfp);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
278
279
280
281
282
283
  		if (bio == NULL)
  			goto confused;
  	}
  
  	length = first_hole << blkbits;
  	if (bio_add_page(bio, page, length, 0) < length) {
eed25cd5b   Mike Christie   mpage: use bio op...
284
  		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
285
286
  		goto alloc_new;
  	}
38c8e6180   Miquel van Smoorenburg   do_mpage_readpage...
287
288
289
290
  	relative_block = block_in_file - *first_logical_block;
  	nblocks = map_bh->b_size >> blkbits;
  	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
  	    (first_hole != blocks_per_page))
eed25cd5b   Mike Christie   mpage: use bio op...
291
  		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
292
293
294
295
296
297
298
  	else
  		*last_block_in_bio = blocks[blocks_per_page - 1];
  out:
  	return bio;
  
  confused:
  	if (bio)
eed25cd5b   Mike Christie   mpage: use bio op...
299
  		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
300
301
302
303
304
305
  	if (!PageUptodate(page))
  	        block_read_full_page(page, get_block);
  	else
  		unlock_page(page);
  	goto out;
  }
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
306
  /**
78a4a50a8   Randy Dunlap   docbook: fix file...
307
   * mpage_readpages - populate an address space with some pages & start reads against them
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
308
309
310
   * @mapping: the address_space
   * @pages: The address of a list_head which contains the target pages.  These
   *   pages have their ->index populated and are otherwise uninitialised.
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
311
312
   *   The page at @pages->prev has the lowest file offset, and reads should be
   *   issued in @pages->prev to @pages->next order.
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
313
314
315
316
317
318
319
320
321
322
323
324
325
326
   * @nr_pages: The number of pages at *@pages
   * @get_block: The filesystem's block mapper function.
   *
   * This function walks the pages and the blocks within each page, building and
   * emitting large BIOs.
   *
   * If anything unusual happens, such as:
   *
   * - encountering a page which has buffers
   * - encountering a page which has a non-hole after a hole
   * - encountering a page with non-contiguous blocks
   *
   * then this code just gives up and calls the buffer_head-based read function.
   * It does handle a page which has holes at the end - that is a common case:
ea1754a08   Kirill A. Shutemov   mm, fs: remove re...
327
   * the end-of-file on blocksize < PAGE_SIZE setups.
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
328
329
330
331
332
333
334
335
336
337
   *
   * BH_Boundary explanation:
   *
   * There is a problem.  The mpage read code assembles several pages, gets all
   * their disk mappings, and then submits them all.  That's fine, but obtaining
   * the disk mappings may require I/O.  Reads of indirect blocks, for example.
   *
   * So an mpage read of the first 16 blocks of an ext2 file will cause I/O to be
   * submitted in the following order:
   * 	12 0 1 2 3 4 5 6 7 8 9 10 11 13 14 15 16
78a4a50a8   Randy Dunlap   docbook: fix file...
338
   *
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
339
340
341
342
343
344
345
346
347
348
   * because the indirect block has to be read to get the mappings of blocks
   * 13,14,15,16.  Obviously, this impacts performance.
   *
   * So what we do it to allow the filesystem's get_block() function to set
   * BH_Boundary when it maps block 11.  BH_Boundary says: mapping of the block
   * after this one will require I/O against a block which is probably close to
   * this one.  So you should push what I/O you have currently accumulated.
   *
   * This all causes the disk requests to be issued in the correct order.
   */
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
349
350
351
352
353
354
355
  int
  mpage_readpages(struct address_space *mapping, struct list_head *pages,
  				unsigned nr_pages, get_block_t get_block)
  {
  	struct bio *bio = NULL;
  	unsigned page_idx;
  	sector_t last_block_in_bio = 0;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
356
357
  	struct buffer_head map_bh;
  	unsigned long first_logical_block = 0;
8a5c743e3   Michal Hocko   mm, memcg: use co...
358
  	gfp_t gfp = readahead_gfp_mask(mapping);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
359

79ffab343   Aneesh Kumar K.V   ext4: Properly in...
360
361
  	map_bh.b_state = 0;
  	map_bh.b_size = 0;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
362
  	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
02c43638e   Andrew Morton   fs/mpage.c:mpage_...
363
  		struct page *page = lru_to_page(pages);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
364
365
366
  
  		prefetchw(&page->flags);
  		list_del(&page->lru);
eb2be1893   Nick Piggin   mm: buffered writ...
367
  		if (!add_to_page_cache_lru(page, mapping,
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
368
369
  					page->index,
  					gfp)) {
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
370
371
  			bio = do_mpage_readpage(bio, page,
  					nr_pages - page_idx,
fa30bd058   Badari Pulavarty   [PATCH] map multi...
372
373
  					&last_block_in_bio, &map_bh,
  					&first_logical_block,
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
374
  					get_block, gfp);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
375
  		}
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
376
  		put_page(page);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
377
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
378
379
  	BUG_ON(!list_empty(pages));
  	if (bio)
eed25cd5b   Mike Christie   mpage: use bio op...
380
  		mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
381
382
383
384
385
386
387
388
389
390
391
  	return 0;
  }
  EXPORT_SYMBOL(mpage_readpages);
  
  /*
   * This isn't called much at all
   */
  int mpage_readpage(struct page *page, get_block_t get_block)
  {
  	struct bio *bio = NULL;
  	sector_t last_block_in_bio = 0;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
392
393
  	struct buffer_head map_bh;
  	unsigned long first_logical_block = 0;
c62d25556   Michal Hocko   mm, fs: introduce...
394
  	gfp_t gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
395

79ffab343   Aneesh Kumar K.V   ext4: Properly in...
396
397
  	map_bh.b_state = 0;
  	map_bh.b_size = 0;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
398
  	bio = do_mpage_readpage(bio, page, 1, &last_block_in_bio,
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
399
  			&map_bh, &first_logical_block, get_block, gfp);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
400
  	if (bio)
eed25cd5b   Mike Christie   mpage: use bio op...
401
  		mpage_bio_submit(REQ_OP_READ, 0, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
  	return 0;
  }
  EXPORT_SYMBOL(mpage_readpage);
  
  /*
   * Writing is not so simple.
   *
   * If the page has buffers then they will be used for obtaining the disk
   * mapping.  We only support pages which are fully mapped-and-dirty, with a
   * special case for pages which are unmapped at the end: end-of-file.
   *
   * If the page has no buffers (preferred) then the page is mapped here.
   *
   * If all blocks are found to be contiguous then the page can go into the
   * BIO.  Otherwise fall back to the mapping's writepage().
   * 
   * FIXME: This code wants an estimate of how many pages are still to be
   * written, so it can intelligently allocate a suitably-sized BIO.  For now,
   * just allocate full-size (16-page) BIOs.
   */
0ea971801   Miklos Szeredi   consolidate gener...
422

ced117c73   Dmitri Vorobiev   Remove two unneed...
423
424
425
426
427
428
  struct mpage_data {
  	struct bio *bio;
  	sector_t last_block_in_bio;
  	get_block_t *get_block;
  	unsigned use_writepage;
  };
90768eee4   Matthew Wilcox   fs/mpage.c: facto...
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
  /*
   * We have our BIO, so we can now mark the buffers clean.  Make
   * sure to only clean buffers which we know we'll be writing.
   */
  static void clean_buffers(struct page *page, unsigned first_unmapped)
  {
  	unsigned buffer_counter = 0;
  	struct buffer_head *bh, *head;
  	if (!page_has_buffers(page))
  		return;
  	head = page_buffers(page);
  	bh = head;
  
  	do {
  		if (buffer_counter++ == first_unmapped)
  			break;
  		clear_buffer_dirty(bh);
  		bh = bh->b_this_page;
  	} while (bh != head);
  
  	/*
  	 * we cannot drop the bh if the page is not uptodate or a concurrent
  	 * readpage would fail to serialize with the bh and it would read from
  	 * disk before we reach the platter.
  	 */
  	if (buffer_heads_over_limit && PageUptodate(page))
  		try_to_free_buffers(page);
  }
ced117c73   Dmitri Vorobiev   Remove two unneed...
457
  static int __mpage_writepage(struct page *page, struct writeback_control *wbc,
29a814d2e   Alex Tomas   vfs: add hooks fo...
458
  		      void *data)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
459
  {
0ea971801   Miklos Szeredi   consolidate gener...
460
461
  	struct mpage_data *mpd = data;
  	struct bio *bio = mpd->bio;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
462
463
464
465
  	struct address_space *mapping = page->mapping;
  	struct inode *inode = page->mapping->host;
  	const unsigned blkbits = inode->i_blkbits;
  	unsigned long end_index;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
466
  	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
467
468
469
470
471
472
473
474
475
476
477
478
  	sector_t last_block;
  	sector_t block_in_file;
  	sector_t blocks[MAX_BUF_PER_PAGE];
  	unsigned page_block;
  	unsigned first_unmapped = blocks_per_page;
  	struct block_device *bdev = NULL;
  	int boundary = 0;
  	sector_t boundary_block = 0;
  	struct block_device *boundary_bdev = NULL;
  	int length;
  	struct buffer_head map_bh;
  	loff_t i_size = i_size_read(inode);
0ea971801   Miklos Szeredi   consolidate gener...
479
  	int ret = 0;
7637241e6   Jens Axboe   writeback: add wb...
480
  	int op_flags = wbc_to_write_flags(wbc);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
  
  	if (page_has_buffers(page)) {
  		struct buffer_head *head = page_buffers(page);
  		struct buffer_head *bh = head;
  
  		/* If they're all mapped and dirty, do it */
  		page_block = 0;
  		do {
  			BUG_ON(buffer_locked(bh));
  			if (!buffer_mapped(bh)) {
  				/*
  				 * unmapped dirty buffers are created by
  				 * __set_page_dirty_buffers -> mmapped data
  				 */
  				if (buffer_dirty(bh))
  					goto confused;
  				if (first_unmapped == blocks_per_page)
  					first_unmapped = page_block;
  				continue;
  			}
  
  			if (first_unmapped != blocks_per_page)
  				goto confused;	/* hole -> non-hole */
  
  			if (!buffer_dirty(bh) || !buffer_uptodate(bh))
  				goto confused;
  			if (page_block) {
  				if (bh->b_blocknr != blocks[page_block-1] + 1)
  					goto confused;
  			}
  			blocks[page_block++] = bh->b_blocknr;
  			boundary = buffer_boundary(bh);
  			if (boundary) {
  				boundary_block = bh->b_blocknr;
  				boundary_bdev = bh->b_bdev;
  			}
  			bdev = bh->b_bdev;
  		} while ((bh = bh->b_this_page) != head);
  
  		if (first_unmapped)
  			goto page_is_mapped;
  
  		/*
  		 * Page has buffers, but they are all unmapped. The page was
  		 * created by pagein or read over a hole which was handled by
  		 * block_read_full_page().  If this address_space is also
  		 * using mpage_readpages then this can rarely happen.
  		 */
  		goto confused;
  	}
  
  	/*
  	 * The page has no buffers: map it to disk
  	 */
  	BUG_ON(!PageUptodate(page));
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
536
  	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
537
538
539
540
541
  	last_block = (i_size - 1) >> blkbits;
  	map_bh.b_page = page;
  	for (page_block = 0; page_block < blocks_per_page; ) {
  
  		map_bh.b_state = 0;
b0cf2321c   Badari Pulavarty   [PATCH] pass b_si...
542
  		map_bh.b_size = 1 << blkbits;
0ea971801   Miklos Szeredi   consolidate gener...
543
  		if (mpd->get_block(inode, block_in_file, &map_bh, 1))
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
544
545
  			goto confused;
  		if (buffer_new(&map_bh))
e64855c6c   Jan Kara   fs: Add helper to...
546
  			clean_bdev_bh_alias(&map_bh);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
  		if (buffer_boundary(&map_bh)) {
  			boundary_block = map_bh.b_blocknr;
  			boundary_bdev = map_bh.b_bdev;
  		}
  		if (page_block) {
  			if (map_bh.b_blocknr != blocks[page_block-1] + 1)
  				goto confused;
  		}
  		blocks[page_block++] = map_bh.b_blocknr;
  		boundary = buffer_boundary(&map_bh);
  		bdev = map_bh.b_bdev;
  		if (block_in_file == last_block)
  			break;
  		block_in_file++;
  	}
  	BUG_ON(page_block == 0);
  
  	first_unmapped = page_block;
  
  page_is_mapped:
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
567
  	end_index = i_size >> PAGE_SHIFT;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
568
569
570
  	if (page->index >= end_index) {
  		/*
  		 * The page straddles i_size.  It must be zeroed out on each
2a61aa401   Adam Buchbinder   Fix misspellings ...
571
  		 * and every writepage invocation because it may be mmapped.
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
572
573
574
575
576
  		 * "A file is mapped in multiples of the page size.  For a file
  		 * that is not a multiple of the page size, the remaining memory
  		 * is zeroed when mapped, and writes to that region are not
  		 * written out to the file."
  		 */
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
577
  		unsigned offset = i_size & (PAGE_SIZE - 1);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
578
579
580
  
  		if (page->index > end_index || !offset)
  			goto confused;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
581
  		zero_user_segment(page, offset, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
582
583
584
585
586
  	}
  
  	/*
  	 * This page will go to BIO.  Do we need to send this BIO off first?
  	 */
0ea971801   Miklos Szeredi   consolidate gener...
587
  	if (bio && mpd->last_block_in_bio != blocks[0] - 1)
eed25cd5b   Mike Christie   mpage: use bio op...
588
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
589
590
591
  
  alloc_new:
  	if (bio == NULL) {
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
592
593
594
595
596
597
598
  		if (first_unmapped == blocks_per_page) {
  			if (!bdev_write_page(bdev, blocks[0] << (blkbits - 9),
  								page, wbc)) {
  				clean_buffers(page, first_unmapped);
  				goto out;
  			}
  		}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
599
  		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
b54ffb73c   Kent Overstreet   block: remove bio...
600
  				BIO_MAX_PAGES, GFP_NOFS|__GFP_HIGH);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
601
602
  		if (bio == NULL)
  			goto confused;
429b3fb02   Tejun Heo   mpage: make __mpa...
603

b16b1deb5   Tejun Heo   writeback: make w...
604
  		wbc_init_bio(wbc, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
605
606
607
608
609
610
611
  	}
  
  	/*
  	 * Must try to add the page before marking the buffer clean or
  	 * the confused fail path above (OOM) will be very confused when
  	 * it finds all bh marked clean (i.e. it will not write anything)
  	 */
2a8149081   Tejun Heo   writeback: implem...
612
  	wbc_account_io(wbc, page, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
613
614
  	length = first_unmapped << blkbits;
  	if (bio_add_page(bio, page, length, 0) < length) {
eed25cd5b   Mike Christie   mpage: use bio op...
615
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
616
617
  		goto alloc_new;
  	}
90768eee4   Matthew Wilcox   fs/mpage.c: facto...
618
  	clean_buffers(page, first_unmapped);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
619
620
621
622
623
  
  	BUG_ON(PageWriteback(page));
  	set_page_writeback(page);
  	unlock_page(page);
  	if (boundary || (first_unmapped != blocks_per_page)) {
eed25cd5b   Mike Christie   mpage: use bio op...
624
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
625
626
627
628
629
  		if (boundary_block) {
  			write_boundary_block(boundary_bdev,
  					boundary_block, 1 << blkbits);
  		}
  	} else {
0ea971801   Miklos Szeredi   consolidate gener...
630
  		mpd->last_block_in_bio = blocks[blocks_per_page - 1];
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
631
632
633
634
635
  	}
  	goto out;
  
  confused:
  	if (bio)
eed25cd5b   Mike Christie   mpage: use bio op...
636
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
637

0ea971801   Miklos Szeredi   consolidate gener...
638
639
  	if (mpd->use_writepage) {
  		ret = mapping->a_ops->writepage(page, wbc);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
640
  	} else {
0ea971801   Miklos Szeredi   consolidate gener...
641
  		ret = -EAGAIN;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
642
643
644
645
646
  		goto out;
  	}
  	/*
  	 * The caller has a ref on the inode, so *mapping is stable
  	 */
0ea971801   Miklos Szeredi   consolidate gener...
647
  	mapping_set_error(mapping, ret);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
648
  out:
0ea971801   Miklos Szeredi   consolidate gener...
649
650
  	mpd->bio = bio;
  	return ret;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
651
652
653
  }
  
  /**
78a4a50a8   Randy Dunlap   docbook: fix file...
654
   * mpage_writepages - walk the list of dirty pages of the given address space & writepage() all of them
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
   * @mapping: address space structure to write
   * @wbc: subtract the number of written pages from *@wbc->nr_to_write
   * @get_block: the filesystem's block mapper function.
   *             If this is NULL then use a_ops->writepage.  Otherwise, go
   *             direct-to-BIO.
   *
   * This is a library function, which implements the writepages()
   * address_space_operation.
   *
   * If a page is already under I/O, generic_writepages() skips it, even
   * if it's dirty.  This is desirable behaviour for memory-cleaning writeback,
   * but it is INCORRECT for data-integrity system calls such as fsync().  fsync()
   * and msync() need to guarantee that all the data which was dirty at the time
   * the call was made get new I/O started against them.  If wbc->sync_mode is
   * WB_SYNC_ALL then we were called for data integrity and we must wait for
   * existing IO to complete.
   */
  int
  mpage_writepages(struct address_space *mapping,
  		struct writeback_control *wbc, get_block_t get_block)
  {
2ed1a6bcf   Jens Axboe   fs: make mpage re...
676
  	struct blk_plug plug;
0ea971801   Miklos Szeredi   consolidate gener...
677
  	int ret;
2ed1a6bcf   Jens Axboe   fs: make mpage re...
678
  	blk_start_plug(&plug);
0ea971801   Miklos Szeredi   consolidate gener...
679
680
681
682
683
684
685
686
687
688
689
  	if (!get_block)
  		ret = generic_writepages(mapping, wbc);
  	else {
  		struct mpage_data mpd = {
  			.bio = NULL,
  			.last_block_in_bio = 0,
  			.get_block = get_block,
  			.use_writepage = 1,
  		};
  
  		ret = write_cache_pages(mapping, wbc, __mpage_writepage, &mpd);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
690
  		if (mpd.bio) {
eed25cd5b   Mike Christie   mpage: use bio op...
691
  			int op_flags = (wbc->sync_mode == WB_SYNC_ALL ?
70fd76140   Christoph Hellwig   block,fs: use REQ...
692
  				  REQ_SYNC : 0);
eed25cd5b   Mike Christie   mpage: use bio op...
693
  			mpage_bio_submit(REQ_OP_WRITE, op_flags, mpd.bio);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
694
  		}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
695
  	}
2ed1a6bcf   Jens Axboe   fs: make mpage re...
696
  	blk_finish_plug(&plug);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
697
698
699
  	return ret;
  }
  EXPORT_SYMBOL(mpage_writepages);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
700
701
702
703
  
  int mpage_writepage(struct page *page, get_block_t get_block,
  	struct writeback_control *wbc)
  {
0ea971801   Miklos Szeredi   consolidate gener...
704
705
706
707
708
709
710
  	struct mpage_data mpd = {
  		.bio = NULL,
  		.last_block_in_bio = 0,
  		.get_block = get_block,
  		.use_writepage = 0,
  	};
  	int ret = __mpage_writepage(page, wbc, &mpd);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
711
  	if (mpd.bio) {
eed25cd5b   Mike Christie   mpage: use bio op...
712
  		int op_flags = (wbc->sync_mode == WB_SYNC_ALL ?
70fd76140   Christoph Hellwig   block,fs: use REQ...
713
  			  REQ_SYNC : 0);
eed25cd5b   Mike Christie   mpage: use bio op...
714
  		mpage_bio_submit(REQ_OP_WRITE, op_flags, mpd.bio);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
715
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
716
717
718
  	return ret;
  }
  EXPORT_SYMBOL(mpage_writepage);