Blame view

fs/mpage.c 21.1 KB
b24413180   Greg Kroah-Hartman   License cleanup: ...
1
  // SPDX-License-Identifier: GPL-2.0
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
2
3
4
5
6
7
8
9
  /*
   * 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...
10
   * 15May2002	Andrew Morton
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
11
12
13
14
15
16
   *		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...
17
  #include <linux/export.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
18
19
  #include <linux/mm.h>
  #include <linux/kdev_t.h>
5a0e3ad6a   Tejun Heo   include cleanup: ...
20
  #include <linux/gfp.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
21
22
23
24
25
26
27
  #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_...
28
  #include <linux/mm_inline.h>
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
29
30
31
  #include <linux/writeback.h>
  #include <linux/backing-dev.h>
  #include <linux/pagevec.h>
c515e1fd3   Dan Magenheimer   mm/fs: add hooks ...
32
  #include <linux/cleancache.h>
4db96b71e   Akinobu Mita   vfs: guard end of...
33
  #include "internal.h"
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
34
35
36
37
38
39
40
41
42
43
44
45
46
  
  /*
   * 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...
47
  static void mpage_end_io(struct bio *bio)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
48
  {
2c30c71bd   Kent Overstreet   block: Convert va...
49
50
  	struct bio_vec *bv;
  	int i;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
51

2c30c71bd   Kent Overstreet   block: Convert va...
52
53
  	bio_for_each_segment_all(bv, bio, i) {
  		struct page *page = bv->bv_page;
3f289dcb4   Tejun Heo   block: make bdev_...
54
55
  		page_endio(page, bio_op(bio),
  			   blk_status_to_errno(bio->bi_status));
2c30c71bd   Kent Overstreet   block: Convert va...
56
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
57
  	bio_put(bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
58
  }
eed25cd5b   Mike Christie   mpage: use bio op...
59
  static struct bio *mpage_bio_submit(int op, int op_flags, struct bio *bio)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
60
  {
c32b0d4b3   Hai Shan   fs/mpage.c: conso...
61
  	bio->bi_end_io = mpage_end_io;
eed25cd5b   Mike Christie   mpage: use bio op...
62
63
  	bio_set_op_attrs(bio, op, op_flags);
  	guard_bio_eod(op, bio);
4e49ea4a3   Mike Christie   block/fs/drivers:...
64
  	submit_bio(bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
65
66
67
68
69
70
  	return NULL;
  }
  
  static struct bio *
  mpage_alloc(struct block_device *bdev,
  		sector_t first_sector, int nr_vecs,
dd0fc66fb   Al Viro   [PATCH] gfp flags...
71
  		gfp_t gfp_flags)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
72
73
  {
  	struct bio *bio;
8a5c743e3   Michal Hocko   mm, memcg: use co...
74
75
  	/* Restrict the given (page cache) mask for slab allocations */
  	gfp_flags &= GFP_KERNEL;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
76
77
78
79
80
81
82
83
  	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) {
74d46992e   Christoph Hellwig   block: replace bi...
84
  		bio_set_dev(bio, bdev);
4f024f379   Kent Overstreet   block: Abstract o...
85
  		bio->bi_iter.bi_sector = first_sector;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
  	}
  	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...
112
  		if (inode->i_blkbits == PAGE_SHIFT &&
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
113
114
115
116
  		    buffer_uptodate(bh)) {
  			SetPageUptodate(page);    
  			return;
  		}
93407472a   Fabian Frederick   fs: add i_blocksi...
117
  		create_empty_buffers(page, i_blocksize(inode), 0);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
118
119
120
121
122
123
124
125
126
127
128
129
130
131
  	}
  	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);
  }
357c12065   Jens Axboe   mpage: add argume...
132
133
134
135
  struct mpage_readpage_args {
  	struct bio *bio;
  	struct page *page;
  	unsigned int nr_pages;
74c8164e1   Jens Axboe   mpage: mpage_read...
136
  	bool is_readahead;
357c12065   Jens Axboe   mpage: add argume...
137
138
139
140
  	sector_t last_block_in_bio;
  	struct buffer_head map_bh;
  	unsigned long first_logical_block;
  	get_block_t *get_block;
357c12065   Jens Axboe   mpage: add argume...
141
  };
fa30bd058   Badari Pulavarty   [PATCH] map multi...
142
143
144
145
146
147
148
149
150
  /*
   * 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.
   */
357c12065   Jens Axboe   mpage: add argume...
151
  static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
152
  {
357c12065   Jens Axboe   mpage: add argume...
153
  	struct page *page = args->page;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
154
155
  	struct inode *inode = page->mapping->host;
  	const unsigned blkbits = inode->i_blkbits;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
156
  	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
157
  	const unsigned blocksize = 1 << blkbits;
357c12065   Jens Axboe   mpage: add argume...
158
  	struct buffer_head *map_bh = &args->map_bh;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
159
160
  	sector_t block_in_file;
  	sector_t last_block;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
161
  	sector_t last_block_in_file;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
162
163
164
165
  	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
166
167
  	int length;
  	int fully_mapped = 1;
74c8164e1   Jens Axboe   mpage: mpage_read...
168
  	int op_flags;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
169
170
  	unsigned nblocks;
  	unsigned relative_block;
74c8164e1   Jens Axboe   mpage: mpage_read...
171
172
173
174
175
176
177
178
179
  	gfp_t gfp;
  
  	if (args->is_readahead) {
  		op_flags = REQ_RAHEAD;
  		gfp = readahead_gfp_mask(page->mapping);
  	} else {
  		op_flags = 0;
  		gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
180
181
182
  
  	if (page_has_buffers(page))
  		goto confused;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
183
  	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
357c12065   Jens Axboe   mpage: add argume...
184
  	last_block = block_in_file + args->nr_pages * blocks_per_page;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
185
186
187
188
189
190
191
192
193
  	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;
357c12065   Jens Axboe   mpage: add argume...
194
195
196
197
  	if (buffer_mapped(map_bh) &&
  			block_in_file > args->first_logical_block &&
  			block_in_file < (args->first_logical_block + nblocks)) {
  		unsigned map_offset = block_in_file - args->first_logical_block;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
  		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
222

1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
223
  		if (block_in_file < last_block) {
fa30bd058   Badari Pulavarty   [PATCH] map multi...
224
  			map_bh->b_size = (last_block-block_in_file) << blkbits;
357c12065   Jens Axboe   mpage: add argume...
225
  			if (args->get_block(inode, block_in_file, map_bh, 0))
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
226
  				goto confused;
357c12065   Jens Axboe   mpage: add argume...
227
  			args->first_logical_block = block_in_file;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
228
  		}
fa30bd058   Badari Pulavarty   [PATCH] map multi...
229
  		if (!buffer_mapped(map_bh)) {
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
230
231
232
  			fully_mapped = 0;
  			if (first_hole == blocks_per_page)
  				first_hole = page_block;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
233
234
  			page_block++;
  			block_in_file++;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
235
236
237
238
239
240
241
242
243
  			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...
244
245
  		if (buffer_uptodate(map_bh)) {
  			map_buffer_to_page(page, map_bh, page_block);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
246
247
248
249
250
251
252
  			goto confused;
  		}
  	
  		if (first_hole != blocks_per_page)
  			goto confused;		/* hole -> non-hole */
  
  		/* Contiguous blocks? */
fa30bd058   Badari Pulavarty   [PATCH] map multi...
253
  		if (page_block && blocks[page_block-1] != map_bh->b_blocknr-1)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
254
  			goto confused;
fa30bd058   Badari Pulavarty   [PATCH] map multi...
255
256
257
258
259
260
261
262
263
264
265
266
  		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
267
268
269
  	}
  
  	if (first_hole != blocks_per_page) {
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
270
  		zero_user_segment(page, first_hole << blkbits, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
271
272
273
274
275
276
277
278
  		if (first_hole == 0) {
  			SetPageUptodate(page);
  			unlock_page(page);
  			goto out;
  		}
  	} else if (fully_mapped) {
  		SetPageMappedToDisk(page);
  	}
c515e1fd3   Dan Magenheimer   mm/fs: add hooks ...
279
280
281
282
283
  	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
284
285
286
  	/*
  	 * This page will go to BIO.  Do we need to send this BIO off first?
  	 */
357c12065   Jens Axboe   mpage: add argume...
287
  	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
74c8164e1   Jens Axboe   mpage: mpage_read...
288
  		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
289
290
  
  alloc_new:
357c12065   Jens Axboe   mpage: add argume...
291
  	if (args->bio == NULL) {
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
292
293
294
295
296
  		if (first_hole == blocks_per_page) {
  			if (!bdev_read_page(bdev, blocks[0] << (blkbits - 9),
  								page))
  				goto out;
  		}
357c12065   Jens Axboe   mpage: add argume...
297
298
299
  		args->bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
  					min_t(int, args->nr_pages,
  					      BIO_MAX_PAGES),
74c8164e1   Jens Axboe   mpage: mpage_read...
300
  					gfp);
357c12065   Jens Axboe   mpage: add argume...
301
  		if (args->bio == NULL)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
302
303
304
305
  			goto confused;
  	}
  
  	length = first_hole << blkbits;
357c12065   Jens Axboe   mpage: add argume...
306
  	if (bio_add_page(args->bio, page, length, 0) < length) {
74c8164e1   Jens Axboe   mpage: mpage_read...
307
  		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
308
309
  		goto alloc_new;
  	}
357c12065   Jens Axboe   mpage: add argume...
310
  	relative_block = block_in_file - args->first_logical_block;
38c8e6180   Miquel van Smoorenburg   do_mpage_readpage...
311
312
313
  	nblocks = map_bh->b_size >> blkbits;
  	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
  	    (first_hole != blocks_per_page))
74c8164e1   Jens Axboe   mpage: mpage_read...
314
  		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
315
  	else
357c12065   Jens Axboe   mpage: add argume...
316
  		args->last_block_in_bio = blocks[blocks_per_page - 1];
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
317
  out:
357c12065   Jens Axboe   mpage: add argume...
318
  	return args->bio;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
319
320
  
  confused:
357c12065   Jens Axboe   mpage: add argume...
321
  	if (args->bio)
74c8164e1   Jens Axboe   mpage: mpage_read...
322
  		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
323
  	if (!PageUptodate(page))
357c12065   Jens Axboe   mpage: add argume...
324
  		block_read_full_page(page, args->get_block);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
325
326
327
328
  	else
  		unlock_page(page);
  	goto out;
  }
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
329
  /**
78a4a50a8   Randy Dunlap   docbook: fix file...
330
   * mpage_readpages - populate an address space with some pages & start reads against them
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
331
332
333
   * @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: ...
334
335
   *   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: ...
336
337
338
339
340
341
342
343
344
345
346
347
348
349
   * @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...
350
   * the end-of-file on blocksize < PAGE_SIZE setups.
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
351
352
353
354
355
356
357
358
359
   *
   * 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:
0117d4272   Mauro Carvalho Chehab   fs: add a blank l...
360
   *
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
361
   * 	12 0 1 2 3 4 5 6 7 8 9 10 11 13 14 15 16
78a4a50a8   Randy Dunlap   docbook: fix file...
362
   *
67be2dd1b   Martin Waitz   [PATCH] DocBook: ...
363
364
365
366
367
368
369
370
371
372
   * 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
373
374
375
376
  int
  mpage_readpages(struct address_space *mapping, struct list_head *pages,
  				unsigned nr_pages, get_block_t get_block)
  {
357c12065   Jens Axboe   mpage: add argume...
377
378
  	struct mpage_readpage_args args = {
  		.get_block = get_block,
74c8164e1   Jens Axboe   mpage: mpage_read...
379
  		.is_readahead = true,
357c12065   Jens Axboe   mpage: add argume...
380
  	};
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
381
  	unsigned page_idx;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
382

1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
383
  	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
02c43638e   Andrew Morton   fs/mpage.c:mpage_...
384
  		struct page *page = lru_to_page(pages);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
385
386
387
  
  		prefetchw(&page->flags);
  		list_del(&page->lru);
eb2be1893   Nick Piggin   mm: buffered writ...
388
  		if (!add_to_page_cache_lru(page, mapping,
063d99b4f   Michal Hocko   mm, fs: obey gfp_...
389
  					page->index,
74c8164e1   Jens Axboe   mpage: mpage_read...
390
  					readahead_gfp_mask(mapping))) {
357c12065   Jens Axboe   mpage: add argume...
391
392
393
  			args.page = page;
  			args.nr_pages = nr_pages - page_idx;
  			args.bio = do_mpage_readpage(&args);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
394
  		}
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
395
  		put_page(page);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
396
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
397
  	BUG_ON(!list_empty(pages));
357c12065   Jens Axboe   mpage: add argume...
398
  	if (args.bio)
74c8164e1   Jens Axboe   mpage: mpage_read...
399
  		mpage_bio_submit(REQ_OP_READ, REQ_RAHEAD, args.bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
400
401
402
403
404
405
406
407
408
  	return 0;
  }
  EXPORT_SYMBOL(mpage_readpages);
  
  /*
   * This isn't called much at all
   */
  int mpage_readpage(struct page *page, get_block_t get_block)
  {
357c12065   Jens Axboe   mpage: add argume...
409
410
411
412
  	struct mpage_readpage_args args = {
  		.page = page,
  		.nr_pages = 1,
  		.get_block = get_block,
357c12065   Jens Axboe   mpage: add argume...
413
414
415
416
417
  	};
  
  	args.bio = do_mpage_readpage(&args);
  	if (args.bio)
  		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
  	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...
438

ced117c73   Dmitri Vorobiev   Remove two unneed...
439
440
441
442
443
444
  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...
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
  /*
   * 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);
  }
f892760aa   Matthew Wilcox   fs/mpage.c: fix m...
473
474
475
476
477
478
479
480
481
  /*
   * For situations where we want to clean all buffers attached to a page.
   * We don't need to calculate how many buffers are attached to the page,
   * we just need to specify a number larger than the maximum number of buffers.
   */
  void clean_page_buffers(struct page *page)
  {
  	clean_buffers(page, ~0U);
  }
ced117c73   Dmitri Vorobiev   Remove two unneed...
482
  static int __mpage_writepage(struct page *page, struct writeback_control *wbc,
29a814d2e   Alex Tomas   vfs: add hooks fo...
483
  		      void *data)
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
484
  {
0ea971801   Miklos Szeredi   consolidate gener...
485
486
  	struct mpage_data *mpd = data;
  	struct bio *bio = mpd->bio;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
487
488
489
490
  	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...
491
  	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
492
493
494
495
496
497
498
499
500
501
502
503
  	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...
504
  	int ret = 0;
7637241e6   Jens Axboe   writeback: add wb...
505
  	int op_flags = wbc_to_write_flags(wbc);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
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
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
  
  	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...
561
  	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
562
563
564
565
566
  	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...
567
  		map_bh.b_size = 1 << blkbits;
0ea971801   Miklos Szeredi   consolidate gener...
568
  		if (mpd->get_block(inode, block_in_file, &map_bh, 1))
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
569
570
  			goto confused;
  		if (buffer_new(&map_bh))
e64855c6c   Jan Kara   fs: Add helper to...
571
  			clean_bdev_bh_alias(&map_bh);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
  		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...
592
  	end_index = i_size >> PAGE_SHIFT;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
593
594
595
  	if (page->index >= end_index) {
  		/*
  		 * The page straddles i_size.  It must be zeroed out on each
2a61aa401   Adam Buchbinder   Fix misspellings ...
596
  		 * and every writepage invocation because it may be mmapped.
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
597
598
599
600
601
  		 * "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...
602
  		unsigned offset = i_size & (PAGE_SIZE - 1);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
603
604
605
  
  		if (page->index > end_index || !offset)
  			goto confused;
09cbfeaf1   Kirill A. Shutemov   mm, fs: get rid o...
606
  		zero_user_segment(page, offset, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
607
608
609
610
611
  	}
  
  	/*
  	 * This page will go to BIO.  Do we need to send this BIO off first?
  	 */
0ea971801   Miklos Szeredi   consolidate gener...
612
  	if (bio && mpd->last_block_in_bio != blocks[0] - 1)
eed25cd5b   Mike Christie   mpage: use bio op...
613
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
614
615
616
  
  alloc_new:
  	if (bio == NULL) {
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
617
618
  		if (first_unmapped == blocks_per_page) {
  			if (!bdev_write_page(bdev, blocks[0] << (blkbits - 9),
f892760aa   Matthew Wilcox   fs/mpage.c: fix m...
619
  								page, wbc))
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
620
  				goto out;
47a191fd3   Matthew Wilcox   fs/block_dev.c: a...
621
  		}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
622
  		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
b54ffb73c   Kent Overstreet   block: remove bio...
623
  				BIO_MAX_PAGES, GFP_NOFS|__GFP_HIGH);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
624
625
  		if (bio == NULL)
  			goto confused;
429b3fb02   Tejun Heo   mpage: make __mpa...
626

b16b1deb5   Tejun Heo   writeback: make w...
627
  		wbc_init_bio(wbc, bio);
8e8f92988   Jens Axboe   fs: add support f...
628
  		bio->bi_write_hint = inode->i_write_hint;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
629
630
631
632
633
634
635
  	}
  
  	/*
  	 * 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...
636
  	wbc_account_io(wbc, page, PAGE_SIZE);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
637
638
  	length = first_unmapped << blkbits;
  	if (bio_add_page(bio, page, length, 0) < length) {
eed25cd5b   Mike Christie   mpage: use bio op...
639
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
640
641
  		goto alloc_new;
  	}
90768eee4   Matthew Wilcox   fs/mpage.c: facto...
642
  	clean_buffers(page, first_unmapped);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
643
644
645
646
647
  
  	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...
648
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
649
650
651
652
653
  		if (boundary_block) {
  			write_boundary_block(boundary_bdev,
  					boundary_block, 1 << blkbits);
  		}
  	} else {
0ea971801   Miklos Szeredi   consolidate gener...
654
  		mpd->last_block_in_bio = blocks[blocks_per_page - 1];
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
655
656
657
658
659
  	}
  	goto out;
  
  confused:
  	if (bio)
eed25cd5b   Mike Christie   mpage: use bio op...
660
  		bio = mpage_bio_submit(REQ_OP_WRITE, op_flags, bio);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
661

0ea971801   Miklos Szeredi   consolidate gener...
662
663
  	if (mpd->use_writepage) {
  		ret = mapping->a_ops->writepage(page, wbc);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
664
  	} else {
0ea971801   Miklos Szeredi   consolidate gener...
665
  		ret = -EAGAIN;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
666
667
668
669
670
  		goto out;
  	}
  	/*
  	 * The caller has a ref on the inode, so *mapping is stable
  	 */
0ea971801   Miklos Szeredi   consolidate gener...
671
  	mapping_set_error(mapping, ret);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
672
  out:
0ea971801   Miklos Szeredi   consolidate gener...
673
674
  	mpd->bio = bio;
  	return ret;
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
675
676
677
  }
  
  /**
78a4a50a8   Randy Dunlap   docbook: fix file...
678
   * 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
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
   * @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...
700
  	struct blk_plug plug;
0ea971801   Miklos Szeredi   consolidate gener...
701
  	int ret;
2ed1a6bcf   Jens Axboe   fs: make mpage re...
702
  	blk_start_plug(&plug);
0ea971801   Miklos Szeredi   consolidate gener...
703
704
705
706
707
708
709
710
711
712
713
  	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...
714
  		if (mpd.bio) {
eed25cd5b   Mike Christie   mpage: use bio op...
715
  			int op_flags = (wbc->sync_mode == WB_SYNC_ALL ?
70fd76140   Christoph Hellwig   block,fs: use REQ...
716
  				  REQ_SYNC : 0);
eed25cd5b   Mike Christie   mpage: use bio op...
717
  			mpage_bio_submit(REQ_OP_WRITE, op_flags, mpd.bio);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
718
  		}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
719
  	}
2ed1a6bcf   Jens Axboe   fs: make mpage re...
720
  	blk_finish_plug(&plug);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
721
722
723
  	return ret;
  }
  EXPORT_SYMBOL(mpage_writepages);
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
724
725
726
727
  
  int mpage_writepage(struct page *page, get_block_t get_block,
  	struct writeback_control *wbc)
  {
0ea971801   Miklos Szeredi   consolidate gener...
728
729
730
731
732
733
734
  	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...
735
  	if (mpd.bio) {
eed25cd5b   Mike Christie   mpage: use bio op...
736
  		int op_flags = (wbc->sync_mode == WB_SYNC_ALL ?
70fd76140   Christoph Hellwig   block,fs: use REQ...
737
  			  REQ_SYNC : 0);
eed25cd5b   Mike Christie   mpage: use bio op...
738
  		mpage_bio_submit(REQ_OP_WRITE, op_flags, mpd.bio);
5948edbcb   Roman Pen   fs/mpage.c: forgo...
739
  	}
1da177e4c   Linus Torvalds   Linux-2.6.12-rc2
740
741
742
  	return ret;
  }
  EXPORT_SYMBOL(mpage_writepage);