Blame view
fs/buffer.c
89.5 KB
457c89965
|
1 |
// SPDX-License-Identifier: GPL-2.0-only |
1da177e4c
|
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
/* * linux/fs/buffer.c * * Copyright (C) 1991, 1992, 2002 Linus Torvalds */ /* * Start bdflush() with kernel_thread not syscall - Paul Gortmaker, 12/95 * * Removed a lot of unnecessary code and simplified things now that * the buffer cache isn't our primary cache - Andrew Tridgell 12/96 * * Speed up hash, lru, and free list operations. Use gfp() for allocating * hash table, use SLAB cache for buffer heads. SMP threading. -DaveM * * Added 32k buffer block sizes - these are required older ARM systems. - RMK * * async buffer flushing, 1999 Andrea Arcangeli <andrea@suse.de> */ |
1da177e4c
|
21 |
#include <linux/kernel.h> |
f361bf4a6
|
22 |
#include <linux/sched/signal.h> |
1da177e4c
|
23 24 |
#include <linux/syscalls.h> #include <linux/fs.h> |
ae259a9c8
|
25 |
#include <linux/iomap.h> |
1da177e4c
|
26 27 28 |
#include <linux/mm.h> #include <linux/percpu.h> #include <linux/slab.h> |
16f7e0fe2
|
29 |
#include <linux/capability.h> |
1da177e4c
|
30 31 32 33 |
#include <linux/blkdev.h> #include <linux/file.h> #include <linux/quotaops.h> #include <linux/highmem.h> |
630d9c472
|
34 |
#include <linux/export.h> |
bafc0dba1
|
35 |
#include <linux/backing-dev.h> |
1da177e4c
|
36 37 38 39 |
#include <linux/writeback.h> #include <linux/hash.h> #include <linux/suspend.h> #include <linux/buffer_head.h> |
55e829af0
|
40 |
#include <linux/task_io_accounting_ops.h> |
1da177e4c
|
41 |
#include <linux/bio.h> |
1da177e4c
|
42 43 44 |
#include <linux/cpu.h> #include <linux/bitops.h> #include <linux/mpage.h> |
fb1c8f93d
|
45 |
#include <linux/bit_spinlock.h> |
29f3ad7d8
|
46 |
#include <linux/pagevec.h> |
f745c6f5f
|
47 |
#include <linux/sched/mm.h> |
5305cb830
|
48 |
#include <trace/events/block.h> |
1da177e4c
|
49 50 |
static int fsync_buffers_list(spinlock_t *lock, struct list_head *list); |
2a222ca99
|
51 |
static int submit_bh_wbc(int op, int op_flags, struct buffer_head *bh, |
8e8f92988
|
52 |
enum rw_hint hint, struct writeback_control *wbc); |
1da177e4c
|
53 54 |
#define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers) |
f0059afd3
|
55 56 |
inline void touch_buffer(struct buffer_head *bh) { |
5305cb830
|
57 |
trace_block_touch_buffer(bh); |
f0059afd3
|
58 59 60 |
mark_page_accessed(bh->b_page); } EXPORT_SYMBOL(touch_buffer); |
fc9b52cd8
|
61 |
void __lock_buffer(struct buffer_head *bh) |
1da177e4c
|
62 |
{ |
743162013
|
63 |
wait_on_bit_lock_io(&bh->b_state, BH_Lock, TASK_UNINTERRUPTIBLE); |
1da177e4c
|
64 65 |
} EXPORT_SYMBOL(__lock_buffer); |
fc9b52cd8
|
66 |
void unlock_buffer(struct buffer_head *bh) |
1da177e4c
|
67 |
{ |
51b07fc3c
|
68 |
clear_bit_unlock(BH_Lock, &bh->b_state); |
4e857c58e
|
69 |
smp_mb__after_atomic(); |
1da177e4c
|
70 71 |
wake_up_bit(&bh->b_state, BH_Lock); } |
1fe72eaa0
|
72 |
EXPORT_SYMBOL(unlock_buffer); |
1da177e4c
|
73 74 |
/* |
b45972265
|
75 76 77 78 79 80 81 82 83 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 |
* Returns if the page has dirty or writeback buffers. If all the buffers * are unlocked and clean then the PageDirty information is stale. If * any of the pages are locked, it is assumed they are locked for IO. */ void buffer_check_dirty_writeback(struct page *page, bool *dirty, bool *writeback) { struct buffer_head *head, *bh; *dirty = false; *writeback = false; BUG_ON(!PageLocked(page)); if (!page_has_buffers(page)) return; if (PageWriteback(page)) *writeback = true; head = page_buffers(page); bh = head; do { if (buffer_locked(bh)) *writeback = true; if (buffer_dirty(bh)) *dirty = true; bh = bh->b_this_page; } while (bh != head); } EXPORT_SYMBOL(buffer_check_dirty_writeback); /* |
1da177e4c
|
109 110 111 112 113 114 |
* Block until a buffer comes unlocked. This doesn't stop it * from becoming locked again - you have to lock it yourself * if you want to preserve its state. */ void __wait_on_buffer(struct buffer_head * bh) { |
743162013
|
115 |
wait_on_bit_io(&bh->b_state, BH_Lock, TASK_UNINTERRUPTIBLE); |
1da177e4c
|
116 |
} |
1fe72eaa0
|
117 |
EXPORT_SYMBOL(__wait_on_buffer); |
1da177e4c
|
118 119 120 121 122 |
static void __clear_page_buffers(struct page *page) { ClearPagePrivate(page); |
4c21e2f24
|
123 |
set_page_private(page, 0); |
09cbfeaf1
|
124 |
put_page(page); |
1da177e4c
|
125 |
} |
b744c2ac4
|
126 |
static void buffer_io_error(struct buffer_head *bh, char *msg) |
1da177e4c
|
127 |
{ |
432f16e64
|
128 129 |
if (!test_bit(BH_Quiet, &bh->b_state)) printk_ratelimited(KERN_ERR |
a1c6f0573
|
130 131 132 |
"Buffer I/O error on dev %pg, logical block %llu%s ", bh->b_bdev, (unsigned long long)bh->b_blocknr, msg); |
1da177e4c
|
133 134 135 |
} /* |
68671f35f
|
136 137 138 139 140 141 |
* End-of-IO handler helper function which does not touch the bh after * unlocking it. * Note: unlock_buffer() sort-of does touch the bh after unlocking it, but * a race there is benign: unlock_buffer() only use the bh's address for * hashing after unlocking the buffer, so it doesn't actually touch the bh * itself. |
1da177e4c
|
142 |
*/ |
68671f35f
|
143 |
static void __end_buffer_read_notouch(struct buffer_head *bh, int uptodate) |
1da177e4c
|
144 145 146 147 |
{ if (uptodate) { set_buffer_uptodate(bh); } else { |
70246286e
|
148 |
/* This happens, due to failed read-ahead attempts. */ |
1da177e4c
|
149 150 151 |
clear_buffer_uptodate(bh); } unlock_buffer(bh); |
68671f35f
|
152 153 154 155 156 157 158 159 160 |
} /* * Default synchronous end-of-IO handler.. Just mark it up-to-date and * unlock the buffer. This is what ll_rw_block uses too. */ void end_buffer_read_sync(struct buffer_head *bh, int uptodate) { __end_buffer_read_notouch(bh, uptodate); |
1da177e4c
|
161 162 |
put_bh(bh); } |
1fe72eaa0
|
163 |
EXPORT_SYMBOL(end_buffer_read_sync); |
1da177e4c
|
164 165 166 |
void end_buffer_write_sync(struct buffer_head *bh, int uptodate) { |
1da177e4c
|
167 168 169 |
if (uptodate) { set_buffer_uptodate(bh); } else { |
432f16e64
|
170 |
buffer_io_error(bh, ", lost sync page write"); |
87354e5de
|
171 |
mark_buffer_write_io_error(bh); |
1da177e4c
|
172 173 174 175 176 |
clear_buffer_uptodate(bh); } unlock_buffer(bh); put_bh(bh); } |
1fe72eaa0
|
177 |
EXPORT_SYMBOL(end_buffer_write_sync); |
1da177e4c
|
178 179 |
/* |
1da177e4c
|
180 181 182 183 184 |
* Various filesystems appear to want __find_get_block to be non-blocking. * But it's the page lock which protects the buffers. To get around this, * we get exclusion from try_to_free_buffers with the blockdev mapping's * private_lock. * |
b93b01631
|
185 |
* Hack idea: for the blockdev mapping, private_lock contention |
1da177e4c
|
186 |
* may be quite high. This code could TryLock the page, and if that |
b93b01631
|
187 |
* succeeds, there is no need to take private_lock. |
1da177e4c
|
188 189 |
*/ static struct buffer_head * |
385fd4c59
|
190 |
__find_get_block_slow(struct block_device *bdev, sector_t block) |
1da177e4c
|
191 192 193 194 195 196 197 198 199 |
{ struct inode *bd_inode = bdev->bd_inode; struct address_space *bd_mapping = bd_inode->i_mapping; struct buffer_head *ret = NULL; pgoff_t index; struct buffer_head *bh; struct buffer_head *head; struct page *page; int all_mapped = 1; |
43636c804
|
200 |
static DEFINE_RATELIMIT_STATE(last_warned, HZ, 1); |
1da177e4c
|
201 |
|
09cbfeaf1
|
202 |
index = block >> (PAGE_SHIFT - bd_inode->i_blkbits); |
2457aec63
|
203 |
page = find_get_page_flags(bd_mapping, index, FGP_ACCESSED); |
1da177e4c
|
204 205 206 207 208 209 210 211 212 |
if (!page) goto out; spin_lock(&bd_mapping->private_lock); if (!page_has_buffers(page)) goto out_unlock; head = page_buffers(page); bh = head; do { |
97f76d3d1
|
213 214 215 |
if (!buffer_mapped(bh)) all_mapped = 0; else if (bh->b_blocknr == block) { |
1da177e4c
|
216 217 218 219 |
ret = bh; get_bh(bh); goto out_unlock; } |
1da177e4c
|
220 221 222 223 224 225 226 227 |
bh = bh->b_this_page; } while (bh != head); /* we might be here because some of the buffers on this page are * not mapped. This is due to various races between * file io on the block device and getblk. It gets dealt with * elsewhere, don't buffer_error if we had some unmapped buffers */ |
43636c804
|
228 229 230 231 232 233 234 235 236 237 |
ratelimit_set_flags(&last_warned, RATELIMIT_MSG_ON_RELEASE); if (all_mapped && __ratelimit(&last_warned)) { printk("__find_get_block_slow() failed. block=%llu, " "b_blocknr=%llu, b_state=0x%08lx, b_size=%zu, " "device %pg blocksize: %d ", (unsigned long long)block, (unsigned long long)bh->b_blocknr, bh->b_state, bh->b_size, bdev, 1 << bd_inode->i_blkbits); |
1da177e4c
|
238 239 240 |
} out_unlock: spin_unlock(&bd_mapping->private_lock); |
09cbfeaf1
|
241 |
put_page(page); |
1da177e4c
|
242 243 244 |
out: return ret; } |
1da177e4c
|
245 |
/* |
1da177e4c
|
246 247 248 249 250 |
* I/O completion handler for block_read_full_page() - pages * which come unlocked at the end of I/O. */ static void end_buffer_async_read(struct buffer_head *bh, int uptodate) { |
1da177e4c
|
251 |
unsigned long flags; |
a39722034
|
252 |
struct buffer_head *first; |
1da177e4c
|
253 254 255 256 257 258 259 260 261 262 263 |
struct buffer_head *tmp; struct page *page; int page_uptodate = 1; BUG_ON(!buffer_async_read(bh)); page = bh->b_page; if (uptodate) { set_buffer_uptodate(bh); } else { clear_buffer_uptodate(bh); |
432f16e64
|
264 |
buffer_io_error(bh, ", async page read"); |
1da177e4c
|
265 266 267 268 269 270 271 272 |
SetPageError(page); } /* * Be _very_ careful from here on. Bad things can happen if * two buffer heads end IO at almost the same time and both * decide that the page is now completely done. */ |
a39722034
|
273 274 275 |
first = page_buffers(page); local_irq_save(flags); bit_spin_lock(BH_Uptodate_Lock, &first->b_state); |
1da177e4c
|
276 277 278 279 280 281 282 283 284 285 286 287 |
clear_buffer_async_read(bh); unlock_buffer(bh); tmp = bh; do { if (!buffer_uptodate(tmp)) page_uptodate = 0; if (buffer_async_read(tmp)) { BUG_ON(!buffer_locked(tmp)); goto still_busy; } tmp = tmp->b_this_page; } while (tmp != bh); |
a39722034
|
288 289 |
bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); local_irq_restore(flags); |
1da177e4c
|
290 291 292 293 294 295 296 297 298 299 300 |
/* * If none of the buffers had errors and they are all * uptodate then we can set the page uptodate. */ if (page_uptodate && !PageError(page)) SetPageUptodate(page); unlock_page(page); return; still_busy: |
a39722034
|
301 302 |
bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); local_irq_restore(flags); |
1da177e4c
|
303 304 305 306 307 308 309 |
return; } /* * Completion handler for block_write_full_page() - pages which are unlocked * during I/O, and which have PageWriteback cleared upon I/O completion. */ |
35c80d5f4
|
310 |
void end_buffer_async_write(struct buffer_head *bh, int uptodate) |
1da177e4c
|
311 |
{ |
1da177e4c
|
312 |
unsigned long flags; |
a39722034
|
313 |
struct buffer_head *first; |
1da177e4c
|
314 315 316 317 318 319 320 321 322 |
struct buffer_head *tmp; struct page *page; BUG_ON(!buffer_async_write(bh)); page = bh->b_page; if (uptodate) { set_buffer_uptodate(bh); } else { |
432f16e64
|
323 |
buffer_io_error(bh, ", lost async page write"); |
87354e5de
|
324 |
mark_buffer_write_io_error(bh); |
1da177e4c
|
325 326 327 |
clear_buffer_uptodate(bh); SetPageError(page); } |
a39722034
|
328 329 330 |
first = page_buffers(page); local_irq_save(flags); bit_spin_lock(BH_Uptodate_Lock, &first->b_state); |
1da177e4c
|
331 332 333 334 335 336 337 338 339 340 |
clear_buffer_async_write(bh); unlock_buffer(bh); tmp = bh->b_this_page; while (tmp != bh) { if (buffer_async_write(tmp)) { BUG_ON(!buffer_locked(tmp)); goto still_busy; } tmp = tmp->b_this_page; } |
a39722034
|
341 342 |
bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); local_irq_restore(flags); |
1da177e4c
|
343 344 345 346 |
end_page_writeback(page); return; still_busy: |
a39722034
|
347 348 |
bit_spin_unlock(BH_Uptodate_Lock, &first->b_state); local_irq_restore(flags); |
1da177e4c
|
349 350 |
return; } |
1fe72eaa0
|
351 |
EXPORT_SYMBOL(end_buffer_async_write); |
1da177e4c
|
352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 |
/* * If a page's buffers are under async readin (end_buffer_async_read * completion) then there is a possibility that another thread of * control could lock one of the buffers after it has completed * but while some of the other buffers have not completed. This * locked buffer would confuse end_buffer_async_read() into not unlocking * the page. So the absence of BH_Async_Read tells end_buffer_async_read() * that this buffer is not under async I/O. * * The page comes unlocked when it has no locked buffer_async buffers * left. * * PageLocked prevents anyone starting new async I/O reads any of * the buffers. * * PageWriteback is used to prevent simultaneous writeout of the same * page. * * PageLocked prevents anyone from starting writeback of a page which is * under read I/O (PageWriteback is only ever set against a locked page). */ static void mark_buffer_async_read(struct buffer_head *bh) { bh->b_end_io = end_buffer_async_read; set_buffer_async_read(bh); } |
1fe72eaa0
|
379 380 |
static void mark_buffer_async_write_endio(struct buffer_head *bh, bh_end_io_t *handler) |
1da177e4c
|
381 |
{ |
35c80d5f4
|
382 |
bh->b_end_io = handler; |
1da177e4c
|
383 384 |
set_buffer_async_write(bh); } |
35c80d5f4
|
385 386 387 388 389 |
void mark_buffer_async_write(struct buffer_head *bh) { mark_buffer_async_write_endio(bh, end_buffer_async_write); } |
1da177e4c
|
390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 |
EXPORT_SYMBOL(mark_buffer_async_write); /* * fs/buffer.c contains helper functions for buffer-backed address space's * fsync functions. A common requirement for buffer-based filesystems is * that certain data from the backing blockdev needs to be written out for * a successful fsync(). For example, ext2 indirect blocks need to be * written back and waited upon before fsync() returns. * * The functions mark_buffer_inode_dirty(), fsync_inode_buffers(), * inode_has_buffers() and invalidate_inode_buffers() are provided for the * management of a list of dependent buffers at ->i_mapping->private_list. * * Locking is a little subtle: try_to_free_buffers() will remove buffers * from their controlling inode's queue when they are being freed. But * try_to_free_buffers() will be operating against the *blockdev* mapping * at the time, not against the S_ISREG file which depends on those buffers. * So the locking for private_list is via the private_lock in the address_space * which backs the buffers. Which is different from the address_space * against which the buffers are listed. So for a particular address_space, * mapping->private_lock does *not* protect mapping->private_list! In fact, * mapping->private_list will always be protected by the backing blockdev's * ->private_lock. * * Which introduces a requirement: all buffers on an address_space's * ->private_list must be from the same address_space: the blockdev's. * * address_spaces which do not place buffers at ->private_list via these * utility functions are free to use private_lock and private_list for * whatever they want. The only requirement is that list_empty(private_list) * be true at clear_inode() time. * * FIXME: clear_inode should not call invalidate_inode_buffers(). The * filesystems should do that. invalidate_inode_buffers() should just go * BUG_ON(!list_empty). * * FIXME: mark_buffer_dirty_inode() is a data-plane operation. It should * take an address_space, not an inode. And it should be called * mark_buffer_dirty_fsync() to clearly define why those buffers are being * queued up. * * FIXME: mark_buffer_dirty_inode() doesn't need to add the buffer to the * list if it is already on a list. Because if the buffer is on a list, * it *must* already be on the right one. If not, the filesystem is being * silly. This will save a ton of locking. But first we have to ensure * that buffers are taken *off* the old inode's list when they are freed * (presumably in truncate). That requires careful auditing of all * filesystems (do it inside bforget()). It could also be done by bringing * b_inode back. */ /* * The buffer's backing address_space's private_lock must be held */ |
dbacefc9c
|
445 |
static void __remove_assoc_queue(struct buffer_head *bh) |
1da177e4c
|
446 447 |
{ list_del_init(&bh->b_assoc_buffers); |
58ff407be
|
448 |
WARN_ON(!bh->b_assoc_map); |
58ff407be
|
449 |
bh->b_assoc_map = NULL; |
1da177e4c
|
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 |
} int inode_has_buffers(struct inode *inode) { return !list_empty(&inode->i_data.private_list); } /* * osync is designed to support O_SYNC io. It waits synchronously for * all already-submitted IO to complete, but does not queue any new * writes to the disk. * * To do O_SYNC writes, just queue the buffer writes with ll_rw_block as * you dirty the buffers, and then use osync_inode_buffers to wait for * completion. Any other dirty buffers which are not yet queued for * write will not be flushed to disk by the osync. */ static int osync_buffers_list(spinlock_t *lock, struct list_head *list) { struct buffer_head *bh; struct list_head *p; int err = 0; spin_lock(lock); repeat: list_for_each_prev(p, list) { bh = BH_ENTRY(p); if (buffer_locked(bh)) { get_bh(bh); spin_unlock(lock); wait_on_buffer(bh); if (!buffer_uptodate(bh)) err = -EIO; brelse(bh); spin_lock(lock); goto repeat; } } spin_unlock(lock); return err; } |
08fdc8a01
|
491 |
void emergency_thaw_bdev(struct super_block *sb) |
c2d754385
|
492 |
{ |
01a05b337
|
493 |
while (sb->s_bdev && !thaw_bdev(sb->s_bdev, sb)) |
a1c6f0573
|
494 495 |
printk(KERN_WARNING "Emergency Thaw on %pg ", sb->s_bdev); |
01a05b337
|
496 |
} |
c2d754385
|
497 |
|
1da177e4c
|
498 |
/** |
78a4a50a8
|
499 |
* sync_mapping_buffers - write out & wait upon a mapping's "associated" buffers |
67be2dd1b
|
500 |
* @mapping: the mapping which wants those buffers written |
1da177e4c
|
501 502 503 504 |
* * Starts I/O against the buffers at mapping->private_list, and waits upon * that I/O. * |
67be2dd1b
|
505 506 507 |
* Basically, this is a convenience function for fsync(). * @mapping is a file or directory which needs those buffers to be written for * a successful fsync(). |
1da177e4c
|
508 509 510 |
*/ int sync_mapping_buffers(struct address_space *mapping) { |
252aa6f5b
|
511 |
struct address_space *buffer_mapping = mapping->private_data; |
1da177e4c
|
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 |
if (buffer_mapping == NULL || list_empty(&mapping->private_list)) return 0; return fsync_buffers_list(&buffer_mapping->private_lock, &mapping->private_list); } EXPORT_SYMBOL(sync_mapping_buffers); /* * Called when we've recently written block `bblock', and it is known that * `bblock' was for a buffer_boundary() buffer. This means that the block at * `bblock + 1' is probably a dirty indirect block. Hunt it down and, if it's * dirty, schedule it for IO. So that indirects merge nicely with their data. */ void write_boundary_block(struct block_device *bdev, sector_t bblock, unsigned blocksize) { struct buffer_head *bh = __find_get_block(bdev, bblock + 1, blocksize); if (bh) { if (buffer_dirty(bh)) |
dfec8a14f
|
533 |
ll_rw_block(REQ_OP_WRITE, 0, 1, &bh); |
1da177e4c
|
534 535 536 537 538 539 540 541 542 543 |
put_bh(bh); } } void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode) { struct address_space *mapping = inode->i_mapping; struct address_space *buffer_mapping = bh->b_page->mapping; mark_buffer_dirty(bh); |
252aa6f5b
|
544 545 |
if (!mapping->private_data) { mapping->private_data = buffer_mapping; |
1da177e4c
|
546 |
} else { |
252aa6f5b
|
547 |
BUG_ON(mapping->private_data != buffer_mapping); |
1da177e4c
|
548 |
} |
535ee2fbf
|
549 |
if (!bh->b_assoc_map) { |
1da177e4c
|
550 551 552 |
spin_lock(&buffer_mapping->private_lock); list_move_tail(&bh->b_assoc_buffers, &mapping->private_list); |
58ff407be
|
553 |
bh->b_assoc_map = mapping; |
1da177e4c
|
554 555 556 557 558 559 |
spin_unlock(&buffer_mapping->private_lock); } } EXPORT_SYMBOL(mark_buffer_dirty_inode); /* |
ec82e1c1c
|
560 |
* Mark the page dirty, and set it dirty in the page cache, and mark the inode |
787d2214c
|
561 562 563 564 |
* dirty. * * If warn is true, then emit a warning if the page is not uptodate and has * not been truncated. |
c4843a759
|
565 |
* |
81f8c3a46
|
566 |
* The caller must hold lock_page_memcg(). |
787d2214c
|
567 |
*/ |
f82b37641
|
568 |
void __set_page_dirty(struct page *page, struct address_space *mapping, |
62cccb8c8
|
569 |
int warn) |
787d2214c
|
570 |
{ |
227d53b39
|
571 |
unsigned long flags; |
b93b01631
|
572 |
xa_lock_irqsave(&mapping->i_pages, flags); |
787d2214c
|
573 574 |
if (page->mapping) { /* Race with truncate? */ WARN_ON_ONCE(warn && !PageUptodate(page)); |
62cccb8c8
|
575 |
account_page_dirtied(page, mapping); |
ec82e1c1c
|
576 577 |
__xa_set_mark(&mapping->i_pages, page_index(page), PAGECACHE_TAG_DIRTY); |
787d2214c
|
578 |
} |
b93b01631
|
579 |
xa_unlock_irqrestore(&mapping->i_pages, flags); |
787d2214c
|
580 |
} |
f82b37641
|
581 |
EXPORT_SYMBOL_GPL(__set_page_dirty); |
787d2214c
|
582 583 |
/* |
1da177e4c
|
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 |
* Add a page to the dirty page list. * * It is a sad fact of life that this function is called from several places * deeply under spinlocking. It may not sleep. * * If the page has buffers, the uptodate buffers are set dirty, to preserve * dirty-state coherency between the page and the buffers. It the page does * not have buffers then when they are later attached they will all be set * dirty. * * The buffers are dirtied before the page is dirtied. There's a small race * window in which a writepage caller may see the page cleanness but not the * buffer dirtiness. That's fine. If this code were to set the page dirty * before the buffers, a concurrent writepage caller could clear the page dirty * bit, see a bunch of clean buffers and we'd end up with dirty buffers/clean * page on the dirty page list. * * We use private_lock to lock against try_to_free_buffers while using the * page's buffer list. Also use this to protect against clean buffers being * added to the page after it was set dirty. * * FIXME: may need to call ->reservepage here as well. That's rather up to the * address_space though. */ int __set_page_dirty_buffers(struct page *page) { |
a8e7d49aa
|
610 |
int newly_dirty; |
787d2214c
|
611 |
struct address_space *mapping = page_mapping(page); |
ebf7a227d
|
612 613 614 |
if (unlikely(!mapping)) return !TestSetPageDirty(page); |
1da177e4c
|
615 616 617 618 619 620 621 622 623 624 625 |
spin_lock(&mapping->private_lock); if (page_has_buffers(page)) { struct buffer_head *head = page_buffers(page); struct buffer_head *bh = head; do { set_buffer_dirty(bh); bh = bh->b_this_page; } while (bh != head); } |
c4843a759
|
626 |
/* |
81f8c3a46
|
627 628 |
* Lock out page->mem_cgroup migration to keep PageDirty * synchronized with per-memcg dirty page counters. |
c4843a759
|
629 |
*/ |
62cccb8c8
|
630 |
lock_page_memcg(page); |
a8e7d49aa
|
631 |
newly_dirty = !TestSetPageDirty(page); |
1da177e4c
|
632 |
spin_unlock(&mapping->private_lock); |
a8e7d49aa
|
633 |
if (newly_dirty) |
62cccb8c8
|
634 |
__set_page_dirty(page, mapping, 1); |
c4843a759
|
635 |
|
62cccb8c8
|
636 |
unlock_page_memcg(page); |
c4843a759
|
637 638 639 |
if (newly_dirty) __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); |
a8e7d49aa
|
640 |
return newly_dirty; |
1da177e4c
|
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 |
} EXPORT_SYMBOL(__set_page_dirty_buffers); /* * Write out and wait upon a list of buffers. * * We have conflicting pressures: we want to make sure that all * initially dirty buffers get waited on, but that any subsequently * dirtied buffers don't. After all, we don't want fsync to last * forever if somebody is actively writing to the file. * * Do this in two main stages: first we copy dirty buffers to a * temporary inode list, queueing the writes as we go. Then we clean * up, waiting for those writes to complete. * * During this second stage, any subsequent updates to the file may end * up refiling the buffer on the original inode's dirty list again, so * there is a chance we will end up with a buffer queued for write but * not yet completed on that list. So, as a final cleanup we go through * the osync code to catch these locked, dirty buffers without requeuing * any newly dirty buffers for write. */ static int fsync_buffers_list(spinlock_t *lock, struct list_head *list) { struct buffer_head *bh; struct list_head tmp; |
7eaceacca
|
667 |
struct address_space *mapping; |
1da177e4c
|
668 |
int err = 0, err2; |
4ee2491ed
|
669 |
struct blk_plug plug; |
1da177e4c
|
670 671 |
INIT_LIST_HEAD(&tmp); |
4ee2491ed
|
672 |
blk_start_plug(&plug); |
1da177e4c
|
673 674 675 676 |
spin_lock(lock); while (!list_empty(list)) { bh = BH_ENTRY(list->next); |
535ee2fbf
|
677 |
mapping = bh->b_assoc_map; |
58ff407be
|
678 |
__remove_assoc_queue(bh); |
535ee2fbf
|
679 680 681 |
/* Avoid race with mark_buffer_dirty_inode() which does * a lockless check and we rely on seeing the dirty bit */ smp_mb(); |
1da177e4c
|
682 683 |
if (buffer_dirty(bh) || buffer_locked(bh)) { list_add(&bh->b_assoc_buffers, &tmp); |
535ee2fbf
|
684 |
bh->b_assoc_map = mapping; |
1da177e4c
|
685 686 687 688 689 |
if (buffer_dirty(bh)) { get_bh(bh); spin_unlock(lock); /* * Ensure any pending I/O completes so that |
9cb569d60
|
690 691 692 693 |
* write_dirty_buffer() actually writes the * current contents - it is a noop if I/O is * still in flight on potentially older * contents. |
1da177e4c
|
694 |
*/ |
70fd76140
|
695 |
write_dirty_buffer(bh, REQ_SYNC); |
9cf6b720f
|
696 697 698 699 700 701 702 |
/* * Kick off IO for the previous mapping. Note * that we will not run the very last mapping, * wait_on_buffer() will do that for us * through sync_buffer(). */ |
1da177e4c
|
703 704 705 706 707 |
brelse(bh); spin_lock(lock); } } } |
4ee2491ed
|
708 709 710 |
spin_unlock(lock); blk_finish_plug(&plug); spin_lock(lock); |
1da177e4c
|
711 712 |
while (!list_empty(&tmp)) { bh = BH_ENTRY(tmp.prev); |
1da177e4c
|
713 |
get_bh(bh); |
535ee2fbf
|
714 715 716 717 718 719 720 |
mapping = bh->b_assoc_map; __remove_assoc_queue(bh); /* Avoid race with mark_buffer_dirty_inode() which does * a lockless check and we rely on seeing the dirty bit */ smp_mb(); if (buffer_dirty(bh)) { list_add(&bh->b_assoc_buffers, |
e3892296d
|
721 |
&mapping->private_list); |
535ee2fbf
|
722 723 |
bh->b_assoc_map = mapping; } |
1da177e4c
|
724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 |
spin_unlock(lock); wait_on_buffer(bh); if (!buffer_uptodate(bh)) err = -EIO; brelse(bh); spin_lock(lock); } spin_unlock(lock); err2 = osync_buffers_list(lock, list); if (err) return err; else return err2; } /* * Invalidate any and all dirty buffers on a given inode. We are * probably unmounting the fs, but that doesn't mean we have already * done a sync(). Just drop the buffers from the inode list. * * NOTE: we take the inode's blockdev's mapping's private_lock. Which * assumes that all the buffers are against the blockdev. Not true * for reiserfs. */ void invalidate_inode_buffers(struct inode *inode) { if (inode_has_buffers(inode)) { struct address_space *mapping = &inode->i_data; struct list_head *list = &mapping->private_list; |
252aa6f5b
|
754 |
struct address_space *buffer_mapping = mapping->private_data; |
1da177e4c
|
755 756 757 758 759 760 761 |
spin_lock(&buffer_mapping->private_lock); while (!list_empty(list)) __remove_assoc_queue(BH_ENTRY(list->next)); spin_unlock(&buffer_mapping->private_lock); } } |
52b19ac99
|
762 |
EXPORT_SYMBOL(invalidate_inode_buffers); |
1da177e4c
|
763 764 765 766 767 768 769 770 771 772 773 774 775 776 |
/* * Remove any clean buffers from the inode's buffer list. This is called * when we're trying to free the inode itself. Those buffers can pin it. * * Returns true if all buffers were removed. */ int remove_inode_buffers(struct inode *inode) { int ret = 1; if (inode_has_buffers(inode)) { struct address_space *mapping = &inode->i_data; struct list_head *list = &mapping->private_list; |
252aa6f5b
|
777 |
struct address_space *buffer_mapping = mapping->private_data; |
1da177e4c
|
778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 |
spin_lock(&buffer_mapping->private_lock); while (!list_empty(list)) { struct buffer_head *bh = BH_ENTRY(list->next); if (buffer_dirty(bh)) { ret = 0; break; } __remove_assoc_queue(bh); } spin_unlock(&buffer_mapping->private_lock); } return ret; } /* * Create the appropriate buffers when given a page for data area and * the size of each buffer.. Use the bh->b_this_page linked list to * follow the buffers created. Return NULL if unable to create more * buffers. * * The retry flag is used to differentiate async IO (paging, swapping) * which may not fail from ordinary buffer allocations. */ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size, |
640ab98fb
|
803 |
bool retry) |
1da177e4c
|
804 805 |
{ struct buffer_head *bh, *head; |
f745c6f5f
|
806 |
gfp_t gfp = GFP_NOFS | __GFP_ACCOUNT; |
1da177e4c
|
807 |
long offset; |
f745c6f5f
|
808 |
struct mem_cgroup *memcg; |
1da177e4c
|
809 |
|
640ab98fb
|
810 811 |
if (retry) gfp |= __GFP_NOFAIL; |
f745c6f5f
|
812 813 |
memcg = get_mem_cgroup_from_page(page); memalloc_use_memcg(memcg); |
1da177e4c
|
814 815 816 |
head = NULL; offset = PAGE_SIZE; while ((offset -= size) >= 0) { |
640ab98fb
|
817 |
bh = alloc_buffer_head(gfp); |
1da177e4c
|
818 819 |
if (!bh) goto no_grow; |
1da177e4c
|
820 821 822 |
bh->b_this_page = head; bh->b_blocknr = -1; head = bh; |
1da177e4c
|
823 824 825 826 |
bh->b_size = size; /* Link the buffer to its page */ set_bh_page(bh, page, offset); |
1da177e4c
|
827 |
} |
f745c6f5f
|
828 829 830 |
out: memalloc_unuse_memcg(); mem_cgroup_put(memcg); |
1da177e4c
|
831 832 833 834 835 836 837 838 839 840 841 842 |
return head; /* * In case anything failed, we just free everything we got. */ no_grow: if (head) { do { bh = head; head = head->b_this_page; free_buffer_head(bh); } while (head); } |
f745c6f5f
|
843 |
goto out; |
1da177e4c
|
844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 |
} EXPORT_SYMBOL_GPL(alloc_page_buffers); static inline void link_dev_buffers(struct page *page, struct buffer_head *head) { struct buffer_head *bh, *tail; bh = head; do { tail = bh; bh = bh->b_this_page; } while (bh); tail->b_this_page = head; attach_page_buffers(page, head); } |
bbec0270b
|
860 861 862 863 864 865 866 867 868 869 870 |
static sector_t blkdev_max_block(struct block_device *bdev, unsigned int size) { sector_t retval = ~((sector_t)0); loff_t sz = i_size_read(bdev->bd_inode); if (sz) { unsigned int sizebits = blksize_bits(size); retval = (sz >> sizebits); } return retval; } |
1da177e4c
|
871 872 873 |
/* * Initialise the state of a blockdev page's buffers. */ |
676ce6d5c
|
874 |
static sector_t |
1da177e4c
|
875 876 877 878 879 880 |
init_page_buffers(struct page *page, struct block_device *bdev, sector_t block, int size) { struct buffer_head *head = page_buffers(page); struct buffer_head *bh = head; int uptodate = PageUptodate(page); |
bbec0270b
|
881 |
sector_t end_block = blkdev_max_block(I_BDEV(bdev->bd_inode), size); |
1da177e4c
|
882 883 884 |
do { if (!buffer_mapped(bh)) { |
01950a349
|
885 886 |
bh->b_end_io = NULL; bh->b_private = NULL; |
1da177e4c
|
887 888 889 890 |
bh->b_bdev = bdev; bh->b_blocknr = block; if (uptodate) set_buffer_uptodate(bh); |
080399aaa
|
891 892 |
if (block < end_block) set_buffer_mapped(bh); |
1da177e4c
|
893 894 895 896 |
} block++; bh = bh->b_this_page; } while (bh != head); |
676ce6d5c
|
897 898 899 900 901 |
/* * Caller needs to validate requested block against end of device. */ return end_block; |
1da177e4c
|
902 903 904 905 906 |
} /* * Create the page-cache page that contains the requested block. * |
676ce6d5c
|
907 |
* This is used purely for blockdev mappings. |
1da177e4c
|
908 |
*/ |
676ce6d5c
|
909 |
static int |
1da177e4c
|
910 |
grow_dev_page(struct block_device *bdev, sector_t block, |
3b5e6454a
|
911 |
pgoff_t index, int size, int sizebits, gfp_t gfp) |
1da177e4c
|
912 913 914 915 |
{ struct inode *inode = bdev->bd_inode; struct page *page; struct buffer_head *bh; |
676ce6d5c
|
916 917 |
sector_t end_block; int ret = 0; /* Will call free_more_memory() */ |
84235de39
|
918 |
gfp_t gfp_mask; |
1da177e4c
|
919 |
|
c62d25556
|
920 |
gfp_mask = mapping_gfp_constraint(inode->i_mapping, ~__GFP_FS) | gfp; |
3b5e6454a
|
921 |
|
84235de39
|
922 923 924 925 926 927 928 929 930 |
/* * XXX: __getblk_slow() can not really deal with failure and * will endlessly loop on improvised global reclaim. Prefer * looping in the allocator rather than here, at least that * code knows what it's doing. */ gfp_mask |= __GFP_NOFAIL; page = find_or_create_page(inode->i_mapping, index, gfp_mask); |
1da177e4c
|
931 |
|
e827f9235
|
932 |
BUG_ON(!PageLocked(page)); |
1da177e4c
|
933 934 935 936 |
if (page_has_buffers(page)) { bh = page_buffers(page); if (bh->b_size == size) { |
676ce6d5c
|
937 |
end_block = init_page_buffers(page, bdev, |
f2d5a9443
|
938 939 |
(sector_t)index << sizebits, size); |
676ce6d5c
|
940 |
goto done; |
1da177e4c
|
941 942 943 944 945 946 947 948 |
} if (!try_to_free_buffers(page)) goto failed; } /* * Allocate some buffers for this page */ |
94dc24c0c
|
949 |
bh = alloc_page_buffers(page, size, true); |
1da177e4c
|
950 951 952 953 954 955 956 957 |
/* * Link the page to the buffers and initialise them. Take the * lock to be atomic wrt __find_get_block(), which does not * run under the page lock. */ spin_lock(&inode->i_mapping->private_lock); link_dev_buffers(page, bh); |
f2d5a9443
|
958 959 |
end_block = init_page_buffers(page, bdev, (sector_t)index << sizebits, size); |
1da177e4c
|
960 |
spin_unlock(&inode->i_mapping->private_lock); |
676ce6d5c
|
961 962 |
done: ret = (block < end_block) ? 1 : -ENXIO; |
1da177e4c
|
963 |
failed: |
1da177e4c
|
964 |
unlock_page(page); |
09cbfeaf1
|
965 |
put_page(page); |
676ce6d5c
|
966 |
return ret; |
1da177e4c
|
967 968 969 970 971 |
} /* * Create buffers for the specified block device block's page. If * that page was dirty, the buffers are set dirty also. |
1da177e4c
|
972 |
*/ |
858119e15
|
973 |
static int |
3b5e6454a
|
974 |
grow_buffers(struct block_device *bdev, sector_t block, int size, gfp_t gfp) |
1da177e4c
|
975 |
{ |
1da177e4c
|
976 977 978 979 980 981 982 983 984 |
pgoff_t index; int sizebits; sizebits = -1; do { sizebits++; } while ((size << sizebits) < PAGE_SIZE); index = block >> sizebits; |
1da177e4c
|
985 |
|
e56579338
|
986 987 988 989 990 |
/* * Check for a block which wants to lie outside our maximum possible * pagecache index. (this comparison is done using sector_t types). */ if (unlikely(index != block >> sizebits)) { |
e56579338
|
991 |
printk(KERN_ERR "%s: requested out-of-range block %llu for " |
a1c6f0573
|
992 993 |
"device %pg ", |
8e24eea72
|
994 |
__func__, (unsigned long long)block, |
a1c6f0573
|
995 |
bdev); |
e56579338
|
996 997 |
return -EIO; } |
676ce6d5c
|
998 |
|
1da177e4c
|
999 |
/* Create a page with the proper size buffers.. */ |
3b5e6454a
|
1000 |
return grow_dev_page(bdev, block, index, size, sizebits, gfp); |
1da177e4c
|
1001 |
} |
0026ba400
|
1002 |
static struct buffer_head * |
3b5e6454a
|
1003 1004 |
__getblk_slow(struct block_device *bdev, sector_t block, unsigned size, gfp_t gfp) |
1da177e4c
|
1005 1006 |
{ /* Size must be multiple of hard sectorsize */ |
e1defc4ff
|
1007 |
if (unlikely(size & (bdev_logical_block_size(bdev)-1) || |
1da177e4c
|
1008 1009 1010 1011 |
(size < 512 || size > PAGE_SIZE))) { printk(KERN_ERR "getblk(): invalid block size %d requested ", size); |
e1defc4ff
|
1012 1013 1014 |
printk(KERN_ERR "logical block size: %d ", bdev_logical_block_size(bdev)); |
1da177e4c
|
1015 1016 1017 1018 |
dump_stack(); return NULL; } |
676ce6d5c
|
1019 1020 1021 |
for (;;) { struct buffer_head *bh; int ret; |
1da177e4c
|
1022 1023 1024 1025 |
bh = __find_get_block(bdev, block, size); if (bh) return bh; |
676ce6d5c
|
1026 |
|
3b5e6454a
|
1027 |
ret = grow_buffers(bdev, block, size, gfp); |
676ce6d5c
|
1028 1029 |
if (ret < 0) return NULL; |
1da177e4c
|
1030 1031 1032 1033 1034 1035 1036 |
} } /* * The relationship between dirty buffers and dirty pages: * * Whenever a page has any dirty buffers, the page's dirty bit is set, and |
ec82e1c1c
|
1037 |
* the page is tagged dirty in the page cache. |
1da177e4c
|
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 |
* * At all times, the dirtiness of the buffers represents the dirtiness of * subsections of the page. If the page has buffers, the page dirty bit is * merely a hint about the true dirty state. * * When a page is set dirty in its entirety, all its buffers are marked dirty * (if the page has buffers). * * When a buffer is marked dirty, its page is dirtied, but the page's other * buffers are not. * * Also. When blockdev buffers are explicitly read with bread(), they * individually become uptodate. But their backing page remains not * uptodate - even if all of its buffers are uptodate. A subsequent * block_read_full_page() against that page will discover all the uptodate * buffers, will set the page uptodate and will perform no I/O. */ /** * mark_buffer_dirty - mark a buffer_head as needing writeout |
67be2dd1b
|
1058 |
* @bh: the buffer_head to mark dirty |
1da177e4c
|
1059 |
* |
ec82e1c1c
|
1060 1061 1062 |
* mark_buffer_dirty() will set the dirty bit against the buffer, then set * its backing page dirty, then tag the page as dirty in the page cache * and then attach the address_space's inode to its superblock's dirty |
1da177e4c
|
1063 1064 1065 |
* inode list. * * mark_buffer_dirty() is atomic. It takes bh->b_page->mapping->private_lock, |
b93b01631
|
1066 |
* i_pages lock and mapping->host->i_lock. |
1da177e4c
|
1067 |
*/ |
fc9b52cd8
|
1068 |
void mark_buffer_dirty(struct buffer_head *bh) |
1da177e4c
|
1069 |
{ |
787d2214c
|
1070 |
WARN_ON_ONCE(!buffer_uptodate(bh)); |
1be62dc19
|
1071 |
|
5305cb830
|
1072 |
trace_block_dirty_buffer(bh); |
1be62dc19
|
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 |
/* * Very *carefully* optimize the it-is-already-dirty case. * * Don't let the final "is it dirty" escape to before we * perhaps modified the buffer. */ if (buffer_dirty(bh)) { smp_mb(); if (buffer_dirty(bh)) return; } |
a8e7d49aa
|
1084 1085 |
if (!test_set_buffer_dirty(bh)) { struct page *page = bh->b_page; |
c4843a759
|
1086 |
struct address_space *mapping = NULL; |
c4843a759
|
1087 |
|
62cccb8c8
|
1088 |
lock_page_memcg(page); |
8e9d78ede
|
1089 |
if (!TestSetPageDirty(page)) { |
c4843a759
|
1090 |
mapping = page_mapping(page); |
8e9d78ede
|
1091 |
if (mapping) |
62cccb8c8
|
1092 |
__set_page_dirty(page, mapping, 0); |
8e9d78ede
|
1093 |
} |
62cccb8c8
|
1094 |
unlock_page_memcg(page); |
c4843a759
|
1095 1096 |
if (mapping) __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); |
a8e7d49aa
|
1097 |
} |
1da177e4c
|
1098 |
} |
1fe72eaa0
|
1099 |
EXPORT_SYMBOL(mark_buffer_dirty); |
1da177e4c
|
1100 |
|
87354e5de
|
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 |
void mark_buffer_write_io_error(struct buffer_head *bh) { set_buffer_write_io_error(bh); /* FIXME: do we need to set this in both places? */ if (bh->b_page && bh->b_page->mapping) mapping_set_error(bh->b_page->mapping, -EIO); if (bh->b_assoc_map) mapping_set_error(bh->b_assoc_map, -EIO); } EXPORT_SYMBOL(mark_buffer_write_io_error); |
1da177e4c
|
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 |
/* * Decrement a buffer_head's reference count. If all buffers against a page * have zero reference count, are clean and unlocked, and if the page is clean * and unlocked then try_to_free_buffers() may strip the buffers from the page * in preparation for freeing it (sometimes, rarely, buffers are removed from * a page but it ends up not being freed, and buffers may later be reattached). */ void __brelse(struct buffer_head * buf) { if (atomic_read(&buf->b_count)) { put_bh(buf); return; } |
5c752ad9f
|
1124 1125 |
WARN(1, KERN_ERR "VFS: brelse: Trying to free free buffer "); |
1da177e4c
|
1126 |
} |
1fe72eaa0
|
1127 |
EXPORT_SYMBOL(__brelse); |
1da177e4c
|
1128 1129 1130 1131 1132 1133 1134 1135 |
/* * bforget() is like brelse(), except it discards any * potentially dirty data. */ void __bforget(struct buffer_head *bh) { clear_buffer_dirty(bh); |
535ee2fbf
|
1136 |
if (bh->b_assoc_map) { |
1da177e4c
|
1137 1138 1139 1140 |
struct address_space *buffer_mapping = bh->b_page->mapping; spin_lock(&buffer_mapping->private_lock); list_del_init(&bh->b_assoc_buffers); |
58ff407be
|
1141 |
bh->b_assoc_map = NULL; |
1da177e4c
|
1142 1143 1144 1145 |
spin_unlock(&buffer_mapping->private_lock); } __brelse(bh); } |
1fe72eaa0
|
1146 |
EXPORT_SYMBOL(__bforget); |
1da177e4c
|
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 |
static struct buffer_head *__bread_slow(struct buffer_head *bh) { lock_buffer(bh); if (buffer_uptodate(bh)) { unlock_buffer(bh); return bh; } else { get_bh(bh); bh->b_end_io = end_buffer_read_sync; |
2a222ca99
|
1157 |
submit_bh(REQ_OP_READ, 0, bh); |
1da177e4c
|
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 |
wait_on_buffer(bh); if (buffer_uptodate(bh)) return bh; } brelse(bh); return NULL; } /* * Per-cpu buffer LRU implementation. To reduce the cost of __find_get_block(). * The bhs[] array is sorted - newest buffer is at bhs[0]. Buffers have their * refcount elevated by one when they're in an LRU. A buffer can only appear * once in a particular CPU's LRU. A single buffer can be present in multiple * CPU's LRUs at the same time. * * This is a transparent caching front-end to sb_bread(), sb_getblk() and * sb_find_get_block(). * * The LRUs themselves only need locking against invalidate_bh_lrus. We use * a local interrupt disable for that. */ |
86cf78d73
|
1179 |
#define BH_LRU_SIZE 16 |
1da177e4c
|
1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 |
struct bh_lru { struct buffer_head *bhs[BH_LRU_SIZE]; }; static DEFINE_PER_CPU(struct bh_lru, bh_lrus) = {{ NULL }}; #ifdef CONFIG_SMP #define bh_lru_lock() local_irq_disable() #define bh_lru_unlock() local_irq_enable() #else #define bh_lru_lock() preempt_disable() #define bh_lru_unlock() preempt_enable() #endif static inline void check_irqs_on(void) { #ifdef irqs_disabled BUG_ON(irqs_disabled()); #endif } /* |
241f01fbe
|
1203 1204 1205 |
* Install a buffer_head into this cpu's LRU. If not already in the LRU, it is * inserted at the front, and the buffer_head at the back if any is evicted. * Or, if already in the LRU it is moved to the front. |
1da177e4c
|
1206 1207 1208 |
*/ static void bh_lru_install(struct buffer_head *bh) { |
241f01fbe
|
1209 1210 1211 |
struct buffer_head *evictee = bh; struct bh_lru *b; int i; |
1da177e4c
|
1212 1213 1214 |
check_irqs_on(); bh_lru_lock(); |
1da177e4c
|
1215 |
|
241f01fbe
|
1216 1217 1218 1219 1220 1221 |
b = this_cpu_ptr(&bh_lrus); for (i = 0; i < BH_LRU_SIZE; i++) { swap(evictee, b->bhs[i]); if (evictee == bh) { bh_lru_unlock(); return; |
1da177e4c
|
1222 |
} |
1da177e4c
|
1223 |
} |
1da177e4c
|
1224 |
|
241f01fbe
|
1225 1226 1227 |
get_bh(bh); bh_lru_unlock(); brelse(evictee); |
1da177e4c
|
1228 1229 1230 1231 1232 |
} /* * Look up the bh in this cpu's LRU. If it's there, move it to the head. */ |
858119e15
|
1233 |
static struct buffer_head * |
3991d3bd1
|
1234 |
lookup_bh_lru(struct block_device *bdev, sector_t block, unsigned size) |
1da177e4c
|
1235 1236 |
{ struct buffer_head *ret = NULL; |
3991d3bd1
|
1237 |
unsigned int i; |
1da177e4c
|
1238 1239 1240 |
check_irqs_on(); bh_lru_lock(); |
1da177e4c
|
1241 |
for (i = 0; i < BH_LRU_SIZE; i++) { |
c7b92516a
|
1242 |
struct buffer_head *bh = __this_cpu_read(bh_lrus.bhs[i]); |
1da177e4c
|
1243 |
|
9470dd5d3
|
1244 1245 |
if (bh && bh->b_blocknr == block && bh->b_bdev == bdev && bh->b_size == size) { |
1da177e4c
|
1246 1247 |
if (i) { while (i) { |
c7b92516a
|
1248 1249 |
__this_cpu_write(bh_lrus.bhs[i], __this_cpu_read(bh_lrus.bhs[i - 1])); |
1da177e4c
|
1250 1251 |
i--; } |
c7b92516a
|
1252 |
__this_cpu_write(bh_lrus.bhs[0], bh); |
1da177e4c
|
1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 |
} get_bh(bh); ret = bh; break; } } bh_lru_unlock(); return ret; } /* * Perform a pagecache lookup for the matching buffer. If it's there, refresh * it in the LRU and mark it as accessed. If it is not present then return * NULL */ struct buffer_head * |
3991d3bd1
|
1269 |
__find_get_block(struct block_device *bdev, sector_t block, unsigned size) |
1da177e4c
|
1270 1271 1272 1273 |
{ struct buffer_head *bh = lookup_bh_lru(bdev, block, size); if (bh == NULL) { |
2457aec63
|
1274 |
/* __find_get_block_slow will mark the page accessed */ |
385fd4c59
|
1275 |
bh = __find_get_block_slow(bdev, block); |
1da177e4c
|
1276 1277 |
if (bh) bh_lru_install(bh); |
2457aec63
|
1278 |
} else |
1da177e4c
|
1279 |
touch_buffer(bh); |
2457aec63
|
1280 |
|
1da177e4c
|
1281 1282 1283 1284 1285 |
return bh; } EXPORT_SYMBOL(__find_get_block); /* |
3b5e6454a
|
1286 |
* __getblk_gfp() will locate (and, if necessary, create) the buffer_head |
1da177e4c
|
1287 1288 1289 |
* which corresponds to the passed block_device, block and size. The * returned buffer has its reference count incremented. * |
3b5e6454a
|
1290 1291 |
* __getblk_gfp() will lock up the machine if grow_dev_page's * try_to_free_buffers() attempt is failing. FIXME, perhaps? |
1da177e4c
|
1292 1293 |
*/ struct buffer_head * |
3b5e6454a
|
1294 1295 |
__getblk_gfp(struct block_device *bdev, sector_t block, unsigned size, gfp_t gfp) |
1da177e4c
|
1296 1297 1298 1299 1300 |
{ struct buffer_head *bh = __find_get_block(bdev, block, size); might_sleep(); if (bh == NULL) |
3b5e6454a
|
1301 |
bh = __getblk_slow(bdev, block, size, gfp); |
1da177e4c
|
1302 1303 |
return bh; } |
3b5e6454a
|
1304 |
EXPORT_SYMBOL(__getblk_gfp); |
1da177e4c
|
1305 1306 1307 1308 |
/* * Do async read-ahead on a buffer.. */ |
3991d3bd1
|
1309 |
void __breadahead(struct block_device *bdev, sector_t block, unsigned size) |
1da177e4c
|
1310 1311 |
{ struct buffer_head *bh = __getblk(bdev, block, size); |
a3e713b5f
|
1312 |
if (likely(bh)) { |
70246286e
|
1313 |
ll_rw_block(REQ_OP_READ, REQ_RAHEAD, 1, &bh); |
a3e713b5f
|
1314 1315 |
brelse(bh); } |
1da177e4c
|
1316 1317 1318 1319 |
} EXPORT_SYMBOL(__breadahead); /** |
3b5e6454a
|
1320 |
* __bread_gfp() - reads a specified block and returns the bh |
67be2dd1b
|
1321 |
* @bdev: the block_device to read from |
1da177e4c
|
1322 1323 |
* @block: number of block * @size: size (in bytes) to read |
3b5e6454a
|
1324 1325 |
* @gfp: page allocation flag * |
1da177e4c
|
1326 |
* Reads a specified block, and returns buffer head that contains it. |
3b5e6454a
|
1327 1328 |
* The page cache can be allocated from non-movable area * not to prevent page migration if you set gfp to zero. |
1da177e4c
|
1329 1330 1331 |
* It returns NULL if the block was unreadable. */ struct buffer_head * |
3b5e6454a
|
1332 1333 |
__bread_gfp(struct block_device *bdev, sector_t block, unsigned size, gfp_t gfp) |
1da177e4c
|
1334 |
{ |
3b5e6454a
|
1335 |
struct buffer_head *bh = __getblk_gfp(bdev, block, size, gfp); |
1da177e4c
|
1336 |
|
a3e713b5f
|
1337 |
if (likely(bh) && !buffer_uptodate(bh)) |
1da177e4c
|
1338 1339 1340 |
bh = __bread_slow(bh); return bh; } |
3b5e6454a
|
1341 |
EXPORT_SYMBOL(__bread_gfp); |
1da177e4c
|
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 |
/* * invalidate_bh_lrus() is called rarely - but not only at unmount. * This doesn't race because it runs in each cpu either in irq * or with preempt disabled. */ static void invalidate_bh_lru(void *arg) { struct bh_lru *b = &get_cpu_var(bh_lrus); int i; for (i = 0; i < BH_LRU_SIZE; i++) { brelse(b->bhs[i]); b->bhs[i] = NULL; } put_cpu_var(bh_lrus); } |
42be35d03
|
1359 1360 1361 1362 1363 |
static bool has_bh_in_lru(int cpu, void *dummy) { struct bh_lru *b = per_cpu_ptr(&bh_lrus, cpu); int i; |
1da177e4c
|
1364 |
|
42be35d03
|
1365 1366 1367 1368 1369 1370 1371 |
for (i = 0; i < BH_LRU_SIZE; i++) { if (b->bhs[i]) return 1; } return 0; } |
f9a14399a
|
1372 |
void invalidate_bh_lrus(void) |
1da177e4c
|
1373 |
{ |
42be35d03
|
1374 |
on_each_cpu_cond(has_bh_in_lru, invalidate_bh_lru, NULL, 1, GFP_KERNEL); |
1da177e4c
|
1375 |
} |
9db5579be
|
1376 |
EXPORT_SYMBOL_GPL(invalidate_bh_lrus); |
1da177e4c
|
1377 1378 1379 1380 1381 |
void set_bh_page(struct buffer_head *bh, struct page *page, unsigned long offset) { bh->b_page = page; |
e827f9235
|
1382 |
BUG_ON(offset >= PAGE_SIZE); |
1da177e4c
|
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 |
if (PageHighMem(page)) /* * This catches illegal uses and preserves the offset: */ bh->b_data = (char *)(0 + offset); else bh->b_data = page_address(page) + offset; } EXPORT_SYMBOL(set_bh_page); /* * Called when truncating a buffer on a page completely. */ |
e7470ee89
|
1396 1397 1398 1399 1400 |
/* Bits that are cleared during an invalidate */ #define BUFFER_FLAGS_DISCARD \ (1 << BH_Mapped | 1 << BH_New | 1 << BH_Req | \ 1 << BH_Delay | 1 << BH_Unwritten) |
858119e15
|
1401 |
static void discard_buffer(struct buffer_head * bh) |
1da177e4c
|
1402 |
{ |
e7470ee89
|
1403 |
unsigned long b_state, b_state_old; |
1da177e4c
|
1404 1405 1406 |
lock_buffer(bh); clear_buffer_dirty(bh); bh->b_bdev = NULL; |
e7470ee89
|
1407 1408 1409 1410 1411 1412 1413 1414 |
b_state = bh->b_state; for (;;) { b_state_old = cmpxchg(&bh->b_state, b_state, (b_state & ~BUFFER_FLAGS_DISCARD)); if (b_state_old == b_state) break; b_state = b_state_old; } |
1da177e4c
|
1415 1416 1417 1418 |
unlock_buffer(bh); } /** |
814e1d25a
|
1419 |
* block_invalidatepage - invalidate part or all of a buffer-backed page |
1da177e4c
|
1420 1421 |
* * @page: the page which is affected |
d47992f86
|
1422 1423 |
* @offset: start of the range to invalidate * @length: length of the range to invalidate |
1da177e4c
|
1424 1425 |
* * block_invalidatepage() is called when all or part of the page has become |
814e1d25a
|
1426 |
* invalidated by a truncate operation. |
1da177e4c
|
1427 1428 1429 1430 1431 1432 1433 |
* * block_invalidatepage() does not have to release all buffers, but it must * ensure that no dirty buffer is left outside @offset and that no I/O * is underway against any of the blocks which are outside the truncation * point. Because the caller is about to free (and possibly reuse) those * blocks on-disk. */ |
d47992f86
|
1434 1435 |
void block_invalidatepage(struct page *page, unsigned int offset, unsigned int length) |
1da177e4c
|
1436 1437 1438 |
{ struct buffer_head *head, *bh, *next; unsigned int curr_off = 0; |
d47992f86
|
1439 |
unsigned int stop = length + offset; |
1da177e4c
|
1440 1441 1442 1443 |
BUG_ON(!PageLocked(page)); if (!page_has_buffers(page)) goto out; |
d47992f86
|
1444 1445 1446 |
/* * Check for overflow */ |
09cbfeaf1
|
1447 |
BUG_ON(stop > PAGE_SIZE || stop < length); |
d47992f86
|
1448 |
|
1da177e4c
|
1449 1450 1451 1452 1453 1454 1455 |
head = page_buffers(page); bh = head; do { unsigned int next_off = curr_off + bh->b_size; next = bh->b_this_page; /* |
d47992f86
|
1456 1457 1458 1459 1460 1461 |
* Are we still fully in range ? */ if (next_off > stop) goto out; /* |
1da177e4c
|
1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 |
* is this block fully invalidated? */ if (offset <= curr_off) discard_buffer(bh); curr_off = next_off; bh = next; } while (bh != head); /* * We release buffers only if the entire page is being invalidated. * The get_block cached value has been unconditionally invalidated, * so real IO is not possible anymore. */ |
3172485f4
|
1475 |
if (length == PAGE_SIZE) |
2ff28e22b
|
1476 |
try_to_release_page(page, 0); |
1da177e4c
|
1477 |
out: |
2ff28e22b
|
1478 |
return; |
1da177e4c
|
1479 1480 |
} EXPORT_SYMBOL(block_invalidatepage); |
d47992f86
|
1481 |
|
1da177e4c
|
1482 1483 1484 1485 1486 1487 1488 1489 1490 |
/* * We attach and possibly dirty the buffers atomically wrt * __set_page_dirty_buffers() via private_lock. try_to_free_buffers * is already excluded via the page lock. */ void create_empty_buffers(struct page *page, unsigned long blocksize, unsigned long b_state) { struct buffer_head *bh, *head, *tail; |
640ab98fb
|
1491 |
head = alloc_page_buffers(page, blocksize, true); |
1da177e4c
|
1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 |
bh = head; do { bh->b_state |= b_state; tail = bh; bh = bh->b_this_page; } while (bh); tail->b_this_page = head; spin_lock(&page->mapping->private_lock); if (PageUptodate(page) || PageDirty(page)) { bh = head; do { if (PageDirty(page)) set_buffer_dirty(bh); if (PageUptodate(page)) set_buffer_uptodate(bh); bh = bh->b_this_page; } while (bh != head); } attach_page_buffers(page, head); spin_unlock(&page->mapping->private_lock); } EXPORT_SYMBOL(create_empty_buffers); |
29f3ad7d8
|
1515 1516 1517 1518 1519 |
/** * clean_bdev_aliases: clean a range of buffers in block device * @bdev: Block device to clean buffers in * @block: Start of a range of blocks to clean * @len: Number of blocks to clean |
1da177e4c
|
1520 |
* |
29f3ad7d8
|
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 |
* We are taking a range of blocks for data and we don't want writeback of any * buffer-cache aliases starting from return from this function and until the * moment when something will explicitly mark the buffer dirty (hopefully that * will not happen until we will free that block ;-) We don't even need to mark * it not-uptodate - nobody can expect anything from a newly allocated buffer * anyway. We used to use unmap_buffer() for such invalidation, but that was * wrong. We definitely don't want to mark the alias unmapped, for example - it * would confuse anyone who might pick it with bread() afterwards... * * Also.. Note that bforget() doesn't lock the buffer. So there can be * writeout I/O going on against recently-freed buffers. We don't wait on that * I/O in bforget() - it's more efficient to wait on the I/O only if we really * need to. That happens here. |
1da177e4c
|
1534 |
*/ |
29f3ad7d8
|
1535 |
void clean_bdev_aliases(struct block_device *bdev, sector_t block, sector_t len) |
1da177e4c
|
1536 |
{ |
29f3ad7d8
|
1537 1538 1539 1540 1541 |
struct inode *bd_inode = bdev->bd_inode; struct address_space *bd_mapping = bd_inode->i_mapping; struct pagevec pvec; pgoff_t index = block >> (PAGE_SHIFT - bd_inode->i_blkbits); pgoff_t end; |
c10f778dd
|
1542 |
int i, count; |
29f3ad7d8
|
1543 1544 |
struct buffer_head *bh; struct buffer_head *head; |
1da177e4c
|
1545 |
|
29f3ad7d8
|
1546 |
end = (block + len - 1) >> (PAGE_SHIFT - bd_inode->i_blkbits); |
866798201
|
1547 |
pagevec_init(&pvec); |
397162ffa
|
1548 |
while (pagevec_lookup_range(&pvec, bd_mapping, &index, end)) { |
c10f778dd
|
1549 1550 |
count = pagevec_count(&pvec); for (i = 0; i < count; i++) { |
29f3ad7d8
|
1551 |
struct page *page = pvec.pages[i]; |
1da177e4c
|
1552 |
|
29f3ad7d8
|
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 |
if (!page_has_buffers(page)) continue; /* * We use page lock instead of bd_mapping->private_lock * to pin buffers here since we can afford to sleep and * it scales better than a global spinlock lock. */ lock_page(page); /* Recheck when the page is locked which pins bhs */ if (!page_has_buffers(page)) goto unlock_page; head = page_buffers(page); bh = head; do { |
6c006a9d9
|
1567 |
if (!buffer_mapped(bh) || (bh->b_blocknr < block)) |
29f3ad7d8
|
1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 |
goto next; if (bh->b_blocknr >= block + len) break; clear_buffer_dirty(bh); wait_on_buffer(bh); clear_buffer_req(bh); next: bh = bh->b_this_page; } while (bh != head); unlock_page: unlock_page(page); } pagevec_release(&pvec); cond_resched(); |
c10f778dd
|
1582 1583 1584 |
/* End of range already reached? */ if (index > end || !index) break; |
1da177e4c
|
1585 1586 |
} } |
29f3ad7d8
|
1587 |
EXPORT_SYMBOL(clean_bdev_aliases); |
1da177e4c
|
1588 1589 |
/* |
45bce8f3e
|
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 |
* Size is a power-of-two in the range 512..PAGE_SIZE, * and the case we care about most is PAGE_SIZE. * * So this *could* possibly be written with those * constraints in mind (relevant mostly if some * architecture has a slow bit-scan instruction) */ static inline int block_size_bits(unsigned int blocksize) { return ilog2(blocksize); } static struct buffer_head *create_page_buffers(struct page *page, struct inode *inode, unsigned int b_state) { BUG_ON(!PageLocked(page)); if (!page_has_buffers(page)) |
6aa7de059
|
1607 1608 |
create_empty_buffers(page, 1 << READ_ONCE(inode->i_blkbits), b_state); |
45bce8f3e
|
1609 1610 1611 1612 |
return page_buffers(page); } /* |
1da177e4c
|
1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 |
* NOTE! All mapped/uptodate combinations are valid: * * Mapped Uptodate Meaning * * No No "unknown" - must do get_block() * No Yes "hole" - zero-filled * Yes No "allocated" - allocated on disk, not read in * Yes Yes "valid" - allocated and up-to-date in memory. * * "Dirty" is valid only with the last case (mapped+uptodate). */ /* * While block_write_full_page is writing back the dirty buffers under * the page lock, whoever dirtied the buffers may decide to clean them * again at any time. We handle that by only looking at the buffer * state inside lock_buffer(). * * If block_write_full_page() is called for regular writeback * (wbc->sync_mode == WB_SYNC_NONE) then it will redirty a page which has a * locked buffer. This only can happen if someone has written the buffer * directly, with submit_bh(). At the address_space level PageWriteback * prevents this contention from occurring. |
6e34eeddf
|
1636 1637 |
* * If block_write_full_page() is called with wbc->sync_mode == |
70fd76140
|
1638 |
* WB_SYNC_ALL, the writes are posted using REQ_SYNC; this |
721a9602e
|
1639 |
* causes the writes to be flagged as synchronous writes. |
1da177e4c
|
1640 |
*/ |
b4bba3890
|
1641 |
int __block_write_full_page(struct inode *inode, struct page *page, |
35c80d5f4
|
1642 1643 |
get_block_t *get_block, struct writeback_control *wbc, bh_end_io_t *handler) |
1da177e4c
|
1644 1645 1646 1647 |
{ int err; sector_t block; sector_t last_block; |
f0fbd5fc0
|
1648 |
struct buffer_head *bh, *head; |
45bce8f3e
|
1649 |
unsigned int blocksize, bbits; |
1da177e4c
|
1650 |
int nr_underway = 0; |
7637241e6
|
1651 |
int write_flags = wbc_to_write_flags(wbc); |
1da177e4c
|
1652 |
|
45bce8f3e
|
1653 |
head = create_page_buffers(page, inode, |
1da177e4c
|
1654 |
(1 << BH_Dirty)|(1 << BH_Uptodate)); |
1da177e4c
|
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 |
/* * Be very careful. We have no exclusion from __set_page_dirty_buffers * here, and the (potentially unmapped) buffers may become dirty at * any time. If a buffer becomes dirty here after we've inspected it * then we just miss that fact, and the page stays dirty. * * Buffers outside i_size may be dirtied by __set_page_dirty_buffers; * handle that here by just cleaning them. */ |
1da177e4c
|
1665 |
bh = head; |
45bce8f3e
|
1666 1667 |
blocksize = bh->b_size; bbits = block_size_bits(blocksize); |
09cbfeaf1
|
1668 |
block = (sector_t)page->index << (PAGE_SHIFT - bbits); |
45bce8f3e
|
1669 |
last_block = (i_size_read(inode) - 1) >> bbits; |
1da177e4c
|
1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 |
/* * Get all the dirty buffers mapped to disk addresses and * handle any aliases from the underlying blockdev's mapping. */ do { if (block > last_block) { /* * mapped buffers outside i_size will occur, because * this page can be outside i_size when there is a * truncate in progress. */ /* * The buffer was zeroed by block_write_full_page() */ clear_buffer_dirty(bh); set_buffer_uptodate(bh); |
29a814d2e
|
1687 1688 |
} else if ((!buffer_mapped(bh) || buffer_delay(bh)) && buffer_dirty(bh)) { |
b0cf2321c
|
1689 |
WARN_ON(bh->b_size != blocksize); |
1da177e4c
|
1690 1691 1692 |
err = get_block(inode, block, bh, 1); if (err) goto recover; |
29a814d2e
|
1693 |
clear_buffer_delay(bh); |
1da177e4c
|
1694 1695 1696 |
if (buffer_new(bh)) { /* blockdev mappings never come here */ clear_buffer_new(bh); |
e64855c6c
|
1697 |
clean_bdev_bh_alias(bh); |
1da177e4c
|
1698 1699 1700 1701 1702 1703 1704 |
} } bh = bh->b_this_page; block++; } while (bh != head); do { |
1da177e4c
|
1705 1706 1707 1708 1709 |
if (!buffer_mapped(bh)) continue; /* * If it's a fully non-blocking write attempt and we cannot * lock the buffer then redirty the page. Note that this can |
5b0830cb9
|
1710 1711 1712 |
* potentially cause a busy-wait loop from writeback threads * and kswapd activity, but those code paths have their own * higher-level throttling. |
1da177e4c
|
1713 |
*/ |
1b430beee
|
1714 |
if (wbc->sync_mode != WB_SYNC_NONE) { |
1da177e4c
|
1715 |
lock_buffer(bh); |
ca5de404f
|
1716 |
} else if (!trylock_buffer(bh)) { |
1da177e4c
|
1717 1718 1719 1720 |
redirty_page_for_writepage(wbc, page); continue; } if (test_clear_buffer_dirty(bh)) { |
35c80d5f4
|
1721 |
mark_buffer_async_write_endio(bh, handler); |
1da177e4c
|
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 |
} else { unlock_buffer(bh); } } while ((bh = bh->b_this_page) != head); /* * The page and its buffers are protected by PageWriteback(), so we can * drop the bh refcounts early. */ BUG_ON(PageWriteback(page)); set_page_writeback(page); |
1da177e4c
|
1733 1734 1735 1736 |
do { struct buffer_head *next = bh->b_this_page; if (buffer_async_write(bh)) { |
8e8f92988
|
1737 1738 |
submit_bh_wbc(REQ_OP_WRITE, write_flags, bh, inode->i_write_hint, wbc); |
1da177e4c
|
1739 1740 |
nr_underway++; } |
1da177e4c
|
1741 1742 |
bh = next; } while (bh != head); |
05937baae
|
1743 |
unlock_page(page); |
1da177e4c
|
1744 1745 1746 1747 1748 1749 1750 1751 1752 |
err = 0; done: if (nr_underway == 0) { /* * The page was marked dirty, but the buffers were * clean. Someone wrote them back by hand with * ll_rw_block/submit_bh. A rare case. */ |
1da177e4c
|
1753 |
end_page_writeback(page); |
3d67f2d7c
|
1754 |
|
1da177e4c
|
1755 1756 1757 1758 |
/* * The page and buffer_heads can be released at any time from * here on. */ |
1da177e4c
|
1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 |
} return err; recover: /* * ENOSPC, or some other error. We may already have added some * blocks to the file, so we need to write these out to avoid * exposing stale data. * The page is currently locked and not marked for writeback */ bh = head; /* Recovery: lock and submit the mapped buffers */ do { |
29a814d2e
|
1772 1773 |
if (buffer_mapped(bh) && buffer_dirty(bh) && !buffer_delay(bh)) { |
1da177e4c
|
1774 |
lock_buffer(bh); |
35c80d5f4
|
1775 |
mark_buffer_async_write_endio(bh, handler); |
1da177e4c
|
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 |
} else { /* * The buffer may have been set dirty during * attachment to a dirty page. */ clear_buffer_dirty(bh); } } while ((bh = bh->b_this_page) != head); SetPageError(page); BUG_ON(PageWriteback(page)); |
7e4c3690b
|
1786 |
mapping_set_error(page->mapping, err); |
1da177e4c
|
1787 |
set_page_writeback(page); |
1da177e4c
|
1788 1789 1790 1791 |
do { struct buffer_head *next = bh->b_this_page; if (buffer_async_write(bh)) { clear_buffer_dirty(bh); |
8e8f92988
|
1792 1793 |
submit_bh_wbc(REQ_OP_WRITE, write_flags, bh, inode->i_write_hint, wbc); |
1da177e4c
|
1794 1795 |
nr_underway++; } |
1da177e4c
|
1796 1797 |
bh = next; } while (bh != head); |
ffda9d302
|
1798 |
unlock_page(page); |
1da177e4c
|
1799 1800 |
goto done; } |
b4bba3890
|
1801 |
EXPORT_SYMBOL(__block_write_full_page); |
1da177e4c
|
1802 |
|
afddba49d
|
1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 |
/* * If a page has any new buffers, zero them out here, and mark them uptodate * and dirty so they'll be written out (in order to prevent uninitialised * block data from leaking). And clear the new bit. */ void page_zero_new_buffers(struct page *page, unsigned from, unsigned to) { unsigned int block_start, block_end; struct buffer_head *head, *bh; BUG_ON(!PageLocked(page)); if (!page_has_buffers(page)) return; bh = head = page_buffers(page); block_start = 0; do { block_end = block_start + bh->b_size; if (buffer_new(bh)) { if (block_end > from && block_start < to) { if (!PageUptodate(page)) { unsigned start, size; start = max(from, block_start); size = min(to, block_end) - start; |
eebd2aa35
|
1829 |
zero_user(page, start, size); |
afddba49d
|
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 |
set_buffer_uptodate(bh); } clear_buffer_new(bh); mark_buffer_dirty(bh); } } block_start = block_end; bh = bh->b_this_page; } while (bh != head); } EXPORT_SYMBOL(page_zero_new_buffers); |
ae259a9c8
|
1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 |
static void iomap_to_bh(struct inode *inode, sector_t block, struct buffer_head *bh, struct iomap *iomap) { loff_t offset = block << inode->i_blkbits; bh->b_bdev = iomap->bdev; /* * Block points to offset in file we need to map, iomap contains * the offset at which the map starts. If the map ends before the * current block, then do not map the buffer and let the caller * handle it. */ BUG_ON(offset >= iomap->offset + iomap->length); switch (iomap->type) { case IOMAP_HOLE: /* * If the buffer is not up to date or beyond the current EOF, * we need to mark it as new to ensure sub-block zeroing is * executed if necessary. */ if (!buffer_uptodate(bh) || (offset >= i_size_read(inode))) set_buffer_new(bh); break; case IOMAP_DELALLOC: if (!buffer_uptodate(bh) || (offset >= i_size_read(inode))) set_buffer_new(bh); set_buffer_uptodate(bh); set_buffer_mapped(bh); set_buffer_delay(bh); break; case IOMAP_UNWRITTEN: /* |
3d7b6b21f
|
1880 1881 1882 |
* For unwritten regions, we always need to ensure that regions * in the block we are not writing to are zeroed. Mark the * buffer as new to ensure this. |
ae259a9c8
|
1883 1884 1885 1886 1887 |
*/ set_buffer_new(bh); set_buffer_unwritten(bh); /* FALLTHRU */ case IOMAP_MAPPED: |
3d7b6b21f
|
1888 1889 |
if ((iomap->flags & IOMAP_F_NEW) || offset >= i_size_read(inode)) |
ae259a9c8
|
1890 |
set_buffer_new(bh); |
19fe5f643
|
1891 1892 |
bh->b_blocknr = (iomap->addr + offset - iomap->offset) >> inode->i_blkbits; |
ae259a9c8
|
1893 1894 1895 1896 1897 1898 1899 |
set_buffer_mapped(bh); break; } } int __block_write_begin_int(struct page *page, loff_t pos, unsigned len, get_block_t *get_block, struct iomap *iomap) |
1da177e4c
|
1900 |
{ |
09cbfeaf1
|
1901 |
unsigned from = pos & (PAGE_SIZE - 1); |
ebdec241d
|
1902 |
unsigned to = from + len; |
6e1db88d5
|
1903 |
struct inode *inode = page->mapping->host; |
1da177e4c
|
1904 1905 1906 1907 1908 1909 1910 |
unsigned block_start, block_end; sector_t block; int err = 0; unsigned blocksize, bbits; struct buffer_head *bh, *head, *wait[2], **wait_bh=wait; BUG_ON(!PageLocked(page)); |
09cbfeaf1
|
1911 1912 |
BUG_ON(from > PAGE_SIZE); BUG_ON(to > PAGE_SIZE); |
1da177e4c
|
1913 |
BUG_ON(from > to); |
45bce8f3e
|
1914 1915 1916 |
head = create_page_buffers(page, inode, 0); blocksize = head->b_size; bbits = block_size_bits(blocksize); |
1da177e4c
|
1917 |
|
09cbfeaf1
|
1918 |
block = (sector_t)page->index << (PAGE_SHIFT - bbits); |
1da177e4c
|
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 |
for(bh = head, block_start = 0; bh != head || !block_start; block++, block_start=block_end, bh = bh->b_this_page) { block_end = block_start + blocksize; if (block_end <= from || block_start >= to) { if (PageUptodate(page)) { if (!buffer_uptodate(bh)) set_buffer_uptodate(bh); } continue; } if (buffer_new(bh)) clear_buffer_new(bh); if (!buffer_mapped(bh)) { |
b0cf2321c
|
1933 |
WARN_ON(bh->b_size != blocksize); |
ae259a9c8
|
1934 1935 1936 1937 1938 1939 1940 |
if (get_block) { err = get_block(inode, block, bh, 1); if (err) break; } else { iomap_to_bh(inode, block, bh, iomap); } |
1da177e4c
|
1941 |
if (buffer_new(bh)) { |
e64855c6c
|
1942 |
clean_bdev_bh_alias(bh); |
1da177e4c
|
1943 |
if (PageUptodate(page)) { |
637aff46f
|
1944 |
clear_buffer_new(bh); |
1da177e4c
|
1945 |
set_buffer_uptodate(bh); |
637aff46f
|
1946 |
mark_buffer_dirty(bh); |
1da177e4c
|
1947 1948 |
continue; } |
eebd2aa35
|
1949 1950 1951 1952 |
if (block_end > to || block_start < from) zero_user_segments(page, to, block_end, block_start, from); |
1da177e4c
|
1953 1954 1955 1956 1957 1958 1959 1960 1961 |
continue; } } if (PageUptodate(page)) { if (!buffer_uptodate(bh)) set_buffer_uptodate(bh); continue; } if (!buffer_uptodate(bh) && !buffer_delay(bh) && |
33a266dda
|
1962 |
!buffer_unwritten(bh) && |
1da177e4c
|
1963 |
(block_start < from || block_end > to)) { |
dfec8a14f
|
1964 |
ll_rw_block(REQ_OP_READ, 0, 1, &bh); |
1da177e4c
|
1965 1966 1967 1968 1969 1970 1971 1972 1973 |
*wait_bh++=bh; } } /* * If we issued read requests - let them complete. */ while(wait_bh > wait) { wait_on_buffer(*--wait_bh); if (!buffer_uptodate(*wait_bh)) |
f3ddbdc62
|
1974 |
err = -EIO; |
1da177e4c
|
1975 |
} |
f9f07b6c1
|
1976 |
if (unlikely(err)) |
afddba49d
|
1977 |
page_zero_new_buffers(page, from, to); |
1da177e4c
|
1978 1979 |
return err; } |
ae259a9c8
|
1980 1981 1982 1983 1984 1985 |
int __block_write_begin(struct page *page, loff_t pos, unsigned len, get_block_t *get_block) { return __block_write_begin_int(page, pos, len, get_block, NULL); } |
ebdec241d
|
1986 |
EXPORT_SYMBOL(__block_write_begin); |
1da177e4c
|
1987 1988 1989 1990 1991 1992 1993 1994 |
static int __block_commit_write(struct inode *inode, struct page *page, unsigned from, unsigned to) { unsigned block_start, block_end; int partial = 0; unsigned blocksize; struct buffer_head *bh, *head; |
45bce8f3e
|
1995 1996 |
bh = head = page_buffers(page); blocksize = bh->b_size; |
1da177e4c
|
1997 |
|
45bce8f3e
|
1998 1999 |
block_start = 0; do { |
1da177e4c
|
2000 2001 2002 2003 2004 2005 2006 2007 |
block_end = block_start + blocksize; if (block_end <= from || block_start >= to) { if (!buffer_uptodate(bh)) partial = 1; } else { set_buffer_uptodate(bh); mark_buffer_dirty(bh); } |
afddba49d
|
2008 |
clear_buffer_new(bh); |
45bce8f3e
|
2009 2010 2011 2012 |
block_start = block_end; bh = bh->b_this_page; } while (bh != head); |
1da177e4c
|
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 |
/* * If this is a partial write which happened to make all buffers * uptodate then we can optimize away a bogus readpage() for * the next read(). Here we 'discover' whether the page went * uptodate as a result of this (potentially partial) write. */ if (!partial) SetPageUptodate(page); return 0; } /* |
155130a4f
|
2026 2027 2028 |
* block_write_begin takes care of the basic task of block allocation and * bringing partial write blocks uptodate first. * |
7bb46a673
|
2029 |
* The filesystem needs to handle block truncation upon failure. |
afddba49d
|
2030 |
*/ |
155130a4f
|
2031 2032 |
int block_write_begin(struct address_space *mapping, loff_t pos, unsigned len, unsigned flags, struct page **pagep, get_block_t *get_block) |
afddba49d
|
2033 |
{ |
09cbfeaf1
|
2034 |
pgoff_t index = pos >> PAGE_SHIFT; |
afddba49d
|
2035 |
struct page *page; |
6e1db88d5
|
2036 |
int status; |
afddba49d
|
2037 |
|
6e1db88d5
|
2038 2039 2040 |
page = grab_cache_page_write_begin(mapping, index, flags); if (!page) return -ENOMEM; |
afddba49d
|
2041 |
|
6e1db88d5
|
2042 |
status = __block_write_begin(page, pos, len, get_block); |
afddba49d
|
2043 |
if (unlikely(status)) { |
6e1db88d5
|
2044 |
unlock_page(page); |
09cbfeaf1
|
2045 |
put_page(page); |
6e1db88d5
|
2046 |
page = NULL; |
afddba49d
|
2047 |
} |
6e1db88d5
|
2048 |
*pagep = page; |
afddba49d
|
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 |
return status; } EXPORT_SYMBOL(block_write_begin); int block_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) { struct inode *inode = mapping->host; unsigned start; |
09cbfeaf1
|
2059 |
start = pos & (PAGE_SIZE - 1); |
afddba49d
|
2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 |
if (unlikely(copied < len)) { /* * The buffers that were written will now be uptodate, so we * don't have to worry about a readpage reading them and * overwriting a partial write. However if we have encountered * a short write and only partially written into a buffer, it * will not be marked uptodate, so a readpage might come in and * destroy our partial write. * * Do the simplest thing, and just treat any short write to a * non uptodate page as a zero-length write, and force the * caller to redo the whole thing. */ if (!PageUptodate(page)) copied = 0; page_zero_new_buffers(page, start+copied, start+len); } flush_dcache_page(page); /* This could be a short (even 0-length) commit */ __block_commit_write(inode, page, start, start+copied); return copied; } EXPORT_SYMBOL(block_write_end); int generic_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) { |
8af54f291
|
2092 2093 2094 |
struct inode *inode = mapping->host; loff_t old_size = inode->i_size; bool i_size_changed = false; |
afddba49d
|
2095 |
copied = block_write_end(file, mapping, pos, len, copied, page, fsdata); |
8af54f291
|
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 |
/* * No need to use i_size_read() here, the i_size cannot change under us * because we hold i_rwsem. * * But it's important to update i_size while still holding page lock: * page writeout could otherwise come in and zero beyond i_size. */ if (pos + copied > inode->i_size) { i_size_write(inode, pos + copied); i_size_changed = true; } unlock_page(page); |
7a77dad7e
|
2110 |
put_page(page); |
8af54f291
|
2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 |
if (old_size < pos) pagecache_isize_extended(inode, old_size, pos); /* * Don't mark the inode dirty under page lock. First, it unnecessarily * makes the holding time of page lock longer. Second, it forces lock * ordering of page lock and transaction start for journaling * filesystems. */ if (i_size_changed) mark_inode_dirty(inode); |
26ddb1f4f
|
2122 |
return copied; |
afddba49d
|
2123 2124 2125 2126 |
} EXPORT_SYMBOL(generic_write_end); /* |
8ab22b9ab
|
2127 2128 2129 2130 2131 2132 |
* block_is_partially_uptodate checks whether buffers within a page are * uptodate or not. * * Returns true if all buffers which correspond to a file portion * we want to read are uptodate. */ |
c186afb4d
|
2133 2134 |
int block_is_partially_uptodate(struct page *page, unsigned long from, unsigned long count) |
8ab22b9ab
|
2135 |
{ |
8ab22b9ab
|
2136 2137 2138 2139 2140 2141 2142 |
unsigned block_start, block_end, blocksize; unsigned to; struct buffer_head *bh, *head; int ret = 1; if (!page_has_buffers(page)) return 0; |
45bce8f3e
|
2143 2144 |
head = page_buffers(page); blocksize = head->b_size; |
09cbfeaf1
|
2145 |
to = min_t(unsigned, PAGE_SIZE - from, count); |
8ab22b9ab
|
2146 |
to = from + to; |
09cbfeaf1
|
2147 |
if (from < blocksize && to > PAGE_SIZE - blocksize) |
8ab22b9ab
|
2148 |
return 0; |
8ab22b9ab
|
2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 |
bh = head; block_start = 0; do { block_end = block_start + blocksize; if (block_end > from && block_start < to) { if (!buffer_uptodate(bh)) { ret = 0; break; } if (block_end >= to) break; } block_start = block_end; bh = bh->b_this_page; } while (bh != head); return ret; } EXPORT_SYMBOL(block_is_partially_uptodate); /* |
1da177e4c
|
2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 |
* Generic "read page" function for block devices that have the normal * get_block functionality. This is most of the block device filesystems. * Reads the page asynchronously --- the unlock_buffer() and * set/clear_buffer_uptodate() functions propagate buffer state into the * page struct once IO has completed. */ int block_read_full_page(struct page *page, get_block_t *get_block) { struct inode *inode = page->mapping->host; sector_t iblock, lblock; struct buffer_head *bh, *head, *arr[MAX_BUF_PER_PAGE]; |
45bce8f3e
|
2181 |
unsigned int blocksize, bbits; |
1da177e4c
|
2182 2183 |
int nr, i; int fully_mapped = 1; |
45bce8f3e
|
2184 2185 2186 |
head = create_page_buffers(page, inode, 0); blocksize = head->b_size; bbits = block_size_bits(blocksize); |
1da177e4c
|
2187 |
|
09cbfeaf1
|
2188 |
iblock = (sector_t)page->index << (PAGE_SHIFT - bbits); |
45bce8f3e
|
2189 |
lblock = (i_size_read(inode)+blocksize-1) >> bbits; |
1da177e4c
|
2190 2191 2192 2193 2194 2195 2196 2197 2198 |
bh = head; nr = 0; i = 0; do { if (buffer_uptodate(bh)) continue; if (!buffer_mapped(bh)) { |
c64610ba5
|
2199 |
int err = 0; |
1da177e4c
|
2200 2201 |
fully_mapped = 0; if (iblock < lblock) { |
b0cf2321c
|
2202 |
WARN_ON(bh->b_size != blocksize); |
c64610ba5
|
2203 2204 |
err = get_block(inode, iblock, bh, 0); if (err) |
1da177e4c
|
2205 2206 2207 |
SetPageError(page); } if (!buffer_mapped(bh)) { |
eebd2aa35
|
2208 |
zero_user(page, i * blocksize, blocksize); |
c64610ba5
|
2209 2210 |
if (!err) set_buffer_uptodate(bh); |
1da177e4c
|
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 |
continue; } /* * get_block() might have updated the buffer * synchronously */ if (buffer_uptodate(bh)) continue; } arr[nr++] = bh; } while (i++, iblock++, (bh = bh->b_this_page) != head); if (fully_mapped) SetPageMappedToDisk(page); if (!nr) { /* * All buffers are uptodate - we can set the page uptodate * as well. But not if get_block() returned an error. */ if (!PageError(page)) SetPageUptodate(page); unlock_page(page); return 0; } /* Stage two: lock the buffers */ for (i = 0; i < nr; i++) { bh = arr[i]; lock_buffer(bh); mark_buffer_async_read(bh); } /* * Stage 3: start the IO. Check for uptodateness * inside the buffer lock in case another process reading * the underlying blockdev brought it uptodate (the sct fix). */ for (i = 0; i < nr; i++) { bh = arr[i]; if (buffer_uptodate(bh)) end_buffer_async_read(bh, 1); else |
2a222ca99
|
2254 |
submit_bh(REQ_OP_READ, 0, bh); |
1da177e4c
|
2255 2256 2257 |
} return 0; } |
1fe72eaa0
|
2258 |
EXPORT_SYMBOL(block_read_full_page); |
1da177e4c
|
2259 2260 |
/* utility function for filesystems that need to do work on expanding |
89e107877
|
2261 |
* truncates. Uses filesystem pagecache writes to allow the filesystem to |
1da177e4c
|
2262 2263 |
* deal with the hole. */ |
89e107877
|
2264 |
int generic_cont_expand_simple(struct inode *inode, loff_t size) |
1da177e4c
|
2265 2266 2267 |
{ struct address_space *mapping = inode->i_mapping; struct page *page; |
89e107877
|
2268 |
void *fsdata; |
1da177e4c
|
2269 |
int err; |
c08d3b0e3
|
2270 2271 |
err = inode_newsize_ok(inode, size); if (err) |
1da177e4c
|
2272 |
goto out; |
89e107877
|
2273 |
err = pagecache_write_begin(NULL, mapping, size, 0, |
c718a9751
|
2274 |
AOP_FLAG_CONT_EXPAND, &page, &fsdata); |
89e107877
|
2275 |
if (err) |
05eb0b51f
|
2276 |
goto out; |
05eb0b51f
|
2277 |
|
89e107877
|
2278 2279 |
err = pagecache_write_end(NULL, mapping, size, 0, 0, page, fsdata); BUG_ON(err > 0); |
05eb0b51f
|
2280 |
|
1da177e4c
|
2281 2282 2283 |
out: return err; } |
1fe72eaa0
|
2284 |
EXPORT_SYMBOL(generic_cont_expand_simple); |
1da177e4c
|
2285 |
|
f1e3af72c
|
2286 2287 |
static int cont_expand_zero(struct file *file, struct address_space *mapping, loff_t pos, loff_t *bytes) |
1da177e4c
|
2288 |
{ |
1da177e4c
|
2289 |
struct inode *inode = mapping->host; |
93407472a
|
2290 |
unsigned int blocksize = i_blocksize(inode); |
89e107877
|
2291 2292 2293 2294 2295 2296 |
struct page *page; void *fsdata; pgoff_t index, curidx; loff_t curpos; unsigned zerofrom, offset, len; int err = 0; |
1da177e4c
|
2297 |
|
09cbfeaf1
|
2298 2299 |
index = pos >> PAGE_SHIFT; offset = pos & ~PAGE_MASK; |
89e107877
|
2300 |
|
09cbfeaf1
|
2301 2302 |
while (index > (curidx = (curpos = *bytes)>>PAGE_SHIFT)) { zerofrom = curpos & ~PAGE_MASK; |
1da177e4c
|
2303 2304 2305 2306 |
if (zerofrom & (blocksize-1)) { *bytes |= (blocksize-1); (*bytes)++; } |
09cbfeaf1
|
2307 |
len = PAGE_SIZE - zerofrom; |
1da177e4c
|
2308 |
|
c718a9751
|
2309 2310 |
err = pagecache_write_begin(file, mapping, curpos, len, 0, &page, &fsdata); |
89e107877
|
2311 2312 |
if (err) goto out; |
eebd2aa35
|
2313 |
zero_user(page, zerofrom, len); |
89e107877
|
2314 2315 2316 2317 2318 2319 |
err = pagecache_write_end(file, mapping, curpos, len, len, page, fsdata); if (err < 0) goto out; BUG_ON(err != len); err = 0; |
061e97469
|
2320 2321 |
balance_dirty_pages_ratelimited(mapping); |
c2ca0fcd2
|
2322 |
|
08d405c8b
|
2323 |
if (fatal_signal_pending(current)) { |
c2ca0fcd2
|
2324 2325 2326 |
err = -EINTR; goto out; } |
89e107877
|
2327 |
} |
1da177e4c
|
2328 |
|
89e107877
|
2329 2330 |
/* page covers the boundary, find the boundary offset */ if (index == curidx) { |
09cbfeaf1
|
2331 |
zerofrom = curpos & ~PAGE_MASK; |
1da177e4c
|
2332 |
/* if we will expand the thing last block will be filled */ |
89e107877
|
2333 2334 2335 2336 |
if (offset <= zerofrom) { goto out; } if (zerofrom & (blocksize-1)) { |
1da177e4c
|
2337 2338 2339 |
*bytes |= (blocksize-1); (*bytes)++; } |
89e107877
|
2340 |
len = offset - zerofrom; |
1da177e4c
|
2341 |
|
c718a9751
|
2342 2343 |
err = pagecache_write_begin(file, mapping, curpos, len, 0, &page, &fsdata); |
89e107877
|
2344 2345 |
if (err) goto out; |
eebd2aa35
|
2346 |
zero_user(page, zerofrom, len); |
89e107877
|
2347 2348 2349 2350 2351 2352 |
err = pagecache_write_end(file, mapping, curpos, len, len, page, fsdata); if (err < 0) goto out; BUG_ON(err != len); err = 0; |
1da177e4c
|
2353 |
} |
89e107877
|
2354 2355 2356 2357 2358 2359 2360 2361 |
out: return err; } /* * For moronic filesystems that do not allow holes in file. * We may have to extend the file. */ |
282dc1788
|
2362 |
int cont_write_begin(struct file *file, struct address_space *mapping, |
89e107877
|
2363 2364 2365 2366 2367 |
loff_t pos, unsigned len, unsigned flags, struct page **pagep, void **fsdata, get_block_t *get_block, loff_t *bytes) { struct inode *inode = mapping->host; |
93407472a
|
2368 2369 |
unsigned int blocksize = i_blocksize(inode); unsigned int zerofrom; |
89e107877
|
2370 2371 2372 2373 |
int err; err = cont_expand_zero(file, mapping, pos, bytes); if (err) |
155130a4f
|
2374 |
return err; |
89e107877
|
2375 |
|
09cbfeaf1
|
2376 |
zerofrom = *bytes & ~PAGE_MASK; |
89e107877
|
2377 2378 2379 |
if (pos+len > *bytes && zerofrom & (blocksize-1)) { *bytes |= (blocksize-1); (*bytes)++; |
1da177e4c
|
2380 |
} |
1da177e4c
|
2381 |
|
155130a4f
|
2382 |
return block_write_begin(mapping, pos, len, flags, pagep, get_block); |
1da177e4c
|
2383 |
} |
1fe72eaa0
|
2384 |
EXPORT_SYMBOL(cont_write_begin); |
1da177e4c
|
2385 |
|
1da177e4c
|
2386 2387 2388 2389 2390 2391 |
int block_commit_write(struct page *page, unsigned from, unsigned to) { struct inode *inode = page->mapping->host; __block_commit_write(inode,page,from,to); return 0; } |
1fe72eaa0
|
2392 |
EXPORT_SYMBOL(block_commit_write); |
1da177e4c
|
2393 |
|
541716902
|
2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 |
/* * block_page_mkwrite() is not allowed to change the file size as it gets * called from a page fault handler when a page is first dirtied. Hence we must * be careful to check for EOF conditions here. We set the page up correctly * for a written page which means we get ENOSPC checking when writing into * holes and correct delalloc and unwritten extent mapping on filesystems that * support these features. * * We are not allowed to take the i_mutex here so we have to play games to * protect against truncate races as the page could now be beyond EOF. Because |
7bb46a673
|
2404 |
* truncate writes the inode size before removing pages, once we have the |
541716902
|
2405 2406 2407 |
* page lock we can determine safely if the page is beyond EOF. If it is not * beyond EOF, then the page is guaranteed safe against truncation until we * unlock the page. |
ea13a8646
|
2408 |
* |
14da92001
|
2409 |
* Direct callers of this function should protect against filesystem freezing |
5c5000296
|
2410 |
* using sb_start_pagefault() - sb_end_pagefault() functions. |
541716902
|
2411 |
*/ |
5c5000296
|
2412 |
int block_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf, |
24da4fab5
|
2413 |
get_block_t get_block) |
541716902
|
2414 |
{ |
c2ec175c3
|
2415 |
struct page *page = vmf->page; |
496ad9aa8
|
2416 |
struct inode *inode = file_inode(vma->vm_file); |
541716902
|
2417 2418 |
unsigned long end; loff_t size; |
24da4fab5
|
2419 |
int ret; |
541716902
|
2420 2421 2422 2423 |
lock_page(page); size = i_size_read(inode); if ((page->mapping != inode->i_mapping) || |
183363380
|
2424 |
(page_offset(page) > size)) { |
24da4fab5
|
2425 2426 2427 |
/* We overload EFAULT to mean page got truncated */ ret = -EFAULT; goto out_unlock; |
541716902
|
2428 2429 2430 |
} /* page is wholly or partially inside EOF */ |
09cbfeaf1
|
2431 2432 |
if (((page->index + 1) << PAGE_SHIFT) > size) end = size & ~PAGE_MASK; |
541716902
|
2433 |
else |
09cbfeaf1
|
2434 |
end = PAGE_SIZE; |
541716902
|
2435 |
|
ebdec241d
|
2436 |
ret = __block_write_begin(page, 0, end, get_block); |
541716902
|
2437 2438 |
if (!ret) ret = block_commit_write(page, 0, end); |
24da4fab5
|
2439 2440 |
if (unlikely(ret < 0)) goto out_unlock; |
ea13a8646
|
2441 |
set_page_dirty(page); |
1d1d1a767
|
2442 |
wait_for_stable_page(page); |
24da4fab5
|
2443 2444 2445 |
return 0; out_unlock: unlock_page(page); |
541716902
|
2446 |
return ret; |
24da4fab5
|
2447 |
} |
1fe72eaa0
|
2448 |
EXPORT_SYMBOL(block_page_mkwrite); |
1da177e4c
|
2449 2450 |
/* |
03158cd7e
|
2451 |
* nobh_write_begin()'s prereads are special: the buffer_heads are freed |
1da177e4c
|
2452 2453 |
* immediately, while under the page lock. So it needs a special end_io * handler which does not touch the bh after unlocking it. |
1da177e4c
|
2454 2455 2456 |
*/ static void end_buffer_read_nobh(struct buffer_head *bh, int uptodate) { |
68671f35f
|
2457 |
__end_buffer_read_notouch(bh, uptodate); |
1da177e4c
|
2458 2459 2460 |
} /* |
03158cd7e
|
2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 |
* Attach the singly-linked list of buffers created by nobh_write_begin, to * the page (converting it to circular linked list and taking care of page * dirty races). */ static void attach_nobh_buffers(struct page *page, struct buffer_head *head) { struct buffer_head *bh; BUG_ON(!PageLocked(page)); spin_lock(&page->mapping->private_lock); bh = head; do { if (PageDirty(page)) set_buffer_dirty(bh); if (!bh->b_this_page) bh->b_this_page = head; bh = bh->b_this_page; } while (bh != head); attach_page_buffers(page, head); spin_unlock(&page->mapping->private_lock); } /* |
ea0f04e59
|
2485 2486 |
* On entry, the page is fully not uptodate. * On exit the page is fully uptodate in the areas outside (from,to) |
7bb46a673
|
2487 |
* The filesystem needs to handle block truncation upon failure. |
1da177e4c
|
2488 |
*/ |
ea0f04e59
|
2489 |
int nobh_write_begin(struct address_space *mapping, |
03158cd7e
|
2490 2491 |
loff_t pos, unsigned len, unsigned flags, struct page **pagep, void **fsdata, |
1da177e4c
|
2492 2493 |
get_block_t *get_block) { |
03158cd7e
|
2494 |
struct inode *inode = mapping->host; |
1da177e4c
|
2495 2496 |
const unsigned blkbits = inode->i_blkbits; const unsigned blocksize = 1 << blkbits; |
a4b0672db
|
2497 |
struct buffer_head *head, *bh; |
03158cd7e
|
2498 2499 2500 |
struct page *page; pgoff_t index; unsigned from, to; |
1da177e4c
|
2501 |
unsigned block_in_page; |
a4b0672db
|
2502 |
unsigned block_start, block_end; |
1da177e4c
|
2503 |
sector_t block_in_file; |
1da177e4c
|
2504 |
int nr_reads = 0; |
1da177e4c
|
2505 2506 |
int ret = 0; int is_mapped_to_disk = 1; |
1da177e4c
|
2507 |
|
09cbfeaf1
|
2508 2509 |
index = pos >> PAGE_SHIFT; from = pos & (PAGE_SIZE - 1); |
03158cd7e
|
2510 |
to = from + len; |
54566b2c1
|
2511 |
page = grab_cache_page_write_begin(mapping, index, flags); |
03158cd7e
|
2512 2513 2514 2515 2516 2517 |
if (!page) return -ENOMEM; *pagep = page; *fsdata = NULL; if (page_has_buffers(page)) { |
309f77ad9
|
2518 2519 2520 2521 |
ret = __block_write_begin(page, pos, len, get_block); if (unlikely(ret)) goto out_release; return ret; |
03158cd7e
|
2522 |
} |
a4b0672db
|
2523 |
|
1da177e4c
|
2524 2525 |
if (PageMappedToDisk(page)) return 0; |
a4b0672db
|
2526 2527 2528 2529 2530 2531 2532 2533 2534 |
/* * Allocate buffers so that we can keep track of state, and potentially * attach them to the page if an error occurs. In the common case of * no error, they will just be freed again without ever being attached * to the page (which is all OK, because we're under the page lock). * * Be careful: the buffer linked list is a NULL terminated one, rather * than the circular one we're used to. */ |
640ab98fb
|
2535 |
head = alloc_page_buffers(page, blocksize, false); |
03158cd7e
|
2536 2537 2538 2539 |
if (!head) { ret = -ENOMEM; goto out_release; } |
a4b0672db
|
2540 |
|
09cbfeaf1
|
2541 |
block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits); |
1da177e4c
|
2542 2543 2544 2545 2546 2547 |
/* * We loop across all blocks in the page, whether or not they are * part of the affected region. This is so we can discover if the * page is fully mapped-to-disk. */ |
a4b0672db
|
2548 |
for (block_start = 0, block_in_page = 0, bh = head; |
09cbfeaf1
|
2549 |
block_start < PAGE_SIZE; |
a4b0672db
|
2550 |
block_in_page++, block_start += blocksize, bh = bh->b_this_page) { |
1da177e4c
|
2551 |
int create; |
a4b0672db
|
2552 2553 |
block_end = block_start + blocksize; bh->b_state = 0; |
1da177e4c
|
2554 2555 2556 2557 |
create = 1; if (block_start >= to) create = 0; ret = get_block(inode, block_in_file + block_in_page, |
a4b0672db
|
2558 |
bh, create); |
1da177e4c
|
2559 2560 |
if (ret) goto failed; |
a4b0672db
|
2561 |
if (!buffer_mapped(bh)) |
1da177e4c
|
2562 |
is_mapped_to_disk = 0; |
a4b0672db
|
2563 |
if (buffer_new(bh)) |
e64855c6c
|
2564 |
clean_bdev_bh_alias(bh); |
a4b0672db
|
2565 2566 |
if (PageUptodate(page)) { set_buffer_uptodate(bh); |
1da177e4c
|
2567 |
continue; |
a4b0672db
|
2568 2569 |
} if (buffer_new(bh) || !buffer_mapped(bh)) { |
eebd2aa35
|
2570 2571 |
zero_user_segments(page, block_start, from, to, block_end); |
1da177e4c
|
2572 2573 |
continue; } |
a4b0672db
|
2574 |
if (buffer_uptodate(bh)) |
1da177e4c
|
2575 2576 |
continue; /* reiserfs does this */ if (block_start < from || block_end > to) { |
a4b0672db
|
2577 2578 |
lock_buffer(bh); bh->b_end_io = end_buffer_read_nobh; |
2a222ca99
|
2579 |
submit_bh(REQ_OP_READ, 0, bh); |
a4b0672db
|
2580 |
nr_reads++; |
1da177e4c
|
2581 2582 2583 2584 |
} } if (nr_reads) { |
1da177e4c
|
2585 2586 2587 2588 2589 |
/* * The page is locked, so these buffers are protected from * any VM or truncate activity. Hence we don't need to care * for the buffer_head refcounts. */ |
a4b0672db
|
2590 |
for (bh = head; bh; bh = bh->b_this_page) { |
1da177e4c
|
2591 2592 2593 |
wait_on_buffer(bh); if (!buffer_uptodate(bh)) ret = -EIO; |
1da177e4c
|
2594 2595 2596 2597 2598 2599 2600 |
} if (ret) goto failed; } if (is_mapped_to_disk) SetPageMappedToDisk(page); |
1da177e4c
|
2601 |
|
03158cd7e
|
2602 |
*fsdata = head; /* to be released by nobh_write_end */ |
a4b0672db
|
2603 |
|
1da177e4c
|
2604 2605 2606 |
return 0; failed: |
03158cd7e
|
2607 |
BUG_ON(!ret); |
1da177e4c
|
2608 |
/* |
a4b0672db
|
2609 2610 2611 2612 2613 |
* Error recovery is a bit difficult. We need to zero out blocks that * were newly allocated, and dirty them to ensure they get written out. * Buffers need to be attached to the page at this point, otherwise * the handling of potential IO errors during writeout would be hard * (could try doing synchronous writeout, but what if that fails too?) |
1da177e4c
|
2614 |
*/ |
03158cd7e
|
2615 2616 |
attach_nobh_buffers(page, head); page_zero_new_buffers(page, from, to); |
a4b0672db
|
2617 |
|
03158cd7e
|
2618 2619 |
out_release: unlock_page(page); |
09cbfeaf1
|
2620 |
put_page(page); |
03158cd7e
|
2621 |
*pagep = NULL; |
a4b0672db
|
2622 |
|
7bb46a673
|
2623 2624 |
return ret; } |
03158cd7e
|
2625 |
EXPORT_SYMBOL(nobh_write_begin); |
1da177e4c
|
2626 |
|
03158cd7e
|
2627 2628 2629 |
int nobh_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) |
1da177e4c
|
2630 2631 |
{ struct inode *inode = page->mapping->host; |
efdc31319
|
2632 |
struct buffer_head *head = fsdata; |
03158cd7e
|
2633 |
struct buffer_head *bh; |
5b41e74ad
|
2634 |
BUG_ON(fsdata != NULL && page_has_buffers(page)); |
1da177e4c
|
2635 |
|
d4cf109f0
|
2636 |
if (unlikely(copied < len) && head) |
5b41e74ad
|
2637 2638 2639 2640 |
attach_nobh_buffers(page, head); if (page_has_buffers(page)) return generic_write_end(file, mapping, pos, len, copied, page, fsdata); |
a4b0672db
|
2641 |
|
22c8ca78f
|
2642 |
SetPageUptodate(page); |
1da177e4c
|
2643 |
set_page_dirty(page); |
03158cd7e
|
2644 2645 |
if (pos+copied > inode->i_size) { i_size_write(inode, pos+copied); |
1da177e4c
|
2646 2647 |
mark_inode_dirty(inode); } |
03158cd7e
|
2648 2649 |
unlock_page(page); |
09cbfeaf1
|
2650 |
put_page(page); |
03158cd7e
|
2651 |
|
03158cd7e
|
2652 2653 2654 2655 2656 2657 2658 |
while (head) { bh = head; head = head->b_this_page; free_buffer_head(bh); } return copied; |
1da177e4c
|
2659 |
} |
03158cd7e
|
2660 |
EXPORT_SYMBOL(nobh_write_end); |
1da177e4c
|
2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 |
/* * nobh_writepage() - based on block_full_write_page() except * that it tries to operate without attaching bufferheads to * the page. */ int nobh_writepage(struct page *page, get_block_t *get_block, struct writeback_control *wbc) { struct inode * const inode = page->mapping->host; loff_t i_size = i_size_read(inode); |
09cbfeaf1
|
2672 |
const pgoff_t end_index = i_size >> PAGE_SHIFT; |
1da177e4c
|
2673 |
unsigned offset; |
1da177e4c
|
2674 2675 2676 2677 2678 2679 2680 |
int ret; /* Is the page fully inside i_size? */ if (page->index < end_index) goto out; /* Is the page fully outside i_size? (truncate in progress) */ |
09cbfeaf1
|
2681 |
offset = i_size & (PAGE_SIZE-1); |
1da177e4c
|
2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 |
if (page->index >= end_index+1 || !offset) { /* * The page may have dirty, unmapped buffers. For example, * they may have been added in ext3_writepage(). Make them * freeable here, so the page does not leak. */ #if 0 /* Not really sure about this - do we need this ? */ if (page->mapping->a_ops->invalidatepage) page->mapping->a_ops->invalidatepage(page, offset); #endif unlock_page(page); return 0; /* don't care */ } /* * The page straddles i_size. It must be zeroed out on each and every * writepage invocation because it may be mmapped. "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
|
2704 |
zero_user_segment(page, offset, PAGE_SIZE); |
1da177e4c
|
2705 2706 2707 |
out: ret = mpage_writepage(page, get_block, wbc); if (ret == -EAGAIN) |
35c80d5f4
|
2708 2709 |
ret = __block_write_full_page(inode, page, get_block, wbc, end_buffer_async_write); |
1da177e4c
|
2710 2711 2712 |
return ret; } EXPORT_SYMBOL(nobh_writepage); |
03158cd7e
|
2713 2714 |
int nobh_truncate_page(struct address_space *mapping, loff_t from, get_block_t *get_block) |
1da177e4c
|
2715 |
{ |
09cbfeaf1
|
2716 2717 |
pgoff_t index = from >> PAGE_SHIFT; unsigned offset = from & (PAGE_SIZE-1); |
03158cd7e
|
2718 2719 2720 2721 |
unsigned blocksize; sector_t iblock; unsigned length, pos; struct inode *inode = mapping->host; |
1da177e4c
|
2722 |
struct page *page; |
03158cd7e
|
2723 2724 |
struct buffer_head map_bh; int err; |
1da177e4c
|
2725 |
|
93407472a
|
2726 |
blocksize = i_blocksize(inode); |
03158cd7e
|
2727 2728 2729 2730 2731 2732 2733 |
length = offset & (blocksize - 1); /* Block boundary? Nothing to do */ if (!length) return 0; length = blocksize - length; |
09cbfeaf1
|
2734 |
iblock = (sector_t)index << (PAGE_SHIFT - inode->i_blkbits); |
1da177e4c
|
2735 |
|
1da177e4c
|
2736 |
page = grab_cache_page(mapping, index); |
03158cd7e
|
2737 |
err = -ENOMEM; |
1da177e4c
|
2738 2739 |
if (!page) goto out; |
03158cd7e
|
2740 2741 2742 |
if (page_has_buffers(page)) { has_buffers: unlock_page(page); |
09cbfeaf1
|
2743 |
put_page(page); |
03158cd7e
|
2744 2745 2746 2747 2748 2749 2750 2751 2752 |
return block_truncate_page(mapping, from, get_block); } /* Find the buffer that contains "offset" */ pos = blocksize; while (offset >= pos) { iblock++; pos += blocksize; } |
460bcf57b
|
2753 2754 |
map_bh.b_size = blocksize; map_bh.b_state = 0; |
03158cd7e
|
2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 |
err = get_block(inode, iblock, &map_bh, 0); if (err) goto unlock; /* unmapped? It's a hole - nothing to do */ if (!buffer_mapped(&map_bh)) goto unlock; /* Ok, it's mapped. Make sure it's up-to-date */ if (!PageUptodate(page)) { err = mapping->a_ops->readpage(NULL, page); if (err) { |
09cbfeaf1
|
2766 |
put_page(page); |
03158cd7e
|
2767 2768 2769 2770 2771 2772 2773 2774 2775 |
goto out; } lock_page(page); if (!PageUptodate(page)) { err = -EIO; goto unlock; } if (page_has_buffers(page)) goto has_buffers; |
1da177e4c
|
2776 |
} |
eebd2aa35
|
2777 |
zero_user(page, offset, length); |
03158cd7e
|
2778 2779 2780 2781 |
set_page_dirty(page); err = 0; unlock: |
1da177e4c
|
2782 |
unlock_page(page); |
09cbfeaf1
|
2783 |
put_page(page); |
1da177e4c
|
2784 |
out: |
03158cd7e
|
2785 |
return err; |
1da177e4c
|
2786 2787 2788 2789 2790 2791 |
} EXPORT_SYMBOL(nobh_truncate_page); int block_truncate_page(struct address_space *mapping, loff_t from, get_block_t *get_block) { |
09cbfeaf1
|
2792 2793 |
pgoff_t index = from >> PAGE_SHIFT; unsigned offset = from & (PAGE_SIZE-1); |
1da177e4c
|
2794 |
unsigned blocksize; |
54b21a799
|
2795 |
sector_t iblock; |
1da177e4c
|
2796 2797 2798 2799 |
unsigned length, pos; struct inode *inode = mapping->host; struct page *page; struct buffer_head *bh; |
1da177e4c
|
2800 |
int err; |
93407472a
|
2801 |
blocksize = i_blocksize(inode); |
1da177e4c
|
2802 2803 2804 2805 2806 2807 2808 |
length = offset & (blocksize - 1); /* Block boundary? Nothing to do */ if (!length) return 0; length = blocksize - length; |
09cbfeaf1
|
2809 |
iblock = (sector_t)index << (PAGE_SHIFT - inode->i_blkbits); |
1da177e4c
|
2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 |
page = grab_cache_page(mapping, index); err = -ENOMEM; if (!page) goto out; if (!page_has_buffers(page)) create_empty_buffers(page, blocksize, 0); /* Find the buffer that contains "offset" */ bh = page_buffers(page); pos = blocksize; while (offset >= pos) { bh = bh->b_this_page; iblock++; pos += blocksize; } err = 0; if (!buffer_mapped(bh)) { |
b0cf2321c
|
2830 |
WARN_ON(bh->b_size != blocksize); |
1da177e4c
|
2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 |
err = get_block(inode, iblock, bh, 0); if (err) goto unlock; /* unmapped? It's a hole - nothing to do */ if (!buffer_mapped(bh)) goto unlock; } /* Ok, it's mapped. Make sure it's up-to-date */ if (PageUptodate(page)) set_buffer_uptodate(bh); |
33a266dda
|
2842 |
if (!buffer_uptodate(bh) && !buffer_delay(bh) && !buffer_unwritten(bh)) { |
1da177e4c
|
2843 |
err = -EIO; |
dfec8a14f
|
2844 |
ll_rw_block(REQ_OP_READ, 0, 1, &bh); |
1da177e4c
|
2845 2846 2847 2848 2849 |
wait_on_buffer(bh); /* Uhhuh. Read error. Complain and punt. */ if (!buffer_uptodate(bh)) goto unlock; } |
eebd2aa35
|
2850 |
zero_user(page, offset, length); |
1da177e4c
|
2851 2852 2853 2854 2855 |
mark_buffer_dirty(bh); err = 0; unlock: unlock_page(page); |
09cbfeaf1
|
2856 |
put_page(page); |
1da177e4c
|
2857 2858 2859 |
out: return err; } |
1fe72eaa0
|
2860 |
EXPORT_SYMBOL(block_truncate_page); |
1da177e4c
|
2861 2862 2863 2864 |
/* * The generic ->writepage function for buffer-backed address_spaces */ |
1b938c082
|
2865 2866 |
int block_write_full_page(struct page *page, get_block_t *get_block, struct writeback_control *wbc) |
1da177e4c
|
2867 2868 2869 |
{ struct inode * const inode = page->mapping->host; loff_t i_size = i_size_read(inode); |
09cbfeaf1
|
2870 |
const pgoff_t end_index = i_size >> PAGE_SHIFT; |
1da177e4c
|
2871 |
unsigned offset; |
1da177e4c
|
2872 2873 2874 |
/* Is the page fully inside i_size? */ if (page->index < end_index) |
35c80d5f4
|
2875 |
return __block_write_full_page(inode, page, get_block, wbc, |
1b938c082
|
2876 |
end_buffer_async_write); |
1da177e4c
|
2877 2878 |
/* Is the page fully outside i_size? (truncate in progress) */ |
09cbfeaf1
|
2879 |
offset = i_size & (PAGE_SIZE-1); |
1da177e4c
|
2880 2881 2882 2883 2884 2885 |
if (page->index >= end_index+1 || !offset) { /* * The page may have dirty, unmapped buffers. For example, * they may have been added in ext3_writepage(). Make them * freeable here, so the page does not leak. */ |
09cbfeaf1
|
2886 |
do_invalidatepage(page, 0, PAGE_SIZE); |
1da177e4c
|
2887 2888 2889 2890 2891 2892 |
unlock_page(page); return 0; /* don't care */ } /* * The page straddles i_size. It must be zeroed out on each and every |
2a61aa401
|
2893 |
* writepage invocation because it may be mmapped. "A file is mapped |
1da177e4c
|
2894 2895 2896 2897 |
* 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
|
2898 |
zero_user_segment(page, offset, PAGE_SIZE); |
1b938c082
|
2899 2900 |
return __block_write_full_page(inode, page, get_block, wbc, end_buffer_async_write); |
35c80d5f4
|
2901 |
} |
1fe72eaa0
|
2902 |
EXPORT_SYMBOL(block_write_full_page); |
35c80d5f4
|
2903 |
|
1da177e4c
|
2904 2905 2906 |
sector_t generic_block_bmap(struct address_space *mapping, sector_t block, get_block_t *get_block) { |
1da177e4c
|
2907 |
struct inode *inode = mapping->host; |
2a527d685
|
2908 2909 2910 |
struct buffer_head tmp = { .b_size = i_blocksize(inode), }; |
1da177e4c
|
2911 2912 2913 |
get_block(inode, block, &tmp, 0); return tmp.b_blocknr; } |
1fe72eaa0
|
2914 |
EXPORT_SYMBOL(generic_block_bmap); |
1da177e4c
|
2915 |
|
4246a0b63
|
2916 |
static void end_bio_bh_io_sync(struct bio *bio) |
1da177e4c
|
2917 2918 |
{ struct buffer_head *bh = bio->bi_private; |
b7c44ed9d
|
2919 |
if (unlikely(bio_flagged(bio, BIO_QUIET))) |
08bafc034
|
2920 |
set_bit(BH_Quiet, &bh->b_state); |
4e4cbee93
|
2921 |
bh->b_end_io(bh, !bio->bi_status); |
1da177e4c
|
2922 |
bio_put(bio); |
1da177e4c
|
2923 |
} |
57302e0dd
|
2924 2925 |
/* * This allows us to do IO even on the odd last sectors |
59d43914e
|
2926 |
* of a device, even if the block size is some multiple |
57302e0dd
|
2927 2928 2929 2930 2931 2932 2933 2934 2935 |
* of the physical sector size. * * We'll just truncate the bio to the size of the device, * and clear the end of the buffer head manually. * * Truly out-of-range accesses will turn into actual IO * errors, this only handles the "we need to be able to * do IO at the final sector" case. */ |
3fe209c84
|
2936 |
void guard_bio_eod(struct bio *bio) |
57302e0dd
|
2937 2938 |
{ sector_t maxsector; |
67f2519fe
|
2939 2940 2941 2942 2943 2944 2945 2946 2947 |
struct hd_struct *part; rcu_read_lock(); part = __disk_get_part(bio->bi_disk, bio->bi_partno); if (part) maxsector = part_nr_sects_read(part); else maxsector = get_capacity(bio->bi_disk); rcu_read_unlock(); |
57302e0dd
|
2948 |
|
57302e0dd
|
2949 2950 2951 2952 2953 2954 2955 2956 |
if (!maxsector) return; /* * If the *whole* IO is past the end of the device, * let it through, and the IO layer will turn it into * an EIO. */ |
4f024f379
|
2957 |
if (unlikely(bio->bi_iter.bi_sector >= maxsector)) |
57302e0dd
|
2958 |
return; |
4f024f379
|
2959 |
maxsector -= bio->bi_iter.bi_sector; |
59d43914e
|
2960 |
if (likely((bio->bi_iter.bi_size >> 9) <= maxsector)) |
57302e0dd
|
2961 |
return; |
943cd69ef
|
2962 |
bio_truncate(bio, maxsector << 9); |
57302e0dd
|
2963 |
} |
2a222ca99
|
2964 |
static int submit_bh_wbc(int op, int op_flags, struct buffer_head *bh, |
8e8f92988
|
2965 |
enum rw_hint write_hint, struct writeback_control *wbc) |
1da177e4c
|
2966 2967 |
{ struct bio *bio; |
1da177e4c
|
2968 2969 2970 2971 |
BUG_ON(!buffer_locked(bh)); BUG_ON(!buffer_mapped(bh)); BUG_ON(!bh->b_end_io); |
8fb0e3424
|
2972 2973 |
BUG_ON(buffer_delay(bh)); BUG_ON(buffer_unwritten(bh)); |
1da177e4c
|
2974 |
|
48fd4f93a
|
2975 |
/* |
48fd4f93a
|
2976 |
* Only clear out a write error when rewriting |
1da177e4c
|
2977 |
*/ |
2a222ca99
|
2978 |
if (test_set_buffer_req(bh) && (op == REQ_OP_WRITE)) |
1da177e4c
|
2979 2980 2981 2982 2983 2984 2985 |
clear_buffer_write_io_error(bh); /* * from here on down, it's all bio -- do the initial mapping, * submit_bio -> generic_make_request may further map this bio around */ bio = bio_alloc(GFP_NOIO, 1); |
4f024f379
|
2986 |
bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9); |
74d46992e
|
2987 |
bio_set_dev(bio, bh->b_bdev); |
8e8f92988
|
2988 |
bio->bi_write_hint = write_hint; |
1da177e4c
|
2989 |
|
6cf66b4ca
|
2990 2991 |
bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh)); BUG_ON(bio->bi_iter.bi_size != bh->b_size); |
1da177e4c
|
2992 2993 2994 |
bio->bi_end_io = end_bio_bh_io_sync; bio->bi_private = bh; |
877f962c5
|
2995 |
if (buffer_meta(bh)) |
2a222ca99
|
2996 |
op_flags |= REQ_META; |
877f962c5
|
2997 |
if (buffer_prio(bh)) |
2a222ca99
|
2998 2999 |
op_flags |= REQ_PRIO; bio_set_op_attrs(bio, op, op_flags); |
877f962c5
|
3000 |
|
3fe209c84
|
3001 3002 |
/* Take care of bh's that straddle the end of the device */ guard_bio_eod(bio); |
fd42df305
|
3003 3004 |
if (wbc) { wbc_init_bio(wbc, bio); |
34e51a5e1
|
3005 |
wbc_account_cgroup_owner(wbc, bh->b_page, bh->b_size); |
fd42df305
|
3006 |
} |
4e49ea4a3
|
3007 |
submit_bio(bio); |
f6454b049
|
3008 |
return 0; |
1da177e4c
|
3009 |
} |
bafc0dba1
|
3010 |
|
020c2833d
|
3011 |
int submit_bh(int op, int op_flags, struct buffer_head *bh) |
bafc0dba1
|
3012 |
{ |
8e8f92988
|
3013 |
return submit_bh_wbc(op, op_flags, bh, 0, NULL); |
713685111
|
3014 |
} |
1fe72eaa0
|
3015 |
EXPORT_SYMBOL(submit_bh); |
1da177e4c
|
3016 3017 3018 |
/** * ll_rw_block: low-level access to block devices (DEPRECATED) |
dfec8a14f
|
3019 |
* @op: whether to %READ or %WRITE |
ef295ecf0
|
3020 |
* @op_flags: req_flag_bits |
1da177e4c
|
3021 3022 3023 |
* @nr: number of &struct buffer_heads in the array * @bhs: array of pointers to &struct buffer_head * |
a76622362
|
3024 |
* ll_rw_block() takes an array of pointers to &struct buffer_heads, and |
70246286e
|
3025 3026 3027 |
* requests an I/O operation on them, either a %REQ_OP_READ or a %REQ_OP_WRITE. * @op_flags contains flags modifying the detailed I/O behavior, most notably * %REQ_RAHEAD. |
1da177e4c
|
3028 3029 |
* * This function drops any buffer that it cannot get a lock on (with the |
9cb569d60
|
3030 3031 3032 3033 3034 |
* BH_Lock state bit), any buffer that appears to be clean when doing a write * request, and any buffer that appears to be up-to-date when doing read * request. Further it marks as clean buffers that are processed for * writing (the buffer cache won't assume that they are actually clean * until the buffer gets unlocked). |
1da177e4c
|
3035 3036 |
* * ll_rw_block sets b_end_io to simple completion handler that marks |
e227867f1
|
3037 |
* the buffer up-to-date (if appropriate), unlocks the buffer and wakes |
1da177e4c
|
3038 3039 3040 3041 3042 |
* any waiters. * * All of the buffers must be for the same device, and must also be a * multiple of the current approved size for the device. */ |
dfec8a14f
|
3043 |
void ll_rw_block(int op, int op_flags, int nr, struct buffer_head *bhs[]) |
1da177e4c
|
3044 3045 3046 3047 3048 |
{ int i; for (i = 0; i < nr; i++) { struct buffer_head *bh = bhs[i]; |
9cb569d60
|
3049 |
if (!trylock_buffer(bh)) |
1da177e4c
|
3050 |
continue; |
dfec8a14f
|
3051 |
if (op == WRITE) { |
1da177e4c
|
3052 |
if (test_clear_buffer_dirty(bh)) { |
76c3073a8
|
3053 |
bh->b_end_io = end_buffer_write_sync; |
e60e5c50a
|
3054 |
get_bh(bh); |
dfec8a14f
|
3055 |
submit_bh(op, op_flags, bh); |
1da177e4c
|
3056 3057 3058 |
continue; } } else { |
1da177e4c
|
3059 |
if (!buffer_uptodate(bh)) { |
76c3073a8
|
3060 |
bh->b_end_io = end_buffer_read_sync; |
e60e5c50a
|
3061 |
get_bh(bh); |
dfec8a14f
|
3062 |
submit_bh(op, op_flags, bh); |
1da177e4c
|
3063 3064 3065 3066 |
continue; } } unlock_buffer(bh); |
1da177e4c
|
3067 3068 |
} } |
1fe72eaa0
|
3069 |
EXPORT_SYMBOL(ll_rw_block); |
1da177e4c
|
3070 |
|
2a222ca99
|
3071 |
void write_dirty_buffer(struct buffer_head *bh, int op_flags) |
9cb569d60
|
3072 3073 3074 3075 3076 3077 3078 3079 |
{ lock_buffer(bh); if (!test_clear_buffer_dirty(bh)) { unlock_buffer(bh); return; } bh->b_end_io = end_buffer_write_sync; get_bh(bh); |
2a222ca99
|
3080 |
submit_bh(REQ_OP_WRITE, op_flags, bh); |
9cb569d60
|
3081 3082 |
} EXPORT_SYMBOL(write_dirty_buffer); |
1da177e4c
|
3083 3084 3085 3086 3087 |
/* * For a data-integrity writeout, we need to wait upon any in-progress I/O * and then start new I/O and then wait upon it. The caller must have a ref on * the buffer_head. */ |
2a222ca99
|
3088 |
int __sync_dirty_buffer(struct buffer_head *bh, int op_flags) |
1da177e4c
|
3089 3090 3091 3092 3093 3094 3095 3096 |
{ int ret = 0; WARN_ON(atomic_read(&bh->b_count) < 1); lock_buffer(bh); if (test_clear_buffer_dirty(bh)) { get_bh(bh); bh->b_end_io = end_buffer_write_sync; |
2a222ca99
|
3097 |
ret = submit_bh(REQ_OP_WRITE, op_flags, bh); |
1da177e4c
|
3098 |
wait_on_buffer(bh); |
1da177e4c
|
3099 3100 3101 3102 3103 3104 3105 |
if (!ret && !buffer_uptodate(bh)) ret = -EIO; } else { unlock_buffer(bh); } return ret; } |
87e99511e
|
3106 3107 3108 3109 |
EXPORT_SYMBOL(__sync_dirty_buffer); int sync_dirty_buffer(struct buffer_head *bh) { |
70fd76140
|
3110 |
return __sync_dirty_buffer(bh, REQ_SYNC); |
87e99511e
|
3111 |
} |
1fe72eaa0
|
3112 |
EXPORT_SYMBOL(sync_dirty_buffer); |
1da177e4c
|
3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 |
/* * try_to_free_buffers() checks if all the buffers on this particular page * are unused, and releases them if so. * * Exclusion against try_to_free_buffers may be obtained by either * locking the page or by holding its mapping's private_lock. * * If the page is dirty but all the buffers are clean then we need to * be sure to mark the page clean as well. This is because the page * may be against a block device, and a later reattachment of buffers * to a dirty page will set *all* buffers dirty. Which would corrupt * filesystem data on the same device. * * The same applies to regular filesystem pages: if all the buffers are * clean then we set the page clean and proceed. To do that, we require * total exclusion from __set_page_dirty_buffers(). That is obtained with * private_lock. * * try_to_free_buffers() is non-blocking. */ static inline int buffer_busy(struct buffer_head *bh) { return atomic_read(&bh->b_count) | (bh->b_state & ((1 << BH_Dirty) | (1 << BH_Lock))); } static int drop_buffers(struct page *page, struct buffer_head **buffers_to_free) { struct buffer_head *head = page_buffers(page); struct buffer_head *bh; bh = head; do { |
1da177e4c
|
3148 3149 3150 3151 3152 3153 3154 |
if (buffer_busy(bh)) goto failed; bh = bh->b_this_page; } while (bh != head); do { struct buffer_head *next = bh->b_this_page; |
535ee2fbf
|
3155 |
if (bh->b_assoc_map) |
1da177e4c
|
3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 |
__remove_assoc_queue(bh); bh = next; } while (bh != head); *buffers_to_free = head; __clear_page_buffers(page); return 1; failed: return 0; } int try_to_free_buffers(struct page *page) { struct address_space * const mapping = page->mapping; struct buffer_head *buffers_to_free = NULL; int ret = 0; BUG_ON(!PageLocked(page)); |
ecdfc9787
|
3173 |
if (PageWriteback(page)) |
1da177e4c
|
3174 3175 3176 3177 3178 3179 3180 3181 3182 |
return 0; if (mapping == NULL) { /* can this still happen? */ ret = drop_buffers(page, &buffers_to_free); goto out; } spin_lock(&mapping->private_lock); ret = drop_buffers(page, &buffers_to_free); |
ecdfc9787
|
3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 |
/* * If the filesystem writes its buffers by hand (eg ext3) * then we can have clean buffers against a dirty page. We * clean the page here; otherwise the VM will never notice * that the filesystem did any IO at all. * * Also, during truncate, discard_buffer will have marked all * the page's buffers clean. We discover that here and clean * the page also. |
87df7241b
|
3193 3194 3195 3196 |
* * private_lock must be held over this entire operation in order * to synchronise against __set_page_dirty_buffers and prevent the * dirty bit from being lost. |
ecdfc9787
|
3197 |
*/ |
11f81becc
|
3198 3199 |
if (ret) cancel_dirty_page(page); |
87df7241b
|
3200 |
spin_unlock(&mapping->private_lock); |
1da177e4c
|
3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 |
out: if (buffers_to_free) { struct buffer_head *bh = buffers_to_free; do { struct buffer_head *next = bh->b_this_page; free_buffer_head(bh); bh = next; } while (bh != buffers_to_free); } return ret; } EXPORT_SYMBOL(try_to_free_buffers); |
1da177e4c
|
3214 3215 3216 3217 3218 |
/* * There are no bdflush tunables left. But distributions are * still running obsolete flush daemons, so we terminate them here. * * Use of bdflush() is deprecated and will be removed in a future kernel. |
5b0830cb9
|
3219 |
* The `flush-X' kernel threads fully replace bdflush daemons and this call. |
1da177e4c
|
3220 |
*/ |
bdc480e3b
|
3221 |
SYSCALL_DEFINE2(bdflush, int, func, long, data) |
1da177e4c
|
3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 |
{ static int msg_count; if (!capable(CAP_SYS_ADMIN)) return -EPERM; if (msg_count < 5) { msg_count++; printk(KERN_INFO "warning: process `%s' used the obsolete bdflush" " system call ", current->comm); printk(KERN_INFO "Fix your initscripts? "); } if (func == 1) do_exit(0); return 0; } /* * Buffer-head allocation */ |
a0a9b0433
|
3246 |
static struct kmem_cache *bh_cachep __read_mostly; |
1da177e4c
|
3247 3248 3249 3250 3251 |
/* * Once the number of bh's in the machine exceeds this level, we start * stripping them in writeback. */ |
43be594a6
|
3252 |
static unsigned long max_buffer_heads; |
1da177e4c
|
3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 |
int buffer_heads_over_limit; struct bh_accounting { int nr; /* Number of live bh's */ int ratelimit; /* Limit cacheline bouncing */ }; static DEFINE_PER_CPU(struct bh_accounting, bh_accounting) = {0, 0}; static void recalc_bh_state(void) { int i; int tot = 0; |
ee1be8626
|
3267 |
if (__this_cpu_inc_return(bh_accounting.ratelimit) - 1 < 4096) |
1da177e4c
|
3268 |
return; |
c7b92516a
|
3269 |
__this_cpu_write(bh_accounting.ratelimit, 0); |
8a1434268
|
3270 |
for_each_online_cpu(i) |
1da177e4c
|
3271 3272 3273 |
tot += per_cpu(bh_accounting, i).nr; buffer_heads_over_limit = (tot > max_buffer_heads); } |
c7b92516a
|
3274 |
|
dd0fc66fb
|
3275 |
struct buffer_head *alloc_buffer_head(gfp_t gfp_flags) |
1da177e4c
|
3276 |
{ |
019b4d123
|
3277 |
struct buffer_head *ret = kmem_cache_zalloc(bh_cachep, gfp_flags); |
1da177e4c
|
3278 |
if (ret) { |
a35afb830
|
3279 |
INIT_LIST_HEAD(&ret->b_assoc_buffers); |
c7b92516a
|
3280 3281 |
preempt_disable(); __this_cpu_inc(bh_accounting.nr); |
1da177e4c
|
3282 |
recalc_bh_state(); |
c7b92516a
|
3283 |
preempt_enable(); |
1da177e4c
|
3284 3285 3286 3287 3288 3289 3290 3291 3292 |
} return ret; } EXPORT_SYMBOL(alloc_buffer_head); void free_buffer_head(struct buffer_head *bh) { BUG_ON(!list_empty(&bh->b_assoc_buffers)); kmem_cache_free(bh_cachep, bh); |
c7b92516a
|
3293 3294 |
preempt_disable(); __this_cpu_dec(bh_accounting.nr); |
1da177e4c
|
3295 |
recalc_bh_state(); |
c7b92516a
|
3296 |
preempt_enable(); |
1da177e4c
|
3297 3298 |
} EXPORT_SYMBOL(free_buffer_head); |
fc4d24c9b
|
3299 |
static int buffer_exit_cpu_dead(unsigned int cpu) |
1da177e4c
|
3300 3301 3302 3303 3304 3305 3306 3307 |
{ int i; struct bh_lru *b = &per_cpu(bh_lrus, cpu); for (i = 0; i < BH_LRU_SIZE; i++) { brelse(b->bhs[i]); b->bhs[i] = NULL; } |
c7b92516a
|
3308 |
this_cpu_add(bh_accounting.nr, per_cpu(bh_accounting, cpu).nr); |
8a1434268
|
3309 |
per_cpu(bh_accounting, cpu).nr = 0; |
fc4d24c9b
|
3310 |
return 0; |
1da177e4c
|
3311 |
} |
1da177e4c
|
3312 |
|
389d1b083
|
3313 |
/** |
a6b91919e
|
3314 |
* bh_uptodate_or_lock - Test whether the buffer is uptodate |
389d1b083
|
3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 |
* @bh: struct buffer_head * * Return true if the buffer is up-to-date and false, * with the buffer locked, if not. */ int bh_uptodate_or_lock(struct buffer_head *bh) { if (!buffer_uptodate(bh)) { lock_buffer(bh); if (!buffer_uptodate(bh)) return 0; unlock_buffer(bh); } return 1; } EXPORT_SYMBOL(bh_uptodate_or_lock); /** |
a6b91919e
|
3333 |
* bh_submit_read - Submit a locked buffer for reading |
389d1b083
|
3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 |
* @bh: struct buffer_head * * Returns zero on success and -EIO on error. */ int bh_submit_read(struct buffer_head *bh) { BUG_ON(!buffer_locked(bh)); if (buffer_uptodate(bh)) { unlock_buffer(bh); return 0; } get_bh(bh); bh->b_end_io = end_buffer_read_sync; |
2a222ca99
|
3349 |
submit_bh(REQ_OP_READ, 0, bh); |
389d1b083
|
3350 3351 3352 3353 3354 3355 |
wait_on_buffer(bh); if (buffer_uptodate(bh)) return 0; return -EIO; } EXPORT_SYMBOL(bh_submit_read); |
1da177e4c
|
3356 3357 |
void __init buffer_init(void) { |
43be594a6
|
3358 |
unsigned long nrpages; |
fc4d24c9b
|
3359 |
int ret; |
1da177e4c
|
3360 |
|
b98938c37
|
3361 3362 3363 3364 |
bh_cachep = kmem_cache_create("buffer_head", sizeof(struct buffer_head), 0, (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC| SLAB_MEM_SPREAD), |
019b4d123
|
3365 |
NULL); |
1da177e4c
|
3366 3367 3368 3369 3370 3371 |
/* * Limit the bh occupancy to 10% of ZONE_NORMAL */ nrpages = (nr_free_buffer_pages() * 10) / 100; max_buffer_heads = nrpages * (PAGE_SIZE / sizeof(struct buffer_head)); |
fc4d24c9b
|
3372 3373 3374 |
ret = cpuhp_setup_state_nocalls(CPUHP_FS_BUFF_DEAD, "fs/buffer:dead", NULL, buffer_exit_cpu_dead); WARN_ON(ret < 0); |
1da177e4c
|
3375 |
} |