]>
Commit | Line | Data |
---|---|---|
0f9dd46c JB |
1 | /* |
2 | * Copyright (C) 2008 Red Hat. All rights reserved. | |
3 | * | |
4 | * This program is free software; you can redistribute it and/or | |
5 | * modify it under the terms of the GNU General Public | |
6 | * License v2 as published by the Free Software Foundation. | |
7 | * | |
8 | * This program is distributed in the hope that it will be useful, | |
9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
11 | * General Public License for more details. | |
12 | * | |
13 | * You should have received a copy of the GNU General Public | |
14 | * License along with this program; if not, write to the | |
15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
16 | * Boston, MA 021110-1307, USA. | |
17 | */ | |
18 | ||
96303081 | 19 | #include <linux/pagemap.h> |
0f9dd46c | 20 | #include <linux/sched.h> |
96303081 | 21 | #include <linux/math64.h> |
0f9dd46c | 22 | #include "ctree.h" |
fa9c0d79 CM |
23 | #include "free-space-cache.h" |
24 | #include "transaction.h" | |
25 | ||
96303081 JB |
26 | #define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8) |
27 | #define MAX_CACHE_BYTES_PER_GIG (32 * 1024) | |
0f9dd46c | 28 | |
96303081 JB |
29 | static inline unsigned long offset_to_bit(u64 bitmap_start, u64 sectorsize, |
30 | u64 offset) | |
0f9dd46c | 31 | { |
96303081 JB |
32 | BUG_ON(offset < bitmap_start); |
33 | offset -= bitmap_start; | |
34 | return (unsigned long)(div64_u64(offset, sectorsize)); | |
35 | } | |
0f9dd46c | 36 | |
96303081 JB |
37 | static inline unsigned long bytes_to_bits(u64 bytes, u64 sectorsize) |
38 | { | |
39 | return (unsigned long)(div64_u64(bytes, sectorsize)); | |
40 | } | |
0f9dd46c | 41 | |
96303081 JB |
42 | static inline u64 offset_to_bitmap(struct btrfs_block_group_cache *block_group, |
43 | u64 offset) | |
44 | { | |
45 | u64 bitmap_start; | |
46 | u64 bytes_per_bitmap; | |
0f9dd46c | 47 | |
96303081 JB |
48 | bytes_per_bitmap = BITS_PER_BITMAP * block_group->sectorsize; |
49 | bitmap_start = offset - block_group->key.objectid; | |
50 | bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap); | |
51 | bitmap_start *= bytes_per_bitmap; | |
52 | bitmap_start += block_group->key.objectid; | |
0f9dd46c | 53 | |
96303081 | 54 | return bitmap_start; |
0f9dd46c JB |
55 | } |
56 | ||
96303081 JB |
57 | static int tree_insert_offset(struct rb_root *root, u64 offset, |
58 | struct rb_node *node, int bitmap) | |
0f9dd46c JB |
59 | { |
60 | struct rb_node **p = &root->rb_node; | |
61 | struct rb_node *parent = NULL; | |
62 | struct btrfs_free_space *info; | |
63 | ||
64 | while (*p) { | |
65 | parent = *p; | |
96303081 | 66 | info = rb_entry(parent, struct btrfs_free_space, offset_index); |
0f9dd46c | 67 | |
96303081 | 68 | if (offset < info->offset) { |
0f9dd46c | 69 | p = &(*p)->rb_left; |
96303081 | 70 | } else if (offset > info->offset) { |
0f9dd46c | 71 | p = &(*p)->rb_right; |
96303081 JB |
72 | } else { |
73 | /* | |
74 | * we could have a bitmap entry and an extent entry | |
75 | * share the same offset. If this is the case, we want | |
76 | * the extent entry to always be found first if we do a | |
77 | * linear search through the tree, since we want to have | |
78 | * the quickest allocation time, and allocating from an | |
79 | * extent is faster than allocating from a bitmap. So | |
80 | * if we're inserting a bitmap and we find an entry at | |
81 | * this offset, we want to go right, or after this entry | |
82 | * logically. If we are inserting an extent and we've | |
83 | * found a bitmap, we want to go left, or before | |
84 | * logically. | |
85 | */ | |
86 | if (bitmap) { | |
87 | WARN_ON(info->bitmap); | |
88 | p = &(*p)->rb_right; | |
89 | } else { | |
90 | WARN_ON(!info->bitmap); | |
91 | p = &(*p)->rb_left; | |
92 | } | |
93 | } | |
0f9dd46c JB |
94 | } |
95 | ||
96 | rb_link_node(node, parent, p); | |
97 | rb_insert_color(node, root); | |
98 | ||
99 | return 0; | |
100 | } | |
101 | ||
102 | /* | |
70cb0743 JB |
103 | * searches the tree for the given offset. |
104 | * | |
96303081 JB |
105 | * fuzzy - If this is set, then we are trying to make an allocation, and we just |
106 | * want a section that has at least bytes size and comes at or after the given | |
107 | * offset. | |
0f9dd46c | 108 | */ |
96303081 JB |
109 | static struct btrfs_free_space * |
110 | tree_search_offset(struct btrfs_block_group_cache *block_group, | |
111 | u64 offset, int bitmap_only, int fuzzy) | |
0f9dd46c | 112 | { |
96303081 JB |
113 | struct rb_node *n = block_group->free_space_offset.rb_node; |
114 | struct btrfs_free_space *entry, *prev = NULL; | |
115 | ||
116 | /* find entry that is closest to the 'offset' */ | |
117 | while (1) { | |
118 | if (!n) { | |
119 | entry = NULL; | |
120 | break; | |
121 | } | |
0f9dd46c | 122 | |
0f9dd46c | 123 | entry = rb_entry(n, struct btrfs_free_space, offset_index); |
96303081 | 124 | prev = entry; |
0f9dd46c | 125 | |
96303081 | 126 | if (offset < entry->offset) |
0f9dd46c | 127 | n = n->rb_left; |
96303081 | 128 | else if (offset > entry->offset) |
0f9dd46c | 129 | n = n->rb_right; |
96303081 | 130 | else |
0f9dd46c | 131 | break; |
0f9dd46c JB |
132 | } |
133 | ||
96303081 JB |
134 | if (bitmap_only) { |
135 | if (!entry) | |
136 | return NULL; | |
137 | if (entry->bitmap) | |
138 | return entry; | |
0f9dd46c | 139 | |
96303081 JB |
140 | /* |
141 | * bitmap entry and extent entry may share same offset, | |
142 | * in that case, bitmap entry comes after extent entry. | |
143 | */ | |
144 | n = rb_next(n); | |
145 | if (!n) | |
146 | return NULL; | |
147 | entry = rb_entry(n, struct btrfs_free_space, offset_index); | |
148 | if (entry->offset != offset) | |
149 | return NULL; | |
0f9dd46c | 150 | |
96303081 JB |
151 | WARN_ON(!entry->bitmap); |
152 | return entry; | |
153 | } else if (entry) { | |
154 | if (entry->bitmap) { | |
0f9dd46c | 155 | /* |
96303081 JB |
156 | * if previous extent entry covers the offset, |
157 | * we should return it instead of the bitmap entry | |
0f9dd46c | 158 | */ |
96303081 JB |
159 | n = &entry->offset_index; |
160 | while (1) { | |
161 | n = rb_prev(n); | |
162 | if (!n) | |
163 | break; | |
164 | prev = rb_entry(n, struct btrfs_free_space, | |
165 | offset_index); | |
166 | if (!prev->bitmap) { | |
167 | if (prev->offset + prev->bytes > offset) | |
168 | entry = prev; | |
169 | break; | |
170 | } | |
0f9dd46c | 171 | } |
96303081 JB |
172 | } |
173 | return entry; | |
174 | } | |
175 | ||
176 | if (!prev) | |
177 | return NULL; | |
178 | ||
179 | /* find last entry before the 'offset' */ | |
180 | entry = prev; | |
181 | if (entry->offset > offset) { | |
182 | n = rb_prev(&entry->offset_index); | |
183 | if (n) { | |
184 | entry = rb_entry(n, struct btrfs_free_space, | |
185 | offset_index); | |
186 | BUG_ON(entry->offset > offset); | |
0f9dd46c | 187 | } else { |
96303081 JB |
188 | if (fuzzy) |
189 | return entry; | |
190 | else | |
191 | return NULL; | |
0f9dd46c JB |
192 | } |
193 | } | |
194 | ||
96303081 JB |
195 | if (entry->bitmap) { |
196 | n = &entry->offset_index; | |
197 | while (1) { | |
198 | n = rb_prev(n); | |
199 | if (!n) | |
200 | break; | |
201 | prev = rb_entry(n, struct btrfs_free_space, | |
202 | offset_index); | |
203 | if (!prev->bitmap) { | |
204 | if (prev->offset + prev->bytes > offset) | |
205 | return prev; | |
206 | break; | |
207 | } | |
208 | } | |
209 | if (entry->offset + BITS_PER_BITMAP * | |
210 | block_group->sectorsize > offset) | |
211 | return entry; | |
212 | } else if (entry->offset + entry->bytes > offset) | |
213 | return entry; | |
214 | ||
215 | if (!fuzzy) | |
216 | return NULL; | |
217 | ||
218 | while (1) { | |
219 | if (entry->bitmap) { | |
220 | if (entry->offset + BITS_PER_BITMAP * | |
221 | block_group->sectorsize > offset) | |
222 | break; | |
223 | } else { | |
224 | if (entry->offset + entry->bytes > offset) | |
225 | break; | |
226 | } | |
227 | ||
228 | n = rb_next(&entry->offset_index); | |
229 | if (!n) | |
230 | return NULL; | |
231 | entry = rb_entry(n, struct btrfs_free_space, offset_index); | |
232 | } | |
233 | return entry; | |
0f9dd46c JB |
234 | } |
235 | ||
236 | static void unlink_free_space(struct btrfs_block_group_cache *block_group, | |
237 | struct btrfs_free_space *info) | |
238 | { | |
239 | rb_erase(&info->offset_index, &block_group->free_space_offset); | |
96303081 | 240 | block_group->free_extents--; |
817d52f8 | 241 | block_group->free_space -= info->bytes; |
0f9dd46c JB |
242 | } |
243 | ||
244 | static int link_free_space(struct btrfs_block_group_cache *block_group, | |
245 | struct btrfs_free_space *info) | |
246 | { | |
247 | int ret = 0; | |
248 | ||
96303081 | 249 | BUG_ON(!info->bitmap && !info->bytes); |
0f9dd46c | 250 | ret = tree_insert_offset(&block_group->free_space_offset, info->offset, |
96303081 | 251 | &info->offset_index, (info->bitmap != NULL)); |
0f9dd46c JB |
252 | if (ret) |
253 | return ret; | |
254 | ||
817d52f8 | 255 | block_group->free_space += info->bytes; |
96303081 JB |
256 | block_group->free_extents++; |
257 | return ret; | |
258 | } | |
259 | ||
260 | static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) | |
261 | { | |
262 | u64 max_bytes, possible_bytes; | |
263 | ||
264 | /* | |
265 | * The goal is to keep the total amount of memory used per 1gb of space | |
266 | * at or below 32k, so we need to adjust how much memory we allow to be | |
267 | * used by extent based free space tracking | |
268 | */ | |
269 | max_bytes = MAX_CACHE_BYTES_PER_GIG * | |
270 | (div64_u64(block_group->key.offset, 1024 * 1024 * 1024)); | |
271 | ||
272 | possible_bytes = (block_group->total_bitmaps * PAGE_CACHE_SIZE) + | |
273 | (sizeof(struct btrfs_free_space) * | |
274 | block_group->extents_thresh); | |
275 | ||
276 | if (possible_bytes > max_bytes) { | |
277 | int extent_bytes = max_bytes - | |
278 | (block_group->total_bitmaps * PAGE_CACHE_SIZE); | |
279 | ||
280 | if (extent_bytes <= 0) { | |
281 | block_group->extents_thresh = 0; | |
282 | return; | |
283 | } | |
284 | ||
285 | block_group->extents_thresh = extent_bytes / | |
286 | (sizeof(struct btrfs_free_space)); | |
287 | } | |
288 | } | |
289 | ||
817d52f8 JB |
290 | static void bitmap_clear_bits(struct btrfs_block_group_cache *block_group, |
291 | struct btrfs_free_space *info, u64 offset, | |
292 | u64 bytes) | |
96303081 JB |
293 | { |
294 | unsigned long start, end; | |
295 | unsigned long i; | |
296 | ||
817d52f8 JB |
297 | start = offset_to_bit(info->offset, block_group->sectorsize, offset); |
298 | end = start + bytes_to_bits(bytes, block_group->sectorsize); | |
96303081 JB |
299 | BUG_ON(end > BITS_PER_BITMAP); |
300 | ||
301 | for (i = start; i < end; i++) | |
302 | clear_bit(i, info->bitmap); | |
303 | ||
304 | info->bytes -= bytes; | |
817d52f8 | 305 | block_group->free_space -= bytes; |
96303081 JB |
306 | } |
307 | ||
817d52f8 JB |
308 | static void bitmap_set_bits(struct btrfs_block_group_cache *block_group, |
309 | struct btrfs_free_space *info, u64 offset, | |
310 | u64 bytes) | |
96303081 JB |
311 | { |
312 | unsigned long start, end; | |
313 | unsigned long i; | |
314 | ||
817d52f8 JB |
315 | start = offset_to_bit(info->offset, block_group->sectorsize, offset); |
316 | end = start + bytes_to_bits(bytes, block_group->sectorsize); | |
96303081 JB |
317 | BUG_ON(end > BITS_PER_BITMAP); |
318 | ||
319 | for (i = start; i < end; i++) | |
320 | set_bit(i, info->bitmap); | |
321 | ||
322 | info->bytes += bytes; | |
817d52f8 | 323 | block_group->free_space += bytes; |
96303081 JB |
324 | } |
325 | ||
326 | static int search_bitmap(struct btrfs_block_group_cache *block_group, | |
327 | struct btrfs_free_space *bitmap_info, u64 *offset, | |
328 | u64 *bytes) | |
329 | { | |
330 | unsigned long found_bits = 0; | |
331 | unsigned long bits, i; | |
332 | unsigned long next_zero; | |
333 | ||
334 | i = offset_to_bit(bitmap_info->offset, block_group->sectorsize, | |
335 | max_t(u64, *offset, bitmap_info->offset)); | |
336 | bits = bytes_to_bits(*bytes, block_group->sectorsize); | |
337 | ||
338 | for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i); | |
339 | i < BITS_PER_BITMAP; | |
340 | i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i + 1)) { | |
341 | next_zero = find_next_zero_bit(bitmap_info->bitmap, | |
342 | BITS_PER_BITMAP, i); | |
343 | if ((next_zero - i) >= bits) { | |
344 | found_bits = next_zero - i; | |
345 | break; | |
346 | } | |
347 | i = next_zero; | |
348 | } | |
349 | ||
350 | if (found_bits) { | |
351 | *offset = (u64)(i * block_group->sectorsize) + | |
352 | bitmap_info->offset; | |
353 | *bytes = (u64)(found_bits) * block_group->sectorsize; | |
354 | return 0; | |
355 | } | |
356 | ||
357 | return -1; | |
358 | } | |
359 | ||
360 | static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache | |
361 | *block_group, u64 *offset, | |
362 | u64 *bytes, int debug) | |
363 | { | |
364 | struct btrfs_free_space *entry; | |
365 | struct rb_node *node; | |
366 | int ret; | |
367 | ||
368 | if (!block_group->free_space_offset.rb_node) | |
369 | return NULL; | |
370 | ||
371 | entry = tree_search_offset(block_group, | |
372 | offset_to_bitmap(block_group, *offset), | |
373 | 0, 1); | |
374 | if (!entry) | |
375 | return NULL; | |
376 | ||
377 | for (node = &entry->offset_index; node; node = rb_next(node)) { | |
378 | entry = rb_entry(node, struct btrfs_free_space, offset_index); | |
379 | if (entry->bytes < *bytes) | |
380 | continue; | |
381 | ||
382 | if (entry->bitmap) { | |
383 | ret = search_bitmap(block_group, entry, offset, bytes); | |
384 | if (!ret) | |
385 | return entry; | |
386 | continue; | |
387 | } | |
388 | ||
389 | *offset = entry->offset; | |
390 | *bytes = entry->bytes; | |
391 | return entry; | |
392 | } | |
393 | ||
394 | return NULL; | |
395 | } | |
396 | ||
397 | static void add_new_bitmap(struct btrfs_block_group_cache *block_group, | |
398 | struct btrfs_free_space *info, u64 offset) | |
399 | { | |
400 | u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize; | |
401 | int max_bitmaps = (int)div64_u64(block_group->key.offset + | |
402 | bytes_per_bg - 1, bytes_per_bg); | |
403 | BUG_ON(block_group->total_bitmaps >= max_bitmaps); | |
404 | ||
405 | info->offset = offset_to_bitmap(block_group, offset); | |
406 | link_free_space(block_group, info); | |
407 | block_group->total_bitmaps++; | |
408 | ||
409 | recalculate_thresholds(block_group); | |
410 | } | |
411 | ||
412 | static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_group, | |
413 | struct btrfs_free_space *bitmap_info, | |
414 | u64 *offset, u64 *bytes) | |
415 | { | |
416 | u64 end; | |
417 | ||
418 | again: | |
419 | end = bitmap_info->offset + | |
420 | (u64)(BITS_PER_BITMAP * block_group->sectorsize) - 1; | |
421 | ||
422 | if (*offset > bitmap_info->offset && *offset + *bytes > end) { | |
817d52f8 JB |
423 | bitmap_clear_bits(block_group, bitmap_info, *offset, |
424 | end - *offset + 1); | |
96303081 JB |
425 | *bytes -= end - *offset + 1; |
426 | *offset = end + 1; | |
427 | } else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) { | |
817d52f8 | 428 | bitmap_clear_bits(block_group, bitmap_info, *offset, *bytes); |
96303081 JB |
429 | *bytes = 0; |
430 | } | |
431 | ||
432 | if (*bytes) { | |
433 | if (!bitmap_info->bytes) { | |
434 | unlink_free_space(block_group, bitmap_info); | |
435 | kfree(bitmap_info->bitmap); | |
436 | kfree(bitmap_info); | |
437 | block_group->total_bitmaps--; | |
438 | recalculate_thresholds(block_group); | |
439 | } | |
440 | ||
441 | bitmap_info = tree_search_offset(block_group, | |
442 | offset_to_bitmap(block_group, | |
443 | *offset), | |
444 | 1, 0); | |
445 | if (!bitmap_info) | |
446 | return -EINVAL; | |
447 | ||
448 | if (!bitmap_info->bitmap) | |
449 | return -EAGAIN; | |
450 | ||
451 | goto again; | |
452 | } else if (!bitmap_info->bytes) { | |
453 | unlink_free_space(block_group, bitmap_info); | |
454 | kfree(bitmap_info->bitmap); | |
455 | kfree(bitmap_info); | |
456 | block_group->total_bitmaps--; | |
457 | recalculate_thresholds(block_group); | |
458 | } | |
459 | ||
460 | return 0; | |
461 | } | |
462 | ||
463 | static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, | |
464 | struct btrfs_free_space *info) | |
465 | { | |
466 | struct btrfs_free_space *bitmap_info; | |
467 | int added = 0; | |
468 | u64 bytes, offset, end; | |
469 | int ret; | |
470 | ||
471 | /* | |
472 | * If we are below the extents threshold then we can add this as an | |
473 | * extent, and don't have to deal with the bitmap | |
474 | */ | |
475 | if (block_group->free_extents < block_group->extents_thresh && | |
476 | info->bytes > block_group->sectorsize * 4) | |
477 | return 0; | |
478 | ||
479 | /* | |
480 | * some block groups are so tiny they can't be enveloped by a bitmap, so | |
481 | * don't even bother to create a bitmap for this | |
482 | */ | |
483 | if (BITS_PER_BITMAP * block_group->sectorsize > | |
484 | block_group->key.offset) | |
485 | return 0; | |
486 | ||
487 | bytes = info->bytes; | |
488 | offset = info->offset; | |
489 | ||
490 | again: | |
491 | bitmap_info = tree_search_offset(block_group, | |
492 | offset_to_bitmap(block_group, offset), | |
493 | 1, 0); | |
494 | if (!bitmap_info) { | |
495 | BUG_ON(added); | |
496 | goto new_bitmap; | |
497 | } | |
498 | ||
499 | end = bitmap_info->offset + | |
500 | (u64)(BITS_PER_BITMAP * block_group->sectorsize); | |
501 | ||
502 | if (offset >= bitmap_info->offset && offset + bytes > end) { | |
817d52f8 JB |
503 | bitmap_set_bits(block_group, bitmap_info, offset, |
504 | end - offset); | |
96303081 JB |
505 | bytes -= end - offset; |
506 | offset = end; | |
507 | added = 0; | |
508 | } else if (offset >= bitmap_info->offset && offset + bytes <= end) { | |
817d52f8 | 509 | bitmap_set_bits(block_group, bitmap_info, offset, bytes); |
96303081 JB |
510 | bytes = 0; |
511 | } else { | |
512 | BUG(); | |
513 | } | |
514 | ||
515 | if (!bytes) { | |
516 | ret = 1; | |
517 | goto out; | |
518 | } else | |
519 | goto again; | |
520 | ||
521 | new_bitmap: | |
522 | if (info && info->bitmap) { | |
523 | add_new_bitmap(block_group, info, offset); | |
524 | added = 1; | |
525 | info = NULL; | |
526 | goto again; | |
527 | } else { | |
528 | spin_unlock(&block_group->tree_lock); | |
529 | ||
530 | /* no pre-allocated info, allocate a new one */ | |
531 | if (!info) { | |
532 | info = kzalloc(sizeof(struct btrfs_free_space), | |
533 | GFP_NOFS); | |
534 | if (!info) { | |
535 | spin_lock(&block_group->tree_lock); | |
536 | ret = -ENOMEM; | |
537 | goto out; | |
538 | } | |
539 | } | |
540 | ||
541 | /* allocate the bitmap */ | |
542 | info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); | |
543 | spin_lock(&block_group->tree_lock); | |
544 | if (!info->bitmap) { | |
545 | ret = -ENOMEM; | |
546 | goto out; | |
547 | } | |
548 | goto again; | |
549 | } | |
550 | ||
551 | out: | |
552 | if (info) { | |
553 | if (info->bitmap) | |
554 | kfree(info->bitmap); | |
555 | kfree(info); | |
556 | } | |
0f9dd46c JB |
557 | |
558 | return ret; | |
559 | } | |
560 | ||
6226cb0a JB |
561 | int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, |
562 | u64 offset, u64 bytes) | |
0f9dd46c | 563 | { |
96303081 JB |
564 | struct btrfs_free_space *right_info = NULL; |
565 | struct btrfs_free_space *left_info = NULL; | |
0f9dd46c | 566 | struct btrfs_free_space *info = NULL; |
0f9dd46c JB |
567 | int ret = 0; |
568 | ||
6226cb0a JB |
569 | info = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS); |
570 | if (!info) | |
571 | return -ENOMEM; | |
572 | ||
573 | info->offset = offset; | |
574 | info->bytes = bytes; | |
575 | ||
576 | spin_lock(&block_group->tree_lock); | |
577 | ||
0f9dd46c JB |
578 | /* |
579 | * first we want to see if there is free space adjacent to the range we | |
580 | * are adding, if there is remove that struct and add a new one to | |
581 | * cover the entire range | |
582 | */ | |
96303081 JB |
583 | right_info = tree_search_offset(block_group, offset + bytes, 0, 0); |
584 | if (right_info && rb_prev(&right_info->offset_index)) | |
585 | left_info = rb_entry(rb_prev(&right_info->offset_index), | |
586 | struct btrfs_free_space, offset_index); | |
587 | else | |
588 | left_info = tree_search_offset(block_group, offset - 1, 0, 0); | |
0f9dd46c | 589 | |
96303081 JB |
590 | /* |
591 | * If there was no extent directly to the left or right of this new | |
592 | * extent then we know we're going to have to allocate a new extent, so | |
593 | * before we do that see if we need to drop this into a bitmap | |
594 | */ | |
595 | if ((!left_info || left_info->bitmap) && | |
596 | (!right_info || right_info->bitmap)) { | |
597 | ret = insert_into_bitmap(block_group, info); | |
598 | ||
599 | if (ret < 0) { | |
600 | goto out; | |
601 | } else if (ret) { | |
602 | ret = 0; | |
603 | goto out; | |
604 | } | |
605 | } | |
606 | ||
607 | if (right_info && !right_info->bitmap) { | |
0f9dd46c | 608 | unlink_free_space(block_group, right_info); |
6226cb0a JB |
609 | info->bytes += right_info->bytes; |
610 | kfree(right_info); | |
0f9dd46c JB |
611 | } |
612 | ||
96303081 JB |
613 | if (left_info && !left_info->bitmap && |
614 | left_info->offset + left_info->bytes == offset) { | |
0f9dd46c | 615 | unlink_free_space(block_group, left_info); |
6226cb0a JB |
616 | info->offset = left_info->offset; |
617 | info->bytes += left_info->bytes; | |
618 | kfree(left_info); | |
0f9dd46c JB |
619 | } |
620 | ||
0f9dd46c JB |
621 | ret = link_free_space(block_group, info); |
622 | if (ret) | |
623 | kfree(info); | |
96303081 | 624 | out: |
6226cb0a JB |
625 | spin_unlock(&block_group->tree_lock); |
626 | ||
0f9dd46c | 627 | if (ret) { |
96303081 | 628 | printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret); |
c293498b | 629 | BUG_ON(ret == -EEXIST); |
0f9dd46c JB |
630 | } |
631 | ||
0f9dd46c JB |
632 | return ret; |
633 | } | |
634 | ||
6226cb0a JB |
635 | int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, |
636 | u64 offset, u64 bytes) | |
0f9dd46c JB |
637 | { |
638 | struct btrfs_free_space *info; | |
96303081 | 639 | struct btrfs_free_space *next_info = NULL; |
0f9dd46c JB |
640 | int ret = 0; |
641 | ||
6226cb0a JB |
642 | spin_lock(&block_group->tree_lock); |
643 | ||
96303081 JB |
644 | again: |
645 | info = tree_search_offset(block_group, offset, 0, 0); | |
646 | if (!info) { | |
647 | WARN_ON(1); | |
648 | goto out_lock; | |
649 | } | |
650 | ||
651 | if (info->bytes < bytes && rb_next(&info->offset_index)) { | |
652 | u64 end; | |
653 | next_info = rb_entry(rb_next(&info->offset_index), | |
654 | struct btrfs_free_space, | |
655 | offset_index); | |
656 | ||
657 | if (next_info->bitmap) | |
658 | end = next_info->offset + BITS_PER_BITMAP * | |
659 | block_group->sectorsize - 1; | |
660 | else | |
661 | end = next_info->offset + next_info->bytes; | |
662 | ||
663 | if (next_info->bytes < bytes || | |
664 | next_info->offset > offset || offset > end) { | |
665 | printk(KERN_CRIT "Found free space at %llu, size %llu," | |
666 | " trying to use %llu\n", | |
667 | (unsigned long long)info->offset, | |
668 | (unsigned long long)info->bytes, | |
669 | (unsigned long long)bytes); | |
0f9dd46c JB |
670 | WARN_ON(1); |
671 | ret = -EINVAL; | |
96303081 | 672 | goto out_lock; |
0f9dd46c | 673 | } |
0f9dd46c | 674 | |
96303081 JB |
675 | info = next_info; |
676 | } | |
677 | ||
678 | if (info->bytes == bytes) { | |
679 | unlink_free_space(block_group, info); | |
680 | if (info->bitmap) { | |
681 | kfree(info->bitmap); | |
682 | block_group->total_bitmaps--; | |
0f9dd46c | 683 | } |
96303081 JB |
684 | kfree(info); |
685 | goto out_lock; | |
686 | } | |
0f9dd46c | 687 | |
96303081 JB |
688 | if (!info->bitmap && info->offset == offset) { |
689 | unlink_free_space(block_group, info); | |
0f9dd46c JB |
690 | info->offset += bytes; |
691 | info->bytes -= bytes; | |
96303081 JB |
692 | link_free_space(block_group, info); |
693 | goto out_lock; | |
694 | } | |
0f9dd46c | 695 | |
96303081 JB |
696 | if (!info->bitmap && info->offset <= offset && |
697 | info->offset + info->bytes >= offset + bytes) { | |
9b49c9b9 CM |
698 | u64 old_start = info->offset; |
699 | /* | |
700 | * we're freeing space in the middle of the info, | |
701 | * this can happen during tree log replay | |
702 | * | |
703 | * first unlink the old info and then | |
704 | * insert it again after the hole we're creating | |
705 | */ | |
706 | unlink_free_space(block_group, info); | |
707 | if (offset + bytes < info->offset + info->bytes) { | |
708 | u64 old_end = info->offset + info->bytes; | |
709 | ||
710 | info->offset = offset + bytes; | |
711 | info->bytes = old_end - info->offset; | |
712 | ret = link_free_space(block_group, info); | |
96303081 JB |
713 | WARN_ON(ret); |
714 | if (ret) | |
715 | goto out_lock; | |
9b49c9b9 CM |
716 | } else { |
717 | /* the hole we're creating ends at the end | |
718 | * of the info struct, just free the info | |
719 | */ | |
720 | kfree(info); | |
721 | } | |
6226cb0a | 722 | spin_unlock(&block_group->tree_lock); |
96303081 JB |
723 | |
724 | /* step two, insert a new info struct to cover | |
725 | * anything before the hole | |
9b49c9b9 | 726 | */ |
6226cb0a JB |
727 | ret = btrfs_add_free_space(block_group, old_start, |
728 | offset - old_start); | |
96303081 JB |
729 | WARN_ON(ret); |
730 | goto out; | |
0f9dd46c | 731 | } |
96303081 JB |
732 | |
733 | ret = remove_from_bitmap(block_group, info, &offset, &bytes); | |
734 | if (ret == -EAGAIN) | |
735 | goto again; | |
736 | BUG_ON(ret); | |
737 | out_lock: | |
738 | spin_unlock(&block_group->tree_lock); | |
0f9dd46c | 739 | out: |
25179201 JB |
740 | return ret; |
741 | } | |
742 | ||
0f9dd46c JB |
743 | void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, |
744 | u64 bytes) | |
745 | { | |
746 | struct btrfs_free_space *info; | |
747 | struct rb_node *n; | |
748 | int count = 0; | |
749 | ||
750 | for (n = rb_first(&block_group->free_space_offset); n; n = rb_next(n)) { | |
751 | info = rb_entry(n, struct btrfs_free_space, offset_index); | |
752 | if (info->bytes >= bytes) | |
753 | count++; | |
96303081 | 754 | printk(KERN_CRIT "entry offset %llu, bytes %llu, bitmap %s\n", |
21380931 | 755 | (unsigned long long)info->offset, |
96303081 JB |
756 | (unsigned long long)info->bytes, |
757 | (info->bitmap) ? "yes" : "no"); | |
0f9dd46c | 758 | } |
96303081 JB |
759 | printk(KERN_INFO "block group has cluster?: %s\n", |
760 | list_empty(&block_group->cluster_list) ? "no" : "yes"); | |
0f9dd46c JB |
761 | printk(KERN_INFO "%d blocks of free space at or bigger than bytes is" |
762 | "\n", count); | |
763 | } | |
764 | ||
765 | u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group) | |
766 | { | |
767 | struct btrfs_free_space *info; | |
768 | struct rb_node *n; | |
769 | u64 ret = 0; | |
770 | ||
771 | for (n = rb_first(&block_group->free_space_offset); n; | |
772 | n = rb_next(n)) { | |
773 | info = rb_entry(n, struct btrfs_free_space, offset_index); | |
774 | ret += info->bytes; | |
775 | } | |
776 | ||
777 | return ret; | |
778 | } | |
779 | ||
fa9c0d79 CM |
780 | /* |
781 | * for a given cluster, put all of its extents back into the free | |
782 | * space cache. If the block group passed doesn't match the block group | |
783 | * pointed to by the cluster, someone else raced in and freed the | |
784 | * cluster already. In that case, we just return without changing anything | |
785 | */ | |
786 | static int | |
787 | __btrfs_return_cluster_to_free_space( | |
788 | struct btrfs_block_group_cache *block_group, | |
789 | struct btrfs_free_cluster *cluster) | |
790 | { | |
791 | struct btrfs_free_space *entry; | |
792 | struct rb_node *node; | |
96303081 | 793 | bool bitmap; |
fa9c0d79 CM |
794 | |
795 | spin_lock(&cluster->lock); | |
796 | if (cluster->block_group != block_group) | |
797 | goto out; | |
798 | ||
96303081 JB |
799 | bitmap = cluster->points_to_bitmap; |
800 | cluster->block_group = NULL; | |
fa9c0d79 | 801 | cluster->window_start = 0; |
96303081 JB |
802 | list_del_init(&cluster->block_group_list); |
803 | cluster->points_to_bitmap = false; | |
804 | ||
805 | if (bitmap) | |
806 | goto out; | |
807 | ||
fa9c0d79 | 808 | node = rb_first(&cluster->root); |
96303081 | 809 | while (node) { |
fa9c0d79 CM |
810 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
811 | node = rb_next(&entry->offset_index); | |
812 | rb_erase(&entry->offset_index, &cluster->root); | |
96303081 JB |
813 | BUG_ON(entry->bitmap); |
814 | tree_insert_offset(&block_group->free_space_offset, | |
815 | entry->offset, &entry->offset_index, 0); | |
fa9c0d79 | 816 | } |
fa9c0d79 | 817 | cluster->root.rb_node = NULL; |
96303081 | 818 | |
fa9c0d79 CM |
819 | out: |
820 | spin_unlock(&cluster->lock); | |
96303081 | 821 | btrfs_put_block_group(block_group); |
fa9c0d79 CM |
822 | return 0; |
823 | } | |
824 | ||
0f9dd46c JB |
825 | void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) |
826 | { | |
827 | struct btrfs_free_space *info; | |
828 | struct rb_node *node; | |
fa9c0d79 | 829 | struct btrfs_free_cluster *cluster; |
96303081 | 830 | struct list_head *head; |
0f9dd46c | 831 | |
6226cb0a | 832 | spin_lock(&block_group->tree_lock); |
96303081 JB |
833 | while ((head = block_group->cluster_list.next) != |
834 | &block_group->cluster_list) { | |
835 | cluster = list_entry(head, struct btrfs_free_cluster, | |
836 | block_group_list); | |
fa9c0d79 CM |
837 | |
838 | WARN_ON(cluster->block_group != block_group); | |
839 | __btrfs_return_cluster_to_free_space(block_group, cluster); | |
96303081 JB |
840 | if (need_resched()) { |
841 | spin_unlock(&block_group->tree_lock); | |
842 | cond_resched(); | |
843 | spin_lock(&block_group->tree_lock); | |
844 | } | |
fa9c0d79 CM |
845 | } |
846 | ||
96303081 JB |
847 | while ((node = rb_last(&block_group->free_space_offset)) != NULL) { |
848 | info = rb_entry(node, struct btrfs_free_space, offset_index); | |
0f9dd46c | 849 | unlink_free_space(block_group, info); |
96303081 JB |
850 | if (info->bitmap) |
851 | kfree(info->bitmap); | |
0f9dd46c JB |
852 | kfree(info); |
853 | if (need_resched()) { | |
6226cb0a | 854 | spin_unlock(&block_group->tree_lock); |
0f9dd46c | 855 | cond_resched(); |
6226cb0a | 856 | spin_lock(&block_group->tree_lock); |
0f9dd46c JB |
857 | } |
858 | } | |
96303081 | 859 | |
6226cb0a | 860 | spin_unlock(&block_group->tree_lock); |
0f9dd46c JB |
861 | } |
862 | ||
6226cb0a JB |
863 | u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, |
864 | u64 offset, u64 bytes, u64 empty_size) | |
0f9dd46c | 865 | { |
6226cb0a | 866 | struct btrfs_free_space *entry = NULL; |
96303081 | 867 | u64 bytes_search = bytes + empty_size; |
6226cb0a | 868 | u64 ret = 0; |
0f9dd46c | 869 | |
6226cb0a | 870 | spin_lock(&block_group->tree_lock); |
96303081 | 871 | entry = find_free_space(block_group, &offset, &bytes_search, 0); |
6226cb0a | 872 | if (!entry) |
96303081 JB |
873 | goto out; |
874 | ||
875 | ret = offset; | |
876 | if (entry->bitmap) { | |
817d52f8 | 877 | bitmap_clear_bits(block_group, entry, offset, bytes); |
96303081 JB |
878 | if (!entry->bytes) { |
879 | unlink_free_space(block_group, entry); | |
880 | kfree(entry->bitmap); | |
881 | kfree(entry); | |
882 | block_group->total_bitmaps--; | |
883 | recalculate_thresholds(block_group); | |
884 | } | |
885 | } else { | |
6226cb0a | 886 | unlink_free_space(block_group, entry); |
6226cb0a JB |
887 | entry->offset += bytes; |
888 | entry->bytes -= bytes; | |
6226cb0a JB |
889 | if (!entry->bytes) |
890 | kfree(entry); | |
891 | else | |
892 | link_free_space(block_group, entry); | |
893 | } | |
0f9dd46c | 894 | |
96303081 JB |
895 | out: |
896 | spin_unlock(&block_group->tree_lock); | |
817d52f8 | 897 | |
0f9dd46c JB |
898 | return ret; |
899 | } | |
fa9c0d79 CM |
900 | |
901 | /* | |
902 | * given a cluster, put all of its extents back into the free space | |
903 | * cache. If a block group is passed, this function will only free | |
904 | * a cluster that belongs to the passed block group. | |
905 | * | |
906 | * Otherwise, it'll get a reference on the block group pointed to by the | |
907 | * cluster and remove the cluster from it. | |
908 | */ | |
909 | int btrfs_return_cluster_to_free_space( | |
910 | struct btrfs_block_group_cache *block_group, | |
911 | struct btrfs_free_cluster *cluster) | |
912 | { | |
913 | int ret; | |
914 | ||
915 | /* first, get a safe pointer to the block group */ | |
916 | spin_lock(&cluster->lock); | |
917 | if (!block_group) { | |
918 | block_group = cluster->block_group; | |
919 | if (!block_group) { | |
920 | spin_unlock(&cluster->lock); | |
921 | return 0; | |
922 | } | |
923 | } else if (cluster->block_group != block_group) { | |
924 | /* someone else has already freed it don't redo their work */ | |
925 | spin_unlock(&cluster->lock); | |
926 | return 0; | |
927 | } | |
928 | atomic_inc(&block_group->count); | |
929 | spin_unlock(&cluster->lock); | |
930 | ||
931 | /* now return any extents the cluster had on it */ | |
932 | spin_lock(&block_group->tree_lock); | |
933 | ret = __btrfs_return_cluster_to_free_space(block_group, cluster); | |
934 | spin_unlock(&block_group->tree_lock); | |
935 | ||
936 | /* finally drop our ref */ | |
937 | btrfs_put_block_group(block_group); | |
938 | return ret; | |
939 | } | |
940 | ||
96303081 JB |
941 | static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, |
942 | struct btrfs_free_cluster *cluster, | |
943 | u64 bytes, u64 min_start) | |
944 | { | |
945 | struct btrfs_free_space *entry; | |
946 | int err; | |
947 | u64 search_start = cluster->window_start; | |
948 | u64 search_bytes = bytes; | |
949 | u64 ret = 0; | |
950 | ||
951 | spin_lock(&block_group->tree_lock); | |
952 | spin_lock(&cluster->lock); | |
953 | ||
954 | if (!cluster->points_to_bitmap) | |
955 | goto out; | |
956 | ||
957 | if (cluster->block_group != block_group) | |
958 | goto out; | |
959 | ||
960 | entry = tree_search_offset(block_group, search_start, 0, 0); | |
961 | ||
962 | if (!entry || !entry->bitmap) | |
963 | goto out; | |
964 | ||
965 | search_start = min_start; | |
966 | search_bytes = bytes; | |
967 | ||
968 | err = search_bitmap(block_group, entry, &search_start, | |
969 | &search_bytes); | |
970 | if (err) | |
971 | goto out; | |
972 | ||
973 | ret = search_start; | |
817d52f8 | 974 | bitmap_clear_bits(block_group, entry, ret, bytes); |
96303081 JB |
975 | out: |
976 | spin_unlock(&cluster->lock); | |
977 | spin_unlock(&block_group->tree_lock); | |
978 | ||
979 | return ret; | |
980 | } | |
981 | ||
fa9c0d79 CM |
982 | /* |
983 | * given a cluster, try to allocate 'bytes' from it, returns 0 | |
984 | * if it couldn't find anything suitably large, or a logical disk offset | |
985 | * if things worked out | |
986 | */ | |
987 | u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, | |
988 | struct btrfs_free_cluster *cluster, u64 bytes, | |
989 | u64 min_start) | |
990 | { | |
991 | struct btrfs_free_space *entry = NULL; | |
992 | struct rb_node *node; | |
993 | u64 ret = 0; | |
994 | ||
96303081 JB |
995 | if (cluster->points_to_bitmap) |
996 | return btrfs_alloc_from_bitmap(block_group, cluster, bytes, | |
997 | min_start); | |
998 | ||
fa9c0d79 CM |
999 | spin_lock(&cluster->lock); |
1000 | if (bytes > cluster->max_size) | |
1001 | goto out; | |
1002 | ||
1003 | if (cluster->block_group != block_group) | |
1004 | goto out; | |
1005 | ||
1006 | node = rb_first(&cluster->root); | |
1007 | if (!node) | |
1008 | goto out; | |
1009 | ||
1010 | entry = rb_entry(node, struct btrfs_free_space, offset_index); | |
1011 | ||
1012 | while(1) { | |
1013 | if (entry->bytes < bytes || entry->offset < min_start) { | |
1014 | struct rb_node *node; | |
1015 | ||
1016 | node = rb_next(&entry->offset_index); | |
1017 | if (!node) | |
1018 | break; | |
1019 | entry = rb_entry(node, struct btrfs_free_space, | |
1020 | offset_index); | |
1021 | continue; | |
1022 | } | |
1023 | ret = entry->offset; | |
1024 | ||
1025 | entry->offset += bytes; | |
1026 | entry->bytes -= bytes; | |
1027 | ||
1028 | if (entry->bytes == 0) { | |
1029 | rb_erase(&entry->offset_index, &cluster->root); | |
1030 | kfree(entry); | |
1031 | } | |
1032 | break; | |
1033 | } | |
1034 | out: | |
1035 | spin_unlock(&cluster->lock); | |
96303081 | 1036 | |
fa9c0d79 CM |
1037 | return ret; |
1038 | } | |
1039 | ||
96303081 JB |
1040 | static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, |
1041 | struct btrfs_free_space *entry, | |
1042 | struct btrfs_free_cluster *cluster, | |
1043 | u64 offset, u64 bytes, u64 min_bytes) | |
1044 | { | |
1045 | unsigned long next_zero; | |
1046 | unsigned long i; | |
1047 | unsigned long search_bits; | |
1048 | unsigned long total_bits; | |
1049 | unsigned long found_bits; | |
1050 | unsigned long start = 0; | |
1051 | unsigned long total_found = 0; | |
1052 | bool found = false; | |
1053 | ||
1054 | i = offset_to_bit(entry->offset, block_group->sectorsize, | |
1055 | max_t(u64, offset, entry->offset)); | |
1056 | search_bits = bytes_to_bits(min_bytes, block_group->sectorsize); | |
1057 | total_bits = bytes_to_bits(bytes, block_group->sectorsize); | |
1058 | ||
1059 | again: | |
1060 | found_bits = 0; | |
1061 | for (i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i); | |
1062 | i < BITS_PER_BITMAP; | |
1063 | i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i + 1)) { | |
1064 | next_zero = find_next_zero_bit(entry->bitmap, | |
1065 | BITS_PER_BITMAP, i); | |
1066 | if (next_zero - i >= search_bits) { | |
1067 | found_bits = next_zero - i; | |
1068 | break; | |
1069 | } | |
1070 | i = next_zero; | |
1071 | } | |
1072 | ||
1073 | if (!found_bits) | |
1074 | return -1; | |
1075 | ||
1076 | if (!found) { | |
1077 | start = i; | |
1078 | found = true; | |
1079 | } | |
1080 | ||
1081 | total_found += found_bits; | |
1082 | ||
1083 | if (cluster->max_size < found_bits * block_group->sectorsize) | |
1084 | cluster->max_size = found_bits * block_group->sectorsize; | |
1085 | ||
1086 | if (total_found < total_bits) { | |
1087 | i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, next_zero); | |
1088 | if (i - start > total_bits * 2) { | |
1089 | total_found = 0; | |
1090 | cluster->max_size = 0; | |
1091 | found = false; | |
1092 | } | |
1093 | goto again; | |
1094 | } | |
1095 | ||
1096 | cluster->window_start = start * block_group->sectorsize + | |
1097 | entry->offset; | |
1098 | cluster->points_to_bitmap = true; | |
1099 | ||
1100 | return 0; | |
1101 | } | |
1102 | ||
fa9c0d79 CM |
1103 | /* |
1104 | * here we try to find a cluster of blocks in a block group. The goal | |
1105 | * is to find at least bytes free and up to empty_size + bytes free. | |
1106 | * We might not find them all in one contiguous area. | |
1107 | * | |
1108 | * returns zero and sets up cluster if things worked out, otherwise | |
1109 | * it returns -enospc | |
1110 | */ | |
1111 | int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, | |
451d7585 | 1112 | struct btrfs_root *root, |
fa9c0d79 CM |
1113 | struct btrfs_block_group_cache *block_group, |
1114 | struct btrfs_free_cluster *cluster, | |
1115 | u64 offset, u64 bytes, u64 empty_size) | |
1116 | { | |
1117 | struct btrfs_free_space *entry = NULL; | |
1118 | struct rb_node *node; | |
1119 | struct btrfs_free_space *next; | |
96303081 | 1120 | struct btrfs_free_space *last = NULL; |
fa9c0d79 CM |
1121 | u64 min_bytes; |
1122 | u64 window_start; | |
1123 | u64 window_free; | |
1124 | u64 max_extent = 0; | |
96303081 | 1125 | bool found_bitmap = false; |
fa9c0d79 CM |
1126 | int ret; |
1127 | ||
1128 | /* for metadata, allow allocates with more holes */ | |
451d7585 CM |
1129 | if (btrfs_test_opt(root, SSD_SPREAD)) { |
1130 | min_bytes = bytes + empty_size; | |
1131 | } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { | |
fa9c0d79 CM |
1132 | /* |
1133 | * we want to do larger allocations when we are | |
1134 | * flushing out the delayed refs, it helps prevent | |
1135 | * making more work as we go along. | |
1136 | */ | |
1137 | if (trans->transaction->delayed_refs.flushing) | |
1138 | min_bytes = max(bytes, (bytes + empty_size) >> 1); | |
1139 | else | |
1140 | min_bytes = max(bytes, (bytes + empty_size) >> 4); | |
1141 | } else | |
1142 | min_bytes = max(bytes, (bytes + empty_size) >> 2); | |
1143 | ||
1144 | spin_lock(&block_group->tree_lock); | |
1145 | spin_lock(&cluster->lock); | |
1146 | ||
1147 | /* someone already found a cluster, hooray */ | |
1148 | if (cluster->block_group) { | |
1149 | ret = 0; | |
1150 | goto out; | |
1151 | } | |
1152 | again: | |
96303081 | 1153 | entry = tree_search_offset(block_group, offset, found_bitmap, 1); |
fa9c0d79 CM |
1154 | if (!entry) { |
1155 | ret = -ENOSPC; | |
1156 | goto out; | |
1157 | } | |
96303081 JB |
1158 | |
1159 | /* | |
1160 | * If found_bitmap is true, we exhausted our search for extent entries, | |
1161 | * and we just want to search all of the bitmaps that we can find, and | |
1162 | * ignore any extent entries we find. | |
1163 | */ | |
1164 | while (entry->bitmap || found_bitmap || | |
1165 | (!entry->bitmap && entry->bytes < min_bytes)) { | |
1166 | struct rb_node *node = rb_next(&entry->offset_index); | |
1167 | ||
1168 | if (entry->bitmap && entry->bytes > bytes + empty_size) { | |
1169 | ret = btrfs_bitmap_cluster(block_group, entry, cluster, | |
1170 | offset, bytes + empty_size, | |
1171 | min_bytes); | |
1172 | if (!ret) | |
1173 | goto got_it; | |
1174 | } | |
1175 | ||
1176 | if (!node) { | |
1177 | ret = -ENOSPC; | |
1178 | goto out; | |
1179 | } | |
1180 | entry = rb_entry(node, struct btrfs_free_space, offset_index); | |
1181 | } | |
1182 | ||
1183 | /* | |
1184 | * We already searched all the extent entries from the passed in offset | |
1185 | * to the end and didn't find enough space for the cluster, and we also | |
1186 | * didn't find any bitmaps that met our criteria, just go ahead and exit | |
1187 | */ | |
1188 | if (found_bitmap) { | |
1189 | ret = -ENOSPC; | |
1190 | goto out; | |
1191 | } | |
1192 | ||
1193 | cluster->points_to_bitmap = false; | |
fa9c0d79 CM |
1194 | window_start = entry->offset; |
1195 | window_free = entry->bytes; | |
1196 | last = entry; | |
1197 | max_extent = entry->bytes; | |
1198 | ||
96303081 | 1199 | while (1) { |
fa9c0d79 CM |
1200 | /* out window is just right, lets fill it */ |
1201 | if (window_free >= bytes + empty_size) | |
1202 | break; | |
1203 | ||
1204 | node = rb_next(&last->offset_index); | |
1205 | if (!node) { | |
96303081 JB |
1206 | if (found_bitmap) |
1207 | goto again; | |
fa9c0d79 CM |
1208 | ret = -ENOSPC; |
1209 | goto out; | |
1210 | } | |
1211 | next = rb_entry(node, struct btrfs_free_space, offset_index); | |
1212 | ||
96303081 JB |
1213 | /* |
1214 | * we found a bitmap, so if this search doesn't result in a | |
1215 | * cluster, we know to go and search again for the bitmaps and | |
1216 | * start looking for space there | |
1217 | */ | |
1218 | if (next->bitmap) { | |
1219 | if (!found_bitmap) | |
1220 | offset = next->offset; | |
1221 | found_bitmap = true; | |
1222 | last = next; | |
1223 | continue; | |
1224 | } | |
1225 | ||
fa9c0d79 CM |
1226 | /* |
1227 | * we haven't filled the empty size and the window is | |
1228 | * very large. reset and try again | |
1229 | */ | |
c6044801 CM |
1230 | if (next->offset - (last->offset + last->bytes) > 128 * 1024 || |
1231 | next->offset - window_start > (bytes + empty_size) * 2) { | |
fa9c0d79 CM |
1232 | entry = next; |
1233 | window_start = entry->offset; | |
1234 | window_free = entry->bytes; | |
1235 | last = entry; | |
1236 | max_extent = 0; | |
fa9c0d79 CM |
1237 | } else { |
1238 | last = next; | |
1239 | window_free += next->bytes; | |
1240 | if (entry->bytes > max_extent) | |
1241 | max_extent = entry->bytes; | |
1242 | } | |
1243 | } | |
1244 | ||
1245 | cluster->window_start = entry->offset; | |
1246 | ||
1247 | /* | |
1248 | * now we've found our entries, pull them out of the free space | |
1249 | * cache and put them into the cluster rbtree | |
1250 | * | |
1251 | * The cluster includes an rbtree, but only uses the offset index | |
1252 | * of each free space cache entry. | |
1253 | */ | |
96303081 | 1254 | while (1) { |
fa9c0d79 | 1255 | node = rb_next(&entry->offset_index); |
96303081 JB |
1256 | if (entry->bitmap && node) { |
1257 | entry = rb_entry(node, struct btrfs_free_space, | |
1258 | offset_index); | |
1259 | continue; | |
1260 | } else if (entry->bitmap && !node) { | |
1261 | break; | |
1262 | } | |
1263 | ||
1264 | rb_erase(&entry->offset_index, &block_group->free_space_offset); | |
fa9c0d79 | 1265 | ret = tree_insert_offset(&cluster->root, entry->offset, |
96303081 | 1266 | &entry->offset_index, 0); |
fa9c0d79 CM |
1267 | BUG_ON(ret); |
1268 | ||
1269 | if (!node || entry == last) | |
1270 | break; | |
1271 | ||
1272 | entry = rb_entry(node, struct btrfs_free_space, offset_index); | |
1273 | } | |
96303081 | 1274 | |
fa9c0d79 | 1275 | cluster->max_size = max_extent; |
96303081 JB |
1276 | got_it: |
1277 | ret = 0; | |
fa9c0d79 CM |
1278 | atomic_inc(&block_group->count); |
1279 | list_add_tail(&cluster->block_group_list, &block_group->cluster_list); | |
1280 | cluster->block_group = block_group; | |
1281 | out: | |
1282 | spin_unlock(&cluster->lock); | |
1283 | spin_unlock(&block_group->tree_lock); | |
1284 | ||
1285 | return ret; | |
1286 | } | |
1287 | ||
1288 | /* | |
1289 | * simple code to zero out a cluster | |
1290 | */ | |
1291 | void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) | |
1292 | { | |
1293 | spin_lock_init(&cluster->lock); | |
1294 | spin_lock_init(&cluster->refill_lock); | |
1295 | cluster->root.rb_node = NULL; | |
1296 | cluster->max_size = 0; | |
96303081 | 1297 | cluster->points_to_bitmap = false; |
fa9c0d79 CM |
1298 | INIT_LIST_HEAD(&cluster->block_group_list); |
1299 | cluster->block_group = NULL; | |
1300 | } | |
1301 |