Merge branch 'vendor/BINUTILS224'
[dragonfly.git] / sys / vfs / hammer / hammer_dedup.c
1 /*
2  * Copyright (c) 2010 The DragonFly Project.  All rights reserved.
3  *
4  * This code is derived from software contributed to The DragonFly Project
5  * by Ilya Dryomov <idryomov@gmail.com>
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  *
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in
15  *    the documentation and/or other materials provided with the
16  *    distribution.
17  * 3. Neither the name of The DragonFly Project nor the names of its
18  *    contributors may be used to endorse or promote products derived
19  *    from this software without specific, prior written permission.
20  *
21  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
25  * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26  * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
27  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
29  * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
30  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
31  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34
35 #include "hammer.h"
36
37 static __inline int validate_zone(hammer_off_t data_offset);
38
39 int
40 hammer_ioc_dedup(hammer_transaction_t trans, hammer_inode_t ip,
41                  struct hammer_ioc_dedup *dedup)
42 {
43         struct hammer_cursor cursor1, cursor2;
44         int error;
45         int seq;
46
47         /*
48          * Enforce hammer filesystem version requirements
49          */
50         if (trans->hmp->version < HAMMER_VOL_VERSION_FIVE) {
51                 kprintf("hammer: Filesystem must be upgraded to v5 "
52                         "before you can run dedup\n");
53                 return (EOPNOTSUPP);
54         }
55
56         /*
57          * Cursor1, return an error -> candidate goes to pass2 list
58          */
59         error = hammer_init_cursor(trans, &cursor1, NULL, NULL);
60         if (error)
61                 goto done_cursor;
62         cursor1.key_beg = dedup->elm1;
63         cursor1.flags |= HAMMER_CURSOR_BACKEND;
64
65         error = hammer_btree_lookup(&cursor1);
66         if (error)
67                 goto done_cursor;
68         error = hammer_btree_extract(&cursor1, HAMMER_CURSOR_GET_LEAF |
69                                                 HAMMER_CURSOR_GET_DATA);
70         if (error)
71                 goto done_cursor;
72
73         /*
74          * Cursor2, return an error -> candidate goes to pass2 list
75          */
76         error = hammer_init_cursor(trans, &cursor2, NULL, NULL);
77         if (error)
78                 goto done_cursors;
79         cursor2.key_beg = dedup->elm2;
80         cursor2.flags |= HAMMER_CURSOR_BACKEND;
81
82         error = hammer_btree_lookup(&cursor2);
83         if (error)
84                 goto done_cursors;
85         error = hammer_btree_extract(&cursor2, HAMMER_CURSOR_GET_LEAF |
86                                                 HAMMER_CURSOR_GET_DATA);
87         if (error)
88                 goto done_cursors;
89
90         /*
91          * Zone validation. We can't de-dup any of the other zones
92          * (BTREE or META) or bad things will happen.
93          *
94          * Return with error = 0, but set an INVALID_ZONE flag.
95          */
96         error = validate_zone(cursor1.leaf->data_offset) +
97                             validate_zone(cursor2.leaf->data_offset);
98         if (error) {
99                 dedup->head.flags |= HAMMER_IOC_DEDUP_INVALID_ZONE;
100                 error = 0;
101                 goto done_cursors;
102         }
103
104         /*
105          * Comparison checks
106          *
107          * If zones don't match or data_len fields aren't the same
108          * we consider it to be a comparison failure.
109          *
110          * Return with error = 0, but set a CMP_FAILURE flag.
111          */
112         if ((cursor1.leaf->data_offset & HAMMER_OFF_ZONE_MASK) !=
113             (cursor2.leaf->data_offset & HAMMER_OFF_ZONE_MASK)) {
114                 dedup->head.flags |= HAMMER_IOC_DEDUP_CMP_FAILURE;
115                 goto done_cursors;
116         }
117         if (cursor1.leaf->data_len != cursor2.leaf->data_len) {
118                 dedup->head.flags |= HAMMER_IOC_DEDUP_CMP_FAILURE;
119                 goto done_cursors;
120         }
121
122         /* byte-by-byte comparison to be sure */
123         if (bcmp(cursor1.data, cursor2.data, cursor1.leaf->data_len)) {
124                 dedup->head.flags |= HAMMER_IOC_DEDUP_CMP_FAILURE;
125                 goto done_cursors;
126         }
127
128         /*
129          * Upgrade both cursors together to an exclusive lock
130          *
131          * Return an error -> candidate goes to pass2 list
132          */
133         hammer_sync_lock_sh(trans);
134         error = hammer_cursor_upgrade2(&cursor1, &cursor2);
135         if (error) {
136                 hammer_sync_unlock(trans);
137                 goto done_cursors;
138         }
139
140         error = hammer_blockmap_dedup(cursor1.trans,
141                         cursor1.leaf->data_offset, cursor1.leaf->data_len);
142         if (error) {
143                 if (error == ERANGE) {
144                         /*
145                          * Return with error = 0, but set an UNDERFLOW flag
146                          */
147                         dedup->head.flags |= HAMMER_IOC_DEDUP_UNDERFLOW;
148                         error = 0;
149                         goto downgrade_cursors;
150                 } else {
151                         /*
152                          * Return an error -> block goes to pass2 list
153                          */
154                         goto downgrade_cursors;
155                 }
156         }
157
158         /*
159          * The cursor2's cache must be invalidated before calling
160          * hammer_blockmap_free(), otherwise it will not be able to
161          * invalidate the underlying data buffer.
162          */
163         hammer_cursor_invalidate_cache(&cursor2);
164         hammer_blockmap_free(cursor2.trans,
165                         cursor2.leaf->data_offset, cursor2.leaf->data_len);
166
167         hammer_modify_node(cursor2.trans, cursor2.node,
168                         &cursor2.leaf->data_offset, sizeof(hammer_off_t));
169         cursor2.leaf->data_offset = cursor1.leaf->data_offset;
170         hammer_modify_node_done(cursor2.node);
171
172 downgrade_cursors:
173         hammer_cursor_downgrade2(&cursor1, &cursor2);
174         hammer_sync_unlock(trans);
175 done_cursors:
176         hammer_done_cursor(&cursor2);
177 done_cursor:
178         hammer_done_cursor(&cursor1);
179
180         /*
181          * Avoid deadlocking the buffer cache
182          */
183         seq = trans->hmp->flusher.done;
184         while (hammer_flusher_meta_halflimit(trans->hmp) ||
185                hammer_flusher_undo_exhausted(trans, 2)) {
186                 hammer_flusher_wait(trans->hmp, seq);
187                 seq = hammer_flusher_async_one(trans->hmp);
188         }
189         return (error);
190 }
191
192 static __inline int
193 validate_zone(hammer_off_t data_offset)
194 {
195         switch(data_offset & HAMMER_OFF_ZONE_MASK) {
196         case HAMMER_ZONE_LARGE_DATA:
197         case HAMMER_ZONE_SMALL_DATA:
198                 return (0);
199         default:
200                 return (1);
201         }
202 }
203
204 /************************************************************************
205  *                              LIVE DEDUP                              *
206  ************************************************************************
207  *
208  * HAMMER Live Dedup (aka as efficient cp(1) implementation)
209  *
210  * The dedup cache is operated in a LRU fashion and destroyed on
211  * unmount, so essentially this is a live dedup on a cached dataset and
212  * not a full-fledged fs-wide one - we have a batched dedup for that.
213  * We allow duplicate entries in the buffer cache, data blocks are
214  * deduplicated only on their way to media.  By default the cache is
215  * populated on reads only, but it can be populated on writes too.
216  *
217  * The main implementation gotcha is on-media requirement - in order for
218  * a data block to be added to a dedup cache it has to be present on
219  * disk.  This simplifies cache logic a lot - once data is laid out on
220  * media it remains valid on media all the way up to the point where the
221  * related big block the data was stored in is freed - so there is only
222  * one place we need to bother with invalidation code.
223  */
224
225 /*
226  * RB-Tree support for dedup cache structures
227  */
228 RB_GENERATE2(hammer_dedup_crc_rb_tree, hammer_dedup_cache, crc_entry,
229                 hammer_dedup_crc_rb_compare, hammer_crc_t, crc);
230 RB_GENERATE2(hammer_dedup_off_rb_tree, hammer_dedup_cache, off_entry,
231                 hammer_dedup_off_rb_compare, hammer_off_t, data_offset);
232
233 struct hammer_dedup_inval {
234         struct hammer_mount *hmp;
235         hammer_off_t base_offset;
236 };
237
238 static int hammer_dedup_scancmp(hammer_dedup_cache_t dc, void *data);
239 static int hammer_dedup_cache_inval_callback(hammer_dedup_cache_t dc,
240                 void *data);
241 static __inline int _vnode_validate(hammer_dedup_cache_t dcp, void *data,
242                 int *errorp);
243 static __inline int _dev_validate(hammer_dedup_cache_t dcp, void *data,
244                 int *errorp);
245
246 int
247 hammer_dedup_crc_rb_compare(hammer_dedup_cache_t dc1,
248                                 hammer_dedup_cache_t dc2)
249 {
250         if (dc1->crc < dc2->crc)
251                 return (-1);
252         if (dc1->crc > dc2->crc)
253                 return (1);
254
255         return (0);
256 }
257
258 int
259 hammer_dedup_off_rb_compare(hammer_dedup_cache_t dc1,
260                                 hammer_dedup_cache_t dc2)
261 {
262         if (dc1->data_offset < dc2->data_offset)
263                 return (-1);
264         if (dc1->data_offset > dc2->data_offset)
265                 return (1);
266
267         return (0);
268 }
269
270 static int
271 hammer_dedup_scancmp(hammer_dedup_cache_t dc, void *data)
272 {
273         hammer_off_t off = ((struct hammer_dedup_inval *)data)->base_offset;
274
275         if (dc->data_offset < off)
276                 return (-1);
277         if (dc->data_offset >= off + HAMMER_LARGEBLOCK_SIZE)
278                 return (1);
279
280         return (0);
281 }
282
283 static int
284 hammer_dedup_cache_inval_callback(hammer_dedup_cache_t dc, void *data)
285 {
286         hammer_mount_t hmp = ((struct hammer_dedup_inval *)data)->hmp;
287
288         RB_REMOVE(hammer_dedup_crc_rb_tree, &hmp->rb_dedup_crc_root, dc);
289         RB_REMOVE(hammer_dedup_off_rb_tree, &hmp->rb_dedup_off_root, dc);
290         TAILQ_REMOVE(&hmp->dedup_lru_list, dc, lru_entry);
291
292         --hmp->dedup_cache_count;
293         kfree(dc, hmp->m_misc);
294
295         return (0);
296 }
297
298 hammer_dedup_cache_t
299 hammer_dedup_cache_add(hammer_inode_t ip, hammer_btree_leaf_elm_t leaf)
300 {
301         hammer_dedup_cache_t dcp, tmp;
302         hammer_mount_t hmp = ip->hmp;
303
304         if (hmp->dedup_free_cache == NULL) {
305                 tmp = kmalloc(sizeof(*tmp), hmp->m_misc, M_WAITOK | M_ZERO);
306                 if (hmp->dedup_free_cache == NULL)
307                         hmp->dedup_free_cache = tmp;
308                 else
309                         kfree(tmp, hmp->m_misc);
310         }
311
312         KKASSERT(leaf != NULL);
313
314         dcp = RB_LOOKUP(hammer_dedup_crc_rb_tree,
315                                 &hmp->rb_dedup_crc_root, leaf->data_crc);
316         if (dcp != NULL) {
317                 RB_REMOVE(hammer_dedup_off_rb_tree,
318                                 &hmp->rb_dedup_off_root, dcp);
319                 TAILQ_REMOVE(&hmp->dedup_lru_list, dcp, lru_entry);
320                 goto populate;
321         }
322
323         if (hmp->dedup_cache_count < hammer_live_dedup_cache_size) {
324                 dcp = hmp->dedup_free_cache;
325                 hmp->dedup_free_cache = NULL;
326                 ++hmp->dedup_cache_count;
327         } else {
328                 dcp = TAILQ_FIRST(&hmp->dedup_lru_list);
329                 RB_REMOVE(hammer_dedup_crc_rb_tree,
330                                 &hmp->rb_dedup_crc_root, dcp);
331                 RB_REMOVE(hammer_dedup_off_rb_tree,
332                                 &hmp->rb_dedup_off_root, dcp);
333                 TAILQ_REMOVE(&hmp->dedup_lru_list, dcp, lru_entry);
334         }
335
336         dcp->crc = leaf->data_crc;
337         tmp = RB_INSERT(hammer_dedup_crc_rb_tree, &hmp->rb_dedup_crc_root, dcp);
338         KKASSERT(tmp == NULL);
339
340 populate:
341         dcp->hmp = ip->hmp;
342         dcp->obj_id = ip->obj_id;
343         dcp->localization = ip->obj_localization;
344         dcp->file_offset = leaf->base.key - leaf->data_len;
345         dcp->bytes = leaf->data_len;
346         dcp->data_offset = leaf->data_offset;
347
348         tmp = RB_INSERT(hammer_dedup_off_rb_tree, &hmp->rb_dedup_off_root, dcp);
349         KKASSERT(tmp == NULL);
350         TAILQ_INSERT_TAIL(&hmp->dedup_lru_list, dcp, lru_entry);
351
352         return (dcp);
353 }
354
355 __inline hammer_dedup_cache_t
356 hammer_dedup_cache_lookup(hammer_mount_t hmp, hammer_crc_t crc)
357 {
358         hammer_dedup_cache_t dcp;
359
360         dcp = RB_LOOKUP(hammer_dedup_crc_rb_tree,
361                                 &hmp->rb_dedup_crc_root, crc);
362         return dcp;
363 }
364
365 void hammer_dedup_cache_inval(hammer_mount_t hmp, hammer_off_t base_offset)
366 {
367         struct hammer_dedup_inval di;
368
369         di.hmp = hmp;
370         di.base_offset = base_offset;
371
372         RB_SCAN(hammer_dedup_off_rb_tree, &hmp->rb_dedup_off_root,
373                 hammer_dedup_scancmp, hammer_dedup_cache_inval_callback, &di);
374 }
375
376 void
377 hammer_destroy_dedup_cache(hammer_mount_t hmp)
378 {
379         hammer_dedup_cache_t dcp;
380
381         while ((dcp = TAILQ_FIRST(&hmp->dedup_lru_list)) != NULL) {
382                 RB_REMOVE(hammer_dedup_crc_rb_tree, &hmp->rb_dedup_crc_root, dcp);
383                 RB_REMOVE(hammer_dedup_off_rb_tree, &hmp->rb_dedup_off_root, dcp);
384                 TAILQ_REMOVE(&hmp->dedup_lru_list, dcp, lru_entry);
385                 --hmp->dedup_cache_count;
386                 kfree(dcp, hmp->m_misc);
387         }
388
389         KKASSERT(RB_EMPTY(&hmp->rb_dedup_crc_root));
390         KKASSERT(RB_EMPTY(&hmp->rb_dedup_off_root));
391         KKASSERT(TAILQ_EMPTY(&hmp->dedup_lru_list));
392
393         KKASSERT(hmp->dedup_cache_count == 0);
394 }
395
396 int
397 hammer_dedup_validate(hammer_dedup_cache_t dcp, int zone, int bytes,
398                       void *data)
399 {
400         int error;
401
402         /*
403          * Zone validation
404          */
405         if (HAMMER_ZONE_DECODE(dcp->data_offset) != zone)
406                 return (0);
407
408         /*
409          * Block length validation
410          */
411         if (dcp->bytes != bytes)
412                 return (0);
413
414         /*
415          * Byte-by-byte data comparison
416          *
417          * The data we need for validation may already be present in the
418          * buffer cache in two flavours: vnode-based buffer or
419          * block-device-based buffer.  In case vnode-based buffer wasn't
420          * there or if a non-blocking attempt to acquire it failed use
421          * device-based buffer (for large-zone data blocks it will
422          * generate a separate read).
423          *
424          * XXX vnode-based checking is not MP safe so when live-dedup
425          *     is turned on we must always use the device buffer.
426          */
427 #if 0
428         if (hammer_double_buffer) {
429                 error = 1;
430         } else if (_vnode_validate(dcp, data, &error)) {
431                 hammer_live_dedup_vnode_bcmps++;
432                 return (1);
433         } else {
434                 if (error == 3)
435                         hammer_live_dedup_findblk_failures++;
436         }
437
438         /*
439          * If there was an error previously or if double buffering is
440          * enabled.
441          */
442         if (error) {
443                 if (_dev_validate(dcp, data, &error)) {
444                         hammer_live_dedup_device_bcmps++;
445                         return (1);
446                 }
447         }
448 #endif
449         if (_dev_validate(dcp, data, &error)) {
450                 hammer_live_dedup_device_bcmps++;
451                 return (1);
452         }
453
454         return (0);
455 }
456
457 static __inline int
458 _vnode_validate(hammer_dedup_cache_t dcp, void *data, int *errorp)
459 {
460         struct hammer_transaction trans;
461         hammer_inode_t ip;
462         struct vnode *vp;
463         struct buf *bp;
464         off_t dooffset;
465         int result, error;
466
467         result = error = 0;
468         *errorp = 0;
469
470         hammer_simple_transaction(&trans, dcp->hmp);
471
472         ip = hammer_get_inode(&trans, NULL, dcp->obj_id, HAMMER_MAX_TID,
473             dcp->localization, 0, &error);
474         if (ip == NULL) {
475                 kprintf("dedup: unable to find objid %016jx:%08x\n",
476                     (intmax_t)dcp->obj_id, dcp->localization);
477                 *errorp = 1;
478                 goto failed2;
479         }
480
481         error = hammer_get_vnode(ip, &vp);
482         if (error) {
483                 kprintf("dedup: unable to acquire vnode for %016jx:%08x\n",
484                     (intmax_t)dcp->obj_id, dcp->localization);
485                 *errorp = 2;
486                 goto failed;
487         }
488
489         if ((bp = findblk(ip->vp, dcp->file_offset, FINDBLK_NBLOCK)) != NULL) {
490                 bremfree(bp);
491
492                 /* XXX if (mapped to userspace) goto done, *errorp = 4 */
493
494                 if ((bp->b_flags & B_CACHE) == 0 || bp->b_flags & B_DIRTY) {
495                         *errorp = 5;
496                         goto done;
497                 }
498
499                 if (bp->b_bio2.bio_offset != dcp->data_offset) {
500                         error = VOP_BMAP(ip->vp, dcp->file_offset, &dooffset,
501                             NULL, NULL, BUF_CMD_READ);
502                         if (error) {
503                                 *errorp = 6;
504                                 goto done;
505                         }
506
507                         if (dooffset != dcp->data_offset) {
508                                 *errorp = 7;
509                                 goto done;
510                         }
511                         hammer_live_dedup_bmap_saves++;
512                 }
513
514                 if (bcmp(data, bp->b_data, dcp->bytes) == 0)
515                         result = 1;
516
517 done:
518                 bqrelse(bp);
519         } else {
520                 *errorp = 3;
521         }
522         vput(vp);
523
524 failed:
525         hammer_rel_inode(ip, 0);
526 failed2:
527         hammer_done_transaction(&trans);
528         return (result);
529 }
530
531 static __inline int
532 _dev_validate(hammer_dedup_cache_t dcp, void *data, int *errorp)
533 {
534         hammer_buffer_t buffer = NULL;
535         void *ondisk_data;
536         int result, error;
537
538         result = error = 0;
539         *errorp = 0;
540
541         ondisk_data = hammer_bread_ext(dcp->hmp, dcp->data_offset,
542             dcp->bytes, &error, &buffer);
543         if (error) {
544                 *errorp = 1;
545                 goto failed;
546         }
547
548         if (bcmp(data, ondisk_data, dcp->bytes) == 0)
549                 result = 1;
550
551 failed:
552         if (buffer)
553                 hammer_rel_buffer(buffer, 0);
554
555         return (result);
556 }