xref: /linux/fs/bcachefs/btree_node_scan.c (revision ba3193fa8fc8910f724b67a523ec67ee24997d3e)
1 // SPDX-License-Identifier: GPL-2.0
2 
3 #include "bcachefs.h"
4 #include "btree_cache.h"
5 #include "btree_io.h"
6 #include "btree_journal_iter.h"
7 #include "btree_node_scan.h"
8 #include "btree_update_interior.h"
9 #include "buckets.h"
10 #include "error.h"
11 #include "journal_io.h"
12 #include "recovery_passes.h"
13 
14 #include <linux/kthread.h>
15 #include <linux/sort.h>
16 
17 struct find_btree_nodes_worker {
18 	struct closure		*cl;
19 	struct find_btree_nodes	*f;
20 	struct bch_dev		*ca;
21 };
22 
23 static void found_btree_node_to_text(struct printbuf *out, struct bch_fs *c, const struct found_btree_node *n)
24 {
25 	prt_printf(out, "%s l=%u seq=%u cookie=%llx ", bch2_btree_id_str(n->btree_id), n->level, n->seq, n->cookie);
26 	bch2_bpos_to_text(out, n->min_key);
27 	prt_str(out, "-");
28 	bch2_bpos_to_text(out, n->max_key);
29 
30 	if (n->range_updated)
31 		prt_str(out, " range updated");
32 	if (n->overwritten)
33 		prt_str(out, " overwritten");
34 
35 	for (unsigned i = 0; i < n->nr_ptrs; i++) {
36 		prt_char(out, ' ');
37 		bch2_extent_ptr_to_text(out, c, n->ptrs + i);
38 	}
39 }
40 
41 static void found_btree_nodes_to_text(struct printbuf *out, struct bch_fs *c, found_btree_nodes nodes)
42 {
43 	printbuf_indent_add(out, 2);
44 	darray_for_each(nodes, i) {
45 		found_btree_node_to_text(out, c, i);
46 		prt_newline(out);
47 	}
48 	printbuf_indent_sub(out, 2);
49 }
50 
51 static void found_btree_node_to_key(struct bkey_i *k, const struct found_btree_node *f)
52 {
53 	struct bkey_i_btree_ptr_v2 *bp = bkey_btree_ptr_v2_init(k);
54 
55 	set_bkey_val_u64s(&bp->k, sizeof(struct bch_btree_ptr_v2) / sizeof(u64) + f->nr_ptrs);
56 	bp->k.p			= f->max_key;
57 	bp->v.seq		= cpu_to_le64(f->cookie);
58 	bp->v.sectors_written	= 0;
59 	bp->v.flags		= 0;
60 	bp->v.min_key		= f->min_key;
61 	SET_BTREE_PTR_RANGE_UPDATED(&bp->v, f->range_updated);
62 	memcpy(bp->v.start, f->ptrs, sizeof(struct bch_extent_ptr) * f->nr_ptrs);
63 }
64 
65 static bool found_btree_node_is_readable(struct btree_trans *trans,
66 					 const struct found_btree_node *f)
67 {
68 	struct { __BKEY_PADDED(k, BKEY_BTREE_PTR_VAL_U64s_MAX); } k;
69 
70 	found_btree_node_to_key(&k.k, f);
71 
72 	struct btree *b = bch2_btree_node_get_noiter(trans, &k.k, f->btree_id, f->level, false);
73 	bool ret = !IS_ERR_OR_NULL(b);
74 	if (ret)
75 		six_unlock_read(&b->c.lock);
76 
77 	/*
78 	 * We might update this node's range; if that happens, we need the node
79 	 * to be re-read so the read path can trim keys that are no longer in
80 	 * this node
81 	 */
82 	if (b != btree_node_root(trans->c, b))
83 		bch2_btree_node_evict(trans, &k.k);
84 	return ret;
85 }
86 
87 static int found_btree_node_cmp_cookie(const void *_l, const void *_r)
88 {
89 	const struct found_btree_node *l = _l;
90 	const struct found_btree_node *r = _r;
91 
92 	return  cmp_int(l->btree_id,	r->btree_id) ?:
93 		cmp_int(l->level,	r->level) ?:
94 		cmp_int(l->cookie,	r->cookie);
95 }
96 
97 /*
98  * Given two found btree nodes, if their sequence numbers are equal, take the
99  * one that's readable:
100  */
101 static int found_btree_node_cmp_time(const struct found_btree_node *l,
102 				     const struct found_btree_node *r)
103 {
104 	return cmp_int(l->seq, r->seq);
105 }
106 
107 static int found_btree_node_cmp_pos(const void *_l, const void *_r)
108 {
109 	const struct found_btree_node *l = _l;
110 	const struct found_btree_node *r = _r;
111 
112 	return  cmp_int(l->btree_id,	r->btree_id) ?:
113 	       -cmp_int(l->level,	r->level) ?:
114 		bpos_cmp(l->min_key,	r->min_key) ?:
115 	       -found_btree_node_cmp_time(l, r);
116 }
117 
118 static void try_read_btree_node(struct find_btree_nodes *f, struct bch_dev *ca,
119 				struct bio *bio, struct btree_node *bn, u64 offset)
120 {
121 	struct bch_fs *c = container_of(f, struct bch_fs, found_btree_nodes);
122 
123 	bio_reset(bio, ca->disk_sb.bdev, REQ_OP_READ);
124 	bio->bi_iter.bi_sector	= offset;
125 	bch2_bio_map(bio, bn, PAGE_SIZE);
126 
127 	submit_bio_wait(bio);
128 	if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_read,
129 			       "IO error in try_read_btree_node() at %llu: %s",
130 			       offset, bch2_blk_status_to_str(bio->bi_status)))
131 		return;
132 
133 	if (le64_to_cpu(bn->magic) != bset_magic(c))
134 		return;
135 
136 	if (bch2_csum_type_is_encryption(BSET_CSUM_TYPE(&bn->keys))) {
137 		struct nonce nonce = btree_nonce(&bn->keys, 0);
138 		unsigned bytes = (void *) &bn->keys - (void *) &bn->flags;
139 
140 		bch2_encrypt(c, BSET_CSUM_TYPE(&bn->keys), nonce, &bn->flags, bytes);
141 	}
142 
143 	if (btree_id_is_alloc(BTREE_NODE_ID(bn)))
144 		return;
145 
146 	if (BTREE_NODE_LEVEL(bn) >= BTREE_MAX_DEPTH)
147 		return;
148 
149 	rcu_read_lock();
150 	struct found_btree_node n = {
151 		.btree_id	= BTREE_NODE_ID(bn),
152 		.level		= BTREE_NODE_LEVEL(bn),
153 		.seq		= BTREE_NODE_SEQ(bn),
154 		.cookie		= le64_to_cpu(bn->keys.seq),
155 		.min_key	= bn->min_key,
156 		.max_key	= bn->max_key,
157 		.nr_ptrs	= 1,
158 		.ptrs[0].type	= 1 << BCH_EXTENT_ENTRY_ptr,
159 		.ptrs[0].offset	= offset,
160 		.ptrs[0].dev	= ca->dev_idx,
161 		.ptrs[0].gen	= *bucket_gen(ca, sector_to_bucket(ca, offset)),
162 	};
163 	rcu_read_unlock();
164 
165 	if (bch2_trans_run(c, found_btree_node_is_readable(trans, &n))) {
166 		mutex_lock(&f->lock);
167 		if (BSET_BIG_ENDIAN(&bn->keys) != CPU_BIG_ENDIAN) {
168 			bch_err(c, "try_read_btree_node() can't handle endian conversion");
169 			f->ret = -EINVAL;
170 			goto unlock;
171 		}
172 
173 		if (darray_push(&f->nodes, n))
174 			f->ret = -ENOMEM;
175 unlock:
176 		mutex_unlock(&f->lock);
177 	}
178 }
179 
180 static int read_btree_nodes_worker(void *p)
181 {
182 	struct find_btree_nodes_worker *w = p;
183 	struct bch_fs *c = container_of(w->f, struct bch_fs, found_btree_nodes);
184 	struct bch_dev *ca = w->ca;
185 	void *buf = (void *) __get_free_page(GFP_KERNEL);
186 	struct bio *bio = bio_alloc(NULL, 1, 0, GFP_KERNEL);
187 	unsigned long last_print = jiffies;
188 
189 	if (!buf || !bio) {
190 		bch_err(c, "read_btree_nodes_worker: error allocating bio/buf");
191 		w->f->ret = -ENOMEM;
192 		goto err;
193 	}
194 
195 	for (u64 bucket = ca->mi.first_bucket; bucket < ca->mi.nbuckets; bucket++)
196 		for (unsigned bucket_offset = 0;
197 		     bucket_offset + btree_sectors(c) <= ca->mi.bucket_size;
198 		     bucket_offset += btree_sectors(c)) {
199 			if (time_after(jiffies, last_print + HZ * 30)) {
200 				u64 cur_sector = bucket * ca->mi.bucket_size + bucket_offset;
201 				u64 end_sector = ca->mi.nbuckets * ca->mi.bucket_size;
202 
203 				bch_info(ca, "%s: %2u%% done", __func__,
204 					 (unsigned) div64_u64(cur_sector * 100, end_sector));
205 				last_print = jiffies;
206 			}
207 
208 			u64 sector = bucket * ca->mi.bucket_size + bucket_offset;
209 
210 			if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_mi_btree_bitmap &&
211 			    !bch2_dev_btree_bitmap_marked_sectors(ca, sector, btree_sectors(c)))
212 				continue;
213 
214 			try_read_btree_node(w->f, ca, bio, buf, sector);
215 		}
216 err:
217 	bio_put(bio);
218 	free_page((unsigned long) buf);
219 	percpu_ref_get(&ca->io_ref);
220 	closure_put(w->cl);
221 	kfree(w);
222 	return 0;
223 }
224 
225 static int read_btree_nodes(struct find_btree_nodes *f)
226 {
227 	struct bch_fs *c = container_of(f, struct bch_fs, found_btree_nodes);
228 	struct closure cl;
229 	int ret = 0;
230 
231 	closure_init_stack(&cl);
232 
233 	for_each_online_member(c, ca) {
234 		if (!(ca->mi.data_allowed & BIT(BCH_DATA_btree)))
235 			continue;
236 
237 		struct find_btree_nodes_worker *w = kmalloc(sizeof(*w), GFP_KERNEL);
238 		struct task_struct *t;
239 
240 		if (!w) {
241 			percpu_ref_put(&ca->io_ref);
242 			ret = -ENOMEM;
243 			goto err;
244 		}
245 
246 		percpu_ref_get(&ca->io_ref);
247 		closure_get(&cl);
248 		w->cl		= &cl;
249 		w->f		= f;
250 		w->ca		= ca;
251 
252 		t = kthread_run(read_btree_nodes_worker, w, "read_btree_nodes/%s", ca->name);
253 		ret = IS_ERR_OR_NULL(t);
254 		if (ret) {
255 			percpu_ref_put(&ca->io_ref);
256 			closure_put(&cl);
257 			f->ret = ret;
258 			bch_err(c, "error starting kthread: %i", ret);
259 			break;
260 		}
261 	}
262 err:
263 	closure_sync(&cl);
264 	return f->ret ?: ret;
265 }
266 
267 static void bubble_up(struct found_btree_node *n, struct found_btree_node *end)
268 {
269 	while (n + 1 < end &&
270 	       found_btree_node_cmp_pos(n, n + 1) > 0) {
271 		swap(n[0], n[1]);
272 		n++;
273 	}
274 }
275 
276 static int handle_overwrites(struct bch_fs *c,
277 			     struct found_btree_node *start,
278 			     struct found_btree_node *end)
279 {
280 	struct found_btree_node *n;
281 again:
282 	for (n = start + 1;
283 	     n < end &&
284 	     n->btree_id	== start->btree_id &&
285 	     n->level		== start->level &&
286 	     bpos_lt(n->min_key, start->max_key);
287 	     n++)  {
288 		int cmp = found_btree_node_cmp_time(start, n);
289 
290 		if (cmp > 0) {
291 			if (bpos_cmp(start->max_key, n->max_key) >= 0)
292 				n->overwritten = true;
293 			else {
294 				n->range_updated = true;
295 				n->min_key = bpos_successor(start->max_key);
296 				n->range_updated = true;
297 				bubble_up(n, end);
298 				goto again;
299 			}
300 		} else if (cmp < 0) {
301 			BUG_ON(bpos_cmp(n->min_key, start->min_key) <= 0);
302 
303 			start->max_key = bpos_predecessor(n->min_key);
304 			start->range_updated = true;
305 		} else {
306 			struct printbuf buf = PRINTBUF;
307 
308 			prt_str(&buf, "overlapping btree nodes with same seq! halting\n  ");
309 			found_btree_node_to_text(&buf, c, start);
310 			prt_str(&buf, "\n  ");
311 			found_btree_node_to_text(&buf, c, n);
312 			bch_err(c, "%s", buf.buf);
313 			printbuf_exit(&buf);
314 			return -BCH_ERR_fsck_repair_unimplemented;
315 		}
316 	}
317 
318 	return 0;
319 }
320 
321 int bch2_scan_for_btree_nodes(struct bch_fs *c)
322 {
323 	struct find_btree_nodes *f = &c->found_btree_nodes;
324 	struct printbuf buf = PRINTBUF;
325 	size_t dst;
326 	int ret = 0;
327 
328 	if (f->nodes.nr)
329 		return 0;
330 
331 	mutex_init(&f->lock);
332 
333 	ret = read_btree_nodes(f);
334 	if (ret)
335 		return ret;
336 
337 	if (!f->nodes.nr) {
338 		bch_err(c, "%s: no btree nodes found", __func__);
339 		ret = -EINVAL;
340 		goto err;
341 	}
342 
343 	if (0 && c->opts.verbose) {
344 		printbuf_reset(&buf);
345 		prt_printf(&buf, "%s: nodes found:\n", __func__);
346 		found_btree_nodes_to_text(&buf, c, f->nodes);
347 		bch2_print_string_as_lines(KERN_INFO, buf.buf);
348 	}
349 
350 	sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_cookie, NULL);
351 
352 	dst = 0;
353 	darray_for_each(f->nodes, i) {
354 		struct found_btree_node *prev = dst ? f->nodes.data + dst - 1 : NULL;
355 
356 		if (prev &&
357 		    prev->cookie == i->cookie) {
358 			if (prev->nr_ptrs == ARRAY_SIZE(prev->ptrs)) {
359 				bch_err(c, "%s: found too many replicas for btree node", __func__);
360 				ret = -EINVAL;
361 				goto err;
362 			}
363 			prev->ptrs[prev->nr_ptrs++] = i->ptrs[0];
364 		} else {
365 			f->nodes.data[dst++] = *i;
366 		}
367 	}
368 	f->nodes.nr = dst;
369 
370 	sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_pos, NULL);
371 
372 	if (0 && c->opts.verbose) {
373 		printbuf_reset(&buf);
374 		prt_printf(&buf, "%s: nodes after merging replicas:\n", __func__);
375 		found_btree_nodes_to_text(&buf, c, f->nodes);
376 		bch2_print_string_as_lines(KERN_INFO, buf.buf);
377 	}
378 
379 	dst = 0;
380 	darray_for_each(f->nodes, i) {
381 		if (i->overwritten)
382 			continue;
383 
384 		ret = handle_overwrites(c, i, &darray_top(f->nodes));
385 		if (ret)
386 			goto err;
387 
388 		BUG_ON(i->overwritten);
389 		f->nodes.data[dst++] = *i;
390 	}
391 	f->nodes.nr = dst;
392 
393 	if (c->opts.verbose) {
394 		printbuf_reset(&buf);
395 		prt_printf(&buf, "%s: nodes found after overwrites:\n", __func__);
396 		found_btree_nodes_to_text(&buf, c, f->nodes);
397 		bch2_print_string_as_lines(KERN_INFO, buf.buf);
398 	}
399 
400 	eytzinger0_sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_pos, NULL);
401 err:
402 	printbuf_exit(&buf);
403 	return ret;
404 }
405 
406 static int found_btree_node_range_start_cmp(const void *_l, const void *_r)
407 {
408 	const struct found_btree_node *l = _l;
409 	const struct found_btree_node *r = _r;
410 
411 	return  cmp_int(l->btree_id,	r->btree_id) ?:
412 	       -cmp_int(l->level,	r->level) ?:
413 		bpos_cmp(l->max_key,	r->min_key);
414 }
415 
416 #define for_each_found_btree_node_in_range(_f, _search, _idx)				\
417 	for (size_t _idx = eytzinger0_find_gt((_f)->nodes.data, (_f)->nodes.nr,		\
418 					sizeof((_f)->nodes.data[0]),			\
419 					found_btree_node_range_start_cmp, &search);	\
420 	     _idx < (_f)->nodes.nr &&							\
421 	     (_f)->nodes.data[_idx].btree_id == _search.btree_id &&			\
422 	     (_f)->nodes.data[_idx].level == _search.level &&				\
423 	     bpos_lt((_f)->nodes.data[_idx].min_key, _search.max_key);			\
424 	     _idx = eytzinger0_next(_idx, (_f)->nodes.nr))
425 
426 bool bch2_btree_node_is_stale(struct bch_fs *c, struct btree *b)
427 {
428 	struct find_btree_nodes *f = &c->found_btree_nodes;
429 
430 	struct found_btree_node search = {
431 		.btree_id	= b->c.btree_id,
432 		.level		= b->c.level,
433 		.min_key	= b->data->min_key,
434 		.max_key	= b->key.k.p,
435 	};
436 
437 	for_each_found_btree_node_in_range(f, search, idx)
438 		if (f->nodes.data[idx].seq > BTREE_NODE_SEQ(b->data))
439 			return true;
440 	return false;
441 }
442 
443 bool bch2_btree_has_scanned_nodes(struct bch_fs *c, enum btree_id btree)
444 {
445 	struct found_btree_node search = {
446 		.btree_id	= btree,
447 		.level		= 0,
448 		.min_key	= POS_MIN,
449 		.max_key	= SPOS_MAX,
450 	};
451 
452 	for_each_found_btree_node_in_range(&c->found_btree_nodes, search, idx)
453 		return true;
454 	return false;
455 }
456 
457 int bch2_get_scanned_nodes(struct bch_fs *c, enum btree_id btree,
458 			   unsigned level, struct bpos node_min, struct bpos node_max)
459 {
460 	if (btree_id_is_alloc(btree))
461 		return 0;
462 
463 	struct find_btree_nodes *f = &c->found_btree_nodes;
464 
465 	int ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
466 	if (ret)
467 		return ret;
468 
469 	if (c->opts.verbose) {
470 		struct printbuf buf = PRINTBUF;
471 
472 		prt_printf(&buf, "recovering %s l=%u ", bch2_btree_id_str(btree), level);
473 		bch2_bpos_to_text(&buf, node_min);
474 		prt_str(&buf, " - ");
475 		bch2_bpos_to_text(&buf, node_max);
476 
477 		bch_info(c, "%s(): %s", __func__, buf.buf);
478 		printbuf_exit(&buf);
479 	}
480 
481 	struct found_btree_node search = {
482 		.btree_id	= btree,
483 		.level		= level,
484 		.min_key	= node_min,
485 		.max_key	= node_max,
486 	};
487 
488 	for_each_found_btree_node_in_range(f, search, idx) {
489 		struct found_btree_node n = f->nodes.data[idx];
490 
491 		n.range_updated |= bpos_lt(n.min_key, node_min);
492 		n.min_key = bpos_max(n.min_key, node_min);
493 
494 		n.range_updated |= bpos_gt(n.max_key, node_max);
495 		n.max_key = bpos_min(n.max_key, node_max);
496 
497 		struct { __BKEY_PADDED(k, BKEY_BTREE_PTR_VAL_U64s_MAX); } tmp;
498 
499 		found_btree_node_to_key(&tmp.k, &n);
500 
501 		struct printbuf buf = PRINTBUF;
502 		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&tmp.k));
503 		bch_verbose(c, "%s(): recovering %s", __func__, buf.buf);
504 		printbuf_exit(&buf);
505 
506 		BUG_ON(bch2_bkey_invalid(c, bkey_i_to_s_c(&tmp.k), BKEY_TYPE_btree, 0, NULL));
507 
508 		ret = bch2_journal_key_insert(c, btree, level + 1, &tmp.k);
509 		if (ret)
510 			return ret;
511 	}
512 
513 	return 0;
514 }
515 
516 void bch2_find_btree_nodes_exit(struct find_btree_nodes *f)
517 {
518 	darray_exit(&f->nodes);
519 }
520