xref: /openbmc/linux/net/ceph/osdmap.c (revision 65417d9f)
1 // SPDX-License-Identifier: GPL-2.0
2 
3 #include <linux/ceph/ceph_debug.h>
4 
5 #include <linux/module.h>
6 #include <linux/slab.h>
7 #include <asm/div64.h>
8 
9 #include <linux/ceph/libceph.h>
10 #include <linux/ceph/osdmap.h>
11 #include <linux/ceph/decode.h>
12 #include <linux/crush/hash.h>
13 #include <linux/crush/mapper.h>
14 
15 char *ceph_osdmap_state_str(char *str, int len, u32 state)
16 {
17 	if (!len)
18 		return str;
19 
20 	if ((state & CEPH_OSD_EXISTS) && (state & CEPH_OSD_UP))
21 		snprintf(str, len, "exists, up");
22 	else if (state & CEPH_OSD_EXISTS)
23 		snprintf(str, len, "exists");
24 	else if (state & CEPH_OSD_UP)
25 		snprintf(str, len, "up");
26 	else
27 		snprintf(str, len, "doesn't exist");
28 
29 	return str;
30 }
31 
32 /* maps */
33 
34 static int calc_bits_of(unsigned int t)
35 {
36 	int b = 0;
37 	while (t) {
38 		t = t >> 1;
39 		b++;
40 	}
41 	return b;
42 }
43 
44 /*
45  * the foo_mask is the smallest value 2^n-1 that is >= foo.
46  */
47 static void calc_pg_masks(struct ceph_pg_pool_info *pi)
48 {
49 	pi->pg_num_mask = (1 << calc_bits_of(pi->pg_num-1)) - 1;
50 	pi->pgp_num_mask = (1 << calc_bits_of(pi->pgp_num-1)) - 1;
51 }
52 
53 /*
54  * decode crush map
55  */
56 static int crush_decode_uniform_bucket(void **p, void *end,
57 				       struct crush_bucket_uniform *b)
58 {
59 	dout("crush_decode_uniform_bucket %p to %p\n", *p, end);
60 	ceph_decode_need(p, end, (1+b->h.size) * sizeof(u32), bad);
61 	b->item_weight = ceph_decode_32(p);
62 	return 0;
63 bad:
64 	return -EINVAL;
65 }
66 
67 static int crush_decode_list_bucket(void **p, void *end,
68 				    struct crush_bucket_list *b)
69 {
70 	int j;
71 	dout("crush_decode_list_bucket %p to %p\n", *p, end);
72 	b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
73 	if (b->item_weights == NULL)
74 		return -ENOMEM;
75 	b->sum_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
76 	if (b->sum_weights == NULL)
77 		return -ENOMEM;
78 	ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad);
79 	for (j = 0; j < b->h.size; j++) {
80 		b->item_weights[j] = ceph_decode_32(p);
81 		b->sum_weights[j] = ceph_decode_32(p);
82 	}
83 	return 0;
84 bad:
85 	return -EINVAL;
86 }
87 
88 static int crush_decode_tree_bucket(void **p, void *end,
89 				    struct crush_bucket_tree *b)
90 {
91 	int j;
92 	dout("crush_decode_tree_bucket %p to %p\n", *p, end);
93 	ceph_decode_8_safe(p, end, b->num_nodes, bad);
94 	b->node_weights = kcalloc(b->num_nodes, sizeof(u32), GFP_NOFS);
95 	if (b->node_weights == NULL)
96 		return -ENOMEM;
97 	ceph_decode_need(p, end, b->num_nodes * sizeof(u32), bad);
98 	for (j = 0; j < b->num_nodes; j++)
99 		b->node_weights[j] = ceph_decode_32(p);
100 	return 0;
101 bad:
102 	return -EINVAL;
103 }
104 
105 static int crush_decode_straw_bucket(void **p, void *end,
106 				     struct crush_bucket_straw *b)
107 {
108 	int j;
109 	dout("crush_decode_straw_bucket %p to %p\n", *p, end);
110 	b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
111 	if (b->item_weights == NULL)
112 		return -ENOMEM;
113 	b->straws = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
114 	if (b->straws == NULL)
115 		return -ENOMEM;
116 	ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad);
117 	for (j = 0; j < b->h.size; j++) {
118 		b->item_weights[j] = ceph_decode_32(p);
119 		b->straws[j] = ceph_decode_32(p);
120 	}
121 	return 0;
122 bad:
123 	return -EINVAL;
124 }
125 
126 static int crush_decode_straw2_bucket(void **p, void *end,
127 				      struct crush_bucket_straw2 *b)
128 {
129 	int j;
130 	dout("crush_decode_straw2_bucket %p to %p\n", *p, end);
131 	b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
132 	if (b->item_weights == NULL)
133 		return -ENOMEM;
134 	ceph_decode_need(p, end, b->h.size * sizeof(u32), bad);
135 	for (j = 0; j < b->h.size; j++)
136 		b->item_weights[j] = ceph_decode_32(p);
137 	return 0;
138 bad:
139 	return -EINVAL;
140 }
141 
142 static struct crush_choose_arg_map *alloc_choose_arg_map(void)
143 {
144 	struct crush_choose_arg_map *arg_map;
145 
146 	arg_map = kzalloc(sizeof(*arg_map), GFP_NOIO);
147 	if (!arg_map)
148 		return NULL;
149 
150 	RB_CLEAR_NODE(&arg_map->node);
151 	return arg_map;
152 }
153 
154 static void free_choose_arg_map(struct crush_choose_arg_map *arg_map)
155 {
156 	if (arg_map) {
157 		int i, j;
158 
159 		WARN_ON(!RB_EMPTY_NODE(&arg_map->node));
160 
161 		for (i = 0; i < arg_map->size; i++) {
162 			struct crush_choose_arg *arg = &arg_map->args[i];
163 
164 			for (j = 0; j < arg->weight_set_size; j++)
165 				kfree(arg->weight_set[j].weights);
166 			kfree(arg->weight_set);
167 			kfree(arg->ids);
168 		}
169 		kfree(arg_map->args);
170 		kfree(arg_map);
171 	}
172 }
173 
174 DEFINE_RB_FUNCS(choose_arg_map, struct crush_choose_arg_map, choose_args_index,
175 		node);
176 
177 void clear_choose_args(struct crush_map *c)
178 {
179 	while (!RB_EMPTY_ROOT(&c->choose_args)) {
180 		struct crush_choose_arg_map *arg_map =
181 		    rb_entry(rb_first(&c->choose_args),
182 			     struct crush_choose_arg_map, node);
183 
184 		erase_choose_arg_map(&c->choose_args, arg_map);
185 		free_choose_arg_map(arg_map);
186 	}
187 }
188 
189 static u32 *decode_array_32_alloc(void **p, void *end, u32 *plen)
190 {
191 	u32 *a = NULL;
192 	u32 len;
193 	int ret;
194 
195 	ceph_decode_32_safe(p, end, len, e_inval);
196 	if (len) {
197 		u32 i;
198 
199 		a = kmalloc_array(len, sizeof(u32), GFP_NOIO);
200 		if (!a) {
201 			ret = -ENOMEM;
202 			goto fail;
203 		}
204 
205 		ceph_decode_need(p, end, len * sizeof(u32), e_inval);
206 		for (i = 0; i < len; i++)
207 			a[i] = ceph_decode_32(p);
208 	}
209 
210 	*plen = len;
211 	return a;
212 
213 e_inval:
214 	ret = -EINVAL;
215 fail:
216 	kfree(a);
217 	return ERR_PTR(ret);
218 }
219 
220 /*
221  * Assumes @arg is zero-initialized.
222  */
223 static int decode_choose_arg(void **p, void *end, struct crush_choose_arg *arg)
224 {
225 	int ret;
226 
227 	ceph_decode_32_safe(p, end, arg->weight_set_size, e_inval);
228 	if (arg->weight_set_size) {
229 		u32 i;
230 
231 		arg->weight_set = kmalloc_array(arg->weight_set_size,
232 						sizeof(*arg->weight_set),
233 						GFP_NOIO);
234 		if (!arg->weight_set)
235 			return -ENOMEM;
236 
237 		for (i = 0; i < arg->weight_set_size; i++) {
238 			struct crush_weight_set *w = &arg->weight_set[i];
239 
240 			w->weights = decode_array_32_alloc(p, end, &w->size);
241 			if (IS_ERR(w->weights)) {
242 				ret = PTR_ERR(w->weights);
243 				w->weights = NULL;
244 				return ret;
245 			}
246 		}
247 	}
248 
249 	arg->ids = decode_array_32_alloc(p, end, &arg->ids_size);
250 	if (IS_ERR(arg->ids)) {
251 		ret = PTR_ERR(arg->ids);
252 		arg->ids = NULL;
253 		return ret;
254 	}
255 
256 	return 0;
257 
258 e_inval:
259 	return -EINVAL;
260 }
261 
262 static int decode_choose_args(void **p, void *end, struct crush_map *c)
263 {
264 	struct crush_choose_arg_map *arg_map = NULL;
265 	u32 num_choose_arg_maps, num_buckets;
266 	int ret;
267 
268 	ceph_decode_32_safe(p, end, num_choose_arg_maps, e_inval);
269 	while (num_choose_arg_maps--) {
270 		arg_map = alloc_choose_arg_map();
271 		if (!arg_map) {
272 			ret = -ENOMEM;
273 			goto fail;
274 		}
275 
276 		ceph_decode_64_safe(p, end, arg_map->choose_args_index,
277 				    e_inval);
278 		arg_map->size = c->max_buckets;
279 		arg_map->args = kcalloc(arg_map->size, sizeof(*arg_map->args),
280 					GFP_NOIO);
281 		if (!arg_map->args) {
282 			ret = -ENOMEM;
283 			goto fail;
284 		}
285 
286 		ceph_decode_32_safe(p, end, num_buckets, e_inval);
287 		while (num_buckets--) {
288 			struct crush_choose_arg *arg;
289 			u32 bucket_index;
290 
291 			ceph_decode_32_safe(p, end, bucket_index, e_inval);
292 			if (bucket_index >= arg_map->size)
293 				goto e_inval;
294 
295 			arg = &arg_map->args[bucket_index];
296 			ret = decode_choose_arg(p, end, arg);
297 			if (ret)
298 				goto fail;
299 
300 			if (arg->ids_size &&
301 			    arg->ids_size != c->buckets[bucket_index]->size)
302 				goto e_inval;
303 		}
304 
305 		insert_choose_arg_map(&c->choose_args, arg_map);
306 	}
307 
308 	return 0;
309 
310 e_inval:
311 	ret = -EINVAL;
312 fail:
313 	free_choose_arg_map(arg_map);
314 	return ret;
315 }
316 
317 static void crush_finalize(struct crush_map *c)
318 {
319 	__s32 b;
320 
321 	/* Space for the array of pointers to per-bucket workspace */
322 	c->working_size = sizeof(struct crush_work) +
323 	    c->max_buckets * sizeof(struct crush_work_bucket *);
324 
325 	for (b = 0; b < c->max_buckets; b++) {
326 		if (!c->buckets[b])
327 			continue;
328 
329 		switch (c->buckets[b]->alg) {
330 		default:
331 			/*
332 			 * The base case, permutation variables and
333 			 * the pointer to the permutation array.
334 			 */
335 			c->working_size += sizeof(struct crush_work_bucket);
336 			break;
337 		}
338 		/* Every bucket has a permutation array. */
339 		c->working_size += c->buckets[b]->size * sizeof(__u32);
340 	}
341 }
342 
343 static struct crush_map *crush_decode(void *pbyval, void *end)
344 {
345 	struct crush_map *c;
346 	int err;
347 	int i, j;
348 	void **p = &pbyval;
349 	void *start = pbyval;
350 	u32 magic;
351 
352 	dout("crush_decode %p to %p len %d\n", *p, end, (int)(end - *p));
353 
354 	c = kzalloc(sizeof(*c), GFP_NOFS);
355 	if (c == NULL)
356 		return ERR_PTR(-ENOMEM);
357 
358 	c->choose_args = RB_ROOT;
359 
360         /* set tunables to default values */
361         c->choose_local_tries = 2;
362         c->choose_local_fallback_tries = 5;
363         c->choose_total_tries = 19;
364 	c->chooseleaf_descend_once = 0;
365 
366 	ceph_decode_need(p, end, 4*sizeof(u32), bad);
367 	magic = ceph_decode_32(p);
368 	if (magic != CRUSH_MAGIC) {
369 		pr_err("crush_decode magic %x != current %x\n",
370 		       (unsigned int)magic, (unsigned int)CRUSH_MAGIC);
371 		goto bad;
372 	}
373 	c->max_buckets = ceph_decode_32(p);
374 	c->max_rules = ceph_decode_32(p);
375 	c->max_devices = ceph_decode_32(p);
376 
377 	c->buckets = kcalloc(c->max_buckets, sizeof(*c->buckets), GFP_NOFS);
378 	if (c->buckets == NULL)
379 		goto badmem;
380 	c->rules = kcalloc(c->max_rules, sizeof(*c->rules), GFP_NOFS);
381 	if (c->rules == NULL)
382 		goto badmem;
383 
384 	/* buckets */
385 	for (i = 0; i < c->max_buckets; i++) {
386 		int size = 0;
387 		u32 alg;
388 		struct crush_bucket *b;
389 
390 		ceph_decode_32_safe(p, end, alg, bad);
391 		if (alg == 0) {
392 			c->buckets[i] = NULL;
393 			continue;
394 		}
395 		dout("crush_decode bucket %d off %x %p to %p\n",
396 		     i, (int)(*p-start), *p, end);
397 
398 		switch (alg) {
399 		case CRUSH_BUCKET_UNIFORM:
400 			size = sizeof(struct crush_bucket_uniform);
401 			break;
402 		case CRUSH_BUCKET_LIST:
403 			size = sizeof(struct crush_bucket_list);
404 			break;
405 		case CRUSH_BUCKET_TREE:
406 			size = sizeof(struct crush_bucket_tree);
407 			break;
408 		case CRUSH_BUCKET_STRAW:
409 			size = sizeof(struct crush_bucket_straw);
410 			break;
411 		case CRUSH_BUCKET_STRAW2:
412 			size = sizeof(struct crush_bucket_straw2);
413 			break;
414 		default:
415 			goto bad;
416 		}
417 		BUG_ON(size == 0);
418 		b = c->buckets[i] = kzalloc(size, GFP_NOFS);
419 		if (b == NULL)
420 			goto badmem;
421 
422 		ceph_decode_need(p, end, 4*sizeof(u32), bad);
423 		b->id = ceph_decode_32(p);
424 		b->type = ceph_decode_16(p);
425 		b->alg = ceph_decode_8(p);
426 		b->hash = ceph_decode_8(p);
427 		b->weight = ceph_decode_32(p);
428 		b->size = ceph_decode_32(p);
429 
430 		dout("crush_decode bucket size %d off %x %p to %p\n",
431 		     b->size, (int)(*p-start), *p, end);
432 
433 		b->items = kcalloc(b->size, sizeof(__s32), GFP_NOFS);
434 		if (b->items == NULL)
435 			goto badmem;
436 
437 		ceph_decode_need(p, end, b->size*sizeof(u32), bad);
438 		for (j = 0; j < b->size; j++)
439 			b->items[j] = ceph_decode_32(p);
440 
441 		switch (b->alg) {
442 		case CRUSH_BUCKET_UNIFORM:
443 			err = crush_decode_uniform_bucket(p, end,
444 				  (struct crush_bucket_uniform *)b);
445 			if (err < 0)
446 				goto fail;
447 			break;
448 		case CRUSH_BUCKET_LIST:
449 			err = crush_decode_list_bucket(p, end,
450 			       (struct crush_bucket_list *)b);
451 			if (err < 0)
452 				goto fail;
453 			break;
454 		case CRUSH_BUCKET_TREE:
455 			err = crush_decode_tree_bucket(p, end,
456 				(struct crush_bucket_tree *)b);
457 			if (err < 0)
458 				goto fail;
459 			break;
460 		case CRUSH_BUCKET_STRAW:
461 			err = crush_decode_straw_bucket(p, end,
462 				(struct crush_bucket_straw *)b);
463 			if (err < 0)
464 				goto fail;
465 			break;
466 		case CRUSH_BUCKET_STRAW2:
467 			err = crush_decode_straw2_bucket(p, end,
468 				(struct crush_bucket_straw2 *)b);
469 			if (err < 0)
470 				goto fail;
471 			break;
472 		}
473 	}
474 
475 	/* rules */
476 	dout("rule vec is %p\n", c->rules);
477 	for (i = 0; i < c->max_rules; i++) {
478 		u32 yes;
479 		struct crush_rule *r;
480 
481 		ceph_decode_32_safe(p, end, yes, bad);
482 		if (!yes) {
483 			dout("crush_decode NO rule %d off %x %p to %p\n",
484 			     i, (int)(*p-start), *p, end);
485 			c->rules[i] = NULL;
486 			continue;
487 		}
488 
489 		dout("crush_decode rule %d off %x %p to %p\n",
490 		     i, (int)(*p-start), *p, end);
491 
492 		/* len */
493 		ceph_decode_32_safe(p, end, yes, bad);
494 #if BITS_PER_LONG == 32
495 		if (yes > (ULONG_MAX - sizeof(*r))
496 			  / sizeof(struct crush_rule_step))
497 			goto bad;
498 #endif
499 		r = c->rules[i] = kmalloc(sizeof(*r) +
500 					  yes*sizeof(struct crush_rule_step),
501 					  GFP_NOFS);
502 		if (r == NULL)
503 			goto badmem;
504 		dout(" rule %d is at %p\n", i, r);
505 		r->len = yes;
506 		ceph_decode_copy_safe(p, end, &r->mask, 4, bad); /* 4 u8's */
507 		ceph_decode_need(p, end, r->len*3*sizeof(u32), bad);
508 		for (j = 0; j < r->len; j++) {
509 			r->steps[j].op = ceph_decode_32(p);
510 			r->steps[j].arg1 = ceph_decode_32(p);
511 			r->steps[j].arg2 = ceph_decode_32(p);
512 		}
513 	}
514 
515 	ceph_decode_skip_map(p, end, 32, string, bad); /* type_map */
516 	ceph_decode_skip_map(p, end, 32, string, bad); /* name_map */
517 	ceph_decode_skip_map(p, end, 32, string, bad); /* rule_name_map */
518 
519         /* tunables */
520         ceph_decode_need(p, end, 3*sizeof(u32), done);
521         c->choose_local_tries = ceph_decode_32(p);
522         c->choose_local_fallback_tries =  ceph_decode_32(p);
523         c->choose_total_tries = ceph_decode_32(p);
524         dout("crush decode tunable choose_local_tries = %d\n",
525              c->choose_local_tries);
526         dout("crush decode tunable choose_local_fallback_tries = %d\n",
527              c->choose_local_fallback_tries);
528         dout("crush decode tunable choose_total_tries = %d\n",
529              c->choose_total_tries);
530 
531 	ceph_decode_need(p, end, sizeof(u32), done);
532 	c->chooseleaf_descend_once = ceph_decode_32(p);
533 	dout("crush decode tunable chooseleaf_descend_once = %d\n",
534 	     c->chooseleaf_descend_once);
535 
536 	ceph_decode_need(p, end, sizeof(u8), done);
537 	c->chooseleaf_vary_r = ceph_decode_8(p);
538 	dout("crush decode tunable chooseleaf_vary_r = %d\n",
539 	     c->chooseleaf_vary_r);
540 
541 	/* skip straw_calc_version, allowed_bucket_algs */
542 	ceph_decode_need(p, end, sizeof(u8) + sizeof(u32), done);
543 	*p += sizeof(u8) + sizeof(u32);
544 
545 	ceph_decode_need(p, end, sizeof(u8), done);
546 	c->chooseleaf_stable = ceph_decode_8(p);
547 	dout("crush decode tunable chooseleaf_stable = %d\n",
548 	     c->chooseleaf_stable);
549 
550 	if (*p != end) {
551 		/* class_map */
552 		ceph_decode_skip_map(p, end, 32, 32, bad);
553 		/* class_name */
554 		ceph_decode_skip_map(p, end, 32, string, bad);
555 		/* class_bucket */
556 		ceph_decode_skip_map_of_map(p, end, 32, 32, 32, bad);
557 	}
558 
559 	if (*p != end) {
560 		err = decode_choose_args(p, end, c);
561 		if (err)
562 			goto fail;
563 	}
564 
565 done:
566 	crush_finalize(c);
567 	dout("crush_decode success\n");
568 	return c;
569 
570 badmem:
571 	err = -ENOMEM;
572 fail:
573 	dout("crush_decode fail %d\n", err);
574 	crush_destroy(c);
575 	return ERR_PTR(err);
576 
577 bad:
578 	err = -EINVAL;
579 	goto fail;
580 }
581 
582 int ceph_pg_compare(const struct ceph_pg *lhs, const struct ceph_pg *rhs)
583 {
584 	if (lhs->pool < rhs->pool)
585 		return -1;
586 	if (lhs->pool > rhs->pool)
587 		return 1;
588 	if (lhs->seed < rhs->seed)
589 		return -1;
590 	if (lhs->seed > rhs->seed)
591 		return 1;
592 
593 	return 0;
594 }
595 
596 int ceph_spg_compare(const struct ceph_spg *lhs, const struct ceph_spg *rhs)
597 {
598 	int ret;
599 
600 	ret = ceph_pg_compare(&lhs->pgid, &rhs->pgid);
601 	if (ret)
602 		return ret;
603 
604 	if (lhs->shard < rhs->shard)
605 		return -1;
606 	if (lhs->shard > rhs->shard)
607 		return 1;
608 
609 	return 0;
610 }
611 
612 static struct ceph_pg_mapping *alloc_pg_mapping(size_t payload_len)
613 {
614 	struct ceph_pg_mapping *pg;
615 
616 	pg = kmalloc(sizeof(*pg) + payload_len, GFP_NOIO);
617 	if (!pg)
618 		return NULL;
619 
620 	RB_CLEAR_NODE(&pg->node);
621 	return pg;
622 }
623 
624 static void free_pg_mapping(struct ceph_pg_mapping *pg)
625 {
626 	WARN_ON(!RB_EMPTY_NODE(&pg->node));
627 
628 	kfree(pg);
629 }
630 
631 /*
632  * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid
633  * to a set of osds) and primary_temp (explicit primary setting)
634  */
635 DEFINE_RB_FUNCS2(pg_mapping, struct ceph_pg_mapping, pgid, ceph_pg_compare,
636 		 RB_BYPTR, const struct ceph_pg *, node)
637 
638 /*
639  * rbtree of pg pool info
640  */
641 static int __insert_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *new)
642 {
643 	struct rb_node **p = &root->rb_node;
644 	struct rb_node *parent = NULL;
645 	struct ceph_pg_pool_info *pi = NULL;
646 
647 	while (*p) {
648 		parent = *p;
649 		pi = rb_entry(parent, struct ceph_pg_pool_info, node);
650 		if (new->id < pi->id)
651 			p = &(*p)->rb_left;
652 		else if (new->id > pi->id)
653 			p = &(*p)->rb_right;
654 		else
655 			return -EEXIST;
656 	}
657 
658 	rb_link_node(&new->node, parent, p);
659 	rb_insert_color(&new->node, root);
660 	return 0;
661 }
662 
663 static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, u64 id)
664 {
665 	struct ceph_pg_pool_info *pi;
666 	struct rb_node *n = root->rb_node;
667 
668 	while (n) {
669 		pi = rb_entry(n, struct ceph_pg_pool_info, node);
670 		if (id < pi->id)
671 			n = n->rb_left;
672 		else if (id > pi->id)
673 			n = n->rb_right;
674 		else
675 			return pi;
676 	}
677 	return NULL;
678 }
679 
680 struct ceph_pg_pool_info *ceph_pg_pool_by_id(struct ceph_osdmap *map, u64 id)
681 {
682 	return __lookup_pg_pool(&map->pg_pools, id);
683 }
684 
685 const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id)
686 {
687 	struct ceph_pg_pool_info *pi;
688 
689 	if (id == CEPH_NOPOOL)
690 		return NULL;
691 
692 	if (WARN_ON_ONCE(id > (u64) INT_MAX))
693 		return NULL;
694 
695 	pi = __lookup_pg_pool(&map->pg_pools, (int) id);
696 
697 	return pi ? pi->name : NULL;
698 }
699 EXPORT_SYMBOL(ceph_pg_pool_name_by_id);
700 
701 int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name)
702 {
703 	struct rb_node *rbp;
704 
705 	for (rbp = rb_first(&map->pg_pools); rbp; rbp = rb_next(rbp)) {
706 		struct ceph_pg_pool_info *pi =
707 			rb_entry(rbp, struct ceph_pg_pool_info, node);
708 		if (pi->name && strcmp(pi->name, name) == 0)
709 			return pi->id;
710 	}
711 	return -ENOENT;
712 }
713 EXPORT_SYMBOL(ceph_pg_poolid_by_name);
714 
715 static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi)
716 {
717 	rb_erase(&pi->node, root);
718 	kfree(pi->name);
719 	kfree(pi);
720 }
721 
722 static int decode_pool(void **p, void *end, struct ceph_pg_pool_info *pi)
723 {
724 	u8 ev, cv;
725 	unsigned len, num;
726 	void *pool_end;
727 
728 	ceph_decode_need(p, end, 2 + 4, bad);
729 	ev = ceph_decode_8(p);  /* encoding version */
730 	cv = ceph_decode_8(p); /* compat version */
731 	if (ev < 5) {
732 		pr_warn("got v %d < 5 cv %d of ceph_pg_pool\n", ev, cv);
733 		return -EINVAL;
734 	}
735 	if (cv > 9) {
736 		pr_warn("got v %d cv %d > 9 of ceph_pg_pool\n", ev, cv);
737 		return -EINVAL;
738 	}
739 	len = ceph_decode_32(p);
740 	ceph_decode_need(p, end, len, bad);
741 	pool_end = *p + len;
742 
743 	pi->type = ceph_decode_8(p);
744 	pi->size = ceph_decode_8(p);
745 	pi->crush_ruleset = ceph_decode_8(p);
746 	pi->object_hash = ceph_decode_8(p);
747 
748 	pi->pg_num = ceph_decode_32(p);
749 	pi->pgp_num = ceph_decode_32(p);
750 
751 	*p += 4 + 4;  /* skip lpg* */
752 	*p += 4;      /* skip last_change */
753 	*p += 8 + 4;  /* skip snap_seq, snap_epoch */
754 
755 	/* skip snaps */
756 	num = ceph_decode_32(p);
757 	while (num--) {
758 		*p += 8;  /* snapid key */
759 		*p += 1 + 1; /* versions */
760 		len = ceph_decode_32(p);
761 		*p += len;
762 	}
763 
764 	/* skip removed_snaps */
765 	num = ceph_decode_32(p);
766 	*p += num * (8 + 8);
767 
768 	*p += 8;  /* skip auid */
769 	pi->flags = ceph_decode_64(p);
770 	*p += 4;  /* skip crash_replay_interval */
771 
772 	if (ev >= 7)
773 		pi->min_size = ceph_decode_8(p);
774 	else
775 		pi->min_size = pi->size - pi->size / 2;
776 
777 	if (ev >= 8)
778 		*p += 8 + 8;  /* skip quota_max_* */
779 
780 	if (ev >= 9) {
781 		/* skip tiers */
782 		num = ceph_decode_32(p);
783 		*p += num * 8;
784 
785 		*p += 8;  /* skip tier_of */
786 		*p += 1;  /* skip cache_mode */
787 
788 		pi->read_tier = ceph_decode_64(p);
789 		pi->write_tier = ceph_decode_64(p);
790 	} else {
791 		pi->read_tier = -1;
792 		pi->write_tier = -1;
793 	}
794 
795 	if (ev >= 10) {
796 		/* skip properties */
797 		num = ceph_decode_32(p);
798 		while (num--) {
799 			len = ceph_decode_32(p);
800 			*p += len; /* key */
801 			len = ceph_decode_32(p);
802 			*p += len; /* val */
803 		}
804 	}
805 
806 	if (ev >= 11) {
807 		/* skip hit_set_params */
808 		*p += 1 + 1; /* versions */
809 		len = ceph_decode_32(p);
810 		*p += len;
811 
812 		*p += 4; /* skip hit_set_period */
813 		*p += 4; /* skip hit_set_count */
814 	}
815 
816 	if (ev >= 12)
817 		*p += 4; /* skip stripe_width */
818 
819 	if (ev >= 13) {
820 		*p += 8; /* skip target_max_bytes */
821 		*p += 8; /* skip target_max_objects */
822 		*p += 4; /* skip cache_target_dirty_ratio_micro */
823 		*p += 4; /* skip cache_target_full_ratio_micro */
824 		*p += 4; /* skip cache_min_flush_age */
825 		*p += 4; /* skip cache_min_evict_age */
826 	}
827 
828 	if (ev >=  14) {
829 		/* skip erasure_code_profile */
830 		len = ceph_decode_32(p);
831 		*p += len;
832 	}
833 
834 	/*
835 	 * last_force_op_resend_preluminous, will be overridden if the
836 	 * map was encoded with RESEND_ON_SPLIT
837 	 */
838 	if (ev >= 15)
839 		pi->last_force_request_resend = ceph_decode_32(p);
840 	else
841 		pi->last_force_request_resend = 0;
842 
843 	if (ev >= 16)
844 		*p += 4; /* skip min_read_recency_for_promote */
845 
846 	if (ev >= 17)
847 		*p += 8; /* skip expected_num_objects */
848 
849 	if (ev >= 19)
850 		*p += 4; /* skip cache_target_dirty_high_ratio_micro */
851 
852 	if (ev >= 20)
853 		*p += 4; /* skip min_write_recency_for_promote */
854 
855 	if (ev >= 21)
856 		*p += 1; /* skip use_gmt_hitset */
857 
858 	if (ev >= 22)
859 		*p += 1; /* skip fast_read */
860 
861 	if (ev >= 23) {
862 		*p += 4; /* skip hit_set_grade_decay_rate */
863 		*p += 4; /* skip hit_set_search_last_n */
864 	}
865 
866 	if (ev >= 24) {
867 		/* skip opts */
868 		*p += 1 + 1; /* versions */
869 		len = ceph_decode_32(p);
870 		*p += len;
871 	}
872 
873 	if (ev >= 25)
874 		pi->last_force_request_resend = ceph_decode_32(p);
875 
876 	/* ignore the rest */
877 
878 	*p = pool_end;
879 	calc_pg_masks(pi);
880 	return 0;
881 
882 bad:
883 	return -EINVAL;
884 }
885 
886 static int decode_pool_names(void **p, void *end, struct ceph_osdmap *map)
887 {
888 	struct ceph_pg_pool_info *pi;
889 	u32 num, len;
890 	u64 pool;
891 
892 	ceph_decode_32_safe(p, end, num, bad);
893 	dout(" %d pool names\n", num);
894 	while (num--) {
895 		ceph_decode_64_safe(p, end, pool, bad);
896 		ceph_decode_32_safe(p, end, len, bad);
897 		dout("  pool %llu len %d\n", pool, len);
898 		ceph_decode_need(p, end, len, bad);
899 		pi = __lookup_pg_pool(&map->pg_pools, pool);
900 		if (pi) {
901 			char *name = kstrndup(*p, len, GFP_NOFS);
902 
903 			if (!name)
904 				return -ENOMEM;
905 			kfree(pi->name);
906 			pi->name = name;
907 			dout("  name is %s\n", pi->name);
908 		}
909 		*p += len;
910 	}
911 	return 0;
912 
913 bad:
914 	return -EINVAL;
915 }
916 
917 /*
918  * osd map
919  */
920 struct ceph_osdmap *ceph_osdmap_alloc(void)
921 {
922 	struct ceph_osdmap *map;
923 
924 	map = kzalloc(sizeof(*map), GFP_NOIO);
925 	if (!map)
926 		return NULL;
927 
928 	map->pg_pools = RB_ROOT;
929 	map->pool_max = -1;
930 	map->pg_temp = RB_ROOT;
931 	map->primary_temp = RB_ROOT;
932 	map->pg_upmap = RB_ROOT;
933 	map->pg_upmap_items = RB_ROOT;
934 	mutex_init(&map->crush_workspace_mutex);
935 
936 	return map;
937 }
938 
939 void ceph_osdmap_destroy(struct ceph_osdmap *map)
940 {
941 	dout("osdmap_destroy %p\n", map);
942 	if (map->crush)
943 		crush_destroy(map->crush);
944 	while (!RB_EMPTY_ROOT(&map->pg_temp)) {
945 		struct ceph_pg_mapping *pg =
946 			rb_entry(rb_first(&map->pg_temp),
947 				 struct ceph_pg_mapping, node);
948 		erase_pg_mapping(&map->pg_temp, pg);
949 		free_pg_mapping(pg);
950 	}
951 	while (!RB_EMPTY_ROOT(&map->primary_temp)) {
952 		struct ceph_pg_mapping *pg =
953 			rb_entry(rb_first(&map->primary_temp),
954 				 struct ceph_pg_mapping, node);
955 		erase_pg_mapping(&map->primary_temp, pg);
956 		free_pg_mapping(pg);
957 	}
958 	while (!RB_EMPTY_ROOT(&map->pg_upmap)) {
959 		struct ceph_pg_mapping *pg =
960 			rb_entry(rb_first(&map->pg_upmap),
961 				 struct ceph_pg_mapping, node);
962 		rb_erase(&pg->node, &map->pg_upmap);
963 		kfree(pg);
964 	}
965 	while (!RB_EMPTY_ROOT(&map->pg_upmap_items)) {
966 		struct ceph_pg_mapping *pg =
967 			rb_entry(rb_first(&map->pg_upmap_items),
968 				 struct ceph_pg_mapping, node);
969 		rb_erase(&pg->node, &map->pg_upmap_items);
970 		kfree(pg);
971 	}
972 	while (!RB_EMPTY_ROOT(&map->pg_pools)) {
973 		struct ceph_pg_pool_info *pi =
974 			rb_entry(rb_first(&map->pg_pools),
975 				 struct ceph_pg_pool_info, node);
976 		__remove_pg_pool(&map->pg_pools, pi);
977 	}
978 	kfree(map->osd_state);
979 	kfree(map->osd_weight);
980 	kfree(map->osd_addr);
981 	kfree(map->osd_primary_affinity);
982 	kfree(map->crush_workspace);
983 	kfree(map);
984 }
985 
986 /*
987  * Adjust max_osd value, (re)allocate arrays.
988  *
989  * The new elements are properly initialized.
990  */
991 static int osdmap_set_max_osd(struct ceph_osdmap *map, int max)
992 {
993 	u32 *state;
994 	u32 *weight;
995 	struct ceph_entity_addr *addr;
996 	int i;
997 
998 	state = krealloc(map->osd_state, max*sizeof(*state), GFP_NOFS);
999 	if (!state)
1000 		return -ENOMEM;
1001 	map->osd_state = state;
1002 
1003 	weight = krealloc(map->osd_weight, max*sizeof(*weight), GFP_NOFS);
1004 	if (!weight)
1005 		return -ENOMEM;
1006 	map->osd_weight = weight;
1007 
1008 	addr = krealloc(map->osd_addr, max*sizeof(*addr), GFP_NOFS);
1009 	if (!addr)
1010 		return -ENOMEM;
1011 	map->osd_addr = addr;
1012 
1013 	for (i = map->max_osd; i < max; i++) {
1014 		map->osd_state[i] = 0;
1015 		map->osd_weight[i] = CEPH_OSD_OUT;
1016 		memset(map->osd_addr + i, 0, sizeof(*map->osd_addr));
1017 	}
1018 
1019 	if (map->osd_primary_affinity) {
1020 		u32 *affinity;
1021 
1022 		affinity = krealloc(map->osd_primary_affinity,
1023 				    max*sizeof(*affinity), GFP_NOFS);
1024 		if (!affinity)
1025 			return -ENOMEM;
1026 		map->osd_primary_affinity = affinity;
1027 
1028 		for (i = map->max_osd; i < max; i++)
1029 			map->osd_primary_affinity[i] =
1030 			    CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1031 	}
1032 
1033 	map->max_osd = max;
1034 
1035 	return 0;
1036 }
1037 
1038 static int osdmap_set_crush(struct ceph_osdmap *map, struct crush_map *crush)
1039 {
1040 	void *workspace;
1041 	size_t work_size;
1042 
1043 	if (IS_ERR(crush))
1044 		return PTR_ERR(crush);
1045 
1046 	work_size = crush_work_size(crush, CEPH_PG_MAX_SIZE);
1047 	dout("%s work_size %zu bytes\n", __func__, work_size);
1048 	workspace = kmalloc(work_size, GFP_NOIO);
1049 	if (!workspace) {
1050 		crush_destroy(crush);
1051 		return -ENOMEM;
1052 	}
1053 	crush_init_workspace(crush, workspace);
1054 
1055 	if (map->crush)
1056 		crush_destroy(map->crush);
1057 	kfree(map->crush_workspace);
1058 	map->crush = crush;
1059 	map->crush_workspace = workspace;
1060 	return 0;
1061 }
1062 
1063 #define OSDMAP_WRAPPER_COMPAT_VER	7
1064 #define OSDMAP_CLIENT_DATA_COMPAT_VER	1
1065 
1066 /*
1067  * Return 0 or error.  On success, *v is set to 0 for old (v6) osdmaps,
1068  * to struct_v of the client_data section for new (v7 and above)
1069  * osdmaps.
1070  */
1071 static int get_osdmap_client_data_v(void **p, void *end,
1072 				    const char *prefix, u8 *v)
1073 {
1074 	u8 struct_v;
1075 
1076 	ceph_decode_8_safe(p, end, struct_v, e_inval);
1077 	if (struct_v >= 7) {
1078 		u8 struct_compat;
1079 
1080 		ceph_decode_8_safe(p, end, struct_compat, e_inval);
1081 		if (struct_compat > OSDMAP_WRAPPER_COMPAT_VER) {
1082 			pr_warn("got v %d cv %d > %d of %s ceph_osdmap\n",
1083 				struct_v, struct_compat,
1084 				OSDMAP_WRAPPER_COMPAT_VER, prefix);
1085 			return -EINVAL;
1086 		}
1087 		*p += 4; /* ignore wrapper struct_len */
1088 
1089 		ceph_decode_8_safe(p, end, struct_v, e_inval);
1090 		ceph_decode_8_safe(p, end, struct_compat, e_inval);
1091 		if (struct_compat > OSDMAP_CLIENT_DATA_COMPAT_VER) {
1092 			pr_warn("got v %d cv %d > %d of %s ceph_osdmap client data\n",
1093 				struct_v, struct_compat,
1094 				OSDMAP_CLIENT_DATA_COMPAT_VER, prefix);
1095 			return -EINVAL;
1096 		}
1097 		*p += 4; /* ignore client data struct_len */
1098 	} else {
1099 		u16 version;
1100 
1101 		*p -= 1;
1102 		ceph_decode_16_safe(p, end, version, e_inval);
1103 		if (version < 6) {
1104 			pr_warn("got v %d < 6 of %s ceph_osdmap\n",
1105 				version, prefix);
1106 			return -EINVAL;
1107 		}
1108 
1109 		/* old osdmap enconding */
1110 		struct_v = 0;
1111 	}
1112 
1113 	*v = struct_v;
1114 	return 0;
1115 
1116 e_inval:
1117 	return -EINVAL;
1118 }
1119 
1120 static int __decode_pools(void **p, void *end, struct ceph_osdmap *map,
1121 			  bool incremental)
1122 {
1123 	u32 n;
1124 
1125 	ceph_decode_32_safe(p, end, n, e_inval);
1126 	while (n--) {
1127 		struct ceph_pg_pool_info *pi;
1128 		u64 pool;
1129 		int ret;
1130 
1131 		ceph_decode_64_safe(p, end, pool, e_inval);
1132 
1133 		pi = __lookup_pg_pool(&map->pg_pools, pool);
1134 		if (!incremental || !pi) {
1135 			pi = kzalloc(sizeof(*pi), GFP_NOFS);
1136 			if (!pi)
1137 				return -ENOMEM;
1138 
1139 			pi->id = pool;
1140 
1141 			ret = __insert_pg_pool(&map->pg_pools, pi);
1142 			if (ret) {
1143 				kfree(pi);
1144 				return ret;
1145 			}
1146 		}
1147 
1148 		ret = decode_pool(p, end, pi);
1149 		if (ret)
1150 			return ret;
1151 	}
1152 
1153 	return 0;
1154 
1155 e_inval:
1156 	return -EINVAL;
1157 }
1158 
1159 static int decode_pools(void **p, void *end, struct ceph_osdmap *map)
1160 {
1161 	return __decode_pools(p, end, map, false);
1162 }
1163 
1164 static int decode_new_pools(void **p, void *end, struct ceph_osdmap *map)
1165 {
1166 	return __decode_pools(p, end, map, true);
1167 }
1168 
1169 typedef struct ceph_pg_mapping *(*decode_mapping_fn_t)(void **, void *, bool);
1170 
1171 static int decode_pg_mapping(void **p, void *end, struct rb_root *mapping_root,
1172 			     decode_mapping_fn_t fn, bool incremental)
1173 {
1174 	u32 n;
1175 
1176 	WARN_ON(!incremental && !fn);
1177 
1178 	ceph_decode_32_safe(p, end, n, e_inval);
1179 	while (n--) {
1180 		struct ceph_pg_mapping *pg;
1181 		struct ceph_pg pgid;
1182 		int ret;
1183 
1184 		ret = ceph_decode_pgid(p, end, &pgid);
1185 		if (ret)
1186 			return ret;
1187 
1188 		pg = lookup_pg_mapping(mapping_root, &pgid);
1189 		if (pg) {
1190 			WARN_ON(!incremental);
1191 			erase_pg_mapping(mapping_root, pg);
1192 			free_pg_mapping(pg);
1193 		}
1194 
1195 		if (fn) {
1196 			pg = fn(p, end, incremental);
1197 			if (IS_ERR(pg))
1198 				return PTR_ERR(pg);
1199 
1200 			if (pg) {
1201 				pg->pgid = pgid; /* struct */
1202 				insert_pg_mapping(mapping_root, pg);
1203 			}
1204 		}
1205 	}
1206 
1207 	return 0;
1208 
1209 e_inval:
1210 	return -EINVAL;
1211 }
1212 
1213 static struct ceph_pg_mapping *__decode_pg_temp(void **p, void *end,
1214 						bool incremental)
1215 {
1216 	struct ceph_pg_mapping *pg;
1217 	u32 len, i;
1218 
1219 	ceph_decode_32_safe(p, end, len, e_inval);
1220 	if (len == 0 && incremental)
1221 		return NULL;	/* new_pg_temp: [] to remove */
1222 	if (len > (SIZE_MAX - sizeof(*pg)) / sizeof(u32))
1223 		return ERR_PTR(-EINVAL);
1224 
1225 	ceph_decode_need(p, end, len * sizeof(u32), e_inval);
1226 	pg = alloc_pg_mapping(len * sizeof(u32));
1227 	if (!pg)
1228 		return ERR_PTR(-ENOMEM);
1229 
1230 	pg->pg_temp.len = len;
1231 	for (i = 0; i < len; i++)
1232 		pg->pg_temp.osds[i] = ceph_decode_32(p);
1233 
1234 	return pg;
1235 
1236 e_inval:
1237 	return ERR_PTR(-EINVAL);
1238 }
1239 
1240 static int decode_pg_temp(void **p, void *end, struct ceph_osdmap *map)
1241 {
1242 	return decode_pg_mapping(p, end, &map->pg_temp, __decode_pg_temp,
1243 				 false);
1244 }
1245 
1246 static int decode_new_pg_temp(void **p, void *end, struct ceph_osdmap *map)
1247 {
1248 	return decode_pg_mapping(p, end, &map->pg_temp, __decode_pg_temp,
1249 				 true);
1250 }
1251 
1252 static struct ceph_pg_mapping *__decode_primary_temp(void **p, void *end,
1253 						     bool incremental)
1254 {
1255 	struct ceph_pg_mapping *pg;
1256 	u32 osd;
1257 
1258 	ceph_decode_32_safe(p, end, osd, e_inval);
1259 	if (osd == (u32)-1 && incremental)
1260 		return NULL;	/* new_primary_temp: -1 to remove */
1261 
1262 	pg = alloc_pg_mapping(0);
1263 	if (!pg)
1264 		return ERR_PTR(-ENOMEM);
1265 
1266 	pg->primary_temp.osd = osd;
1267 	return pg;
1268 
1269 e_inval:
1270 	return ERR_PTR(-EINVAL);
1271 }
1272 
1273 static int decode_primary_temp(void **p, void *end, struct ceph_osdmap *map)
1274 {
1275 	return decode_pg_mapping(p, end, &map->primary_temp,
1276 				 __decode_primary_temp, false);
1277 }
1278 
1279 static int decode_new_primary_temp(void **p, void *end,
1280 				   struct ceph_osdmap *map)
1281 {
1282 	return decode_pg_mapping(p, end, &map->primary_temp,
1283 				 __decode_primary_temp, true);
1284 }
1285 
1286 u32 ceph_get_primary_affinity(struct ceph_osdmap *map, int osd)
1287 {
1288 	BUG_ON(osd >= map->max_osd);
1289 
1290 	if (!map->osd_primary_affinity)
1291 		return CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1292 
1293 	return map->osd_primary_affinity[osd];
1294 }
1295 
1296 static int set_primary_affinity(struct ceph_osdmap *map, int osd, u32 aff)
1297 {
1298 	BUG_ON(osd >= map->max_osd);
1299 
1300 	if (!map->osd_primary_affinity) {
1301 		int i;
1302 
1303 		map->osd_primary_affinity = kmalloc(map->max_osd*sizeof(u32),
1304 						    GFP_NOFS);
1305 		if (!map->osd_primary_affinity)
1306 			return -ENOMEM;
1307 
1308 		for (i = 0; i < map->max_osd; i++)
1309 			map->osd_primary_affinity[i] =
1310 			    CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1311 	}
1312 
1313 	map->osd_primary_affinity[osd] = aff;
1314 
1315 	return 0;
1316 }
1317 
1318 static int decode_primary_affinity(void **p, void *end,
1319 				   struct ceph_osdmap *map)
1320 {
1321 	u32 len, i;
1322 
1323 	ceph_decode_32_safe(p, end, len, e_inval);
1324 	if (len == 0) {
1325 		kfree(map->osd_primary_affinity);
1326 		map->osd_primary_affinity = NULL;
1327 		return 0;
1328 	}
1329 	if (len != map->max_osd)
1330 		goto e_inval;
1331 
1332 	ceph_decode_need(p, end, map->max_osd*sizeof(u32), e_inval);
1333 
1334 	for (i = 0; i < map->max_osd; i++) {
1335 		int ret;
1336 
1337 		ret = set_primary_affinity(map, i, ceph_decode_32(p));
1338 		if (ret)
1339 			return ret;
1340 	}
1341 
1342 	return 0;
1343 
1344 e_inval:
1345 	return -EINVAL;
1346 }
1347 
1348 static int decode_new_primary_affinity(void **p, void *end,
1349 				       struct ceph_osdmap *map)
1350 {
1351 	u32 n;
1352 
1353 	ceph_decode_32_safe(p, end, n, e_inval);
1354 	while (n--) {
1355 		u32 osd, aff;
1356 		int ret;
1357 
1358 		ceph_decode_32_safe(p, end, osd, e_inval);
1359 		ceph_decode_32_safe(p, end, aff, e_inval);
1360 
1361 		ret = set_primary_affinity(map, osd, aff);
1362 		if (ret)
1363 			return ret;
1364 
1365 		pr_info("osd%d primary-affinity 0x%x\n", osd, aff);
1366 	}
1367 
1368 	return 0;
1369 
1370 e_inval:
1371 	return -EINVAL;
1372 }
1373 
1374 static struct ceph_pg_mapping *__decode_pg_upmap(void **p, void *end,
1375 						 bool __unused)
1376 {
1377 	return __decode_pg_temp(p, end, false);
1378 }
1379 
1380 static int decode_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1381 {
1382 	return decode_pg_mapping(p, end, &map->pg_upmap, __decode_pg_upmap,
1383 				 false);
1384 }
1385 
1386 static int decode_new_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1387 {
1388 	return decode_pg_mapping(p, end, &map->pg_upmap, __decode_pg_upmap,
1389 				 true);
1390 }
1391 
1392 static int decode_old_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1393 {
1394 	return decode_pg_mapping(p, end, &map->pg_upmap, NULL, true);
1395 }
1396 
1397 static struct ceph_pg_mapping *__decode_pg_upmap_items(void **p, void *end,
1398 						       bool __unused)
1399 {
1400 	struct ceph_pg_mapping *pg;
1401 	u32 len, i;
1402 
1403 	ceph_decode_32_safe(p, end, len, e_inval);
1404 	if (len > (SIZE_MAX - sizeof(*pg)) / (2 * sizeof(u32)))
1405 		return ERR_PTR(-EINVAL);
1406 
1407 	ceph_decode_need(p, end, 2 * len * sizeof(u32), e_inval);
1408 	pg = alloc_pg_mapping(2 * len * sizeof(u32));
1409 	if (!pg)
1410 		return ERR_PTR(-ENOMEM);
1411 
1412 	pg->pg_upmap_items.len = len;
1413 	for (i = 0; i < len; i++) {
1414 		pg->pg_upmap_items.from_to[i][0] = ceph_decode_32(p);
1415 		pg->pg_upmap_items.from_to[i][1] = ceph_decode_32(p);
1416 	}
1417 
1418 	return pg;
1419 
1420 e_inval:
1421 	return ERR_PTR(-EINVAL);
1422 }
1423 
1424 static int decode_pg_upmap_items(void **p, void *end, struct ceph_osdmap *map)
1425 {
1426 	return decode_pg_mapping(p, end, &map->pg_upmap_items,
1427 				 __decode_pg_upmap_items, false);
1428 }
1429 
1430 static int decode_new_pg_upmap_items(void **p, void *end,
1431 				     struct ceph_osdmap *map)
1432 {
1433 	return decode_pg_mapping(p, end, &map->pg_upmap_items,
1434 				 __decode_pg_upmap_items, true);
1435 }
1436 
1437 static int decode_old_pg_upmap_items(void **p, void *end,
1438 				     struct ceph_osdmap *map)
1439 {
1440 	return decode_pg_mapping(p, end, &map->pg_upmap_items, NULL, true);
1441 }
1442 
1443 /*
1444  * decode a full map.
1445  */
1446 static int osdmap_decode(void **p, void *end, struct ceph_osdmap *map)
1447 {
1448 	u8 struct_v;
1449 	u32 epoch = 0;
1450 	void *start = *p;
1451 	u32 max;
1452 	u32 len, i;
1453 	int err;
1454 
1455 	dout("%s %p to %p len %d\n", __func__, *p, end, (int)(end - *p));
1456 
1457 	err = get_osdmap_client_data_v(p, end, "full", &struct_v);
1458 	if (err)
1459 		goto bad;
1460 
1461 	/* fsid, epoch, created, modified */
1462 	ceph_decode_need(p, end, sizeof(map->fsid) + sizeof(u32) +
1463 			 sizeof(map->created) + sizeof(map->modified), e_inval);
1464 	ceph_decode_copy(p, &map->fsid, sizeof(map->fsid));
1465 	epoch = map->epoch = ceph_decode_32(p);
1466 	ceph_decode_copy(p, &map->created, sizeof(map->created));
1467 	ceph_decode_copy(p, &map->modified, sizeof(map->modified));
1468 
1469 	/* pools */
1470 	err = decode_pools(p, end, map);
1471 	if (err)
1472 		goto bad;
1473 
1474 	/* pool_name */
1475 	err = decode_pool_names(p, end, map);
1476 	if (err)
1477 		goto bad;
1478 
1479 	ceph_decode_32_safe(p, end, map->pool_max, e_inval);
1480 
1481 	ceph_decode_32_safe(p, end, map->flags, e_inval);
1482 
1483 	/* max_osd */
1484 	ceph_decode_32_safe(p, end, max, e_inval);
1485 
1486 	/* (re)alloc osd arrays */
1487 	err = osdmap_set_max_osd(map, max);
1488 	if (err)
1489 		goto bad;
1490 
1491 	/* osd_state, osd_weight, osd_addrs->client_addr */
1492 	ceph_decode_need(p, end, 3*sizeof(u32) +
1493 			 map->max_osd*((struct_v >= 5 ? sizeof(u32) :
1494 							sizeof(u8)) +
1495 				       sizeof(*map->osd_weight) +
1496 				       sizeof(*map->osd_addr)), e_inval);
1497 
1498 	if (ceph_decode_32(p) != map->max_osd)
1499 		goto e_inval;
1500 
1501 	if (struct_v >= 5) {
1502 		for (i = 0; i < map->max_osd; i++)
1503 			map->osd_state[i] = ceph_decode_32(p);
1504 	} else {
1505 		for (i = 0; i < map->max_osd; i++)
1506 			map->osd_state[i] = ceph_decode_8(p);
1507 	}
1508 
1509 	if (ceph_decode_32(p) != map->max_osd)
1510 		goto e_inval;
1511 
1512 	for (i = 0; i < map->max_osd; i++)
1513 		map->osd_weight[i] = ceph_decode_32(p);
1514 
1515 	if (ceph_decode_32(p) != map->max_osd)
1516 		goto e_inval;
1517 
1518 	ceph_decode_copy(p, map->osd_addr, map->max_osd*sizeof(*map->osd_addr));
1519 	for (i = 0; i < map->max_osd; i++)
1520 		ceph_decode_addr(&map->osd_addr[i]);
1521 
1522 	/* pg_temp */
1523 	err = decode_pg_temp(p, end, map);
1524 	if (err)
1525 		goto bad;
1526 
1527 	/* primary_temp */
1528 	if (struct_v >= 1) {
1529 		err = decode_primary_temp(p, end, map);
1530 		if (err)
1531 			goto bad;
1532 	}
1533 
1534 	/* primary_affinity */
1535 	if (struct_v >= 2) {
1536 		err = decode_primary_affinity(p, end, map);
1537 		if (err)
1538 			goto bad;
1539 	} else {
1540 		WARN_ON(map->osd_primary_affinity);
1541 	}
1542 
1543 	/* crush */
1544 	ceph_decode_32_safe(p, end, len, e_inval);
1545 	err = osdmap_set_crush(map, crush_decode(*p, min(*p + len, end)));
1546 	if (err)
1547 		goto bad;
1548 
1549 	*p += len;
1550 	if (struct_v >= 3) {
1551 		/* erasure_code_profiles */
1552 		ceph_decode_skip_map_of_map(p, end, string, string, string,
1553 					    e_inval);
1554 	}
1555 
1556 	if (struct_v >= 4) {
1557 		err = decode_pg_upmap(p, end, map);
1558 		if (err)
1559 			goto bad;
1560 
1561 		err = decode_pg_upmap_items(p, end, map);
1562 		if (err)
1563 			goto bad;
1564 	} else {
1565 		WARN_ON(!RB_EMPTY_ROOT(&map->pg_upmap));
1566 		WARN_ON(!RB_EMPTY_ROOT(&map->pg_upmap_items));
1567 	}
1568 
1569 	/* ignore the rest */
1570 	*p = end;
1571 
1572 	dout("full osdmap epoch %d max_osd %d\n", map->epoch, map->max_osd);
1573 	return 0;
1574 
1575 e_inval:
1576 	err = -EINVAL;
1577 bad:
1578 	pr_err("corrupt full osdmap (%d) epoch %d off %d (%p of %p-%p)\n",
1579 	       err, epoch, (int)(*p - start), *p, start, end);
1580 	print_hex_dump(KERN_DEBUG, "osdmap: ",
1581 		       DUMP_PREFIX_OFFSET, 16, 1,
1582 		       start, end - start, true);
1583 	return err;
1584 }
1585 
1586 /*
1587  * Allocate and decode a full map.
1588  */
1589 struct ceph_osdmap *ceph_osdmap_decode(void **p, void *end)
1590 {
1591 	struct ceph_osdmap *map;
1592 	int ret;
1593 
1594 	map = ceph_osdmap_alloc();
1595 	if (!map)
1596 		return ERR_PTR(-ENOMEM);
1597 
1598 	ret = osdmap_decode(p, end, map);
1599 	if (ret) {
1600 		ceph_osdmap_destroy(map);
1601 		return ERR_PTR(ret);
1602 	}
1603 
1604 	return map;
1605 }
1606 
1607 /*
1608  * Encoding order is (new_up_client, new_state, new_weight).  Need to
1609  * apply in the (new_weight, new_state, new_up_client) order, because
1610  * an incremental map may look like e.g.
1611  *
1612  *     new_up_client: { osd=6, addr=... } # set osd_state and addr
1613  *     new_state: { osd=6, xorstate=EXISTS } # clear osd_state
1614  */
1615 static int decode_new_up_state_weight(void **p, void *end, u8 struct_v,
1616 				      struct ceph_osdmap *map)
1617 {
1618 	void *new_up_client;
1619 	void *new_state;
1620 	void *new_weight_end;
1621 	u32 len;
1622 
1623 	new_up_client = *p;
1624 	ceph_decode_32_safe(p, end, len, e_inval);
1625 	len *= sizeof(u32) + sizeof(struct ceph_entity_addr);
1626 	ceph_decode_need(p, end, len, e_inval);
1627 	*p += len;
1628 
1629 	new_state = *p;
1630 	ceph_decode_32_safe(p, end, len, e_inval);
1631 	len *= sizeof(u32) + (struct_v >= 5 ? sizeof(u32) : sizeof(u8));
1632 	ceph_decode_need(p, end, len, e_inval);
1633 	*p += len;
1634 
1635 	/* new_weight */
1636 	ceph_decode_32_safe(p, end, len, e_inval);
1637 	while (len--) {
1638 		s32 osd;
1639 		u32 w;
1640 
1641 		ceph_decode_need(p, end, 2*sizeof(u32), e_inval);
1642 		osd = ceph_decode_32(p);
1643 		w = ceph_decode_32(p);
1644 		BUG_ON(osd >= map->max_osd);
1645 		pr_info("osd%d weight 0x%x %s\n", osd, w,
1646 		     w == CEPH_OSD_IN ? "(in)" :
1647 		     (w == CEPH_OSD_OUT ? "(out)" : ""));
1648 		map->osd_weight[osd] = w;
1649 
1650 		/*
1651 		 * If we are marking in, set the EXISTS, and clear the
1652 		 * AUTOOUT and NEW bits.
1653 		 */
1654 		if (w) {
1655 			map->osd_state[osd] |= CEPH_OSD_EXISTS;
1656 			map->osd_state[osd] &= ~(CEPH_OSD_AUTOOUT |
1657 						 CEPH_OSD_NEW);
1658 		}
1659 	}
1660 	new_weight_end = *p;
1661 
1662 	/* new_state (up/down) */
1663 	*p = new_state;
1664 	len = ceph_decode_32(p);
1665 	while (len--) {
1666 		s32 osd;
1667 		u32 xorstate;
1668 		int ret;
1669 
1670 		osd = ceph_decode_32(p);
1671 		if (struct_v >= 5)
1672 			xorstate = ceph_decode_32(p);
1673 		else
1674 			xorstate = ceph_decode_8(p);
1675 		if (xorstate == 0)
1676 			xorstate = CEPH_OSD_UP;
1677 		BUG_ON(osd >= map->max_osd);
1678 		if ((map->osd_state[osd] & CEPH_OSD_UP) &&
1679 		    (xorstate & CEPH_OSD_UP))
1680 			pr_info("osd%d down\n", osd);
1681 		if ((map->osd_state[osd] & CEPH_OSD_EXISTS) &&
1682 		    (xorstate & CEPH_OSD_EXISTS)) {
1683 			pr_info("osd%d does not exist\n", osd);
1684 			ret = set_primary_affinity(map, osd,
1685 						   CEPH_OSD_DEFAULT_PRIMARY_AFFINITY);
1686 			if (ret)
1687 				return ret;
1688 			memset(map->osd_addr + osd, 0, sizeof(*map->osd_addr));
1689 			map->osd_state[osd] = 0;
1690 		} else {
1691 			map->osd_state[osd] ^= xorstate;
1692 		}
1693 	}
1694 
1695 	/* new_up_client */
1696 	*p = new_up_client;
1697 	len = ceph_decode_32(p);
1698 	while (len--) {
1699 		s32 osd;
1700 		struct ceph_entity_addr addr;
1701 
1702 		osd = ceph_decode_32(p);
1703 		ceph_decode_copy(p, &addr, sizeof(addr));
1704 		ceph_decode_addr(&addr);
1705 		BUG_ON(osd >= map->max_osd);
1706 		pr_info("osd%d up\n", osd);
1707 		map->osd_state[osd] |= CEPH_OSD_EXISTS | CEPH_OSD_UP;
1708 		map->osd_addr[osd] = addr;
1709 	}
1710 
1711 	*p = new_weight_end;
1712 	return 0;
1713 
1714 e_inval:
1715 	return -EINVAL;
1716 }
1717 
1718 /*
1719  * decode and apply an incremental map update.
1720  */
1721 struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
1722 					     struct ceph_osdmap *map)
1723 {
1724 	struct ceph_fsid fsid;
1725 	u32 epoch = 0;
1726 	struct ceph_timespec modified;
1727 	s32 len;
1728 	u64 pool;
1729 	__s64 new_pool_max;
1730 	__s32 new_flags, max;
1731 	void *start = *p;
1732 	int err;
1733 	u8 struct_v;
1734 
1735 	dout("%s %p to %p len %d\n", __func__, *p, end, (int)(end - *p));
1736 
1737 	err = get_osdmap_client_data_v(p, end, "inc", &struct_v);
1738 	if (err)
1739 		goto bad;
1740 
1741 	/* fsid, epoch, modified, new_pool_max, new_flags */
1742 	ceph_decode_need(p, end, sizeof(fsid) + sizeof(u32) + sizeof(modified) +
1743 			 sizeof(u64) + sizeof(u32), e_inval);
1744 	ceph_decode_copy(p, &fsid, sizeof(fsid));
1745 	epoch = ceph_decode_32(p);
1746 	BUG_ON(epoch != map->epoch+1);
1747 	ceph_decode_copy(p, &modified, sizeof(modified));
1748 	new_pool_max = ceph_decode_64(p);
1749 	new_flags = ceph_decode_32(p);
1750 
1751 	/* full map? */
1752 	ceph_decode_32_safe(p, end, len, e_inval);
1753 	if (len > 0) {
1754 		dout("apply_incremental full map len %d, %p to %p\n",
1755 		     len, *p, end);
1756 		return ceph_osdmap_decode(p, min(*p+len, end));
1757 	}
1758 
1759 	/* new crush? */
1760 	ceph_decode_32_safe(p, end, len, e_inval);
1761 	if (len > 0) {
1762 		err = osdmap_set_crush(map,
1763 				       crush_decode(*p, min(*p + len, end)));
1764 		if (err)
1765 			goto bad;
1766 		*p += len;
1767 	}
1768 
1769 	/* new flags? */
1770 	if (new_flags >= 0)
1771 		map->flags = new_flags;
1772 	if (new_pool_max >= 0)
1773 		map->pool_max = new_pool_max;
1774 
1775 	/* new max? */
1776 	ceph_decode_32_safe(p, end, max, e_inval);
1777 	if (max >= 0) {
1778 		err = osdmap_set_max_osd(map, max);
1779 		if (err)
1780 			goto bad;
1781 	}
1782 
1783 	map->epoch++;
1784 	map->modified = modified;
1785 
1786 	/* new_pools */
1787 	err = decode_new_pools(p, end, map);
1788 	if (err)
1789 		goto bad;
1790 
1791 	/* new_pool_names */
1792 	err = decode_pool_names(p, end, map);
1793 	if (err)
1794 		goto bad;
1795 
1796 	/* old_pool */
1797 	ceph_decode_32_safe(p, end, len, e_inval);
1798 	while (len--) {
1799 		struct ceph_pg_pool_info *pi;
1800 
1801 		ceph_decode_64_safe(p, end, pool, e_inval);
1802 		pi = __lookup_pg_pool(&map->pg_pools, pool);
1803 		if (pi)
1804 			__remove_pg_pool(&map->pg_pools, pi);
1805 	}
1806 
1807 	/* new_up_client, new_state, new_weight */
1808 	err = decode_new_up_state_weight(p, end, struct_v, map);
1809 	if (err)
1810 		goto bad;
1811 
1812 	/* new_pg_temp */
1813 	err = decode_new_pg_temp(p, end, map);
1814 	if (err)
1815 		goto bad;
1816 
1817 	/* new_primary_temp */
1818 	if (struct_v >= 1) {
1819 		err = decode_new_primary_temp(p, end, map);
1820 		if (err)
1821 			goto bad;
1822 	}
1823 
1824 	/* new_primary_affinity */
1825 	if (struct_v >= 2) {
1826 		err = decode_new_primary_affinity(p, end, map);
1827 		if (err)
1828 			goto bad;
1829 	}
1830 
1831 	if (struct_v >= 3) {
1832 		/* new_erasure_code_profiles */
1833 		ceph_decode_skip_map_of_map(p, end, string, string, string,
1834 					    e_inval);
1835 		/* old_erasure_code_profiles */
1836 		ceph_decode_skip_set(p, end, string, e_inval);
1837 	}
1838 
1839 	if (struct_v >= 4) {
1840 		err = decode_new_pg_upmap(p, end, map);
1841 		if (err)
1842 			goto bad;
1843 
1844 		err = decode_old_pg_upmap(p, end, map);
1845 		if (err)
1846 			goto bad;
1847 
1848 		err = decode_new_pg_upmap_items(p, end, map);
1849 		if (err)
1850 			goto bad;
1851 
1852 		err = decode_old_pg_upmap_items(p, end, map);
1853 		if (err)
1854 			goto bad;
1855 	}
1856 
1857 	/* ignore the rest */
1858 	*p = end;
1859 
1860 	dout("inc osdmap epoch %d max_osd %d\n", map->epoch, map->max_osd);
1861 	return map;
1862 
1863 e_inval:
1864 	err = -EINVAL;
1865 bad:
1866 	pr_err("corrupt inc osdmap (%d) epoch %d off %d (%p of %p-%p)\n",
1867 	       err, epoch, (int)(*p - start), *p, start, end);
1868 	print_hex_dump(KERN_DEBUG, "osdmap: ",
1869 		       DUMP_PREFIX_OFFSET, 16, 1,
1870 		       start, end - start, true);
1871 	return ERR_PTR(err);
1872 }
1873 
1874 void ceph_oloc_copy(struct ceph_object_locator *dest,
1875 		    const struct ceph_object_locator *src)
1876 {
1877 	ceph_oloc_destroy(dest);
1878 
1879 	dest->pool = src->pool;
1880 	if (src->pool_ns)
1881 		dest->pool_ns = ceph_get_string(src->pool_ns);
1882 	else
1883 		dest->pool_ns = NULL;
1884 }
1885 EXPORT_SYMBOL(ceph_oloc_copy);
1886 
1887 void ceph_oloc_destroy(struct ceph_object_locator *oloc)
1888 {
1889 	ceph_put_string(oloc->pool_ns);
1890 }
1891 EXPORT_SYMBOL(ceph_oloc_destroy);
1892 
1893 void ceph_oid_copy(struct ceph_object_id *dest,
1894 		   const struct ceph_object_id *src)
1895 {
1896 	ceph_oid_destroy(dest);
1897 
1898 	if (src->name != src->inline_name) {
1899 		/* very rare, see ceph_object_id definition */
1900 		dest->name = kmalloc(src->name_len + 1,
1901 				     GFP_NOIO | __GFP_NOFAIL);
1902 	} else {
1903 		dest->name = dest->inline_name;
1904 	}
1905 	memcpy(dest->name, src->name, src->name_len + 1);
1906 	dest->name_len = src->name_len;
1907 }
1908 EXPORT_SYMBOL(ceph_oid_copy);
1909 
1910 static __printf(2, 0)
1911 int oid_printf_vargs(struct ceph_object_id *oid, const char *fmt, va_list ap)
1912 {
1913 	int len;
1914 
1915 	WARN_ON(!ceph_oid_empty(oid));
1916 
1917 	len = vsnprintf(oid->inline_name, sizeof(oid->inline_name), fmt, ap);
1918 	if (len >= sizeof(oid->inline_name))
1919 		return len;
1920 
1921 	oid->name_len = len;
1922 	return 0;
1923 }
1924 
1925 /*
1926  * If oid doesn't fit into inline buffer, BUG.
1927  */
1928 void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...)
1929 {
1930 	va_list ap;
1931 
1932 	va_start(ap, fmt);
1933 	BUG_ON(oid_printf_vargs(oid, fmt, ap));
1934 	va_end(ap);
1935 }
1936 EXPORT_SYMBOL(ceph_oid_printf);
1937 
1938 static __printf(3, 0)
1939 int oid_aprintf_vargs(struct ceph_object_id *oid, gfp_t gfp,
1940 		      const char *fmt, va_list ap)
1941 {
1942 	va_list aq;
1943 	int len;
1944 
1945 	va_copy(aq, ap);
1946 	len = oid_printf_vargs(oid, fmt, aq);
1947 	va_end(aq);
1948 
1949 	if (len) {
1950 		char *external_name;
1951 
1952 		external_name = kmalloc(len + 1, gfp);
1953 		if (!external_name)
1954 			return -ENOMEM;
1955 
1956 		oid->name = external_name;
1957 		WARN_ON(vsnprintf(oid->name, len + 1, fmt, ap) != len);
1958 		oid->name_len = len;
1959 	}
1960 
1961 	return 0;
1962 }
1963 
1964 /*
1965  * If oid doesn't fit into inline buffer, allocate.
1966  */
1967 int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp,
1968 		     const char *fmt, ...)
1969 {
1970 	va_list ap;
1971 	int ret;
1972 
1973 	va_start(ap, fmt);
1974 	ret = oid_aprintf_vargs(oid, gfp, fmt, ap);
1975 	va_end(ap);
1976 
1977 	return ret;
1978 }
1979 EXPORT_SYMBOL(ceph_oid_aprintf);
1980 
1981 void ceph_oid_destroy(struct ceph_object_id *oid)
1982 {
1983 	if (oid->name != oid->inline_name)
1984 		kfree(oid->name);
1985 }
1986 EXPORT_SYMBOL(ceph_oid_destroy);
1987 
1988 /*
1989  * osds only
1990  */
1991 static bool __osds_equal(const struct ceph_osds *lhs,
1992 			 const struct ceph_osds *rhs)
1993 {
1994 	if (lhs->size == rhs->size &&
1995 	    !memcmp(lhs->osds, rhs->osds, rhs->size * sizeof(rhs->osds[0])))
1996 		return true;
1997 
1998 	return false;
1999 }
2000 
2001 /*
2002  * osds + primary
2003  */
2004 static bool osds_equal(const struct ceph_osds *lhs,
2005 		       const struct ceph_osds *rhs)
2006 {
2007 	if (__osds_equal(lhs, rhs) &&
2008 	    lhs->primary == rhs->primary)
2009 		return true;
2010 
2011 	return false;
2012 }
2013 
2014 static bool osds_valid(const struct ceph_osds *set)
2015 {
2016 	/* non-empty set */
2017 	if (set->size > 0 && set->primary >= 0)
2018 		return true;
2019 
2020 	/* empty can_shift_osds set */
2021 	if (!set->size && set->primary == -1)
2022 		return true;
2023 
2024 	/* empty !can_shift_osds set - all NONE */
2025 	if (set->size > 0 && set->primary == -1) {
2026 		int i;
2027 
2028 		for (i = 0; i < set->size; i++) {
2029 			if (set->osds[i] != CRUSH_ITEM_NONE)
2030 				break;
2031 		}
2032 		if (i == set->size)
2033 			return true;
2034 	}
2035 
2036 	return false;
2037 }
2038 
2039 void ceph_osds_copy(struct ceph_osds *dest, const struct ceph_osds *src)
2040 {
2041 	memcpy(dest->osds, src->osds, src->size * sizeof(src->osds[0]));
2042 	dest->size = src->size;
2043 	dest->primary = src->primary;
2044 }
2045 
2046 bool ceph_pg_is_split(const struct ceph_pg *pgid, u32 old_pg_num,
2047 		      u32 new_pg_num)
2048 {
2049 	int old_bits = calc_bits_of(old_pg_num);
2050 	int old_mask = (1 << old_bits) - 1;
2051 	int n;
2052 
2053 	WARN_ON(pgid->seed >= old_pg_num);
2054 	if (new_pg_num <= old_pg_num)
2055 		return false;
2056 
2057 	for (n = 1; ; n++) {
2058 		int next_bit = n << (old_bits - 1);
2059 		u32 s = next_bit | pgid->seed;
2060 
2061 		if (s < old_pg_num || s == pgid->seed)
2062 			continue;
2063 		if (s >= new_pg_num)
2064 			break;
2065 
2066 		s = ceph_stable_mod(s, old_pg_num, old_mask);
2067 		if (s == pgid->seed)
2068 			return true;
2069 	}
2070 
2071 	return false;
2072 }
2073 
2074 bool ceph_is_new_interval(const struct ceph_osds *old_acting,
2075 			  const struct ceph_osds *new_acting,
2076 			  const struct ceph_osds *old_up,
2077 			  const struct ceph_osds *new_up,
2078 			  int old_size,
2079 			  int new_size,
2080 			  int old_min_size,
2081 			  int new_min_size,
2082 			  u32 old_pg_num,
2083 			  u32 new_pg_num,
2084 			  bool old_sort_bitwise,
2085 			  bool new_sort_bitwise,
2086 			  bool old_recovery_deletes,
2087 			  bool new_recovery_deletes,
2088 			  const struct ceph_pg *pgid)
2089 {
2090 	return !osds_equal(old_acting, new_acting) ||
2091 	       !osds_equal(old_up, new_up) ||
2092 	       old_size != new_size ||
2093 	       old_min_size != new_min_size ||
2094 	       ceph_pg_is_split(pgid, old_pg_num, new_pg_num) ||
2095 	       old_sort_bitwise != new_sort_bitwise ||
2096 	       old_recovery_deletes != new_recovery_deletes;
2097 }
2098 
2099 static int calc_pg_rank(int osd, const struct ceph_osds *acting)
2100 {
2101 	int i;
2102 
2103 	for (i = 0; i < acting->size; i++) {
2104 		if (acting->osds[i] == osd)
2105 			return i;
2106 	}
2107 
2108 	return -1;
2109 }
2110 
2111 static bool primary_changed(const struct ceph_osds *old_acting,
2112 			    const struct ceph_osds *new_acting)
2113 {
2114 	if (!old_acting->size && !new_acting->size)
2115 		return false; /* both still empty */
2116 
2117 	if (!old_acting->size ^ !new_acting->size)
2118 		return true; /* was empty, now not, or vice versa */
2119 
2120 	if (old_acting->primary != new_acting->primary)
2121 		return true; /* primary changed */
2122 
2123 	if (calc_pg_rank(old_acting->primary, old_acting) !=
2124 	    calc_pg_rank(new_acting->primary, new_acting))
2125 		return true;
2126 
2127 	return false; /* same primary (tho replicas may have changed) */
2128 }
2129 
2130 bool ceph_osds_changed(const struct ceph_osds *old_acting,
2131 		       const struct ceph_osds *new_acting,
2132 		       bool any_change)
2133 {
2134 	if (primary_changed(old_acting, new_acting))
2135 		return true;
2136 
2137 	if (any_change && !__osds_equal(old_acting, new_acting))
2138 		return true;
2139 
2140 	return false;
2141 }
2142 
2143 /*
2144  * calculate file layout from given offset, length.
2145  * fill in correct oid, logical length, and object extent
2146  * offset, length.
2147  *
2148  * for now, we write only a single su, until we can
2149  * pass a stride back to the caller.
2150  */
2151 int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
2152 				   u64 off, u64 len,
2153 				   u64 *ono,
2154 				   u64 *oxoff, u64 *oxlen)
2155 {
2156 	u32 osize = layout->object_size;
2157 	u32 su = layout->stripe_unit;
2158 	u32 sc = layout->stripe_count;
2159 	u32 bl, stripeno, stripepos, objsetno;
2160 	u32 su_per_object;
2161 	u64 t, su_offset;
2162 
2163 	dout("mapping %llu~%llu  osize %u fl_su %u\n", off, len,
2164 	     osize, su);
2165 	if (su == 0 || sc == 0)
2166 		goto invalid;
2167 	su_per_object = osize / su;
2168 	if (su_per_object == 0)
2169 		goto invalid;
2170 	dout("osize %u / su %u = su_per_object %u\n", osize, su,
2171 	     su_per_object);
2172 
2173 	if ((su & ~PAGE_MASK) != 0)
2174 		goto invalid;
2175 
2176 	/* bl = *off / su; */
2177 	t = off;
2178 	do_div(t, su);
2179 	bl = t;
2180 	dout("off %llu / su %u = bl %u\n", off, su, bl);
2181 
2182 	stripeno = bl / sc;
2183 	stripepos = bl % sc;
2184 	objsetno = stripeno / su_per_object;
2185 
2186 	*ono = objsetno * sc + stripepos;
2187 	dout("objset %u * sc %u = ono %u\n", objsetno, sc, (unsigned int)*ono);
2188 
2189 	/* *oxoff = *off % layout->fl_stripe_unit;  # offset in su */
2190 	t = off;
2191 	su_offset = do_div(t, su);
2192 	*oxoff = su_offset + (stripeno % su_per_object) * su;
2193 
2194 	/*
2195 	 * Calculate the length of the extent being written to the selected
2196 	 * object. This is the minimum of the full length requested (len) or
2197 	 * the remainder of the current stripe being written to.
2198 	 */
2199 	*oxlen = min_t(u64, len, su - su_offset);
2200 
2201 	dout(" obj extent %llu~%llu\n", *oxoff, *oxlen);
2202 	return 0;
2203 
2204 invalid:
2205 	dout(" invalid layout\n");
2206 	*ono = 0;
2207 	*oxoff = 0;
2208 	*oxlen = 0;
2209 	return -EINVAL;
2210 }
2211 EXPORT_SYMBOL(ceph_calc_file_object_mapping);
2212 
2213 /*
2214  * Map an object into a PG.
2215  *
2216  * Should only be called with target_oid and target_oloc (as opposed to
2217  * base_oid and base_oloc), since tiering isn't taken into account.
2218  */
2219 int __ceph_object_locator_to_pg(struct ceph_pg_pool_info *pi,
2220 				const struct ceph_object_id *oid,
2221 				const struct ceph_object_locator *oloc,
2222 				struct ceph_pg *raw_pgid)
2223 {
2224 	WARN_ON(pi->id != oloc->pool);
2225 
2226 	if (!oloc->pool_ns) {
2227 		raw_pgid->pool = oloc->pool;
2228 		raw_pgid->seed = ceph_str_hash(pi->object_hash, oid->name,
2229 					     oid->name_len);
2230 		dout("%s %s -> raw_pgid %llu.%x\n", __func__, oid->name,
2231 		     raw_pgid->pool, raw_pgid->seed);
2232 	} else {
2233 		char stack_buf[256];
2234 		char *buf = stack_buf;
2235 		int nsl = oloc->pool_ns->len;
2236 		size_t total = nsl + 1 + oid->name_len;
2237 
2238 		if (total > sizeof(stack_buf)) {
2239 			buf = kmalloc(total, GFP_NOIO);
2240 			if (!buf)
2241 				return -ENOMEM;
2242 		}
2243 		memcpy(buf, oloc->pool_ns->str, nsl);
2244 		buf[nsl] = '\037';
2245 		memcpy(buf + nsl + 1, oid->name, oid->name_len);
2246 		raw_pgid->pool = oloc->pool;
2247 		raw_pgid->seed = ceph_str_hash(pi->object_hash, buf, total);
2248 		if (buf != stack_buf)
2249 			kfree(buf);
2250 		dout("%s %s ns %.*s -> raw_pgid %llu.%x\n", __func__,
2251 		     oid->name, nsl, oloc->pool_ns->str,
2252 		     raw_pgid->pool, raw_pgid->seed);
2253 	}
2254 	return 0;
2255 }
2256 
2257 int ceph_object_locator_to_pg(struct ceph_osdmap *osdmap,
2258 			      const struct ceph_object_id *oid,
2259 			      const struct ceph_object_locator *oloc,
2260 			      struct ceph_pg *raw_pgid)
2261 {
2262 	struct ceph_pg_pool_info *pi;
2263 
2264 	pi = ceph_pg_pool_by_id(osdmap, oloc->pool);
2265 	if (!pi)
2266 		return -ENOENT;
2267 
2268 	return __ceph_object_locator_to_pg(pi, oid, oloc, raw_pgid);
2269 }
2270 EXPORT_SYMBOL(ceph_object_locator_to_pg);
2271 
2272 /*
2273  * Map a raw PG (full precision ps) into an actual PG.
2274  */
2275 static void raw_pg_to_pg(struct ceph_pg_pool_info *pi,
2276 			 const struct ceph_pg *raw_pgid,
2277 			 struct ceph_pg *pgid)
2278 {
2279 	pgid->pool = raw_pgid->pool;
2280 	pgid->seed = ceph_stable_mod(raw_pgid->seed, pi->pg_num,
2281 				     pi->pg_num_mask);
2282 }
2283 
2284 /*
2285  * Map a raw PG (full precision ps) into a placement ps (placement
2286  * seed).  Include pool id in that value so that different pools don't
2287  * use the same seeds.
2288  */
2289 static u32 raw_pg_to_pps(struct ceph_pg_pool_info *pi,
2290 			 const struct ceph_pg *raw_pgid)
2291 {
2292 	if (pi->flags & CEPH_POOL_FLAG_HASHPSPOOL) {
2293 		/* hash pool id and seed so that pool PGs do not overlap */
2294 		return crush_hash32_2(CRUSH_HASH_RJENKINS1,
2295 				      ceph_stable_mod(raw_pgid->seed,
2296 						      pi->pgp_num,
2297 						      pi->pgp_num_mask),
2298 				      raw_pgid->pool);
2299 	} else {
2300 		/*
2301 		 * legacy behavior: add ps and pool together.  this is
2302 		 * not a great approach because the PGs from each pool
2303 		 * will overlap on top of each other: 0.5 == 1.4 ==
2304 		 * 2.3 == ...
2305 		 */
2306 		return ceph_stable_mod(raw_pgid->seed, pi->pgp_num,
2307 				       pi->pgp_num_mask) +
2308 		       (unsigned)raw_pgid->pool;
2309 	}
2310 }
2311 
2312 /*
2313  * Magic value used for a "default" fallback choose_args, used if the
2314  * crush_choose_arg_map passed to do_crush() does not exist.  If this
2315  * also doesn't exist, fall back to canonical weights.
2316  */
2317 #define CEPH_DEFAULT_CHOOSE_ARGS	-1
2318 
2319 static int do_crush(struct ceph_osdmap *map, int ruleno, int x,
2320 		    int *result, int result_max,
2321 		    const __u32 *weight, int weight_max,
2322 		    s64 choose_args_index)
2323 {
2324 	struct crush_choose_arg_map *arg_map;
2325 	int r;
2326 
2327 	BUG_ON(result_max > CEPH_PG_MAX_SIZE);
2328 
2329 	arg_map = lookup_choose_arg_map(&map->crush->choose_args,
2330 					choose_args_index);
2331 	if (!arg_map)
2332 		arg_map = lookup_choose_arg_map(&map->crush->choose_args,
2333 						CEPH_DEFAULT_CHOOSE_ARGS);
2334 
2335 	mutex_lock(&map->crush_workspace_mutex);
2336 	r = crush_do_rule(map->crush, ruleno, x, result, result_max,
2337 			  weight, weight_max, map->crush_workspace,
2338 			  arg_map ? arg_map->args : NULL);
2339 	mutex_unlock(&map->crush_workspace_mutex);
2340 
2341 	return r;
2342 }
2343 
2344 static void remove_nonexistent_osds(struct ceph_osdmap *osdmap,
2345 				    struct ceph_pg_pool_info *pi,
2346 				    struct ceph_osds *set)
2347 {
2348 	int i;
2349 
2350 	if (ceph_can_shift_osds(pi)) {
2351 		int removed = 0;
2352 
2353 		/* shift left */
2354 		for (i = 0; i < set->size; i++) {
2355 			if (!ceph_osd_exists(osdmap, set->osds[i])) {
2356 				removed++;
2357 				continue;
2358 			}
2359 			if (removed)
2360 				set->osds[i - removed] = set->osds[i];
2361 		}
2362 		set->size -= removed;
2363 	} else {
2364 		/* set dne devices to NONE */
2365 		for (i = 0; i < set->size; i++) {
2366 			if (!ceph_osd_exists(osdmap, set->osds[i]))
2367 				set->osds[i] = CRUSH_ITEM_NONE;
2368 		}
2369 	}
2370 }
2371 
2372 /*
2373  * Calculate raw set (CRUSH output) for given PG and filter out
2374  * nonexistent OSDs.  ->primary is undefined for a raw set.
2375  *
2376  * Placement seed (CRUSH input) is returned through @ppps.
2377  */
2378 static void pg_to_raw_osds(struct ceph_osdmap *osdmap,
2379 			   struct ceph_pg_pool_info *pi,
2380 			   const struct ceph_pg *raw_pgid,
2381 			   struct ceph_osds *raw,
2382 			   u32 *ppps)
2383 {
2384 	u32 pps = raw_pg_to_pps(pi, raw_pgid);
2385 	int ruleno;
2386 	int len;
2387 
2388 	ceph_osds_init(raw);
2389 	if (ppps)
2390 		*ppps = pps;
2391 
2392 	ruleno = crush_find_rule(osdmap->crush, pi->crush_ruleset, pi->type,
2393 				 pi->size);
2394 	if (ruleno < 0) {
2395 		pr_err("no crush rule: pool %lld ruleset %d type %d size %d\n",
2396 		       pi->id, pi->crush_ruleset, pi->type, pi->size);
2397 		return;
2398 	}
2399 
2400 	if (pi->size > ARRAY_SIZE(raw->osds)) {
2401 		pr_err_ratelimited("pool %lld ruleset %d type %d too wide: size %d > %zu\n",
2402 		       pi->id, pi->crush_ruleset, pi->type, pi->size,
2403 		       ARRAY_SIZE(raw->osds));
2404 		return;
2405 	}
2406 
2407 	len = do_crush(osdmap, ruleno, pps, raw->osds, pi->size,
2408 		       osdmap->osd_weight, osdmap->max_osd, pi->id);
2409 	if (len < 0) {
2410 		pr_err("error %d from crush rule %d: pool %lld ruleset %d type %d size %d\n",
2411 		       len, ruleno, pi->id, pi->crush_ruleset, pi->type,
2412 		       pi->size);
2413 		return;
2414 	}
2415 
2416 	raw->size = len;
2417 	remove_nonexistent_osds(osdmap, pi, raw);
2418 }
2419 
2420 /* apply pg_upmap[_items] mappings */
2421 static void apply_upmap(struct ceph_osdmap *osdmap,
2422 			const struct ceph_pg *pgid,
2423 			struct ceph_osds *raw)
2424 {
2425 	struct ceph_pg_mapping *pg;
2426 	int i, j;
2427 
2428 	pg = lookup_pg_mapping(&osdmap->pg_upmap, pgid);
2429 	if (pg) {
2430 		/* make sure targets aren't marked out */
2431 		for (i = 0; i < pg->pg_upmap.len; i++) {
2432 			int osd = pg->pg_upmap.osds[i];
2433 
2434 			if (osd != CRUSH_ITEM_NONE &&
2435 			    osd < osdmap->max_osd &&
2436 			    osdmap->osd_weight[osd] == 0) {
2437 				/* reject/ignore explicit mapping */
2438 				return;
2439 			}
2440 		}
2441 		for (i = 0; i < pg->pg_upmap.len; i++)
2442 			raw->osds[i] = pg->pg_upmap.osds[i];
2443 		raw->size = pg->pg_upmap.len;
2444 		/* check and apply pg_upmap_items, if any */
2445 	}
2446 
2447 	pg = lookup_pg_mapping(&osdmap->pg_upmap_items, pgid);
2448 	if (pg) {
2449 		/*
2450 		 * Note: this approach does not allow a bidirectional swap,
2451 		 * e.g., [[1,2],[2,1]] applied to [0,1,2] -> [0,2,1].
2452 		 */
2453 		for (i = 0; i < pg->pg_upmap_items.len; i++) {
2454 			int from = pg->pg_upmap_items.from_to[i][0];
2455 			int to = pg->pg_upmap_items.from_to[i][1];
2456 			int pos = -1;
2457 			bool exists = false;
2458 
2459 			/* make sure replacement doesn't already appear */
2460 			for (j = 0; j < raw->size; j++) {
2461 				int osd = raw->osds[j];
2462 
2463 				if (osd == to) {
2464 					exists = true;
2465 					break;
2466 				}
2467 				/* ignore mapping if target is marked out */
2468 				if (osd == from && pos < 0 &&
2469 				    !(to != CRUSH_ITEM_NONE &&
2470 				      to < osdmap->max_osd &&
2471 				      osdmap->osd_weight[to] == 0)) {
2472 					pos = j;
2473 				}
2474 			}
2475 			if (!exists && pos >= 0)
2476 				raw->osds[pos] = to;
2477 		}
2478 	}
2479 }
2480 
2481 /*
2482  * Given raw set, calculate up set and up primary.  By definition of an
2483  * up set, the result won't contain nonexistent or down OSDs.
2484  *
2485  * This is done in-place - on return @set is the up set.  If it's
2486  * empty, ->primary will remain undefined.
2487  */
2488 static void raw_to_up_osds(struct ceph_osdmap *osdmap,
2489 			   struct ceph_pg_pool_info *pi,
2490 			   struct ceph_osds *set)
2491 {
2492 	int i;
2493 
2494 	/* ->primary is undefined for a raw set */
2495 	BUG_ON(set->primary != -1);
2496 
2497 	if (ceph_can_shift_osds(pi)) {
2498 		int removed = 0;
2499 
2500 		/* shift left */
2501 		for (i = 0; i < set->size; i++) {
2502 			if (ceph_osd_is_down(osdmap, set->osds[i])) {
2503 				removed++;
2504 				continue;
2505 			}
2506 			if (removed)
2507 				set->osds[i - removed] = set->osds[i];
2508 		}
2509 		set->size -= removed;
2510 		if (set->size > 0)
2511 			set->primary = set->osds[0];
2512 	} else {
2513 		/* set down/dne devices to NONE */
2514 		for (i = set->size - 1; i >= 0; i--) {
2515 			if (ceph_osd_is_down(osdmap, set->osds[i]))
2516 				set->osds[i] = CRUSH_ITEM_NONE;
2517 			else
2518 				set->primary = set->osds[i];
2519 		}
2520 	}
2521 }
2522 
2523 static void apply_primary_affinity(struct ceph_osdmap *osdmap,
2524 				   struct ceph_pg_pool_info *pi,
2525 				   u32 pps,
2526 				   struct ceph_osds *up)
2527 {
2528 	int i;
2529 	int pos = -1;
2530 
2531 	/*
2532 	 * Do we have any non-default primary_affinity values for these
2533 	 * osds?
2534 	 */
2535 	if (!osdmap->osd_primary_affinity)
2536 		return;
2537 
2538 	for (i = 0; i < up->size; i++) {
2539 		int osd = up->osds[i];
2540 
2541 		if (osd != CRUSH_ITEM_NONE &&
2542 		    osdmap->osd_primary_affinity[osd] !=
2543 					CEPH_OSD_DEFAULT_PRIMARY_AFFINITY) {
2544 			break;
2545 		}
2546 	}
2547 	if (i == up->size)
2548 		return;
2549 
2550 	/*
2551 	 * Pick the primary.  Feed both the seed (for the pg) and the
2552 	 * osd into the hash/rng so that a proportional fraction of an
2553 	 * osd's pgs get rejected as primary.
2554 	 */
2555 	for (i = 0; i < up->size; i++) {
2556 		int osd = up->osds[i];
2557 		u32 aff;
2558 
2559 		if (osd == CRUSH_ITEM_NONE)
2560 			continue;
2561 
2562 		aff = osdmap->osd_primary_affinity[osd];
2563 		if (aff < CEPH_OSD_MAX_PRIMARY_AFFINITY &&
2564 		    (crush_hash32_2(CRUSH_HASH_RJENKINS1,
2565 				    pps, osd) >> 16) >= aff) {
2566 			/*
2567 			 * We chose not to use this primary.  Note it
2568 			 * anyway as a fallback in case we don't pick
2569 			 * anyone else, but keep looking.
2570 			 */
2571 			if (pos < 0)
2572 				pos = i;
2573 		} else {
2574 			pos = i;
2575 			break;
2576 		}
2577 	}
2578 	if (pos < 0)
2579 		return;
2580 
2581 	up->primary = up->osds[pos];
2582 
2583 	if (ceph_can_shift_osds(pi) && pos > 0) {
2584 		/* move the new primary to the front */
2585 		for (i = pos; i > 0; i--)
2586 			up->osds[i] = up->osds[i - 1];
2587 		up->osds[0] = up->primary;
2588 	}
2589 }
2590 
2591 /*
2592  * Get pg_temp and primary_temp mappings for given PG.
2593  *
2594  * Note that a PG may have none, only pg_temp, only primary_temp or
2595  * both pg_temp and primary_temp mappings.  This means @temp isn't
2596  * always a valid OSD set on return: in the "only primary_temp" case,
2597  * @temp will have its ->primary >= 0 but ->size == 0.
2598  */
2599 static void get_temp_osds(struct ceph_osdmap *osdmap,
2600 			  struct ceph_pg_pool_info *pi,
2601 			  const struct ceph_pg *pgid,
2602 			  struct ceph_osds *temp)
2603 {
2604 	struct ceph_pg_mapping *pg;
2605 	int i;
2606 
2607 	ceph_osds_init(temp);
2608 
2609 	/* pg_temp? */
2610 	pg = lookup_pg_mapping(&osdmap->pg_temp, pgid);
2611 	if (pg) {
2612 		for (i = 0; i < pg->pg_temp.len; i++) {
2613 			if (ceph_osd_is_down(osdmap, pg->pg_temp.osds[i])) {
2614 				if (ceph_can_shift_osds(pi))
2615 					continue;
2616 
2617 				temp->osds[temp->size++] = CRUSH_ITEM_NONE;
2618 			} else {
2619 				temp->osds[temp->size++] = pg->pg_temp.osds[i];
2620 			}
2621 		}
2622 
2623 		/* apply pg_temp's primary */
2624 		for (i = 0; i < temp->size; i++) {
2625 			if (temp->osds[i] != CRUSH_ITEM_NONE) {
2626 				temp->primary = temp->osds[i];
2627 				break;
2628 			}
2629 		}
2630 	}
2631 
2632 	/* primary_temp? */
2633 	pg = lookup_pg_mapping(&osdmap->primary_temp, pgid);
2634 	if (pg)
2635 		temp->primary = pg->primary_temp.osd;
2636 }
2637 
2638 /*
2639  * Map a PG to its acting set as well as its up set.
2640  *
2641  * Acting set is used for data mapping purposes, while up set can be
2642  * recorded for detecting interval changes and deciding whether to
2643  * resend a request.
2644  */
2645 void ceph_pg_to_up_acting_osds(struct ceph_osdmap *osdmap,
2646 			       struct ceph_pg_pool_info *pi,
2647 			       const struct ceph_pg *raw_pgid,
2648 			       struct ceph_osds *up,
2649 			       struct ceph_osds *acting)
2650 {
2651 	struct ceph_pg pgid;
2652 	u32 pps;
2653 
2654 	WARN_ON(pi->id != raw_pgid->pool);
2655 	raw_pg_to_pg(pi, raw_pgid, &pgid);
2656 
2657 	pg_to_raw_osds(osdmap, pi, raw_pgid, up, &pps);
2658 	apply_upmap(osdmap, &pgid, up);
2659 	raw_to_up_osds(osdmap, pi, up);
2660 	apply_primary_affinity(osdmap, pi, pps, up);
2661 	get_temp_osds(osdmap, pi, &pgid, acting);
2662 	if (!acting->size) {
2663 		memcpy(acting->osds, up->osds, up->size * sizeof(up->osds[0]));
2664 		acting->size = up->size;
2665 		if (acting->primary == -1)
2666 			acting->primary = up->primary;
2667 	}
2668 	WARN_ON(!osds_valid(up) || !osds_valid(acting));
2669 }
2670 
2671 bool ceph_pg_to_primary_shard(struct ceph_osdmap *osdmap,
2672 			      struct ceph_pg_pool_info *pi,
2673 			      const struct ceph_pg *raw_pgid,
2674 			      struct ceph_spg *spgid)
2675 {
2676 	struct ceph_pg pgid;
2677 	struct ceph_osds up, acting;
2678 	int i;
2679 
2680 	WARN_ON(pi->id != raw_pgid->pool);
2681 	raw_pg_to_pg(pi, raw_pgid, &pgid);
2682 
2683 	if (ceph_can_shift_osds(pi)) {
2684 		spgid->pgid = pgid; /* struct */
2685 		spgid->shard = CEPH_SPG_NOSHARD;
2686 		return true;
2687 	}
2688 
2689 	ceph_pg_to_up_acting_osds(osdmap, pi, &pgid, &up, &acting);
2690 	for (i = 0; i < acting.size; i++) {
2691 		if (acting.osds[i] == acting.primary) {
2692 			spgid->pgid = pgid; /* struct */
2693 			spgid->shard = i;
2694 			return true;
2695 		}
2696 	}
2697 
2698 	return false;
2699 }
2700 
2701 /*
2702  * Return acting primary for given PG, or -1 if none.
2703  */
2704 int ceph_pg_to_acting_primary(struct ceph_osdmap *osdmap,
2705 			      const struct ceph_pg *raw_pgid)
2706 {
2707 	struct ceph_pg_pool_info *pi;
2708 	struct ceph_osds up, acting;
2709 
2710 	pi = ceph_pg_pool_by_id(osdmap, raw_pgid->pool);
2711 	if (!pi)
2712 		return -1;
2713 
2714 	ceph_pg_to_up_acting_osds(osdmap, pi, raw_pgid, &up, &acting);
2715 	return acting.primary;
2716 }
2717 EXPORT_SYMBOL(ceph_pg_to_acting_primary);
2718