xref: /openbmc/linux/net/ceph/mon_client.c (revision 8e9356c6)
1 #include <linux/ceph/ceph_debug.h>
2 
3 #include <linux/module.h>
4 #include <linux/types.h>
5 #include <linux/slab.h>
6 #include <linux/random.h>
7 #include <linux/sched.h>
8 
9 #include <linux/ceph/mon_client.h>
10 #include <linux/ceph/libceph.h>
11 #include <linux/ceph/debugfs.h>
12 #include <linux/ceph/decode.h>
13 #include <linux/ceph/auth.h>
14 
15 /*
16  * Interact with Ceph monitor cluster.  Handle requests for new map
17  * versions, and periodically resend as needed.  Also implement
18  * statfs() and umount().
19  *
20  * A small cluster of Ceph "monitors" are responsible for managing critical
21  * cluster configuration and state information.  An odd number (e.g., 3, 5)
22  * of cmon daemons use a modified version of the Paxos part-time parliament
23  * algorithm to manage the MDS map (mds cluster membership), OSD map, and
24  * list of clients who have mounted the file system.
25  *
26  * We maintain an open, active session with a monitor at all times in order to
27  * receive timely MDSMap updates.  We periodically send a keepalive byte on the
28  * TCP socket to ensure we detect a failure.  If the connection does break, we
29  * randomly hunt for a new monitor.  Once the connection is reestablished, we
30  * resend any outstanding requests.
31  */
32 
33 static const struct ceph_connection_operations mon_con_ops;
34 
35 static int __validate_auth(struct ceph_mon_client *monc);
36 
37 /*
38  * Decode a monmap blob (e.g., during mount).
39  */
40 struct ceph_monmap *ceph_monmap_decode(void *p, void *end)
41 {
42 	struct ceph_monmap *m = NULL;
43 	int i, err = -EINVAL;
44 	struct ceph_fsid fsid;
45 	u32 epoch, num_mon;
46 	u16 version;
47 	u32 len;
48 
49 	ceph_decode_32_safe(&p, end, len, bad);
50 	ceph_decode_need(&p, end, len, bad);
51 
52 	dout("monmap_decode %p %p len %d\n", p, end, (int)(end-p));
53 
54 	ceph_decode_16_safe(&p, end, version, bad);
55 
56 	ceph_decode_need(&p, end, sizeof(fsid) + 2*sizeof(u32), bad);
57 	ceph_decode_copy(&p, &fsid, sizeof(fsid));
58 	epoch = ceph_decode_32(&p);
59 
60 	num_mon = ceph_decode_32(&p);
61 	ceph_decode_need(&p, end, num_mon*sizeof(m->mon_inst[0]), bad);
62 
63 	if (num_mon >= CEPH_MAX_MON)
64 		goto bad;
65 	m = kmalloc(sizeof(*m) + sizeof(m->mon_inst[0])*num_mon, GFP_NOFS);
66 	if (m == NULL)
67 		return ERR_PTR(-ENOMEM);
68 	m->fsid = fsid;
69 	m->epoch = epoch;
70 	m->num_mon = num_mon;
71 	ceph_decode_copy(&p, m->mon_inst, num_mon*sizeof(m->mon_inst[0]));
72 	for (i = 0; i < num_mon; i++)
73 		ceph_decode_addr(&m->mon_inst[i].addr);
74 
75 	dout("monmap_decode epoch %d, num_mon %d\n", m->epoch,
76 	     m->num_mon);
77 	for (i = 0; i < m->num_mon; i++)
78 		dout("monmap_decode  mon%d is %s\n", i,
79 		     ceph_pr_addr(&m->mon_inst[i].addr.in_addr));
80 	return m;
81 
82 bad:
83 	dout("monmap_decode failed with %d\n", err);
84 	kfree(m);
85 	return ERR_PTR(err);
86 }
87 
88 /*
89  * return true if *addr is included in the monmap.
90  */
91 int ceph_monmap_contains(struct ceph_monmap *m, struct ceph_entity_addr *addr)
92 {
93 	int i;
94 
95 	for (i = 0; i < m->num_mon; i++)
96 		if (memcmp(addr, &m->mon_inst[i].addr, sizeof(*addr)) == 0)
97 			return 1;
98 	return 0;
99 }
100 
101 /*
102  * Send an auth request.
103  */
104 static void __send_prepared_auth_request(struct ceph_mon_client *monc, int len)
105 {
106 	monc->pending_auth = 1;
107 	monc->m_auth->front.iov_len = len;
108 	monc->m_auth->hdr.front_len = cpu_to_le32(len);
109 	ceph_msg_revoke(monc->m_auth);
110 	ceph_msg_get(monc->m_auth);  /* keep our ref */
111 	ceph_con_send(&monc->con, monc->m_auth);
112 }
113 
114 /*
115  * Close monitor session, if any.
116  */
117 static void __close_session(struct ceph_mon_client *monc)
118 {
119 	dout("__close_session closing mon%d\n", monc->cur_mon);
120 	ceph_msg_revoke(monc->m_auth);
121 	ceph_msg_revoke_incoming(monc->m_auth_reply);
122 	ceph_msg_revoke(monc->m_subscribe);
123 	ceph_msg_revoke_incoming(monc->m_subscribe_ack);
124 	ceph_con_close(&monc->con);
125 	monc->cur_mon = -1;
126 	monc->pending_auth = 0;
127 	ceph_auth_reset(monc->auth);
128 }
129 
130 /*
131  * Open a session with a (new) monitor.
132  */
133 static int __open_session(struct ceph_mon_client *monc)
134 {
135 	char r;
136 	int ret;
137 
138 	if (monc->cur_mon < 0) {
139 		get_random_bytes(&r, 1);
140 		monc->cur_mon = r % monc->monmap->num_mon;
141 		dout("open_session num=%d r=%d -> mon%d\n",
142 		     monc->monmap->num_mon, r, monc->cur_mon);
143 		monc->sub_sent = 0;
144 		monc->sub_renew_after = jiffies;  /* i.e., expired */
145 		monc->want_next_osdmap = !!monc->want_next_osdmap;
146 
147 		dout("open_session mon%d opening\n", monc->cur_mon);
148 		ceph_con_open(&monc->con,
149 			      CEPH_ENTITY_TYPE_MON, monc->cur_mon,
150 			      &monc->monmap->mon_inst[monc->cur_mon].addr);
151 
152 		/* initiatiate authentication handshake */
153 		ret = ceph_auth_build_hello(monc->auth,
154 					    monc->m_auth->front.iov_base,
155 					    monc->m_auth->front_alloc_len);
156 		__send_prepared_auth_request(monc, ret);
157 	} else {
158 		dout("open_session mon%d already open\n", monc->cur_mon);
159 	}
160 	return 0;
161 }
162 
163 static bool __sub_expired(struct ceph_mon_client *monc)
164 {
165 	return time_after_eq(jiffies, monc->sub_renew_after);
166 }
167 
168 /*
169  * Reschedule delayed work timer.
170  */
171 static void __schedule_delayed(struct ceph_mon_client *monc)
172 {
173 	unsigned int delay;
174 
175 	if (monc->cur_mon < 0 || __sub_expired(monc))
176 		delay = 10 * HZ;
177 	else
178 		delay = 20 * HZ;
179 	dout("__schedule_delayed after %u\n", delay);
180 	schedule_delayed_work(&monc->delayed_work, delay);
181 }
182 
183 /*
184  * Send subscribe request for mdsmap and/or osdmap.
185  */
186 static void __send_subscribe(struct ceph_mon_client *monc)
187 {
188 	dout("__send_subscribe sub_sent=%u exp=%u want_osd=%d\n",
189 	     (unsigned int)monc->sub_sent, __sub_expired(monc),
190 	     monc->want_next_osdmap);
191 	if ((__sub_expired(monc) && !monc->sub_sent) ||
192 	    monc->want_next_osdmap == 1) {
193 		struct ceph_msg *msg = monc->m_subscribe;
194 		struct ceph_mon_subscribe_item *i;
195 		void *p, *end;
196 		int num;
197 
198 		p = msg->front.iov_base;
199 		end = p + msg->front_alloc_len;
200 
201 		num = 1 + !!monc->want_next_osdmap + !!monc->want_mdsmap;
202 		ceph_encode_32(&p, num);
203 
204 		if (monc->want_next_osdmap) {
205 			dout("__send_subscribe to 'osdmap' %u\n",
206 			     (unsigned int)monc->have_osdmap);
207 			ceph_encode_string(&p, end, "osdmap", 6);
208 			i = p;
209 			i->have = cpu_to_le64(monc->have_osdmap);
210 			i->onetime = 1;
211 			p += sizeof(*i);
212 			monc->want_next_osdmap = 2;  /* requested */
213 		}
214 		if (monc->want_mdsmap) {
215 			dout("__send_subscribe to 'mdsmap' %u+\n",
216 			     (unsigned int)monc->have_mdsmap);
217 			ceph_encode_string(&p, end, "mdsmap", 6);
218 			i = p;
219 			i->have = cpu_to_le64(monc->have_mdsmap);
220 			i->onetime = 0;
221 			p += sizeof(*i);
222 		}
223 		ceph_encode_string(&p, end, "monmap", 6);
224 		i = p;
225 		i->have = 0;
226 		i->onetime = 0;
227 		p += sizeof(*i);
228 
229 		msg->front.iov_len = p - msg->front.iov_base;
230 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
231 		ceph_msg_revoke(msg);
232 		ceph_con_send(&monc->con, ceph_msg_get(msg));
233 
234 		monc->sub_sent = jiffies | 1;  /* never 0 */
235 	}
236 }
237 
238 static void handle_subscribe_ack(struct ceph_mon_client *monc,
239 				 struct ceph_msg *msg)
240 {
241 	unsigned int seconds;
242 	struct ceph_mon_subscribe_ack *h = msg->front.iov_base;
243 
244 	if (msg->front.iov_len < sizeof(*h))
245 		goto bad;
246 	seconds = le32_to_cpu(h->duration);
247 
248 	mutex_lock(&monc->mutex);
249 	if (monc->hunting) {
250 		pr_info("mon%d %s session established\n",
251 			monc->cur_mon,
252 			ceph_pr_addr(&monc->con.peer_addr.in_addr));
253 		monc->hunting = false;
254 	}
255 	dout("handle_subscribe_ack after %d seconds\n", seconds);
256 	monc->sub_renew_after = monc->sub_sent + (seconds >> 1)*HZ - 1;
257 	monc->sub_sent = 0;
258 	mutex_unlock(&monc->mutex);
259 	return;
260 bad:
261 	pr_err("got corrupt subscribe-ack msg\n");
262 	ceph_msg_dump(msg);
263 }
264 
265 /*
266  * Keep track of which maps we have
267  */
268 int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 got)
269 {
270 	mutex_lock(&monc->mutex);
271 	monc->have_mdsmap = got;
272 	mutex_unlock(&monc->mutex);
273 	return 0;
274 }
275 EXPORT_SYMBOL(ceph_monc_got_mdsmap);
276 
277 int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 got)
278 {
279 	mutex_lock(&monc->mutex);
280 	monc->have_osdmap = got;
281 	monc->want_next_osdmap = 0;
282 	mutex_unlock(&monc->mutex);
283 	return 0;
284 }
285 
286 /*
287  * Register interest in the next osdmap
288  */
289 void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc)
290 {
291 	dout("request_next_osdmap have %u\n", monc->have_osdmap);
292 	mutex_lock(&monc->mutex);
293 	if (!monc->want_next_osdmap)
294 		monc->want_next_osdmap = 1;
295 	if (monc->want_next_osdmap < 2)
296 		__send_subscribe(monc);
297 	mutex_unlock(&monc->mutex);
298 }
299 
300 /*
301  *
302  */
303 int ceph_monc_open_session(struct ceph_mon_client *monc)
304 {
305 	mutex_lock(&monc->mutex);
306 	__open_session(monc);
307 	__schedule_delayed(monc);
308 	mutex_unlock(&monc->mutex);
309 	return 0;
310 }
311 EXPORT_SYMBOL(ceph_monc_open_session);
312 
313 /*
314  * We require the fsid and global_id in order to initialize our
315  * debugfs dir.
316  */
317 static bool have_debugfs_info(struct ceph_mon_client *monc)
318 {
319 	dout("have_debugfs_info fsid %d globalid %lld\n",
320 	     (int)monc->client->have_fsid, monc->auth->global_id);
321 	return monc->client->have_fsid && monc->auth->global_id > 0;
322 }
323 
324 /*
325  * The monitor responds with mount ack indicate mount success.  The
326  * included client ticket allows the client to talk to MDSs and OSDs.
327  */
328 static void ceph_monc_handle_map(struct ceph_mon_client *monc,
329 				 struct ceph_msg *msg)
330 {
331 	struct ceph_client *client = monc->client;
332 	struct ceph_monmap *monmap = NULL, *old = monc->monmap;
333 	void *p, *end;
334 	int had_debugfs_info, init_debugfs = 0;
335 
336 	mutex_lock(&monc->mutex);
337 
338 	had_debugfs_info = have_debugfs_info(monc);
339 
340 	dout("handle_monmap\n");
341 	p = msg->front.iov_base;
342 	end = p + msg->front.iov_len;
343 
344 	monmap = ceph_monmap_decode(p, end);
345 	if (IS_ERR(monmap)) {
346 		pr_err("problem decoding monmap, %d\n",
347 		       (int)PTR_ERR(monmap));
348 		goto out;
349 	}
350 
351 	if (ceph_check_fsid(monc->client, &monmap->fsid) < 0) {
352 		kfree(monmap);
353 		goto out;
354 	}
355 
356 	client->monc.monmap = monmap;
357 	kfree(old);
358 
359 	if (!client->have_fsid) {
360 		client->have_fsid = true;
361 		if (!had_debugfs_info && have_debugfs_info(monc)) {
362 			pr_info("client%lld fsid %pU\n",
363 				ceph_client_id(monc->client),
364 				&monc->client->fsid);
365 			init_debugfs = 1;
366 		}
367 		mutex_unlock(&monc->mutex);
368 
369 		if (init_debugfs) {
370 			/*
371 			 * do debugfs initialization without mutex to avoid
372 			 * creating a locking dependency
373 			 */
374 			ceph_debugfs_client_init(monc->client);
375 		}
376 
377 		goto out_unlocked;
378 	}
379 out:
380 	mutex_unlock(&monc->mutex);
381 out_unlocked:
382 	wake_up_all(&client->auth_wq);
383 }
384 
385 /*
386  * generic requests (e.g., statfs, poolop)
387  */
388 static struct ceph_mon_generic_request *__lookup_generic_req(
389 	struct ceph_mon_client *monc, u64 tid)
390 {
391 	struct ceph_mon_generic_request *req;
392 	struct rb_node *n = monc->generic_request_tree.rb_node;
393 
394 	while (n) {
395 		req = rb_entry(n, struct ceph_mon_generic_request, node);
396 		if (tid < req->tid)
397 			n = n->rb_left;
398 		else if (tid > req->tid)
399 			n = n->rb_right;
400 		else
401 			return req;
402 	}
403 	return NULL;
404 }
405 
406 static void __insert_generic_request(struct ceph_mon_client *monc,
407 			    struct ceph_mon_generic_request *new)
408 {
409 	struct rb_node **p = &monc->generic_request_tree.rb_node;
410 	struct rb_node *parent = NULL;
411 	struct ceph_mon_generic_request *req = NULL;
412 
413 	while (*p) {
414 		parent = *p;
415 		req = rb_entry(parent, struct ceph_mon_generic_request, node);
416 		if (new->tid < req->tid)
417 			p = &(*p)->rb_left;
418 		else if (new->tid > req->tid)
419 			p = &(*p)->rb_right;
420 		else
421 			BUG();
422 	}
423 
424 	rb_link_node(&new->node, parent, p);
425 	rb_insert_color(&new->node, &monc->generic_request_tree);
426 }
427 
428 static void release_generic_request(struct kref *kref)
429 {
430 	struct ceph_mon_generic_request *req =
431 		container_of(kref, struct ceph_mon_generic_request, kref);
432 
433 	if (req->reply)
434 		ceph_msg_put(req->reply);
435 	if (req->request)
436 		ceph_msg_put(req->request);
437 
438 	kfree(req);
439 }
440 
441 static void put_generic_request(struct ceph_mon_generic_request *req)
442 {
443 	kref_put(&req->kref, release_generic_request);
444 }
445 
446 static void get_generic_request(struct ceph_mon_generic_request *req)
447 {
448 	kref_get(&req->kref);
449 }
450 
451 static struct ceph_msg *get_generic_reply(struct ceph_connection *con,
452 					 struct ceph_msg_header *hdr,
453 					 int *skip)
454 {
455 	struct ceph_mon_client *monc = con->private;
456 	struct ceph_mon_generic_request *req;
457 	u64 tid = le64_to_cpu(hdr->tid);
458 	struct ceph_msg *m;
459 
460 	mutex_lock(&monc->mutex);
461 	req = __lookup_generic_req(monc, tid);
462 	if (!req) {
463 		dout("get_generic_reply %lld dne\n", tid);
464 		*skip = 1;
465 		m = NULL;
466 	} else {
467 		dout("get_generic_reply %lld got %p\n", tid, req->reply);
468 		*skip = 0;
469 		m = ceph_msg_get(req->reply);
470 		/*
471 		 * we don't need to track the connection reading into
472 		 * this reply because we only have one open connection
473 		 * at a time, ever.
474 		 */
475 	}
476 	mutex_unlock(&monc->mutex);
477 	return m;
478 }
479 
480 static int do_generic_request(struct ceph_mon_client *monc,
481 			      struct ceph_mon_generic_request *req)
482 {
483 	int err;
484 
485 	/* register request */
486 	mutex_lock(&monc->mutex);
487 	req->tid = ++monc->last_tid;
488 	req->request->hdr.tid = cpu_to_le64(req->tid);
489 	__insert_generic_request(monc, req);
490 	monc->num_generic_requests++;
491 	ceph_con_send(&monc->con, ceph_msg_get(req->request));
492 	mutex_unlock(&monc->mutex);
493 
494 	err = wait_for_completion_interruptible(&req->completion);
495 
496 	mutex_lock(&monc->mutex);
497 	rb_erase(&req->node, &monc->generic_request_tree);
498 	monc->num_generic_requests--;
499 	mutex_unlock(&monc->mutex);
500 
501 	if (!err)
502 		err = req->result;
503 	return err;
504 }
505 
506 /*
507  * statfs
508  */
509 static void handle_statfs_reply(struct ceph_mon_client *monc,
510 				struct ceph_msg *msg)
511 {
512 	struct ceph_mon_generic_request *req;
513 	struct ceph_mon_statfs_reply *reply = msg->front.iov_base;
514 	u64 tid = le64_to_cpu(msg->hdr.tid);
515 
516 	if (msg->front.iov_len != sizeof(*reply))
517 		goto bad;
518 	dout("handle_statfs_reply %p tid %llu\n", msg, tid);
519 
520 	mutex_lock(&monc->mutex);
521 	req = __lookup_generic_req(monc, tid);
522 	if (req) {
523 		*(struct ceph_statfs *)req->buf = reply->st;
524 		req->result = 0;
525 		get_generic_request(req);
526 	}
527 	mutex_unlock(&monc->mutex);
528 	if (req) {
529 		complete_all(&req->completion);
530 		put_generic_request(req);
531 	}
532 	return;
533 
534 bad:
535 	pr_err("corrupt generic reply, tid %llu\n", tid);
536 	ceph_msg_dump(msg);
537 }
538 
539 /*
540  * Do a synchronous statfs().
541  */
542 int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf)
543 {
544 	struct ceph_mon_generic_request *req;
545 	struct ceph_mon_statfs *h;
546 	int err;
547 
548 	req = kzalloc(sizeof(*req), GFP_NOFS);
549 	if (!req)
550 		return -ENOMEM;
551 
552 	kref_init(&req->kref);
553 	req->buf = buf;
554 	req->buf_len = sizeof(*buf);
555 	init_completion(&req->completion);
556 
557 	err = -ENOMEM;
558 	req->request = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), GFP_NOFS,
559 				    true);
560 	if (!req->request)
561 		goto out;
562 	req->reply = ceph_msg_new(CEPH_MSG_STATFS_REPLY, 1024, GFP_NOFS,
563 				  true);
564 	if (!req->reply)
565 		goto out;
566 
567 	/* fill out request */
568 	h = req->request->front.iov_base;
569 	h->monhdr.have_version = 0;
570 	h->monhdr.session_mon = cpu_to_le16(-1);
571 	h->monhdr.session_mon_tid = 0;
572 	h->fsid = monc->monmap->fsid;
573 
574 	err = do_generic_request(monc, req);
575 
576 out:
577 	kref_put(&req->kref, release_generic_request);
578 	return err;
579 }
580 EXPORT_SYMBOL(ceph_monc_do_statfs);
581 
582 /*
583  * pool ops
584  */
585 static int get_poolop_reply_buf(const char *src, size_t src_len,
586 				char *dst, size_t dst_len)
587 {
588 	u32 buf_len;
589 
590 	if (src_len != sizeof(u32) + dst_len)
591 		return -EINVAL;
592 
593 	buf_len = le32_to_cpu(*(u32 *)src);
594 	if (buf_len != dst_len)
595 		return -EINVAL;
596 
597 	memcpy(dst, src + sizeof(u32), dst_len);
598 	return 0;
599 }
600 
601 static void handle_poolop_reply(struct ceph_mon_client *monc,
602 				struct ceph_msg *msg)
603 {
604 	struct ceph_mon_generic_request *req;
605 	struct ceph_mon_poolop_reply *reply = msg->front.iov_base;
606 	u64 tid = le64_to_cpu(msg->hdr.tid);
607 
608 	if (msg->front.iov_len < sizeof(*reply))
609 		goto bad;
610 	dout("handle_poolop_reply %p tid %llu\n", msg, tid);
611 
612 	mutex_lock(&monc->mutex);
613 	req = __lookup_generic_req(monc, tid);
614 	if (req) {
615 		if (req->buf_len &&
616 		    get_poolop_reply_buf(msg->front.iov_base + sizeof(*reply),
617 				     msg->front.iov_len - sizeof(*reply),
618 				     req->buf, req->buf_len) < 0) {
619 			mutex_unlock(&monc->mutex);
620 			goto bad;
621 		}
622 		req->result = le32_to_cpu(reply->reply_code);
623 		get_generic_request(req);
624 	}
625 	mutex_unlock(&monc->mutex);
626 	if (req) {
627 		complete(&req->completion);
628 		put_generic_request(req);
629 	}
630 	return;
631 
632 bad:
633 	pr_err("corrupt generic reply, tid %llu\n", tid);
634 	ceph_msg_dump(msg);
635 }
636 
637 /*
638  * Do a synchronous pool op.
639  */
640 static int do_poolop(struct ceph_mon_client *monc, u32 op,
641 			u32 pool, u64 snapid,
642 			char *buf, int len)
643 {
644 	struct ceph_mon_generic_request *req;
645 	struct ceph_mon_poolop *h;
646 	int err;
647 
648 	req = kzalloc(sizeof(*req), GFP_NOFS);
649 	if (!req)
650 		return -ENOMEM;
651 
652 	kref_init(&req->kref);
653 	req->buf = buf;
654 	req->buf_len = len;
655 	init_completion(&req->completion);
656 
657 	err = -ENOMEM;
658 	req->request = ceph_msg_new(CEPH_MSG_POOLOP, sizeof(*h), GFP_NOFS,
659 				    true);
660 	if (!req->request)
661 		goto out;
662 	req->reply = ceph_msg_new(CEPH_MSG_POOLOP_REPLY, 1024, GFP_NOFS,
663 				  true);
664 	if (!req->reply)
665 		goto out;
666 
667 	/* fill out request */
668 	req->request->hdr.version = cpu_to_le16(2);
669 	h = req->request->front.iov_base;
670 	h->monhdr.have_version = 0;
671 	h->monhdr.session_mon = cpu_to_le16(-1);
672 	h->monhdr.session_mon_tid = 0;
673 	h->fsid = monc->monmap->fsid;
674 	h->pool = cpu_to_le32(pool);
675 	h->op = cpu_to_le32(op);
676 	h->auid = 0;
677 	h->snapid = cpu_to_le64(snapid);
678 	h->name_len = 0;
679 
680 	err = do_generic_request(monc, req);
681 
682 out:
683 	kref_put(&req->kref, release_generic_request);
684 	return err;
685 }
686 
687 int ceph_monc_create_snapid(struct ceph_mon_client *monc,
688 			    u32 pool, u64 *snapid)
689 {
690 	return do_poolop(monc,  POOL_OP_CREATE_UNMANAGED_SNAP,
691 				   pool, 0, (char *)snapid, sizeof(*snapid));
692 
693 }
694 EXPORT_SYMBOL(ceph_monc_create_snapid);
695 
696 int ceph_monc_delete_snapid(struct ceph_mon_client *monc,
697 			    u32 pool, u64 snapid)
698 {
699 	return do_poolop(monc,  POOL_OP_CREATE_UNMANAGED_SNAP,
700 				   pool, snapid, NULL, 0);
701 
702 }
703 
704 /*
705  * Resend pending generic requests.
706  */
707 static void __resend_generic_request(struct ceph_mon_client *monc)
708 {
709 	struct ceph_mon_generic_request *req;
710 	struct rb_node *p;
711 
712 	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
713 		req = rb_entry(p, struct ceph_mon_generic_request, node);
714 		ceph_msg_revoke(req->request);
715 		ceph_msg_revoke_incoming(req->reply);
716 		ceph_con_send(&monc->con, ceph_msg_get(req->request));
717 	}
718 }
719 
720 /*
721  * Delayed work.  If we haven't mounted yet, retry.  Otherwise,
722  * renew/retry subscription as needed (in case it is timing out, or we
723  * got an ENOMEM).  And keep the monitor connection alive.
724  */
725 static void delayed_work(struct work_struct *work)
726 {
727 	struct ceph_mon_client *monc =
728 		container_of(work, struct ceph_mon_client, delayed_work.work);
729 
730 	dout("monc delayed_work\n");
731 	mutex_lock(&monc->mutex);
732 	if (monc->hunting) {
733 		__close_session(monc);
734 		__open_session(monc);  /* continue hunting */
735 	} else {
736 		ceph_con_keepalive(&monc->con);
737 
738 		__validate_auth(monc);
739 
740 		if (ceph_auth_is_authenticated(monc->auth))
741 			__send_subscribe(monc);
742 	}
743 	__schedule_delayed(monc);
744 	mutex_unlock(&monc->mutex);
745 }
746 
747 /*
748  * On startup, we build a temporary monmap populated with the IPs
749  * provided by mount(2).
750  */
751 static int build_initial_monmap(struct ceph_mon_client *monc)
752 {
753 	struct ceph_options *opt = monc->client->options;
754 	struct ceph_entity_addr *mon_addr = opt->mon_addr;
755 	int num_mon = opt->num_mon;
756 	int i;
757 
758 	/* build initial monmap */
759 	monc->monmap = kzalloc(sizeof(*monc->monmap) +
760 			       num_mon*sizeof(monc->monmap->mon_inst[0]),
761 			       GFP_KERNEL);
762 	if (!monc->monmap)
763 		return -ENOMEM;
764 	for (i = 0; i < num_mon; i++) {
765 		monc->monmap->mon_inst[i].addr = mon_addr[i];
766 		monc->monmap->mon_inst[i].addr.nonce = 0;
767 		monc->monmap->mon_inst[i].name.type =
768 			CEPH_ENTITY_TYPE_MON;
769 		monc->monmap->mon_inst[i].name.num = cpu_to_le64(i);
770 	}
771 	monc->monmap->num_mon = num_mon;
772 	return 0;
773 }
774 
775 int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl)
776 {
777 	int err = 0;
778 
779 	dout("init\n");
780 	memset(monc, 0, sizeof(*monc));
781 	monc->client = cl;
782 	monc->monmap = NULL;
783 	mutex_init(&monc->mutex);
784 
785 	err = build_initial_monmap(monc);
786 	if (err)
787 		goto out;
788 
789 	/* connection */
790 	/* authentication */
791 	monc->auth = ceph_auth_init(cl->options->name,
792 				    cl->options->key);
793 	if (IS_ERR(monc->auth)) {
794 		err = PTR_ERR(monc->auth);
795 		goto out_monmap;
796 	}
797 	monc->auth->want_keys =
798 		CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON |
799 		CEPH_ENTITY_TYPE_OSD | CEPH_ENTITY_TYPE_MDS;
800 
801 	/* msgs */
802 	err = -ENOMEM;
803 	monc->m_subscribe_ack = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE_ACK,
804 				     sizeof(struct ceph_mon_subscribe_ack),
805 				     GFP_NOFS, true);
806 	if (!monc->m_subscribe_ack)
807 		goto out_auth;
808 
809 	monc->m_subscribe = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, GFP_NOFS,
810 					 true);
811 	if (!monc->m_subscribe)
812 		goto out_subscribe_ack;
813 
814 	monc->m_auth_reply = ceph_msg_new(CEPH_MSG_AUTH_REPLY, 4096, GFP_NOFS,
815 					  true);
816 	if (!monc->m_auth_reply)
817 		goto out_subscribe;
818 
819 	monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, GFP_NOFS, true);
820 	monc->pending_auth = 0;
821 	if (!monc->m_auth)
822 		goto out_auth_reply;
823 
824 	ceph_con_init(&monc->con, monc, &mon_con_ops,
825 		      &monc->client->msgr);
826 
827 	monc->cur_mon = -1;
828 	monc->hunting = true;
829 	monc->sub_renew_after = jiffies;
830 	monc->sub_sent = 0;
831 
832 	INIT_DELAYED_WORK(&monc->delayed_work, delayed_work);
833 	monc->generic_request_tree = RB_ROOT;
834 	monc->num_generic_requests = 0;
835 	monc->last_tid = 0;
836 
837 	monc->have_mdsmap = 0;
838 	monc->have_osdmap = 0;
839 	monc->want_next_osdmap = 1;
840 	return 0;
841 
842 out_auth_reply:
843 	ceph_msg_put(monc->m_auth_reply);
844 out_subscribe:
845 	ceph_msg_put(monc->m_subscribe);
846 out_subscribe_ack:
847 	ceph_msg_put(monc->m_subscribe_ack);
848 out_auth:
849 	ceph_auth_destroy(monc->auth);
850 out_monmap:
851 	kfree(monc->monmap);
852 out:
853 	return err;
854 }
855 EXPORT_SYMBOL(ceph_monc_init);
856 
857 void ceph_monc_stop(struct ceph_mon_client *monc)
858 {
859 	dout("stop\n");
860 	cancel_delayed_work_sync(&monc->delayed_work);
861 
862 	mutex_lock(&monc->mutex);
863 	__close_session(monc);
864 
865 	mutex_unlock(&monc->mutex);
866 
867 	/*
868 	 * flush msgr queue before we destroy ourselves to ensure that:
869 	 *  - any work that references our embedded con is finished.
870 	 *  - any osd_client or other work that may reference an authorizer
871 	 *    finishes before we shut down the auth subsystem.
872 	 */
873 	ceph_msgr_flush();
874 
875 	ceph_auth_destroy(monc->auth);
876 
877 	ceph_msg_put(monc->m_auth);
878 	ceph_msg_put(monc->m_auth_reply);
879 	ceph_msg_put(monc->m_subscribe);
880 	ceph_msg_put(monc->m_subscribe_ack);
881 
882 	kfree(monc->monmap);
883 }
884 EXPORT_SYMBOL(ceph_monc_stop);
885 
886 static void handle_auth_reply(struct ceph_mon_client *monc,
887 			      struct ceph_msg *msg)
888 {
889 	int ret;
890 	int was_auth = 0;
891 	int had_debugfs_info, init_debugfs = 0;
892 
893 	mutex_lock(&monc->mutex);
894 	had_debugfs_info = have_debugfs_info(monc);
895 	was_auth = ceph_auth_is_authenticated(monc->auth);
896 	monc->pending_auth = 0;
897 	ret = ceph_handle_auth_reply(monc->auth, msg->front.iov_base,
898 				     msg->front.iov_len,
899 				     monc->m_auth->front.iov_base,
900 				     monc->m_auth->front_alloc_len);
901 	if (ret < 0) {
902 		monc->client->auth_err = ret;
903 		wake_up_all(&monc->client->auth_wq);
904 	} else if (ret > 0) {
905 		__send_prepared_auth_request(monc, ret);
906 	} else if (!was_auth && ceph_auth_is_authenticated(monc->auth)) {
907 		dout("authenticated, starting session\n");
908 
909 		monc->client->msgr.inst.name.type = CEPH_ENTITY_TYPE_CLIENT;
910 		monc->client->msgr.inst.name.num =
911 					cpu_to_le64(monc->auth->global_id);
912 
913 		__send_subscribe(monc);
914 		__resend_generic_request(monc);
915 	}
916 
917 	if (!had_debugfs_info && have_debugfs_info(monc)) {
918 		pr_info("client%lld fsid %pU\n",
919 			ceph_client_id(monc->client),
920 			&monc->client->fsid);
921 		init_debugfs = 1;
922 	}
923 	mutex_unlock(&monc->mutex);
924 
925 	if (init_debugfs) {
926 		/*
927 		 * do debugfs initialization without mutex to avoid
928 		 * creating a locking dependency
929 		 */
930 		ceph_debugfs_client_init(monc->client);
931 	}
932 }
933 
934 static int __validate_auth(struct ceph_mon_client *monc)
935 {
936 	int ret;
937 
938 	if (monc->pending_auth)
939 		return 0;
940 
941 	ret = ceph_build_auth(monc->auth, monc->m_auth->front.iov_base,
942 			      monc->m_auth->front_alloc_len);
943 	if (ret <= 0)
944 		return ret; /* either an error, or no need to authenticate */
945 	__send_prepared_auth_request(monc, ret);
946 	return 0;
947 }
948 
949 int ceph_monc_validate_auth(struct ceph_mon_client *monc)
950 {
951 	int ret;
952 
953 	mutex_lock(&monc->mutex);
954 	ret = __validate_auth(monc);
955 	mutex_unlock(&monc->mutex);
956 	return ret;
957 }
958 EXPORT_SYMBOL(ceph_monc_validate_auth);
959 
960 /*
961  * handle incoming message
962  */
963 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
964 {
965 	struct ceph_mon_client *monc = con->private;
966 	int type = le16_to_cpu(msg->hdr.type);
967 
968 	if (!monc)
969 		return;
970 
971 	switch (type) {
972 	case CEPH_MSG_AUTH_REPLY:
973 		handle_auth_reply(monc, msg);
974 		break;
975 
976 	case CEPH_MSG_MON_SUBSCRIBE_ACK:
977 		handle_subscribe_ack(monc, msg);
978 		break;
979 
980 	case CEPH_MSG_STATFS_REPLY:
981 		handle_statfs_reply(monc, msg);
982 		break;
983 
984 	case CEPH_MSG_POOLOP_REPLY:
985 		handle_poolop_reply(monc, msg);
986 		break;
987 
988 	case CEPH_MSG_MON_MAP:
989 		ceph_monc_handle_map(monc, msg);
990 		break;
991 
992 	case CEPH_MSG_OSD_MAP:
993 		ceph_osdc_handle_map(&monc->client->osdc, msg);
994 		break;
995 
996 	default:
997 		/* can the chained handler handle it? */
998 		if (monc->client->extra_mon_dispatch &&
999 		    monc->client->extra_mon_dispatch(monc->client, msg) == 0)
1000 			break;
1001 
1002 		pr_err("received unknown message type %d %s\n", type,
1003 		       ceph_msg_type_name(type));
1004 	}
1005 	ceph_msg_put(msg);
1006 }
1007 
1008 /*
1009  * Allocate memory for incoming message
1010  */
1011 static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con,
1012 				      struct ceph_msg_header *hdr,
1013 				      int *skip)
1014 {
1015 	struct ceph_mon_client *monc = con->private;
1016 	int type = le16_to_cpu(hdr->type);
1017 	int front_len = le32_to_cpu(hdr->front_len);
1018 	struct ceph_msg *m = NULL;
1019 
1020 	*skip = 0;
1021 
1022 	switch (type) {
1023 	case CEPH_MSG_MON_SUBSCRIBE_ACK:
1024 		m = ceph_msg_get(monc->m_subscribe_ack);
1025 		break;
1026 	case CEPH_MSG_POOLOP_REPLY:
1027 	case CEPH_MSG_STATFS_REPLY:
1028 		return get_generic_reply(con, hdr, skip);
1029 	case CEPH_MSG_AUTH_REPLY:
1030 		m = ceph_msg_get(monc->m_auth_reply);
1031 		break;
1032 	case CEPH_MSG_MON_MAP:
1033 	case CEPH_MSG_MDS_MAP:
1034 	case CEPH_MSG_OSD_MAP:
1035 		m = ceph_msg_new(type, front_len, GFP_NOFS, false);
1036 		if (!m)
1037 			return NULL;	/* ENOMEM--return skip == 0 */
1038 		break;
1039 	}
1040 
1041 	if (!m) {
1042 		pr_info("alloc_msg unknown type %d\n", type);
1043 		*skip = 1;
1044 	}
1045 	return m;
1046 }
1047 
1048 /*
1049  * If the monitor connection resets, pick a new monitor and resubmit
1050  * any pending requests.
1051  */
1052 static void mon_fault(struct ceph_connection *con)
1053 {
1054 	struct ceph_mon_client *monc = con->private;
1055 
1056 	if (!monc)
1057 		return;
1058 
1059 	dout("mon_fault\n");
1060 	mutex_lock(&monc->mutex);
1061 	if (!con->private)
1062 		goto out;
1063 
1064 	if (!monc->hunting)
1065 		pr_info("mon%d %s session lost, "
1066 			"hunting for new mon\n", monc->cur_mon,
1067 			ceph_pr_addr(&monc->con.peer_addr.in_addr));
1068 
1069 	__close_session(monc);
1070 	if (!monc->hunting) {
1071 		/* start hunting */
1072 		monc->hunting = true;
1073 		__open_session(monc);
1074 	} else {
1075 		/* already hunting, let's wait a bit */
1076 		__schedule_delayed(monc);
1077 	}
1078 out:
1079 	mutex_unlock(&monc->mutex);
1080 }
1081 
1082 /*
1083  * We can ignore refcounting on the connection struct, as all references
1084  * will come from the messenger workqueue, which is drained prior to
1085  * mon_client destruction.
1086  */
1087 static struct ceph_connection *con_get(struct ceph_connection *con)
1088 {
1089 	return con;
1090 }
1091 
1092 static void con_put(struct ceph_connection *con)
1093 {
1094 }
1095 
1096 static const struct ceph_connection_operations mon_con_ops = {
1097 	.get = con_get,
1098 	.put = con_put,
1099 	.dispatch = dispatch,
1100 	.fault = mon_fault,
1101 	.alloc_msg = mon_alloc_msg,
1102 };
1103