xref: /openbmc/linux/net/ceph/mon_client.c (revision a8da474e)
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 		/* send an initial keepalive to ensure our timestamp is
153 		 * valid by the time we are in an OPENED state */
154 		ceph_con_keepalive(&monc->con);
155 
156 		/* initiatiate authentication handshake */
157 		ret = ceph_auth_build_hello(monc->auth,
158 					    monc->m_auth->front.iov_base,
159 					    monc->m_auth->front_alloc_len);
160 		__send_prepared_auth_request(monc, ret);
161 	} else {
162 		dout("open_session mon%d already open\n", monc->cur_mon);
163 	}
164 	return 0;
165 }
166 
167 static bool __sub_expired(struct ceph_mon_client *monc)
168 {
169 	return time_after_eq(jiffies, monc->sub_renew_after);
170 }
171 
172 /*
173  * Reschedule delayed work timer.
174  */
175 static void __schedule_delayed(struct ceph_mon_client *monc)
176 {
177 	struct ceph_options *opt = monc->client->options;
178 	unsigned long delay;
179 
180 	if (monc->cur_mon < 0 || __sub_expired(monc)) {
181 		delay = 10 * HZ;
182 	} else {
183 		delay = 20 * HZ;
184 		if (opt->monc_ping_timeout > 0)
185 			delay = min(delay, opt->monc_ping_timeout / 3);
186 	}
187 	dout("__schedule_delayed after %lu\n", delay);
188 	schedule_delayed_work(&monc->delayed_work,
189 			      round_jiffies_relative(delay));
190 }
191 
192 /*
193  * Send subscribe request for mdsmap and/or osdmap.
194  */
195 static void __send_subscribe(struct ceph_mon_client *monc)
196 {
197 	dout("__send_subscribe sub_sent=%u exp=%u want_osd=%d\n",
198 	     (unsigned int)monc->sub_sent, __sub_expired(monc),
199 	     monc->want_next_osdmap);
200 	if ((__sub_expired(monc) && !monc->sub_sent) ||
201 	    monc->want_next_osdmap == 1) {
202 		struct ceph_msg *msg = monc->m_subscribe;
203 		struct ceph_mon_subscribe_item *i;
204 		void *p, *end;
205 		int num;
206 
207 		p = msg->front.iov_base;
208 		end = p + msg->front_alloc_len;
209 
210 		num = 1 + !!monc->want_next_osdmap + !!monc->want_mdsmap;
211 		ceph_encode_32(&p, num);
212 
213 		if (monc->want_next_osdmap) {
214 			dout("__send_subscribe to 'osdmap' %u\n",
215 			     (unsigned int)monc->have_osdmap);
216 			ceph_encode_string(&p, end, "osdmap", 6);
217 			i = p;
218 			i->have = cpu_to_le64(monc->have_osdmap);
219 			i->onetime = 1;
220 			p += sizeof(*i);
221 			monc->want_next_osdmap = 2;  /* requested */
222 		}
223 		if (monc->want_mdsmap) {
224 			dout("__send_subscribe to 'mdsmap' %u+\n",
225 			     (unsigned int)monc->have_mdsmap);
226 			ceph_encode_string(&p, end, "mdsmap", 6);
227 			i = p;
228 			i->have = cpu_to_le64(monc->have_mdsmap);
229 			i->onetime = 0;
230 			p += sizeof(*i);
231 		}
232 		ceph_encode_string(&p, end, "monmap", 6);
233 		i = p;
234 		i->have = 0;
235 		i->onetime = 0;
236 		p += sizeof(*i);
237 
238 		msg->front.iov_len = p - msg->front.iov_base;
239 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
240 		ceph_msg_revoke(msg);
241 		ceph_con_send(&monc->con, ceph_msg_get(msg));
242 
243 		monc->sub_sent = jiffies | 1;  /* never 0 */
244 	}
245 }
246 
247 static void handle_subscribe_ack(struct ceph_mon_client *monc,
248 				 struct ceph_msg *msg)
249 {
250 	unsigned int seconds;
251 	struct ceph_mon_subscribe_ack *h = msg->front.iov_base;
252 
253 	if (msg->front.iov_len < sizeof(*h))
254 		goto bad;
255 	seconds = le32_to_cpu(h->duration);
256 
257 	mutex_lock(&monc->mutex);
258 	if (monc->hunting) {
259 		pr_info("mon%d %s session established\n",
260 			monc->cur_mon,
261 			ceph_pr_addr(&monc->con.peer_addr.in_addr));
262 		monc->hunting = false;
263 	}
264 	dout("handle_subscribe_ack after %d seconds\n", seconds);
265 	monc->sub_renew_after = monc->sub_sent + (seconds >> 1)*HZ - 1;
266 	monc->sub_sent = 0;
267 	mutex_unlock(&monc->mutex);
268 	return;
269 bad:
270 	pr_err("got corrupt subscribe-ack msg\n");
271 	ceph_msg_dump(msg);
272 }
273 
274 /*
275  * Keep track of which maps we have
276  */
277 int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 got)
278 {
279 	mutex_lock(&monc->mutex);
280 	monc->have_mdsmap = got;
281 	mutex_unlock(&monc->mutex);
282 	return 0;
283 }
284 EXPORT_SYMBOL(ceph_monc_got_mdsmap);
285 
286 int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 got)
287 {
288 	mutex_lock(&monc->mutex);
289 	monc->have_osdmap = got;
290 	monc->want_next_osdmap = 0;
291 	mutex_unlock(&monc->mutex);
292 	return 0;
293 }
294 
295 /*
296  * Register interest in the next osdmap
297  */
298 void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc)
299 {
300 	dout("request_next_osdmap have %u\n", monc->have_osdmap);
301 	mutex_lock(&monc->mutex);
302 	if (!monc->want_next_osdmap)
303 		monc->want_next_osdmap = 1;
304 	if (monc->want_next_osdmap < 2)
305 		__send_subscribe(monc);
306 	mutex_unlock(&monc->mutex);
307 }
308 EXPORT_SYMBOL(ceph_monc_request_next_osdmap);
309 
310 /*
311  * Wait for an osdmap with a given epoch.
312  *
313  * @epoch: epoch to wait for
314  * @timeout: in jiffies, 0 means "wait forever"
315  */
316 int ceph_monc_wait_osdmap(struct ceph_mon_client *monc, u32 epoch,
317 			  unsigned long timeout)
318 {
319 	unsigned long started = jiffies;
320 	long ret;
321 
322 	mutex_lock(&monc->mutex);
323 	while (monc->have_osdmap < epoch) {
324 		mutex_unlock(&monc->mutex);
325 
326 		if (timeout && time_after_eq(jiffies, started + timeout))
327 			return -ETIMEDOUT;
328 
329 		ret = wait_event_interruptible_timeout(monc->client->auth_wq,
330 						monc->have_osdmap >= epoch,
331 						ceph_timeout_jiffies(timeout));
332 		if (ret < 0)
333 			return ret;
334 
335 		mutex_lock(&monc->mutex);
336 	}
337 
338 	mutex_unlock(&monc->mutex);
339 	return 0;
340 }
341 EXPORT_SYMBOL(ceph_monc_wait_osdmap);
342 
343 /*
344  *
345  */
346 int ceph_monc_open_session(struct ceph_mon_client *monc)
347 {
348 	mutex_lock(&monc->mutex);
349 	__open_session(monc);
350 	__schedule_delayed(monc);
351 	mutex_unlock(&monc->mutex);
352 	return 0;
353 }
354 EXPORT_SYMBOL(ceph_monc_open_session);
355 
356 /*
357  * We require the fsid and global_id in order to initialize our
358  * debugfs dir.
359  */
360 static bool have_debugfs_info(struct ceph_mon_client *monc)
361 {
362 	dout("have_debugfs_info fsid %d globalid %lld\n",
363 	     (int)monc->client->have_fsid, monc->auth->global_id);
364 	return monc->client->have_fsid && monc->auth->global_id > 0;
365 }
366 
367 /*
368  * The monitor responds with mount ack indicate mount success.  The
369  * included client ticket allows the client to talk to MDSs and OSDs.
370  */
371 static void ceph_monc_handle_map(struct ceph_mon_client *monc,
372 				 struct ceph_msg *msg)
373 {
374 	struct ceph_client *client = monc->client;
375 	struct ceph_monmap *monmap = NULL, *old = monc->monmap;
376 	void *p, *end;
377 	int had_debugfs_info, init_debugfs = 0;
378 
379 	mutex_lock(&monc->mutex);
380 
381 	had_debugfs_info = have_debugfs_info(monc);
382 
383 	dout("handle_monmap\n");
384 	p = msg->front.iov_base;
385 	end = p + msg->front.iov_len;
386 
387 	monmap = ceph_monmap_decode(p, end);
388 	if (IS_ERR(monmap)) {
389 		pr_err("problem decoding monmap, %d\n",
390 		       (int)PTR_ERR(monmap));
391 		goto out;
392 	}
393 
394 	if (ceph_check_fsid(monc->client, &monmap->fsid) < 0) {
395 		kfree(monmap);
396 		goto out;
397 	}
398 
399 	client->monc.monmap = monmap;
400 	kfree(old);
401 
402 	if (!client->have_fsid) {
403 		client->have_fsid = true;
404 		if (!had_debugfs_info && have_debugfs_info(monc)) {
405 			pr_info("client%lld fsid %pU\n",
406 				ceph_client_id(monc->client),
407 				&monc->client->fsid);
408 			init_debugfs = 1;
409 		}
410 		mutex_unlock(&monc->mutex);
411 
412 		if (init_debugfs) {
413 			/*
414 			 * do debugfs initialization without mutex to avoid
415 			 * creating a locking dependency
416 			 */
417 			ceph_debugfs_client_init(monc->client);
418 		}
419 
420 		goto out_unlocked;
421 	}
422 out:
423 	mutex_unlock(&monc->mutex);
424 out_unlocked:
425 	wake_up_all(&client->auth_wq);
426 }
427 
428 /*
429  * generic requests (currently statfs, mon_get_version)
430  */
431 static struct ceph_mon_generic_request *__lookup_generic_req(
432 	struct ceph_mon_client *monc, u64 tid)
433 {
434 	struct ceph_mon_generic_request *req;
435 	struct rb_node *n = monc->generic_request_tree.rb_node;
436 
437 	while (n) {
438 		req = rb_entry(n, struct ceph_mon_generic_request, node);
439 		if (tid < req->tid)
440 			n = n->rb_left;
441 		else if (tid > req->tid)
442 			n = n->rb_right;
443 		else
444 			return req;
445 	}
446 	return NULL;
447 }
448 
449 static void __insert_generic_request(struct ceph_mon_client *monc,
450 			    struct ceph_mon_generic_request *new)
451 {
452 	struct rb_node **p = &monc->generic_request_tree.rb_node;
453 	struct rb_node *parent = NULL;
454 	struct ceph_mon_generic_request *req = NULL;
455 
456 	while (*p) {
457 		parent = *p;
458 		req = rb_entry(parent, struct ceph_mon_generic_request, node);
459 		if (new->tid < req->tid)
460 			p = &(*p)->rb_left;
461 		else if (new->tid > req->tid)
462 			p = &(*p)->rb_right;
463 		else
464 			BUG();
465 	}
466 
467 	rb_link_node(&new->node, parent, p);
468 	rb_insert_color(&new->node, &monc->generic_request_tree);
469 }
470 
471 static void release_generic_request(struct kref *kref)
472 {
473 	struct ceph_mon_generic_request *req =
474 		container_of(kref, struct ceph_mon_generic_request, kref);
475 
476 	if (req->reply)
477 		ceph_msg_put(req->reply);
478 	if (req->request)
479 		ceph_msg_put(req->request);
480 
481 	kfree(req);
482 }
483 
484 static void put_generic_request(struct ceph_mon_generic_request *req)
485 {
486 	kref_put(&req->kref, release_generic_request);
487 }
488 
489 static void get_generic_request(struct ceph_mon_generic_request *req)
490 {
491 	kref_get(&req->kref);
492 }
493 
494 static struct ceph_msg *get_generic_reply(struct ceph_connection *con,
495 					 struct ceph_msg_header *hdr,
496 					 int *skip)
497 {
498 	struct ceph_mon_client *monc = con->private;
499 	struct ceph_mon_generic_request *req;
500 	u64 tid = le64_to_cpu(hdr->tid);
501 	struct ceph_msg *m;
502 
503 	mutex_lock(&monc->mutex);
504 	req = __lookup_generic_req(monc, tid);
505 	if (!req) {
506 		dout("get_generic_reply %lld dne\n", tid);
507 		*skip = 1;
508 		m = NULL;
509 	} else {
510 		dout("get_generic_reply %lld got %p\n", tid, req->reply);
511 		*skip = 0;
512 		m = ceph_msg_get(req->reply);
513 		/*
514 		 * we don't need to track the connection reading into
515 		 * this reply because we only have one open connection
516 		 * at a time, ever.
517 		 */
518 	}
519 	mutex_unlock(&monc->mutex);
520 	return m;
521 }
522 
523 static int __do_generic_request(struct ceph_mon_client *monc, u64 tid,
524 				struct ceph_mon_generic_request *req)
525 {
526 	int err;
527 
528 	/* register request */
529 	req->tid = tid != 0 ? tid : ++monc->last_tid;
530 	req->request->hdr.tid = cpu_to_le64(req->tid);
531 	__insert_generic_request(monc, req);
532 	monc->num_generic_requests++;
533 	ceph_con_send(&monc->con, ceph_msg_get(req->request));
534 	mutex_unlock(&monc->mutex);
535 
536 	err = wait_for_completion_interruptible(&req->completion);
537 
538 	mutex_lock(&monc->mutex);
539 	rb_erase(&req->node, &monc->generic_request_tree);
540 	monc->num_generic_requests--;
541 
542 	if (!err)
543 		err = req->result;
544 	return err;
545 }
546 
547 static int do_generic_request(struct ceph_mon_client *monc,
548 			      struct ceph_mon_generic_request *req)
549 {
550 	int err;
551 
552 	mutex_lock(&monc->mutex);
553 	err = __do_generic_request(monc, 0, req);
554 	mutex_unlock(&monc->mutex);
555 
556 	return err;
557 }
558 
559 /*
560  * statfs
561  */
562 static void handle_statfs_reply(struct ceph_mon_client *monc,
563 				struct ceph_msg *msg)
564 {
565 	struct ceph_mon_generic_request *req;
566 	struct ceph_mon_statfs_reply *reply = msg->front.iov_base;
567 	u64 tid = le64_to_cpu(msg->hdr.tid);
568 
569 	if (msg->front.iov_len != sizeof(*reply))
570 		goto bad;
571 	dout("handle_statfs_reply %p tid %llu\n", msg, tid);
572 
573 	mutex_lock(&monc->mutex);
574 	req = __lookup_generic_req(monc, tid);
575 	if (req) {
576 		*(struct ceph_statfs *)req->buf = reply->st;
577 		req->result = 0;
578 		get_generic_request(req);
579 	}
580 	mutex_unlock(&monc->mutex);
581 	if (req) {
582 		complete_all(&req->completion);
583 		put_generic_request(req);
584 	}
585 	return;
586 
587 bad:
588 	pr_err("corrupt statfs reply, tid %llu\n", tid);
589 	ceph_msg_dump(msg);
590 }
591 
592 /*
593  * Do a synchronous statfs().
594  */
595 int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf)
596 {
597 	struct ceph_mon_generic_request *req;
598 	struct ceph_mon_statfs *h;
599 	int err;
600 
601 	req = kzalloc(sizeof(*req), GFP_NOFS);
602 	if (!req)
603 		return -ENOMEM;
604 
605 	kref_init(&req->kref);
606 	req->buf = buf;
607 	init_completion(&req->completion);
608 
609 	err = -ENOMEM;
610 	req->request = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), GFP_NOFS,
611 				    true);
612 	if (!req->request)
613 		goto out;
614 	req->reply = ceph_msg_new(CEPH_MSG_STATFS_REPLY, 1024, GFP_NOFS,
615 				  true);
616 	if (!req->reply)
617 		goto out;
618 
619 	/* fill out request */
620 	h = req->request->front.iov_base;
621 	h->monhdr.have_version = 0;
622 	h->monhdr.session_mon = cpu_to_le16(-1);
623 	h->monhdr.session_mon_tid = 0;
624 	h->fsid = monc->monmap->fsid;
625 
626 	err = do_generic_request(monc, req);
627 
628 out:
629 	put_generic_request(req);
630 	return err;
631 }
632 EXPORT_SYMBOL(ceph_monc_do_statfs);
633 
634 static void handle_get_version_reply(struct ceph_mon_client *monc,
635 				     struct ceph_msg *msg)
636 {
637 	struct ceph_mon_generic_request *req;
638 	u64 tid = le64_to_cpu(msg->hdr.tid);
639 	void *p = msg->front.iov_base;
640 	void *end = p + msg->front_alloc_len;
641 	u64 handle;
642 
643 	dout("%s %p tid %llu\n", __func__, msg, tid);
644 
645 	ceph_decode_need(&p, end, 2*sizeof(u64), bad);
646 	handle = ceph_decode_64(&p);
647 	if (tid != 0 && tid != handle)
648 		goto bad;
649 
650 	mutex_lock(&monc->mutex);
651 	req = __lookup_generic_req(monc, handle);
652 	if (req) {
653 		*(u64 *)req->buf = ceph_decode_64(&p);
654 		req->result = 0;
655 		get_generic_request(req);
656 	}
657 	mutex_unlock(&monc->mutex);
658 	if (req) {
659 		complete_all(&req->completion);
660 		put_generic_request(req);
661 	}
662 
663 	return;
664 bad:
665 	pr_err("corrupt mon_get_version reply, tid %llu\n", tid);
666 	ceph_msg_dump(msg);
667 }
668 
669 /*
670  * Send MMonGetVersion and wait for the reply.
671  *
672  * @what: one of "mdsmap", "osdmap" or "monmap"
673  */
674 int ceph_monc_do_get_version(struct ceph_mon_client *monc, const char *what,
675 			     u64 *newest)
676 {
677 	struct ceph_mon_generic_request *req;
678 	void *p, *end;
679 	u64 tid;
680 	int err;
681 
682 	req = kzalloc(sizeof(*req), GFP_NOFS);
683 	if (!req)
684 		return -ENOMEM;
685 
686 	kref_init(&req->kref);
687 	req->buf = newest;
688 	init_completion(&req->completion);
689 
690 	req->request = ceph_msg_new(CEPH_MSG_MON_GET_VERSION,
691 				    sizeof(u64) + sizeof(u32) + strlen(what),
692 				    GFP_NOFS, true);
693 	if (!req->request) {
694 		err = -ENOMEM;
695 		goto out;
696 	}
697 
698 	req->reply = ceph_msg_new(CEPH_MSG_MON_GET_VERSION_REPLY, 1024,
699 				  GFP_NOFS, true);
700 	if (!req->reply) {
701 		err = -ENOMEM;
702 		goto out;
703 	}
704 
705 	p = req->request->front.iov_base;
706 	end = p + req->request->front_alloc_len;
707 
708 	/* fill out request */
709 	mutex_lock(&monc->mutex);
710 	tid = ++monc->last_tid;
711 	ceph_encode_64(&p, tid); /* handle */
712 	ceph_encode_string(&p, end, what, strlen(what));
713 
714 	err = __do_generic_request(monc, tid, req);
715 
716 	mutex_unlock(&monc->mutex);
717 out:
718 	put_generic_request(req);
719 	return err;
720 }
721 EXPORT_SYMBOL(ceph_monc_do_get_version);
722 
723 /*
724  * Resend pending generic requests.
725  */
726 static void __resend_generic_request(struct ceph_mon_client *monc)
727 {
728 	struct ceph_mon_generic_request *req;
729 	struct rb_node *p;
730 
731 	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
732 		req = rb_entry(p, struct ceph_mon_generic_request, node);
733 		ceph_msg_revoke(req->request);
734 		ceph_msg_revoke_incoming(req->reply);
735 		ceph_con_send(&monc->con, ceph_msg_get(req->request));
736 	}
737 }
738 
739 /*
740  * Delayed work.  If we haven't mounted yet, retry.  Otherwise,
741  * renew/retry subscription as needed (in case it is timing out, or we
742  * got an ENOMEM).  And keep the monitor connection alive.
743  */
744 static void delayed_work(struct work_struct *work)
745 {
746 	struct ceph_mon_client *monc =
747 		container_of(work, struct ceph_mon_client, delayed_work.work);
748 
749 	dout("monc delayed_work\n");
750 	mutex_lock(&monc->mutex);
751 	if (monc->hunting) {
752 		__close_session(monc);
753 		__open_session(monc);  /* continue hunting */
754 	} else {
755 		struct ceph_options *opt = monc->client->options;
756 		int is_auth = ceph_auth_is_authenticated(monc->auth);
757 		if (ceph_con_keepalive_expired(&monc->con,
758 					       opt->monc_ping_timeout)) {
759 			dout("monc keepalive timeout\n");
760 			is_auth = 0;
761 			__close_session(monc);
762 			monc->hunting = true;
763 			__open_session(monc);
764 		}
765 
766 		if (!monc->hunting) {
767 			ceph_con_keepalive(&monc->con);
768 			__validate_auth(monc);
769 		}
770 
771 		if (is_auth)
772 			__send_subscribe(monc);
773 	}
774 	__schedule_delayed(monc);
775 	mutex_unlock(&monc->mutex);
776 }
777 
778 /*
779  * On startup, we build a temporary monmap populated with the IPs
780  * provided by mount(2).
781  */
782 static int build_initial_monmap(struct ceph_mon_client *monc)
783 {
784 	struct ceph_options *opt = monc->client->options;
785 	struct ceph_entity_addr *mon_addr = opt->mon_addr;
786 	int num_mon = opt->num_mon;
787 	int i;
788 
789 	/* build initial monmap */
790 	monc->monmap = kzalloc(sizeof(*monc->monmap) +
791 			       num_mon*sizeof(monc->monmap->mon_inst[0]),
792 			       GFP_KERNEL);
793 	if (!monc->monmap)
794 		return -ENOMEM;
795 	for (i = 0; i < num_mon; i++) {
796 		monc->monmap->mon_inst[i].addr = mon_addr[i];
797 		monc->monmap->mon_inst[i].addr.nonce = 0;
798 		monc->monmap->mon_inst[i].name.type =
799 			CEPH_ENTITY_TYPE_MON;
800 		monc->monmap->mon_inst[i].name.num = cpu_to_le64(i);
801 	}
802 	monc->monmap->num_mon = num_mon;
803 	return 0;
804 }
805 
806 int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl)
807 {
808 	int err = 0;
809 
810 	dout("init\n");
811 	memset(monc, 0, sizeof(*monc));
812 	monc->client = cl;
813 	monc->monmap = NULL;
814 	mutex_init(&monc->mutex);
815 
816 	err = build_initial_monmap(monc);
817 	if (err)
818 		goto out;
819 
820 	/* connection */
821 	/* authentication */
822 	monc->auth = ceph_auth_init(cl->options->name,
823 				    cl->options->key);
824 	if (IS_ERR(monc->auth)) {
825 		err = PTR_ERR(monc->auth);
826 		goto out_monmap;
827 	}
828 	monc->auth->want_keys =
829 		CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON |
830 		CEPH_ENTITY_TYPE_OSD | CEPH_ENTITY_TYPE_MDS;
831 
832 	/* msgs */
833 	err = -ENOMEM;
834 	monc->m_subscribe_ack = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE_ACK,
835 				     sizeof(struct ceph_mon_subscribe_ack),
836 				     GFP_NOFS, true);
837 	if (!monc->m_subscribe_ack)
838 		goto out_auth;
839 
840 	monc->m_subscribe = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, GFP_NOFS,
841 					 true);
842 	if (!monc->m_subscribe)
843 		goto out_subscribe_ack;
844 
845 	monc->m_auth_reply = ceph_msg_new(CEPH_MSG_AUTH_REPLY, 4096, GFP_NOFS,
846 					  true);
847 	if (!monc->m_auth_reply)
848 		goto out_subscribe;
849 
850 	monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, GFP_NOFS, true);
851 	monc->pending_auth = 0;
852 	if (!monc->m_auth)
853 		goto out_auth_reply;
854 
855 	ceph_con_init(&monc->con, monc, &mon_con_ops,
856 		      &monc->client->msgr);
857 
858 	monc->cur_mon = -1;
859 	monc->hunting = true;
860 	monc->sub_renew_after = jiffies;
861 	monc->sub_sent = 0;
862 
863 	INIT_DELAYED_WORK(&monc->delayed_work, delayed_work);
864 	monc->generic_request_tree = RB_ROOT;
865 	monc->num_generic_requests = 0;
866 	monc->last_tid = 0;
867 
868 	monc->have_mdsmap = 0;
869 	monc->have_osdmap = 0;
870 	monc->want_next_osdmap = 1;
871 	return 0;
872 
873 out_auth_reply:
874 	ceph_msg_put(monc->m_auth_reply);
875 out_subscribe:
876 	ceph_msg_put(monc->m_subscribe);
877 out_subscribe_ack:
878 	ceph_msg_put(monc->m_subscribe_ack);
879 out_auth:
880 	ceph_auth_destroy(monc->auth);
881 out_monmap:
882 	kfree(monc->monmap);
883 out:
884 	return err;
885 }
886 EXPORT_SYMBOL(ceph_monc_init);
887 
888 void ceph_monc_stop(struct ceph_mon_client *monc)
889 {
890 	dout("stop\n");
891 	cancel_delayed_work_sync(&monc->delayed_work);
892 
893 	mutex_lock(&monc->mutex);
894 	__close_session(monc);
895 
896 	mutex_unlock(&monc->mutex);
897 
898 	/*
899 	 * flush msgr queue before we destroy ourselves to ensure that:
900 	 *  - any work that references our embedded con is finished.
901 	 *  - any osd_client or other work that may reference an authorizer
902 	 *    finishes before we shut down the auth subsystem.
903 	 */
904 	ceph_msgr_flush();
905 
906 	ceph_auth_destroy(monc->auth);
907 
908 	ceph_msg_put(monc->m_auth);
909 	ceph_msg_put(monc->m_auth_reply);
910 	ceph_msg_put(monc->m_subscribe);
911 	ceph_msg_put(monc->m_subscribe_ack);
912 
913 	kfree(monc->monmap);
914 }
915 EXPORT_SYMBOL(ceph_monc_stop);
916 
917 static void handle_auth_reply(struct ceph_mon_client *monc,
918 			      struct ceph_msg *msg)
919 {
920 	int ret;
921 	int was_auth = 0;
922 	int had_debugfs_info, init_debugfs = 0;
923 
924 	mutex_lock(&monc->mutex);
925 	had_debugfs_info = have_debugfs_info(monc);
926 	was_auth = ceph_auth_is_authenticated(monc->auth);
927 	monc->pending_auth = 0;
928 	ret = ceph_handle_auth_reply(monc->auth, msg->front.iov_base,
929 				     msg->front.iov_len,
930 				     monc->m_auth->front.iov_base,
931 				     monc->m_auth->front_alloc_len);
932 	if (ret < 0) {
933 		monc->client->auth_err = ret;
934 		wake_up_all(&monc->client->auth_wq);
935 	} else if (ret > 0) {
936 		__send_prepared_auth_request(monc, ret);
937 	} else if (!was_auth && ceph_auth_is_authenticated(monc->auth)) {
938 		dout("authenticated, starting session\n");
939 
940 		monc->client->msgr.inst.name.type = CEPH_ENTITY_TYPE_CLIENT;
941 		monc->client->msgr.inst.name.num =
942 					cpu_to_le64(monc->auth->global_id);
943 
944 		__send_subscribe(monc);
945 		__resend_generic_request(monc);
946 	}
947 
948 	if (!had_debugfs_info && have_debugfs_info(monc)) {
949 		pr_info("client%lld fsid %pU\n",
950 			ceph_client_id(monc->client),
951 			&monc->client->fsid);
952 		init_debugfs = 1;
953 	}
954 	mutex_unlock(&monc->mutex);
955 
956 	if (init_debugfs) {
957 		/*
958 		 * do debugfs initialization without mutex to avoid
959 		 * creating a locking dependency
960 		 */
961 		ceph_debugfs_client_init(monc->client);
962 	}
963 }
964 
965 static int __validate_auth(struct ceph_mon_client *monc)
966 {
967 	int ret;
968 
969 	if (monc->pending_auth)
970 		return 0;
971 
972 	ret = ceph_build_auth(monc->auth, monc->m_auth->front.iov_base,
973 			      monc->m_auth->front_alloc_len);
974 	if (ret <= 0)
975 		return ret; /* either an error, or no need to authenticate */
976 	__send_prepared_auth_request(monc, ret);
977 	return 0;
978 }
979 
980 int ceph_monc_validate_auth(struct ceph_mon_client *monc)
981 {
982 	int ret;
983 
984 	mutex_lock(&monc->mutex);
985 	ret = __validate_auth(monc);
986 	mutex_unlock(&monc->mutex);
987 	return ret;
988 }
989 EXPORT_SYMBOL(ceph_monc_validate_auth);
990 
991 /*
992  * handle incoming message
993  */
994 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
995 {
996 	struct ceph_mon_client *monc = con->private;
997 	int type = le16_to_cpu(msg->hdr.type);
998 
999 	if (!monc)
1000 		return;
1001 
1002 	switch (type) {
1003 	case CEPH_MSG_AUTH_REPLY:
1004 		handle_auth_reply(monc, msg);
1005 		break;
1006 
1007 	case CEPH_MSG_MON_SUBSCRIBE_ACK:
1008 		handle_subscribe_ack(monc, msg);
1009 		break;
1010 
1011 	case CEPH_MSG_STATFS_REPLY:
1012 		handle_statfs_reply(monc, msg);
1013 		break;
1014 
1015 	case CEPH_MSG_MON_GET_VERSION_REPLY:
1016 		handle_get_version_reply(monc, msg);
1017 		break;
1018 
1019 	case CEPH_MSG_MON_MAP:
1020 		ceph_monc_handle_map(monc, msg);
1021 		break;
1022 
1023 	case CEPH_MSG_OSD_MAP:
1024 		ceph_osdc_handle_map(&monc->client->osdc, msg);
1025 		break;
1026 
1027 	default:
1028 		/* can the chained handler handle it? */
1029 		if (monc->client->extra_mon_dispatch &&
1030 		    monc->client->extra_mon_dispatch(monc->client, msg) == 0)
1031 			break;
1032 
1033 		pr_err("received unknown message type %d %s\n", type,
1034 		       ceph_msg_type_name(type));
1035 	}
1036 	ceph_msg_put(msg);
1037 }
1038 
1039 /*
1040  * Allocate memory for incoming message
1041  */
1042 static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con,
1043 				      struct ceph_msg_header *hdr,
1044 				      int *skip)
1045 {
1046 	struct ceph_mon_client *monc = con->private;
1047 	int type = le16_to_cpu(hdr->type);
1048 	int front_len = le32_to_cpu(hdr->front_len);
1049 	struct ceph_msg *m = NULL;
1050 
1051 	*skip = 0;
1052 
1053 	switch (type) {
1054 	case CEPH_MSG_MON_SUBSCRIBE_ACK:
1055 		m = ceph_msg_get(monc->m_subscribe_ack);
1056 		break;
1057 	case CEPH_MSG_STATFS_REPLY:
1058 		return get_generic_reply(con, hdr, skip);
1059 	case CEPH_MSG_AUTH_REPLY:
1060 		m = ceph_msg_get(monc->m_auth_reply);
1061 		break;
1062 	case CEPH_MSG_MON_GET_VERSION_REPLY:
1063 		if (le64_to_cpu(hdr->tid) != 0)
1064 			return get_generic_reply(con, hdr, skip);
1065 
1066 		/*
1067 		 * Older OSDs don't set reply tid even if the orignal
1068 		 * request had a non-zero tid.  Workaround this weirdness
1069 		 * by falling through to the allocate case.
1070 		 */
1071 	case CEPH_MSG_MON_MAP:
1072 	case CEPH_MSG_MDS_MAP:
1073 	case CEPH_MSG_OSD_MAP:
1074 		m = ceph_msg_new(type, front_len, GFP_NOFS, false);
1075 		if (!m)
1076 			return NULL;	/* ENOMEM--return skip == 0 */
1077 		break;
1078 	}
1079 
1080 	if (!m) {
1081 		pr_info("alloc_msg unknown type %d\n", type);
1082 		*skip = 1;
1083 	} else if (front_len > m->front_alloc_len) {
1084 		pr_warn("mon_alloc_msg front %d > prealloc %d (%u#%llu)\n",
1085 			front_len, m->front_alloc_len,
1086 			(unsigned int)con->peer_name.type,
1087 			le64_to_cpu(con->peer_name.num));
1088 		ceph_msg_put(m);
1089 		m = ceph_msg_new(type, front_len, GFP_NOFS, false);
1090 	}
1091 
1092 	return m;
1093 }
1094 
1095 /*
1096  * If the monitor connection resets, pick a new monitor and resubmit
1097  * any pending requests.
1098  */
1099 static void mon_fault(struct ceph_connection *con)
1100 {
1101 	struct ceph_mon_client *monc = con->private;
1102 
1103 	if (!monc)
1104 		return;
1105 
1106 	dout("mon_fault\n");
1107 	mutex_lock(&monc->mutex);
1108 	if (!con->private)
1109 		goto out;
1110 
1111 	if (!monc->hunting)
1112 		pr_info("mon%d %s session lost, "
1113 			"hunting for new mon\n", monc->cur_mon,
1114 			ceph_pr_addr(&monc->con.peer_addr.in_addr));
1115 
1116 	__close_session(monc);
1117 	if (!monc->hunting) {
1118 		/* start hunting */
1119 		monc->hunting = true;
1120 		__open_session(monc);
1121 	} else {
1122 		/* already hunting, let's wait a bit */
1123 		__schedule_delayed(monc);
1124 	}
1125 out:
1126 	mutex_unlock(&monc->mutex);
1127 }
1128 
1129 /*
1130  * We can ignore refcounting on the connection struct, as all references
1131  * will come from the messenger workqueue, which is drained prior to
1132  * mon_client destruction.
1133  */
1134 static struct ceph_connection *con_get(struct ceph_connection *con)
1135 {
1136 	return con;
1137 }
1138 
1139 static void con_put(struct ceph_connection *con)
1140 {
1141 }
1142 
1143 static const struct ceph_connection_operations mon_con_ops = {
1144 	.get = con_get,
1145 	.put = con_put,
1146 	.dispatch = dispatch,
1147 	.fault = mon_fault,
1148 	.alloc_msg = mon_alloc_msg,
1149 };
1150