]>
Commit | Line | Data |
---|---|---|
2f2dc053 SW |
1 | #include "ceph_debug.h" |
2 | ||
3 | #include <linux/wait.h> | |
5a0e3ad6 | 4 | #include <linux/slab.h> |
2f2dc053 SW |
5 | #include <linux/sched.h> |
6 | ||
7 | #include "mds_client.h" | |
8 | #include "mon_client.h" | |
9 | #include "super.h" | |
10 | #include "messenger.h" | |
11 | #include "decode.h" | |
4e7a5dcd | 12 | #include "auth.h" |
93cea5be | 13 | #include "pagelist.h" |
2f2dc053 SW |
14 | |
15 | /* | |
16 | * A cluster of MDS (metadata server) daemons is responsible for | |
17 | * managing the file system namespace (the directory hierarchy and | |
18 | * inodes) and for coordinating shared access to storage. Metadata is | |
19 | * partitioning hierarchically across a number of servers, and that | |
20 | * partition varies over time as the cluster adjusts the distribution | |
21 | * in order to balance load. | |
22 | * | |
23 | * The MDS client is primarily responsible to managing synchronous | |
24 | * metadata requests for operations like open, unlink, and so forth. | |
25 | * If there is a MDS failure, we find out about it when we (possibly | |
26 | * request and) receive a new MDS map, and can resubmit affected | |
27 | * requests. | |
28 | * | |
29 | * For the most part, though, we take advantage of a lossless | |
30 | * communications channel to the MDS, and do not need to worry about | |
31 | * timing out or resubmitting requests. | |
32 | * | |
33 | * We maintain a stateful "session" with each MDS we interact with. | |
34 | * Within each session, we sent periodic heartbeat messages to ensure | |
35 | * any capabilities or leases we have been issues remain valid. If | |
36 | * the session times out and goes stale, our leases and capabilities | |
37 | * are no longer valid. | |
38 | */ | |
39 | ||
20cb34ae SW |
40 | struct ceph_reconnect_state { |
41 | struct ceph_pagelist *pagelist; | |
42 | bool flock; | |
43 | }; | |
44 | ||
2f2dc053 SW |
45 | static void __wake_requests(struct ceph_mds_client *mdsc, |
46 | struct list_head *head); | |
47 | ||
9e32789f | 48 | static const struct ceph_connection_operations mds_con_ops; |
2f2dc053 SW |
49 | |
50 | ||
51 | /* | |
52 | * mds reply parsing | |
53 | */ | |
54 | ||
55 | /* | |
56 | * parse individual inode info | |
57 | */ | |
58 | static int parse_reply_info_in(void **p, void *end, | |
59 | struct ceph_mds_reply_info_in *info) | |
60 | { | |
61 | int err = -EIO; | |
62 | ||
63 | info->in = *p; | |
64 | *p += sizeof(struct ceph_mds_reply_inode) + | |
65 | sizeof(*info->in->fragtree.splits) * | |
66 | le32_to_cpu(info->in->fragtree.nsplits); | |
67 | ||
68 | ceph_decode_32_safe(p, end, info->symlink_len, bad); | |
69 | ceph_decode_need(p, end, info->symlink_len, bad); | |
70 | info->symlink = *p; | |
71 | *p += info->symlink_len; | |
72 | ||
73 | ceph_decode_32_safe(p, end, info->xattr_len, bad); | |
74 | ceph_decode_need(p, end, info->xattr_len, bad); | |
75 | info->xattr_data = *p; | |
76 | *p += info->xattr_len; | |
77 | return 0; | |
78 | bad: | |
79 | return err; | |
80 | } | |
81 | ||
82 | /* | |
83 | * parse a normal reply, which may contain a (dir+)dentry and/or a | |
84 | * target inode. | |
85 | */ | |
86 | static int parse_reply_info_trace(void **p, void *end, | |
87 | struct ceph_mds_reply_info_parsed *info) | |
88 | { | |
89 | int err; | |
90 | ||
91 | if (info->head->is_dentry) { | |
92 | err = parse_reply_info_in(p, end, &info->diri); | |
93 | if (err < 0) | |
94 | goto out_bad; | |
95 | ||
96 | if (unlikely(*p + sizeof(*info->dirfrag) > end)) | |
97 | goto bad; | |
98 | info->dirfrag = *p; | |
99 | *p += sizeof(*info->dirfrag) + | |
100 | sizeof(u32)*le32_to_cpu(info->dirfrag->ndist); | |
101 | if (unlikely(*p > end)) | |
102 | goto bad; | |
103 | ||
104 | ceph_decode_32_safe(p, end, info->dname_len, bad); | |
105 | ceph_decode_need(p, end, info->dname_len, bad); | |
106 | info->dname = *p; | |
107 | *p += info->dname_len; | |
108 | info->dlease = *p; | |
109 | *p += sizeof(*info->dlease); | |
110 | } | |
111 | ||
112 | if (info->head->is_target) { | |
113 | err = parse_reply_info_in(p, end, &info->targeti); | |
114 | if (err < 0) | |
115 | goto out_bad; | |
116 | } | |
117 | ||
118 | if (unlikely(*p != end)) | |
119 | goto bad; | |
120 | return 0; | |
121 | ||
122 | bad: | |
123 | err = -EIO; | |
124 | out_bad: | |
125 | pr_err("problem parsing mds trace %d\n", err); | |
126 | return err; | |
127 | } | |
128 | ||
129 | /* | |
130 | * parse readdir results | |
131 | */ | |
132 | static int parse_reply_info_dir(void **p, void *end, | |
133 | struct ceph_mds_reply_info_parsed *info) | |
134 | { | |
135 | u32 num, i = 0; | |
136 | int err; | |
137 | ||
138 | info->dir_dir = *p; | |
139 | if (*p + sizeof(*info->dir_dir) > end) | |
140 | goto bad; | |
141 | *p += sizeof(*info->dir_dir) + | |
142 | sizeof(u32)*le32_to_cpu(info->dir_dir->ndist); | |
143 | if (*p > end) | |
144 | goto bad; | |
145 | ||
146 | ceph_decode_need(p, end, sizeof(num) + 2, bad); | |
c89136ea SW |
147 | num = ceph_decode_32(p); |
148 | info->dir_end = ceph_decode_8(p); | |
149 | info->dir_complete = ceph_decode_8(p); | |
2f2dc053 SW |
150 | if (num == 0) |
151 | goto done; | |
152 | ||
153 | /* alloc large array */ | |
154 | info->dir_nr = num; | |
155 | info->dir_in = kcalloc(num, sizeof(*info->dir_in) + | |
156 | sizeof(*info->dir_dname) + | |
157 | sizeof(*info->dir_dname_len) + | |
158 | sizeof(*info->dir_dlease), | |
159 | GFP_NOFS); | |
160 | if (info->dir_in == NULL) { | |
161 | err = -ENOMEM; | |
162 | goto out_bad; | |
163 | } | |
164 | info->dir_dname = (void *)(info->dir_in + num); | |
165 | info->dir_dname_len = (void *)(info->dir_dname + num); | |
166 | info->dir_dlease = (void *)(info->dir_dname_len + num); | |
167 | ||
168 | while (num) { | |
169 | /* dentry */ | |
170 | ceph_decode_need(p, end, sizeof(u32)*2, bad); | |
c89136ea | 171 | info->dir_dname_len[i] = ceph_decode_32(p); |
2f2dc053 SW |
172 | ceph_decode_need(p, end, info->dir_dname_len[i], bad); |
173 | info->dir_dname[i] = *p; | |
174 | *p += info->dir_dname_len[i]; | |
175 | dout("parsed dir dname '%.*s'\n", info->dir_dname_len[i], | |
176 | info->dir_dname[i]); | |
177 | info->dir_dlease[i] = *p; | |
178 | *p += sizeof(struct ceph_mds_reply_lease); | |
179 | ||
180 | /* inode */ | |
181 | err = parse_reply_info_in(p, end, &info->dir_in[i]); | |
182 | if (err < 0) | |
183 | goto out_bad; | |
184 | i++; | |
185 | num--; | |
186 | } | |
187 | ||
188 | done: | |
189 | if (*p != end) | |
190 | goto bad; | |
191 | return 0; | |
192 | ||
193 | bad: | |
194 | err = -EIO; | |
195 | out_bad: | |
196 | pr_err("problem parsing dir contents %d\n", err); | |
197 | return err; | |
198 | } | |
199 | ||
200 | /* | |
201 | * parse entire mds reply | |
202 | */ | |
203 | static int parse_reply_info(struct ceph_msg *msg, | |
204 | struct ceph_mds_reply_info_parsed *info) | |
205 | { | |
206 | void *p, *end; | |
207 | u32 len; | |
208 | int err; | |
209 | ||
210 | info->head = msg->front.iov_base; | |
211 | p = msg->front.iov_base + sizeof(struct ceph_mds_reply_head); | |
212 | end = p + msg->front.iov_len - sizeof(struct ceph_mds_reply_head); | |
213 | ||
214 | /* trace */ | |
215 | ceph_decode_32_safe(&p, end, len, bad); | |
216 | if (len > 0) { | |
217 | err = parse_reply_info_trace(&p, p+len, info); | |
218 | if (err < 0) | |
219 | goto out_bad; | |
220 | } | |
221 | ||
222 | /* dir content */ | |
223 | ceph_decode_32_safe(&p, end, len, bad); | |
224 | if (len > 0) { | |
225 | err = parse_reply_info_dir(&p, p+len, info); | |
226 | if (err < 0) | |
227 | goto out_bad; | |
228 | } | |
229 | ||
230 | /* snap blob */ | |
231 | ceph_decode_32_safe(&p, end, len, bad); | |
232 | info->snapblob_len = len; | |
233 | info->snapblob = p; | |
234 | p += len; | |
235 | ||
236 | if (p != end) | |
237 | goto bad; | |
238 | return 0; | |
239 | ||
240 | bad: | |
241 | err = -EIO; | |
242 | out_bad: | |
243 | pr_err("mds parse_reply err %d\n", err); | |
244 | return err; | |
245 | } | |
246 | ||
247 | static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info) | |
248 | { | |
249 | kfree(info->dir_in); | |
250 | } | |
251 | ||
252 | ||
253 | /* | |
254 | * sessions | |
255 | */ | |
256 | static const char *session_state_name(int s) | |
257 | { | |
258 | switch (s) { | |
259 | case CEPH_MDS_SESSION_NEW: return "new"; | |
260 | case CEPH_MDS_SESSION_OPENING: return "opening"; | |
261 | case CEPH_MDS_SESSION_OPEN: return "open"; | |
262 | case CEPH_MDS_SESSION_HUNG: return "hung"; | |
263 | case CEPH_MDS_SESSION_CLOSING: return "closing"; | |
44ca18f2 | 264 | case CEPH_MDS_SESSION_RESTARTING: return "restarting"; |
2f2dc053 SW |
265 | case CEPH_MDS_SESSION_RECONNECTING: return "reconnecting"; |
266 | default: return "???"; | |
267 | } | |
268 | } | |
269 | ||
270 | static struct ceph_mds_session *get_session(struct ceph_mds_session *s) | |
271 | { | |
272 | if (atomic_inc_not_zero(&s->s_ref)) { | |
273 | dout("mdsc get_session %p %d -> %d\n", s, | |
274 | atomic_read(&s->s_ref)-1, atomic_read(&s->s_ref)); | |
275 | return s; | |
276 | } else { | |
277 | dout("mdsc get_session %p 0 -- FAIL", s); | |
278 | return NULL; | |
279 | } | |
280 | } | |
281 | ||
282 | void ceph_put_mds_session(struct ceph_mds_session *s) | |
283 | { | |
284 | dout("mdsc put_session %p %d -> %d\n", s, | |
285 | atomic_read(&s->s_ref), atomic_read(&s->s_ref)-1); | |
4e7a5dcd SW |
286 | if (atomic_dec_and_test(&s->s_ref)) { |
287 | if (s->s_authorizer) | |
288 | s->s_mdsc->client->monc.auth->ops->destroy_authorizer( | |
289 | s->s_mdsc->client->monc.auth, s->s_authorizer); | |
2f2dc053 | 290 | kfree(s); |
4e7a5dcd | 291 | } |
2f2dc053 SW |
292 | } |
293 | ||
294 | /* | |
295 | * called under mdsc->mutex | |
296 | */ | |
297 | struct ceph_mds_session *__ceph_lookup_mds_session(struct ceph_mds_client *mdsc, | |
298 | int mds) | |
299 | { | |
300 | struct ceph_mds_session *session; | |
301 | ||
302 | if (mds >= mdsc->max_sessions || mdsc->sessions[mds] == NULL) | |
303 | return NULL; | |
304 | session = mdsc->sessions[mds]; | |
305 | dout("lookup_mds_session %p %d\n", session, | |
306 | atomic_read(&session->s_ref)); | |
307 | get_session(session); | |
308 | return session; | |
309 | } | |
310 | ||
311 | static bool __have_session(struct ceph_mds_client *mdsc, int mds) | |
312 | { | |
313 | if (mds >= mdsc->max_sessions) | |
314 | return false; | |
315 | return mdsc->sessions[mds]; | |
316 | } | |
317 | ||
2600d2dd SW |
318 | static int __verify_registered_session(struct ceph_mds_client *mdsc, |
319 | struct ceph_mds_session *s) | |
320 | { | |
321 | if (s->s_mds >= mdsc->max_sessions || | |
322 | mdsc->sessions[s->s_mds] != s) | |
323 | return -ENOENT; | |
324 | return 0; | |
325 | } | |
326 | ||
2f2dc053 SW |
327 | /* |
328 | * create+register a new session for given mds. | |
329 | * called under mdsc->mutex. | |
330 | */ | |
331 | static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc, | |
332 | int mds) | |
333 | { | |
334 | struct ceph_mds_session *s; | |
335 | ||
336 | s = kzalloc(sizeof(*s), GFP_NOFS); | |
4736b009 DC |
337 | if (!s) |
338 | return ERR_PTR(-ENOMEM); | |
2f2dc053 SW |
339 | s->s_mdsc = mdsc; |
340 | s->s_mds = mds; | |
341 | s->s_state = CEPH_MDS_SESSION_NEW; | |
342 | s->s_ttl = 0; | |
343 | s->s_seq = 0; | |
344 | mutex_init(&s->s_mutex); | |
345 | ||
346 | ceph_con_init(mdsc->client->msgr, &s->s_con); | |
347 | s->s_con.private = s; | |
348 | s->s_con.ops = &mds_con_ops; | |
349 | s->s_con.peer_name.type = CEPH_ENTITY_TYPE_MDS; | |
350 | s->s_con.peer_name.num = cpu_to_le64(mds); | |
2f2dc053 SW |
351 | |
352 | spin_lock_init(&s->s_cap_lock); | |
353 | s->s_cap_gen = 0; | |
354 | s->s_cap_ttl = 0; | |
355 | s->s_renew_requested = 0; | |
356 | s->s_renew_seq = 0; | |
357 | INIT_LIST_HEAD(&s->s_caps); | |
358 | s->s_nr_caps = 0; | |
5dacf091 | 359 | s->s_trim_caps = 0; |
2f2dc053 SW |
360 | atomic_set(&s->s_ref, 1); |
361 | INIT_LIST_HEAD(&s->s_waiting); | |
362 | INIT_LIST_HEAD(&s->s_unsafe); | |
363 | s->s_num_cap_releases = 0; | |
7c1332b8 | 364 | s->s_cap_iterator = NULL; |
2f2dc053 SW |
365 | INIT_LIST_HEAD(&s->s_cap_releases); |
366 | INIT_LIST_HEAD(&s->s_cap_releases_done); | |
367 | INIT_LIST_HEAD(&s->s_cap_flushing); | |
368 | INIT_LIST_HEAD(&s->s_cap_snaps_flushing); | |
369 | ||
370 | dout("register_session mds%d\n", mds); | |
371 | if (mds >= mdsc->max_sessions) { | |
372 | int newmax = 1 << get_count_order(mds+1); | |
373 | struct ceph_mds_session **sa; | |
374 | ||
375 | dout("register_session realloc to %d\n", newmax); | |
376 | sa = kcalloc(newmax, sizeof(void *), GFP_NOFS); | |
377 | if (sa == NULL) | |
42ce56e5 | 378 | goto fail_realloc; |
2f2dc053 SW |
379 | if (mdsc->sessions) { |
380 | memcpy(sa, mdsc->sessions, | |
381 | mdsc->max_sessions * sizeof(void *)); | |
382 | kfree(mdsc->sessions); | |
383 | } | |
384 | mdsc->sessions = sa; | |
385 | mdsc->max_sessions = newmax; | |
386 | } | |
387 | mdsc->sessions[mds] = s; | |
388 | atomic_inc(&s->s_ref); /* one ref to sessions[], one to caller */ | |
42ce56e5 SW |
389 | |
390 | ceph_con_open(&s->s_con, ceph_mdsmap_get_addr(mdsc->mdsmap, mds)); | |
391 | ||
2f2dc053 | 392 | return s; |
42ce56e5 SW |
393 | |
394 | fail_realloc: | |
395 | kfree(s); | |
396 | return ERR_PTR(-ENOMEM); | |
2f2dc053 SW |
397 | } |
398 | ||
399 | /* | |
400 | * called under mdsc->mutex | |
401 | */ | |
2600d2dd | 402 | static void __unregister_session(struct ceph_mds_client *mdsc, |
42ce56e5 | 403 | struct ceph_mds_session *s) |
2f2dc053 | 404 | { |
2600d2dd SW |
405 | dout("__unregister_session mds%d %p\n", s->s_mds, s); |
406 | BUG_ON(mdsc->sessions[s->s_mds] != s); | |
42ce56e5 SW |
407 | mdsc->sessions[s->s_mds] = NULL; |
408 | ceph_con_close(&s->s_con); | |
409 | ceph_put_mds_session(s); | |
2f2dc053 SW |
410 | } |
411 | ||
412 | /* | |
413 | * drop session refs in request. | |
414 | * | |
415 | * should be last request ref, or hold mdsc->mutex | |
416 | */ | |
417 | static void put_request_session(struct ceph_mds_request *req) | |
418 | { | |
419 | if (req->r_session) { | |
420 | ceph_put_mds_session(req->r_session); | |
421 | req->r_session = NULL; | |
422 | } | |
423 | } | |
424 | ||
153c8e6b | 425 | void ceph_mdsc_release_request(struct kref *kref) |
2f2dc053 | 426 | { |
153c8e6b SW |
427 | struct ceph_mds_request *req = container_of(kref, |
428 | struct ceph_mds_request, | |
429 | r_kref); | |
430 | if (req->r_request) | |
431 | ceph_msg_put(req->r_request); | |
432 | if (req->r_reply) { | |
433 | ceph_msg_put(req->r_reply); | |
434 | destroy_reply_info(&req->r_reply_info); | |
435 | } | |
436 | if (req->r_inode) { | |
437 | ceph_put_cap_refs(ceph_inode(req->r_inode), | |
438 | CEPH_CAP_PIN); | |
439 | iput(req->r_inode); | |
440 | } | |
441 | if (req->r_locked_dir) | |
442 | ceph_put_cap_refs(ceph_inode(req->r_locked_dir), | |
443 | CEPH_CAP_PIN); | |
444 | if (req->r_target_inode) | |
445 | iput(req->r_target_inode); | |
446 | if (req->r_dentry) | |
447 | dput(req->r_dentry); | |
448 | if (req->r_old_dentry) { | |
449 | ceph_put_cap_refs( | |
450 | ceph_inode(req->r_old_dentry->d_parent->d_inode), | |
451 | CEPH_CAP_PIN); | |
452 | dput(req->r_old_dentry); | |
2f2dc053 | 453 | } |
153c8e6b SW |
454 | kfree(req->r_path1); |
455 | kfree(req->r_path2); | |
456 | put_request_session(req); | |
37151668 | 457 | ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation); |
153c8e6b | 458 | kfree(req); |
2f2dc053 SW |
459 | } |
460 | ||
461 | /* | |
462 | * lookup session, bump ref if found. | |
463 | * | |
464 | * called under mdsc->mutex. | |
465 | */ | |
466 | static struct ceph_mds_request *__lookup_request(struct ceph_mds_client *mdsc, | |
467 | u64 tid) | |
468 | { | |
469 | struct ceph_mds_request *req; | |
44ca18f2 SW |
470 | struct rb_node *n = mdsc->request_tree.rb_node; |
471 | ||
472 | while (n) { | |
473 | req = rb_entry(n, struct ceph_mds_request, r_node); | |
474 | if (tid < req->r_tid) | |
475 | n = n->rb_left; | |
476 | else if (tid > req->r_tid) | |
477 | n = n->rb_right; | |
478 | else { | |
479 | ceph_mdsc_get_request(req); | |
480 | return req; | |
481 | } | |
482 | } | |
483 | return NULL; | |
484 | } | |
485 | ||
486 | static void __insert_request(struct ceph_mds_client *mdsc, | |
487 | struct ceph_mds_request *new) | |
488 | { | |
489 | struct rb_node **p = &mdsc->request_tree.rb_node; | |
490 | struct rb_node *parent = NULL; | |
491 | struct ceph_mds_request *req = NULL; | |
492 | ||
493 | while (*p) { | |
494 | parent = *p; | |
495 | req = rb_entry(parent, struct ceph_mds_request, r_node); | |
496 | if (new->r_tid < req->r_tid) | |
497 | p = &(*p)->rb_left; | |
498 | else if (new->r_tid > req->r_tid) | |
499 | p = &(*p)->rb_right; | |
500 | else | |
501 | BUG(); | |
502 | } | |
503 | ||
504 | rb_link_node(&new->r_node, parent, p); | |
505 | rb_insert_color(&new->r_node, &mdsc->request_tree); | |
2f2dc053 SW |
506 | } |
507 | ||
508 | /* | |
509 | * Register an in-flight request, and assign a tid. Link to directory | |
510 | * are modifying (if any). | |
511 | * | |
512 | * Called under mdsc->mutex. | |
513 | */ | |
514 | static void __register_request(struct ceph_mds_client *mdsc, | |
515 | struct ceph_mds_request *req, | |
516 | struct inode *dir) | |
517 | { | |
518 | req->r_tid = ++mdsc->last_tid; | |
519 | if (req->r_num_caps) | |
37151668 YS |
520 | ceph_reserve_caps(mdsc, &req->r_caps_reservation, |
521 | req->r_num_caps); | |
2f2dc053 SW |
522 | dout("__register_request %p tid %lld\n", req, req->r_tid); |
523 | ceph_mdsc_get_request(req); | |
44ca18f2 | 524 | __insert_request(mdsc, req); |
2f2dc053 SW |
525 | |
526 | if (dir) { | |
527 | struct ceph_inode_info *ci = ceph_inode(dir); | |
528 | ||
529 | spin_lock(&ci->i_unsafe_lock); | |
530 | req->r_unsafe_dir = dir; | |
531 | list_add_tail(&req->r_unsafe_dir_item, &ci->i_unsafe_dirops); | |
532 | spin_unlock(&ci->i_unsafe_lock); | |
533 | } | |
534 | } | |
535 | ||
536 | static void __unregister_request(struct ceph_mds_client *mdsc, | |
537 | struct ceph_mds_request *req) | |
538 | { | |
539 | dout("__unregister_request %p tid %lld\n", req, req->r_tid); | |
44ca18f2 | 540 | rb_erase(&req->r_node, &mdsc->request_tree); |
80fc7314 | 541 | RB_CLEAR_NODE(&req->r_node); |
2f2dc053 SW |
542 | |
543 | if (req->r_unsafe_dir) { | |
544 | struct ceph_inode_info *ci = ceph_inode(req->r_unsafe_dir); | |
545 | ||
546 | spin_lock(&ci->i_unsafe_lock); | |
547 | list_del_init(&req->r_unsafe_dir_item); | |
548 | spin_unlock(&ci->i_unsafe_lock); | |
549 | } | |
94aa8ae1 SW |
550 | |
551 | ceph_mdsc_put_request(req); | |
2f2dc053 SW |
552 | } |
553 | ||
554 | /* | |
555 | * Choose mds to send request to next. If there is a hint set in the | |
556 | * request (e.g., due to a prior forward hint from the mds), use that. | |
557 | * Otherwise, consult frag tree and/or caps to identify the | |
558 | * appropriate mds. If all else fails, choose randomly. | |
559 | * | |
560 | * Called under mdsc->mutex. | |
561 | */ | |
562 | static int __choose_mds(struct ceph_mds_client *mdsc, | |
563 | struct ceph_mds_request *req) | |
564 | { | |
565 | struct inode *inode; | |
566 | struct ceph_inode_info *ci; | |
567 | struct ceph_cap *cap; | |
568 | int mode = req->r_direct_mode; | |
569 | int mds = -1; | |
570 | u32 hash = req->r_direct_hash; | |
571 | bool is_hash = req->r_direct_is_hash; | |
572 | ||
573 | /* | |
574 | * is there a specific mds we should try? ignore hint if we have | |
575 | * no session and the mds is not up (active or recovering). | |
576 | */ | |
577 | if (req->r_resend_mds >= 0 && | |
578 | (__have_session(mdsc, req->r_resend_mds) || | |
579 | ceph_mdsmap_get_state(mdsc->mdsmap, req->r_resend_mds) > 0)) { | |
580 | dout("choose_mds using resend_mds mds%d\n", | |
581 | req->r_resend_mds); | |
582 | return req->r_resend_mds; | |
583 | } | |
584 | ||
585 | if (mode == USE_RANDOM_MDS) | |
586 | goto random; | |
587 | ||
588 | inode = NULL; | |
589 | if (req->r_inode) { | |
590 | inode = req->r_inode; | |
591 | } else if (req->r_dentry) { | |
592 | if (req->r_dentry->d_inode) { | |
593 | inode = req->r_dentry->d_inode; | |
594 | } else { | |
595 | inode = req->r_dentry->d_parent->d_inode; | |
596 | hash = req->r_dentry->d_name.hash; | |
597 | is_hash = true; | |
598 | } | |
599 | } | |
600 | dout("__choose_mds %p is_hash=%d (%d) mode %d\n", inode, (int)is_hash, | |
601 | (int)hash, mode); | |
602 | if (!inode) | |
603 | goto random; | |
604 | ci = ceph_inode(inode); | |
605 | ||
606 | if (is_hash && S_ISDIR(inode->i_mode)) { | |
607 | struct ceph_inode_frag frag; | |
608 | int found; | |
609 | ||
610 | ceph_choose_frag(ci, hash, &frag, &found); | |
611 | if (found) { | |
612 | if (mode == USE_ANY_MDS && frag.ndist > 0) { | |
613 | u8 r; | |
614 | ||
615 | /* choose a random replica */ | |
616 | get_random_bytes(&r, 1); | |
617 | r %= frag.ndist; | |
618 | mds = frag.dist[r]; | |
619 | dout("choose_mds %p %llx.%llx " | |
620 | "frag %u mds%d (%d/%d)\n", | |
621 | inode, ceph_vinop(inode), | |
622 | frag.frag, frag.mds, | |
623 | (int)r, frag.ndist); | |
624 | return mds; | |
625 | } | |
626 | ||
627 | /* since this file/dir wasn't known to be | |
628 | * replicated, then we want to look for the | |
629 | * authoritative mds. */ | |
630 | mode = USE_AUTH_MDS; | |
631 | if (frag.mds >= 0) { | |
632 | /* choose auth mds */ | |
633 | mds = frag.mds; | |
634 | dout("choose_mds %p %llx.%llx " | |
635 | "frag %u mds%d (auth)\n", | |
636 | inode, ceph_vinop(inode), frag.frag, mds); | |
637 | return mds; | |
638 | } | |
639 | } | |
640 | } | |
641 | ||
642 | spin_lock(&inode->i_lock); | |
643 | cap = NULL; | |
644 | if (mode == USE_AUTH_MDS) | |
645 | cap = ci->i_auth_cap; | |
646 | if (!cap && !RB_EMPTY_ROOT(&ci->i_caps)) | |
647 | cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node); | |
648 | if (!cap) { | |
649 | spin_unlock(&inode->i_lock); | |
650 | goto random; | |
651 | } | |
652 | mds = cap->session->s_mds; | |
653 | dout("choose_mds %p %llx.%llx mds%d (%scap %p)\n", | |
654 | inode, ceph_vinop(inode), mds, | |
655 | cap == ci->i_auth_cap ? "auth " : "", cap); | |
656 | spin_unlock(&inode->i_lock); | |
657 | return mds; | |
658 | ||
659 | random: | |
660 | mds = ceph_mdsmap_get_random_mds(mdsc->mdsmap); | |
661 | dout("choose_mds chose random mds%d\n", mds); | |
662 | return mds; | |
663 | } | |
664 | ||
665 | ||
666 | /* | |
667 | * session messages | |
668 | */ | |
669 | static struct ceph_msg *create_session_msg(u32 op, u64 seq) | |
670 | { | |
671 | struct ceph_msg *msg; | |
672 | struct ceph_mds_session_head *h; | |
673 | ||
34d23762 | 674 | msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS); |
a79832f2 | 675 | if (!msg) { |
2f2dc053 | 676 | pr_err("create_session_msg ENOMEM creating msg\n"); |
a79832f2 | 677 | return NULL; |
2f2dc053 SW |
678 | } |
679 | h = msg->front.iov_base; | |
680 | h->op = cpu_to_le32(op); | |
681 | h->seq = cpu_to_le64(seq); | |
682 | return msg; | |
683 | } | |
684 | ||
685 | /* | |
686 | * send session open request. | |
687 | * | |
688 | * called under mdsc->mutex | |
689 | */ | |
690 | static int __open_session(struct ceph_mds_client *mdsc, | |
691 | struct ceph_mds_session *session) | |
692 | { | |
693 | struct ceph_msg *msg; | |
694 | int mstate; | |
695 | int mds = session->s_mds; | |
2f2dc053 SW |
696 | |
697 | /* wait for mds to go active? */ | |
698 | mstate = ceph_mdsmap_get_state(mdsc->mdsmap, mds); | |
699 | dout("open_session to mds%d (%s)\n", mds, | |
700 | ceph_mds_state_name(mstate)); | |
701 | session->s_state = CEPH_MDS_SESSION_OPENING; | |
702 | session->s_renew_requested = jiffies; | |
703 | ||
704 | /* send connect message */ | |
705 | msg = create_session_msg(CEPH_SESSION_REQUEST_OPEN, session->s_seq); | |
a79832f2 SW |
706 | if (!msg) |
707 | return -ENOMEM; | |
2f2dc053 | 708 | ceph_con_send(&session->s_con, msg); |
2f2dc053 SW |
709 | return 0; |
710 | } | |
711 | ||
ed0552a1 SW |
712 | /* |
713 | * open sessions for any export targets for the given mds | |
714 | * | |
715 | * called under mdsc->mutex | |
716 | */ | |
717 | static void __open_export_target_sessions(struct ceph_mds_client *mdsc, | |
718 | struct ceph_mds_session *session) | |
719 | { | |
720 | struct ceph_mds_info *mi; | |
721 | struct ceph_mds_session *ts; | |
722 | int i, mds = session->s_mds; | |
723 | int target; | |
724 | ||
725 | if (mds >= mdsc->mdsmap->m_max_mds) | |
726 | return; | |
727 | mi = &mdsc->mdsmap->m_info[mds]; | |
728 | dout("open_export_target_sessions for mds%d (%d targets)\n", | |
729 | session->s_mds, mi->num_export_targets); | |
730 | ||
731 | for (i = 0; i < mi->num_export_targets; i++) { | |
732 | target = mi->export_targets[i]; | |
733 | ts = __ceph_lookup_mds_session(mdsc, target); | |
734 | if (!ts) { | |
735 | ts = register_session(mdsc, target); | |
736 | if (IS_ERR(ts)) | |
737 | return; | |
738 | } | |
739 | if (session->s_state == CEPH_MDS_SESSION_NEW || | |
740 | session->s_state == CEPH_MDS_SESSION_CLOSING) | |
741 | __open_session(mdsc, session); | |
742 | else | |
743 | dout(" mds%d target mds%d %p is %s\n", session->s_mds, | |
744 | i, ts, session_state_name(ts->s_state)); | |
745 | ceph_put_mds_session(ts); | |
746 | } | |
747 | } | |
748 | ||
154f42c2 SW |
749 | void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc, |
750 | struct ceph_mds_session *session) | |
751 | { | |
752 | mutex_lock(&mdsc->mutex); | |
753 | __open_export_target_sessions(mdsc, session); | |
754 | mutex_unlock(&mdsc->mutex); | |
755 | } | |
756 | ||
2f2dc053 SW |
757 | /* |
758 | * session caps | |
759 | */ | |
760 | ||
761 | /* | |
762 | * Free preallocated cap messages assigned to this session | |
763 | */ | |
764 | static void cleanup_cap_releases(struct ceph_mds_session *session) | |
765 | { | |
766 | struct ceph_msg *msg; | |
767 | ||
768 | spin_lock(&session->s_cap_lock); | |
769 | while (!list_empty(&session->s_cap_releases)) { | |
770 | msg = list_first_entry(&session->s_cap_releases, | |
771 | struct ceph_msg, list_head); | |
772 | list_del_init(&msg->list_head); | |
773 | ceph_msg_put(msg); | |
774 | } | |
775 | while (!list_empty(&session->s_cap_releases_done)) { | |
776 | msg = list_first_entry(&session->s_cap_releases_done, | |
777 | struct ceph_msg, list_head); | |
778 | list_del_init(&msg->list_head); | |
779 | ceph_msg_put(msg); | |
780 | } | |
781 | spin_unlock(&session->s_cap_lock); | |
782 | } | |
783 | ||
784 | /* | |
f818a736 SW |
785 | * Helper to safely iterate over all caps associated with a session, with |
786 | * special care taken to handle a racing __ceph_remove_cap(). | |
2f2dc053 | 787 | * |
f818a736 | 788 | * Caller must hold session s_mutex. |
2f2dc053 SW |
789 | */ |
790 | static int iterate_session_caps(struct ceph_mds_session *session, | |
791 | int (*cb)(struct inode *, struct ceph_cap *, | |
792 | void *), void *arg) | |
793 | { | |
7c1332b8 SW |
794 | struct list_head *p; |
795 | struct ceph_cap *cap; | |
796 | struct inode *inode, *last_inode = NULL; | |
797 | struct ceph_cap *old_cap = NULL; | |
2f2dc053 SW |
798 | int ret; |
799 | ||
800 | dout("iterate_session_caps %p mds%d\n", session, session->s_mds); | |
801 | spin_lock(&session->s_cap_lock); | |
7c1332b8 SW |
802 | p = session->s_caps.next; |
803 | while (p != &session->s_caps) { | |
804 | cap = list_entry(p, struct ceph_cap, session_caps); | |
2f2dc053 | 805 | inode = igrab(&cap->ci->vfs_inode); |
7c1332b8 SW |
806 | if (!inode) { |
807 | p = p->next; | |
2f2dc053 | 808 | continue; |
7c1332b8 SW |
809 | } |
810 | session->s_cap_iterator = cap; | |
2f2dc053 | 811 | spin_unlock(&session->s_cap_lock); |
7c1332b8 SW |
812 | |
813 | if (last_inode) { | |
814 | iput(last_inode); | |
815 | last_inode = NULL; | |
816 | } | |
817 | if (old_cap) { | |
37151668 | 818 | ceph_put_cap(session->s_mdsc, old_cap); |
7c1332b8 SW |
819 | old_cap = NULL; |
820 | } | |
821 | ||
2f2dc053 | 822 | ret = cb(inode, cap, arg); |
7c1332b8 SW |
823 | last_inode = inode; |
824 | ||
2f2dc053 | 825 | spin_lock(&session->s_cap_lock); |
7c1332b8 SW |
826 | p = p->next; |
827 | if (cap->ci == NULL) { | |
828 | dout("iterate_session_caps finishing cap %p removal\n", | |
829 | cap); | |
830 | BUG_ON(cap->session != session); | |
831 | list_del_init(&cap->session_caps); | |
832 | session->s_nr_caps--; | |
833 | cap->session = NULL; | |
834 | old_cap = cap; /* put_cap it w/o locks held */ | |
835 | } | |
5dacf091 SW |
836 | if (ret < 0) |
837 | goto out; | |
2f2dc053 | 838 | } |
5dacf091 SW |
839 | ret = 0; |
840 | out: | |
7c1332b8 | 841 | session->s_cap_iterator = NULL; |
2f2dc053 | 842 | spin_unlock(&session->s_cap_lock); |
7c1332b8 SW |
843 | |
844 | if (last_inode) | |
845 | iput(last_inode); | |
846 | if (old_cap) | |
37151668 | 847 | ceph_put_cap(session->s_mdsc, old_cap); |
7c1332b8 | 848 | |
5dacf091 | 849 | return ret; |
2f2dc053 SW |
850 | } |
851 | ||
852 | static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, | |
6c99f254 | 853 | void *arg) |
2f2dc053 SW |
854 | { |
855 | struct ceph_inode_info *ci = ceph_inode(inode); | |
6c99f254 SW |
856 | int drop = 0; |
857 | ||
2f2dc053 SW |
858 | dout("removing cap %p, ci is %p, inode is %p\n", |
859 | cap, ci, &ci->vfs_inode); | |
6c99f254 SW |
860 | spin_lock(&inode->i_lock); |
861 | __ceph_remove_cap(cap); | |
862 | if (!__ceph_is_any_real_caps(ci)) { | |
863 | struct ceph_mds_client *mdsc = | |
864 | &ceph_sb_to_client(inode->i_sb)->mdsc; | |
865 | ||
866 | spin_lock(&mdsc->cap_dirty_lock); | |
867 | if (!list_empty(&ci->i_dirty_item)) { | |
868 | pr_info(" dropping dirty %s state for %p %lld\n", | |
869 | ceph_cap_string(ci->i_dirty_caps), | |
870 | inode, ceph_ino(inode)); | |
871 | ci->i_dirty_caps = 0; | |
872 | list_del_init(&ci->i_dirty_item); | |
873 | drop = 1; | |
874 | } | |
875 | if (!list_empty(&ci->i_flushing_item)) { | |
876 | pr_info(" dropping dirty+flushing %s state for %p %lld\n", | |
877 | ceph_cap_string(ci->i_flushing_caps), | |
878 | inode, ceph_ino(inode)); | |
879 | ci->i_flushing_caps = 0; | |
880 | list_del_init(&ci->i_flushing_item); | |
881 | mdsc->num_cap_flushing--; | |
882 | drop = 1; | |
883 | } | |
884 | if (drop && ci->i_wrbuffer_ref) { | |
885 | pr_info(" dropping dirty data for %p %lld\n", | |
886 | inode, ceph_ino(inode)); | |
887 | ci->i_wrbuffer_ref = 0; | |
888 | ci->i_wrbuffer_ref_head = 0; | |
889 | drop++; | |
890 | } | |
891 | spin_unlock(&mdsc->cap_dirty_lock); | |
892 | } | |
893 | spin_unlock(&inode->i_lock); | |
894 | while (drop--) | |
895 | iput(inode); | |
2f2dc053 SW |
896 | return 0; |
897 | } | |
898 | ||
899 | /* | |
900 | * caller must hold session s_mutex | |
901 | */ | |
902 | static void remove_session_caps(struct ceph_mds_session *session) | |
903 | { | |
904 | dout("remove_session_caps on %p\n", session); | |
905 | iterate_session_caps(session, remove_session_caps_cb, NULL); | |
906 | BUG_ON(session->s_nr_caps > 0); | |
6c99f254 | 907 | BUG_ON(!list_empty(&session->s_cap_flushing)); |
2f2dc053 SW |
908 | cleanup_cap_releases(session); |
909 | } | |
910 | ||
911 | /* | |
912 | * wake up any threads waiting on this session's caps. if the cap is | |
913 | * old (didn't get renewed on the client reconnect), remove it now. | |
914 | * | |
915 | * caller must hold s_mutex. | |
916 | */ | |
917 | static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap, | |
918 | void *arg) | |
919 | { | |
0dc2570f SW |
920 | struct ceph_inode_info *ci = ceph_inode(inode); |
921 | ||
03066f23 | 922 | wake_up_all(&ci->i_cap_wq); |
0dc2570f SW |
923 | if (arg) { |
924 | spin_lock(&inode->i_lock); | |
925 | ci->i_wanted_max_size = 0; | |
926 | ci->i_requested_max_size = 0; | |
927 | spin_unlock(&inode->i_lock); | |
928 | } | |
2f2dc053 SW |
929 | return 0; |
930 | } | |
931 | ||
0dc2570f SW |
932 | static void wake_up_session_caps(struct ceph_mds_session *session, |
933 | int reconnect) | |
2f2dc053 SW |
934 | { |
935 | dout("wake_up_session_caps %p mds%d\n", session, session->s_mds); | |
0dc2570f SW |
936 | iterate_session_caps(session, wake_up_session_cb, |
937 | (void *)(unsigned long)reconnect); | |
2f2dc053 SW |
938 | } |
939 | ||
940 | /* | |
941 | * Send periodic message to MDS renewing all currently held caps. The | |
942 | * ack will reset the expiration for all caps from this session. | |
943 | * | |
944 | * caller holds s_mutex | |
945 | */ | |
946 | static int send_renew_caps(struct ceph_mds_client *mdsc, | |
947 | struct ceph_mds_session *session) | |
948 | { | |
949 | struct ceph_msg *msg; | |
950 | int state; | |
951 | ||
952 | if (time_after_eq(jiffies, session->s_cap_ttl) && | |
953 | time_after_eq(session->s_cap_ttl, session->s_renew_requested)) | |
954 | pr_info("mds%d caps stale\n", session->s_mds); | |
e4cb4cb8 | 955 | session->s_renew_requested = jiffies; |
2f2dc053 SW |
956 | |
957 | /* do not try to renew caps until a recovering mds has reconnected | |
958 | * with its clients. */ | |
959 | state = ceph_mdsmap_get_state(mdsc->mdsmap, session->s_mds); | |
960 | if (state < CEPH_MDS_STATE_RECONNECT) { | |
961 | dout("send_renew_caps ignoring mds%d (%s)\n", | |
962 | session->s_mds, ceph_mds_state_name(state)); | |
963 | return 0; | |
964 | } | |
965 | ||
966 | dout("send_renew_caps to mds%d (%s)\n", session->s_mds, | |
967 | ceph_mds_state_name(state)); | |
2f2dc053 SW |
968 | msg = create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS, |
969 | ++session->s_renew_seq); | |
a79832f2 SW |
970 | if (!msg) |
971 | return -ENOMEM; | |
2f2dc053 SW |
972 | ceph_con_send(&session->s_con, msg); |
973 | return 0; | |
974 | } | |
975 | ||
976 | /* | |
977 | * Note new cap ttl, and any transition from stale -> not stale (fresh?). | |
0dc2570f SW |
978 | * |
979 | * Called under session->s_mutex | |
2f2dc053 SW |
980 | */ |
981 | static void renewed_caps(struct ceph_mds_client *mdsc, | |
982 | struct ceph_mds_session *session, int is_renew) | |
983 | { | |
984 | int was_stale; | |
985 | int wake = 0; | |
986 | ||
987 | spin_lock(&session->s_cap_lock); | |
988 | was_stale = is_renew && (session->s_cap_ttl == 0 || | |
989 | time_after_eq(jiffies, session->s_cap_ttl)); | |
990 | ||
991 | session->s_cap_ttl = session->s_renew_requested + | |
992 | mdsc->mdsmap->m_session_timeout*HZ; | |
993 | ||
994 | if (was_stale) { | |
995 | if (time_before(jiffies, session->s_cap_ttl)) { | |
996 | pr_info("mds%d caps renewed\n", session->s_mds); | |
997 | wake = 1; | |
998 | } else { | |
999 | pr_info("mds%d caps still stale\n", session->s_mds); | |
1000 | } | |
1001 | } | |
1002 | dout("renewed_caps mds%d ttl now %lu, was %s, now %s\n", | |
1003 | session->s_mds, session->s_cap_ttl, was_stale ? "stale" : "fresh", | |
1004 | time_before(jiffies, session->s_cap_ttl) ? "stale" : "fresh"); | |
1005 | spin_unlock(&session->s_cap_lock); | |
1006 | ||
1007 | if (wake) | |
0dc2570f | 1008 | wake_up_session_caps(session, 0); |
2f2dc053 SW |
1009 | } |
1010 | ||
1011 | /* | |
1012 | * send a session close request | |
1013 | */ | |
1014 | static int request_close_session(struct ceph_mds_client *mdsc, | |
1015 | struct ceph_mds_session *session) | |
1016 | { | |
1017 | struct ceph_msg *msg; | |
2f2dc053 SW |
1018 | |
1019 | dout("request_close_session mds%d state %s seq %lld\n", | |
1020 | session->s_mds, session_state_name(session->s_state), | |
1021 | session->s_seq); | |
1022 | msg = create_session_msg(CEPH_SESSION_REQUEST_CLOSE, session->s_seq); | |
a79832f2 SW |
1023 | if (!msg) |
1024 | return -ENOMEM; | |
1025 | ceph_con_send(&session->s_con, msg); | |
1026 | return 0; | |
2f2dc053 SW |
1027 | } |
1028 | ||
1029 | /* | |
1030 | * Called with s_mutex held. | |
1031 | */ | |
1032 | static int __close_session(struct ceph_mds_client *mdsc, | |
1033 | struct ceph_mds_session *session) | |
1034 | { | |
1035 | if (session->s_state >= CEPH_MDS_SESSION_CLOSING) | |
1036 | return 0; | |
1037 | session->s_state = CEPH_MDS_SESSION_CLOSING; | |
1038 | return request_close_session(mdsc, session); | |
1039 | } | |
1040 | ||
1041 | /* | |
1042 | * Trim old(er) caps. | |
1043 | * | |
1044 | * Because we can't cache an inode without one or more caps, we do | |
1045 | * this indirectly: if a cap is unused, we prune its aliases, at which | |
1046 | * point the inode will hopefully get dropped to. | |
1047 | * | |
1048 | * Yes, this is a bit sloppy. Our only real goal here is to respond to | |
1049 | * memory pressure from the MDS, though, so it needn't be perfect. | |
1050 | */ | |
1051 | static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) | |
1052 | { | |
1053 | struct ceph_mds_session *session = arg; | |
1054 | struct ceph_inode_info *ci = ceph_inode(inode); | |
1055 | int used, oissued, mine; | |
1056 | ||
1057 | if (session->s_trim_caps <= 0) | |
1058 | return -1; | |
1059 | ||
1060 | spin_lock(&inode->i_lock); | |
1061 | mine = cap->issued | cap->implemented; | |
1062 | used = __ceph_caps_used(ci); | |
1063 | oissued = __ceph_caps_issued_other(ci, cap); | |
1064 | ||
1065 | dout("trim_caps_cb %p cap %p mine %s oissued %s used %s\n", | |
1066 | inode, cap, ceph_cap_string(mine), ceph_cap_string(oissued), | |
1067 | ceph_cap_string(used)); | |
1068 | if (ci->i_dirty_caps) | |
1069 | goto out; /* dirty caps */ | |
1070 | if ((used & ~oissued) & mine) | |
1071 | goto out; /* we need these caps */ | |
1072 | ||
1073 | session->s_trim_caps--; | |
1074 | if (oissued) { | |
1075 | /* we aren't the only cap.. just remove us */ | |
7c1332b8 | 1076 | __ceph_remove_cap(cap); |
2f2dc053 SW |
1077 | } else { |
1078 | /* try to drop referring dentries */ | |
1079 | spin_unlock(&inode->i_lock); | |
1080 | d_prune_aliases(inode); | |
1081 | dout("trim_caps_cb %p cap %p pruned, count now %d\n", | |
1082 | inode, cap, atomic_read(&inode->i_count)); | |
1083 | return 0; | |
1084 | } | |
1085 | ||
1086 | out: | |
1087 | spin_unlock(&inode->i_lock); | |
1088 | return 0; | |
1089 | } | |
1090 | ||
1091 | /* | |
1092 | * Trim session cap count down to some max number. | |
1093 | */ | |
1094 | static int trim_caps(struct ceph_mds_client *mdsc, | |
1095 | struct ceph_mds_session *session, | |
1096 | int max_caps) | |
1097 | { | |
1098 | int trim_caps = session->s_nr_caps - max_caps; | |
1099 | ||
1100 | dout("trim_caps mds%d start: %d / %d, trim %d\n", | |
1101 | session->s_mds, session->s_nr_caps, max_caps, trim_caps); | |
1102 | if (trim_caps > 0) { | |
1103 | session->s_trim_caps = trim_caps; | |
1104 | iterate_session_caps(session, trim_caps_cb, session); | |
1105 | dout("trim_caps mds%d done: %d / %d, trimmed %d\n", | |
1106 | session->s_mds, session->s_nr_caps, max_caps, | |
1107 | trim_caps - session->s_trim_caps); | |
5dacf091 | 1108 | session->s_trim_caps = 0; |
2f2dc053 SW |
1109 | } |
1110 | return 0; | |
1111 | } | |
1112 | ||
1113 | /* | |
1114 | * Allocate cap_release messages. If there is a partially full message | |
1115 | * in the queue, try to allocate enough to cover it's remainder, so that | |
1116 | * we can send it immediately. | |
1117 | * | |
1118 | * Called under s_mutex. | |
1119 | */ | |
2b2300d6 | 1120 | int ceph_add_cap_releases(struct ceph_mds_client *mdsc, |
ee6b272b | 1121 | struct ceph_mds_session *session) |
2f2dc053 | 1122 | { |
38e8883e | 1123 | struct ceph_msg *msg, *partial = NULL; |
2f2dc053 SW |
1124 | struct ceph_mds_cap_release *head; |
1125 | int err = -ENOMEM; | |
ee6b272b | 1126 | int extra = mdsc->client->mount_args->cap_release_safety; |
38e8883e | 1127 | int num; |
2f2dc053 | 1128 | |
38e8883e SW |
1129 | dout("add_cap_releases %p mds%d extra %d\n", session, session->s_mds, |
1130 | extra); | |
2f2dc053 SW |
1131 | |
1132 | spin_lock(&session->s_cap_lock); | |
1133 | ||
1134 | if (!list_empty(&session->s_cap_releases)) { | |
1135 | msg = list_first_entry(&session->s_cap_releases, | |
1136 | struct ceph_msg, | |
1137 | list_head); | |
1138 | head = msg->front.iov_base; | |
38e8883e SW |
1139 | num = le32_to_cpu(head->num); |
1140 | if (num) { | |
1141 | dout(" partial %p with (%d/%d)\n", msg, num, | |
1142 | (int)CEPH_CAPS_PER_RELEASE); | |
1143 | extra += CEPH_CAPS_PER_RELEASE - num; | |
1144 | partial = msg; | |
1145 | } | |
2f2dc053 | 1146 | } |
2f2dc053 SW |
1147 | while (session->s_num_cap_releases < session->s_nr_caps + extra) { |
1148 | spin_unlock(&session->s_cap_lock); | |
34d23762 YS |
1149 | msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPRELEASE, PAGE_CACHE_SIZE, |
1150 | GFP_NOFS); | |
2f2dc053 SW |
1151 | if (!msg) |
1152 | goto out_unlocked; | |
1153 | dout("add_cap_releases %p msg %p now %d\n", session, msg, | |
1154 | (int)msg->front.iov_len); | |
1155 | head = msg->front.iov_base; | |
1156 | head->num = cpu_to_le32(0); | |
1157 | msg->front.iov_len = sizeof(*head); | |
1158 | spin_lock(&session->s_cap_lock); | |
1159 | list_add(&msg->list_head, &session->s_cap_releases); | |
1160 | session->s_num_cap_releases += CEPH_CAPS_PER_RELEASE; | |
1161 | } | |
1162 | ||
38e8883e SW |
1163 | if (partial) { |
1164 | head = partial->front.iov_base; | |
1165 | num = le32_to_cpu(head->num); | |
1166 | dout(" queueing partial %p with %d/%d\n", partial, num, | |
1167 | (int)CEPH_CAPS_PER_RELEASE); | |
1168 | list_move_tail(&partial->list_head, | |
1169 | &session->s_cap_releases_done); | |
1170 | session->s_num_cap_releases -= CEPH_CAPS_PER_RELEASE - num; | |
2f2dc053 SW |
1171 | } |
1172 | err = 0; | |
1173 | spin_unlock(&session->s_cap_lock); | |
1174 | out_unlocked: | |
1175 | return err; | |
1176 | } | |
1177 | ||
1178 | /* | |
1179 | * flush all dirty inode data to disk. | |
1180 | * | |
1181 | * returns true if we've flushed through want_flush_seq | |
1182 | */ | |
1183 | static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq) | |
1184 | { | |
1185 | int mds, ret = 1; | |
1186 | ||
1187 | dout("check_cap_flush want %lld\n", want_flush_seq); | |
1188 | mutex_lock(&mdsc->mutex); | |
1189 | for (mds = 0; ret && mds < mdsc->max_sessions; mds++) { | |
1190 | struct ceph_mds_session *session = mdsc->sessions[mds]; | |
1191 | ||
1192 | if (!session) | |
1193 | continue; | |
1194 | get_session(session); | |
1195 | mutex_unlock(&mdsc->mutex); | |
1196 | ||
1197 | mutex_lock(&session->s_mutex); | |
1198 | if (!list_empty(&session->s_cap_flushing)) { | |
1199 | struct ceph_inode_info *ci = | |
1200 | list_entry(session->s_cap_flushing.next, | |
1201 | struct ceph_inode_info, | |
1202 | i_flushing_item); | |
1203 | struct inode *inode = &ci->vfs_inode; | |
1204 | ||
1205 | spin_lock(&inode->i_lock); | |
1206 | if (ci->i_cap_flush_seq <= want_flush_seq) { | |
1207 | dout("check_cap_flush still flushing %p " | |
1208 | "seq %lld <= %lld to mds%d\n", inode, | |
1209 | ci->i_cap_flush_seq, want_flush_seq, | |
1210 | session->s_mds); | |
1211 | ret = 0; | |
1212 | } | |
1213 | spin_unlock(&inode->i_lock); | |
1214 | } | |
1215 | mutex_unlock(&session->s_mutex); | |
1216 | ceph_put_mds_session(session); | |
1217 | ||
1218 | if (!ret) | |
1219 | return ret; | |
1220 | mutex_lock(&mdsc->mutex); | |
1221 | } | |
1222 | ||
1223 | mutex_unlock(&mdsc->mutex); | |
1224 | dout("check_cap_flush ok, flushed thru %lld\n", want_flush_seq); | |
1225 | return ret; | |
1226 | } | |
1227 | ||
1228 | /* | |
1229 | * called under s_mutex | |
1230 | */ | |
3d7ded4d SW |
1231 | void ceph_send_cap_releases(struct ceph_mds_client *mdsc, |
1232 | struct ceph_mds_session *session) | |
2f2dc053 SW |
1233 | { |
1234 | struct ceph_msg *msg; | |
1235 | ||
1236 | dout("send_cap_releases mds%d\n", session->s_mds); | |
0f8605f2 SW |
1237 | spin_lock(&session->s_cap_lock); |
1238 | while (!list_empty(&session->s_cap_releases_done)) { | |
2f2dc053 SW |
1239 | msg = list_first_entry(&session->s_cap_releases_done, |
1240 | struct ceph_msg, list_head); | |
1241 | list_del_init(&msg->list_head); | |
1242 | spin_unlock(&session->s_cap_lock); | |
1243 | msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); | |
1244 | dout("send_cap_releases mds%d %p\n", session->s_mds, msg); | |
1245 | ceph_con_send(&session->s_con, msg); | |
0f8605f2 | 1246 | spin_lock(&session->s_cap_lock); |
2f2dc053 SW |
1247 | } |
1248 | spin_unlock(&session->s_cap_lock); | |
1249 | } | |
1250 | ||
e01a5946 SW |
1251 | static void discard_cap_releases(struct ceph_mds_client *mdsc, |
1252 | struct ceph_mds_session *session) | |
1253 | { | |
1254 | struct ceph_msg *msg; | |
1255 | struct ceph_mds_cap_release *head; | |
1256 | unsigned num; | |
1257 | ||
1258 | dout("discard_cap_releases mds%d\n", session->s_mds); | |
1259 | spin_lock(&session->s_cap_lock); | |
1260 | ||
1261 | /* zero out the in-progress message */ | |
1262 | msg = list_first_entry(&session->s_cap_releases, | |
1263 | struct ceph_msg, list_head); | |
1264 | head = msg->front.iov_base; | |
1265 | num = le32_to_cpu(head->num); | |
1266 | dout("discard_cap_releases mds%d %p %u\n", session->s_mds, msg, num); | |
1267 | head->num = cpu_to_le32(0); | |
1268 | session->s_num_cap_releases += num; | |
1269 | ||
1270 | /* requeue completed messages */ | |
1271 | while (!list_empty(&session->s_cap_releases_done)) { | |
1272 | msg = list_first_entry(&session->s_cap_releases_done, | |
1273 | struct ceph_msg, list_head); | |
1274 | list_del_init(&msg->list_head); | |
1275 | ||
1276 | head = msg->front.iov_base; | |
1277 | num = le32_to_cpu(head->num); | |
1278 | dout("discard_cap_releases mds%d %p %u\n", session->s_mds, msg, | |
1279 | num); | |
1280 | session->s_num_cap_releases += num; | |
1281 | head->num = cpu_to_le32(0); | |
1282 | msg->front.iov_len = sizeof(*head); | |
1283 | list_add(&msg->list_head, &session->s_cap_releases); | |
1284 | } | |
1285 | ||
1286 | spin_unlock(&session->s_cap_lock); | |
1287 | } | |
1288 | ||
2f2dc053 SW |
1289 | /* |
1290 | * requests | |
1291 | */ | |
1292 | ||
1293 | /* | |
1294 | * Create an mds request. | |
1295 | */ | |
1296 | struct ceph_mds_request * | |
1297 | ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode) | |
1298 | { | |
1299 | struct ceph_mds_request *req = kzalloc(sizeof(*req), GFP_NOFS); | |
1300 | ||
1301 | if (!req) | |
1302 | return ERR_PTR(-ENOMEM); | |
1303 | ||
b4556396 | 1304 | mutex_init(&req->r_fill_mutex); |
37151668 | 1305 | req->r_mdsc = mdsc; |
2f2dc053 SW |
1306 | req->r_started = jiffies; |
1307 | req->r_resend_mds = -1; | |
1308 | INIT_LIST_HEAD(&req->r_unsafe_dir_item); | |
1309 | req->r_fmode = -1; | |
153c8e6b | 1310 | kref_init(&req->r_kref); |
2f2dc053 SW |
1311 | INIT_LIST_HEAD(&req->r_wait); |
1312 | init_completion(&req->r_completion); | |
1313 | init_completion(&req->r_safe_completion); | |
1314 | INIT_LIST_HEAD(&req->r_unsafe_item); | |
1315 | ||
1316 | req->r_op = op; | |
1317 | req->r_direct_mode = mode; | |
1318 | return req; | |
1319 | } | |
1320 | ||
1321 | /* | |
44ca18f2 | 1322 | * return oldest (lowest) request, tid in request tree, 0 if none. |
2f2dc053 SW |
1323 | * |
1324 | * called under mdsc->mutex. | |
1325 | */ | |
44ca18f2 SW |
1326 | static struct ceph_mds_request *__get_oldest_req(struct ceph_mds_client *mdsc) |
1327 | { | |
1328 | if (RB_EMPTY_ROOT(&mdsc->request_tree)) | |
1329 | return NULL; | |
1330 | return rb_entry(rb_first(&mdsc->request_tree), | |
1331 | struct ceph_mds_request, r_node); | |
1332 | } | |
1333 | ||
2f2dc053 SW |
1334 | static u64 __get_oldest_tid(struct ceph_mds_client *mdsc) |
1335 | { | |
44ca18f2 SW |
1336 | struct ceph_mds_request *req = __get_oldest_req(mdsc); |
1337 | ||
1338 | if (req) | |
1339 | return req->r_tid; | |
1340 | return 0; | |
2f2dc053 SW |
1341 | } |
1342 | ||
1343 | /* | |
1344 | * Build a dentry's path. Allocate on heap; caller must kfree. Based | |
1345 | * on build_path_from_dentry in fs/cifs/dir.c. | |
1346 | * | |
1347 | * If @stop_on_nosnap, generate path relative to the first non-snapped | |
1348 | * inode. | |
1349 | * | |
1350 | * Encode hidden .snap dirs as a double /, i.e. | |
1351 | * foo/.snap/bar -> foo//bar | |
1352 | */ | |
1353 | char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *base, | |
1354 | int stop_on_nosnap) | |
1355 | { | |
1356 | struct dentry *temp; | |
1357 | char *path; | |
1358 | int len, pos; | |
1359 | ||
1360 | if (dentry == NULL) | |
1361 | return ERR_PTR(-EINVAL); | |
1362 | ||
1363 | retry: | |
1364 | len = 0; | |
1365 | for (temp = dentry; !IS_ROOT(temp);) { | |
1366 | struct inode *inode = temp->d_inode; | |
1367 | if (inode && ceph_snap(inode) == CEPH_SNAPDIR) | |
1368 | len++; /* slash only */ | |
1369 | else if (stop_on_nosnap && inode && | |
1370 | ceph_snap(inode) == CEPH_NOSNAP) | |
1371 | break; | |
1372 | else | |
1373 | len += 1 + temp->d_name.len; | |
1374 | temp = temp->d_parent; | |
1375 | if (temp == NULL) { | |
6c99f254 | 1376 | pr_err("build_path corrupt dentry %p\n", dentry); |
2f2dc053 SW |
1377 | return ERR_PTR(-EINVAL); |
1378 | } | |
1379 | } | |
1380 | if (len) | |
1381 | len--; /* no leading '/' */ | |
1382 | ||
1383 | path = kmalloc(len+1, GFP_NOFS); | |
1384 | if (path == NULL) | |
1385 | return ERR_PTR(-ENOMEM); | |
1386 | pos = len; | |
1387 | path[pos] = 0; /* trailing null */ | |
1388 | for (temp = dentry; !IS_ROOT(temp) && pos != 0; ) { | |
1389 | struct inode *inode = temp->d_inode; | |
1390 | ||
1391 | if (inode && ceph_snap(inode) == CEPH_SNAPDIR) { | |
104648ad | 1392 | dout("build_path path+%d: %p SNAPDIR\n", |
2f2dc053 SW |
1393 | pos, temp); |
1394 | } else if (stop_on_nosnap && inode && | |
1395 | ceph_snap(inode) == CEPH_NOSNAP) { | |
1396 | break; | |
1397 | } else { | |
1398 | pos -= temp->d_name.len; | |
1399 | if (pos < 0) | |
1400 | break; | |
1401 | strncpy(path + pos, temp->d_name.name, | |
1402 | temp->d_name.len); | |
2f2dc053 SW |
1403 | } |
1404 | if (pos) | |
1405 | path[--pos] = '/'; | |
1406 | temp = temp->d_parent; | |
1407 | if (temp == NULL) { | |
104648ad | 1408 | pr_err("build_path corrupt dentry\n"); |
2f2dc053 SW |
1409 | kfree(path); |
1410 | return ERR_PTR(-EINVAL); | |
1411 | } | |
1412 | } | |
1413 | if (pos != 0) { | |
104648ad | 1414 | pr_err("build_path did not end path lookup where " |
2f2dc053 SW |
1415 | "expected, namelen is %d, pos is %d\n", len, pos); |
1416 | /* presumably this is only possible if racing with a | |
1417 | rename of one of the parent directories (we can not | |
1418 | lock the dentries above us to prevent this, but | |
1419 | retrying should be harmless) */ | |
1420 | kfree(path); | |
1421 | goto retry; | |
1422 | } | |
1423 | ||
1424 | *base = ceph_ino(temp->d_inode); | |
1425 | *plen = len; | |
104648ad | 1426 | dout("build_path on %p %d built %llx '%.*s'\n", |
2f2dc053 SW |
1427 | dentry, atomic_read(&dentry->d_count), *base, len, path); |
1428 | return path; | |
1429 | } | |
1430 | ||
1431 | static int build_dentry_path(struct dentry *dentry, | |
1432 | const char **ppath, int *ppathlen, u64 *pino, | |
1433 | int *pfreepath) | |
1434 | { | |
1435 | char *path; | |
1436 | ||
1437 | if (ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP) { | |
1438 | *pino = ceph_ino(dentry->d_parent->d_inode); | |
1439 | *ppath = dentry->d_name.name; | |
1440 | *ppathlen = dentry->d_name.len; | |
1441 | return 0; | |
1442 | } | |
1443 | path = ceph_mdsc_build_path(dentry, ppathlen, pino, 1); | |
1444 | if (IS_ERR(path)) | |
1445 | return PTR_ERR(path); | |
1446 | *ppath = path; | |
1447 | *pfreepath = 1; | |
1448 | return 0; | |
1449 | } | |
1450 | ||
1451 | static int build_inode_path(struct inode *inode, | |
1452 | const char **ppath, int *ppathlen, u64 *pino, | |
1453 | int *pfreepath) | |
1454 | { | |
1455 | struct dentry *dentry; | |
1456 | char *path; | |
1457 | ||
1458 | if (ceph_snap(inode) == CEPH_NOSNAP) { | |
1459 | *pino = ceph_ino(inode); | |
1460 | *ppathlen = 0; | |
1461 | return 0; | |
1462 | } | |
1463 | dentry = d_find_alias(inode); | |
1464 | path = ceph_mdsc_build_path(dentry, ppathlen, pino, 1); | |
1465 | dput(dentry); | |
1466 | if (IS_ERR(path)) | |
1467 | return PTR_ERR(path); | |
1468 | *ppath = path; | |
1469 | *pfreepath = 1; | |
1470 | return 0; | |
1471 | } | |
1472 | ||
1473 | /* | |
1474 | * request arguments may be specified via an inode *, a dentry *, or | |
1475 | * an explicit ino+path. | |
1476 | */ | |
1477 | static int set_request_path_attr(struct inode *rinode, struct dentry *rdentry, | |
1478 | const char *rpath, u64 rino, | |
1479 | const char **ppath, int *pathlen, | |
1480 | u64 *ino, int *freepath) | |
1481 | { | |
1482 | int r = 0; | |
1483 | ||
1484 | if (rinode) { | |
1485 | r = build_inode_path(rinode, ppath, pathlen, ino, freepath); | |
1486 | dout(" inode %p %llx.%llx\n", rinode, ceph_ino(rinode), | |
1487 | ceph_snap(rinode)); | |
1488 | } else if (rdentry) { | |
1489 | r = build_dentry_path(rdentry, ppath, pathlen, ino, freepath); | |
1490 | dout(" dentry %p %llx/%.*s\n", rdentry, *ino, *pathlen, | |
1491 | *ppath); | |
1492 | } else if (rpath) { | |
1493 | *ino = rino; | |
1494 | *ppath = rpath; | |
1495 | *pathlen = strlen(rpath); | |
1496 | dout(" path %.*s\n", *pathlen, rpath); | |
1497 | } | |
1498 | ||
1499 | return r; | |
1500 | } | |
1501 | ||
1502 | /* | |
1503 | * called under mdsc->mutex | |
1504 | */ | |
1505 | static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, | |
1506 | struct ceph_mds_request *req, | |
1507 | int mds) | |
1508 | { | |
1509 | struct ceph_msg *msg; | |
1510 | struct ceph_mds_request_head *head; | |
1511 | const char *path1 = NULL; | |
1512 | const char *path2 = NULL; | |
1513 | u64 ino1 = 0, ino2 = 0; | |
1514 | int pathlen1 = 0, pathlen2 = 0; | |
1515 | int freepath1 = 0, freepath2 = 0; | |
1516 | int len; | |
1517 | u16 releases; | |
1518 | void *p, *end; | |
1519 | int ret; | |
1520 | ||
1521 | ret = set_request_path_attr(req->r_inode, req->r_dentry, | |
1522 | req->r_path1, req->r_ino1.ino, | |
1523 | &path1, &pathlen1, &ino1, &freepath1); | |
1524 | if (ret < 0) { | |
1525 | msg = ERR_PTR(ret); | |
1526 | goto out; | |
1527 | } | |
1528 | ||
1529 | ret = set_request_path_attr(NULL, req->r_old_dentry, | |
1530 | req->r_path2, req->r_ino2.ino, | |
1531 | &path2, &pathlen2, &ino2, &freepath2); | |
1532 | if (ret < 0) { | |
1533 | msg = ERR_PTR(ret); | |
1534 | goto out_free1; | |
1535 | } | |
1536 | ||
1537 | len = sizeof(*head) + | |
ac8839d7 | 1538 | pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64)); |
2f2dc053 SW |
1539 | |
1540 | /* calculate (max) length for cap releases */ | |
1541 | len += sizeof(struct ceph_mds_request_release) * | |
1542 | (!!req->r_inode_drop + !!req->r_dentry_drop + | |
1543 | !!req->r_old_inode_drop + !!req->r_old_dentry_drop); | |
1544 | if (req->r_dentry_drop) | |
1545 | len += req->r_dentry->d_name.len; | |
1546 | if (req->r_old_dentry_drop) | |
1547 | len += req->r_old_dentry->d_name.len; | |
1548 | ||
34d23762 | 1549 | msg = ceph_msg_new(CEPH_MSG_CLIENT_REQUEST, len, GFP_NOFS); |
a79832f2 SW |
1550 | if (!msg) { |
1551 | msg = ERR_PTR(-ENOMEM); | |
2f2dc053 | 1552 | goto out_free2; |
a79832f2 | 1553 | } |
2f2dc053 | 1554 | |
6df058c0 SW |
1555 | msg->hdr.tid = cpu_to_le64(req->r_tid); |
1556 | ||
2f2dc053 SW |
1557 | head = msg->front.iov_base; |
1558 | p = msg->front.iov_base + sizeof(*head); | |
1559 | end = msg->front.iov_base + msg->front.iov_len; | |
1560 | ||
1561 | head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch); | |
1562 | head->op = cpu_to_le32(req->r_op); | |
1563 | head->caller_uid = cpu_to_le32(current_fsuid()); | |
1564 | head->caller_gid = cpu_to_le32(current_fsgid()); | |
1565 | head->args = req->r_args; | |
1566 | ||
1567 | ceph_encode_filepath(&p, end, ino1, path1); | |
1568 | ceph_encode_filepath(&p, end, ino2, path2); | |
1569 | ||
e979cf50 SW |
1570 | /* make note of release offset, in case we need to replay */ |
1571 | req->r_request_release_offset = p - msg->front.iov_base; | |
1572 | ||
2f2dc053 SW |
1573 | /* cap releases */ |
1574 | releases = 0; | |
1575 | if (req->r_inode_drop) | |
1576 | releases += ceph_encode_inode_release(&p, | |
1577 | req->r_inode ? req->r_inode : req->r_dentry->d_inode, | |
1578 | mds, req->r_inode_drop, req->r_inode_unless, 0); | |
1579 | if (req->r_dentry_drop) | |
1580 | releases += ceph_encode_dentry_release(&p, req->r_dentry, | |
1581 | mds, req->r_dentry_drop, req->r_dentry_unless); | |
1582 | if (req->r_old_dentry_drop) | |
1583 | releases += ceph_encode_dentry_release(&p, req->r_old_dentry, | |
1584 | mds, req->r_old_dentry_drop, req->r_old_dentry_unless); | |
1585 | if (req->r_old_inode_drop) | |
1586 | releases += ceph_encode_inode_release(&p, | |
1587 | req->r_old_dentry->d_inode, | |
1588 | mds, req->r_old_inode_drop, req->r_old_inode_unless, 0); | |
1589 | head->num_releases = cpu_to_le16(releases); | |
1590 | ||
1591 | BUG_ON(p > end); | |
1592 | msg->front.iov_len = p - msg->front.iov_base; | |
1593 | msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); | |
1594 | ||
1595 | msg->pages = req->r_pages; | |
1596 | msg->nr_pages = req->r_num_pages; | |
1597 | msg->hdr.data_len = cpu_to_le32(req->r_data_len); | |
1598 | msg->hdr.data_off = cpu_to_le16(0); | |
1599 | ||
1600 | out_free2: | |
1601 | if (freepath2) | |
1602 | kfree((char *)path2); | |
1603 | out_free1: | |
1604 | if (freepath1) | |
1605 | kfree((char *)path1); | |
1606 | out: | |
1607 | return msg; | |
1608 | } | |
1609 | ||
1610 | /* | |
1611 | * called under mdsc->mutex if error, under no mutex if | |
1612 | * success. | |
1613 | */ | |
1614 | static void complete_request(struct ceph_mds_client *mdsc, | |
1615 | struct ceph_mds_request *req) | |
1616 | { | |
1617 | if (req->r_callback) | |
1618 | req->r_callback(mdsc, req); | |
1619 | else | |
03066f23 | 1620 | complete_all(&req->r_completion); |
2f2dc053 SW |
1621 | } |
1622 | ||
1623 | /* | |
1624 | * called under mdsc->mutex | |
1625 | */ | |
1626 | static int __prepare_send_request(struct ceph_mds_client *mdsc, | |
1627 | struct ceph_mds_request *req, | |
1628 | int mds) | |
1629 | { | |
1630 | struct ceph_mds_request_head *rhead; | |
1631 | struct ceph_msg *msg; | |
1632 | int flags = 0; | |
1633 | ||
1634 | req->r_mds = mds; | |
1635 | req->r_attempts++; | |
e55b71f8 GF |
1636 | if (req->r_inode) { |
1637 | struct ceph_cap *cap = | |
1638 | ceph_get_cap_for_mds(ceph_inode(req->r_inode), mds); | |
1639 | ||
1640 | if (cap) | |
1641 | req->r_sent_on_mseq = cap->mseq; | |
1642 | else | |
1643 | req->r_sent_on_mseq = -1; | |
1644 | } | |
2f2dc053 SW |
1645 | dout("prepare_send_request %p tid %lld %s (attempt %d)\n", req, |
1646 | req->r_tid, ceph_mds_op_name(req->r_op), req->r_attempts); | |
1647 | ||
01a92f17 SW |
1648 | if (req->r_got_unsafe) { |
1649 | /* | |
1650 | * Replay. Do not regenerate message (and rebuild | |
1651 | * paths, etc.); just use the original message. | |
1652 | * Rebuilding paths will break for renames because | |
1653 | * d_move mangles the src name. | |
1654 | */ | |
1655 | msg = req->r_request; | |
1656 | rhead = msg->front.iov_base; | |
1657 | ||
1658 | flags = le32_to_cpu(rhead->flags); | |
1659 | flags |= CEPH_MDS_FLAG_REPLAY; | |
1660 | rhead->flags = cpu_to_le32(flags); | |
1661 | ||
1662 | if (req->r_target_inode) | |
1663 | rhead->ino = cpu_to_le64(ceph_ino(req->r_target_inode)); | |
1664 | ||
1665 | rhead->num_retry = req->r_attempts - 1; | |
e979cf50 SW |
1666 | |
1667 | /* remove cap/dentry releases from message */ | |
1668 | rhead->num_releases = 0; | |
1669 | msg->hdr.front_len = cpu_to_le32(req->r_request_release_offset); | |
1670 | msg->front.iov_len = req->r_request_release_offset; | |
01a92f17 SW |
1671 | return 0; |
1672 | } | |
1673 | ||
2f2dc053 SW |
1674 | if (req->r_request) { |
1675 | ceph_msg_put(req->r_request); | |
1676 | req->r_request = NULL; | |
1677 | } | |
1678 | msg = create_request_message(mdsc, req, mds); | |
1679 | if (IS_ERR(msg)) { | |
e1518c7c | 1680 | req->r_err = PTR_ERR(msg); |
2f2dc053 | 1681 | complete_request(mdsc, req); |
a79832f2 | 1682 | return PTR_ERR(msg); |
2f2dc053 SW |
1683 | } |
1684 | req->r_request = msg; | |
1685 | ||
1686 | rhead = msg->front.iov_base; | |
2f2dc053 SW |
1687 | rhead->oldest_client_tid = cpu_to_le64(__get_oldest_tid(mdsc)); |
1688 | if (req->r_got_unsafe) | |
1689 | flags |= CEPH_MDS_FLAG_REPLAY; | |
1690 | if (req->r_locked_dir) | |
1691 | flags |= CEPH_MDS_FLAG_WANT_DENTRY; | |
1692 | rhead->flags = cpu_to_le32(flags); | |
1693 | rhead->num_fwd = req->r_num_fwd; | |
1694 | rhead->num_retry = req->r_attempts - 1; | |
01a92f17 | 1695 | rhead->ino = 0; |
2f2dc053 SW |
1696 | |
1697 | dout(" r_locked_dir = %p\n", req->r_locked_dir); | |
2f2dc053 SW |
1698 | return 0; |
1699 | } | |
1700 | ||
1701 | /* | |
1702 | * send request, or put it on the appropriate wait list. | |
1703 | */ | |
1704 | static int __do_request(struct ceph_mds_client *mdsc, | |
1705 | struct ceph_mds_request *req) | |
1706 | { | |
1707 | struct ceph_mds_session *session = NULL; | |
1708 | int mds = -1; | |
1709 | int err = -EAGAIN; | |
1710 | ||
e1518c7c | 1711 | if (req->r_err || req->r_got_result) |
2f2dc053 SW |
1712 | goto out; |
1713 | ||
1714 | if (req->r_timeout && | |
1715 | time_after_eq(jiffies, req->r_started + req->r_timeout)) { | |
1716 | dout("do_request timed out\n"); | |
1717 | err = -EIO; | |
1718 | goto finish; | |
1719 | } | |
1720 | ||
1721 | mds = __choose_mds(mdsc, req); | |
1722 | if (mds < 0 || | |
1723 | ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) { | |
1724 | dout("do_request no mds or not active, waiting for map\n"); | |
1725 | list_add(&req->r_wait, &mdsc->waiting_for_map); | |
1726 | goto out; | |
1727 | } | |
1728 | ||
1729 | /* get, open session */ | |
1730 | session = __ceph_lookup_mds_session(mdsc, mds); | |
9c423956 | 1731 | if (!session) { |
2f2dc053 | 1732 | session = register_session(mdsc, mds); |
9c423956 SW |
1733 | if (IS_ERR(session)) { |
1734 | err = PTR_ERR(session); | |
1735 | goto finish; | |
1736 | } | |
1737 | } | |
2f2dc053 SW |
1738 | dout("do_request mds%d session %p state %s\n", mds, session, |
1739 | session_state_name(session->s_state)); | |
1740 | if (session->s_state != CEPH_MDS_SESSION_OPEN && | |
1741 | session->s_state != CEPH_MDS_SESSION_HUNG) { | |
1742 | if (session->s_state == CEPH_MDS_SESSION_NEW || | |
1743 | session->s_state == CEPH_MDS_SESSION_CLOSING) | |
1744 | __open_session(mdsc, session); | |
1745 | list_add(&req->r_wait, &session->s_waiting); | |
1746 | goto out_session; | |
1747 | } | |
1748 | ||
1749 | /* send request */ | |
1750 | req->r_session = get_session(session); | |
1751 | req->r_resend_mds = -1; /* forget any previous mds hint */ | |
1752 | ||
1753 | if (req->r_request_started == 0) /* note request start time */ | |
1754 | req->r_request_started = jiffies; | |
1755 | ||
1756 | err = __prepare_send_request(mdsc, req, mds); | |
1757 | if (!err) { | |
1758 | ceph_msg_get(req->r_request); | |
1759 | ceph_con_send(&session->s_con, req->r_request); | |
1760 | } | |
1761 | ||
1762 | out_session: | |
1763 | ceph_put_mds_session(session); | |
1764 | out: | |
1765 | return err; | |
1766 | ||
1767 | finish: | |
e1518c7c | 1768 | req->r_err = err; |
2f2dc053 SW |
1769 | complete_request(mdsc, req); |
1770 | goto out; | |
1771 | } | |
1772 | ||
1773 | /* | |
1774 | * called under mdsc->mutex | |
1775 | */ | |
1776 | static void __wake_requests(struct ceph_mds_client *mdsc, | |
1777 | struct list_head *head) | |
1778 | { | |
1779 | struct ceph_mds_request *req, *nreq; | |
1780 | ||
1781 | list_for_each_entry_safe(req, nreq, head, r_wait) { | |
1782 | list_del_init(&req->r_wait); | |
1783 | __do_request(mdsc, req); | |
1784 | } | |
1785 | } | |
1786 | ||
1787 | /* | |
1788 | * Wake up threads with requests pending for @mds, so that they can | |
29790f26 | 1789 | * resubmit their requests to a possibly different mds. |
2f2dc053 | 1790 | */ |
29790f26 | 1791 | static void kick_requests(struct ceph_mds_client *mdsc, int mds) |
2f2dc053 | 1792 | { |
44ca18f2 SW |
1793 | struct ceph_mds_request *req; |
1794 | struct rb_node *p; | |
2f2dc053 SW |
1795 | |
1796 | dout("kick_requests mds%d\n", mds); | |
44ca18f2 SW |
1797 | for (p = rb_first(&mdsc->request_tree); p; p = rb_next(p)) { |
1798 | req = rb_entry(p, struct ceph_mds_request, r_node); | |
1799 | if (req->r_got_unsafe) | |
1800 | continue; | |
1801 | if (req->r_session && | |
1802 | req->r_session->s_mds == mds) { | |
1803 | dout(" kicking tid %llu\n", req->r_tid); | |
1804 | put_request_session(req); | |
1805 | __do_request(mdsc, req); | |
2f2dc053 SW |
1806 | } |
1807 | } | |
1808 | } | |
1809 | ||
1810 | void ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, | |
1811 | struct ceph_mds_request *req) | |
1812 | { | |
1813 | dout("submit_request on %p\n", req); | |
1814 | mutex_lock(&mdsc->mutex); | |
1815 | __register_request(mdsc, req, NULL); | |
1816 | __do_request(mdsc, req); | |
1817 | mutex_unlock(&mdsc->mutex); | |
1818 | } | |
1819 | ||
1820 | /* | |
1821 | * Synchrously perform an mds request. Take care of all of the | |
1822 | * session setup, forwarding, retry details. | |
1823 | */ | |
1824 | int ceph_mdsc_do_request(struct ceph_mds_client *mdsc, | |
1825 | struct inode *dir, | |
1826 | struct ceph_mds_request *req) | |
1827 | { | |
1828 | int err; | |
1829 | ||
1830 | dout("do_request on %p\n", req); | |
1831 | ||
1832 | /* take CAP_PIN refs for r_inode, r_locked_dir, r_old_dentry */ | |
1833 | if (req->r_inode) | |
1834 | ceph_get_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN); | |
1835 | if (req->r_locked_dir) | |
1836 | ceph_get_cap_refs(ceph_inode(req->r_locked_dir), CEPH_CAP_PIN); | |
1837 | if (req->r_old_dentry) | |
1838 | ceph_get_cap_refs( | |
1839 | ceph_inode(req->r_old_dentry->d_parent->d_inode), | |
1840 | CEPH_CAP_PIN); | |
1841 | ||
1842 | /* issue */ | |
1843 | mutex_lock(&mdsc->mutex); | |
1844 | __register_request(mdsc, req, dir); | |
1845 | __do_request(mdsc, req); | |
1846 | ||
e1518c7c SW |
1847 | if (req->r_err) { |
1848 | err = req->r_err; | |
1849 | __unregister_request(mdsc, req); | |
1850 | dout("do_request early error %d\n", err); | |
1851 | goto out; | |
2f2dc053 SW |
1852 | } |
1853 | ||
e1518c7c SW |
1854 | /* wait */ |
1855 | mutex_unlock(&mdsc->mutex); | |
1856 | dout("do_request waiting\n"); | |
1857 | if (req->r_timeout) { | |
aa91647c | 1858 | err = (long)wait_for_completion_killable_timeout( |
e1518c7c SW |
1859 | &req->r_completion, req->r_timeout); |
1860 | if (err == 0) | |
1861 | err = -EIO; | |
1862 | } else { | |
aa91647c | 1863 | err = wait_for_completion_killable(&req->r_completion); |
e1518c7c SW |
1864 | } |
1865 | dout("do_request waited, got %d\n", err); | |
1866 | mutex_lock(&mdsc->mutex); | |
5b1daecd | 1867 | |
e1518c7c SW |
1868 | /* only abort if we didn't race with a real reply */ |
1869 | if (req->r_got_result) { | |
1870 | err = le32_to_cpu(req->r_reply_info.head->result); | |
1871 | } else if (err < 0) { | |
1872 | dout("aborted request %lld with %d\n", req->r_tid, err); | |
b4556396 SW |
1873 | |
1874 | /* | |
1875 | * ensure we aren't running concurrently with | |
1876 | * ceph_fill_trace or ceph_readdir_prepopulate, which | |
1877 | * rely on locks (dir mutex) held by our caller. | |
1878 | */ | |
1879 | mutex_lock(&req->r_fill_mutex); | |
e1518c7c SW |
1880 | req->r_err = err; |
1881 | req->r_aborted = true; | |
b4556396 | 1882 | mutex_unlock(&req->r_fill_mutex); |
5b1daecd | 1883 | |
e1518c7c | 1884 | if (req->r_locked_dir && |
167c9e35 SW |
1885 | (req->r_op & CEPH_MDS_OP_WRITE)) |
1886 | ceph_invalidate_dir_request(req); | |
2f2dc053 | 1887 | } else { |
e1518c7c | 1888 | err = req->r_err; |
2f2dc053 | 1889 | } |
2f2dc053 | 1890 | |
e1518c7c SW |
1891 | out: |
1892 | mutex_unlock(&mdsc->mutex); | |
2f2dc053 SW |
1893 | dout("do_request %p done, result %d\n", req, err); |
1894 | return err; | |
1895 | } | |
1896 | ||
167c9e35 SW |
1897 | /* |
1898 | * Invalidate dir I_COMPLETE, dentry lease state on an aborted MDS | |
1899 | * namespace request. | |
1900 | */ | |
1901 | void ceph_invalidate_dir_request(struct ceph_mds_request *req) | |
1902 | { | |
1903 | struct inode *inode = req->r_locked_dir; | |
1904 | struct ceph_inode_info *ci = ceph_inode(inode); | |
1905 | ||
1906 | dout("invalidate_dir_request %p (I_COMPLETE, lease(s))\n", inode); | |
1907 | spin_lock(&inode->i_lock); | |
1908 | ci->i_ceph_flags &= ~CEPH_I_COMPLETE; | |
1909 | ci->i_release_count++; | |
1910 | spin_unlock(&inode->i_lock); | |
1911 | ||
1912 | if (req->r_dentry) | |
1913 | ceph_invalidate_dentry_lease(req->r_dentry); | |
1914 | if (req->r_old_dentry) | |
1915 | ceph_invalidate_dentry_lease(req->r_old_dentry); | |
1916 | } | |
1917 | ||
2f2dc053 SW |
1918 | /* |
1919 | * Handle mds reply. | |
1920 | * | |
1921 | * We take the session mutex and parse and process the reply immediately. | |
1922 | * This preserves the logical ordering of replies, capabilities, etc., sent | |
1923 | * by the MDS as they are applied to our local cache. | |
1924 | */ | |
1925 | static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg) | |
1926 | { | |
1927 | struct ceph_mds_client *mdsc = session->s_mdsc; | |
1928 | struct ceph_mds_request *req; | |
1929 | struct ceph_mds_reply_head *head = msg->front.iov_base; | |
1930 | struct ceph_mds_reply_info_parsed *rinfo; /* parsed reply info */ | |
1931 | u64 tid; | |
1932 | int err, result; | |
2600d2dd | 1933 | int mds = session->s_mds; |
2f2dc053 | 1934 | |
2f2dc053 SW |
1935 | if (msg->front.iov_len < sizeof(*head)) { |
1936 | pr_err("mdsc_handle_reply got corrupt (short) reply\n"); | |
9ec7cab1 | 1937 | ceph_msg_dump(msg); |
2f2dc053 SW |
1938 | return; |
1939 | } | |
1940 | ||
1941 | /* get request, session */ | |
6df058c0 | 1942 | tid = le64_to_cpu(msg->hdr.tid); |
2f2dc053 SW |
1943 | mutex_lock(&mdsc->mutex); |
1944 | req = __lookup_request(mdsc, tid); | |
1945 | if (!req) { | |
1946 | dout("handle_reply on unknown tid %llu\n", tid); | |
1947 | mutex_unlock(&mdsc->mutex); | |
1948 | return; | |
1949 | } | |
1950 | dout("handle_reply %p\n", req); | |
2f2dc053 SW |
1951 | |
1952 | /* correct session? */ | |
d96d6049 | 1953 | if (req->r_session != session) { |
2f2dc053 SW |
1954 | pr_err("mdsc_handle_reply got %llu on session mds%d" |
1955 | " not mds%d\n", tid, session->s_mds, | |
1956 | req->r_session ? req->r_session->s_mds : -1); | |
1957 | mutex_unlock(&mdsc->mutex); | |
1958 | goto out; | |
1959 | } | |
1960 | ||
1961 | /* dup? */ | |
1962 | if ((req->r_got_unsafe && !head->safe) || | |
1963 | (req->r_got_safe && head->safe)) { | |
1964 | pr_warning("got a dup %s reply on %llu from mds%d\n", | |
1965 | head->safe ? "safe" : "unsafe", tid, mds); | |
1966 | mutex_unlock(&mdsc->mutex); | |
1967 | goto out; | |
1968 | } | |
85792d0d SW |
1969 | if (req->r_got_safe && !head->safe) { |
1970 | pr_warning("got unsafe after safe on %llu from mds%d\n", | |
1971 | tid, mds); | |
1972 | mutex_unlock(&mdsc->mutex); | |
1973 | goto out; | |
1974 | } | |
2f2dc053 SW |
1975 | |
1976 | result = le32_to_cpu(head->result); | |
1977 | ||
1978 | /* | |
e55b71f8 GF |
1979 | * Handle an ESTALE |
1980 | * if we're not talking to the authority, send to them | |
1981 | * if the authority has changed while we weren't looking, | |
1982 | * send to new authority | |
1983 | * Otherwise we just have to return an ESTALE | |
2f2dc053 SW |
1984 | */ |
1985 | if (result == -ESTALE) { | |
e55b71f8 GF |
1986 | dout("got ESTALE on request %llu", req->r_tid); |
1987 | if (!req->r_inode) ; //do nothing; not an authority problem | |
1988 | else if (req->r_direct_mode != USE_AUTH_MDS) { | |
1989 | dout("not using auth, setting for that now"); | |
1990 | req->r_direct_mode = USE_AUTH_MDS; | |
2f2dc053 SW |
1991 | __do_request(mdsc, req); |
1992 | mutex_unlock(&mdsc->mutex); | |
1993 | goto out; | |
e55b71f8 GF |
1994 | } else { |
1995 | struct ceph_inode_info *ci = ceph_inode(req->r_inode); | |
1996 | struct ceph_cap *cap = | |
1997 | ceph_get_cap_for_mds(ci, req->r_mds);; | |
1998 | ||
1999 | dout("already using auth"); | |
2000 | if ((!cap || cap != ci->i_auth_cap) || | |
2001 | (cap->mseq != req->r_sent_on_mseq)) { | |
2002 | dout("but cap changed, so resending"); | |
2003 | __do_request(mdsc, req); | |
2004 | mutex_unlock(&mdsc->mutex); | |
2005 | goto out; | |
2006 | } | |
2f2dc053 | 2007 | } |
e55b71f8 | 2008 | dout("have to return ESTALE on request %llu", req->r_tid); |
2f2dc053 SW |
2009 | } |
2010 | ||
e55b71f8 | 2011 | |
2f2dc053 SW |
2012 | if (head->safe) { |
2013 | req->r_got_safe = true; | |
2014 | __unregister_request(mdsc, req); | |
03066f23 | 2015 | complete_all(&req->r_safe_completion); |
2f2dc053 SW |
2016 | |
2017 | if (req->r_got_unsafe) { | |
2018 | /* | |
2019 | * We already handled the unsafe response, now do the | |
2020 | * cleanup. No need to examine the response; the MDS | |
2021 | * doesn't include any result info in the safe | |
2022 | * response. And even if it did, there is nothing | |
2023 | * useful we could do with a revised return value. | |
2024 | */ | |
2025 | dout("got safe reply %llu, mds%d\n", tid, mds); | |
2026 | list_del_init(&req->r_unsafe_item); | |
2027 | ||
2028 | /* last unsafe request during umount? */ | |
44ca18f2 | 2029 | if (mdsc->stopping && !__get_oldest_req(mdsc)) |
03066f23 | 2030 | complete_all(&mdsc->safe_umount_waiters); |
2f2dc053 SW |
2031 | mutex_unlock(&mdsc->mutex); |
2032 | goto out; | |
2033 | } | |
e1518c7c | 2034 | } else { |
2f2dc053 SW |
2035 | req->r_got_unsafe = true; |
2036 | list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe); | |
2037 | } | |
2038 | ||
2039 | dout("handle_reply tid %lld result %d\n", tid, result); | |
2040 | rinfo = &req->r_reply_info; | |
2041 | err = parse_reply_info(msg, rinfo); | |
2042 | mutex_unlock(&mdsc->mutex); | |
2043 | ||
2044 | mutex_lock(&session->s_mutex); | |
2045 | if (err < 0) { | |
2046 | pr_err("mdsc_handle_reply got corrupt reply mds%d\n", mds); | |
9ec7cab1 | 2047 | ceph_msg_dump(msg); |
2f2dc053 SW |
2048 | goto out_err; |
2049 | } | |
2050 | ||
2051 | /* snap trace */ | |
2052 | if (rinfo->snapblob_len) { | |
2053 | down_write(&mdsc->snap_rwsem); | |
2054 | ceph_update_snap_trace(mdsc, rinfo->snapblob, | |
2055 | rinfo->snapblob + rinfo->snapblob_len, | |
2056 | le32_to_cpu(head->op) == CEPH_MDS_OP_RMSNAP); | |
2057 | downgrade_write(&mdsc->snap_rwsem); | |
2058 | } else { | |
2059 | down_read(&mdsc->snap_rwsem); | |
2060 | } | |
2061 | ||
2062 | /* insert trace into our cache */ | |
b4556396 | 2063 | mutex_lock(&req->r_fill_mutex); |
2f2dc053 SW |
2064 | err = ceph_fill_trace(mdsc->client->sb, req, req->r_session); |
2065 | if (err == 0) { | |
2066 | if (result == 0 && rinfo->dir_nr) | |
2067 | ceph_readdir_prepopulate(req, req->r_session); | |
37151668 | 2068 | ceph_unreserve_caps(mdsc, &req->r_caps_reservation); |
2f2dc053 | 2069 | } |
b4556396 | 2070 | mutex_unlock(&req->r_fill_mutex); |
2f2dc053 SW |
2071 | |
2072 | up_read(&mdsc->snap_rwsem); | |
2073 | out_err: | |
e1518c7c SW |
2074 | mutex_lock(&mdsc->mutex); |
2075 | if (!req->r_aborted) { | |
2076 | if (err) { | |
2077 | req->r_err = err; | |
2078 | } else { | |
2079 | req->r_reply = msg; | |
2080 | ceph_msg_get(msg); | |
2081 | req->r_got_result = true; | |
2082 | } | |
2f2dc053 | 2083 | } else { |
e1518c7c | 2084 | dout("reply arrived after request %lld was aborted\n", tid); |
2f2dc053 | 2085 | } |
e1518c7c | 2086 | mutex_unlock(&mdsc->mutex); |
2f2dc053 | 2087 | |
ee6b272b | 2088 | ceph_add_cap_releases(mdsc, req->r_session); |
2f2dc053 SW |
2089 | mutex_unlock(&session->s_mutex); |
2090 | ||
2091 | /* kick calling process */ | |
2092 | complete_request(mdsc, req); | |
2093 | out: | |
2094 | ceph_mdsc_put_request(req); | |
2095 | return; | |
2096 | } | |
2097 | ||
2098 | ||
2099 | ||
2100 | /* | |
2101 | * handle mds notification that our request has been forwarded. | |
2102 | */ | |
2600d2dd SW |
2103 | static void handle_forward(struct ceph_mds_client *mdsc, |
2104 | struct ceph_mds_session *session, | |
2105 | struct ceph_msg *msg) | |
2f2dc053 SW |
2106 | { |
2107 | struct ceph_mds_request *req; | |
a1ea787c | 2108 | u64 tid = le64_to_cpu(msg->hdr.tid); |
2f2dc053 SW |
2109 | u32 next_mds; |
2110 | u32 fwd_seq; | |
2f2dc053 SW |
2111 | int err = -EINVAL; |
2112 | void *p = msg->front.iov_base; | |
2113 | void *end = p + msg->front.iov_len; | |
2f2dc053 | 2114 | |
a1ea787c | 2115 | ceph_decode_need(&p, end, 2*sizeof(u32), bad); |
c89136ea SW |
2116 | next_mds = ceph_decode_32(&p); |
2117 | fwd_seq = ceph_decode_32(&p); | |
2f2dc053 SW |
2118 | |
2119 | mutex_lock(&mdsc->mutex); | |
2120 | req = __lookup_request(mdsc, tid); | |
2121 | if (!req) { | |
2a8e5e36 | 2122 | dout("forward tid %llu to mds%d - req dne\n", tid, next_mds); |
2f2dc053 SW |
2123 | goto out; /* dup reply? */ |
2124 | } | |
2125 | ||
2a8e5e36 SW |
2126 | if (req->r_aborted) { |
2127 | dout("forward tid %llu aborted, unregistering\n", tid); | |
2128 | __unregister_request(mdsc, req); | |
2129 | } else if (fwd_seq <= req->r_num_fwd) { | |
2130 | dout("forward tid %llu to mds%d - old seq %d <= %d\n", | |
2f2dc053 SW |
2131 | tid, next_mds, req->r_num_fwd, fwd_seq); |
2132 | } else { | |
2133 | /* resend. forward race not possible; mds would drop */ | |
2a8e5e36 SW |
2134 | dout("forward tid %llu to mds%d (we resend)\n", tid, next_mds); |
2135 | BUG_ON(req->r_err); | |
2136 | BUG_ON(req->r_got_result); | |
2f2dc053 SW |
2137 | req->r_num_fwd = fwd_seq; |
2138 | req->r_resend_mds = next_mds; | |
2139 | put_request_session(req); | |
2140 | __do_request(mdsc, req); | |
2141 | } | |
2142 | ceph_mdsc_put_request(req); | |
2143 | out: | |
2144 | mutex_unlock(&mdsc->mutex); | |
2145 | return; | |
2146 | ||
2147 | bad: | |
2148 | pr_err("mdsc_handle_forward decode error err=%d\n", err); | |
2149 | } | |
2150 | ||
2151 | /* | |
2152 | * handle a mds session control message | |
2153 | */ | |
2154 | static void handle_session(struct ceph_mds_session *session, | |
2155 | struct ceph_msg *msg) | |
2156 | { | |
2157 | struct ceph_mds_client *mdsc = session->s_mdsc; | |
2158 | u32 op; | |
2159 | u64 seq; | |
2600d2dd | 2160 | int mds = session->s_mds; |
2f2dc053 SW |
2161 | struct ceph_mds_session_head *h = msg->front.iov_base; |
2162 | int wake = 0; | |
2163 | ||
2f2dc053 SW |
2164 | /* decode */ |
2165 | if (msg->front.iov_len != sizeof(*h)) | |
2166 | goto bad; | |
2167 | op = le32_to_cpu(h->op); | |
2168 | seq = le64_to_cpu(h->seq); | |
2169 | ||
2170 | mutex_lock(&mdsc->mutex); | |
2600d2dd SW |
2171 | if (op == CEPH_SESSION_CLOSE) |
2172 | __unregister_session(mdsc, session); | |
2f2dc053 SW |
2173 | /* FIXME: this ttl calculation is generous */ |
2174 | session->s_ttl = jiffies + HZ*mdsc->mdsmap->m_session_autoclose; | |
2175 | mutex_unlock(&mdsc->mutex); | |
2176 | ||
2177 | mutex_lock(&session->s_mutex); | |
2178 | ||
2179 | dout("handle_session mds%d %s %p state %s seq %llu\n", | |
2180 | mds, ceph_session_op_name(op), session, | |
2181 | session_state_name(session->s_state), seq); | |
2182 | ||
2183 | if (session->s_state == CEPH_MDS_SESSION_HUNG) { | |
2184 | session->s_state = CEPH_MDS_SESSION_OPEN; | |
2185 | pr_info("mds%d came back\n", session->s_mds); | |
2186 | } | |
2187 | ||
2188 | switch (op) { | |
2189 | case CEPH_SESSION_OPEN: | |
29790f26 SW |
2190 | if (session->s_state == CEPH_MDS_SESSION_RECONNECTING) |
2191 | pr_info("mds%d reconnect success\n", session->s_mds); | |
2f2dc053 SW |
2192 | session->s_state = CEPH_MDS_SESSION_OPEN; |
2193 | renewed_caps(mdsc, session, 0); | |
2194 | wake = 1; | |
2195 | if (mdsc->stopping) | |
2196 | __close_session(mdsc, session); | |
2197 | break; | |
2198 | ||
2199 | case CEPH_SESSION_RENEWCAPS: | |
2200 | if (session->s_renew_seq == seq) | |
2201 | renewed_caps(mdsc, session, 1); | |
2202 | break; | |
2203 | ||
2204 | case CEPH_SESSION_CLOSE: | |
29790f26 SW |
2205 | if (session->s_state == CEPH_MDS_SESSION_RECONNECTING) |
2206 | pr_info("mds%d reconnect denied\n", session->s_mds); | |
2f2dc053 SW |
2207 | remove_session_caps(session); |
2208 | wake = 1; /* for good measure */ | |
03066f23 | 2209 | complete_all(&mdsc->session_close_waiters); |
29790f26 | 2210 | kick_requests(mdsc, mds); |
2f2dc053 SW |
2211 | break; |
2212 | ||
2213 | case CEPH_SESSION_STALE: | |
2214 | pr_info("mds%d caps went stale, renewing\n", | |
2215 | session->s_mds); | |
2216 | spin_lock(&session->s_cap_lock); | |
2217 | session->s_cap_gen++; | |
2218 | session->s_cap_ttl = 0; | |
2219 | spin_unlock(&session->s_cap_lock); | |
2220 | send_renew_caps(mdsc, session); | |
2221 | break; | |
2222 | ||
2223 | case CEPH_SESSION_RECALL_STATE: | |
2224 | trim_caps(mdsc, session, le32_to_cpu(h->max_caps)); | |
2225 | break; | |
2226 | ||
2227 | default: | |
2228 | pr_err("mdsc_handle_session bad op %d mds%d\n", op, mds); | |
2229 | WARN_ON(1); | |
2230 | } | |
2231 | ||
2232 | mutex_unlock(&session->s_mutex); | |
2233 | if (wake) { | |
2234 | mutex_lock(&mdsc->mutex); | |
2235 | __wake_requests(mdsc, &session->s_waiting); | |
2236 | mutex_unlock(&mdsc->mutex); | |
2237 | } | |
2238 | return; | |
2239 | ||
2240 | bad: | |
2241 | pr_err("mdsc_handle_session corrupt message mds%d len %d\n", mds, | |
2242 | (int)msg->front.iov_len); | |
9ec7cab1 | 2243 | ceph_msg_dump(msg); |
2f2dc053 SW |
2244 | return; |
2245 | } | |
2246 | ||
2247 | ||
2248 | /* | |
2249 | * called under session->mutex. | |
2250 | */ | |
2251 | static void replay_unsafe_requests(struct ceph_mds_client *mdsc, | |
2252 | struct ceph_mds_session *session) | |
2253 | { | |
2254 | struct ceph_mds_request *req, *nreq; | |
2255 | int err; | |
2256 | ||
2257 | dout("replay_unsafe_requests mds%d\n", session->s_mds); | |
2258 | ||
2259 | mutex_lock(&mdsc->mutex); | |
2260 | list_for_each_entry_safe(req, nreq, &session->s_unsafe, r_unsafe_item) { | |
2261 | err = __prepare_send_request(mdsc, req, session->s_mds); | |
2262 | if (!err) { | |
2263 | ceph_msg_get(req->r_request); | |
2264 | ceph_con_send(&session->s_con, req->r_request); | |
2265 | } | |
2266 | } | |
2267 | mutex_unlock(&mdsc->mutex); | |
2268 | } | |
2269 | ||
2270 | /* | |
2271 | * Encode information about a cap for a reconnect with the MDS. | |
2272 | */ | |
2f2dc053 SW |
2273 | static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, |
2274 | void *arg) | |
2275 | { | |
20cb34ae SW |
2276 | union { |
2277 | struct ceph_mds_cap_reconnect v2; | |
2278 | struct ceph_mds_cap_reconnect_v1 v1; | |
2279 | } rec; | |
2280 | size_t reclen; | |
2f2dc053 | 2281 | struct ceph_inode_info *ci; |
20cb34ae SW |
2282 | struct ceph_reconnect_state *recon_state = arg; |
2283 | struct ceph_pagelist *pagelist = recon_state->pagelist; | |
2f2dc053 SW |
2284 | char *path; |
2285 | int pathlen, err; | |
2286 | u64 pathbase; | |
2287 | struct dentry *dentry; | |
2288 | ||
2289 | ci = cap->ci; | |
2290 | ||
2291 | dout(" adding %p ino %llx.%llx cap %p %lld %s\n", | |
2292 | inode, ceph_vinop(inode), cap, cap->cap_id, | |
2293 | ceph_cap_string(cap->issued)); | |
93cea5be SW |
2294 | err = ceph_pagelist_encode_64(pagelist, ceph_ino(inode)); |
2295 | if (err) | |
2296 | return err; | |
2f2dc053 SW |
2297 | |
2298 | dentry = d_find_alias(inode); | |
2299 | if (dentry) { | |
2300 | path = ceph_mdsc_build_path(dentry, &pathlen, &pathbase, 0); | |
2301 | if (IS_ERR(path)) { | |
2302 | err = PTR_ERR(path); | |
2303 | BUG_ON(err); | |
2304 | } | |
2305 | } else { | |
2306 | path = NULL; | |
2307 | pathlen = 0; | |
2308 | } | |
93cea5be SW |
2309 | err = ceph_pagelist_encode_string(pagelist, path, pathlen); |
2310 | if (err) | |
2311 | goto out; | |
2f2dc053 | 2312 | |
2f2dc053 SW |
2313 | spin_lock(&inode->i_lock); |
2314 | cap->seq = 0; /* reset cap seq */ | |
2315 | cap->issue_seq = 0; /* and issue_seq */ | |
20cb34ae SW |
2316 | |
2317 | if (recon_state->flock) { | |
2318 | rec.v2.cap_id = cpu_to_le64(cap->cap_id); | |
2319 | rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); | |
2320 | rec.v2.issued = cpu_to_le32(cap->issued); | |
2321 | rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); | |
2322 | rec.v2.pathbase = cpu_to_le64(pathbase); | |
2323 | rec.v2.flock_len = 0; | |
2324 | reclen = sizeof(rec.v2); | |
2325 | } else { | |
2326 | rec.v1.cap_id = cpu_to_le64(cap->cap_id); | |
2327 | rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); | |
2328 | rec.v1.issued = cpu_to_le32(cap->issued); | |
2329 | rec.v1.size = cpu_to_le64(inode->i_size); | |
2330 | ceph_encode_timespec(&rec.v1.mtime, &inode->i_mtime); | |
2331 | ceph_encode_timespec(&rec.v1.atime, &inode->i_atime); | |
2332 | rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); | |
2333 | rec.v1.pathbase = cpu_to_le64(pathbase); | |
2334 | reclen = sizeof(rec.v1); | |
2335 | } | |
2f2dc053 SW |
2336 | spin_unlock(&inode->i_lock); |
2337 | ||
20cb34ae | 2338 | err = ceph_pagelist_append(pagelist, &rec, reclen); |
93cea5be SW |
2339 | |
2340 | out: | |
2f2dc053 SW |
2341 | kfree(path); |
2342 | dput(dentry); | |
93cea5be | 2343 | return err; |
2f2dc053 SW |
2344 | } |
2345 | ||
2346 | ||
2347 | /* | |
2348 | * If an MDS fails and recovers, clients need to reconnect in order to | |
2349 | * reestablish shared state. This includes all caps issued through | |
2350 | * this session _and_ the snap_realm hierarchy. Because it's not | |
2351 | * clear which snap realms the mds cares about, we send everything we | |
2352 | * know about.. that ensures we'll then get any new info the | |
2353 | * recovering MDS might have. | |
2354 | * | |
2355 | * This is a relatively heavyweight operation, but it's rare. | |
2356 | * | |
2357 | * called with mdsc->mutex held. | |
2358 | */ | |
34b6c855 SW |
2359 | static void send_mds_reconnect(struct ceph_mds_client *mdsc, |
2360 | struct ceph_mds_session *session) | |
2f2dc053 | 2361 | { |
2f2dc053 | 2362 | struct ceph_msg *reply; |
a105f00c | 2363 | struct rb_node *p; |
34b6c855 | 2364 | int mds = session->s_mds; |
9abf82b8 | 2365 | int err = -ENOMEM; |
93cea5be | 2366 | struct ceph_pagelist *pagelist; |
20cb34ae | 2367 | struct ceph_reconnect_state recon_state; |
2f2dc053 | 2368 | |
34b6c855 | 2369 | pr_info("mds%d reconnect start\n", mds); |
2f2dc053 | 2370 | |
93cea5be SW |
2371 | pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS); |
2372 | if (!pagelist) | |
2373 | goto fail_nopagelist; | |
2374 | ceph_pagelist_init(pagelist); | |
2375 | ||
34d23762 | 2376 | reply = ceph_msg_new(CEPH_MSG_CLIENT_RECONNECT, 0, GFP_NOFS); |
a79832f2 | 2377 | if (!reply) |
93cea5be | 2378 | goto fail_nomsg; |
93cea5be | 2379 | |
34b6c855 SW |
2380 | mutex_lock(&session->s_mutex); |
2381 | session->s_state = CEPH_MDS_SESSION_RECONNECTING; | |
2382 | session->s_seq = 0; | |
2f2dc053 | 2383 | |
34b6c855 SW |
2384 | ceph_con_open(&session->s_con, |
2385 | ceph_mdsmap_get_addr(mdsc->mdsmap, mds)); | |
2f2dc053 | 2386 | |
34b6c855 SW |
2387 | /* replay unsafe requests */ |
2388 | replay_unsafe_requests(mdsc, session); | |
2f2dc053 SW |
2389 | |
2390 | down_read(&mdsc->snap_rwsem); | |
2391 | ||
2f2dc053 SW |
2392 | dout("session %p state %s\n", session, |
2393 | session_state_name(session->s_state)); | |
2394 | ||
e01a5946 SW |
2395 | /* drop old cap expires; we're about to reestablish that state */ |
2396 | discard_cap_releases(mdsc, session); | |
2397 | ||
2f2dc053 | 2398 | /* traverse this session's caps */ |
93cea5be SW |
2399 | err = ceph_pagelist_encode_32(pagelist, session->s_nr_caps); |
2400 | if (err) | |
2401 | goto fail; | |
20cb34ae SW |
2402 | |
2403 | recon_state.pagelist = pagelist; | |
2404 | recon_state.flock = session->s_con.peer_features & CEPH_FEATURE_FLOCK; | |
2405 | err = iterate_session_caps(session, encode_caps_cb, &recon_state); | |
2f2dc053 | 2406 | if (err < 0) |
9abf82b8 | 2407 | goto fail; |
2f2dc053 SW |
2408 | |
2409 | /* | |
2410 | * snaprealms. we provide mds with the ino, seq (version), and | |
2411 | * parent for all of our realms. If the mds has any newer info, | |
2412 | * it will tell us. | |
2413 | */ | |
a105f00c SW |
2414 | for (p = rb_first(&mdsc->snap_realms); p; p = rb_next(p)) { |
2415 | struct ceph_snap_realm *realm = | |
2416 | rb_entry(p, struct ceph_snap_realm, node); | |
93cea5be | 2417 | struct ceph_mds_snaprealm_reconnect sr_rec; |
2f2dc053 SW |
2418 | |
2419 | dout(" adding snap realm %llx seq %lld parent %llx\n", | |
2420 | realm->ino, realm->seq, realm->parent_ino); | |
93cea5be SW |
2421 | sr_rec.ino = cpu_to_le64(realm->ino); |
2422 | sr_rec.seq = cpu_to_le64(realm->seq); | |
2423 | sr_rec.parent = cpu_to_le64(realm->parent_ino); | |
2424 | err = ceph_pagelist_append(pagelist, &sr_rec, sizeof(sr_rec)); | |
2425 | if (err) | |
2426 | goto fail; | |
2f2dc053 | 2427 | } |
2f2dc053 | 2428 | |
93cea5be | 2429 | reply->pagelist = pagelist; |
20cb34ae SW |
2430 | if (recon_state.flock) |
2431 | reply->hdr.version = cpu_to_le16(2); | |
93cea5be SW |
2432 | reply->hdr.data_len = cpu_to_le32(pagelist->length); |
2433 | reply->nr_pages = calc_pages_for(0, pagelist->length); | |
2f2dc053 SW |
2434 | ceph_con_send(&session->s_con, reply); |
2435 | ||
9abf82b8 SW |
2436 | mutex_unlock(&session->s_mutex); |
2437 | ||
2438 | mutex_lock(&mdsc->mutex); | |
2439 | __wake_requests(mdsc, &session->s_waiting); | |
2440 | mutex_unlock(&mdsc->mutex); | |
2441 | ||
2f2dc053 | 2442 | up_read(&mdsc->snap_rwsem); |
2f2dc053 SW |
2443 | return; |
2444 | ||
93cea5be | 2445 | fail: |
2f2dc053 | 2446 | ceph_msg_put(reply); |
9abf82b8 SW |
2447 | up_read(&mdsc->snap_rwsem); |
2448 | mutex_unlock(&session->s_mutex); | |
93cea5be SW |
2449 | fail_nomsg: |
2450 | ceph_pagelist_release(pagelist); | |
2451 | kfree(pagelist); | |
2452 | fail_nopagelist: | |
9abf82b8 | 2453 | pr_err("error %d preparing reconnect for mds%d\n", err, mds); |
9abf82b8 | 2454 | return; |
2f2dc053 SW |
2455 | } |
2456 | ||
2457 | ||
2458 | /* | |
2459 | * compare old and new mdsmaps, kicking requests | |
2460 | * and closing out old connections as necessary | |
2461 | * | |
2462 | * called under mdsc->mutex. | |
2463 | */ | |
2464 | static void check_new_map(struct ceph_mds_client *mdsc, | |
2465 | struct ceph_mdsmap *newmap, | |
2466 | struct ceph_mdsmap *oldmap) | |
2467 | { | |
2468 | int i; | |
2469 | int oldstate, newstate; | |
2470 | struct ceph_mds_session *s; | |
2471 | ||
2472 | dout("check_new_map new %u old %u\n", | |
2473 | newmap->m_epoch, oldmap->m_epoch); | |
2474 | ||
2475 | for (i = 0; i < oldmap->m_max_mds && i < mdsc->max_sessions; i++) { | |
2476 | if (mdsc->sessions[i] == NULL) | |
2477 | continue; | |
2478 | s = mdsc->sessions[i]; | |
2479 | oldstate = ceph_mdsmap_get_state(oldmap, i); | |
2480 | newstate = ceph_mdsmap_get_state(newmap, i); | |
2481 | ||
0deb01c9 | 2482 | dout("check_new_map mds%d state %s%s -> %s%s (session %s)\n", |
2f2dc053 | 2483 | i, ceph_mds_state_name(oldstate), |
0deb01c9 | 2484 | ceph_mdsmap_is_laggy(oldmap, i) ? " (laggy)" : "", |
2f2dc053 | 2485 | ceph_mds_state_name(newstate), |
0deb01c9 | 2486 | ceph_mdsmap_is_laggy(newmap, i) ? " (laggy)" : "", |
2f2dc053 SW |
2487 | session_state_name(s->s_state)); |
2488 | ||
2489 | if (memcmp(ceph_mdsmap_get_addr(oldmap, i), | |
2490 | ceph_mdsmap_get_addr(newmap, i), | |
2491 | sizeof(struct ceph_entity_addr))) { | |
2492 | if (s->s_state == CEPH_MDS_SESSION_OPENING) { | |
2493 | /* the session never opened, just close it | |
2494 | * out now */ | |
2495 | __wake_requests(mdsc, &s->s_waiting); | |
2600d2dd | 2496 | __unregister_session(mdsc, s); |
2f2dc053 SW |
2497 | } else { |
2498 | /* just close it */ | |
2499 | mutex_unlock(&mdsc->mutex); | |
2500 | mutex_lock(&s->s_mutex); | |
2501 | mutex_lock(&mdsc->mutex); | |
2502 | ceph_con_close(&s->s_con); | |
2503 | mutex_unlock(&s->s_mutex); | |
2504 | s->s_state = CEPH_MDS_SESSION_RESTARTING; | |
2505 | } | |
2506 | ||
2507 | /* kick any requests waiting on the recovering mds */ | |
29790f26 | 2508 | kick_requests(mdsc, i); |
2f2dc053 SW |
2509 | } else if (oldstate == newstate) { |
2510 | continue; /* nothing new with this mds */ | |
2511 | } | |
2512 | ||
2513 | /* | |
2514 | * send reconnect? | |
2515 | */ | |
2516 | if (s->s_state == CEPH_MDS_SESSION_RESTARTING && | |
34b6c855 SW |
2517 | newstate >= CEPH_MDS_STATE_RECONNECT) { |
2518 | mutex_unlock(&mdsc->mutex); | |
2519 | send_mds_reconnect(mdsc, s); | |
2520 | mutex_lock(&mdsc->mutex); | |
2521 | } | |
2f2dc053 SW |
2522 | |
2523 | /* | |
29790f26 | 2524 | * kick request on any mds that has gone active. |
2f2dc053 SW |
2525 | */ |
2526 | if (oldstate < CEPH_MDS_STATE_ACTIVE && | |
2527 | newstate >= CEPH_MDS_STATE_ACTIVE) { | |
29790f26 SW |
2528 | if (oldstate != CEPH_MDS_STATE_CREATING && |
2529 | oldstate != CEPH_MDS_STATE_STARTING) | |
2530 | pr_info("mds%d recovery completed\n", s->s_mds); | |
2531 | kick_requests(mdsc, i); | |
2f2dc053 | 2532 | ceph_kick_flushing_caps(mdsc, s); |
0dc2570f | 2533 | wake_up_session_caps(s, 1); |
2f2dc053 SW |
2534 | } |
2535 | } | |
cb170a22 SW |
2536 | |
2537 | for (i = 0; i < newmap->m_max_mds && i < mdsc->max_sessions; i++) { | |
2538 | s = mdsc->sessions[i]; | |
2539 | if (!s) | |
2540 | continue; | |
2541 | if (!ceph_mdsmap_is_laggy(newmap, i)) | |
2542 | continue; | |
2543 | if (s->s_state == CEPH_MDS_SESSION_OPEN || | |
2544 | s->s_state == CEPH_MDS_SESSION_HUNG || | |
2545 | s->s_state == CEPH_MDS_SESSION_CLOSING) { | |
2546 | dout(" connecting to export targets of laggy mds%d\n", | |
2547 | i); | |
2548 | __open_export_target_sessions(mdsc, s); | |
2549 | } | |
2550 | } | |
2f2dc053 SW |
2551 | } |
2552 | ||
2553 | ||
2554 | ||
2555 | /* | |
2556 | * leases | |
2557 | */ | |
2558 | ||
2559 | /* | |
2560 | * caller must hold session s_mutex, dentry->d_lock | |
2561 | */ | |
2562 | void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry) | |
2563 | { | |
2564 | struct ceph_dentry_info *di = ceph_dentry(dentry); | |
2565 | ||
2566 | ceph_put_mds_session(di->lease_session); | |
2567 | di->lease_session = NULL; | |
2568 | } | |
2569 | ||
2600d2dd SW |
2570 | static void handle_lease(struct ceph_mds_client *mdsc, |
2571 | struct ceph_mds_session *session, | |
2572 | struct ceph_msg *msg) | |
2f2dc053 SW |
2573 | { |
2574 | struct super_block *sb = mdsc->client->sb; | |
2575 | struct inode *inode; | |
2f2dc053 SW |
2576 | struct ceph_inode_info *ci; |
2577 | struct dentry *parent, *dentry; | |
2578 | struct ceph_dentry_info *di; | |
2600d2dd | 2579 | int mds = session->s_mds; |
2f2dc053 | 2580 | struct ceph_mds_lease *h = msg->front.iov_base; |
1e5ea23d | 2581 | u32 seq; |
2f2dc053 SW |
2582 | struct ceph_vino vino; |
2583 | int mask; | |
2584 | struct qstr dname; | |
2585 | int release = 0; | |
2586 | ||
2f2dc053 SW |
2587 | dout("handle_lease from mds%d\n", mds); |
2588 | ||
2589 | /* decode */ | |
2590 | if (msg->front.iov_len < sizeof(*h) + sizeof(u32)) | |
2591 | goto bad; | |
2592 | vino.ino = le64_to_cpu(h->ino); | |
2593 | vino.snap = CEPH_NOSNAP; | |
2594 | mask = le16_to_cpu(h->mask); | |
1e5ea23d | 2595 | seq = le32_to_cpu(h->seq); |
2f2dc053 SW |
2596 | dname.name = (void *)h + sizeof(*h) + sizeof(u32); |
2597 | dname.len = msg->front.iov_len - sizeof(*h) - sizeof(u32); | |
2598 | if (dname.len != get_unaligned_le32(h+1)) | |
2599 | goto bad; | |
2600 | ||
2f2dc053 SW |
2601 | mutex_lock(&session->s_mutex); |
2602 | session->s_seq++; | |
2603 | ||
2604 | /* lookup inode */ | |
2605 | inode = ceph_find_inode(sb, vino); | |
1e5ea23d SW |
2606 | dout("handle_lease %s, mask %d, ino %llx %p %.*s\n", |
2607 | ceph_lease_op_name(h->action), mask, vino.ino, inode, | |
2608 | dname.len, dname.name); | |
2f2dc053 SW |
2609 | if (inode == NULL) { |
2610 | dout("handle_lease no inode %llx\n", vino.ino); | |
2611 | goto release; | |
2612 | } | |
2613 | ci = ceph_inode(inode); | |
2614 | ||
2615 | /* dentry */ | |
2616 | parent = d_find_alias(inode); | |
2617 | if (!parent) { | |
2618 | dout("no parent dentry on inode %p\n", inode); | |
2619 | WARN_ON(1); | |
2620 | goto release; /* hrm... */ | |
2621 | } | |
2622 | dname.hash = full_name_hash(dname.name, dname.len); | |
2623 | dentry = d_lookup(parent, &dname); | |
2624 | dput(parent); | |
2625 | if (!dentry) | |
2626 | goto release; | |
2627 | ||
2628 | spin_lock(&dentry->d_lock); | |
2629 | di = ceph_dentry(dentry); | |
2630 | switch (h->action) { | |
2631 | case CEPH_MDS_LEASE_REVOKE: | |
2632 | if (di && di->lease_session == session) { | |
1e5ea23d SW |
2633 | if (ceph_seq_cmp(di->lease_seq, seq) > 0) |
2634 | h->seq = cpu_to_le32(di->lease_seq); | |
2f2dc053 SW |
2635 | __ceph_mdsc_drop_dentry_lease(dentry); |
2636 | } | |
2637 | release = 1; | |
2638 | break; | |
2639 | ||
2640 | case CEPH_MDS_LEASE_RENEW: | |
2641 | if (di && di->lease_session == session && | |
2642 | di->lease_gen == session->s_cap_gen && | |
2643 | di->lease_renew_from && | |
2644 | di->lease_renew_after == 0) { | |
2645 | unsigned long duration = | |
2646 | le32_to_cpu(h->duration_ms) * HZ / 1000; | |
2647 | ||
1e5ea23d | 2648 | di->lease_seq = seq; |
2f2dc053 SW |
2649 | dentry->d_time = di->lease_renew_from + duration; |
2650 | di->lease_renew_after = di->lease_renew_from + | |
2651 | (duration >> 1); | |
2652 | di->lease_renew_from = 0; | |
2653 | } | |
2654 | break; | |
2655 | } | |
2656 | spin_unlock(&dentry->d_lock); | |
2657 | dput(dentry); | |
2658 | ||
2659 | if (!release) | |
2660 | goto out; | |
2661 | ||
2662 | release: | |
2663 | /* let's just reuse the same message */ | |
2664 | h->action = CEPH_MDS_LEASE_REVOKE_ACK; | |
2665 | ceph_msg_get(msg); | |
2666 | ceph_con_send(&session->s_con, msg); | |
2667 | ||
2668 | out: | |
2669 | iput(inode); | |
2670 | mutex_unlock(&session->s_mutex); | |
2f2dc053 SW |
2671 | return; |
2672 | ||
2673 | bad: | |
2674 | pr_err("corrupt lease message\n"); | |
9ec7cab1 | 2675 | ceph_msg_dump(msg); |
2f2dc053 SW |
2676 | } |
2677 | ||
2678 | void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, | |
2679 | struct inode *inode, | |
2680 | struct dentry *dentry, char action, | |
2681 | u32 seq) | |
2682 | { | |
2683 | struct ceph_msg *msg; | |
2684 | struct ceph_mds_lease *lease; | |
2685 | int len = sizeof(*lease) + sizeof(u32); | |
2686 | int dnamelen = 0; | |
2687 | ||
2688 | dout("lease_send_msg inode %p dentry %p %s to mds%d\n", | |
2689 | inode, dentry, ceph_lease_op_name(action), session->s_mds); | |
2690 | dnamelen = dentry->d_name.len; | |
2691 | len += dnamelen; | |
2692 | ||
34d23762 | 2693 | msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, GFP_NOFS); |
a79832f2 | 2694 | if (!msg) |
2f2dc053 SW |
2695 | return; |
2696 | lease = msg->front.iov_base; | |
2697 | lease->action = action; | |
dd1c9057 | 2698 | lease->mask = cpu_to_le16(1); |
2f2dc053 SW |
2699 | lease->ino = cpu_to_le64(ceph_vino(inode).ino); |
2700 | lease->first = lease->last = cpu_to_le64(ceph_vino(inode).snap); | |
2701 | lease->seq = cpu_to_le32(seq); | |
2702 | put_unaligned_le32(dnamelen, lease + 1); | |
2703 | memcpy((void *)(lease + 1) + 4, dentry->d_name.name, dnamelen); | |
2704 | ||
2705 | /* | |
2706 | * if this is a preemptive lease RELEASE, no need to | |
2707 | * flush request stream, since the actual request will | |
2708 | * soon follow. | |
2709 | */ | |
2710 | msg->more_to_follow = (action == CEPH_MDS_LEASE_RELEASE); | |
2711 | ||
2712 | ceph_con_send(&session->s_con, msg); | |
2713 | } | |
2714 | ||
2715 | /* | |
2716 | * Preemptively release a lease we expect to invalidate anyway. | |
2717 | * Pass @inode always, @dentry is optional. | |
2718 | */ | |
2719 | void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode, | |
2720 | struct dentry *dentry, int mask) | |
2721 | { | |
2722 | struct ceph_dentry_info *di; | |
2723 | struct ceph_mds_session *session; | |
2724 | u32 seq; | |
2725 | ||
2726 | BUG_ON(inode == NULL); | |
2727 | BUG_ON(dentry == NULL); | |
dd1c9057 | 2728 | BUG_ON(mask == 0); |
2f2dc053 SW |
2729 | |
2730 | /* is dentry lease valid? */ | |
2731 | spin_lock(&dentry->d_lock); | |
2732 | di = ceph_dentry(dentry); | |
2733 | if (!di || !di->lease_session || | |
2734 | di->lease_session->s_mds < 0 || | |
2735 | di->lease_gen != di->lease_session->s_cap_gen || | |
2736 | !time_before(jiffies, dentry->d_time)) { | |
2737 | dout("lease_release inode %p dentry %p -- " | |
2738 | "no lease on %d\n", | |
2739 | inode, dentry, mask); | |
2740 | spin_unlock(&dentry->d_lock); | |
2741 | return; | |
2742 | } | |
2743 | ||
2744 | /* we do have a lease on this dentry; note mds and seq */ | |
2745 | session = ceph_get_mds_session(di->lease_session); | |
2746 | seq = di->lease_seq; | |
2747 | __ceph_mdsc_drop_dentry_lease(dentry); | |
2748 | spin_unlock(&dentry->d_lock); | |
2749 | ||
2750 | dout("lease_release inode %p dentry %p mask %d to mds%d\n", | |
2751 | inode, dentry, mask, session->s_mds); | |
2752 | ceph_mdsc_lease_send_msg(session, inode, dentry, | |
2753 | CEPH_MDS_LEASE_RELEASE, seq); | |
2754 | ceph_put_mds_session(session); | |
2755 | } | |
2756 | ||
2757 | /* | |
2758 | * drop all leases (and dentry refs) in preparation for umount | |
2759 | */ | |
2760 | static void drop_leases(struct ceph_mds_client *mdsc) | |
2761 | { | |
2762 | int i; | |
2763 | ||
2764 | dout("drop_leases\n"); | |
2765 | mutex_lock(&mdsc->mutex); | |
2766 | for (i = 0; i < mdsc->max_sessions; i++) { | |
2767 | struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i); | |
2768 | if (!s) | |
2769 | continue; | |
2770 | mutex_unlock(&mdsc->mutex); | |
2771 | mutex_lock(&s->s_mutex); | |
2772 | mutex_unlock(&s->s_mutex); | |
2773 | ceph_put_mds_session(s); | |
2774 | mutex_lock(&mdsc->mutex); | |
2775 | } | |
2776 | mutex_unlock(&mdsc->mutex); | |
2777 | } | |
2778 | ||
2779 | ||
2780 | ||
2781 | /* | |
2782 | * delayed work -- periodically trim expired leases, renew caps with mds | |
2783 | */ | |
2784 | static void schedule_delayed(struct ceph_mds_client *mdsc) | |
2785 | { | |
2786 | int delay = 5; | |
2787 | unsigned hz = round_jiffies_relative(HZ * delay); | |
2788 | schedule_delayed_work(&mdsc->delayed_work, hz); | |
2789 | } | |
2790 | ||
2791 | static void delayed_work(struct work_struct *work) | |
2792 | { | |
2793 | int i; | |
2794 | struct ceph_mds_client *mdsc = | |
2795 | container_of(work, struct ceph_mds_client, delayed_work.work); | |
2796 | int renew_interval; | |
2797 | int renew_caps; | |
2798 | ||
2799 | dout("mdsc delayed_work\n"); | |
afcdaea3 | 2800 | ceph_check_delayed_caps(mdsc); |
2f2dc053 SW |
2801 | |
2802 | mutex_lock(&mdsc->mutex); | |
2803 | renew_interval = mdsc->mdsmap->m_session_timeout >> 2; | |
2804 | renew_caps = time_after_eq(jiffies, HZ*renew_interval + | |
2805 | mdsc->last_renew_caps); | |
2806 | if (renew_caps) | |
2807 | mdsc->last_renew_caps = jiffies; | |
2808 | ||
2809 | for (i = 0; i < mdsc->max_sessions; i++) { | |
2810 | struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i); | |
2811 | if (s == NULL) | |
2812 | continue; | |
2813 | if (s->s_state == CEPH_MDS_SESSION_CLOSING) { | |
2814 | dout("resending session close request for mds%d\n", | |
2815 | s->s_mds); | |
2816 | request_close_session(mdsc, s); | |
2817 | ceph_put_mds_session(s); | |
2818 | continue; | |
2819 | } | |
2820 | if (s->s_ttl && time_after(jiffies, s->s_ttl)) { | |
2821 | if (s->s_state == CEPH_MDS_SESSION_OPEN) { | |
2822 | s->s_state = CEPH_MDS_SESSION_HUNG; | |
2823 | pr_info("mds%d hung\n", s->s_mds); | |
2824 | } | |
2825 | } | |
2826 | if (s->s_state < CEPH_MDS_SESSION_OPEN) { | |
2827 | /* this mds is failed or recovering, just wait */ | |
2828 | ceph_put_mds_session(s); | |
2829 | continue; | |
2830 | } | |
2831 | mutex_unlock(&mdsc->mutex); | |
2832 | ||
2833 | mutex_lock(&s->s_mutex); | |
2834 | if (renew_caps) | |
2835 | send_renew_caps(mdsc, s); | |
2836 | else | |
2837 | ceph_con_keepalive(&s->s_con); | |
ee6b272b | 2838 | ceph_add_cap_releases(mdsc, s); |
aab53dd9 SW |
2839 | if (s->s_state == CEPH_MDS_SESSION_OPEN || |
2840 | s->s_state == CEPH_MDS_SESSION_HUNG) | |
3d7ded4d | 2841 | ceph_send_cap_releases(mdsc, s); |
2f2dc053 SW |
2842 | mutex_unlock(&s->s_mutex); |
2843 | ceph_put_mds_session(s); | |
2844 | ||
2845 | mutex_lock(&mdsc->mutex); | |
2846 | } | |
2847 | mutex_unlock(&mdsc->mutex); | |
2848 | ||
2849 | schedule_delayed(mdsc); | |
2850 | } | |
2851 | ||
2852 | ||
5f44f142 | 2853 | int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client) |
2f2dc053 SW |
2854 | { |
2855 | mdsc->client = client; | |
2856 | mutex_init(&mdsc->mutex); | |
2857 | mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS); | |
2d06eeb8 CR |
2858 | if (mdsc->mdsmap == NULL) |
2859 | return -ENOMEM; | |
2860 | ||
2f2dc053 SW |
2861 | init_completion(&mdsc->safe_umount_waiters); |
2862 | init_completion(&mdsc->session_close_waiters); | |
2863 | INIT_LIST_HEAD(&mdsc->waiting_for_map); | |
2864 | mdsc->sessions = NULL; | |
2865 | mdsc->max_sessions = 0; | |
2866 | mdsc->stopping = 0; | |
2867 | init_rwsem(&mdsc->snap_rwsem); | |
a105f00c | 2868 | mdsc->snap_realms = RB_ROOT; |
2f2dc053 SW |
2869 | INIT_LIST_HEAD(&mdsc->snap_empty); |
2870 | spin_lock_init(&mdsc->snap_empty_lock); | |
2871 | mdsc->last_tid = 0; | |
44ca18f2 | 2872 | mdsc->request_tree = RB_ROOT; |
2f2dc053 SW |
2873 | INIT_DELAYED_WORK(&mdsc->delayed_work, delayed_work); |
2874 | mdsc->last_renew_caps = jiffies; | |
2875 | INIT_LIST_HEAD(&mdsc->cap_delay_list); | |
2876 | spin_lock_init(&mdsc->cap_delay_lock); | |
2877 | INIT_LIST_HEAD(&mdsc->snap_flush_list); | |
2878 | spin_lock_init(&mdsc->snap_flush_lock); | |
2879 | mdsc->cap_flush_seq = 0; | |
2880 | INIT_LIST_HEAD(&mdsc->cap_dirty); | |
2881 | mdsc->num_cap_flushing = 0; | |
2882 | spin_lock_init(&mdsc->cap_dirty_lock); | |
2883 | init_waitqueue_head(&mdsc->cap_flushing_wq); | |
2884 | spin_lock_init(&mdsc->dentry_lru_lock); | |
2885 | INIT_LIST_HEAD(&mdsc->dentry_lru); | |
2d06eeb8 | 2886 | |
37151668 YS |
2887 | ceph_caps_init(mdsc); |
2888 | ceph_adjust_min_caps(mdsc, client->min_caps); | |
2889 | ||
5f44f142 | 2890 | return 0; |
2f2dc053 SW |
2891 | } |
2892 | ||
2893 | /* | |
2894 | * Wait for safe replies on open mds requests. If we time out, drop | |
2895 | * all requests from the tree to avoid dangling dentry refs. | |
2896 | */ | |
2897 | static void wait_requests(struct ceph_mds_client *mdsc) | |
2898 | { | |
2899 | struct ceph_mds_request *req; | |
2900 | struct ceph_client *client = mdsc->client; | |
2901 | ||
2902 | mutex_lock(&mdsc->mutex); | |
44ca18f2 | 2903 | if (__get_oldest_req(mdsc)) { |
2f2dc053 | 2904 | mutex_unlock(&mdsc->mutex); |
44ca18f2 | 2905 | |
2f2dc053 SW |
2906 | dout("wait_requests waiting for requests\n"); |
2907 | wait_for_completion_timeout(&mdsc->safe_umount_waiters, | |
6b805185 | 2908 | client->mount_args->mount_timeout * HZ); |
2f2dc053 SW |
2909 | |
2910 | /* tear down remaining requests */ | |
44ca18f2 SW |
2911 | mutex_lock(&mdsc->mutex); |
2912 | while ((req = __get_oldest_req(mdsc))) { | |
2f2dc053 SW |
2913 | dout("wait_requests timed out on tid %llu\n", |
2914 | req->r_tid); | |
44ca18f2 | 2915 | __unregister_request(mdsc, req); |
2f2dc053 SW |
2916 | } |
2917 | } | |
2918 | mutex_unlock(&mdsc->mutex); | |
2919 | dout("wait_requests done\n"); | |
2920 | } | |
2921 | ||
2922 | /* | |
2923 | * called before mount is ro, and before dentries are torn down. | |
2924 | * (hmm, does this still race with new lookups?) | |
2925 | */ | |
2926 | void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc) | |
2927 | { | |
2928 | dout("pre_umount\n"); | |
2929 | mdsc->stopping = 1; | |
2930 | ||
2931 | drop_leases(mdsc); | |
afcdaea3 | 2932 | ceph_flush_dirty_caps(mdsc); |
2f2dc053 | 2933 | wait_requests(mdsc); |
17c688c3 SW |
2934 | |
2935 | /* | |
2936 | * wait for reply handlers to drop their request refs and | |
2937 | * their inode/dcache refs | |
2938 | */ | |
2939 | ceph_msgr_flush(); | |
2f2dc053 SW |
2940 | } |
2941 | ||
2942 | /* | |
2943 | * wait for all write mds requests to flush. | |
2944 | */ | |
2945 | static void wait_unsafe_requests(struct ceph_mds_client *mdsc, u64 want_tid) | |
2946 | { | |
80fc7314 | 2947 | struct ceph_mds_request *req = NULL, *nextreq; |
44ca18f2 | 2948 | struct rb_node *n; |
2f2dc053 SW |
2949 | |
2950 | mutex_lock(&mdsc->mutex); | |
2951 | dout("wait_unsafe_requests want %lld\n", want_tid); | |
80fc7314 | 2952 | restart: |
44ca18f2 SW |
2953 | req = __get_oldest_req(mdsc); |
2954 | while (req && req->r_tid <= want_tid) { | |
80fc7314 SW |
2955 | /* find next request */ |
2956 | n = rb_next(&req->r_node); | |
2957 | if (n) | |
2958 | nextreq = rb_entry(n, struct ceph_mds_request, r_node); | |
2959 | else | |
2960 | nextreq = NULL; | |
44ca18f2 SW |
2961 | if ((req->r_op & CEPH_MDS_OP_WRITE)) { |
2962 | /* write op */ | |
2963 | ceph_mdsc_get_request(req); | |
80fc7314 SW |
2964 | if (nextreq) |
2965 | ceph_mdsc_get_request(nextreq); | |
44ca18f2 SW |
2966 | mutex_unlock(&mdsc->mutex); |
2967 | dout("wait_unsafe_requests wait on %llu (want %llu)\n", | |
2968 | req->r_tid, want_tid); | |
2969 | wait_for_completion(&req->r_safe_completion); | |
2970 | mutex_lock(&mdsc->mutex); | |
44ca18f2 | 2971 | ceph_mdsc_put_request(req); |
80fc7314 SW |
2972 | if (!nextreq) |
2973 | break; /* next dne before, so we're done! */ | |
2974 | if (RB_EMPTY_NODE(&nextreq->r_node)) { | |
2975 | /* next request was removed from tree */ | |
2976 | ceph_mdsc_put_request(nextreq); | |
2977 | goto restart; | |
2978 | } | |
2979 | ceph_mdsc_put_request(nextreq); /* won't go away */ | |
44ca18f2 | 2980 | } |
80fc7314 | 2981 | req = nextreq; |
2f2dc053 SW |
2982 | } |
2983 | mutex_unlock(&mdsc->mutex); | |
2984 | dout("wait_unsafe_requests done\n"); | |
2985 | } | |
2986 | ||
2987 | void ceph_mdsc_sync(struct ceph_mds_client *mdsc) | |
2988 | { | |
2989 | u64 want_tid, want_flush; | |
2990 | ||
56b7cf95 SW |
2991 | if (mdsc->client->mount_state == CEPH_MOUNT_SHUTDOWN) |
2992 | return; | |
2993 | ||
2f2dc053 SW |
2994 | dout("sync\n"); |
2995 | mutex_lock(&mdsc->mutex); | |
2996 | want_tid = mdsc->last_tid; | |
2997 | want_flush = mdsc->cap_flush_seq; | |
2998 | mutex_unlock(&mdsc->mutex); | |
2999 | dout("sync want tid %lld flush_seq %lld\n", want_tid, want_flush); | |
3000 | ||
afcdaea3 | 3001 | ceph_flush_dirty_caps(mdsc); |
2f2dc053 SW |
3002 | |
3003 | wait_unsafe_requests(mdsc, want_tid); | |
3004 | wait_event(mdsc->cap_flushing_wq, check_cap_flush(mdsc, want_flush)); | |
3005 | } | |
3006 | ||
3007 | ||
3008 | /* | |
3009 | * called after sb is ro. | |
3010 | */ | |
3011 | void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc) | |
3012 | { | |
3013 | struct ceph_mds_session *session; | |
3014 | int i; | |
3015 | int n; | |
3016 | struct ceph_client *client = mdsc->client; | |
6b805185 | 3017 | unsigned long started, timeout = client->mount_args->mount_timeout * HZ; |
2f2dc053 SW |
3018 | |
3019 | dout("close_sessions\n"); | |
3020 | ||
3021 | mutex_lock(&mdsc->mutex); | |
3022 | ||
3023 | /* close sessions */ | |
3024 | started = jiffies; | |
3025 | while (time_before(jiffies, started + timeout)) { | |
3026 | dout("closing sessions\n"); | |
3027 | n = 0; | |
3028 | for (i = 0; i < mdsc->max_sessions; i++) { | |
3029 | session = __ceph_lookup_mds_session(mdsc, i); | |
3030 | if (!session) | |
3031 | continue; | |
3032 | mutex_unlock(&mdsc->mutex); | |
3033 | mutex_lock(&session->s_mutex); | |
3034 | __close_session(mdsc, session); | |
3035 | mutex_unlock(&session->s_mutex); | |
3036 | ceph_put_mds_session(session); | |
3037 | mutex_lock(&mdsc->mutex); | |
3038 | n++; | |
3039 | } | |
3040 | if (n == 0) | |
3041 | break; | |
3042 | ||
3043 | if (client->mount_state == CEPH_MOUNT_SHUTDOWN) | |
3044 | break; | |
3045 | ||
3046 | dout("waiting for sessions to close\n"); | |
3047 | mutex_unlock(&mdsc->mutex); | |
3048 | wait_for_completion_timeout(&mdsc->session_close_waiters, | |
3049 | timeout); | |
3050 | mutex_lock(&mdsc->mutex); | |
3051 | } | |
3052 | ||
3053 | /* tear down remaining sessions */ | |
3054 | for (i = 0; i < mdsc->max_sessions; i++) { | |
3055 | if (mdsc->sessions[i]) { | |
3056 | session = get_session(mdsc->sessions[i]); | |
2600d2dd | 3057 | __unregister_session(mdsc, session); |
2f2dc053 SW |
3058 | mutex_unlock(&mdsc->mutex); |
3059 | mutex_lock(&session->s_mutex); | |
3060 | remove_session_caps(session); | |
3061 | mutex_unlock(&session->s_mutex); | |
3062 | ceph_put_mds_session(session); | |
3063 | mutex_lock(&mdsc->mutex); | |
3064 | } | |
3065 | } | |
3066 | ||
3067 | WARN_ON(!list_empty(&mdsc->cap_delay_list)); | |
3068 | ||
3069 | mutex_unlock(&mdsc->mutex); | |
3070 | ||
3071 | ceph_cleanup_empty_realms(mdsc); | |
3072 | ||
3073 | cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */ | |
3074 | ||
3075 | dout("stopped\n"); | |
3076 | } | |
3077 | ||
3078 | void ceph_mdsc_stop(struct ceph_mds_client *mdsc) | |
3079 | { | |
3080 | dout("stop\n"); | |
3081 | cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */ | |
3082 | if (mdsc->mdsmap) | |
3083 | ceph_mdsmap_destroy(mdsc->mdsmap); | |
3084 | kfree(mdsc->sessions); | |
37151668 | 3085 | ceph_caps_finalize(mdsc); |
2f2dc053 SW |
3086 | } |
3087 | ||
3088 | ||
3089 | /* | |
3090 | * handle mds map update. | |
3091 | */ | |
3092 | void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg) | |
3093 | { | |
3094 | u32 epoch; | |
3095 | u32 maplen; | |
3096 | void *p = msg->front.iov_base; | |
3097 | void *end = p + msg->front.iov_len; | |
3098 | struct ceph_mdsmap *newmap, *oldmap; | |
3099 | struct ceph_fsid fsid; | |
3100 | int err = -EINVAL; | |
3101 | ||
3102 | ceph_decode_need(&p, end, sizeof(fsid)+2*sizeof(u32), bad); | |
3103 | ceph_decode_copy(&p, &fsid, sizeof(fsid)); | |
0743304d SW |
3104 | if (ceph_check_fsid(mdsc->client, &fsid) < 0) |
3105 | return; | |
c89136ea SW |
3106 | epoch = ceph_decode_32(&p); |
3107 | maplen = ceph_decode_32(&p); | |
2f2dc053 SW |
3108 | dout("handle_map epoch %u len %d\n", epoch, (int)maplen); |
3109 | ||
3110 | /* do we need it? */ | |
3111 | ceph_monc_got_mdsmap(&mdsc->client->monc, epoch); | |
3112 | mutex_lock(&mdsc->mutex); | |
3113 | if (mdsc->mdsmap && epoch <= mdsc->mdsmap->m_epoch) { | |
3114 | dout("handle_map epoch %u <= our %u\n", | |
3115 | epoch, mdsc->mdsmap->m_epoch); | |
3116 | mutex_unlock(&mdsc->mutex); | |
3117 | return; | |
3118 | } | |
3119 | ||
3120 | newmap = ceph_mdsmap_decode(&p, end); | |
3121 | if (IS_ERR(newmap)) { | |
3122 | err = PTR_ERR(newmap); | |
3123 | goto bad_unlock; | |
3124 | } | |
3125 | ||
3126 | /* swap into place */ | |
3127 | if (mdsc->mdsmap) { | |
3128 | oldmap = mdsc->mdsmap; | |
3129 | mdsc->mdsmap = newmap; | |
3130 | check_new_map(mdsc, newmap, oldmap); | |
3131 | ceph_mdsmap_destroy(oldmap); | |
3132 | } else { | |
3133 | mdsc->mdsmap = newmap; /* first mds map */ | |
3134 | } | |
3135 | mdsc->client->sb->s_maxbytes = mdsc->mdsmap->m_max_file_size; | |
3136 | ||
3137 | __wake_requests(mdsc, &mdsc->waiting_for_map); | |
3138 | ||
3139 | mutex_unlock(&mdsc->mutex); | |
3140 | schedule_delayed(mdsc); | |
3141 | return; | |
3142 | ||
3143 | bad_unlock: | |
3144 | mutex_unlock(&mdsc->mutex); | |
3145 | bad: | |
3146 | pr_err("error decoding mdsmap %d\n", err); | |
3147 | return; | |
3148 | } | |
3149 | ||
3150 | static struct ceph_connection *con_get(struct ceph_connection *con) | |
3151 | { | |
3152 | struct ceph_mds_session *s = con->private; | |
3153 | ||
3154 | if (get_session(s)) { | |
2600d2dd | 3155 | dout("mdsc con_get %p ok (%d)\n", s, atomic_read(&s->s_ref)); |
2f2dc053 SW |
3156 | return con; |
3157 | } | |
3158 | dout("mdsc con_get %p FAIL\n", s); | |
3159 | return NULL; | |
3160 | } | |
3161 | ||
3162 | static void con_put(struct ceph_connection *con) | |
3163 | { | |
3164 | struct ceph_mds_session *s = con->private; | |
3165 | ||
2f2dc053 | 3166 | ceph_put_mds_session(s); |
2600d2dd | 3167 | dout("mdsc con_put %p (%d)\n", s, atomic_read(&s->s_ref)); |
2f2dc053 SW |
3168 | } |
3169 | ||
3170 | /* | |
3171 | * if the client is unresponsive for long enough, the mds will kill | |
3172 | * the session entirely. | |
3173 | */ | |
3174 | static void peer_reset(struct ceph_connection *con) | |
3175 | { | |
3176 | struct ceph_mds_session *s = con->private; | |
7e70f0ed | 3177 | struct ceph_mds_client *mdsc = s->s_mdsc; |
2f2dc053 | 3178 | |
7e70f0ed SW |
3179 | pr_warning("mds%d closed our session\n", s->s_mds); |
3180 | send_mds_reconnect(mdsc, s); | |
2f2dc053 SW |
3181 | } |
3182 | ||
3183 | static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) | |
3184 | { | |
3185 | struct ceph_mds_session *s = con->private; | |
3186 | struct ceph_mds_client *mdsc = s->s_mdsc; | |
3187 | int type = le16_to_cpu(msg->hdr.type); | |
3188 | ||
2600d2dd SW |
3189 | mutex_lock(&mdsc->mutex); |
3190 | if (__verify_registered_session(mdsc, s) < 0) { | |
3191 | mutex_unlock(&mdsc->mutex); | |
3192 | goto out; | |
3193 | } | |
3194 | mutex_unlock(&mdsc->mutex); | |
3195 | ||
2f2dc053 SW |
3196 | switch (type) { |
3197 | case CEPH_MSG_MDS_MAP: | |
3198 | ceph_mdsc_handle_map(mdsc, msg); | |
3199 | break; | |
3200 | case CEPH_MSG_CLIENT_SESSION: | |
3201 | handle_session(s, msg); | |
3202 | break; | |
3203 | case CEPH_MSG_CLIENT_REPLY: | |
3204 | handle_reply(s, msg); | |
3205 | break; | |
3206 | case CEPH_MSG_CLIENT_REQUEST_FORWARD: | |
2600d2dd | 3207 | handle_forward(mdsc, s, msg); |
2f2dc053 SW |
3208 | break; |
3209 | case CEPH_MSG_CLIENT_CAPS: | |
3210 | ceph_handle_caps(s, msg); | |
3211 | break; | |
3212 | case CEPH_MSG_CLIENT_SNAP: | |
2600d2dd | 3213 | ceph_handle_snap(mdsc, s, msg); |
2f2dc053 SW |
3214 | break; |
3215 | case CEPH_MSG_CLIENT_LEASE: | |
2600d2dd | 3216 | handle_lease(mdsc, s, msg); |
2f2dc053 SW |
3217 | break; |
3218 | ||
3219 | default: | |
3220 | pr_err("received unknown message type %d %s\n", type, | |
3221 | ceph_msg_type_name(type)); | |
3222 | } | |
2600d2dd | 3223 | out: |
2f2dc053 SW |
3224 | ceph_msg_put(msg); |
3225 | } | |
3226 | ||
4e7a5dcd SW |
3227 | /* |
3228 | * authentication | |
3229 | */ | |
3230 | static int get_authorizer(struct ceph_connection *con, | |
3231 | void **buf, int *len, int *proto, | |
3232 | void **reply_buf, int *reply_len, int force_new) | |
3233 | { | |
3234 | struct ceph_mds_session *s = con->private; | |
3235 | struct ceph_mds_client *mdsc = s->s_mdsc; | |
3236 | struct ceph_auth_client *ac = mdsc->client->monc.auth; | |
3237 | int ret = 0; | |
3238 | ||
3239 | if (force_new && s->s_authorizer) { | |
3240 | ac->ops->destroy_authorizer(ac, s->s_authorizer); | |
3241 | s->s_authorizer = NULL; | |
3242 | } | |
3243 | if (s->s_authorizer == NULL) { | |
3244 | if (ac->ops->create_authorizer) { | |
3245 | ret = ac->ops->create_authorizer( | |
3246 | ac, CEPH_ENTITY_TYPE_MDS, | |
3247 | &s->s_authorizer, | |
3248 | &s->s_authorizer_buf, | |
3249 | &s->s_authorizer_buf_len, | |
3250 | &s->s_authorizer_reply_buf, | |
3251 | &s->s_authorizer_reply_buf_len); | |
3252 | if (ret) | |
3253 | return ret; | |
3254 | } | |
3255 | } | |
3256 | ||
3257 | *proto = ac->protocol; | |
3258 | *buf = s->s_authorizer_buf; | |
3259 | *len = s->s_authorizer_buf_len; | |
3260 | *reply_buf = s->s_authorizer_reply_buf; | |
3261 | *reply_len = s->s_authorizer_reply_buf_len; | |
3262 | return 0; | |
3263 | } | |
3264 | ||
3265 | ||
3266 | static int verify_authorizer_reply(struct ceph_connection *con, int len) | |
3267 | { | |
3268 | struct ceph_mds_session *s = con->private; | |
3269 | struct ceph_mds_client *mdsc = s->s_mdsc; | |
3270 | struct ceph_auth_client *ac = mdsc->client->monc.auth; | |
3271 | ||
3272 | return ac->ops->verify_authorizer_reply(ac, s->s_authorizer, len); | |
3273 | } | |
3274 | ||
9bd2e6f8 SW |
3275 | static int invalidate_authorizer(struct ceph_connection *con) |
3276 | { | |
3277 | struct ceph_mds_session *s = con->private; | |
3278 | struct ceph_mds_client *mdsc = s->s_mdsc; | |
3279 | struct ceph_auth_client *ac = mdsc->client->monc.auth; | |
3280 | ||
3281 | if (ac->ops->invalidate_authorizer) | |
3282 | ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS); | |
3283 | ||
3284 | return ceph_monc_validate_auth(&mdsc->client->monc); | |
3285 | } | |
3286 | ||
9e32789f | 3287 | static const struct ceph_connection_operations mds_con_ops = { |
2f2dc053 SW |
3288 | .get = con_get, |
3289 | .put = con_put, | |
3290 | .dispatch = dispatch, | |
4e7a5dcd SW |
3291 | .get_authorizer = get_authorizer, |
3292 | .verify_authorizer_reply = verify_authorizer_reply, | |
9bd2e6f8 | 3293 | .invalidate_authorizer = invalidate_authorizer, |
2f2dc053 | 3294 | .peer_reset = peer_reset, |
2f2dc053 SW |
3295 | }; |
3296 | ||
3297 | ||
3298 | ||
3299 | ||
3300 | /* eof */ |