1 /******************************************************************************
2 *******************************************************************************
3 **
4 ** Copyright (C) 2005-2010 Red Hat, Inc. All rights reserved.
5 **
6 ** This copyrighted material is made available to anyone wishing to use,
7 ** modify, copy, or redistribute it subject to the terms and conditions
8 ** of the GNU General Public License v.2.
9 **
10 *******************************************************************************
11 ******************************************************************************/
12
13 /* Central locking logic has four stages:
14
15 dlm_lock()
16 dlm_unlock()
17
18 request_lock(ls, lkb)
19 convert_lock(ls, lkb)
20 unlock_lock(ls, lkb)
21 cancel_lock(ls, lkb)
22
23 _request_lock(r, lkb)
24 _convert_lock(r, lkb)
25 _unlock_lock(r, lkb)
26 _cancel_lock(r, lkb)
27
28 do_request(r, lkb)
29 do_convert(r, lkb)
30 do_unlock(r, lkb)
31 do_cancel(r, lkb)
32
33 Stage 1 (lock, unlock) is mainly about checking input args and
34 splitting into one of the four main operations:
35
36 dlm_lock = request_lock
37 dlm_lock+CONVERT = convert_lock
38 dlm_unlock = unlock_lock
39 dlm_unlock+CANCEL = cancel_lock
40
41 Stage 2, xxxx_lock(), just finds and locks the relevant rsb which is
42 provided to the next stage.
43
44 Stage 3, _xxxx_lock(), determines if the operation is local or remote.
45 When remote, it calls send_xxxx(), when local it calls do_xxxx().
46
47 Stage 4, do_xxxx(), is the guts of the operation. It manipulates the
48 given rsb and lkb and queues callbacks.
49
50 For remote operations, send_xxxx() results in the corresponding do_xxxx()
51 function being executed on the remote node. The connecting send/receive
52 calls on local (L) and remote (R) nodes:
53
54 L: send_xxxx() -> R: receive_xxxx()
55 R: do_xxxx()
56 L: receive_xxxx_reply() <- R: send_xxxx_reply()
57 */
58 #include <linux/types.h>
59 #include <linux/rbtree.h>
60 #include <linux/slab.h>
61 #include "dlm_internal.h"
62 #include <linux/dlm_device.h>
63 #include "memory.h"
64 #include "lowcomms.h"
65 #include "requestqueue.h"
66 #include "util.h"
67 #include "dir.h"
68 #include "member.h"
69 #include "lockspace.h"
70 #include "ast.h"
71 #include "lock.h"
72 #include "rcom.h"
73 #include "recover.h"
74 #include "lvb_table.h"
75 #include "user.h"
76 #include "config.h"
77
78 static int send_request(struct dlm_rsb *r, struct dlm_lkb *lkb);
79 static int send_convert(struct dlm_rsb *r, struct dlm_lkb *lkb);
80 static int send_unlock(struct dlm_rsb *r, struct dlm_lkb *lkb);
81 static int send_cancel(struct dlm_rsb *r, struct dlm_lkb *lkb);
82 static int send_grant(struct dlm_rsb *r, struct dlm_lkb *lkb);
83 static int send_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int mode);
84 static int send_lookup(struct dlm_rsb *r, struct dlm_lkb *lkb);
85 static int send_remove(struct dlm_rsb *r);
86 static int _request_lock(struct dlm_rsb *r, struct dlm_lkb *lkb);
87 static int _cancel_lock(struct dlm_rsb *r, struct dlm_lkb *lkb);
88 static void __receive_convert_reply(struct dlm_rsb *r, struct dlm_lkb *lkb,
89 struct dlm_message *ms);
90 static int receive_extralen(struct dlm_message *ms);
91 static void do_purge(struct dlm_ls *ls, int nodeid, int pid);
92 static void del_timeout(struct dlm_lkb *lkb);
93
94 /*
95 * Lock compatibilty matrix - thanks Steve
96 * UN = Unlocked state. Not really a state, used as a flag
97 * PD = Padding. Used to make the matrix a nice power of two in size
98 * Other states are the same as the VMS DLM.
99 * Usage: matrix[grmode+1][rqmode+1] (although m[rq+1][gr+1] is the same)
100 */
101
102 static const int __dlm_compat_matrix[8][8] = {
103 /* UN NL CR CW PR PW EX PD */
104 {1, 1, 1, 1, 1, 1, 1, 0}, /* UN */
105 {1, 1, 1, 1, 1, 1, 1, 0}, /* NL */
106 {1, 1, 1, 1, 1, 1, 0, 0}, /* CR */
107 {1, 1, 1, 1, 0, 0, 0, 0}, /* CW */
108 {1, 1, 1, 0, 1, 0, 0, 0}, /* PR */
109 {1, 1, 1, 0, 0, 0, 0, 0}, /* PW */
110 {1, 1, 0, 0, 0, 0, 0, 0}, /* EX */
111 {0, 0, 0, 0, 0, 0, 0, 0} /* PD */
112 };
113
114 /*
115 * This defines the direction of transfer of LVB data.
116 * Granted mode is the row; requested mode is the column.
117 * Usage: matrix[grmode+1][rqmode+1]
118 * 1 = LVB is returned to the caller
119 * 0 = LVB is written to the resource
120 * -1 = nothing happens to the LVB
121 */
122
123 const int dlm_lvb_operations[8][8] = {
124 /* UN NL CR CW PR PW EX PD*/
125 { -1, 1, 1, 1, 1, 1, 1, -1 }, /* UN */
126 { -1, 1, 1, 1, 1, 1, 1, 0 }, /* NL */
127 { -1, -1, 1, 1, 1, 1, 1, 0 }, /* CR */
128 { -1, -1, -1, 1, 1, 1, 1, 0 }, /* CW */
129 { -1, -1, -1, -1, 1, 1, 1, 0 }, /* PR */
130 { -1, 0, 0, 0, 0, 0, 1, 0 }, /* PW */
131 { -1, 0, 0, 0, 0, 0, 0, 0 }, /* EX */
132 { -1, 0, 0, 0, 0, 0, 0, 0 } /* PD */
133 };
134
135 #define modes_compat(gr, rq) \
136 __dlm_compat_matrix[(gr)->lkb_grmode + 1][(rq)->lkb_rqmode + 1]
137
dlm_modes_compat(int mode1,int mode2)138 int dlm_modes_compat(int mode1, int mode2)
139 {
140 return __dlm_compat_matrix[mode1 + 1][mode2 + 1];
141 }
142
143 /*
144 * Compatibility matrix for conversions with QUECVT set.
145 * Granted mode is the row; requested mode is the column.
146 * Usage: matrix[grmode+1][rqmode+1]
147 */
148
149 static const int __quecvt_compat_matrix[8][8] = {
150 /* UN NL CR CW PR PW EX PD */
151 {0, 0, 0, 0, 0, 0, 0, 0}, /* UN */
152 {0, 0, 1, 1, 1, 1, 1, 0}, /* NL */
153 {0, 0, 0, 1, 1, 1, 1, 0}, /* CR */
154 {0, 0, 0, 0, 1, 1, 1, 0}, /* CW */
155 {0, 0, 0, 1, 0, 1, 1, 0}, /* PR */
156 {0, 0, 0, 0, 0, 0, 1, 0}, /* PW */
157 {0, 0, 0, 0, 0, 0, 0, 0}, /* EX */
158 {0, 0, 0, 0, 0, 0, 0, 0} /* PD */
159 };
160
dlm_print_lkb(struct dlm_lkb * lkb)161 void dlm_print_lkb(struct dlm_lkb *lkb)
162 {
163 printk(KERN_ERR "lkb: nodeid %d id %x remid %x exflags %x flags %x\n"
164 " status %d rqmode %d grmode %d wait_type %d\n",
165 lkb->lkb_nodeid, lkb->lkb_id, lkb->lkb_remid, lkb->lkb_exflags,
166 lkb->lkb_flags, lkb->lkb_status, lkb->lkb_rqmode,
167 lkb->lkb_grmode, lkb->lkb_wait_type);
168 }
169
dlm_print_rsb(struct dlm_rsb * r)170 static void dlm_print_rsb(struct dlm_rsb *r)
171 {
172 printk(KERN_ERR "rsb: nodeid %d flags %lx first %x rlc %d name %s\n",
173 r->res_nodeid, r->res_flags, r->res_first_lkid,
174 r->res_recover_locks_count, r->res_name);
175 }
176
dlm_dump_rsb(struct dlm_rsb * r)177 void dlm_dump_rsb(struct dlm_rsb *r)
178 {
179 struct dlm_lkb *lkb;
180
181 dlm_print_rsb(r);
182
183 printk(KERN_ERR "rsb: root_list empty %d recover_list empty %d\n",
184 list_empty(&r->res_root_list), list_empty(&r->res_recover_list));
185 printk(KERN_ERR "rsb lookup list\n");
186 list_for_each_entry(lkb, &r->res_lookup, lkb_rsb_lookup)
187 dlm_print_lkb(lkb);
188 printk(KERN_ERR "rsb grant queue:\n");
189 list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue)
190 dlm_print_lkb(lkb);
191 printk(KERN_ERR "rsb convert queue:\n");
192 list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue)
193 dlm_print_lkb(lkb);
194 printk(KERN_ERR "rsb wait queue:\n");
195 list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue)
196 dlm_print_lkb(lkb);
197 }
198
199 /* Threads cannot use the lockspace while it's being recovered */
200
dlm_lock_recovery(struct dlm_ls * ls)201 static inline void dlm_lock_recovery(struct dlm_ls *ls)
202 {
203 down_read(&ls->ls_in_recovery);
204 }
205
dlm_unlock_recovery(struct dlm_ls * ls)206 void dlm_unlock_recovery(struct dlm_ls *ls)
207 {
208 up_read(&ls->ls_in_recovery);
209 }
210
dlm_lock_recovery_try(struct dlm_ls * ls)211 int dlm_lock_recovery_try(struct dlm_ls *ls)
212 {
213 return down_read_trylock(&ls->ls_in_recovery);
214 }
215
can_be_queued(struct dlm_lkb * lkb)216 static inline int can_be_queued(struct dlm_lkb *lkb)
217 {
218 return !(lkb->lkb_exflags & DLM_LKF_NOQUEUE);
219 }
220
force_blocking_asts(struct dlm_lkb * lkb)221 static inline int force_blocking_asts(struct dlm_lkb *lkb)
222 {
223 return (lkb->lkb_exflags & DLM_LKF_NOQUEUEBAST);
224 }
225
is_demoted(struct dlm_lkb * lkb)226 static inline int is_demoted(struct dlm_lkb *lkb)
227 {
228 return (lkb->lkb_sbflags & DLM_SBF_DEMOTED);
229 }
230
is_altmode(struct dlm_lkb * lkb)231 static inline int is_altmode(struct dlm_lkb *lkb)
232 {
233 return (lkb->lkb_sbflags & DLM_SBF_ALTMODE);
234 }
235
is_granted(struct dlm_lkb * lkb)236 static inline int is_granted(struct dlm_lkb *lkb)
237 {
238 return (lkb->lkb_status == DLM_LKSTS_GRANTED);
239 }
240
is_remote(struct dlm_rsb * r)241 static inline int is_remote(struct dlm_rsb *r)
242 {
243 DLM_ASSERT(r->res_nodeid >= 0, dlm_print_rsb(r););
244 return !!r->res_nodeid;
245 }
246
is_process_copy(struct dlm_lkb * lkb)247 static inline int is_process_copy(struct dlm_lkb *lkb)
248 {
249 return (lkb->lkb_nodeid && !(lkb->lkb_flags & DLM_IFL_MSTCPY));
250 }
251
is_master_copy(struct dlm_lkb * lkb)252 static inline int is_master_copy(struct dlm_lkb *lkb)
253 {
254 if (lkb->lkb_flags & DLM_IFL_MSTCPY)
255 DLM_ASSERT(lkb->lkb_nodeid, dlm_print_lkb(lkb););
256 return (lkb->lkb_flags & DLM_IFL_MSTCPY) ? 1 : 0;
257 }
258
middle_conversion(struct dlm_lkb * lkb)259 static inline int middle_conversion(struct dlm_lkb *lkb)
260 {
261 if ((lkb->lkb_grmode==DLM_LOCK_PR && lkb->lkb_rqmode==DLM_LOCK_CW) ||
262 (lkb->lkb_rqmode==DLM_LOCK_PR && lkb->lkb_grmode==DLM_LOCK_CW))
263 return 1;
264 return 0;
265 }
266
down_conversion(struct dlm_lkb * lkb)267 static inline int down_conversion(struct dlm_lkb *lkb)
268 {
269 return (!middle_conversion(lkb) && lkb->lkb_rqmode < lkb->lkb_grmode);
270 }
271
is_overlap_unlock(struct dlm_lkb * lkb)272 static inline int is_overlap_unlock(struct dlm_lkb *lkb)
273 {
274 return lkb->lkb_flags & DLM_IFL_OVERLAP_UNLOCK;
275 }
276
is_overlap_cancel(struct dlm_lkb * lkb)277 static inline int is_overlap_cancel(struct dlm_lkb *lkb)
278 {
279 return lkb->lkb_flags & DLM_IFL_OVERLAP_CANCEL;
280 }
281
is_overlap(struct dlm_lkb * lkb)282 static inline int is_overlap(struct dlm_lkb *lkb)
283 {
284 return (lkb->lkb_flags & (DLM_IFL_OVERLAP_UNLOCK |
285 DLM_IFL_OVERLAP_CANCEL));
286 }
287
queue_cast(struct dlm_rsb * r,struct dlm_lkb * lkb,int rv)288 static void queue_cast(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
289 {
290 if (is_master_copy(lkb))
291 return;
292
293 del_timeout(lkb);
294
295 DLM_ASSERT(lkb->lkb_lksb, dlm_print_lkb(lkb););
296
297 /* if the operation was a cancel, then return -DLM_ECANCEL, if a
298 timeout caused the cancel then return -ETIMEDOUT */
299 if (rv == -DLM_ECANCEL && (lkb->lkb_flags & DLM_IFL_TIMEOUT_CANCEL)) {
300 lkb->lkb_flags &= ~DLM_IFL_TIMEOUT_CANCEL;
301 rv = -ETIMEDOUT;
302 }
303
304 if (rv == -DLM_ECANCEL && (lkb->lkb_flags & DLM_IFL_DEADLOCK_CANCEL)) {
305 lkb->lkb_flags &= ~DLM_IFL_DEADLOCK_CANCEL;
306 rv = -EDEADLK;
307 }
308
309 dlm_add_cb(lkb, DLM_CB_CAST, lkb->lkb_grmode, rv, lkb->lkb_sbflags);
310 }
311
queue_cast_overlap(struct dlm_rsb * r,struct dlm_lkb * lkb)312 static inline void queue_cast_overlap(struct dlm_rsb *r, struct dlm_lkb *lkb)
313 {
314 queue_cast(r, lkb,
315 is_overlap_unlock(lkb) ? -DLM_EUNLOCK : -DLM_ECANCEL);
316 }
317
queue_bast(struct dlm_rsb * r,struct dlm_lkb * lkb,int rqmode)318 static void queue_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int rqmode)
319 {
320 if (is_master_copy(lkb)) {
321 send_bast(r, lkb, rqmode);
322 } else {
323 dlm_add_cb(lkb, DLM_CB_BAST, rqmode, 0, 0);
324 }
325 }
326
327 /*
328 * Basic operations on rsb's and lkb's
329 */
330
pre_rsb_struct(struct dlm_ls * ls)331 static int pre_rsb_struct(struct dlm_ls *ls)
332 {
333 struct dlm_rsb *r1, *r2;
334 int count = 0;
335
336 spin_lock(&ls->ls_new_rsb_spin);
337 if (ls->ls_new_rsb_count > dlm_config.ci_new_rsb_count / 2) {
338 spin_unlock(&ls->ls_new_rsb_spin);
339 return 0;
340 }
341 spin_unlock(&ls->ls_new_rsb_spin);
342
343 r1 = dlm_allocate_rsb(ls);
344 r2 = dlm_allocate_rsb(ls);
345
346 spin_lock(&ls->ls_new_rsb_spin);
347 if (r1) {
348 list_add(&r1->res_hashchain, &ls->ls_new_rsb);
349 ls->ls_new_rsb_count++;
350 }
351 if (r2) {
352 list_add(&r2->res_hashchain, &ls->ls_new_rsb);
353 ls->ls_new_rsb_count++;
354 }
355 count = ls->ls_new_rsb_count;
356 spin_unlock(&ls->ls_new_rsb_spin);
357
358 if (!count)
359 return -ENOMEM;
360 return 0;
361 }
362
363 /* If ls->ls_new_rsb is empty, return -EAGAIN, so the caller can
364 unlock any spinlocks, go back and call pre_rsb_struct again.
365 Otherwise, take an rsb off the list and return it. */
366
get_rsb_struct(struct dlm_ls * ls,char * name,int len,struct dlm_rsb ** r_ret)367 static int get_rsb_struct(struct dlm_ls *ls, char *name, int len,
368 struct dlm_rsb **r_ret)
369 {
370 struct dlm_rsb *r;
371 int count;
372
373 spin_lock(&ls->ls_new_rsb_spin);
374 if (list_empty(&ls->ls_new_rsb)) {
375 count = ls->ls_new_rsb_count;
376 spin_unlock(&ls->ls_new_rsb_spin);
377 log_debug(ls, "find_rsb retry %d %d %s",
378 count, dlm_config.ci_new_rsb_count, name);
379 return -EAGAIN;
380 }
381
382 r = list_first_entry(&ls->ls_new_rsb, struct dlm_rsb, res_hashchain);
383 list_del(&r->res_hashchain);
384 /* Convert the empty list_head to a NULL rb_node for tree usage: */
385 memset(&r->res_hashnode, 0, sizeof(struct rb_node));
386 ls->ls_new_rsb_count--;
387 spin_unlock(&ls->ls_new_rsb_spin);
388
389 r->res_ls = ls;
390 r->res_length = len;
391 memcpy(r->res_name, name, len);
392 mutex_init(&r->res_mutex);
393
394 INIT_LIST_HEAD(&r->res_lookup);
395 INIT_LIST_HEAD(&r->res_grantqueue);
396 INIT_LIST_HEAD(&r->res_convertqueue);
397 INIT_LIST_HEAD(&r->res_waitqueue);
398 INIT_LIST_HEAD(&r->res_root_list);
399 INIT_LIST_HEAD(&r->res_recover_list);
400
401 *r_ret = r;
402 return 0;
403 }
404
rsb_cmp(struct dlm_rsb * r,const char * name,int nlen)405 static int rsb_cmp(struct dlm_rsb *r, const char *name, int nlen)
406 {
407 char maxname[DLM_RESNAME_MAXLEN];
408
409 memset(maxname, 0, DLM_RESNAME_MAXLEN);
410 memcpy(maxname, name, nlen);
411 return memcmp(r->res_name, maxname, DLM_RESNAME_MAXLEN);
412 }
413
dlm_search_rsb_tree(struct rb_root * tree,char * name,int len,unsigned int flags,struct dlm_rsb ** r_ret)414 int dlm_search_rsb_tree(struct rb_root *tree, char *name, int len,
415 unsigned int flags, struct dlm_rsb **r_ret)
416 {
417 struct rb_node *node = tree->rb_node;
418 struct dlm_rsb *r;
419 int error = 0;
420 int rc;
421
422 while (node) {
423 r = rb_entry(node, struct dlm_rsb, res_hashnode);
424 rc = rsb_cmp(r, name, len);
425 if (rc < 0)
426 node = node->rb_left;
427 else if (rc > 0)
428 node = node->rb_right;
429 else
430 goto found;
431 }
432 *r_ret = NULL;
433 return -EBADR;
434
435 found:
436 if (r->res_nodeid && (flags & R_MASTER))
437 error = -ENOTBLK;
438 *r_ret = r;
439 return error;
440 }
441
rsb_insert(struct dlm_rsb * rsb,struct rb_root * tree)442 static int rsb_insert(struct dlm_rsb *rsb, struct rb_root *tree)
443 {
444 struct rb_node **newn = &tree->rb_node;
445 struct rb_node *parent = NULL;
446 int rc;
447
448 while (*newn) {
449 struct dlm_rsb *cur = rb_entry(*newn, struct dlm_rsb,
450 res_hashnode);
451
452 parent = *newn;
453 rc = rsb_cmp(cur, rsb->res_name, rsb->res_length);
454 if (rc < 0)
455 newn = &parent->rb_left;
456 else if (rc > 0)
457 newn = &parent->rb_right;
458 else {
459 log_print("rsb_insert match");
460 dlm_dump_rsb(rsb);
461 dlm_dump_rsb(cur);
462 return -EEXIST;
463 }
464 }
465
466 rb_link_node(&rsb->res_hashnode, parent, newn);
467 rb_insert_color(&rsb->res_hashnode, tree);
468 return 0;
469 }
470
_search_rsb(struct dlm_ls * ls,char * name,int len,int b,unsigned int flags,struct dlm_rsb ** r_ret)471 static int _search_rsb(struct dlm_ls *ls, char *name, int len, int b,
472 unsigned int flags, struct dlm_rsb **r_ret)
473 {
474 struct dlm_rsb *r;
475 int error;
476
477 error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, flags, &r);
478 if (!error) {
479 kref_get(&r->res_ref);
480 goto out;
481 }
482 error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, flags, &r);
483 if (error)
484 goto out;
485
486 rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
487 error = rsb_insert(r, &ls->ls_rsbtbl[b].keep);
488 if (error)
489 return error;
490
491 if (dlm_no_directory(ls))
492 goto out;
493
494 if (r->res_nodeid == -1) {
495 rsb_clear_flag(r, RSB_MASTER_UNCERTAIN);
496 r->res_first_lkid = 0;
497 } else if (r->res_nodeid > 0) {
498 rsb_set_flag(r, RSB_MASTER_UNCERTAIN);
499 r->res_first_lkid = 0;
500 } else {
501 DLM_ASSERT(r->res_nodeid == 0, dlm_print_rsb(r););
502 DLM_ASSERT(!rsb_flag(r, RSB_MASTER_UNCERTAIN),);
503 }
504 out:
505 *r_ret = r;
506 return error;
507 }
508
509 /*
510 * Find rsb in rsbtbl and potentially create/add one
511 *
512 * Delaying the release of rsb's has a similar benefit to applications keeping
513 * NL locks on an rsb, but without the guarantee that the cached master value
514 * will still be valid when the rsb is reused. Apps aren't always smart enough
515 * to keep NL locks on an rsb that they may lock again shortly; this can lead
516 * to excessive master lookups and removals if we don't delay the release.
517 *
518 * Searching for an rsb means looking through both the normal list and toss
519 * list. When found on the toss list the rsb is moved to the normal list with
520 * ref count of 1; when found on normal list the ref count is incremented.
521 */
522
find_rsb(struct dlm_ls * ls,char * name,int namelen,unsigned int flags,struct dlm_rsb ** r_ret)523 static int find_rsb(struct dlm_ls *ls, char *name, int namelen,
524 unsigned int flags, struct dlm_rsb **r_ret)
525 {
526 struct dlm_rsb *r = NULL;
527 uint32_t hash, bucket;
528 int error;
529
530 if (namelen > DLM_RESNAME_MAXLEN) {
531 error = -EINVAL;
532 goto out;
533 }
534
535 if (dlm_no_directory(ls))
536 flags |= R_CREATE;
537
538 hash = jhash(name, namelen, 0);
539 bucket = hash & (ls->ls_rsbtbl_size - 1);
540
541 retry:
542 if (flags & R_CREATE) {
543 error = pre_rsb_struct(ls);
544 if (error < 0)
545 goto out;
546 }
547
548 spin_lock(&ls->ls_rsbtbl[bucket].lock);
549
550 error = _search_rsb(ls, name, namelen, bucket, flags, &r);
551 if (!error)
552 goto out_unlock;
553
554 if (error == -EBADR && !(flags & R_CREATE))
555 goto out_unlock;
556
557 /* the rsb was found but wasn't a master copy */
558 if (error == -ENOTBLK)
559 goto out_unlock;
560
561 error = get_rsb_struct(ls, name, namelen, &r);
562 if (error == -EAGAIN) {
563 spin_unlock(&ls->ls_rsbtbl[bucket].lock);
564 goto retry;
565 }
566 if (error)
567 goto out_unlock;
568
569 r->res_hash = hash;
570 r->res_bucket = bucket;
571 r->res_nodeid = -1;
572 kref_init(&r->res_ref);
573
574 /* With no directory, the master can be set immediately */
575 if (dlm_no_directory(ls)) {
576 int nodeid = dlm_dir_nodeid(r);
577 if (nodeid == dlm_our_nodeid())
578 nodeid = 0;
579 r->res_nodeid = nodeid;
580 }
581 error = rsb_insert(r, &ls->ls_rsbtbl[bucket].keep);
582 out_unlock:
583 spin_unlock(&ls->ls_rsbtbl[bucket].lock);
584 out:
585 *r_ret = r;
586 return error;
587 }
588
589 /* This is only called to add a reference when the code already holds
590 a valid reference to the rsb, so there's no need for locking. */
591
hold_rsb(struct dlm_rsb * r)592 static inline void hold_rsb(struct dlm_rsb *r)
593 {
594 kref_get(&r->res_ref);
595 }
596
dlm_hold_rsb(struct dlm_rsb * r)597 void dlm_hold_rsb(struct dlm_rsb *r)
598 {
599 hold_rsb(r);
600 }
601
toss_rsb(struct kref * kref)602 static void toss_rsb(struct kref *kref)
603 {
604 struct dlm_rsb *r = container_of(kref, struct dlm_rsb, res_ref);
605 struct dlm_ls *ls = r->res_ls;
606
607 DLM_ASSERT(list_empty(&r->res_root_list), dlm_print_rsb(r););
608 kref_init(&r->res_ref);
609 rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[r->res_bucket].keep);
610 rsb_insert(r, &ls->ls_rsbtbl[r->res_bucket].toss);
611 r->res_toss_time = jiffies;
612 if (r->res_lvbptr) {
613 dlm_free_lvb(r->res_lvbptr);
614 r->res_lvbptr = NULL;
615 }
616 }
617
618 /* When all references to the rsb are gone it's transferred to
619 the tossed list for later disposal. */
620
put_rsb(struct dlm_rsb * r)621 static void put_rsb(struct dlm_rsb *r)
622 {
623 struct dlm_ls *ls = r->res_ls;
624 uint32_t bucket = r->res_bucket;
625
626 spin_lock(&ls->ls_rsbtbl[bucket].lock);
627 kref_put(&r->res_ref, toss_rsb);
628 spin_unlock(&ls->ls_rsbtbl[bucket].lock);
629 }
630
dlm_put_rsb(struct dlm_rsb * r)631 void dlm_put_rsb(struct dlm_rsb *r)
632 {
633 put_rsb(r);
634 }
635
636 /* See comment for unhold_lkb */
637
unhold_rsb(struct dlm_rsb * r)638 static void unhold_rsb(struct dlm_rsb *r)
639 {
640 int rv;
641 rv = kref_put(&r->res_ref, toss_rsb);
642 DLM_ASSERT(!rv, dlm_dump_rsb(r););
643 }
644
kill_rsb(struct kref * kref)645 static void kill_rsb(struct kref *kref)
646 {
647 struct dlm_rsb *r = container_of(kref, struct dlm_rsb, res_ref);
648
649 /* All work is done after the return from kref_put() so we
650 can release the write_lock before the remove and free. */
651
652 DLM_ASSERT(list_empty(&r->res_lookup), dlm_dump_rsb(r););
653 DLM_ASSERT(list_empty(&r->res_grantqueue), dlm_dump_rsb(r););
654 DLM_ASSERT(list_empty(&r->res_convertqueue), dlm_dump_rsb(r););
655 DLM_ASSERT(list_empty(&r->res_waitqueue), dlm_dump_rsb(r););
656 DLM_ASSERT(list_empty(&r->res_root_list), dlm_dump_rsb(r););
657 DLM_ASSERT(list_empty(&r->res_recover_list), dlm_dump_rsb(r););
658 }
659
660 /* Attaching/detaching lkb's from rsb's is for rsb reference counting.
661 The rsb must exist as long as any lkb's for it do. */
662
attach_lkb(struct dlm_rsb * r,struct dlm_lkb * lkb)663 static void attach_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb)
664 {
665 hold_rsb(r);
666 lkb->lkb_resource = r;
667 }
668
detach_lkb(struct dlm_lkb * lkb)669 static void detach_lkb(struct dlm_lkb *lkb)
670 {
671 if (lkb->lkb_resource) {
672 put_rsb(lkb->lkb_resource);
673 lkb->lkb_resource = NULL;
674 }
675 }
676
create_lkb(struct dlm_ls * ls,struct dlm_lkb ** lkb_ret)677 static int create_lkb(struct dlm_ls *ls, struct dlm_lkb **lkb_ret)
678 {
679 struct dlm_lkb *lkb;
680 int rv, id;
681
682 lkb = dlm_allocate_lkb(ls);
683 if (!lkb)
684 return -ENOMEM;
685
686 lkb->lkb_nodeid = -1;
687 lkb->lkb_grmode = DLM_LOCK_IV;
688 kref_init(&lkb->lkb_ref);
689 INIT_LIST_HEAD(&lkb->lkb_ownqueue);
690 INIT_LIST_HEAD(&lkb->lkb_rsb_lookup);
691 INIT_LIST_HEAD(&lkb->lkb_time_list);
692 INIT_LIST_HEAD(&lkb->lkb_cb_list);
693 mutex_init(&lkb->lkb_cb_mutex);
694 INIT_WORK(&lkb->lkb_cb_work, dlm_callback_work);
695
696 retry:
697 rv = idr_pre_get(&ls->ls_lkbidr, GFP_NOFS);
698 if (!rv)
699 return -ENOMEM;
700
701 spin_lock(&ls->ls_lkbidr_spin);
702 rv = idr_get_new_above(&ls->ls_lkbidr, lkb, 1, &id);
703 if (!rv)
704 lkb->lkb_id = id;
705 spin_unlock(&ls->ls_lkbidr_spin);
706
707 if (rv == -EAGAIN)
708 goto retry;
709
710 if (rv < 0) {
711 log_error(ls, "create_lkb idr error %d", rv);
712 return rv;
713 }
714
715 *lkb_ret = lkb;
716 return 0;
717 }
718
find_lkb(struct dlm_ls * ls,uint32_t lkid,struct dlm_lkb ** lkb_ret)719 static int find_lkb(struct dlm_ls *ls, uint32_t lkid, struct dlm_lkb **lkb_ret)
720 {
721 struct dlm_lkb *lkb;
722
723 spin_lock(&ls->ls_lkbidr_spin);
724 lkb = idr_find(&ls->ls_lkbidr, lkid);
725 if (lkb)
726 kref_get(&lkb->lkb_ref);
727 spin_unlock(&ls->ls_lkbidr_spin);
728
729 *lkb_ret = lkb;
730 return lkb ? 0 : -ENOENT;
731 }
732
kill_lkb(struct kref * kref)733 static void kill_lkb(struct kref *kref)
734 {
735 struct dlm_lkb *lkb = container_of(kref, struct dlm_lkb, lkb_ref);
736
737 /* All work is done after the return from kref_put() so we
738 can release the write_lock before the detach_lkb */
739
740 DLM_ASSERT(!lkb->lkb_status, dlm_print_lkb(lkb););
741 }
742
743 /* __put_lkb() is used when an lkb may not have an rsb attached to
744 it so we need to provide the lockspace explicitly */
745
__put_lkb(struct dlm_ls * ls,struct dlm_lkb * lkb)746 static int __put_lkb(struct dlm_ls *ls, struct dlm_lkb *lkb)
747 {
748 uint32_t lkid = lkb->lkb_id;
749
750 spin_lock(&ls->ls_lkbidr_spin);
751 if (kref_put(&lkb->lkb_ref, kill_lkb)) {
752 idr_remove(&ls->ls_lkbidr, lkid);
753 spin_unlock(&ls->ls_lkbidr_spin);
754
755 detach_lkb(lkb);
756
757 /* for local/process lkbs, lvbptr points to caller's lksb */
758 if (lkb->lkb_lvbptr && is_master_copy(lkb))
759 dlm_free_lvb(lkb->lkb_lvbptr);
760 dlm_free_lkb(lkb);
761 return 1;
762 } else {
763 spin_unlock(&ls->ls_lkbidr_spin);
764 return 0;
765 }
766 }
767
dlm_put_lkb(struct dlm_lkb * lkb)768 int dlm_put_lkb(struct dlm_lkb *lkb)
769 {
770 struct dlm_ls *ls;
771
772 DLM_ASSERT(lkb->lkb_resource, dlm_print_lkb(lkb););
773 DLM_ASSERT(lkb->lkb_resource->res_ls, dlm_print_lkb(lkb););
774
775 ls = lkb->lkb_resource->res_ls;
776 return __put_lkb(ls, lkb);
777 }
778
779 /* This is only called to add a reference when the code already holds
780 a valid reference to the lkb, so there's no need for locking. */
781
hold_lkb(struct dlm_lkb * lkb)782 static inline void hold_lkb(struct dlm_lkb *lkb)
783 {
784 kref_get(&lkb->lkb_ref);
785 }
786
787 /* This is called when we need to remove a reference and are certain
788 it's not the last ref. e.g. del_lkb is always called between a
789 find_lkb/put_lkb and is always the inverse of a previous add_lkb.
790 put_lkb would work fine, but would involve unnecessary locking */
791
unhold_lkb(struct dlm_lkb * lkb)792 static inline void unhold_lkb(struct dlm_lkb *lkb)
793 {
794 int rv;
795 rv = kref_put(&lkb->lkb_ref, kill_lkb);
796 DLM_ASSERT(!rv, dlm_print_lkb(lkb););
797 }
798
lkb_add_ordered(struct list_head * new,struct list_head * head,int mode)799 static void lkb_add_ordered(struct list_head *new, struct list_head *head,
800 int mode)
801 {
802 struct dlm_lkb *lkb = NULL;
803
804 list_for_each_entry(lkb, head, lkb_statequeue)
805 if (lkb->lkb_rqmode < mode)
806 break;
807
808 __list_add(new, lkb->lkb_statequeue.prev, &lkb->lkb_statequeue);
809 }
810
811 /* add/remove lkb to rsb's grant/convert/wait queue */
812
add_lkb(struct dlm_rsb * r,struct dlm_lkb * lkb,int status)813 static void add_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb, int status)
814 {
815 kref_get(&lkb->lkb_ref);
816
817 DLM_ASSERT(!lkb->lkb_status, dlm_print_lkb(lkb););
818
819 lkb->lkb_timestamp = ktime_get();
820
821 lkb->lkb_status = status;
822
823 switch (status) {
824 case DLM_LKSTS_WAITING:
825 if (lkb->lkb_exflags & DLM_LKF_HEADQUE)
826 list_add(&lkb->lkb_statequeue, &r->res_waitqueue);
827 else
828 list_add_tail(&lkb->lkb_statequeue, &r->res_waitqueue);
829 break;
830 case DLM_LKSTS_GRANTED:
831 /* convention says granted locks kept in order of grmode */
832 lkb_add_ordered(&lkb->lkb_statequeue, &r->res_grantqueue,
833 lkb->lkb_grmode);
834 break;
835 case DLM_LKSTS_CONVERT:
836 if (lkb->lkb_exflags & DLM_LKF_HEADQUE)
837 list_add(&lkb->lkb_statequeue, &r->res_convertqueue);
838 else
839 list_add_tail(&lkb->lkb_statequeue,
840 &r->res_convertqueue);
841 break;
842 default:
843 DLM_ASSERT(0, dlm_print_lkb(lkb); printk("sts=%d\n", status););
844 }
845 }
846
del_lkb(struct dlm_rsb * r,struct dlm_lkb * lkb)847 static void del_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb)
848 {
849 lkb->lkb_status = 0;
850 list_del(&lkb->lkb_statequeue);
851 unhold_lkb(lkb);
852 }
853
move_lkb(struct dlm_rsb * r,struct dlm_lkb * lkb,int sts)854 static void move_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb, int sts)
855 {
856 hold_lkb(lkb);
857 del_lkb(r, lkb);
858 add_lkb(r, lkb, sts);
859 unhold_lkb(lkb);
860 }
861
msg_reply_type(int mstype)862 static int msg_reply_type(int mstype)
863 {
864 switch (mstype) {
865 case DLM_MSG_REQUEST:
866 return DLM_MSG_REQUEST_REPLY;
867 case DLM_MSG_CONVERT:
868 return DLM_MSG_CONVERT_REPLY;
869 case DLM_MSG_UNLOCK:
870 return DLM_MSG_UNLOCK_REPLY;
871 case DLM_MSG_CANCEL:
872 return DLM_MSG_CANCEL_REPLY;
873 case DLM_MSG_LOOKUP:
874 return DLM_MSG_LOOKUP_REPLY;
875 }
876 return -1;
877 }
878
nodeid_warned(int nodeid,int num_nodes,int * warned)879 static int nodeid_warned(int nodeid, int num_nodes, int *warned)
880 {
881 int i;
882
883 for (i = 0; i < num_nodes; i++) {
884 if (!warned[i]) {
885 warned[i] = nodeid;
886 return 0;
887 }
888 if (warned[i] == nodeid)
889 return 1;
890 }
891 return 0;
892 }
893
dlm_scan_waiters(struct dlm_ls * ls)894 void dlm_scan_waiters(struct dlm_ls *ls)
895 {
896 struct dlm_lkb *lkb;
897 ktime_t zero = ktime_set(0, 0);
898 s64 us;
899 s64 debug_maxus = 0;
900 u32 debug_scanned = 0;
901 u32 debug_expired = 0;
902 int num_nodes = 0;
903 int *warned = NULL;
904
905 if (!dlm_config.ci_waitwarn_us)
906 return;
907
908 mutex_lock(&ls->ls_waiters_mutex);
909
910 list_for_each_entry(lkb, &ls->ls_waiters, lkb_wait_reply) {
911 if (ktime_equal(lkb->lkb_wait_time, zero))
912 continue;
913
914 debug_scanned++;
915
916 us = ktime_to_us(ktime_sub(ktime_get(), lkb->lkb_wait_time));
917
918 if (us < dlm_config.ci_waitwarn_us)
919 continue;
920
921 lkb->lkb_wait_time = zero;
922
923 debug_expired++;
924 if (us > debug_maxus)
925 debug_maxus = us;
926
927 if (!num_nodes) {
928 num_nodes = ls->ls_num_nodes;
929 warned = kzalloc(num_nodes * sizeof(int), GFP_KERNEL);
930 }
931 if (!warned)
932 continue;
933 if (nodeid_warned(lkb->lkb_wait_nodeid, num_nodes, warned))
934 continue;
935
936 log_error(ls, "waitwarn %x %lld %d us check connection to "
937 "node %d", lkb->lkb_id, (long long)us,
938 dlm_config.ci_waitwarn_us, lkb->lkb_wait_nodeid);
939 }
940 mutex_unlock(&ls->ls_waiters_mutex);
941 kfree(warned);
942
943 if (debug_expired)
944 log_debug(ls, "scan_waiters %u warn %u over %d us max %lld us",
945 debug_scanned, debug_expired,
946 dlm_config.ci_waitwarn_us, (long long)debug_maxus);
947 }
948
949 /* add/remove lkb from global waiters list of lkb's waiting for
950 a reply from a remote node */
951
add_to_waiters(struct dlm_lkb * lkb,int mstype,int to_nodeid)952 static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)
953 {
954 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
955 int error = 0;
956
957 mutex_lock(&ls->ls_waiters_mutex);
958
959 if (is_overlap_unlock(lkb) ||
960 (is_overlap_cancel(lkb) && (mstype == DLM_MSG_CANCEL))) {
961 error = -EINVAL;
962 goto out;
963 }
964
965 if (lkb->lkb_wait_type || is_overlap_cancel(lkb)) {
966 switch (mstype) {
967 case DLM_MSG_UNLOCK:
968 lkb->lkb_flags |= DLM_IFL_OVERLAP_UNLOCK;
969 break;
970 case DLM_MSG_CANCEL:
971 lkb->lkb_flags |= DLM_IFL_OVERLAP_CANCEL;
972 break;
973 default:
974 error = -EBUSY;
975 goto out;
976 }
977 lkb->lkb_wait_count++;
978 hold_lkb(lkb);
979
980 log_debug(ls, "addwait %x cur %d overlap %d count %d f %x",
981 lkb->lkb_id, lkb->lkb_wait_type, mstype,
982 lkb->lkb_wait_count, lkb->lkb_flags);
983 goto out;
984 }
985
986 DLM_ASSERT(!lkb->lkb_wait_count,
987 dlm_print_lkb(lkb);
988 printk("wait_count %d\n", lkb->lkb_wait_count););
989
990 lkb->lkb_wait_count++;
991 lkb->lkb_wait_type = mstype;
992 lkb->lkb_wait_time = ktime_get();
993 lkb->lkb_wait_nodeid = to_nodeid; /* for debugging */
994 hold_lkb(lkb);
995 list_add(&lkb->lkb_wait_reply, &ls->ls_waiters);
996 out:
997 if (error)
998 log_error(ls, "addwait error %x %d flags %x %d %d %s",
999 lkb->lkb_id, error, lkb->lkb_flags, mstype,
1000 lkb->lkb_wait_type, lkb->lkb_resource->res_name);
1001 mutex_unlock(&ls->ls_waiters_mutex);
1002 return error;
1003 }
1004
1005 /* We clear the RESEND flag because we might be taking an lkb off the waiters
1006 list as part of process_requestqueue (e.g. a lookup that has an optimized
1007 request reply on the requestqueue) between dlm_recover_waiters_pre() which
1008 set RESEND and dlm_recover_waiters_post() */
1009
_remove_from_waiters(struct dlm_lkb * lkb,int mstype,struct dlm_message * ms)1010 static int _remove_from_waiters(struct dlm_lkb *lkb, int mstype,
1011 struct dlm_message *ms)
1012 {
1013 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
1014 int overlap_done = 0;
1015
1016 if (is_overlap_unlock(lkb) && (mstype == DLM_MSG_UNLOCK_REPLY)) {
1017 log_debug(ls, "remwait %x unlock_reply overlap", lkb->lkb_id);
1018 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_UNLOCK;
1019 overlap_done = 1;
1020 goto out_del;
1021 }
1022
1023 if (is_overlap_cancel(lkb) && (mstype == DLM_MSG_CANCEL_REPLY)) {
1024 log_debug(ls, "remwait %x cancel_reply overlap", lkb->lkb_id);
1025 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
1026 overlap_done = 1;
1027 goto out_del;
1028 }
1029
1030 /* Cancel state was preemptively cleared by a successful convert,
1031 see next comment, nothing to do. */
1032
1033 if ((mstype == DLM_MSG_CANCEL_REPLY) &&
1034 (lkb->lkb_wait_type != DLM_MSG_CANCEL)) {
1035 log_debug(ls, "remwait %x cancel_reply wait_type %d",
1036 lkb->lkb_id, lkb->lkb_wait_type);
1037 return -1;
1038 }
1039
1040 /* Remove for the convert reply, and premptively remove for the
1041 cancel reply. A convert has been granted while there's still
1042 an outstanding cancel on it (the cancel is moot and the result
1043 in the cancel reply should be 0). We preempt the cancel reply
1044 because the app gets the convert result and then can follow up
1045 with another op, like convert. This subsequent op would see the
1046 lingering state of the cancel and fail with -EBUSY. */
1047
1048 if ((mstype == DLM_MSG_CONVERT_REPLY) &&
1049 (lkb->lkb_wait_type == DLM_MSG_CONVERT) &&
1050 is_overlap_cancel(lkb) && ms && !ms->m_result) {
1051 log_debug(ls, "remwait %x convert_reply zap overlap_cancel",
1052 lkb->lkb_id);
1053 lkb->lkb_wait_type = 0;
1054 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
1055 lkb->lkb_wait_count--;
1056 goto out_del;
1057 }
1058
1059 /* N.B. type of reply may not always correspond to type of original
1060 msg due to lookup->request optimization, verify others? */
1061
1062 if (lkb->lkb_wait_type) {
1063 lkb->lkb_wait_type = 0;
1064 goto out_del;
1065 }
1066
1067 log_error(ls, "remwait error %x reply %d flags %x no wait_type",
1068 lkb->lkb_id, mstype, lkb->lkb_flags);
1069 return -1;
1070
1071 out_del:
1072 /* the force-unlock/cancel has completed and we haven't recvd a reply
1073 to the op that was in progress prior to the unlock/cancel; we
1074 give up on any reply to the earlier op. FIXME: not sure when/how
1075 this would happen */
1076
1077 if (overlap_done && lkb->lkb_wait_type) {
1078 log_error(ls, "remwait error %x reply %d wait_type %d overlap",
1079 lkb->lkb_id, mstype, lkb->lkb_wait_type);
1080 lkb->lkb_wait_count--;
1081 lkb->lkb_wait_type = 0;
1082 }
1083
1084 DLM_ASSERT(lkb->lkb_wait_count, dlm_print_lkb(lkb););
1085
1086 lkb->lkb_flags &= ~DLM_IFL_RESEND;
1087 lkb->lkb_wait_count--;
1088 if (!lkb->lkb_wait_count)
1089 list_del_init(&lkb->lkb_wait_reply);
1090 unhold_lkb(lkb);
1091 return 0;
1092 }
1093
remove_from_waiters(struct dlm_lkb * lkb,int mstype)1094 static int remove_from_waiters(struct dlm_lkb *lkb, int mstype)
1095 {
1096 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
1097 int error;
1098
1099 mutex_lock(&ls->ls_waiters_mutex);
1100 error = _remove_from_waiters(lkb, mstype, NULL);
1101 mutex_unlock(&ls->ls_waiters_mutex);
1102 return error;
1103 }
1104
1105 /* Handles situations where we might be processing a "fake" or "stub" reply in
1106 which we can't try to take waiters_mutex again. */
1107
remove_from_waiters_ms(struct dlm_lkb * lkb,struct dlm_message * ms)1108 static int remove_from_waiters_ms(struct dlm_lkb *lkb, struct dlm_message *ms)
1109 {
1110 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
1111 int error;
1112
1113 if (ms->m_flags != DLM_IFL_STUB_MS)
1114 mutex_lock(&ls->ls_waiters_mutex);
1115 error = _remove_from_waiters(lkb, ms->m_type, ms);
1116 if (ms->m_flags != DLM_IFL_STUB_MS)
1117 mutex_unlock(&ls->ls_waiters_mutex);
1118 return error;
1119 }
1120
dir_remove(struct dlm_rsb * r)1121 static void dir_remove(struct dlm_rsb *r)
1122 {
1123 int to_nodeid;
1124
1125 if (dlm_no_directory(r->res_ls))
1126 return;
1127
1128 to_nodeid = dlm_dir_nodeid(r);
1129 if (to_nodeid != dlm_our_nodeid())
1130 send_remove(r);
1131 else
1132 dlm_dir_remove_entry(r->res_ls, to_nodeid,
1133 r->res_name, r->res_length);
1134 }
1135
1136 /* FIXME: make this more efficient */
1137
shrink_bucket(struct dlm_ls * ls,int b)1138 static int shrink_bucket(struct dlm_ls *ls, int b)
1139 {
1140 struct rb_node *n;
1141 struct dlm_rsb *r;
1142 int count = 0, found;
1143
1144 for (;;) {
1145 found = 0;
1146 spin_lock(&ls->ls_rsbtbl[b].lock);
1147 for (n = rb_first(&ls->ls_rsbtbl[b].toss); n; n = rb_next(n)) {
1148 r = rb_entry(n, struct dlm_rsb, res_hashnode);
1149 if (!time_after_eq(jiffies, r->res_toss_time +
1150 dlm_config.ci_toss_secs * HZ))
1151 continue;
1152 found = 1;
1153 break;
1154 }
1155
1156 if (!found) {
1157 spin_unlock(&ls->ls_rsbtbl[b].lock);
1158 break;
1159 }
1160
1161 if (kref_put(&r->res_ref, kill_rsb)) {
1162 rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
1163 spin_unlock(&ls->ls_rsbtbl[b].lock);
1164
1165 if (is_master(r))
1166 dir_remove(r);
1167 dlm_free_rsb(r);
1168 count++;
1169 } else {
1170 spin_unlock(&ls->ls_rsbtbl[b].lock);
1171 log_error(ls, "tossed rsb in use %s", r->res_name);
1172 }
1173 }
1174
1175 return count;
1176 }
1177
dlm_scan_rsbs(struct dlm_ls * ls)1178 void dlm_scan_rsbs(struct dlm_ls *ls)
1179 {
1180 int i;
1181
1182 for (i = 0; i < ls->ls_rsbtbl_size; i++) {
1183 shrink_bucket(ls, i);
1184 if (dlm_locking_stopped(ls))
1185 break;
1186 cond_resched();
1187 }
1188 }
1189
add_timeout(struct dlm_lkb * lkb)1190 static void add_timeout(struct dlm_lkb *lkb)
1191 {
1192 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
1193
1194 if (is_master_copy(lkb))
1195 return;
1196
1197 if (test_bit(LSFL_TIMEWARN, &ls->ls_flags) &&
1198 !(lkb->lkb_exflags & DLM_LKF_NODLCKWT)) {
1199 lkb->lkb_flags |= DLM_IFL_WATCH_TIMEWARN;
1200 goto add_it;
1201 }
1202 if (lkb->lkb_exflags & DLM_LKF_TIMEOUT)
1203 goto add_it;
1204 return;
1205
1206 add_it:
1207 DLM_ASSERT(list_empty(&lkb->lkb_time_list), dlm_print_lkb(lkb););
1208 mutex_lock(&ls->ls_timeout_mutex);
1209 hold_lkb(lkb);
1210 list_add_tail(&lkb->lkb_time_list, &ls->ls_timeout);
1211 mutex_unlock(&ls->ls_timeout_mutex);
1212 }
1213
del_timeout(struct dlm_lkb * lkb)1214 static void del_timeout(struct dlm_lkb *lkb)
1215 {
1216 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
1217
1218 mutex_lock(&ls->ls_timeout_mutex);
1219 if (!list_empty(&lkb->lkb_time_list)) {
1220 list_del_init(&lkb->lkb_time_list);
1221 unhold_lkb(lkb);
1222 }
1223 mutex_unlock(&ls->ls_timeout_mutex);
1224 }
1225
1226 /* FIXME: is it safe to look at lkb_exflags, lkb_flags, lkb_timestamp, and
1227 lkb_lksb_timeout without lock_rsb? Note: we can't lock timeout_mutex
1228 and then lock rsb because of lock ordering in add_timeout. We may need
1229 to specify some special timeout-related bits in the lkb that are just to
1230 be accessed under the timeout_mutex. */
1231
dlm_scan_timeout(struct dlm_ls * ls)1232 void dlm_scan_timeout(struct dlm_ls *ls)
1233 {
1234 struct dlm_rsb *r;
1235 struct dlm_lkb *lkb;
1236 int do_cancel, do_warn;
1237 s64 wait_us;
1238
1239 for (;;) {
1240 if (dlm_locking_stopped(ls))
1241 break;
1242
1243 do_cancel = 0;
1244 do_warn = 0;
1245 mutex_lock(&ls->ls_timeout_mutex);
1246 list_for_each_entry(lkb, &ls->ls_timeout, lkb_time_list) {
1247
1248 wait_us = ktime_to_us(ktime_sub(ktime_get(),
1249 lkb->lkb_timestamp));
1250
1251 if ((lkb->lkb_exflags & DLM_LKF_TIMEOUT) &&
1252 wait_us >= (lkb->lkb_timeout_cs * 10000))
1253 do_cancel = 1;
1254
1255 if ((lkb->lkb_flags & DLM_IFL_WATCH_TIMEWARN) &&
1256 wait_us >= dlm_config.ci_timewarn_cs * 10000)
1257 do_warn = 1;
1258
1259 if (!do_cancel && !do_warn)
1260 continue;
1261 hold_lkb(lkb);
1262 break;
1263 }
1264 mutex_unlock(&ls->ls_timeout_mutex);
1265
1266 if (!do_cancel && !do_warn)
1267 break;
1268
1269 r = lkb->lkb_resource;
1270 hold_rsb(r);
1271 lock_rsb(r);
1272
1273 if (do_warn) {
1274 /* clear flag so we only warn once */
1275 lkb->lkb_flags &= ~DLM_IFL_WATCH_TIMEWARN;
1276 if (!(lkb->lkb_exflags & DLM_LKF_TIMEOUT))
1277 del_timeout(lkb);
1278 dlm_timeout_warn(lkb);
1279 }
1280
1281 if (do_cancel) {
1282 log_debug(ls, "timeout cancel %x node %d %s",
1283 lkb->lkb_id, lkb->lkb_nodeid, r->res_name);
1284 lkb->lkb_flags &= ~DLM_IFL_WATCH_TIMEWARN;
1285 lkb->lkb_flags |= DLM_IFL_TIMEOUT_CANCEL;
1286 del_timeout(lkb);
1287 _cancel_lock(r, lkb);
1288 }
1289
1290 unlock_rsb(r);
1291 unhold_rsb(r);
1292 dlm_put_lkb(lkb);
1293 }
1294 }
1295
1296 /* This is only called by dlm_recoverd, and we rely on dlm_ls_stop() stopping
1297 dlm_recoverd before checking/setting ls_recover_begin. */
1298
dlm_adjust_timeouts(struct dlm_ls * ls)1299 void dlm_adjust_timeouts(struct dlm_ls *ls)
1300 {
1301 struct dlm_lkb *lkb;
1302 u64 adj_us = jiffies_to_usecs(jiffies - ls->ls_recover_begin);
1303
1304 ls->ls_recover_begin = 0;
1305 mutex_lock(&ls->ls_timeout_mutex);
1306 list_for_each_entry(lkb, &ls->ls_timeout, lkb_time_list)
1307 lkb->lkb_timestamp = ktime_add_us(lkb->lkb_timestamp, adj_us);
1308 mutex_unlock(&ls->ls_timeout_mutex);
1309
1310 if (!dlm_config.ci_waitwarn_us)
1311 return;
1312
1313 mutex_lock(&ls->ls_waiters_mutex);
1314 list_for_each_entry(lkb, &ls->ls_waiters, lkb_wait_reply) {
1315 if (ktime_to_us(lkb->lkb_wait_time))
1316 lkb->lkb_wait_time = ktime_get();
1317 }
1318 mutex_unlock(&ls->ls_waiters_mutex);
1319 }
1320
1321 /* lkb is master or local copy */
1322
set_lvb_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1323 static void set_lvb_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1324 {
1325 int b, len = r->res_ls->ls_lvblen;
1326
1327 /* b=1 lvb returned to caller
1328 b=0 lvb written to rsb or invalidated
1329 b=-1 do nothing */
1330
1331 b = dlm_lvb_operations[lkb->lkb_grmode + 1][lkb->lkb_rqmode + 1];
1332
1333 if (b == 1) {
1334 if (!lkb->lkb_lvbptr)
1335 return;
1336
1337 if (!(lkb->lkb_exflags & DLM_LKF_VALBLK))
1338 return;
1339
1340 if (!r->res_lvbptr)
1341 return;
1342
1343 memcpy(lkb->lkb_lvbptr, r->res_lvbptr, len);
1344 lkb->lkb_lvbseq = r->res_lvbseq;
1345
1346 } else if (b == 0) {
1347 if (lkb->lkb_exflags & DLM_LKF_IVVALBLK) {
1348 rsb_set_flag(r, RSB_VALNOTVALID);
1349 return;
1350 }
1351
1352 if (!lkb->lkb_lvbptr)
1353 return;
1354
1355 if (!(lkb->lkb_exflags & DLM_LKF_VALBLK))
1356 return;
1357
1358 if (!r->res_lvbptr)
1359 r->res_lvbptr = dlm_allocate_lvb(r->res_ls);
1360
1361 if (!r->res_lvbptr)
1362 return;
1363
1364 memcpy(r->res_lvbptr, lkb->lkb_lvbptr, len);
1365 r->res_lvbseq++;
1366 lkb->lkb_lvbseq = r->res_lvbseq;
1367 rsb_clear_flag(r, RSB_VALNOTVALID);
1368 }
1369
1370 if (rsb_flag(r, RSB_VALNOTVALID))
1371 lkb->lkb_sbflags |= DLM_SBF_VALNOTVALID;
1372 }
1373
set_lvb_unlock(struct dlm_rsb * r,struct dlm_lkb * lkb)1374 static void set_lvb_unlock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1375 {
1376 if (lkb->lkb_grmode < DLM_LOCK_PW)
1377 return;
1378
1379 if (lkb->lkb_exflags & DLM_LKF_IVVALBLK) {
1380 rsb_set_flag(r, RSB_VALNOTVALID);
1381 return;
1382 }
1383
1384 if (!lkb->lkb_lvbptr)
1385 return;
1386
1387 if (!(lkb->lkb_exflags & DLM_LKF_VALBLK))
1388 return;
1389
1390 if (!r->res_lvbptr)
1391 r->res_lvbptr = dlm_allocate_lvb(r->res_ls);
1392
1393 if (!r->res_lvbptr)
1394 return;
1395
1396 memcpy(r->res_lvbptr, lkb->lkb_lvbptr, r->res_ls->ls_lvblen);
1397 r->res_lvbseq++;
1398 rsb_clear_flag(r, RSB_VALNOTVALID);
1399 }
1400
1401 /* lkb is process copy (pc) */
1402
set_lvb_lock_pc(struct dlm_rsb * r,struct dlm_lkb * lkb,struct dlm_message * ms)1403 static void set_lvb_lock_pc(struct dlm_rsb *r, struct dlm_lkb *lkb,
1404 struct dlm_message *ms)
1405 {
1406 int b;
1407
1408 if (!lkb->lkb_lvbptr)
1409 return;
1410
1411 if (!(lkb->lkb_exflags & DLM_LKF_VALBLK))
1412 return;
1413
1414 b = dlm_lvb_operations[lkb->lkb_grmode + 1][lkb->lkb_rqmode + 1];
1415 if (b == 1) {
1416 int len = receive_extralen(ms);
1417 if (len > DLM_RESNAME_MAXLEN)
1418 len = DLM_RESNAME_MAXLEN;
1419 memcpy(lkb->lkb_lvbptr, ms->m_extra, len);
1420 lkb->lkb_lvbseq = ms->m_lvbseq;
1421 }
1422 }
1423
1424 /* Manipulate lkb's on rsb's convert/granted/waiting queues
1425 remove_lock -- used for unlock, removes lkb from granted
1426 revert_lock -- used for cancel, moves lkb from convert to granted
1427 grant_lock -- used for request and convert, adds lkb to granted or
1428 moves lkb from convert or waiting to granted
1429
1430 Each of these is used for master or local copy lkb's. There is
1431 also a _pc() variation used to make the corresponding change on
1432 a process copy (pc) lkb. */
1433
_remove_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1434 static void _remove_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1435 {
1436 del_lkb(r, lkb);
1437 lkb->lkb_grmode = DLM_LOCK_IV;
1438 /* this unhold undoes the original ref from create_lkb()
1439 so this leads to the lkb being freed */
1440 unhold_lkb(lkb);
1441 }
1442
remove_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1443 static void remove_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1444 {
1445 set_lvb_unlock(r, lkb);
1446 _remove_lock(r, lkb);
1447 }
1448
remove_lock_pc(struct dlm_rsb * r,struct dlm_lkb * lkb)1449 static void remove_lock_pc(struct dlm_rsb *r, struct dlm_lkb *lkb)
1450 {
1451 _remove_lock(r, lkb);
1452 }
1453
1454 /* returns: 0 did nothing
1455 1 moved lock to granted
1456 -1 removed lock */
1457
revert_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1458 static int revert_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1459 {
1460 int rv = 0;
1461
1462 lkb->lkb_rqmode = DLM_LOCK_IV;
1463
1464 switch (lkb->lkb_status) {
1465 case DLM_LKSTS_GRANTED:
1466 break;
1467 case DLM_LKSTS_CONVERT:
1468 move_lkb(r, lkb, DLM_LKSTS_GRANTED);
1469 rv = 1;
1470 break;
1471 case DLM_LKSTS_WAITING:
1472 del_lkb(r, lkb);
1473 lkb->lkb_grmode = DLM_LOCK_IV;
1474 /* this unhold undoes the original ref from create_lkb()
1475 so this leads to the lkb being freed */
1476 unhold_lkb(lkb);
1477 rv = -1;
1478 break;
1479 default:
1480 log_print("invalid status for revert %d", lkb->lkb_status);
1481 }
1482 return rv;
1483 }
1484
revert_lock_pc(struct dlm_rsb * r,struct dlm_lkb * lkb)1485 static int revert_lock_pc(struct dlm_rsb *r, struct dlm_lkb *lkb)
1486 {
1487 return revert_lock(r, lkb);
1488 }
1489
_grant_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1490 static void _grant_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1491 {
1492 if (lkb->lkb_grmode != lkb->lkb_rqmode) {
1493 lkb->lkb_grmode = lkb->lkb_rqmode;
1494 if (lkb->lkb_status)
1495 move_lkb(r, lkb, DLM_LKSTS_GRANTED);
1496 else
1497 add_lkb(r, lkb, DLM_LKSTS_GRANTED);
1498 }
1499
1500 lkb->lkb_rqmode = DLM_LOCK_IV;
1501 }
1502
grant_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)1503 static void grant_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
1504 {
1505 set_lvb_lock(r, lkb);
1506 _grant_lock(r, lkb);
1507 lkb->lkb_highbast = 0;
1508 }
1509
grant_lock_pc(struct dlm_rsb * r,struct dlm_lkb * lkb,struct dlm_message * ms)1510 static void grant_lock_pc(struct dlm_rsb *r, struct dlm_lkb *lkb,
1511 struct dlm_message *ms)
1512 {
1513 set_lvb_lock_pc(r, lkb, ms);
1514 _grant_lock(r, lkb);
1515 }
1516
1517 /* called by grant_pending_locks() which means an async grant message must
1518 be sent to the requesting node in addition to granting the lock if the
1519 lkb belongs to a remote node. */
1520
grant_lock_pending(struct dlm_rsb * r,struct dlm_lkb * lkb)1521 static void grant_lock_pending(struct dlm_rsb *r, struct dlm_lkb *lkb)
1522 {
1523 grant_lock(r, lkb);
1524 if (is_master_copy(lkb))
1525 send_grant(r, lkb);
1526 else
1527 queue_cast(r, lkb, 0);
1528 }
1529
1530 /* The special CONVDEADLK, ALTPR and ALTCW flags allow the master to
1531 change the granted/requested modes. We're munging things accordingly in
1532 the process copy.
1533 CONVDEADLK: our grmode may have been forced down to NL to resolve a
1534 conversion deadlock
1535 ALTPR/ALTCW: our rqmode may have been changed to PR or CW to become
1536 compatible with other granted locks */
1537
munge_demoted(struct dlm_lkb * lkb)1538 static void munge_demoted(struct dlm_lkb *lkb)
1539 {
1540 if (lkb->lkb_rqmode == DLM_LOCK_IV || lkb->lkb_grmode == DLM_LOCK_IV) {
1541 log_print("munge_demoted %x invalid modes gr %d rq %d",
1542 lkb->lkb_id, lkb->lkb_grmode, lkb->lkb_rqmode);
1543 return;
1544 }
1545
1546 lkb->lkb_grmode = DLM_LOCK_NL;
1547 }
1548
munge_altmode(struct dlm_lkb * lkb,struct dlm_message * ms)1549 static void munge_altmode(struct dlm_lkb *lkb, struct dlm_message *ms)
1550 {
1551 if (ms->m_type != DLM_MSG_REQUEST_REPLY &&
1552 ms->m_type != DLM_MSG_GRANT) {
1553 log_print("munge_altmode %x invalid reply type %d",
1554 lkb->lkb_id, ms->m_type);
1555 return;
1556 }
1557
1558 if (lkb->lkb_exflags & DLM_LKF_ALTPR)
1559 lkb->lkb_rqmode = DLM_LOCK_PR;
1560 else if (lkb->lkb_exflags & DLM_LKF_ALTCW)
1561 lkb->lkb_rqmode = DLM_LOCK_CW;
1562 else {
1563 log_print("munge_altmode invalid exflags %x", lkb->lkb_exflags);
1564 dlm_print_lkb(lkb);
1565 }
1566 }
1567
first_in_list(struct dlm_lkb * lkb,struct list_head * head)1568 static inline int first_in_list(struct dlm_lkb *lkb, struct list_head *head)
1569 {
1570 struct dlm_lkb *first = list_entry(head->next, struct dlm_lkb,
1571 lkb_statequeue);
1572 if (lkb->lkb_id == first->lkb_id)
1573 return 1;
1574
1575 return 0;
1576 }
1577
1578 /* Check if the given lkb conflicts with another lkb on the queue. */
1579
queue_conflict(struct list_head * head,struct dlm_lkb * lkb)1580 static int queue_conflict(struct list_head *head, struct dlm_lkb *lkb)
1581 {
1582 struct dlm_lkb *this;
1583
1584 list_for_each_entry(this, head, lkb_statequeue) {
1585 if (this == lkb)
1586 continue;
1587 if (!modes_compat(this, lkb))
1588 return 1;
1589 }
1590 return 0;
1591 }
1592
1593 /*
1594 * "A conversion deadlock arises with a pair of lock requests in the converting
1595 * queue for one resource. The granted mode of each lock blocks the requested
1596 * mode of the other lock."
1597 *
1598 * Part 2: if the granted mode of lkb is preventing an earlier lkb in the
1599 * convert queue from being granted, then deadlk/demote lkb.
1600 *
1601 * Example:
1602 * Granted Queue: empty
1603 * Convert Queue: NL->EX (first lock)
1604 * PR->EX (second lock)
1605 *
1606 * The first lock can't be granted because of the granted mode of the second
1607 * lock and the second lock can't be granted because it's not first in the
1608 * list. We either cancel lkb's conversion (PR->EX) and return EDEADLK, or we
1609 * demote the granted mode of lkb (from PR to NL) if it has the CONVDEADLK
1610 * flag set and return DEMOTED in the lksb flags.
1611 *
1612 * Originally, this function detected conv-deadlk in a more limited scope:
1613 * - if !modes_compat(lkb1, lkb2) && !modes_compat(lkb2, lkb1), or
1614 * - if lkb1 was the first entry in the queue (not just earlier), and was
1615 * blocked by the granted mode of lkb2, and there was nothing on the
1616 * granted queue preventing lkb1 from being granted immediately, i.e.
1617 * lkb2 was the only thing preventing lkb1 from being granted.
1618 *
1619 * That second condition meant we'd only say there was conv-deadlk if
1620 * resolving it (by demotion) would lead to the first lock on the convert
1621 * queue being granted right away. It allowed conversion deadlocks to exist
1622 * between locks on the convert queue while they couldn't be granted anyway.
1623 *
1624 * Now, we detect and take action on conversion deadlocks immediately when
1625 * they're created, even if they may not be immediately consequential. If
1626 * lkb1 exists anywhere in the convert queue and lkb2 comes in with a granted
1627 * mode that would prevent lkb1's conversion from being granted, we do a
1628 * deadlk/demote on lkb2 right away and don't let it onto the convert queue.
1629 * I think this means that the lkb_is_ahead condition below should always
1630 * be zero, i.e. there will never be conv-deadlk between two locks that are
1631 * both already on the convert queue.
1632 */
1633
conversion_deadlock_detect(struct dlm_rsb * r,struct dlm_lkb * lkb2)1634 static int conversion_deadlock_detect(struct dlm_rsb *r, struct dlm_lkb *lkb2)
1635 {
1636 struct dlm_lkb *lkb1;
1637 int lkb_is_ahead = 0;
1638
1639 list_for_each_entry(lkb1, &r->res_convertqueue, lkb_statequeue) {
1640 if (lkb1 == lkb2) {
1641 lkb_is_ahead = 1;
1642 continue;
1643 }
1644
1645 if (!lkb_is_ahead) {
1646 if (!modes_compat(lkb2, lkb1))
1647 return 1;
1648 } else {
1649 if (!modes_compat(lkb2, lkb1) &&
1650 !modes_compat(lkb1, lkb2))
1651 return 1;
1652 }
1653 }
1654 return 0;
1655 }
1656
1657 /*
1658 * Return 1 if the lock can be granted, 0 otherwise.
1659 * Also detect and resolve conversion deadlocks.
1660 *
1661 * lkb is the lock to be granted
1662 *
1663 * now is 1 if the function is being called in the context of the
1664 * immediate request, it is 0 if called later, after the lock has been
1665 * queued.
1666 *
1667 * References are from chapter 6 of "VAXcluster Principles" by Roy Davis
1668 */
1669
_can_be_granted(struct dlm_rsb * r,struct dlm_lkb * lkb,int now)1670 static int _can_be_granted(struct dlm_rsb *r, struct dlm_lkb *lkb, int now)
1671 {
1672 int8_t conv = (lkb->lkb_grmode != DLM_LOCK_IV);
1673
1674 /*
1675 * 6-10: Version 5.4 introduced an option to address the phenomenon of
1676 * a new request for a NL mode lock being blocked.
1677 *
1678 * 6-11: If the optional EXPEDITE flag is used with the new NL mode
1679 * request, then it would be granted. In essence, the use of this flag
1680 * tells the Lock Manager to expedite theis request by not considering
1681 * what may be in the CONVERTING or WAITING queues... As of this
1682 * writing, the EXPEDITE flag can be used only with new requests for NL
1683 * mode locks. This flag is not valid for conversion requests.
1684 *
1685 * A shortcut. Earlier checks return an error if EXPEDITE is used in a
1686 * conversion or used with a non-NL requested mode. We also know an
1687 * EXPEDITE request is always granted immediately, so now must always
1688 * be 1. The full condition to grant an expedite request: (now &&
1689 * !conv && lkb->rqmode == DLM_LOCK_NL && (flags & EXPEDITE)) can
1690 * therefore be shortened to just checking the flag.
1691 */
1692
1693 if (lkb->lkb_exflags & DLM_LKF_EXPEDITE)
1694 return 1;
1695
1696 /*
1697 * A shortcut. Without this, !queue_conflict(grantqueue, lkb) would be
1698 * added to the remaining conditions.
1699 */
1700
1701 if (queue_conflict(&r->res_grantqueue, lkb))
1702 goto out;
1703
1704 /*
1705 * 6-3: By default, a conversion request is immediately granted if the
1706 * requested mode is compatible with the modes of all other granted
1707 * locks
1708 */
1709
1710 if (queue_conflict(&r->res_convertqueue, lkb))
1711 goto out;
1712
1713 /*
1714 * 6-5: But the default algorithm for deciding whether to grant or
1715 * queue conversion requests does not by itself guarantee that such
1716 * requests are serviced on a "first come first serve" basis. This, in
1717 * turn, can lead to a phenomenon known as "indefinate postponement".
1718 *
1719 * 6-7: This issue is dealt with by using the optional QUECVT flag with
1720 * the system service employed to request a lock conversion. This flag
1721 * forces certain conversion requests to be queued, even if they are
1722 * compatible with the granted modes of other locks on the same
1723 * resource. Thus, the use of this flag results in conversion requests
1724 * being ordered on a "first come first servce" basis.
1725 *
1726 * DCT: This condition is all about new conversions being able to occur
1727 * "in place" while the lock remains on the granted queue (assuming
1728 * nothing else conflicts.) IOW if QUECVT isn't set, a conversion
1729 * doesn't _have_ to go onto the convert queue where it's processed in
1730 * order. The "now" variable is necessary to distinguish converts
1731 * being received and processed for the first time now, because once a
1732 * convert is moved to the conversion queue the condition below applies
1733 * requiring fifo granting.
1734 */
1735
1736 if (now && conv && !(lkb->lkb_exflags & DLM_LKF_QUECVT))
1737 return 1;
1738
1739 /*
1740 * Even if the convert is compat with all granted locks,
1741 * QUECVT forces it behind other locks on the convert queue.
1742 */
1743
1744 if (now && conv && (lkb->lkb_exflags & DLM_LKF_QUECVT)) {
1745 if (list_empty(&r->res_convertqueue))
1746 return 1;
1747 else
1748 goto out;
1749 }
1750
1751 /*
1752 * The NOORDER flag is set to avoid the standard vms rules on grant
1753 * order.
1754 */
1755
1756 if (lkb->lkb_exflags & DLM_LKF_NOORDER)
1757 return 1;
1758
1759 /*
1760 * 6-3: Once in that queue [CONVERTING], a conversion request cannot be
1761 * granted until all other conversion requests ahead of it are granted
1762 * and/or canceled.
1763 */
1764
1765 if (!now && conv && first_in_list(lkb, &r->res_convertqueue))
1766 return 1;
1767
1768 /*
1769 * 6-4: By default, a new request is immediately granted only if all
1770 * three of the following conditions are satisfied when the request is
1771 * issued:
1772 * - The queue of ungranted conversion requests for the resource is
1773 * empty.
1774 * - The queue of ungranted new requests for the resource is empty.
1775 * - The mode of the new request is compatible with the most
1776 * restrictive mode of all granted locks on the resource.
1777 */
1778
1779 if (now && !conv && list_empty(&r->res_convertqueue) &&
1780 list_empty(&r->res_waitqueue))
1781 return 1;
1782
1783 /*
1784 * 6-4: Once a lock request is in the queue of ungranted new requests,
1785 * it cannot be granted until the queue of ungranted conversion
1786 * requests is empty, all ungranted new requests ahead of it are
1787 * granted and/or canceled, and it is compatible with the granted mode
1788 * of the most restrictive lock granted on the resource.
1789 */
1790
1791 if (!now && !conv && list_empty(&r->res_convertqueue) &&
1792 first_in_list(lkb, &r->res_waitqueue))
1793 return 1;
1794 out:
1795 return 0;
1796 }
1797
can_be_granted(struct dlm_rsb * r,struct dlm_lkb * lkb,int now,int * err)1798 static int can_be_granted(struct dlm_rsb *r, struct dlm_lkb *lkb, int now,
1799 int *err)
1800 {
1801 int rv;
1802 int8_t alt = 0, rqmode = lkb->lkb_rqmode;
1803 int8_t is_convert = (lkb->lkb_grmode != DLM_LOCK_IV);
1804
1805 if (err)
1806 *err = 0;
1807
1808 rv = _can_be_granted(r, lkb, now);
1809 if (rv)
1810 goto out;
1811
1812 /*
1813 * The CONVDEADLK flag is non-standard and tells the dlm to resolve
1814 * conversion deadlocks by demoting grmode to NL, otherwise the dlm
1815 * cancels one of the locks.
1816 */
1817
1818 if (is_convert && can_be_queued(lkb) &&
1819 conversion_deadlock_detect(r, lkb)) {
1820 if (lkb->lkb_exflags & DLM_LKF_CONVDEADLK) {
1821 lkb->lkb_grmode = DLM_LOCK_NL;
1822 lkb->lkb_sbflags |= DLM_SBF_DEMOTED;
1823 } else if (!(lkb->lkb_exflags & DLM_LKF_NODLCKWT)) {
1824 if (err)
1825 *err = -EDEADLK;
1826 else {
1827 log_print("can_be_granted deadlock %x now %d",
1828 lkb->lkb_id, now);
1829 dlm_dump_rsb(r);
1830 }
1831 }
1832 goto out;
1833 }
1834
1835 /*
1836 * The ALTPR and ALTCW flags are non-standard and tell the dlm to try
1837 * to grant a request in a mode other than the normal rqmode. It's a
1838 * simple way to provide a big optimization to applications that can
1839 * use them.
1840 */
1841
1842 if (rqmode != DLM_LOCK_PR && (lkb->lkb_exflags & DLM_LKF_ALTPR))
1843 alt = DLM_LOCK_PR;
1844 else if (rqmode != DLM_LOCK_CW && (lkb->lkb_exflags & DLM_LKF_ALTCW))
1845 alt = DLM_LOCK_CW;
1846
1847 if (alt) {
1848 lkb->lkb_rqmode = alt;
1849 rv = _can_be_granted(r, lkb, now);
1850 if (rv)
1851 lkb->lkb_sbflags |= DLM_SBF_ALTMODE;
1852 else
1853 lkb->lkb_rqmode = rqmode;
1854 }
1855 out:
1856 return rv;
1857 }
1858
1859 /* FIXME: I don't think that can_be_granted() can/will demote or find deadlock
1860 for locks pending on the convert list. Once verified (watch for these
1861 log_prints), we should be able to just call _can_be_granted() and not
1862 bother with the demote/deadlk cases here (and there's no easy way to deal
1863 with a deadlk here, we'd have to generate something like grant_lock with
1864 the deadlk error.) */
1865
1866 /* Returns the highest requested mode of all blocked conversions; sets
1867 cw if there's a blocked conversion to DLM_LOCK_CW. */
1868
grant_pending_convert(struct dlm_rsb * r,int high,int * cw)1869 static int grant_pending_convert(struct dlm_rsb *r, int high, int *cw)
1870 {
1871 struct dlm_lkb *lkb, *s;
1872 int hi, demoted, quit, grant_restart, demote_restart;
1873 int deadlk;
1874
1875 quit = 0;
1876 restart:
1877 grant_restart = 0;
1878 demote_restart = 0;
1879 hi = DLM_LOCK_IV;
1880
1881 list_for_each_entry_safe(lkb, s, &r->res_convertqueue, lkb_statequeue) {
1882 demoted = is_demoted(lkb);
1883 deadlk = 0;
1884
1885 if (can_be_granted(r, lkb, 0, &deadlk)) {
1886 grant_lock_pending(r, lkb);
1887 grant_restart = 1;
1888 continue;
1889 }
1890
1891 if (!demoted && is_demoted(lkb)) {
1892 log_print("WARN: pending demoted %x node %d %s",
1893 lkb->lkb_id, lkb->lkb_nodeid, r->res_name);
1894 demote_restart = 1;
1895 continue;
1896 }
1897
1898 if (deadlk) {
1899 log_print("WARN: pending deadlock %x node %d %s",
1900 lkb->lkb_id, lkb->lkb_nodeid, r->res_name);
1901 dlm_dump_rsb(r);
1902 continue;
1903 }
1904
1905 hi = max_t(int, lkb->lkb_rqmode, hi);
1906
1907 if (cw && lkb->lkb_rqmode == DLM_LOCK_CW)
1908 *cw = 1;
1909 }
1910
1911 if (grant_restart)
1912 goto restart;
1913 if (demote_restart && !quit) {
1914 quit = 1;
1915 goto restart;
1916 }
1917
1918 return max_t(int, high, hi);
1919 }
1920
grant_pending_wait(struct dlm_rsb * r,int high,int * cw)1921 static int grant_pending_wait(struct dlm_rsb *r, int high, int *cw)
1922 {
1923 struct dlm_lkb *lkb, *s;
1924
1925 list_for_each_entry_safe(lkb, s, &r->res_waitqueue, lkb_statequeue) {
1926 if (can_be_granted(r, lkb, 0, NULL))
1927 grant_lock_pending(r, lkb);
1928 else {
1929 high = max_t(int, lkb->lkb_rqmode, high);
1930 if (lkb->lkb_rqmode == DLM_LOCK_CW)
1931 *cw = 1;
1932 }
1933 }
1934
1935 return high;
1936 }
1937
1938 /* cw of 1 means there's a lock with a rqmode of DLM_LOCK_CW that's blocked
1939 on either the convert or waiting queue.
1940 high is the largest rqmode of all locks blocked on the convert or
1941 waiting queue. */
1942
lock_requires_bast(struct dlm_lkb * gr,int high,int cw)1943 static int lock_requires_bast(struct dlm_lkb *gr, int high, int cw)
1944 {
1945 if (gr->lkb_grmode == DLM_LOCK_PR && cw) {
1946 if (gr->lkb_highbast < DLM_LOCK_EX)
1947 return 1;
1948 return 0;
1949 }
1950
1951 if (gr->lkb_highbast < high &&
1952 !__dlm_compat_matrix[gr->lkb_grmode+1][high+1])
1953 return 1;
1954 return 0;
1955 }
1956
grant_pending_locks(struct dlm_rsb * r)1957 static void grant_pending_locks(struct dlm_rsb *r)
1958 {
1959 struct dlm_lkb *lkb, *s;
1960 int high = DLM_LOCK_IV;
1961 int cw = 0;
1962
1963 DLM_ASSERT(is_master(r), dlm_dump_rsb(r););
1964
1965 high = grant_pending_convert(r, high, &cw);
1966 high = grant_pending_wait(r, high, &cw);
1967
1968 if (high == DLM_LOCK_IV)
1969 return;
1970
1971 /*
1972 * If there are locks left on the wait/convert queue then send blocking
1973 * ASTs to granted locks based on the largest requested mode (high)
1974 * found above.
1975 */
1976
1977 list_for_each_entry_safe(lkb, s, &r->res_grantqueue, lkb_statequeue) {
1978 if (lkb->lkb_bastfn && lock_requires_bast(lkb, high, cw)) {
1979 if (cw && high == DLM_LOCK_PR &&
1980 lkb->lkb_grmode == DLM_LOCK_PR)
1981 queue_bast(r, lkb, DLM_LOCK_CW);
1982 else
1983 queue_bast(r, lkb, high);
1984 lkb->lkb_highbast = high;
1985 }
1986 }
1987 }
1988
modes_require_bast(struct dlm_lkb * gr,struct dlm_lkb * rq)1989 static int modes_require_bast(struct dlm_lkb *gr, struct dlm_lkb *rq)
1990 {
1991 if ((gr->lkb_grmode == DLM_LOCK_PR && rq->lkb_rqmode == DLM_LOCK_CW) ||
1992 (gr->lkb_grmode == DLM_LOCK_CW && rq->lkb_rqmode == DLM_LOCK_PR)) {
1993 if (gr->lkb_highbast < DLM_LOCK_EX)
1994 return 1;
1995 return 0;
1996 }
1997
1998 if (gr->lkb_highbast < rq->lkb_rqmode && !modes_compat(gr, rq))
1999 return 1;
2000 return 0;
2001 }
2002
send_bast_queue(struct dlm_rsb * r,struct list_head * head,struct dlm_lkb * lkb)2003 static void send_bast_queue(struct dlm_rsb *r, struct list_head *head,
2004 struct dlm_lkb *lkb)
2005 {
2006 struct dlm_lkb *gr;
2007
2008 list_for_each_entry(gr, head, lkb_statequeue) {
2009 /* skip self when sending basts to convertqueue */
2010 if (gr == lkb)
2011 continue;
2012 if (gr->lkb_bastfn && modes_require_bast(gr, lkb)) {
2013 queue_bast(r, gr, lkb->lkb_rqmode);
2014 gr->lkb_highbast = lkb->lkb_rqmode;
2015 }
2016 }
2017 }
2018
send_blocking_asts(struct dlm_rsb * r,struct dlm_lkb * lkb)2019 static void send_blocking_asts(struct dlm_rsb *r, struct dlm_lkb *lkb)
2020 {
2021 send_bast_queue(r, &r->res_grantqueue, lkb);
2022 }
2023
send_blocking_asts_all(struct dlm_rsb * r,struct dlm_lkb * lkb)2024 static void send_blocking_asts_all(struct dlm_rsb *r, struct dlm_lkb *lkb)
2025 {
2026 send_bast_queue(r, &r->res_grantqueue, lkb);
2027 send_bast_queue(r, &r->res_convertqueue, lkb);
2028 }
2029
2030 /* set_master(r, lkb) -- set the master nodeid of a resource
2031
2032 The purpose of this function is to set the nodeid field in the given
2033 lkb using the nodeid field in the given rsb. If the rsb's nodeid is
2034 known, it can just be copied to the lkb and the function will return
2035 0. If the rsb's nodeid is _not_ known, it needs to be looked up
2036 before it can be copied to the lkb.
2037
2038 When the rsb nodeid is being looked up remotely, the initial lkb
2039 causing the lookup is kept on the ls_waiters list waiting for the
2040 lookup reply. Other lkb's waiting for the same rsb lookup are kept
2041 on the rsb's res_lookup list until the master is verified.
2042
2043 Return values:
2044 0: nodeid is set in rsb/lkb and the caller should go ahead and use it
2045 1: the rsb master is not available and the lkb has been placed on
2046 a wait queue
2047 */
2048
set_master(struct dlm_rsb * r,struct dlm_lkb * lkb)2049 static int set_master(struct dlm_rsb *r, struct dlm_lkb *lkb)
2050 {
2051 struct dlm_ls *ls = r->res_ls;
2052 int i, error, dir_nodeid, ret_nodeid, our_nodeid = dlm_our_nodeid();
2053
2054 if (rsb_flag(r, RSB_MASTER_UNCERTAIN)) {
2055 rsb_clear_flag(r, RSB_MASTER_UNCERTAIN);
2056 r->res_first_lkid = lkb->lkb_id;
2057 lkb->lkb_nodeid = r->res_nodeid;
2058 return 0;
2059 }
2060
2061 if (r->res_first_lkid && r->res_first_lkid != lkb->lkb_id) {
2062 list_add_tail(&lkb->lkb_rsb_lookup, &r->res_lookup);
2063 return 1;
2064 }
2065
2066 if (r->res_nodeid == 0) {
2067 lkb->lkb_nodeid = 0;
2068 return 0;
2069 }
2070
2071 if (r->res_nodeid > 0) {
2072 lkb->lkb_nodeid = r->res_nodeid;
2073 return 0;
2074 }
2075
2076 DLM_ASSERT(r->res_nodeid == -1, dlm_dump_rsb(r););
2077
2078 dir_nodeid = dlm_dir_nodeid(r);
2079
2080 if (dir_nodeid != our_nodeid) {
2081 r->res_first_lkid = lkb->lkb_id;
2082 send_lookup(r, lkb);
2083 return 1;
2084 }
2085
2086 for (i = 0; i < 2; i++) {
2087 /* It's possible for dlm_scand to remove an old rsb for
2088 this same resource from the toss list, us to create
2089 a new one, look up the master locally, and find it
2090 already exists just before dlm_scand does the
2091 dir_remove() on the previous rsb. */
2092
2093 error = dlm_dir_lookup(ls, our_nodeid, r->res_name,
2094 r->res_length, &ret_nodeid);
2095 if (!error)
2096 break;
2097 log_debug(ls, "dir_lookup error %d %s", error, r->res_name);
2098 schedule();
2099 }
2100 if (error && error != -EEXIST)
2101 return error;
2102
2103 if (ret_nodeid == our_nodeid) {
2104 r->res_first_lkid = 0;
2105 r->res_nodeid = 0;
2106 lkb->lkb_nodeid = 0;
2107 } else {
2108 r->res_first_lkid = lkb->lkb_id;
2109 r->res_nodeid = ret_nodeid;
2110 lkb->lkb_nodeid = ret_nodeid;
2111 }
2112 return 0;
2113 }
2114
process_lookup_list(struct dlm_rsb * r)2115 static void process_lookup_list(struct dlm_rsb *r)
2116 {
2117 struct dlm_lkb *lkb, *safe;
2118
2119 list_for_each_entry_safe(lkb, safe, &r->res_lookup, lkb_rsb_lookup) {
2120 list_del_init(&lkb->lkb_rsb_lookup);
2121 _request_lock(r, lkb);
2122 schedule();
2123 }
2124 }
2125
2126 /* confirm_master -- confirm (or deny) an rsb's master nodeid */
2127
confirm_master(struct dlm_rsb * r,int error)2128 static void confirm_master(struct dlm_rsb *r, int error)
2129 {
2130 struct dlm_lkb *lkb;
2131
2132 if (!r->res_first_lkid)
2133 return;
2134
2135 switch (error) {
2136 case 0:
2137 case -EINPROGRESS:
2138 r->res_first_lkid = 0;
2139 process_lookup_list(r);
2140 break;
2141
2142 case -EAGAIN:
2143 case -EBADR:
2144 case -ENOTBLK:
2145 /* the remote request failed and won't be retried (it was
2146 a NOQUEUE, or has been canceled/unlocked); make a waiting
2147 lkb the first_lkid */
2148
2149 r->res_first_lkid = 0;
2150
2151 if (!list_empty(&r->res_lookup)) {
2152 lkb = list_entry(r->res_lookup.next, struct dlm_lkb,
2153 lkb_rsb_lookup);
2154 list_del_init(&lkb->lkb_rsb_lookup);
2155 r->res_first_lkid = lkb->lkb_id;
2156 _request_lock(r, lkb);
2157 }
2158 break;
2159
2160 default:
2161 log_error(r->res_ls, "confirm_master unknown error %d", error);
2162 }
2163 }
2164
set_lock_args(int mode,struct dlm_lksb * lksb,uint32_t flags,int namelen,unsigned long timeout_cs,void (* ast)(void * astparam),void * astparam,void (* bast)(void * astparam,int mode),struct dlm_args * args)2165 static int set_lock_args(int mode, struct dlm_lksb *lksb, uint32_t flags,
2166 int namelen, unsigned long timeout_cs,
2167 void (*ast) (void *astparam),
2168 void *astparam,
2169 void (*bast) (void *astparam, int mode),
2170 struct dlm_args *args)
2171 {
2172 int rv = -EINVAL;
2173
2174 /* check for invalid arg usage */
2175
2176 if (mode < 0 || mode > DLM_LOCK_EX)
2177 goto out;
2178
2179 if (!(flags & DLM_LKF_CONVERT) && (namelen > DLM_RESNAME_MAXLEN))
2180 goto out;
2181
2182 if (flags & DLM_LKF_CANCEL)
2183 goto out;
2184
2185 if (flags & DLM_LKF_QUECVT && !(flags & DLM_LKF_CONVERT))
2186 goto out;
2187
2188 if (flags & DLM_LKF_CONVDEADLK && !(flags & DLM_LKF_CONVERT))
2189 goto out;
2190
2191 if (flags & DLM_LKF_CONVDEADLK && flags & DLM_LKF_NOQUEUE)
2192 goto out;
2193
2194 if (flags & DLM_LKF_EXPEDITE && flags & DLM_LKF_CONVERT)
2195 goto out;
2196
2197 if (flags & DLM_LKF_EXPEDITE && flags & DLM_LKF_QUECVT)
2198 goto out;
2199
2200 if (flags & DLM_LKF_EXPEDITE && flags & DLM_LKF_NOQUEUE)
2201 goto out;
2202
2203 if (flags & DLM_LKF_EXPEDITE && mode != DLM_LOCK_NL)
2204 goto out;
2205
2206 if (!ast || !lksb)
2207 goto out;
2208
2209 if (flags & DLM_LKF_VALBLK && !lksb->sb_lvbptr)
2210 goto out;
2211
2212 if (flags & DLM_LKF_CONVERT && !lksb->sb_lkid)
2213 goto out;
2214
2215 /* these args will be copied to the lkb in validate_lock_args,
2216 it cannot be done now because when converting locks, fields in
2217 an active lkb cannot be modified before locking the rsb */
2218
2219 args->flags = flags;
2220 args->astfn = ast;
2221 args->astparam = astparam;
2222 args->bastfn = bast;
2223 args->timeout = timeout_cs;
2224 args->mode = mode;
2225 args->lksb = lksb;
2226 rv = 0;
2227 out:
2228 return rv;
2229 }
2230
set_unlock_args(uint32_t flags,void * astarg,struct dlm_args * args)2231 static int set_unlock_args(uint32_t flags, void *astarg, struct dlm_args *args)
2232 {
2233 if (flags & ~(DLM_LKF_CANCEL | DLM_LKF_VALBLK | DLM_LKF_IVVALBLK |
2234 DLM_LKF_FORCEUNLOCK))
2235 return -EINVAL;
2236
2237 if (flags & DLM_LKF_CANCEL && flags & DLM_LKF_FORCEUNLOCK)
2238 return -EINVAL;
2239
2240 args->flags = flags;
2241 args->astparam = astarg;
2242 return 0;
2243 }
2244
validate_lock_args(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_args * args)2245 static int validate_lock_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
2246 struct dlm_args *args)
2247 {
2248 int rv = -EINVAL;
2249
2250 if (args->flags & DLM_LKF_CONVERT) {
2251 if (lkb->lkb_flags & DLM_IFL_MSTCPY)
2252 goto out;
2253
2254 if (args->flags & DLM_LKF_QUECVT &&
2255 !__quecvt_compat_matrix[lkb->lkb_grmode+1][args->mode+1])
2256 goto out;
2257
2258 rv = -EBUSY;
2259 if (lkb->lkb_status != DLM_LKSTS_GRANTED)
2260 goto out;
2261
2262 if (lkb->lkb_wait_type)
2263 goto out;
2264
2265 if (is_overlap(lkb))
2266 goto out;
2267 }
2268
2269 lkb->lkb_exflags = args->flags;
2270 lkb->lkb_sbflags = 0;
2271 lkb->lkb_astfn = args->astfn;
2272 lkb->lkb_astparam = args->astparam;
2273 lkb->lkb_bastfn = args->bastfn;
2274 lkb->lkb_rqmode = args->mode;
2275 lkb->lkb_lksb = args->lksb;
2276 lkb->lkb_lvbptr = args->lksb->sb_lvbptr;
2277 lkb->lkb_ownpid = (int) current->pid;
2278 lkb->lkb_timeout_cs = args->timeout;
2279 rv = 0;
2280 out:
2281 if (rv)
2282 log_debug(ls, "validate_lock_args %d %x %x %x %d %d %s",
2283 rv, lkb->lkb_id, lkb->lkb_flags, args->flags,
2284 lkb->lkb_status, lkb->lkb_wait_type,
2285 lkb->lkb_resource->res_name);
2286 return rv;
2287 }
2288
2289 /* when dlm_unlock() sees -EBUSY with CANCEL/FORCEUNLOCK it returns 0
2290 for success */
2291
2292 /* note: it's valid for lkb_nodeid/res_nodeid to be -1 when we get here
2293 because there may be a lookup in progress and it's valid to do
2294 cancel/unlockf on it */
2295
validate_unlock_args(struct dlm_lkb * lkb,struct dlm_args * args)2296 static int validate_unlock_args(struct dlm_lkb *lkb, struct dlm_args *args)
2297 {
2298 struct dlm_ls *ls = lkb->lkb_resource->res_ls;
2299 int rv = -EINVAL;
2300
2301 if (lkb->lkb_flags & DLM_IFL_MSTCPY) {
2302 log_error(ls, "unlock on MSTCPY %x", lkb->lkb_id);
2303 dlm_print_lkb(lkb);
2304 goto out;
2305 }
2306
2307 /* an lkb may still exist even though the lock is EOL'ed due to a
2308 cancel, unlock or failed noqueue request; an app can't use these
2309 locks; return same error as if the lkid had not been found at all */
2310
2311 if (lkb->lkb_flags & DLM_IFL_ENDOFLIFE) {
2312 log_debug(ls, "unlock on ENDOFLIFE %x", lkb->lkb_id);
2313 rv = -ENOENT;
2314 goto out;
2315 }
2316
2317 /* an lkb may be waiting for an rsb lookup to complete where the
2318 lookup was initiated by another lock */
2319
2320 if (!list_empty(&lkb->lkb_rsb_lookup)) {
2321 if (args->flags & (DLM_LKF_CANCEL | DLM_LKF_FORCEUNLOCK)) {
2322 log_debug(ls, "unlock on rsb_lookup %x", lkb->lkb_id);
2323 list_del_init(&lkb->lkb_rsb_lookup);
2324 queue_cast(lkb->lkb_resource, lkb,
2325 args->flags & DLM_LKF_CANCEL ?
2326 -DLM_ECANCEL : -DLM_EUNLOCK);
2327 unhold_lkb(lkb); /* undoes create_lkb() */
2328 }
2329 /* caller changes -EBUSY to 0 for CANCEL and FORCEUNLOCK */
2330 rv = -EBUSY;
2331 goto out;
2332 }
2333
2334 /* cancel not allowed with another cancel/unlock in progress */
2335
2336 if (args->flags & DLM_LKF_CANCEL) {
2337 if (lkb->lkb_exflags & DLM_LKF_CANCEL)
2338 goto out;
2339
2340 if (is_overlap(lkb))
2341 goto out;
2342
2343 /* don't let scand try to do a cancel */
2344 del_timeout(lkb);
2345
2346 if (lkb->lkb_flags & DLM_IFL_RESEND) {
2347 lkb->lkb_flags |= DLM_IFL_OVERLAP_CANCEL;
2348 rv = -EBUSY;
2349 goto out;
2350 }
2351
2352 /* there's nothing to cancel */
2353 if (lkb->lkb_status == DLM_LKSTS_GRANTED &&
2354 !lkb->lkb_wait_type) {
2355 rv = -EBUSY;
2356 goto out;
2357 }
2358
2359 switch (lkb->lkb_wait_type) {
2360 case DLM_MSG_LOOKUP:
2361 case DLM_MSG_REQUEST:
2362 lkb->lkb_flags |= DLM_IFL_OVERLAP_CANCEL;
2363 rv = -EBUSY;
2364 goto out;
2365 case DLM_MSG_UNLOCK:
2366 case DLM_MSG_CANCEL:
2367 goto out;
2368 }
2369 /* add_to_waiters() will set OVERLAP_CANCEL */
2370 goto out_ok;
2371 }
2372
2373 /* do we need to allow a force-unlock if there's a normal unlock
2374 already in progress? in what conditions could the normal unlock
2375 fail such that we'd want to send a force-unlock to be sure? */
2376
2377 if (args->flags & DLM_LKF_FORCEUNLOCK) {
2378 if (lkb->lkb_exflags & DLM_LKF_FORCEUNLOCK)
2379 goto out;
2380
2381 if (is_overlap_unlock(lkb))
2382 goto out;
2383
2384 /* don't let scand try to do a cancel */
2385 del_timeout(lkb);
2386
2387 if (lkb->lkb_flags & DLM_IFL_RESEND) {
2388 lkb->lkb_flags |= DLM_IFL_OVERLAP_UNLOCK;
2389 rv = -EBUSY;
2390 goto out;
2391 }
2392
2393 switch (lkb->lkb_wait_type) {
2394 case DLM_MSG_LOOKUP:
2395 case DLM_MSG_REQUEST:
2396 lkb->lkb_flags |= DLM_IFL_OVERLAP_UNLOCK;
2397 rv = -EBUSY;
2398 goto out;
2399 case DLM_MSG_UNLOCK:
2400 goto out;
2401 }
2402 /* add_to_waiters() will set OVERLAP_UNLOCK */
2403 goto out_ok;
2404 }
2405
2406 /* normal unlock not allowed if there's any op in progress */
2407 rv = -EBUSY;
2408 if (lkb->lkb_wait_type || lkb->lkb_wait_count)
2409 goto out;
2410
2411 out_ok:
2412 /* an overlapping op shouldn't blow away exflags from other op */
2413 lkb->lkb_exflags |= args->flags;
2414 lkb->lkb_sbflags = 0;
2415 lkb->lkb_astparam = args->astparam;
2416 rv = 0;
2417 out:
2418 if (rv)
2419 log_debug(ls, "validate_unlock_args %d %x %x %x %x %d %s", rv,
2420 lkb->lkb_id, lkb->lkb_flags, lkb->lkb_exflags,
2421 args->flags, lkb->lkb_wait_type,
2422 lkb->lkb_resource->res_name);
2423 return rv;
2424 }
2425
2426 /*
2427 * Four stage 4 varieties:
2428 * do_request(), do_convert(), do_unlock(), do_cancel()
2429 * These are called on the master node for the given lock and
2430 * from the central locking logic.
2431 */
2432
do_request(struct dlm_rsb * r,struct dlm_lkb * lkb)2433 static int do_request(struct dlm_rsb *r, struct dlm_lkb *lkb)
2434 {
2435 int error = 0;
2436
2437 if (can_be_granted(r, lkb, 1, NULL)) {
2438 grant_lock(r, lkb);
2439 queue_cast(r, lkb, 0);
2440 goto out;
2441 }
2442
2443 if (can_be_queued(lkb)) {
2444 error = -EINPROGRESS;
2445 add_lkb(r, lkb, DLM_LKSTS_WAITING);
2446 add_timeout(lkb);
2447 goto out;
2448 }
2449
2450 error = -EAGAIN;
2451 queue_cast(r, lkb, -EAGAIN);
2452 out:
2453 return error;
2454 }
2455
do_request_effects(struct dlm_rsb * r,struct dlm_lkb * lkb,int error)2456 static void do_request_effects(struct dlm_rsb *r, struct dlm_lkb *lkb,
2457 int error)
2458 {
2459 switch (error) {
2460 case -EAGAIN:
2461 if (force_blocking_asts(lkb))
2462 send_blocking_asts_all(r, lkb);
2463 break;
2464 case -EINPROGRESS:
2465 send_blocking_asts(r, lkb);
2466 break;
2467 }
2468 }
2469
do_convert(struct dlm_rsb * r,struct dlm_lkb * lkb)2470 static int do_convert(struct dlm_rsb *r, struct dlm_lkb *lkb)
2471 {
2472 int error = 0;
2473 int deadlk = 0;
2474
2475 /* changing an existing lock may allow others to be granted */
2476
2477 if (can_be_granted(r, lkb, 1, &deadlk)) {
2478 grant_lock(r, lkb);
2479 queue_cast(r, lkb, 0);
2480 goto out;
2481 }
2482
2483 /* can_be_granted() detected that this lock would block in a conversion
2484 deadlock, so we leave it on the granted queue and return EDEADLK in
2485 the ast for the convert. */
2486
2487 if (deadlk) {
2488 /* it's left on the granted queue */
2489 revert_lock(r, lkb);
2490 queue_cast(r, lkb, -EDEADLK);
2491 error = -EDEADLK;
2492 goto out;
2493 }
2494
2495 /* is_demoted() means the can_be_granted() above set the grmode
2496 to NL, and left us on the granted queue. This auto-demotion
2497 (due to CONVDEADLK) might mean other locks, and/or this lock, are
2498 now grantable. We have to try to grant other converting locks
2499 before we try again to grant this one. */
2500
2501 if (is_demoted(lkb)) {
2502 grant_pending_convert(r, DLM_LOCK_IV, NULL);
2503 if (_can_be_granted(r, lkb, 1)) {
2504 grant_lock(r, lkb);
2505 queue_cast(r, lkb, 0);
2506 goto out;
2507 }
2508 /* else fall through and move to convert queue */
2509 }
2510
2511 if (can_be_queued(lkb)) {
2512 error = -EINPROGRESS;
2513 del_lkb(r, lkb);
2514 add_lkb(r, lkb, DLM_LKSTS_CONVERT);
2515 add_timeout(lkb);
2516 goto out;
2517 }
2518
2519 error = -EAGAIN;
2520 queue_cast(r, lkb, -EAGAIN);
2521 out:
2522 return error;
2523 }
2524
do_convert_effects(struct dlm_rsb * r,struct dlm_lkb * lkb,int error)2525 static void do_convert_effects(struct dlm_rsb *r, struct dlm_lkb *lkb,
2526 int error)
2527 {
2528 switch (error) {
2529 case 0:
2530 grant_pending_locks(r);
2531 /* grant_pending_locks also sends basts */
2532 break;
2533 case -EAGAIN:
2534 if (force_blocking_asts(lkb))
2535 send_blocking_asts_all(r, lkb);
2536 break;
2537 case -EINPROGRESS:
2538 send_blocking_asts(r, lkb);
2539 break;
2540 }
2541 }
2542
do_unlock(struct dlm_rsb * r,struct dlm_lkb * lkb)2543 static int do_unlock(struct dlm_rsb *r, struct dlm_lkb *lkb)
2544 {
2545 remove_lock(r, lkb);
2546 queue_cast(r, lkb, -DLM_EUNLOCK);
2547 return -DLM_EUNLOCK;
2548 }
2549
do_unlock_effects(struct dlm_rsb * r,struct dlm_lkb * lkb,int error)2550 static void do_unlock_effects(struct dlm_rsb *r, struct dlm_lkb *lkb,
2551 int error)
2552 {
2553 grant_pending_locks(r);
2554 }
2555
2556 /* returns: 0 did nothing, -DLM_ECANCEL canceled lock */
2557
do_cancel(struct dlm_rsb * r,struct dlm_lkb * lkb)2558 static int do_cancel(struct dlm_rsb *r, struct dlm_lkb *lkb)
2559 {
2560 int error;
2561
2562 error = revert_lock(r, lkb);
2563 if (error) {
2564 queue_cast(r, lkb, -DLM_ECANCEL);
2565 return -DLM_ECANCEL;
2566 }
2567 return 0;
2568 }
2569
do_cancel_effects(struct dlm_rsb * r,struct dlm_lkb * lkb,int error)2570 static void do_cancel_effects(struct dlm_rsb *r, struct dlm_lkb *lkb,
2571 int error)
2572 {
2573 if (error)
2574 grant_pending_locks(r);
2575 }
2576
2577 /*
2578 * Four stage 3 varieties:
2579 * _request_lock(), _convert_lock(), _unlock_lock(), _cancel_lock()
2580 */
2581
2582 /* add a new lkb to a possibly new rsb, called by requesting process */
2583
_request_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)2584 static int _request_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
2585 {
2586 int error;
2587
2588 /* set_master: sets lkb nodeid from r */
2589
2590 error = set_master(r, lkb);
2591 if (error < 0)
2592 goto out;
2593 if (error) {
2594 error = 0;
2595 goto out;
2596 }
2597
2598 if (is_remote(r)) {
2599 /* receive_request() calls do_request() on remote node */
2600 error = send_request(r, lkb);
2601 } else {
2602 error = do_request(r, lkb);
2603 /* for remote locks the request_reply is sent
2604 between do_request and do_request_effects */
2605 do_request_effects(r, lkb, error);
2606 }
2607 out:
2608 return error;
2609 }
2610
2611 /* change some property of an existing lkb, e.g. mode */
2612
_convert_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)2613 static int _convert_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
2614 {
2615 int error;
2616
2617 if (is_remote(r)) {
2618 /* receive_convert() calls do_convert() on remote node */
2619 error = send_convert(r, lkb);
2620 } else {
2621 error = do_convert(r, lkb);
2622 /* for remote locks the convert_reply is sent
2623 between do_convert and do_convert_effects */
2624 do_convert_effects(r, lkb, error);
2625 }
2626
2627 return error;
2628 }
2629
2630 /* remove an existing lkb from the granted queue */
2631
_unlock_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)2632 static int _unlock_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
2633 {
2634 int error;
2635
2636 if (is_remote(r)) {
2637 /* receive_unlock() calls do_unlock() on remote node */
2638 error = send_unlock(r, lkb);
2639 } else {
2640 error = do_unlock(r, lkb);
2641 /* for remote locks the unlock_reply is sent
2642 between do_unlock and do_unlock_effects */
2643 do_unlock_effects(r, lkb, error);
2644 }
2645
2646 return error;
2647 }
2648
2649 /* remove an existing lkb from the convert or wait queue */
2650
_cancel_lock(struct dlm_rsb * r,struct dlm_lkb * lkb)2651 static int _cancel_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
2652 {
2653 int error;
2654
2655 if (is_remote(r)) {
2656 /* receive_cancel() calls do_cancel() on remote node */
2657 error = send_cancel(r, lkb);
2658 } else {
2659 error = do_cancel(r, lkb);
2660 /* for remote locks the cancel_reply is sent
2661 between do_cancel and do_cancel_effects */
2662 do_cancel_effects(r, lkb, error);
2663 }
2664
2665 return error;
2666 }
2667
2668 /*
2669 * Four stage 2 varieties:
2670 * request_lock(), convert_lock(), unlock_lock(), cancel_lock()
2671 */
2672
request_lock(struct dlm_ls * ls,struct dlm_lkb * lkb,char * name,int len,struct dlm_args * args)2673 static int request_lock(struct dlm_ls *ls, struct dlm_lkb *lkb, char *name,
2674 int len, struct dlm_args *args)
2675 {
2676 struct dlm_rsb *r;
2677 int error;
2678
2679 error = validate_lock_args(ls, lkb, args);
2680 if (error)
2681 goto out;
2682
2683 error = find_rsb(ls, name, len, R_CREATE, &r);
2684 if (error)
2685 goto out;
2686
2687 lock_rsb(r);
2688
2689 attach_lkb(r, lkb);
2690 lkb->lkb_lksb->sb_lkid = lkb->lkb_id;
2691
2692 error = _request_lock(r, lkb);
2693
2694 unlock_rsb(r);
2695 put_rsb(r);
2696
2697 out:
2698 return error;
2699 }
2700
convert_lock(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_args * args)2701 static int convert_lock(struct dlm_ls *ls, struct dlm_lkb *lkb,
2702 struct dlm_args *args)
2703 {
2704 struct dlm_rsb *r;
2705 int error;
2706
2707 r = lkb->lkb_resource;
2708
2709 hold_rsb(r);
2710 lock_rsb(r);
2711
2712 error = validate_lock_args(ls, lkb, args);
2713 if (error)
2714 goto out;
2715
2716 error = _convert_lock(r, lkb);
2717 out:
2718 unlock_rsb(r);
2719 put_rsb(r);
2720 return error;
2721 }
2722
unlock_lock(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_args * args)2723 static int unlock_lock(struct dlm_ls *ls, struct dlm_lkb *lkb,
2724 struct dlm_args *args)
2725 {
2726 struct dlm_rsb *r;
2727 int error;
2728
2729 r = lkb->lkb_resource;
2730
2731 hold_rsb(r);
2732 lock_rsb(r);
2733
2734 error = validate_unlock_args(lkb, args);
2735 if (error)
2736 goto out;
2737
2738 error = _unlock_lock(r, lkb);
2739 out:
2740 unlock_rsb(r);
2741 put_rsb(r);
2742 return error;
2743 }
2744
cancel_lock(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_args * args)2745 static int cancel_lock(struct dlm_ls *ls, struct dlm_lkb *lkb,
2746 struct dlm_args *args)
2747 {
2748 struct dlm_rsb *r;
2749 int error;
2750
2751 r = lkb->lkb_resource;
2752
2753 hold_rsb(r);
2754 lock_rsb(r);
2755
2756 error = validate_unlock_args(lkb, args);
2757 if (error)
2758 goto out;
2759
2760 error = _cancel_lock(r, lkb);
2761 out:
2762 unlock_rsb(r);
2763 put_rsb(r);
2764 return error;
2765 }
2766
2767 /*
2768 * Two stage 1 varieties: dlm_lock() and dlm_unlock()
2769 */
2770
dlm_lock(dlm_lockspace_t * lockspace,int mode,struct dlm_lksb * lksb,uint32_t flags,void * name,unsigned int namelen,uint32_t parent_lkid,void (* ast)(void * astarg),void * astarg,void (* bast)(void * astarg,int mode))2771 int dlm_lock(dlm_lockspace_t *lockspace,
2772 int mode,
2773 struct dlm_lksb *lksb,
2774 uint32_t flags,
2775 void *name,
2776 unsigned int namelen,
2777 uint32_t parent_lkid,
2778 void (*ast) (void *astarg),
2779 void *astarg,
2780 void (*bast) (void *astarg, int mode))
2781 {
2782 struct dlm_ls *ls;
2783 struct dlm_lkb *lkb;
2784 struct dlm_args args;
2785 int error, convert = flags & DLM_LKF_CONVERT;
2786
2787 ls = dlm_find_lockspace_local(lockspace);
2788 if (!ls)
2789 return -EINVAL;
2790
2791 dlm_lock_recovery(ls);
2792
2793 if (convert)
2794 error = find_lkb(ls, lksb->sb_lkid, &lkb);
2795 else
2796 error = create_lkb(ls, &lkb);
2797
2798 if (error)
2799 goto out;
2800
2801 error = set_lock_args(mode, lksb, flags, namelen, 0, ast,
2802 astarg, bast, &args);
2803 if (error)
2804 goto out_put;
2805
2806 if (convert)
2807 error = convert_lock(ls, lkb, &args);
2808 else
2809 error = request_lock(ls, lkb, name, namelen, &args);
2810
2811 if (error == -EINPROGRESS)
2812 error = 0;
2813 out_put:
2814 if (convert || error)
2815 __put_lkb(ls, lkb);
2816 if (error == -EAGAIN || error == -EDEADLK)
2817 error = 0;
2818 out:
2819 dlm_unlock_recovery(ls);
2820 dlm_put_lockspace(ls);
2821 return error;
2822 }
2823
dlm_unlock(dlm_lockspace_t * lockspace,uint32_t lkid,uint32_t flags,struct dlm_lksb * lksb,void * astarg)2824 int dlm_unlock(dlm_lockspace_t *lockspace,
2825 uint32_t lkid,
2826 uint32_t flags,
2827 struct dlm_lksb *lksb,
2828 void *astarg)
2829 {
2830 struct dlm_ls *ls;
2831 struct dlm_lkb *lkb;
2832 struct dlm_args args;
2833 int error;
2834
2835 ls = dlm_find_lockspace_local(lockspace);
2836 if (!ls)
2837 return -EINVAL;
2838
2839 dlm_lock_recovery(ls);
2840
2841 error = find_lkb(ls, lkid, &lkb);
2842 if (error)
2843 goto out;
2844
2845 error = set_unlock_args(flags, astarg, &args);
2846 if (error)
2847 goto out_put;
2848
2849 if (flags & DLM_LKF_CANCEL)
2850 error = cancel_lock(ls, lkb, &args);
2851 else
2852 error = unlock_lock(ls, lkb, &args);
2853
2854 if (error == -DLM_EUNLOCK || error == -DLM_ECANCEL)
2855 error = 0;
2856 if (error == -EBUSY && (flags & (DLM_LKF_CANCEL | DLM_LKF_FORCEUNLOCK)))
2857 error = 0;
2858 out_put:
2859 dlm_put_lkb(lkb);
2860 out:
2861 dlm_unlock_recovery(ls);
2862 dlm_put_lockspace(ls);
2863 return error;
2864 }
2865
2866 /*
2867 * send/receive routines for remote operations and replies
2868 *
2869 * send_args
2870 * send_common
2871 * send_request receive_request
2872 * send_convert receive_convert
2873 * send_unlock receive_unlock
2874 * send_cancel receive_cancel
2875 * send_grant receive_grant
2876 * send_bast receive_bast
2877 * send_lookup receive_lookup
2878 * send_remove receive_remove
2879 *
2880 * send_common_reply
2881 * receive_request_reply send_request_reply
2882 * receive_convert_reply send_convert_reply
2883 * receive_unlock_reply send_unlock_reply
2884 * receive_cancel_reply send_cancel_reply
2885 * receive_lookup_reply send_lookup_reply
2886 */
2887
_create_message(struct dlm_ls * ls,int mb_len,int to_nodeid,int mstype,struct dlm_message ** ms_ret,struct dlm_mhandle ** mh_ret)2888 static int _create_message(struct dlm_ls *ls, int mb_len,
2889 int to_nodeid, int mstype,
2890 struct dlm_message **ms_ret,
2891 struct dlm_mhandle **mh_ret)
2892 {
2893 struct dlm_message *ms;
2894 struct dlm_mhandle *mh;
2895 char *mb;
2896
2897 /* get_buffer gives us a message handle (mh) that we need to
2898 pass into lowcomms_commit and a message buffer (mb) that we
2899 write our data into */
2900
2901 mh = dlm_lowcomms_get_buffer(to_nodeid, mb_len, GFP_NOFS, &mb);
2902 if (!mh)
2903 return -ENOBUFS;
2904
2905 memset(mb, 0, mb_len);
2906
2907 ms = (struct dlm_message *) mb;
2908
2909 ms->m_header.h_version = (DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
2910 ms->m_header.h_lockspace = ls->ls_global_id;
2911 ms->m_header.h_nodeid = dlm_our_nodeid();
2912 ms->m_header.h_length = mb_len;
2913 ms->m_header.h_cmd = DLM_MSG;
2914
2915 ms->m_type = mstype;
2916
2917 *mh_ret = mh;
2918 *ms_ret = ms;
2919 return 0;
2920 }
2921
create_message(struct dlm_rsb * r,struct dlm_lkb * lkb,int to_nodeid,int mstype,struct dlm_message ** ms_ret,struct dlm_mhandle ** mh_ret)2922 static int create_message(struct dlm_rsb *r, struct dlm_lkb *lkb,
2923 int to_nodeid, int mstype,
2924 struct dlm_message **ms_ret,
2925 struct dlm_mhandle **mh_ret)
2926 {
2927 int mb_len = sizeof(struct dlm_message);
2928
2929 switch (mstype) {
2930 case DLM_MSG_REQUEST:
2931 case DLM_MSG_LOOKUP:
2932 case DLM_MSG_REMOVE:
2933 mb_len += r->res_length;
2934 break;
2935 case DLM_MSG_CONVERT:
2936 case DLM_MSG_UNLOCK:
2937 case DLM_MSG_REQUEST_REPLY:
2938 case DLM_MSG_CONVERT_REPLY:
2939 case DLM_MSG_GRANT:
2940 if (lkb && lkb->lkb_lvbptr)
2941 mb_len += r->res_ls->ls_lvblen;
2942 break;
2943 }
2944
2945 return _create_message(r->res_ls, mb_len, to_nodeid, mstype,
2946 ms_ret, mh_ret);
2947 }
2948
2949 /* further lowcomms enhancements or alternate implementations may make
2950 the return value from this function useful at some point */
2951
send_message(struct dlm_mhandle * mh,struct dlm_message * ms)2952 static int send_message(struct dlm_mhandle *mh, struct dlm_message *ms)
2953 {
2954 dlm_message_out(ms);
2955 dlm_lowcomms_commit_buffer(mh);
2956 return 0;
2957 }
2958
send_args(struct dlm_rsb * r,struct dlm_lkb * lkb,struct dlm_message * ms)2959 static void send_args(struct dlm_rsb *r, struct dlm_lkb *lkb,
2960 struct dlm_message *ms)
2961 {
2962 ms->m_nodeid = lkb->lkb_nodeid;
2963 ms->m_pid = lkb->lkb_ownpid;
2964 ms->m_lkid = lkb->lkb_id;
2965 ms->m_remid = lkb->lkb_remid;
2966 ms->m_exflags = lkb->lkb_exflags;
2967 ms->m_sbflags = lkb->lkb_sbflags;
2968 ms->m_flags = lkb->lkb_flags;
2969 ms->m_lvbseq = lkb->lkb_lvbseq;
2970 ms->m_status = lkb->lkb_status;
2971 ms->m_grmode = lkb->lkb_grmode;
2972 ms->m_rqmode = lkb->lkb_rqmode;
2973 ms->m_hash = r->res_hash;
2974
2975 /* m_result and m_bastmode are set from function args,
2976 not from lkb fields */
2977
2978 if (lkb->lkb_bastfn)
2979 ms->m_asts |= DLM_CB_BAST;
2980 if (lkb->lkb_astfn)
2981 ms->m_asts |= DLM_CB_CAST;
2982
2983 /* compare with switch in create_message; send_remove() doesn't
2984 use send_args() */
2985
2986 switch (ms->m_type) {
2987 case DLM_MSG_REQUEST:
2988 case DLM_MSG_LOOKUP:
2989 memcpy(ms->m_extra, r->res_name, r->res_length);
2990 break;
2991 case DLM_MSG_CONVERT:
2992 case DLM_MSG_UNLOCK:
2993 case DLM_MSG_REQUEST_REPLY:
2994 case DLM_MSG_CONVERT_REPLY:
2995 case DLM_MSG_GRANT:
2996 if (!lkb->lkb_lvbptr)
2997 break;
2998 memcpy(ms->m_extra, lkb->lkb_lvbptr, r->res_ls->ls_lvblen);
2999 break;
3000 }
3001 }
3002
send_common(struct dlm_rsb * r,struct dlm_lkb * lkb,int mstype)3003 static int send_common(struct dlm_rsb *r, struct dlm_lkb *lkb, int mstype)
3004 {
3005 struct dlm_message *ms;
3006 struct dlm_mhandle *mh;
3007 int to_nodeid, error;
3008
3009 to_nodeid = r->res_nodeid;
3010
3011 error = add_to_waiters(lkb, mstype, to_nodeid);
3012 if (error)
3013 return error;
3014
3015 error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh);
3016 if (error)
3017 goto fail;
3018
3019 send_args(r, lkb, ms);
3020
3021 error = send_message(mh, ms);
3022 if (error)
3023 goto fail;
3024 return 0;
3025
3026 fail:
3027 remove_from_waiters(lkb, msg_reply_type(mstype));
3028 return error;
3029 }
3030
send_request(struct dlm_rsb * r,struct dlm_lkb * lkb)3031 static int send_request(struct dlm_rsb *r, struct dlm_lkb *lkb)
3032 {
3033 return send_common(r, lkb, DLM_MSG_REQUEST);
3034 }
3035
send_convert(struct dlm_rsb * r,struct dlm_lkb * lkb)3036 static int send_convert(struct dlm_rsb *r, struct dlm_lkb *lkb)
3037 {
3038 int error;
3039
3040 error = send_common(r, lkb, DLM_MSG_CONVERT);
3041
3042 /* down conversions go without a reply from the master */
3043 if (!error && down_conversion(lkb)) {
3044 remove_from_waiters(lkb, DLM_MSG_CONVERT_REPLY);
3045 r->res_ls->ls_stub_ms.m_flags = DLM_IFL_STUB_MS;
3046 r->res_ls->ls_stub_ms.m_type = DLM_MSG_CONVERT_REPLY;
3047 r->res_ls->ls_stub_ms.m_result = 0;
3048 __receive_convert_reply(r, lkb, &r->res_ls->ls_stub_ms);
3049 }
3050
3051 return error;
3052 }
3053
3054 /* FIXME: if this lkb is the only lock we hold on the rsb, then set
3055 MASTER_UNCERTAIN to force the next request on the rsb to confirm
3056 that the master is still correct. */
3057
send_unlock(struct dlm_rsb * r,struct dlm_lkb * lkb)3058 static int send_unlock(struct dlm_rsb *r, struct dlm_lkb *lkb)
3059 {
3060 return send_common(r, lkb, DLM_MSG_UNLOCK);
3061 }
3062
send_cancel(struct dlm_rsb * r,struct dlm_lkb * lkb)3063 static int send_cancel(struct dlm_rsb *r, struct dlm_lkb *lkb)
3064 {
3065 return send_common(r, lkb, DLM_MSG_CANCEL);
3066 }
3067
send_grant(struct dlm_rsb * r,struct dlm_lkb * lkb)3068 static int send_grant(struct dlm_rsb *r, struct dlm_lkb *lkb)
3069 {
3070 struct dlm_message *ms;
3071 struct dlm_mhandle *mh;
3072 int to_nodeid, error;
3073
3074 to_nodeid = lkb->lkb_nodeid;
3075
3076 error = create_message(r, lkb, to_nodeid, DLM_MSG_GRANT, &ms, &mh);
3077 if (error)
3078 goto out;
3079
3080 send_args(r, lkb, ms);
3081
3082 ms->m_result = 0;
3083
3084 error = send_message(mh, ms);
3085 out:
3086 return error;
3087 }
3088
send_bast(struct dlm_rsb * r,struct dlm_lkb * lkb,int mode)3089 static int send_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int mode)
3090 {
3091 struct dlm_message *ms;
3092 struct dlm_mhandle *mh;
3093 int to_nodeid, error;
3094
3095 to_nodeid = lkb->lkb_nodeid;
3096
3097 error = create_message(r, NULL, to_nodeid, DLM_MSG_BAST, &ms, &mh);
3098 if (error)
3099 goto out;
3100
3101 send_args(r, lkb, ms);
3102
3103 ms->m_bastmode = mode;
3104
3105 error = send_message(mh, ms);
3106 out:
3107 return error;
3108 }
3109
send_lookup(struct dlm_rsb * r,struct dlm_lkb * lkb)3110 static int send_lookup(struct dlm_rsb *r, struct dlm_lkb *lkb)
3111 {
3112 struct dlm_message *ms;
3113 struct dlm_mhandle *mh;
3114 int to_nodeid, error;
3115
3116 to_nodeid = dlm_dir_nodeid(r);
3117
3118 error = add_to_waiters(lkb, DLM_MSG_LOOKUP, to_nodeid);
3119 if (error)
3120 return error;
3121
3122 error = create_message(r, NULL, to_nodeid, DLM_MSG_LOOKUP, &ms, &mh);
3123 if (error)
3124 goto fail;
3125
3126 send_args(r, lkb, ms);
3127
3128 error = send_message(mh, ms);
3129 if (error)
3130 goto fail;
3131 return 0;
3132
3133 fail:
3134 remove_from_waiters(lkb, DLM_MSG_LOOKUP_REPLY);
3135 return error;
3136 }
3137
send_remove(struct dlm_rsb * r)3138 static int send_remove(struct dlm_rsb *r)
3139 {
3140 struct dlm_message *ms;
3141 struct dlm_mhandle *mh;
3142 int to_nodeid, error;
3143
3144 to_nodeid = dlm_dir_nodeid(r);
3145
3146 error = create_message(r, NULL, to_nodeid, DLM_MSG_REMOVE, &ms, &mh);
3147 if (error)
3148 goto out;
3149
3150 memcpy(ms->m_extra, r->res_name, r->res_length);
3151 ms->m_hash = r->res_hash;
3152
3153 error = send_message(mh, ms);
3154 out:
3155 return error;
3156 }
3157
send_common_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,int mstype,int rv)3158 static int send_common_reply(struct dlm_rsb *r, struct dlm_lkb *lkb,
3159 int mstype, int rv)
3160 {
3161 struct dlm_message *ms;
3162 struct dlm_mhandle *mh;
3163 int to_nodeid, error;
3164
3165 to_nodeid = lkb->lkb_nodeid;
3166
3167 error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh);
3168 if (error)
3169 goto out;
3170
3171 send_args(r, lkb, ms);
3172
3173 ms->m_result = rv;
3174
3175 error = send_message(mh, ms);
3176 out:
3177 return error;
3178 }
3179
send_request_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,int rv)3180 static int send_request_reply(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
3181 {
3182 return send_common_reply(r, lkb, DLM_MSG_REQUEST_REPLY, rv);
3183 }
3184
send_convert_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,int rv)3185 static int send_convert_reply(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
3186 {
3187 return send_common_reply(r, lkb, DLM_MSG_CONVERT_REPLY, rv);
3188 }
3189
send_unlock_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,int rv)3190 static int send_unlock_reply(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
3191 {
3192 return send_common_reply(r, lkb, DLM_MSG_UNLOCK_REPLY, rv);
3193 }
3194
send_cancel_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,int rv)3195 static int send_cancel_reply(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
3196 {
3197 return send_common_reply(r, lkb, DLM_MSG_CANCEL_REPLY, rv);
3198 }
3199
send_lookup_reply(struct dlm_ls * ls,struct dlm_message * ms_in,int ret_nodeid,int rv)3200 static int send_lookup_reply(struct dlm_ls *ls, struct dlm_message *ms_in,
3201 int ret_nodeid, int rv)
3202 {
3203 struct dlm_rsb *r = &ls->ls_stub_rsb;
3204 struct dlm_message *ms;
3205 struct dlm_mhandle *mh;
3206 int error, nodeid = ms_in->m_header.h_nodeid;
3207
3208 error = create_message(r, NULL, nodeid, DLM_MSG_LOOKUP_REPLY, &ms, &mh);
3209 if (error)
3210 goto out;
3211
3212 ms->m_lkid = ms_in->m_lkid;
3213 ms->m_result = rv;
3214 ms->m_nodeid = ret_nodeid;
3215
3216 error = send_message(mh, ms);
3217 out:
3218 return error;
3219 }
3220
3221 /* which args we save from a received message depends heavily on the type
3222 of message, unlike the send side where we can safely send everything about
3223 the lkb for any type of message */
3224
receive_flags(struct dlm_lkb * lkb,struct dlm_message * ms)3225 static void receive_flags(struct dlm_lkb *lkb, struct dlm_message *ms)
3226 {
3227 lkb->lkb_exflags = ms->m_exflags;
3228 lkb->lkb_sbflags = ms->m_sbflags;
3229 lkb->lkb_flags = (lkb->lkb_flags & 0xFFFF0000) |
3230 (ms->m_flags & 0x0000FFFF);
3231 }
3232
receive_flags_reply(struct dlm_lkb * lkb,struct dlm_message * ms)3233 static void receive_flags_reply(struct dlm_lkb *lkb, struct dlm_message *ms)
3234 {
3235 if (ms->m_flags == DLM_IFL_STUB_MS)
3236 return;
3237
3238 lkb->lkb_sbflags = ms->m_sbflags;
3239 lkb->lkb_flags = (lkb->lkb_flags & 0xFFFF0000) |
3240 (ms->m_flags & 0x0000FFFF);
3241 }
3242
receive_extralen(struct dlm_message * ms)3243 static int receive_extralen(struct dlm_message *ms)
3244 {
3245 return (ms->m_header.h_length - sizeof(struct dlm_message));
3246 }
3247
receive_lvb(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_message * ms)3248 static int receive_lvb(struct dlm_ls *ls, struct dlm_lkb *lkb,
3249 struct dlm_message *ms)
3250 {
3251 int len;
3252
3253 if (lkb->lkb_exflags & DLM_LKF_VALBLK) {
3254 if (!lkb->lkb_lvbptr)
3255 lkb->lkb_lvbptr = dlm_allocate_lvb(ls);
3256 if (!lkb->lkb_lvbptr)
3257 return -ENOMEM;
3258 len = receive_extralen(ms);
3259 if (len > DLM_RESNAME_MAXLEN)
3260 len = DLM_RESNAME_MAXLEN;
3261 memcpy(lkb->lkb_lvbptr, ms->m_extra, len);
3262 }
3263 return 0;
3264 }
3265
fake_bastfn(void * astparam,int mode)3266 static void fake_bastfn(void *astparam, int mode)
3267 {
3268 log_print("fake_bastfn should not be called");
3269 }
3270
fake_astfn(void * astparam)3271 static void fake_astfn(void *astparam)
3272 {
3273 log_print("fake_astfn should not be called");
3274 }
3275
receive_request_args(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_message * ms)3276 static int receive_request_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
3277 struct dlm_message *ms)
3278 {
3279 lkb->lkb_nodeid = ms->m_header.h_nodeid;
3280 lkb->lkb_ownpid = ms->m_pid;
3281 lkb->lkb_remid = ms->m_lkid;
3282 lkb->lkb_grmode = DLM_LOCK_IV;
3283 lkb->lkb_rqmode = ms->m_rqmode;
3284
3285 lkb->lkb_bastfn = (ms->m_asts & DLM_CB_BAST) ? &fake_bastfn : NULL;
3286 lkb->lkb_astfn = (ms->m_asts & DLM_CB_CAST) ? &fake_astfn : NULL;
3287
3288 if (lkb->lkb_exflags & DLM_LKF_VALBLK) {
3289 /* lkb was just created so there won't be an lvb yet */
3290 lkb->lkb_lvbptr = dlm_allocate_lvb(ls);
3291 if (!lkb->lkb_lvbptr)
3292 return -ENOMEM;
3293 }
3294
3295 return 0;
3296 }
3297
receive_convert_args(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_message * ms)3298 static int receive_convert_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
3299 struct dlm_message *ms)
3300 {
3301 if (lkb->lkb_status != DLM_LKSTS_GRANTED)
3302 return -EBUSY;
3303
3304 if (receive_lvb(ls, lkb, ms))
3305 return -ENOMEM;
3306
3307 lkb->lkb_rqmode = ms->m_rqmode;
3308 lkb->lkb_lvbseq = ms->m_lvbseq;
3309
3310 return 0;
3311 }
3312
receive_unlock_args(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_message * ms)3313 static int receive_unlock_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
3314 struct dlm_message *ms)
3315 {
3316 if (receive_lvb(ls, lkb, ms))
3317 return -ENOMEM;
3318 return 0;
3319 }
3320
3321 /* We fill in the stub-lkb fields with the info that send_xxxx_reply()
3322 uses to send a reply and that the remote end uses to process the reply. */
3323
setup_stub_lkb(struct dlm_ls * ls,struct dlm_message * ms)3324 static void setup_stub_lkb(struct dlm_ls *ls, struct dlm_message *ms)
3325 {
3326 struct dlm_lkb *lkb = &ls->ls_stub_lkb;
3327 lkb->lkb_nodeid = ms->m_header.h_nodeid;
3328 lkb->lkb_remid = ms->m_lkid;
3329 }
3330
3331 /* This is called after the rsb is locked so that we can safely inspect
3332 fields in the lkb. */
3333
validate_message(struct dlm_lkb * lkb,struct dlm_message * ms)3334 static int validate_message(struct dlm_lkb *lkb, struct dlm_message *ms)
3335 {
3336 int from = ms->m_header.h_nodeid;
3337 int error = 0;
3338
3339 switch (ms->m_type) {
3340 case DLM_MSG_CONVERT:
3341 case DLM_MSG_UNLOCK:
3342 case DLM_MSG_CANCEL:
3343 if (!is_master_copy(lkb) || lkb->lkb_nodeid != from)
3344 error = -EINVAL;
3345 break;
3346
3347 case DLM_MSG_CONVERT_REPLY:
3348 case DLM_MSG_UNLOCK_REPLY:
3349 case DLM_MSG_CANCEL_REPLY:
3350 case DLM_MSG_GRANT:
3351 case DLM_MSG_BAST:
3352 if (!is_process_copy(lkb) || lkb->lkb_nodeid != from)
3353 error = -EINVAL;
3354 break;
3355
3356 case DLM_MSG_REQUEST_REPLY:
3357 if (!is_process_copy(lkb))
3358 error = -EINVAL;
3359 else if (lkb->lkb_nodeid != -1 && lkb->lkb_nodeid != from)
3360 error = -EINVAL;
3361 break;
3362
3363 default:
3364 error = -EINVAL;
3365 }
3366
3367 if (error)
3368 log_error(lkb->lkb_resource->res_ls,
3369 "ignore invalid message %d from %d %x %x %x %d",
3370 ms->m_type, from, lkb->lkb_id, lkb->lkb_remid,
3371 lkb->lkb_flags, lkb->lkb_nodeid);
3372 return error;
3373 }
3374
receive_request(struct dlm_ls * ls,struct dlm_message * ms)3375 static void receive_request(struct dlm_ls *ls, struct dlm_message *ms)
3376 {
3377 struct dlm_lkb *lkb;
3378 struct dlm_rsb *r;
3379 int error, namelen;
3380
3381 error = create_lkb(ls, &lkb);
3382 if (error)
3383 goto fail;
3384
3385 receive_flags(lkb, ms);
3386 lkb->lkb_flags |= DLM_IFL_MSTCPY;
3387 error = receive_request_args(ls, lkb, ms);
3388 if (error) {
3389 __put_lkb(ls, lkb);
3390 goto fail;
3391 }
3392
3393 namelen = receive_extralen(ms);
3394
3395 error = find_rsb(ls, ms->m_extra, namelen, R_MASTER, &r);
3396 if (error) {
3397 __put_lkb(ls, lkb);
3398 goto fail;
3399 }
3400
3401 lock_rsb(r);
3402
3403 attach_lkb(r, lkb);
3404 error = do_request(r, lkb);
3405 send_request_reply(r, lkb, error);
3406 do_request_effects(r, lkb, error);
3407
3408 unlock_rsb(r);
3409 put_rsb(r);
3410
3411 if (error == -EINPROGRESS)
3412 error = 0;
3413 if (error)
3414 dlm_put_lkb(lkb);
3415 return;
3416
3417 fail:
3418 setup_stub_lkb(ls, ms);
3419 send_request_reply(&ls->ls_stub_rsb, &ls->ls_stub_lkb, error);
3420 }
3421
receive_convert(struct dlm_ls * ls,struct dlm_message * ms)3422 static void receive_convert(struct dlm_ls *ls, struct dlm_message *ms)
3423 {
3424 struct dlm_lkb *lkb;
3425 struct dlm_rsb *r;
3426 int error, reply = 1;
3427
3428 error = find_lkb(ls, ms->m_remid, &lkb);
3429 if (error)
3430 goto fail;
3431
3432 r = lkb->lkb_resource;
3433
3434 hold_rsb(r);
3435 lock_rsb(r);
3436
3437 error = validate_message(lkb, ms);
3438 if (error)
3439 goto out;
3440
3441 receive_flags(lkb, ms);
3442
3443 error = receive_convert_args(ls, lkb, ms);
3444 if (error) {
3445 send_convert_reply(r, lkb, error);
3446 goto out;
3447 }
3448
3449 reply = !down_conversion(lkb);
3450
3451 error = do_convert(r, lkb);
3452 if (reply)
3453 send_convert_reply(r, lkb, error);
3454 do_convert_effects(r, lkb, error);
3455 out:
3456 unlock_rsb(r);
3457 put_rsb(r);
3458 dlm_put_lkb(lkb);
3459 return;
3460
3461 fail:
3462 setup_stub_lkb(ls, ms);
3463 send_convert_reply(&ls->ls_stub_rsb, &ls->ls_stub_lkb, error);
3464 }
3465
receive_unlock(struct dlm_ls * ls,struct dlm_message * ms)3466 static void receive_unlock(struct dlm_ls *ls, struct dlm_message *ms)
3467 {
3468 struct dlm_lkb *lkb;
3469 struct dlm_rsb *r;
3470 int error;
3471
3472 error = find_lkb(ls, ms->m_remid, &lkb);
3473 if (error)
3474 goto fail;
3475
3476 r = lkb->lkb_resource;
3477
3478 hold_rsb(r);
3479 lock_rsb(r);
3480
3481 error = validate_message(lkb, ms);
3482 if (error)
3483 goto out;
3484
3485 receive_flags(lkb, ms);
3486
3487 error = receive_unlock_args(ls, lkb, ms);
3488 if (error) {
3489 send_unlock_reply(r, lkb, error);
3490 goto out;
3491 }
3492
3493 error = do_unlock(r, lkb);
3494 send_unlock_reply(r, lkb, error);
3495 do_unlock_effects(r, lkb, error);
3496 out:
3497 unlock_rsb(r);
3498 put_rsb(r);
3499 dlm_put_lkb(lkb);
3500 return;
3501
3502 fail:
3503 setup_stub_lkb(ls, ms);
3504 send_unlock_reply(&ls->ls_stub_rsb, &ls->ls_stub_lkb, error);
3505 }
3506
receive_cancel(struct dlm_ls * ls,struct dlm_message * ms)3507 static void receive_cancel(struct dlm_ls *ls, struct dlm_message *ms)
3508 {
3509 struct dlm_lkb *lkb;
3510 struct dlm_rsb *r;
3511 int error;
3512
3513 error = find_lkb(ls, ms->m_remid, &lkb);
3514 if (error)
3515 goto fail;
3516
3517 receive_flags(lkb, ms);
3518
3519 r = lkb->lkb_resource;
3520
3521 hold_rsb(r);
3522 lock_rsb(r);
3523
3524 error = validate_message(lkb, ms);
3525 if (error)
3526 goto out;
3527
3528 error = do_cancel(r, lkb);
3529 send_cancel_reply(r, lkb, error);
3530 do_cancel_effects(r, lkb, error);
3531 out:
3532 unlock_rsb(r);
3533 put_rsb(r);
3534 dlm_put_lkb(lkb);
3535 return;
3536
3537 fail:
3538 setup_stub_lkb(ls, ms);
3539 send_cancel_reply(&ls->ls_stub_rsb, &ls->ls_stub_lkb, error);
3540 }
3541
receive_grant(struct dlm_ls * ls,struct dlm_message * ms)3542 static void receive_grant(struct dlm_ls *ls, struct dlm_message *ms)
3543 {
3544 struct dlm_lkb *lkb;
3545 struct dlm_rsb *r;
3546 int error;
3547
3548 error = find_lkb(ls, ms->m_remid, &lkb);
3549 if (error) {
3550 log_debug(ls, "receive_grant from %d no lkb %x",
3551 ms->m_header.h_nodeid, ms->m_remid);
3552 return;
3553 }
3554
3555 r = lkb->lkb_resource;
3556
3557 hold_rsb(r);
3558 lock_rsb(r);
3559
3560 error = validate_message(lkb, ms);
3561 if (error)
3562 goto out;
3563
3564 receive_flags_reply(lkb, ms);
3565 if (is_altmode(lkb))
3566 munge_altmode(lkb, ms);
3567 grant_lock_pc(r, lkb, ms);
3568 queue_cast(r, lkb, 0);
3569 out:
3570 unlock_rsb(r);
3571 put_rsb(r);
3572 dlm_put_lkb(lkb);
3573 }
3574
receive_bast(struct dlm_ls * ls,struct dlm_message * ms)3575 static void receive_bast(struct dlm_ls *ls, struct dlm_message *ms)
3576 {
3577 struct dlm_lkb *lkb;
3578 struct dlm_rsb *r;
3579 int error;
3580
3581 error = find_lkb(ls, ms->m_remid, &lkb);
3582 if (error) {
3583 log_debug(ls, "receive_bast from %d no lkb %x",
3584 ms->m_header.h_nodeid, ms->m_remid);
3585 return;
3586 }
3587
3588 r = lkb->lkb_resource;
3589
3590 hold_rsb(r);
3591 lock_rsb(r);
3592
3593 error = validate_message(lkb, ms);
3594 if (error)
3595 goto out;
3596
3597 queue_bast(r, lkb, ms->m_bastmode);
3598 out:
3599 unlock_rsb(r);
3600 put_rsb(r);
3601 dlm_put_lkb(lkb);
3602 }
3603
receive_lookup(struct dlm_ls * ls,struct dlm_message * ms)3604 static void receive_lookup(struct dlm_ls *ls, struct dlm_message *ms)
3605 {
3606 int len, error, ret_nodeid, dir_nodeid, from_nodeid, our_nodeid;
3607
3608 from_nodeid = ms->m_header.h_nodeid;
3609 our_nodeid = dlm_our_nodeid();
3610
3611 len = receive_extralen(ms);
3612
3613 dir_nodeid = dlm_hash2nodeid(ls, ms->m_hash);
3614 if (dir_nodeid != our_nodeid) {
3615 log_error(ls, "lookup dir_nodeid %d from %d",
3616 dir_nodeid, from_nodeid);
3617 error = -EINVAL;
3618 ret_nodeid = -1;
3619 goto out;
3620 }
3621
3622 error = dlm_dir_lookup(ls, from_nodeid, ms->m_extra, len, &ret_nodeid);
3623
3624 /* Optimization: we're master so treat lookup as a request */
3625 if (!error && ret_nodeid == our_nodeid) {
3626 receive_request(ls, ms);
3627 return;
3628 }
3629 out:
3630 send_lookup_reply(ls, ms, ret_nodeid, error);
3631 }
3632
receive_remove(struct dlm_ls * ls,struct dlm_message * ms)3633 static void receive_remove(struct dlm_ls *ls, struct dlm_message *ms)
3634 {
3635 int len, dir_nodeid, from_nodeid;
3636
3637 from_nodeid = ms->m_header.h_nodeid;
3638
3639 len = receive_extralen(ms);
3640
3641 dir_nodeid = dlm_hash2nodeid(ls, ms->m_hash);
3642 if (dir_nodeid != dlm_our_nodeid()) {
3643 log_error(ls, "remove dir entry dir_nodeid %d from %d",
3644 dir_nodeid, from_nodeid);
3645 return;
3646 }
3647
3648 dlm_dir_remove_entry(ls, from_nodeid, ms->m_extra, len);
3649 }
3650
receive_purge(struct dlm_ls * ls,struct dlm_message * ms)3651 static void receive_purge(struct dlm_ls *ls, struct dlm_message *ms)
3652 {
3653 do_purge(ls, ms->m_nodeid, ms->m_pid);
3654 }
3655
receive_request_reply(struct dlm_ls * ls,struct dlm_message * ms)3656 static void receive_request_reply(struct dlm_ls *ls, struct dlm_message *ms)
3657 {
3658 struct dlm_lkb *lkb;
3659 struct dlm_rsb *r;
3660 int error, mstype, result;
3661
3662 error = find_lkb(ls, ms->m_remid, &lkb);
3663 if (error) {
3664 log_debug(ls, "receive_request_reply from %d no lkb %x",
3665 ms->m_header.h_nodeid, ms->m_remid);
3666 return;
3667 }
3668
3669 r = lkb->lkb_resource;
3670 hold_rsb(r);
3671 lock_rsb(r);
3672
3673 error = validate_message(lkb, ms);
3674 if (error)
3675 goto out;
3676
3677 mstype = lkb->lkb_wait_type;
3678 error = remove_from_waiters(lkb, DLM_MSG_REQUEST_REPLY);
3679 if (error)
3680 goto out;
3681
3682 /* Optimization: the dir node was also the master, so it took our
3683 lookup as a request and sent request reply instead of lookup reply */
3684 if (mstype == DLM_MSG_LOOKUP) {
3685 r->res_nodeid = ms->m_header.h_nodeid;
3686 lkb->lkb_nodeid = r->res_nodeid;
3687 }
3688
3689 /* this is the value returned from do_request() on the master */
3690 result = ms->m_result;
3691
3692 switch (result) {
3693 case -EAGAIN:
3694 /* request would block (be queued) on remote master */
3695 queue_cast(r, lkb, -EAGAIN);
3696 confirm_master(r, -EAGAIN);
3697 unhold_lkb(lkb); /* undoes create_lkb() */
3698 break;
3699
3700 case -EINPROGRESS:
3701 case 0:
3702 /* request was queued or granted on remote master */
3703 receive_flags_reply(lkb, ms);
3704 lkb->lkb_remid = ms->m_lkid;
3705 if (is_altmode(lkb))
3706 munge_altmode(lkb, ms);
3707 if (result) {
3708 add_lkb(r, lkb, DLM_LKSTS_WAITING);
3709 add_timeout(lkb);
3710 } else {
3711 grant_lock_pc(r, lkb, ms);
3712 queue_cast(r, lkb, 0);
3713 }
3714 confirm_master(r, result);
3715 break;
3716
3717 case -EBADR:
3718 case -ENOTBLK:
3719 /* find_rsb failed to find rsb or rsb wasn't master */
3720 log_debug(ls, "receive_request_reply %x %x master diff %d %d",
3721 lkb->lkb_id, lkb->lkb_flags, r->res_nodeid, result);
3722 r->res_nodeid = -1;
3723 lkb->lkb_nodeid = -1;
3724
3725 if (is_overlap(lkb)) {
3726 /* we'll ignore error in cancel/unlock reply */
3727 queue_cast_overlap(r, lkb);
3728 confirm_master(r, result);
3729 unhold_lkb(lkb); /* undoes create_lkb() */
3730 } else
3731 _request_lock(r, lkb);
3732 break;
3733
3734 default:
3735 log_error(ls, "receive_request_reply %x error %d",
3736 lkb->lkb_id, result);
3737 }
3738
3739 if (is_overlap_unlock(lkb) && (result == 0 || result == -EINPROGRESS)) {
3740 log_debug(ls, "receive_request_reply %x result %d unlock",
3741 lkb->lkb_id, result);
3742 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_UNLOCK;
3743 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
3744 send_unlock(r, lkb);
3745 } else if (is_overlap_cancel(lkb) && (result == -EINPROGRESS)) {
3746 log_debug(ls, "receive_request_reply %x cancel", lkb->lkb_id);
3747 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_UNLOCK;
3748 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
3749 send_cancel(r, lkb);
3750 } else {
3751 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
3752 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_UNLOCK;
3753 }
3754 out:
3755 unlock_rsb(r);
3756 put_rsb(r);
3757 dlm_put_lkb(lkb);
3758 }
3759
__receive_convert_reply(struct dlm_rsb * r,struct dlm_lkb * lkb,struct dlm_message * ms)3760 static void __receive_convert_reply(struct dlm_rsb *r, struct dlm_lkb *lkb,
3761 struct dlm_message *ms)
3762 {
3763 /* this is the value returned from do_convert() on the master */
3764 switch (ms->m_result) {
3765 case -EAGAIN:
3766 /* convert would block (be queued) on remote master */
3767 queue_cast(r, lkb, -EAGAIN);
3768 break;
3769
3770 case -EDEADLK:
3771 receive_flags_reply(lkb, ms);
3772 revert_lock_pc(r, lkb);
3773 queue_cast(r, lkb, -EDEADLK);
3774 break;
3775
3776 case -EINPROGRESS:
3777 /* convert was queued on remote master */
3778 receive_flags_reply(lkb, ms);
3779 if (is_demoted(lkb))
3780 munge_demoted(lkb);
3781 del_lkb(r, lkb);
3782 add_lkb(r, lkb, DLM_LKSTS_CONVERT);
3783 add_timeout(lkb);
3784 break;
3785
3786 case 0:
3787 /* convert was granted on remote master */
3788 receive_flags_reply(lkb, ms);
3789 if (is_demoted(lkb))
3790 munge_demoted(lkb);
3791 grant_lock_pc(r, lkb, ms);
3792 queue_cast(r, lkb, 0);
3793 break;
3794
3795 default:
3796 log_error(r->res_ls, "receive_convert_reply %x error %d",
3797 lkb->lkb_id, ms->m_result);
3798 }
3799 }
3800
_receive_convert_reply(struct dlm_lkb * lkb,struct dlm_message * ms)3801 static void _receive_convert_reply(struct dlm_lkb *lkb, struct dlm_message *ms)
3802 {
3803 struct dlm_rsb *r = lkb->lkb_resource;
3804 int error;
3805
3806 hold_rsb(r);
3807 lock_rsb(r);
3808
3809 error = validate_message(lkb, ms);
3810 if (error)
3811 goto out;
3812
3813 /* stub reply can happen with waiters_mutex held */
3814 error = remove_from_waiters_ms(lkb, ms);
3815 if (error)
3816 goto out;
3817
3818 __receive_convert_reply(r, lkb, ms);
3819 out:
3820 unlock_rsb(r);
3821 put_rsb(r);
3822 }
3823
receive_convert_reply(struct dlm_ls * ls,struct dlm_message * ms)3824 static void receive_convert_reply(struct dlm_ls *ls, struct dlm_message *ms)
3825 {
3826 struct dlm_lkb *lkb;
3827 int error;
3828
3829 error = find_lkb(ls, ms->m_remid, &lkb);
3830 if (error) {
3831 log_debug(ls, "receive_convert_reply from %d no lkb %x",
3832 ms->m_header.h_nodeid, ms->m_remid);
3833 return;
3834 }
3835
3836 _receive_convert_reply(lkb, ms);
3837 dlm_put_lkb(lkb);
3838 }
3839
_receive_unlock_reply(struct dlm_lkb * lkb,struct dlm_message * ms)3840 static void _receive_unlock_reply(struct dlm_lkb *lkb, struct dlm_message *ms)
3841 {
3842 struct dlm_rsb *r = lkb->lkb_resource;
3843 int error;
3844
3845 hold_rsb(r);
3846 lock_rsb(r);
3847
3848 error = validate_message(lkb, ms);
3849 if (error)
3850 goto out;
3851
3852 /* stub reply can happen with waiters_mutex held */
3853 error = remove_from_waiters_ms(lkb, ms);
3854 if (error)
3855 goto out;
3856
3857 /* this is the value returned from do_unlock() on the master */
3858
3859 switch (ms->m_result) {
3860 case -DLM_EUNLOCK:
3861 receive_flags_reply(lkb, ms);
3862 remove_lock_pc(r, lkb);
3863 queue_cast(r, lkb, -DLM_EUNLOCK);
3864 break;
3865 case -ENOENT:
3866 break;
3867 default:
3868 log_error(r->res_ls, "receive_unlock_reply %x error %d",
3869 lkb->lkb_id, ms->m_result);
3870 }
3871 out:
3872 unlock_rsb(r);
3873 put_rsb(r);
3874 }
3875
receive_unlock_reply(struct dlm_ls * ls,struct dlm_message * ms)3876 static void receive_unlock_reply(struct dlm_ls *ls, struct dlm_message *ms)
3877 {
3878 struct dlm_lkb *lkb;
3879 int error;
3880
3881 error = find_lkb(ls, ms->m_remid, &lkb);
3882 if (error) {
3883 log_debug(ls, "receive_unlock_reply from %d no lkb %x",
3884 ms->m_header.h_nodeid, ms->m_remid);
3885 return;
3886 }
3887
3888 _receive_unlock_reply(lkb, ms);
3889 dlm_put_lkb(lkb);
3890 }
3891
_receive_cancel_reply(struct dlm_lkb * lkb,struct dlm_message * ms)3892 static void _receive_cancel_reply(struct dlm_lkb *lkb, struct dlm_message *ms)
3893 {
3894 struct dlm_rsb *r = lkb->lkb_resource;
3895 int error;
3896
3897 hold_rsb(r);
3898 lock_rsb(r);
3899
3900 error = validate_message(lkb, ms);
3901 if (error)
3902 goto out;
3903
3904 /* stub reply can happen with waiters_mutex held */
3905 error = remove_from_waiters_ms(lkb, ms);
3906 if (error)
3907 goto out;
3908
3909 /* this is the value returned from do_cancel() on the master */
3910
3911 switch (ms->m_result) {
3912 case -DLM_ECANCEL:
3913 receive_flags_reply(lkb, ms);
3914 revert_lock_pc(r, lkb);
3915 queue_cast(r, lkb, -DLM_ECANCEL);
3916 break;
3917 case 0:
3918 break;
3919 default:
3920 log_error(r->res_ls, "receive_cancel_reply %x error %d",
3921 lkb->lkb_id, ms->m_result);
3922 }
3923 out:
3924 unlock_rsb(r);
3925 put_rsb(r);
3926 }
3927
receive_cancel_reply(struct dlm_ls * ls,struct dlm_message * ms)3928 static void receive_cancel_reply(struct dlm_ls *ls, struct dlm_message *ms)
3929 {
3930 struct dlm_lkb *lkb;
3931 int error;
3932
3933 error = find_lkb(ls, ms->m_remid, &lkb);
3934 if (error) {
3935 log_debug(ls, "receive_cancel_reply from %d no lkb %x",
3936 ms->m_header.h_nodeid, ms->m_remid);
3937 return;
3938 }
3939
3940 _receive_cancel_reply(lkb, ms);
3941 dlm_put_lkb(lkb);
3942 }
3943
receive_lookup_reply(struct dlm_ls * ls,struct dlm_message * ms)3944 static void receive_lookup_reply(struct dlm_ls *ls, struct dlm_message *ms)
3945 {
3946 struct dlm_lkb *lkb;
3947 struct dlm_rsb *r;
3948 int error, ret_nodeid;
3949
3950 error = find_lkb(ls, ms->m_lkid, &lkb);
3951 if (error) {
3952 log_error(ls, "receive_lookup_reply no lkb");
3953 return;
3954 }
3955
3956 /* ms->m_result is the value returned by dlm_dir_lookup on dir node
3957 FIXME: will a non-zero error ever be returned? */
3958
3959 r = lkb->lkb_resource;
3960 hold_rsb(r);
3961 lock_rsb(r);
3962
3963 error = remove_from_waiters(lkb, DLM_MSG_LOOKUP_REPLY);
3964 if (error)
3965 goto out;
3966
3967 ret_nodeid = ms->m_nodeid;
3968 if (ret_nodeid == dlm_our_nodeid()) {
3969 r->res_nodeid = 0;
3970 ret_nodeid = 0;
3971 r->res_first_lkid = 0;
3972 } else {
3973 /* set_master() will copy res_nodeid to lkb_nodeid */
3974 r->res_nodeid = ret_nodeid;
3975 }
3976
3977 if (is_overlap(lkb)) {
3978 log_debug(ls, "receive_lookup_reply %x unlock %x",
3979 lkb->lkb_id, lkb->lkb_flags);
3980 queue_cast_overlap(r, lkb);
3981 unhold_lkb(lkb); /* undoes create_lkb() */
3982 goto out_list;
3983 }
3984
3985 _request_lock(r, lkb);
3986
3987 out_list:
3988 if (!ret_nodeid)
3989 process_lookup_list(r);
3990 out:
3991 unlock_rsb(r);
3992 put_rsb(r);
3993 dlm_put_lkb(lkb);
3994 }
3995
_receive_message(struct dlm_ls * ls,struct dlm_message * ms)3996 static void _receive_message(struct dlm_ls *ls, struct dlm_message *ms)
3997 {
3998 if (!dlm_is_member(ls, ms->m_header.h_nodeid)) {
3999 log_debug(ls, "ignore non-member message %d from %d %x %x %d",
4000 ms->m_type, ms->m_header.h_nodeid, ms->m_lkid,
4001 ms->m_remid, ms->m_result);
4002 return;
4003 }
4004
4005 switch (ms->m_type) {
4006
4007 /* messages sent to a master node */
4008
4009 case DLM_MSG_REQUEST:
4010 receive_request(ls, ms);
4011 break;
4012
4013 case DLM_MSG_CONVERT:
4014 receive_convert(ls, ms);
4015 break;
4016
4017 case DLM_MSG_UNLOCK:
4018 receive_unlock(ls, ms);
4019 break;
4020
4021 case DLM_MSG_CANCEL:
4022 receive_cancel(ls, ms);
4023 break;
4024
4025 /* messages sent from a master node (replies to above) */
4026
4027 case DLM_MSG_REQUEST_REPLY:
4028 receive_request_reply(ls, ms);
4029 break;
4030
4031 case DLM_MSG_CONVERT_REPLY:
4032 receive_convert_reply(ls, ms);
4033 break;
4034
4035 case DLM_MSG_UNLOCK_REPLY:
4036 receive_unlock_reply(ls, ms);
4037 break;
4038
4039 case DLM_MSG_CANCEL_REPLY:
4040 receive_cancel_reply(ls, ms);
4041 break;
4042
4043 /* messages sent from a master node (only two types of async msg) */
4044
4045 case DLM_MSG_GRANT:
4046 receive_grant(ls, ms);
4047 break;
4048
4049 case DLM_MSG_BAST:
4050 receive_bast(ls, ms);
4051 break;
4052
4053 /* messages sent to a dir node */
4054
4055 case DLM_MSG_LOOKUP:
4056 receive_lookup(ls, ms);
4057 break;
4058
4059 case DLM_MSG_REMOVE:
4060 receive_remove(ls, ms);
4061 break;
4062
4063 /* messages sent from a dir node (remove has no reply) */
4064
4065 case DLM_MSG_LOOKUP_REPLY:
4066 receive_lookup_reply(ls, ms);
4067 break;
4068
4069 /* other messages */
4070
4071 case DLM_MSG_PURGE:
4072 receive_purge(ls, ms);
4073 break;
4074
4075 default:
4076 log_error(ls, "unknown message type %d", ms->m_type);
4077 }
4078 }
4079
4080 /* If the lockspace is in recovery mode (locking stopped), then normal
4081 messages are saved on the requestqueue for processing after recovery is
4082 done. When not in recovery mode, we wait for dlm_recoverd to drain saved
4083 messages off the requestqueue before we process new ones. This occurs right
4084 after recovery completes when we transition from saving all messages on
4085 requestqueue, to processing all the saved messages, to processing new
4086 messages as they arrive. */
4087
dlm_receive_message(struct dlm_ls * ls,struct dlm_message * ms,int nodeid)4088 static void dlm_receive_message(struct dlm_ls *ls, struct dlm_message *ms,
4089 int nodeid)
4090 {
4091 if (dlm_locking_stopped(ls)) {
4092 dlm_add_requestqueue(ls, nodeid, ms);
4093 } else {
4094 dlm_wait_requestqueue(ls);
4095 _receive_message(ls, ms);
4096 }
4097 }
4098
4099 /* This is called by dlm_recoverd to process messages that were saved on
4100 the requestqueue. */
4101
dlm_receive_message_saved(struct dlm_ls * ls,struct dlm_message * ms)4102 void dlm_receive_message_saved(struct dlm_ls *ls, struct dlm_message *ms)
4103 {
4104 _receive_message(ls, ms);
4105 }
4106
4107 /* This is called by the midcomms layer when something is received for
4108 the lockspace. It could be either a MSG (normal message sent as part of
4109 standard locking activity) or an RCOM (recovery message sent as part of
4110 lockspace recovery). */
4111
dlm_receive_buffer(union dlm_packet * p,int nodeid)4112 void dlm_receive_buffer(union dlm_packet *p, int nodeid)
4113 {
4114 struct dlm_header *hd = &p->header;
4115 struct dlm_ls *ls;
4116 int type = 0;
4117
4118 switch (hd->h_cmd) {
4119 case DLM_MSG:
4120 dlm_message_in(&p->message);
4121 type = p->message.m_type;
4122 break;
4123 case DLM_RCOM:
4124 dlm_rcom_in(&p->rcom);
4125 type = p->rcom.rc_type;
4126 break;
4127 default:
4128 log_print("invalid h_cmd %d from %u", hd->h_cmd, nodeid);
4129 return;
4130 }
4131
4132 if (hd->h_nodeid != nodeid) {
4133 log_print("invalid h_nodeid %d from %d lockspace %x",
4134 hd->h_nodeid, nodeid, hd->h_lockspace);
4135 return;
4136 }
4137
4138 ls = dlm_find_lockspace_global(hd->h_lockspace);
4139 if (!ls) {
4140 if (dlm_config.ci_log_debug)
4141 log_print("invalid lockspace %x from %d cmd %d type %d",
4142 hd->h_lockspace, nodeid, hd->h_cmd, type);
4143
4144 if (hd->h_cmd == DLM_RCOM && type == DLM_RCOM_STATUS)
4145 dlm_send_ls_not_ready(nodeid, &p->rcom);
4146 return;
4147 }
4148
4149 /* this rwsem allows dlm_ls_stop() to wait for all dlm_recv threads to
4150 be inactive (in this ls) before transitioning to recovery mode */
4151
4152 down_read(&ls->ls_recv_active);
4153 if (hd->h_cmd == DLM_MSG)
4154 dlm_receive_message(ls, &p->message, nodeid);
4155 else
4156 dlm_receive_rcom(ls, &p->rcom, nodeid);
4157 up_read(&ls->ls_recv_active);
4158
4159 dlm_put_lockspace(ls);
4160 }
4161
recover_convert_waiter(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_message * ms_stub)4162 static void recover_convert_waiter(struct dlm_ls *ls, struct dlm_lkb *lkb,
4163 struct dlm_message *ms_stub)
4164 {
4165 if (middle_conversion(lkb)) {
4166 hold_lkb(lkb);
4167 memset(ms_stub, 0, sizeof(struct dlm_message));
4168 ms_stub->m_flags = DLM_IFL_STUB_MS;
4169 ms_stub->m_type = DLM_MSG_CONVERT_REPLY;
4170 ms_stub->m_result = -EINPROGRESS;
4171 ms_stub->m_header.h_nodeid = lkb->lkb_nodeid;
4172 _receive_convert_reply(lkb, ms_stub);
4173
4174 /* Same special case as in receive_rcom_lock_args() */
4175 lkb->lkb_grmode = DLM_LOCK_IV;
4176 rsb_set_flag(lkb->lkb_resource, RSB_RECOVER_CONVERT);
4177 unhold_lkb(lkb);
4178
4179 } else if (lkb->lkb_rqmode >= lkb->lkb_grmode) {
4180 lkb->lkb_flags |= DLM_IFL_RESEND;
4181 }
4182
4183 /* lkb->lkb_rqmode < lkb->lkb_grmode shouldn't happen since down
4184 conversions are async; there's no reply from the remote master */
4185 }
4186
4187 /* A waiting lkb needs recovery if the master node has failed, or
4188 the master node is changing (only when no directory is used) */
4189
waiter_needs_recovery(struct dlm_ls * ls,struct dlm_lkb * lkb)4190 static int waiter_needs_recovery(struct dlm_ls *ls, struct dlm_lkb *lkb)
4191 {
4192 if (dlm_is_removed(ls, lkb->lkb_nodeid))
4193 return 1;
4194
4195 if (!dlm_no_directory(ls))
4196 return 0;
4197
4198 if (dlm_dir_nodeid(lkb->lkb_resource) != lkb->lkb_nodeid)
4199 return 1;
4200
4201 return 0;
4202 }
4203
4204 /* Recovery for locks that are waiting for replies from nodes that are now
4205 gone. We can just complete unlocks and cancels by faking a reply from the
4206 dead node. Requests and up-conversions we flag to be resent after
4207 recovery. Down-conversions can just be completed with a fake reply like
4208 unlocks. Conversions between PR and CW need special attention. */
4209
dlm_recover_waiters_pre(struct dlm_ls * ls)4210 void dlm_recover_waiters_pre(struct dlm_ls *ls)
4211 {
4212 struct dlm_lkb *lkb, *safe;
4213 struct dlm_message *ms_stub;
4214 int wait_type, stub_unlock_result, stub_cancel_result;
4215
4216 ms_stub = kmalloc(sizeof(struct dlm_message), GFP_KERNEL);
4217 if (!ms_stub) {
4218 log_error(ls, "dlm_recover_waiters_pre no mem");
4219 return;
4220 }
4221
4222 mutex_lock(&ls->ls_waiters_mutex);
4223
4224 list_for_each_entry_safe(lkb, safe, &ls->ls_waiters, lkb_wait_reply) {
4225
4226 /* exclude debug messages about unlocks because there can be so
4227 many and they aren't very interesting */
4228
4229 if (lkb->lkb_wait_type != DLM_MSG_UNLOCK) {
4230 log_debug(ls, "recover_waiter %x nodeid %d "
4231 "msg %d to %d", lkb->lkb_id, lkb->lkb_nodeid,
4232 lkb->lkb_wait_type, lkb->lkb_wait_nodeid);
4233 }
4234
4235 /* all outstanding lookups, regardless of destination will be
4236 resent after recovery is done */
4237
4238 if (lkb->lkb_wait_type == DLM_MSG_LOOKUP) {
4239 lkb->lkb_flags |= DLM_IFL_RESEND;
4240 continue;
4241 }
4242
4243 if (!waiter_needs_recovery(ls, lkb))
4244 continue;
4245
4246 wait_type = lkb->lkb_wait_type;
4247 stub_unlock_result = -DLM_EUNLOCK;
4248 stub_cancel_result = -DLM_ECANCEL;
4249
4250 /* Main reply may have been received leaving a zero wait_type,
4251 but a reply for the overlapping op may not have been
4252 received. In that case we need to fake the appropriate
4253 reply for the overlap op. */
4254
4255 if (!wait_type) {
4256 if (is_overlap_cancel(lkb)) {
4257 wait_type = DLM_MSG_CANCEL;
4258 if (lkb->lkb_grmode == DLM_LOCK_IV)
4259 stub_cancel_result = 0;
4260 }
4261 if (is_overlap_unlock(lkb)) {
4262 wait_type = DLM_MSG_UNLOCK;
4263 if (lkb->lkb_grmode == DLM_LOCK_IV)
4264 stub_unlock_result = -ENOENT;
4265 }
4266
4267 log_debug(ls, "rwpre overlap %x %x %d %d %d",
4268 lkb->lkb_id, lkb->lkb_flags, wait_type,
4269 stub_cancel_result, stub_unlock_result);
4270 }
4271
4272 switch (wait_type) {
4273
4274 case DLM_MSG_REQUEST:
4275 lkb->lkb_flags |= DLM_IFL_RESEND;
4276 break;
4277
4278 case DLM_MSG_CONVERT:
4279 recover_convert_waiter(ls, lkb, ms_stub);
4280 break;
4281
4282 case DLM_MSG_UNLOCK:
4283 hold_lkb(lkb);
4284 memset(ms_stub, 0, sizeof(struct dlm_message));
4285 ms_stub->m_flags = DLM_IFL_STUB_MS;
4286 ms_stub->m_type = DLM_MSG_UNLOCK_REPLY;
4287 ms_stub->m_result = stub_unlock_result;
4288 ms_stub->m_header.h_nodeid = lkb->lkb_nodeid;
4289 _receive_unlock_reply(lkb, ms_stub);
4290 dlm_put_lkb(lkb);
4291 break;
4292
4293 case DLM_MSG_CANCEL:
4294 hold_lkb(lkb);
4295 memset(ms_stub, 0, sizeof(struct dlm_message));
4296 ms_stub->m_flags = DLM_IFL_STUB_MS;
4297 ms_stub->m_type = DLM_MSG_CANCEL_REPLY;
4298 ms_stub->m_result = stub_cancel_result;
4299 ms_stub->m_header.h_nodeid = lkb->lkb_nodeid;
4300 _receive_cancel_reply(lkb, ms_stub);
4301 dlm_put_lkb(lkb);
4302 break;
4303
4304 default:
4305 log_error(ls, "invalid lkb wait_type %d %d",
4306 lkb->lkb_wait_type, wait_type);
4307 }
4308 schedule();
4309 }
4310 mutex_unlock(&ls->ls_waiters_mutex);
4311 kfree(ms_stub);
4312 }
4313
find_resend_waiter(struct dlm_ls * ls)4314 static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls)
4315 {
4316 struct dlm_lkb *lkb;
4317 int found = 0;
4318
4319 mutex_lock(&ls->ls_waiters_mutex);
4320 list_for_each_entry(lkb, &ls->ls_waiters, lkb_wait_reply) {
4321 if (lkb->lkb_flags & DLM_IFL_RESEND) {
4322 hold_lkb(lkb);
4323 found = 1;
4324 break;
4325 }
4326 }
4327 mutex_unlock(&ls->ls_waiters_mutex);
4328
4329 if (!found)
4330 lkb = NULL;
4331 return lkb;
4332 }
4333
4334 /* Deal with lookups and lkb's marked RESEND from _pre. We may now be the
4335 master or dir-node for r. Processing the lkb may result in it being placed
4336 back on waiters. */
4337
4338 /* We do this after normal locking has been enabled and any saved messages
4339 (in requestqueue) have been processed. We should be confident that at
4340 this point we won't get or process a reply to any of these waiting
4341 operations. But, new ops may be coming in on the rsbs/locks here from
4342 userspace or remotely. */
4343
4344 /* there may have been an overlap unlock/cancel prior to recovery or after
4345 recovery. if before, the lkb may still have a pos wait_count; if after, the
4346 overlap flag would just have been set and nothing new sent. we can be
4347 confident here than any replies to either the initial op or overlap ops
4348 prior to recovery have been received. */
4349
dlm_recover_waiters_post(struct dlm_ls * ls)4350 int dlm_recover_waiters_post(struct dlm_ls *ls)
4351 {
4352 struct dlm_lkb *lkb;
4353 struct dlm_rsb *r;
4354 int error = 0, mstype, err, oc, ou;
4355
4356 while (1) {
4357 if (dlm_locking_stopped(ls)) {
4358 log_debug(ls, "recover_waiters_post aborted");
4359 error = -EINTR;
4360 break;
4361 }
4362
4363 lkb = find_resend_waiter(ls);
4364 if (!lkb)
4365 break;
4366
4367 r = lkb->lkb_resource;
4368 hold_rsb(r);
4369 lock_rsb(r);
4370
4371 mstype = lkb->lkb_wait_type;
4372 oc = is_overlap_cancel(lkb);
4373 ou = is_overlap_unlock(lkb);
4374 err = 0;
4375
4376 log_debug(ls, "recover_waiter %x nodeid %d msg %d r_nodeid %d",
4377 lkb->lkb_id, lkb->lkb_nodeid, mstype, r->res_nodeid);
4378
4379 /* At this point we assume that we won't get a reply to any
4380 previous op or overlap op on this lock. First, do a big
4381 remove_from_waiters() for all previous ops. */
4382
4383 lkb->lkb_flags &= ~DLM_IFL_RESEND;
4384 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_UNLOCK;
4385 lkb->lkb_flags &= ~DLM_IFL_OVERLAP_CANCEL;
4386 lkb->lkb_wait_type = 0;
4387 lkb->lkb_wait_count = 0;
4388 mutex_lock(&ls->ls_waiters_mutex);
4389 list_del_init(&lkb->lkb_wait_reply);
4390 mutex_unlock(&ls->ls_waiters_mutex);
4391 unhold_lkb(lkb); /* for waiters list */
4392
4393 if (oc || ou) {
4394 /* do an unlock or cancel instead of resending */
4395 switch (mstype) {
4396 case DLM_MSG_LOOKUP:
4397 case DLM_MSG_REQUEST:
4398 queue_cast(r, lkb, ou ? -DLM_EUNLOCK :
4399 -DLM_ECANCEL);
4400 unhold_lkb(lkb); /* undoes create_lkb() */
4401 break;
4402 case DLM_MSG_CONVERT:
4403 if (oc) {
4404 queue_cast(r, lkb, -DLM_ECANCEL);
4405 } else {
4406 lkb->lkb_exflags |= DLM_LKF_FORCEUNLOCK;
4407 _unlock_lock(r, lkb);
4408 }
4409 break;
4410 default:
4411 err = 1;
4412 }
4413 } else {
4414 switch (mstype) {
4415 case DLM_MSG_LOOKUP:
4416 case DLM_MSG_REQUEST:
4417 _request_lock(r, lkb);
4418 if (is_master(r))
4419 confirm_master(r, 0);
4420 break;
4421 case DLM_MSG_CONVERT:
4422 _convert_lock(r, lkb);
4423 break;
4424 default:
4425 err = 1;
4426 }
4427 }
4428
4429 if (err)
4430 log_error(ls, "recover_waiters_post %x %d %x %d %d",
4431 lkb->lkb_id, mstype, lkb->lkb_flags, oc, ou);
4432 unlock_rsb(r);
4433 put_rsb(r);
4434 dlm_put_lkb(lkb);
4435 }
4436
4437 return error;
4438 }
4439
purge_queue(struct dlm_rsb * r,struct list_head * queue,int (* test)(struct dlm_ls * ls,struct dlm_lkb * lkb))4440 static void purge_queue(struct dlm_rsb *r, struct list_head *queue,
4441 int (*test)(struct dlm_ls *ls, struct dlm_lkb *lkb))
4442 {
4443 struct dlm_ls *ls = r->res_ls;
4444 struct dlm_lkb *lkb, *safe;
4445
4446 list_for_each_entry_safe(lkb, safe, queue, lkb_statequeue) {
4447 if (test(ls, lkb)) {
4448 rsb_set_flag(r, RSB_LOCKS_PURGED);
4449 del_lkb(r, lkb);
4450 /* this put should free the lkb */
4451 if (!dlm_put_lkb(lkb))
4452 log_error(ls, "purged lkb not released");
4453 }
4454 }
4455 }
4456
purge_dead_test(struct dlm_ls * ls,struct dlm_lkb * lkb)4457 static int purge_dead_test(struct dlm_ls *ls, struct dlm_lkb *lkb)
4458 {
4459 return (is_master_copy(lkb) && dlm_is_removed(ls, lkb->lkb_nodeid));
4460 }
4461
purge_mstcpy_test(struct dlm_ls * ls,struct dlm_lkb * lkb)4462 static int purge_mstcpy_test(struct dlm_ls *ls, struct dlm_lkb *lkb)
4463 {
4464 return is_master_copy(lkb);
4465 }
4466
purge_dead_locks(struct dlm_rsb * r)4467 static void purge_dead_locks(struct dlm_rsb *r)
4468 {
4469 purge_queue(r, &r->res_grantqueue, &purge_dead_test);
4470 purge_queue(r, &r->res_convertqueue, &purge_dead_test);
4471 purge_queue(r, &r->res_waitqueue, &purge_dead_test);
4472 }
4473
dlm_purge_mstcpy_locks(struct dlm_rsb * r)4474 void dlm_purge_mstcpy_locks(struct dlm_rsb *r)
4475 {
4476 purge_queue(r, &r->res_grantqueue, &purge_mstcpy_test);
4477 purge_queue(r, &r->res_convertqueue, &purge_mstcpy_test);
4478 purge_queue(r, &r->res_waitqueue, &purge_mstcpy_test);
4479 }
4480
4481 /* Get rid of locks held by nodes that are gone. */
4482
dlm_purge_locks(struct dlm_ls * ls)4483 int dlm_purge_locks(struct dlm_ls *ls)
4484 {
4485 struct dlm_rsb *r;
4486
4487 log_debug(ls, "dlm_purge_locks");
4488
4489 down_write(&ls->ls_root_sem);
4490 list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
4491 hold_rsb(r);
4492 lock_rsb(r);
4493 if (is_master(r))
4494 purge_dead_locks(r);
4495 unlock_rsb(r);
4496 unhold_rsb(r);
4497
4498 schedule();
4499 }
4500 up_write(&ls->ls_root_sem);
4501
4502 return 0;
4503 }
4504
find_purged_rsb(struct dlm_ls * ls,int bucket)4505 static struct dlm_rsb *find_purged_rsb(struct dlm_ls *ls, int bucket)
4506 {
4507 struct rb_node *n;
4508 struct dlm_rsb *r, *r_ret = NULL;
4509
4510 spin_lock(&ls->ls_rsbtbl[bucket].lock);
4511 for (n = rb_first(&ls->ls_rsbtbl[bucket].keep); n; n = rb_next(n)) {
4512 r = rb_entry(n, struct dlm_rsb, res_hashnode);
4513 if (!rsb_flag(r, RSB_LOCKS_PURGED))
4514 continue;
4515 hold_rsb(r);
4516 rsb_clear_flag(r, RSB_LOCKS_PURGED);
4517 r_ret = r;
4518 break;
4519 }
4520 spin_unlock(&ls->ls_rsbtbl[bucket].lock);
4521 return r_ret;
4522 }
4523
dlm_grant_after_purge(struct dlm_ls * ls)4524 void dlm_grant_after_purge(struct dlm_ls *ls)
4525 {
4526 struct dlm_rsb *r;
4527 int bucket = 0;
4528
4529 while (1) {
4530 r = find_purged_rsb(ls, bucket);
4531 if (!r) {
4532 if (bucket == ls->ls_rsbtbl_size - 1)
4533 break;
4534 bucket++;
4535 continue;
4536 }
4537 lock_rsb(r);
4538 if (is_master(r)) {
4539 grant_pending_locks(r);
4540 confirm_master(r, 0);
4541 }
4542 unlock_rsb(r);
4543 put_rsb(r);
4544 schedule();
4545 }
4546 }
4547
search_remid_list(struct list_head * head,int nodeid,uint32_t remid)4548 static struct dlm_lkb *search_remid_list(struct list_head *head, int nodeid,
4549 uint32_t remid)
4550 {
4551 struct dlm_lkb *lkb;
4552
4553 list_for_each_entry(lkb, head, lkb_statequeue) {
4554 if (lkb->lkb_nodeid == nodeid && lkb->lkb_remid == remid)
4555 return lkb;
4556 }
4557 return NULL;
4558 }
4559
search_remid(struct dlm_rsb * r,int nodeid,uint32_t remid)4560 static struct dlm_lkb *search_remid(struct dlm_rsb *r, int nodeid,
4561 uint32_t remid)
4562 {
4563 struct dlm_lkb *lkb;
4564
4565 lkb = search_remid_list(&r->res_grantqueue, nodeid, remid);
4566 if (lkb)
4567 return lkb;
4568 lkb = search_remid_list(&r->res_convertqueue, nodeid, remid);
4569 if (lkb)
4570 return lkb;
4571 lkb = search_remid_list(&r->res_waitqueue, nodeid, remid);
4572 if (lkb)
4573 return lkb;
4574 return NULL;
4575 }
4576
4577 /* needs at least dlm_rcom + rcom_lock */
receive_rcom_lock_args(struct dlm_ls * ls,struct dlm_lkb * lkb,struct dlm_rsb * r,struct dlm_rcom * rc)4578 static int receive_rcom_lock_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
4579 struct dlm_rsb *r, struct dlm_rcom *rc)
4580 {
4581 struct rcom_lock *rl = (struct rcom_lock *) rc->rc_buf;
4582
4583 lkb->lkb_nodeid = rc->rc_header.h_nodeid;
4584 lkb->lkb_ownpid = le32_to_cpu(rl->rl_ownpid);
4585 lkb->lkb_remid = le32_to_cpu(rl->rl_lkid);
4586 lkb->lkb_exflags = le32_to_cpu(rl->rl_exflags);
4587 lkb->lkb_flags = le32_to_cpu(rl->rl_flags) & 0x0000FFFF;
4588 lkb->lkb_flags |= DLM_IFL_MSTCPY;
4589 lkb->lkb_lvbseq = le32_to_cpu(rl->rl_lvbseq);
4590 lkb->lkb_rqmode = rl->rl_rqmode;
4591 lkb->lkb_grmode = rl->rl_grmode;
4592 /* don't set lkb_status because add_lkb wants to itself */
4593
4594 lkb->lkb_bastfn = (rl->rl_asts & DLM_CB_BAST) ? &fake_bastfn : NULL;
4595 lkb->lkb_astfn = (rl->rl_asts & DLM_CB_CAST) ? &fake_astfn : NULL;
4596
4597 if (lkb->lkb_exflags & DLM_LKF_VALBLK) {
4598 int lvblen = rc->rc_header.h_length - sizeof(struct dlm_rcom) -
4599 sizeof(struct rcom_lock);
4600 if (lvblen > ls->ls_lvblen)
4601 return -EINVAL;
4602 lkb->lkb_lvbptr = dlm_allocate_lvb(ls);
4603 if (!lkb->lkb_lvbptr)
4604 return -ENOMEM;
4605 memcpy(lkb->lkb_lvbptr, rl->rl_lvb, lvblen);
4606 }
4607
4608 /* Conversions between PR and CW (middle modes) need special handling.
4609 The real granted mode of these converting locks cannot be determined
4610 until all locks have been rebuilt on the rsb (recover_conversion) */
4611
4612 if (rl->rl_wait_type == cpu_to_le16(DLM_MSG_CONVERT) &&
4613 middle_conversion(lkb)) {
4614 rl->rl_status = DLM_LKSTS_CONVERT;
4615 lkb->lkb_grmode = DLM_LOCK_IV;
4616 rsb_set_flag(r, RSB_RECOVER_CONVERT);
4617 }
4618
4619 return 0;
4620 }
4621
4622 /* This lkb may have been recovered in a previous aborted recovery so we need
4623 to check if the rsb already has an lkb with the given remote nodeid/lkid.
4624 If so we just send back a standard reply. If not, we create a new lkb with
4625 the given values and send back our lkid. We send back our lkid by sending
4626 back the rcom_lock struct we got but with the remid field filled in. */
4627
4628 /* needs at least dlm_rcom + rcom_lock */
dlm_recover_master_copy(struct dlm_ls * ls,struct dlm_rcom * rc)4629 int dlm_recover_master_copy(struct dlm_ls *ls, struct dlm_rcom *rc)
4630 {
4631 struct rcom_lock *rl = (struct rcom_lock *) rc->rc_buf;
4632 struct dlm_rsb *r;
4633 struct dlm_lkb *lkb;
4634 int error;
4635
4636 if (rl->rl_parent_lkid) {
4637 error = -EOPNOTSUPP;
4638 goto out;
4639 }
4640
4641 error = find_rsb(ls, rl->rl_name, le16_to_cpu(rl->rl_namelen),
4642 R_MASTER, &r);
4643 if (error)
4644 goto out;
4645
4646 lock_rsb(r);
4647
4648 lkb = search_remid(r, rc->rc_header.h_nodeid, le32_to_cpu(rl->rl_lkid));
4649 if (lkb) {
4650 error = -EEXIST;
4651 goto out_remid;
4652 }
4653
4654 error = create_lkb(ls, &lkb);
4655 if (error)
4656 goto out_unlock;
4657
4658 error = receive_rcom_lock_args(ls, lkb, r, rc);
4659 if (error) {
4660 __put_lkb(ls, lkb);
4661 goto out_unlock;
4662 }
4663
4664 attach_lkb(r, lkb);
4665 add_lkb(r, lkb, rl->rl_status);
4666 error = 0;
4667
4668 out_remid:
4669 /* this is the new value returned to the lock holder for
4670 saving in its process-copy lkb */
4671 rl->rl_remid = cpu_to_le32(lkb->lkb_id);
4672
4673 out_unlock:
4674 unlock_rsb(r);
4675 put_rsb(r);
4676 out:
4677 if (error)
4678 log_debug(ls, "recover_master_copy %d %x", error,
4679 le32_to_cpu(rl->rl_lkid));
4680 rl->rl_result = cpu_to_le32(error);
4681 return error;
4682 }
4683
4684 /* needs at least dlm_rcom + rcom_lock */
dlm_recover_process_copy(struct dlm_ls * ls,struct dlm_rcom * rc)4685 int dlm_recover_process_copy(struct dlm_ls *ls, struct dlm_rcom *rc)
4686 {
4687 struct rcom_lock *rl = (struct rcom_lock *) rc->rc_buf;
4688 struct dlm_rsb *r;
4689 struct dlm_lkb *lkb;
4690 int error;
4691
4692 error = find_lkb(ls, le32_to_cpu(rl->rl_lkid), &lkb);
4693 if (error) {
4694 log_error(ls, "recover_process_copy no lkid %x",
4695 le32_to_cpu(rl->rl_lkid));
4696 return error;
4697 }
4698
4699 DLM_ASSERT(is_process_copy(lkb), dlm_print_lkb(lkb););
4700
4701 error = le32_to_cpu(rl->rl_result);
4702
4703 r = lkb->lkb_resource;
4704 hold_rsb(r);
4705 lock_rsb(r);
4706
4707 switch (error) {
4708 case -EBADR:
4709 /* There's a chance the new master received our lock before
4710 dlm_recover_master_reply(), this wouldn't happen if we did
4711 a barrier between recover_masters and recover_locks. */
4712 log_debug(ls, "master copy not ready %x r %lx %s", lkb->lkb_id,
4713 (unsigned long)r, r->res_name);
4714 dlm_send_rcom_lock(r, lkb);
4715 goto out;
4716 case -EEXIST:
4717 log_debug(ls, "master copy exists %x", lkb->lkb_id);
4718 /* fall through */
4719 case 0:
4720 lkb->lkb_remid = le32_to_cpu(rl->rl_remid);
4721 break;
4722 default:
4723 log_error(ls, "dlm_recover_process_copy unknown error %d %x",
4724 error, lkb->lkb_id);
4725 }
4726
4727 /* an ack for dlm_recover_locks() which waits for replies from
4728 all the locks it sends to new masters */
4729 dlm_recovered_lock(r);
4730 out:
4731 unlock_rsb(r);
4732 put_rsb(r);
4733 dlm_put_lkb(lkb);
4734
4735 return 0;
4736 }
4737
dlm_user_request(struct dlm_ls * ls,struct dlm_user_args * ua,int mode,uint32_t flags,void * name,unsigned int namelen,unsigned long timeout_cs)4738 int dlm_user_request(struct dlm_ls *ls, struct dlm_user_args *ua,
4739 int mode, uint32_t flags, void *name, unsigned int namelen,
4740 unsigned long timeout_cs)
4741 {
4742 struct dlm_lkb *lkb;
4743 struct dlm_args args;
4744 int error;
4745
4746 dlm_lock_recovery(ls);
4747
4748 error = create_lkb(ls, &lkb);
4749 if (error) {
4750 kfree(ua);
4751 goto out;
4752 }
4753
4754 if (flags & DLM_LKF_VALBLK) {
4755 ua->lksb.sb_lvbptr = kzalloc(DLM_USER_LVB_LEN, GFP_NOFS);
4756 if (!ua->lksb.sb_lvbptr) {
4757 kfree(ua);
4758 __put_lkb(ls, lkb);
4759 error = -ENOMEM;
4760 goto out;
4761 }
4762 }
4763
4764 /* After ua is attached to lkb it will be freed by dlm_free_lkb().
4765 When DLM_IFL_USER is set, the dlm knows that this is a userspace
4766 lock and that lkb_astparam is the dlm_user_args structure. */
4767
4768 error = set_lock_args(mode, &ua->lksb, flags, namelen, timeout_cs,
4769 fake_astfn, ua, fake_bastfn, &args);
4770 lkb->lkb_flags |= DLM_IFL_USER;
4771
4772 if (error) {
4773 __put_lkb(ls, lkb);
4774 goto out;
4775 }
4776
4777 error = request_lock(ls, lkb, name, namelen, &args);
4778
4779 switch (error) {
4780 case 0:
4781 break;
4782 case -EINPROGRESS:
4783 error = 0;
4784 break;
4785 case -EAGAIN:
4786 error = 0;
4787 /* fall through */
4788 default:
4789 __put_lkb(ls, lkb);
4790 goto out;
4791 }
4792
4793 /* add this new lkb to the per-process list of locks */
4794 spin_lock(&ua->proc->locks_spin);
4795 hold_lkb(lkb);
4796 list_add_tail(&lkb->lkb_ownqueue, &ua->proc->locks);
4797 spin_unlock(&ua->proc->locks_spin);
4798 out:
4799 dlm_unlock_recovery(ls);
4800 return error;
4801 }
4802
dlm_user_convert(struct dlm_ls * ls,struct dlm_user_args * ua_tmp,int mode,uint32_t flags,uint32_t lkid,char * lvb_in,unsigned long timeout_cs)4803 int dlm_user_convert(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
4804 int mode, uint32_t flags, uint32_t lkid, char *lvb_in,
4805 unsigned long timeout_cs)
4806 {
4807 struct dlm_lkb *lkb;
4808 struct dlm_args args;
4809 struct dlm_user_args *ua;
4810 int error;
4811
4812 dlm_lock_recovery(ls);
4813
4814 error = find_lkb(ls, lkid, &lkb);
4815 if (error)
4816 goto out;
4817
4818 /* user can change the params on its lock when it converts it, or
4819 add an lvb that didn't exist before */
4820
4821 ua = lkb->lkb_ua;
4822
4823 if (flags & DLM_LKF_VALBLK && !ua->lksb.sb_lvbptr) {
4824 ua->lksb.sb_lvbptr = kzalloc(DLM_USER_LVB_LEN, GFP_NOFS);
4825 if (!ua->lksb.sb_lvbptr) {
4826 error = -ENOMEM;
4827 goto out_put;
4828 }
4829 }
4830 if (lvb_in && ua->lksb.sb_lvbptr)
4831 memcpy(ua->lksb.sb_lvbptr, lvb_in, DLM_USER_LVB_LEN);
4832
4833 ua->xid = ua_tmp->xid;
4834 ua->castparam = ua_tmp->castparam;
4835 ua->castaddr = ua_tmp->castaddr;
4836 ua->bastparam = ua_tmp->bastparam;
4837 ua->bastaddr = ua_tmp->bastaddr;
4838 ua->user_lksb = ua_tmp->user_lksb;
4839
4840 error = set_lock_args(mode, &ua->lksb, flags, 0, timeout_cs,
4841 fake_astfn, ua, fake_bastfn, &args);
4842 if (error)
4843 goto out_put;
4844
4845 error = convert_lock(ls, lkb, &args);
4846
4847 if (error == -EINPROGRESS || error == -EAGAIN || error == -EDEADLK)
4848 error = 0;
4849 out_put:
4850 dlm_put_lkb(lkb);
4851 out:
4852 dlm_unlock_recovery(ls);
4853 kfree(ua_tmp);
4854 return error;
4855 }
4856
dlm_user_unlock(struct dlm_ls * ls,struct dlm_user_args * ua_tmp,uint32_t flags,uint32_t lkid,char * lvb_in)4857 int dlm_user_unlock(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
4858 uint32_t flags, uint32_t lkid, char *lvb_in)
4859 {
4860 struct dlm_lkb *lkb;
4861 struct dlm_args args;
4862 struct dlm_user_args *ua;
4863 int error;
4864
4865 dlm_lock_recovery(ls);
4866
4867 error = find_lkb(ls, lkid, &lkb);
4868 if (error)
4869 goto out;
4870
4871 ua = lkb->lkb_ua;
4872
4873 if (lvb_in && ua->lksb.sb_lvbptr)
4874 memcpy(ua->lksb.sb_lvbptr, lvb_in, DLM_USER_LVB_LEN);
4875 if (ua_tmp->castparam)
4876 ua->castparam = ua_tmp->castparam;
4877 ua->user_lksb = ua_tmp->user_lksb;
4878
4879 error = set_unlock_args(flags, ua, &args);
4880 if (error)
4881 goto out_put;
4882
4883 error = unlock_lock(ls, lkb, &args);
4884
4885 if (error == -DLM_EUNLOCK)
4886 error = 0;
4887 /* from validate_unlock_args() */
4888 if (error == -EBUSY && (flags & DLM_LKF_FORCEUNLOCK))
4889 error = 0;
4890 if (error)
4891 goto out_put;
4892
4893 spin_lock(&ua->proc->locks_spin);
4894 /* dlm_user_add_cb() may have already taken lkb off the proc list */
4895 if (!list_empty(&lkb->lkb_ownqueue))
4896 list_move(&lkb->lkb_ownqueue, &ua->proc->unlocking);
4897 spin_unlock(&ua->proc->locks_spin);
4898 out_put:
4899 dlm_put_lkb(lkb);
4900 out:
4901 dlm_unlock_recovery(ls);
4902 kfree(ua_tmp);
4903 return error;
4904 }
4905
dlm_user_cancel(struct dlm_ls * ls,struct dlm_user_args * ua_tmp,uint32_t flags,uint32_t lkid)4906 int dlm_user_cancel(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
4907 uint32_t flags, uint32_t lkid)
4908 {
4909 struct dlm_lkb *lkb;
4910 struct dlm_args args;
4911 struct dlm_user_args *ua;
4912 int error;
4913
4914 dlm_lock_recovery(ls);
4915
4916 error = find_lkb(ls, lkid, &lkb);
4917 if (error)
4918 goto out;
4919
4920 ua = lkb->lkb_ua;
4921 if (ua_tmp->castparam)
4922 ua->castparam = ua_tmp->castparam;
4923 ua->user_lksb = ua_tmp->user_lksb;
4924
4925 error = set_unlock_args(flags, ua, &args);
4926 if (error)
4927 goto out_put;
4928
4929 error = cancel_lock(ls, lkb, &args);
4930
4931 if (error == -DLM_ECANCEL)
4932 error = 0;
4933 /* from validate_unlock_args() */
4934 if (error == -EBUSY)
4935 error = 0;
4936 out_put:
4937 dlm_put_lkb(lkb);
4938 out:
4939 dlm_unlock_recovery(ls);
4940 kfree(ua_tmp);
4941 return error;
4942 }
4943
dlm_user_deadlock(struct dlm_ls * ls,uint32_t flags,uint32_t lkid)4944 int dlm_user_deadlock(struct dlm_ls *ls, uint32_t flags, uint32_t lkid)
4945 {
4946 struct dlm_lkb *lkb;
4947 struct dlm_args args;
4948 struct dlm_user_args *ua;
4949 struct dlm_rsb *r;
4950 int error;
4951
4952 dlm_lock_recovery(ls);
4953
4954 error = find_lkb(ls, lkid, &lkb);
4955 if (error)
4956 goto out;
4957
4958 ua = lkb->lkb_ua;
4959
4960 error = set_unlock_args(flags, ua, &args);
4961 if (error)
4962 goto out_put;
4963
4964 /* same as cancel_lock(), but set DEADLOCK_CANCEL after lock_rsb */
4965
4966 r = lkb->lkb_resource;
4967 hold_rsb(r);
4968 lock_rsb(r);
4969
4970 error = validate_unlock_args(lkb, &args);
4971 if (error)
4972 goto out_r;
4973 lkb->lkb_flags |= DLM_IFL_DEADLOCK_CANCEL;
4974
4975 error = _cancel_lock(r, lkb);
4976 out_r:
4977 unlock_rsb(r);
4978 put_rsb(r);
4979
4980 if (error == -DLM_ECANCEL)
4981 error = 0;
4982 /* from validate_unlock_args() */
4983 if (error == -EBUSY)
4984 error = 0;
4985 out_put:
4986 dlm_put_lkb(lkb);
4987 out:
4988 dlm_unlock_recovery(ls);
4989 return error;
4990 }
4991
4992 /* lkb's that are removed from the waiters list by revert are just left on the
4993 orphans list with the granted orphan locks, to be freed by purge */
4994
orphan_proc_lock(struct dlm_ls * ls,struct dlm_lkb * lkb)4995 static int orphan_proc_lock(struct dlm_ls *ls, struct dlm_lkb *lkb)
4996 {
4997 struct dlm_args args;
4998 int error;
4999
5000 hold_lkb(lkb);
5001 mutex_lock(&ls->ls_orphans_mutex);
5002 list_add_tail(&lkb->lkb_ownqueue, &ls->ls_orphans);
5003 mutex_unlock(&ls->ls_orphans_mutex);
5004
5005 set_unlock_args(0, lkb->lkb_ua, &args);
5006
5007 error = cancel_lock(ls, lkb, &args);
5008 if (error == -DLM_ECANCEL)
5009 error = 0;
5010 return error;
5011 }
5012
5013 /* The force flag allows the unlock to go ahead even if the lkb isn't granted.
5014 Regardless of what rsb queue the lock is on, it's removed and freed. */
5015
unlock_proc_lock(struct dlm_ls * ls,struct dlm_lkb * lkb)5016 static int unlock_proc_lock(struct dlm_ls *ls, struct dlm_lkb *lkb)
5017 {
5018 struct dlm_args args;
5019 int error;
5020
5021 set_unlock_args(DLM_LKF_FORCEUNLOCK, lkb->lkb_ua, &args);
5022
5023 error = unlock_lock(ls, lkb, &args);
5024 if (error == -DLM_EUNLOCK)
5025 error = 0;
5026 return error;
5027 }
5028
5029 /* We have to release clear_proc_locks mutex before calling unlock_proc_lock()
5030 (which does lock_rsb) due to deadlock with receiving a message that does
5031 lock_rsb followed by dlm_user_add_cb() */
5032
del_proc_lock(struct dlm_ls * ls,struct dlm_user_proc * proc)5033 static struct dlm_lkb *del_proc_lock(struct dlm_ls *ls,
5034 struct dlm_user_proc *proc)
5035 {
5036 struct dlm_lkb *lkb = NULL;
5037
5038 mutex_lock(&ls->ls_clear_proc_locks);
5039 if (list_empty(&proc->locks))
5040 goto out;
5041
5042 lkb = list_entry(proc->locks.next, struct dlm_lkb, lkb_ownqueue);
5043 list_del_init(&lkb->lkb_ownqueue);
5044
5045 if (lkb->lkb_exflags & DLM_LKF_PERSISTENT)
5046 lkb->lkb_flags |= DLM_IFL_ORPHAN;
5047 else
5048 lkb->lkb_flags |= DLM_IFL_DEAD;
5049 out:
5050 mutex_unlock(&ls->ls_clear_proc_locks);
5051 return lkb;
5052 }
5053
5054 /* The ls_clear_proc_locks mutex protects against dlm_user_add_cb() which
5055 1) references lkb->ua which we free here and 2) adds lkbs to proc->asts,
5056 which we clear here. */
5057
5058 /* proc CLOSING flag is set so no more device_reads should look at proc->asts
5059 list, and no more device_writes should add lkb's to proc->locks list; so we
5060 shouldn't need to take asts_spin or locks_spin here. this assumes that
5061 device reads/writes/closes are serialized -- FIXME: we may need to serialize
5062 them ourself. */
5063
dlm_clear_proc_locks(struct dlm_ls * ls,struct dlm_user_proc * proc)5064 void dlm_clear_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
5065 {
5066 struct dlm_lkb *lkb, *safe;
5067
5068 dlm_lock_recovery(ls);
5069
5070 while (1) {
5071 lkb = del_proc_lock(ls, proc);
5072 if (!lkb)
5073 break;
5074 del_timeout(lkb);
5075 if (lkb->lkb_exflags & DLM_LKF_PERSISTENT)
5076 orphan_proc_lock(ls, lkb);
5077 else
5078 unlock_proc_lock(ls, lkb);
5079
5080 /* this removes the reference for the proc->locks list
5081 added by dlm_user_request, it may result in the lkb
5082 being freed */
5083
5084 dlm_put_lkb(lkb);
5085 }
5086
5087 mutex_lock(&ls->ls_clear_proc_locks);
5088
5089 /* in-progress unlocks */
5090 list_for_each_entry_safe(lkb, safe, &proc->unlocking, lkb_ownqueue) {
5091 list_del_init(&lkb->lkb_ownqueue);
5092 lkb->lkb_flags |= DLM_IFL_DEAD;
5093 dlm_put_lkb(lkb);
5094 }
5095
5096 list_for_each_entry_safe(lkb, safe, &proc->asts, lkb_cb_list) {
5097 memset(&lkb->lkb_callbacks, 0,
5098 sizeof(struct dlm_callback) * DLM_CALLBACKS_SIZE);
5099 list_del_init(&lkb->lkb_cb_list);
5100 dlm_put_lkb(lkb);
5101 }
5102
5103 mutex_unlock(&ls->ls_clear_proc_locks);
5104 dlm_unlock_recovery(ls);
5105 }
5106
purge_proc_locks(struct dlm_ls * ls,struct dlm_user_proc * proc)5107 static void purge_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
5108 {
5109 struct dlm_lkb *lkb, *safe;
5110
5111 while (1) {
5112 lkb = NULL;
5113 spin_lock(&proc->locks_spin);
5114 if (!list_empty(&proc->locks)) {
5115 lkb = list_entry(proc->locks.next, struct dlm_lkb,
5116 lkb_ownqueue);
5117 list_del_init(&lkb->lkb_ownqueue);
5118 }
5119 spin_unlock(&proc->locks_spin);
5120
5121 if (!lkb)
5122 break;
5123
5124 lkb->lkb_flags |= DLM_IFL_DEAD;
5125 unlock_proc_lock(ls, lkb);
5126 dlm_put_lkb(lkb); /* ref from proc->locks list */
5127 }
5128
5129 spin_lock(&proc->locks_spin);
5130 list_for_each_entry_safe(lkb, safe, &proc->unlocking, lkb_ownqueue) {
5131 list_del_init(&lkb->lkb_ownqueue);
5132 lkb->lkb_flags |= DLM_IFL_DEAD;
5133 dlm_put_lkb(lkb);
5134 }
5135 spin_unlock(&proc->locks_spin);
5136
5137 spin_lock(&proc->asts_spin);
5138 list_for_each_entry_safe(lkb, safe, &proc->asts, lkb_cb_list) {
5139 memset(&lkb->lkb_callbacks, 0,
5140 sizeof(struct dlm_callback) * DLM_CALLBACKS_SIZE);
5141 list_del_init(&lkb->lkb_cb_list);
5142 dlm_put_lkb(lkb);
5143 }
5144 spin_unlock(&proc->asts_spin);
5145 }
5146
5147 /* pid of 0 means purge all orphans */
5148
do_purge(struct dlm_ls * ls,int nodeid,int pid)5149 static void do_purge(struct dlm_ls *ls, int nodeid, int pid)
5150 {
5151 struct dlm_lkb *lkb, *safe;
5152
5153 mutex_lock(&ls->ls_orphans_mutex);
5154 list_for_each_entry_safe(lkb, safe, &ls->ls_orphans, lkb_ownqueue) {
5155 if (pid && lkb->lkb_ownpid != pid)
5156 continue;
5157 unlock_proc_lock(ls, lkb);
5158 list_del_init(&lkb->lkb_ownqueue);
5159 dlm_put_lkb(lkb);
5160 }
5161 mutex_unlock(&ls->ls_orphans_mutex);
5162 }
5163
send_purge(struct dlm_ls * ls,int nodeid,int pid)5164 static int send_purge(struct dlm_ls *ls, int nodeid, int pid)
5165 {
5166 struct dlm_message *ms;
5167 struct dlm_mhandle *mh;
5168 int error;
5169
5170 error = _create_message(ls, sizeof(struct dlm_message), nodeid,
5171 DLM_MSG_PURGE, &ms, &mh);
5172 if (error)
5173 return error;
5174 ms->m_nodeid = nodeid;
5175 ms->m_pid = pid;
5176
5177 return send_message(mh, ms);
5178 }
5179
dlm_user_purge(struct dlm_ls * ls,struct dlm_user_proc * proc,int nodeid,int pid)5180 int dlm_user_purge(struct dlm_ls *ls, struct dlm_user_proc *proc,
5181 int nodeid, int pid)
5182 {
5183 int error = 0;
5184
5185 if (nodeid != dlm_our_nodeid()) {
5186 error = send_purge(ls, nodeid, pid);
5187 } else {
5188 dlm_lock_recovery(ls);
5189 if (pid == current->pid)
5190 purge_proc_locks(ls, proc);
5191 else
5192 do_purge(ls, nodeid, pid);
5193 dlm_unlock_recovery(ls);
5194 }
5195 return error;
5196 }
5197
5198