1/*
2 * Copyright (C) 2015, SUSE
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2, or (at your option)
7 * any later version.
8 *
9 */
10
11
12#include <linux/module.h>
13#include <linux/kthread.h>
14#include <linux/dlm.h>
15#include <linux/sched.h>
16#include <linux/raid/md_p.h>
17#include "md.h"
18#include "md-bitmap.h"
19#include "md-cluster.h"
20
21#define LVB_SIZE 64
22#define NEW_DEV_TIMEOUT 5000
23
24struct dlm_lock_resource {
25 dlm_lockspace_t *ls;
26 struct dlm_lksb lksb;
27 char *name; /* lock name. */
28 uint32_t flags; /* flags to pass to dlm_lock() */
29 wait_queue_head_t sync_locking; /* wait queue for synchronized locking */
30 bool sync_locking_done;
31 void (*bast)(void *arg, int mode); /* blocking AST function pointer*/
32 struct mddev *mddev; /* pointing back to mddev. */
33 int mode;
34};
35
36struct resync_info {
37 __le64 lo;
38 __le64 hi;
39};
40
41/* md_cluster_info flags */
42#define MD_CLUSTER_WAITING_FOR_NEWDISK 1
43#define MD_CLUSTER_SUSPEND_READ_BALANCING 2
44#define MD_CLUSTER_BEGIN_JOIN_CLUSTER 3
45
46/* Lock the send communication. This is done through
47 * bit manipulation as opposed to a mutex in order to
48 * accomodate lock and hold. See next comment.
49 */
50#define MD_CLUSTER_SEND_LOCK 4
51/* If cluster operations (such as adding a disk) must lock the
52 * communication channel, so as to perform extra operations
53 * (update metadata) and no other operation is allowed on the
54 * MD. Token needs to be locked and held until the operation
55 * completes witha md_update_sb(), which would eventually release
56 * the lock.
57 */
58#define MD_CLUSTER_SEND_LOCKED_ALREADY 5
59/* We should receive message after node joined cluster and
60 * set up all the related infos such as bitmap and personality */
61#define MD_CLUSTER_ALREADY_IN_CLUSTER 6
62#define MD_CLUSTER_PENDING_RECV_EVENT 7
63#define MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD 8
64
65struct md_cluster_info {
66 struct mddev *mddev; /* the md device which md_cluster_info belongs to */
67 /* dlm lock space and resources for clustered raid. */
68 dlm_lockspace_t *lockspace;
69 int slot_number;
70 struct completion completion;
71 struct mutex recv_mutex;
72 struct dlm_lock_resource *bitmap_lockres;
73 struct dlm_lock_resource **other_bitmap_lockres;
74 struct dlm_lock_resource *resync_lockres;
75 struct list_head suspend_list;
76
77 spinlock_t suspend_lock;
78 /* record the region which write should be suspended */
79 sector_t suspend_lo;
80 sector_t suspend_hi;
81 int suspend_from; /* the slot which broadcast suspend_lo/hi */
82
83 struct md_thread *recovery_thread;
84 unsigned long recovery_map;
85 /* communication loc resources */
86 struct dlm_lock_resource *ack_lockres;
87 struct dlm_lock_resource *message_lockres;
88 struct dlm_lock_resource *token_lockres;
89 struct dlm_lock_resource *no_new_dev_lockres;
90 struct md_thread *recv_thread;
91 struct completion newdisk_completion;
92 wait_queue_head_t wait;
93 unsigned long state;
94 /* record the region in RESYNCING message */
95 sector_t sync_low;
96 sector_t sync_hi;
97};
98
99enum msg_type {
100 METADATA_UPDATED = 0,
101 RESYNCING,
102 NEWDISK,
103 REMOVE,
104 RE_ADD,
105 BITMAP_NEEDS_SYNC,
106 CHANGE_CAPACITY,
107 BITMAP_RESIZE,
108};
109
110struct cluster_msg {
111 __le32 type;
112 __le32 slot;
113 /* TODO: Unionize this for smaller footprint */
114 __le64 low;
115 __le64 high;
116 char uuid[16];
117 __le32 raid_slot;
118};
119
120static void sync_ast(void *arg)
121{
122 struct dlm_lock_resource *res;
123
124 res = arg;
125 res->sync_locking_done = true;
126 wake_up(&res->sync_locking);
127}
128
129static int dlm_lock_sync(struct dlm_lock_resource *res, int mode)
130{
131 int ret = 0;
132
133 ret = dlm_lock(res->ls, mode, &res->lksb,
134 res->flags, res->name, strlen(res->name),
135 0, sync_ast, res, res->bast);
136 if (ret)
137 return ret;
138 wait_event(res->sync_locking, res->sync_locking_done);
139 res->sync_locking_done = false;
140 if (res->lksb.sb_status == 0)
141 res->mode = mode;
142 return res->lksb.sb_status;
143}
144
145static int dlm_unlock_sync(struct dlm_lock_resource *res)
146{
147 return dlm_lock_sync(res, DLM_LOCK_NL);
148}
149
150/*
151 * An variation of dlm_lock_sync, which make lock request could
152 * be interrupted
153 */
154static int dlm_lock_sync_interruptible(struct dlm_lock_resource *res, int mode,
155 struct mddev *mddev)
156{
157 int ret = 0;
158
159 ret = dlm_lock(res->ls, mode, &res->lksb,
160 res->flags, res->name, strlen(res->name),
161 0, sync_ast, res, res->bast);
162 if (ret)
163 return ret;
164
165 wait_event(res->sync_locking, res->sync_locking_done
166 || kthread_should_stop()
167 || test_bit(MD_CLOSING, &mddev->flags));
168 if (!res->sync_locking_done) {
169 /*
170 * the convert queue contains the lock request when request is
171 * interrupted, and sync_ast could still be run, so need to
172 * cancel the request and reset completion
173 */
174 ret = dlm_unlock(res->ls, res->lksb.sb_lkid, DLM_LKF_CANCEL,
175 &res->lksb, res);
176 res->sync_locking_done = false;
177 if (unlikely(ret != 0))
178 pr_info("failed to cancel previous lock request "
179 "%s return %d\n", res->name, ret);
180 return -EPERM;
181 } else
182 res->sync_locking_done = false;
183 if (res->lksb.sb_status == 0)
184 res->mode = mode;
185 return res->lksb.sb_status;
186}
187
188static struct dlm_lock_resource *lockres_init(struct mddev *mddev,
189 char *name, void (*bastfn)(void *arg, int mode), int with_lvb)
190{
191 struct dlm_lock_resource *res = NULL;
192 int ret, namelen;
193 struct md_cluster_info *cinfo = mddev->cluster_info;
194
195 res = kzalloc(sizeof(struct dlm_lock_resource), GFP_KERNEL);
196 if (!res)
197 return NULL;
198 init_waitqueue_head(&res->sync_locking);
199 res->sync_locking_done = false;
200 res->ls = cinfo->lockspace;
201 res->mddev = mddev;
202 res->mode = DLM_LOCK_IV;
203 namelen = strlen(name);
204 res->name = kzalloc(namelen + 1, GFP_KERNEL);
205 if (!res->name) {
206 pr_err("md-cluster: Unable to allocate resource name for resource %s\n", name);
207 goto out_err;
208 }
209 strlcpy(res->name, name, namelen + 1);
210 if (with_lvb) {
211 res->lksb.sb_lvbptr = kzalloc(LVB_SIZE, GFP_KERNEL);
212 if (!res->lksb.sb_lvbptr) {
213 pr_err("md-cluster: Unable to allocate LVB for resource %s\n", name);
214 goto out_err;
215 }
216 res->flags = DLM_LKF_VALBLK;
217 }
218
219 if (bastfn)
220 res->bast = bastfn;
221
222 res->flags |= DLM_LKF_EXPEDITE;
223
224 ret = dlm_lock_sync(res, DLM_LOCK_NL);
225 if (ret) {
226 pr_err("md-cluster: Unable to lock NL on new lock resource %s\n", name);
227 goto out_err;
228 }
229 res->flags &= ~DLM_LKF_EXPEDITE;
230 res->flags |= DLM_LKF_CONVERT;
231
232 return res;
233out_err:
234 kfree(res->lksb.sb_lvbptr);
235 kfree(res->name);
236 kfree(res);
237 return NULL;
238}
239
240static void lockres_free(struct dlm_lock_resource *res)
241{
242 int ret = 0;
243
244 if (!res)
245 return;
246
247 /*
248 * use FORCEUNLOCK flag, so we can unlock even the lock is on the
249 * waiting or convert queue
250 */
251 ret = dlm_unlock(res->ls, res->lksb.sb_lkid, DLM_LKF_FORCEUNLOCK,
252 &res->lksb, res);
253 if (unlikely(ret != 0))
254 pr_err("failed to unlock %s return %d\n", res->name, ret);
255 else
256 wait_event(res->sync_locking, res->sync_locking_done);
257
258 kfree(res->name);
259 kfree(res->lksb.sb_lvbptr);
260 kfree(res);
261}
262
263static void add_resync_info(struct dlm_lock_resource *lockres,
264 sector_t lo, sector_t hi)
265{
266 struct resync_info *ri;
267
268 ri = (struct resync_info *)lockres->lksb.sb_lvbptr;
269 ri->lo = cpu_to_le64(lo);
270 ri->hi = cpu_to_le64(hi);
271}
272
273static int read_resync_info(struct mddev *mddev,
274 struct dlm_lock_resource *lockres)
275{
276 struct resync_info ri;
277 struct md_cluster_info *cinfo = mddev->cluster_info;
278 int ret = 0;
279
280 dlm_lock_sync(lockres, DLM_LOCK_CR);
281 memcpy(&ri, lockres->lksb.sb_lvbptr, sizeof(struct resync_info));
282 if (le64_to_cpu(ri.hi) > 0) {
283 cinfo->suspend_hi = le64_to_cpu(ri.hi);
284 cinfo->suspend_lo = le64_to_cpu(ri.lo);
285 ret = 1;
286 }
287 dlm_unlock_sync(lockres);
288 return ret;
289}
290
291static void recover_bitmaps(struct md_thread *thread)
292{
293 struct mddev *mddev = thread->mddev;
294 struct md_cluster_info *cinfo = mddev->cluster_info;
295 struct dlm_lock_resource *bm_lockres;
296 char str[64];
297 int slot, ret;
298 sector_t lo, hi;
299
300 while (cinfo->recovery_map) {
301 slot = fls64((u64)cinfo->recovery_map) - 1;
302
303 snprintf(str, 64, "bitmap%04d", slot);
304 bm_lockres = lockres_init(mddev, str, NULL, 1);
305 if (!bm_lockres) {
306 pr_err("md-cluster: Cannot initialize bitmaps\n");
307 goto clear_bit;
308 }
309
310 ret = dlm_lock_sync_interruptible(bm_lockres, DLM_LOCK_PW, mddev);
311 if (ret) {
312 pr_err("md-cluster: Could not DLM lock %s: %d\n",
313 str, ret);
314 goto clear_bit;
315 }
316 ret = md_bitmap_copy_from_slot(mddev, slot, &lo, &hi, true);
317 if (ret) {
318 pr_err("md-cluster: Could not copy data from bitmap %d\n", slot);
319 goto clear_bit;
320 }
321
322 /* Clear suspend_area associated with the bitmap */
323 spin_lock_irq(&cinfo->suspend_lock);
324 cinfo->suspend_hi = 0;
325 cinfo->suspend_lo = 0;
326 cinfo->suspend_from = -1;
327 spin_unlock_irq(&cinfo->suspend_lock);
328
329 /* Kick off a reshape if needed */
330 if (test_bit(MD_RESYNCING_REMOTE, &mddev->recovery) &&
331 test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
332 mddev->reshape_position != MaxSector)
333 md_wakeup_thread(mddev->sync_thread);
334
335 if (hi > 0) {
336 if (lo < mddev->recovery_cp)
337 mddev->recovery_cp = lo;
338 /* wake up thread to continue resync in case resync
339 * is not finished */
340 if (mddev->recovery_cp != MaxSector) {
341 /*
342 * clear the REMOTE flag since we will launch
343 * resync thread in current node.
344 */
345 clear_bit(MD_RESYNCING_REMOTE,
346 &mddev->recovery);
347 set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
348 md_wakeup_thread(mddev->thread);
349 }
350 }
351clear_bit:
352 lockres_free(bm_lockres);
353 clear_bit(slot, &cinfo->recovery_map);
354 }
355}
356
357static void recover_prep(void *arg)
358{
359 struct mddev *mddev = arg;
360 struct md_cluster_info *cinfo = mddev->cluster_info;
361 set_bit(MD_CLUSTER_SUSPEND_READ_BALANCING, &cinfo->state);
362}
363
364static void __recover_slot(struct mddev *mddev, int slot)
365{
366 struct md_cluster_info *cinfo = mddev->cluster_info;
367
368 set_bit(slot, &cinfo->recovery_map);
369 if (!cinfo->recovery_thread) {
370 cinfo->recovery_thread = md_register_thread(recover_bitmaps,
371 mddev, "recover");
372 if (!cinfo->recovery_thread) {
373 pr_warn("md-cluster: Could not create recovery thread\n");
374 return;
375 }
376 }
377 md_wakeup_thread(cinfo->recovery_thread);
378}
379
380static void recover_slot(void *arg, struct dlm_slot *slot)
381{
382 struct mddev *mddev = arg;
383 struct md_cluster_info *cinfo = mddev->cluster_info;
384
385 pr_info("md-cluster: %s Node %d/%d down. My slot: %d. Initiating recovery.\n",
386 mddev->bitmap_info.cluster_name,
387 slot->nodeid, slot->slot,
388 cinfo->slot_number);
389 /* deduct one since dlm slot starts from one while the num of
390 * cluster-md begins with 0 */
391 __recover_slot(mddev, slot->slot - 1);
392}
393
394static void recover_done(void *arg, struct dlm_slot *slots,
395 int num_slots, int our_slot,
396 uint32_t generation)
397{
398 struct mddev *mddev = arg;
399 struct md_cluster_info *cinfo = mddev->cluster_info;
400
401 cinfo->slot_number = our_slot;
402 /* completion is only need to be complete when node join cluster,
403 * it doesn't need to run during another node's failure */
404 if (test_bit(MD_CLUSTER_BEGIN_JOIN_CLUSTER, &cinfo->state)) {
405 complete(&cinfo->completion);
406 clear_bit(MD_CLUSTER_BEGIN_JOIN_CLUSTER, &cinfo->state);
407 }
408 clear_bit(MD_CLUSTER_SUSPEND_READ_BALANCING, &cinfo->state);
409}
410
411/* the ops is called when node join the cluster, and do lock recovery
412 * if node failure occurs */
413static const struct dlm_lockspace_ops md_ls_ops = {
414 .recover_prep = recover_prep,
415 .recover_slot = recover_slot,
416 .recover_done = recover_done,
417};
418
419/*
420 * The BAST function for the ack lock resource
421 * This function wakes up the receive thread in
422 * order to receive and process the message.
423 */
424static void ack_bast(void *arg, int mode)
425{
426 struct dlm_lock_resource *res = arg;
427 struct md_cluster_info *cinfo = res->mddev->cluster_info;
428
429 if (mode == DLM_LOCK_EX) {
430 if (test_bit(MD_CLUSTER_ALREADY_IN_CLUSTER, &cinfo->state))
431 md_wakeup_thread(cinfo->recv_thread);
432 else
433 set_bit(MD_CLUSTER_PENDING_RECV_EVENT, &cinfo->state);
434 }
435}
436
437static void remove_suspend_info(struct mddev *mddev, int slot)
438{
439 struct md_cluster_info *cinfo = mddev->cluster_info;
440 mddev->pers->quiesce(mddev, 1);
441 spin_lock_irq(&cinfo->suspend_lock);
442 cinfo->suspend_hi = 0;
443 cinfo->suspend_lo = 0;
444 spin_unlock_irq(&cinfo->suspend_lock);
445 mddev->pers->quiesce(mddev, 0);
446}
447
448static void process_suspend_info(struct mddev *mddev,
449 int slot, sector_t lo, sector_t hi)
450{
451 struct md_cluster_info *cinfo = mddev->cluster_info;
452 struct mdp_superblock_1 *sb = NULL;
453 struct md_rdev *rdev;
454
455 if (!hi) {
456 /*
457 * clear the REMOTE flag since resync or recovery is finished
458 * in remote node.
459 */
460 clear_bit(MD_RESYNCING_REMOTE, &mddev->recovery);
461 remove_suspend_info(mddev, slot);
462 set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
463 md_wakeup_thread(mddev->thread);
464 return;
465 }
466
467 rdev_for_each(rdev, mddev)
468 if (rdev->raid_disk > -1 && !test_bit(Faulty, &rdev->flags)) {
469 sb = page_address(rdev->sb_page);
470 break;
471 }
472
473 /*
474 * The bitmaps are not same for different nodes
475 * if RESYNCING is happening in one node, then
476 * the node which received the RESYNCING message
477 * probably will perform resync with the region
478 * [lo, hi] again, so we could reduce resync time
479 * a lot if we can ensure that the bitmaps among
480 * different nodes are match up well.
481 *
482 * sync_low/hi is used to record the region which
483 * arrived in the previous RESYNCING message,
484 *
485 * Call md_bitmap_sync_with_cluster to clear NEEDED_MASK
486 * and set RESYNC_MASK since resync thread is running
487 * in another node, so we don't need to do the resync
488 * again with the same section.
489 *
490 * Skip md_bitmap_sync_with_cluster in case reshape
491 * happening, because reshaping region is small and
492 * we don't want to trigger lots of WARN.
493 */
494 if (sb && !(le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE))
495 md_bitmap_sync_with_cluster(mddev, cinfo->sync_low,
496 cinfo->sync_hi, lo, hi);
497 cinfo->sync_low = lo;
498 cinfo->sync_hi = hi;
499
500 mddev->pers->quiesce(mddev, 1);
501 spin_lock_irq(&cinfo->suspend_lock);
502 cinfo->suspend_from = slot;
503 cinfo->suspend_lo = lo;
504 cinfo->suspend_hi = hi;
505 spin_unlock_irq(&cinfo->suspend_lock);
506 mddev->pers->quiesce(mddev, 0);
507}
508
509static void process_add_new_disk(struct mddev *mddev, struct cluster_msg *cmsg)
510{
511 char disk_uuid[64];
512 struct md_cluster_info *cinfo = mddev->cluster_info;
513 char event_name[] = "EVENT=ADD_DEVICE";
514 char raid_slot[16];
515 char *envp[] = {event_name, disk_uuid, raid_slot, NULL};
516 int len;
517
518 len = snprintf(disk_uuid, 64, "DEVICE_UUID=");
519 sprintf(disk_uuid + len, "%pU", cmsg->uuid);
520 snprintf(raid_slot, 16, "RAID_DISK=%d", le32_to_cpu(cmsg->raid_slot));
521 pr_info("%s:%d Sending kobject change with %s and %s\n", __func__, __LINE__, disk_uuid, raid_slot);
522 init_completion(&cinfo->newdisk_completion);
523 set_bit(MD_CLUSTER_WAITING_FOR_NEWDISK, &cinfo->state);
524 kobject_uevent_env(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE, envp);
525 wait_for_completion_timeout(&cinfo->newdisk_completion,
526 NEW_DEV_TIMEOUT);
527 clear_bit(MD_CLUSTER_WAITING_FOR_NEWDISK, &cinfo->state);
528}
529
530
531static void process_metadata_update(struct mddev *mddev, struct cluster_msg *msg)
532{
533 int got_lock = 0;
534 struct md_cluster_info *cinfo = mddev->cluster_info;
535 mddev->good_device_nr = le32_to_cpu(msg->raid_slot);
536
537 dlm_lock_sync(cinfo->no_new_dev_lockres, DLM_LOCK_CR);
538 wait_event(mddev->thread->wqueue,
539 (got_lock = mddev_trylock(mddev)) ||
540 test_bit(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state));
541 md_reload_sb(mddev, mddev->good_device_nr);
542 if (got_lock)
543 mddev_unlock(mddev);
544}
545
546static void process_remove_disk(struct mddev *mddev, struct cluster_msg *msg)
547{
548 struct md_rdev *rdev;
549
550 rcu_read_lock();
551 rdev = md_find_rdev_nr_rcu(mddev, le32_to_cpu(msg->raid_slot));
552 if (rdev) {
553 set_bit(ClusterRemove, &rdev->flags);
554 set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
555 md_wakeup_thread(mddev->thread);
556 }
557 else
558 pr_warn("%s: %d Could not find disk(%d) to REMOVE\n",
559 __func__, __LINE__, le32_to_cpu(msg->raid_slot));
560 rcu_read_unlock();
561}
562
563static void process_readd_disk(struct mddev *mddev, struct cluster_msg *msg)
564{
565 struct md_rdev *rdev;
566
567 rcu_read_lock();
568 rdev = md_find_rdev_nr_rcu(mddev, le32_to_cpu(msg->raid_slot));
569 if (rdev && test_bit(Faulty, &rdev->flags))
570 clear_bit(Faulty, &rdev->flags);
571 else
572 pr_warn("%s: %d Could not find disk(%d) which is faulty",
573 __func__, __LINE__, le32_to_cpu(msg->raid_slot));
574 rcu_read_unlock();
575}
576
577static int process_recvd_msg(struct mddev *mddev, struct cluster_msg *msg)
578{
579 int ret = 0;
580
581 if (WARN(mddev->cluster_info->slot_number - 1 == le32_to_cpu(msg->slot),
582 "node %d received it's own msg\n", le32_to_cpu(msg->slot)))
583 return -1;
584 switch (le32_to_cpu(msg->type)) {
585 case METADATA_UPDATED:
586 process_metadata_update(mddev, msg);
587 break;
588 case CHANGE_CAPACITY:
589 set_capacity(mddev->gendisk, mddev->array_sectors);
590 revalidate_disk(mddev->gendisk);
591 break;
592 case RESYNCING:
593 set_bit(MD_RESYNCING_REMOTE, &mddev->recovery);
594 process_suspend_info(mddev, le32_to_cpu(msg->slot),
595 le64_to_cpu(msg->low),
596 le64_to_cpu(msg->high));
597 break;
598 case NEWDISK:
599 process_add_new_disk(mddev, msg);
600 break;
601 case REMOVE:
602 process_remove_disk(mddev, msg);
603 break;
604 case RE_ADD:
605 process_readd_disk(mddev, msg);
606 break;
607 case BITMAP_NEEDS_SYNC:
608 __recover_slot(mddev, le32_to_cpu(msg->slot));
609 break;
610 case BITMAP_RESIZE:
611 if (le64_to_cpu(msg->high) != mddev->pers->size(mddev, 0, 0))
612 ret = md_bitmap_resize(mddev->bitmap,
613 le64_to_cpu(msg->high), 0, 0);
614 break;
615 default:
616 ret = -1;
617 pr_warn("%s:%d Received unknown message from %d\n",
618 __func__, __LINE__, msg->slot);
619 }
620 return ret;
621}
622
623/*
624 * thread for receiving message
625 */
626static void recv_daemon(struct md_thread *thread)
627{
628 struct md_cluster_info *cinfo = thread->mddev->cluster_info;
629 struct dlm_lock_resource *ack_lockres = cinfo->ack_lockres;
630 struct dlm_lock_resource *message_lockres = cinfo->message_lockres;
631 struct cluster_msg msg;
632 int ret;
633
634 mutex_lock(&cinfo->recv_mutex);
635 /*get CR on Message*/
636 if (dlm_lock_sync(message_lockres, DLM_LOCK_CR)) {
637 pr_err("md/raid1:failed to get CR on MESSAGE\n");
638 mutex_unlock(&cinfo->recv_mutex);
639 return;
640 }
641
642 /* read lvb and wake up thread to process this message_lockres */
643 memcpy(&msg, message_lockres->lksb.sb_lvbptr, sizeof(struct cluster_msg));
644 ret = process_recvd_msg(thread->mddev, &msg);
645 if (ret)
646 goto out;
647
648 /*release CR on ack_lockres*/
649 ret = dlm_unlock_sync(ack_lockres);
650 if (unlikely(ret != 0))
651 pr_info("unlock ack failed return %d\n", ret);
652 /*up-convert to PR on message_lockres*/
653 ret = dlm_lock_sync(message_lockres, DLM_LOCK_PR);
654 if (unlikely(ret != 0))
655 pr_info("lock PR on msg failed return %d\n", ret);
656 /*get CR on ack_lockres again*/
657 ret = dlm_lock_sync(ack_lockres, DLM_LOCK_CR);
658 if (unlikely(ret != 0))
659 pr_info("lock CR on ack failed return %d\n", ret);
660out:
661 /*release CR on message_lockres*/
662 ret = dlm_unlock_sync(message_lockres);
663 if (unlikely(ret != 0))
664 pr_info("unlock msg failed return %d\n", ret);
665 mutex_unlock(&cinfo->recv_mutex);
666}
667
668/* lock_token()
669 * Takes the lock on the TOKEN lock resource so no other
670 * node can communicate while the operation is underway.
671 */
672static int lock_token(struct md_cluster_info *cinfo, bool mddev_locked)
673{
674 int error, set_bit = 0;
675 struct mddev *mddev = cinfo->mddev;
676
677 /*
678 * If resync thread run after raid1d thread, then process_metadata_update
679 * could not continue if raid1d held reconfig_mutex (and raid1d is blocked
680 * since another node already got EX on Token and waitting the EX of Ack),
681 * so let resync wake up thread in case flag is set.
682 */
683 if (mddev_locked && !test_bit(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD,
684 &cinfo->state)) {
685 error = test_and_set_bit_lock(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD,
686 &cinfo->state);
687 WARN_ON_ONCE(error);
688 md_wakeup_thread(mddev->thread);
689 set_bit = 1;
690 }
691 error = dlm_lock_sync(cinfo->token_lockres, DLM_LOCK_EX);
692 if (set_bit)
693 clear_bit_unlock(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state);
694
695 if (error)
696 pr_err("md-cluster(%s:%d): failed to get EX on TOKEN (%d)\n",
697 __func__, __LINE__, error);
698
699 /* Lock the receive sequence */
700 mutex_lock(&cinfo->recv_mutex);
701 return error;
702}
703
704/* lock_comm()
705 * Sets the MD_CLUSTER_SEND_LOCK bit to lock the send channel.
706 */
707static int lock_comm(struct md_cluster_info *cinfo, bool mddev_locked)
708{
709 wait_event(cinfo->wait,
710 !test_and_set_bit(MD_CLUSTER_SEND_LOCK, &cinfo->state));
711
712 return lock_token(cinfo, mddev_locked);
713}
714
715static void unlock_comm(struct md_cluster_info *cinfo)
716{
717 WARN_ON(cinfo->token_lockres->mode != DLM_LOCK_EX);
718 mutex_unlock(&cinfo->recv_mutex);
719 dlm_unlock_sync(cinfo->token_lockres);
720 clear_bit(MD_CLUSTER_SEND_LOCK, &cinfo->state);
721 wake_up(&cinfo->wait);
722}
723
724/* __sendmsg()
725 * This function performs the actual sending of the message. This function is
726 * usually called after performing the encompassing operation
727 * The function:
728 * 1. Grabs the message lockresource in EX mode
729 * 2. Copies the message to the message LVB
730 * 3. Downconverts message lockresource to CW
731 * 4. Upconverts ack lock resource from CR to EX. This forces the BAST on other nodes
732 * and the other nodes read the message. The thread will wait here until all other
733 * nodes have released ack lock resource.
734 * 5. Downconvert ack lockresource to CR
735 */
736static int __sendmsg(struct md_cluster_info *cinfo, struct cluster_msg *cmsg)
737{
738 int error;
739 int slot = cinfo->slot_number - 1;
740
741 cmsg->slot = cpu_to_le32(slot);
742 /*get EX on Message*/
743 error = dlm_lock_sync(cinfo->message_lockres, DLM_LOCK_EX);
744 if (error) {
745 pr_err("md-cluster: failed to get EX on MESSAGE (%d)\n", error);
746 goto failed_message;
747 }
748
749 memcpy(cinfo->message_lockres->lksb.sb_lvbptr, (void *)cmsg,
750 sizeof(struct cluster_msg));
751 /*down-convert EX to CW on Message*/
752 error = dlm_lock_sync(cinfo->message_lockres, DLM_LOCK_CW);
753 if (error) {
754 pr_err("md-cluster: failed to convert EX to CW on MESSAGE(%d)\n",
755 error);
756 goto failed_ack;
757 }
758
759 /*up-convert CR to EX on Ack*/
760 error = dlm_lock_sync(cinfo->ack_lockres, DLM_LOCK_EX);
761 if (error) {
762 pr_err("md-cluster: failed to convert CR to EX on ACK(%d)\n",
763 error);
764 goto failed_ack;
765 }
766
767 /*down-convert EX to CR on Ack*/
768 error = dlm_lock_sync(cinfo->ack_lockres, DLM_LOCK_CR);
769 if (error) {
770 pr_err("md-cluster: failed to convert EX to CR on ACK(%d)\n",
771 error);
772 goto failed_ack;
773 }
774
775failed_ack:
776 error = dlm_unlock_sync(cinfo->message_lockres);
777 if (unlikely(error != 0)) {
778 pr_err("md-cluster: failed convert to NL on MESSAGE(%d)\n",
779 error);
780 /* in case the message can't be released due to some reason */
781 goto failed_ack;
782 }
783failed_message:
784 return error;
785}
786
787static int sendmsg(struct md_cluster_info *cinfo, struct cluster_msg *cmsg,
788 bool mddev_locked)
789{
790 int ret;
791
792 lock_comm(cinfo, mddev_locked);
793 ret = __sendmsg(cinfo, cmsg);
794 unlock_comm(cinfo);
795 return ret;
796}
797
798static int gather_all_resync_info(struct mddev *mddev, int total_slots)
799{
800 struct md_cluster_info *cinfo = mddev->cluster_info;
801 int i, ret = 0;
802 struct dlm_lock_resource *bm_lockres;
803 char str[64];
804 sector_t lo, hi;
805
806
807 for (i = 0; i < total_slots; i++) {
808 memset(str, '\0', 64);
809 snprintf(str, 64, "bitmap%04d", i);
810 bm_lockres = lockres_init(mddev, str, NULL, 1);
811 if (!bm_lockres)
812 return -ENOMEM;
813 if (i == (cinfo->slot_number - 1)) {
814 lockres_free(bm_lockres);
815 continue;
816 }
817
818 bm_lockres->flags |= DLM_LKF_NOQUEUE;
819 ret = dlm_lock_sync(bm_lockres, DLM_LOCK_PW);
820 if (ret == -EAGAIN) {
821 if (read_resync_info(mddev, bm_lockres)) {
822 pr_info("%s:%d Resync[%llu..%llu] in progress on %d\n",
823 __func__, __LINE__,
824 (unsigned long long) cinfo->suspend_lo,
825 (unsigned long long) cinfo->suspend_hi,
826 i);
827 cinfo->suspend_from = i;
828 }
829 ret = 0;
830 lockres_free(bm_lockres);
831 continue;
832 }
833 if (ret) {
834 lockres_free(bm_lockres);
835 goto out;
836 }
837
838 /* Read the disk bitmap sb and check if it needs recovery */
839 ret = md_bitmap_copy_from_slot(mddev, i, &lo, &hi, false);
840 if (ret) {
841 pr_warn("md-cluster: Could not gather bitmaps from slot %d", i);
842 lockres_free(bm_lockres);
843 continue;
844 }
845 if ((hi > 0) && (lo < mddev->recovery_cp)) {
846 set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
847 mddev->recovery_cp = lo;
848 md_check_recovery(mddev);
849 }
850
851 lockres_free(bm_lockres);
852 }
853out:
854 return ret;
855}
856
857static int join(struct mddev *mddev, int nodes)
858{
859 struct md_cluster_info *cinfo;
860 int ret, ops_rv;
861 char str[64];
862
863 cinfo = kzalloc(sizeof(struct md_cluster_info), GFP_KERNEL);
864 if (!cinfo)
865 return -ENOMEM;
866
867 INIT_LIST_HEAD(&cinfo->suspend_list);
868 spin_lock_init(&cinfo->suspend_lock);
869 init_completion(&cinfo->completion);
870 set_bit(MD_CLUSTER_BEGIN_JOIN_CLUSTER, &cinfo->state);
871 init_waitqueue_head(&cinfo->wait);
872 mutex_init(&cinfo->recv_mutex);
873
874 mddev->cluster_info = cinfo;
875 cinfo->mddev = mddev;
876
877 memset(str, 0, 64);
878 sprintf(str, "%pU", mddev->uuid);
879 ret = dlm_new_lockspace(str, mddev->bitmap_info.cluster_name,
880 DLM_LSFL_FS, LVB_SIZE,
881 &md_ls_ops, mddev, &ops_rv, &cinfo->lockspace);
882 if (ret)
883 goto err;
884 wait_for_completion(&cinfo->completion);
885 if (nodes < cinfo->slot_number) {
886 pr_err("md-cluster: Slot allotted(%d) is greater than available slots(%d).",
887 cinfo->slot_number, nodes);
888 ret = -ERANGE;
889 goto err;
890 }
891 /* Initiate the communication resources */
892 ret = -ENOMEM;
893 cinfo->recv_thread = md_register_thread(recv_daemon, mddev, "cluster_recv");
894 if (!cinfo->recv_thread) {
895 pr_err("md-cluster: cannot allocate memory for recv_thread!\n");
896 goto err;
897 }
898 cinfo->message_lockres = lockres_init(mddev, "message", NULL, 1);
899 if (!cinfo->message_lockres)
900 goto err;
901 cinfo->token_lockres = lockres_init(mddev, "token", NULL, 0);
902 if (!cinfo->token_lockres)
903 goto err;
904 cinfo->no_new_dev_lockres = lockres_init(mddev, "no-new-dev", NULL, 0);
905 if (!cinfo->no_new_dev_lockres)
906 goto err;
907
908 ret = dlm_lock_sync(cinfo->token_lockres, DLM_LOCK_EX);
909 if (ret) {
910 ret = -EAGAIN;
911 pr_err("md-cluster: can't join cluster to avoid lock issue\n");
912 goto err;
913 }
914 cinfo->ack_lockres = lockres_init(mddev, "ack", ack_bast, 0);
915 if (!cinfo->ack_lockres) {
916 ret = -ENOMEM;
917 goto err;
918 }
919 /* get sync CR lock on ACK. */
920 if (dlm_lock_sync(cinfo->ack_lockres, DLM_LOCK_CR))
921 pr_err("md-cluster: failed to get a sync CR lock on ACK!(%d)\n",
922 ret);
923 dlm_unlock_sync(cinfo->token_lockres);
924 /* get sync CR lock on no-new-dev. */
925 if (dlm_lock_sync(cinfo->no_new_dev_lockres, DLM_LOCK_CR))
926 pr_err("md-cluster: failed to get a sync CR lock on no-new-dev!(%d)\n", ret);
927
928
929 pr_info("md-cluster: Joined cluster %s slot %d\n", str, cinfo->slot_number);
930 snprintf(str, 64, "bitmap%04d", cinfo->slot_number - 1);
931 cinfo->bitmap_lockres = lockres_init(mddev, str, NULL, 1);
932 if (!cinfo->bitmap_lockres) {
933 ret = -ENOMEM;
934 goto err;
935 }
936 if (dlm_lock_sync(cinfo->bitmap_lockres, DLM_LOCK_PW)) {
937 pr_err("Failed to get bitmap lock\n");
938 ret = -EINVAL;
939 goto err;
940 }
941
942 cinfo->resync_lockres = lockres_init(mddev, "resync", NULL, 0);
943 if (!cinfo->resync_lockres) {
944 ret = -ENOMEM;
945 goto err;
946 }
947
948 return 0;
949err:
950 set_bit(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state);
951 md_unregister_thread(&cinfo->recovery_thread);
952 md_unregister_thread(&cinfo->recv_thread);
953 lockres_free(cinfo->message_lockres);
954 lockres_free(cinfo->token_lockres);
955 lockres_free(cinfo->ack_lockres);
956 lockres_free(cinfo->no_new_dev_lockres);
957 lockres_free(cinfo->resync_lockres);
958 lockres_free(cinfo->bitmap_lockres);
959 if (cinfo->lockspace)
960 dlm_release_lockspace(cinfo->lockspace, 2);
961 mddev->cluster_info = NULL;
962 kfree(cinfo);
963 return ret;
964}
965
966static void load_bitmaps(struct mddev *mddev, int total_slots)
967{
968 struct md_cluster_info *cinfo = mddev->cluster_info;
969
970 /* load all the node's bitmap info for resync */
971 if (gather_all_resync_info(mddev, total_slots))
972 pr_err("md-cluster: failed to gather all resyn infos\n");
973 set_bit(MD_CLUSTER_ALREADY_IN_CLUSTER, &cinfo->state);
974 /* wake up recv thread in case something need to be handled */
975 if (test_and_clear_bit(MD_CLUSTER_PENDING_RECV_EVENT, &cinfo->state))
976 md_wakeup_thread(cinfo->recv_thread);
977}
978
979static void resync_bitmap(struct mddev *mddev)
980{
981 struct md_cluster_info *cinfo = mddev->cluster_info;
982 struct cluster_msg cmsg = {0};
983 int err;
984
985 cmsg.type = cpu_to_le32(BITMAP_NEEDS_SYNC);
986 err = sendmsg(cinfo, &cmsg, 1);
987 if (err)
988 pr_err("%s:%d: failed to send BITMAP_NEEDS_SYNC message (%d)\n",
989 __func__, __LINE__, err);
990}
991
992static void unlock_all_bitmaps(struct mddev *mddev);
993static int leave(struct mddev *mddev)
994{
995 struct md_cluster_info *cinfo = mddev->cluster_info;
996
997 if (!cinfo)
998 return 0;
999
1000 /*
1001 * BITMAP_NEEDS_SYNC message should be sent when node
1002 * is leaving the cluster with dirty bitmap, also we
1003 * can only deliver it when dlm connection is available.
1004 *
1005 * Also, we should send BITMAP_NEEDS_SYNC message in
1006 * case reshaping is interrupted.
1007 */
1008 if ((cinfo->slot_number > 0 && mddev->recovery_cp != MaxSector) ||
1009 (mddev->reshape_position != MaxSector &&
1010 test_bit(MD_CLOSING, &mddev->flags)))
1011 resync_bitmap(mddev);
1012
1013 set_bit(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state);
1014 md_unregister_thread(&cinfo->recovery_thread);
1015 md_unregister_thread(&cinfo->recv_thread);
1016 lockres_free(cinfo->message_lockres);
1017 lockres_free(cinfo->token_lockres);
1018 lockres_free(cinfo->ack_lockres);
1019 lockres_free(cinfo->no_new_dev_lockres);
1020 lockres_free(cinfo->resync_lockres);
1021 lockres_free(cinfo->bitmap_lockres);
1022 unlock_all_bitmaps(mddev);
1023 dlm_release_lockspace(cinfo->lockspace, 2);
1024 kfree(cinfo);
1025 return 0;
1026}
1027
1028/* slot_number(): Returns the MD slot number to use
1029 * DLM starts the slot numbers from 1, wheras cluster-md
1030 * wants the number to be from zero, so we deduct one
1031 */
1032static int slot_number(struct mddev *mddev)
1033{
1034 struct md_cluster_info *cinfo = mddev->cluster_info;
1035
1036 return cinfo->slot_number - 1;
1037}
1038
1039/*
1040 * Check if the communication is already locked, else lock the communication
1041 * channel.
1042 * If it is already locked, token is in EX mode, and hence lock_token()
1043 * should not be called.
1044 */
1045static int metadata_update_start(struct mddev *mddev)
1046{
1047 struct md_cluster_info *cinfo = mddev->cluster_info;
1048 int ret;
1049
1050 /*
1051 * metadata_update_start is always called with the protection of
1052 * reconfig_mutex, so set WAITING_FOR_TOKEN here.
1053 */
1054 ret = test_and_set_bit_lock(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD,
1055 &cinfo->state);
1056 WARN_ON_ONCE(ret);
1057 md_wakeup_thread(mddev->thread);
1058
1059 wait_event(cinfo->wait,
1060 !test_and_set_bit(MD_CLUSTER_SEND_LOCK, &cinfo->state) ||
1061 test_and_clear_bit(MD_CLUSTER_SEND_LOCKED_ALREADY, &cinfo->state));
1062
1063 /* If token is already locked, return 0 */
1064 if (cinfo->token_lockres->mode == DLM_LOCK_EX) {
1065 clear_bit_unlock(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state);
1066 return 0;
1067 }
1068
1069 ret = lock_token(cinfo, 1);
1070 clear_bit_unlock(MD_CLUSTER_HOLDING_MUTEX_FOR_RECVD, &cinfo->state);
1071 return ret;
1072}
1073
1074static int metadata_update_finish(struct mddev *mddev)
1075{
1076 struct md_cluster_info *cinfo = mddev->cluster_info;
1077 struct cluster_msg cmsg;
1078 struct md_rdev *rdev;
1079 int ret = 0;
1080 int raid_slot = -1;
1081
1082 memset(&cmsg, 0, sizeof(cmsg));
1083 cmsg.type = cpu_to_le32(METADATA_UPDATED);
1084 /* Pick up a good active device number to send.
1085 */
1086 rdev_for_each(rdev, mddev)
1087 if (rdev->raid_disk > -1 && !test_bit(Faulty, &rdev->flags)) {
1088 raid_slot = rdev->desc_nr;
1089 break;
1090 }
1091 if (raid_slot >= 0) {
1092 cmsg.raid_slot = cpu_to_le32(raid_slot);
1093 ret = __sendmsg(cinfo, &cmsg);
1094 } else
1095 pr_warn("md-cluster: No good device id found to send\n");
1096 clear_bit(MD_CLUSTER_SEND_LOCKED_ALREADY, &cinfo->state);
1097 unlock_comm(cinfo);
1098 return ret;
1099}
1100
1101static void metadata_update_cancel(struct mddev *mddev)
1102{
1103 struct md_cluster_info *cinfo = mddev->cluster_info;
1104 clear_bit(MD_CLUSTER_SEND_LOCKED_ALREADY, &cinfo->state);
1105 unlock_comm(cinfo);
1106}
1107
1108static int update_bitmap_size(struct mddev *mddev, sector_t size)
1109{
1110 struct md_cluster_info *cinfo = mddev->cluster_info;
1111 struct cluster_msg cmsg = {0};
1112 int ret;
1113
1114 cmsg.type = cpu_to_le32(BITMAP_RESIZE);
1115 cmsg.high = cpu_to_le64(size);
1116 ret = sendmsg(cinfo, &cmsg, 0);
1117 if (ret)
1118 pr_err("%s:%d: failed to send BITMAP_RESIZE message (%d)\n",
1119 __func__, __LINE__, ret);
1120 return ret;
1121}
1122
1123static int resize_bitmaps(struct mddev *mddev, sector_t newsize, sector_t oldsize)
1124{
1125 struct bitmap_counts *counts;
1126 char str[64];
1127 struct dlm_lock_resource *bm_lockres;
1128 struct bitmap *bitmap = mddev->bitmap;
1129 unsigned long my_pages = bitmap->counts.pages;
1130 int i, rv;
1131
1132 /*
1133 * We need to ensure all the nodes can grow to a larger
1134 * bitmap size before make the reshaping.
1135 */
1136 rv = update_bitmap_size(mddev, newsize);
1137 if (rv)
1138 return rv;
1139
1140 for (i = 0; i < mddev->bitmap_info.nodes; i++) {
1141 if (i == md_cluster_ops->slot_number(mddev))
1142 continue;
1143
1144 bitmap = get_bitmap_from_slot(mddev, i);
1145 if (IS_ERR(bitmap)) {
1146 pr_err("can't get bitmap from slot %d\n", i);
1147 goto out;
1148 }
1149 counts = &bitmap->counts;
1150
1151 /*
1152 * If we can hold the bitmap lock of one node then
1153 * the slot is not occupied, update the pages.
1154 */
1155 snprintf(str, 64, "bitmap%04d", i);
1156 bm_lockres = lockres_init(mddev, str, NULL, 1);
1157 if (!bm_lockres) {
1158 pr_err("Cannot initialize %s lock\n", str);
1159 goto out;
1160 }
1161 bm_lockres->flags |= DLM_LKF_NOQUEUE;
1162 rv = dlm_lock_sync(bm_lockres, DLM_LOCK_PW);
1163 if (!rv)
1164 counts->pages = my_pages;
1165 lockres_free(bm_lockres);
1166
1167 if (my_pages != counts->pages)
1168 /*
1169 * Let's revert the bitmap size if one node
1170 * can't resize bitmap
1171 */
1172 goto out;
1173 }
1174
1175 return 0;
1176out:
1177 md_bitmap_free(bitmap);
1178 update_bitmap_size(mddev, oldsize);
1179 return -1;
1180}
1181
1182/*
1183 * return 0 if all the bitmaps have the same sync_size
1184 */
1185static int cluster_check_sync_size(struct mddev *mddev)
1186{
1187 int i, rv;
1188 bitmap_super_t *sb;
1189 unsigned long my_sync_size, sync_size = 0;
1190 int node_num = mddev->bitmap_info.nodes;
1191 int current_slot = md_cluster_ops->slot_number(mddev);
1192 struct bitmap *bitmap = mddev->bitmap;
1193 char str[64];
1194 struct dlm_lock_resource *bm_lockres;
1195
1196 sb = kmap_atomic(bitmap->storage.sb_page);
1197 my_sync_size = sb->sync_size;
1198 kunmap_atomic(sb);
1199
1200 for (i = 0; i < node_num; i++) {
1201 if (i == current_slot)
1202 continue;
1203
1204 bitmap = get_bitmap_from_slot(mddev, i);
1205 if (IS_ERR(bitmap)) {
1206 pr_err("can't get bitmap from slot %d\n", i);
1207 return -1;
1208 }
1209
1210 /*
1211 * If we can hold the bitmap lock of one node then
1212 * the slot is not occupied, update the sb.
1213 */
1214 snprintf(str, 64, "bitmap%04d", i);
1215 bm_lockres = lockres_init(mddev, str, NULL, 1);
1216 if (!bm_lockres) {
1217 pr_err("md-cluster: Cannot initialize %s\n", str);
1218 md_bitmap_free(bitmap);
1219 return -1;
1220 }
1221 bm_lockres->flags |= DLM_LKF_NOQUEUE;
1222 rv = dlm_lock_sync(bm_lockres, DLM_LOCK_PW);
1223 if (!rv)
1224 md_bitmap_update_sb(bitmap);
1225 lockres_free(bm_lockres);
1226
1227 sb = kmap_atomic(bitmap->storage.sb_page);
1228 if (sync_size == 0)
1229 sync_size = sb->sync_size;
1230 else if (sync_size != sb->sync_size) {
1231 kunmap_atomic(sb);
1232 md_bitmap_free(bitmap);
1233 return -1;
1234 }
1235 kunmap_atomic(sb);
1236 md_bitmap_free(bitmap);
1237 }
1238
1239 return (my_sync_size == sync_size) ? 0 : -1;
1240}
1241
1242/*
1243 * Update the size for cluster raid is a little more complex, we perform it
1244 * by the steps:
1245 * 1. hold token lock and update superblock in initiator node.
1246 * 2. send METADATA_UPDATED msg to other nodes.
1247 * 3. The initiator node continues to check each bitmap's sync_size, if all
1248 * bitmaps have the same value of sync_size, then we can set capacity and
1249 * let other nodes to perform it. If one node can't update sync_size
1250 * accordingly, we need to revert to previous value.
1251 */
1252static void update_size(struct mddev *mddev, sector_t old_dev_sectors)
1253{
1254 struct md_cluster_info *cinfo = mddev->cluster_info;
1255 struct cluster_msg cmsg;
1256 struct md_rdev *rdev;
1257 int ret = 0;
1258 int raid_slot = -1;
1259
1260 md_update_sb(mddev, 1);
1261 lock_comm(cinfo, 1);
1262
1263 memset(&cmsg, 0, sizeof(cmsg));
1264 cmsg.type = cpu_to_le32(METADATA_UPDATED);
1265 rdev_for_each(rdev, mddev)
1266 if (rdev->raid_disk >= 0 && !test_bit(Faulty, &rdev->flags)) {
1267 raid_slot = rdev->desc_nr;
1268 break;
1269 }
1270 if (raid_slot >= 0) {
1271 cmsg.raid_slot = cpu_to_le32(raid_slot);
1272 /*
1273 * We can only change capiticy after all the nodes can do it,
1274 * so need to wait after other nodes already received the msg
1275 * and handled the change
1276 */
1277 ret = __sendmsg(cinfo, &cmsg);
1278 if (ret) {
1279 pr_err("%s:%d: failed to send METADATA_UPDATED msg\n",
1280 __func__, __LINE__);
1281 unlock_comm(cinfo);
1282 return;
1283 }
1284 } else {
1285 pr_err("md-cluster: No good device id found to send\n");
1286 unlock_comm(cinfo);
1287 return;
1288 }
1289
1290 /*
1291 * check the sync_size from other node's bitmap, if sync_size
1292 * have already updated in other nodes as expected, send an
1293 * empty metadata msg to permit the change of capacity
1294 */
1295 if (cluster_check_sync_size(mddev) == 0) {
1296 memset(&cmsg, 0, sizeof(cmsg));
1297 cmsg.type = cpu_to_le32(CHANGE_CAPACITY);
1298 ret = __sendmsg(cinfo, &cmsg);
1299 if (ret)
1300 pr_err("%s:%d: failed to send CHANGE_CAPACITY msg\n",
1301 __func__, __LINE__);
1302 set_capacity(mddev->gendisk, mddev->array_sectors);
1303 revalidate_disk(mddev->gendisk);
1304 } else {
1305 /* revert to previous sectors */
1306 ret = mddev->pers->resize(mddev, old_dev_sectors);
1307 if (!ret)
1308 revalidate_disk(mddev->gendisk);
1309 ret = __sendmsg(cinfo, &cmsg);
1310 if (ret)
1311 pr_err("%s:%d: failed to send METADATA_UPDATED msg\n",
1312 __func__, __LINE__);
1313 }
1314 unlock_comm(cinfo);
1315}
1316
1317static int resync_start(struct mddev *mddev)
1318{
1319 struct md_cluster_info *cinfo = mddev->cluster_info;
1320 return dlm_lock_sync_interruptible(cinfo->resync_lockres, DLM_LOCK_EX, mddev);
1321}
1322
1323static void resync_info_get(struct mddev *mddev, sector_t *lo, sector_t *hi)
1324{
1325 struct md_cluster_info *cinfo = mddev->cluster_info;
1326
1327 spin_lock_irq(&cinfo->suspend_lock);
1328 *lo = cinfo->suspend_lo;
1329 *hi = cinfo->suspend_hi;
1330 spin_unlock_irq(&cinfo->suspend_lock);
1331}
1332
1333static int resync_info_update(struct mddev *mddev, sector_t lo, sector_t hi)
1334{
1335 struct md_cluster_info *cinfo = mddev->cluster_info;
1336 struct resync_info ri;
1337 struct cluster_msg cmsg = {0};
1338
1339 /* do not send zero again, if we have sent before */
1340 if (hi == 0) {
1341 memcpy(&ri, cinfo->bitmap_lockres->lksb.sb_lvbptr, sizeof(struct resync_info));
1342 if (le64_to_cpu(ri.hi) == 0)
1343 return 0;
1344 }
1345
1346 add_resync_info(cinfo->bitmap_lockres, lo, hi);
1347 /* Re-acquire the lock to refresh LVB */
1348 dlm_lock_sync(cinfo->bitmap_lockres, DLM_LOCK_PW);
1349 cmsg.type = cpu_to_le32(RESYNCING);
1350 cmsg.low = cpu_to_le64(lo);
1351 cmsg.high = cpu_to_le64(hi);
1352
1353 /*
1354 * mddev_lock is held if resync_info_update is called from
1355 * resync_finish (md_reap_sync_thread -> resync_finish)
1356 */
1357 if (lo == 0 && hi == 0)
1358 return sendmsg(cinfo, &cmsg, 1);
1359 else
1360 return sendmsg(cinfo, &cmsg, 0);
1361}
1362
1363static int resync_finish(struct mddev *mddev)
1364{
1365 struct md_cluster_info *cinfo = mddev->cluster_info;
1366 int ret = 0;
1367
1368 clear_bit(MD_RESYNCING_REMOTE, &mddev->recovery);
1369
1370 /*
1371 * If resync thread is interrupted so we can't say resync is finished,
1372 * another node will launch resync thread to continue.
1373 */
1374 if (!test_bit(MD_CLOSING, &mddev->flags))
1375 ret = resync_info_update(mddev, 0, 0);
1376 dlm_unlock_sync(cinfo->resync_lockres);
1377 return ret;
1378}
1379
1380static int area_resyncing(struct mddev *mddev, int direction,
1381 sector_t lo, sector_t hi)
1382{
1383 struct md_cluster_info *cinfo = mddev->cluster_info;
1384 int ret = 0;
1385
1386 if ((direction == READ) &&
1387 test_bit(MD_CLUSTER_SUSPEND_READ_BALANCING, &cinfo->state))
1388 return 1;
1389
1390 spin_lock_irq(&cinfo->suspend_lock);
1391 if (hi > cinfo->suspend_lo && lo < cinfo->suspend_hi)
1392 ret = 1;
1393 spin_unlock_irq(&cinfo->suspend_lock);
1394 return ret;
1395}
1396
1397/* add_new_disk() - initiates a disk add
1398 * However, if this fails before writing md_update_sb(),
1399 * add_new_disk_cancel() must be called to release token lock
1400 */
1401static int add_new_disk(struct mddev *mddev, struct md_rdev *rdev)
1402{
1403 struct md_cluster_info *cinfo = mddev->cluster_info;
1404 struct cluster_msg cmsg;
1405 int ret = 0;
1406 struct mdp_superblock_1 *sb = page_address(rdev->sb_page);
1407 char *uuid = sb->device_uuid;
1408
1409 memset(&cmsg, 0, sizeof(cmsg));
1410 cmsg.type = cpu_to_le32(NEWDISK);
1411 memcpy(cmsg.uuid, uuid, 16);
1412 cmsg.raid_slot = cpu_to_le32(rdev->desc_nr);
1413 lock_comm(cinfo, 1);
1414 ret = __sendmsg(cinfo, &cmsg);
1415 if (ret) {
1416 unlock_comm(cinfo);
1417 return ret;
1418 }
1419 cinfo->no_new_dev_lockres->flags |= DLM_LKF_NOQUEUE;
1420 ret = dlm_lock_sync(cinfo->no_new_dev_lockres, DLM_LOCK_EX);
1421 cinfo->no_new_dev_lockres->flags &= ~DLM_LKF_NOQUEUE;
1422 /* Some node does not "see" the device */
1423 if (ret == -EAGAIN)
1424 ret = -ENOENT;
1425 if (ret)
1426 unlock_comm(cinfo);
1427 else {
1428 dlm_lock_sync(cinfo->no_new_dev_lockres, DLM_LOCK_CR);
1429 /* Since MD_CHANGE_DEVS will be set in add_bound_rdev which
1430 * will run soon after add_new_disk, the below path will be
1431 * invoked:
1432 * md_wakeup_thread(mddev->thread)
1433 * -> conf->thread (raid1d)
1434 * -> md_check_recovery -> md_update_sb
1435 * -> metadata_update_start/finish
1436 * MD_CLUSTER_SEND_LOCKED_ALREADY will be cleared eventually.
1437 *
1438 * For other failure cases, metadata_update_cancel and
1439 * add_new_disk_cancel also clear below bit as well.
1440 * */
1441 set_bit(MD_CLUSTER_SEND_LOCKED_ALREADY, &cinfo->state);
1442 wake_up(&cinfo->wait);
1443 }
1444 return ret;
1445}
1446
1447static void add_new_disk_cancel(struct mddev *mddev)
1448{
1449 struct md_cluster_info *cinfo = mddev->cluster_info;
1450 clear_bit(MD_CLUSTER_SEND_LOCKED_ALREADY, &cinfo->state);
1451 unlock_comm(cinfo);
1452}
1453
1454static int new_disk_ack(struct mddev *mddev, bool ack)
1455{
1456 struct md_cluster_info *cinfo = mddev->cluster_info;
1457
1458 if (!test_bit(MD_CLUSTER_WAITING_FOR_NEWDISK, &cinfo->state)) {
1459 pr_warn("md-cluster(%s): Spurious cluster confirmation\n", mdname(mddev));
1460 return -EINVAL;
1461 }
1462
1463 if (ack)
1464 dlm_unlock_sync(cinfo->no_new_dev_lockres);
1465 complete(&cinfo->newdisk_completion);
1466 return 0;
1467}
1468
1469static int remove_disk(struct mddev *mddev, struct md_rdev *rdev)
1470{
1471 struct cluster_msg cmsg = {0};
1472 struct md_cluster_info *cinfo = mddev->cluster_info;
1473 cmsg.type = cpu_to_le32(REMOVE);
1474 cmsg.raid_slot = cpu_to_le32(rdev->desc_nr);
1475 return sendmsg(cinfo, &cmsg, 1);
1476}
1477
1478static int lock_all_bitmaps(struct mddev *mddev)
1479{
1480 int slot, my_slot, ret, held = 1, i = 0;
1481 char str[64];
1482 struct md_cluster_info *cinfo = mddev->cluster_info;
1483
1484 cinfo->other_bitmap_lockres =
1485 kcalloc(mddev->bitmap_info.nodes - 1,
1486 sizeof(struct dlm_lock_resource *), GFP_KERNEL);
1487 if (!cinfo->other_bitmap_lockres) {
1488 pr_err("md: can't alloc mem for other bitmap locks\n");
1489 return 0;
1490 }
1491
1492 my_slot = slot_number(mddev);
1493 for (slot = 0; slot < mddev->bitmap_info.nodes; slot++) {
1494 if (slot == my_slot)
1495 continue;
1496
1497 memset(str, '\0', 64);
1498 snprintf(str, 64, "bitmap%04d", slot);
1499 cinfo->other_bitmap_lockres[i] = lockres_init(mddev, str, NULL, 1);
1500 if (!cinfo->other_bitmap_lockres[i])
1501 return -ENOMEM;
1502
1503 cinfo->other_bitmap_lockres[i]->flags |= DLM_LKF_NOQUEUE;
1504 ret = dlm_lock_sync(cinfo->other_bitmap_lockres[i], DLM_LOCK_PW);
1505 if (ret)
1506 held = -1;
1507 i++;
1508 }
1509
1510 return held;
1511}
1512
1513static void unlock_all_bitmaps(struct mddev *mddev)
1514{
1515 struct md_cluster_info *cinfo = mddev->cluster_info;
1516 int i;
1517
1518 /* release other node's bitmap lock if they are existed */
1519 if (cinfo->other_bitmap_lockres) {
1520 for (i = 0; i < mddev->bitmap_info.nodes - 1; i++) {
1521 if (cinfo->other_bitmap_lockres[i]) {
1522 lockres_free(cinfo->other_bitmap_lockres[i]);
1523 }
1524 }
1525 kfree(cinfo->other_bitmap_lockres);
1526 }
1527}
1528
1529static int gather_bitmaps(struct md_rdev *rdev)
1530{
1531 int sn, err;
1532 sector_t lo, hi;
1533 struct cluster_msg cmsg = {0};
1534 struct mddev *mddev = rdev->mddev;
1535 struct md_cluster_info *cinfo = mddev->cluster_info;
1536
1537 cmsg.type = cpu_to_le32(RE_ADD);
1538 cmsg.raid_slot = cpu_to_le32(rdev->desc_nr);
1539 err = sendmsg(cinfo, &cmsg, 1);
1540 if (err)
1541 goto out;
1542
1543 for (sn = 0; sn < mddev->bitmap_info.nodes; sn++) {
1544 if (sn == (cinfo->slot_number - 1))
1545 continue;
1546 err = md_bitmap_copy_from_slot(mddev, sn, &lo, &hi, false);
1547 if (err) {
1548 pr_warn("md-cluster: Could not gather bitmaps from slot %d", sn);
1549 goto out;
1550 }
1551 if ((hi > 0) && (lo < mddev->recovery_cp))
1552 mddev->recovery_cp = lo;
1553 }
1554out:
1555 return err;
1556}
1557
1558static struct md_cluster_operations cluster_ops = {
1559 .join = join,
1560 .leave = leave,
1561 .slot_number = slot_number,
1562 .resync_start = resync_start,
1563 .resync_finish = resync_finish,
1564 .resync_info_update = resync_info_update,
1565 .resync_info_get = resync_info_get,
1566 .metadata_update_start = metadata_update_start,
1567 .metadata_update_finish = metadata_update_finish,
1568 .metadata_update_cancel = metadata_update_cancel,
1569 .area_resyncing = area_resyncing,
1570 .add_new_disk = add_new_disk,
1571 .add_new_disk_cancel = add_new_disk_cancel,
1572 .new_disk_ack = new_disk_ack,
1573 .remove_disk = remove_disk,
1574 .load_bitmaps = load_bitmaps,
1575 .gather_bitmaps = gather_bitmaps,
1576 .resize_bitmaps = resize_bitmaps,
1577 .lock_all_bitmaps = lock_all_bitmaps,
1578 .unlock_all_bitmaps = unlock_all_bitmaps,
1579 .update_size = update_size,
1580};
1581
1582static int __init cluster_init(void)
1583{
1584 pr_warn("md-cluster: support raid1 and raid10 (limited support)\n");
1585 pr_info("Registering Cluster MD functions\n");
1586 register_md_cluster_operations(&cluster_ops, THIS_MODULE);
1587 return 0;
1588}
1589
1590static void cluster_exit(void)
1591{
1592 unregister_md_cluster_operations();
1593}
1594
1595module_init(cluster_init);
1596module_exit(cluster_exit);
1597MODULE_AUTHOR("SUSE");
1598MODULE_LICENSE("GPL");
1599MODULE_DESCRIPTION("Clustering support for MD");
1600