aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/md/dm-kcopyd.c
blob: 924f5f0084c27191604907eaa9983a2bf687b6c1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
/*
 * Copyright (C) 2002 Sistina Software (UK) Limited.
 * Copyright (C) 2006 Red Hat GmbH
 *
 * This file is released under the GPL.
 *
 * Kcopyd provides a simple interface for copying an area of one
 * block-device to one or more other block-devices, with an asynchronous
 * completion notification.
 */

#include <linux/types.h>
#include <asm/atomic.h>
#include <linux/blkdev.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/list.h>
#include <linux/mempool.h>
#include <linux/module.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/workqueue.h>
#include <linux/mutex.h>
#include <linux/device-mapper.h>
#include <linux/dm-kcopyd.h>

#include "dm.h"

/*-----------------------------------------------------------------
 * Each kcopyd client has its own little pool of preallocated
 * pages for kcopyd io.
 *---------------------------------------------------------------*/
struct dm_kcopyd_client {
	spinlock_t lock;
	struct page_list *pages;
	unsigned int nr_pages;
	unsigned int nr_free_pages;

	/*
	 * Block devices to unplug.
	 * Non-NULL pointer means that a block device has some pending requests
	 * and needs to be unplugged.
	 */
	struct block_device *unplug[2];

	struct dm_io_client *io_client;

	wait_queue_head_t destroyq;
	atomic_t nr_jobs;

	mempool_t *job_pool;

	struct workqueue_struct *kcopyd_wq;
	struct work_struct kcopyd_work;

/*
 * We maintain three lists of jobs:
 *
 * i)   jobs waiting for pages
 * ii)  jobs that have pages, and are waiting for the io to be issued.
 * iii) jobs that have completed.
 *
 * All three of these are protected by job_lock.
 */
	spinlock_t job_lock;
	struct list_head complete_jobs;
	struct list_head io_jobs;
	struct list_head pages_jobs;
};

static void wake(struct dm_kcopyd_client *kc)
{
	queue_work(kc->kcopyd_wq, &kc->kcopyd_work);
}

static struct page_list *alloc_pl(void)
{
	struct page_list *pl;

	pl = kmalloc(sizeof(*pl), GFP_KERNEL);
	if (!pl)
		return NULL;

	pl->page = alloc_page(GFP_KERNEL);
	if (!pl->page) {
		kfree(pl);
		return NULL;
	}

	return pl;
}

static void free_pl(struct page_list *pl)
{
	__free_page(pl->page);
	kfree(pl);
}

static int kcopyd_get_pages(struct dm_kcopyd_client *kc,
			    unsigned int nr, struct page_list **pages)
{
	struct page_list *pl;

	spin_lock(&kc->lock);
	if (kc->nr_free_pages < nr) {
		spin_unlock(&kc->lock);
		return -ENOMEM;
	}

	kc->nr_free_pages -= nr;
	for (*pages = pl = kc->pages; --nr; pl = pl->next)
		;

	kc->pages = pl->next;
	pl->next = NULL;

	spin_unlock(&kc->lock);

	return 0;
}

static void kcopyd_put_pages(struct dm_kcopyd_client *kc, struct page_list *pl)
{
	struct page_list *cursor;

	spin_lock(&kc->lock);
	for (cursor = pl; cursor->next; cursor = cursor->next)
		kc->nr_free_pages++;

	kc->nr_free_pages++;
	cursor->next = kc->pages;
	kc->pages = pl;
	spin_unlock(&kc->lock);
}

/*
 * These three functions resize the page pool.
 */
static void drop_pages(struct page_list *pl)
{
	struct page_list *next;

	while (pl) {
		next = pl->next;
		free_pl(pl);
		pl = next;
	}
}

static int client_alloc_pages(struct dm_kcopyd_client *kc, unsigned int nr)
{
	unsigned int i;
	struct page_list *pl = NULL, *next;

	for (i = 0; i < nr; i++) {
		next = alloc_pl();
		if (!next) {
			if (pl)
				drop_pages(pl);
			return -ENOMEM;
		}
		next->next = pl;
		pl = next;
	}

	kcopyd_put_pages(kc, pl);
	kc->nr_pages += nr;
	return 0;
}

static void client_free_pages(struct dm_kcopyd_client *kc)
{
	BUG_ON(kc->nr_free_pages != kc->nr_pages);
	drop_pages(kc->pages);
	kc->pages = NULL;
	kc->nr_free_pages = kc->nr_pages = 0;
}

/*-----------------------------------------------------------------
 * kcopyd_jobs need to be allocated by the *clients* of kcopyd,
 * for this reason we use a mempool to prevent the client from
 * ever having to do io (which could cause a deadlock).
 *---------------------------------------------------------------*/
struct kcopyd_job {
	struct dm_kcopyd_client *kc;
	struct list_head list;
	unsigned long flags;

	/*
	 * Error state of the job.
	 */
	int read_err;
	unsigned long write_err;

	/*
	 * Either READ or WRITE
	 */
	int rw;
	struct dm_io_region source;

	/*
	 * The destinations for the transfer.
	 */
	unsigned int num_dests;
	struct dm_io_region dests[DM_KCOPYD_MAX_REGIONS];

	sector_t offset;
	unsigned int nr_pages;
	struct page_list *pages;

	/*
	 * Set this to ensure you are notified when the job has
	 * completed.  'context' is for callback to use.
	 */
	dm_kcopyd_notify_fn fn;
	void *context;

	/*
	 * These fields are only used if the job has been split
	 * into more manageable parts.
	 */
	struct mutex lock;
	atomic_t sub_jobs;
	sector_t progress;
};

/* FIXME: this should scale with the number of pages */
#define MIN_JOBS 512

static struct kmem_cache *_job_cache;

int __init dm_kcopyd_init(void)
{
	_job_cache = KMEM_CACHE(kcopyd_job, 0);
	if (!_job_cache)
		return -ENOMEM;

	return 0;
}

void dm_kcopyd_exit(void)
{
	kmem_cache_destroy(_job_cache);
	_job_cache = NULL;
}

/*
 * Functions to push and pop a job onto the head of a given job
 * list.
 */
static struct kcopyd_job *pop(struct list_head *jobs,
			      struct dm_kcopyd_client *kc)
{
	struct kcopyd_job *job = NULL;
	unsigned long flags;

	spin_lock_irqsave(&kc->job_lock, flags);

	if (!list_empty(jobs)) {
		job = list_entry(jobs->next, struct kcopyd_job, list);
		list_del(&job->list);
	}
	spin_unlock_irqrestore(&kc->job_lock, flags);

	return job;
}

static void push(struct list_head *jobs, struct kcopyd_job *job)
{
	unsigned long flags;
	struct dm_kcopyd_client *kc = job->kc;

	spin_lock_irqsave(&kc->job_lock, flags);
	list_add_tail(&job->list, jobs);
	spin_unlock_irqrestore(&kc->job_lock, flags);
}


static void push_head(struct list_head *jobs, struct kcopyd_job *job)
{
	unsigned long flags;
	struct dm_kcopyd_client *kc = job->kc;

	spin_lock_irqsave(&kc->job_lock, flags);
	list_add(&job->list, jobs);
	spin_unlock_irqrestore(&kc->job_lock, flags);
}

/*
 * These three functions process 1 item from the corresponding
 * job list.
 *
 * They return:
 * < 0: error
 *   0: success
 * > 0: can't process yet.
 */
static int run_complete_job(struct kcopyd_job *job)
{
	void *context = job->context;
	int read_err = job->read_err;
	unsigned long write_err = job->write_err;
	dm_kcopyd_notify_fn fn = job->fn;
	struct dm_kcopyd_client *kc = job->kc;

	if (job->pages)
		kcopyd_put_pages(kc, job->pages);
	mempool_free(job, kc->job_pool);
	fn(read_err, write_err, context);

	if (atomic_dec_and_test(&kc->nr_jobs))
		wake_up(&kc->destroyq);

	return 0;
}

/*
 * Unplug the block device at the specified index.
 */
static void unplug(struct dm_kcopyd_client *kc, int rw)
{
	if (kc->unplug[rw] != NULL) {
		blk_unplug(bdev_get_queue(kc->unplug[rw]));
		kc->unplug[rw] = NULL;
	}
}

/*
 * Prepare block device unplug. If there's another device
 * to be unplugged at the same array index, we unplug that
 * device first.
 */
static void prepare_unplug(struct dm_kcopyd_client *kc, int rw,
			   struct block_device *bdev)
{
	if (likely(kc->unplug[rw] == bdev))
		return;
	unplug(kc, rw);
	kc->unplug[rw] = bdev;
}

static void complete_io(unsigned long error, void *context)
{
	struct kcopyd_job *job = (struct kcopyd_job *) context;
	struct dm_kcopyd_client *kc = job->kc;

	if (error) {
		if (job->rw == WRITE)
			job->write_err |= error;
		else
			job->read_err = 1;

		if (!test_bit(DM_KCOPYD_IGNORE_ERROR, &job->flags)) {
			push(&kc->complete_jobs, job);
			wake(kc);
			return;
		}
	}

	if (job->rw == WRITE)
		push(&kc->complete_jobs, job);

	else {
		job->rw = WRITE;
		push(&kc->io_jobs, job);
	}

	wake(kc);
}

/*
 * Request io on as many buffer heads as we can currently get for
 * a particular job.
 */
static int run_io_job(struct kcopyd_job *job)
{
	int r;
	struct dm_io_request io_req = {
		.bi_rw = job->rw,
		.mem.type = DM_IO_PAGE_LIST,
		.mem.ptr.pl = job->pages,
		.mem.offset = job->offset,
		.notify.fn = complete_io,
		.notify.context = job,
		.client = job->kc->io_client,
	};

	if (job->rw == READ) {
		r = dm_io(&io_req, 1, &job->source, NULL);
		prepare_unplug(job->kc, READ, job->source.bdev);
	} else {
		if (job->num_dests > 1)
			io_req.bi_rw |= REQ_UNPLUG;
		r = dm_io(&io_req, job->num_dests, job->dests, NULL);
		if (!(io_req.bi_rw & REQ_UNPLUG))
			prepare_unplug(job->kc, WRITE, job->dests[0].bdev);
	}

	return r;
}

static int run_pages_job(struct kcopyd_job *job)
{
	int r;

	job->nr_pages = dm_div_up(job->dests[0].count + job->offset,
				  PAGE_SIZE >> 9);
	r = kcopyd_get_pages(job->kc, job->nr_pages, &job->pages);
	if (!r) {
		/* this job is ready for io */
		push(&job->kc->io_jobs, job);
		return 0;
	}

	if (r == -ENOMEM)
		/* can't complete now */
		return 1;

	return r;
}

/*
 * Run through a list for as long as possible.  Returns the count
 * of successful jobs.
 */
static int process_jobs(struct list_head *jobs, struct dm_kcopyd_client *kc,
			int (*fn) (struct kcopyd_job *))
{
	struct kcopyd_job *job;
	int r, count = 0;

	while ((job = pop(jobs, kc))) {

		r = fn(job);

		if (r < 0) {
			/* error this rogue job */
			if (job->rw == WRITE)
				job->write_err = (unsigned long) -1L;
			else
				job->read_err = 1;
			push(&kc->complete_jobs, job);
			break;
		}

		if (r > 0) {
			/*
			 * We couldn't service this job ATM, so
			 * push this job back onto the list.
			 */
			push_head(jobs, job);
			break;
		}

		count++;
	}

	return count;
}

/*
 * kcopyd does this every time it's woken up.
 */
static void do_work(struct work_struct *work)
{
	struct dm_kcopyd_client *kc = container_of(work,
					struct dm_kcopyd_client, kcopyd_work);

	/*
	 * The order that these are called is *very* important.
	 * complete jobs can free some pages for pages jobs.
	 * Pages jobs when successful will jump onto the io jobs
	 * list.  io jobs call wake when they complete and it all
	 * starts again.
	 *
	 * Note that io_jobs add block devices to the unplug array,
	 * this array is cleared with "unplug" calls. It is thus
	 * forbidden to run complete_jobs after io_jobs and before
	 * unplug because the block device could be destroyed in
	 * job completion callback.
	 */
	process_jobs(&kc->complete_jobs, kc, run_complete_job);
	process_jobs(&kc->pages_jobs, kc, run_pages_job);
	process_jobs(&kc->io_jobs, kc, run_io_job);
	unplug(kc, READ);
	unplug(kc, WRITE);
}

/*
 * If we are copying a small region we just dispatch a single job
 * to do the copy, otherwise the io has to be split up into many
 * jobs.
 */
static void dispatch_job(struct kcopyd_job *job)
{
	struct dm_kcopyd_client *kc = job->kc;
	atomic_inc(&kc->nr_jobs);
	if (unlikely(!job->source.count))
		push(&kc->complete_jobs, job);
	else
		push(&kc->pages_jobs, job);
	wake(kc);
}

#define SUB_JOB_SIZE 128
static void segment_complete(int read_err, unsigned long write_err,
			     void *context)
{
	/* FIXME: tidy this function */
	sector_t progress = 0;
	sector_t count = 0;
	struct kcopyd_job *job = (struct kcopyd_job *) context;
	struct dm_kcopyd_client *kc = job->kc;

	mutex_lock(&job->lock);

	/* update the error */
	if (read_err)
		job->read_err = 1;

	if (write_err)
		job->write_err |= write_err;

	/*
	 * Only dispatch more work if there hasn't been an error.
	 */
	if ((!job->read_err && !job->write_err) ||
	    test_bit(DM_KCOPYD_IGNORE_ERROR, &job->flags)) {
		/* get the next chunk of work */
		progress = job->progress;
		count = job->source.count - progress;
		if (count) {
			if (count > SUB_JOB_SIZE)
				count = SUB_JOB_SIZE;

			job->progress += count;
		}
	}
	mutex_unlock(&job->lock);

	if (count) {
		int i;
		struct kcopyd_job *sub_job = mempool_alloc(kc->job_pool,
							   GFP_NOIO);

		*sub_job = *job;
		sub_job->source.sector += progress;
		sub_job->source.count = count;

		for (i = 0; i < job->num_dests; i++) {
			sub_job->dests[i].sector += progress;
			sub_job->dests[i].count = count;
		}

		sub_job->fn = segment_complete;
		sub_job->context = job;
		dispatch_job(sub_job);

	} else if (atomic_dec_and_test(&job->sub_jobs)) {

		/*
		 * Queue the completion callback to the kcopyd thread.
		 *
		 * Some callers assume that all the completions are called
		 * from a single thread and don't race with each other.
		 *
		 * We must not call the callback directly here because this
		 * code may not be executing in the thread.
		 */
		push(&kc->complete_jobs, job);
		wake(kc);
	}
}

/*
 * Create some little jobs that will do the move between
 * them.
 */
#define SPLIT_COUNT 8
static void split_job(struct kcopyd_job *job)
{
	int i;

	atomic_inc(&job->kc->nr_jobs);

	atomic_set(&job->sub_jobs, SPLIT_COUNT);
	for (i = 0; i < SPLIT_COUNT; i++)
		segment_complete(0, 0u, job);
}

int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
		   unsigned int num_dests, struct dm_io_region *dests,
		   unsigned int flags, dm_kcopyd_notify_fn fn, void *context)
{
	struct kcopyd_job *job;

	/*
	 * Allocate a new job.
	 */
	job = mempool_alloc(kc->job_pool, GFP_NOIO);

	/*
	 * set up for the read.
	 */
	job->kc = kc;
	job->flags = flags;
	job->read_err = 0;
	job->write_err = 0;
	job->rw = READ;

	job->source = *from;

	job->num_dests = num_dests;
	memcpy(&job->dests, dests, sizeof(*dests) * num_dests);

	job->offset = 0;
	job->nr_pages = 0;
	job->pages = NULL;

	job->fn = fn;
	job->context = context;

	if (job->source.count < SUB_JOB_SIZE)
		dispatch_job(job);

	else {
		mutex_init(&job->lock);
		job->progress = 0;
		split_job(job);
	}

	return 0;
}
EXPORT_SYMBOL(dm_kcopyd_copy);

/*
 * Cancels a kcopyd job, eg. someone might be deactivating a
 * mirror.
 */
#if 0
int kcopyd_cancel(struct kcopyd_job *job, int block)
{
	/* FIXME: finish */
	return -1;
}
#endif  /*  0  */

/*-----------------------------------------------------------------
 * Client setup
 *---------------------------------------------------------------*/
int dm_kcopyd_client_create(unsigned int nr_pages,
			    struct dm_kcopyd_client **result)
{
	int r = -ENOMEM;
	struct dm_kcopyd_client *kc;

	kc = kmalloc(sizeof(*kc), GFP_KERNEL);
	if (!kc)
		return -ENOMEM;

	spin_lock_init(&kc->lock);
	spin_lock_init(&kc->job_lock);
	INIT_LIST_HEAD(&kc->complete_jobs);
	INIT_LIST_HEAD(&kc->io_jobs);
	INIT_LIST_HEAD(&kc->pages_jobs);

	memset(kc->unplug, 0, sizeof(kc->unplug));

	kc->job_pool = mempool_create_slab_pool(MIN_JOBS, _job_cache);
	if (!kc->job_pool)
		goto bad_slab;

	INIT_WORK(&kc->kcopyd_work, do_work);
	kc->kcopyd_wq = alloc_workqueue("kcopyd",
					WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0);
	if (!kc->kcopyd_wq)
		goto bad_workqueue;

	kc->pages = NULL;
	kc->nr_pages = kc->nr_free_pages = 0;
	r = client_alloc_pages(kc, nr_pages);
	if (r)
		goto bad_client_pages;

	kc->io_client = dm_io_client_create(nr_pages);
	if (IS_ERR(kc->io_client)) {
		r = PTR_ERR(kc->io_client);
		goto bad_io_client;
	}

	init_waitqueue_head(&kc->destroyq);
	atomic_set(&kc->nr_jobs, 0);

	*result = kc;
	return 0;

bad_io_client:
	client_free_pages(kc);
bad_client_pages:
	destroy_workqueue(kc->kcopyd_wq);
bad_workqueue:
	mempool_destroy(kc->job_pool);
bad_slab:
	kfree(kc);

	return r;
}
EXPORT_SYMBOL(dm_kcopyd_client_create);

void dm_kcopyd_client_destroy(struct dm_kcopyd_client *kc)
{
	/* Wait for completion of all jobs submitted by this client. */
	wait_event(kc->destroyq, !atomic_read(&kc->nr_jobs));

	BUG_ON(!list_empty(&kc->complete_jobs));
	BUG_ON(!list_empty(&kc->io_jobs));
	BUG_ON(!list_empty(&kc->pages_jobs));
	destroy_workqueue(kc->kcopyd_wq);
	dm_io_client_destroy(kc->io_client);
	client_free_pages(kc);
	mempool_destroy(kc->job_pool);
	kfree(kc);
}
EXPORT_SYMBOL(dm_kcopyd_client_destroy);