Loading...
1/*
2 rbd.c -- Export ceph rados objects as a Linux block device
3
4
5 based on drivers/block/osdblk.c:
6
7 Copyright 2009 Red Hat, Inc.
8
9 This program is free software; you can redistribute it and/or modify
10 it under the terms of the GNU General Public License as published by
11 the Free Software Foundation.
12
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
17
18 You should have received a copy of the GNU General Public License
19 along with this program; see the file COPYING. If not, write to
20 the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
21
22
23
24 For usage instructions, please refer to:
25
26 Documentation/ABI/testing/sysfs-bus-rbd
27
28 */
29
30#include <linux/ceph/libceph.h>
31#include <linux/ceph/osd_client.h>
32#include <linux/ceph/mon_client.h>
33#include <linux/ceph/decode.h>
34#include <linux/parser.h>
35
36#include <linux/kernel.h>
37#include <linux/device.h>
38#include <linux/module.h>
39#include <linux/fs.h>
40#include <linux/blkdev.h>
41
42#include "rbd_types.h"
43
44/*
45 * The basic unit of block I/O is a sector. It is interpreted in a
46 * number of contexts in Linux (blk, bio, genhd), but the default is
47 * universally 512 bytes. These symbols are just slightly more
48 * meaningful than the bare numbers they represent.
49 */
50#define SECTOR_SHIFT 9
51#define SECTOR_SIZE (1ULL << SECTOR_SHIFT)
52
53#define RBD_DRV_NAME "rbd"
54#define RBD_DRV_NAME_LONG "rbd (rados block device)"
55
56#define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */
57
58#define RBD_MAX_MD_NAME_LEN (RBD_MAX_OBJ_NAME_LEN + sizeof(RBD_SUFFIX))
59#define RBD_MAX_POOL_NAME_LEN 64
60#define RBD_MAX_SNAP_NAME_LEN 32
61#define RBD_MAX_OPT_LEN 1024
62
63#define RBD_SNAP_HEAD_NAME "-"
64
65/*
66 * An RBD device name will be "rbd#", where the "rbd" comes from
67 * RBD_DRV_NAME above, and # is a unique integer identifier.
68 * MAX_INT_FORMAT_WIDTH is used in ensuring DEV_NAME_LEN is big
69 * enough to hold all possible device names.
70 */
71#define DEV_NAME_LEN 32
72#define MAX_INT_FORMAT_WIDTH ((5 * sizeof (int)) / 2 + 1)
73
74#define RBD_NOTIFY_TIMEOUT_DEFAULT 10
75
76/*
77 * block device image metadata (in-memory version)
78 */
79struct rbd_image_header {
80 u64 image_size;
81 char block_name[32];
82 __u8 obj_order;
83 __u8 crypt_type;
84 __u8 comp_type;
85 struct ceph_snap_context *snapc;
86 size_t snap_names_len;
87 u64 snap_seq;
88 u32 total_snaps;
89
90 char *snap_names;
91 u64 *snap_sizes;
92
93 u64 obj_version;
94};
95
96struct rbd_options {
97 int notify_timeout;
98};
99
100/*
101 * an instance of the client. multiple devices may share an rbd client.
102 */
103struct rbd_client {
104 struct ceph_client *client;
105 struct rbd_options *rbd_opts;
106 struct kref kref;
107 struct list_head node;
108};
109
110/*
111 * a request completion status
112 */
113struct rbd_req_status {
114 int done;
115 int rc;
116 u64 bytes;
117};
118
119/*
120 * a collection of requests
121 */
122struct rbd_req_coll {
123 int total;
124 int num_done;
125 struct kref kref;
126 struct rbd_req_status status[0];
127};
128
129/*
130 * a single io request
131 */
132struct rbd_request {
133 struct request *rq; /* blk layer request */
134 struct bio *bio; /* cloned bio */
135 struct page **pages; /* list of used pages */
136 u64 len;
137 int coll_index;
138 struct rbd_req_coll *coll;
139};
140
141struct rbd_snap {
142 struct device dev;
143 const char *name;
144 u64 size;
145 struct list_head node;
146 u64 id;
147};
148
149/*
150 * a single device
151 */
152struct rbd_device {
153 int id; /* blkdev unique id */
154
155 int major; /* blkdev assigned major */
156 struct gendisk *disk; /* blkdev's gendisk and rq */
157 struct request_queue *q;
158
159 struct rbd_client *rbd_client;
160
161 char name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */
162
163 spinlock_t lock; /* queue lock */
164
165 struct rbd_image_header header;
166 char obj[RBD_MAX_OBJ_NAME_LEN]; /* rbd image name */
167 int obj_len;
168 char obj_md_name[RBD_MAX_MD_NAME_LEN]; /* hdr nm. */
169 char pool_name[RBD_MAX_POOL_NAME_LEN];
170 int poolid;
171
172 struct ceph_osd_event *watch_event;
173 struct ceph_osd_request *watch_request;
174
175 /* protects updating the header */
176 struct rw_semaphore header_rwsem;
177 char snap_name[RBD_MAX_SNAP_NAME_LEN];
178 u64 snap_id; /* current snapshot id */
179 int read_only;
180
181 struct list_head node;
182
183 /* list of snapshots */
184 struct list_head snaps;
185
186 /* sysfs related */
187 struct device dev;
188};
189
190static DEFINE_MUTEX(ctl_mutex); /* Serialize open/close/setup/teardown */
191
192static LIST_HEAD(rbd_dev_list); /* devices */
193static DEFINE_SPINLOCK(rbd_dev_list_lock);
194
195static LIST_HEAD(rbd_client_list); /* clients */
196static DEFINE_SPINLOCK(rbd_client_list_lock);
197
198static int __rbd_init_snaps_header(struct rbd_device *rbd_dev);
199static void rbd_dev_release(struct device *dev);
200static ssize_t rbd_snap_add(struct device *dev,
201 struct device_attribute *attr,
202 const char *buf,
203 size_t count);
204static void __rbd_remove_snap_dev(struct rbd_device *rbd_dev,
205 struct rbd_snap *snap);
206
207static ssize_t rbd_add(struct bus_type *bus, const char *buf,
208 size_t count);
209static ssize_t rbd_remove(struct bus_type *bus, const char *buf,
210 size_t count);
211
212static struct bus_attribute rbd_bus_attrs[] = {
213 __ATTR(add, S_IWUSR, NULL, rbd_add),
214 __ATTR(remove, S_IWUSR, NULL, rbd_remove),
215 __ATTR_NULL
216};
217
218static struct bus_type rbd_bus_type = {
219 .name = "rbd",
220 .bus_attrs = rbd_bus_attrs,
221};
222
223static void rbd_root_dev_release(struct device *dev)
224{
225}
226
227static struct device rbd_root_dev = {
228 .init_name = "rbd",
229 .release = rbd_root_dev_release,
230};
231
232
233static struct device *rbd_get_dev(struct rbd_device *rbd_dev)
234{
235 return get_device(&rbd_dev->dev);
236}
237
238static void rbd_put_dev(struct rbd_device *rbd_dev)
239{
240 put_device(&rbd_dev->dev);
241}
242
243static int __rbd_refresh_header(struct rbd_device *rbd_dev);
244
245static int rbd_open(struct block_device *bdev, fmode_t mode)
246{
247 struct rbd_device *rbd_dev = bdev->bd_disk->private_data;
248
249 rbd_get_dev(rbd_dev);
250
251 set_device_ro(bdev, rbd_dev->read_only);
252
253 if ((mode & FMODE_WRITE) && rbd_dev->read_only)
254 return -EROFS;
255
256 return 0;
257}
258
259static int rbd_release(struct gendisk *disk, fmode_t mode)
260{
261 struct rbd_device *rbd_dev = disk->private_data;
262
263 rbd_put_dev(rbd_dev);
264
265 return 0;
266}
267
268static const struct block_device_operations rbd_bd_ops = {
269 .owner = THIS_MODULE,
270 .open = rbd_open,
271 .release = rbd_release,
272};
273
274/*
275 * Initialize an rbd client instance.
276 * We own *opt.
277 */
278static struct rbd_client *rbd_client_create(struct ceph_options *opt,
279 struct rbd_options *rbd_opts)
280{
281 struct rbd_client *rbdc;
282 int ret = -ENOMEM;
283
284 dout("rbd_client_create\n");
285 rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL);
286 if (!rbdc)
287 goto out_opt;
288
289 kref_init(&rbdc->kref);
290 INIT_LIST_HEAD(&rbdc->node);
291
292 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
293
294 rbdc->client = ceph_create_client(opt, rbdc, 0, 0);
295 if (IS_ERR(rbdc->client))
296 goto out_mutex;
297 opt = NULL; /* Now rbdc->client is responsible for opt */
298
299 ret = ceph_open_session(rbdc->client);
300 if (ret < 0)
301 goto out_err;
302
303 rbdc->rbd_opts = rbd_opts;
304
305 spin_lock(&rbd_client_list_lock);
306 list_add_tail(&rbdc->node, &rbd_client_list);
307 spin_unlock(&rbd_client_list_lock);
308
309 mutex_unlock(&ctl_mutex);
310
311 dout("rbd_client_create created %p\n", rbdc);
312 return rbdc;
313
314out_err:
315 ceph_destroy_client(rbdc->client);
316out_mutex:
317 mutex_unlock(&ctl_mutex);
318 kfree(rbdc);
319out_opt:
320 if (opt)
321 ceph_destroy_options(opt);
322 return ERR_PTR(ret);
323}
324
325/*
326 * Find a ceph client with specific addr and configuration.
327 */
328static struct rbd_client *__rbd_client_find(struct ceph_options *opt)
329{
330 struct rbd_client *client_node;
331
332 if (opt->flags & CEPH_OPT_NOSHARE)
333 return NULL;
334
335 list_for_each_entry(client_node, &rbd_client_list, node)
336 if (ceph_compare_options(opt, client_node->client) == 0)
337 return client_node;
338 return NULL;
339}
340
341/*
342 * mount options
343 */
344enum {
345 Opt_notify_timeout,
346 Opt_last_int,
347 /* int args above */
348 Opt_last_string,
349 /* string args above */
350};
351
352static match_table_t rbdopt_tokens = {
353 {Opt_notify_timeout, "notify_timeout=%d"},
354 /* int args above */
355 /* string args above */
356 {-1, NULL}
357};
358
359static int parse_rbd_opts_token(char *c, void *private)
360{
361 struct rbd_options *rbdopt = private;
362 substring_t argstr[MAX_OPT_ARGS];
363 int token, intval, ret;
364
365 token = match_token(c, rbdopt_tokens, argstr);
366 if (token < 0)
367 return -EINVAL;
368
369 if (token < Opt_last_int) {
370 ret = match_int(&argstr[0], &intval);
371 if (ret < 0) {
372 pr_err("bad mount option arg (not int) "
373 "at '%s'\n", c);
374 return ret;
375 }
376 dout("got int token %d val %d\n", token, intval);
377 } else if (token > Opt_last_int && token < Opt_last_string) {
378 dout("got string token %d val %s\n", token,
379 argstr[0].from);
380 } else {
381 dout("got token %d\n", token);
382 }
383
384 switch (token) {
385 case Opt_notify_timeout:
386 rbdopt->notify_timeout = intval;
387 break;
388 default:
389 BUG_ON(token);
390 }
391 return 0;
392}
393
394/*
395 * Get a ceph client with specific addr and configuration, if one does
396 * not exist create it.
397 */
398static struct rbd_client *rbd_get_client(const char *mon_addr,
399 size_t mon_addr_len,
400 char *options)
401{
402 struct rbd_client *rbdc;
403 struct ceph_options *opt;
404 struct rbd_options *rbd_opts;
405
406 rbd_opts = kzalloc(sizeof(*rbd_opts), GFP_KERNEL);
407 if (!rbd_opts)
408 return ERR_PTR(-ENOMEM);
409
410 rbd_opts->notify_timeout = RBD_NOTIFY_TIMEOUT_DEFAULT;
411
412 opt = ceph_parse_options(options, mon_addr,
413 mon_addr + mon_addr_len,
414 parse_rbd_opts_token, rbd_opts);
415 if (IS_ERR(opt)) {
416 kfree(rbd_opts);
417 return ERR_CAST(opt);
418 }
419
420 spin_lock(&rbd_client_list_lock);
421 rbdc = __rbd_client_find(opt);
422 if (rbdc) {
423 /* using an existing client */
424 kref_get(&rbdc->kref);
425 spin_unlock(&rbd_client_list_lock);
426
427 ceph_destroy_options(opt);
428 kfree(rbd_opts);
429
430 return rbdc;
431 }
432 spin_unlock(&rbd_client_list_lock);
433
434 rbdc = rbd_client_create(opt, rbd_opts);
435
436 if (IS_ERR(rbdc))
437 kfree(rbd_opts);
438
439 return rbdc;
440}
441
442/*
443 * Destroy ceph client
444 *
445 * Caller must hold rbd_client_list_lock.
446 */
447static void rbd_client_release(struct kref *kref)
448{
449 struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref);
450
451 dout("rbd_release_client %p\n", rbdc);
452 spin_lock(&rbd_client_list_lock);
453 list_del(&rbdc->node);
454 spin_unlock(&rbd_client_list_lock);
455
456 ceph_destroy_client(rbdc->client);
457 kfree(rbdc->rbd_opts);
458 kfree(rbdc);
459}
460
461/*
462 * Drop reference to ceph client node. If it's not referenced anymore, release
463 * it.
464 */
465static void rbd_put_client(struct rbd_device *rbd_dev)
466{
467 kref_put(&rbd_dev->rbd_client->kref, rbd_client_release);
468 rbd_dev->rbd_client = NULL;
469}
470
471/*
472 * Destroy requests collection
473 */
474static void rbd_coll_release(struct kref *kref)
475{
476 struct rbd_req_coll *coll =
477 container_of(kref, struct rbd_req_coll, kref);
478
479 dout("rbd_coll_release %p\n", coll);
480 kfree(coll);
481}
482
483/*
484 * Create a new header structure, translate header format from the on-disk
485 * header.
486 */
487static int rbd_header_from_disk(struct rbd_image_header *header,
488 struct rbd_image_header_ondisk *ondisk,
489 u32 allocated_snaps,
490 gfp_t gfp_flags)
491{
492 u32 i, snap_count;
493
494 if (memcmp(ondisk, RBD_HEADER_TEXT, sizeof(RBD_HEADER_TEXT)))
495 return -ENXIO;
496
497 snap_count = le32_to_cpu(ondisk->snap_count);
498 if (snap_count > (UINT_MAX - sizeof(struct ceph_snap_context))
499 / sizeof (*ondisk))
500 return -EINVAL;
501 header->snapc = kmalloc(sizeof(struct ceph_snap_context) +
502 snap_count * sizeof(u64),
503 gfp_flags);
504 if (!header->snapc)
505 return -ENOMEM;
506
507 header->snap_names_len = le64_to_cpu(ondisk->snap_names_len);
508 if (snap_count) {
509 header->snap_names = kmalloc(header->snap_names_len,
510 gfp_flags);
511 if (!header->snap_names)
512 goto err_snapc;
513 header->snap_sizes = kmalloc(snap_count * sizeof(u64),
514 gfp_flags);
515 if (!header->snap_sizes)
516 goto err_names;
517 } else {
518 header->snap_names = NULL;
519 header->snap_sizes = NULL;
520 }
521 memcpy(header->block_name, ondisk->block_name,
522 sizeof(ondisk->block_name));
523
524 header->image_size = le64_to_cpu(ondisk->image_size);
525 header->obj_order = ondisk->options.order;
526 header->crypt_type = ondisk->options.crypt_type;
527 header->comp_type = ondisk->options.comp_type;
528
529 atomic_set(&header->snapc->nref, 1);
530 header->snap_seq = le64_to_cpu(ondisk->snap_seq);
531 header->snapc->num_snaps = snap_count;
532 header->total_snaps = snap_count;
533
534 if (snap_count && allocated_snaps == snap_count) {
535 for (i = 0; i < snap_count; i++) {
536 header->snapc->snaps[i] =
537 le64_to_cpu(ondisk->snaps[i].id);
538 header->snap_sizes[i] =
539 le64_to_cpu(ondisk->snaps[i].image_size);
540 }
541
542 /* copy snapshot names */
543 memcpy(header->snap_names, &ondisk->snaps[i],
544 header->snap_names_len);
545 }
546
547 return 0;
548
549err_names:
550 kfree(header->snap_names);
551err_snapc:
552 kfree(header->snapc);
553 return -ENOMEM;
554}
555
556static int snap_by_name(struct rbd_image_header *header, const char *snap_name,
557 u64 *seq, u64 *size)
558{
559 int i;
560 char *p = header->snap_names;
561
562 for (i = 0; i < header->total_snaps; i++) {
563 if (!strcmp(snap_name, p)) {
564
565 /* Found it. Pass back its id and/or size */
566
567 if (seq)
568 *seq = header->snapc->snaps[i];
569 if (size)
570 *size = header->snap_sizes[i];
571 return i;
572 }
573 p += strlen(p) + 1; /* Skip ahead to the next name */
574 }
575 return -ENOENT;
576}
577
578static int rbd_header_set_snap(struct rbd_device *dev, u64 *size)
579{
580 struct rbd_image_header *header = &dev->header;
581 struct ceph_snap_context *snapc = header->snapc;
582 int ret = -ENOENT;
583
584 BUILD_BUG_ON(sizeof (dev->snap_name) < sizeof (RBD_SNAP_HEAD_NAME));
585
586 down_write(&dev->header_rwsem);
587
588 if (!memcmp(dev->snap_name, RBD_SNAP_HEAD_NAME,
589 sizeof (RBD_SNAP_HEAD_NAME))) {
590 if (header->total_snaps)
591 snapc->seq = header->snap_seq;
592 else
593 snapc->seq = 0;
594 dev->snap_id = CEPH_NOSNAP;
595 dev->read_only = 0;
596 if (size)
597 *size = header->image_size;
598 } else {
599 ret = snap_by_name(header, dev->snap_name, &snapc->seq, size);
600 if (ret < 0)
601 goto done;
602 dev->snap_id = snapc->seq;
603 dev->read_only = 1;
604 }
605
606 ret = 0;
607done:
608 up_write(&dev->header_rwsem);
609 return ret;
610}
611
612static void rbd_header_free(struct rbd_image_header *header)
613{
614 kfree(header->snapc);
615 kfree(header->snap_names);
616 kfree(header->snap_sizes);
617}
618
619/*
620 * get the actual striped segment name, offset and length
621 */
622static u64 rbd_get_segment(struct rbd_image_header *header,
623 const char *block_name,
624 u64 ofs, u64 len,
625 char *seg_name, u64 *segofs)
626{
627 u64 seg = ofs >> header->obj_order;
628
629 if (seg_name)
630 snprintf(seg_name, RBD_MAX_SEG_NAME_LEN,
631 "%s.%012llx", block_name, seg);
632
633 ofs = ofs & ((1 << header->obj_order) - 1);
634 len = min_t(u64, len, (1 << header->obj_order) - ofs);
635
636 if (segofs)
637 *segofs = ofs;
638
639 return len;
640}
641
642static int rbd_get_num_segments(struct rbd_image_header *header,
643 u64 ofs, u64 len)
644{
645 u64 start_seg = ofs >> header->obj_order;
646 u64 end_seg = (ofs + len - 1) >> header->obj_order;
647 return end_seg - start_seg + 1;
648}
649
650/*
651 * returns the size of an object in the image
652 */
653static u64 rbd_obj_bytes(struct rbd_image_header *header)
654{
655 return 1 << header->obj_order;
656}
657
658/*
659 * bio helpers
660 */
661
662static void bio_chain_put(struct bio *chain)
663{
664 struct bio *tmp;
665
666 while (chain) {
667 tmp = chain;
668 chain = chain->bi_next;
669 bio_put(tmp);
670 }
671}
672
673/*
674 * zeros a bio chain, starting at specific offset
675 */
676static void zero_bio_chain(struct bio *chain, int start_ofs)
677{
678 struct bio_vec *bv;
679 unsigned long flags;
680 void *buf;
681 int i;
682 int pos = 0;
683
684 while (chain) {
685 bio_for_each_segment(bv, chain, i) {
686 if (pos + bv->bv_len > start_ofs) {
687 int remainder = max(start_ofs - pos, 0);
688 buf = bvec_kmap_irq(bv, &flags);
689 memset(buf + remainder, 0,
690 bv->bv_len - remainder);
691 bvec_kunmap_irq(buf, &flags);
692 }
693 pos += bv->bv_len;
694 }
695
696 chain = chain->bi_next;
697 }
698}
699
700/*
701 * bio_chain_clone - clone a chain of bios up to a certain length.
702 * might return a bio_pair that will need to be released.
703 */
704static struct bio *bio_chain_clone(struct bio **old, struct bio **next,
705 struct bio_pair **bp,
706 int len, gfp_t gfpmask)
707{
708 struct bio *tmp, *old_chain = *old, *new_chain = NULL, *tail = NULL;
709 int total = 0;
710
711 if (*bp) {
712 bio_pair_release(*bp);
713 *bp = NULL;
714 }
715
716 while (old_chain && (total < len)) {
717 tmp = bio_kmalloc(gfpmask, old_chain->bi_max_vecs);
718 if (!tmp)
719 goto err_out;
720
721 if (total + old_chain->bi_size > len) {
722 struct bio_pair *bp;
723
724 /*
725 * this split can only happen with a single paged bio,
726 * split_bio will BUG_ON if this is not the case
727 */
728 dout("bio_chain_clone split! total=%d remaining=%d"
729 "bi_size=%d\n",
730 (int)total, (int)len-total,
731 (int)old_chain->bi_size);
732
733 /* split the bio. We'll release it either in the next
734 call, or it will have to be released outside */
735 bp = bio_split(old_chain, (len - total) / SECTOR_SIZE);
736 if (!bp)
737 goto err_out;
738
739 __bio_clone(tmp, &bp->bio1);
740
741 *next = &bp->bio2;
742 } else {
743 __bio_clone(tmp, old_chain);
744 *next = old_chain->bi_next;
745 }
746
747 tmp->bi_bdev = NULL;
748 gfpmask &= ~__GFP_WAIT;
749 tmp->bi_next = NULL;
750
751 if (!new_chain) {
752 new_chain = tail = tmp;
753 } else {
754 tail->bi_next = tmp;
755 tail = tmp;
756 }
757 old_chain = old_chain->bi_next;
758
759 total += tmp->bi_size;
760 }
761
762 BUG_ON(total < len);
763
764 if (tail)
765 tail->bi_next = NULL;
766
767 *old = old_chain;
768
769 return new_chain;
770
771err_out:
772 dout("bio_chain_clone with err\n");
773 bio_chain_put(new_chain);
774 return NULL;
775}
776
777/*
778 * helpers for osd request op vectors.
779 */
780static int rbd_create_rw_ops(struct ceph_osd_req_op **ops,
781 int num_ops,
782 int opcode,
783 u32 payload_len)
784{
785 *ops = kzalloc(sizeof(struct ceph_osd_req_op) * (num_ops + 1),
786 GFP_NOIO);
787 if (!*ops)
788 return -ENOMEM;
789 (*ops)[0].op = opcode;
790 /*
791 * op extent offset and length will be set later on
792 * in calc_raw_layout()
793 */
794 (*ops)[0].payload_len = payload_len;
795 return 0;
796}
797
798static void rbd_destroy_ops(struct ceph_osd_req_op *ops)
799{
800 kfree(ops);
801}
802
803static void rbd_coll_end_req_index(struct request *rq,
804 struct rbd_req_coll *coll,
805 int index,
806 int ret, u64 len)
807{
808 struct request_queue *q;
809 int min, max, i;
810
811 dout("rbd_coll_end_req_index %p index %d ret %d len %lld\n",
812 coll, index, ret, len);
813
814 if (!rq)
815 return;
816
817 if (!coll) {
818 blk_end_request(rq, ret, len);
819 return;
820 }
821
822 q = rq->q;
823
824 spin_lock_irq(q->queue_lock);
825 coll->status[index].done = 1;
826 coll->status[index].rc = ret;
827 coll->status[index].bytes = len;
828 max = min = coll->num_done;
829 while (max < coll->total && coll->status[max].done)
830 max++;
831
832 for (i = min; i<max; i++) {
833 __blk_end_request(rq, coll->status[i].rc,
834 coll->status[i].bytes);
835 coll->num_done++;
836 kref_put(&coll->kref, rbd_coll_release);
837 }
838 spin_unlock_irq(q->queue_lock);
839}
840
841static void rbd_coll_end_req(struct rbd_request *req,
842 int ret, u64 len)
843{
844 rbd_coll_end_req_index(req->rq, req->coll, req->coll_index, ret, len);
845}
846
847/*
848 * Send ceph osd request
849 */
850static int rbd_do_request(struct request *rq,
851 struct rbd_device *dev,
852 struct ceph_snap_context *snapc,
853 u64 snapid,
854 const char *obj, u64 ofs, u64 len,
855 struct bio *bio,
856 struct page **pages,
857 int num_pages,
858 int flags,
859 struct ceph_osd_req_op *ops,
860 int num_reply,
861 struct rbd_req_coll *coll,
862 int coll_index,
863 void (*rbd_cb)(struct ceph_osd_request *req,
864 struct ceph_msg *msg),
865 struct ceph_osd_request **linger_req,
866 u64 *ver)
867{
868 struct ceph_osd_request *req;
869 struct ceph_file_layout *layout;
870 int ret;
871 u64 bno;
872 struct timespec mtime = CURRENT_TIME;
873 struct rbd_request *req_data;
874 struct ceph_osd_request_head *reqhead;
875 struct ceph_osd_client *osdc;
876
877 req_data = kzalloc(sizeof(*req_data), GFP_NOIO);
878 if (!req_data) {
879 if (coll)
880 rbd_coll_end_req_index(rq, coll, coll_index,
881 -ENOMEM, len);
882 return -ENOMEM;
883 }
884
885 if (coll) {
886 req_data->coll = coll;
887 req_data->coll_index = coll_index;
888 }
889
890 dout("rbd_do_request obj=%s ofs=%lld len=%lld\n", obj, len, ofs);
891
892 down_read(&dev->header_rwsem);
893
894 osdc = &dev->rbd_client->client->osdc;
895 req = ceph_osdc_alloc_request(osdc, flags, snapc, ops,
896 false, GFP_NOIO, pages, bio);
897 if (!req) {
898 up_read(&dev->header_rwsem);
899 ret = -ENOMEM;
900 goto done_pages;
901 }
902
903 req->r_callback = rbd_cb;
904
905 req_data->rq = rq;
906 req_data->bio = bio;
907 req_data->pages = pages;
908 req_data->len = len;
909
910 req->r_priv = req_data;
911
912 reqhead = req->r_request->front.iov_base;
913 reqhead->snapid = cpu_to_le64(CEPH_NOSNAP);
914
915 strncpy(req->r_oid, obj, sizeof(req->r_oid));
916 req->r_oid_len = strlen(req->r_oid);
917
918 layout = &req->r_file_layout;
919 memset(layout, 0, sizeof(*layout));
920 layout->fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
921 layout->fl_stripe_count = cpu_to_le32(1);
922 layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
923 layout->fl_pg_pool = cpu_to_le32(dev->poolid);
924 ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
925 req, ops);
926
927 ceph_osdc_build_request(req, ofs, &len,
928 ops,
929 snapc,
930 &mtime,
931 req->r_oid, req->r_oid_len);
932 up_read(&dev->header_rwsem);
933
934 if (linger_req) {
935 ceph_osdc_set_request_linger(osdc, req);
936 *linger_req = req;
937 }
938
939 ret = ceph_osdc_start_request(osdc, req, false);
940 if (ret < 0)
941 goto done_err;
942
943 if (!rbd_cb) {
944 ret = ceph_osdc_wait_request(osdc, req);
945 if (ver)
946 *ver = le64_to_cpu(req->r_reassert_version.version);
947 dout("reassert_ver=%lld\n",
948 le64_to_cpu(req->r_reassert_version.version));
949 ceph_osdc_put_request(req);
950 }
951 return ret;
952
953done_err:
954 bio_chain_put(req_data->bio);
955 ceph_osdc_put_request(req);
956done_pages:
957 rbd_coll_end_req(req_data, ret, len);
958 kfree(req_data);
959 return ret;
960}
961
962/*
963 * Ceph osd op callback
964 */
965static void rbd_req_cb(struct ceph_osd_request *req, struct ceph_msg *msg)
966{
967 struct rbd_request *req_data = req->r_priv;
968 struct ceph_osd_reply_head *replyhead;
969 struct ceph_osd_op *op;
970 __s32 rc;
971 u64 bytes;
972 int read_op;
973
974 /* parse reply */
975 replyhead = msg->front.iov_base;
976 WARN_ON(le32_to_cpu(replyhead->num_ops) == 0);
977 op = (void *)(replyhead + 1);
978 rc = le32_to_cpu(replyhead->result);
979 bytes = le64_to_cpu(op->extent.length);
980 read_op = (le16_to_cpu(op->op) == CEPH_OSD_OP_READ);
981
982 dout("rbd_req_cb bytes=%lld readop=%d rc=%d\n", bytes, read_op, rc);
983
984 if (rc == -ENOENT && read_op) {
985 zero_bio_chain(req_data->bio, 0);
986 rc = 0;
987 } else if (rc == 0 && read_op && bytes < req_data->len) {
988 zero_bio_chain(req_data->bio, bytes);
989 bytes = req_data->len;
990 }
991
992 rbd_coll_end_req(req_data, rc, bytes);
993
994 if (req_data->bio)
995 bio_chain_put(req_data->bio);
996
997 ceph_osdc_put_request(req);
998 kfree(req_data);
999}
1000
1001static void rbd_simple_req_cb(struct ceph_osd_request *req, struct ceph_msg *msg)
1002{
1003 ceph_osdc_put_request(req);
1004}
1005
1006/*
1007 * Do a synchronous ceph osd operation
1008 */
1009static int rbd_req_sync_op(struct rbd_device *dev,
1010 struct ceph_snap_context *snapc,
1011 u64 snapid,
1012 int opcode,
1013 int flags,
1014 struct ceph_osd_req_op *orig_ops,
1015 int num_reply,
1016 const char *obj,
1017 u64 ofs, u64 len,
1018 char *buf,
1019 struct ceph_osd_request **linger_req,
1020 u64 *ver)
1021{
1022 int ret;
1023 struct page **pages;
1024 int num_pages;
1025 struct ceph_osd_req_op *ops = orig_ops;
1026 u32 payload_len;
1027
1028 num_pages = calc_pages_for(ofs , len);
1029 pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
1030 if (IS_ERR(pages))
1031 return PTR_ERR(pages);
1032
1033 if (!orig_ops) {
1034 payload_len = (flags & CEPH_OSD_FLAG_WRITE ? len : 0);
1035 ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len);
1036 if (ret < 0)
1037 goto done;
1038
1039 if ((flags & CEPH_OSD_FLAG_WRITE) && buf) {
1040 ret = ceph_copy_to_page_vector(pages, buf, ofs, len);
1041 if (ret < 0)
1042 goto done_ops;
1043 }
1044 }
1045
1046 ret = rbd_do_request(NULL, dev, snapc, snapid,
1047 obj, ofs, len, NULL,
1048 pages, num_pages,
1049 flags,
1050 ops,
1051 2,
1052 NULL, 0,
1053 NULL,
1054 linger_req, ver);
1055 if (ret < 0)
1056 goto done_ops;
1057
1058 if ((flags & CEPH_OSD_FLAG_READ) && buf)
1059 ret = ceph_copy_from_page_vector(pages, buf, ofs, ret);
1060
1061done_ops:
1062 if (!orig_ops)
1063 rbd_destroy_ops(ops);
1064done:
1065 ceph_release_page_vector(pages, num_pages);
1066 return ret;
1067}
1068
1069/*
1070 * Do an asynchronous ceph osd operation
1071 */
1072static int rbd_do_op(struct request *rq,
1073 struct rbd_device *rbd_dev ,
1074 struct ceph_snap_context *snapc,
1075 u64 snapid,
1076 int opcode, int flags, int num_reply,
1077 u64 ofs, u64 len,
1078 struct bio *bio,
1079 struct rbd_req_coll *coll,
1080 int coll_index)
1081{
1082 char *seg_name;
1083 u64 seg_ofs;
1084 u64 seg_len;
1085 int ret;
1086 struct ceph_osd_req_op *ops;
1087 u32 payload_len;
1088
1089 seg_name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO);
1090 if (!seg_name)
1091 return -ENOMEM;
1092
1093 seg_len = rbd_get_segment(&rbd_dev->header,
1094 rbd_dev->header.block_name,
1095 ofs, len,
1096 seg_name, &seg_ofs);
1097
1098 payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0);
1099
1100 ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len);
1101 if (ret < 0)
1102 goto done;
1103
1104 /* we've taken care of segment sizes earlier when we
1105 cloned the bios. We should never have a segment
1106 truncated at this point */
1107 BUG_ON(seg_len < len);
1108
1109 ret = rbd_do_request(rq, rbd_dev, snapc, snapid,
1110 seg_name, seg_ofs, seg_len,
1111 bio,
1112 NULL, 0,
1113 flags,
1114 ops,
1115 num_reply,
1116 coll, coll_index,
1117 rbd_req_cb, 0, NULL);
1118
1119 rbd_destroy_ops(ops);
1120done:
1121 kfree(seg_name);
1122 return ret;
1123}
1124
1125/*
1126 * Request async osd write
1127 */
1128static int rbd_req_write(struct request *rq,
1129 struct rbd_device *rbd_dev,
1130 struct ceph_snap_context *snapc,
1131 u64 ofs, u64 len,
1132 struct bio *bio,
1133 struct rbd_req_coll *coll,
1134 int coll_index)
1135{
1136 return rbd_do_op(rq, rbd_dev, snapc, CEPH_NOSNAP,
1137 CEPH_OSD_OP_WRITE,
1138 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
1139 2,
1140 ofs, len, bio, coll, coll_index);
1141}
1142
1143/*
1144 * Request async osd read
1145 */
1146static int rbd_req_read(struct request *rq,
1147 struct rbd_device *rbd_dev,
1148 u64 snapid,
1149 u64 ofs, u64 len,
1150 struct bio *bio,
1151 struct rbd_req_coll *coll,
1152 int coll_index)
1153{
1154 return rbd_do_op(rq, rbd_dev, NULL,
1155 snapid,
1156 CEPH_OSD_OP_READ,
1157 CEPH_OSD_FLAG_READ,
1158 2,
1159 ofs, len, bio, coll, coll_index);
1160}
1161
1162/*
1163 * Request sync osd read
1164 */
1165static int rbd_req_sync_read(struct rbd_device *dev,
1166 struct ceph_snap_context *snapc,
1167 u64 snapid,
1168 const char *obj,
1169 u64 ofs, u64 len,
1170 char *buf,
1171 u64 *ver)
1172{
1173 return rbd_req_sync_op(dev, NULL,
1174 snapid,
1175 CEPH_OSD_OP_READ,
1176 CEPH_OSD_FLAG_READ,
1177 NULL,
1178 1, obj, ofs, len, buf, NULL, ver);
1179}
1180
1181/*
1182 * Request sync osd watch
1183 */
1184static int rbd_req_sync_notify_ack(struct rbd_device *dev,
1185 u64 ver,
1186 u64 notify_id,
1187 const char *obj)
1188{
1189 struct ceph_osd_req_op *ops;
1190 struct page **pages = NULL;
1191 int ret;
1192
1193 ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_NOTIFY_ACK, 0);
1194 if (ret < 0)
1195 return ret;
1196
1197 ops[0].watch.ver = cpu_to_le64(dev->header.obj_version);
1198 ops[0].watch.cookie = notify_id;
1199 ops[0].watch.flag = 0;
1200
1201 ret = rbd_do_request(NULL, dev, NULL, CEPH_NOSNAP,
1202 obj, 0, 0, NULL,
1203 pages, 0,
1204 CEPH_OSD_FLAG_READ,
1205 ops,
1206 1,
1207 NULL, 0,
1208 rbd_simple_req_cb, 0, NULL);
1209
1210 rbd_destroy_ops(ops);
1211 return ret;
1212}
1213
1214static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
1215{
1216 struct rbd_device *dev = (struct rbd_device *)data;
1217 int rc;
1218
1219 if (!dev)
1220 return;
1221
1222 dout("rbd_watch_cb %s notify_id=%lld opcode=%d\n", dev->obj_md_name,
1223 notify_id, (int)opcode);
1224 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
1225 rc = __rbd_refresh_header(dev);
1226 mutex_unlock(&ctl_mutex);
1227 if (rc)
1228 pr_warning(RBD_DRV_NAME "%d got notification but failed to "
1229 " update snaps: %d\n", dev->major, rc);
1230
1231 rbd_req_sync_notify_ack(dev, ver, notify_id, dev->obj_md_name);
1232}
1233
1234/*
1235 * Request sync osd watch
1236 */
1237static int rbd_req_sync_watch(struct rbd_device *dev,
1238 const char *obj,
1239 u64 ver)
1240{
1241 struct ceph_osd_req_op *ops;
1242 struct ceph_osd_client *osdc = &dev->rbd_client->client->osdc;
1243
1244 int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_WATCH, 0);
1245 if (ret < 0)
1246 return ret;
1247
1248 ret = ceph_osdc_create_event(osdc, rbd_watch_cb, 0,
1249 (void *)dev, &dev->watch_event);
1250 if (ret < 0)
1251 goto fail;
1252
1253 ops[0].watch.ver = cpu_to_le64(ver);
1254 ops[0].watch.cookie = cpu_to_le64(dev->watch_event->cookie);
1255 ops[0].watch.flag = 1;
1256
1257 ret = rbd_req_sync_op(dev, NULL,
1258 CEPH_NOSNAP,
1259 0,
1260 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
1261 ops,
1262 1, obj, 0, 0, NULL,
1263 &dev->watch_request, NULL);
1264
1265 if (ret < 0)
1266 goto fail_event;
1267
1268 rbd_destroy_ops(ops);
1269 return 0;
1270
1271fail_event:
1272 ceph_osdc_cancel_event(dev->watch_event);
1273 dev->watch_event = NULL;
1274fail:
1275 rbd_destroy_ops(ops);
1276 return ret;
1277}
1278
1279/*
1280 * Request sync osd unwatch
1281 */
1282static int rbd_req_sync_unwatch(struct rbd_device *dev,
1283 const char *obj)
1284{
1285 struct ceph_osd_req_op *ops;
1286
1287 int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_WATCH, 0);
1288 if (ret < 0)
1289 return ret;
1290
1291 ops[0].watch.ver = 0;
1292 ops[0].watch.cookie = cpu_to_le64(dev->watch_event->cookie);
1293 ops[0].watch.flag = 0;
1294
1295 ret = rbd_req_sync_op(dev, NULL,
1296 CEPH_NOSNAP,
1297 0,
1298 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
1299 ops,
1300 1, obj, 0, 0, NULL, NULL, NULL);
1301
1302 rbd_destroy_ops(ops);
1303 ceph_osdc_cancel_event(dev->watch_event);
1304 dev->watch_event = NULL;
1305 return ret;
1306}
1307
1308struct rbd_notify_info {
1309 struct rbd_device *dev;
1310};
1311
1312static void rbd_notify_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
1313{
1314 struct rbd_device *dev = (struct rbd_device *)data;
1315 if (!dev)
1316 return;
1317
1318 dout("rbd_notify_cb %s notify_id=%lld opcode=%d\n", dev->obj_md_name,
1319 notify_id, (int)opcode);
1320}
1321
1322/*
1323 * Request sync osd notify
1324 */
1325static int rbd_req_sync_notify(struct rbd_device *dev,
1326 const char *obj)
1327{
1328 struct ceph_osd_req_op *ops;
1329 struct ceph_osd_client *osdc = &dev->rbd_client->client->osdc;
1330 struct ceph_osd_event *event;
1331 struct rbd_notify_info info;
1332 int payload_len = sizeof(u32) + sizeof(u32);
1333 int ret;
1334
1335 ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_NOTIFY, payload_len);
1336 if (ret < 0)
1337 return ret;
1338
1339 info.dev = dev;
1340
1341 ret = ceph_osdc_create_event(osdc, rbd_notify_cb, 1,
1342 (void *)&info, &event);
1343 if (ret < 0)
1344 goto fail;
1345
1346 ops[0].watch.ver = 1;
1347 ops[0].watch.flag = 1;
1348 ops[0].watch.cookie = event->cookie;
1349 ops[0].watch.prot_ver = RADOS_NOTIFY_VER;
1350 ops[0].watch.timeout = 12;
1351
1352 ret = rbd_req_sync_op(dev, NULL,
1353 CEPH_NOSNAP,
1354 0,
1355 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
1356 ops,
1357 1, obj, 0, 0, NULL, NULL, NULL);
1358 if (ret < 0)
1359 goto fail_event;
1360
1361 ret = ceph_osdc_wait_event(event, CEPH_OSD_TIMEOUT_DEFAULT);
1362 dout("ceph_osdc_wait_event returned %d\n", ret);
1363 rbd_destroy_ops(ops);
1364 return 0;
1365
1366fail_event:
1367 ceph_osdc_cancel_event(event);
1368fail:
1369 rbd_destroy_ops(ops);
1370 return ret;
1371}
1372
1373/*
1374 * Request sync osd read
1375 */
1376static int rbd_req_sync_exec(struct rbd_device *dev,
1377 const char *obj,
1378 const char *cls,
1379 const char *method,
1380 const char *data,
1381 int len,
1382 u64 *ver)
1383{
1384 struct ceph_osd_req_op *ops;
1385 int cls_len = strlen(cls);
1386 int method_len = strlen(method);
1387 int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_CALL,
1388 cls_len + method_len + len);
1389 if (ret < 0)
1390 return ret;
1391
1392 ops[0].cls.class_name = cls;
1393 ops[0].cls.class_len = (__u8)cls_len;
1394 ops[0].cls.method_name = method;
1395 ops[0].cls.method_len = (__u8)method_len;
1396 ops[0].cls.argc = 0;
1397 ops[0].cls.indata = data;
1398 ops[0].cls.indata_len = len;
1399
1400 ret = rbd_req_sync_op(dev, NULL,
1401 CEPH_NOSNAP,
1402 0,
1403 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
1404 ops,
1405 1, obj, 0, 0, NULL, NULL, ver);
1406
1407 rbd_destroy_ops(ops);
1408
1409 dout("cls_exec returned %d\n", ret);
1410 return ret;
1411}
1412
1413static struct rbd_req_coll *rbd_alloc_coll(int num_reqs)
1414{
1415 struct rbd_req_coll *coll =
1416 kzalloc(sizeof(struct rbd_req_coll) +
1417 sizeof(struct rbd_req_status) * num_reqs,
1418 GFP_ATOMIC);
1419
1420 if (!coll)
1421 return NULL;
1422 coll->total = num_reqs;
1423 kref_init(&coll->kref);
1424 return coll;
1425}
1426
1427/*
1428 * block device queue callback
1429 */
1430static void rbd_rq_fn(struct request_queue *q)
1431{
1432 struct rbd_device *rbd_dev = q->queuedata;
1433 struct request *rq;
1434 struct bio_pair *bp = NULL;
1435
1436 while ((rq = blk_fetch_request(q))) {
1437 struct bio *bio;
1438 struct bio *rq_bio, *next_bio = NULL;
1439 bool do_write;
1440 int size, op_size = 0;
1441 u64 ofs;
1442 int num_segs, cur_seg = 0;
1443 struct rbd_req_coll *coll;
1444
1445 /* peek at request from block layer */
1446 if (!rq)
1447 break;
1448
1449 dout("fetched request\n");
1450
1451 /* filter out block requests we don't understand */
1452 if ((rq->cmd_type != REQ_TYPE_FS)) {
1453 __blk_end_request_all(rq, 0);
1454 continue;
1455 }
1456
1457 /* deduce our operation (read, write) */
1458 do_write = (rq_data_dir(rq) == WRITE);
1459
1460 size = blk_rq_bytes(rq);
1461 ofs = blk_rq_pos(rq) * SECTOR_SIZE;
1462 rq_bio = rq->bio;
1463 if (do_write && rbd_dev->read_only) {
1464 __blk_end_request_all(rq, -EROFS);
1465 continue;
1466 }
1467
1468 spin_unlock_irq(q->queue_lock);
1469
1470 dout("%s 0x%x bytes at 0x%llx\n",
1471 do_write ? "write" : "read",
1472 size, blk_rq_pos(rq) * SECTOR_SIZE);
1473
1474 num_segs = rbd_get_num_segments(&rbd_dev->header, ofs, size);
1475 coll = rbd_alloc_coll(num_segs);
1476 if (!coll) {
1477 spin_lock_irq(q->queue_lock);
1478 __blk_end_request_all(rq, -ENOMEM);
1479 continue;
1480 }
1481
1482 do {
1483 /* a bio clone to be passed down to OSD req */
1484 dout("rq->bio->bi_vcnt=%d\n", rq->bio->bi_vcnt);
1485 op_size = rbd_get_segment(&rbd_dev->header,
1486 rbd_dev->header.block_name,
1487 ofs, size,
1488 NULL, NULL);
1489 kref_get(&coll->kref);
1490 bio = bio_chain_clone(&rq_bio, &next_bio, &bp,
1491 op_size, GFP_ATOMIC);
1492 if (!bio) {
1493 rbd_coll_end_req_index(rq, coll, cur_seg,
1494 -ENOMEM, op_size);
1495 goto next_seg;
1496 }
1497
1498
1499 /* init OSD command: write or read */
1500 if (do_write)
1501 rbd_req_write(rq, rbd_dev,
1502 rbd_dev->header.snapc,
1503 ofs,
1504 op_size, bio,
1505 coll, cur_seg);
1506 else
1507 rbd_req_read(rq, rbd_dev,
1508 rbd_dev->snap_id,
1509 ofs,
1510 op_size, bio,
1511 coll, cur_seg);
1512
1513next_seg:
1514 size -= op_size;
1515 ofs += op_size;
1516
1517 cur_seg++;
1518 rq_bio = next_bio;
1519 } while (size > 0);
1520 kref_put(&coll->kref, rbd_coll_release);
1521
1522 if (bp)
1523 bio_pair_release(bp);
1524 spin_lock_irq(q->queue_lock);
1525 }
1526}
1527
1528/*
1529 * a queue callback. Makes sure that we don't create a bio that spans across
1530 * multiple osd objects. One exception would be with a single page bios,
1531 * which we handle later at bio_chain_clone
1532 */
1533static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd,
1534 struct bio_vec *bvec)
1535{
1536 struct rbd_device *rbd_dev = q->queuedata;
1537 unsigned int chunk_sectors;
1538 sector_t sector;
1539 unsigned int bio_sectors;
1540 int max;
1541
1542 chunk_sectors = 1 << (rbd_dev->header.obj_order - SECTOR_SHIFT);
1543 sector = bmd->bi_sector + get_start_sect(bmd->bi_bdev);
1544 bio_sectors = bmd->bi_size >> SECTOR_SHIFT;
1545
1546 max = (chunk_sectors - ((sector & (chunk_sectors - 1))
1547 + bio_sectors)) << SECTOR_SHIFT;
1548 if (max < 0)
1549 max = 0; /* bio_add cannot handle a negative return */
1550 if (max <= bvec->bv_len && bio_sectors == 0)
1551 return bvec->bv_len;
1552 return max;
1553}
1554
1555static void rbd_free_disk(struct rbd_device *rbd_dev)
1556{
1557 struct gendisk *disk = rbd_dev->disk;
1558
1559 if (!disk)
1560 return;
1561
1562 rbd_header_free(&rbd_dev->header);
1563
1564 if (disk->flags & GENHD_FL_UP)
1565 del_gendisk(disk);
1566 if (disk->queue)
1567 blk_cleanup_queue(disk->queue);
1568 put_disk(disk);
1569}
1570
1571/*
1572 * reload the ondisk the header
1573 */
1574static int rbd_read_header(struct rbd_device *rbd_dev,
1575 struct rbd_image_header *header)
1576{
1577 ssize_t rc;
1578 struct rbd_image_header_ondisk *dh;
1579 u32 snap_count = 0;
1580 u64 ver;
1581 size_t len;
1582
1583 /*
1584 * First reads the fixed-size header to determine the number
1585 * of snapshots, then re-reads it, along with all snapshot
1586 * records as well as their stored names.
1587 */
1588 len = sizeof (*dh);
1589 while (1) {
1590 dh = kmalloc(len, GFP_KERNEL);
1591 if (!dh)
1592 return -ENOMEM;
1593
1594 rc = rbd_req_sync_read(rbd_dev,
1595 NULL, CEPH_NOSNAP,
1596 rbd_dev->obj_md_name,
1597 0, len,
1598 (char *)dh, &ver);
1599 if (rc < 0)
1600 goto out_dh;
1601
1602 rc = rbd_header_from_disk(header, dh, snap_count, GFP_KERNEL);
1603 if (rc < 0) {
1604 if (rc == -ENXIO)
1605 pr_warning("unrecognized header format"
1606 " for image %s", rbd_dev->obj);
1607 goto out_dh;
1608 }
1609
1610 if (snap_count == header->total_snaps)
1611 break;
1612
1613 snap_count = header->total_snaps;
1614 len = sizeof (*dh) +
1615 snap_count * sizeof(struct rbd_image_snap_ondisk) +
1616 header->snap_names_len;
1617
1618 rbd_header_free(header);
1619 kfree(dh);
1620 }
1621 header->obj_version = ver;
1622
1623out_dh:
1624 kfree(dh);
1625 return rc;
1626}
1627
1628/*
1629 * create a snapshot
1630 */
1631static int rbd_header_add_snap(struct rbd_device *dev,
1632 const char *snap_name,
1633 gfp_t gfp_flags)
1634{
1635 int name_len = strlen(snap_name);
1636 u64 new_snapid;
1637 int ret;
1638 void *data, *p, *e;
1639 u64 ver;
1640 struct ceph_mon_client *monc;
1641
1642 /* we should create a snapshot only if we're pointing at the head */
1643 if (dev->snap_id != CEPH_NOSNAP)
1644 return -EINVAL;
1645
1646 monc = &dev->rbd_client->client->monc;
1647 ret = ceph_monc_create_snapid(monc, dev->poolid, &new_snapid);
1648 dout("created snapid=%lld\n", new_snapid);
1649 if (ret < 0)
1650 return ret;
1651
1652 data = kmalloc(name_len + 16, gfp_flags);
1653 if (!data)
1654 return -ENOMEM;
1655
1656 p = data;
1657 e = data + name_len + 16;
1658
1659 ceph_encode_string_safe(&p, e, snap_name, name_len, bad);
1660 ceph_encode_64_safe(&p, e, new_snapid, bad);
1661
1662 ret = rbd_req_sync_exec(dev, dev->obj_md_name, "rbd", "snap_add",
1663 data, p - data, &ver);
1664
1665 kfree(data);
1666
1667 if (ret < 0)
1668 return ret;
1669
1670 down_write(&dev->header_rwsem);
1671 dev->header.snapc->seq = new_snapid;
1672 up_write(&dev->header_rwsem);
1673
1674 return 0;
1675bad:
1676 return -ERANGE;
1677}
1678
1679static void __rbd_remove_all_snaps(struct rbd_device *rbd_dev)
1680{
1681 struct rbd_snap *snap;
1682
1683 while (!list_empty(&rbd_dev->snaps)) {
1684 snap = list_first_entry(&rbd_dev->snaps, struct rbd_snap, node);
1685 __rbd_remove_snap_dev(rbd_dev, snap);
1686 }
1687}
1688
1689/*
1690 * only read the first part of the ondisk header, without the snaps info
1691 */
1692static int __rbd_refresh_header(struct rbd_device *rbd_dev)
1693{
1694 int ret;
1695 struct rbd_image_header h;
1696 u64 snap_seq;
1697 int follow_seq = 0;
1698
1699 ret = rbd_read_header(rbd_dev, &h);
1700 if (ret < 0)
1701 return ret;
1702
1703 /* resized? */
1704 set_capacity(rbd_dev->disk, h.image_size / SECTOR_SIZE);
1705
1706 down_write(&rbd_dev->header_rwsem);
1707
1708 snap_seq = rbd_dev->header.snapc->seq;
1709 if (rbd_dev->header.total_snaps &&
1710 rbd_dev->header.snapc->snaps[0] == snap_seq)
1711 /* pointing at the head, will need to follow that
1712 if head moves */
1713 follow_seq = 1;
1714
1715 kfree(rbd_dev->header.snapc);
1716 kfree(rbd_dev->header.snap_names);
1717 kfree(rbd_dev->header.snap_sizes);
1718
1719 rbd_dev->header.total_snaps = h.total_snaps;
1720 rbd_dev->header.snapc = h.snapc;
1721 rbd_dev->header.snap_names = h.snap_names;
1722 rbd_dev->header.snap_names_len = h.snap_names_len;
1723 rbd_dev->header.snap_sizes = h.snap_sizes;
1724 if (follow_seq)
1725 rbd_dev->header.snapc->seq = rbd_dev->header.snapc->snaps[0];
1726 else
1727 rbd_dev->header.snapc->seq = snap_seq;
1728
1729 ret = __rbd_init_snaps_header(rbd_dev);
1730
1731 up_write(&rbd_dev->header_rwsem);
1732
1733 return ret;
1734}
1735
1736static int rbd_init_disk(struct rbd_device *rbd_dev)
1737{
1738 struct gendisk *disk;
1739 struct request_queue *q;
1740 int rc;
1741 u64 segment_size;
1742 u64 total_size = 0;
1743
1744 /* contact OSD, request size info about the object being mapped */
1745 rc = rbd_read_header(rbd_dev, &rbd_dev->header);
1746 if (rc)
1747 return rc;
1748
1749 /* no need to lock here, as rbd_dev is not registered yet */
1750 rc = __rbd_init_snaps_header(rbd_dev);
1751 if (rc)
1752 return rc;
1753
1754 rc = rbd_header_set_snap(rbd_dev, &total_size);
1755 if (rc)
1756 return rc;
1757
1758 /* create gendisk info */
1759 rc = -ENOMEM;
1760 disk = alloc_disk(RBD_MINORS_PER_MAJOR);
1761 if (!disk)
1762 goto out;
1763
1764 snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
1765 rbd_dev->id);
1766 disk->major = rbd_dev->major;
1767 disk->first_minor = 0;
1768 disk->fops = &rbd_bd_ops;
1769 disk->private_data = rbd_dev;
1770
1771 /* init rq */
1772 rc = -ENOMEM;
1773 q = blk_init_queue(rbd_rq_fn, &rbd_dev->lock);
1774 if (!q)
1775 goto out_disk;
1776
1777 /* We use the default size, but let's be explicit about it. */
1778 blk_queue_physical_block_size(q, SECTOR_SIZE);
1779
1780 /* set io sizes to object size */
1781 segment_size = rbd_obj_bytes(&rbd_dev->header);
1782 blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
1783 blk_queue_max_segment_size(q, segment_size);
1784 blk_queue_io_min(q, segment_size);
1785 blk_queue_io_opt(q, segment_size);
1786
1787 blk_queue_merge_bvec(q, rbd_merge_bvec);
1788 disk->queue = q;
1789
1790 q->queuedata = rbd_dev;
1791
1792 rbd_dev->disk = disk;
1793 rbd_dev->q = q;
1794
1795 /* finally, announce the disk to the world */
1796 set_capacity(disk, total_size / SECTOR_SIZE);
1797 add_disk(disk);
1798
1799 pr_info("%s: added with size 0x%llx\n",
1800 disk->disk_name, (unsigned long long)total_size);
1801 return 0;
1802
1803out_disk:
1804 put_disk(disk);
1805out:
1806 return rc;
1807}
1808
1809/*
1810 sysfs
1811*/
1812
1813static struct rbd_device *dev_to_rbd_dev(struct device *dev)
1814{
1815 return container_of(dev, struct rbd_device, dev);
1816}
1817
1818static ssize_t rbd_size_show(struct device *dev,
1819 struct device_attribute *attr, char *buf)
1820{
1821 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1822
1823 return sprintf(buf, "%llu\n", (unsigned long long)rbd_dev->header.image_size);
1824}
1825
1826static ssize_t rbd_major_show(struct device *dev,
1827 struct device_attribute *attr, char *buf)
1828{
1829 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1830
1831 return sprintf(buf, "%d\n", rbd_dev->major);
1832}
1833
1834static ssize_t rbd_client_id_show(struct device *dev,
1835 struct device_attribute *attr, char *buf)
1836{
1837 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1838
1839 return sprintf(buf, "client%lld\n",
1840 ceph_client_id(rbd_dev->rbd_client->client));
1841}
1842
1843static ssize_t rbd_pool_show(struct device *dev,
1844 struct device_attribute *attr, char *buf)
1845{
1846 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1847
1848 return sprintf(buf, "%s\n", rbd_dev->pool_name);
1849}
1850
1851static ssize_t rbd_name_show(struct device *dev,
1852 struct device_attribute *attr, char *buf)
1853{
1854 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1855
1856 return sprintf(buf, "%s\n", rbd_dev->obj);
1857}
1858
1859static ssize_t rbd_snap_show(struct device *dev,
1860 struct device_attribute *attr,
1861 char *buf)
1862{
1863 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1864
1865 return sprintf(buf, "%s\n", rbd_dev->snap_name);
1866}
1867
1868static ssize_t rbd_image_refresh(struct device *dev,
1869 struct device_attribute *attr,
1870 const char *buf,
1871 size_t size)
1872{
1873 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
1874 int rc;
1875 int ret = size;
1876
1877 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
1878
1879 rc = __rbd_refresh_header(rbd_dev);
1880 if (rc < 0)
1881 ret = rc;
1882
1883 mutex_unlock(&ctl_mutex);
1884 return ret;
1885}
1886
1887static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
1888static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
1889static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
1890static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
1891static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
1892static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
1893static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
1894static DEVICE_ATTR(create_snap, S_IWUSR, NULL, rbd_snap_add);
1895
1896static struct attribute *rbd_attrs[] = {
1897 &dev_attr_size.attr,
1898 &dev_attr_major.attr,
1899 &dev_attr_client_id.attr,
1900 &dev_attr_pool.attr,
1901 &dev_attr_name.attr,
1902 &dev_attr_current_snap.attr,
1903 &dev_attr_refresh.attr,
1904 &dev_attr_create_snap.attr,
1905 NULL
1906};
1907
1908static struct attribute_group rbd_attr_group = {
1909 .attrs = rbd_attrs,
1910};
1911
1912static const struct attribute_group *rbd_attr_groups[] = {
1913 &rbd_attr_group,
1914 NULL
1915};
1916
1917static void rbd_sysfs_dev_release(struct device *dev)
1918{
1919}
1920
1921static struct device_type rbd_device_type = {
1922 .name = "rbd",
1923 .groups = rbd_attr_groups,
1924 .release = rbd_sysfs_dev_release,
1925};
1926
1927
1928/*
1929 sysfs - snapshots
1930*/
1931
1932static ssize_t rbd_snap_size_show(struct device *dev,
1933 struct device_attribute *attr,
1934 char *buf)
1935{
1936 struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
1937
1938 return sprintf(buf, "%llu\n", (unsigned long long)snap->size);
1939}
1940
1941static ssize_t rbd_snap_id_show(struct device *dev,
1942 struct device_attribute *attr,
1943 char *buf)
1944{
1945 struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
1946
1947 return sprintf(buf, "%llu\n", (unsigned long long)snap->id);
1948}
1949
1950static DEVICE_ATTR(snap_size, S_IRUGO, rbd_snap_size_show, NULL);
1951static DEVICE_ATTR(snap_id, S_IRUGO, rbd_snap_id_show, NULL);
1952
1953static struct attribute *rbd_snap_attrs[] = {
1954 &dev_attr_snap_size.attr,
1955 &dev_attr_snap_id.attr,
1956 NULL,
1957};
1958
1959static struct attribute_group rbd_snap_attr_group = {
1960 .attrs = rbd_snap_attrs,
1961};
1962
1963static void rbd_snap_dev_release(struct device *dev)
1964{
1965 struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
1966 kfree(snap->name);
1967 kfree(snap);
1968}
1969
1970static const struct attribute_group *rbd_snap_attr_groups[] = {
1971 &rbd_snap_attr_group,
1972 NULL
1973};
1974
1975static struct device_type rbd_snap_device_type = {
1976 .groups = rbd_snap_attr_groups,
1977 .release = rbd_snap_dev_release,
1978};
1979
1980static void __rbd_remove_snap_dev(struct rbd_device *rbd_dev,
1981 struct rbd_snap *snap)
1982{
1983 list_del(&snap->node);
1984 device_unregister(&snap->dev);
1985}
1986
1987static int rbd_register_snap_dev(struct rbd_device *rbd_dev,
1988 struct rbd_snap *snap,
1989 struct device *parent)
1990{
1991 struct device *dev = &snap->dev;
1992 int ret;
1993
1994 dev->type = &rbd_snap_device_type;
1995 dev->parent = parent;
1996 dev->release = rbd_snap_dev_release;
1997 dev_set_name(dev, "snap_%s", snap->name);
1998 ret = device_register(dev);
1999
2000 return ret;
2001}
2002
2003static int __rbd_add_snap_dev(struct rbd_device *rbd_dev,
2004 int i, const char *name,
2005 struct rbd_snap **snapp)
2006{
2007 int ret;
2008 struct rbd_snap *snap = kzalloc(sizeof(*snap), GFP_KERNEL);
2009 if (!snap)
2010 return -ENOMEM;
2011 snap->name = kstrdup(name, GFP_KERNEL);
2012 snap->size = rbd_dev->header.snap_sizes[i];
2013 snap->id = rbd_dev->header.snapc->snaps[i];
2014 if (device_is_registered(&rbd_dev->dev)) {
2015 ret = rbd_register_snap_dev(rbd_dev, snap,
2016 &rbd_dev->dev);
2017 if (ret < 0)
2018 goto err;
2019 }
2020 *snapp = snap;
2021 return 0;
2022err:
2023 kfree(snap->name);
2024 kfree(snap);
2025 return ret;
2026}
2027
2028/*
2029 * search for the previous snap in a null delimited string list
2030 */
2031const char *rbd_prev_snap_name(const char *name, const char *start)
2032{
2033 if (name < start + 2)
2034 return NULL;
2035
2036 name -= 2;
2037 while (*name) {
2038 if (name == start)
2039 return start;
2040 name--;
2041 }
2042 return name + 1;
2043}
2044
2045/*
2046 * compare the old list of snapshots that we have to what's in the header
2047 * and update it accordingly. Note that the header holds the snapshots
2048 * in a reverse order (from newest to oldest) and we need to go from
2049 * older to new so that we don't get a duplicate snap name when
2050 * doing the process (e.g., removed snapshot and recreated a new
2051 * one with the same name.
2052 */
2053static int __rbd_init_snaps_header(struct rbd_device *rbd_dev)
2054{
2055 const char *name, *first_name;
2056 int i = rbd_dev->header.total_snaps;
2057 struct rbd_snap *snap, *old_snap = NULL;
2058 int ret;
2059 struct list_head *p, *n;
2060
2061 first_name = rbd_dev->header.snap_names;
2062 name = first_name + rbd_dev->header.snap_names_len;
2063
2064 list_for_each_prev_safe(p, n, &rbd_dev->snaps) {
2065 u64 cur_id;
2066
2067 old_snap = list_entry(p, struct rbd_snap, node);
2068
2069 if (i)
2070 cur_id = rbd_dev->header.snapc->snaps[i - 1];
2071
2072 if (!i || old_snap->id < cur_id) {
2073 /* old_snap->id was skipped, thus was removed */
2074 __rbd_remove_snap_dev(rbd_dev, old_snap);
2075 continue;
2076 }
2077 if (old_snap->id == cur_id) {
2078 /* we have this snapshot already */
2079 i--;
2080 name = rbd_prev_snap_name(name, first_name);
2081 continue;
2082 }
2083 for (; i > 0;
2084 i--, name = rbd_prev_snap_name(name, first_name)) {
2085 if (!name) {
2086 WARN_ON(1);
2087 return -EINVAL;
2088 }
2089 cur_id = rbd_dev->header.snapc->snaps[i];
2090 /* snapshot removal? handle it above */
2091 if (cur_id >= old_snap->id)
2092 break;
2093 /* a new snapshot */
2094 ret = __rbd_add_snap_dev(rbd_dev, i - 1, name, &snap);
2095 if (ret < 0)
2096 return ret;
2097
2098 /* note that we add it backward so using n and not p */
2099 list_add(&snap->node, n);
2100 p = &snap->node;
2101 }
2102 }
2103 /* we're done going over the old snap list, just add what's left */
2104 for (; i > 0; i--) {
2105 name = rbd_prev_snap_name(name, first_name);
2106 if (!name) {
2107 WARN_ON(1);
2108 return -EINVAL;
2109 }
2110 ret = __rbd_add_snap_dev(rbd_dev, i - 1, name, &snap);
2111 if (ret < 0)
2112 return ret;
2113 list_add(&snap->node, &rbd_dev->snaps);
2114 }
2115
2116 return 0;
2117}
2118
2119static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
2120{
2121 int ret;
2122 struct device *dev;
2123 struct rbd_snap *snap;
2124
2125 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
2126 dev = &rbd_dev->dev;
2127
2128 dev->bus = &rbd_bus_type;
2129 dev->type = &rbd_device_type;
2130 dev->parent = &rbd_root_dev;
2131 dev->release = rbd_dev_release;
2132 dev_set_name(dev, "%d", rbd_dev->id);
2133 ret = device_register(dev);
2134 if (ret < 0)
2135 goto out;
2136
2137 list_for_each_entry(snap, &rbd_dev->snaps, node) {
2138 ret = rbd_register_snap_dev(rbd_dev, snap,
2139 &rbd_dev->dev);
2140 if (ret < 0)
2141 break;
2142 }
2143out:
2144 mutex_unlock(&ctl_mutex);
2145 return ret;
2146}
2147
2148static void rbd_bus_del_dev(struct rbd_device *rbd_dev)
2149{
2150 device_unregister(&rbd_dev->dev);
2151}
2152
2153static int rbd_init_watch_dev(struct rbd_device *rbd_dev)
2154{
2155 int ret, rc;
2156
2157 do {
2158 ret = rbd_req_sync_watch(rbd_dev, rbd_dev->obj_md_name,
2159 rbd_dev->header.obj_version);
2160 if (ret == -ERANGE) {
2161 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
2162 rc = __rbd_refresh_header(rbd_dev);
2163 mutex_unlock(&ctl_mutex);
2164 if (rc < 0)
2165 return rc;
2166 }
2167 } while (ret == -ERANGE);
2168
2169 return ret;
2170}
2171
2172static atomic64_t rbd_id_max = ATOMIC64_INIT(0);
2173
2174/*
2175 * Get a unique rbd identifier for the given new rbd_dev, and add
2176 * the rbd_dev to the global list. The minimum rbd id is 1.
2177 */
2178static void rbd_id_get(struct rbd_device *rbd_dev)
2179{
2180 rbd_dev->id = atomic64_inc_return(&rbd_id_max);
2181
2182 spin_lock(&rbd_dev_list_lock);
2183 list_add_tail(&rbd_dev->node, &rbd_dev_list);
2184 spin_unlock(&rbd_dev_list_lock);
2185}
2186
2187/*
2188 * Remove an rbd_dev from the global list, and record that its
2189 * identifier is no longer in use.
2190 */
2191static void rbd_id_put(struct rbd_device *rbd_dev)
2192{
2193 struct list_head *tmp;
2194 int rbd_id = rbd_dev->id;
2195 int max_id;
2196
2197 BUG_ON(rbd_id < 1);
2198
2199 spin_lock(&rbd_dev_list_lock);
2200 list_del_init(&rbd_dev->node);
2201
2202 /*
2203 * If the id being "put" is not the current maximum, there
2204 * is nothing special we need to do.
2205 */
2206 if (rbd_id != atomic64_read(&rbd_id_max)) {
2207 spin_unlock(&rbd_dev_list_lock);
2208 return;
2209 }
2210
2211 /*
2212 * We need to update the current maximum id. Search the
2213 * list to find out what it is. We're more likely to find
2214 * the maximum at the end, so search the list backward.
2215 */
2216 max_id = 0;
2217 list_for_each_prev(tmp, &rbd_dev_list) {
2218 struct rbd_device *rbd_dev;
2219
2220 rbd_dev = list_entry(tmp, struct rbd_device, node);
2221 if (rbd_id > max_id)
2222 max_id = rbd_id;
2223 }
2224 spin_unlock(&rbd_dev_list_lock);
2225
2226 /*
2227 * The max id could have been updated by rbd_id_get(), in
2228 * which case it now accurately reflects the new maximum.
2229 * Be careful not to overwrite the maximum value in that
2230 * case.
2231 */
2232 atomic64_cmpxchg(&rbd_id_max, rbd_id, max_id);
2233}
2234
2235/*
2236 * Skips over white space at *buf, and updates *buf to point to the
2237 * first found non-space character (if any). Returns the length of
2238 * the token (string of non-white space characters) found. Note
2239 * that *buf must be terminated with '\0'.
2240 */
2241static inline size_t next_token(const char **buf)
2242{
2243 /*
2244 * These are the characters that produce nonzero for
2245 * isspace() in the "C" and "POSIX" locales.
2246 */
2247 const char *spaces = " \f\n\r\t\v";
2248
2249 *buf += strspn(*buf, spaces); /* Find start of token */
2250
2251 return strcspn(*buf, spaces); /* Return token length */
2252}
2253
2254/*
2255 * Finds the next token in *buf, and if the provided token buffer is
2256 * big enough, copies the found token into it. The result, if
2257 * copied, is guaranteed to be terminated with '\0'. Note that *buf
2258 * must be terminated with '\0' on entry.
2259 *
2260 * Returns the length of the token found (not including the '\0').
2261 * Return value will be 0 if no token is found, and it will be >=
2262 * token_size if the token would not fit.
2263 *
2264 * The *buf pointer will be updated to point beyond the end of the
2265 * found token. Note that this occurs even if the token buffer is
2266 * too small to hold it.
2267 */
2268static inline size_t copy_token(const char **buf,
2269 char *token,
2270 size_t token_size)
2271{
2272 size_t len;
2273
2274 len = next_token(buf);
2275 if (len < token_size) {
2276 memcpy(token, *buf, len);
2277 *(token + len) = '\0';
2278 }
2279 *buf += len;
2280
2281 return len;
2282}
2283
2284/*
2285 * This fills in the pool_name, obj, obj_len, snap_name, obj_len,
2286 * rbd_dev, rbd_md_name, and name fields of the given rbd_dev, based
2287 * on the list of monitor addresses and other options provided via
2288 * /sys/bus/rbd/add.
2289 */
2290static int rbd_add_parse_args(struct rbd_device *rbd_dev,
2291 const char *buf,
2292 const char **mon_addrs,
2293 size_t *mon_addrs_size,
2294 char *options,
2295 size_t options_size)
2296{
2297 size_t len;
2298
2299 /* The first four tokens are required */
2300
2301 len = next_token(&buf);
2302 if (!len)
2303 return -EINVAL;
2304 *mon_addrs_size = len + 1;
2305 *mon_addrs = buf;
2306
2307 buf += len;
2308
2309 len = copy_token(&buf, options, options_size);
2310 if (!len || len >= options_size)
2311 return -EINVAL;
2312
2313 len = copy_token(&buf, rbd_dev->pool_name, sizeof (rbd_dev->pool_name));
2314 if (!len || len >= sizeof (rbd_dev->pool_name))
2315 return -EINVAL;
2316
2317 len = copy_token(&buf, rbd_dev->obj, sizeof (rbd_dev->obj));
2318 if (!len || len >= sizeof (rbd_dev->obj))
2319 return -EINVAL;
2320
2321 /* We have the object length in hand, save it. */
2322
2323 rbd_dev->obj_len = len;
2324
2325 BUILD_BUG_ON(RBD_MAX_MD_NAME_LEN
2326 < RBD_MAX_OBJ_NAME_LEN + sizeof (RBD_SUFFIX));
2327 sprintf(rbd_dev->obj_md_name, "%s%s", rbd_dev->obj, RBD_SUFFIX);
2328
2329 /*
2330 * The snapshot name is optional, but it's an error if it's
2331 * too long. If no snapshot is supplied, fill in the default.
2332 */
2333 len = copy_token(&buf, rbd_dev->snap_name, sizeof (rbd_dev->snap_name));
2334 if (!len)
2335 memcpy(rbd_dev->snap_name, RBD_SNAP_HEAD_NAME,
2336 sizeof (RBD_SNAP_HEAD_NAME));
2337 else if (len >= sizeof (rbd_dev->snap_name))
2338 return -EINVAL;
2339
2340 return 0;
2341}
2342
2343static ssize_t rbd_add(struct bus_type *bus,
2344 const char *buf,
2345 size_t count)
2346{
2347 struct rbd_device *rbd_dev;
2348 const char *mon_addrs = NULL;
2349 size_t mon_addrs_size = 0;
2350 char *options = NULL;
2351 struct ceph_osd_client *osdc;
2352 int rc = -ENOMEM;
2353
2354 if (!try_module_get(THIS_MODULE))
2355 return -ENODEV;
2356
2357 rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL);
2358 if (!rbd_dev)
2359 goto err_nomem;
2360 options = kmalloc(count, GFP_KERNEL);
2361 if (!options)
2362 goto err_nomem;
2363
2364 /* static rbd_device initialization */
2365 spin_lock_init(&rbd_dev->lock);
2366 INIT_LIST_HEAD(&rbd_dev->node);
2367 INIT_LIST_HEAD(&rbd_dev->snaps);
2368 init_rwsem(&rbd_dev->header_rwsem);
2369
2370 init_rwsem(&rbd_dev->header_rwsem);
2371
2372 /* generate unique id: find highest unique id, add one */
2373 rbd_id_get(rbd_dev);
2374
2375 /* Fill in the device name, now that we have its id. */
2376 BUILD_BUG_ON(DEV_NAME_LEN
2377 < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
2378 sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->id);
2379
2380 /* parse add command */
2381 rc = rbd_add_parse_args(rbd_dev, buf, &mon_addrs, &mon_addrs_size,
2382 options, count);
2383 if (rc)
2384 goto err_put_id;
2385
2386 rbd_dev->rbd_client = rbd_get_client(mon_addrs, mon_addrs_size - 1,
2387 options);
2388 if (IS_ERR(rbd_dev->rbd_client)) {
2389 rc = PTR_ERR(rbd_dev->rbd_client);
2390 goto err_put_id;
2391 }
2392
2393 /* pick the pool */
2394 osdc = &rbd_dev->rbd_client->client->osdc;
2395 rc = ceph_pg_poolid_by_name(osdc->osdmap, rbd_dev->pool_name);
2396 if (rc < 0)
2397 goto err_out_client;
2398 rbd_dev->poolid = rc;
2399
2400 /* register our block device */
2401 rc = register_blkdev(0, rbd_dev->name);
2402 if (rc < 0)
2403 goto err_out_client;
2404 rbd_dev->major = rc;
2405
2406 rc = rbd_bus_add_dev(rbd_dev);
2407 if (rc)
2408 goto err_out_blkdev;
2409
2410 /*
2411 * At this point cleanup in the event of an error is the job
2412 * of the sysfs code (initiated by rbd_bus_del_dev()).
2413 *
2414 * Set up and announce blkdev mapping.
2415 */
2416 rc = rbd_init_disk(rbd_dev);
2417 if (rc)
2418 goto err_out_bus;
2419
2420 rc = rbd_init_watch_dev(rbd_dev);
2421 if (rc)
2422 goto err_out_bus;
2423
2424 return count;
2425
2426err_out_bus:
2427 /* this will also clean up rest of rbd_dev stuff */
2428
2429 rbd_bus_del_dev(rbd_dev);
2430 kfree(options);
2431 return rc;
2432
2433err_out_blkdev:
2434 unregister_blkdev(rbd_dev->major, rbd_dev->name);
2435err_out_client:
2436 rbd_put_client(rbd_dev);
2437err_put_id:
2438 rbd_id_put(rbd_dev);
2439err_nomem:
2440 kfree(options);
2441 kfree(rbd_dev);
2442
2443 dout("Error adding device %s\n", buf);
2444 module_put(THIS_MODULE);
2445
2446 return (ssize_t) rc;
2447}
2448
2449static struct rbd_device *__rbd_get_dev(unsigned long id)
2450{
2451 struct list_head *tmp;
2452 struct rbd_device *rbd_dev;
2453
2454 spin_lock(&rbd_dev_list_lock);
2455 list_for_each(tmp, &rbd_dev_list) {
2456 rbd_dev = list_entry(tmp, struct rbd_device, node);
2457 if (rbd_dev->id == id) {
2458 spin_unlock(&rbd_dev_list_lock);
2459 return rbd_dev;
2460 }
2461 }
2462 spin_unlock(&rbd_dev_list_lock);
2463 return NULL;
2464}
2465
2466static void rbd_dev_release(struct device *dev)
2467{
2468 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
2469
2470 if (rbd_dev->watch_request) {
2471 struct ceph_client *client = rbd_dev->rbd_client->client;
2472
2473 ceph_osdc_unregister_linger_request(&client->osdc,
2474 rbd_dev->watch_request);
2475 }
2476 if (rbd_dev->watch_event)
2477 rbd_req_sync_unwatch(rbd_dev, rbd_dev->obj_md_name);
2478
2479 rbd_put_client(rbd_dev);
2480
2481 /* clean up and free blkdev */
2482 rbd_free_disk(rbd_dev);
2483 unregister_blkdev(rbd_dev->major, rbd_dev->name);
2484
2485 /* done with the id, and with the rbd_dev */
2486 rbd_id_put(rbd_dev);
2487 kfree(rbd_dev);
2488
2489 /* release module ref */
2490 module_put(THIS_MODULE);
2491}
2492
2493static ssize_t rbd_remove(struct bus_type *bus,
2494 const char *buf,
2495 size_t count)
2496{
2497 struct rbd_device *rbd_dev = NULL;
2498 int target_id, rc;
2499 unsigned long ul;
2500 int ret = count;
2501
2502 rc = strict_strtoul(buf, 10, &ul);
2503 if (rc)
2504 return rc;
2505
2506 /* convert to int; abort if we lost anything in the conversion */
2507 target_id = (int) ul;
2508 if (target_id != ul)
2509 return -EINVAL;
2510
2511 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
2512
2513 rbd_dev = __rbd_get_dev(target_id);
2514 if (!rbd_dev) {
2515 ret = -ENOENT;
2516 goto done;
2517 }
2518
2519 __rbd_remove_all_snaps(rbd_dev);
2520 rbd_bus_del_dev(rbd_dev);
2521
2522done:
2523 mutex_unlock(&ctl_mutex);
2524 return ret;
2525}
2526
2527static ssize_t rbd_snap_add(struct device *dev,
2528 struct device_attribute *attr,
2529 const char *buf,
2530 size_t count)
2531{
2532 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
2533 int ret;
2534 char *name = kmalloc(count + 1, GFP_KERNEL);
2535 if (!name)
2536 return -ENOMEM;
2537
2538 snprintf(name, count, "%s", buf);
2539
2540 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
2541
2542 ret = rbd_header_add_snap(rbd_dev,
2543 name, GFP_KERNEL);
2544 if (ret < 0)
2545 goto err_unlock;
2546
2547 ret = __rbd_refresh_header(rbd_dev);
2548 if (ret < 0)
2549 goto err_unlock;
2550
2551 /* shouldn't hold ctl_mutex when notifying.. notify might
2552 trigger a watch callback that would need to get that mutex */
2553 mutex_unlock(&ctl_mutex);
2554
2555 /* make a best effort, don't error if failed */
2556 rbd_req_sync_notify(rbd_dev, rbd_dev->obj_md_name);
2557
2558 ret = count;
2559 kfree(name);
2560 return ret;
2561
2562err_unlock:
2563 mutex_unlock(&ctl_mutex);
2564 kfree(name);
2565 return ret;
2566}
2567
2568/*
2569 * create control files in sysfs
2570 * /sys/bus/rbd/...
2571 */
2572static int rbd_sysfs_init(void)
2573{
2574 int ret;
2575
2576 ret = device_register(&rbd_root_dev);
2577 if (ret < 0)
2578 return ret;
2579
2580 ret = bus_register(&rbd_bus_type);
2581 if (ret < 0)
2582 device_unregister(&rbd_root_dev);
2583
2584 return ret;
2585}
2586
2587static void rbd_sysfs_cleanup(void)
2588{
2589 bus_unregister(&rbd_bus_type);
2590 device_unregister(&rbd_root_dev);
2591}
2592
2593int __init rbd_init(void)
2594{
2595 int rc;
2596
2597 rc = rbd_sysfs_init();
2598 if (rc)
2599 return rc;
2600 pr_info("loaded " RBD_DRV_NAME_LONG "\n");
2601 return 0;
2602}
2603
2604void __exit rbd_exit(void)
2605{
2606 rbd_sysfs_cleanup();
2607}
2608
2609module_init(rbd_init);
2610module_exit(rbd_exit);
2611
2612MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
2613MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
2614MODULE_DESCRIPTION("rados block device");
2615
2616/* following authorship retained from original osdblk.c */
2617MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");
2618
2619MODULE_LICENSE("GPL");