2 rbd.c -- Export ceph rados objects as a Linux block device
5 based on drivers/block/osdblk.c:
7 Copyright 2009 Red Hat, Inc.
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.
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.
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.
24 For usage instructions, please refer to:
26 Documentation/ABI/testing/sysfs-bus-rbd
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>
36 #include <linux/kernel.h>
37 #include <linux/device.h>
38 #include <linux/module.h>
40 #include <linux/blkdev.h>
42 #include "rbd_types.h"
44 #define RBD_DEBUG /* Activate rbd_assert() calls */
47 * The basic unit of block I/O is a sector. It is interpreted in a
48 * number of contexts in Linux (blk, bio, genhd), but the default is
49 * universally 512 bytes. These symbols are just slightly more
50 * meaningful than the bare numbers they represent.
52 #define SECTOR_SHIFT 9
53 #define SECTOR_SIZE (1ULL << SECTOR_SHIFT)
55 #define RBD_DRV_NAME "rbd"
56 #define RBD_DRV_NAME_LONG "rbd (rados block device)"
58 #define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */
60 #define RBD_SNAP_DEV_NAME_PREFIX "snap_"
61 #define RBD_MAX_SNAP_NAME_LEN \
62 (NAME_MAX - (sizeof (RBD_SNAP_DEV_NAME_PREFIX) - 1))
64 #define RBD_MAX_SNAP_COUNT 510 /* allows max snapc to fit in 4KB */
66 #define RBD_SNAP_HEAD_NAME "-"
68 /* This allows a single page to hold an image name sent by OSD */
69 #define RBD_IMAGE_NAME_LEN_MAX (PAGE_SIZE - sizeof (__le32) - 1)
70 #define RBD_IMAGE_ID_LEN_MAX 64
72 #define RBD_OBJ_PREFIX_LEN_MAX 64
76 #define RBD_FEATURE_LAYERING (1<<0)
77 #define RBD_FEATURE_STRIPINGV2 (1<<1)
78 #define RBD_FEATURES_ALL \
79 (RBD_FEATURE_LAYERING | RBD_FEATURE_STRIPINGV2)
81 /* Features supported by this (client software) implementation. */
83 #define RBD_FEATURES_SUPPORTED (RBD_FEATURES_ALL)
86 * An RBD device name will be "rbd#", where the "rbd" comes from
87 * RBD_DRV_NAME above, and # is a unique integer identifier.
88 * MAX_INT_FORMAT_WIDTH is used in ensuring DEV_NAME_LEN is big
89 * enough to hold all possible device names.
91 #define DEV_NAME_LEN 32
92 #define MAX_INT_FORMAT_WIDTH ((5 * sizeof (int)) / 2 + 1)
95 * block device image metadata (in-memory version)
97 struct rbd_image_header {
98 /* These four fields never change for a given rbd image */
105 /* The remaining fields need to be updated occasionally */
107 struct ceph_snap_context *snapc;
118 * An rbd image specification.
120 * The tuple (pool_id, image_id, snap_id) is sufficient to uniquely
121 * identify an image. Each rbd_dev structure includes a pointer to
122 * an rbd_spec structure that encapsulates this identity.
124 * Each of the id's in an rbd_spec has an associated name. For a
125 * user-mapped image, the names are supplied and the id's associated
126 * with them are looked up. For a layered image, a parent image is
127 * defined by the tuple, and the names are looked up.
129 * An rbd_dev structure contains a parent_spec pointer which is
130 * non-null if the image it represents is a child in a layered
131 * image. This pointer will refer to the rbd_spec structure used
132 * by the parent rbd_dev for its own identity (i.e., the structure
133 * is shared between the parent and child).
135 * Since these structures are populated once, during the discovery
136 * phase of image construction, they are effectively immutable so
137 * we make no effort to synchronize access to them.
139 * Note that code herein does not assume the image name is known (it
140 * could be a null pointer).
144 const char *pool_name;
146 const char *image_id;
147 const char *image_name;
150 const char *snap_name;
156 * an instance of the client. multiple devices may share an rbd client.
159 struct ceph_client *client;
161 struct list_head node;
164 struct rbd_img_request;
165 typedef void (*rbd_img_callback_t)(struct rbd_img_request *);
167 #define BAD_WHICH U32_MAX /* Good which or bad which, which? */
169 struct rbd_obj_request;
170 typedef void (*rbd_obj_callback_t)(struct rbd_obj_request *);
172 enum obj_request_type {
173 OBJ_REQUEST_NODATA, OBJ_REQUEST_BIO, OBJ_REQUEST_PAGES
177 OBJ_REQ_DONE, /* completion flag: not done = 0, done = 1 */
178 OBJ_REQ_IMG_DATA, /* object usage: standalone = 0, image = 1 */
179 OBJ_REQ_KNOWN, /* EXISTS flag valid: no = 0, yes = 1 */
180 OBJ_REQ_EXISTS, /* target exists: no = 0, yes = 1 */
183 struct rbd_obj_request {
184 const char *object_name;
185 u64 offset; /* object start byte */
186 u64 length; /* bytes from offset */
190 * An object request associated with an image will have its
191 * img_data flag set; a standalone object request will not.
193 * A standalone object request will have which == BAD_WHICH
194 * and a null obj_request pointer.
196 * An object request initiated in support of a layered image
197 * object (to check for its existence before a write) will
198 * have which == BAD_WHICH and a non-null obj_request pointer.
200 * Finally, an object request for rbd image data will have
201 * which != BAD_WHICH, and will have a non-null img_request
202 * pointer. The value of which will be in the range
203 * 0..(img_request->obj_request_count-1).
206 struct rbd_obj_request *obj_request; /* STAT op */
208 struct rbd_img_request *img_request;
210 /* links for img_request->obj_requests list */
211 struct list_head links;
214 u32 which; /* posn image request list */
216 enum obj_request_type type;
218 struct bio *bio_list;
224 struct page **copyup_pages;
226 struct ceph_osd_request *osd_req;
228 u64 xferred; /* bytes transferred */
232 rbd_obj_callback_t callback;
233 struct completion completion;
239 IMG_REQ_WRITE, /* I/O direction: read = 0, write = 1 */
240 IMG_REQ_CHILD, /* initiator: block = 0, child image = 1 */
241 IMG_REQ_LAYERED, /* ENOENT handling: normal = 0, layered = 1 */
244 struct rbd_img_request {
245 struct rbd_device *rbd_dev;
246 u64 offset; /* starting image byte offset */
247 u64 length; /* byte count from offset */
250 u64 snap_id; /* for reads */
251 struct ceph_snap_context *snapc; /* for writes */
254 struct request *rq; /* block request */
255 struct rbd_obj_request *obj_request; /* obj req initiator */
257 struct page **copyup_pages;
258 spinlock_t completion_lock;/* protects next_completion */
260 rbd_img_callback_t callback;
261 u64 xferred;/* aggregate bytes transferred */
262 int result; /* first nonzero obj_request result */
264 u32 obj_request_count;
265 struct list_head obj_requests; /* rbd_obj_request structs */
270 #define for_each_obj_request(ireq, oreq) \
271 list_for_each_entry(oreq, &(ireq)->obj_requests, links)
272 #define for_each_obj_request_from(ireq, oreq) \
273 list_for_each_entry_from(oreq, &(ireq)->obj_requests, links)
274 #define for_each_obj_request_safe(ireq, oreq, n) \
275 list_for_each_entry_safe_reverse(oreq, n, &(ireq)->obj_requests, links)
280 struct list_head node;
295 int dev_id; /* blkdev unique id */
297 int major; /* blkdev assigned major */
298 struct gendisk *disk; /* blkdev's gendisk and rq */
300 u32 image_format; /* Either 1 or 2 */
301 struct rbd_client *rbd_client;
303 char name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */
305 spinlock_t lock; /* queue, flags, open_count */
307 struct rbd_image_header header;
308 unsigned long flags; /* possibly lock protected */
309 struct rbd_spec *spec;
313 struct ceph_file_layout layout;
315 struct ceph_osd_event *watch_event;
316 struct rbd_obj_request *watch_request;
318 struct rbd_spec *parent_spec;
320 struct rbd_device *parent;
322 /* protects updating the header */
323 struct rw_semaphore header_rwsem;
325 struct rbd_mapping mapping;
327 struct list_head node;
329 /* list of snapshots */
330 struct list_head snaps;
334 unsigned long open_count; /* protected by lock */
338 * Flag bits for rbd_dev->flags. If atomicity is required,
339 * rbd_dev->lock is used to protect access.
341 * Currently, only the "removing" flag (which is coupled with the
342 * "open_count" field) requires atomic access.
345 RBD_DEV_FLAG_EXISTS, /* mapped snapshot has not been deleted */
346 RBD_DEV_FLAG_REMOVING, /* this mapping is being removed */
349 static DEFINE_MUTEX(ctl_mutex); /* Serialize open/close/setup/teardown */
351 static LIST_HEAD(rbd_dev_list); /* devices */
352 static DEFINE_SPINLOCK(rbd_dev_list_lock);
354 static LIST_HEAD(rbd_client_list); /* clients */
355 static DEFINE_SPINLOCK(rbd_client_list_lock);
357 static int rbd_img_request_submit(struct rbd_img_request *img_request);
359 static int rbd_dev_snaps_update(struct rbd_device *rbd_dev);
361 static void rbd_dev_release(struct device *dev);
362 static void rbd_snap_destroy(struct rbd_snap *snap);
364 static ssize_t rbd_add(struct bus_type *bus, const char *buf,
366 static ssize_t rbd_remove(struct bus_type *bus, const char *buf,
368 static int rbd_dev_image_probe(struct rbd_device *rbd_dev);
370 static struct bus_attribute rbd_bus_attrs[] = {
371 __ATTR(add, S_IWUSR, NULL, rbd_add),
372 __ATTR(remove, S_IWUSR, NULL, rbd_remove),
376 static struct bus_type rbd_bus_type = {
378 .bus_attrs = rbd_bus_attrs,
381 static void rbd_root_dev_release(struct device *dev)
385 static struct device rbd_root_dev = {
387 .release = rbd_root_dev_release,
390 static __printf(2, 3)
391 void rbd_warn(struct rbd_device *rbd_dev, const char *fmt, ...)
393 struct va_format vaf;
401 printk(KERN_WARNING "%s: %pV\n", RBD_DRV_NAME, &vaf);
402 else if (rbd_dev->disk)
403 printk(KERN_WARNING "%s: %s: %pV\n",
404 RBD_DRV_NAME, rbd_dev->disk->disk_name, &vaf);
405 else if (rbd_dev->spec && rbd_dev->spec->image_name)
406 printk(KERN_WARNING "%s: image %s: %pV\n",
407 RBD_DRV_NAME, rbd_dev->spec->image_name, &vaf);
408 else if (rbd_dev->spec && rbd_dev->spec->image_id)
409 printk(KERN_WARNING "%s: id %s: %pV\n",
410 RBD_DRV_NAME, rbd_dev->spec->image_id, &vaf);
412 printk(KERN_WARNING "%s: rbd_dev %p: %pV\n",
413 RBD_DRV_NAME, rbd_dev, &vaf);
418 #define rbd_assert(expr) \
419 if (unlikely(!(expr))) { \
420 printk(KERN_ERR "\nAssertion failure in %s() " \
422 "\trbd_assert(%s);\n\n", \
423 __func__, __LINE__, #expr); \
426 #else /* !RBD_DEBUG */
427 # define rbd_assert(expr) ((void) 0)
428 #endif /* !RBD_DEBUG */
430 static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request);
431 static void rbd_img_parent_read(struct rbd_obj_request *obj_request);
432 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);
434 static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver);
435 static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev, u64 *hver);
437 static int rbd_open(struct block_device *bdev, fmode_t mode)
439 struct rbd_device *rbd_dev = bdev->bd_disk->private_data;
440 bool removing = false;
442 if ((mode & FMODE_WRITE) && rbd_dev->mapping.read_only)
445 spin_lock_irq(&rbd_dev->lock);
446 if (test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags))
449 rbd_dev->open_count++;
450 spin_unlock_irq(&rbd_dev->lock);
454 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
455 (void) get_device(&rbd_dev->dev);
456 set_device_ro(bdev, rbd_dev->mapping.read_only);
457 mutex_unlock(&ctl_mutex);
462 static int rbd_release(struct gendisk *disk, fmode_t mode)
464 struct rbd_device *rbd_dev = disk->private_data;
465 unsigned long open_count_before;
467 spin_lock_irq(&rbd_dev->lock);
468 open_count_before = rbd_dev->open_count--;
469 spin_unlock_irq(&rbd_dev->lock);
470 rbd_assert(open_count_before > 0);
472 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
473 put_device(&rbd_dev->dev);
474 mutex_unlock(&ctl_mutex);
479 static const struct block_device_operations rbd_bd_ops = {
480 .owner = THIS_MODULE,
482 .release = rbd_release,
486 * Initialize an rbd client instance.
489 static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
491 struct rbd_client *rbdc;
494 dout("%s:\n", __func__);
495 rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL);
499 kref_init(&rbdc->kref);
500 INIT_LIST_HEAD(&rbdc->node);
502 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
504 rbdc->client = ceph_create_client(ceph_opts, rbdc, 0, 0);
505 if (IS_ERR(rbdc->client))
507 ceph_opts = NULL; /* Now rbdc->client is responsible for ceph_opts */
509 ret = ceph_open_session(rbdc->client);
513 spin_lock(&rbd_client_list_lock);
514 list_add_tail(&rbdc->node, &rbd_client_list);
515 spin_unlock(&rbd_client_list_lock);
517 mutex_unlock(&ctl_mutex);
518 dout("%s: rbdc %p\n", __func__, rbdc);
523 ceph_destroy_client(rbdc->client);
525 mutex_unlock(&ctl_mutex);
529 ceph_destroy_options(ceph_opts);
530 dout("%s: error %d\n", __func__, ret);
535 static struct rbd_client *__rbd_get_client(struct rbd_client *rbdc)
537 kref_get(&rbdc->kref);
543 * Find a ceph client with specific addr and configuration. If
544 * found, bump its reference count.
546 static struct rbd_client *rbd_client_find(struct ceph_options *ceph_opts)
548 struct rbd_client *client_node;
551 if (ceph_opts->flags & CEPH_OPT_NOSHARE)
554 spin_lock(&rbd_client_list_lock);
555 list_for_each_entry(client_node, &rbd_client_list, node) {
556 if (!ceph_compare_options(ceph_opts, client_node->client)) {
557 __rbd_get_client(client_node);
563 spin_unlock(&rbd_client_list_lock);
565 return found ? client_node : NULL;
575 /* string args above */
578 /* Boolean args above */
582 static match_table_t rbd_opts_tokens = {
584 /* string args above */
585 {Opt_read_only, "read_only"},
586 {Opt_read_only, "ro"}, /* Alternate spelling */
587 {Opt_read_write, "read_write"},
588 {Opt_read_write, "rw"}, /* Alternate spelling */
589 /* Boolean args above */
597 #define RBD_READ_ONLY_DEFAULT false
599 static int parse_rbd_opts_token(char *c, void *private)
601 struct rbd_options *rbd_opts = private;
602 substring_t argstr[MAX_OPT_ARGS];
603 int token, intval, ret;
605 token = match_token(c, rbd_opts_tokens, argstr);
609 if (token < Opt_last_int) {
610 ret = match_int(&argstr[0], &intval);
612 pr_err("bad mount option arg (not int) "
616 dout("got int token %d val %d\n", token, intval);
617 } else if (token > Opt_last_int && token < Opt_last_string) {
618 dout("got string token %d val %s\n", token,
620 } else if (token > Opt_last_string && token < Opt_last_bool) {
621 dout("got Boolean token %d\n", token);
623 dout("got token %d\n", token);
628 rbd_opts->read_only = true;
631 rbd_opts->read_only = false;
641 * Get a ceph client with specific addr and configuration, if one does
642 * not exist create it.
644 static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts)
646 struct rbd_client *rbdc;
648 rbdc = rbd_client_find(ceph_opts);
649 if (rbdc) /* using an existing client */
650 ceph_destroy_options(ceph_opts);
652 rbdc = rbd_client_create(ceph_opts);
658 * Destroy ceph client
660 * Caller must hold rbd_client_list_lock.
662 static void rbd_client_release(struct kref *kref)
664 struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref);
666 dout("%s: rbdc %p\n", __func__, rbdc);
667 spin_lock(&rbd_client_list_lock);
668 list_del(&rbdc->node);
669 spin_unlock(&rbd_client_list_lock);
671 ceph_destroy_client(rbdc->client);
675 /* Caller has to fill in snapc->seq and snapc->snaps[0..snap_count-1] */
677 static struct ceph_snap_context *rbd_snap_context_create(u32 snap_count)
679 struct ceph_snap_context *snapc;
682 size = sizeof (struct ceph_snap_context);
683 size += snap_count * sizeof (snapc->snaps[0]);
684 snapc = kzalloc(size, GFP_KERNEL);
688 atomic_set(&snapc->nref, 1);
689 snapc->num_snaps = snap_count;
694 static inline void rbd_snap_context_get(struct ceph_snap_context *snapc)
696 (void)ceph_get_snap_context(snapc);
699 static inline void rbd_snap_context_put(struct ceph_snap_context *snapc)
701 ceph_put_snap_context(snapc);
705 * Drop reference to ceph client node. If it's not referenced anymore, release
708 static void rbd_put_client(struct rbd_client *rbdc)
711 kref_put(&rbdc->kref, rbd_client_release);
714 static bool rbd_image_format_valid(u32 image_format)
716 return image_format == 1 || image_format == 2;
719 static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
724 /* The header has to start with the magic rbd header text */
725 if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT)))
728 /* The bio layer requires at least sector-sized I/O */
730 if (ondisk->options.order < SECTOR_SHIFT)
733 /* If we use u64 in a few spots we may be able to loosen this */
735 if (ondisk->options.order > 8 * sizeof (int) - 1)
739 * The size of a snapshot header has to fit in a size_t, and
740 * that limits the number of snapshots.
742 snap_count = le32_to_cpu(ondisk->snap_count);
743 size = SIZE_MAX - sizeof (struct ceph_snap_context);
744 if (snap_count > size / sizeof (__le64))
748 * Not only that, but the size of the entire the snapshot
749 * header must also be representable in a size_t.
751 size -= snap_count * sizeof (__le64);
752 if ((u64) size < le64_to_cpu(ondisk->snap_names_len))
759 * Create a new header structure, translate header format from the on-disk
762 static int rbd_header_from_disk(struct rbd_image_header *header,
763 struct rbd_image_header_ondisk *ondisk)
770 memset(header, 0, sizeof (*header));
772 snap_count = le32_to_cpu(ondisk->snap_count);
774 len = strnlen(ondisk->object_prefix, sizeof (ondisk->object_prefix));
775 header->object_prefix = kmalloc(len + 1, GFP_KERNEL);
776 if (!header->object_prefix)
778 memcpy(header->object_prefix, ondisk->object_prefix, len);
779 header->object_prefix[len] = '\0';
782 u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len);
784 /* Save a copy of the snapshot names */
786 if (snap_names_len > (u64) SIZE_MAX)
788 header->snap_names = kmalloc(snap_names_len, GFP_KERNEL);
789 if (!header->snap_names)
792 * Note that rbd_dev_v1_header_read() guarantees
793 * the ondisk buffer we're working with has
794 * snap_names_len bytes beyond the end of the
795 * snapshot id array, this memcpy() is safe.
797 memcpy(header->snap_names, &ondisk->snaps[snap_count],
800 /* Record each snapshot's size */
802 size = snap_count * sizeof (*header->snap_sizes);
803 header->snap_sizes = kmalloc(size, GFP_KERNEL);
804 if (!header->snap_sizes)
806 for (i = 0; i < snap_count; i++)
807 header->snap_sizes[i] =
808 le64_to_cpu(ondisk->snaps[i].image_size);
810 header->snap_names = NULL;
811 header->snap_sizes = NULL;
814 header->features = 0; /* No features support in v1 images */
815 header->obj_order = ondisk->options.order;
816 header->crypt_type = ondisk->options.crypt_type;
817 header->comp_type = ondisk->options.comp_type;
819 /* Allocate and fill in the snapshot context */
821 header->image_size = le64_to_cpu(ondisk->image_size);
823 header->snapc = rbd_snap_context_create(snap_count);
826 header->snapc->seq = le64_to_cpu(ondisk->snap_seq);
827 for (i = 0; i < snap_count; i++)
828 header->snapc->snaps[i] = le64_to_cpu(ondisk->snaps[i].id);
833 kfree(header->snap_sizes);
834 header->snap_sizes = NULL;
835 kfree(header->snap_names);
836 header->snap_names = NULL;
837 kfree(header->object_prefix);
838 header->object_prefix = NULL;
843 static const char *rbd_snap_name(struct rbd_device *rbd_dev, u64 snap_id)
845 struct rbd_snap *snap;
847 if (snap_id == CEPH_NOSNAP)
848 return RBD_SNAP_HEAD_NAME;
850 list_for_each_entry(snap, &rbd_dev->snaps, node)
851 if (snap_id == snap->id)
857 static struct rbd_snap *snap_by_name(struct rbd_device *rbd_dev,
858 const char *snap_name)
860 struct rbd_snap *snap;
862 list_for_each_entry(snap, &rbd_dev->snaps, node)
863 if (!strcmp(snap_name, snap->name))
869 static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
871 if (!memcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME,
872 sizeof (RBD_SNAP_HEAD_NAME))) {
873 rbd_dev->mapping.size = rbd_dev->header.image_size;
874 rbd_dev->mapping.features = rbd_dev->header.features;
876 struct rbd_snap *snap;
878 snap = snap_by_name(rbd_dev, rbd_dev->spec->snap_name);
881 rbd_dev->mapping.size = snap->size;
882 rbd_dev->mapping.features = snap->features;
883 rbd_dev->mapping.read_only = true;
889 static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
891 rbd_dev->mapping.size = 0;
892 rbd_dev->mapping.features = 0;
893 rbd_dev->mapping.read_only = true;
896 static void rbd_header_free(struct rbd_image_header *header)
898 kfree(header->object_prefix);
899 header->object_prefix = NULL;
900 kfree(header->snap_sizes);
901 header->snap_sizes = NULL;
902 kfree(header->snap_names);
903 header->snap_names = NULL;
904 rbd_snap_context_put(header->snapc);
905 header->snapc = NULL;
908 static const char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset)
914 name = kmalloc(MAX_OBJ_NAME_SIZE + 1, GFP_NOIO);
917 segment = offset >> rbd_dev->header.obj_order;
918 ret = snprintf(name, MAX_OBJ_NAME_SIZE + 1, "%s.%012llx",
919 rbd_dev->header.object_prefix, segment);
920 if (ret < 0 || ret > MAX_OBJ_NAME_SIZE) {
921 pr_err("error formatting segment name for #%llu (%d)\n",
930 static u64 rbd_segment_offset(struct rbd_device *rbd_dev, u64 offset)
932 u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
934 return offset & (segment_size - 1);
937 static u64 rbd_segment_length(struct rbd_device *rbd_dev,
938 u64 offset, u64 length)
940 u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
942 offset &= segment_size - 1;
944 rbd_assert(length <= U64_MAX - offset);
945 if (offset + length > segment_size)
946 length = segment_size - offset;
952 * returns the size of an object in the image
954 static u64 rbd_obj_bytes(struct rbd_image_header *header)
956 return 1 << header->obj_order;
963 static void bio_chain_put(struct bio *chain)
969 chain = chain->bi_next;
975 * zeros a bio chain, starting at specific offset
977 static void zero_bio_chain(struct bio *chain, int start_ofs)
986 bio_for_each_segment(bv, chain, i) {
987 if (pos + bv->bv_len > start_ofs) {
988 int remainder = max(start_ofs - pos, 0);
989 buf = bvec_kmap_irq(bv, &flags);
990 memset(buf + remainder, 0,
991 bv->bv_len - remainder);
992 bvec_kunmap_irq(buf, &flags);
997 chain = chain->bi_next;
1002 * similar to zero_bio_chain(), zeros data defined by a page array,
1003 * starting at the given byte offset from the start of the array and
1004 * continuing up to the given end offset. The pages array is
1005 * assumed to be big enough to hold all bytes up to the end.
1007 static void zero_pages(struct page **pages, u64 offset, u64 end)
1009 struct page **page = &pages[offset >> PAGE_SHIFT];
1011 rbd_assert(end > offset);
1012 rbd_assert(end - offset <= (u64)SIZE_MAX);
1013 while (offset < end) {
1016 unsigned long flags;
1019 page_offset = (size_t)(offset & ~PAGE_MASK);
1020 length = min(PAGE_SIZE - page_offset, (size_t)(end - offset));
1021 local_irq_save(flags);
1022 kaddr = kmap_atomic(*page);
1023 memset(kaddr + page_offset, 0, length);
1024 kunmap_atomic(kaddr);
1025 local_irq_restore(flags);
1033 * Clone a portion of a bio, starting at the given byte offset
1034 * and continuing for the number of bytes indicated.
1036 static struct bio *bio_clone_range(struct bio *bio_src,
1037 unsigned int offset,
1045 unsigned short end_idx;
1046 unsigned short vcnt;
1049 /* Handle the easy case for the caller */
1051 if (!offset && len == bio_src->bi_size)
1052 return bio_clone(bio_src, gfpmask);
1054 if (WARN_ON_ONCE(!len))
1056 if (WARN_ON_ONCE(len > bio_src->bi_size))
1058 if (WARN_ON_ONCE(offset > bio_src->bi_size - len))
1061 /* Find first affected segment... */
1064 __bio_for_each_segment(bv, bio_src, idx, 0) {
1065 if (resid < bv->bv_len)
1067 resid -= bv->bv_len;
1071 /* ...and the last affected segment */
1074 __bio_for_each_segment(bv, bio_src, end_idx, idx) {
1075 if (resid <= bv->bv_len)
1077 resid -= bv->bv_len;
1079 vcnt = end_idx - idx + 1;
1081 /* Build the clone */
1083 bio = bio_alloc(gfpmask, (unsigned int) vcnt);
1085 return NULL; /* ENOMEM */
1087 bio->bi_bdev = bio_src->bi_bdev;
1088 bio->bi_sector = bio_src->bi_sector + (offset >> SECTOR_SHIFT);
1089 bio->bi_rw = bio_src->bi_rw;
1090 bio->bi_flags |= 1 << BIO_CLONED;
1093 * Copy over our part of the bio_vec, then update the first
1094 * and last (or only) entries.
1096 memcpy(&bio->bi_io_vec[0], &bio_src->bi_io_vec[idx],
1097 vcnt * sizeof (struct bio_vec));
1098 bio->bi_io_vec[0].bv_offset += voff;
1100 bio->bi_io_vec[0].bv_len -= voff;
1101 bio->bi_io_vec[vcnt - 1].bv_len = resid;
1103 bio->bi_io_vec[0].bv_len = len;
1106 bio->bi_vcnt = vcnt;
1114 * Clone a portion of a bio chain, starting at the given byte offset
1115 * into the first bio in the source chain and continuing for the
1116 * number of bytes indicated. The result is another bio chain of
1117 * exactly the given length, or a null pointer on error.
1119 * The bio_src and offset parameters are both in-out. On entry they
1120 * refer to the first source bio and the offset into that bio where
1121 * the start of data to be cloned is located.
1123 * On return, bio_src is updated to refer to the bio in the source
1124 * chain that contains first un-cloned byte, and *offset will
1125 * contain the offset of that byte within that bio.
1127 static struct bio *bio_chain_clone_range(struct bio **bio_src,
1128 unsigned int *offset,
1132 struct bio *bi = *bio_src;
1133 unsigned int off = *offset;
1134 struct bio *chain = NULL;
1137 /* Build up a chain of clone bios up to the limit */
1139 if (!bi || off >= bi->bi_size || !len)
1140 return NULL; /* Nothing to clone */
1144 unsigned int bi_size;
1148 rbd_warn(NULL, "bio_chain exhausted with %u left", len);
1149 goto out_err; /* EINVAL; ran out of bio's */
1151 bi_size = min_t(unsigned int, bi->bi_size - off, len);
1152 bio = bio_clone_range(bi, off, bi_size, gfpmask);
1154 goto out_err; /* ENOMEM */
1157 end = &bio->bi_next;
1160 if (off == bi->bi_size) {
1171 bio_chain_put(chain);
1177 * The default/initial value for all object request flags is 0. For
1178 * each flag, once its value is set to 1 it is never reset to 0
1181 static void obj_request_img_data_set(struct rbd_obj_request *obj_request)
1183 if (test_and_set_bit(OBJ_REQ_IMG_DATA, &obj_request->flags)) {
1184 struct rbd_device *rbd_dev;
1186 rbd_dev = obj_request->img_request->rbd_dev;
1187 rbd_warn(rbd_dev, "obj_request %p already marked img_data\n",
1192 static bool obj_request_img_data_test(struct rbd_obj_request *obj_request)
1195 return test_bit(OBJ_REQ_IMG_DATA, &obj_request->flags) != 0;
1198 static void obj_request_done_set(struct rbd_obj_request *obj_request)
1200 if (test_and_set_bit(OBJ_REQ_DONE, &obj_request->flags)) {
1201 struct rbd_device *rbd_dev = NULL;
1203 if (obj_request_img_data_test(obj_request))
1204 rbd_dev = obj_request->img_request->rbd_dev;
1205 rbd_warn(rbd_dev, "obj_request %p already marked done\n",
1210 static bool obj_request_done_test(struct rbd_obj_request *obj_request)
1213 return test_bit(OBJ_REQ_DONE, &obj_request->flags) != 0;
1217 * This sets the KNOWN flag after (possibly) setting the EXISTS
1218 * flag. The latter is set based on the "exists" value provided.
1220 * Note that for our purposes once an object exists it never goes
1221 * away again. It's possible that the response from two existence
1222 * checks are separated by the creation of the target object, and
1223 * the first ("doesn't exist") response arrives *after* the second
1224 * ("does exist"). In that case we ignore the second one.
1226 static void obj_request_existence_set(struct rbd_obj_request *obj_request,
1230 set_bit(OBJ_REQ_EXISTS, &obj_request->flags);
1231 set_bit(OBJ_REQ_KNOWN, &obj_request->flags);
1235 static bool obj_request_known_test(struct rbd_obj_request *obj_request)
1238 return test_bit(OBJ_REQ_KNOWN, &obj_request->flags) != 0;
1241 static bool obj_request_exists_test(struct rbd_obj_request *obj_request)
1244 return test_bit(OBJ_REQ_EXISTS, &obj_request->flags) != 0;
1247 static void rbd_obj_request_get(struct rbd_obj_request *obj_request)
1249 dout("%s: obj %p (was %d)\n", __func__, obj_request,
1250 atomic_read(&obj_request->kref.refcount));
1251 kref_get(&obj_request->kref);
1254 static void rbd_obj_request_destroy(struct kref *kref);
1255 static void rbd_obj_request_put(struct rbd_obj_request *obj_request)
1257 rbd_assert(obj_request != NULL);
1258 dout("%s: obj %p (was %d)\n", __func__, obj_request,
1259 atomic_read(&obj_request->kref.refcount));
1260 kref_put(&obj_request->kref, rbd_obj_request_destroy);
1263 static void rbd_img_request_get(struct rbd_img_request *img_request)
1265 dout("%s: img %p (was %d)\n", __func__, img_request,
1266 atomic_read(&img_request->kref.refcount));
1267 kref_get(&img_request->kref);
1270 static void rbd_img_request_destroy(struct kref *kref);
1271 static void rbd_img_request_put(struct rbd_img_request *img_request)
1273 rbd_assert(img_request != NULL);
1274 dout("%s: img %p (was %d)\n", __func__, img_request,
1275 atomic_read(&img_request->kref.refcount));
1276 kref_put(&img_request->kref, rbd_img_request_destroy);
1279 static inline void rbd_img_obj_request_add(struct rbd_img_request *img_request,
1280 struct rbd_obj_request *obj_request)
1282 rbd_assert(obj_request->img_request == NULL);
1284 /* Image request now owns object's original reference */
1285 obj_request->img_request = img_request;
1286 obj_request->which = img_request->obj_request_count;
1287 rbd_assert(!obj_request_img_data_test(obj_request));
1288 obj_request_img_data_set(obj_request);
1289 rbd_assert(obj_request->which != BAD_WHICH);
1290 img_request->obj_request_count++;
1291 list_add_tail(&obj_request->links, &img_request->obj_requests);
1292 dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request,
1293 obj_request->which);
1296 static inline void rbd_img_obj_request_del(struct rbd_img_request *img_request,
1297 struct rbd_obj_request *obj_request)
1299 rbd_assert(obj_request->which != BAD_WHICH);
1301 dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request,
1302 obj_request->which);
1303 list_del(&obj_request->links);
1304 rbd_assert(img_request->obj_request_count > 0);
1305 img_request->obj_request_count--;
1306 rbd_assert(obj_request->which == img_request->obj_request_count);
1307 obj_request->which = BAD_WHICH;
1308 rbd_assert(obj_request_img_data_test(obj_request));
1309 rbd_assert(obj_request->img_request == img_request);
1310 obj_request->img_request = NULL;
1311 obj_request->callback = NULL;
1312 rbd_obj_request_put(obj_request);
1315 static bool obj_request_type_valid(enum obj_request_type type)
1318 case OBJ_REQUEST_NODATA:
1319 case OBJ_REQUEST_BIO:
1320 case OBJ_REQUEST_PAGES:
1327 static int rbd_obj_request_submit(struct ceph_osd_client *osdc,
1328 struct rbd_obj_request *obj_request)
1330 dout("%s: osdc %p obj %p\n", __func__, osdc, obj_request);
1332 return ceph_osdc_start_request(osdc, obj_request->osd_req, false);
1335 static void rbd_img_request_complete(struct rbd_img_request *img_request)
1338 dout("%s: img %p\n", __func__, img_request);
1341 * If no error occurred, compute the aggregate transfer
1342 * count for the image request. We could instead use
1343 * atomic64_cmpxchg() to update it as each object request
1344 * completes; not clear which way is better off hand.
1346 if (!img_request->result) {
1347 struct rbd_obj_request *obj_request;
1350 for_each_obj_request(img_request, obj_request)
1351 xferred += obj_request->xferred;
1352 img_request->xferred = xferred;
1355 if (img_request->callback)
1356 img_request->callback(img_request);
1358 rbd_img_request_put(img_request);
1361 /* Caller is responsible for rbd_obj_request_destroy(obj_request) */
1363 static int rbd_obj_request_wait(struct rbd_obj_request *obj_request)
1365 dout("%s: obj %p\n", __func__, obj_request);
1367 return wait_for_completion_interruptible(&obj_request->completion);
1371 * The default/initial value for all image request flags is 0. Each
1372 * is conditionally set to 1 at image request initialization time
1373 * and currently never change thereafter.
1375 static void img_request_write_set(struct rbd_img_request *img_request)
1377 set_bit(IMG_REQ_WRITE, &img_request->flags);
1381 static bool img_request_write_test(struct rbd_img_request *img_request)
1384 return test_bit(IMG_REQ_WRITE, &img_request->flags) != 0;
1387 static void img_request_child_set(struct rbd_img_request *img_request)
1389 set_bit(IMG_REQ_CHILD, &img_request->flags);
1393 static bool img_request_child_test(struct rbd_img_request *img_request)
1396 return test_bit(IMG_REQ_CHILD, &img_request->flags) != 0;
1399 static void img_request_layered_set(struct rbd_img_request *img_request)
1401 set_bit(IMG_REQ_LAYERED, &img_request->flags);
1405 static bool img_request_layered_test(struct rbd_img_request *img_request)
1408 return test_bit(IMG_REQ_LAYERED, &img_request->flags) != 0;
1412 rbd_img_obj_request_read_callback(struct rbd_obj_request *obj_request)
1414 u64 xferred = obj_request->xferred;
1415 u64 length = obj_request->length;
1417 dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
1418 obj_request, obj_request->img_request, obj_request->result,
1421 * ENOENT means a hole in the image. We zero-fill the
1422 * entire length of the request. A short read also implies
1423 * zero-fill to the end of the request. Either way we
1424 * update the xferred count to indicate the whole request
1427 rbd_assert(obj_request->type != OBJ_REQUEST_NODATA);
1428 if (obj_request->result == -ENOENT) {
1429 if (obj_request->type == OBJ_REQUEST_BIO)
1430 zero_bio_chain(obj_request->bio_list, 0);
1432 zero_pages(obj_request->pages, 0, length);
1433 obj_request->result = 0;
1434 obj_request->xferred = length;
1435 } else if (xferred < length && !obj_request->result) {
1436 if (obj_request->type == OBJ_REQUEST_BIO)
1437 zero_bio_chain(obj_request->bio_list, xferred);
1439 zero_pages(obj_request->pages, xferred, length);
1440 obj_request->xferred = length;
1442 obj_request_done_set(obj_request);
1445 static void rbd_obj_request_complete(struct rbd_obj_request *obj_request)
1447 dout("%s: obj %p cb %p\n", __func__, obj_request,
1448 obj_request->callback);
1449 if (obj_request->callback)
1450 obj_request->callback(obj_request);
1452 complete_all(&obj_request->completion);
1455 static void rbd_osd_trivial_callback(struct rbd_obj_request *obj_request)
1457 dout("%s: obj %p\n", __func__, obj_request);
1458 obj_request_done_set(obj_request);
1461 static void rbd_osd_read_callback(struct rbd_obj_request *obj_request)
1463 struct rbd_img_request *img_request = NULL;
1464 struct rbd_device *rbd_dev = NULL;
1465 bool layered = false;
1467 if (obj_request_img_data_test(obj_request)) {
1468 img_request = obj_request->img_request;
1469 layered = img_request && img_request_layered_test(img_request);
1470 rbd_dev = img_request->rbd_dev;
1473 dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
1474 obj_request, img_request, obj_request->result,
1475 obj_request->xferred, obj_request->length);
1476 if (layered && obj_request->result == -ENOENT &&
1477 obj_request->img_offset < rbd_dev->parent_overlap)
1478 rbd_img_parent_read(obj_request);
1479 else if (img_request)
1480 rbd_img_obj_request_read_callback(obj_request);
1482 obj_request_done_set(obj_request);
1485 static void rbd_osd_write_callback(struct rbd_obj_request *obj_request)
1487 dout("%s: obj %p result %d %llu\n", __func__, obj_request,
1488 obj_request->result, obj_request->length);
1490 * There is no such thing as a successful short write. Set
1491 * it to our originally-requested length.
1493 obj_request->xferred = obj_request->length;
1494 obj_request_done_set(obj_request);
1498 * For a simple stat call there's nothing to do. We'll do more if
1499 * this is part of a write sequence for a layered image.
1501 static void rbd_osd_stat_callback(struct rbd_obj_request *obj_request)
1503 dout("%s: obj %p\n", __func__, obj_request);
1504 obj_request_done_set(obj_request);
1507 static void rbd_osd_req_callback(struct ceph_osd_request *osd_req,
1508 struct ceph_msg *msg)
1510 struct rbd_obj_request *obj_request = osd_req->r_priv;
1513 dout("%s: osd_req %p msg %p\n", __func__, osd_req, msg);
1514 rbd_assert(osd_req == obj_request->osd_req);
1515 if (obj_request_img_data_test(obj_request)) {
1516 rbd_assert(obj_request->img_request);
1517 rbd_assert(obj_request->which != BAD_WHICH);
1519 rbd_assert(obj_request->which == BAD_WHICH);
1522 if (osd_req->r_result < 0)
1523 obj_request->result = osd_req->r_result;
1524 obj_request->version = le64_to_cpu(osd_req->r_reassert_version.version);
1526 BUG_ON(osd_req->r_num_ops > 2);
1529 * We support a 64-bit length, but ultimately it has to be
1530 * passed to blk_end_request(), which takes an unsigned int.
1532 obj_request->xferred = osd_req->r_reply_op_len[0];
1533 rbd_assert(obj_request->xferred < (u64)UINT_MAX);
1534 opcode = osd_req->r_ops[0].op;
1536 case CEPH_OSD_OP_READ:
1537 rbd_osd_read_callback(obj_request);
1539 case CEPH_OSD_OP_WRITE:
1540 rbd_osd_write_callback(obj_request);
1542 case CEPH_OSD_OP_STAT:
1543 rbd_osd_stat_callback(obj_request);
1545 case CEPH_OSD_OP_CALL:
1546 case CEPH_OSD_OP_NOTIFY_ACK:
1547 case CEPH_OSD_OP_WATCH:
1548 rbd_osd_trivial_callback(obj_request);
1551 rbd_warn(NULL, "%s: unsupported op %hu\n",
1552 obj_request->object_name, (unsigned short) opcode);
1556 if (obj_request_done_test(obj_request))
1557 rbd_obj_request_complete(obj_request);
1560 static void rbd_osd_req_format_read(struct rbd_obj_request *obj_request)
1562 struct rbd_img_request *img_request = obj_request->img_request;
1563 struct ceph_osd_request *osd_req = obj_request->osd_req;
1566 rbd_assert(osd_req != NULL);
1568 snap_id = img_request ? img_request->snap_id : CEPH_NOSNAP;
1569 ceph_osdc_build_request(osd_req, obj_request->offset,
1570 NULL, snap_id, NULL);
1573 static void rbd_osd_req_format_write(struct rbd_obj_request *obj_request)
1575 struct rbd_img_request *img_request = obj_request->img_request;
1576 struct ceph_osd_request *osd_req = obj_request->osd_req;
1577 struct ceph_snap_context *snapc;
1578 struct timespec mtime = CURRENT_TIME;
1580 rbd_assert(osd_req != NULL);
1582 snapc = img_request ? img_request->snapc : NULL;
1583 ceph_osdc_build_request(osd_req, obj_request->offset,
1584 snapc, CEPH_NOSNAP, &mtime);
1587 static struct ceph_osd_request *rbd_osd_req_create(
1588 struct rbd_device *rbd_dev,
1590 struct rbd_obj_request *obj_request)
1592 struct ceph_snap_context *snapc = NULL;
1593 struct ceph_osd_client *osdc;
1594 struct ceph_osd_request *osd_req;
1596 if (obj_request_img_data_test(obj_request)) {
1597 struct rbd_img_request *img_request = obj_request->img_request;
1599 rbd_assert(write_request ==
1600 img_request_write_test(img_request));
1602 snapc = img_request->snapc;
1605 /* Allocate and initialize the request, for the single op */
1607 osdc = &rbd_dev->rbd_client->client->osdc;
1608 osd_req = ceph_osdc_alloc_request(osdc, snapc, 1, false, GFP_ATOMIC);
1610 return NULL; /* ENOMEM */
1613 osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK;
1615 osd_req->r_flags = CEPH_OSD_FLAG_READ;
1617 osd_req->r_callback = rbd_osd_req_callback;
1618 osd_req->r_priv = obj_request;
1620 osd_req->r_oid_len = strlen(obj_request->object_name);
1621 rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid));
1622 memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len);
1624 osd_req->r_file_layout = rbd_dev->layout; /* struct */
1630 * Create a copyup osd request based on the information in the
1631 * object request supplied. A copyup request has two osd ops,
1632 * a copyup method call, and a "normal" write request.
1634 static struct ceph_osd_request *
1635 rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request)
1637 struct rbd_img_request *img_request;
1638 struct ceph_snap_context *snapc;
1639 struct rbd_device *rbd_dev;
1640 struct ceph_osd_client *osdc;
1641 struct ceph_osd_request *osd_req;
1643 rbd_assert(obj_request_img_data_test(obj_request));
1644 img_request = obj_request->img_request;
1645 rbd_assert(img_request);
1646 rbd_assert(img_request_write_test(img_request));
1648 /* Allocate and initialize the request, for the two ops */
1650 snapc = img_request->snapc;
1651 rbd_dev = img_request->rbd_dev;
1652 osdc = &rbd_dev->rbd_client->client->osdc;
1653 osd_req = ceph_osdc_alloc_request(osdc, snapc, 2, false, GFP_ATOMIC);
1655 return NULL; /* ENOMEM */
1657 osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK;
1658 osd_req->r_callback = rbd_osd_req_callback;
1659 osd_req->r_priv = obj_request;
1661 osd_req->r_oid_len = strlen(obj_request->object_name);
1662 rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid));
1663 memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len);
1665 osd_req->r_file_layout = rbd_dev->layout; /* struct */
1671 static void rbd_osd_req_destroy(struct ceph_osd_request *osd_req)
1673 ceph_osdc_put_request(osd_req);
1676 /* object_name is assumed to be a non-null pointer and NUL-terminated */
1678 static struct rbd_obj_request *rbd_obj_request_create(const char *object_name,
1679 u64 offset, u64 length,
1680 enum obj_request_type type)
1682 struct rbd_obj_request *obj_request;
1686 rbd_assert(obj_request_type_valid(type));
1688 size = strlen(object_name) + 1;
1689 obj_request = kzalloc(sizeof (*obj_request) + size, GFP_KERNEL);
1693 name = (char *)(obj_request + 1);
1694 obj_request->object_name = memcpy(name, object_name, size);
1695 obj_request->offset = offset;
1696 obj_request->length = length;
1697 obj_request->flags = 0;
1698 obj_request->which = BAD_WHICH;
1699 obj_request->type = type;
1700 INIT_LIST_HEAD(&obj_request->links);
1701 init_completion(&obj_request->completion);
1702 kref_init(&obj_request->kref);
1704 dout("%s: \"%s\" %llu/%llu %d -> obj %p\n", __func__, object_name,
1705 offset, length, (int)type, obj_request);
1710 static void rbd_obj_request_destroy(struct kref *kref)
1712 struct rbd_obj_request *obj_request;
1714 obj_request = container_of(kref, struct rbd_obj_request, kref);
1716 dout("%s: obj %p\n", __func__, obj_request);
1718 rbd_assert(obj_request->img_request == NULL);
1719 rbd_assert(obj_request->which == BAD_WHICH);
1721 if (obj_request->osd_req)
1722 rbd_osd_req_destroy(obj_request->osd_req);
1724 rbd_assert(obj_request_type_valid(obj_request->type));
1725 switch (obj_request->type) {
1726 case OBJ_REQUEST_NODATA:
1727 break; /* Nothing to do */
1728 case OBJ_REQUEST_BIO:
1729 if (obj_request->bio_list)
1730 bio_chain_put(obj_request->bio_list);
1732 case OBJ_REQUEST_PAGES:
1733 if (obj_request->pages)
1734 ceph_release_page_vector(obj_request->pages,
1735 obj_request->page_count);
1743 * Caller is responsible for filling in the list of object requests
1744 * that comprises the image request, and the Linux request pointer
1745 * (if there is one).
1747 static struct rbd_img_request *rbd_img_request_create(
1748 struct rbd_device *rbd_dev,
1749 u64 offset, u64 length,
1753 struct rbd_img_request *img_request;
1755 img_request = kmalloc(sizeof (*img_request), GFP_ATOMIC);
1759 if (write_request) {
1760 down_read(&rbd_dev->header_rwsem);
1761 rbd_snap_context_get(rbd_dev->header.snapc);
1762 up_read(&rbd_dev->header_rwsem);
1765 img_request->rq = NULL;
1766 img_request->rbd_dev = rbd_dev;
1767 img_request->offset = offset;
1768 img_request->length = length;
1769 img_request->flags = 0;
1770 if (write_request) {
1771 img_request_write_set(img_request);
1772 img_request->snapc = rbd_dev->header.snapc;
1774 img_request->snap_id = rbd_dev->spec->snap_id;
1777 img_request_child_set(img_request);
1778 if (rbd_dev->parent_spec)
1779 img_request_layered_set(img_request);
1780 spin_lock_init(&img_request->completion_lock);
1781 img_request->next_completion = 0;
1782 img_request->callback = NULL;
1783 img_request->result = 0;
1784 img_request->obj_request_count = 0;
1785 INIT_LIST_HEAD(&img_request->obj_requests);
1786 kref_init(&img_request->kref);
1788 rbd_img_request_get(img_request); /* Avoid a warning */
1789 rbd_img_request_put(img_request); /* TEMPORARY */
1791 dout("%s: rbd_dev %p %s %llu/%llu -> img %p\n", __func__, rbd_dev,
1792 write_request ? "write" : "read", offset, length,
1798 static void rbd_img_request_destroy(struct kref *kref)
1800 struct rbd_img_request *img_request;
1801 struct rbd_obj_request *obj_request;
1802 struct rbd_obj_request *next_obj_request;
1804 img_request = container_of(kref, struct rbd_img_request, kref);
1806 dout("%s: img %p\n", __func__, img_request);
1808 for_each_obj_request_safe(img_request, obj_request, next_obj_request)
1809 rbd_img_obj_request_del(img_request, obj_request);
1810 rbd_assert(img_request->obj_request_count == 0);
1812 if (img_request_write_test(img_request))
1813 rbd_snap_context_put(img_request->snapc);
1815 if (img_request_child_test(img_request))
1816 rbd_obj_request_put(img_request->obj_request);
1821 static bool rbd_img_obj_end_request(struct rbd_obj_request *obj_request)
1823 struct rbd_img_request *img_request;
1824 unsigned int xferred;
1828 rbd_assert(obj_request_img_data_test(obj_request));
1829 img_request = obj_request->img_request;
1831 rbd_assert(obj_request->xferred <= (u64)UINT_MAX);
1832 xferred = (unsigned int)obj_request->xferred;
1833 result = obj_request->result;
1835 struct rbd_device *rbd_dev = img_request->rbd_dev;
1837 rbd_warn(rbd_dev, "%s %llx at %llx (%llx)\n",
1838 img_request_write_test(img_request) ? "write" : "read",
1839 obj_request->length, obj_request->img_offset,
1840 obj_request->offset);
1841 rbd_warn(rbd_dev, " result %d xferred %x\n",
1843 if (!img_request->result)
1844 img_request->result = result;
1847 /* Image object requests don't own their page array */
1849 if (obj_request->type == OBJ_REQUEST_PAGES) {
1850 obj_request->pages = NULL;
1851 obj_request->page_count = 0;
1854 if (img_request_child_test(img_request)) {
1855 rbd_assert(img_request->obj_request != NULL);
1856 more = obj_request->which < img_request->obj_request_count - 1;
1858 rbd_assert(img_request->rq != NULL);
1859 more = blk_end_request(img_request->rq, result, xferred);
1865 static void rbd_img_obj_callback(struct rbd_obj_request *obj_request)
1867 struct rbd_img_request *img_request;
1868 u32 which = obj_request->which;
1871 rbd_assert(obj_request_img_data_test(obj_request));
1872 img_request = obj_request->img_request;
1874 dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
1875 rbd_assert(img_request != NULL);
1876 rbd_assert(img_request->obj_request_count > 0);
1877 rbd_assert(which != BAD_WHICH);
1878 rbd_assert(which < img_request->obj_request_count);
1879 rbd_assert(which >= img_request->next_completion);
1881 spin_lock_irq(&img_request->completion_lock);
1882 if (which != img_request->next_completion)
1885 for_each_obj_request_from(img_request, obj_request) {
1887 rbd_assert(which < img_request->obj_request_count);
1889 if (!obj_request_done_test(obj_request))
1891 more = rbd_img_obj_end_request(obj_request);
1895 rbd_assert(more ^ (which == img_request->obj_request_count));
1896 img_request->next_completion = which;
1898 spin_unlock_irq(&img_request->completion_lock);
1901 rbd_img_request_complete(img_request);
1905 * Split up an image request into one or more object requests, each
1906 * to a different object. The "type" parameter indicates whether
1907 * "data_desc" is the pointer to the head of a list of bio
1908 * structures, or the base of a page array. In either case this
1909 * function assumes data_desc describes memory sufficient to hold
1910 * all data described by the image request.
1912 static int rbd_img_request_fill(struct rbd_img_request *img_request,
1913 enum obj_request_type type,
1916 struct rbd_device *rbd_dev = img_request->rbd_dev;
1917 struct rbd_obj_request *obj_request = NULL;
1918 struct rbd_obj_request *next_obj_request;
1919 bool write_request = img_request_write_test(img_request);
1920 struct bio *bio_list;
1921 unsigned int bio_offset = 0;
1922 struct page **pages;
1927 dout("%s: img %p type %d data_desc %p\n", __func__, img_request,
1928 (int)type, data_desc);
1930 opcode = write_request ? CEPH_OSD_OP_WRITE : CEPH_OSD_OP_READ;
1931 img_offset = img_request->offset;
1932 resid = img_request->length;
1933 rbd_assert(resid > 0);
1935 if (type == OBJ_REQUEST_BIO) {
1936 bio_list = data_desc;
1937 rbd_assert(img_offset == bio_list->bi_sector << SECTOR_SHIFT);
1939 rbd_assert(type == OBJ_REQUEST_PAGES);
1944 struct ceph_osd_request *osd_req;
1945 const char *object_name;
1949 object_name = rbd_segment_name(rbd_dev, img_offset);
1952 offset = rbd_segment_offset(rbd_dev, img_offset);
1953 length = rbd_segment_length(rbd_dev, img_offset, resid);
1954 obj_request = rbd_obj_request_create(object_name,
1955 offset, length, type);
1956 kfree(object_name); /* object request has its own copy */
1960 if (type == OBJ_REQUEST_BIO) {
1961 unsigned int clone_size;
1963 rbd_assert(length <= (u64)UINT_MAX);
1964 clone_size = (unsigned int)length;
1965 obj_request->bio_list =
1966 bio_chain_clone_range(&bio_list,
1970 if (!obj_request->bio_list)
1973 unsigned int page_count;
1975 obj_request->pages = pages;
1976 page_count = (u32)calc_pages_for(offset, length);
1977 obj_request->page_count = page_count;
1978 if ((offset + length) & ~PAGE_MASK)
1979 page_count--; /* more on last page */
1980 pages += page_count;
1983 osd_req = rbd_osd_req_create(rbd_dev, write_request,
1987 obj_request->osd_req = osd_req;
1988 obj_request->callback = rbd_img_obj_callback;
1990 osd_req_op_extent_init(osd_req, 0, opcode, offset, length,
1992 if (type == OBJ_REQUEST_BIO)
1993 osd_req_op_extent_osd_data_bio(osd_req, 0,
1994 obj_request->bio_list, length);
1996 osd_req_op_extent_osd_data_pages(osd_req, 0,
1997 obj_request->pages, length,
1998 offset & ~PAGE_MASK, false, false);
2001 rbd_osd_req_format_write(obj_request);
2003 rbd_osd_req_format_read(obj_request);
2005 obj_request->img_offset = img_offset;
2006 rbd_img_obj_request_add(img_request, obj_request);
2008 img_offset += length;
2015 rbd_obj_request_put(obj_request);
2017 for_each_obj_request_safe(img_request, obj_request, next_obj_request)
2018 rbd_obj_request_put(obj_request);
2024 rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
2026 struct rbd_img_request *img_request;
2027 struct rbd_device *rbd_dev;
2031 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
2032 rbd_assert(obj_request_img_data_test(obj_request));
2033 img_request = obj_request->img_request;
2034 rbd_assert(img_request);
2036 rbd_dev = img_request->rbd_dev;
2037 rbd_assert(rbd_dev);
2038 length = (u64)1 << rbd_dev->header.obj_order;
2039 page_count = (u32)calc_pages_for(0, length);
2041 rbd_assert(obj_request->copyup_pages);
2042 ceph_release_page_vector(obj_request->copyup_pages, page_count);
2043 obj_request->copyup_pages = NULL;
2046 * We want the transfer count to reflect the size of the
2047 * original write request. There is no such thing as a
2048 * successful short write, so if the request was successful
2049 * we can just set it to the originally-requested length.
2051 if (!obj_request->result)
2052 obj_request->xferred = obj_request->length;
2054 /* Finish up with the normal image object callback */
2056 rbd_img_obj_callback(obj_request);
2060 rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
2062 struct rbd_obj_request *orig_request;
2063 struct ceph_osd_request *osd_req;
2064 struct ceph_osd_client *osdc;
2065 struct rbd_device *rbd_dev;
2066 struct page **pages;
2071 rbd_assert(img_request_child_test(img_request));
2073 /* First get what we need from the image request */
2075 pages = img_request->copyup_pages;
2076 rbd_assert(pages != NULL);
2077 img_request->copyup_pages = NULL;
2079 orig_request = img_request->obj_request;
2080 rbd_assert(orig_request != NULL);
2081 rbd_assert(orig_request->type == OBJ_REQUEST_BIO);
2082 result = img_request->result;
2083 obj_size = img_request->length;
2084 xferred = img_request->xferred;
2086 rbd_dev = img_request->rbd_dev;
2087 rbd_assert(rbd_dev);
2088 rbd_assert(obj_size == (u64)1 << rbd_dev->header.obj_order);
2090 rbd_img_request_put(img_request);
2095 /* Allocate the new copyup osd request for the original request */
2098 rbd_assert(!orig_request->osd_req);
2099 osd_req = rbd_osd_req_create_copyup(orig_request);
2102 orig_request->osd_req = osd_req;
2103 orig_request->copyup_pages = pages;
2105 /* Initialize the copyup op */
2107 osd_req_op_cls_init(osd_req, 0, CEPH_OSD_OP_CALL, "rbd", "copyup");
2108 osd_req_op_cls_request_data_pages(osd_req, 0, pages, obj_size, 0,
2111 /* Then the original write request op */
2113 osd_req_op_extent_init(osd_req, 1, CEPH_OSD_OP_WRITE,
2114 orig_request->offset,
2115 orig_request->length, 0, 0);
2116 osd_req_op_extent_osd_data_bio(osd_req, 1, orig_request->bio_list,
2117 orig_request->length);
2119 rbd_osd_req_format_write(orig_request);
2121 /* All set, send it off. */
2123 orig_request->callback = rbd_img_obj_copyup_callback;
2124 osdc = &rbd_dev->rbd_client->client->osdc;
2125 result = rbd_obj_request_submit(osdc, orig_request);
2129 /* Record the error code and complete the request */
2131 orig_request->result = result;
2132 orig_request->xferred = 0;
2133 obj_request_done_set(orig_request);
2134 rbd_obj_request_complete(orig_request);
2138 * Read from the parent image the range of data that covers the
2139 * entire target of the given object request. This is used for
2140 * satisfying a layered image write request when the target of an
2141 * object request from the image request does not exist.
2143 * A page array big enough to hold the returned data is allocated
2144 * and supplied to rbd_img_request_fill() as the "data descriptor."
2145 * When the read completes, this page array will be transferred to
2146 * the original object request for the copyup operation.
2148 * If an error occurs, record it as the result of the original
2149 * object request and mark it done so it gets completed.
2151 static int rbd_img_obj_parent_read_full(struct rbd_obj_request *obj_request)
2153 struct rbd_img_request *img_request = NULL;
2154 struct rbd_img_request *parent_request = NULL;
2155 struct rbd_device *rbd_dev;
2158 struct page **pages = NULL;
2162 rbd_assert(obj_request_img_data_test(obj_request));
2163 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
2165 img_request = obj_request->img_request;
2166 rbd_assert(img_request != NULL);
2167 rbd_dev = img_request->rbd_dev;
2168 rbd_assert(rbd_dev->parent != NULL);
2171 * First things first. The original osd request is of no
2172 * use to use any more, we'll need a new one that can hold
2173 * the two ops in a copyup request. We'll get that later,
2174 * but for now we can release the old one.
2176 rbd_osd_req_destroy(obj_request->osd_req);
2177 obj_request->osd_req = NULL;
2180 * Determine the byte range covered by the object in the
2181 * child image to which the original request was to be sent.
2183 img_offset = obj_request->img_offset - obj_request->offset;
2184 length = (u64)1 << rbd_dev->header.obj_order;
2187 * There is no defined parent data beyond the parent
2188 * overlap, so limit what we read at that boundary if
2191 if (img_offset + length > rbd_dev->parent_overlap) {
2192 rbd_assert(img_offset < rbd_dev->parent_overlap);
2193 length = rbd_dev->parent_overlap - img_offset;
2197 * Allocate a page array big enough to receive the data read
2200 page_count = (u32)calc_pages_for(0, length);
2201 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2202 if (IS_ERR(pages)) {
2203 result = PTR_ERR(pages);
2209 parent_request = rbd_img_request_create(rbd_dev->parent,
2212 if (!parent_request)
2214 rbd_obj_request_get(obj_request);
2215 parent_request->obj_request = obj_request;
2217 result = rbd_img_request_fill(parent_request, OBJ_REQUEST_PAGES, pages);
2220 parent_request->copyup_pages = pages;
2222 parent_request->callback = rbd_img_obj_parent_read_full_callback;
2223 result = rbd_img_request_submit(parent_request);
2227 parent_request->copyup_pages = NULL;
2228 parent_request->obj_request = NULL;
2229 rbd_obj_request_put(obj_request);
2232 ceph_release_page_vector(pages, page_count);
2234 rbd_img_request_put(parent_request);
2235 obj_request->result = result;
2236 obj_request->xferred = 0;
2237 obj_request_done_set(obj_request);
2242 static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
2244 struct rbd_obj_request *orig_request;
2247 rbd_assert(!obj_request_img_data_test(obj_request));
2250 * All we need from the object request is the original
2251 * request and the result of the STAT op. Grab those, then
2252 * we're done with the request.
2254 orig_request = obj_request->obj_request;
2255 obj_request->obj_request = NULL;
2256 rbd_assert(orig_request);
2257 rbd_assert(orig_request->img_request);
2259 result = obj_request->result;
2260 obj_request->result = 0;
2262 dout("%s: obj %p for obj %p result %d %llu/%llu\n", __func__,
2263 obj_request, orig_request, result,
2264 obj_request->xferred, obj_request->length);
2265 rbd_obj_request_put(obj_request);
2267 rbd_assert(orig_request);
2268 rbd_assert(orig_request->img_request);
2271 * Our only purpose here is to determine whether the object
2272 * exists, and we don't want to treat the non-existence as
2273 * an error. If something else comes back, transfer the
2274 * error to the original request and complete it now.
2277 obj_request_existence_set(orig_request, true);
2278 } else if (result == -ENOENT) {
2279 obj_request_existence_set(orig_request, false);
2280 } else if (result) {
2281 orig_request->result = result;
2286 * Resubmit the original request now that we have recorded
2287 * whether the target object exists.
2289 orig_request->result = rbd_img_obj_request_submit(orig_request);
2291 if (orig_request->result)
2292 rbd_obj_request_complete(orig_request);
2293 rbd_obj_request_put(orig_request);
2296 static int rbd_img_obj_exists_submit(struct rbd_obj_request *obj_request)
2298 struct rbd_obj_request *stat_request;
2299 struct rbd_device *rbd_dev;
2300 struct ceph_osd_client *osdc;
2301 struct page **pages = NULL;
2307 * The response data for a STAT call consists of:
2314 size = sizeof (__le64) + sizeof (__le32) + sizeof (__le32);
2315 page_count = (u32)calc_pages_for(0, size);
2316 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2318 return PTR_ERR(pages);
2321 stat_request = rbd_obj_request_create(obj_request->object_name, 0, 0,
2326 rbd_obj_request_get(obj_request);
2327 stat_request->obj_request = obj_request;
2328 stat_request->pages = pages;
2329 stat_request->page_count = page_count;
2331 rbd_assert(obj_request->img_request);
2332 rbd_dev = obj_request->img_request->rbd_dev;
2333 stat_request->osd_req = rbd_osd_req_create(rbd_dev, false,
2335 if (!stat_request->osd_req)
2337 stat_request->callback = rbd_img_obj_exists_callback;
2339 osd_req_op_init(stat_request->osd_req, 0, CEPH_OSD_OP_STAT);
2340 osd_req_op_raw_data_in_pages(stat_request->osd_req, 0, pages, size, 0,
2342 rbd_osd_req_format_read(stat_request);
2344 osdc = &rbd_dev->rbd_client->client->osdc;
2345 ret = rbd_obj_request_submit(osdc, stat_request);
2348 rbd_obj_request_put(obj_request);
2353 static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request)
2355 struct rbd_img_request *img_request;
2356 struct rbd_device *rbd_dev;
2359 rbd_assert(obj_request_img_data_test(obj_request));
2361 img_request = obj_request->img_request;
2362 rbd_assert(img_request);
2363 rbd_dev = img_request->rbd_dev;
2366 * Only writes to layered images need special handling.
2367 * Reads and non-layered writes are simple object requests.
2368 * Layered writes that start beyond the end of the overlap
2369 * with the parent have no parent data, so they too are
2370 * simple object requests. Finally, if the target object is
2371 * known to already exist, its parent data has already been
2372 * copied, so a write to the object can also be handled as a
2373 * simple object request.
2375 if (!img_request_write_test(img_request) ||
2376 !img_request_layered_test(img_request) ||
2377 rbd_dev->parent_overlap <= obj_request->img_offset ||
2378 ((known = obj_request_known_test(obj_request)) &&
2379 obj_request_exists_test(obj_request))) {
2381 struct rbd_device *rbd_dev;
2382 struct ceph_osd_client *osdc;
2384 rbd_dev = obj_request->img_request->rbd_dev;
2385 osdc = &rbd_dev->rbd_client->client->osdc;
2387 return rbd_obj_request_submit(osdc, obj_request);
2391 * It's a layered write. The target object might exist but
2392 * we may not know that yet. If we know it doesn't exist,
2393 * start by reading the data for the full target object from
2394 * the parent so we can use it for a copyup to the target.
2397 return rbd_img_obj_parent_read_full(obj_request);
2399 /* We don't know whether the target exists. Go find out. */
2401 return rbd_img_obj_exists_submit(obj_request);
2404 static int rbd_img_request_submit(struct rbd_img_request *img_request)
2406 struct rbd_obj_request *obj_request;
2407 struct rbd_obj_request *next_obj_request;
2409 dout("%s: img %p\n", __func__, img_request);
2410 for_each_obj_request_safe(img_request, obj_request, next_obj_request) {
2413 ret = rbd_img_obj_request_submit(obj_request);
2421 static void rbd_img_parent_read_callback(struct rbd_img_request *img_request)
2423 struct rbd_obj_request *obj_request;
2424 struct rbd_device *rbd_dev;
2427 rbd_assert(img_request_child_test(img_request));
2429 obj_request = img_request->obj_request;
2430 rbd_assert(obj_request);
2431 rbd_assert(obj_request->img_request);
2433 obj_request->result = img_request->result;
2434 if (obj_request->result)
2438 * We need to zero anything beyond the parent overlap
2439 * boundary. Since rbd_img_obj_request_read_callback()
2440 * will zero anything beyond the end of a short read, an
2441 * easy way to do this is to pretend the data from the
2442 * parent came up short--ending at the overlap boundary.
2444 rbd_assert(obj_request->img_offset < U64_MAX - obj_request->length);
2445 obj_end = obj_request->img_offset + obj_request->length;
2446 rbd_dev = obj_request->img_request->rbd_dev;
2447 if (obj_end > rbd_dev->parent_overlap) {
2450 if (obj_request->img_offset < rbd_dev->parent_overlap)
2451 xferred = rbd_dev->parent_overlap -
2452 obj_request->img_offset;
2454 obj_request->xferred = min(img_request->xferred, xferred);
2456 obj_request->xferred = img_request->xferred;
2459 rbd_img_obj_request_read_callback(obj_request);
2460 rbd_obj_request_complete(obj_request);
2463 static void rbd_img_parent_read(struct rbd_obj_request *obj_request)
2465 struct rbd_device *rbd_dev;
2466 struct rbd_img_request *img_request;
2469 rbd_assert(obj_request_img_data_test(obj_request));
2470 rbd_assert(obj_request->img_request != NULL);
2471 rbd_assert(obj_request->result == (s32) -ENOENT);
2472 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
2474 rbd_dev = obj_request->img_request->rbd_dev;
2475 rbd_assert(rbd_dev->parent != NULL);
2476 /* rbd_read_finish(obj_request, obj_request->length); */
2477 img_request = rbd_img_request_create(rbd_dev->parent,
2478 obj_request->img_offset,
2479 obj_request->length,
2485 rbd_obj_request_get(obj_request);
2486 img_request->obj_request = obj_request;
2488 result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
2489 obj_request->bio_list);
2493 img_request->callback = rbd_img_parent_read_callback;
2494 result = rbd_img_request_submit(img_request);
2501 rbd_img_request_put(img_request);
2502 obj_request->result = result;
2503 obj_request->xferred = 0;
2504 obj_request_done_set(obj_request);
2507 static int rbd_obj_notify_ack(struct rbd_device *rbd_dev,
2508 u64 ver, u64 notify_id)
2510 struct rbd_obj_request *obj_request;
2511 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2514 obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
2515 OBJ_REQUEST_NODATA);
2520 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2521 if (!obj_request->osd_req)
2523 obj_request->callback = rbd_obj_request_put;
2525 osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_NOTIFY_ACK,
2527 rbd_osd_req_format_read(obj_request);
2529 ret = rbd_obj_request_submit(osdc, obj_request);
2532 rbd_obj_request_put(obj_request);
2537 static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
2539 struct rbd_device *rbd_dev = (struct rbd_device *)data;
2545 dout("%s: \"%s\" notify_id %llu opcode %u\n", __func__,
2546 rbd_dev->header_name, (unsigned long long) notify_id,
2547 (unsigned int) opcode);
2548 (void)rbd_dev_refresh(rbd_dev, &hver);
2550 rbd_obj_notify_ack(rbd_dev, hver, notify_id);
2554 * Request sync osd watch/unwatch. The value of "start" determines
2555 * whether a watch request is being initiated or torn down.
2557 static int rbd_dev_header_watch_sync(struct rbd_device *rbd_dev, int start)
2559 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2560 struct rbd_obj_request *obj_request;
2563 rbd_assert(start ^ !!rbd_dev->watch_event);
2564 rbd_assert(start ^ !!rbd_dev->watch_request);
2567 ret = ceph_osdc_create_event(osdc, rbd_watch_cb, rbd_dev,
2568 &rbd_dev->watch_event);
2571 rbd_assert(rbd_dev->watch_event != NULL);
2575 obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
2576 OBJ_REQUEST_NODATA);
2580 obj_request->osd_req = rbd_osd_req_create(rbd_dev, true, obj_request);
2581 if (!obj_request->osd_req)
2585 ceph_osdc_set_request_linger(osdc, obj_request->osd_req);
2587 ceph_osdc_unregister_linger_request(osdc,
2588 rbd_dev->watch_request->osd_req);
2590 osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH,
2591 rbd_dev->watch_event->cookie,
2592 rbd_dev->header.obj_version, start);
2593 rbd_osd_req_format_write(obj_request);
2595 ret = rbd_obj_request_submit(osdc, obj_request);
2598 ret = rbd_obj_request_wait(obj_request);
2601 ret = obj_request->result;
2606 * A watch request is set to linger, so the underlying osd
2607 * request won't go away until we unregister it. We retain
2608 * a pointer to the object request during that time (in
2609 * rbd_dev->watch_request), so we'll keep a reference to
2610 * it. We'll drop that reference (below) after we've
2614 rbd_dev->watch_request = obj_request;
2619 /* We have successfully torn down the watch request */
2621 rbd_obj_request_put(rbd_dev->watch_request);
2622 rbd_dev->watch_request = NULL;
2624 /* Cancel the event if we're tearing down, or on error */
2625 ceph_osdc_cancel_event(rbd_dev->watch_event);
2626 rbd_dev->watch_event = NULL;
2628 rbd_obj_request_put(obj_request);
2634 * Synchronous osd object method call. Returns the number of bytes
2635 * returned in the outbound buffer, or a negative error code.
2637 static int rbd_obj_method_sync(struct rbd_device *rbd_dev,
2638 const char *object_name,
2639 const char *class_name,
2640 const char *method_name,
2641 const void *outbound,
2642 size_t outbound_size,
2644 size_t inbound_size,
2647 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2648 struct rbd_obj_request *obj_request;
2649 struct page **pages;
2654 * Method calls are ultimately read operations. The result
2655 * should placed into the inbound buffer provided. They
2656 * also supply outbound data--parameters for the object
2657 * method. Currently if this is present it will be a
2660 page_count = (u32)calc_pages_for(0, inbound_size);
2661 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2663 return PTR_ERR(pages);
2666 obj_request = rbd_obj_request_create(object_name, 0, inbound_size,
2671 obj_request->pages = pages;
2672 obj_request->page_count = page_count;
2674 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2675 if (!obj_request->osd_req)
2678 osd_req_op_cls_init(obj_request->osd_req, 0, CEPH_OSD_OP_CALL,
2679 class_name, method_name);
2680 if (outbound_size) {
2681 struct ceph_pagelist *pagelist;
2683 pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
2687 ceph_pagelist_init(pagelist);
2688 ceph_pagelist_append(pagelist, outbound, outbound_size);
2689 osd_req_op_cls_request_data_pagelist(obj_request->osd_req, 0,
2692 osd_req_op_cls_response_data_pages(obj_request->osd_req, 0,
2693 obj_request->pages, inbound_size,
2695 rbd_osd_req_format_read(obj_request);
2697 ret = rbd_obj_request_submit(osdc, obj_request);
2700 ret = rbd_obj_request_wait(obj_request);
2704 ret = obj_request->result;
2708 rbd_assert(obj_request->xferred < (u64)INT_MAX);
2709 ret = (int)obj_request->xferred;
2710 ceph_copy_from_page_vector(pages, inbound, 0, obj_request->xferred);
2712 *version = obj_request->version;
2715 rbd_obj_request_put(obj_request);
2717 ceph_release_page_vector(pages, page_count);
2722 static void rbd_request_fn(struct request_queue *q)
2723 __releases(q->queue_lock) __acquires(q->queue_lock)
2725 struct rbd_device *rbd_dev = q->queuedata;
2726 bool read_only = rbd_dev->mapping.read_only;
2730 while ((rq = blk_fetch_request(q))) {
2731 bool write_request = rq_data_dir(rq) == WRITE;
2732 struct rbd_img_request *img_request;
2736 /* Ignore any non-FS requests that filter through. */
2738 if (rq->cmd_type != REQ_TYPE_FS) {
2739 dout("%s: non-fs request type %d\n", __func__,
2740 (int) rq->cmd_type);
2741 __blk_end_request_all(rq, 0);
2745 /* Ignore/skip any zero-length requests */
2747 offset = (u64) blk_rq_pos(rq) << SECTOR_SHIFT;
2748 length = (u64) blk_rq_bytes(rq);
2751 dout("%s: zero-length request\n", __func__);
2752 __blk_end_request_all(rq, 0);
2756 spin_unlock_irq(q->queue_lock);
2758 /* Disallow writes to a read-only device */
2760 if (write_request) {
2764 rbd_assert(rbd_dev->spec->snap_id == CEPH_NOSNAP);
2768 * Quit early if the mapped snapshot no longer
2769 * exists. It's still possible the snapshot will
2770 * have disappeared by the time our request arrives
2771 * at the osd, but there's no sense in sending it if
2774 if (!test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags)) {
2775 dout("request for non-existent snapshot");
2776 rbd_assert(rbd_dev->spec->snap_id != CEPH_NOSNAP);
2782 if (offset && length > U64_MAX - offset + 1) {
2783 rbd_warn(rbd_dev, "bad request range (%llu~%llu)\n",
2785 goto end_request; /* Shouldn't happen */
2789 img_request = rbd_img_request_create(rbd_dev, offset, length,
2790 write_request, false);
2794 img_request->rq = rq;
2796 result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
2799 result = rbd_img_request_submit(img_request);
2801 rbd_img_request_put(img_request);
2803 spin_lock_irq(q->queue_lock);
2805 rbd_warn(rbd_dev, "%s %llx at %llx result %d\n",
2806 write_request ? "write" : "read",
2807 length, offset, result);
2809 __blk_end_request_all(rq, result);
2815 * a queue callback. Makes sure that we don't create a bio that spans across
2816 * multiple osd objects. One exception would be with a single page bios,
2817 * which we handle later at bio_chain_clone_range()
2819 static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd,
2820 struct bio_vec *bvec)
2822 struct rbd_device *rbd_dev = q->queuedata;
2823 sector_t sector_offset;
2824 sector_t sectors_per_obj;
2825 sector_t obj_sector_offset;
2829 * Find how far into its rbd object the partition-relative
2830 * bio start sector is to offset relative to the enclosing
2833 sector_offset = get_start_sect(bmd->bi_bdev) + bmd->bi_sector;
2834 sectors_per_obj = 1 << (rbd_dev->header.obj_order - SECTOR_SHIFT);
2835 obj_sector_offset = sector_offset & (sectors_per_obj - 1);
2838 * Compute the number of bytes from that offset to the end
2839 * of the object. Account for what's already used by the bio.
2841 ret = (int) (sectors_per_obj - obj_sector_offset) << SECTOR_SHIFT;
2842 if (ret > bmd->bi_size)
2843 ret -= bmd->bi_size;
2848 * Don't send back more than was asked for. And if the bio
2849 * was empty, let the whole thing through because: "Note
2850 * that a block device *must* allow a single page to be
2851 * added to an empty bio."
2853 rbd_assert(bvec->bv_len <= PAGE_SIZE);
2854 if (ret > (int) bvec->bv_len || !bmd->bi_size)
2855 ret = (int) bvec->bv_len;
2860 static void rbd_free_disk(struct rbd_device *rbd_dev)
2862 struct gendisk *disk = rbd_dev->disk;
2867 rbd_dev->disk = NULL;
2868 if (disk->flags & GENHD_FL_UP) {
2871 blk_cleanup_queue(disk->queue);
2876 static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
2877 const char *object_name,
2878 u64 offset, u64 length,
2879 void *buf, u64 *version)
2882 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2883 struct rbd_obj_request *obj_request;
2884 struct page **pages = NULL;
2889 page_count = (u32) calc_pages_for(offset, length);
2890 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2892 ret = PTR_ERR(pages);
2895 obj_request = rbd_obj_request_create(object_name, offset, length,
2900 obj_request->pages = pages;
2901 obj_request->page_count = page_count;
2903 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2904 if (!obj_request->osd_req)
2907 osd_req_op_extent_init(obj_request->osd_req, 0, CEPH_OSD_OP_READ,
2908 offset, length, 0, 0);
2909 osd_req_op_extent_osd_data_pages(obj_request->osd_req, 0,
2911 obj_request->length,
2912 obj_request->offset & ~PAGE_MASK,
2914 rbd_osd_req_format_read(obj_request);
2916 ret = rbd_obj_request_submit(osdc, obj_request);
2919 ret = rbd_obj_request_wait(obj_request);
2923 ret = obj_request->result;
2927 rbd_assert(obj_request->xferred <= (u64) SIZE_MAX);
2928 size = (size_t) obj_request->xferred;
2929 ceph_copy_from_page_vector(pages, buf, 0, size);
2930 rbd_assert(size <= (size_t) INT_MAX);
2933 *version = obj_request->version;
2936 rbd_obj_request_put(obj_request);
2938 ceph_release_page_vector(pages, page_count);
2944 * Read the complete header for the given rbd device.
2946 * Returns a pointer to a dynamically-allocated buffer containing
2947 * the complete and validated header. Caller can pass the address
2948 * of a variable that will be filled in with the version of the
2949 * header object at the time it was read.
2951 * Returns a pointer-coded errno if a failure occurs.
2953 static struct rbd_image_header_ondisk *
2954 rbd_dev_v1_header_read(struct rbd_device *rbd_dev, u64 *version)
2956 struct rbd_image_header_ondisk *ondisk = NULL;
2963 * The complete header will include an array of its 64-bit
2964 * snapshot ids, followed by the names of those snapshots as
2965 * a contiguous block of NUL-terminated strings. Note that
2966 * the number of snapshots could change by the time we read
2967 * it in, in which case we re-read it.
2974 size = sizeof (*ondisk);
2975 size += snap_count * sizeof (struct rbd_image_snap_ondisk);
2977 ondisk = kmalloc(size, GFP_KERNEL);
2979 return ERR_PTR(-ENOMEM);
2981 ret = rbd_obj_read_sync(rbd_dev, rbd_dev->header_name,
2982 0, size, ondisk, version);
2985 if ((size_t)ret < size) {
2987 rbd_warn(rbd_dev, "short header read (want %zd got %d)",
2991 if (!rbd_dev_ondisk_valid(ondisk)) {
2993 rbd_warn(rbd_dev, "invalid header");
2997 names_size = le64_to_cpu(ondisk->snap_names_len);
2998 want_count = snap_count;
2999 snap_count = le32_to_cpu(ondisk->snap_count);
3000 } while (snap_count != want_count);
3007 return ERR_PTR(ret);
3011 * reload the ondisk the header
3013 static int rbd_read_header(struct rbd_device *rbd_dev,
3014 struct rbd_image_header *header)
3016 struct rbd_image_header_ondisk *ondisk;
3020 ondisk = rbd_dev_v1_header_read(rbd_dev, &ver);
3022 return PTR_ERR(ondisk);
3023 ret = rbd_header_from_disk(header, ondisk);
3025 header->obj_version = ver;
3031 static void rbd_remove_all_snaps(struct rbd_device *rbd_dev)
3033 struct rbd_snap *snap;
3034 struct rbd_snap *next;
3036 list_for_each_entry_safe(snap, next, &rbd_dev->snaps, node) {
3037 list_del(&snap->node);
3038 rbd_snap_destroy(snap);
3042 static void rbd_update_mapping_size(struct rbd_device *rbd_dev)
3044 if (rbd_dev->spec->snap_id != CEPH_NOSNAP)
3047 if (rbd_dev->mapping.size != rbd_dev->header.image_size) {
3050 rbd_dev->mapping.size = rbd_dev->header.image_size;
3051 size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE;
3052 dout("setting size to %llu sectors", (unsigned long long)size);
3053 set_capacity(rbd_dev->disk, size);
3058 * only read the first part of the ondisk header, without the snaps info
3060 static int rbd_dev_v1_refresh(struct rbd_device *rbd_dev, u64 *hver)
3063 struct rbd_image_header h;
3065 ret = rbd_read_header(rbd_dev, &h);
3069 down_write(&rbd_dev->header_rwsem);
3071 /* Update image size, and check for resize of mapped image */
3072 rbd_dev->header.image_size = h.image_size;
3073 rbd_update_mapping_size(rbd_dev);
3075 /* rbd_dev->header.object_prefix shouldn't change */
3076 kfree(rbd_dev->header.snap_sizes);
3077 kfree(rbd_dev->header.snap_names);
3078 /* osd requests may still refer to snapc */
3079 rbd_snap_context_put(rbd_dev->header.snapc);
3082 *hver = h.obj_version;
3083 rbd_dev->header.obj_version = h.obj_version;
3084 rbd_dev->header.image_size = h.image_size;
3085 rbd_dev->header.snapc = h.snapc;
3086 rbd_dev->header.snap_names = h.snap_names;
3087 rbd_dev->header.snap_sizes = h.snap_sizes;
3088 /* Free the extra copy of the object prefix */
3089 if (strcmp(rbd_dev->header.object_prefix, h.object_prefix))
3090 rbd_warn(rbd_dev, "object prefix changed (ignoring)");
3091 kfree(h.object_prefix);
3093 ret = rbd_dev_snaps_update(rbd_dev);
3095 up_write(&rbd_dev->header_rwsem);
3100 static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver)
3104 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
3105 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
3106 if (rbd_dev->image_format == 1)
3107 ret = rbd_dev_v1_refresh(rbd_dev, hver);
3109 ret = rbd_dev_v2_refresh(rbd_dev, hver);
3110 mutex_unlock(&ctl_mutex);
3111 revalidate_disk(rbd_dev->disk);
3113 rbd_warn(rbd_dev, "got notification but failed to "
3114 " update snaps: %d\n", ret);
3119 static int rbd_init_disk(struct rbd_device *rbd_dev)
3121 struct gendisk *disk;
3122 struct request_queue *q;
3125 /* create gendisk info */
3126 disk = alloc_disk(RBD_MINORS_PER_MAJOR);
3130 snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
3132 disk->major = rbd_dev->major;
3133 disk->first_minor = 0;
3134 disk->fops = &rbd_bd_ops;
3135 disk->private_data = rbd_dev;
3137 q = blk_init_queue(rbd_request_fn, &rbd_dev->lock);
3141 /* We use the default size, but let's be explicit about it. */
3142 blk_queue_physical_block_size(q, SECTOR_SIZE);
3144 /* set io sizes to object size */
3145 segment_size = rbd_obj_bytes(&rbd_dev->header);
3146 blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
3147 blk_queue_max_segment_size(q, segment_size);
3148 blk_queue_io_min(q, segment_size);
3149 blk_queue_io_opt(q, segment_size);
3151 blk_queue_merge_bvec(q, rbd_merge_bvec);
3154 q->queuedata = rbd_dev;
3156 rbd_dev->disk = disk;
3169 static struct rbd_device *dev_to_rbd_dev(struct device *dev)
3171 return container_of(dev, struct rbd_device, dev);
3174 static ssize_t rbd_size_show(struct device *dev,
3175 struct device_attribute *attr, char *buf)
3177 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3179 return sprintf(buf, "%llu\n",
3180 (unsigned long long)rbd_dev->mapping.size);
3184 * Note this shows the features for whatever's mapped, which is not
3185 * necessarily the base image.
3187 static ssize_t rbd_features_show(struct device *dev,
3188 struct device_attribute *attr, char *buf)
3190 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3192 return sprintf(buf, "0x%016llx\n",
3193 (unsigned long long)rbd_dev->mapping.features);
3196 static ssize_t rbd_major_show(struct device *dev,
3197 struct device_attribute *attr, char *buf)
3199 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3202 return sprintf(buf, "%d\n", rbd_dev->major);
3204 return sprintf(buf, "(none)\n");
3208 static ssize_t rbd_client_id_show(struct device *dev,
3209 struct device_attribute *attr, char *buf)
3211 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3213 return sprintf(buf, "client%lld\n",
3214 ceph_client_id(rbd_dev->rbd_client->client));
3217 static ssize_t rbd_pool_show(struct device *dev,
3218 struct device_attribute *attr, char *buf)
3220 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3222 return sprintf(buf, "%s\n", rbd_dev->spec->pool_name);
3225 static ssize_t rbd_pool_id_show(struct device *dev,
3226 struct device_attribute *attr, char *buf)
3228 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3230 return sprintf(buf, "%llu\n",
3231 (unsigned long long) rbd_dev->spec->pool_id);
3234 static ssize_t rbd_name_show(struct device *dev,
3235 struct device_attribute *attr, char *buf)
3237 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3239 if (rbd_dev->spec->image_name)
3240 return sprintf(buf, "%s\n", rbd_dev->spec->image_name);
3242 return sprintf(buf, "(unknown)\n");
3245 static ssize_t rbd_image_id_show(struct device *dev,
3246 struct device_attribute *attr, char *buf)
3248 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3250 return sprintf(buf, "%s\n", rbd_dev->spec->image_id);
3254 * Shows the name of the currently-mapped snapshot (or
3255 * RBD_SNAP_HEAD_NAME for the base image).
3257 static ssize_t rbd_snap_show(struct device *dev,
3258 struct device_attribute *attr,
3261 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3263 return sprintf(buf, "%s\n", rbd_dev->spec->snap_name);
3267 * For an rbd v2 image, shows the pool id, image id, and snapshot id
3268 * for the parent image. If there is no parent, simply shows
3269 * "(no parent image)".
3271 static ssize_t rbd_parent_show(struct device *dev,
3272 struct device_attribute *attr,
3275 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3276 struct rbd_spec *spec = rbd_dev->parent_spec;
3281 return sprintf(buf, "(no parent image)\n");
3283 count = sprintf(bufp, "pool_id %llu\npool_name %s\n",
3284 (unsigned long long) spec->pool_id, spec->pool_name);
3289 count = sprintf(bufp, "image_id %s\nimage_name %s\n", spec->image_id,
3290 spec->image_name ? spec->image_name : "(unknown)");
3295 count = sprintf(bufp, "snap_id %llu\nsnap_name %s\n",
3296 (unsigned long long) spec->snap_id, spec->snap_name);
3301 count = sprintf(bufp, "overlap %llu\n", rbd_dev->parent_overlap);
3306 return (ssize_t) (bufp - buf);
3309 static ssize_t rbd_image_refresh(struct device *dev,
3310 struct device_attribute *attr,
3314 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3317 ret = rbd_dev_refresh(rbd_dev, NULL);
3319 return ret < 0 ? ret : size;
3322 static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
3323 static DEVICE_ATTR(features, S_IRUGO, rbd_features_show, NULL);
3324 static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
3325 static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
3326 static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
3327 static DEVICE_ATTR(pool_id, S_IRUGO, rbd_pool_id_show, NULL);
3328 static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
3329 static DEVICE_ATTR(image_id, S_IRUGO, rbd_image_id_show, NULL);
3330 static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
3331 static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
3332 static DEVICE_ATTR(parent, S_IRUGO, rbd_parent_show, NULL);
3334 static struct attribute *rbd_attrs[] = {
3335 &dev_attr_size.attr,
3336 &dev_attr_features.attr,
3337 &dev_attr_major.attr,
3338 &dev_attr_client_id.attr,
3339 &dev_attr_pool.attr,
3340 &dev_attr_pool_id.attr,
3341 &dev_attr_name.attr,
3342 &dev_attr_image_id.attr,
3343 &dev_attr_current_snap.attr,
3344 &dev_attr_parent.attr,
3345 &dev_attr_refresh.attr,
3349 static struct attribute_group rbd_attr_group = {
3353 static const struct attribute_group *rbd_attr_groups[] = {
3358 static void rbd_sysfs_dev_release(struct device *dev)
3362 static struct device_type rbd_device_type = {
3364 .groups = rbd_attr_groups,
3365 .release = rbd_sysfs_dev_release,
3368 static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec)
3370 kref_get(&spec->kref);
3375 static void rbd_spec_free(struct kref *kref);
3376 static void rbd_spec_put(struct rbd_spec *spec)
3379 kref_put(&spec->kref, rbd_spec_free);
3382 static struct rbd_spec *rbd_spec_alloc(void)
3384 struct rbd_spec *spec;
3386 spec = kzalloc(sizeof (*spec), GFP_KERNEL);
3389 kref_init(&spec->kref);
3394 static void rbd_spec_free(struct kref *kref)
3396 struct rbd_spec *spec = container_of(kref, struct rbd_spec, kref);
3398 kfree(spec->pool_name);
3399 kfree(spec->image_id);
3400 kfree(spec->image_name);
3401 kfree(spec->snap_name);
3405 static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc,
3406 struct rbd_spec *spec)
3408 struct rbd_device *rbd_dev;
3410 rbd_dev = kzalloc(sizeof (*rbd_dev), GFP_KERNEL);
3414 spin_lock_init(&rbd_dev->lock);
3416 INIT_LIST_HEAD(&rbd_dev->node);
3417 INIT_LIST_HEAD(&rbd_dev->snaps);
3418 init_rwsem(&rbd_dev->header_rwsem);
3420 rbd_dev->spec = spec;
3421 rbd_dev->rbd_client = rbdc;
3423 /* Initialize the layout used for all rbd requests */
3425 rbd_dev->layout.fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
3426 rbd_dev->layout.fl_stripe_count = cpu_to_le32(1);
3427 rbd_dev->layout.fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
3428 rbd_dev->layout.fl_pg_pool = cpu_to_le32((u32) spec->pool_id);
3433 static void rbd_dev_destroy(struct rbd_device *rbd_dev)
3435 rbd_put_client(rbd_dev->rbd_client);
3436 rbd_spec_put(rbd_dev->spec);
3440 static void rbd_snap_destroy(struct rbd_snap *snap)
3446 static struct rbd_snap *rbd_snap_create(struct rbd_device *rbd_dev,
3447 const char *snap_name,
3448 u64 snap_id, u64 snap_size,
3451 struct rbd_snap *snap;
3453 snap = kzalloc(sizeof (*snap), GFP_KERNEL);
3455 return ERR_PTR(-ENOMEM);
3457 snap->name = snap_name;
3459 snap->size = snap_size;
3460 snap->features = snap_features;
3466 * Returns a dynamically-allocated snapshot name if successful, or a
3467 * pointer-coded error otherwise.
3469 static char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
3470 u64 *snap_size, u64 *snap_features)
3475 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3477 /* Skip over names until we find the one we are looking for */
3479 snap_name = rbd_dev->header.snap_names;
3480 for (i = 0; i < which; i++)
3481 snap_name += strlen(snap_name) + 1;
3483 snap_name = kstrdup(snap_name, GFP_KERNEL);
3485 return ERR_PTR(-ENOMEM);
3487 *snap_size = rbd_dev->header.snap_sizes[which];
3488 *snap_features = 0; /* No features for v1 */
3494 * Get the size and object order for an image snapshot, or if
3495 * snap_id is CEPH_NOSNAP, gets this information for the base
3498 static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
3499 u8 *order, u64 *snap_size)
3501 __le64 snapid = cpu_to_le64(snap_id);
3506 } __attribute__ ((packed)) size_buf = { 0 };
3508 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3510 &snapid, sizeof (snapid),
3511 &size_buf, sizeof (size_buf), NULL);
3512 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3515 if (ret < sizeof (size_buf))
3519 *order = size_buf.order;
3520 *snap_size = le64_to_cpu(size_buf.size);
3522 dout(" snap_id 0x%016llx order = %u, snap_size = %llu\n",
3523 (unsigned long long)snap_id, (unsigned int)*order,
3524 (unsigned long long)*snap_size);
3529 static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
3531 return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
3532 &rbd_dev->header.obj_order,
3533 &rbd_dev->header.image_size);
3536 static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
3542 reply_buf = kzalloc(RBD_OBJ_PREFIX_LEN_MAX, GFP_KERNEL);
3546 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3547 "rbd", "get_object_prefix", NULL, 0,
3548 reply_buf, RBD_OBJ_PREFIX_LEN_MAX, NULL);
3549 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3554 rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
3555 p + ret, NULL, GFP_NOIO);
3558 if (IS_ERR(rbd_dev->header.object_prefix)) {
3559 ret = PTR_ERR(rbd_dev->header.object_prefix);
3560 rbd_dev->header.object_prefix = NULL;
3562 dout(" object_prefix = %s\n", rbd_dev->header.object_prefix);
3570 static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
3573 __le64 snapid = cpu_to_le64(snap_id);
3577 } __attribute__ ((packed)) features_buf = { 0 };
3581 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3582 "rbd", "get_features",
3583 &snapid, sizeof (snapid),
3584 &features_buf, sizeof (features_buf), NULL);
3585 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3588 if (ret < sizeof (features_buf))
3591 incompat = le64_to_cpu(features_buf.incompat);
3592 if (incompat & ~RBD_FEATURES_SUPPORTED)
3595 *snap_features = le64_to_cpu(features_buf.features);
3597 dout(" snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
3598 (unsigned long long)snap_id,
3599 (unsigned long long)*snap_features,
3600 (unsigned long long)le64_to_cpu(features_buf.incompat));
3605 static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
3607 return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
3608 &rbd_dev->header.features);
3611 static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
3613 struct rbd_spec *parent_spec;
3615 void *reply_buf = NULL;
3623 parent_spec = rbd_spec_alloc();
3627 size = sizeof (__le64) + /* pool_id */
3628 sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX + /* image_id */
3629 sizeof (__le64) + /* snap_id */
3630 sizeof (__le64); /* overlap */
3631 reply_buf = kmalloc(size, GFP_KERNEL);
3637 snapid = cpu_to_le64(CEPH_NOSNAP);
3638 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3639 "rbd", "get_parent",
3640 &snapid, sizeof (snapid),
3641 reply_buf, size, NULL);
3642 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3647 end = reply_buf + ret;
3649 ceph_decode_64_safe(&p, end, parent_spec->pool_id, out_err);
3650 if (parent_spec->pool_id == CEPH_NOPOOL)
3651 goto out; /* No parent? No problem. */
3653 /* The ceph file layout needs to fit pool id in 32 bits */
3656 if (parent_spec->pool_id > (u64)U32_MAX) {
3657 rbd_warn(NULL, "parent pool id too large (%llu > %u)\n",
3658 (unsigned long long)parent_spec->pool_id, U32_MAX);
3662 image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
3663 if (IS_ERR(image_id)) {
3664 ret = PTR_ERR(image_id);
3667 parent_spec->image_id = image_id;
3668 ceph_decode_64_safe(&p, end, parent_spec->snap_id, out_err);
3669 ceph_decode_64_safe(&p, end, overlap, out_err);
3671 rbd_dev->parent_overlap = overlap;
3672 rbd_dev->parent_spec = parent_spec;
3673 parent_spec = NULL; /* rbd_dev now owns this */
3678 rbd_spec_put(parent_spec);
3683 static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev)
3687 __le64 stripe_count;
3688 } __attribute__ ((packed)) striping_info_buf = { 0 };
3689 size_t size = sizeof (striping_info_buf);
3696 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3697 "rbd", "get_stripe_unit_count", NULL, 0,
3698 (char *)&striping_info_buf, size, NULL);
3699 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3706 * We don't actually support the "fancy striping" feature
3707 * (STRIPINGV2) yet, but if the striping sizes are the
3708 * defaults the behavior is the same as before. So find
3709 * out, and only fail if the image has non-default values.
3712 obj_size = (u64)1 << rbd_dev->header.obj_order;
3713 p = &striping_info_buf;
3714 stripe_unit = ceph_decode_64(&p);
3715 if (stripe_unit != obj_size) {
3716 rbd_warn(rbd_dev, "unsupported stripe unit "
3717 "(got %llu want %llu)",
3718 stripe_unit, obj_size);
3721 stripe_count = ceph_decode_64(&p);
3722 if (stripe_count != 1) {
3723 rbd_warn(rbd_dev, "unsupported stripe count "
3724 "(got %llu want 1)", stripe_count);
3727 rbd_dev->header.stripe_unit = stripe_unit;
3728 rbd_dev->header.stripe_count = stripe_count;
3733 static char *rbd_dev_image_name(struct rbd_device *rbd_dev)
3735 size_t image_id_size;
3740 void *reply_buf = NULL;
3742 char *image_name = NULL;
3745 rbd_assert(!rbd_dev->spec->image_name);
3747 len = strlen(rbd_dev->spec->image_id);
3748 image_id_size = sizeof (__le32) + len;
3749 image_id = kmalloc(image_id_size, GFP_KERNEL);
3754 end = image_id + image_id_size;
3755 ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len);
3757 size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX;
3758 reply_buf = kmalloc(size, GFP_KERNEL);
3762 ret = rbd_obj_method_sync(rbd_dev, RBD_DIRECTORY,
3763 "rbd", "dir_get_name",
3764 image_id, image_id_size,
3765 reply_buf, size, NULL);
3769 end = reply_buf + ret;
3771 image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL);
3772 if (IS_ERR(image_name))
3775 dout("%s: name is %s len is %zd\n", __func__, image_name, len);
3784 * When an rbd image has a parent image, it is identified by the
3785 * pool, image, and snapshot ids (not names). This function fills
3786 * in the names for those ids. (It's OK if we can't figure out the
3787 * name for an image id, but the pool and snapshot ids should always
3788 * exist and have names.) All names in an rbd spec are dynamically
3791 * When an image being mapped (not a parent) is probed, we have the
3792 * pool name and pool id, image name and image id, and the snapshot
3793 * name. The only thing we're missing is the snapshot id.
3795 * The set of snapshots for an image is not known until they have
3796 * been read by rbd_dev_snaps_update(), so we can't completely fill
3797 * in this information until after that has been called.
3799 static int rbd_dev_spec_update(struct rbd_device *rbd_dev)
3801 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3802 struct rbd_spec *spec = rbd_dev->spec;
3803 const char *pool_name;
3804 const char *image_name;
3805 const char *snap_name;
3809 * An image being mapped will have the pool name (etc.), but
3810 * we need to look up the snapshot id.
3812 if (spec->pool_name) {
3813 if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
3814 struct rbd_snap *snap;
3816 snap = snap_by_name(rbd_dev, spec->snap_name);
3819 spec->snap_id = snap->id;
3821 spec->snap_id = CEPH_NOSNAP;
3827 /* Get the pool name; we have to make our own copy of this */
3829 pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id);
3831 rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id);
3834 pool_name = kstrdup(pool_name, GFP_KERNEL);
3838 /* Fetch the image name; tolerate failure here */
3840 image_name = rbd_dev_image_name(rbd_dev);
3842 rbd_warn(rbd_dev, "unable to get image name");
3844 /* Look up the snapshot name, and make a copy */
3846 snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
3848 rbd_warn(rbd_dev, "no snapshot with id %llu", spec->snap_id);
3852 snap_name = kstrdup(snap_name, GFP_KERNEL);
3858 spec->pool_name = pool_name;
3859 spec->image_name = image_name;
3860 spec->snap_name = snap_name;
3870 static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev, u64 *ver)
3879 struct ceph_snap_context *snapc;
3883 * We'll need room for the seq value (maximum snapshot id),
3884 * snapshot count, and array of that many snapshot ids.
3885 * For now we have a fixed upper limit on the number we're
3886 * prepared to receive.
3888 size = sizeof (__le64) + sizeof (__le32) +
3889 RBD_MAX_SNAP_COUNT * sizeof (__le64);
3890 reply_buf = kzalloc(size, GFP_KERNEL);
3894 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3895 "rbd", "get_snapcontext", NULL, 0,
3896 reply_buf, size, ver);
3897 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3902 end = reply_buf + ret;
3904 ceph_decode_64_safe(&p, end, seq, out);
3905 ceph_decode_32_safe(&p, end, snap_count, out);
3908 * Make sure the reported number of snapshot ids wouldn't go
3909 * beyond the end of our buffer. But before checking that,
3910 * make sure the computed size of the snapshot context we
3911 * allocate is representable in a size_t.
3913 if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
3918 if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
3922 snapc = rbd_snap_context_create(snap_count);
3928 for (i = 0; i < snap_count; i++)
3929 snapc->snaps[i] = ceph_decode_64(&p);
3931 rbd_dev->header.snapc = snapc;
3933 dout(" snap context seq = %llu, snap_count = %u\n",
3934 (unsigned long long)seq, (unsigned int)snap_count);
3941 static char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
3951 size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
3952 reply_buf = kmalloc(size, GFP_KERNEL);
3954 return ERR_PTR(-ENOMEM);
3956 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3957 snap_id = cpu_to_le64(rbd_dev->header.snapc->snaps[which]);
3958 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3959 "rbd", "get_snapshot_name",
3960 &snap_id, sizeof (snap_id),
3961 reply_buf, size, NULL);
3962 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3964 snap_name = ERR_PTR(ret);
3969 end = reply_buf + ret;
3970 snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
3971 if (IS_ERR(snap_name))
3974 dout(" snap_id 0x%016llx snap_name = %s\n",
3975 (unsigned long long)le64_to_cpu(snap_id), snap_name);
3982 static char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
3983 u64 *snap_size, u64 *snap_features)
3991 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3992 snap_id = rbd_dev->header.snapc->snaps[which];
3993 ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size);
3997 ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, &features);
4001 snap_name = rbd_dev_v2_snap_name(rbd_dev, which);
4002 if (!IS_ERR(snap_name)) {
4004 *snap_features = features;
4009 return ERR_PTR(ret);
4012 static char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
4013 u64 *snap_size, u64 *snap_features)
4015 if (rbd_dev->image_format == 1)
4016 return rbd_dev_v1_snap_info(rbd_dev, which,
4017 snap_size, snap_features);
4018 if (rbd_dev->image_format == 2)
4019 return rbd_dev_v2_snap_info(rbd_dev, which,
4020 snap_size, snap_features);
4021 return ERR_PTR(-EINVAL);
4024 static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev, u64 *hver)
4028 down_write(&rbd_dev->header_rwsem);
4030 ret = rbd_dev_v2_image_size(rbd_dev);
4033 rbd_update_mapping_size(rbd_dev);
4035 ret = rbd_dev_v2_snap_context(rbd_dev, hver);
4036 dout("rbd_dev_v2_snap_context returned %d\n", ret);
4039 ret = rbd_dev_snaps_update(rbd_dev);
4040 dout("rbd_dev_snaps_update returned %d\n", ret);
4044 up_write(&rbd_dev->header_rwsem);
4050 * Scan the rbd device's current snapshot list and compare it to the
4051 * newly-received snapshot context. Remove any existing snapshots
4052 * not present in the new snapshot context. Add a new snapshot for
4053 * any snaphots in the snapshot context not in the current list.
4054 * And verify there are no changes to snapshots we already know
4057 * Assumes the snapshots in the snapshot context are sorted by
4058 * snapshot id, highest id first. (Snapshots in the rbd_dev's list
4059 * are also maintained in that order.)
4061 * Note that any error occurs while updating the snapshot list
4062 * aborts the update, and the entire list is cleared. The snapshot
4063 * list becomes inconsistent at that point anyway, so it might as
4066 static int rbd_dev_snaps_update(struct rbd_device *rbd_dev)
4068 struct ceph_snap_context *snapc = rbd_dev->header.snapc;
4069 const u32 snap_count = snapc->num_snaps;
4070 struct list_head *head = &rbd_dev->snaps;
4071 struct list_head *links = head->next;
4075 dout("%s: snap count is %u\n", __func__, (unsigned int)snap_count);
4076 while (index < snap_count || links != head) {
4078 struct rbd_snap *snap;
4081 u64 snap_features = 0;
4083 snap_id = index < snap_count ? snapc->snaps[index]
4085 snap = links != head ? list_entry(links, struct rbd_snap, node)
4087 rbd_assert(!snap || snap->id != CEPH_NOSNAP);
4089 if (snap_id == CEPH_NOSNAP || (snap && snap->id > snap_id)) {
4090 struct list_head *next = links->next;
4093 * A previously-existing snapshot is not in
4094 * the new snap context.
4096 * If the now-missing snapshot is the one
4097 * the image represents, clear its existence
4098 * flag so we can avoid sending any more
4101 if (rbd_dev->spec->snap_id == snap->id)
4102 clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
4103 dout("removing %ssnap id %llu\n",
4104 rbd_dev->spec->snap_id == snap->id ?
4106 (unsigned long long)snap->id);
4108 list_del(&snap->node);
4109 rbd_snap_destroy(snap);
4111 /* Done with this list entry; advance */
4117 snap_name = rbd_dev_snap_info(rbd_dev, index,
4118 &snap_size, &snap_features);
4119 if (IS_ERR(snap_name)) {
4120 ret = PTR_ERR(snap_name);
4121 dout("failed to get snap info, error %d\n", ret);
4125 dout("entry %u: snap_id = %llu\n", (unsigned int)snap_count,
4126 (unsigned long long)snap_id);
4127 if (!snap || (snap_id != CEPH_NOSNAP && snap->id < snap_id)) {
4128 struct rbd_snap *new_snap;
4130 /* We haven't seen this snapshot before */
4132 new_snap = rbd_snap_create(rbd_dev, snap_name,
4133 snap_id, snap_size, snap_features);
4134 if (IS_ERR(new_snap)) {
4135 ret = PTR_ERR(new_snap);
4136 dout(" failed to add dev, error %d\n", ret);
4140 /* New goes before existing, or at end of list */
4142 dout(" added dev%s\n", snap ? "" : " at end\n");
4144 list_add_tail(&new_snap->node, &snap->node);
4146 list_add_tail(&new_snap->node, head);
4148 /* Already have this one */
4150 dout(" already present\n");
4152 rbd_assert(snap->size == snap_size);
4153 rbd_assert(!strcmp(snap->name, snap_name));
4154 rbd_assert(snap->features == snap_features);
4156 /* Done with this list entry; advance */
4158 links = links->next;
4161 /* Advance to the next entry in the snapshot context */
4165 dout("%s: done\n", __func__);
4169 rbd_remove_all_snaps(rbd_dev);
4174 static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
4179 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
4181 dev = &rbd_dev->dev;
4182 dev->bus = &rbd_bus_type;
4183 dev->type = &rbd_device_type;
4184 dev->parent = &rbd_root_dev;
4185 dev->release = rbd_dev_release;
4186 dev_set_name(dev, "%d", rbd_dev->dev_id);
4187 ret = device_register(dev);
4189 mutex_unlock(&ctl_mutex);
4194 static void rbd_bus_del_dev(struct rbd_device *rbd_dev)
4196 device_unregister(&rbd_dev->dev);
4199 static atomic64_t rbd_dev_id_max = ATOMIC64_INIT(0);
4202 * Get a unique rbd identifier for the given new rbd_dev, and add
4203 * the rbd_dev to the global list. The minimum rbd id is 1.
4205 static void rbd_dev_id_get(struct rbd_device *rbd_dev)
4207 rbd_dev->dev_id = atomic64_inc_return(&rbd_dev_id_max);
4209 spin_lock(&rbd_dev_list_lock);
4210 list_add_tail(&rbd_dev->node, &rbd_dev_list);
4211 spin_unlock(&rbd_dev_list_lock);
4212 dout("rbd_dev %p given dev id %llu\n", rbd_dev,
4213 (unsigned long long) rbd_dev->dev_id);
4217 * Remove an rbd_dev from the global list, and record that its
4218 * identifier is no longer in use.
4220 static void rbd_dev_id_put(struct rbd_device *rbd_dev)
4222 struct list_head *tmp;
4223 int rbd_id = rbd_dev->dev_id;
4226 rbd_assert(rbd_id > 0);
4228 dout("rbd_dev %p released dev id %llu\n", rbd_dev,
4229 (unsigned long long) rbd_dev->dev_id);
4230 spin_lock(&rbd_dev_list_lock);
4231 list_del_init(&rbd_dev->node);
4234 * If the id being "put" is not the current maximum, there
4235 * is nothing special we need to do.
4237 if (rbd_id != atomic64_read(&rbd_dev_id_max)) {
4238 spin_unlock(&rbd_dev_list_lock);
4243 * We need to update the current maximum id. Search the
4244 * list to find out what it is. We're more likely to find
4245 * the maximum at the end, so search the list backward.
4248 list_for_each_prev(tmp, &rbd_dev_list) {
4249 struct rbd_device *rbd_dev;
4251 rbd_dev = list_entry(tmp, struct rbd_device, node);
4252 if (rbd_dev->dev_id > max_id)
4253 max_id = rbd_dev->dev_id;
4255 spin_unlock(&rbd_dev_list_lock);
4258 * The max id could have been updated by rbd_dev_id_get(), in
4259 * which case it now accurately reflects the new maximum.
4260 * Be careful not to overwrite the maximum value in that
4263 atomic64_cmpxchg(&rbd_dev_id_max, rbd_id, max_id);
4264 dout(" max dev id has been reset\n");
4268 * Skips over white space at *buf, and updates *buf to point to the
4269 * first found non-space character (if any). Returns the length of
4270 * the token (string of non-white space characters) found. Note
4271 * that *buf must be terminated with '\0'.
4273 static inline size_t next_token(const char **buf)
4276 * These are the characters that produce nonzero for
4277 * isspace() in the "C" and "POSIX" locales.
4279 const char *spaces = " \f\n\r\t\v";
4281 *buf += strspn(*buf, spaces); /* Find start of token */
4283 return strcspn(*buf, spaces); /* Return token length */
4287 * Finds the next token in *buf, and if the provided token buffer is
4288 * big enough, copies the found token into it. The result, if
4289 * copied, is guaranteed to be terminated with '\0'. Note that *buf
4290 * must be terminated with '\0' on entry.
4292 * Returns the length of the token found (not including the '\0').
4293 * Return value will be 0 if no token is found, and it will be >=
4294 * token_size if the token would not fit.
4296 * The *buf pointer will be updated to point beyond the end of the
4297 * found token. Note that this occurs even if the token buffer is
4298 * too small to hold it.
4300 static inline size_t copy_token(const char **buf,
4306 len = next_token(buf);
4307 if (len < token_size) {
4308 memcpy(token, *buf, len);
4309 *(token + len) = '\0';
4317 * Finds the next token in *buf, dynamically allocates a buffer big
4318 * enough to hold a copy of it, and copies the token into the new
4319 * buffer. The copy is guaranteed to be terminated with '\0'. Note
4320 * that a duplicate buffer is created even for a zero-length token.
4322 * Returns a pointer to the newly-allocated duplicate, or a null
4323 * pointer if memory for the duplicate was not available. If
4324 * the lenp argument is a non-null pointer, the length of the token
4325 * (not including the '\0') is returned in *lenp.
4327 * If successful, the *buf pointer will be updated to point beyond
4328 * the end of the found token.
4330 * Note: uses GFP_KERNEL for allocation.
4332 static inline char *dup_token(const char **buf, size_t *lenp)
4337 len = next_token(buf);
4338 dup = kmemdup(*buf, len + 1, GFP_KERNEL);
4341 *(dup + len) = '\0';
4351 * Parse the options provided for an "rbd add" (i.e., rbd image
4352 * mapping) request. These arrive via a write to /sys/bus/rbd/add,
4353 * and the data written is passed here via a NUL-terminated buffer.
4354 * Returns 0 if successful or an error code otherwise.
4356 * The information extracted from these options is recorded in
4357 * the other parameters which return dynamically-allocated
4360 * The address of a pointer that will refer to a ceph options
4361 * structure. Caller must release the returned pointer using
4362 * ceph_destroy_options() when it is no longer needed.
4364 * Address of an rbd options pointer. Fully initialized by
4365 * this function; caller must release with kfree().
4367 * Address of an rbd image specification pointer. Fully
4368 * initialized by this function based on parsed options.
4369 * Caller must release with rbd_spec_put().
4371 * The options passed take this form:
4372 * <mon_addrs> <options> <pool_name> <image_name> [<snap_id>]
4375 * A comma-separated list of one or more monitor addresses.
4376 * A monitor address is an ip address, optionally followed
4377 * by a port number (separated by a colon).
4378 * I.e.: ip1[:port1][,ip2[:port2]...]
4380 * A comma-separated list of ceph and/or rbd options.
4382 * The name of the rados pool containing the rbd image.
4384 * The name of the image in that pool to map.
4386 * An optional snapshot id. If provided, the mapping will
4387 * present data from the image at the time that snapshot was
4388 * created. The image head is used if no snapshot id is
4389 * provided. Snapshot mappings are always read-only.
4391 static int rbd_add_parse_args(const char *buf,
4392 struct ceph_options **ceph_opts,
4393 struct rbd_options **opts,
4394 struct rbd_spec **rbd_spec)
4398 const char *mon_addrs;
4400 size_t mon_addrs_size;
4401 struct rbd_spec *spec = NULL;
4402 struct rbd_options *rbd_opts = NULL;
4403 struct ceph_options *copts;
4406 /* The first four tokens are required */
4408 len = next_token(&buf);
4410 rbd_warn(NULL, "no monitor address(es) provided");
4414 mon_addrs_size = len + 1;
4418 options = dup_token(&buf, NULL);
4422 rbd_warn(NULL, "no options provided");
4426 spec = rbd_spec_alloc();
4430 spec->pool_name = dup_token(&buf, NULL);
4431 if (!spec->pool_name)
4433 if (!*spec->pool_name) {
4434 rbd_warn(NULL, "no pool name provided");
4438 spec->image_name = dup_token(&buf, NULL);
4439 if (!spec->image_name)
4441 if (!*spec->image_name) {
4442 rbd_warn(NULL, "no image name provided");
4447 * Snapshot name is optional; default is to use "-"
4448 * (indicating the head/no snapshot).
4450 len = next_token(&buf);
4452 buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */
4453 len = sizeof (RBD_SNAP_HEAD_NAME) - 1;
4454 } else if (len > RBD_MAX_SNAP_NAME_LEN) {
4455 ret = -ENAMETOOLONG;
4458 snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
4461 *(snap_name + len) = '\0';
4462 spec->snap_name = snap_name;
4464 /* Initialize all rbd options to the defaults */
4466 rbd_opts = kzalloc(sizeof (*rbd_opts), GFP_KERNEL);
4470 rbd_opts->read_only = RBD_READ_ONLY_DEFAULT;
4472 copts = ceph_parse_options(options, mon_addrs,
4473 mon_addrs + mon_addrs_size - 1,
4474 parse_rbd_opts_token, rbd_opts);
4475 if (IS_ERR(copts)) {
4476 ret = PTR_ERR(copts);
4497 * An rbd format 2 image has a unique identifier, distinct from the
4498 * name given to it by the user. Internally, that identifier is
4499 * what's used to specify the names of objects related to the image.
4501 * A special "rbd id" object is used to map an rbd image name to its
4502 * id. If that object doesn't exist, then there is no v2 rbd image
4503 * with the supplied name.
4505 * This function will record the given rbd_dev's image_id field if
4506 * it can be determined, and in that case will return 0. If any
4507 * errors occur a negative errno will be returned and the rbd_dev's
4508 * image_id field will be unchanged (and should be NULL).
4510 static int rbd_dev_image_id(struct rbd_device *rbd_dev)
4519 * When probing a parent image, the image id is already
4520 * known (and the image name likely is not). There's no
4521 * need to fetch the image id again in this case. We
4522 * do still need to set the image format though.
4524 if (rbd_dev->spec->image_id) {
4525 rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1;
4531 * First, see if the format 2 image id file exists, and if
4532 * so, get the image's persistent id from it.
4534 size = sizeof (RBD_ID_PREFIX) + strlen(rbd_dev->spec->image_name);
4535 object_name = kmalloc(size, GFP_NOIO);
4538 sprintf(object_name, "%s%s", RBD_ID_PREFIX, rbd_dev->spec->image_name);
4539 dout("rbd id object name is %s\n", object_name);
4541 /* Response will be an encoded string, which includes a length */
4543 size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
4544 response = kzalloc(size, GFP_NOIO);
4550 /* If it doesn't exist we'll assume it's a format 1 image */
4552 ret = rbd_obj_method_sync(rbd_dev, object_name,
4553 "rbd", "get_id", NULL, 0,
4554 response, RBD_IMAGE_ID_LEN_MAX, NULL);
4555 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
4556 if (ret == -ENOENT) {
4557 image_id = kstrdup("", GFP_KERNEL);
4558 ret = image_id ? 0 : -ENOMEM;
4560 rbd_dev->image_format = 1;
4561 } else if (ret > sizeof (__le32)) {
4564 image_id = ceph_extract_encoded_string(&p, p + ret,
4566 ret = IS_ERR(image_id) ? PTR_ERR(image_id) : 0;
4568 rbd_dev->image_format = 2;
4574 rbd_dev->spec->image_id = image_id;
4575 dout("image_id is %s\n", image_id);
4584 static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
4588 /* Populate rbd image metadata */
4590 ret = rbd_read_header(rbd_dev, &rbd_dev->header);
4594 /* Version 1 images have no parent (no layering) */
4596 rbd_dev->parent_spec = NULL;
4597 rbd_dev->parent_overlap = 0;
4599 dout("discovered version 1 image, header name is %s\n",
4600 rbd_dev->header_name);
4605 kfree(rbd_dev->header_name);
4606 rbd_dev->header_name = NULL;
4607 kfree(rbd_dev->spec->image_id);
4608 rbd_dev->spec->image_id = NULL;
4613 static int rbd_dev_v2_probe(struct rbd_device *rbd_dev)
4618 ret = rbd_dev_v2_image_size(rbd_dev);
4622 /* Get the object prefix (a.k.a. block_name) for the image */
4624 ret = rbd_dev_v2_object_prefix(rbd_dev);
4628 /* Get the and check features for the image */
4630 ret = rbd_dev_v2_features(rbd_dev);
4634 /* If the image supports layering, get the parent info */
4636 if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
4637 ret = rbd_dev_v2_parent_info(rbd_dev);
4640 rbd_warn(rbd_dev, "WARNING: kernel support for "
4641 "layered rbd images is EXPERIMENTAL!");
4644 /* If the image supports fancy striping, get its parameters */
4646 if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) {
4647 ret = rbd_dev_v2_striping_info(rbd_dev);
4652 /* crypto and compression type aren't (yet) supported for v2 images */
4654 rbd_dev->header.crypt_type = 0;
4655 rbd_dev->header.comp_type = 0;
4657 /* Get the snapshot context, plus the header version */
4659 ret = rbd_dev_v2_snap_context(rbd_dev, &ver);
4662 rbd_dev->header.obj_version = ver;
4664 dout("discovered version 2 image, header name is %s\n",
4665 rbd_dev->header_name);
4669 rbd_dev->parent_overlap = 0;
4670 rbd_spec_put(rbd_dev->parent_spec);
4671 rbd_dev->parent_spec = NULL;
4672 kfree(rbd_dev->header_name);
4673 rbd_dev->header_name = NULL;
4674 kfree(rbd_dev->header.object_prefix);
4675 rbd_dev->header.object_prefix = NULL;
4680 static int rbd_dev_probe_parent(struct rbd_device *rbd_dev)
4682 struct rbd_device *parent = NULL;
4683 struct rbd_spec *parent_spec;
4684 struct rbd_client *rbdc;
4687 if (!rbd_dev->parent_spec)
4690 * We need to pass a reference to the client and the parent
4691 * spec when creating the parent rbd_dev. Images related by
4692 * parent/child relationships always share both.
4694 parent_spec = rbd_spec_get(rbd_dev->parent_spec);
4695 rbdc = __rbd_get_client(rbd_dev->rbd_client);
4698 parent = rbd_dev_create(rbdc, parent_spec);
4702 ret = rbd_dev_image_probe(parent);
4705 rbd_dev->parent = parent;
4710 rbd_spec_put(rbd_dev->parent_spec);
4711 kfree(rbd_dev->header_name);
4712 rbd_dev_destroy(parent);
4714 rbd_put_client(rbdc);
4715 rbd_spec_put(parent_spec);
4721 static int rbd_dev_probe_finish(struct rbd_device *rbd_dev)
4726 ret = rbd_dev_header_watch_sync(rbd_dev, 1);
4730 ret = rbd_dev_mapping_set(rbd_dev);
4734 /* generate unique id: find highest unique id, add one */
4735 rbd_dev_id_get(rbd_dev);
4737 /* Fill in the device name, now that we have its id. */
4738 BUILD_BUG_ON(DEV_NAME_LEN
4739 < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
4740 sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->dev_id);
4742 /* Get our block major device number. */
4744 ret = register_blkdev(0, rbd_dev->name);
4747 rbd_dev->major = ret;
4749 /* Set up the blkdev mapping. */
4751 ret = rbd_init_disk(rbd_dev);
4753 goto err_out_blkdev;
4755 ret = rbd_bus_add_dev(rbd_dev);
4759 /* Everything's ready. Announce the disk to the world. */
4761 set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
4762 set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
4763 add_disk(rbd_dev->disk);
4765 pr_info("%s: added with size 0x%llx\n", rbd_dev->disk->disk_name,
4766 (unsigned long long) rbd_dev->mapping.size);
4771 rbd_free_disk(rbd_dev);
4773 unregister_blkdev(rbd_dev->major, rbd_dev->name);
4775 rbd_dev_id_put(rbd_dev);
4776 tmp = rbd_dev_header_watch_sync(rbd_dev, 0);
4778 rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret);
4779 rbd_dev_mapping_clear(rbd_dev);
4784 static int rbd_dev_header_name(struct rbd_device *rbd_dev)
4786 struct rbd_spec *spec = rbd_dev->spec;
4789 /* Record the header object name for this rbd image. */
4791 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
4793 if (rbd_dev->image_format == 1)
4794 size = strlen(spec->image_name) + sizeof (RBD_SUFFIX);
4796 size = sizeof (RBD_HEADER_PREFIX) + strlen(spec->image_id);
4798 rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
4799 if (!rbd_dev->header_name)
4802 if (rbd_dev->image_format == 1)
4803 sprintf(rbd_dev->header_name, "%s%s",
4804 spec->image_name, RBD_SUFFIX);
4806 sprintf(rbd_dev->header_name, "%s%s",
4807 RBD_HEADER_PREFIX, spec->image_id);
4812 * Probe for the existence of the header object for the given rbd
4813 * device. For format 2 images this includes determining the image
4816 static int rbd_dev_image_probe(struct rbd_device *rbd_dev)
4821 * Get the id from the image id object. If it's not a
4822 * format 2 image, we'll get ENOENT back, and we'll assume
4823 * it's a format 1 image.
4825 ret = rbd_dev_image_id(rbd_dev);
4828 rbd_assert(rbd_dev->spec->image_id);
4829 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
4831 ret = rbd_dev_header_name(rbd_dev);
4833 goto err_out_format;
4835 if (rbd_dev->image_format == 1)
4836 ret = rbd_dev_v1_probe(rbd_dev);
4838 ret = rbd_dev_v2_probe(rbd_dev);
4840 goto out_header_name;
4842 ret = rbd_dev_snaps_update(rbd_dev);
4844 goto out_header_name;
4846 ret = rbd_dev_spec_update(rbd_dev);
4850 ret = rbd_dev_probe_parent(rbd_dev);
4854 ret = rbd_dev_probe_finish(rbd_dev);
4856 goto err_out_parent;
4860 rbd_dev_remove_parent(rbd_dev);
4861 rbd_header_free(&rbd_dev->header);
4863 rbd_remove_all_snaps(rbd_dev);
4865 kfree(rbd_dev->header_name);
4866 rbd_dev->header_name = NULL;
4868 rbd_dev->image_format = 0;
4869 kfree(rbd_dev->spec->image_id);
4870 rbd_dev->spec->image_id = NULL;
4872 dout("probe failed, returning %d\n", ret);
4877 static ssize_t rbd_add(struct bus_type *bus,
4881 struct rbd_device *rbd_dev = NULL;
4882 struct ceph_options *ceph_opts = NULL;
4883 struct rbd_options *rbd_opts = NULL;
4884 struct rbd_spec *spec = NULL;
4885 struct rbd_client *rbdc;
4886 struct ceph_osd_client *osdc;
4889 if (!try_module_get(THIS_MODULE))
4892 /* parse add command */
4893 rc = rbd_add_parse_args(buf, &ceph_opts, &rbd_opts, &spec);
4895 goto err_out_module;
4897 rbdc = rbd_get_client(ceph_opts);
4902 ceph_opts = NULL; /* rbd_dev client now owns this */
4905 osdc = &rbdc->client->osdc;
4906 rc = ceph_pg_poolid_by_name(osdc->osdmap, spec->pool_name);
4908 goto err_out_client;
4909 spec->pool_id = (u64)rc;
4911 /* The ceph file layout needs to fit pool id in 32 bits */
4913 if (spec->pool_id > (u64)U32_MAX) {
4914 rbd_warn(NULL, "pool id too large (%llu > %u)\n",
4915 (unsigned long long)spec->pool_id, U32_MAX);
4917 goto err_out_client;
4920 rbd_dev = rbd_dev_create(rbdc, spec);
4922 goto err_out_client;
4923 rbdc = NULL; /* rbd_dev now owns this */
4924 spec = NULL; /* rbd_dev now owns this */
4926 rbd_dev->mapping.read_only = rbd_opts->read_only;
4928 rbd_opts = NULL; /* done with this */
4930 rc = rbd_dev_image_probe(rbd_dev);
4932 goto err_out_rbd_dev;
4936 kfree(rbd_dev->header_name);
4937 rbd_dev_destroy(rbd_dev);
4939 rbd_put_client(rbdc);
4942 ceph_destroy_options(ceph_opts);
4946 module_put(THIS_MODULE);
4948 dout("Error adding device %s\n", buf);
4953 static struct rbd_device *__rbd_get_dev(unsigned long dev_id)
4955 struct list_head *tmp;
4956 struct rbd_device *rbd_dev;
4958 spin_lock(&rbd_dev_list_lock);
4959 list_for_each(tmp, &rbd_dev_list) {
4960 rbd_dev = list_entry(tmp, struct rbd_device, node);
4961 if (rbd_dev->dev_id == dev_id) {
4962 spin_unlock(&rbd_dev_list_lock);
4966 spin_unlock(&rbd_dev_list_lock);
4970 static void rbd_dev_release(struct device *dev)
4972 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
4974 /* clean up and free blkdev */
4975 rbd_free_disk(rbd_dev);
4976 unregister_blkdev(rbd_dev->major, rbd_dev->name);
4978 /* release allocated disk header fields */
4979 rbd_header_free(&rbd_dev->header);
4981 /* done with the id, and with the rbd_dev */
4982 rbd_dev_id_put(rbd_dev);
4983 rbd_dev_mapping_clear(rbd_dev);
4984 rbd_assert(rbd_dev->rbd_client != NULL);
4985 rbd_spec_put(rbd_dev->parent_spec);
4986 kfree(rbd_dev->header_name);
4987 rbd_dev_destroy(rbd_dev);
4989 /* release module ref */
4990 module_put(THIS_MODULE);
4993 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev)
4995 while (rbd_dev->parent) {
4996 struct rbd_device *first = rbd_dev;
4997 struct rbd_device *second = first->parent;
4998 struct rbd_device *third;
5002 * Follow to the parent with no grandparent and
5005 while (second && (third = second->parent)) {
5010 ret = rbd_dev_header_watch_sync(rbd_dev, 0);
5013 "failed to cancel watch event (%d)\n", ret);
5014 rbd_remove_all_snaps(second);
5015 rbd_bus_del_dev(second);
5016 first->parent = NULL;
5017 first->parent_overlap = 0;
5019 rbd_assert(first->parent_spec);
5020 rbd_spec_put(first->parent_spec);
5021 first->parent_spec = NULL;
5025 static ssize_t rbd_remove(struct bus_type *bus,
5029 struct rbd_device *rbd_dev = NULL;
5034 ret = strict_strtoul(buf, 10, &ul);
5038 /* convert to int; abort if we lost anything in the conversion */
5039 target_id = (int) ul;
5040 if (target_id != ul)
5043 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
5045 rbd_dev = __rbd_get_dev(target_id);
5051 spin_lock_irq(&rbd_dev->lock);
5052 if (rbd_dev->open_count)
5055 set_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags);
5056 spin_unlock_irq(&rbd_dev->lock);
5060 ret = rbd_dev_header_watch_sync(rbd_dev, 0);
5062 rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret);
5063 clear_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags);
5069 rbd_dev_remove_parent(rbd_dev);
5071 rbd_remove_all_snaps(rbd_dev);
5072 rbd_bus_del_dev(rbd_dev);
5074 mutex_unlock(&ctl_mutex);
5080 * create control files in sysfs
5083 static int rbd_sysfs_init(void)
5087 ret = device_register(&rbd_root_dev);
5091 ret = bus_register(&rbd_bus_type);
5093 device_unregister(&rbd_root_dev);
5098 static void rbd_sysfs_cleanup(void)
5100 bus_unregister(&rbd_bus_type);
5101 device_unregister(&rbd_root_dev);
5104 static int __init rbd_init(void)
5108 if (!libceph_compatible(NULL)) {
5109 rbd_warn(NULL, "libceph incompatibility (quitting)");
5113 rc = rbd_sysfs_init();
5116 pr_info("loaded " RBD_DRV_NAME_LONG "\n");
5120 static void __exit rbd_exit(void)
5122 rbd_sysfs_cleanup();
5125 module_init(rbd_init);
5126 module_exit(rbd_exit);
5128 MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
5129 MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
5130 MODULE_DESCRIPTION("rados block device");
5132 /* following authorship retained from original osdblk.c */
5133 MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");
5135 MODULE_LICENSE("GPL");