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_device_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);
676 * Drop reference to ceph client node. If it's not referenced anymore, release
679 static void rbd_put_client(struct rbd_client *rbdc)
682 kref_put(&rbdc->kref, rbd_client_release);
685 static bool rbd_image_format_valid(u32 image_format)
687 return image_format == 1 || image_format == 2;
690 static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
695 /* The header has to start with the magic rbd header text */
696 if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT)))
699 /* The bio layer requires at least sector-sized I/O */
701 if (ondisk->options.order < SECTOR_SHIFT)
704 /* If we use u64 in a few spots we may be able to loosen this */
706 if (ondisk->options.order > 8 * sizeof (int) - 1)
710 * The size of a snapshot header has to fit in a size_t, and
711 * that limits the number of snapshots.
713 snap_count = le32_to_cpu(ondisk->snap_count);
714 size = SIZE_MAX - sizeof (struct ceph_snap_context);
715 if (snap_count > size / sizeof (__le64))
719 * Not only that, but the size of the entire the snapshot
720 * header must also be representable in a size_t.
722 size -= snap_count * sizeof (__le64);
723 if ((u64) size < le64_to_cpu(ondisk->snap_names_len))
730 * Create a new header structure, translate header format from the on-disk
733 static int rbd_header_from_disk(struct rbd_image_header *header,
734 struct rbd_image_header_ondisk *ondisk)
741 memset(header, 0, sizeof (*header));
743 snap_count = le32_to_cpu(ondisk->snap_count);
745 len = strnlen(ondisk->object_prefix, sizeof (ondisk->object_prefix));
746 header->object_prefix = kmalloc(len + 1, GFP_KERNEL);
747 if (!header->object_prefix)
749 memcpy(header->object_prefix, ondisk->object_prefix, len);
750 header->object_prefix[len] = '\0';
753 u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len);
755 /* Save a copy of the snapshot names */
757 if (snap_names_len > (u64) SIZE_MAX)
759 header->snap_names = kmalloc(snap_names_len, GFP_KERNEL);
760 if (!header->snap_names)
763 * Note that rbd_dev_v1_header_read() guarantees
764 * the ondisk buffer we're working with has
765 * snap_names_len bytes beyond the end of the
766 * snapshot id array, this memcpy() is safe.
768 memcpy(header->snap_names, &ondisk->snaps[snap_count],
771 /* Record each snapshot's size */
773 size = snap_count * sizeof (*header->snap_sizes);
774 header->snap_sizes = kmalloc(size, GFP_KERNEL);
775 if (!header->snap_sizes)
777 for (i = 0; i < snap_count; i++)
778 header->snap_sizes[i] =
779 le64_to_cpu(ondisk->snaps[i].image_size);
781 header->snap_names = NULL;
782 header->snap_sizes = NULL;
785 header->features = 0; /* No features support in v1 images */
786 header->obj_order = ondisk->options.order;
787 header->crypt_type = ondisk->options.crypt_type;
788 header->comp_type = ondisk->options.comp_type;
790 /* Allocate and fill in the snapshot context */
792 header->image_size = le64_to_cpu(ondisk->image_size);
794 header->snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
797 header->snapc->seq = le64_to_cpu(ondisk->snap_seq);
798 for (i = 0; i < snap_count; i++)
799 header->snapc->snaps[i] = le64_to_cpu(ondisk->snaps[i].id);
804 kfree(header->snap_sizes);
805 header->snap_sizes = NULL;
806 kfree(header->snap_names);
807 header->snap_names = NULL;
808 kfree(header->object_prefix);
809 header->object_prefix = NULL;
814 static const char *rbd_snap_name(struct rbd_device *rbd_dev, u64 snap_id)
816 struct rbd_snap *snap;
818 if (snap_id == CEPH_NOSNAP)
819 return RBD_SNAP_HEAD_NAME;
821 list_for_each_entry(snap, &rbd_dev->snaps, node)
822 if (snap_id == snap->id)
828 static struct rbd_snap *snap_by_name(struct rbd_device *rbd_dev,
829 const char *snap_name)
831 struct rbd_snap *snap;
833 list_for_each_entry(snap, &rbd_dev->snaps, node)
834 if (!strcmp(snap_name, snap->name))
840 static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
842 if (!memcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME,
843 sizeof (RBD_SNAP_HEAD_NAME))) {
844 rbd_dev->mapping.size = rbd_dev->header.image_size;
845 rbd_dev->mapping.features = rbd_dev->header.features;
847 struct rbd_snap *snap;
849 snap = snap_by_name(rbd_dev, rbd_dev->spec->snap_name);
852 rbd_dev->mapping.size = snap->size;
853 rbd_dev->mapping.features = snap->features;
854 rbd_dev->mapping.read_only = true;
860 static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
862 rbd_dev->mapping.size = 0;
863 rbd_dev->mapping.features = 0;
864 rbd_dev->mapping.read_only = true;
867 static void rbd_dev_clear_mapping(struct rbd_device *rbd_dev)
869 rbd_dev->mapping.size = 0;
870 rbd_dev->mapping.features = 0;
871 rbd_dev->mapping.read_only = true;
874 static const char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset)
880 name = kmalloc(MAX_OBJ_NAME_SIZE + 1, GFP_NOIO);
883 segment = offset >> rbd_dev->header.obj_order;
884 ret = snprintf(name, MAX_OBJ_NAME_SIZE + 1, "%s.%012llx",
885 rbd_dev->header.object_prefix, segment);
886 if (ret < 0 || ret > MAX_OBJ_NAME_SIZE) {
887 pr_err("error formatting segment name for #%llu (%d)\n",
896 static u64 rbd_segment_offset(struct rbd_device *rbd_dev, u64 offset)
898 u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
900 return offset & (segment_size - 1);
903 static u64 rbd_segment_length(struct rbd_device *rbd_dev,
904 u64 offset, u64 length)
906 u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
908 offset &= segment_size - 1;
910 rbd_assert(length <= U64_MAX - offset);
911 if (offset + length > segment_size)
912 length = segment_size - offset;
918 * returns the size of an object in the image
920 static u64 rbd_obj_bytes(struct rbd_image_header *header)
922 return 1 << header->obj_order;
929 static void bio_chain_put(struct bio *chain)
935 chain = chain->bi_next;
941 * zeros a bio chain, starting at specific offset
943 static void zero_bio_chain(struct bio *chain, int start_ofs)
952 bio_for_each_segment(bv, chain, i) {
953 if (pos + bv->bv_len > start_ofs) {
954 int remainder = max(start_ofs - pos, 0);
955 buf = bvec_kmap_irq(bv, &flags);
956 memset(buf + remainder, 0,
957 bv->bv_len - remainder);
958 bvec_kunmap_irq(buf, &flags);
963 chain = chain->bi_next;
968 * similar to zero_bio_chain(), zeros data defined by a page array,
969 * starting at the given byte offset from the start of the array and
970 * continuing up to the given end offset. The pages array is
971 * assumed to be big enough to hold all bytes up to the end.
973 static void zero_pages(struct page **pages, u64 offset, u64 end)
975 struct page **page = &pages[offset >> PAGE_SHIFT];
977 rbd_assert(end > offset);
978 rbd_assert(end - offset <= (u64)SIZE_MAX);
979 while (offset < end) {
985 page_offset = (size_t)(offset & ~PAGE_MASK);
986 length = min(PAGE_SIZE - page_offset, (size_t)(end - offset));
987 local_irq_save(flags);
988 kaddr = kmap_atomic(*page);
989 memset(kaddr + page_offset, 0, length);
990 kunmap_atomic(kaddr);
991 local_irq_restore(flags);
999 * Clone a portion of a bio, starting at the given byte offset
1000 * and continuing for the number of bytes indicated.
1002 static struct bio *bio_clone_range(struct bio *bio_src,
1003 unsigned int offset,
1011 unsigned short end_idx;
1012 unsigned short vcnt;
1015 /* Handle the easy case for the caller */
1017 if (!offset && len == bio_src->bi_size)
1018 return bio_clone(bio_src, gfpmask);
1020 if (WARN_ON_ONCE(!len))
1022 if (WARN_ON_ONCE(len > bio_src->bi_size))
1024 if (WARN_ON_ONCE(offset > bio_src->bi_size - len))
1027 /* Find first affected segment... */
1030 __bio_for_each_segment(bv, bio_src, idx, 0) {
1031 if (resid < bv->bv_len)
1033 resid -= bv->bv_len;
1037 /* ...and the last affected segment */
1040 __bio_for_each_segment(bv, bio_src, end_idx, idx) {
1041 if (resid <= bv->bv_len)
1043 resid -= bv->bv_len;
1045 vcnt = end_idx - idx + 1;
1047 /* Build the clone */
1049 bio = bio_alloc(gfpmask, (unsigned int) vcnt);
1051 return NULL; /* ENOMEM */
1053 bio->bi_bdev = bio_src->bi_bdev;
1054 bio->bi_sector = bio_src->bi_sector + (offset >> SECTOR_SHIFT);
1055 bio->bi_rw = bio_src->bi_rw;
1056 bio->bi_flags |= 1 << BIO_CLONED;
1059 * Copy over our part of the bio_vec, then update the first
1060 * and last (or only) entries.
1062 memcpy(&bio->bi_io_vec[0], &bio_src->bi_io_vec[idx],
1063 vcnt * sizeof (struct bio_vec));
1064 bio->bi_io_vec[0].bv_offset += voff;
1066 bio->bi_io_vec[0].bv_len -= voff;
1067 bio->bi_io_vec[vcnt - 1].bv_len = resid;
1069 bio->bi_io_vec[0].bv_len = len;
1072 bio->bi_vcnt = vcnt;
1080 * Clone a portion of a bio chain, starting at the given byte offset
1081 * into the first bio in the source chain and continuing for the
1082 * number of bytes indicated. The result is another bio chain of
1083 * exactly the given length, or a null pointer on error.
1085 * The bio_src and offset parameters are both in-out. On entry they
1086 * refer to the first source bio and the offset into that bio where
1087 * the start of data to be cloned is located.
1089 * On return, bio_src is updated to refer to the bio in the source
1090 * chain that contains first un-cloned byte, and *offset will
1091 * contain the offset of that byte within that bio.
1093 static struct bio *bio_chain_clone_range(struct bio **bio_src,
1094 unsigned int *offset,
1098 struct bio *bi = *bio_src;
1099 unsigned int off = *offset;
1100 struct bio *chain = NULL;
1103 /* Build up a chain of clone bios up to the limit */
1105 if (!bi || off >= bi->bi_size || !len)
1106 return NULL; /* Nothing to clone */
1110 unsigned int bi_size;
1114 rbd_warn(NULL, "bio_chain exhausted with %u left", len);
1115 goto out_err; /* EINVAL; ran out of bio's */
1117 bi_size = min_t(unsigned int, bi->bi_size - off, len);
1118 bio = bio_clone_range(bi, off, bi_size, gfpmask);
1120 goto out_err; /* ENOMEM */
1123 end = &bio->bi_next;
1126 if (off == bi->bi_size) {
1137 bio_chain_put(chain);
1143 * The default/initial value for all object request flags is 0. For
1144 * each flag, once its value is set to 1 it is never reset to 0
1147 static void obj_request_img_data_set(struct rbd_obj_request *obj_request)
1149 if (test_and_set_bit(OBJ_REQ_IMG_DATA, &obj_request->flags)) {
1150 struct rbd_device *rbd_dev;
1152 rbd_dev = obj_request->img_request->rbd_dev;
1153 rbd_warn(rbd_dev, "obj_request %p already marked img_data\n",
1158 static bool obj_request_img_data_test(struct rbd_obj_request *obj_request)
1161 return test_bit(OBJ_REQ_IMG_DATA, &obj_request->flags) != 0;
1164 static void obj_request_done_set(struct rbd_obj_request *obj_request)
1166 if (test_and_set_bit(OBJ_REQ_DONE, &obj_request->flags)) {
1167 struct rbd_device *rbd_dev = NULL;
1169 if (obj_request_img_data_test(obj_request))
1170 rbd_dev = obj_request->img_request->rbd_dev;
1171 rbd_warn(rbd_dev, "obj_request %p already marked done\n",
1176 static bool obj_request_done_test(struct rbd_obj_request *obj_request)
1179 return test_bit(OBJ_REQ_DONE, &obj_request->flags) != 0;
1183 * This sets the KNOWN flag after (possibly) setting the EXISTS
1184 * flag. The latter is set based on the "exists" value provided.
1186 * Note that for our purposes once an object exists it never goes
1187 * away again. It's possible that the response from two existence
1188 * checks are separated by the creation of the target object, and
1189 * the first ("doesn't exist") response arrives *after* the second
1190 * ("does exist"). In that case we ignore the second one.
1192 static void obj_request_existence_set(struct rbd_obj_request *obj_request,
1196 set_bit(OBJ_REQ_EXISTS, &obj_request->flags);
1197 set_bit(OBJ_REQ_KNOWN, &obj_request->flags);
1201 static bool obj_request_known_test(struct rbd_obj_request *obj_request)
1204 return test_bit(OBJ_REQ_KNOWN, &obj_request->flags) != 0;
1207 static bool obj_request_exists_test(struct rbd_obj_request *obj_request)
1210 return test_bit(OBJ_REQ_EXISTS, &obj_request->flags) != 0;
1213 static void rbd_obj_request_get(struct rbd_obj_request *obj_request)
1215 dout("%s: obj %p (was %d)\n", __func__, obj_request,
1216 atomic_read(&obj_request->kref.refcount));
1217 kref_get(&obj_request->kref);
1220 static void rbd_obj_request_destroy(struct kref *kref);
1221 static void rbd_obj_request_put(struct rbd_obj_request *obj_request)
1223 rbd_assert(obj_request != NULL);
1224 dout("%s: obj %p (was %d)\n", __func__, obj_request,
1225 atomic_read(&obj_request->kref.refcount));
1226 kref_put(&obj_request->kref, rbd_obj_request_destroy);
1229 static void rbd_img_request_get(struct rbd_img_request *img_request)
1231 dout("%s: img %p (was %d)\n", __func__, img_request,
1232 atomic_read(&img_request->kref.refcount));
1233 kref_get(&img_request->kref);
1236 static void rbd_img_request_destroy(struct kref *kref);
1237 static void rbd_img_request_put(struct rbd_img_request *img_request)
1239 rbd_assert(img_request != NULL);
1240 dout("%s: img %p (was %d)\n", __func__, img_request,
1241 atomic_read(&img_request->kref.refcount));
1242 kref_put(&img_request->kref, rbd_img_request_destroy);
1245 static inline void rbd_img_obj_request_add(struct rbd_img_request *img_request,
1246 struct rbd_obj_request *obj_request)
1248 rbd_assert(obj_request->img_request == NULL);
1250 /* Image request now owns object's original reference */
1251 obj_request->img_request = img_request;
1252 obj_request->which = img_request->obj_request_count;
1253 rbd_assert(!obj_request_img_data_test(obj_request));
1254 obj_request_img_data_set(obj_request);
1255 rbd_assert(obj_request->which != BAD_WHICH);
1256 img_request->obj_request_count++;
1257 list_add_tail(&obj_request->links, &img_request->obj_requests);
1258 dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request,
1259 obj_request->which);
1262 static inline void rbd_img_obj_request_del(struct rbd_img_request *img_request,
1263 struct rbd_obj_request *obj_request)
1265 rbd_assert(obj_request->which != BAD_WHICH);
1267 dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request,
1268 obj_request->which);
1269 list_del(&obj_request->links);
1270 rbd_assert(img_request->obj_request_count > 0);
1271 img_request->obj_request_count--;
1272 rbd_assert(obj_request->which == img_request->obj_request_count);
1273 obj_request->which = BAD_WHICH;
1274 rbd_assert(obj_request_img_data_test(obj_request));
1275 rbd_assert(obj_request->img_request == img_request);
1276 obj_request->img_request = NULL;
1277 obj_request->callback = NULL;
1278 rbd_obj_request_put(obj_request);
1281 static bool obj_request_type_valid(enum obj_request_type type)
1284 case OBJ_REQUEST_NODATA:
1285 case OBJ_REQUEST_BIO:
1286 case OBJ_REQUEST_PAGES:
1293 static int rbd_obj_request_submit(struct ceph_osd_client *osdc,
1294 struct rbd_obj_request *obj_request)
1296 dout("%s: osdc %p obj %p\n", __func__, osdc, obj_request);
1298 return ceph_osdc_start_request(osdc, obj_request->osd_req, false);
1301 static void rbd_img_request_complete(struct rbd_img_request *img_request)
1304 dout("%s: img %p\n", __func__, img_request);
1307 * If no error occurred, compute the aggregate transfer
1308 * count for the image request. We could instead use
1309 * atomic64_cmpxchg() to update it as each object request
1310 * completes; not clear which way is better off hand.
1312 if (!img_request->result) {
1313 struct rbd_obj_request *obj_request;
1316 for_each_obj_request(img_request, obj_request)
1317 xferred += obj_request->xferred;
1318 img_request->xferred = xferred;
1321 if (img_request->callback)
1322 img_request->callback(img_request);
1324 rbd_img_request_put(img_request);
1327 /* Caller is responsible for rbd_obj_request_destroy(obj_request) */
1329 static int rbd_obj_request_wait(struct rbd_obj_request *obj_request)
1331 dout("%s: obj %p\n", __func__, obj_request);
1333 return wait_for_completion_interruptible(&obj_request->completion);
1337 * The default/initial value for all image request flags is 0. Each
1338 * is conditionally set to 1 at image request initialization time
1339 * and currently never change thereafter.
1341 static void img_request_write_set(struct rbd_img_request *img_request)
1343 set_bit(IMG_REQ_WRITE, &img_request->flags);
1347 static bool img_request_write_test(struct rbd_img_request *img_request)
1350 return test_bit(IMG_REQ_WRITE, &img_request->flags) != 0;
1353 static void img_request_child_set(struct rbd_img_request *img_request)
1355 set_bit(IMG_REQ_CHILD, &img_request->flags);
1359 static bool img_request_child_test(struct rbd_img_request *img_request)
1362 return test_bit(IMG_REQ_CHILD, &img_request->flags) != 0;
1365 static void img_request_layered_set(struct rbd_img_request *img_request)
1367 set_bit(IMG_REQ_LAYERED, &img_request->flags);
1371 static bool img_request_layered_test(struct rbd_img_request *img_request)
1374 return test_bit(IMG_REQ_LAYERED, &img_request->flags) != 0;
1378 rbd_img_obj_request_read_callback(struct rbd_obj_request *obj_request)
1380 u64 xferred = obj_request->xferred;
1381 u64 length = obj_request->length;
1383 dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
1384 obj_request, obj_request->img_request, obj_request->result,
1387 * ENOENT means a hole in the image. We zero-fill the
1388 * entire length of the request. A short read also implies
1389 * zero-fill to the end of the request. Either way we
1390 * update the xferred count to indicate the whole request
1393 rbd_assert(obj_request->type != OBJ_REQUEST_NODATA);
1394 if (obj_request->result == -ENOENT) {
1395 if (obj_request->type == OBJ_REQUEST_BIO)
1396 zero_bio_chain(obj_request->bio_list, 0);
1398 zero_pages(obj_request->pages, 0, length);
1399 obj_request->result = 0;
1400 obj_request->xferred = length;
1401 } else if (xferred < length && !obj_request->result) {
1402 if (obj_request->type == OBJ_REQUEST_BIO)
1403 zero_bio_chain(obj_request->bio_list, xferred);
1405 zero_pages(obj_request->pages, xferred, length);
1406 obj_request->xferred = length;
1408 obj_request_done_set(obj_request);
1411 static void rbd_obj_request_complete(struct rbd_obj_request *obj_request)
1413 dout("%s: obj %p cb %p\n", __func__, obj_request,
1414 obj_request->callback);
1415 if (obj_request->callback)
1416 obj_request->callback(obj_request);
1418 complete_all(&obj_request->completion);
1421 static void rbd_osd_trivial_callback(struct rbd_obj_request *obj_request)
1423 dout("%s: obj %p\n", __func__, obj_request);
1424 obj_request_done_set(obj_request);
1427 static void rbd_osd_read_callback(struct rbd_obj_request *obj_request)
1429 struct rbd_img_request *img_request = NULL;
1430 struct rbd_device *rbd_dev = NULL;
1431 bool layered = false;
1433 if (obj_request_img_data_test(obj_request)) {
1434 img_request = obj_request->img_request;
1435 layered = img_request && img_request_layered_test(img_request);
1436 rbd_dev = img_request->rbd_dev;
1439 dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
1440 obj_request, img_request, obj_request->result,
1441 obj_request->xferred, obj_request->length);
1442 if (layered && obj_request->result == -ENOENT &&
1443 obj_request->img_offset < rbd_dev->parent_overlap)
1444 rbd_img_parent_read(obj_request);
1445 else if (img_request)
1446 rbd_img_obj_request_read_callback(obj_request);
1448 obj_request_done_set(obj_request);
1451 static void rbd_osd_write_callback(struct rbd_obj_request *obj_request)
1453 dout("%s: obj %p result %d %llu\n", __func__, obj_request,
1454 obj_request->result, obj_request->length);
1456 * There is no such thing as a successful short write. Set
1457 * it to our originally-requested length.
1459 obj_request->xferred = obj_request->length;
1460 obj_request_done_set(obj_request);
1464 * For a simple stat call there's nothing to do. We'll do more if
1465 * this is part of a write sequence for a layered image.
1467 static void rbd_osd_stat_callback(struct rbd_obj_request *obj_request)
1469 dout("%s: obj %p\n", __func__, obj_request);
1470 obj_request_done_set(obj_request);
1473 static void rbd_osd_req_callback(struct ceph_osd_request *osd_req,
1474 struct ceph_msg *msg)
1476 struct rbd_obj_request *obj_request = osd_req->r_priv;
1479 dout("%s: osd_req %p msg %p\n", __func__, osd_req, msg);
1480 rbd_assert(osd_req == obj_request->osd_req);
1481 if (obj_request_img_data_test(obj_request)) {
1482 rbd_assert(obj_request->img_request);
1483 rbd_assert(obj_request->which != BAD_WHICH);
1485 rbd_assert(obj_request->which == BAD_WHICH);
1488 if (osd_req->r_result < 0)
1489 obj_request->result = osd_req->r_result;
1490 obj_request->version = le64_to_cpu(osd_req->r_reassert_version.version);
1492 BUG_ON(osd_req->r_num_ops > 2);
1495 * We support a 64-bit length, but ultimately it has to be
1496 * passed to blk_end_request(), which takes an unsigned int.
1498 obj_request->xferred = osd_req->r_reply_op_len[0];
1499 rbd_assert(obj_request->xferred < (u64)UINT_MAX);
1500 opcode = osd_req->r_ops[0].op;
1502 case CEPH_OSD_OP_READ:
1503 rbd_osd_read_callback(obj_request);
1505 case CEPH_OSD_OP_WRITE:
1506 rbd_osd_write_callback(obj_request);
1508 case CEPH_OSD_OP_STAT:
1509 rbd_osd_stat_callback(obj_request);
1511 case CEPH_OSD_OP_CALL:
1512 case CEPH_OSD_OP_NOTIFY_ACK:
1513 case CEPH_OSD_OP_WATCH:
1514 rbd_osd_trivial_callback(obj_request);
1517 rbd_warn(NULL, "%s: unsupported op %hu\n",
1518 obj_request->object_name, (unsigned short) opcode);
1522 if (obj_request_done_test(obj_request))
1523 rbd_obj_request_complete(obj_request);
1526 static void rbd_osd_req_format_read(struct rbd_obj_request *obj_request)
1528 struct rbd_img_request *img_request = obj_request->img_request;
1529 struct ceph_osd_request *osd_req = obj_request->osd_req;
1532 rbd_assert(osd_req != NULL);
1534 snap_id = img_request ? img_request->snap_id : CEPH_NOSNAP;
1535 ceph_osdc_build_request(osd_req, obj_request->offset,
1536 NULL, snap_id, NULL);
1539 static void rbd_osd_req_format_write(struct rbd_obj_request *obj_request)
1541 struct rbd_img_request *img_request = obj_request->img_request;
1542 struct ceph_osd_request *osd_req = obj_request->osd_req;
1543 struct ceph_snap_context *snapc;
1544 struct timespec mtime = CURRENT_TIME;
1546 rbd_assert(osd_req != NULL);
1548 snapc = img_request ? img_request->snapc : NULL;
1549 ceph_osdc_build_request(osd_req, obj_request->offset,
1550 snapc, CEPH_NOSNAP, &mtime);
1553 static struct ceph_osd_request *rbd_osd_req_create(
1554 struct rbd_device *rbd_dev,
1556 struct rbd_obj_request *obj_request)
1558 struct ceph_snap_context *snapc = NULL;
1559 struct ceph_osd_client *osdc;
1560 struct ceph_osd_request *osd_req;
1562 if (obj_request_img_data_test(obj_request)) {
1563 struct rbd_img_request *img_request = obj_request->img_request;
1565 rbd_assert(write_request ==
1566 img_request_write_test(img_request));
1568 snapc = img_request->snapc;
1571 /* Allocate and initialize the request, for the single op */
1573 osdc = &rbd_dev->rbd_client->client->osdc;
1574 osd_req = ceph_osdc_alloc_request(osdc, snapc, 1, false, GFP_ATOMIC);
1576 return NULL; /* ENOMEM */
1579 osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK;
1581 osd_req->r_flags = CEPH_OSD_FLAG_READ;
1583 osd_req->r_callback = rbd_osd_req_callback;
1584 osd_req->r_priv = obj_request;
1586 osd_req->r_oid_len = strlen(obj_request->object_name);
1587 rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid));
1588 memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len);
1590 osd_req->r_file_layout = rbd_dev->layout; /* struct */
1596 * Create a copyup osd request based on the information in the
1597 * object request supplied. A copyup request has two osd ops,
1598 * a copyup method call, and a "normal" write request.
1600 static struct ceph_osd_request *
1601 rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request)
1603 struct rbd_img_request *img_request;
1604 struct ceph_snap_context *snapc;
1605 struct rbd_device *rbd_dev;
1606 struct ceph_osd_client *osdc;
1607 struct ceph_osd_request *osd_req;
1609 rbd_assert(obj_request_img_data_test(obj_request));
1610 img_request = obj_request->img_request;
1611 rbd_assert(img_request);
1612 rbd_assert(img_request_write_test(img_request));
1614 /* Allocate and initialize the request, for the two ops */
1616 snapc = img_request->snapc;
1617 rbd_dev = img_request->rbd_dev;
1618 osdc = &rbd_dev->rbd_client->client->osdc;
1619 osd_req = ceph_osdc_alloc_request(osdc, snapc, 2, false, GFP_ATOMIC);
1621 return NULL; /* ENOMEM */
1623 osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK;
1624 osd_req->r_callback = rbd_osd_req_callback;
1625 osd_req->r_priv = obj_request;
1627 osd_req->r_oid_len = strlen(obj_request->object_name);
1628 rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid));
1629 memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len);
1631 osd_req->r_file_layout = rbd_dev->layout; /* struct */
1637 static void rbd_osd_req_destroy(struct ceph_osd_request *osd_req)
1639 ceph_osdc_put_request(osd_req);
1642 /* object_name is assumed to be a non-null pointer and NUL-terminated */
1644 static struct rbd_obj_request *rbd_obj_request_create(const char *object_name,
1645 u64 offset, u64 length,
1646 enum obj_request_type type)
1648 struct rbd_obj_request *obj_request;
1652 rbd_assert(obj_request_type_valid(type));
1654 size = strlen(object_name) + 1;
1655 obj_request = kzalloc(sizeof (*obj_request) + size, GFP_KERNEL);
1659 name = (char *)(obj_request + 1);
1660 obj_request->object_name = memcpy(name, object_name, size);
1661 obj_request->offset = offset;
1662 obj_request->length = length;
1663 obj_request->flags = 0;
1664 obj_request->which = BAD_WHICH;
1665 obj_request->type = type;
1666 INIT_LIST_HEAD(&obj_request->links);
1667 init_completion(&obj_request->completion);
1668 kref_init(&obj_request->kref);
1670 dout("%s: \"%s\" %llu/%llu %d -> obj %p\n", __func__, object_name,
1671 offset, length, (int)type, obj_request);
1676 static void rbd_obj_request_destroy(struct kref *kref)
1678 struct rbd_obj_request *obj_request;
1680 obj_request = container_of(kref, struct rbd_obj_request, kref);
1682 dout("%s: obj %p\n", __func__, obj_request);
1684 rbd_assert(obj_request->img_request == NULL);
1685 rbd_assert(obj_request->which == BAD_WHICH);
1687 if (obj_request->osd_req)
1688 rbd_osd_req_destroy(obj_request->osd_req);
1690 rbd_assert(obj_request_type_valid(obj_request->type));
1691 switch (obj_request->type) {
1692 case OBJ_REQUEST_NODATA:
1693 break; /* Nothing to do */
1694 case OBJ_REQUEST_BIO:
1695 if (obj_request->bio_list)
1696 bio_chain_put(obj_request->bio_list);
1698 case OBJ_REQUEST_PAGES:
1699 if (obj_request->pages)
1700 ceph_release_page_vector(obj_request->pages,
1701 obj_request->page_count);
1709 * Caller is responsible for filling in the list of object requests
1710 * that comprises the image request, and the Linux request pointer
1711 * (if there is one).
1713 static struct rbd_img_request *rbd_img_request_create(
1714 struct rbd_device *rbd_dev,
1715 u64 offset, u64 length,
1719 struct rbd_img_request *img_request;
1721 img_request = kmalloc(sizeof (*img_request), GFP_ATOMIC);
1725 if (write_request) {
1726 down_read(&rbd_dev->header_rwsem);
1727 ceph_get_snap_context(rbd_dev->header.snapc);
1728 up_read(&rbd_dev->header_rwsem);
1731 img_request->rq = NULL;
1732 img_request->rbd_dev = rbd_dev;
1733 img_request->offset = offset;
1734 img_request->length = length;
1735 img_request->flags = 0;
1736 if (write_request) {
1737 img_request_write_set(img_request);
1738 img_request->snapc = rbd_dev->header.snapc;
1740 img_request->snap_id = rbd_dev->spec->snap_id;
1743 img_request_child_set(img_request);
1744 if (rbd_dev->parent_spec)
1745 img_request_layered_set(img_request);
1746 spin_lock_init(&img_request->completion_lock);
1747 img_request->next_completion = 0;
1748 img_request->callback = NULL;
1749 img_request->result = 0;
1750 img_request->obj_request_count = 0;
1751 INIT_LIST_HEAD(&img_request->obj_requests);
1752 kref_init(&img_request->kref);
1754 rbd_img_request_get(img_request); /* Avoid a warning */
1755 rbd_img_request_put(img_request); /* TEMPORARY */
1757 dout("%s: rbd_dev %p %s %llu/%llu -> img %p\n", __func__, rbd_dev,
1758 write_request ? "write" : "read", offset, length,
1764 static void rbd_img_request_destroy(struct kref *kref)
1766 struct rbd_img_request *img_request;
1767 struct rbd_obj_request *obj_request;
1768 struct rbd_obj_request *next_obj_request;
1770 img_request = container_of(kref, struct rbd_img_request, kref);
1772 dout("%s: img %p\n", __func__, img_request);
1774 for_each_obj_request_safe(img_request, obj_request, next_obj_request)
1775 rbd_img_obj_request_del(img_request, obj_request);
1776 rbd_assert(img_request->obj_request_count == 0);
1778 if (img_request_write_test(img_request))
1779 ceph_put_snap_context(img_request->snapc);
1781 if (img_request_child_test(img_request))
1782 rbd_obj_request_put(img_request->obj_request);
1787 static bool rbd_img_obj_end_request(struct rbd_obj_request *obj_request)
1789 struct rbd_img_request *img_request;
1790 unsigned int xferred;
1794 rbd_assert(obj_request_img_data_test(obj_request));
1795 img_request = obj_request->img_request;
1797 rbd_assert(obj_request->xferred <= (u64)UINT_MAX);
1798 xferred = (unsigned int)obj_request->xferred;
1799 result = obj_request->result;
1801 struct rbd_device *rbd_dev = img_request->rbd_dev;
1803 rbd_warn(rbd_dev, "%s %llx at %llx (%llx)\n",
1804 img_request_write_test(img_request) ? "write" : "read",
1805 obj_request->length, obj_request->img_offset,
1806 obj_request->offset);
1807 rbd_warn(rbd_dev, " result %d xferred %x\n",
1809 if (!img_request->result)
1810 img_request->result = result;
1813 /* Image object requests don't own their page array */
1815 if (obj_request->type == OBJ_REQUEST_PAGES) {
1816 obj_request->pages = NULL;
1817 obj_request->page_count = 0;
1820 if (img_request_child_test(img_request)) {
1821 rbd_assert(img_request->obj_request != NULL);
1822 more = obj_request->which < img_request->obj_request_count - 1;
1824 rbd_assert(img_request->rq != NULL);
1825 more = blk_end_request(img_request->rq, result, xferred);
1831 static void rbd_img_obj_callback(struct rbd_obj_request *obj_request)
1833 struct rbd_img_request *img_request;
1834 u32 which = obj_request->which;
1837 rbd_assert(obj_request_img_data_test(obj_request));
1838 img_request = obj_request->img_request;
1840 dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
1841 rbd_assert(img_request != NULL);
1842 rbd_assert(img_request->obj_request_count > 0);
1843 rbd_assert(which != BAD_WHICH);
1844 rbd_assert(which < img_request->obj_request_count);
1845 rbd_assert(which >= img_request->next_completion);
1847 spin_lock_irq(&img_request->completion_lock);
1848 if (which != img_request->next_completion)
1851 for_each_obj_request_from(img_request, obj_request) {
1853 rbd_assert(which < img_request->obj_request_count);
1855 if (!obj_request_done_test(obj_request))
1857 more = rbd_img_obj_end_request(obj_request);
1861 rbd_assert(more ^ (which == img_request->obj_request_count));
1862 img_request->next_completion = which;
1864 spin_unlock_irq(&img_request->completion_lock);
1867 rbd_img_request_complete(img_request);
1871 * Split up an image request into one or more object requests, each
1872 * to a different object. The "type" parameter indicates whether
1873 * "data_desc" is the pointer to the head of a list of bio
1874 * structures, or the base of a page array. In either case this
1875 * function assumes data_desc describes memory sufficient to hold
1876 * all data described by the image request.
1878 static int rbd_img_request_fill(struct rbd_img_request *img_request,
1879 enum obj_request_type type,
1882 struct rbd_device *rbd_dev = img_request->rbd_dev;
1883 struct rbd_obj_request *obj_request = NULL;
1884 struct rbd_obj_request *next_obj_request;
1885 bool write_request = img_request_write_test(img_request);
1886 struct bio *bio_list;
1887 unsigned int bio_offset = 0;
1888 struct page **pages;
1893 dout("%s: img %p type %d data_desc %p\n", __func__, img_request,
1894 (int)type, data_desc);
1896 opcode = write_request ? CEPH_OSD_OP_WRITE : CEPH_OSD_OP_READ;
1897 img_offset = img_request->offset;
1898 resid = img_request->length;
1899 rbd_assert(resid > 0);
1901 if (type == OBJ_REQUEST_BIO) {
1902 bio_list = data_desc;
1903 rbd_assert(img_offset == bio_list->bi_sector << SECTOR_SHIFT);
1905 rbd_assert(type == OBJ_REQUEST_PAGES);
1910 struct ceph_osd_request *osd_req;
1911 const char *object_name;
1915 object_name = rbd_segment_name(rbd_dev, img_offset);
1918 offset = rbd_segment_offset(rbd_dev, img_offset);
1919 length = rbd_segment_length(rbd_dev, img_offset, resid);
1920 obj_request = rbd_obj_request_create(object_name,
1921 offset, length, type);
1922 kfree(object_name); /* object request has its own copy */
1926 if (type == OBJ_REQUEST_BIO) {
1927 unsigned int clone_size;
1929 rbd_assert(length <= (u64)UINT_MAX);
1930 clone_size = (unsigned int)length;
1931 obj_request->bio_list =
1932 bio_chain_clone_range(&bio_list,
1936 if (!obj_request->bio_list)
1939 unsigned int page_count;
1941 obj_request->pages = pages;
1942 page_count = (u32)calc_pages_for(offset, length);
1943 obj_request->page_count = page_count;
1944 if ((offset + length) & ~PAGE_MASK)
1945 page_count--; /* more on last page */
1946 pages += page_count;
1949 osd_req = rbd_osd_req_create(rbd_dev, write_request,
1953 obj_request->osd_req = osd_req;
1954 obj_request->callback = rbd_img_obj_callback;
1956 osd_req_op_extent_init(osd_req, 0, opcode, offset, length,
1958 if (type == OBJ_REQUEST_BIO)
1959 osd_req_op_extent_osd_data_bio(osd_req, 0,
1960 obj_request->bio_list, length);
1962 osd_req_op_extent_osd_data_pages(osd_req, 0,
1963 obj_request->pages, length,
1964 offset & ~PAGE_MASK, false, false);
1967 rbd_osd_req_format_write(obj_request);
1969 rbd_osd_req_format_read(obj_request);
1971 obj_request->img_offset = img_offset;
1972 rbd_img_obj_request_add(img_request, obj_request);
1974 img_offset += length;
1981 rbd_obj_request_put(obj_request);
1983 for_each_obj_request_safe(img_request, obj_request, next_obj_request)
1984 rbd_obj_request_put(obj_request);
1990 rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
1992 struct rbd_img_request *img_request;
1993 struct rbd_device *rbd_dev;
1997 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
1998 rbd_assert(obj_request_img_data_test(obj_request));
1999 img_request = obj_request->img_request;
2000 rbd_assert(img_request);
2002 rbd_dev = img_request->rbd_dev;
2003 rbd_assert(rbd_dev);
2004 length = (u64)1 << rbd_dev->header.obj_order;
2005 page_count = (u32)calc_pages_for(0, length);
2007 rbd_assert(obj_request->copyup_pages);
2008 ceph_release_page_vector(obj_request->copyup_pages, page_count);
2009 obj_request->copyup_pages = NULL;
2012 * We want the transfer count to reflect the size of the
2013 * original write request. There is no such thing as a
2014 * successful short write, so if the request was successful
2015 * we can just set it to the originally-requested length.
2017 if (!obj_request->result)
2018 obj_request->xferred = obj_request->length;
2020 /* Finish up with the normal image object callback */
2022 rbd_img_obj_callback(obj_request);
2026 rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
2028 struct rbd_obj_request *orig_request;
2029 struct ceph_osd_request *osd_req;
2030 struct ceph_osd_client *osdc;
2031 struct rbd_device *rbd_dev;
2032 struct page **pages;
2037 rbd_assert(img_request_child_test(img_request));
2039 /* First get what we need from the image request */
2041 pages = img_request->copyup_pages;
2042 rbd_assert(pages != NULL);
2043 img_request->copyup_pages = NULL;
2045 orig_request = img_request->obj_request;
2046 rbd_assert(orig_request != NULL);
2047 rbd_assert(orig_request->type == OBJ_REQUEST_BIO);
2048 result = img_request->result;
2049 obj_size = img_request->length;
2050 xferred = img_request->xferred;
2052 rbd_dev = img_request->rbd_dev;
2053 rbd_assert(rbd_dev);
2054 rbd_assert(obj_size == (u64)1 << rbd_dev->header.obj_order);
2056 rbd_img_request_put(img_request);
2061 /* Allocate the new copyup osd request for the original request */
2064 rbd_assert(!orig_request->osd_req);
2065 osd_req = rbd_osd_req_create_copyup(orig_request);
2068 orig_request->osd_req = osd_req;
2069 orig_request->copyup_pages = pages;
2071 /* Initialize the copyup op */
2073 osd_req_op_cls_init(osd_req, 0, CEPH_OSD_OP_CALL, "rbd", "copyup");
2074 osd_req_op_cls_request_data_pages(osd_req, 0, pages, obj_size, 0,
2077 /* Then the original write request op */
2079 osd_req_op_extent_init(osd_req, 1, CEPH_OSD_OP_WRITE,
2080 orig_request->offset,
2081 orig_request->length, 0, 0);
2082 osd_req_op_extent_osd_data_bio(osd_req, 1, orig_request->bio_list,
2083 orig_request->length);
2085 rbd_osd_req_format_write(orig_request);
2087 /* All set, send it off. */
2089 orig_request->callback = rbd_img_obj_copyup_callback;
2090 osdc = &rbd_dev->rbd_client->client->osdc;
2091 result = rbd_obj_request_submit(osdc, orig_request);
2095 /* Record the error code and complete the request */
2097 orig_request->result = result;
2098 orig_request->xferred = 0;
2099 obj_request_done_set(orig_request);
2100 rbd_obj_request_complete(orig_request);
2104 * Read from the parent image the range of data that covers the
2105 * entire target of the given object request. This is used for
2106 * satisfying a layered image write request when the target of an
2107 * object request from the image request does not exist.
2109 * A page array big enough to hold the returned data is allocated
2110 * and supplied to rbd_img_request_fill() as the "data descriptor."
2111 * When the read completes, this page array will be transferred to
2112 * the original object request for the copyup operation.
2114 * If an error occurs, record it as the result of the original
2115 * object request and mark it done so it gets completed.
2117 static int rbd_img_obj_parent_read_full(struct rbd_obj_request *obj_request)
2119 struct rbd_img_request *img_request = NULL;
2120 struct rbd_img_request *parent_request = NULL;
2121 struct rbd_device *rbd_dev;
2124 struct page **pages = NULL;
2128 rbd_assert(obj_request_img_data_test(obj_request));
2129 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
2131 img_request = obj_request->img_request;
2132 rbd_assert(img_request != NULL);
2133 rbd_dev = img_request->rbd_dev;
2134 rbd_assert(rbd_dev->parent != NULL);
2137 * First things first. The original osd request is of no
2138 * use to use any more, we'll need a new one that can hold
2139 * the two ops in a copyup request. We'll get that later,
2140 * but for now we can release the old one.
2142 rbd_osd_req_destroy(obj_request->osd_req);
2143 obj_request->osd_req = NULL;
2146 * Determine the byte range covered by the object in the
2147 * child image to which the original request was to be sent.
2149 img_offset = obj_request->img_offset - obj_request->offset;
2150 length = (u64)1 << rbd_dev->header.obj_order;
2153 * There is no defined parent data beyond the parent
2154 * overlap, so limit what we read at that boundary if
2157 if (img_offset + length > rbd_dev->parent_overlap) {
2158 rbd_assert(img_offset < rbd_dev->parent_overlap);
2159 length = rbd_dev->parent_overlap - img_offset;
2163 * Allocate a page array big enough to receive the data read
2166 page_count = (u32)calc_pages_for(0, length);
2167 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2168 if (IS_ERR(pages)) {
2169 result = PTR_ERR(pages);
2175 parent_request = rbd_img_request_create(rbd_dev->parent,
2178 if (!parent_request)
2180 rbd_obj_request_get(obj_request);
2181 parent_request->obj_request = obj_request;
2183 result = rbd_img_request_fill(parent_request, OBJ_REQUEST_PAGES, pages);
2186 parent_request->copyup_pages = pages;
2188 parent_request->callback = rbd_img_obj_parent_read_full_callback;
2189 result = rbd_img_request_submit(parent_request);
2193 parent_request->copyup_pages = NULL;
2194 parent_request->obj_request = NULL;
2195 rbd_obj_request_put(obj_request);
2198 ceph_release_page_vector(pages, page_count);
2200 rbd_img_request_put(parent_request);
2201 obj_request->result = result;
2202 obj_request->xferred = 0;
2203 obj_request_done_set(obj_request);
2208 static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
2210 struct rbd_obj_request *orig_request;
2213 rbd_assert(!obj_request_img_data_test(obj_request));
2216 * All we need from the object request is the original
2217 * request and the result of the STAT op. Grab those, then
2218 * we're done with the request.
2220 orig_request = obj_request->obj_request;
2221 obj_request->obj_request = NULL;
2222 rbd_assert(orig_request);
2223 rbd_assert(orig_request->img_request);
2225 result = obj_request->result;
2226 obj_request->result = 0;
2228 dout("%s: obj %p for obj %p result %d %llu/%llu\n", __func__,
2229 obj_request, orig_request, result,
2230 obj_request->xferred, obj_request->length);
2231 rbd_obj_request_put(obj_request);
2233 rbd_assert(orig_request);
2234 rbd_assert(orig_request->img_request);
2237 * Our only purpose here is to determine whether the object
2238 * exists, and we don't want to treat the non-existence as
2239 * an error. If something else comes back, transfer the
2240 * error to the original request and complete it now.
2243 obj_request_existence_set(orig_request, true);
2244 } else if (result == -ENOENT) {
2245 obj_request_existence_set(orig_request, false);
2246 } else if (result) {
2247 orig_request->result = result;
2252 * Resubmit the original request now that we have recorded
2253 * whether the target object exists.
2255 orig_request->result = rbd_img_obj_request_submit(orig_request);
2257 if (orig_request->result)
2258 rbd_obj_request_complete(orig_request);
2259 rbd_obj_request_put(orig_request);
2262 static int rbd_img_obj_exists_submit(struct rbd_obj_request *obj_request)
2264 struct rbd_obj_request *stat_request;
2265 struct rbd_device *rbd_dev;
2266 struct ceph_osd_client *osdc;
2267 struct page **pages = NULL;
2273 * The response data for a STAT call consists of:
2280 size = sizeof (__le64) + sizeof (__le32) + sizeof (__le32);
2281 page_count = (u32)calc_pages_for(0, size);
2282 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2284 return PTR_ERR(pages);
2287 stat_request = rbd_obj_request_create(obj_request->object_name, 0, 0,
2292 rbd_obj_request_get(obj_request);
2293 stat_request->obj_request = obj_request;
2294 stat_request->pages = pages;
2295 stat_request->page_count = page_count;
2297 rbd_assert(obj_request->img_request);
2298 rbd_dev = obj_request->img_request->rbd_dev;
2299 stat_request->osd_req = rbd_osd_req_create(rbd_dev, false,
2301 if (!stat_request->osd_req)
2303 stat_request->callback = rbd_img_obj_exists_callback;
2305 osd_req_op_init(stat_request->osd_req, 0, CEPH_OSD_OP_STAT);
2306 osd_req_op_raw_data_in_pages(stat_request->osd_req, 0, pages, size, 0,
2308 rbd_osd_req_format_read(stat_request);
2310 osdc = &rbd_dev->rbd_client->client->osdc;
2311 ret = rbd_obj_request_submit(osdc, stat_request);
2314 rbd_obj_request_put(obj_request);
2319 static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request)
2321 struct rbd_img_request *img_request;
2322 struct rbd_device *rbd_dev;
2325 rbd_assert(obj_request_img_data_test(obj_request));
2327 img_request = obj_request->img_request;
2328 rbd_assert(img_request);
2329 rbd_dev = img_request->rbd_dev;
2332 * Only writes to layered images need special handling.
2333 * Reads and non-layered writes are simple object requests.
2334 * Layered writes that start beyond the end of the overlap
2335 * with the parent have no parent data, so they too are
2336 * simple object requests. Finally, if the target object is
2337 * known to already exist, its parent data has already been
2338 * copied, so a write to the object can also be handled as a
2339 * simple object request.
2341 if (!img_request_write_test(img_request) ||
2342 !img_request_layered_test(img_request) ||
2343 rbd_dev->parent_overlap <= obj_request->img_offset ||
2344 ((known = obj_request_known_test(obj_request)) &&
2345 obj_request_exists_test(obj_request))) {
2347 struct rbd_device *rbd_dev;
2348 struct ceph_osd_client *osdc;
2350 rbd_dev = obj_request->img_request->rbd_dev;
2351 osdc = &rbd_dev->rbd_client->client->osdc;
2353 return rbd_obj_request_submit(osdc, obj_request);
2357 * It's a layered write. The target object might exist but
2358 * we may not know that yet. If we know it doesn't exist,
2359 * start by reading the data for the full target object from
2360 * the parent so we can use it for a copyup to the target.
2363 return rbd_img_obj_parent_read_full(obj_request);
2365 /* We don't know whether the target exists. Go find out. */
2367 return rbd_img_obj_exists_submit(obj_request);
2370 static int rbd_img_request_submit(struct rbd_img_request *img_request)
2372 struct rbd_obj_request *obj_request;
2373 struct rbd_obj_request *next_obj_request;
2375 dout("%s: img %p\n", __func__, img_request);
2376 for_each_obj_request_safe(img_request, obj_request, next_obj_request) {
2379 ret = rbd_img_obj_request_submit(obj_request);
2387 static void rbd_img_parent_read_callback(struct rbd_img_request *img_request)
2389 struct rbd_obj_request *obj_request;
2390 struct rbd_device *rbd_dev;
2393 rbd_assert(img_request_child_test(img_request));
2395 obj_request = img_request->obj_request;
2396 rbd_assert(obj_request);
2397 rbd_assert(obj_request->img_request);
2399 obj_request->result = img_request->result;
2400 if (obj_request->result)
2404 * We need to zero anything beyond the parent overlap
2405 * boundary. Since rbd_img_obj_request_read_callback()
2406 * will zero anything beyond the end of a short read, an
2407 * easy way to do this is to pretend the data from the
2408 * parent came up short--ending at the overlap boundary.
2410 rbd_assert(obj_request->img_offset < U64_MAX - obj_request->length);
2411 obj_end = obj_request->img_offset + obj_request->length;
2412 rbd_dev = obj_request->img_request->rbd_dev;
2413 if (obj_end > rbd_dev->parent_overlap) {
2416 if (obj_request->img_offset < rbd_dev->parent_overlap)
2417 xferred = rbd_dev->parent_overlap -
2418 obj_request->img_offset;
2420 obj_request->xferred = min(img_request->xferred, xferred);
2422 obj_request->xferred = img_request->xferred;
2425 rbd_img_obj_request_read_callback(obj_request);
2426 rbd_obj_request_complete(obj_request);
2429 static void rbd_img_parent_read(struct rbd_obj_request *obj_request)
2431 struct rbd_device *rbd_dev;
2432 struct rbd_img_request *img_request;
2435 rbd_assert(obj_request_img_data_test(obj_request));
2436 rbd_assert(obj_request->img_request != NULL);
2437 rbd_assert(obj_request->result == (s32) -ENOENT);
2438 rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
2440 rbd_dev = obj_request->img_request->rbd_dev;
2441 rbd_assert(rbd_dev->parent != NULL);
2442 /* rbd_read_finish(obj_request, obj_request->length); */
2443 img_request = rbd_img_request_create(rbd_dev->parent,
2444 obj_request->img_offset,
2445 obj_request->length,
2451 rbd_obj_request_get(obj_request);
2452 img_request->obj_request = obj_request;
2454 result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
2455 obj_request->bio_list);
2459 img_request->callback = rbd_img_parent_read_callback;
2460 result = rbd_img_request_submit(img_request);
2467 rbd_img_request_put(img_request);
2468 obj_request->result = result;
2469 obj_request->xferred = 0;
2470 obj_request_done_set(obj_request);
2473 static int rbd_obj_notify_ack(struct rbd_device *rbd_dev,
2474 u64 ver, u64 notify_id)
2476 struct rbd_obj_request *obj_request;
2477 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2480 obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
2481 OBJ_REQUEST_NODATA);
2486 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2487 if (!obj_request->osd_req)
2489 obj_request->callback = rbd_obj_request_put;
2491 osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_NOTIFY_ACK,
2493 rbd_osd_req_format_read(obj_request);
2495 ret = rbd_obj_request_submit(osdc, obj_request);
2498 rbd_obj_request_put(obj_request);
2503 static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
2505 struct rbd_device *rbd_dev = (struct rbd_device *)data;
2511 dout("%s: \"%s\" notify_id %llu opcode %u\n", __func__,
2512 rbd_dev->header_name, (unsigned long long) notify_id,
2513 (unsigned int) opcode);
2514 (void)rbd_dev_refresh(rbd_dev, &hver);
2516 rbd_obj_notify_ack(rbd_dev, hver, notify_id);
2520 * Request sync osd watch/unwatch. The value of "start" determines
2521 * whether a watch request is being initiated or torn down.
2523 static int rbd_dev_header_watch_sync(struct rbd_device *rbd_dev, int start)
2525 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2526 struct rbd_obj_request *obj_request;
2529 rbd_assert(start ^ !!rbd_dev->watch_event);
2530 rbd_assert(start ^ !!rbd_dev->watch_request);
2533 ret = ceph_osdc_create_event(osdc, rbd_watch_cb, rbd_dev,
2534 &rbd_dev->watch_event);
2537 rbd_assert(rbd_dev->watch_event != NULL);
2541 obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
2542 OBJ_REQUEST_NODATA);
2546 obj_request->osd_req = rbd_osd_req_create(rbd_dev, true, obj_request);
2547 if (!obj_request->osd_req)
2551 ceph_osdc_set_request_linger(osdc, obj_request->osd_req);
2553 ceph_osdc_unregister_linger_request(osdc,
2554 rbd_dev->watch_request->osd_req);
2556 osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH,
2557 rbd_dev->watch_event->cookie,
2558 rbd_dev->header.obj_version, start);
2559 rbd_osd_req_format_write(obj_request);
2561 ret = rbd_obj_request_submit(osdc, obj_request);
2564 ret = rbd_obj_request_wait(obj_request);
2567 ret = obj_request->result;
2572 * A watch request is set to linger, so the underlying osd
2573 * request won't go away until we unregister it. We retain
2574 * a pointer to the object request during that time (in
2575 * rbd_dev->watch_request), so we'll keep a reference to
2576 * it. We'll drop that reference (below) after we've
2580 rbd_dev->watch_request = obj_request;
2585 /* We have successfully torn down the watch request */
2587 rbd_obj_request_put(rbd_dev->watch_request);
2588 rbd_dev->watch_request = NULL;
2590 /* Cancel the event if we're tearing down, or on error */
2591 ceph_osdc_cancel_event(rbd_dev->watch_event);
2592 rbd_dev->watch_event = NULL;
2594 rbd_obj_request_put(obj_request);
2600 * Synchronous osd object method call. Returns the number of bytes
2601 * returned in the outbound buffer, or a negative error code.
2603 static int rbd_obj_method_sync(struct rbd_device *rbd_dev,
2604 const char *object_name,
2605 const char *class_name,
2606 const char *method_name,
2607 const void *outbound,
2608 size_t outbound_size,
2610 size_t inbound_size,
2613 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2614 struct rbd_obj_request *obj_request;
2615 struct page **pages;
2620 * Method calls are ultimately read operations. The result
2621 * should placed into the inbound buffer provided. They
2622 * also supply outbound data--parameters for the object
2623 * method. Currently if this is present it will be a
2626 page_count = (u32)calc_pages_for(0, inbound_size);
2627 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2629 return PTR_ERR(pages);
2632 obj_request = rbd_obj_request_create(object_name, 0, inbound_size,
2637 obj_request->pages = pages;
2638 obj_request->page_count = page_count;
2640 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2641 if (!obj_request->osd_req)
2644 osd_req_op_cls_init(obj_request->osd_req, 0, CEPH_OSD_OP_CALL,
2645 class_name, method_name);
2646 if (outbound_size) {
2647 struct ceph_pagelist *pagelist;
2649 pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
2653 ceph_pagelist_init(pagelist);
2654 ceph_pagelist_append(pagelist, outbound, outbound_size);
2655 osd_req_op_cls_request_data_pagelist(obj_request->osd_req, 0,
2658 osd_req_op_cls_response_data_pages(obj_request->osd_req, 0,
2659 obj_request->pages, inbound_size,
2661 rbd_osd_req_format_read(obj_request);
2663 ret = rbd_obj_request_submit(osdc, obj_request);
2666 ret = rbd_obj_request_wait(obj_request);
2670 ret = obj_request->result;
2674 rbd_assert(obj_request->xferred < (u64)INT_MAX);
2675 ret = (int)obj_request->xferred;
2676 ceph_copy_from_page_vector(pages, inbound, 0, obj_request->xferred);
2678 *version = obj_request->version;
2681 rbd_obj_request_put(obj_request);
2683 ceph_release_page_vector(pages, page_count);
2688 static void rbd_request_fn(struct request_queue *q)
2689 __releases(q->queue_lock) __acquires(q->queue_lock)
2691 struct rbd_device *rbd_dev = q->queuedata;
2692 bool read_only = rbd_dev->mapping.read_only;
2696 while ((rq = blk_fetch_request(q))) {
2697 bool write_request = rq_data_dir(rq) == WRITE;
2698 struct rbd_img_request *img_request;
2702 /* Ignore any non-FS requests that filter through. */
2704 if (rq->cmd_type != REQ_TYPE_FS) {
2705 dout("%s: non-fs request type %d\n", __func__,
2706 (int) rq->cmd_type);
2707 __blk_end_request_all(rq, 0);
2711 /* Ignore/skip any zero-length requests */
2713 offset = (u64) blk_rq_pos(rq) << SECTOR_SHIFT;
2714 length = (u64) blk_rq_bytes(rq);
2717 dout("%s: zero-length request\n", __func__);
2718 __blk_end_request_all(rq, 0);
2722 spin_unlock_irq(q->queue_lock);
2724 /* Disallow writes to a read-only device */
2726 if (write_request) {
2730 rbd_assert(rbd_dev->spec->snap_id == CEPH_NOSNAP);
2734 * Quit early if the mapped snapshot no longer
2735 * exists. It's still possible the snapshot will
2736 * have disappeared by the time our request arrives
2737 * at the osd, but there's no sense in sending it if
2740 if (!test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags)) {
2741 dout("request for non-existent snapshot");
2742 rbd_assert(rbd_dev->spec->snap_id != CEPH_NOSNAP);
2748 if (offset && length > U64_MAX - offset + 1) {
2749 rbd_warn(rbd_dev, "bad request range (%llu~%llu)\n",
2751 goto end_request; /* Shouldn't happen */
2755 img_request = rbd_img_request_create(rbd_dev, offset, length,
2756 write_request, false);
2760 img_request->rq = rq;
2762 result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
2765 result = rbd_img_request_submit(img_request);
2767 rbd_img_request_put(img_request);
2769 spin_lock_irq(q->queue_lock);
2771 rbd_warn(rbd_dev, "%s %llx at %llx result %d\n",
2772 write_request ? "write" : "read",
2773 length, offset, result);
2775 __blk_end_request_all(rq, result);
2781 * a queue callback. Makes sure that we don't create a bio that spans across
2782 * multiple osd objects. One exception would be with a single page bios,
2783 * which we handle later at bio_chain_clone_range()
2785 static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd,
2786 struct bio_vec *bvec)
2788 struct rbd_device *rbd_dev = q->queuedata;
2789 sector_t sector_offset;
2790 sector_t sectors_per_obj;
2791 sector_t obj_sector_offset;
2795 * Find how far into its rbd object the partition-relative
2796 * bio start sector is to offset relative to the enclosing
2799 sector_offset = get_start_sect(bmd->bi_bdev) + bmd->bi_sector;
2800 sectors_per_obj = 1 << (rbd_dev->header.obj_order - SECTOR_SHIFT);
2801 obj_sector_offset = sector_offset & (sectors_per_obj - 1);
2804 * Compute the number of bytes from that offset to the end
2805 * of the object. Account for what's already used by the bio.
2807 ret = (int) (sectors_per_obj - obj_sector_offset) << SECTOR_SHIFT;
2808 if (ret > bmd->bi_size)
2809 ret -= bmd->bi_size;
2814 * Don't send back more than was asked for. And if the bio
2815 * was empty, let the whole thing through because: "Note
2816 * that a block device *must* allow a single page to be
2817 * added to an empty bio."
2819 rbd_assert(bvec->bv_len <= PAGE_SIZE);
2820 if (ret > (int) bvec->bv_len || !bmd->bi_size)
2821 ret = (int) bvec->bv_len;
2826 static void rbd_free_disk(struct rbd_device *rbd_dev)
2828 struct gendisk *disk = rbd_dev->disk;
2833 rbd_dev->disk = NULL;
2834 if (disk->flags & GENHD_FL_UP) {
2837 blk_cleanup_queue(disk->queue);
2842 static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
2843 const char *object_name,
2844 u64 offset, u64 length,
2845 void *buf, u64 *version)
2848 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2849 struct rbd_obj_request *obj_request;
2850 struct page **pages = NULL;
2855 page_count = (u32) calc_pages_for(offset, length);
2856 pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
2858 ret = PTR_ERR(pages);
2861 obj_request = rbd_obj_request_create(object_name, offset, length,
2866 obj_request->pages = pages;
2867 obj_request->page_count = page_count;
2869 obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
2870 if (!obj_request->osd_req)
2873 osd_req_op_extent_init(obj_request->osd_req, 0, CEPH_OSD_OP_READ,
2874 offset, length, 0, 0);
2875 osd_req_op_extent_osd_data_pages(obj_request->osd_req, 0,
2877 obj_request->length,
2878 obj_request->offset & ~PAGE_MASK,
2880 rbd_osd_req_format_read(obj_request);
2882 ret = rbd_obj_request_submit(osdc, obj_request);
2885 ret = rbd_obj_request_wait(obj_request);
2889 ret = obj_request->result;
2893 rbd_assert(obj_request->xferred <= (u64) SIZE_MAX);
2894 size = (size_t) obj_request->xferred;
2895 ceph_copy_from_page_vector(pages, buf, 0, size);
2896 rbd_assert(size <= (size_t) INT_MAX);
2899 *version = obj_request->version;
2902 rbd_obj_request_put(obj_request);
2904 ceph_release_page_vector(pages, page_count);
2910 * Read the complete header for the given rbd device.
2912 * Returns a pointer to a dynamically-allocated buffer containing
2913 * the complete and validated header. Caller can pass the address
2914 * of a variable that will be filled in with the version of the
2915 * header object at the time it was read.
2917 * Returns a pointer-coded errno if a failure occurs.
2919 static struct rbd_image_header_ondisk *
2920 rbd_dev_v1_header_read(struct rbd_device *rbd_dev, u64 *version)
2922 struct rbd_image_header_ondisk *ondisk = NULL;
2929 * The complete header will include an array of its 64-bit
2930 * snapshot ids, followed by the names of those snapshots as
2931 * a contiguous block of NUL-terminated strings. Note that
2932 * the number of snapshots could change by the time we read
2933 * it in, in which case we re-read it.
2940 size = sizeof (*ondisk);
2941 size += snap_count * sizeof (struct rbd_image_snap_ondisk);
2943 ondisk = kmalloc(size, GFP_KERNEL);
2945 return ERR_PTR(-ENOMEM);
2947 ret = rbd_obj_read_sync(rbd_dev, rbd_dev->header_name,
2948 0, size, ondisk, version);
2951 if ((size_t)ret < size) {
2953 rbd_warn(rbd_dev, "short header read (want %zd got %d)",
2957 if (!rbd_dev_ondisk_valid(ondisk)) {
2959 rbd_warn(rbd_dev, "invalid header");
2963 names_size = le64_to_cpu(ondisk->snap_names_len);
2964 want_count = snap_count;
2965 snap_count = le32_to_cpu(ondisk->snap_count);
2966 } while (snap_count != want_count);
2973 return ERR_PTR(ret);
2977 * reload the ondisk the header
2979 static int rbd_read_header(struct rbd_device *rbd_dev,
2980 struct rbd_image_header *header)
2982 struct rbd_image_header_ondisk *ondisk;
2986 ondisk = rbd_dev_v1_header_read(rbd_dev, &ver);
2988 return PTR_ERR(ondisk);
2989 ret = rbd_header_from_disk(header, ondisk);
2991 header->obj_version = ver;
2997 static void rbd_remove_all_snaps(struct rbd_device *rbd_dev)
2999 struct rbd_snap *snap;
3000 struct rbd_snap *next;
3002 list_for_each_entry_safe(snap, next, &rbd_dev->snaps, node) {
3003 list_del(&snap->node);
3004 rbd_snap_destroy(snap);
3008 static void rbd_update_mapping_size(struct rbd_device *rbd_dev)
3010 if (rbd_dev->spec->snap_id != CEPH_NOSNAP)
3013 if (rbd_dev->mapping.size != rbd_dev->header.image_size) {
3016 rbd_dev->mapping.size = rbd_dev->header.image_size;
3017 size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE;
3018 dout("setting size to %llu sectors", (unsigned long long)size);
3019 set_capacity(rbd_dev->disk, size);
3024 * only read the first part of the ondisk header, without the snaps info
3026 static int rbd_dev_v1_refresh(struct rbd_device *rbd_dev, u64 *hver)
3029 struct rbd_image_header h;
3031 ret = rbd_read_header(rbd_dev, &h);
3035 down_write(&rbd_dev->header_rwsem);
3037 /* Update image size, and check for resize of mapped image */
3038 rbd_dev->header.image_size = h.image_size;
3039 rbd_update_mapping_size(rbd_dev);
3041 /* rbd_dev->header.object_prefix shouldn't change */
3042 kfree(rbd_dev->header.snap_sizes);
3043 kfree(rbd_dev->header.snap_names);
3044 /* osd requests may still refer to snapc */
3045 ceph_put_snap_context(rbd_dev->header.snapc);
3048 *hver = h.obj_version;
3049 rbd_dev->header.obj_version = h.obj_version;
3050 rbd_dev->header.image_size = h.image_size;
3051 rbd_dev->header.snapc = h.snapc;
3052 rbd_dev->header.snap_names = h.snap_names;
3053 rbd_dev->header.snap_sizes = h.snap_sizes;
3054 /* Free the extra copy of the object prefix */
3055 if (strcmp(rbd_dev->header.object_prefix, h.object_prefix))
3056 rbd_warn(rbd_dev, "object prefix changed (ignoring)");
3057 kfree(h.object_prefix);
3059 ret = rbd_dev_snaps_update(rbd_dev);
3061 up_write(&rbd_dev->header_rwsem);
3066 static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver)
3071 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
3072 image_size = rbd_dev->header.image_size;
3073 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
3074 if (rbd_dev->image_format == 1)
3075 ret = rbd_dev_v1_refresh(rbd_dev, hver);
3077 ret = rbd_dev_v2_refresh(rbd_dev, hver);
3078 mutex_unlock(&ctl_mutex);
3080 rbd_warn(rbd_dev, "got notification but failed to "
3081 " update snaps: %d\n", ret);
3082 if (image_size != rbd_dev->header.image_size)
3083 revalidate_disk(rbd_dev->disk);
3088 static int rbd_init_disk(struct rbd_device *rbd_dev)
3090 struct gendisk *disk;
3091 struct request_queue *q;
3094 /* create gendisk info */
3095 disk = alloc_disk(RBD_MINORS_PER_MAJOR);
3099 snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
3101 disk->major = rbd_dev->major;
3102 disk->first_minor = 0;
3103 disk->fops = &rbd_bd_ops;
3104 disk->private_data = rbd_dev;
3106 q = blk_init_queue(rbd_request_fn, &rbd_dev->lock);
3110 /* We use the default size, but let's be explicit about it. */
3111 blk_queue_physical_block_size(q, SECTOR_SIZE);
3113 /* set io sizes to object size */
3114 segment_size = rbd_obj_bytes(&rbd_dev->header);
3115 blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
3116 blk_queue_max_segment_size(q, segment_size);
3117 blk_queue_io_min(q, segment_size);
3118 blk_queue_io_opt(q, segment_size);
3120 blk_queue_merge_bvec(q, rbd_merge_bvec);
3123 q->queuedata = rbd_dev;
3125 rbd_dev->disk = disk;
3138 static struct rbd_device *dev_to_rbd_dev(struct device *dev)
3140 return container_of(dev, struct rbd_device, dev);
3143 static ssize_t rbd_size_show(struct device *dev,
3144 struct device_attribute *attr, char *buf)
3146 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3148 return sprintf(buf, "%llu\n",
3149 (unsigned long long)rbd_dev->mapping.size);
3153 * Note this shows the features for whatever's mapped, which is not
3154 * necessarily the base image.
3156 static ssize_t rbd_features_show(struct device *dev,
3157 struct device_attribute *attr, char *buf)
3159 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3161 return sprintf(buf, "0x%016llx\n",
3162 (unsigned long long)rbd_dev->mapping.features);
3165 static ssize_t rbd_major_show(struct device *dev,
3166 struct device_attribute *attr, char *buf)
3168 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3171 return sprintf(buf, "%d\n", rbd_dev->major);
3173 return sprintf(buf, "(none)\n");
3177 static ssize_t rbd_client_id_show(struct device *dev,
3178 struct device_attribute *attr, char *buf)
3180 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3182 return sprintf(buf, "client%lld\n",
3183 ceph_client_id(rbd_dev->rbd_client->client));
3186 static ssize_t rbd_pool_show(struct device *dev,
3187 struct device_attribute *attr, char *buf)
3189 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3191 return sprintf(buf, "%s\n", rbd_dev->spec->pool_name);
3194 static ssize_t rbd_pool_id_show(struct device *dev,
3195 struct device_attribute *attr, char *buf)
3197 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3199 return sprintf(buf, "%llu\n",
3200 (unsigned long long) rbd_dev->spec->pool_id);
3203 static ssize_t rbd_name_show(struct device *dev,
3204 struct device_attribute *attr, char *buf)
3206 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3208 if (rbd_dev->spec->image_name)
3209 return sprintf(buf, "%s\n", rbd_dev->spec->image_name);
3211 return sprintf(buf, "(unknown)\n");
3214 static ssize_t rbd_image_id_show(struct device *dev,
3215 struct device_attribute *attr, char *buf)
3217 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3219 return sprintf(buf, "%s\n", rbd_dev->spec->image_id);
3223 * Shows the name of the currently-mapped snapshot (or
3224 * RBD_SNAP_HEAD_NAME for the base image).
3226 static ssize_t rbd_snap_show(struct device *dev,
3227 struct device_attribute *attr,
3230 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3232 return sprintf(buf, "%s\n", rbd_dev->spec->snap_name);
3236 * For an rbd v2 image, shows the pool id, image id, and snapshot id
3237 * for the parent image. If there is no parent, simply shows
3238 * "(no parent image)".
3240 static ssize_t rbd_parent_show(struct device *dev,
3241 struct device_attribute *attr,
3244 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3245 struct rbd_spec *spec = rbd_dev->parent_spec;
3250 return sprintf(buf, "(no parent image)\n");
3252 count = sprintf(bufp, "pool_id %llu\npool_name %s\n",
3253 (unsigned long long) spec->pool_id, spec->pool_name);
3258 count = sprintf(bufp, "image_id %s\nimage_name %s\n", spec->image_id,
3259 spec->image_name ? spec->image_name : "(unknown)");
3264 count = sprintf(bufp, "snap_id %llu\nsnap_name %s\n",
3265 (unsigned long long) spec->snap_id, spec->snap_name);
3270 count = sprintf(bufp, "overlap %llu\n", rbd_dev->parent_overlap);
3275 return (ssize_t) (bufp - buf);
3278 static ssize_t rbd_image_refresh(struct device *dev,
3279 struct device_attribute *attr,
3283 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
3286 ret = rbd_dev_refresh(rbd_dev, NULL);
3288 return ret < 0 ? ret : size;
3291 static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
3292 static DEVICE_ATTR(features, S_IRUGO, rbd_features_show, NULL);
3293 static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
3294 static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
3295 static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
3296 static DEVICE_ATTR(pool_id, S_IRUGO, rbd_pool_id_show, NULL);
3297 static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
3298 static DEVICE_ATTR(image_id, S_IRUGO, rbd_image_id_show, NULL);
3299 static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
3300 static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
3301 static DEVICE_ATTR(parent, S_IRUGO, rbd_parent_show, NULL);
3303 static struct attribute *rbd_attrs[] = {
3304 &dev_attr_size.attr,
3305 &dev_attr_features.attr,
3306 &dev_attr_major.attr,
3307 &dev_attr_client_id.attr,
3308 &dev_attr_pool.attr,
3309 &dev_attr_pool_id.attr,
3310 &dev_attr_name.attr,
3311 &dev_attr_image_id.attr,
3312 &dev_attr_current_snap.attr,
3313 &dev_attr_parent.attr,
3314 &dev_attr_refresh.attr,
3318 static struct attribute_group rbd_attr_group = {
3322 static const struct attribute_group *rbd_attr_groups[] = {
3327 static void rbd_sysfs_dev_release(struct device *dev)
3331 static struct device_type rbd_device_type = {
3333 .groups = rbd_attr_groups,
3334 .release = rbd_sysfs_dev_release,
3337 static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec)
3339 kref_get(&spec->kref);
3344 static void rbd_spec_free(struct kref *kref);
3345 static void rbd_spec_put(struct rbd_spec *spec)
3348 kref_put(&spec->kref, rbd_spec_free);
3351 static struct rbd_spec *rbd_spec_alloc(void)
3353 struct rbd_spec *spec;
3355 spec = kzalloc(sizeof (*spec), GFP_KERNEL);
3358 kref_init(&spec->kref);
3363 static void rbd_spec_free(struct kref *kref)
3365 struct rbd_spec *spec = container_of(kref, struct rbd_spec, kref);
3367 kfree(spec->pool_name);
3368 kfree(spec->image_id);
3369 kfree(spec->image_name);
3370 kfree(spec->snap_name);
3374 static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc,
3375 struct rbd_spec *spec)
3377 struct rbd_device *rbd_dev;
3379 rbd_dev = kzalloc(sizeof (*rbd_dev), GFP_KERNEL);
3383 spin_lock_init(&rbd_dev->lock);
3385 INIT_LIST_HEAD(&rbd_dev->node);
3386 INIT_LIST_HEAD(&rbd_dev->snaps);
3387 init_rwsem(&rbd_dev->header_rwsem);
3389 rbd_dev->spec = spec;
3390 rbd_dev->rbd_client = rbdc;
3392 /* Initialize the layout used for all rbd requests */
3394 rbd_dev->layout.fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
3395 rbd_dev->layout.fl_stripe_count = cpu_to_le32(1);
3396 rbd_dev->layout.fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
3397 rbd_dev->layout.fl_pg_pool = cpu_to_le32((u32) spec->pool_id);
3402 static void rbd_dev_destroy(struct rbd_device *rbd_dev)
3404 rbd_put_client(rbd_dev->rbd_client);
3405 rbd_spec_put(rbd_dev->spec);
3409 static void rbd_snap_destroy(struct rbd_snap *snap)
3415 static struct rbd_snap *rbd_snap_create(struct rbd_device *rbd_dev,
3416 const char *snap_name,
3417 u64 snap_id, u64 snap_size,
3420 struct rbd_snap *snap;
3422 snap = kzalloc(sizeof (*snap), GFP_KERNEL);
3424 return ERR_PTR(-ENOMEM);
3426 snap->name = snap_name;
3428 snap->size = snap_size;
3429 snap->features = snap_features;
3435 * Returns a dynamically-allocated snapshot name if successful, or a
3436 * pointer-coded error otherwise.
3438 static const char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
3439 u64 *snap_size, u64 *snap_features)
3441 const char *snap_name;
3444 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3446 /* Skip over names until we find the one we are looking for */
3448 snap_name = rbd_dev->header.snap_names;
3449 for (i = 0; i < which; i++)
3450 snap_name += strlen(snap_name) + 1;
3452 snap_name = kstrdup(snap_name, GFP_KERNEL);
3454 return ERR_PTR(-ENOMEM);
3456 *snap_size = rbd_dev->header.snap_sizes[which];
3457 *snap_features = 0; /* No features for v1 */
3463 * Get the size and object order for an image snapshot, or if
3464 * snap_id is CEPH_NOSNAP, gets this information for the base
3467 static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
3468 u8 *order, u64 *snap_size)
3470 __le64 snapid = cpu_to_le64(snap_id);
3475 } __attribute__ ((packed)) size_buf = { 0 };
3477 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3479 &snapid, sizeof (snapid),
3480 &size_buf, sizeof (size_buf), NULL);
3481 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3484 if (ret < sizeof (size_buf))
3488 *order = size_buf.order;
3489 *snap_size = le64_to_cpu(size_buf.size);
3491 dout(" snap_id 0x%016llx order = %u, snap_size = %llu\n",
3492 (unsigned long long)snap_id, (unsigned int)*order,
3493 (unsigned long long)*snap_size);
3498 static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
3500 return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
3501 &rbd_dev->header.obj_order,
3502 &rbd_dev->header.image_size);
3505 static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
3511 reply_buf = kzalloc(RBD_OBJ_PREFIX_LEN_MAX, GFP_KERNEL);
3515 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3516 "rbd", "get_object_prefix", NULL, 0,
3517 reply_buf, RBD_OBJ_PREFIX_LEN_MAX, NULL);
3518 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3523 rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
3524 p + ret, NULL, GFP_NOIO);
3527 if (IS_ERR(rbd_dev->header.object_prefix)) {
3528 ret = PTR_ERR(rbd_dev->header.object_prefix);
3529 rbd_dev->header.object_prefix = NULL;
3531 dout(" object_prefix = %s\n", rbd_dev->header.object_prefix);
3539 static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
3542 __le64 snapid = cpu_to_le64(snap_id);
3546 } __attribute__ ((packed)) features_buf = { 0 };
3550 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3551 "rbd", "get_features",
3552 &snapid, sizeof (snapid),
3553 &features_buf, sizeof (features_buf), NULL);
3554 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3557 if (ret < sizeof (features_buf))
3560 incompat = le64_to_cpu(features_buf.incompat);
3561 if (incompat & ~RBD_FEATURES_SUPPORTED)
3564 *snap_features = le64_to_cpu(features_buf.features);
3566 dout(" snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
3567 (unsigned long long)snap_id,
3568 (unsigned long long)*snap_features,
3569 (unsigned long long)le64_to_cpu(features_buf.incompat));
3574 static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
3576 return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
3577 &rbd_dev->header.features);
3580 static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
3582 struct rbd_spec *parent_spec;
3584 void *reply_buf = NULL;
3592 parent_spec = rbd_spec_alloc();
3596 size = sizeof (__le64) + /* pool_id */
3597 sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX + /* image_id */
3598 sizeof (__le64) + /* snap_id */
3599 sizeof (__le64); /* overlap */
3600 reply_buf = kmalloc(size, GFP_KERNEL);
3606 snapid = cpu_to_le64(CEPH_NOSNAP);
3607 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3608 "rbd", "get_parent",
3609 &snapid, sizeof (snapid),
3610 reply_buf, size, NULL);
3611 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3616 end = reply_buf + ret;
3618 ceph_decode_64_safe(&p, end, parent_spec->pool_id, out_err);
3619 if (parent_spec->pool_id == CEPH_NOPOOL)
3620 goto out; /* No parent? No problem. */
3622 /* The ceph file layout needs to fit pool id in 32 bits */
3625 if (parent_spec->pool_id > (u64)U32_MAX) {
3626 rbd_warn(NULL, "parent pool id too large (%llu > %u)\n",
3627 (unsigned long long)parent_spec->pool_id, U32_MAX);
3631 image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
3632 if (IS_ERR(image_id)) {
3633 ret = PTR_ERR(image_id);
3636 parent_spec->image_id = image_id;
3637 ceph_decode_64_safe(&p, end, parent_spec->snap_id, out_err);
3638 ceph_decode_64_safe(&p, end, overlap, out_err);
3640 rbd_dev->parent_overlap = overlap;
3641 rbd_dev->parent_spec = parent_spec;
3642 parent_spec = NULL; /* rbd_dev now owns this */
3647 rbd_spec_put(parent_spec);
3652 static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev)
3656 __le64 stripe_count;
3657 } __attribute__ ((packed)) striping_info_buf = { 0 };
3658 size_t size = sizeof (striping_info_buf);
3665 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3666 "rbd", "get_stripe_unit_count", NULL, 0,
3667 (char *)&striping_info_buf, size, NULL);
3668 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3675 * We don't actually support the "fancy striping" feature
3676 * (STRIPINGV2) yet, but if the striping sizes are the
3677 * defaults the behavior is the same as before. So find
3678 * out, and only fail if the image has non-default values.
3681 obj_size = (u64)1 << rbd_dev->header.obj_order;
3682 p = &striping_info_buf;
3683 stripe_unit = ceph_decode_64(&p);
3684 if (stripe_unit != obj_size) {
3685 rbd_warn(rbd_dev, "unsupported stripe unit "
3686 "(got %llu want %llu)",
3687 stripe_unit, obj_size);
3690 stripe_count = ceph_decode_64(&p);
3691 if (stripe_count != 1) {
3692 rbd_warn(rbd_dev, "unsupported stripe count "
3693 "(got %llu want 1)", stripe_count);
3696 rbd_dev->header.stripe_unit = stripe_unit;
3697 rbd_dev->header.stripe_count = stripe_count;
3702 static char *rbd_dev_image_name(struct rbd_device *rbd_dev)
3704 size_t image_id_size;
3709 void *reply_buf = NULL;
3711 char *image_name = NULL;
3714 rbd_assert(!rbd_dev->spec->image_name);
3716 len = strlen(rbd_dev->spec->image_id);
3717 image_id_size = sizeof (__le32) + len;
3718 image_id = kmalloc(image_id_size, GFP_KERNEL);
3723 end = image_id + image_id_size;
3724 ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len);
3726 size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX;
3727 reply_buf = kmalloc(size, GFP_KERNEL);
3731 ret = rbd_obj_method_sync(rbd_dev, RBD_DIRECTORY,
3732 "rbd", "dir_get_name",
3733 image_id, image_id_size,
3734 reply_buf, size, NULL);
3738 end = reply_buf + ret;
3740 image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL);
3741 if (IS_ERR(image_name))
3744 dout("%s: name is %s len is %zd\n", __func__, image_name, len);
3753 * When an rbd image has a parent image, it is identified by the
3754 * pool, image, and snapshot ids (not names). This function fills
3755 * in the names for those ids. (It's OK if we can't figure out the
3756 * name for an image id, but the pool and snapshot ids should always
3757 * exist and have names.) All names in an rbd spec are dynamically
3760 * When an image being mapped (not a parent) is probed, we have the
3761 * pool name and pool id, image name and image id, and the snapshot
3762 * name. The only thing we're missing is the snapshot id.
3764 * The set of snapshots for an image is not known until they have
3765 * been read by rbd_dev_snaps_update(), so we can't completely fill
3766 * in this information until after that has been called.
3768 static int rbd_dev_spec_update(struct rbd_device *rbd_dev)
3770 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3771 struct rbd_spec *spec = rbd_dev->spec;
3772 const char *pool_name;
3773 const char *image_name;
3774 const char *snap_name;
3778 * An image being mapped will have the pool name (etc.), but
3779 * we need to look up the snapshot id.
3781 if (spec->pool_name) {
3782 if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
3783 struct rbd_snap *snap;
3785 snap = snap_by_name(rbd_dev, spec->snap_name);
3788 spec->snap_id = snap->id;
3790 spec->snap_id = CEPH_NOSNAP;
3796 /* Get the pool name; we have to make our own copy of this */
3798 pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id);
3800 rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id);
3803 pool_name = kstrdup(pool_name, GFP_KERNEL);
3807 /* Fetch the image name; tolerate failure here */
3809 image_name = rbd_dev_image_name(rbd_dev);
3811 rbd_warn(rbd_dev, "unable to get image name");
3813 /* Look up the snapshot name, and make a copy */
3815 snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
3817 rbd_warn(rbd_dev, "no snapshot with id %llu", spec->snap_id);
3821 snap_name = kstrdup(snap_name, GFP_KERNEL);
3827 spec->pool_name = pool_name;
3828 spec->image_name = image_name;
3829 spec->snap_name = snap_name;
3839 static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev, u64 *ver)
3848 struct ceph_snap_context *snapc;
3852 * We'll need room for the seq value (maximum snapshot id),
3853 * snapshot count, and array of that many snapshot ids.
3854 * For now we have a fixed upper limit on the number we're
3855 * prepared to receive.
3857 size = sizeof (__le64) + sizeof (__le32) +
3858 RBD_MAX_SNAP_COUNT * sizeof (__le64);
3859 reply_buf = kzalloc(size, GFP_KERNEL);
3863 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3864 "rbd", "get_snapcontext", NULL, 0,
3865 reply_buf, size, ver);
3866 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3871 end = reply_buf + ret;
3873 ceph_decode_64_safe(&p, end, seq, out);
3874 ceph_decode_32_safe(&p, end, snap_count, out);
3877 * Make sure the reported number of snapshot ids wouldn't go
3878 * beyond the end of our buffer. But before checking that,
3879 * make sure the computed size of the snapshot context we
3880 * allocate is representable in a size_t.
3882 if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
3887 if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
3891 snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
3897 for (i = 0; i < snap_count; i++)
3898 snapc->snaps[i] = ceph_decode_64(&p);
3900 rbd_dev->header.snapc = snapc;
3902 dout(" snap context seq = %llu, snap_count = %u\n",
3903 (unsigned long long)seq, (unsigned int)snap_count);
3910 static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
3920 size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
3921 reply_buf = kmalloc(size, GFP_KERNEL);
3923 return ERR_PTR(-ENOMEM);
3925 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3926 snap_id = cpu_to_le64(rbd_dev->header.snapc->snaps[which]);
3927 ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
3928 "rbd", "get_snapshot_name",
3929 &snap_id, sizeof (snap_id),
3930 reply_buf, size, NULL);
3931 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
3933 snap_name = ERR_PTR(ret);
3938 end = reply_buf + ret;
3939 snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
3940 if (IS_ERR(snap_name))
3943 dout(" snap_id 0x%016llx snap_name = %s\n",
3944 (unsigned long long)le64_to_cpu(snap_id), snap_name);
3951 static const char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
3952 u64 *snap_size, u64 *snap_features)
3957 const char *snap_name;
3960 rbd_assert(which < rbd_dev->header.snapc->num_snaps);
3961 snap_id = rbd_dev->header.snapc->snaps[which];
3962 ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size);
3966 ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, &features);
3970 snap_name = rbd_dev_v2_snap_name(rbd_dev, which);
3971 if (!IS_ERR(snap_name)) {
3973 *snap_features = features;
3978 return ERR_PTR(ret);
3981 static const char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
3982 u64 *snap_size, u64 *snap_features)
3984 if (rbd_dev->image_format == 1)
3985 return rbd_dev_v1_snap_info(rbd_dev, which,
3986 snap_size, snap_features);
3987 if (rbd_dev->image_format == 2)
3988 return rbd_dev_v2_snap_info(rbd_dev, which,
3989 snap_size, snap_features);
3990 return ERR_PTR(-EINVAL);
3993 static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev, u64 *hver)
3997 down_write(&rbd_dev->header_rwsem);
3999 ret = rbd_dev_v2_image_size(rbd_dev);
4002 rbd_update_mapping_size(rbd_dev);
4004 ret = rbd_dev_v2_snap_context(rbd_dev, hver);
4005 dout("rbd_dev_v2_snap_context returned %d\n", ret);
4008 ret = rbd_dev_snaps_update(rbd_dev);
4009 dout("rbd_dev_snaps_update returned %d\n", ret);
4013 up_write(&rbd_dev->header_rwsem);
4019 * Scan the rbd device's current snapshot list and compare it to the
4020 * newly-received snapshot context. Remove any existing snapshots
4021 * not present in the new snapshot context. Add a new snapshot for
4022 * any snaphots in the snapshot context not in the current list.
4023 * And verify there are no changes to snapshots we already know
4026 * Assumes the snapshots in the snapshot context are sorted by
4027 * snapshot id, highest id first. (Snapshots in the rbd_dev's list
4028 * are also maintained in that order.)
4030 * Note that any error occurs while updating the snapshot list
4031 * aborts the update, and the entire list is cleared. The snapshot
4032 * list becomes inconsistent at that point anyway, so it might as
4035 static int rbd_dev_snaps_update(struct rbd_device *rbd_dev)
4037 struct ceph_snap_context *snapc = rbd_dev->header.snapc;
4038 const u32 snap_count = snapc->num_snaps;
4039 struct list_head *head = &rbd_dev->snaps;
4040 struct list_head *links = head->next;
4044 dout("%s: snap count is %u\n", __func__, (unsigned int)snap_count);
4045 while (index < snap_count || links != head) {
4047 struct rbd_snap *snap;
4048 const char *snap_name;
4050 u64 snap_features = 0;
4052 snap_id = index < snap_count ? snapc->snaps[index]
4054 snap = links != head ? list_entry(links, struct rbd_snap, node)
4056 rbd_assert(!snap || snap->id != CEPH_NOSNAP);
4058 if (snap_id == CEPH_NOSNAP || (snap && snap->id > snap_id)) {
4059 struct list_head *next = links->next;
4062 * A previously-existing snapshot is not in
4063 * the new snap context.
4065 * If the now-missing snapshot is the one
4066 * the image represents, clear its existence
4067 * flag so we can avoid sending any more
4070 if (rbd_dev->spec->snap_id == snap->id)
4071 clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
4072 dout("removing %ssnap id %llu\n",
4073 rbd_dev->spec->snap_id == snap->id ?
4075 (unsigned long long)snap->id);
4077 list_del(&snap->node);
4078 rbd_snap_destroy(snap);
4080 /* Done with this list entry; advance */
4086 snap_name = rbd_dev_snap_info(rbd_dev, index,
4087 &snap_size, &snap_features);
4088 if (IS_ERR(snap_name)) {
4089 ret = PTR_ERR(snap_name);
4090 dout("failed to get snap info, error %d\n", ret);
4094 dout("entry %u: snap_id = %llu\n", (unsigned int)snap_count,
4095 (unsigned long long)snap_id);
4096 if (!snap || (snap_id != CEPH_NOSNAP && snap->id < snap_id)) {
4097 struct rbd_snap *new_snap;
4099 /* We haven't seen this snapshot before */
4101 new_snap = rbd_snap_create(rbd_dev, snap_name,
4102 snap_id, snap_size, snap_features);
4103 if (IS_ERR(new_snap)) {
4104 ret = PTR_ERR(new_snap);
4105 dout(" failed to add dev, error %d\n", ret);
4109 /* New goes before existing, or at end of list */
4111 dout(" added dev%s\n", snap ? "" : " at end\n");
4113 list_add_tail(&new_snap->node, &snap->node);
4115 list_add_tail(&new_snap->node, head);
4117 /* Already have this one */
4119 dout(" already present\n");
4121 rbd_assert(snap->size == snap_size);
4122 rbd_assert(!strcmp(snap->name, snap_name));
4123 rbd_assert(snap->features == snap_features);
4125 /* Done with this list entry; advance */
4127 links = links->next;
4130 /* Advance to the next entry in the snapshot context */
4134 dout("%s: done\n", __func__);
4138 rbd_remove_all_snaps(rbd_dev);
4143 static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
4148 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
4150 dev = &rbd_dev->dev;
4151 dev->bus = &rbd_bus_type;
4152 dev->type = &rbd_device_type;
4153 dev->parent = &rbd_root_dev;
4154 dev->release = rbd_dev_device_release;
4155 dev_set_name(dev, "%d", rbd_dev->dev_id);
4156 ret = device_register(dev);
4158 mutex_unlock(&ctl_mutex);
4163 static void rbd_bus_del_dev(struct rbd_device *rbd_dev)
4165 device_unregister(&rbd_dev->dev);
4168 static atomic64_t rbd_dev_id_max = ATOMIC64_INIT(0);
4171 * Get a unique rbd identifier for the given new rbd_dev, and add
4172 * the rbd_dev to the global list. The minimum rbd id is 1.
4174 static void rbd_dev_id_get(struct rbd_device *rbd_dev)
4176 rbd_dev->dev_id = atomic64_inc_return(&rbd_dev_id_max);
4178 spin_lock(&rbd_dev_list_lock);
4179 list_add_tail(&rbd_dev->node, &rbd_dev_list);
4180 spin_unlock(&rbd_dev_list_lock);
4181 dout("rbd_dev %p given dev id %llu\n", rbd_dev,
4182 (unsigned long long) rbd_dev->dev_id);
4186 * Remove an rbd_dev from the global list, and record that its
4187 * identifier is no longer in use.
4189 static void rbd_dev_id_put(struct rbd_device *rbd_dev)
4191 struct list_head *tmp;
4192 int rbd_id = rbd_dev->dev_id;
4195 rbd_assert(rbd_id > 0);
4197 dout("rbd_dev %p released dev id %llu\n", rbd_dev,
4198 (unsigned long long) rbd_dev->dev_id);
4199 spin_lock(&rbd_dev_list_lock);
4200 list_del_init(&rbd_dev->node);
4203 * If the id being "put" is not the current maximum, there
4204 * is nothing special we need to do.
4206 if (rbd_id != atomic64_read(&rbd_dev_id_max)) {
4207 spin_unlock(&rbd_dev_list_lock);
4212 * We need to update the current maximum id. Search the
4213 * list to find out what it is. We're more likely to find
4214 * the maximum at the end, so search the list backward.
4217 list_for_each_prev(tmp, &rbd_dev_list) {
4218 struct rbd_device *rbd_dev;
4220 rbd_dev = list_entry(tmp, struct rbd_device, node);
4221 if (rbd_dev->dev_id > max_id)
4222 max_id = rbd_dev->dev_id;
4224 spin_unlock(&rbd_dev_list_lock);
4227 * The max id could have been updated by rbd_dev_id_get(), in
4228 * which case it now accurately reflects the new maximum.
4229 * Be careful not to overwrite the maximum value in that
4232 atomic64_cmpxchg(&rbd_dev_id_max, rbd_id, max_id);
4233 dout(" max dev id has been reset\n");
4237 * Skips over white space at *buf, and updates *buf to point to the
4238 * first found non-space character (if any). Returns the length of
4239 * the token (string of non-white space characters) found. Note
4240 * that *buf must be terminated with '\0'.
4242 static inline size_t next_token(const char **buf)
4245 * These are the characters that produce nonzero for
4246 * isspace() in the "C" and "POSIX" locales.
4248 const char *spaces = " \f\n\r\t\v";
4250 *buf += strspn(*buf, spaces); /* Find start of token */
4252 return strcspn(*buf, spaces); /* Return token length */
4256 * Finds the next token in *buf, and if the provided token buffer is
4257 * big enough, copies the found token into it. The result, if
4258 * copied, is guaranteed to be terminated with '\0'. Note that *buf
4259 * must be terminated with '\0' on entry.
4261 * Returns the length of the token found (not including the '\0').
4262 * Return value will be 0 if no token is found, and it will be >=
4263 * token_size if the token would not fit.
4265 * The *buf pointer will be updated to point beyond the end of the
4266 * found token. Note that this occurs even if the token buffer is
4267 * too small to hold it.
4269 static inline size_t copy_token(const char **buf,
4275 len = next_token(buf);
4276 if (len < token_size) {
4277 memcpy(token, *buf, len);
4278 *(token + len) = '\0';
4286 * Finds the next token in *buf, dynamically allocates a buffer big
4287 * enough to hold a copy of it, and copies the token into the new
4288 * buffer. The copy is guaranteed to be terminated with '\0'. Note
4289 * that a duplicate buffer is created even for a zero-length token.
4291 * Returns a pointer to the newly-allocated duplicate, or a null
4292 * pointer if memory for the duplicate was not available. If
4293 * the lenp argument is a non-null pointer, the length of the token
4294 * (not including the '\0') is returned in *lenp.
4296 * If successful, the *buf pointer will be updated to point beyond
4297 * the end of the found token.
4299 * Note: uses GFP_KERNEL for allocation.
4301 static inline char *dup_token(const char **buf, size_t *lenp)
4306 len = next_token(buf);
4307 dup = kmemdup(*buf, len + 1, GFP_KERNEL);
4310 *(dup + len) = '\0';
4320 * Parse the options provided for an "rbd add" (i.e., rbd image
4321 * mapping) request. These arrive via a write to /sys/bus/rbd/add,
4322 * and the data written is passed here via a NUL-terminated buffer.
4323 * Returns 0 if successful or an error code otherwise.
4325 * The information extracted from these options is recorded in
4326 * the other parameters which return dynamically-allocated
4329 * The address of a pointer that will refer to a ceph options
4330 * structure. Caller must release the returned pointer using
4331 * ceph_destroy_options() when it is no longer needed.
4333 * Address of an rbd options pointer. Fully initialized by
4334 * this function; caller must release with kfree().
4336 * Address of an rbd image specification pointer. Fully
4337 * initialized by this function based on parsed options.
4338 * Caller must release with rbd_spec_put().
4340 * The options passed take this form:
4341 * <mon_addrs> <options> <pool_name> <image_name> [<snap_id>]
4344 * A comma-separated list of one or more monitor addresses.
4345 * A monitor address is an ip address, optionally followed
4346 * by a port number (separated by a colon).
4347 * I.e.: ip1[:port1][,ip2[:port2]...]
4349 * A comma-separated list of ceph and/or rbd options.
4351 * The name of the rados pool containing the rbd image.
4353 * The name of the image in that pool to map.
4355 * An optional snapshot id. If provided, the mapping will
4356 * present data from the image at the time that snapshot was
4357 * created. The image head is used if no snapshot id is
4358 * provided. Snapshot mappings are always read-only.
4360 static int rbd_add_parse_args(const char *buf,
4361 struct ceph_options **ceph_opts,
4362 struct rbd_options **opts,
4363 struct rbd_spec **rbd_spec)
4367 const char *mon_addrs;
4369 size_t mon_addrs_size;
4370 struct rbd_spec *spec = NULL;
4371 struct rbd_options *rbd_opts = NULL;
4372 struct ceph_options *copts;
4375 /* The first four tokens are required */
4377 len = next_token(&buf);
4379 rbd_warn(NULL, "no monitor address(es) provided");
4383 mon_addrs_size = len + 1;
4387 options = dup_token(&buf, NULL);
4391 rbd_warn(NULL, "no options provided");
4395 spec = rbd_spec_alloc();
4399 spec->pool_name = dup_token(&buf, NULL);
4400 if (!spec->pool_name)
4402 if (!*spec->pool_name) {
4403 rbd_warn(NULL, "no pool name provided");
4407 spec->image_name = dup_token(&buf, NULL);
4408 if (!spec->image_name)
4410 if (!*spec->image_name) {
4411 rbd_warn(NULL, "no image name provided");
4416 * Snapshot name is optional; default is to use "-"
4417 * (indicating the head/no snapshot).
4419 len = next_token(&buf);
4421 buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */
4422 len = sizeof (RBD_SNAP_HEAD_NAME) - 1;
4423 } else if (len > RBD_MAX_SNAP_NAME_LEN) {
4424 ret = -ENAMETOOLONG;
4427 snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
4430 *(snap_name + len) = '\0';
4431 spec->snap_name = snap_name;
4433 /* Initialize all rbd options to the defaults */
4435 rbd_opts = kzalloc(sizeof (*rbd_opts), GFP_KERNEL);
4439 rbd_opts->read_only = RBD_READ_ONLY_DEFAULT;
4441 copts = ceph_parse_options(options, mon_addrs,
4442 mon_addrs + mon_addrs_size - 1,
4443 parse_rbd_opts_token, rbd_opts);
4444 if (IS_ERR(copts)) {
4445 ret = PTR_ERR(copts);
4466 * An rbd format 2 image has a unique identifier, distinct from the
4467 * name given to it by the user. Internally, that identifier is
4468 * what's used to specify the names of objects related to the image.
4470 * A special "rbd id" object is used to map an rbd image name to its
4471 * id. If that object doesn't exist, then there is no v2 rbd image
4472 * with the supplied name.
4474 * This function will record the given rbd_dev's image_id field if
4475 * it can be determined, and in that case will return 0. If any
4476 * errors occur a negative errno will be returned and the rbd_dev's
4477 * image_id field will be unchanged (and should be NULL).
4479 static int rbd_dev_image_id(struct rbd_device *rbd_dev)
4488 * When probing a parent image, the image id is already
4489 * known (and the image name likely is not). There's no
4490 * need to fetch the image id again in this case. We
4491 * do still need to set the image format though.
4493 if (rbd_dev->spec->image_id) {
4494 rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1;
4500 * First, see if the format 2 image id file exists, and if
4501 * so, get the image's persistent id from it.
4503 size = sizeof (RBD_ID_PREFIX) + strlen(rbd_dev->spec->image_name);
4504 object_name = kmalloc(size, GFP_NOIO);
4507 sprintf(object_name, "%s%s", RBD_ID_PREFIX, rbd_dev->spec->image_name);
4508 dout("rbd id object name is %s\n", object_name);
4510 /* Response will be an encoded string, which includes a length */
4512 size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
4513 response = kzalloc(size, GFP_NOIO);
4519 /* If it doesn't exist we'll assume it's a format 1 image */
4521 ret = rbd_obj_method_sync(rbd_dev, object_name,
4522 "rbd", "get_id", NULL, 0,
4523 response, RBD_IMAGE_ID_LEN_MAX, NULL);
4524 dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
4525 if (ret == -ENOENT) {
4526 image_id = kstrdup("", GFP_KERNEL);
4527 ret = image_id ? 0 : -ENOMEM;
4529 rbd_dev->image_format = 1;
4530 } else if (ret > sizeof (__le32)) {
4533 image_id = ceph_extract_encoded_string(&p, p + ret,
4535 ret = IS_ERR(image_id) ? PTR_ERR(image_id) : 0;
4537 rbd_dev->image_format = 2;
4543 rbd_dev->spec->image_id = image_id;
4544 dout("image_id is %s\n", image_id);
4553 /* Undo whatever state changes are made by v1 or v2 image probe */
4555 static void rbd_dev_unprobe(struct rbd_device *rbd_dev)
4557 struct rbd_image_header *header;
4559 rbd_dev_remove_parent(rbd_dev);
4560 rbd_spec_put(rbd_dev->parent_spec);
4561 rbd_dev->parent_spec = NULL;
4562 rbd_dev->parent_overlap = 0;
4564 /* Free dynamic fields from the header, then zero it out */
4566 header = &rbd_dev->header;
4567 ceph_put_snap_context(header->snapc);
4568 kfree(header->snap_sizes);
4569 kfree(header->snap_names);
4570 kfree(header->object_prefix);
4571 memset(header, 0, sizeof (*header));
4574 static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
4578 /* Populate rbd image metadata */
4580 ret = rbd_read_header(rbd_dev, &rbd_dev->header);
4584 /* Version 1 images have no parent (no layering) */
4586 rbd_dev->parent_spec = NULL;
4587 rbd_dev->parent_overlap = 0;
4589 dout("discovered version 1 image, header name is %s\n",
4590 rbd_dev->header_name);
4595 kfree(rbd_dev->header_name);
4596 rbd_dev->header_name = NULL;
4597 kfree(rbd_dev->spec->image_id);
4598 rbd_dev->spec->image_id = NULL;
4603 static int rbd_dev_v2_probe(struct rbd_device *rbd_dev)
4608 ret = rbd_dev_v2_image_size(rbd_dev);
4612 /* Get the object prefix (a.k.a. block_name) for the image */
4614 ret = rbd_dev_v2_object_prefix(rbd_dev);
4618 /* Get the and check features for the image */
4620 ret = rbd_dev_v2_features(rbd_dev);
4624 /* If the image supports layering, get the parent info */
4626 if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
4627 ret = rbd_dev_v2_parent_info(rbd_dev);
4632 * Don't print a warning for parent images. We can
4633 * tell this point because we won't know its pool
4634 * name yet (just its pool id).
4636 if (rbd_dev->spec->pool_name)
4637 rbd_warn(rbd_dev, "WARNING: kernel layering "
4638 "is EXPERIMENTAL!");
4641 /* If the image supports fancy striping, get its parameters */
4643 if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) {
4644 ret = rbd_dev_v2_striping_info(rbd_dev);
4649 /* crypto and compression type aren't (yet) supported for v2 images */
4651 rbd_dev->header.crypt_type = 0;
4652 rbd_dev->header.comp_type = 0;
4654 /* Get the snapshot context, plus the header version */
4656 ret = rbd_dev_v2_snap_context(rbd_dev, &ver);
4659 rbd_dev->header.obj_version = ver;
4661 dout("discovered version 2 image, header name is %s\n",
4662 rbd_dev->header_name);
4666 rbd_dev->parent_overlap = 0;
4667 rbd_spec_put(rbd_dev->parent_spec);
4668 rbd_dev->parent_spec = NULL;
4669 kfree(rbd_dev->header_name);
4670 rbd_dev->header_name = NULL;
4671 kfree(rbd_dev->header.object_prefix);
4672 rbd_dev->header.object_prefix = NULL;
4677 static int rbd_dev_probe_parent(struct rbd_device *rbd_dev)
4679 struct rbd_device *parent = NULL;
4680 struct rbd_spec *parent_spec;
4681 struct rbd_client *rbdc;
4684 if (!rbd_dev->parent_spec)
4687 * We need to pass a reference to the client and the parent
4688 * spec when creating the parent rbd_dev. Images related by
4689 * parent/child relationships always share both.
4691 parent_spec = rbd_spec_get(rbd_dev->parent_spec);
4692 rbdc = __rbd_get_client(rbd_dev->rbd_client);
4695 parent = rbd_dev_create(rbdc, parent_spec);
4699 ret = rbd_dev_image_probe(parent);
4702 rbd_dev->parent = parent;
4707 rbd_spec_put(rbd_dev->parent_spec);
4708 kfree(rbd_dev->header_name);
4709 rbd_dev_destroy(parent);
4711 rbd_put_client(rbdc);
4712 rbd_spec_put(parent_spec);
4718 static int rbd_dev_device_setup(struct rbd_device *rbd_dev)
4722 ret = rbd_dev_mapping_set(rbd_dev);
4726 /* generate unique id: find highest unique id, add one */
4727 rbd_dev_id_get(rbd_dev);
4729 /* Fill in the device name, now that we have its id. */
4730 BUILD_BUG_ON(DEV_NAME_LEN
4731 < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
4732 sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->dev_id);
4734 /* Get our block major device number. */
4736 ret = register_blkdev(0, rbd_dev->name);
4739 rbd_dev->major = ret;
4741 /* Set up the blkdev mapping. */
4743 ret = rbd_init_disk(rbd_dev);
4745 goto err_out_blkdev;
4747 ret = rbd_bus_add_dev(rbd_dev);
4751 /* Everything's ready. Announce the disk to the world. */
4753 set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
4754 set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
4755 add_disk(rbd_dev->disk);
4757 pr_info("%s: added with size 0x%llx\n", rbd_dev->disk->disk_name,
4758 (unsigned long long) rbd_dev->mapping.size);
4763 rbd_free_disk(rbd_dev);
4765 unregister_blkdev(rbd_dev->major, rbd_dev->name);
4767 rbd_dev_id_put(rbd_dev);
4768 rbd_dev_mapping_clear(rbd_dev);
4773 static int rbd_dev_header_name(struct rbd_device *rbd_dev)
4775 struct rbd_spec *spec = rbd_dev->spec;
4778 /* Record the header object name for this rbd image. */
4780 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
4782 if (rbd_dev->image_format == 1)
4783 size = strlen(spec->image_name) + sizeof (RBD_SUFFIX);
4785 size = sizeof (RBD_HEADER_PREFIX) + strlen(spec->image_id);
4787 rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
4788 if (!rbd_dev->header_name)
4791 if (rbd_dev->image_format == 1)
4792 sprintf(rbd_dev->header_name, "%s%s",
4793 spec->image_name, RBD_SUFFIX);
4795 sprintf(rbd_dev->header_name, "%s%s",
4796 RBD_HEADER_PREFIX, spec->image_id);
4800 static void rbd_dev_image_release(struct rbd_device *rbd_dev)
4804 rbd_remove_all_snaps(rbd_dev);
4805 rbd_dev_unprobe(rbd_dev);
4806 ret = rbd_dev_header_watch_sync(rbd_dev, 0);
4808 rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret);
4809 kfree(rbd_dev->header_name);
4810 rbd_dev->header_name = NULL;
4811 rbd_dev->image_format = 0;
4812 kfree(rbd_dev->spec->image_id);
4813 rbd_dev->spec->image_id = NULL;
4815 rbd_dev_destroy(rbd_dev);
4819 * Probe for the existence of the header object for the given rbd
4820 * device. For format 2 images this includes determining the image
4823 static int rbd_dev_image_probe(struct rbd_device *rbd_dev)
4829 * Get the id from the image id object. If it's not a
4830 * format 2 image, we'll get ENOENT back, and we'll assume
4831 * it's a format 1 image.
4833 ret = rbd_dev_image_id(rbd_dev);
4836 rbd_assert(rbd_dev->spec->image_id);
4837 rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
4839 ret = rbd_dev_header_name(rbd_dev);
4841 goto err_out_format;
4843 ret = rbd_dev_header_watch_sync(rbd_dev, 1);
4845 goto out_header_name;
4847 if (rbd_dev->image_format == 1)
4848 ret = rbd_dev_v1_probe(rbd_dev);
4850 ret = rbd_dev_v2_probe(rbd_dev);
4854 ret = rbd_dev_snaps_update(rbd_dev);
4858 ret = rbd_dev_spec_update(rbd_dev);
4862 ret = rbd_dev_probe_parent(rbd_dev);
4867 rbd_remove_all_snaps(rbd_dev);
4869 rbd_dev_unprobe(rbd_dev);
4871 tmp = rbd_dev_header_watch_sync(rbd_dev, 0);
4873 rbd_warn(rbd_dev, "unable to tear down watch request\n");
4875 kfree(rbd_dev->header_name);
4876 rbd_dev->header_name = NULL;
4878 rbd_dev->image_format = 0;
4879 kfree(rbd_dev->spec->image_id);
4880 rbd_dev->spec->image_id = NULL;
4882 dout("probe failed, returning %d\n", ret);
4887 static ssize_t rbd_add(struct bus_type *bus,
4891 struct rbd_device *rbd_dev = NULL;
4892 struct ceph_options *ceph_opts = NULL;
4893 struct rbd_options *rbd_opts = NULL;
4894 struct rbd_spec *spec = NULL;
4895 struct rbd_client *rbdc;
4896 struct ceph_osd_client *osdc;
4899 if (!try_module_get(THIS_MODULE))
4902 /* parse add command */
4903 rc = rbd_add_parse_args(buf, &ceph_opts, &rbd_opts, &spec);
4905 goto err_out_module;
4907 rbdc = rbd_get_client(ceph_opts);
4912 ceph_opts = NULL; /* rbd_dev client now owns this */
4915 osdc = &rbdc->client->osdc;
4916 rc = ceph_pg_poolid_by_name(osdc->osdmap, spec->pool_name);
4918 goto err_out_client;
4919 spec->pool_id = (u64)rc;
4921 /* The ceph file layout needs to fit pool id in 32 bits */
4923 if (spec->pool_id > (u64)U32_MAX) {
4924 rbd_warn(NULL, "pool id too large (%llu > %u)\n",
4925 (unsigned long long)spec->pool_id, U32_MAX);
4927 goto err_out_client;
4930 rbd_dev = rbd_dev_create(rbdc, spec);
4932 goto err_out_client;
4933 rbdc = NULL; /* rbd_dev now owns this */
4934 spec = NULL; /* rbd_dev now owns this */
4936 rbd_dev->mapping.read_only = rbd_opts->read_only;
4938 rbd_opts = NULL; /* done with this */
4940 rc = rbd_dev_image_probe(rbd_dev);
4942 goto err_out_rbd_dev;
4944 rc = rbd_dev_device_setup(rbd_dev);
4948 rbd_dev_image_release(rbd_dev);
4950 rbd_dev_destroy(rbd_dev);
4952 rbd_put_client(rbdc);
4955 ceph_destroy_options(ceph_opts);
4959 module_put(THIS_MODULE);
4961 dout("Error adding device %s\n", buf);
4966 static struct rbd_device *__rbd_get_dev(unsigned long dev_id)
4968 struct list_head *tmp;
4969 struct rbd_device *rbd_dev;
4971 spin_lock(&rbd_dev_list_lock);
4972 list_for_each(tmp, &rbd_dev_list) {
4973 rbd_dev = list_entry(tmp, struct rbd_device, node);
4974 if (rbd_dev->dev_id == dev_id) {
4975 spin_unlock(&rbd_dev_list_lock);
4979 spin_unlock(&rbd_dev_list_lock);
4983 static void rbd_dev_device_release(struct device *dev)
4985 struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
4987 rbd_free_disk(rbd_dev);
4988 clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
4989 rbd_dev_clear_mapping(rbd_dev);
4990 unregister_blkdev(rbd_dev->major, rbd_dev->name);
4992 rbd_dev_id_put(rbd_dev);
4993 rbd_dev_mapping_clear(rbd_dev);
4996 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev)
4998 while (rbd_dev->parent) {
4999 struct rbd_device *first = rbd_dev;
5000 struct rbd_device *second = first->parent;
5001 struct rbd_device *third;
5004 * Follow to the parent with no grandparent and
5007 while (second && (third = second->parent)) {
5012 rbd_dev_image_release(second);
5013 first->parent = NULL;
5014 first->parent_overlap = 0;
5016 rbd_assert(first->parent_spec);
5017 rbd_spec_put(first->parent_spec);
5018 first->parent_spec = NULL;
5022 static ssize_t rbd_remove(struct bus_type *bus,
5026 struct rbd_device *rbd_dev = NULL;
5031 ret = strict_strtoul(buf, 10, &ul);
5035 /* convert to int; abort if we lost anything in the conversion */
5036 target_id = (int) ul;
5037 if (target_id != ul)
5040 mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
5042 rbd_dev = __rbd_get_dev(target_id);
5048 spin_lock_irq(&rbd_dev->lock);
5049 if (rbd_dev->open_count)
5052 set_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags);
5053 spin_unlock_irq(&rbd_dev->lock);
5057 rbd_bus_del_dev(rbd_dev);
5058 rbd_dev_image_release(rbd_dev);
5059 module_put(THIS_MODULE);
5061 mutex_unlock(&ctl_mutex);
5067 * create control files in sysfs
5070 static int rbd_sysfs_init(void)
5074 ret = device_register(&rbd_root_dev);
5078 ret = bus_register(&rbd_bus_type);
5080 device_unregister(&rbd_root_dev);
5085 static void rbd_sysfs_cleanup(void)
5087 bus_unregister(&rbd_bus_type);
5088 device_unregister(&rbd_root_dev);
5091 static int __init rbd_init(void)
5095 if (!libceph_compatible(NULL)) {
5096 rbd_warn(NULL, "libceph incompatibility (quitting)");
5100 rc = rbd_sysfs_init();
5103 pr_info("loaded " RBD_DRV_NAME_LONG "\n");
5107 static void __exit rbd_exit(void)
5109 rbd_sysfs_cleanup();
5112 module_init(rbd_init);
5113 module_exit(rbd_exit);
5115 MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
5116 MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
5117 MODULE_DESCRIPTION("rados block device");
5119 /* following authorship retained from original osdblk.c */
5120 MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");
5122 MODULE_LICENSE("GPL");