Commit | Line | Data |
---|---|---|
e2a58ee5 | 1 | |
602adf40 YS |
2 | /* |
3 | rbd.c -- Export ceph rados objects as a Linux block device | |
4 | ||
5 | ||
6 | based on drivers/block/osdblk.c: | |
7 | ||
8 | Copyright 2009 Red Hat, Inc. | |
9 | ||
10 | This program is free software; you can redistribute it and/or modify | |
11 | it under the terms of the GNU General Public License as published by | |
12 | the Free Software Foundation. | |
13 | ||
14 | This program is distributed in the hope that it will be useful, | |
15 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
16 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
17 | GNU General Public License for more details. | |
18 | ||
19 | You should have received a copy of the GNU General Public License | |
20 | along with this program; see the file COPYING. If not, write to | |
21 | the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. | |
22 | ||
23 | ||
24 | ||
dfc5606d | 25 | For usage instructions, please refer to: |
602adf40 | 26 | |
dfc5606d | 27 | Documentation/ABI/testing/sysfs-bus-rbd |
602adf40 YS |
28 | |
29 | */ | |
30 | ||
31 | #include <linux/ceph/libceph.h> | |
32 | #include <linux/ceph/osd_client.h> | |
33 | #include <linux/ceph/mon_client.h> | |
34 | #include <linux/ceph/decode.h> | |
59c2be1e | 35 | #include <linux/parser.h> |
30d1cff8 | 36 | #include <linux/bsearch.h> |
602adf40 YS |
37 | |
38 | #include <linux/kernel.h> | |
39 | #include <linux/device.h> | |
40 | #include <linux/module.h> | |
41 | #include <linux/fs.h> | |
42 | #include <linux/blkdev.h> | |
1c2a9dfe | 43 | #include <linux/slab.h> |
602adf40 YS |
44 | |
45 | #include "rbd_types.h" | |
46 | ||
aafb230e AE |
47 | #define RBD_DEBUG /* Activate rbd_assert() calls */ |
48 | ||
593a9e7b AE |
49 | /* |
50 | * The basic unit of block I/O is a sector. It is interpreted in a | |
51 | * number of contexts in Linux (blk, bio, genhd), but the default is | |
52 | * universally 512 bytes. These symbols are just slightly more | |
53 | * meaningful than the bare numbers they represent. | |
54 | */ | |
55 | #define SECTOR_SHIFT 9 | |
56 | #define SECTOR_SIZE (1ULL << SECTOR_SHIFT) | |
57 | ||
f0f8cef5 AE |
58 | #define RBD_DRV_NAME "rbd" |
59 | #define RBD_DRV_NAME_LONG "rbd (rados block device)" | |
602adf40 YS |
60 | |
61 | #define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */ | |
62 | ||
d4b125e9 AE |
63 | #define RBD_SNAP_DEV_NAME_PREFIX "snap_" |
64 | #define RBD_MAX_SNAP_NAME_LEN \ | |
65 | (NAME_MAX - (sizeof (RBD_SNAP_DEV_NAME_PREFIX) - 1)) | |
66 | ||
35d489f9 | 67 | #define RBD_MAX_SNAP_COUNT 510 /* allows max snapc to fit in 4KB */ |
602adf40 YS |
68 | |
69 | #define RBD_SNAP_HEAD_NAME "-" | |
70 | ||
9682fc6d AE |
71 | #define BAD_SNAP_INDEX U32_MAX /* invalid index into snap array */ |
72 | ||
9e15b77d AE |
73 | /* This allows a single page to hold an image name sent by OSD */ |
74 | #define RBD_IMAGE_NAME_LEN_MAX (PAGE_SIZE - sizeof (__le32) - 1) | |
1e130199 | 75 | #define RBD_IMAGE_ID_LEN_MAX 64 |
9e15b77d | 76 | |
1e130199 | 77 | #define RBD_OBJ_PREFIX_LEN_MAX 64 |
589d30e0 | 78 | |
d889140c AE |
79 | /* Feature bits */ |
80 | ||
5cbf6f12 AE |
81 | #define RBD_FEATURE_LAYERING (1<<0) |
82 | #define RBD_FEATURE_STRIPINGV2 (1<<1) | |
83 | #define RBD_FEATURES_ALL \ | |
84 | (RBD_FEATURE_LAYERING | RBD_FEATURE_STRIPINGV2) | |
d889140c AE |
85 | |
86 | /* Features supported by this (client software) implementation. */ | |
87 | ||
770eba6e | 88 | #define RBD_FEATURES_SUPPORTED (RBD_FEATURES_ALL) |
d889140c | 89 | |
81a89793 AE |
90 | /* |
91 | * An RBD device name will be "rbd#", where the "rbd" comes from | |
92 | * RBD_DRV_NAME above, and # is a unique integer identifier. | |
93 | * MAX_INT_FORMAT_WIDTH is used in ensuring DEV_NAME_LEN is big | |
94 | * enough to hold all possible device names. | |
95 | */ | |
602adf40 | 96 | #define DEV_NAME_LEN 32 |
81a89793 | 97 | #define MAX_INT_FORMAT_WIDTH ((5 * sizeof (int)) / 2 + 1) |
602adf40 YS |
98 | |
99 | /* | |
100 | * block device image metadata (in-memory version) | |
101 | */ | |
102 | struct rbd_image_header { | |
f35a4dee | 103 | /* These six fields never change for a given rbd image */ |
849b4260 | 104 | char *object_prefix; |
602adf40 YS |
105 | __u8 obj_order; |
106 | __u8 crypt_type; | |
107 | __u8 comp_type; | |
f35a4dee AE |
108 | u64 stripe_unit; |
109 | u64 stripe_count; | |
110 | u64 features; /* Might be changeable someday? */ | |
602adf40 | 111 | |
f84344f3 AE |
112 | /* The remaining fields need to be updated occasionally */ |
113 | u64 image_size; | |
114 | struct ceph_snap_context *snapc; | |
f35a4dee AE |
115 | char *snap_names; /* format 1 only */ |
116 | u64 *snap_sizes; /* format 1 only */ | |
59c2be1e YS |
117 | }; |
118 | ||
0d7dbfce AE |
119 | /* |
120 | * An rbd image specification. | |
121 | * | |
122 | * The tuple (pool_id, image_id, snap_id) is sufficient to uniquely | |
c66c6e0c AE |
123 | * identify an image. Each rbd_dev structure includes a pointer to |
124 | * an rbd_spec structure that encapsulates this identity. | |
125 | * | |
126 | * Each of the id's in an rbd_spec has an associated name. For a | |
127 | * user-mapped image, the names are supplied and the id's associated | |
128 | * with them are looked up. For a layered image, a parent image is | |
129 | * defined by the tuple, and the names are looked up. | |
130 | * | |
131 | * An rbd_dev structure contains a parent_spec pointer which is | |
132 | * non-null if the image it represents is a child in a layered | |
133 | * image. This pointer will refer to the rbd_spec structure used | |
134 | * by the parent rbd_dev for its own identity (i.e., the structure | |
135 | * is shared between the parent and child). | |
136 | * | |
137 | * Since these structures are populated once, during the discovery | |
138 | * phase of image construction, they are effectively immutable so | |
139 | * we make no effort to synchronize access to them. | |
140 | * | |
141 | * Note that code herein does not assume the image name is known (it | |
142 | * could be a null pointer). | |
0d7dbfce AE |
143 | */ |
144 | struct rbd_spec { | |
145 | u64 pool_id; | |
ecb4dc22 | 146 | const char *pool_name; |
0d7dbfce | 147 | |
ecb4dc22 AE |
148 | const char *image_id; |
149 | const char *image_name; | |
0d7dbfce AE |
150 | |
151 | u64 snap_id; | |
ecb4dc22 | 152 | const char *snap_name; |
0d7dbfce AE |
153 | |
154 | struct kref kref; | |
155 | }; | |
156 | ||
602adf40 | 157 | /* |
f0f8cef5 | 158 | * an instance of the client. multiple devices may share an rbd client. |
602adf40 YS |
159 | */ |
160 | struct rbd_client { | |
161 | struct ceph_client *client; | |
162 | struct kref kref; | |
163 | struct list_head node; | |
164 | }; | |
165 | ||
bf0d5f50 AE |
166 | struct rbd_img_request; |
167 | typedef void (*rbd_img_callback_t)(struct rbd_img_request *); | |
168 | ||
169 | #define BAD_WHICH U32_MAX /* Good which or bad which, which? */ | |
170 | ||
171 | struct rbd_obj_request; | |
172 | typedef void (*rbd_obj_callback_t)(struct rbd_obj_request *); | |
173 | ||
9969ebc5 AE |
174 | enum obj_request_type { |
175 | OBJ_REQUEST_NODATA, OBJ_REQUEST_BIO, OBJ_REQUEST_PAGES | |
176 | }; | |
bf0d5f50 | 177 | |
926f9b3f AE |
178 | enum obj_req_flags { |
179 | OBJ_REQ_DONE, /* completion flag: not done = 0, done = 1 */ | |
6365d33a | 180 | OBJ_REQ_IMG_DATA, /* object usage: standalone = 0, image = 1 */ |
5679c59f AE |
181 | OBJ_REQ_KNOWN, /* EXISTS flag valid: no = 0, yes = 1 */ |
182 | OBJ_REQ_EXISTS, /* target exists: no = 0, yes = 1 */ | |
926f9b3f AE |
183 | }; |
184 | ||
bf0d5f50 AE |
185 | struct rbd_obj_request { |
186 | const char *object_name; | |
187 | u64 offset; /* object start byte */ | |
188 | u64 length; /* bytes from offset */ | |
926f9b3f | 189 | unsigned long flags; |
bf0d5f50 | 190 | |
c5b5ef6c AE |
191 | /* |
192 | * An object request associated with an image will have its | |
193 | * img_data flag set; a standalone object request will not. | |
194 | * | |
195 | * A standalone object request will have which == BAD_WHICH | |
196 | * and a null obj_request pointer. | |
197 | * | |
198 | * An object request initiated in support of a layered image | |
199 | * object (to check for its existence before a write) will | |
200 | * have which == BAD_WHICH and a non-null obj_request pointer. | |
201 | * | |
202 | * Finally, an object request for rbd image data will have | |
203 | * which != BAD_WHICH, and will have a non-null img_request | |
204 | * pointer. The value of which will be in the range | |
205 | * 0..(img_request->obj_request_count-1). | |
206 | */ | |
207 | union { | |
208 | struct rbd_obj_request *obj_request; /* STAT op */ | |
209 | struct { | |
210 | struct rbd_img_request *img_request; | |
211 | u64 img_offset; | |
212 | /* links for img_request->obj_requests list */ | |
213 | struct list_head links; | |
214 | }; | |
215 | }; | |
bf0d5f50 AE |
216 | u32 which; /* posn image request list */ |
217 | ||
218 | enum obj_request_type type; | |
788e2df3 AE |
219 | union { |
220 | struct bio *bio_list; | |
221 | struct { | |
222 | struct page **pages; | |
223 | u32 page_count; | |
224 | }; | |
225 | }; | |
0eefd470 | 226 | struct page **copyup_pages; |
bf0d5f50 AE |
227 | |
228 | struct ceph_osd_request *osd_req; | |
229 | ||
230 | u64 xferred; /* bytes transferred */ | |
1b83bef2 | 231 | int result; |
bf0d5f50 AE |
232 | |
233 | rbd_obj_callback_t callback; | |
788e2df3 | 234 | struct completion completion; |
bf0d5f50 AE |
235 | |
236 | struct kref kref; | |
237 | }; | |
238 | ||
0c425248 | 239 | enum img_req_flags { |
9849e986 AE |
240 | IMG_REQ_WRITE, /* I/O direction: read = 0, write = 1 */ |
241 | IMG_REQ_CHILD, /* initiator: block = 0, child image = 1 */ | |
d0b2e944 | 242 | IMG_REQ_LAYERED, /* ENOENT handling: normal = 0, layered = 1 */ |
0c425248 AE |
243 | }; |
244 | ||
bf0d5f50 | 245 | struct rbd_img_request { |
bf0d5f50 AE |
246 | struct rbd_device *rbd_dev; |
247 | u64 offset; /* starting image byte offset */ | |
248 | u64 length; /* byte count from offset */ | |
0c425248 | 249 | unsigned long flags; |
bf0d5f50 | 250 | union { |
9849e986 | 251 | u64 snap_id; /* for reads */ |
bf0d5f50 | 252 | struct ceph_snap_context *snapc; /* for writes */ |
9849e986 AE |
253 | }; |
254 | union { | |
255 | struct request *rq; /* block request */ | |
256 | struct rbd_obj_request *obj_request; /* obj req initiator */ | |
bf0d5f50 | 257 | }; |
3d7efd18 | 258 | struct page **copyup_pages; |
bf0d5f50 AE |
259 | spinlock_t completion_lock;/* protects next_completion */ |
260 | u32 next_completion; | |
261 | rbd_img_callback_t callback; | |
55f27e09 | 262 | u64 xferred;/* aggregate bytes transferred */ |
a5a337d4 | 263 | int result; /* first nonzero obj_request result */ |
bf0d5f50 AE |
264 | |
265 | u32 obj_request_count; | |
266 | struct list_head obj_requests; /* rbd_obj_request structs */ | |
267 | ||
268 | struct kref kref; | |
269 | }; | |
270 | ||
271 | #define for_each_obj_request(ireq, oreq) \ | |
ef06f4d3 | 272 | list_for_each_entry(oreq, &(ireq)->obj_requests, links) |
bf0d5f50 | 273 | #define for_each_obj_request_from(ireq, oreq) \ |
ef06f4d3 | 274 | list_for_each_entry_from(oreq, &(ireq)->obj_requests, links) |
bf0d5f50 | 275 | #define for_each_obj_request_safe(ireq, oreq, n) \ |
ef06f4d3 | 276 | list_for_each_entry_safe_reverse(oreq, n, &(ireq)->obj_requests, links) |
bf0d5f50 | 277 | |
f84344f3 | 278 | struct rbd_mapping { |
99c1f08f | 279 | u64 size; |
34b13184 | 280 | u64 features; |
f84344f3 AE |
281 | bool read_only; |
282 | }; | |
283 | ||
602adf40 YS |
284 | /* |
285 | * a single device | |
286 | */ | |
287 | struct rbd_device { | |
de71a297 | 288 | int dev_id; /* blkdev unique id */ |
602adf40 YS |
289 | |
290 | int major; /* blkdev assigned major */ | |
291 | struct gendisk *disk; /* blkdev's gendisk and rq */ | |
602adf40 | 292 | |
a30b71b9 | 293 | u32 image_format; /* Either 1 or 2 */ |
602adf40 YS |
294 | struct rbd_client *rbd_client; |
295 | ||
296 | char name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */ | |
297 | ||
b82d167b | 298 | spinlock_t lock; /* queue, flags, open_count */ |
602adf40 YS |
299 | |
300 | struct rbd_image_header header; | |
b82d167b | 301 | unsigned long flags; /* possibly lock protected */ |
0d7dbfce | 302 | struct rbd_spec *spec; |
602adf40 | 303 | |
0d7dbfce | 304 | char *header_name; |
971f839a | 305 | |
0903e875 AE |
306 | struct ceph_file_layout layout; |
307 | ||
59c2be1e | 308 | struct ceph_osd_event *watch_event; |
975241af | 309 | struct rbd_obj_request *watch_request; |
59c2be1e | 310 | |
86b00e0d AE |
311 | struct rbd_spec *parent_spec; |
312 | u64 parent_overlap; | |
2f82ee54 | 313 | struct rbd_device *parent; |
86b00e0d | 314 | |
c666601a JD |
315 | /* protects updating the header */ |
316 | struct rw_semaphore header_rwsem; | |
f84344f3 AE |
317 | |
318 | struct rbd_mapping mapping; | |
602adf40 YS |
319 | |
320 | struct list_head node; | |
dfc5606d | 321 | |
dfc5606d YS |
322 | /* sysfs related */ |
323 | struct device dev; | |
b82d167b | 324 | unsigned long open_count; /* protected by lock */ |
dfc5606d YS |
325 | }; |
326 | ||
b82d167b AE |
327 | /* |
328 | * Flag bits for rbd_dev->flags. If atomicity is required, | |
329 | * rbd_dev->lock is used to protect access. | |
330 | * | |
331 | * Currently, only the "removing" flag (which is coupled with the | |
332 | * "open_count" field) requires atomic access. | |
333 | */ | |
6d292906 AE |
334 | enum rbd_dev_flags { |
335 | RBD_DEV_FLAG_EXISTS, /* mapped snapshot has not been deleted */ | |
b82d167b | 336 | RBD_DEV_FLAG_REMOVING, /* this mapping is being removed */ |
6d292906 AE |
337 | }; |
338 | ||
602adf40 | 339 | static DEFINE_MUTEX(ctl_mutex); /* Serialize open/close/setup/teardown */ |
e124a82f | 340 | |
602adf40 | 341 | static LIST_HEAD(rbd_dev_list); /* devices */ |
e124a82f AE |
342 | static DEFINE_SPINLOCK(rbd_dev_list_lock); |
343 | ||
432b8587 AE |
344 | static LIST_HEAD(rbd_client_list); /* clients */ |
345 | static DEFINE_SPINLOCK(rbd_client_list_lock); | |
602adf40 | 346 | |
78c2a44a AE |
347 | /* Slab caches for frequently-allocated structures */ |
348 | ||
1c2a9dfe | 349 | static struct kmem_cache *rbd_img_request_cache; |
868311b1 | 350 | static struct kmem_cache *rbd_obj_request_cache; |
78c2a44a | 351 | static struct kmem_cache *rbd_segment_name_cache; |
1c2a9dfe | 352 | |
3d7efd18 AE |
353 | static int rbd_img_request_submit(struct rbd_img_request *img_request); |
354 | ||
200a6a8b | 355 | static void rbd_dev_device_release(struct device *dev); |
dfc5606d | 356 | |
f0f8cef5 AE |
357 | static ssize_t rbd_add(struct bus_type *bus, const char *buf, |
358 | size_t count); | |
359 | static ssize_t rbd_remove(struct bus_type *bus, const char *buf, | |
360 | size_t count); | |
51344a38 | 361 | static int rbd_dev_image_probe(struct rbd_device *rbd_dev, bool read_only); |
f0f8cef5 AE |
362 | |
363 | static struct bus_attribute rbd_bus_attrs[] = { | |
364 | __ATTR(add, S_IWUSR, NULL, rbd_add), | |
365 | __ATTR(remove, S_IWUSR, NULL, rbd_remove), | |
366 | __ATTR_NULL | |
367 | }; | |
368 | ||
369 | static struct bus_type rbd_bus_type = { | |
370 | .name = "rbd", | |
371 | .bus_attrs = rbd_bus_attrs, | |
372 | }; | |
373 | ||
374 | static void rbd_root_dev_release(struct device *dev) | |
375 | { | |
376 | } | |
377 | ||
378 | static struct device rbd_root_dev = { | |
379 | .init_name = "rbd", | |
380 | .release = rbd_root_dev_release, | |
381 | }; | |
382 | ||
06ecc6cb AE |
383 | static __printf(2, 3) |
384 | void rbd_warn(struct rbd_device *rbd_dev, const char *fmt, ...) | |
385 | { | |
386 | struct va_format vaf; | |
387 | va_list args; | |
388 | ||
389 | va_start(args, fmt); | |
390 | vaf.fmt = fmt; | |
391 | vaf.va = &args; | |
392 | ||
393 | if (!rbd_dev) | |
394 | printk(KERN_WARNING "%s: %pV\n", RBD_DRV_NAME, &vaf); | |
395 | else if (rbd_dev->disk) | |
396 | printk(KERN_WARNING "%s: %s: %pV\n", | |
397 | RBD_DRV_NAME, rbd_dev->disk->disk_name, &vaf); | |
398 | else if (rbd_dev->spec && rbd_dev->spec->image_name) | |
399 | printk(KERN_WARNING "%s: image %s: %pV\n", | |
400 | RBD_DRV_NAME, rbd_dev->spec->image_name, &vaf); | |
401 | else if (rbd_dev->spec && rbd_dev->spec->image_id) | |
402 | printk(KERN_WARNING "%s: id %s: %pV\n", | |
403 | RBD_DRV_NAME, rbd_dev->spec->image_id, &vaf); | |
404 | else /* punt */ | |
405 | printk(KERN_WARNING "%s: rbd_dev %p: %pV\n", | |
406 | RBD_DRV_NAME, rbd_dev, &vaf); | |
407 | va_end(args); | |
408 | } | |
409 | ||
aafb230e AE |
410 | #ifdef RBD_DEBUG |
411 | #define rbd_assert(expr) \ | |
412 | if (unlikely(!(expr))) { \ | |
413 | printk(KERN_ERR "\nAssertion failure in %s() " \ | |
414 | "at line %d:\n\n" \ | |
415 | "\trbd_assert(%s);\n\n", \ | |
416 | __func__, __LINE__, #expr); \ | |
417 | BUG(); \ | |
418 | } | |
419 | #else /* !RBD_DEBUG */ | |
420 | # define rbd_assert(expr) ((void) 0) | |
421 | #endif /* !RBD_DEBUG */ | |
dfc5606d | 422 | |
b454e36d | 423 | static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request); |
05a46afd AE |
424 | static void rbd_img_parent_read(struct rbd_obj_request *obj_request); |
425 | static void rbd_dev_remove_parent(struct rbd_device *rbd_dev); | |
8b3e1a56 | 426 | |
cc4a38bd AE |
427 | static int rbd_dev_refresh(struct rbd_device *rbd_dev); |
428 | static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev); | |
54cac61f AE |
429 | static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, |
430 | u64 snap_id); | |
2ad3d716 AE |
431 | static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id, |
432 | u8 *order, u64 *snap_size); | |
433 | static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id, | |
434 | u64 *snap_features); | |
435 | static u64 rbd_snap_id_by_name(struct rbd_device *rbd_dev, const char *name); | |
59c2be1e | 436 | |
602adf40 YS |
437 | static int rbd_open(struct block_device *bdev, fmode_t mode) |
438 | { | |
f0f8cef5 | 439 | struct rbd_device *rbd_dev = bdev->bd_disk->private_data; |
b82d167b | 440 | bool removing = false; |
602adf40 | 441 | |
f84344f3 | 442 | if ((mode & FMODE_WRITE) && rbd_dev->mapping.read_only) |
602adf40 YS |
443 | return -EROFS; |
444 | ||
a14ea269 | 445 | spin_lock_irq(&rbd_dev->lock); |
b82d167b AE |
446 | if (test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags)) |
447 | removing = true; | |
448 | else | |
449 | rbd_dev->open_count++; | |
a14ea269 | 450 | spin_unlock_irq(&rbd_dev->lock); |
b82d167b AE |
451 | if (removing) |
452 | return -ENOENT; | |
453 | ||
42382b70 | 454 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); |
c3e946ce | 455 | (void) get_device(&rbd_dev->dev); |
f84344f3 | 456 | set_device_ro(bdev, rbd_dev->mapping.read_only); |
42382b70 | 457 | mutex_unlock(&ctl_mutex); |
340c7a2b | 458 | |
602adf40 YS |
459 | return 0; |
460 | } | |
461 | ||
dfc5606d YS |
462 | static int rbd_release(struct gendisk *disk, fmode_t mode) |
463 | { | |
464 | struct rbd_device *rbd_dev = disk->private_data; | |
b82d167b AE |
465 | unsigned long open_count_before; |
466 | ||
a14ea269 | 467 | spin_lock_irq(&rbd_dev->lock); |
b82d167b | 468 | open_count_before = rbd_dev->open_count--; |
a14ea269 | 469 | spin_unlock_irq(&rbd_dev->lock); |
b82d167b | 470 | rbd_assert(open_count_before > 0); |
dfc5606d | 471 | |
42382b70 | 472 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); |
c3e946ce | 473 | put_device(&rbd_dev->dev); |
42382b70 | 474 | mutex_unlock(&ctl_mutex); |
dfc5606d YS |
475 | |
476 | return 0; | |
477 | } | |
478 | ||
602adf40 YS |
479 | static const struct block_device_operations rbd_bd_ops = { |
480 | .owner = THIS_MODULE, | |
481 | .open = rbd_open, | |
dfc5606d | 482 | .release = rbd_release, |
602adf40 YS |
483 | }; |
484 | ||
485 | /* | |
486 | * Initialize an rbd client instance. | |
43ae4701 | 487 | * We own *ceph_opts. |
602adf40 | 488 | */ |
f8c38929 | 489 | static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts) |
602adf40 YS |
490 | { |
491 | struct rbd_client *rbdc; | |
492 | int ret = -ENOMEM; | |
493 | ||
37206ee5 | 494 | dout("%s:\n", __func__); |
602adf40 YS |
495 | rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL); |
496 | if (!rbdc) | |
497 | goto out_opt; | |
498 | ||
499 | kref_init(&rbdc->kref); | |
500 | INIT_LIST_HEAD(&rbdc->node); | |
501 | ||
bc534d86 AE |
502 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); |
503 | ||
43ae4701 | 504 | rbdc->client = ceph_create_client(ceph_opts, rbdc, 0, 0); |
602adf40 | 505 | if (IS_ERR(rbdc->client)) |
bc534d86 | 506 | goto out_mutex; |
43ae4701 | 507 | ceph_opts = NULL; /* Now rbdc->client is responsible for ceph_opts */ |
602adf40 YS |
508 | |
509 | ret = ceph_open_session(rbdc->client); | |
510 | if (ret < 0) | |
511 | goto out_err; | |
512 | ||
432b8587 | 513 | spin_lock(&rbd_client_list_lock); |
602adf40 | 514 | list_add_tail(&rbdc->node, &rbd_client_list); |
432b8587 | 515 | spin_unlock(&rbd_client_list_lock); |
602adf40 | 516 | |
bc534d86 | 517 | mutex_unlock(&ctl_mutex); |
37206ee5 | 518 | dout("%s: rbdc %p\n", __func__, rbdc); |
bc534d86 | 519 | |
602adf40 YS |
520 | return rbdc; |
521 | ||
522 | out_err: | |
523 | ceph_destroy_client(rbdc->client); | |
bc534d86 AE |
524 | out_mutex: |
525 | mutex_unlock(&ctl_mutex); | |
602adf40 YS |
526 | kfree(rbdc); |
527 | out_opt: | |
43ae4701 AE |
528 | if (ceph_opts) |
529 | ceph_destroy_options(ceph_opts); | |
37206ee5 AE |
530 | dout("%s: error %d\n", __func__, ret); |
531 | ||
28f259b7 | 532 | return ERR_PTR(ret); |
602adf40 YS |
533 | } |
534 | ||
2f82ee54 AE |
535 | static struct rbd_client *__rbd_get_client(struct rbd_client *rbdc) |
536 | { | |
537 | kref_get(&rbdc->kref); | |
538 | ||
539 | return rbdc; | |
540 | } | |
541 | ||
602adf40 | 542 | /* |
1f7ba331 AE |
543 | * Find a ceph client with specific addr and configuration. If |
544 | * found, bump its reference count. | |
602adf40 | 545 | */ |
1f7ba331 | 546 | static struct rbd_client *rbd_client_find(struct ceph_options *ceph_opts) |
602adf40 YS |
547 | { |
548 | struct rbd_client *client_node; | |
1f7ba331 | 549 | bool found = false; |
602adf40 | 550 | |
43ae4701 | 551 | if (ceph_opts->flags & CEPH_OPT_NOSHARE) |
602adf40 YS |
552 | return NULL; |
553 | ||
1f7ba331 AE |
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)) { | |
2f82ee54 AE |
557 | __rbd_get_client(client_node); |
558 | ||
1f7ba331 AE |
559 | found = true; |
560 | break; | |
561 | } | |
562 | } | |
563 | spin_unlock(&rbd_client_list_lock); | |
564 | ||
565 | return found ? client_node : NULL; | |
602adf40 YS |
566 | } |
567 | ||
59c2be1e YS |
568 | /* |
569 | * mount options | |
570 | */ | |
571 | enum { | |
59c2be1e YS |
572 | Opt_last_int, |
573 | /* int args above */ | |
574 | Opt_last_string, | |
575 | /* string args above */ | |
cc0538b6 AE |
576 | Opt_read_only, |
577 | Opt_read_write, | |
578 | /* Boolean args above */ | |
579 | Opt_last_bool, | |
59c2be1e YS |
580 | }; |
581 | ||
43ae4701 | 582 | static match_table_t rbd_opts_tokens = { |
59c2be1e YS |
583 | /* int args above */ |
584 | /* string args above */ | |
be466c1c | 585 | {Opt_read_only, "read_only"}, |
cc0538b6 AE |
586 | {Opt_read_only, "ro"}, /* Alternate spelling */ |
587 | {Opt_read_write, "read_write"}, | |
588 | {Opt_read_write, "rw"}, /* Alternate spelling */ | |
589 | /* Boolean args above */ | |
59c2be1e YS |
590 | {-1, NULL} |
591 | }; | |
592 | ||
98571b5a AE |
593 | struct rbd_options { |
594 | bool read_only; | |
595 | }; | |
596 | ||
597 | #define RBD_READ_ONLY_DEFAULT false | |
598 | ||
59c2be1e YS |
599 | static int parse_rbd_opts_token(char *c, void *private) |
600 | { | |
43ae4701 | 601 | struct rbd_options *rbd_opts = private; |
59c2be1e YS |
602 | substring_t argstr[MAX_OPT_ARGS]; |
603 | int token, intval, ret; | |
604 | ||
43ae4701 | 605 | token = match_token(c, rbd_opts_tokens, argstr); |
59c2be1e YS |
606 | if (token < 0) |
607 | return -EINVAL; | |
608 | ||
609 | if (token < Opt_last_int) { | |
610 | ret = match_int(&argstr[0], &intval); | |
611 | if (ret < 0) { | |
612 | pr_err("bad mount option arg (not int) " | |
613 | "at '%s'\n", c); | |
614 | return ret; | |
615 | } | |
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, | |
619 | argstr[0].from); | |
cc0538b6 AE |
620 | } else if (token > Opt_last_string && token < Opt_last_bool) { |
621 | dout("got Boolean token %d\n", token); | |
59c2be1e YS |
622 | } else { |
623 | dout("got token %d\n", token); | |
624 | } | |
625 | ||
626 | switch (token) { | |
cc0538b6 AE |
627 | case Opt_read_only: |
628 | rbd_opts->read_only = true; | |
629 | break; | |
630 | case Opt_read_write: | |
631 | rbd_opts->read_only = false; | |
632 | break; | |
59c2be1e | 633 | default: |
aafb230e AE |
634 | rbd_assert(false); |
635 | break; | |
59c2be1e YS |
636 | } |
637 | return 0; | |
638 | } | |
639 | ||
602adf40 YS |
640 | /* |
641 | * Get a ceph client with specific addr and configuration, if one does | |
642 | * not exist create it. | |
643 | */ | |
9d3997fd | 644 | static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts) |
602adf40 | 645 | { |
f8c38929 | 646 | struct rbd_client *rbdc; |
59c2be1e | 647 | |
1f7ba331 | 648 | rbdc = rbd_client_find(ceph_opts); |
9d3997fd | 649 | if (rbdc) /* using an existing client */ |
43ae4701 | 650 | ceph_destroy_options(ceph_opts); |
9d3997fd | 651 | else |
f8c38929 | 652 | rbdc = rbd_client_create(ceph_opts); |
602adf40 | 653 | |
9d3997fd | 654 | return rbdc; |
602adf40 YS |
655 | } |
656 | ||
657 | /* | |
658 | * Destroy ceph client | |
d23a4b3f | 659 | * |
432b8587 | 660 | * Caller must hold rbd_client_list_lock. |
602adf40 YS |
661 | */ |
662 | static void rbd_client_release(struct kref *kref) | |
663 | { | |
664 | struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref); | |
665 | ||
37206ee5 | 666 | dout("%s: rbdc %p\n", __func__, rbdc); |
cd9d9f5d | 667 | spin_lock(&rbd_client_list_lock); |
602adf40 | 668 | list_del(&rbdc->node); |
cd9d9f5d | 669 | spin_unlock(&rbd_client_list_lock); |
602adf40 YS |
670 | |
671 | ceph_destroy_client(rbdc->client); | |
672 | kfree(rbdc); | |
673 | } | |
674 | ||
675 | /* | |
676 | * Drop reference to ceph client node. If it's not referenced anymore, release | |
677 | * it. | |
678 | */ | |
9d3997fd | 679 | static void rbd_put_client(struct rbd_client *rbdc) |
602adf40 | 680 | { |
c53d5893 AE |
681 | if (rbdc) |
682 | kref_put(&rbdc->kref, rbd_client_release); | |
602adf40 YS |
683 | } |
684 | ||
a30b71b9 AE |
685 | static bool rbd_image_format_valid(u32 image_format) |
686 | { | |
687 | return image_format == 1 || image_format == 2; | |
688 | } | |
689 | ||
8e94af8e AE |
690 | static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk) |
691 | { | |
103a150f AE |
692 | size_t size; |
693 | u32 snap_count; | |
694 | ||
695 | /* The header has to start with the magic rbd header text */ | |
696 | if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT))) | |
697 | return false; | |
698 | ||
db2388b6 AE |
699 | /* The bio layer requires at least sector-sized I/O */ |
700 | ||
701 | if (ondisk->options.order < SECTOR_SHIFT) | |
702 | return false; | |
703 | ||
704 | /* If we use u64 in a few spots we may be able to loosen this */ | |
705 | ||
706 | if (ondisk->options.order > 8 * sizeof (int) - 1) | |
707 | return false; | |
708 | ||
103a150f AE |
709 | /* |
710 | * The size of a snapshot header has to fit in a size_t, and | |
711 | * that limits the number of snapshots. | |
712 | */ | |
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)) | |
716 | return false; | |
717 | ||
718 | /* | |
719 | * Not only that, but the size of the entire the snapshot | |
720 | * header must also be representable in a size_t. | |
721 | */ | |
722 | size -= snap_count * sizeof (__le64); | |
723 | if ((u64) size < le64_to_cpu(ondisk->snap_names_len)) | |
724 | return false; | |
725 | ||
726 | return true; | |
8e94af8e AE |
727 | } |
728 | ||
602adf40 | 729 | /* |
bb23e37a AE |
730 | * Fill an rbd image header with information from the given format 1 |
731 | * on-disk header. | |
602adf40 | 732 | */ |
662518b1 | 733 | static int rbd_header_from_disk(struct rbd_device *rbd_dev, |
4156d998 | 734 | struct rbd_image_header_ondisk *ondisk) |
602adf40 | 735 | { |
662518b1 | 736 | struct rbd_image_header *header = &rbd_dev->header; |
bb23e37a AE |
737 | bool first_time = header->object_prefix == NULL; |
738 | struct ceph_snap_context *snapc; | |
739 | char *object_prefix = NULL; | |
740 | char *snap_names = NULL; | |
741 | u64 *snap_sizes = NULL; | |
ccece235 | 742 | u32 snap_count; |
d2bb24e5 | 743 | size_t size; |
bb23e37a | 744 | int ret = -ENOMEM; |
621901d6 | 745 | u32 i; |
602adf40 | 746 | |
bb23e37a AE |
747 | /* Allocate this now to avoid having to handle failure below */ |
748 | ||
749 | if (first_time) { | |
750 | size_t len; | |
751 | ||
752 | len = strnlen(ondisk->object_prefix, | |
753 | sizeof (ondisk->object_prefix)); | |
754 | object_prefix = kmalloc(len + 1, GFP_KERNEL); | |
755 | if (!object_prefix) | |
756 | return -ENOMEM; | |
757 | memcpy(object_prefix, ondisk->object_prefix, len); | |
758 | object_prefix[len] = '\0'; | |
759 | } | |
103a150f | 760 | |
bb23e37a | 761 | /* Allocate the snapshot context and fill it in */ |
00f1f36f | 762 | |
bb23e37a AE |
763 | snap_count = le32_to_cpu(ondisk->snap_count); |
764 | snapc = ceph_create_snap_context(snap_count, GFP_KERNEL); | |
765 | if (!snapc) | |
766 | goto out_err; | |
767 | snapc->seq = le64_to_cpu(ondisk->snap_seq); | |
602adf40 | 768 | if (snap_count) { |
bb23e37a | 769 | struct rbd_image_snap_ondisk *snaps; |
f785cc1d AE |
770 | u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len); |
771 | ||
bb23e37a | 772 | /* We'll keep a copy of the snapshot names... */ |
621901d6 | 773 | |
bb23e37a AE |
774 | if (snap_names_len > (u64)SIZE_MAX) |
775 | goto out_2big; | |
776 | snap_names = kmalloc(snap_names_len, GFP_KERNEL); | |
777 | if (!snap_names) | |
6a52325f AE |
778 | goto out_err; |
779 | ||
bb23e37a | 780 | /* ...as well as the array of their sizes. */ |
621901d6 | 781 | |
d2bb24e5 | 782 | size = snap_count * sizeof (*header->snap_sizes); |
bb23e37a AE |
783 | snap_sizes = kmalloc(size, GFP_KERNEL); |
784 | if (!snap_sizes) | |
6a52325f | 785 | goto out_err; |
bb23e37a AE |
786 | |
787 | /* | |
788 | * Copy the names, and fill in each snapshot's id | |
789 | * and size. | |
790 | * | |
791 | * Note that rbd_dev_v1_header_read() guarantees the | |
792 | * ondisk buffer we're working with has | |
793 | * snap_names_len bytes beyond the end of the | |
794 | * snapshot id array, this memcpy() is safe. | |
795 | */ | |
796 | memcpy(snap_names, &ondisk->snaps[snap_count], snap_names_len); | |
797 | snaps = ondisk->snaps; | |
798 | for (i = 0; i < snap_count; i++) { | |
799 | snapc->snaps[i] = le64_to_cpu(snaps[i].id); | |
800 | snap_sizes[i] = le64_to_cpu(snaps[i].image_size); | |
801 | } | |
602adf40 | 802 | } |
849b4260 | 803 | |
bb23e37a AE |
804 | /* We won't fail any more, fill in the header */ |
805 | ||
662518b1 | 806 | down_write(&rbd_dev->header_rwsem); |
bb23e37a AE |
807 | if (first_time) { |
808 | header->object_prefix = object_prefix; | |
809 | header->obj_order = ondisk->options.order; | |
810 | header->crypt_type = ondisk->options.crypt_type; | |
811 | header->comp_type = ondisk->options.comp_type; | |
812 | /* The rest aren't used for format 1 images */ | |
813 | header->stripe_unit = 0; | |
814 | header->stripe_count = 0; | |
815 | header->features = 0; | |
662518b1 AE |
816 | } else { |
817 | ceph_put_snap_context(header->snapc); | |
818 | kfree(header->snap_names); | |
819 | kfree(header->snap_sizes); | |
bb23e37a | 820 | } |
6a52325f | 821 | |
bb23e37a | 822 | /* The remaining fields always get updated (when we refresh) */ |
621901d6 | 823 | |
f84344f3 | 824 | header->image_size = le64_to_cpu(ondisk->image_size); |
bb23e37a AE |
825 | header->snapc = snapc; |
826 | header->snap_names = snap_names; | |
827 | header->snap_sizes = snap_sizes; | |
602adf40 | 828 | |
662518b1 AE |
829 | /* Make sure mapping size is consistent with header info */ |
830 | ||
831 | if (rbd_dev->spec->snap_id == CEPH_NOSNAP || first_time) | |
832 | if (rbd_dev->mapping.size != header->image_size) | |
833 | rbd_dev->mapping.size = header->image_size; | |
834 | ||
835 | up_write(&rbd_dev->header_rwsem); | |
836 | ||
602adf40 | 837 | return 0; |
bb23e37a AE |
838 | out_2big: |
839 | ret = -EIO; | |
6a52325f | 840 | out_err: |
bb23e37a AE |
841 | kfree(snap_sizes); |
842 | kfree(snap_names); | |
843 | ceph_put_snap_context(snapc); | |
844 | kfree(object_prefix); | |
845 | ||
846 | return ret; | |
602adf40 YS |
847 | } |
848 | ||
9682fc6d AE |
849 | static const char *_rbd_dev_v1_snap_name(struct rbd_device *rbd_dev, u32 which) |
850 | { | |
851 | const char *snap_name; | |
852 | ||
853 | rbd_assert(which < rbd_dev->header.snapc->num_snaps); | |
854 | ||
855 | /* Skip over names until we find the one we are looking for */ | |
856 | ||
857 | snap_name = rbd_dev->header.snap_names; | |
858 | while (which--) | |
859 | snap_name += strlen(snap_name) + 1; | |
860 | ||
861 | return kstrdup(snap_name, GFP_KERNEL); | |
862 | } | |
863 | ||
30d1cff8 AE |
864 | /* |
865 | * Snapshot id comparison function for use with qsort()/bsearch(). | |
866 | * Note that result is for snapshots in *descending* order. | |
867 | */ | |
868 | static int snapid_compare_reverse(const void *s1, const void *s2) | |
869 | { | |
870 | u64 snap_id1 = *(u64 *)s1; | |
871 | u64 snap_id2 = *(u64 *)s2; | |
872 | ||
873 | if (snap_id1 < snap_id2) | |
874 | return 1; | |
875 | return snap_id1 == snap_id2 ? 0 : -1; | |
876 | } | |
877 | ||
878 | /* | |
879 | * Search a snapshot context to see if the given snapshot id is | |
880 | * present. | |
881 | * | |
882 | * Returns the position of the snapshot id in the array if it's found, | |
883 | * or BAD_SNAP_INDEX otherwise. | |
884 | * | |
885 | * Note: The snapshot array is in kept sorted (by the osd) in | |
886 | * reverse order, highest snapshot id first. | |
887 | */ | |
9682fc6d AE |
888 | static u32 rbd_dev_snap_index(struct rbd_device *rbd_dev, u64 snap_id) |
889 | { | |
890 | struct ceph_snap_context *snapc = rbd_dev->header.snapc; | |
30d1cff8 | 891 | u64 *found; |
9682fc6d | 892 | |
30d1cff8 AE |
893 | found = bsearch(&snap_id, &snapc->snaps, snapc->num_snaps, |
894 | sizeof (snap_id), snapid_compare_reverse); | |
9682fc6d | 895 | |
30d1cff8 | 896 | return found ? (u32)(found - &snapc->snaps[0]) : BAD_SNAP_INDEX; |
9682fc6d AE |
897 | } |
898 | ||
2ad3d716 AE |
899 | static const char *rbd_dev_v1_snap_name(struct rbd_device *rbd_dev, |
900 | u64 snap_id) | |
9e15b77d | 901 | { |
54cac61f | 902 | u32 which; |
9e15b77d | 903 | |
54cac61f AE |
904 | which = rbd_dev_snap_index(rbd_dev, snap_id); |
905 | if (which == BAD_SNAP_INDEX) | |
906 | return NULL; | |
907 | ||
908 | return _rbd_dev_v1_snap_name(rbd_dev, which); | |
909 | } | |
910 | ||
911 | static const char *rbd_snap_name(struct rbd_device *rbd_dev, u64 snap_id) | |
912 | { | |
9e15b77d AE |
913 | if (snap_id == CEPH_NOSNAP) |
914 | return RBD_SNAP_HEAD_NAME; | |
915 | ||
54cac61f AE |
916 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); |
917 | if (rbd_dev->image_format == 1) | |
918 | return rbd_dev_v1_snap_name(rbd_dev, snap_id); | |
9e15b77d | 919 | |
54cac61f | 920 | return rbd_dev_v2_snap_name(rbd_dev, snap_id); |
9e15b77d AE |
921 | } |
922 | ||
2ad3d716 AE |
923 | static int rbd_snap_size(struct rbd_device *rbd_dev, u64 snap_id, |
924 | u64 *snap_size) | |
602adf40 | 925 | { |
2ad3d716 AE |
926 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); |
927 | if (snap_id == CEPH_NOSNAP) { | |
928 | *snap_size = rbd_dev->header.image_size; | |
929 | } else if (rbd_dev->image_format == 1) { | |
930 | u32 which; | |
602adf40 | 931 | |
2ad3d716 AE |
932 | which = rbd_dev_snap_index(rbd_dev, snap_id); |
933 | if (which == BAD_SNAP_INDEX) | |
934 | return -ENOENT; | |
e86924a8 | 935 | |
2ad3d716 AE |
936 | *snap_size = rbd_dev->header.snap_sizes[which]; |
937 | } else { | |
938 | u64 size = 0; | |
939 | int ret; | |
940 | ||
941 | ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size); | |
942 | if (ret) | |
943 | return ret; | |
944 | ||
945 | *snap_size = size; | |
946 | } | |
947 | return 0; | |
602adf40 YS |
948 | } |
949 | ||
2ad3d716 AE |
950 | static int rbd_snap_features(struct rbd_device *rbd_dev, u64 snap_id, |
951 | u64 *snap_features) | |
602adf40 | 952 | { |
2ad3d716 AE |
953 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); |
954 | if (snap_id == CEPH_NOSNAP) { | |
955 | *snap_features = rbd_dev->header.features; | |
956 | } else if (rbd_dev->image_format == 1) { | |
957 | *snap_features = 0; /* No features for format 1 */ | |
602adf40 | 958 | } else { |
2ad3d716 AE |
959 | u64 features = 0; |
960 | int ret; | |
8b0241f8 | 961 | |
2ad3d716 AE |
962 | ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, &features); |
963 | if (ret) | |
964 | return ret; | |
965 | ||
966 | *snap_features = features; | |
967 | } | |
968 | return 0; | |
969 | } | |
970 | ||
971 | static int rbd_dev_mapping_set(struct rbd_device *rbd_dev) | |
972 | { | |
8f4b7d98 | 973 | u64 snap_id = rbd_dev->spec->snap_id; |
2ad3d716 AE |
974 | u64 size = 0; |
975 | u64 features = 0; | |
976 | int ret; | |
977 | ||
2ad3d716 AE |
978 | ret = rbd_snap_size(rbd_dev, snap_id, &size); |
979 | if (ret) | |
980 | return ret; | |
981 | ret = rbd_snap_features(rbd_dev, snap_id, &features); | |
982 | if (ret) | |
983 | return ret; | |
984 | ||
985 | rbd_dev->mapping.size = size; | |
986 | rbd_dev->mapping.features = features; | |
987 | ||
8b0241f8 | 988 | return 0; |
602adf40 YS |
989 | } |
990 | ||
d1cf5788 AE |
991 | static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev) |
992 | { | |
993 | rbd_dev->mapping.size = 0; | |
994 | rbd_dev->mapping.features = 0; | |
d1cf5788 AE |
995 | } |
996 | ||
98571b5a | 997 | static const char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset) |
602adf40 | 998 | { |
65ccfe21 AE |
999 | char *name; |
1000 | u64 segment; | |
1001 | int ret; | |
602adf40 | 1002 | |
78c2a44a | 1003 | name = kmem_cache_alloc(rbd_segment_name_cache, GFP_NOIO); |
65ccfe21 AE |
1004 | if (!name) |
1005 | return NULL; | |
1006 | segment = offset >> rbd_dev->header.obj_order; | |
2fd82b9e | 1007 | ret = snprintf(name, MAX_OBJ_NAME_SIZE + 1, "%s.%012llx", |
65ccfe21 | 1008 | rbd_dev->header.object_prefix, segment); |
2fd82b9e | 1009 | if (ret < 0 || ret > MAX_OBJ_NAME_SIZE) { |
65ccfe21 AE |
1010 | pr_err("error formatting segment name for #%llu (%d)\n", |
1011 | segment, ret); | |
1012 | kfree(name); | |
1013 | name = NULL; | |
1014 | } | |
602adf40 | 1015 | |
65ccfe21 AE |
1016 | return name; |
1017 | } | |
602adf40 | 1018 | |
78c2a44a AE |
1019 | static void rbd_segment_name_free(const char *name) |
1020 | { | |
1021 | /* The explicit cast here is needed to drop the const qualifier */ | |
1022 | ||
1023 | kmem_cache_free(rbd_segment_name_cache, (void *)name); | |
1024 | } | |
1025 | ||
65ccfe21 AE |
1026 | static u64 rbd_segment_offset(struct rbd_device *rbd_dev, u64 offset) |
1027 | { | |
1028 | u64 segment_size = (u64) 1 << rbd_dev->header.obj_order; | |
602adf40 | 1029 | |
65ccfe21 AE |
1030 | return offset & (segment_size - 1); |
1031 | } | |
1032 | ||
1033 | static u64 rbd_segment_length(struct rbd_device *rbd_dev, | |
1034 | u64 offset, u64 length) | |
1035 | { | |
1036 | u64 segment_size = (u64) 1 << rbd_dev->header.obj_order; | |
1037 | ||
1038 | offset &= segment_size - 1; | |
1039 | ||
aafb230e | 1040 | rbd_assert(length <= U64_MAX - offset); |
65ccfe21 AE |
1041 | if (offset + length > segment_size) |
1042 | length = segment_size - offset; | |
1043 | ||
1044 | return length; | |
602adf40 YS |
1045 | } |
1046 | ||
029bcbd8 JD |
1047 | /* |
1048 | * returns the size of an object in the image | |
1049 | */ | |
1050 | static u64 rbd_obj_bytes(struct rbd_image_header *header) | |
1051 | { | |
1052 | return 1 << header->obj_order; | |
1053 | } | |
1054 | ||
602adf40 YS |
1055 | /* |
1056 | * bio helpers | |
1057 | */ | |
1058 | ||
1059 | static void bio_chain_put(struct bio *chain) | |
1060 | { | |
1061 | struct bio *tmp; | |
1062 | ||
1063 | while (chain) { | |
1064 | tmp = chain; | |
1065 | chain = chain->bi_next; | |
1066 | bio_put(tmp); | |
1067 | } | |
1068 | } | |
1069 | ||
1070 | /* | |
1071 | * zeros a bio chain, starting at specific offset | |
1072 | */ | |
1073 | static void zero_bio_chain(struct bio *chain, int start_ofs) | |
1074 | { | |
1075 | struct bio_vec *bv; | |
1076 | unsigned long flags; | |
1077 | void *buf; | |
1078 | int i; | |
1079 | int pos = 0; | |
1080 | ||
1081 | while (chain) { | |
1082 | bio_for_each_segment(bv, chain, i) { | |
1083 | if (pos + bv->bv_len > start_ofs) { | |
1084 | int remainder = max(start_ofs - pos, 0); | |
1085 | buf = bvec_kmap_irq(bv, &flags); | |
1086 | memset(buf + remainder, 0, | |
1087 | bv->bv_len - remainder); | |
85b5aaa6 | 1088 | bvec_kunmap_irq(buf, &flags); |
602adf40 YS |
1089 | } |
1090 | pos += bv->bv_len; | |
1091 | } | |
1092 | ||
1093 | chain = chain->bi_next; | |
1094 | } | |
1095 | } | |
1096 | ||
b9434c5b AE |
1097 | /* |
1098 | * similar to zero_bio_chain(), zeros data defined by a page array, | |
1099 | * starting at the given byte offset from the start of the array and | |
1100 | * continuing up to the given end offset. The pages array is | |
1101 | * assumed to be big enough to hold all bytes up to the end. | |
1102 | */ | |
1103 | static void zero_pages(struct page **pages, u64 offset, u64 end) | |
1104 | { | |
1105 | struct page **page = &pages[offset >> PAGE_SHIFT]; | |
1106 | ||
1107 | rbd_assert(end > offset); | |
1108 | rbd_assert(end - offset <= (u64)SIZE_MAX); | |
1109 | while (offset < end) { | |
1110 | size_t page_offset; | |
1111 | size_t length; | |
1112 | unsigned long flags; | |
1113 | void *kaddr; | |
1114 | ||
1115 | page_offset = (size_t)(offset & ~PAGE_MASK); | |
1116 | length = min(PAGE_SIZE - page_offset, (size_t)(end - offset)); | |
1117 | local_irq_save(flags); | |
1118 | kaddr = kmap_atomic(*page); | |
1119 | memset(kaddr + page_offset, 0, length); | |
1120 | kunmap_atomic(kaddr); | |
1121 | local_irq_restore(flags); | |
1122 | ||
1123 | offset += length; | |
1124 | page++; | |
1125 | } | |
1126 | } | |
1127 | ||
602adf40 | 1128 | /* |
f7760dad AE |
1129 | * Clone a portion of a bio, starting at the given byte offset |
1130 | * and continuing for the number of bytes indicated. | |
602adf40 | 1131 | */ |
f7760dad AE |
1132 | static struct bio *bio_clone_range(struct bio *bio_src, |
1133 | unsigned int offset, | |
1134 | unsigned int len, | |
1135 | gfp_t gfpmask) | |
602adf40 | 1136 | { |
f7760dad AE |
1137 | struct bio_vec *bv; |
1138 | unsigned int resid; | |
1139 | unsigned short idx; | |
1140 | unsigned int voff; | |
1141 | unsigned short end_idx; | |
1142 | unsigned short vcnt; | |
1143 | struct bio *bio; | |
1144 | ||
1145 | /* Handle the easy case for the caller */ | |
1146 | ||
1147 | if (!offset && len == bio_src->bi_size) | |
1148 | return bio_clone(bio_src, gfpmask); | |
1149 | ||
1150 | if (WARN_ON_ONCE(!len)) | |
1151 | return NULL; | |
1152 | if (WARN_ON_ONCE(len > bio_src->bi_size)) | |
1153 | return NULL; | |
1154 | if (WARN_ON_ONCE(offset > bio_src->bi_size - len)) | |
1155 | return NULL; | |
1156 | ||
1157 | /* Find first affected segment... */ | |
1158 | ||
1159 | resid = offset; | |
1160 | __bio_for_each_segment(bv, bio_src, idx, 0) { | |
1161 | if (resid < bv->bv_len) | |
1162 | break; | |
1163 | resid -= bv->bv_len; | |
602adf40 | 1164 | } |
f7760dad | 1165 | voff = resid; |
602adf40 | 1166 | |
f7760dad | 1167 | /* ...and the last affected segment */ |
602adf40 | 1168 | |
f7760dad AE |
1169 | resid += len; |
1170 | __bio_for_each_segment(bv, bio_src, end_idx, idx) { | |
1171 | if (resid <= bv->bv_len) | |
1172 | break; | |
1173 | resid -= bv->bv_len; | |
1174 | } | |
1175 | vcnt = end_idx - idx + 1; | |
1176 | ||
1177 | /* Build the clone */ | |
1178 | ||
1179 | bio = bio_alloc(gfpmask, (unsigned int) vcnt); | |
1180 | if (!bio) | |
1181 | return NULL; /* ENOMEM */ | |
602adf40 | 1182 | |
f7760dad AE |
1183 | bio->bi_bdev = bio_src->bi_bdev; |
1184 | bio->bi_sector = bio_src->bi_sector + (offset >> SECTOR_SHIFT); | |
1185 | bio->bi_rw = bio_src->bi_rw; | |
1186 | bio->bi_flags |= 1 << BIO_CLONED; | |
1187 | ||
1188 | /* | |
1189 | * Copy over our part of the bio_vec, then update the first | |
1190 | * and last (or only) entries. | |
1191 | */ | |
1192 | memcpy(&bio->bi_io_vec[0], &bio_src->bi_io_vec[idx], | |
1193 | vcnt * sizeof (struct bio_vec)); | |
1194 | bio->bi_io_vec[0].bv_offset += voff; | |
1195 | if (vcnt > 1) { | |
1196 | bio->bi_io_vec[0].bv_len -= voff; | |
1197 | bio->bi_io_vec[vcnt - 1].bv_len = resid; | |
1198 | } else { | |
1199 | bio->bi_io_vec[0].bv_len = len; | |
602adf40 YS |
1200 | } |
1201 | ||
f7760dad AE |
1202 | bio->bi_vcnt = vcnt; |
1203 | bio->bi_size = len; | |
1204 | bio->bi_idx = 0; | |
1205 | ||
1206 | return bio; | |
1207 | } | |
1208 | ||
1209 | /* | |
1210 | * Clone a portion of a bio chain, starting at the given byte offset | |
1211 | * into the first bio in the source chain and continuing for the | |
1212 | * number of bytes indicated. The result is another bio chain of | |
1213 | * exactly the given length, or a null pointer on error. | |
1214 | * | |
1215 | * The bio_src and offset parameters are both in-out. On entry they | |
1216 | * refer to the first source bio and the offset into that bio where | |
1217 | * the start of data to be cloned is located. | |
1218 | * | |
1219 | * On return, bio_src is updated to refer to the bio in the source | |
1220 | * chain that contains first un-cloned byte, and *offset will | |
1221 | * contain the offset of that byte within that bio. | |
1222 | */ | |
1223 | static struct bio *bio_chain_clone_range(struct bio **bio_src, | |
1224 | unsigned int *offset, | |
1225 | unsigned int len, | |
1226 | gfp_t gfpmask) | |
1227 | { | |
1228 | struct bio *bi = *bio_src; | |
1229 | unsigned int off = *offset; | |
1230 | struct bio *chain = NULL; | |
1231 | struct bio **end; | |
1232 | ||
1233 | /* Build up a chain of clone bios up to the limit */ | |
1234 | ||
1235 | if (!bi || off >= bi->bi_size || !len) | |
1236 | return NULL; /* Nothing to clone */ | |
602adf40 | 1237 | |
f7760dad AE |
1238 | end = &chain; |
1239 | while (len) { | |
1240 | unsigned int bi_size; | |
1241 | struct bio *bio; | |
1242 | ||
f5400b7a AE |
1243 | if (!bi) { |
1244 | rbd_warn(NULL, "bio_chain exhausted with %u left", len); | |
f7760dad | 1245 | goto out_err; /* EINVAL; ran out of bio's */ |
f5400b7a | 1246 | } |
f7760dad AE |
1247 | bi_size = min_t(unsigned int, bi->bi_size - off, len); |
1248 | bio = bio_clone_range(bi, off, bi_size, gfpmask); | |
1249 | if (!bio) | |
1250 | goto out_err; /* ENOMEM */ | |
1251 | ||
1252 | *end = bio; | |
1253 | end = &bio->bi_next; | |
602adf40 | 1254 | |
f7760dad AE |
1255 | off += bi_size; |
1256 | if (off == bi->bi_size) { | |
1257 | bi = bi->bi_next; | |
1258 | off = 0; | |
1259 | } | |
1260 | len -= bi_size; | |
1261 | } | |
1262 | *bio_src = bi; | |
1263 | *offset = off; | |
1264 | ||
1265 | return chain; | |
1266 | out_err: | |
1267 | bio_chain_put(chain); | |
602adf40 | 1268 | |
602adf40 YS |
1269 | return NULL; |
1270 | } | |
1271 | ||
926f9b3f AE |
1272 | /* |
1273 | * The default/initial value for all object request flags is 0. For | |
1274 | * each flag, once its value is set to 1 it is never reset to 0 | |
1275 | * again. | |
1276 | */ | |
57acbaa7 | 1277 | static void obj_request_img_data_set(struct rbd_obj_request *obj_request) |
926f9b3f | 1278 | { |
57acbaa7 | 1279 | if (test_and_set_bit(OBJ_REQ_IMG_DATA, &obj_request->flags)) { |
926f9b3f AE |
1280 | struct rbd_device *rbd_dev; |
1281 | ||
57acbaa7 AE |
1282 | rbd_dev = obj_request->img_request->rbd_dev; |
1283 | rbd_warn(rbd_dev, "obj_request %p already marked img_data\n", | |
926f9b3f AE |
1284 | obj_request); |
1285 | } | |
1286 | } | |
1287 | ||
57acbaa7 | 1288 | static bool obj_request_img_data_test(struct rbd_obj_request *obj_request) |
926f9b3f AE |
1289 | { |
1290 | smp_mb(); | |
57acbaa7 | 1291 | return test_bit(OBJ_REQ_IMG_DATA, &obj_request->flags) != 0; |
926f9b3f AE |
1292 | } |
1293 | ||
57acbaa7 | 1294 | static void obj_request_done_set(struct rbd_obj_request *obj_request) |
6365d33a | 1295 | { |
57acbaa7 AE |
1296 | if (test_and_set_bit(OBJ_REQ_DONE, &obj_request->flags)) { |
1297 | struct rbd_device *rbd_dev = NULL; | |
6365d33a | 1298 | |
57acbaa7 AE |
1299 | if (obj_request_img_data_test(obj_request)) |
1300 | rbd_dev = obj_request->img_request->rbd_dev; | |
1301 | rbd_warn(rbd_dev, "obj_request %p already marked done\n", | |
6365d33a AE |
1302 | obj_request); |
1303 | } | |
1304 | } | |
1305 | ||
57acbaa7 | 1306 | static bool obj_request_done_test(struct rbd_obj_request *obj_request) |
6365d33a AE |
1307 | { |
1308 | smp_mb(); | |
57acbaa7 | 1309 | return test_bit(OBJ_REQ_DONE, &obj_request->flags) != 0; |
6365d33a AE |
1310 | } |
1311 | ||
5679c59f AE |
1312 | /* |
1313 | * This sets the KNOWN flag after (possibly) setting the EXISTS | |
1314 | * flag. The latter is set based on the "exists" value provided. | |
1315 | * | |
1316 | * Note that for our purposes once an object exists it never goes | |
1317 | * away again. It's possible that the response from two existence | |
1318 | * checks are separated by the creation of the target object, and | |
1319 | * the first ("doesn't exist") response arrives *after* the second | |
1320 | * ("does exist"). In that case we ignore the second one. | |
1321 | */ | |
1322 | static void obj_request_existence_set(struct rbd_obj_request *obj_request, | |
1323 | bool exists) | |
1324 | { | |
1325 | if (exists) | |
1326 | set_bit(OBJ_REQ_EXISTS, &obj_request->flags); | |
1327 | set_bit(OBJ_REQ_KNOWN, &obj_request->flags); | |
1328 | smp_mb(); | |
1329 | } | |
1330 | ||
1331 | static bool obj_request_known_test(struct rbd_obj_request *obj_request) | |
1332 | { | |
1333 | smp_mb(); | |
1334 | return test_bit(OBJ_REQ_KNOWN, &obj_request->flags) != 0; | |
1335 | } | |
1336 | ||
1337 | static bool obj_request_exists_test(struct rbd_obj_request *obj_request) | |
1338 | { | |
1339 | smp_mb(); | |
1340 | return test_bit(OBJ_REQ_EXISTS, &obj_request->flags) != 0; | |
1341 | } | |
1342 | ||
bf0d5f50 AE |
1343 | static void rbd_obj_request_get(struct rbd_obj_request *obj_request) |
1344 | { | |
37206ee5 AE |
1345 | dout("%s: obj %p (was %d)\n", __func__, obj_request, |
1346 | atomic_read(&obj_request->kref.refcount)); | |
bf0d5f50 AE |
1347 | kref_get(&obj_request->kref); |
1348 | } | |
1349 | ||
1350 | static void rbd_obj_request_destroy(struct kref *kref); | |
1351 | static void rbd_obj_request_put(struct rbd_obj_request *obj_request) | |
1352 | { | |
1353 | rbd_assert(obj_request != NULL); | |
37206ee5 AE |
1354 | dout("%s: obj %p (was %d)\n", __func__, obj_request, |
1355 | atomic_read(&obj_request->kref.refcount)); | |
bf0d5f50 AE |
1356 | kref_put(&obj_request->kref, rbd_obj_request_destroy); |
1357 | } | |
1358 | ||
1359 | static void rbd_img_request_get(struct rbd_img_request *img_request) | |
1360 | { | |
37206ee5 AE |
1361 | dout("%s: img %p (was %d)\n", __func__, img_request, |
1362 | atomic_read(&img_request->kref.refcount)); | |
bf0d5f50 AE |
1363 | kref_get(&img_request->kref); |
1364 | } | |
1365 | ||
1366 | static void rbd_img_request_destroy(struct kref *kref); | |
1367 | static void rbd_img_request_put(struct rbd_img_request *img_request) | |
1368 | { | |
1369 | rbd_assert(img_request != NULL); | |
37206ee5 AE |
1370 | dout("%s: img %p (was %d)\n", __func__, img_request, |
1371 | atomic_read(&img_request->kref.refcount)); | |
bf0d5f50 AE |
1372 | kref_put(&img_request->kref, rbd_img_request_destroy); |
1373 | } | |
1374 | ||
1375 | static inline void rbd_img_obj_request_add(struct rbd_img_request *img_request, | |
1376 | struct rbd_obj_request *obj_request) | |
1377 | { | |
25dcf954 AE |
1378 | rbd_assert(obj_request->img_request == NULL); |
1379 | ||
b155e86c | 1380 | /* Image request now owns object's original reference */ |
bf0d5f50 | 1381 | obj_request->img_request = img_request; |
25dcf954 | 1382 | obj_request->which = img_request->obj_request_count; |
6365d33a AE |
1383 | rbd_assert(!obj_request_img_data_test(obj_request)); |
1384 | obj_request_img_data_set(obj_request); | |
bf0d5f50 | 1385 | rbd_assert(obj_request->which != BAD_WHICH); |
25dcf954 AE |
1386 | img_request->obj_request_count++; |
1387 | list_add_tail(&obj_request->links, &img_request->obj_requests); | |
37206ee5 AE |
1388 | dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request, |
1389 | obj_request->which); | |
bf0d5f50 AE |
1390 | } |
1391 | ||
1392 | static inline void rbd_img_obj_request_del(struct rbd_img_request *img_request, | |
1393 | struct rbd_obj_request *obj_request) | |
1394 | { | |
1395 | rbd_assert(obj_request->which != BAD_WHICH); | |
25dcf954 | 1396 | |
37206ee5 AE |
1397 | dout("%s: img %p obj %p w=%u\n", __func__, img_request, obj_request, |
1398 | obj_request->which); | |
bf0d5f50 | 1399 | list_del(&obj_request->links); |
25dcf954 AE |
1400 | rbd_assert(img_request->obj_request_count > 0); |
1401 | img_request->obj_request_count--; | |
1402 | rbd_assert(obj_request->which == img_request->obj_request_count); | |
1403 | obj_request->which = BAD_WHICH; | |
6365d33a | 1404 | rbd_assert(obj_request_img_data_test(obj_request)); |
bf0d5f50 | 1405 | rbd_assert(obj_request->img_request == img_request); |
bf0d5f50 | 1406 | obj_request->img_request = NULL; |
25dcf954 | 1407 | obj_request->callback = NULL; |
bf0d5f50 AE |
1408 | rbd_obj_request_put(obj_request); |
1409 | } | |
1410 | ||
1411 | static bool obj_request_type_valid(enum obj_request_type type) | |
1412 | { | |
1413 | switch (type) { | |
9969ebc5 | 1414 | case OBJ_REQUEST_NODATA: |
bf0d5f50 | 1415 | case OBJ_REQUEST_BIO: |
788e2df3 | 1416 | case OBJ_REQUEST_PAGES: |
bf0d5f50 AE |
1417 | return true; |
1418 | default: | |
1419 | return false; | |
1420 | } | |
1421 | } | |
1422 | ||
bf0d5f50 AE |
1423 | static int rbd_obj_request_submit(struct ceph_osd_client *osdc, |
1424 | struct rbd_obj_request *obj_request) | |
1425 | { | |
37206ee5 AE |
1426 | dout("%s: osdc %p obj %p\n", __func__, osdc, obj_request); |
1427 | ||
bf0d5f50 AE |
1428 | return ceph_osdc_start_request(osdc, obj_request->osd_req, false); |
1429 | } | |
1430 | ||
1431 | static void rbd_img_request_complete(struct rbd_img_request *img_request) | |
1432 | { | |
55f27e09 | 1433 | |
37206ee5 | 1434 | dout("%s: img %p\n", __func__, img_request); |
55f27e09 AE |
1435 | |
1436 | /* | |
1437 | * If no error occurred, compute the aggregate transfer | |
1438 | * count for the image request. We could instead use | |
1439 | * atomic64_cmpxchg() to update it as each object request | |
1440 | * completes; not clear which way is better off hand. | |
1441 | */ | |
1442 | if (!img_request->result) { | |
1443 | struct rbd_obj_request *obj_request; | |
1444 | u64 xferred = 0; | |
1445 | ||
1446 | for_each_obj_request(img_request, obj_request) | |
1447 | xferred += obj_request->xferred; | |
1448 | img_request->xferred = xferred; | |
1449 | } | |
1450 | ||
bf0d5f50 AE |
1451 | if (img_request->callback) |
1452 | img_request->callback(img_request); | |
1453 | else | |
1454 | rbd_img_request_put(img_request); | |
1455 | } | |
1456 | ||
788e2df3 AE |
1457 | /* Caller is responsible for rbd_obj_request_destroy(obj_request) */ |
1458 | ||
1459 | static int rbd_obj_request_wait(struct rbd_obj_request *obj_request) | |
1460 | { | |
37206ee5 AE |
1461 | dout("%s: obj %p\n", __func__, obj_request); |
1462 | ||
788e2df3 AE |
1463 | return wait_for_completion_interruptible(&obj_request->completion); |
1464 | } | |
1465 | ||
0c425248 AE |
1466 | /* |
1467 | * The default/initial value for all image request flags is 0. Each | |
1468 | * is conditionally set to 1 at image request initialization time | |
1469 | * and currently never change thereafter. | |
1470 | */ | |
1471 | static void img_request_write_set(struct rbd_img_request *img_request) | |
1472 | { | |
1473 | set_bit(IMG_REQ_WRITE, &img_request->flags); | |
1474 | smp_mb(); | |
1475 | } | |
1476 | ||
1477 | static bool img_request_write_test(struct rbd_img_request *img_request) | |
1478 | { | |
1479 | smp_mb(); | |
1480 | return test_bit(IMG_REQ_WRITE, &img_request->flags) != 0; | |
1481 | } | |
1482 | ||
9849e986 AE |
1483 | static void img_request_child_set(struct rbd_img_request *img_request) |
1484 | { | |
1485 | set_bit(IMG_REQ_CHILD, &img_request->flags); | |
1486 | smp_mb(); | |
1487 | } | |
1488 | ||
1489 | static bool img_request_child_test(struct rbd_img_request *img_request) | |
1490 | { | |
1491 | smp_mb(); | |
1492 | return test_bit(IMG_REQ_CHILD, &img_request->flags) != 0; | |
1493 | } | |
1494 | ||
d0b2e944 AE |
1495 | static void img_request_layered_set(struct rbd_img_request *img_request) |
1496 | { | |
1497 | set_bit(IMG_REQ_LAYERED, &img_request->flags); | |
1498 | smp_mb(); | |
1499 | } | |
1500 | ||
1501 | static bool img_request_layered_test(struct rbd_img_request *img_request) | |
1502 | { | |
1503 | smp_mb(); | |
1504 | return test_bit(IMG_REQ_LAYERED, &img_request->flags) != 0; | |
1505 | } | |
1506 | ||
6e2a4505 AE |
1507 | static void |
1508 | rbd_img_obj_request_read_callback(struct rbd_obj_request *obj_request) | |
1509 | { | |
b9434c5b AE |
1510 | u64 xferred = obj_request->xferred; |
1511 | u64 length = obj_request->length; | |
1512 | ||
6e2a4505 AE |
1513 | dout("%s: obj %p img %p result %d %llu/%llu\n", __func__, |
1514 | obj_request, obj_request->img_request, obj_request->result, | |
b9434c5b | 1515 | xferred, length); |
6e2a4505 AE |
1516 | /* |
1517 | * ENOENT means a hole in the image. We zero-fill the | |
1518 | * entire length of the request. A short read also implies | |
1519 | * zero-fill to the end of the request. Either way we | |
1520 | * update the xferred count to indicate the whole request | |
1521 | * was satisfied. | |
1522 | */ | |
b9434c5b | 1523 | rbd_assert(obj_request->type != OBJ_REQUEST_NODATA); |
6e2a4505 | 1524 | if (obj_request->result == -ENOENT) { |
b9434c5b AE |
1525 | if (obj_request->type == OBJ_REQUEST_BIO) |
1526 | zero_bio_chain(obj_request->bio_list, 0); | |
1527 | else | |
1528 | zero_pages(obj_request->pages, 0, length); | |
6e2a4505 | 1529 | obj_request->result = 0; |
b9434c5b AE |
1530 | obj_request->xferred = length; |
1531 | } else if (xferred < length && !obj_request->result) { | |
1532 | if (obj_request->type == OBJ_REQUEST_BIO) | |
1533 | zero_bio_chain(obj_request->bio_list, xferred); | |
1534 | else | |
1535 | zero_pages(obj_request->pages, xferred, length); | |
1536 | obj_request->xferred = length; | |
6e2a4505 AE |
1537 | } |
1538 | obj_request_done_set(obj_request); | |
1539 | } | |
1540 | ||
bf0d5f50 AE |
1541 | static void rbd_obj_request_complete(struct rbd_obj_request *obj_request) |
1542 | { | |
37206ee5 AE |
1543 | dout("%s: obj %p cb %p\n", __func__, obj_request, |
1544 | obj_request->callback); | |
bf0d5f50 AE |
1545 | if (obj_request->callback) |
1546 | obj_request->callback(obj_request); | |
788e2df3 AE |
1547 | else |
1548 | complete_all(&obj_request->completion); | |
bf0d5f50 AE |
1549 | } |
1550 | ||
c47f9371 | 1551 | static void rbd_osd_trivial_callback(struct rbd_obj_request *obj_request) |
39bf2c5d AE |
1552 | { |
1553 | dout("%s: obj %p\n", __func__, obj_request); | |
1554 | obj_request_done_set(obj_request); | |
1555 | } | |
1556 | ||
c47f9371 | 1557 | static void rbd_osd_read_callback(struct rbd_obj_request *obj_request) |
bf0d5f50 | 1558 | { |
57acbaa7 | 1559 | struct rbd_img_request *img_request = NULL; |
a9e8ba2c | 1560 | struct rbd_device *rbd_dev = NULL; |
57acbaa7 AE |
1561 | bool layered = false; |
1562 | ||
1563 | if (obj_request_img_data_test(obj_request)) { | |
1564 | img_request = obj_request->img_request; | |
1565 | layered = img_request && img_request_layered_test(img_request); | |
a9e8ba2c | 1566 | rbd_dev = img_request->rbd_dev; |
57acbaa7 | 1567 | } |
8b3e1a56 AE |
1568 | |
1569 | dout("%s: obj %p img %p result %d %llu/%llu\n", __func__, | |
1570 | obj_request, img_request, obj_request->result, | |
1571 | obj_request->xferred, obj_request->length); | |
a9e8ba2c AE |
1572 | if (layered && obj_request->result == -ENOENT && |
1573 | obj_request->img_offset < rbd_dev->parent_overlap) | |
8b3e1a56 AE |
1574 | rbd_img_parent_read(obj_request); |
1575 | else if (img_request) | |
6e2a4505 AE |
1576 | rbd_img_obj_request_read_callback(obj_request); |
1577 | else | |
1578 | obj_request_done_set(obj_request); | |
bf0d5f50 AE |
1579 | } |
1580 | ||
c47f9371 | 1581 | static void rbd_osd_write_callback(struct rbd_obj_request *obj_request) |
bf0d5f50 | 1582 | { |
1b83bef2 SW |
1583 | dout("%s: obj %p result %d %llu\n", __func__, obj_request, |
1584 | obj_request->result, obj_request->length); | |
1585 | /* | |
8b3e1a56 AE |
1586 | * There is no such thing as a successful short write. Set |
1587 | * it to our originally-requested length. | |
1b83bef2 SW |
1588 | */ |
1589 | obj_request->xferred = obj_request->length; | |
07741308 | 1590 | obj_request_done_set(obj_request); |
bf0d5f50 AE |
1591 | } |
1592 | ||
fbfab539 AE |
1593 | /* |
1594 | * For a simple stat call there's nothing to do. We'll do more if | |
1595 | * this is part of a write sequence for a layered image. | |
1596 | */ | |
c47f9371 | 1597 | static void rbd_osd_stat_callback(struct rbd_obj_request *obj_request) |
fbfab539 | 1598 | { |
37206ee5 | 1599 | dout("%s: obj %p\n", __func__, obj_request); |
fbfab539 AE |
1600 | obj_request_done_set(obj_request); |
1601 | } | |
1602 | ||
bf0d5f50 AE |
1603 | static void rbd_osd_req_callback(struct ceph_osd_request *osd_req, |
1604 | struct ceph_msg *msg) | |
1605 | { | |
1606 | struct rbd_obj_request *obj_request = osd_req->r_priv; | |
bf0d5f50 AE |
1607 | u16 opcode; |
1608 | ||
37206ee5 | 1609 | dout("%s: osd_req %p msg %p\n", __func__, osd_req, msg); |
bf0d5f50 | 1610 | rbd_assert(osd_req == obj_request->osd_req); |
57acbaa7 AE |
1611 | if (obj_request_img_data_test(obj_request)) { |
1612 | rbd_assert(obj_request->img_request); | |
1613 | rbd_assert(obj_request->which != BAD_WHICH); | |
1614 | } else { | |
1615 | rbd_assert(obj_request->which == BAD_WHICH); | |
1616 | } | |
bf0d5f50 | 1617 | |
1b83bef2 SW |
1618 | if (osd_req->r_result < 0) |
1619 | obj_request->result = osd_req->r_result; | |
bf0d5f50 | 1620 | |
0eefd470 | 1621 | BUG_ON(osd_req->r_num_ops > 2); |
bf0d5f50 | 1622 | |
c47f9371 AE |
1623 | /* |
1624 | * We support a 64-bit length, but ultimately it has to be | |
1625 | * passed to blk_end_request(), which takes an unsigned int. | |
1626 | */ | |
1b83bef2 | 1627 | obj_request->xferred = osd_req->r_reply_op_len[0]; |
8b3e1a56 | 1628 | rbd_assert(obj_request->xferred < (u64)UINT_MAX); |
79528734 | 1629 | opcode = osd_req->r_ops[0].op; |
bf0d5f50 AE |
1630 | switch (opcode) { |
1631 | case CEPH_OSD_OP_READ: | |
c47f9371 | 1632 | rbd_osd_read_callback(obj_request); |
bf0d5f50 AE |
1633 | break; |
1634 | case CEPH_OSD_OP_WRITE: | |
c47f9371 | 1635 | rbd_osd_write_callback(obj_request); |
bf0d5f50 | 1636 | break; |
fbfab539 | 1637 | case CEPH_OSD_OP_STAT: |
c47f9371 | 1638 | rbd_osd_stat_callback(obj_request); |
fbfab539 | 1639 | break; |
36be9a76 | 1640 | case CEPH_OSD_OP_CALL: |
b8d70035 | 1641 | case CEPH_OSD_OP_NOTIFY_ACK: |
9969ebc5 | 1642 | case CEPH_OSD_OP_WATCH: |
c47f9371 | 1643 | rbd_osd_trivial_callback(obj_request); |
9969ebc5 | 1644 | break; |
bf0d5f50 AE |
1645 | default: |
1646 | rbd_warn(NULL, "%s: unsupported op %hu\n", | |
1647 | obj_request->object_name, (unsigned short) opcode); | |
1648 | break; | |
1649 | } | |
1650 | ||
07741308 | 1651 | if (obj_request_done_test(obj_request)) |
bf0d5f50 AE |
1652 | rbd_obj_request_complete(obj_request); |
1653 | } | |
1654 | ||
9d4df01f | 1655 | static void rbd_osd_req_format_read(struct rbd_obj_request *obj_request) |
430c28c3 AE |
1656 | { |
1657 | struct rbd_img_request *img_request = obj_request->img_request; | |
8c042b0d | 1658 | struct ceph_osd_request *osd_req = obj_request->osd_req; |
9d4df01f | 1659 | u64 snap_id; |
430c28c3 | 1660 | |
8c042b0d | 1661 | rbd_assert(osd_req != NULL); |
430c28c3 | 1662 | |
9d4df01f | 1663 | snap_id = img_request ? img_request->snap_id : CEPH_NOSNAP; |
8c042b0d | 1664 | ceph_osdc_build_request(osd_req, obj_request->offset, |
9d4df01f AE |
1665 | NULL, snap_id, NULL); |
1666 | } | |
1667 | ||
1668 | static void rbd_osd_req_format_write(struct rbd_obj_request *obj_request) | |
1669 | { | |
1670 | struct rbd_img_request *img_request = obj_request->img_request; | |
1671 | struct ceph_osd_request *osd_req = obj_request->osd_req; | |
1672 | struct ceph_snap_context *snapc; | |
1673 | struct timespec mtime = CURRENT_TIME; | |
1674 | ||
1675 | rbd_assert(osd_req != NULL); | |
1676 | ||
1677 | snapc = img_request ? img_request->snapc : NULL; | |
1678 | ceph_osdc_build_request(osd_req, obj_request->offset, | |
1679 | snapc, CEPH_NOSNAP, &mtime); | |
430c28c3 AE |
1680 | } |
1681 | ||
bf0d5f50 AE |
1682 | static struct ceph_osd_request *rbd_osd_req_create( |
1683 | struct rbd_device *rbd_dev, | |
1684 | bool write_request, | |
430c28c3 | 1685 | struct rbd_obj_request *obj_request) |
bf0d5f50 | 1686 | { |
bf0d5f50 AE |
1687 | struct ceph_snap_context *snapc = NULL; |
1688 | struct ceph_osd_client *osdc; | |
1689 | struct ceph_osd_request *osd_req; | |
bf0d5f50 | 1690 | |
6365d33a AE |
1691 | if (obj_request_img_data_test(obj_request)) { |
1692 | struct rbd_img_request *img_request = obj_request->img_request; | |
1693 | ||
0c425248 AE |
1694 | rbd_assert(write_request == |
1695 | img_request_write_test(img_request)); | |
1696 | if (write_request) | |
bf0d5f50 | 1697 | snapc = img_request->snapc; |
bf0d5f50 AE |
1698 | } |
1699 | ||
1700 | /* Allocate and initialize the request, for the single op */ | |
1701 | ||
1702 | osdc = &rbd_dev->rbd_client->client->osdc; | |
1703 | osd_req = ceph_osdc_alloc_request(osdc, snapc, 1, false, GFP_ATOMIC); | |
1704 | if (!osd_req) | |
1705 | return NULL; /* ENOMEM */ | |
bf0d5f50 | 1706 | |
430c28c3 | 1707 | if (write_request) |
bf0d5f50 | 1708 | osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK; |
430c28c3 | 1709 | else |
bf0d5f50 | 1710 | osd_req->r_flags = CEPH_OSD_FLAG_READ; |
bf0d5f50 AE |
1711 | |
1712 | osd_req->r_callback = rbd_osd_req_callback; | |
1713 | osd_req->r_priv = obj_request; | |
1714 | ||
1715 | osd_req->r_oid_len = strlen(obj_request->object_name); | |
1716 | rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid)); | |
1717 | memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len); | |
1718 | ||
1719 | osd_req->r_file_layout = rbd_dev->layout; /* struct */ | |
1720 | ||
bf0d5f50 AE |
1721 | return osd_req; |
1722 | } | |
1723 | ||
0eefd470 AE |
1724 | /* |
1725 | * Create a copyup osd request based on the information in the | |
1726 | * object request supplied. A copyup request has two osd ops, | |
1727 | * a copyup method call, and a "normal" write request. | |
1728 | */ | |
1729 | static struct ceph_osd_request * | |
1730 | rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) | |
1731 | { | |
1732 | struct rbd_img_request *img_request; | |
1733 | struct ceph_snap_context *snapc; | |
1734 | struct rbd_device *rbd_dev; | |
1735 | struct ceph_osd_client *osdc; | |
1736 | struct ceph_osd_request *osd_req; | |
1737 | ||
1738 | rbd_assert(obj_request_img_data_test(obj_request)); | |
1739 | img_request = obj_request->img_request; | |
1740 | rbd_assert(img_request); | |
1741 | rbd_assert(img_request_write_test(img_request)); | |
1742 | ||
1743 | /* Allocate and initialize the request, for the two ops */ | |
1744 | ||
1745 | snapc = img_request->snapc; | |
1746 | rbd_dev = img_request->rbd_dev; | |
1747 | osdc = &rbd_dev->rbd_client->client->osdc; | |
1748 | osd_req = ceph_osdc_alloc_request(osdc, snapc, 2, false, GFP_ATOMIC); | |
1749 | if (!osd_req) | |
1750 | return NULL; /* ENOMEM */ | |
1751 | ||
1752 | osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK; | |
1753 | osd_req->r_callback = rbd_osd_req_callback; | |
1754 | osd_req->r_priv = obj_request; | |
1755 | ||
1756 | osd_req->r_oid_len = strlen(obj_request->object_name); | |
1757 | rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid)); | |
1758 | memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len); | |
1759 | ||
1760 | osd_req->r_file_layout = rbd_dev->layout; /* struct */ | |
1761 | ||
1762 | return osd_req; | |
1763 | } | |
1764 | ||
1765 | ||
bf0d5f50 AE |
1766 | static void rbd_osd_req_destroy(struct ceph_osd_request *osd_req) |
1767 | { | |
1768 | ceph_osdc_put_request(osd_req); | |
1769 | } | |
1770 | ||
1771 | /* object_name is assumed to be a non-null pointer and NUL-terminated */ | |
1772 | ||
1773 | static struct rbd_obj_request *rbd_obj_request_create(const char *object_name, | |
1774 | u64 offset, u64 length, | |
1775 | enum obj_request_type type) | |
1776 | { | |
1777 | struct rbd_obj_request *obj_request; | |
1778 | size_t size; | |
1779 | char *name; | |
1780 | ||
1781 | rbd_assert(obj_request_type_valid(type)); | |
1782 | ||
1783 | size = strlen(object_name) + 1; | |
f907ad55 AE |
1784 | name = kmalloc(size, GFP_KERNEL); |
1785 | if (!name) | |
bf0d5f50 AE |
1786 | return NULL; |
1787 | ||
868311b1 | 1788 | obj_request = kmem_cache_zalloc(rbd_obj_request_cache, GFP_KERNEL); |
f907ad55 AE |
1789 | if (!obj_request) { |
1790 | kfree(name); | |
1791 | return NULL; | |
1792 | } | |
1793 | ||
bf0d5f50 AE |
1794 | obj_request->object_name = memcpy(name, object_name, size); |
1795 | obj_request->offset = offset; | |
1796 | obj_request->length = length; | |
926f9b3f | 1797 | obj_request->flags = 0; |
bf0d5f50 AE |
1798 | obj_request->which = BAD_WHICH; |
1799 | obj_request->type = type; | |
1800 | INIT_LIST_HEAD(&obj_request->links); | |
788e2df3 | 1801 | init_completion(&obj_request->completion); |
bf0d5f50 AE |
1802 | kref_init(&obj_request->kref); |
1803 | ||
37206ee5 AE |
1804 | dout("%s: \"%s\" %llu/%llu %d -> obj %p\n", __func__, object_name, |
1805 | offset, length, (int)type, obj_request); | |
1806 | ||
bf0d5f50 AE |
1807 | return obj_request; |
1808 | } | |
1809 | ||
1810 | static void rbd_obj_request_destroy(struct kref *kref) | |
1811 | { | |
1812 | struct rbd_obj_request *obj_request; | |
1813 | ||
1814 | obj_request = container_of(kref, struct rbd_obj_request, kref); | |
1815 | ||
37206ee5 AE |
1816 | dout("%s: obj %p\n", __func__, obj_request); |
1817 | ||
bf0d5f50 AE |
1818 | rbd_assert(obj_request->img_request == NULL); |
1819 | rbd_assert(obj_request->which == BAD_WHICH); | |
1820 | ||
1821 | if (obj_request->osd_req) | |
1822 | rbd_osd_req_destroy(obj_request->osd_req); | |
1823 | ||
1824 | rbd_assert(obj_request_type_valid(obj_request->type)); | |
1825 | switch (obj_request->type) { | |
9969ebc5 AE |
1826 | case OBJ_REQUEST_NODATA: |
1827 | break; /* Nothing to do */ | |
bf0d5f50 AE |
1828 | case OBJ_REQUEST_BIO: |
1829 | if (obj_request->bio_list) | |
1830 | bio_chain_put(obj_request->bio_list); | |
1831 | break; | |
788e2df3 AE |
1832 | case OBJ_REQUEST_PAGES: |
1833 | if (obj_request->pages) | |
1834 | ceph_release_page_vector(obj_request->pages, | |
1835 | obj_request->page_count); | |
1836 | break; | |
bf0d5f50 AE |
1837 | } |
1838 | ||
f907ad55 | 1839 | kfree(obj_request->object_name); |
868311b1 AE |
1840 | obj_request->object_name = NULL; |
1841 | kmem_cache_free(rbd_obj_request_cache, obj_request); | |
bf0d5f50 AE |
1842 | } |
1843 | ||
1844 | /* | |
1845 | * Caller is responsible for filling in the list of object requests | |
1846 | * that comprises the image request, and the Linux request pointer | |
1847 | * (if there is one). | |
1848 | */ | |
cc344fa1 AE |
1849 | static struct rbd_img_request *rbd_img_request_create( |
1850 | struct rbd_device *rbd_dev, | |
bf0d5f50 | 1851 | u64 offset, u64 length, |
9849e986 AE |
1852 | bool write_request, |
1853 | bool child_request) | |
bf0d5f50 AE |
1854 | { |
1855 | struct rbd_img_request *img_request; | |
bf0d5f50 | 1856 | |
1c2a9dfe | 1857 | img_request = kmem_cache_alloc(rbd_img_request_cache, GFP_ATOMIC); |
bf0d5f50 AE |
1858 | if (!img_request) |
1859 | return NULL; | |
1860 | ||
1861 | if (write_request) { | |
1862 | down_read(&rbd_dev->header_rwsem); | |
812164f8 | 1863 | ceph_get_snap_context(rbd_dev->header.snapc); |
bf0d5f50 | 1864 | up_read(&rbd_dev->header_rwsem); |
bf0d5f50 AE |
1865 | } |
1866 | ||
1867 | img_request->rq = NULL; | |
1868 | img_request->rbd_dev = rbd_dev; | |
1869 | img_request->offset = offset; | |
1870 | img_request->length = length; | |
0c425248 AE |
1871 | img_request->flags = 0; |
1872 | if (write_request) { | |
1873 | img_request_write_set(img_request); | |
468521c1 | 1874 | img_request->snapc = rbd_dev->header.snapc; |
0c425248 | 1875 | } else { |
bf0d5f50 | 1876 | img_request->snap_id = rbd_dev->spec->snap_id; |
0c425248 | 1877 | } |
9849e986 AE |
1878 | if (child_request) |
1879 | img_request_child_set(img_request); | |
d0b2e944 AE |
1880 | if (rbd_dev->parent_spec) |
1881 | img_request_layered_set(img_request); | |
bf0d5f50 AE |
1882 | spin_lock_init(&img_request->completion_lock); |
1883 | img_request->next_completion = 0; | |
1884 | img_request->callback = NULL; | |
a5a337d4 | 1885 | img_request->result = 0; |
bf0d5f50 AE |
1886 | img_request->obj_request_count = 0; |
1887 | INIT_LIST_HEAD(&img_request->obj_requests); | |
1888 | kref_init(&img_request->kref); | |
1889 | ||
1890 | rbd_img_request_get(img_request); /* Avoid a warning */ | |
1891 | rbd_img_request_put(img_request); /* TEMPORARY */ | |
1892 | ||
37206ee5 AE |
1893 | dout("%s: rbd_dev %p %s %llu/%llu -> img %p\n", __func__, rbd_dev, |
1894 | write_request ? "write" : "read", offset, length, | |
1895 | img_request); | |
1896 | ||
bf0d5f50 AE |
1897 | return img_request; |
1898 | } | |
1899 | ||
1900 | static void rbd_img_request_destroy(struct kref *kref) | |
1901 | { | |
1902 | struct rbd_img_request *img_request; | |
1903 | struct rbd_obj_request *obj_request; | |
1904 | struct rbd_obj_request *next_obj_request; | |
1905 | ||
1906 | img_request = container_of(kref, struct rbd_img_request, kref); | |
1907 | ||
37206ee5 AE |
1908 | dout("%s: img %p\n", __func__, img_request); |
1909 | ||
bf0d5f50 AE |
1910 | for_each_obj_request_safe(img_request, obj_request, next_obj_request) |
1911 | rbd_img_obj_request_del(img_request, obj_request); | |
25dcf954 | 1912 | rbd_assert(img_request->obj_request_count == 0); |
bf0d5f50 | 1913 | |
0c425248 | 1914 | if (img_request_write_test(img_request)) |
812164f8 | 1915 | ceph_put_snap_context(img_request->snapc); |
bf0d5f50 | 1916 | |
8b3e1a56 AE |
1917 | if (img_request_child_test(img_request)) |
1918 | rbd_obj_request_put(img_request->obj_request); | |
1919 | ||
1c2a9dfe | 1920 | kmem_cache_free(rbd_img_request_cache, img_request); |
bf0d5f50 AE |
1921 | } |
1922 | ||
1217857f AE |
1923 | static bool rbd_img_obj_end_request(struct rbd_obj_request *obj_request) |
1924 | { | |
6365d33a | 1925 | struct rbd_img_request *img_request; |
1217857f AE |
1926 | unsigned int xferred; |
1927 | int result; | |
8b3e1a56 | 1928 | bool more; |
1217857f | 1929 | |
6365d33a AE |
1930 | rbd_assert(obj_request_img_data_test(obj_request)); |
1931 | img_request = obj_request->img_request; | |
1932 | ||
1217857f AE |
1933 | rbd_assert(obj_request->xferred <= (u64)UINT_MAX); |
1934 | xferred = (unsigned int)obj_request->xferred; | |
1935 | result = obj_request->result; | |
1936 | if (result) { | |
1937 | struct rbd_device *rbd_dev = img_request->rbd_dev; | |
1938 | ||
1939 | rbd_warn(rbd_dev, "%s %llx at %llx (%llx)\n", | |
1940 | img_request_write_test(img_request) ? "write" : "read", | |
1941 | obj_request->length, obj_request->img_offset, | |
1942 | obj_request->offset); | |
1943 | rbd_warn(rbd_dev, " result %d xferred %x\n", | |
1944 | result, xferred); | |
1945 | if (!img_request->result) | |
1946 | img_request->result = result; | |
1947 | } | |
1948 | ||
f1a4739f AE |
1949 | /* Image object requests don't own their page array */ |
1950 | ||
1951 | if (obj_request->type == OBJ_REQUEST_PAGES) { | |
1952 | obj_request->pages = NULL; | |
1953 | obj_request->page_count = 0; | |
1954 | } | |
1955 | ||
8b3e1a56 AE |
1956 | if (img_request_child_test(img_request)) { |
1957 | rbd_assert(img_request->obj_request != NULL); | |
1958 | more = obj_request->which < img_request->obj_request_count - 1; | |
1959 | } else { | |
1960 | rbd_assert(img_request->rq != NULL); | |
1961 | more = blk_end_request(img_request->rq, result, xferred); | |
1962 | } | |
1963 | ||
1964 | return more; | |
1217857f AE |
1965 | } |
1966 | ||
2169238d AE |
1967 | static void rbd_img_obj_callback(struct rbd_obj_request *obj_request) |
1968 | { | |
1969 | struct rbd_img_request *img_request; | |
1970 | u32 which = obj_request->which; | |
1971 | bool more = true; | |
1972 | ||
6365d33a | 1973 | rbd_assert(obj_request_img_data_test(obj_request)); |
2169238d AE |
1974 | img_request = obj_request->img_request; |
1975 | ||
1976 | dout("%s: img %p obj %p\n", __func__, img_request, obj_request); | |
1977 | rbd_assert(img_request != NULL); | |
2169238d AE |
1978 | rbd_assert(img_request->obj_request_count > 0); |
1979 | rbd_assert(which != BAD_WHICH); | |
1980 | rbd_assert(which < img_request->obj_request_count); | |
1981 | rbd_assert(which >= img_request->next_completion); | |
1982 | ||
1983 | spin_lock_irq(&img_request->completion_lock); | |
1984 | if (which != img_request->next_completion) | |
1985 | goto out; | |
1986 | ||
1987 | for_each_obj_request_from(img_request, obj_request) { | |
2169238d AE |
1988 | rbd_assert(more); |
1989 | rbd_assert(which < img_request->obj_request_count); | |
1990 | ||
1991 | if (!obj_request_done_test(obj_request)) | |
1992 | break; | |
1217857f | 1993 | more = rbd_img_obj_end_request(obj_request); |
2169238d AE |
1994 | which++; |
1995 | } | |
1996 | ||
1997 | rbd_assert(more ^ (which == img_request->obj_request_count)); | |
1998 | img_request->next_completion = which; | |
1999 | out: | |
2000 | spin_unlock_irq(&img_request->completion_lock); | |
2001 | ||
2002 | if (!more) | |
2003 | rbd_img_request_complete(img_request); | |
2004 | } | |
2005 | ||
f1a4739f AE |
2006 | /* |
2007 | * Split up an image request into one or more object requests, each | |
2008 | * to a different object. The "type" parameter indicates whether | |
2009 | * "data_desc" is the pointer to the head of a list of bio | |
2010 | * structures, or the base of a page array. In either case this | |
2011 | * function assumes data_desc describes memory sufficient to hold | |
2012 | * all data described by the image request. | |
2013 | */ | |
2014 | static int rbd_img_request_fill(struct rbd_img_request *img_request, | |
2015 | enum obj_request_type type, | |
2016 | void *data_desc) | |
bf0d5f50 AE |
2017 | { |
2018 | struct rbd_device *rbd_dev = img_request->rbd_dev; | |
2019 | struct rbd_obj_request *obj_request = NULL; | |
2020 | struct rbd_obj_request *next_obj_request; | |
0c425248 | 2021 | bool write_request = img_request_write_test(img_request); |
f1a4739f AE |
2022 | struct bio *bio_list; |
2023 | unsigned int bio_offset = 0; | |
2024 | struct page **pages; | |
7da22d29 | 2025 | u64 img_offset; |
bf0d5f50 AE |
2026 | u64 resid; |
2027 | u16 opcode; | |
2028 | ||
f1a4739f AE |
2029 | dout("%s: img %p type %d data_desc %p\n", __func__, img_request, |
2030 | (int)type, data_desc); | |
37206ee5 | 2031 | |
430c28c3 | 2032 | opcode = write_request ? CEPH_OSD_OP_WRITE : CEPH_OSD_OP_READ; |
7da22d29 | 2033 | img_offset = img_request->offset; |
bf0d5f50 | 2034 | resid = img_request->length; |
4dda41d3 | 2035 | rbd_assert(resid > 0); |
f1a4739f AE |
2036 | |
2037 | if (type == OBJ_REQUEST_BIO) { | |
2038 | bio_list = data_desc; | |
2039 | rbd_assert(img_offset == bio_list->bi_sector << SECTOR_SHIFT); | |
2040 | } else { | |
2041 | rbd_assert(type == OBJ_REQUEST_PAGES); | |
2042 | pages = data_desc; | |
2043 | } | |
2044 | ||
bf0d5f50 | 2045 | while (resid) { |
2fa12320 | 2046 | struct ceph_osd_request *osd_req; |
bf0d5f50 | 2047 | const char *object_name; |
bf0d5f50 AE |
2048 | u64 offset; |
2049 | u64 length; | |
2050 | ||
7da22d29 | 2051 | object_name = rbd_segment_name(rbd_dev, img_offset); |
bf0d5f50 AE |
2052 | if (!object_name) |
2053 | goto out_unwind; | |
7da22d29 AE |
2054 | offset = rbd_segment_offset(rbd_dev, img_offset); |
2055 | length = rbd_segment_length(rbd_dev, img_offset, resid); | |
bf0d5f50 | 2056 | obj_request = rbd_obj_request_create(object_name, |
f1a4739f | 2057 | offset, length, type); |
78c2a44a AE |
2058 | /* object request has its own copy of the object name */ |
2059 | rbd_segment_name_free(object_name); | |
bf0d5f50 AE |
2060 | if (!obj_request) |
2061 | goto out_unwind; | |
2062 | ||
f1a4739f AE |
2063 | if (type == OBJ_REQUEST_BIO) { |
2064 | unsigned int clone_size; | |
2065 | ||
2066 | rbd_assert(length <= (u64)UINT_MAX); | |
2067 | clone_size = (unsigned int)length; | |
2068 | obj_request->bio_list = | |
2069 | bio_chain_clone_range(&bio_list, | |
2070 | &bio_offset, | |
2071 | clone_size, | |
2072 | GFP_ATOMIC); | |
2073 | if (!obj_request->bio_list) | |
2074 | goto out_partial; | |
2075 | } else { | |
2076 | unsigned int page_count; | |
2077 | ||
2078 | obj_request->pages = pages; | |
2079 | page_count = (u32)calc_pages_for(offset, length); | |
2080 | obj_request->page_count = page_count; | |
2081 | if ((offset + length) & ~PAGE_MASK) | |
2082 | page_count--; /* more on last page */ | |
2083 | pages += page_count; | |
2084 | } | |
bf0d5f50 | 2085 | |
2fa12320 AE |
2086 | osd_req = rbd_osd_req_create(rbd_dev, write_request, |
2087 | obj_request); | |
2088 | if (!osd_req) | |
bf0d5f50 | 2089 | goto out_partial; |
2fa12320 | 2090 | obj_request->osd_req = osd_req; |
2169238d | 2091 | obj_request->callback = rbd_img_obj_callback; |
430c28c3 | 2092 | |
2fa12320 AE |
2093 | osd_req_op_extent_init(osd_req, 0, opcode, offset, length, |
2094 | 0, 0); | |
f1a4739f AE |
2095 | if (type == OBJ_REQUEST_BIO) |
2096 | osd_req_op_extent_osd_data_bio(osd_req, 0, | |
2097 | obj_request->bio_list, length); | |
2098 | else | |
2099 | osd_req_op_extent_osd_data_pages(osd_req, 0, | |
2100 | obj_request->pages, length, | |
2101 | offset & ~PAGE_MASK, false, false); | |
9d4df01f AE |
2102 | |
2103 | if (write_request) | |
2104 | rbd_osd_req_format_write(obj_request); | |
2105 | else | |
2106 | rbd_osd_req_format_read(obj_request); | |
430c28c3 | 2107 | |
7da22d29 | 2108 | obj_request->img_offset = img_offset; |
bf0d5f50 AE |
2109 | rbd_img_obj_request_add(img_request, obj_request); |
2110 | ||
7da22d29 | 2111 | img_offset += length; |
bf0d5f50 AE |
2112 | resid -= length; |
2113 | } | |
2114 | ||
2115 | return 0; | |
2116 | ||
2117 | out_partial: | |
2118 | rbd_obj_request_put(obj_request); | |
2119 | out_unwind: | |
2120 | for_each_obj_request_safe(img_request, obj_request, next_obj_request) | |
2121 | rbd_obj_request_put(obj_request); | |
2122 | ||
2123 | return -ENOMEM; | |
2124 | } | |
2125 | ||
0eefd470 AE |
2126 | static void |
2127 | rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request) | |
2128 | { | |
2129 | struct rbd_img_request *img_request; | |
2130 | struct rbd_device *rbd_dev; | |
2131 | u64 length; | |
2132 | u32 page_count; | |
2133 | ||
2134 | rbd_assert(obj_request->type == OBJ_REQUEST_BIO); | |
2135 | rbd_assert(obj_request_img_data_test(obj_request)); | |
2136 | img_request = obj_request->img_request; | |
2137 | rbd_assert(img_request); | |
2138 | ||
2139 | rbd_dev = img_request->rbd_dev; | |
2140 | rbd_assert(rbd_dev); | |
2141 | length = (u64)1 << rbd_dev->header.obj_order; | |
2142 | page_count = (u32)calc_pages_for(0, length); | |
2143 | ||
2144 | rbd_assert(obj_request->copyup_pages); | |
2145 | ceph_release_page_vector(obj_request->copyup_pages, page_count); | |
2146 | obj_request->copyup_pages = NULL; | |
2147 | ||
2148 | /* | |
2149 | * We want the transfer count to reflect the size of the | |
2150 | * original write request. There is no such thing as a | |
2151 | * successful short write, so if the request was successful | |
2152 | * we can just set it to the originally-requested length. | |
2153 | */ | |
2154 | if (!obj_request->result) | |
2155 | obj_request->xferred = obj_request->length; | |
2156 | ||
2157 | /* Finish up with the normal image object callback */ | |
2158 | ||
2159 | rbd_img_obj_callback(obj_request); | |
2160 | } | |
2161 | ||
3d7efd18 AE |
2162 | static void |
2163 | rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request) | |
2164 | { | |
2165 | struct rbd_obj_request *orig_request; | |
0eefd470 AE |
2166 | struct ceph_osd_request *osd_req; |
2167 | struct ceph_osd_client *osdc; | |
2168 | struct rbd_device *rbd_dev; | |
3d7efd18 | 2169 | struct page **pages; |
3d7efd18 AE |
2170 | int result; |
2171 | u64 obj_size; | |
2172 | u64 xferred; | |
2173 | ||
2174 | rbd_assert(img_request_child_test(img_request)); | |
2175 | ||
2176 | /* First get what we need from the image request */ | |
2177 | ||
2178 | pages = img_request->copyup_pages; | |
2179 | rbd_assert(pages != NULL); | |
2180 | img_request->copyup_pages = NULL; | |
2181 | ||
2182 | orig_request = img_request->obj_request; | |
2183 | rbd_assert(orig_request != NULL); | |
0eefd470 | 2184 | rbd_assert(orig_request->type == OBJ_REQUEST_BIO); |
3d7efd18 AE |
2185 | result = img_request->result; |
2186 | obj_size = img_request->length; | |
2187 | xferred = img_request->xferred; | |
2188 | ||
0eefd470 AE |
2189 | rbd_dev = img_request->rbd_dev; |
2190 | rbd_assert(rbd_dev); | |
2191 | rbd_assert(obj_size == (u64)1 << rbd_dev->header.obj_order); | |
2192 | ||
3d7efd18 AE |
2193 | rbd_img_request_put(img_request); |
2194 | ||
0eefd470 AE |
2195 | if (result) |
2196 | goto out_err; | |
2197 | ||
2198 | /* Allocate the new copyup osd request for the original request */ | |
2199 | ||
2200 | result = -ENOMEM; | |
2201 | rbd_assert(!orig_request->osd_req); | |
2202 | osd_req = rbd_osd_req_create_copyup(orig_request); | |
2203 | if (!osd_req) | |
2204 | goto out_err; | |
2205 | orig_request->osd_req = osd_req; | |
2206 | orig_request->copyup_pages = pages; | |
3d7efd18 | 2207 | |
0eefd470 | 2208 | /* Initialize the copyup op */ |
3d7efd18 | 2209 | |
0eefd470 AE |
2210 | osd_req_op_cls_init(osd_req, 0, CEPH_OSD_OP_CALL, "rbd", "copyup"); |
2211 | osd_req_op_cls_request_data_pages(osd_req, 0, pages, obj_size, 0, | |
2212 | false, false); | |
3d7efd18 | 2213 | |
0eefd470 AE |
2214 | /* Then the original write request op */ |
2215 | ||
2216 | osd_req_op_extent_init(osd_req, 1, CEPH_OSD_OP_WRITE, | |
2217 | orig_request->offset, | |
2218 | orig_request->length, 0, 0); | |
2219 | osd_req_op_extent_osd_data_bio(osd_req, 1, orig_request->bio_list, | |
2220 | orig_request->length); | |
2221 | ||
2222 | rbd_osd_req_format_write(orig_request); | |
2223 | ||
2224 | /* All set, send it off. */ | |
2225 | ||
2226 | orig_request->callback = rbd_img_obj_copyup_callback; | |
2227 | osdc = &rbd_dev->rbd_client->client->osdc; | |
2228 | result = rbd_obj_request_submit(osdc, orig_request); | |
2229 | if (!result) | |
2230 | return; | |
2231 | out_err: | |
2232 | /* Record the error code and complete the request */ | |
2233 | ||
2234 | orig_request->result = result; | |
2235 | orig_request->xferred = 0; | |
2236 | obj_request_done_set(orig_request); | |
2237 | rbd_obj_request_complete(orig_request); | |
3d7efd18 AE |
2238 | } |
2239 | ||
2240 | /* | |
2241 | * Read from the parent image the range of data that covers the | |
2242 | * entire target of the given object request. This is used for | |
2243 | * satisfying a layered image write request when the target of an | |
2244 | * object request from the image request does not exist. | |
2245 | * | |
2246 | * A page array big enough to hold the returned data is allocated | |
2247 | * and supplied to rbd_img_request_fill() as the "data descriptor." | |
2248 | * When the read completes, this page array will be transferred to | |
2249 | * the original object request for the copyup operation. | |
2250 | * | |
2251 | * If an error occurs, record it as the result of the original | |
2252 | * object request and mark it done so it gets completed. | |
2253 | */ | |
2254 | static int rbd_img_obj_parent_read_full(struct rbd_obj_request *obj_request) | |
2255 | { | |
2256 | struct rbd_img_request *img_request = NULL; | |
2257 | struct rbd_img_request *parent_request = NULL; | |
2258 | struct rbd_device *rbd_dev; | |
2259 | u64 img_offset; | |
2260 | u64 length; | |
2261 | struct page **pages = NULL; | |
2262 | u32 page_count; | |
2263 | int result; | |
2264 | ||
2265 | rbd_assert(obj_request_img_data_test(obj_request)); | |
2266 | rbd_assert(obj_request->type == OBJ_REQUEST_BIO); | |
2267 | ||
2268 | img_request = obj_request->img_request; | |
2269 | rbd_assert(img_request != NULL); | |
2270 | rbd_dev = img_request->rbd_dev; | |
2271 | rbd_assert(rbd_dev->parent != NULL); | |
2272 | ||
0eefd470 AE |
2273 | /* |
2274 | * First things first. The original osd request is of no | |
2275 | * use to use any more, we'll need a new one that can hold | |
2276 | * the two ops in a copyup request. We'll get that later, | |
2277 | * but for now we can release the old one. | |
2278 | */ | |
2279 | rbd_osd_req_destroy(obj_request->osd_req); | |
2280 | obj_request->osd_req = NULL; | |
2281 | ||
3d7efd18 AE |
2282 | /* |
2283 | * Determine the byte range covered by the object in the | |
2284 | * child image to which the original request was to be sent. | |
2285 | */ | |
2286 | img_offset = obj_request->img_offset - obj_request->offset; | |
2287 | length = (u64)1 << rbd_dev->header.obj_order; | |
2288 | ||
a9e8ba2c AE |
2289 | /* |
2290 | * There is no defined parent data beyond the parent | |
2291 | * overlap, so limit what we read at that boundary if | |
2292 | * necessary. | |
2293 | */ | |
2294 | if (img_offset + length > rbd_dev->parent_overlap) { | |
2295 | rbd_assert(img_offset < rbd_dev->parent_overlap); | |
2296 | length = rbd_dev->parent_overlap - img_offset; | |
2297 | } | |
2298 | ||
3d7efd18 AE |
2299 | /* |
2300 | * Allocate a page array big enough to receive the data read | |
2301 | * from the parent. | |
2302 | */ | |
2303 | page_count = (u32)calc_pages_for(0, length); | |
2304 | pages = ceph_alloc_page_vector(page_count, GFP_KERNEL); | |
2305 | if (IS_ERR(pages)) { | |
2306 | result = PTR_ERR(pages); | |
2307 | pages = NULL; | |
2308 | goto out_err; | |
2309 | } | |
2310 | ||
2311 | result = -ENOMEM; | |
2312 | parent_request = rbd_img_request_create(rbd_dev->parent, | |
2313 | img_offset, length, | |
2314 | false, true); | |
2315 | if (!parent_request) | |
2316 | goto out_err; | |
2317 | rbd_obj_request_get(obj_request); | |
2318 | parent_request->obj_request = obj_request; | |
2319 | ||
2320 | result = rbd_img_request_fill(parent_request, OBJ_REQUEST_PAGES, pages); | |
2321 | if (result) | |
2322 | goto out_err; | |
2323 | parent_request->copyup_pages = pages; | |
2324 | ||
2325 | parent_request->callback = rbd_img_obj_parent_read_full_callback; | |
2326 | result = rbd_img_request_submit(parent_request); | |
2327 | if (!result) | |
2328 | return 0; | |
2329 | ||
2330 | parent_request->copyup_pages = NULL; | |
2331 | parent_request->obj_request = NULL; | |
2332 | rbd_obj_request_put(obj_request); | |
2333 | out_err: | |
2334 | if (pages) | |
2335 | ceph_release_page_vector(pages, page_count); | |
2336 | if (parent_request) | |
2337 | rbd_img_request_put(parent_request); | |
2338 | obj_request->result = result; | |
2339 | obj_request->xferred = 0; | |
2340 | obj_request_done_set(obj_request); | |
2341 | ||
2342 | return result; | |
2343 | } | |
2344 | ||
c5b5ef6c AE |
2345 | static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request) |
2346 | { | |
c5b5ef6c AE |
2347 | struct rbd_obj_request *orig_request; |
2348 | int result; | |
2349 | ||
2350 | rbd_assert(!obj_request_img_data_test(obj_request)); | |
2351 | ||
2352 | /* | |
2353 | * All we need from the object request is the original | |
2354 | * request and the result of the STAT op. Grab those, then | |
2355 | * we're done with the request. | |
2356 | */ | |
2357 | orig_request = obj_request->obj_request; | |
2358 | obj_request->obj_request = NULL; | |
2359 | rbd_assert(orig_request); | |
2360 | rbd_assert(orig_request->img_request); | |
2361 | ||
2362 | result = obj_request->result; | |
2363 | obj_request->result = 0; | |
2364 | ||
2365 | dout("%s: obj %p for obj %p result %d %llu/%llu\n", __func__, | |
2366 | obj_request, orig_request, result, | |
2367 | obj_request->xferred, obj_request->length); | |
2368 | rbd_obj_request_put(obj_request); | |
2369 | ||
2370 | rbd_assert(orig_request); | |
2371 | rbd_assert(orig_request->img_request); | |
c5b5ef6c AE |
2372 | |
2373 | /* | |
2374 | * Our only purpose here is to determine whether the object | |
2375 | * exists, and we don't want to treat the non-existence as | |
2376 | * an error. If something else comes back, transfer the | |
2377 | * error to the original request and complete it now. | |
2378 | */ | |
2379 | if (!result) { | |
2380 | obj_request_existence_set(orig_request, true); | |
2381 | } else if (result == -ENOENT) { | |
2382 | obj_request_existence_set(orig_request, false); | |
2383 | } else if (result) { | |
2384 | orig_request->result = result; | |
3d7efd18 | 2385 | goto out; |
c5b5ef6c AE |
2386 | } |
2387 | ||
2388 | /* | |
2389 | * Resubmit the original request now that we have recorded | |
2390 | * whether the target object exists. | |
2391 | */ | |
b454e36d | 2392 | orig_request->result = rbd_img_obj_request_submit(orig_request); |
3d7efd18 | 2393 | out: |
c5b5ef6c AE |
2394 | if (orig_request->result) |
2395 | rbd_obj_request_complete(orig_request); | |
2396 | rbd_obj_request_put(orig_request); | |
2397 | } | |
2398 | ||
2399 | static int rbd_img_obj_exists_submit(struct rbd_obj_request *obj_request) | |
2400 | { | |
2401 | struct rbd_obj_request *stat_request; | |
2402 | struct rbd_device *rbd_dev; | |
2403 | struct ceph_osd_client *osdc; | |
2404 | struct page **pages = NULL; | |
2405 | u32 page_count; | |
2406 | size_t size; | |
2407 | int ret; | |
2408 | ||
2409 | /* | |
2410 | * The response data for a STAT call consists of: | |
2411 | * le64 length; | |
2412 | * struct { | |
2413 | * le32 tv_sec; | |
2414 | * le32 tv_nsec; | |
2415 | * } mtime; | |
2416 | */ | |
2417 | size = sizeof (__le64) + sizeof (__le32) + sizeof (__le32); | |
2418 | page_count = (u32)calc_pages_for(0, size); | |
2419 | pages = ceph_alloc_page_vector(page_count, GFP_KERNEL); | |
2420 | if (IS_ERR(pages)) | |
2421 | return PTR_ERR(pages); | |
2422 | ||
2423 | ret = -ENOMEM; | |
2424 | stat_request = rbd_obj_request_create(obj_request->object_name, 0, 0, | |
2425 | OBJ_REQUEST_PAGES); | |
2426 | if (!stat_request) | |
2427 | goto out; | |
2428 | ||
2429 | rbd_obj_request_get(obj_request); | |
2430 | stat_request->obj_request = obj_request; | |
2431 | stat_request->pages = pages; | |
2432 | stat_request->page_count = page_count; | |
2433 | ||
2434 | rbd_assert(obj_request->img_request); | |
2435 | rbd_dev = obj_request->img_request->rbd_dev; | |
2436 | stat_request->osd_req = rbd_osd_req_create(rbd_dev, false, | |
2437 | stat_request); | |
2438 | if (!stat_request->osd_req) | |
2439 | goto out; | |
2440 | stat_request->callback = rbd_img_obj_exists_callback; | |
2441 | ||
2442 | osd_req_op_init(stat_request->osd_req, 0, CEPH_OSD_OP_STAT); | |
2443 | osd_req_op_raw_data_in_pages(stat_request->osd_req, 0, pages, size, 0, | |
2444 | false, false); | |
9d4df01f | 2445 | rbd_osd_req_format_read(stat_request); |
c5b5ef6c AE |
2446 | |
2447 | osdc = &rbd_dev->rbd_client->client->osdc; | |
2448 | ret = rbd_obj_request_submit(osdc, stat_request); | |
2449 | out: | |
2450 | if (ret) | |
2451 | rbd_obj_request_put(obj_request); | |
2452 | ||
2453 | return ret; | |
2454 | } | |
2455 | ||
b454e36d AE |
2456 | static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request) |
2457 | { | |
2458 | struct rbd_img_request *img_request; | |
a9e8ba2c | 2459 | struct rbd_device *rbd_dev; |
3d7efd18 | 2460 | bool known; |
b454e36d AE |
2461 | |
2462 | rbd_assert(obj_request_img_data_test(obj_request)); | |
2463 | ||
2464 | img_request = obj_request->img_request; | |
2465 | rbd_assert(img_request); | |
a9e8ba2c | 2466 | rbd_dev = img_request->rbd_dev; |
b454e36d | 2467 | |
b454e36d | 2468 | /* |
a9e8ba2c AE |
2469 | * Only writes to layered images need special handling. |
2470 | * Reads and non-layered writes are simple object requests. | |
2471 | * Layered writes that start beyond the end of the overlap | |
2472 | * with the parent have no parent data, so they too are | |
2473 | * simple object requests. Finally, if the target object is | |
2474 | * known to already exist, its parent data has already been | |
2475 | * copied, so a write to the object can also be handled as a | |
2476 | * simple object request. | |
b454e36d AE |
2477 | */ |
2478 | if (!img_request_write_test(img_request) || | |
2479 | !img_request_layered_test(img_request) || | |
a9e8ba2c | 2480 | rbd_dev->parent_overlap <= obj_request->img_offset || |
3d7efd18 AE |
2481 | ((known = obj_request_known_test(obj_request)) && |
2482 | obj_request_exists_test(obj_request))) { | |
b454e36d AE |
2483 | |
2484 | struct rbd_device *rbd_dev; | |
2485 | struct ceph_osd_client *osdc; | |
2486 | ||
2487 | rbd_dev = obj_request->img_request->rbd_dev; | |
2488 | osdc = &rbd_dev->rbd_client->client->osdc; | |
2489 | ||
2490 | return rbd_obj_request_submit(osdc, obj_request); | |
2491 | } | |
2492 | ||
2493 | /* | |
3d7efd18 AE |
2494 | * It's a layered write. The target object might exist but |
2495 | * we may not know that yet. If we know it doesn't exist, | |
2496 | * start by reading the data for the full target object from | |
2497 | * the parent so we can use it for a copyup to the target. | |
b454e36d | 2498 | */ |
3d7efd18 AE |
2499 | if (known) |
2500 | return rbd_img_obj_parent_read_full(obj_request); | |
2501 | ||
2502 | /* We don't know whether the target exists. Go find out. */ | |
b454e36d AE |
2503 | |
2504 | return rbd_img_obj_exists_submit(obj_request); | |
2505 | } | |
2506 | ||
bf0d5f50 AE |
2507 | static int rbd_img_request_submit(struct rbd_img_request *img_request) |
2508 | { | |
bf0d5f50 | 2509 | struct rbd_obj_request *obj_request; |
46faeed4 | 2510 | struct rbd_obj_request *next_obj_request; |
bf0d5f50 | 2511 | |
37206ee5 | 2512 | dout("%s: img %p\n", __func__, img_request); |
46faeed4 | 2513 | for_each_obj_request_safe(img_request, obj_request, next_obj_request) { |
bf0d5f50 AE |
2514 | int ret; |
2515 | ||
b454e36d | 2516 | ret = rbd_img_obj_request_submit(obj_request); |
bf0d5f50 AE |
2517 | if (ret) |
2518 | return ret; | |
bf0d5f50 AE |
2519 | } |
2520 | ||
2521 | return 0; | |
2522 | } | |
8b3e1a56 AE |
2523 | |
2524 | static void rbd_img_parent_read_callback(struct rbd_img_request *img_request) | |
2525 | { | |
2526 | struct rbd_obj_request *obj_request; | |
a9e8ba2c AE |
2527 | struct rbd_device *rbd_dev; |
2528 | u64 obj_end; | |
8b3e1a56 AE |
2529 | |
2530 | rbd_assert(img_request_child_test(img_request)); | |
2531 | ||
2532 | obj_request = img_request->obj_request; | |
a9e8ba2c AE |
2533 | rbd_assert(obj_request); |
2534 | rbd_assert(obj_request->img_request); | |
2535 | ||
8b3e1a56 | 2536 | obj_request->result = img_request->result; |
a9e8ba2c AE |
2537 | if (obj_request->result) |
2538 | goto out; | |
2539 | ||
2540 | /* | |
2541 | * We need to zero anything beyond the parent overlap | |
2542 | * boundary. Since rbd_img_obj_request_read_callback() | |
2543 | * will zero anything beyond the end of a short read, an | |
2544 | * easy way to do this is to pretend the data from the | |
2545 | * parent came up short--ending at the overlap boundary. | |
2546 | */ | |
2547 | rbd_assert(obj_request->img_offset < U64_MAX - obj_request->length); | |
2548 | obj_end = obj_request->img_offset + obj_request->length; | |
2549 | rbd_dev = obj_request->img_request->rbd_dev; | |
2550 | if (obj_end > rbd_dev->parent_overlap) { | |
2551 | u64 xferred = 0; | |
2552 | ||
2553 | if (obj_request->img_offset < rbd_dev->parent_overlap) | |
2554 | xferred = rbd_dev->parent_overlap - | |
2555 | obj_request->img_offset; | |
8b3e1a56 | 2556 | |
a9e8ba2c AE |
2557 | obj_request->xferred = min(img_request->xferred, xferred); |
2558 | } else { | |
2559 | obj_request->xferred = img_request->xferred; | |
2560 | } | |
2561 | out: | |
b5b09be3 | 2562 | rbd_img_request_put(img_request); |
8b3e1a56 AE |
2563 | rbd_img_obj_request_read_callback(obj_request); |
2564 | rbd_obj_request_complete(obj_request); | |
2565 | } | |
2566 | ||
2567 | static void rbd_img_parent_read(struct rbd_obj_request *obj_request) | |
2568 | { | |
2569 | struct rbd_device *rbd_dev; | |
2570 | struct rbd_img_request *img_request; | |
2571 | int result; | |
2572 | ||
2573 | rbd_assert(obj_request_img_data_test(obj_request)); | |
2574 | rbd_assert(obj_request->img_request != NULL); | |
2575 | rbd_assert(obj_request->result == (s32) -ENOENT); | |
2576 | rbd_assert(obj_request->type == OBJ_REQUEST_BIO); | |
2577 | ||
2578 | rbd_dev = obj_request->img_request->rbd_dev; | |
2579 | rbd_assert(rbd_dev->parent != NULL); | |
2580 | /* rbd_read_finish(obj_request, obj_request->length); */ | |
2581 | img_request = rbd_img_request_create(rbd_dev->parent, | |
2582 | obj_request->img_offset, | |
2583 | obj_request->length, | |
2584 | false, true); | |
2585 | result = -ENOMEM; | |
2586 | if (!img_request) | |
2587 | goto out_err; | |
2588 | ||
2589 | rbd_obj_request_get(obj_request); | |
2590 | img_request->obj_request = obj_request; | |
2591 | ||
f1a4739f AE |
2592 | result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO, |
2593 | obj_request->bio_list); | |
8b3e1a56 AE |
2594 | if (result) |
2595 | goto out_err; | |
2596 | ||
2597 | img_request->callback = rbd_img_parent_read_callback; | |
2598 | result = rbd_img_request_submit(img_request); | |
2599 | if (result) | |
2600 | goto out_err; | |
2601 | ||
2602 | return; | |
2603 | out_err: | |
2604 | if (img_request) | |
2605 | rbd_img_request_put(img_request); | |
2606 | obj_request->result = result; | |
2607 | obj_request->xferred = 0; | |
2608 | obj_request_done_set(obj_request); | |
2609 | } | |
bf0d5f50 | 2610 | |
cc4a38bd | 2611 | static int rbd_obj_notify_ack(struct rbd_device *rbd_dev, u64 notify_id) |
b8d70035 AE |
2612 | { |
2613 | struct rbd_obj_request *obj_request; | |
2169238d | 2614 | struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; |
b8d70035 AE |
2615 | int ret; |
2616 | ||
2617 | obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0, | |
2618 | OBJ_REQUEST_NODATA); | |
2619 | if (!obj_request) | |
2620 | return -ENOMEM; | |
2621 | ||
2622 | ret = -ENOMEM; | |
430c28c3 | 2623 | obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request); |
b8d70035 AE |
2624 | if (!obj_request->osd_req) |
2625 | goto out; | |
2169238d | 2626 | obj_request->callback = rbd_obj_request_put; |
b8d70035 | 2627 | |
c99d2d4a | 2628 | osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_NOTIFY_ACK, |
cc4a38bd | 2629 | notify_id, 0, 0); |
9d4df01f | 2630 | rbd_osd_req_format_read(obj_request); |
430c28c3 | 2631 | |
b8d70035 | 2632 | ret = rbd_obj_request_submit(osdc, obj_request); |
b8d70035 | 2633 | out: |
cf81b60e AE |
2634 | if (ret) |
2635 | rbd_obj_request_put(obj_request); | |
b8d70035 AE |
2636 | |
2637 | return ret; | |
2638 | } | |
2639 | ||
2640 | static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data) | |
2641 | { | |
2642 | struct rbd_device *rbd_dev = (struct rbd_device *)data; | |
e627db08 | 2643 | int ret; |
b8d70035 AE |
2644 | |
2645 | if (!rbd_dev) | |
2646 | return; | |
2647 | ||
37206ee5 | 2648 | dout("%s: \"%s\" notify_id %llu opcode %u\n", __func__, |
cc4a38bd AE |
2649 | rbd_dev->header_name, (unsigned long long)notify_id, |
2650 | (unsigned int)opcode); | |
e627db08 AE |
2651 | ret = rbd_dev_refresh(rbd_dev); |
2652 | if (ret) | |
2653 | rbd_warn(rbd_dev, ": header refresh error (%d)\n", ret); | |
b8d70035 | 2654 | |
cc4a38bd | 2655 | rbd_obj_notify_ack(rbd_dev, notify_id); |
b8d70035 AE |
2656 | } |
2657 | ||
9969ebc5 AE |
2658 | /* |
2659 | * Request sync osd watch/unwatch. The value of "start" determines | |
2660 | * whether a watch request is being initiated or torn down. | |
2661 | */ | |
2662 | static int rbd_dev_header_watch_sync(struct rbd_device *rbd_dev, int start) | |
2663 | { | |
2664 | struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; | |
2665 | struct rbd_obj_request *obj_request; | |
9969ebc5 AE |
2666 | int ret; |
2667 | ||
2668 | rbd_assert(start ^ !!rbd_dev->watch_event); | |
2669 | rbd_assert(start ^ !!rbd_dev->watch_request); | |
2670 | ||
2671 | if (start) { | |
3c663bbd | 2672 | ret = ceph_osdc_create_event(osdc, rbd_watch_cb, rbd_dev, |
9969ebc5 AE |
2673 | &rbd_dev->watch_event); |
2674 | if (ret < 0) | |
2675 | return ret; | |
8eb87565 | 2676 | rbd_assert(rbd_dev->watch_event != NULL); |
9969ebc5 AE |
2677 | } |
2678 | ||
2679 | ret = -ENOMEM; | |
2680 | obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0, | |
2681 | OBJ_REQUEST_NODATA); | |
2682 | if (!obj_request) | |
2683 | goto out_cancel; | |
2684 | ||
430c28c3 AE |
2685 | obj_request->osd_req = rbd_osd_req_create(rbd_dev, true, obj_request); |
2686 | if (!obj_request->osd_req) | |
2687 | goto out_cancel; | |
2688 | ||
8eb87565 | 2689 | if (start) |
975241af | 2690 | ceph_osdc_set_request_linger(osdc, obj_request->osd_req); |
8eb87565 | 2691 | else |
6977c3f9 | 2692 | ceph_osdc_unregister_linger_request(osdc, |
975241af | 2693 | rbd_dev->watch_request->osd_req); |
2169238d AE |
2694 | |
2695 | osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH, | |
b21ebddd | 2696 | rbd_dev->watch_event->cookie, 0, start); |
9d4df01f | 2697 | rbd_osd_req_format_write(obj_request); |
2169238d | 2698 | |
9969ebc5 AE |
2699 | ret = rbd_obj_request_submit(osdc, obj_request); |
2700 | if (ret) | |
2701 | goto out_cancel; | |
2702 | ret = rbd_obj_request_wait(obj_request); | |
2703 | if (ret) | |
2704 | goto out_cancel; | |
9969ebc5 AE |
2705 | ret = obj_request->result; |
2706 | if (ret) | |
2707 | goto out_cancel; | |
2708 | ||
8eb87565 AE |
2709 | /* |
2710 | * A watch request is set to linger, so the underlying osd | |
2711 | * request won't go away until we unregister it. We retain | |
2712 | * a pointer to the object request during that time (in | |
2713 | * rbd_dev->watch_request), so we'll keep a reference to | |
2714 | * it. We'll drop that reference (below) after we've | |
2715 | * unregistered it. | |
2716 | */ | |
2717 | if (start) { | |
2718 | rbd_dev->watch_request = obj_request; | |
2719 | ||
2720 | return 0; | |
2721 | } | |
2722 | ||
2723 | /* We have successfully torn down the watch request */ | |
2724 | ||
2725 | rbd_obj_request_put(rbd_dev->watch_request); | |
2726 | rbd_dev->watch_request = NULL; | |
9969ebc5 AE |
2727 | out_cancel: |
2728 | /* Cancel the event if we're tearing down, or on error */ | |
2729 | ceph_osdc_cancel_event(rbd_dev->watch_event); | |
2730 | rbd_dev->watch_event = NULL; | |
9969ebc5 AE |
2731 | if (obj_request) |
2732 | rbd_obj_request_put(obj_request); | |
2733 | ||
2734 | return ret; | |
2735 | } | |
2736 | ||
36be9a76 | 2737 | /* |
f40eb349 AE |
2738 | * Synchronous osd object method call. Returns the number of bytes |
2739 | * returned in the outbound buffer, or a negative error code. | |
36be9a76 AE |
2740 | */ |
2741 | static int rbd_obj_method_sync(struct rbd_device *rbd_dev, | |
2742 | const char *object_name, | |
2743 | const char *class_name, | |
2744 | const char *method_name, | |
4157976b | 2745 | const void *outbound, |
36be9a76 | 2746 | size_t outbound_size, |
4157976b | 2747 | void *inbound, |
e2a58ee5 | 2748 | size_t inbound_size) |
36be9a76 | 2749 | { |
2169238d | 2750 | struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; |
36be9a76 | 2751 | struct rbd_obj_request *obj_request; |
36be9a76 AE |
2752 | struct page **pages; |
2753 | u32 page_count; | |
2754 | int ret; | |
2755 | ||
2756 | /* | |
6010a451 AE |
2757 | * Method calls are ultimately read operations. The result |
2758 | * should placed into the inbound buffer provided. They | |
2759 | * also supply outbound data--parameters for the object | |
2760 | * method. Currently if this is present it will be a | |
2761 | * snapshot id. | |
36be9a76 | 2762 | */ |
57385b51 | 2763 | page_count = (u32)calc_pages_for(0, inbound_size); |
36be9a76 AE |
2764 | pages = ceph_alloc_page_vector(page_count, GFP_KERNEL); |
2765 | if (IS_ERR(pages)) | |
2766 | return PTR_ERR(pages); | |
2767 | ||
2768 | ret = -ENOMEM; | |
6010a451 | 2769 | obj_request = rbd_obj_request_create(object_name, 0, inbound_size, |
36be9a76 AE |
2770 | OBJ_REQUEST_PAGES); |
2771 | if (!obj_request) | |
2772 | goto out; | |
2773 | ||
2774 | obj_request->pages = pages; | |
2775 | obj_request->page_count = page_count; | |
2776 | ||
430c28c3 | 2777 | obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request); |
36be9a76 AE |
2778 | if (!obj_request->osd_req) |
2779 | goto out; | |
2780 | ||
c99d2d4a | 2781 | osd_req_op_cls_init(obj_request->osd_req, 0, CEPH_OSD_OP_CALL, |
04017e29 AE |
2782 | class_name, method_name); |
2783 | if (outbound_size) { | |
2784 | struct ceph_pagelist *pagelist; | |
2785 | ||
2786 | pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS); | |
2787 | if (!pagelist) | |
2788 | goto out; | |
2789 | ||
2790 | ceph_pagelist_init(pagelist); | |
2791 | ceph_pagelist_append(pagelist, outbound, outbound_size); | |
2792 | osd_req_op_cls_request_data_pagelist(obj_request->osd_req, 0, | |
2793 | pagelist); | |
2794 | } | |
a4ce40a9 AE |
2795 | osd_req_op_cls_response_data_pages(obj_request->osd_req, 0, |
2796 | obj_request->pages, inbound_size, | |
44cd188d | 2797 | 0, false, false); |
9d4df01f | 2798 | rbd_osd_req_format_read(obj_request); |
430c28c3 | 2799 | |
36be9a76 AE |
2800 | ret = rbd_obj_request_submit(osdc, obj_request); |
2801 | if (ret) | |
2802 | goto out; | |
2803 | ret = rbd_obj_request_wait(obj_request); | |
2804 | if (ret) | |
2805 | goto out; | |
2806 | ||
2807 | ret = obj_request->result; | |
2808 | if (ret < 0) | |
2809 | goto out; | |
57385b51 AE |
2810 | |
2811 | rbd_assert(obj_request->xferred < (u64)INT_MAX); | |
2812 | ret = (int)obj_request->xferred; | |
903bb32e | 2813 | ceph_copy_from_page_vector(pages, inbound, 0, obj_request->xferred); |
36be9a76 AE |
2814 | out: |
2815 | if (obj_request) | |
2816 | rbd_obj_request_put(obj_request); | |
2817 | else | |
2818 | ceph_release_page_vector(pages, page_count); | |
2819 | ||
2820 | return ret; | |
2821 | } | |
2822 | ||
bf0d5f50 | 2823 | static void rbd_request_fn(struct request_queue *q) |
cc344fa1 | 2824 | __releases(q->queue_lock) __acquires(q->queue_lock) |
bf0d5f50 AE |
2825 | { |
2826 | struct rbd_device *rbd_dev = q->queuedata; | |
2827 | bool read_only = rbd_dev->mapping.read_only; | |
2828 | struct request *rq; | |
2829 | int result; | |
2830 | ||
2831 | while ((rq = blk_fetch_request(q))) { | |
2832 | bool write_request = rq_data_dir(rq) == WRITE; | |
2833 | struct rbd_img_request *img_request; | |
2834 | u64 offset; | |
2835 | u64 length; | |
2836 | ||
2837 | /* Ignore any non-FS requests that filter through. */ | |
2838 | ||
2839 | if (rq->cmd_type != REQ_TYPE_FS) { | |
4dda41d3 AE |
2840 | dout("%s: non-fs request type %d\n", __func__, |
2841 | (int) rq->cmd_type); | |
2842 | __blk_end_request_all(rq, 0); | |
2843 | continue; | |
2844 | } | |
2845 | ||
2846 | /* Ignore/skip any zero-length requests */ | |
2847 | ||
2848 | offset = (u64) blk_rq_pos(rq) << SECTOR_SHIFT; | |
2849 | length = (u64) blk_rq_bytes(rq); | |
2850 | ||
2851 | if (!length) { | |
2852 | dout("%s: zero-length request\n", __func__); | |
bf0d5f50 AE |
2853 | __blk_end_request_all(rq, 0); |
2854 | continue; | |
2855 | } | |
2856 | ||
2857 | spin_unlock_irq(q->queue_lock); | |
2858 | ||
2859 | /* Disallow writes to a read-only device */ | |
2860 | ||
2861 | if (write_request) { | |
2862 | result = -EROFS; | |
2863 | if (read_only) | |
2864 | goto end_request; | |
2865 | rbd_assert(rbd_dev->spec->snap_id == CEPH_NOSNAP); | |
2866 | } | |
2867 | ||
6d292906 AE |
2868 | /* |
2869 | * Quit early if the mapped snapshot no longer | |
2870 | * exists. It's still possible the snapshot will | |
2871 | * have disappeared by the time our request arrives | |
2872 | * at the osd, but there's no sense in sending it if | |
2873 | * we already know. | |
2874 | */ | |
2875 | if (!test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags)) { | |
bf0d5f50 AE |
2876 | dout("request for non-existent snapshot"); |
2877 | rbd_assert(rbd_dev->spec->snap_id != CEPH_NOSNAP); | |
2878 | result = -ENXIO; | |
2879 | goto end_request; | |
2880 | } | |
2881 | ||
bf0d5f50 | 2882 | result = -EINVAL; |
c0cd10db AE |
2883 | if (offset && length > U64_MAX - offset + 1) { |
2884 | rbd_warn(rbd_dev, "bad request range (%llu~%llu)\n", | |
2885 | offset, length); | |
bf0d5f50 | 2886 | goto end_request; /* Shouldn't happen */ |
c0cd10db | 2887 | } |
bf0d5f50 | 2888 | |
00a653e2 AE |
2889 | result = -EIO; |
2890 | if (offset + length > rbd_dev->mapping.size) { | |
2891 | rbd_warn(rbd_dev, "beyond EOD (%llu~%llu > %llu)\n", | |
2892 | offset, length, rbd_dev->mapping.size); | |
2893 | goto end_request; | |
2894 | } | |
2895 | ||
bf0d5f50 AE |
2896 | result = -ENOMEM; |
2897 | img_request = rbd_img_request_create(rbd_dev, offset, length, | |
9849e986 | 2898 | write_request, false); |
bf0d5f50 AE |
2899 | if (!img_request) |
2900 | goto end_request; | |
2901 | ||
2902 | img_request->rq = rq; | |
2903 | ||
f1a4739f AE |
2904 | result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO, |
2905 | rq->bio); | |
bf0d5f50 AE |
2906 | if (!result) |
2907 | result = rbd_img_request_submit(img_request); | |
2908 | if (result) | |
2909 | rbd_img_request_put(img_request); | |
2910 | end_request: | |
2911 | spin_lock_irq(q->queue_lock); | |
2912 | if (result < 0) { | |
7da22d29 AE |
2913 | rbd_warn(rbd_dev, "%s %llx at %llx result %d\n", |
2914 | write_request ? "write" : "read", | |
2915 | length, offset, result); | |
2916 | ||
bf0d5f50 AE |
2917 | __blk_end_request_all(rq, result); |
2918 | } | |
2919 | } | |
2920 | } | |
2921 | ||
602adf40 YS |
2922 | /* |
2923 | * a queue callback. Makes sure that we don't create a bio that spans across | |
2924 | * multiple osd objects. One exception would be with a single page bios, | |
f7760dad | 2925 | * which we handle later at bio_chain_clone_range() |
602adf40 YS |
2926 | */ |
2927 | static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd, | |
2928 | struct bio_vec *bvec) | |
2929 | { | |
2930 | struct rbd_device *rbd_dev = q->queuedata; | |
e5cfeed2 AE |
2931 | sector_t sector_offset; |
2932 | sector_t sectors_per_obj; | |
2933 | sector_t obj_sector_offset; | |
2934 | int ret; | |
2935 | ||
2936 | /* | |
2937 | * Find how far into its rbd object the partition-relative | |
2938 | * bio start sector is to offset relative to the enclosing | |
2939 | * device. | |
2940 | */ | |
2941 | sector_offset = get_start_sect(bmd->bi_bdev) + bmd->bi_sector; | |
2942 | sectors_per_obj = 1 << (rbd_dev->header.obj_order - SECTOR_SHIFT); | |
2943 | obj_sector_offset = sector_offset & (sectors_per_obj - 1); | |
2944 | ||
2945 | /* | |
2946 | * Compute the number of bytes from that offset to the end | |
2947 | * of the object. Account for what's already used by the bio. | |
2948 | */ | |
2949 | ret = (int) (sectors_per_obj - obj_sector_offset) << SECTOR_SHIFT; | |
2950 | if (ret > bmd->bi_size) | |
2951 | ret -= bmd->bi_size; | |
2952 | else | |
2953 | ret = 0; | |
2954 | ||
2955 | /* | |
2956 | * Don't send back more than was asked for. And if the bio | |
2957 | * was empty, let the whole thing through because: "Note | |
2958 | * that a block device *must* allow a single page to be | |
2959 | * added to an empty bio." | |
2960 | */ | |
2961 | rbd_assert(bvec->bv_len <= PAGE_SIZE); | |
2962 | if (ret > (int) bvec->bv_len || !bmd->bi_size) | |
2963 | ret = (int) bvec->bv_len; | |
2964 | ||
2965 | return ret; | |
602adf40 YS |
2966 | } |
2967 | ||
2968 | static void rbd_free_disk(struct rbd_device *rbd_dev) | |
2969 | { | |
2970 | struct gendisk *disk = rbd_dev->disk; | |
2971 | ||
2972 | if (!disk) | |
2973 | return; | |
2974 | ||
a0cab924 AE |
2975 | rbd_dev->disk = NULL; |
2976 | if (disk->flags & GENHD_FL_UP) { | |
602adf40 | 2977 | del_gendisk(disk); |
a0cab924 AE |
2978 | if (disk->queue) |
2979 | blk_cleanup_queue(disk->queue); | |
2980 | } | |
602adf40 YS |
2981 | put_disk(disk); |
2982 | } | |
2983 | ||
788e2df3 AE |
2984 | static int rbd_obj_read_sync(struct rbd_device *rbd_dev, |
2985 | const char *object_name, | |
7097f8df | 2986 | u64 offset, u64 length, void *buf) |
788e2df3 AE |
2987 | |
2988 | { | |
2169238d | 2989 | struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; |
788e2df3 | 2990 | struct rbd_obj_request *obj_request; |
788e2df3 AE |
2991 | struct page **pages = NULL; |
2992 | u32 page_count; | |
1ceae7ef | 2993 | size_t size; |
788e2df3 AE |
2994 | int ret; |
2995 | ||
2996 | page_count = (u32) calc_pages_for(offset, length); | |
2997 | pages = ceph_alloc_page_vector(page_count, GFP_KERNEL); | |
2998 | if (IS_ERR(pages)) | |
2999 | ret = PTR_ERR(pages); | |
3000 | ||
3001 | ret = -ENOMEM; | |
3002 | obj_request = rbd_obj_request_create(object_name, offset, length, | |
36be9a76 | 3003 | OBJ_REQUEST_PAGES); |
788e2df3 AE |
3004 | if (!obj_request) |
3005 | goto out; | |
3006 | ||
3007 | obj_request->pages = pages; | |
3008 | obj_request->page_count = page_count; | |
3009 | ||
430c28c3 | 3010 | obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request); |
788e2df3 AE |
3011 | if (!obj_request->osd_req) |
3012 | goto out; | |
3013 | ||
c99d2d4a AE |
3014 | osd_req_op_extent_init(obj_request->osd_req, 0, CEPH_OSD_OP_READ, |
3015 | offset, length, 0, 0); | |
406e2c9f | 3016 | osd_req_op_extent_osd_data_pages(obj_request->osd_req, 0, |
a4ce40a9 | 3017 | obj_request->pages, |
44cd188d AE |
3018 | obj_request->length, |
3019 | obj_request->offset & ~PAGE_MASK, | |
3020 | false, false); | |
9d4df01f | 3021 | rbd_osd_req_format_read(obj_request); |
430c28c3 | 3022 | |
788e2df3 AE |
3023 | ret = rbd_obj_request_submit(osdc, obj_request); |
3024 | if (ret) | |
3025 | goto out; | |
3026 | ret = rbd_obj_request_wait(obj_request); | |
3027 | if (ret) | |
3028 | goto out; | |
3029 | ||
3030 | ret = obj_request->result; | |
3031 | if (ret < 0) | |
3032 | goto out; | |
1ceae7ef AE |
3033 | |
3034 | rbd_assert(obj_request->xferred <= (u64) SIZE_MAX); | |
3035 | size = (size_t) obj_request->xferred; | |
903bb32e | 3036 | ceph_copy_from_page_vector(pages, buf, 0, size); |
7097f8df AE |
3037 | rbd_assert(size <= (size_t)INT_MAX); |
3038 | ret = (int)size; | |
788e2df3 AE |
3039 | out: |
3040 | if (obj_request) | |
3041 | rbd_obj_request_put(obj_request); | |
3042 | else | |
3043 | ceph_release_page_vector(pages, page_count); | |
3044 | ||
3045 | return ret; | |
3046 | } | |
3047 | ||
602adf40 | 3048 | /* |
662518b1 AE |
3049 | * Read the complete header for the given rbd device. On successful |
3050 | * return, the rbd_dev->header field will contain up-to-date | |
3051 | * information about the image. | |
602adf40 | 3052 | */ |
662518b1 | 3053 | static int rbd_dev_v1_header_read(struct rbd_device *rbd_dev) |
602adf40 | 3054 | { |
4156d998 | 3055 | struct rbd_image_header_ondisk *ondisk = NULL; |
50f7c4c9 | 3056 | u32 snap_count = 0; |
4156d998 AE |
3057 | u64 names_size = 0; |
3058 | u32 want_count; | |
3059 | int ret; | |
602adf40 | 3060 | |
00f1f36f | 3061 | /* |
4156d998 AE |
3062 | * The complete header will include an array of its 64-bit |
3063 | * snapshot ids, followed by the names of those snapshots as | |
3064 | * a contiguous block of NUL-terminated strings. Note that | |
3065 | * the number of snapshots could change by the time we read | |
3066 | * it in, in which case we re-read it. | |
00f1f36f | 3067 | */ |
4156d998 AE |
3068 | do { |
3069 | size_t size; | |
3070 | ||
3071 | kfree(ondisk); | |
3072 | ||
3073 | size = sizeof (*ondisk); | |
3074 | size += snap_count * sizeof (struct rbd_image_snap_ondisk); | |
3075 | size += names_size; | |
3076 | ondisk = kmalloc(size, GFP_KERNEL); | |
3077 | if (!ondisk) | |
662518b1 | 3078 | return -ENOMEM; |
4156d998 | 3079 | |
788e2df3 | 3080 | ret = rbd_obj_read_sync(rbd_dev, rbd_dev->header_name, |
7097f8df | 3081 | 0, size, ondisk); |
4156d998 | 3082 | if (ret < 0) |
662518b1 | 3083 | goto out; |
c0cd10db | 3084 | if ((size_t)ret < size) { |
4156d998 | 3085 | ret = -ENXIO; |
06ecc6cb AE |
3086 | rbd_warn(rbd_dev, "short header read (want %zd got %d)", |
3087 | size, ret); | |
662518b1 | 3088 | goto out; |
4156d998 AE |
3089 | } |
3090 | if (!rbd_dev_ondisk_valid(ondisk)) { | |
3091 | ret = -ENXIO; | |
06ecc6cb | 3092 | rbd_warn(rbd_dev, "invalid header"); |
662518b1 | 3093 | goto out; |
81e759fb | 3094 | } |
602adf40 | 3095 | |
4156d998 AE |
3096 | names_size = le64_to_cpu(ondisk->snap_names_len); |
3097 | want_count = snap_count; | |
3098 | snap_count = le32_to_cpu(ondisk->snap_count); | |
3099 | } while (snap_count != want_count); | |
00f1f36f | 3100 | |
662518b1 AE |
3101 | ret = rbd_header_from_disk(rbd_dev, ondisk); |
3102 | out: | |
4156d998 AE |
3103 | kfree(ondisk); |
3104 | ||
3105 | return ret; | |
602adf40 YS |
3106 | } |
3107 | ||
602adf40 YS |
3108 | /* |
3109 | * only read the first part of the ondisk header, without the snaps info | |
3110 | */ | |
cc4a38bd | 3111 | static int rbd_dev_v1_refresh(struct rbd_device *rbd_dev) |
602adf40 | 3112 | { |
662518b1 | 3113 | return rbd_dev_v1_header_read(rbd_dev); |
602adf40 YS |
3114 | } |
3115 | ||
15228ede AE |
3116 | /* |
3117 | * Clear the rbd device's EXISTS flag if the snapshot it's mapped to | |
3118 | * has disappeared from the (just updated) snapshot context. | |
3119 | */ | |
3120 | static void rbd_exists_validate(struct rbd_device *rbd_dev) | |
3121 | { | |
3122 | u64 snap_id; | |
3123 | ||
3124 | if (!test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags)) | |
3125 | return; | |
3126 | ||
3127 | snap_id = rbd_dev->spec->snap_id; | |
3128 | if (snap_id == CEPH_NOSNAP) | |
3129 | return; | |
3130 | ||
3131 | if (rbd_dev_snap_index(rbd_dev, snap_id) == BAD_SNAP_INDEX) | |
3132 | clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags); | |
3133 | } | |
3134 | ||
cc4a38bd | 3135 | static int rbd_dev_refresh(struct rbd_device *rbd_dev) |
1fe5e993 | 3136 | { |
e627db08 | 3137 | u64 mapping_size; |
1fe5e993 AE |
3138 | int ret; |
3139 | ||
117973fb | 3140 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); |
e627db08 | 3141 | mapping_size = rbd_dev->mapping.size; |
1fe5e993 | 3142 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); |
117973fb | 3143 | if (rbd_dev->image_format == 1) |
cc4a38bd | 3144 | ret = rbd_dev_v1_refresh(rbd_dev); |
117973fb | 3145 | else |
cc4a38bd | 3146 | ret = rbd_dev_v2_refresh(rbd_dev); |
15228ede AE |
3147 | |
3148 | /* If it's a mapped snapshot, validate its EXISTS flag */ | |
3149 | ||
3150 | rbd_exists_validate(rbd_dev); | |
1fe5e993 | 3151 | mutex_unlock(&ctl_mutex); |
00a653e2 AE |
3152 | if (mapping_size != rbd_dev->mapping.size) { |
3153 | sector_t size; | |
3154 | ||
3155 | size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE; | |
3156 | dout("setting size to %llu sectors", (unsigned long long)size); | |
3157 | set_capacity(rbd_dev->disk, size); | |
a3fbe5d4 | 3158 | revalidate_disk(rbd_dev->disk); |
00a653e2 | 3159 | } |
1fe5e993 AE |
3160 | |
3161 | return ret; | |
3162 | } | |
3163 | ||
602adf40 YS |
3164 | static int rbd_init_disk(struct rbd_device *rbd_dev) |
3165 | { | |
3166 | struct gendisk *disk; | |
3167 | struct request_queue *q; | |
593a9e7b | 3168 | u64 segment_size; |
602adf40 | 3169 | |
602adf40 | 3170 | /* create gendisk info */ |
602adf40 YS |
3171 | disk = alloc_disk(RBD_MINORS_PER_MAJOR); |
3172 | if (!disk) | |
1fcdb8aa | 3173 | return -ENOMEM; |
602adf40 | 3174 | |
f0f8cef5 | 3175 | snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d", |
de71a297 | 3176 | rbd_dev->dev_id); |
602adf40 YS |
3177 | disk->major = rbd_dev->major; |
3178 | disk->first_minor = 0; | |
3179 | disk->fops = &rbd_bd_ops; | |
3180 | disk->private_data = rbd_dev; | |
3181 | ||
bf0d5f50 | 3182 | q = blk_init_queue(rbd_request_fn, &rbd_dev->lock); |
602adf40 YS |
3183 | if (!q) |
3184 | goto out_disk; | |
029bcbd8 | 3185 | |
593a9e7b AE |
3186 | /* We use the default size, but let's be explicit about it. */ |
3187 | blk_queue_physical_block_size(q, SECTOR_SIZE); | |
3188 | ||
029bcbd8 | 3189 | /* set io sizes to object size */ |
593a9e7b AE |
3190 | segment_size = rbd_obj_bytes(&rbd_dev->header); |
3191 | blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE); | |
3192 | blk_queue_max_segment_size(q, segment_size); | |
3193 | blk_queue_io_min(q, segment_size); | |
3194 | blk_queue_io_opt(q, segment_size); | |
029bcbd8 | 3195 | |
602adf40 YS |
3196 | blk_queue_merge_bvec(q, rbd_merge_bvec); |
3197 | disk->queue = q; | |
3198 | ||
3199 | q->queuedata = rbd_dev; | |
3200 | ||
3201 | rbd_dev->disk = disk; | |
602adf40 | 3202 | |
602adf40 | 3203 | return 0; |
602adf40 YS |
3204 | out_disk: |
3205 | put_disk(disk); | |
1fcdb8aa AE |
3206 | |
3207 | return -ENOMEM; | |
602adf40 YS |
3208 | } |
3209 | ||
dfc5606d YS |
3210 | /* |
3211 | sysfs | |
3212 | */ | |
3213 | ||
593a9e7b AE |
3214 | static struct rbd_device *dev_to_rbd_dev(struct device *dev) |
3215 | { | |
3216 | return container_of(dev, struct rbd_device, dev); | |
3217 | } | |
3218 | ||
dfc5606d YS |
3219 | static ssize_t rbd_size_show(struct device *dev, |
3220 | struct device_attribute *attr, char *buf) | |
3221 | { | |
593a9e7b | 3222 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
a51aa0c0 | 3223 | |
fc71d833 AE |
3224 | return sprintf(buf, "%llu\n", |
3225 | (unsigned long long)rbd_dev->mapping.size); | |
dfc5606d YS |
3226 | } |
3227 | ||
34b13184 AE |
3228 | /* |
3229 | * Note this shows the features for whatever's mapped, which is not | |
3230 | * necessarily the base image. | |
3231 | */ | |
3232 | static ssize_t rbd_features_show(struct device *dev, | |
3233 | struct device_attribute *attr, char *buf) | |
3234 | { | |
3235 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); | |
3236 | ||
3237 | return sprintf(buf, "0x%016llx\n", | |
fc71d833 | 3238 | (unsigned long long)rbd_dev->mapping.features); |
34b13184 AE |
3239 | } |
3240 | ||
dfc5606d YS |
3241 | static ssize_t rbd_major_show(struct device *dev, |
3242 | struct device_attribute *attr, char *buf) | |
3243 | { | |
593a9e7b | 3244 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
602adf40 | 3245 | |
fc71d833 AE |
3246 | if (rbd_dev->major) |
3247 | return sprintf(buf, "%d\n", rbd_dev->major); | |
3248 | ||
3249 | return sprintf(buf, "(none)\n"); | |
3250 | ||
dfc5606d YS |
3251 | } |
3252 | ||
3253 | static ssize_t rbd_client_id_show(struct device *dev, | |
3254 | struct device_attribute *attr, char *buf) | |
602adf40 | 3255 | { |
593a9e7b | 3256 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
dfc5606d | 3257 | |
1dbb4399 AE |
3258 | return sprintf(buf, "client%lld\n", |
3259 | ceph_client_id(rbd_dev->rbd_client->client)); | |
602adf40 YS |
3260 | } |
3261 | ||
dfc5606d YS |
3262 | static ssize_t rbd_pool_show(struct device *dev, |
3263 | struct device_attribute *attr, char *buf) | |
602adf40 | 3264 | { |
593a9e7b | 3265 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
dfc5606d | 3266 | |
0d7dbfce | 3267 | return sprintf(buf, "%s\n", rbd_dev->spec->pool_name); |
dfc5606d YS |
3268 | } |
3269 | ||
9bb2f334 AE |
3270 | static ssize_t rbd_pool_id_show(struct device *dev, |
3271 | struct device_attribute *attr, char *buf) | |
3272 | { | |
3273 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); | |
3274 | ||
0d7dbfce | 3275 | return sprintf(buf, "%llu\n", |
fc71d833 | 3276 | (unsigned long long) rbd_dev->spec->pool_id); |
9bb2f334 AE |
3277 | } |
3278 | ||
dfc5606d YS |
3279 | static ssize_t rbd_name_show(struct device *dev, |
3280 | struct device_attribute *attr, char *buf) | |
3281 | { | |
593a9e7b | 3282 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
dfc5606d | 3283 | |
a92ffdf8 AE |
3284 | if (rbd_dev->spec->image_name) |
3285 | return sprintf(buf, "%s\n", rbd_dev->spec->image_name); | |
3286 | ||
3287 | return sprintf(buf, "(unknown)\n"); | |
dfc5606d YS |
3288 | } |
3289 | ||
589d30e0 AE |
3290 | static ssize_t rbd_image_id_show(struct device *dev, |
3291 | struct device_attribute *attr, char *buf) | |
3292 | { | |
3293 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); | |
3294 | ||
0d7dbfce | 3295 | return sprintf(buf, "%s\n", rbd_dev->spec->image_id); |
589d30e0 AE |
3296 | } |
3297 | ||
34b13184 AE |
3298 | /* |
3299 | * Shows the name of the currently-mapped snapshot (or | |
3300 | * RBD_SNAP_HEAD_NAME for the base image). | |
3301 | */ | |
dfc5606d YS |
3302 | static ssize_t rbd_snap_show(struct device *dev, |
3303 | struct device_attribute *attr, | |
3304 | char *buf) | |
3305 | { | |
593a9e7b | 3306 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
dfc5606d | 3307 | |
0d7dbfce | 3308 | return sprintf(buf, "%s\n", rbd_dev->spec->snap_name); |
dfc5606d YS |
3309 | } |
3310 | ||
86b00e0d AE |
3311 | /* |
3312 | * For an rbd v2 image, shows the pool id, image id, and snapshot id | |
3313 | * for the parent image. If there is no parent, simply shows | |
3314 | * "(no parent image)". | |
3315 | */ | |
3316 | static ssize_t rbd_parent_show(struct device *dev, | |
3317 | struct device_attribute *attr, | |
3318 | char *buf) | |
3319 | { | |
3320 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); | |
3321 | struct rbd_spec *spec = rbd_dev->parent_spec; | |
3322 | int count; | |
3323 | char *bufp = buf; | |
3324 | ||
3325 | if (!spec) | |
3326 | return sprintf(buf, "(no parent image)\n"); | |
3327 | ||
3328 | count = sprintf(bufp, "pool_id %llu\npool_name %s\n", | |
3329 | (unsigned long long) spec->pool_id, spec->pool_name); | |
3330 | if (count < 0) | |
3331 | return count; | |
3332 | bufp += count; | |
3333 | ||
3334 | count = sprintf(bufp, "image_id %s\nimage_name %s\n", spec->image_id, | |
3335 | spec->image_name ? spec->image_name : "(unknown)"); | |
3336 | if (count < 0) | |
3337 | return count; | |
3338 | bufp += count; | |
3339 | ||
3340 | count = sprintf(bufp, "snap_id %llu\nsnap_name %s\n", | |
3341 | (unsigned long long) spec->snap_id, spec->snap_name); | |
3342 | if (count < 0) | |
3343 | return count; | |
3344 | bufp += count; | |
3345 | ||
3346 | count = sprintf(bufp, "overlap %llu\n", rbd_dev->parent_overlap); | |
3347 | if (count < 0) | |
3348 | return count; | |
3349 | bufp += count; | |
3350 | ||
3351 | return (ssize_t) (bufp - buf); | |
3352 | } | |
3353 | ||
dfc5606d YS |
3354 | static ssize_t rbd_image_refresh(struct device *dev, |
3355 | struct device_attribute *attr, | |
3356 | const char *buf, | |
3357 | size_t size) | |
3358 | { | |
593a9e7b | 3359 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
b813623a | 3360 | int ret; |
602adf40 | 3361 | |
cc4a38bd | 3362 | ret = rbd_dev_refresh(rbd_dev); |
e627db08 AE |
3363 | if (ret) |
3364 | rbd_warn(rbd_dev, ": manual header refresh error (%d)\n", ret); | |
b813623a AE |
3365 | |
3366 | return ret < 0 ? ret : size; | |
dfc5606d | 3367 | } |
602adf40 | 3368 | |
dfc5606d | 3369 | static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL); |
34b13184 | 3370 | static DEVICE_ATTR(features, S_IRUGO, rbd_features_show, NULL); |
dfc5606d YS |
3371 | static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL); |
3372 | static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL); | |
3373 | static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL); | |
9bb2f334 | 3374 | static DEVICE_ATTR(pool_id, S_IRUGO, rbd_pool_id_show, NULL); |
dfc5606d | 3375 | static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL); |
589d30e0 | 3376 | static DEVICE_ATTR(image_id, S_IRUGO, rbd_image_id_show, NULL); |
dfc5606d YS |
3377 | static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh); |
3378 | static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL); | |
86b00e0d | 3379 | static DEVICE_ATTR(parent, S_IRUGO, rbd_parent_show, NULL); |
dfc5606d YS |
3380 | |
3381 | static struct attribute *rbd_attrs[] = { | |
3382 | &dev_attr_size.attr, | |
34b13184 | 3383 | &dev_attr_features.attr, |
dfc5606d YS |
3384 | &dev_attr_major.attr, |
3385 | &dev_attr_client_id.attr, | |
3386 | &dev_attr_pool.attr, | |
9bb2f334 | 3387 | &dev_attr_pool_id.attr, |
dfc5606d | 3388 | &dev_attr_name.attr, |
589d30e0 | 3389 | &dev_attr_image_id.attr, |
dfc5606d | 3390 | &dev_attr_current_snap.attr, |
86b00e0d | 3391 | &dev_attr_parent.attr, |
dfc5606d | 3392 | &dev_attr_refresh.attr, |
dfc5606d YS |
3393 | NULL |
3394 | }; | |
3395 | ||
3396 | static struct attribute_group rbd_attr_group = { | |
3397 | .attrs = rbd_attrs, | |
3398 | }; | |
3399 | ||
3400 | static const struct attribute_group *rbd_attr_groups[] = { | |
3401 | &rbd_attr_group, | |
3402 | NULL | |
3403 | }; | |
3404 | ||
3405 | static void rbd_sysfs_dev_release(struct device *dev) | |
3406 | { | |
3407 | } | |
3408 | ||
3409 | static struct device_type rbd_device_type = { | |
3410 | .name = "rbd", | |
3411 | .groups = rbd_attr_groups, | |
3412 | .release = rbd_sysfs_dev_release, | |
3413 | }; | |
3414 | ||
8b8fb99c AE |
3415 | static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec) |
3416 | { | |
3417 | kref_get(&spec->kref); | |
3418 | ||
3419 | return spec; | |
3420 | } | |
3421 | ||
3422 | static void rbd_spec_free(struct kref *kref); | |
3423 | static void rbd_spec_put(struct rbd_spec *spec) | |
3424 | { | |
3425 | if (spec) | |
3426 | kref_put(&spec->kref, rbd_spec_free); | |
3427 | } | |
3428 | ||
3429 | static struct rbd_spec *rbd_spec_alloc(void) | |
3430 | { | |
3431 | struct rbd_spec *spec; | |
3432 | ||
3433 | spec = kzalloc(sizeof (*spec), GFP_KERNEL); | |
3434 | if (!spec) | |
3435 | return NULL; | |
3436 | kref_init(&spec->kref); | |
3437 | ||
8b8fb99c AE |
3438 | return spec; |
3439 | } | |
3440 | ||
3441 | static void rbd_spec_free(struct kref *kref) | |
3442 | { | |
3443 | struct rbd_spec *spec = container_of(kref, struct rbd_spec, kref); | |
3444 | ||
3445 | kfree(spec->pool_name); | |
3446 | kfree(spec->image_id); | |
3447 | kfree(spec->image_name); | |
3448 | kfree(spec->snap_name); | |
3449 | kfree(spec); | |
3450 | } | |
3451 | ||
cc344fa1 | 3452 | static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc, |
c53d5893 AE |
3453 | struct rbd_spec *spec) |
3454 | { | |
3455 | struct rbd_device *rbd_dev; | |
3456 | ||
3457 | rbd_dev = kzalloc(sizeof (*rbd_dev), GFP_KERNEL); | |
3458 | if (!rbd_dev) | |
3459 | return NULL; | |
3460 | ||
3461 | spin_lock_init(&rbd_dev->lock); | |
6d292906 | 3462 | rbd_dev->flags = 0; |
c53d5893 | 3463 | INIT_LIST_HEAD(&rbd_dev->node); |
c53d5893 AE |
3464 | init_rwsem(&rbd_dev->header_rwsem); |
3465 | ||
3466 | rbd_dev->spec = spec; | |
3467 | rbd_dev->rbd_client = rbdc; | |
3468 | ||
0903e875 AE |
3469 | /* Initialize the layout used for all rbd requests */ |
3470 | ||
3471 | rbd_dev->layout.fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); | |
3472 | rbd_dev->layout.fl_stripe_count = cpu_to_le32(1); | |
3473 | rbd_dev->layout.fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); | |
3474 | rbd_dev->layout.fl_pg_pool = cpu_to_le32((u32) spec->pool_id); | |
3475 | ||
c53d5893 AE |
3476 | return rbd_dev; |
3477 | } | |
3478 | ||
3479 | static void rbd_dev_destroy(struct rbd_device *rbd_dev) | |
3480 | { | |
c53d5893 AE |
3481 | rbd_put_client(rbd_dev->rbd_client); |
3482 | rbd_spec_put(rbd_dev->spec); | |
3483 | kfree(rbd_dev); | |
3484 | } | |
3485 | ||
9d475de5 AE |
3486 | /* |
3487 | * Get the size and object order for an image snapshot, or if | |
3488 | * snap_id is CEPH_NOSNAP, gets this information for the base | |
3489 | * image. | |
3490 | */ | |
3491 | static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id, | |
3492 | u8 *order, u64 *snap_size) | |
3493 | { | |
3494 | __le64 snapid = cpu_to_le64(snap_id); | |
3495 | int ret; | |
3496 | struct { | |
3497 | u8 order; | |
3498 | __le64 size; | |
3499 | } __attribute__ ((packed)) size_buf = { 0 }; | |
3500 | ||
36be9a76 | 3501 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
9d475de5 | 3502 | "rbd", "get_size", |
4157976b | 3503 | &snapid, sizeof (snapid), |
e2a58ee5 | 3504 | &size_buf, sizeof (size_buf)); |
36be9a76 | 3505 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
9d475de5 AE |
3506 | if (ret < 0) |
3507 | return ret; | |
57385b51 AE |
3508 | if (ret < sizeof (size_buf)) |
3509 | return -ERANGE; | |
9d475de5 | 3510 | |
c86f86e9 AE |
3511 | if (order) |
3512 | *order = size_buf.order; | |
9d475de5 AE |
3513 | *snap_size = le64_to_cpu(size_buf.size); |
3514 | ||
3515 | dout(" snap_id 0x%016llx order = %u, snap_size = %llu\n", | |
57385b51 AE |
3516 | (unsigned long long)snap_id, (unsigned int)*order, |
3517 | (unsigned long long)*snap_size); | |
9d475de5 AE |
3518 | |
3519 | return 0; | |
3520 | } | |
3521 | ||
3522 | static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev) | |
3523 | { | |
3524 | return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP, | |
3525 | &rbd_dev->header.obj_order, | |
3526 | &rbd_dev->header.image_size); | |
3527 | } | |
3528 | ||
1e130199 AE |
3529 | static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev) |
3530 | { | |
3531 | void *reply_buf; | |
3532 | int ret; | |
3533 | void *p; | |
3534 | ||
3535 | reply_buf = kzalloc(RBD_OBJ_PREFIX_LEN_MAX, GFP_KERNEL); | |
3536 | if (!reply_buf) | |
3537 | return -ENOMEM; | |
3538 | ||
36be9a76 | 3539 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
4157976b | 3540 | "rbd", "get_object_prefix", NULL, 0, |
e2a58ee5 | 3541 | reply_buf, RBD_OBJ_PREFIX_LEN_MAX); |
36be9a76 | 3542 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
1e130199 AE |
3543 | if (ret < 0) |
3544 | goto out; | |
3545 | ||
3546 | p = reply_buf; | |
3547 | rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p, | |
57385b51 AE |
3548 | p + ret, NULL, GFP_NOIO); |
3549 | ret = 0; | |
1e130199 AE |
3550 | |
3551 | if (IS_ERR(rbd_dev->header.object_prefix)) { | |
3552 | ret = PTR_ERR(rbd_dev->header.object_prefix); | |
3553 | rbd_dev->header.object_prefix = NULL; | |
3554 | } else { | |
3555 | dout(" object_prefix = %s\n", rbd_dev->header.object_prefix); | |
3556 | } | |
1e130199 AE |
3557 | out: |
3558 | kfree(reply_buf); | |
3559 | ||
3560 | return ret; | |
3561 | } | |
3562 | ||
b1b5402a AE |
3563 | static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id, |
3564 | u64 *snap_features) | |
3565 | { | |
3566 | __le64 snapid = cpu_to_le64(snap_id); | |
3567 | struct { | |
3568 | __le64 features; | |
3569 | __le64 incompat; | |
4157976b | 3570 | } __attribute__ ((packed)) features_buf = { 0 }; |
d889140c | 3571 | u64 incompat; |
b1b5402a AE |
3572 | int ret; |
3573 | ||
36be9a76 | 3574 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
b1b5402a | 3575 | "rbd", "get_features", |
4157976b | 3576 | &snapid, sizeof (snapid), |
e2a58ee5 | 3577 | &features_buf, sizeof (features_buf)); |
36be9a76 | 3578 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
b1b5402a AE |
3579 | if (ret < 0) |
3580 | return ret; | |
57385b51 AE |
3581 | if (ret < sizeof (features_buf)) |
3582 | return -ERANGE; | |
d889140c AE |
3583 | |
3584 | incompat = le64_to_cpu(features_buf.incompat); | |
5cbf6f12 | 3585 | if (incompat & ~RBD_FEATURES_SUPPORTED) |
b8f5c6ed | 3586 | return -ENXIO; |
d889140c | 3587 | |
b1b5402a AE |
3588 | *snap_features = le64_to_cpu(features_buf.features); |
3589 | ||
3590 | dout(" snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n", | |
57385b51 AE |
3591 | (unsigned long long)snap_id, |
3592 | (unsigned long long)*snap_features, | |
3593 | (unsigned long long)le64_to_cpu(features_buf.incompat)); | |
b1b5402a AE |
3594 | |
3595 | return 0; | |
3596 | } | |
3597 | ||
3598 | static int rbd_dev_v2_features(struct rbd_device *rbd_dev) | |
3599 | { | |
3600 | return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP, | |
3601 | &rbd_dev->header.features); | |
3602 | } | |
3603 | ||
86b00e0d AE |
3604 | static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev) |
3605 | { | |
3606 | struct rbd_spec *parent_spec; | |
3607 | size_t size; | |
3608 | void *reply_buf = NULL; | |
3609 | __le64 snapid; | |
3610 | void *p; | |
3611 | void *end; | |
3612 | char *image_id; | |
3613 | u64 overlap; | |
86b00e0d AE |
3614 | int ret; |
3615 | ||
3616 | parent_spec = rbd_spec_alloc(); | |
3617 | if (!parent_spec) | |
3618 | return -ENOMEM; | |
3619 | ||
3620 | size = sizeof (__le64) + /* pool_id */ | |
3621 | sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX + /* image_id */ | |
3622 | sizeof (__le64) + /* snap_id */ | |
3623 | sizeof (__le64); /* overlap */ | |
3624 | reply_buf = kmalloc(size, GFP_KERNEL); | |
3625 | if (!reply_buf) { | |
3626 | ret = -ENOMEM; | |
3627 | goto out_err; | |
3628 | } | |
3629 | ||
3630 | snapid = cpu_to_le64(CEPH_NOSNAP); | |
36be9a76 | 3631 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
86b00e0d | 3632 | "rbd", "get_parent", |
4157976b | 3633 | &snapid, sizeof (snapid), |
e2a58ee5 | 3634 | reply_buf, size); |
36be9a76 | 3635 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
86b00e0d AE |
3636 | if (ret < 0) |
3637 | goto out_err; | |
3638 | ||
86b00e0d | 3639 | p = reply_buf; |
57385b51 AE |
3640 | end = reply_buf + ret; |
3641 | ret = -ERANGE; | |
86b00e0d AE |
3642 | ceph_decode_64_safe(&p, end, parent_spec->pool_id, out_err); |
3643 | if (parent_spec->pool_id == CEPH_NOPOOL) | |
3644 | goto out; /* No parent? No problem. */ | |
3645 | ||
0903e875 AE |
3646 | /* The ceph file layout needs to fit pool id in 32 bits */ |
3647 | ||
3648 | ret = -EIO; | |
c0cd10db AE |
3649 | if (parent_spec->pool_id > (u64)U32_MAX) { |
3650 | rbd_warn(NULL, "parent pool id too large (%llu > %u)\n", | |
3651 | (unsigned long long)parent_spec->pool_id, U32_MAX); | |
57385b51 | 3652 | goto out_err; |
c0cd10db | 3653 | } |
0903e875 | 3654 | |
979ed480 | 3655 | image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL); |
86b00e0d AE |
3656 | if (IS_ERR(image_id)) { |
3657 | ret = PTR_ERR(image_id); | |
3658 | goto out_err; | |
3659 | } | |
3660 | parent_spec->image_id = image_id; | |
3661 | ceph_decode_64_safe(&p, end, parent_spec->snap_id, out_err); | |
3662 | ceph_decode_64_safe(&p, end, overlap, out_err); | |
3663 | ||
3664 | rbd_dev->parent_overlap = overlap; | |
3665 | rbd_dev->parent_spec = parent_spec; | |
3666 | parent_spec = NULL; /* rbd_dev now owns this */ | |
3667 | out: | |
3668 | ret = 0; | |
3669 | out_err: | |
3670 | kfree(reply_buf); | |
3671 | rbd_spec_put(parent_spec); | |
3672 | ||
3673 | return ret; | |
3674 | } | |
3675 | ||
cc070d59 AE |
3676 | static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev) |
3677 | { | |
3678 | struct { | |
3679 | __le64 stripe_unit; | |
3680 | __le64 stripe_count; | |
3681 | } __attribute__ ((packed)) striping_info_buf = { 0 }; | |
3682 | size_t size = sizeof (striping_info_buf); | |
3683 | void *p; | |
3684 | u64 obj_size; | |
3685 | u64 stripe_unit; | |
3686 | u64 stripe_count; | |
3687 | int ret; | |
3688 | ||
3689 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, | |
3690 | "rbd", "get_stripe_unit_count", NULL, 0, | |
e2a58ee5 | 3691 | (char *)&striping_info_buf, size); |
cc070d59 AE |
3692 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
3693 | if (ret < 0) | |
3694 | return ret; | |
3695 | if (ret < size) | |
3696 | return -ERANGE; | |
3697 | ||
3698 | /* | |
3699 | * We don't actually support the "fancy striping" feature | |
3700 | * (STRIPINGV2) yet, but if the striping sizes are the | |
3701 | * defaults the behavior is the same as before. So find | |
3702 | * out, and only fail if the image has non-default values. | |
3703 | */ | |
3704 | ret = -EINVAL; | |
3705 | obj_size = (u64)1 << rbd_dev->header.obj_order; | |
3706 | p = &striping_info_buf; | |
3707 | stripe_unit = ceph_decode_64(&p); | |
3708 | if (stripe_unit != obj_size) { | |
3709 | rbd_warn(rbd_dev, "unsupported stripe unit " | |
3710 | "(got %llu want %llu)", | |
3711 | stripe_unit, obj_size); | |
3712 | return -EINVAL; | |
3713 | } | |
3714 | stripe_count = ceph_decode_64(&p); | |
3715 | if (stripe_count != 1) { | |
3716 | rbd_warn(rbd_dev, "unsupported stripe count " | |
3717 | "(got %llu want 1)", stripe_count); | |
3718 | return -EINVAL; | |
3719 | } | |
500d0c0f AE |
3720 | rbd_dev->header.stripe_unit = stripe_unit; |
3721 | rbd_dev->header.stripe_count = stripe_count; | |
cc070d59 AE |
3722 | |
3723 | return 0; | |
3724 | } | |
3725 | ||
9e15b77d AE |
3726 | static char *rbd_dev_image_name(struct rbd_device *rbd_dev) |
3727 | { | |
3728 | size_t image_id_size; | |
3729 | char *image_id; | |
3730 | void *p; | |
3731 | void *end; | |
3732 | size_t size; | |
3733 | void *reply_buf = NULL; | |
3734 | size_t len = 0; | |
3735 | char *image_name = NULL; | |
3736 | int ret; | |
3737 | ||
3738 | rbd_assert(!rbd_dev->spec->image_name); | |
3739 | ||
69e7a02f AE |
3740 | len = strlen(rbd_dev->spec->image_id); |
3741 | image_id_size = sizeof (__le32) + len; | |
9e15b77d AE |
3742 | image_id = kmalloc(image_id_size, GFP_KERNEL); |
3743 | if (!image_id) | |
3744 | return NULL; | |
3745 | ||
3746 | p = image_id; | |
4157976b | 3747 | end = image_id + image_id_size; |
57385b51 | 3748 | ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len); |
9e15b77d AE |
3749 | |
3750 | size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX; | |
3751 | reply_buf = kmalloc(size, GFP_KERNEL); | |
3752 | if (!reply_buf) | |
3753 | goto out; | |
3754 | ||
36be9a76 | 3755 | ret = rbd_obj_method_sync(rbd_dev, RBD_DIRECTORY, |
9e15b77d AE |
3756 | "rbd", "dir_get_name", |
3757 | image_id, image_id_size, | |
e2a58ee5 | 3758 | reply_buf, size); |
9e15b77d AE |
3759 | if (ret < 0) |
3760 | goto out; | |
3761 | p = reply_buf; | |
f40eb349 AE |
3762 | end = reply_buf + ret; |
3763 | ||
9e15b77d AE |
3764 | image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL); |
3765 | if (IS_ERR(image_name)) | |
3766 | image_name = NULL; | |
3767 | else | |
3768 | dout("%s: name is %s len is %zd\n", __func__, image_name, len); | |
3769 | out: | |
3770 | kfree(reply_buf); | |
3771 | kfree(image_id); | |
3772 | ||
3773 | return image_name; | |
3774 | } | |
3775 | ||
2ad3d716 AE |
3776 | static u64 rbd_v1_snap_id_by_name(struct rbd_device *rbd_dev, const char *name) |
3777 | { | |
3778 | struct ceph_snap_context *snapc = rbd_dev->header.snapc; | |
3779 | const char *snap_name; | |
3780 | u32 which = 0; | |
3781 | ||
3782 | /* Skip over names until we find the one we are looking for */ | |
3783 | ||
3784 | snap_name = rbd_dev->header.snap_names; | |
3785 | while (which < snapc->num_snaps) { | |
3786 | if (!strcmp(name, snap_name)) | |
3787 | return snapc->snaps[which]; | |
3788 | snap_name += strlen(snap_name) + 1; | |
3789 | which++; | |
3790 | } | |
3791 | return CEPH_NOSNAP; | |
3792 | } | |
3793 | ||
3794 | static u64 rbd_v2_snap_id_by_name(struct rbd_device *rbd_dev, const char *name) | |
3795 | { | |
3796 | struct ceph_snap_context *snapc = rbd_dev->header.snapc; | |
3797 | u32 which; | |
3798 | bool found = false; | |
3799 | u64 snap_id; | |
3800 | ||
3801 | for (which = 0; !found && which < snapc->num_snaps; which++) { | |
3802 | const char *snap_name; | |
3803 | ||
3804 | snap_id = snapc->snaps[which]; | |
3805 | snap_name = rbd_dev_v2_snap_name(rbd_dev, snap_id); | |
3806 | if (IS_ERR(snap_name)) | |
3807 | break; | |
3808 | found = !strcmp(name, snap_name); | |
3809 | kfree(snap_name); | |
3810 | } | |
3811 | return found ? snap_id : CEPH_NOSNAP; | |
3812 | } | |
3813 | ||
3814 | /* | |
3815 | * Assumes name is never RBD_SNAP_HEAD_NAME; returns CEPH_NOSNAP if | |
3816 | * no snapshot by that name is found, or if an error occurs. | |
3817 | */ | |
3818 | static u64 rbd_snap_id_by_name(struct rbd_device *rbd_dev, const char *name) | |
3819 | { | |
3820 | if (rbd_dev->image_format == 1) | |
3821 | return rbd_v1_snap_id_by_name(rbd_dev, name); | |
3822 | ||
3823 | return rbd_v2_snap_id_by_name(rbd_dev, name); | |
3824 | } | |
3825 | ||
9e15b77d | 3826 | /* |
2e9f7f1c AE |
3827 | * When an rbd image has a parent image, it is identified by the |
3828 | * pool, image, and snapshot ids (not names). This function fills | |
3829 | * in the names for those ids. (It's OK if we can't figure out the | |
3830 | * name for an image id, but the pool and snapshot ids should always | |
3831 | * exist and have names.) All names in an rbd spec are dynamically | |
3832 | * allocated. | |
e1d4213f AE |
3833 | * |
3834 | * When an image being mapped (not a parent) is probed, we have the | |
3835 | * pool name and pool id, image name and image id, and the snapshot | |
3836 | * name. The only thing we're missing is the snapshot id. | |
9e15b77d | 3837 | */ |
2e9f7f1c | 3838 | static int rbd_dev_spec_update(struct rbd_device *rbd_dev) |
9e15b77d | 3839 | { |
2e9f7f1c AE |
3840 | struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; |
3841 | struct rbd_spec *spec = rbd_dev->spec; | |
3842 | const char *pool_name; | |
3843 | const char *image_name; | |
3844 | const char *snap_name; | |
9e15b77d AE |
3845 | int ret; |
3846 | ||
e1d4213f AE |
3847 | /* |
3848 | * An image being mapped will have the pool name (etc.), but | |
3849 | * we need to look up the snapshot id. | |
3850 | */ | |
2e9f7f1c AE |
3851 | if (spec->pool_name) { |
3852 | if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) { | |
2ad3d716 | 3853 | u64 snap_id; |
e1d4213f | 3854 | |
2ad3d716 AE |
3855 | snap_id = rbd_snap_id_by_name(rbd_dev, spec->snap_name); |
3856 | if (snap_id == CEPH_NOSNAP) | |
e1d4213f | 3857 | return -ENOENT; |
2ad3d716 | 3858 | spec->snap_id = snap_id; |
e1d4213f | 3859 | } else { |
2e9f7f1c | 3860 | spec->snap_id = CEPH_NOSNAP; |
e1d4213f AE |
3861 | } |
3862 | ||
3863 | return 0; | |
3864 | } | |
9e15b77d | 3865 | |
2e9f7f1c | 3866 | /* Get the pool name; we have to make our own copy of this */ |
9e15b77d | 3867 | |
2e9f7f1c AE |
3868 | pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id); |
3869 | if (!pool_name) { | |
3870 | rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id); | |
935dc89f AE |
3871 | return -EIO; |
3872 | } | |
2e9f7f1c AE |
3873 | pool_name = kstrdup(pool_name, GFP_KERNEL); |
3874 | if (!pool_name) | |
9e15b77d AE |
3875 | return -ENOMEM; |
3876 | ||
3877 | /* Fetch the image name; tolerate failure here */ | |
3878 | ||
2e9f7f1c AE |
3879 | image_name = rbd_dev_image_name(rbd_dev); |
3880 | if (!image_name) | |
06ecc6cb | 3881 | rbd_warn(rbd_dev, "unable to get image name"); |
9e15b77d | 3882 | |
2e9f7f1c | 3883 | /* Look up the snapshot name, and make a copy */ |
9e15b77d | 3884 | |
2e9f7f1c | 3885 | snap_name = rbd_snap_name(rbd_dev, spec->snap_id); |
2e9f7f1c AE |
3886 | if (!snap_name) { |
3887 | ret = -ENOMEM; | |
9e15b77d | 3888 | goto out_err; |
2e9f7f1c AE |
3889 | } |
3890 | ||
3891 | spec->pool_name = pool_name; | |
3892 | spec->image_name = image_name; | |
3893 | spec->snap_name = snap_name; | |
9e15b77d AE |
3894 | |
3895 | return 0; | |
3896 | out_err: | |
2e9f7f1c AE |
3897 | kfree(image_name); |
3898 | kfree(pool_name); | |
9e15b77d AE |
3899 | |
3900 | return ret; | |
3901 | } | |
3902 | ||
cc4a38bd | 3903 | static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev) |
35d489f9 AE |
3904 | { |
3905 | size_t size; | |
3906 | int ret; | |
3907 | void *reply_buf; | |
3908 | void *p; | |
3909 | void *end; | |
3910 | u64 seq; | |
3911 | u32 snap_count; | |
3912 | struct ceph_snap_context *snapc; | |
3913 | u32 i; | |
3914 | ||
3915 | /* | |
3916 | * We'll need room for the seq value (maximum snapshot id), | |
3917 | * snapshot count, and array of that many snapshot ids. | |
3918 | * For now we have a fixed upper limit on the number we're | |
3919 | * prepared to receive. | |
3920 | */ | |
3921 | size = sizeof (__le64) + sizeof (__le32) + | |
3922 | RBD_MAX_SNAP_COUNT * sizeof (__le64); | |
3923 | reply_buf = kzalloc(size, GFP_KERNEL); | |
3924 | if (!reply_buf) | |
3925 | return -ENOMEM; | |
3926 | ||
36be9a76 | 3927 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
4157976b | 3928 | "rbd", "get_snapcontext", NULL, 0, |
e2a58ee5 | 3929 | reply_buf, size); |
36be9a76 | 3930 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
35d489f9 AE |
3931 | if (ret < 0) |
3932 | goto out; | |
3933 | ||
35d489f9 | 3934 | p = reply_buf; |
57385b51 AE |
3935 | end = reply_buf + ret; |
3936 | ret = -ERANGE; | |
35d489f9 AE |
3937 | ceph_decode_64_safe(&p, end, seq, out); |
3938 | ceph_decode_32_safe(&p, end, snap_count, out); | |
3939 | ||
3940 | /* | |
3941 | * Make sure the reported number of snapshot ids wouldn't go | |
3942 | * beyond the end of our buffer. But before checking that, | |
3943 | * make sure the computed size of the snapshot context we | |
3944 | * allocate is representable in a size_t. | |
3945 | */ | |
3946 | if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context)) | |
3947 | / sizeof (u64)) { | |
3948 | ret = -EINVAL; | |
3949 | goto out; | |
3950 | } | |
3951 | if (!ceph_has_room(&p, end, snap_count * sizeof (__le64))) | |
3952 | goto out; | |
468521c1 | 3953 | ret = 0; |
35d489f9 | 3954 | |
812164f8 | 3955 | snapc = ceph_create_snap_context(snap_count, GFP_KERNEL); |
35d489f9 AE |
3956 | if (!snapc) { |
3957 | ret = -ENOMEM; | |
3958 | goto out; | |
3959 | } | |
35d489f9 | 3960 | snapc->seq = seq; |
35d489f9 AE |
3961 | for (i = 0; i < snap_count; i++) |
3962 | snapc->snaps[i] = ceph_decode_64(&p); | |
3963 | ||
49ece554 | 3964 | ceph_put_snap_context(rbd_dev->header.snapc); |
35d489f9 AE |
3965 | rbd_dev->header.snapc = snapc; |
3966 | ||
3967 | dout(" snap context seq = %llu, snap_count = %u\n", | |
57385b51 | 3968 | (unsigned long long)seq, (unsigned int)snap_count); |
35d489f9 AE |
3969 | out: |
3970 | kfree(reply_buf); | |
3971 | ||
57385b51 | 3972 | return ret; |
35d489f9 AE |
3973 | } |
3974 | ||
54cac61f AE |
3975 | static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, |
3976 | u64 snap_id) | |
b8b1e2db AE |
3977 | { |
3978 | size_t size; | |
3979 | void *reply_buf; | |
54cac61f | 3980 | __le64 snapid; |
b8b1e2db AE |
3981 | int ret; |
3982 | void *p; | |
3983 | void *end; | |
b8b1e2db AE |
3984 | char *snap_name; |
3985 | ||
3986 | size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN; | |
3987 | reply_buf = kmalloc(size, GFP_KERNEL); | |
3988 | if (!reply_buf) | |
3989 | return ERR_PTR(-ENOMEM); | |
3990 | ||
54cac61f | 3991 | snapid = cpu_to_le64(snap_id); |
36be9a76 | 3992 | ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name, |
b8b1e2db | 3993 | "rbd", "get_snapshot_name", |
54cac61f | 3994 | &snapid, sizeof (snapid), |
e2a58ee5 | 3995 | reply_buf, size); |
36be9a76 | 3996 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
f40eb349 AE |
3997 | if (ret < 0) { |
3998 | snap_name = ERR_PTR(ret); | |
b8b1e2db | 3999 | goto out; |
f40eb349 | 4000 | } |
b8b1e2db AE |
4001 | |
4002 | p = reply_buf; | |
f40eb349 | 4003 | end = reply_buf + ret; |
e5c35534 | 4004 | snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL); |
f40eb349 | 4005 | if (IS_ERR(snap_name)) |
b8b1e2db | 4006 | goto out; |
b8b1e2db | 4007 | |
f40eb349 | 4008 | dout(" snap_id 0x%016llx snap_name = %s\n", |
54cac61f | 4009 | (unsigned long long)snap_id, snap_name); |
b8b1e2db AE |
4010 | out: |
4011 | kfree(reply_buf); | |
4012 | ||
f40eb349 | 4013 | return snap_name; |
b8b1e2db AE |
4014 | } |
4015 | ||
cc4a38bd | 4016 | static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev) |
117973fb AE |
4017 | { |
4018 | int ret; | |
117973fb AE |
4019 | |
4020 | down_write(&rbd_dev->header_rwsem); | |
4021 | ||
117973fb AE |
4022 | ret = rbd_dev_v2_image_size(rbd_dev); |
4023 | if (ret) | |
4024 | goto out; | |
29334ba4 AE |
4025 | if (rbd_dev->spec->snap_id == CEPH_NOSNAP) |
4026 | if (rbd_dev->mapping.size != rbd_dev->header.image_size) | |
4027 | rbd_dev->mapping.size = rbd_dev->header.image_size; | |
117973fb | 4028 | |
cc4a38bd | 4029 | ret = rbd_dev_v2_snap_context(rbd_dev); |
117973fb AE |
4030 | dout("rbd_dev_v2_snap_context returned %d\n", ret); |
4031 | if (ret) | |
4032 | goto out; | |
117973fb AE |
4033 | out: |
4034 | up_write(&rbd_dev->header_rwsem); | |
4035 | ||
4036 | return ret; | |
4037 | } | |
4038 | ||
dfc5606d YS |
4039 | static int rbd_bus_add_dev(struct rbd_device *rbd_dev) |
4040 | { | |
dfc5606d | 4041 | struct device *dev; |
cd789ab9 | 4042 | int ret; |
dfc5606d YS |
4043 | |
4044 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); | |
dfc5606d | 4045 | |
cd789ab9 | 4046 | dev = &rbd_dev->dev; |
dfc5606d YS |
4047 | dev->bus = &rbd_bus_type; |
4048 | dev->type = &rbd_device_type; | |
4049 | dev->parent = &rbd_root_dev; | |
200a6a8b | 4050 | dev->release = rbd_dev_device_release; |
de71a297 | 4051 | dev_set_name(dev, "%d", rbd_dev->dev_id); |
dfc5606d | 4052 | ret = device_register(dev); |
dfc5606d | 4053 | |
dfc5606d | 4054 | mutex_unlock(&ctl_mutex); |
cd789ab9 | 4055 | |
dfc5606d | 4056 | return ret; |
602adf40 YS |
4057 | } |
4058 | ||
dfc5606d YS |
4059 | static void rbd_bus_del_dev(struct rbd_device *rbd_dev) |
4060 | { | |
4061 | device_unregister(&rbd_dev->dev); | |
4062 | } | |
4063 | ||
e2839308 | 4064 | static atomic64_t rbd_dev_id_max = ATOMIC64_INIT(0); |
1ddbe94e AE |
4065 | |
4066 | /* | |
499afd5b AE |
4067 | * Get a unique rbd identifier for the given new rbd_dev, and add |
4068 | * the rbd_dev to the global list. The minimum rbd id is 1. | |
1ddbe94e | 4069 | */ |
e2839308 | 4070 | static void rbd_dev_id_get(struct rbd_device *rbd_dev) |
b7f23c36 | 4071 | { |
e2839308 | 4072 | rbd_dev->dev_id = atomic64_inc_return(&rbd_dev_id_max); |
499afd5b AE |
4073 | |
4074 | spin_lock(&rbd_dev_list_lock); | |
4075 | list_add_tail(&rbd_dev->node, &rbd_dev_list); | |
4076 | spin_unlock(&rbd_dev_list_lock); | |
e2839308 AE |
4077 | dout("rbd_dev %p given dev id %llu\n", rbd_dev, |
4078 | (unsigned long long) rbd_dev->dev_id); | |
1ddbe94e | 4079 | } |
b7f23c36 | 4080 | |
1ddbe94e | 4081 | /* |
499afd5b AE |
4082 | * Remove an rbd_dev from the global list, and record that its |
4083 | * identifier is no longer in use. | |
1ddbe94e | 4084 | */ |
e2839308 | 4085 | static void rbd_dev_id_put(struct rbd_device *rbd_dev) |
1ddbe94e | 4086 | { |
d184f6bf | 4087 | struct list_head *tmp; |
de71a297 | 4088 | int rbd_id = rbd_dev->dev_id; |
d184f6bf AE |
4089 | int max_id; |
4090 | ||
aafb230e | 4091 | rbd_assert(rbd_id > 0); |
499afd5b | 4092 | |
e2839308 AE |
4093 | dout("rbd_dev %p released dev id %llu\n", rbd_dev, |
4094 | (unsigned long long) rbd_dev->dev_id); | |
499afd5b AE |
4095 | spin_lock(&rbd_dev_list_lock); |
4096 | list_del_init(&rbd_dev->node); | |
d184f6bf AE |
4097 | |
4098 | /* | |
4099 | * If the id being "put" is not the current maximum, there | |
4100 | * is nothing special we need to do. | |
4101 | */ | |
e2839308 | 4102 | if (rbd_id != atomic64_read(&rbd_dev_id_max)) { |
d184f6bf AE |
4103 | spin_unlock(&rbd_dev_list_lock); |
4104 | return; | |
4105 | } | |
4106 | ||
4107 | /* | |
4108 | * We need to update the current maximum id. Search the | |
4109 | * list to find out what it is. We're more likely to find | |
4110 | * the maximum at the end, so search the list backward. | |
4111 | */ | |
4112 | max_id = 0; | |
4113 | list_for_each_prev(tmp, &rbd_dev_list) { | |
4114 | struct rbd_device *rbd_dev; | |
4115 | ||
4116 | rbd_dev = list_entry(tmp, struct rbd_device, node); | |
b213e0b1 AE |
4117 | if (rbd_dev->dev_id > max_id) |
4118 | max_id = rbd_dev->dev_id; | |
d184f6bf | 4119 | } |
499afd5b | 4120 | spin_unlock(&rbd_dev_list_lock); |
b7f23c36 | 4121 | |
1ddbe94e | 4122 | /* |
e2839308 | 4123 | * The max id could have been updated by rbd_dev_id_get(), in |
d184f6bf AE |
4124 | * which case it now accurately reflects the new maximum. |
4125 | * Be careful not to overwrite the maximum value in that | |
4126 | * case. | |
1ddbe94e | 4127 | */ |
e2839308 AE |
4128 | atomic64_cmpxchg(&rbd_dev_id_max, rbd_id, max_id); |
4129 | dout(" max dev id has been reset\n"); | |
b7f23c36 AE |
4130 | } |
4131 | ||
e28fff26 AE |
4132 | /* |
4133 | * Skips over white space at *buf, and updates *buf to point to the | |
4134 | * first found non-space character (if any). Returns the length of | |
593a9e7b AE |
4135 | * the token (string of non-white space characters) found. Note |
4136 | * that *buf must be terminated with '\0'. | |
e28fff26 AE |
4137 | */ |
4138 | static inline size_t next_token(const char **buf) | |
4139 | { | |
4140 | /* | |
4141 | * These are the characters that produce nonzero for | |
4142 | * isspace() in the "C" and "POSIX" locales. | |
4143 | */ | |
4144 | const char *spaces = " \f\n\r\t\v"; | |
4145 | ||
4146 | *buf += strspn(*buf, spaces); /* Find start of token */ | |
4147 | ||
4148 | return strcspn(*buf, spaces); /* Return token length */ | |
4149 | } | |
4150 | ||
4151 | /* | |
4152 | * Finds the next token in *buf, and if the provided token buffer is | |
4153 | * big enough, copies the found token into it. The result, if | |
593a9e7b AE |
4154 | * copied, is guaranteed to be terminated with '\0'. Note that *buf |
4155 | * must be terminated with '\0' on entry. | |
e28fff26 AE |
4156 | * |
4157 | * Returns the length of the token found (not including the '\0'). | |
4158 | * Return value will be 0 if no token is found, and it will be >= | |
4159 | * token_size if the token would not fit. | |
4160 | * | |
593a9e7b | 4161 | * The *buf pointer will be updated to point beyond the end of the |
e28fff26 AE |
4162 | * found token. Note that this occurs even if the token buffer is |
4163 | * too small to hold it. | |
4164 | */ | |
4165 | static inline size_t copy_token(const char **buf, | |
4166 | char *token, | |
4167 | size_t token_size) | |
4168 | { | |
4169 | size_t len; | |
4170 | ||
4171 | len = next_token(buf); | |
4172 | if (len < token_size) { | |
4173 | memcpy(token, *buf, len); | |
4174 | *(token + len) = '\0'; | |
4175 | } | |
4176 | *buf += len; | |
4177 | ||
4178 | return len; | |
4179 | } | |
4180 | ||
ea3352f4 AE |
4181 | /* |
4182 | * Finds the next token in *buf, dynamically allocates a buffer big | |
4183 | * enough to hold a copy of it, and copies the token into the new | |
4184 | * buffer. The copy is guaranteed to be terminated with '\0'. Note | |
4185 | * that a duplicate buffer is created even for a zero-length token. | |
4186 | * | |
4187 | * Returns a pointer to the newly-allocated duplicate, or a null | |
4188 | * pointer if memory for the duplicate was not available. If | |
4189 | * the lenp argument is a non-null pointer, the length of the token | |
4190 | * (not including the '\0') is returned in *lenp. | |
4191 | * | |
4192 | * If successful, the *buf pointer will be updated to point beyond | |
4193 | * the end of the found token. | |
4194 | * | |
4195 | * Note: uses GFP_KERNEL for allocation. | |
4196 | */ | |
4197 | static inline char *dup_token(const char **buf, size_t *lenp) | |
4198 | { | |
4199 | char *dup; | |
4200 | size_t len; | |
4201 | ||
4202 | len = next_token(buf); | |
4caf35f9 | 4203 | dup = kmemdup(*buf, len + 1, GFP_KERNEL); |
ea3352f4 AE |
4204 | if (!dup) |
4205 | return NULL; | |
ea3352f4 AE |
4206 | *(dup + len) = '\0'; |
4207 | *buf += len; | |
4208 | ||
4209 | if (lenp) | |
4210 | *lenp = len; | |
4211 | ||
4212 | return dup; | |
4213 | } | |
4214 | ||
a725f65e | 4215 | /* |
859c31df AE |
4216 | * Parse the options provided for an "rbd add" (i.e., rbd image |
4217 | * mapping) request. These arrive via a write to /sys/bus/rbd/add, | |
4218 | * and the data written is passed here via a NUL-terminated buffer. | |
4219 | * Returns 0 if successful or an error code otherwise. | |
d22f76e7 | 4220 | * |
859c31df AE |
4221 | * The information extracted from these options is recorded in |
4222 | * the other parameters which return dynamically-allocated | |
4223 | * structures: | |
4224 | * ceph_opts | |
4225 | * The address of a pointer that will refer to a ceph options | |
4226 | * structure. Caller must release the returned pointer using | |
4227 | * ceph_destroy_options() when it is no longer needed. | |
4228 | * rbd_opts | |
4229 | * Address of an rbd options pointer. Fully initialized by | |
4230 | * this function; caller must release with kfree(). | |
4231 | * spec | |
4232 | * Address of an rbd image specification pointer. Fully | |
4233 | * initialized by this function based on parsed options. | |
4234 | * Caller must release with rbd_spec_put(). | |
4235 | * | |
4236 | * The options passed take this form: | |
4237 | * <mon_addrs> <options> <pool_name> <image_name> [<snap_id>] | |
4238 | * where: | |
4239 | * <mon_addrs> | |
4240 | * A comma-separated list of one or more monitor addresses. | |
4241 | * A monitor address is an ip address, optionally followed | |
4242 | * by a port number (separated by a colon). | |
4243 | * I.e.: ip1[:port1][,ip2[:port2]...] | |
4244 | * <options> | |
4245 | * A comma-separated list of ceph and/or rbd options. | |
4246 | * <pool_name> | |
4247 | * The name of the rados pool containing the rbd image. | |
4248 | * <image_name> | |
4249 | * The name of the image in that pool to map. | |
4250 | * <snap_id> | |
4251 | * An optional snapshot id. If provided, the mapping will | |
4252 | * present data from the image at the time that snapshot was | |
4253 | * created. The image head is used if no snapshot id is | |
4254 | * provided. Snapshot mappings are always read-only. | |
a725f65e | 4255 | */ |
859c31df | 4256 | static int rbd_add_parse_args(const char *buf, |
dc79b113 | 4257 | struct ceph_options **ceph_opts, |
859c31df AE |
4258 | struct rbd_options **opts, |
4259 | struct rbd_spec **rbd_spec) | |
e28fff26 | 4260 | { |
d22f76e7 | 4261 | size_t len; |
859c31df | 4262 | char *options; |
0ddebc0c | 4263 | const char *mon_addrs; |
ecb4dc22 | 4264 | char *snap_name; |
0ddebc0c | 4265 | size_t mon_addrs_size; |
859c31df | 4266 | struct rbd_spec *spec = NULL; |
4e9afeba | 4267 | struct rbd_options *rbd_opts = NULL; |
859c31df | 4268 | struct ceph_options *copts; |
dc79b113 | 4269 | int ret; |
e28fff26 AE |
4270 | |
4271 | /* The first four tokens are required */ | |
4272 | ||
7ef3214a | 4273 | len = next_token(&buf); |
4fb5d671 AE |
4274 | if (!len) { |
4275 | rbd_warn(NULL, "no monitor address(es) provided"); | |
4276 | return -EINVAL; | |
4277 | } | |
0ddebc0c | 4278 | mon_addrs = buf; |
f28e565a | 4279 | mon_addrs_size = len + 1; |
7ef3214a | 4280 | buf += len; |
a725f65e | 4281 | |
dc79b113 | 4282 | ret = -EINVAL; |
f28e565a AE |
4283 | options = dup_token(&buf, NULL); |
4284 | if (!options) | |
dc79b113 | 4285 | return -ENOMEM; |
4fb5d671 AE |
4286 | if (!*options) { |
4287 | rbd_warn(NULL, "no options provided"); | |
4288 | goto out_err; | |
4289 | } | |
e28fff26 | 4290 | |
859c31df AE |
4291 | spec = rbd_spec_alloc(); |
4292 | if (!spec) | |
f28e565a | 4293 | goto out_mem; |
859c31df AE |
4294 | |
4295 | spec->pool_name = dup_token(&buf, NULL); | |
4296 | if (!spec->pool_name) | |
4297 | goto out_mem; | |
4fb5d671 AE |
4298 | if (!*spec->pool_name) { |
4299 | rbd_warn(NULL, "no pool name provided"); | |
4300 | goto out_err; | |
4301 | } | |
e28fff26 | 4302 | |
69e7a02f | 4303 | spec->image_name = dup_token(&buf, NULL); |
859c31df | 4304 | if (!spec->image_name) |
f28e565a | 4305 | goto out_mem; |
4fb5d671 AE |
4306 | if (!*spec->image_name) { |
4307 | rbd_warn(NULL, "no image name provided"); | |
4308 | goto out_err; | |
4309 | } | |
d4b125e9 | 4310 | |
f28e565a AE |
4311 | /* |
4312 | * Snapshot name is optional; default is to use "-" | |
4313 | * (indicating the head/no snapshot). | |
4314 | */ | |
3feeb894 | 4315 | len = next_token(&buf); |
820a5f3e | 4316 | if (!len) { |
3feeb894 AE |
4317 | buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */ |
4318 | len = sizeof (RBD_SNAP_HEAD_NAME) - 1; | |
f28e565a | 4319 | } else if (len > RBD_MAX_SNAP_NAME_LEN) { |
dc79b113 | 4320 | ret = -ENAMETOOLONG; |
f28e565a | 4321 | goto out_err; |
849b4260 | 4322 | } |
ecb4dc22 AE |
4323 | snap_name = kmemdup(buf, len + 1, GFP_KERNEL); |
4324 | if (!snap_name) | |
f28e565a | 4325 | goto out_mem; |
ecb4dc22 AE |
4326 | *(snap_name + len) = '\0'; |
4327 | spec->snap_name = snap_name; | |
e5c35534 | 4328 | |
0ddebc0c | 4329 | /* Initialize all rbd options to the defaults */ |
e28fff26 | 4330 | |
4e9afeba AE |
4331 | rbd_opts = kzalloc(sizeof (*rbd_opts), GFP_KERNEL); |
4332 | if (!rbd_opts) | |
4333 | goto out_mem; | |
4334 | ||
4335 | rbd_opts->read_only = RBD_READ_ONLY_DEFAULT; | |
d22f76e7 | 4336 | |
859c31df | 4337 | copts = ceph_parse_options(options, mon_addrs, |
0ddebc0c | 4338 | mon_addrs + mon_addrs_size - 1, |
4e9afeba | 4339 | parse_rbd_opts_token, rbd_opts); |
859c31df AE |
4340 | if (IS_ERR(copts)) { |
4341 | ret = PTR_ERR(copts); | |
dc79b113 AE |
4342 | goto out_err; |
4343 | } | |
859c31df AE |
4344 | kfree(options); |
4345 | ||
4346 | *ceph_opts = copts; | |
4e9afeba | 4347 | *opts = rbd_opts; |
859c31df | 4348 | *rbd_spec = spec; |
0ddebc0c | 4349 | |
dc79b113 | 4350 | return 0; |
f28e565a | 4351 | out_mem: |
dc79b113 | 4352 | ret = -ENOMEM; |
d22f76e7 | 4353 | out_err: |
859c31df AE |
4354 | kfree(rbd_opts); |
4355 | rbd_spec_put(spec); | |
f28e565a | 4356 | kfree(options); |
d22f76e7 | 4357 | |
dc79b113 | 4358 | return ret; |
a725f65e AE |
4359 | } |
4360 | ||
589d30e0 AE |
4361 | /* |
4362 | * An rbd format 2 image has a unique identifier, distinct from the | |
4363 | * name given to it by the user. Internally, that identifier is | |
4364 | * what's used to specify the names of objects related to the image. | |
4365 | * | |
4366 | * A special "rbd id" object is used to map an rbd image name to its | |
4367 | * id. If that object doesn't exist, then there is no v2 rbd image | |
4368 | * with the supplied name. | |
4369 | * | |
4370 | * This function will record the given rbd_dev's image_id field if | |
4371 | * it can be determined, and in that case will return 0. If any | |
4372 | * errors occur a negative errno will be returned and the rbd_dev's | |
4373 | * image_id field will be unchanged (and should be NULL). | |
4374 | */ | |
4375 | static int rbd_dev_image_id(struct rbd_device *rbd_dev) | |
4376 | { | |
4377 | int ret; | |
4378 | size_t size; | |
4379 | char *object_name; | |
4380 | void *response; | |
c0fba368 | 4381 | char *image_id; |
2f82ee54 | 4382 | |
2c0d0a10 AE |
4383 | /* |
4384 | * When probing a parent image, the image id is already | |
4385 | * known (and the image name likely is not). There's no | |
c0fba368 AE |
4386 | * need to fetch the image id again in this case. We |
4387 | * do still need to set the image format though. | |
2c0d0a10 | 4388 | */ |
c0fba368 AE |
4389 | if (rbd_dev->spec->image_id) { |
4390 | rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1; | |
4391 | ||
2c0d0a10 | 4392 | return 0; |
c0fba368 | 4393 | } |
2c0d0a10 | 4394 | |
589d30e0 AE |
4395 | /* |
4396 | * First, see if the format 2 image id file exists, and if | |
4397 | * so, get the image's persistent id from it. | |
4398 | */ | |
69e7a02f | 4399 | size = sizeof (RBD_ID_PREFIX) + strlen(rbd_dev->spec->image_name); |
589d30e0 AE |
4400 | object_name = kmalloc(size, GFP_NOIO); |
4401 | if (!object_name) | |
4402 | return -ENOMEM; | |
0d7dbfce | 4403 | sprintf(object_name, "%s%s", RBD_ID_PREFIX, rbd_dev->spec->image_name); |
589d30e0 AE |
4404 | dout("rbd id object name is %s\n", object_name); |
4405 | ||
4406 | /* Response will be an encoded string, which includes a length */ | |
4407 | ||
4408 | size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX; | |
4409 | response = kzalloc(size, GFP_NOIO); | |
4410 | if (!response) { | |
4411 | ret = -ENOMEM; | |
4412 | goto out; | |
4413 | } | |
4414 | ||
c0fba368 AE |
4415 | /* If it doesn't exist we'll assume it's a format 1 image */ |
4416 | ||
36be9a76 | 4417 | ret = rbd_obj_method_sync(rbd_dev, object_name, |
4157976b | 4418 | "rbd", "get_id", NULL, 0, |
e2a58ee5 | 4419 | response, RBD_IMAGE_ID_LEN_MAX); |
36be9a76 | 4420 | dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); |
c0fba368 AE |
4421 | if (ret == -ENOENT) { |
4422 | image_id = kstrdup("", GFP_KERNEL); | |
4423 | ret = image_id ? 0 : -ENOMEM; | |
4424 | if (!ret) | |
4425 | rbd_dev->image_format = 1; | |
4426 | } else if (ret > sizeof (__le32)) { | |
4427 | void *p = response; | |
4428 | ||
4429 | image_id = ceph_extract_encoded_string(&p, p + ret, | |
979ed480 | 4430 | NULL, GFP_NOIO); |
c0fba368 AE |
4431 | ret = IS_ERR(image_id) ? PTR_ERR(image_id) : 0; |
4432 | if (!ret) | |
4433 | rbd_dev->image_format = 2; | |
589d30e0 | 4434 | } else { |
c0fba368 AE |
4435 | ret = -EINVAL; |
4436 | } | |
4437 | ||
4438 | if (!ret) { | |
4439 | rbd_dev->spec->image_id = image_id; | |
4440 | dout("image_id is %s\n", image_id); | |
589d30e0 AE |
4441 | } |
4442 | out: | |
4443 | kfree(response); | |
4444 | kfree(object_name); | |
4445 | ||
4446 | return ret; | |
4447 | } | |
4448 | ||
6fd48b3b AE |
4449 | /* Undo whatever state changes are made by v1 or v2 image probe */ |
4450 | ||
4451 | static void rbd_dev_unprobe(struct rbd_device *rbd_dev) | |
4452 | { | |
4453 | struct rbd_image_header *header; | |
4454 | ||
4455 | rbd_dev_remove_parent(rbd_dev); | |
4456 | rbd_spec_put(rbd_dev->parent_spec); | |
4457 | rbd_dev->parent_spec = NULL; | |
4458 | rbd_dev->parent_overlap = 0; | |
4459 | ||
4460 | /* Free dynamic fields from the header, then zero it out */ | |
4461 | ||
4462 | header = &rbd_dev->header; | |
812164f8 | 4463 | ceph_put_snap_context(header->snapc); |
6fd48b3b AE |
4464 | kfree(header->snap_sizes); |
4465 | kfree(header->snap_names); | |
4466 | kfree(header->object_prefix); | |
4467 | memset(header, 0, sizeof (*header)); | |
4468 | } | |
4469 | ||
a30b71b9 AE |
4470 | static int rbd_dev_v1_probe(struct rbd_device *rbd_dev) |
4471 | { | |
30d60ba2 | 4472 | return rbd_dev_v1_header_read(rbd_dev); |
a30b71b9 AE |
4473 | } |
4474 | ||
4475 | static int rbd_dev_v2_probe(struct rbd_device *rbd_dev) | |
4476 | { | |
9d475de5 | 4477 | int ret; |
a30b71b9 | 4478 | |
9d475de5 | 4479 | ret = rbd_dev_v2_image_size(rbd_dev); |
57385b51 | 4480 | if (ret) |
1e130199 AE |
4481 | goto out_err; |
4482 | ||
4483 | /* Get the object prefix (a.k.a. block_name) for the image */ | |
4484 | ||
4485 | ret = rbd_dev_v2_object_prefix(rbd_dev); | |
57385b51 | 4486 | if (ret) |
b1b5402a AE |
4487 | goto out_err; |
4488 | ||
d889140c | 4489 | /* Get the and check features for the image */ |
b1b5402a AE |
4490 | |
4491 | ret = rbd_dev_v2_features(rbd_dev); | |
57385b51 | 4492 | if (ret) |
9d475de5 | 4493 | goto out_err; |
35d489f9 | 4494 | |
86b00e0d AE |
4495 | /* If the image supports layering, get the parent info */ |
4496 | ||
4497 | if (rbd_dev->header.features & RBD_FEATURE_LAYERING) { | |
4498 | ret = rbd_dev_v2_parent_info(rbd_dev); | |
57385b51 | 4499 | if (ret) |
86b00e0d | 4500 | goto out_err; |
96882f55 | 4501 | /* |
c734b796 AE |
4502 | * Print a warning if this image has a parent. |
4503 | * Don't print it if the image now being probed | |
4504 | * is itself a parent. We can tell at this point | |
4505 | * because we won't know its pool name yet (just its | |
4506 | * pool id). | |
96882f55 | 4507 | */ |
c734b796 | 4508 | if (rbd_dev->parent_spec && rbd_dev->spec->pool_name) |
96882f55 AE |
4509 | rbd_warn(rbd_dev, "WARNING: kernel layering " |
4510 | "is EXPERIMENTAL!"); | |
86b00e0d AE |
4511 | } |
4512 | ||
cc070d59 AE |
4513 | /* If the image supports fancy striping, get its parameters */ |
4514 | ||
4515 | if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) { | |
4516 | ret = rbd_dev_v2_striping_info(rbd_dev); | |
4517 | if (ret < 0) | |
4518 | goto out_err; | |
4519 | } | |
4520 | ||
6e14b1a6 AE |
4521 | /* crypto and compression type aren't (yet) supported for v2 images */ |
4522 | ||
4523 | rbd_dev->header.crypt_type = 0; | |
4524 | rbd_dev->header.comp_type = 0; | |
35d489f9 | 4525 | |
6e14b1a6 AE |
4526 | /* Get the snapshot context, plus the header version */ |
4527 | ||
cc4a38bd | 4528 | ret = rbd_dev_v2_snap_context(rbd_dev); |
35d489f9 AE |
4529 | if (ret) |
4530 | goto out_err; | |
6e14b1a6 | 4531 | |
35152979 | 4532 | return 0; |
9d475de5 | 4533 | out_err: |
86b00e0d AE |
4534 | rbd_dev->parent_overlap = 0; |
4535 | rbd_spec_put(rbd_dev->parent_spec); | |
4536 | rbd_dev->parent_spec = NULL; | |
9d475de5 AE |
4537 | kfree(rbd_dev->header_name); |
4538 | rbd_dev->header_name = NULL; | |
1e130199 AE |
4539 | kfree(rbd_dev->header.object_prefix); |
4540 | rbd_dev->header.object_prefix = NULL; | |
9d475de5 AE |
4541 | |
4542 | return ret; | |
a30b71b9 AE |
4543 | } |
4544 | ||
124afba2 | 4545 | static int rbd_dev_probe_parent(struct rbd_device *rbd_dev) |
83a06263 | 4546 | { |
2f82ee54 | 4547 | struct rbd_device *parent = NULL; |
124afba2 AE |
4548 | struct rbd_spec *parent_spec; |
4549 | struct rbd_client *rbdc; | |
4550 | int ret; | |
4551 | ||
4552 | if (!rbd_dev->parent_spec) | |
4553 | return 0; | |
4554 | /* | |
4555 | * We need to pass a reference to the client and the parent | |
4556 | * spec when creating the parent rbd_dev. Images related by | |
4557 | * parent/child relationships always share both. | |
4558 | */ | |
4559 | parent_spec = rbd_spec_get(rbd_dev->parent_spec); | |
4560 | rbdc = __rbd_get_client(rbd_dev->rbd_client); | |
4561 | ||
4562 | ret = -ENOMEM; | |
4563 | parent = rbd_dev_create(rbdc, parent_spec); | |
4564 | if (!parent) | |
4565 | goto out_err; | |
4566 | ||
51344a38 | 4567 | ret = rbd_dev_image_probe(parent, true); |
124afba2 AE |
4568 | if (ret < 0) |
4569 | goto out_err; | |
4570 | rbd_dev->parent = parent; | |
4571 | ||
4572 | return 0; | |
4573 | out_err: | |
4574 | if (parent) { | |
4575 | rbd_spec_put(rbd_dev->parent_spec); | |
4576 | kfree(rbd_dev->header_name); | |
4577 | rbd_dev_destroy(parent); | |
4578 | } else { | |
4579 | rbd_put_client(rbdc); | |
4580 | rbd_spec_put(parent_spec); | |
4581 | } | |
4582 | ||
4583 | return ret; | |
4584 | } | |
4585 | ||
200a6a8b | 4586 | static int rbd_dev_device_setup(struct rbd_device *rbd_dev) |
124afba2 | 4587 | { |
83a06263 | 4588 | int ret; |
d1cf5788 | 4589 | |
83a06263 AE |
4590 | /* generate unique id: find highest unique id, add one */ |
4591 | rbd_dev_id_get(rbd_dev); | |
4592 | ||
4593 | /* Fill in the device name, now that we have its id. */ | |
4594 | BUILD_BUG_ON(DEV_NAME_LEN | |
4595 | < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH); | |
4596 | sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->dev_id); | |
4597 | ||
4598 | /* Get our block major device number. */ | |
4599 | ||
4600 | ret = register_blkdev(0, rbd_dev->name); | |
4601 | if (ret < 0) | |
4602 | goto err_out_id; | |
4603 | rbd_dev->major = ret; | |
4604 | ||
4605 | /* Set up the blkdev mapping. */ | |
4606 | ||
4607 | ret = rbd_init_disk(rbd_dev); | |
4608 | if (ret) | |
4609 | goto err_out_blkdev; | |
4610 | ||
f35a4dee | 4611 | ret = rbd_dev_mapping_set(rbd_dev); |
83a06263 AE |
4612 | if (ret) |
4613 | goto err_out_disk; | |
f35a4dee AE |
4614 | set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE); |
4615 | ||
4616 | ret = rbd_bus_add_dev(rbd_dev); | |
4617 | if (ret) | |
4618 | goto err_out_mapping; | |
83a06263 | 4619 | |
83a06263 AE |
4620 | /* Everything's ready. Announce the disk to the world. */ |
4621 | ||
129b79d4 | 4622 | set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags); |
83a06263 AE |
4623 | add_disk(rbd_dev->disk); |
4624 | ||
4625 | pr_info("%s: added with size 0x%llx\n", rbd_dev->disk->disk_name, | |
4626 | (unsigned long long) rbd_dev->mapping.size); | |
4627 | ||
4628 | return ret; | |
2f82ee54 | 4629 | |
f35a4dee AE |
4630 | err_out_mapping: |
4631 | rbd_dev_mapping_clear(rbd_dev); | |
83a06263 AE |
4632 | err_out_disk: |
4633 | rbd_free_disk(rbd_dev); | |
4634 | err_out_blkdev: | |
4635 | unregister_blkdev(rbd_dev->major, rbd_dev->name); | |
4636 | err_out_id: | |
4637 | rbd_dev_id_put(rbd_dev); | |
d1cf5788 | 4638 | rbd_dev_mapping_clear(rbd_dev); |
83a06263 AE |
4639 | |
4640 | return ret; | |
4641 | } | |
4642 | ||
332bb12d AE |
4643 | static int rbd_dev_header_name(struct rbd_device *rbd_dev) |
4644 | { | |
4645 | struct rbd_spec *spec = rbd_dev->spec; | |
4646 | size_t size; | |
4647 | ||
4648 | /* Record the header object name for this rbd image. */ | |
4649 | ||
4650 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); | |
4651 | ||
4652 | if (rbd_dev->image_format == 1) | |
4653 | size = strlen(spec->image_name) + sizeof (RBD_SUFFIX); | |
4654 | else | |
4655 | size = sizeof (RBD_HEADER_PREFIX) + strlen(spec->image_id); | |
4656 | ||
4657 | rbd_dev->header_name = kmalloc(size, GFP_KERNEL); | |
4658 | if (!rbd_dev->header_name) | |
4659 | return -ENOMEM; | |
4660 | ||
4661 | if (rbd_dev->image_format == 1) | |
4662 | sprintf(rbd_dev->header_name, "%s%s", | |
4663 | spec->image_name, RBD_SUFFIX); | |
4664 | else | |
4665 | sprintf(rbd_dev->header_name, "%s%s", | |
4666 | RBD_HEADER_PREFIX, spec->image_id); | |
4667 | return 0; | |
4668 | } | |
4669 | ||
200a6a8b AE |
4670 | static void rbd_dev_image_release(struct rbd_device *rbd_dev) |
4671 | { | |
6fd48b3b AE |
4672 | int ret; |
4673 | ||
6fd48b3b AE |
4674 | rbd_dev_unprobe(rbd_dev); |
4675 | ret = rbd_dev_header_watch_sync(rbd_dev, 0); | |
4676 | if (ret) | |
4677 | rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret); | |
200a6a8b | 4678 | kfree(rbd_dev->header_name); |
6fd48b3b AE |
4679 | rbd_dev->header_name = NULL; |
4680 | rbd_dev->image_format = 0; | |
4681 | kfree(rbd_dev->spec->image_id); | |
4682 | rbd_dev->spec->image_id = NULL; | |
4683 | ||
200a6a8b AE |
4684 | rbd_dev_destroy(rbd_dev); |
4685 | } | |
4686 | ||
a30b71b9 AE |
4687 | /* |
4688 | * Probe for the existence of the header object for the given rbd | |
4689 | * device. For format 2 images this includes determining the image | |
4690 | * id. | |
4691 | */ | |
51344a38 | 4692 | static int rbd_dev_image_probe(struct rbd_device *rbd_dev, bool read_only) |
a30b71b9 AE |
4693 | { |
4694 | int ret; | |
b644de2b | 4695 | int tmp; |
a30b71b9 AE |
4696 | |
4697 | /* | |
4698 | * Get the id from the image id object. If it's not a | |
4699 | * format 2 image, we'll get ENOENT back, and we'll assume | |
4700 | * it's a format 1 image. | |
4701 | */ | |
4702 | ret = rbd_dev_image_id(rbd_dev); | |
4703 | if (ret) | |
c0fba368 AE |
4704 | return ret; |
4705 | rbd_assert(rbd_dev->spec->image_id); | |
4706 | rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); | |
4707 | ||
332bb12d AE |
4708 | ret = rbd_dev_header_name(rbd_dev); |
4709 | if (ret) | |
4710 | goto err_out_format; | |
4711 | ||
b644de2b AE |
4712 | ret = rbd_dev_header_watch_sync(rbd_dev, 1); |
4713 | if (ret) | |
4714 | goto out_header_name; | |
4715 | ||
c0fba368 | 4716 | if (rbd_dev->image_format == 1) |
a30b71b9 AE |
4717 | ret = rbd_dev_v1_probe(rbd_dev); |
4718 | else | |
4719 | ret = rbd_dev_v2_probe(rbd_dev); | |
5655c4d9 | 4720 | if (ret) |
b644de2b | 4721 | goto err_out_watch; |
83a06263 | 4722 | |
9bb81c9b AE |
4723 | ret = rbd_dev_spec_update(rbd_dev); |
4724 | if (ret) | |
33dca39f | 4725 | goto err_out_probe; |
9bb81c9b | 4726 | |
51344a38 AE |
4727 | /* If we are mapping a snapshot it must be marked read-only */ |
4728 | ||
4729 | if (rbd_dev->spec->snap_id != CEPH_NOSNAP) | |
4730 | read_only = true; | |
4731 | rbd_dev->mapping.read_only = read_only; | |
4732 | ||
9bb81c9b | 4733 | ret = rbd_dev_probe_parent(rbd_dev); |
30d60ba2 AE |
4734 | if (ret) |
4735 | goto err_out_probe; | |
4736 | ||
4737 | dout("discovered format %u image, header name is %s\n", | |
4738 | rbd_dev->image_format, rbd_dev->header_name); | |
83a06263 | 4739 | |
30d60ba2 | 4740 | return 0; |
6fd48b3b AE |
4741 | err_out_probe: |
4742 | rbd_dev_unprobe(rbd_dev); | |
b644de2b AE |
4743 | err_out_watch: |
4744 | tmp = rbd_dev_header_watch_sync(rbd_dev, 0); | |
4745 | if (tmp) | |
4746 | rbd_warn(rbd_dev, "unable to tear down watch request\n"); | |
332bb12d AE |
4747 | out_header_name: |
4748 | kfree(rbd_dev->header_name); | |
4749 | rbd_dev->header_name = NULL; | |
4750 | err_out_format: | |
4751 | rbd_dev->image_format = 0; | |
5655c4d9 AE |
4752 | kfree(rbd_dev->spec->image_id); |
4753 | rbd_dev->spec->image_id = NULL; | |
4754 | ||
4755 | dout("probe failed, returning %d\n", ret); | |
4756 | ||
a30b71b9 AE |
4757 | return ret; |
4758 | } | |
4759 | ||
59c2be1e YS |
4760 | static ssize_t rbd_add(struct bus_type *bus, |
4761 | const char *buf, | |
4762 | size_t count) | |
602adf40 | 4763 | { |
cb8627c7 | 4764 | struct rbd_device *rbd_dev = NULL; |
dc79b113 | 4765 | struct ceph_options *ceph_opts = NULL; |
4e9afeba | 4766 | struct rbd_options *rbd_opts = NULL; |
859c31df | 4767 | struct rbd_spec *spec = NULL; |
9d3997fd | 4768 | struct rbd_client *rbdc; |
27cc2594 | 4769 | struct ceph_osd_client *osdc; |
51344a38 | 4770 | bool read_only; |
27cc2594 | 4771 | int rc = -ENOMEM; |
602adf40 YS |
4772 | |
4773 | if (!try_module_get(THIS_MODULE)) | |
4774 | return -ENODEV; | |
4775 | ||
602adf40 | 4776 | /* parse add command */ |
859c31df | 4777 | rc = rbd_add_parse_args(buf, &ceph_opts, &rbd_opts, &spec); |
dc79b113 | 4778 | if (rc < 0) |
bd4ba655 | 4779 | goto err_out_module; |
51344a38 AE |
4780 | read_only = rbd_opts->read_only; |
4781 | kfree(rbd_opts); | |
4782 | rbd_opts = NULL; /* done with this */ | |
78cea76e | 4783 | |
9d3997fd AE |
4784 | rbdc = rbd_get_client(ceph_opts); |
4785 | if (IS_ERR(rbdc)) { | |
4786 | rc = PTR_ERR(rbdc); | |
0ddebc0c | 4787 | goto err_out_args; |
9d3997fd | 4788 | } |
c53d5893 | 4789 | ceph_opts = NULL; /* rbd_dev client now owns this */ |
602adf40 | 4790 | |
602adf40 | 4791 | /* pick the pool */ |
9d3997fd | 4792 | osdc = &rbdc->client->osdc; |
859c31df | 4793 | rc = ceph_pg_poolid_by_name(osdc->osdmap, spec->pool_name); |
602adf40 YS |
4794 | if (rc < 0) |
4795 | goto err_out_client; | |
c0cd10db | 4796 | spec->pool_id = (u64)rc; |
859c31df | 4797 | |
0903e875 AE |
4798 | /* The ceph file layout needs to fit pool id in 32 bits */ |
4799 | ||
c0cd10db AE |
4800 | if (spec->pool_id > (u64)U32_MAX) { |
4801 | rbd_warn(NULL, "pool id too large (%llu > %u)\n", | |
4802 | (unsigned long long)spec->pool_id, U32_MAX); | |
0903e875 AE |
4803 | rc = -EIO; |
4804 | goto err_out_client; | |
4805 | } | |
4806 | ||
c53d5893 | 4807 | rbd_dev = rbd_dev_create(rbdc, spec); |
bd4ba655 AE |
4808 | if (!rbd_dev) |
4809 | goto err_out_client; | |
c53d5893 AE |
4810 | rbdc = NULL; /* rbd_dev now owns this */ |
4811 | spec = NULL; /* rbd_dev now owns this */ | |
602adf40 | 4812 | |
51344a38 | 4813 | rc = rbd_dev_image_probe(rbd_dev, read_only); |
a30b71b9 | 4814 | if (rc < 0) |
c53d5893 | 4815 | goto err_out_rbd_dev; |
05fd6f6f | 4816 | |
b536f69a AE |
4817 | rc = rbd_dev_device_setup(rbd_dev); |
4818 | if (!rc) | |
4819 | return count; | |
4820 | ||
4821 | rbd_dev_image_release(rbd_dev); | |
c53d5893 AE |
4822 | err_out_rbd_dev: |
4823 | rbd_dev_destroy(rbd_dev); | |
bd4ba655 | 4824 | err_out_client: |
9d3997fd | 4825 | rbd_put_client(rbdc); |
0ddebc0c | 4826 | err_out_args: |
78cea76e AE |
4827 | if (ceph_opts) |
4828 | ceph_destroy_options(ceph_opts); | |
4e9afeba | 4829 | kfree(rbd_opts); |
859c31df | 4830 | rbd_spec_put(spec); |
bd4ba655 AE |
4831 | err_out_module: |
4832 | module_put(THIS_MODULE); | |
27cc2594 | 4833 | |
602adf40 | 4834 | dout("Error adding device %s\n", buf); |
27cc2594 | 4835 | |
c0cd10db | 4836 | return (ssize_t)rc; |
602adf40 YS |
4837 | } |
4838 | ||
de71a297 | 4839 | static struct rbd_device *__rbd_get_dev(unsigned long dev_id) |
602adf40 YS |
4840 | { |
4841 | struct list_head *tmp; | |
4842 | struct rbd_device *rbd_dev; | |
4843 | ||
e124a82f | 4844 | spin_lock(&rbd_dev_list_lock); |
602adf40 YS |
4845 | list_for_each(tmp, &rbd_dev_list) { |
4846 | rbd_dev = list_entry(tmp, struct rbd_device, node); | |
de71a297 | 4847 | if (rbd_dev->dev_id == dev_id) { |
e124a82f | 4848 | spin_unlock(&rbd_dev_list_lock); |
602adf40 | 4849 | return rbd_dev; |
e124a82f | 4850 | } |
602adf40 | 4851 | } |
e124a82f | 4852 | spin_unlock(&rbd_dev_list_lock); |
602adf40 YS |
4853 | return NULL; |
4854 | } | |
4855 | ||
200a6a8b | 4856 | static void rbd_dev_device_release(struct device *dev) |
602adf40 | 4857 | { |
593a9e7b | 4858 | struct rbd_device *rbd_dev = dev_to_rbd_dev(dev); |
602adf40 | 4859 | |
602adf40 | 4860 | rbd_free_disk(rbd_dev); |
200a6a8b | 4861 | clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags); |
6d80b130 | 4862 | rbd_dev_mapping_clear(rbd_dev); |
602adf40 | 4863 | unregister_blkdev(rbd_dev->major, rbd_dev->name); |
200a6a8b | 4864 | rbd_dev->major = 0; |
e2839308 | 4865 | rbd_dev_id_put(rbd_dev); |
d1cf5788 | 4866 | rbd_dev_mapping_clear(rbd_dev); |
602adf40 YS |
4867 | } |
4868 | ||
05a46afd AE |
4869 | static void rbd_dev_remove_parent(struct rbd_device *rbd_dev) |
4870 | { | |
ad945fc1 | 4871 | while (rbd_dev->parent) { |
05a46afd AE |
4872 | struct rbd_device *first = rbd_dev; |
4873 | struct rbd_device *second = first->parent; | |
4874 | struct rbd_device *third; | |
4875 | ||
4876 | /* | |
4877 | * Follow to the parent with no grandparent and | |
4878 | * remove it. | |
4879 | */ | |
4880 | while (second && (third = second->parent)) { | |
4881 | first = second; | |
4882 | second = third; | |
4883 | } | |
ad945fc1 | 4884 | rbd_assert(second); |
8ad42cd0 | 4885 | rbd_dev_image_release(second); |
ad945fc1 AE |
4886 | first->parent = NULL; |
4887 | first->parent_overlap = 0; | |
4888 | ||
4889 | rbd_assert(first->parent_spec); | |
05a46afd AE |
4890 | rbd_spec_put(first->parent_spec); |
4891 | first->parent_spec = NULL; | |
05a46afd AE |
4892 | } |
4893 | } | |
4894 | ||
dfc5606d YS |
4895 | static ssize_t rbd_remove(struct bus_type *bus, |
4896 | const char *buf, | |
4897 | size_t count) | |
602adf40 YS |
4898 | { |
4899 | struct rbd_device *rbd_dev = NULL; | |
0d8189e1 | 4900 | int target_id; |
602adf40 | 4901 | unsigned long ul; |
0d8189e1 | 4902 | int ret; |
602adf40 | 4903 | |
0d8189e1 AE |
4904 | ret = strict_strtoul(buf, 10, &ul); |
4905 | if (ret) | |
4906 | return ret; | |
602adf40 YS |
4907 | |
4908 | /* convert to int; abort if we lost anything in the conversion */ | |
4909 | target_id = (int) ul; | |
4910 | if (target_id != ul) | |
4911 | return -EINVAL; | |
4912 | ||
4913 | mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); | |
4914 | ||
4915 | rbd_dev = __rbd_get_dev(target_id); | |
4916 | if (!rbd_dev) { | |
4917 | ret = -ENOENT; | |
4918 | goto done; | |
42382b70 AE |
4919 | } |
4920 | ||
a14ea269 | 4921 | spin_lock_irq(&rbd_dev->lock); |
b82d167b | 4922 | if (rbd_dev->open_count) |
42382b70 | 4923 | ret = -EBUSY; |
b82d167b AE |
4924 | else |
4925 | set_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags); | |
a14ea269 | 4926 | spin_unlock_irq(&rbd_dev->lock); |
b82d167b | 4927 | if (ret < 0) |
42382b70 | 4928 | goto done; |
0d8189e1 | 4929 | ret = count; |
b480815a | 4930 | rbd_bus_del_dev(rbd_dev); |
8ad42cd0 | 4931 | rbd_dev_image_release(rbd_dev); |
79ab7558 | 4932 | module_put(THIS_MODULE); |
602adf40 YS |
4933 | done: |
4934 | mutex_unlock(&ctl_mutex); | |
aafb230e | 4935 | |
602adf40 YS |
4936 | return ret; |
4937 | } | |
4938 | ||
602adf40 YS |
4939 | /* |
4940 | * create control files in sysfs | |
dfc5606d | 4941 | * /sys/bus/rbd/... |
602adf40 YS |
4942 | */ |
4943 | static int rbd_sysfs_init(void) | |
4944 | { | |
dfc5606d | 4945 | int ret; |
602adf40 | 4946 | |
fed4c143 | 4947 | ret = device_register(&rbd_root_dev); |
21079786 | 4948 | if (ret < 0) |
dfc5606d | 4949 | return ret; |
602adf40 | 4950 | |
fed4c143 AE |
4951 | ret = bus_register(&rbd_bus_type); |
4952 | if (ret < 0) | |
4953 | device_unregister(&rbd_root_dev); | |
602adf40 | 4954 | |
602adf40 YS |
4955 | return ret; |
4956 | } | |
4957 | ||
4958 | static void rbd_sysfs_cleanup(void) | |
4959 | { | |
dfc5606d | 4960 | bus_unregister(&rbd_bus_type); |
fed4c143 | 4961 | device_unregister(&rbd_root_dev); |
602adf40 YS |
4962 | } |
4963 | ||
1c2a9dfe AE |
4964 | static int rbd_slab_init(void) |
4965 | { | |
4966 | rbd_assert(!rbd_img_request_cache); | |
4967 | rbd_img_request_cache = kmem_cache_create("rbd_img_request", | |
4968 | sizeof (struct rbd_img_request), | |
4969 | __alignof__(struct rbd_img_request), | |
4970 | 0, NULL); | |
868311b1 AE |
4971 | if (!rbd_img_request_cache) |
4972 | return -ENOMEM; | |
4973 | ||
4974 | rbd_assert(!rbd_obj_request_cache); | |
4975 | rbd_obj_request_cache = kmem_cache_create("rbd_obj_request", | |
4976 | sizeof (struct rbd_obj_request), | |
4977 | __alignof__(struct rbd_obj_request), | |
4978 | 0, NULL); | |
78c2a44a AE |
4979 | if (!rbd_obj_request_cache) |
4980 | goto out_err; | |
4981 | ||
4982 | rbd_assert(!rbd_segment_name_cache); | |
4983 | rbd_segment_name_cache = kmem_cache_create("rbd_segment_name", | |
4984 | MAX_OBJ_NAME_SIZE + 1, 1, 0, NULL); | |
4985 | if (rbd_segment_name_cache) | |
1c2a9dfe | 4986 | return 0; |
78c2a44a AE |
4987 | out_err: |
4988 | if (rbd_obj_request_cache) { | |
4989 | kmem_cache_destroy(rbd_obj_request_cache); | |
4990 | rbd_obj_request_cache = NULL; | |
4991 | } | |
1c2a9dfe | 4992 | |
868311b1 AE |
4993 | kmem_cache_destroy(rbd_img_request_cache); |
4994 | rbd_img_request_cache = NULL; | |
4995 | ||
1c2a9dfe AE |
4996 | return -ENOMEM; |
4997 | } | |
4998 | ||
4999 | static void rbd_slab_exit(void) | |
5000 | { | |
78c2a44a AE |
5001 | rbd_assert(rbd_segment_name_cache); |
5002 | kmem_cache_destroy(rbd_segment_name_cache); | |
5003 | rbd_segment_name_cache = NULL; | |
5004 | ||
868311b1 AE |
5005 | rbd_assert(rbd_obj_request_cache); |
5006 | kmem_cache_destroy(rbd_obj_request_cache); | |
5007 | rbd_obj_request_cache = NULL; | |
5008 | ||
1c2a9dfe AE |
5009 | rbd_assert(rbd_img_request_cache); |
5010 | kmem_cache_destroy(rbd_img_request_cache); | |
5011 | rbd_img_request_cache = NULL; | |
5012 | } | |
5013 | ||
cc344fa1 | 5014 | static int __init rbd_init(void) |
602adf40 YS |
5015 | { |
5016 | int rc; | |
5017 | ||
1e32d34c AE |
5018 | if (!libceph_compatible(NULL)) { |
5019 | rbd_warn(NULL, "libceph incompatibility (quitting)"); | |
5020 | ||
5021 | return -EINVAL; | |
5022 | } | |
1c2a9dfe | 5023 | rc = rbd_slab_init(); |
602adf40 YS |
5024 | if (rc) |
5025 | return rc; | |
1c2a9dfe AE |
5026 | rc = rbd_sysfs_init(); |
5027 | if (rc) | |
5028 | rbd_slab_exit(); | |
5029 | else | |
5030 | pr_info("loaded " RBD_DRV_NAME_LONG "\n"); | |
5031 | ||
5032 | return rc; | |
602adf40 YS |
5033 | } |
5034 | ||
cc344fa1 | 5035 | static void __exit rbd_exit(void) |
602adf40 YS |
5036 | { |
5037 | rbd_sysfs_cleanup(); | |
1c2a9dfe | 5038 | rbd_slab_exit(); |
602adf40 YS |
5039 | } |
5040 | ||
5041 | module_init(rbd_init); | |
5042 | module_exit(rbd_exit); | |
5043 | ||
5044 | MODULE_AUTHOR("Sage Weil <sage@newdream.net>"); | |
5045 | MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>"); | |
5046 | MODULE_DESCRIPTION("rados block device"); | |
5047 | ||
5048 | /* following authorship retained from original osdblk.c */ | |
5049 | MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>"); | |
5050 | ||
5051 | MODULE_LICENSE("GPL"); |