xref: /openbmc/linux/drivers/block/rbd.c (revision d7955ce4)
1 
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 
25    For usage instructions, please refer to:
26 
27                  Documentation/ABI/testing/sysfs-bus-rbd
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/cls_lock_client.h>
35 #include <linux/ceph/striper.h>
36 #include <linux/ceph/decode.h>
37 #include <linux/fs_parser.h>
38 #include <linux/bsearch.h>
39 
40 #include <linux/kernel.h>
41 #include <linux/device.h>
42 #include <linux/module.h>
43 #include <linux/blk-mq.h>
44 #include <linux/fs.h>
45 #include <linux/blkdev.h>
46 #include <linux/slab.h>
47 #include <linux/idr.h>
48 #include <linux/workqueue.h>
49 
50 #include "rbd_types.h"
51 
52 #define RBD_DEBUG	/* Activate rbd_assert() calls */
53 
54 /*
55  * Increment the given counter and return its updated value.
56  * If the counter is already 0 it will not be incremented.
57  * If the counter is already at its maximum value returns
58  * -EINVAL without updating it.
59  */
60 static int atomic_inc_return_safe(atomic_t *v)
61 {
62 	unsigned int counter;
63 
64 	counter = (unsigned int)atomic_fetch_add_unless(v, 1, 0);
65 	if (counter <= (unsigned int)INT_MAX)
66 		return (int)counter;
67 
68 	atomic_dec(v);
69 
70 	return -EINVAL;
71 }
72 
73 /* Decrement the counter.  Return the resulting value, or -EINVAL */
74 static int atomic_dec_return_safe(atomic_t *v)
75 {
76 	int counter;
77 
78 	counter = atomic_dec_return(v);
79 	if (counter >= 0)
80 		return counter;
81 
82 	atomic_inc(v);
83 
84 	return -EINVAL;
85 }
86 
87 #define RBD_DRV_NAME "rbd"
88 
89 #define RBD_MINORS_PER_MAJOR		256
90 #define RBD_SINGLE_MAJOR_PART_SHIFT	4
91 
92 #define RBD_MAX_PARENT_CHAIN_LEN	16
93 
94 #define RBD_SNAP_DEV_NAME_PREFIX	"snap_"
95 #define RBD_MAX_SNAP_NAME_LEN	\
96 			(NAME_MAX - (sizeof (RBD_SNAP_DEV_NAME_PREFIX) - 1))
97 
98 #define RBD_MAX_SNAP_COUNT	510	/* allows max snapc to fit in 4KB */
99 
100 #define RBD_SNAP_HEAD_NAME	"-"
101 
102 #define	BAD_SNAP_INDEX	U32_MAX		/* invalid index into snap array */
103 
104 /* This allows a single page to hold an image name sent by OSD */
105 #define RBD_IMAGE_NAME_LEN_MAX	(PAGE_SIZE - sizeof (__le32) - 1)
106 #define RBD_IMAGE_ID_LEN_MAX	64
107 
108 #define RBD_OBJ_PREFIX_LEN_MAX	64
109 
110 #define RBD_NOTIFY_TIMEOUT	5	/* seconds */
111 #define RBD_RETRY_DELAY		msecs_to_jiffies(1000)
112 
113 /* Feature bits */
114 
115 #define RBD_FEATURE_LAYERING		(1ULL<<0)
116 #define RBD_FEATURE_STRIPINGV2		(1ULL<<1)
117 #define RBD_FEATURE_EXCLUSIVE_LOCK	(1ULL<<2)
118 #define RBD_FEATURE_OBJECT_MAP		(1ULL<<3)
119 #define RBD_FEATURE_FAST_DIFF		(1ULL<<4)
120 #define RBD_FEATURE_DEEP_FLATTEN	(1ULL<<5)
121 #define RBD_FEATURE_DATA_POOL		(1ULL<<7)
122 #define RBD_FEATURE_OPERATIONS		(1ULL<<8)
123 
124 #define RBD_FEATURES_ALL	(RBD_FEATURE_LAYERING |		\
125 				 RBD_FEATURE_STRIPINGV2 |	\
126 				 RBD_FEATURE_EXCLUSIVE_LOCK |	\
127 				 RBD_FEATURE_OBJECT_MAP |	\
128 				 RBD_FEATURE_FAST_DIFF |	\
129 				 RBD_FEATURE_DEEP_FLATTEN |	\
130 				 RBD_FEATURE_DATA_POOL |	\
131 				 RBD_FEATURE_OPERATIONS)
132 
133 /* Features supported by this (client software) implementation. */
134 
135 #define RBD_FEATURES_SUPPORTED	(RBD_FEATURES_ALL)
136 
137 /*
138  * An RBD device name will be "rbd#", where the "rbd" comes from
139  * RBD_DRV_NAME above, and # is a unique integer identifier.
140  */
141 #define DEV_NAME_LEN		32
142 
143 /*
144  * block device image metadata (in-memory version)
145  */
146 struct rbd_image_header {
147 	/* These six fields never change for a given rbd image */
148 	char *object_prefix;
149 	__u8 obj_order;
150 	u64 stripe_unit;
151 	u64 stripe_count;
152 	s64 data_pool_id;
153 	u64 features;		/* Might be changeable someday? */
154 
155 	/* The remaining fields need to be updated occasionally */
156 	u64 image_size;
157 	struct ceph_snap_context *snapc;
158 	char *snap_names;	/* format 1 only */
159 	u64 *snap_sizes;	/* format 1 only */
160 };
161 
162 /*
163  * An rbd image specification.
164  *
165  * The tuple (pool_id, image_id, snap_id) is sufficient to uniquely
166  * identify an image.  Each rbd_dev structure includes a pointer to
167  * an rbd_spec structure that encapsulates this identity.
168  *
169  * Each of the id's in an rbd_spec has an associated name.  For a
170  * user-mapped image, the names are supplied and the id's associated
171  * with them are looked up.  For a layered image, a parent image is
172  * defined by the tuple, and the names are looked up.
173  *
174  * An rbd_dev structure contains a parent_spec pointer which is
175  * non-null if the image it represents is a child in a layered
176  * image.  This pointer will refer to the rbd_spec structure used
177  * by the parent rbd_dev for its own identity (i.e., the structure
178  * is shared between the parent and child).
179  *
180  * Since these structures are populated once, during the discovery
181  * phase of image construction, they are effectively immutable so
182  * we make no effort to synchronize access to them.
183  *
184  * Note that code herein does not assume the image name is known (it
185  * could be a null pointer).
186  */
187 struct rbd_spec {
188 	u64		pool_id;
189 	const char	*pool_name;
190 	const char	*pool_ns;	/* NULL if default, never "" */
191 
192 	const char	*image_id;
193 	const char	*image_name;
194 
195 	u64		snap_id;
196 	const char	*snap_name;
197 
198 	struct kref	kref;
199 };
200 
201 /*
202  * an instance of the client.  multiple devices may share an rbd client.
203  */
204 struct rbd_client {
205 	struct ceph_client	*client;
206 	struct kref		kref;
207 	struct list_head	node;
208 };
209 
210 struct pending_result {
211 	int			result;		/* first nonzero result */
212 	int			num_pending;
213 };
214 
215 struct rbd_img_request;
216 
217 enum obj_request_type {
218 	OBJ_REQUEST_NODATA = 1,
219 	OBJ_REQUEST_BIO,	/* pointer into provided bio (list) */
220 	OBJ_REQUEST_BVECS,	/* pointer into provided bio_vec array */
221 	OBJ_REQUEST_OWN_BVECS,	/* private bio_vec array, doesn't own pages */
222 };
223 
224 enum obj_operation_type {
225 	OBJ_OP_READ = 1,
226 	OBJ_OP_WRITE,
227 	OBJ_OP_DISCARD,
228 	OBJ_OP_ZEROOUT,
229 };
230 
231 #define RBD_OBJ_FLAG_DELETION			(1U << 0)
232 #define RBD_OBJ_FLAG_COPYUP_ENABLED		(1U << 1)
233 #define RBD_OBJ_FLAG_COPYUP_ZEROS		(1U << 2)
234 #define RBD_OBJ_FLAG_MAY_EXIST			(1U << 3)
235 #define RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT	(1U << 4)
236 
237 enum rbd_obj_read_state {
238 	RBD_OBJ_READ_START = 1,
239 	RBD_OBJ_READ_OBJECT,
240 	RBD_OBJ_READ_PARENT,
241 };
242 
243 /*
244  * Writes go through the following state machine to deal with
245  * layering:
246  *
247  *            . . . . . RBD_OBJ_WRITE_GUARD. . . . . . . . . . . . . .
248  *            .                 |                                    .
249  *            .                 v                                    .
250  *            .    RBD_OBJ_WRITE_READ_FROM_PARENT. . .               .
251  *            .                 |                    .               .
252  *            .                 v                    v (deep-copyup  .
253  *    (image  .   RBD_OBJ_WRITE_COPYUP_EMPTY_SNAPC   .  not needed)  .
254  * flattened) v                 |                    .               .
255  *            .                 v                    .               .
256  *            . . . .RBD_OBJ_WRITE_COPYUP_OPS. . . . .      (copyup  .
257  *                              |                        not needed) v
258  *                              v                                    .
259  *                            done . . . . . . . . . . . . . . . . . .
260  *                              ^
261  *                              |
262  *                     RBD_OBJ_WRITE_FLAT
263  *
264  * Writes start in RBD_OBJ_WRITE_GUARD or _FLAT, depending on whether
265  * assert_exists guard is needed or not (in some cases it's not needed
266  * even if there is a parent).
267  */
268 enum rbd_obj_write_state {
269 	RBD_OBJ_WRITE_START = 1,
270 	RBD_OBJ_WRITE_PRE_OBJECT_MAP,
271 	RBD_OBJ_WRITE_OBJECT,
272 	__RBD_OBJ_WRITE_COPYUP,
273 	RBD_OBJ_WRITE_COPYUP,
274 	RBD_OBJ_WRITE_POST_OBJECT_MAP,
275 };
276 
277 enum rbd_obj_copyup_state {
278 	RBD_OBJ_COPYUP_START = 1,
279 	RBD_OBJ_COPYUP_READ_PARENT,
280 	__RBD_OBJ_COPYUP_OBJECT_MAPS,
281 	RBD_OBJ_COPYUP_OBJECT_MAPS,
282 	__RBD_OBJ_COPYUP_WRITE_OBJECT,
283 	RBD_OBJ_COPYUP_WRITE_OBJECT,
284 };
285 
286 struct rbd_obj_request {
287 	struct ceph_object_extent ex;
288 	unsigned int		flags;	/* RBD_OBJ_FLAG_* */
289 	union {
290 		enum rbd_obj_read_state	 read_state;	/* for reads */
291 		enum rbd_obj_write_state write_state;	/* for writes */
292 	};
293 
294 	struct rbd_img_request	*img_request;
295 	struct ceph_file_extent	*img_extents;
296 	u32			num_img_extents;
297 
298 	union {
299 		struct ceph_bio_iter	bio_pos;
300 		struct {
301 			struct ceph_bvec_iter	bvec_pos;
302 			u32			bvec_count;
303 			u32			bvec_idx;
304 		};
305 	};
306 
307 	enum rbd_obj_copyup_state copyup_state;
308 	struct bio_vec		*copyup_bvecs;
309 	u32			copyup_bvec_count;
310 
311 	struct list_head	osd_reqs;	/* w/ r_private_item */
312 
313 	struct mutex		state_mutex;
314 	struct pending_result	pending;
315 	struct kref		kref;
316 };
317 
318 enum img_req_flags {
319 	IMG_REQ_CHILD,		/* initiator: block = 0, child image = 1 */
320 	IMG_REQ_LAYERED,	/* ENOENT handling: normal = 0, layered = 1 */
321 };
322 
323 enum rbd_img_state {
324 	RBD_IMG_START = 1,
325 	RBD_IMG_EXCLUSIVE_LOCK,
326 	__RBD_IMG_OBJECT_REQUESTS,
327 	RBD_IMG_OBJECT_REQUESTS,
328 };
329 
330 struct rbd_img_request {
331 	struct rbd_device	*rbd_dev;
332 	enum obj_operation_type	op_type;
333 	enum obj_request_type	data_type;
334 	unsigned long		flags;
335 	enum rbd_img_state	state;
336 	union {
337 		u64			snap_id;	/* for reads */
338 		struct ceph_snap_context *snapc;	/* for writes */
339 	};
340 	struct rbd_obj_request	*obj_request;	/* obj req initiator */
341 
342 	struct list_head	lock_item;
343 	struct list_head	object_extents;	/* obj_req.ex structs */
344 
345 	struct mutex		state_mutex;
346 	struct pending_result	pending;
347 	struct work_struct	work;
348 	int			work_result;
349 };
350 
351 #define for_each_obj_request(ireq, oreq) \
352 	list_for_each_entry(oreq, &(ireq)->object_extents, ex.oe_item)
353 #define for_each_obj_request_safe(ireq, oreq, n) \
354 	list_for_each_entry_safe(oreq, n, &(ireq)->object_extents, ex.oe_item)
355 
356 enum rbd_watch_state {
357 	RBD_WATCH_STATE_UNREGISTERED,
358 	RBD_WATCH_STATE_REGISTERED,
359 	RBD_WATCH_STATE_ERROR,
360 };
361 
362 enum rbd_lock_state {
363 	RBD_LOCK_STATE_UNLOCKED,
364 	RBD_LOCK_STATE_LOCKED,
365 	RBD_LOCK_STATE_RELEASING,
366 };
367 
368 /* WatchNotify::ClientId */
369 struct rbd_client_id {
370 	u64 gid;
371 	u64 handle;
372 };
373 
374 struct rbd_mapping {
375 	u64                     size;
376 };
377 
378 /*
379  * a single device
380  */
381 struct rbd_device {
382 	int			dev_id;		/* blkdev unique id */
383 
384 	int			major;		/* blkdev assigned major */
385 	int			minor;
386 	struct gendisk		*disk;		/* blkdev's gendisk and rq */
387 
388 	u32			image_format;	/* Either 1 or 2 */
389 	struct rbd_client	*rbd_client;
390 
391 	char			name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */
392 
393 	spinlock_t		lock;		/* queue, flags, open_count */
394 
395 	struct rbd_image_header	header;
396 	unsigned long		flags;		/* possibly lock protected */
397 	struct rbd_spec		*spec;
398 	struct rbd_options	*opts;
399 	char			*config_info;	/* add{,_single_major} string */
400 
401 	struct ceph_object_id	header_oid;
402 	struct ceph_object_locator header_oloc;
403 
404 	struct ceph_file_layout	layout;		/* used for all rbd requests */
405 
406 	struct mutex		watch_mutex;
407 	enum rbd_watch_state	watch_state;
408 	struct ceph_osd_linger_request *watch_handle;
409 	u64			watch_cookie;
410 	struct delayed_work	watch_dwork;
411 
412 	struct rw_semaphore	lock_rwsem;
413 	enum rbd_lock_state	lock_state;
414 	char			lock_cookie[32];
415 	struct rbd_client_id	owner_cid;
416 	struct work_struct	acquired_lock_work;
417 	struct work_struct	released_lock_work;
418 	struct delayed_work	lock_dwork;
419 	struct work_struct	unlock_work;
420 	spinlock_t		lock_lists_lock;
421 	struct list_head	acquiring_list;
422 	struct list_head	running_list;
423 	struct completion	acquire_wait;
424 	int			acquire_err;
425 	struct completion	releasing_wait;
426 
427 	spinlock_t		object_map_lock;
428 	u8			*object_map;
429 	u64			object_map_size;	/* in objects */
430 	u64			object_map_flags;
431 
432 	struct workqueue_struct	*task_wq;
433 
434 	struct rbd_spec		*parent_spec;
435 	u64			parent_overlap;
436 	atomic_t		parent_ref;
437 	struct rbd_device	*parent;
438 
439 	/* Block layer tags. */
440 	struct blk_mq_tag_set	tag_set;
441 
442 	/* protects updating the header */
443 	struct rw_semaphore     header_rwsem;
444 
445 	struct rbd_mapping	mapping;
446 
447 	struct list_head	node;
448 
449 	/* sysfs related */
450 	struct device		dev;
451 	unsigned long		open_count;	/* protected by lock */
452 };
453 
454 /*
455  * Flag bits for rbd_dev->flags:
456  * - REMOVING (which is coupled with rbd_dev->open_count) is protected
457  *   by rbd_dev->lock
458  */
459 enum rbd_dev_flags {
460 	RBD_DEV_FLAG_EXISTS,	/* rbd_dev_device_setup() ran */
461 	RBD_DEV_FLAG_REMOVING,	/* this mapping is being removed */
462 	RBD_DEV_FLAG_READONLY,  /* -o ro or snapshot */
463 };
464 
465 static DEFINE_MUTEX(client_mutex);	/* Serialize client creation */
466 
467 static LIST_HEAD(rbd_dev_list);    /* devices */
468 static DEFINE_SPINLOCK(rbd_dev_list_lock);
469 
470 static LIST_HEAD(rbd_client_list);		/* clients */
471 static DEFINE_SPINLOCK(rbd_client_list_lock);
472 
473 /* Slab caches for frequently-allocated structures */
474 
475 static struct kmem_cache	*rbd_img_request_cache;
476 static struct kmem_cache	*rbd_obj_request_cache;
477 
478 static int rbd_major;
479 static DEFINE_IDA(rbd_dev_id_ida);
480 
481 static struct workqueue_struct *rbd_wq;
482 
483 static struct ceph_snap_context rbd_empty_snapc = {
484 	.nref = REFCOUNT_INIT(1),
485 };
486 
487 /*
488  * single-major requires >= 0.75 version of userspace rbd utility.
489  */
490 static bool single_major = true;
491 module_param(single_major, bool, 0444);
492 MODULE_PARM_DESC(single_major, "Use a single major number for all rbd devices (default: true)");
493 
494 static ssize_t add_store(const struct bus_type *bus, const char *buf, size_t count);
495 static ssize_t remove_store(const struct bus_type *bus, const char *buf,
496 			    size_t count);
497 static ssize_t add_single_major_store(const struct bus_type *bus, const char *buf,
498 				      size_t count);
499 static ssize_t remove_single_major_store(const struct bus_type *bus, const char *buf,
500 					 size_t count);
501 static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth);
502 
503 static int rbd_dev_id_to_minor(int dev_id)
504 {
505 	return dev_id << RBD_SINGLE_MAJOR_PART_SHIFT;
506 }
507 
508 static int minor_to_rbd_dev_id(int minor)
509 {
510 	return minor >> RBD_SINGLE_MAJOR_PART_SHIFT;
511 }
512 
513 static bool rbd_is_ro(struct rbd_device *rbd_dev)
514 {
515 	return test_bit(RBD_DEV_FLAG_READONLY, &rbd_dev->flags);
516 }
517 
518 static bool rbd_is_snap(struct rbd_device *rbd_dev)
519 {
520 	return rbd_dev->spec->snap_id != CEPH_NOSNAP;
521 }
522 
523 static bool __rbd_is_lock_owner(struct rbd_device *rbd_dev)
524 {
525 	lockdep_assert_held(&rbd_dev->lock_rwsem);
526 
527 	return rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED ||
528 	       rbd_dev->lock_state == RBD_LOCK_STATE_RELEASING;
529 }
530 
531 static bool rbd_is_lock_owner(struct rbd_device *rbd_dev)
532 {
533 	bool is_lock_owner;
534 
535 	down_read(&rbd_dev->lock_rwsem);
536 	is_lock_owner = __rbd_is_lock_owner(rbd_dev);
537 	up_read(&rbd_dev->lock_rwsem);
538 	return is_lock_owner;
539 }
540 
541 static ssize_t supported_features_show(const struct bus_type *bus, char *buf)
542 {
543 	return sprintf(buf, "0x%llx\n", RBD_FEATURES_SUPPORTED);
544 }
545 
546 static BUS_ATTR_WO(add);
547 static BUS_ATTR_WO(remove);
548 static BUS_ATTR_WO(add_single_major);
549 static BUS_ATTR_WO(remove_single_major);
550 static BUS_ATTR_RO(supported_features);
551 
552 static struct attribute *rbd_bus_attrs[] = {
553 	&bus_attr_add.attr,
554 	&bus_attr_remove.attr,
555 	&bus_attr_add_single_major.attr,
556 	&bus_attr_remove_single_major.attr,
557 	&bus_attr_supported_features.attr,
558 	NULL,
559 };
560 
561 static umode_t rbd_bus_is_visible(struct kobject *kobj,
562 				  struct attribute *attr, int index)
563 {
564 	if (!single_major &&
565 	    (attr == &bus_attr_add_single_major.attr ||
566 	     attr == &bus_attr_remove_single_major.attr))
567 		return 0;
568 
569 	return attr->mode;
570 }
571 
572 static const struct attribute_group rbd_bus_group = {
573 	.attrs = rbd_bus_attrs,
574 	.is_visible = rbd_bus_is_visible,
575 };
576 __ATTRIBUTE_GROUPS(rbd_bus);
577 
578 static struct bus_type rbd_bus_type = {
579 	.name		= "rbd",
580 	.bus_groups	= rbd_bus_groups,
581 };
582 
583 static void rbd_root_dev_release(struct device *dev)
584 {
585 }
586 
587 static struct device rbd_root_dev = {
588 	.init_name =    "rbd",
589 	.release =      rbd_root_dev_release,
590 };
591 
592 static __printf(2, 3)
593 void rbd_warn(struct rbd_device *rbd_dev, const char *fmt, ...)
594 {
595 	struct va_format vaf;
596 	va_list args;
597 
598 	va_start(args, fmt);
599 	vaf.fmt = fmt;
600 	vaf.va = &args;
601 
602 	if (!rbd_dev)
603 		printk(KERN_WARNING "%s: %pV\n", RBD_DRV_NAME, &vaf);
604 	else if (rbd_dev->disk)
605 		printk(KERN_WARNING "%s: %s: %pV\n",
606 			RBD_DRV_NAME, rbd_dev->disk->disk_name, &vaf);
607 	else if (rbd_dev->spec && rbd_dev->spec->image_name)
608 		printk(KERN_WARNING "%s: image %s: %pV\n",
609 			RBD_DRV_NAME, rbd_dev->spec->image_name, &vaf);
610 	else if (rbd_dev->spec && rbd_dev->spec->image_id)
611 		printk(KERN_WARNING "%s: id %s: %pV\n",
612 			RBD_DRV_NAME, rbd_dev->spec->image_id, &vaf);
613 	else	/* punt */
614 		printk(KERN_WARNING "%s: rbd_dev %p: %pV\n",
615 			RBD_DRV_NAME, rbd_dev, &vaf);
616 	va_end(args);
617 }
618 
619 #ifdef RBD_DEBUG
620 #define rbd_assert(expr)						\
621 		if (unlikely(!(expr))) {				\
622 			printk(KERN_ERR "\nAssertion failure in %s() "	\
623 						"at line %d:\n\n"	\
624 					"\trbd_assert(%s);\n\n",	\
625 					__func__, __LINE__, #expr);	\
626 			BUG();						\
627 		}
628 #else /* !RBD_DEBUG */
629 #  define rbd_assert(expr)	((void) 0)
630 #endif /* !RBD_DEBUG */
631 
632 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);
633 
634 static int rbd_dev_refresh(struct rbd_device *rbd_dev);
635 static int rbd_dev_v2_header_onetime(struct rbd_device *rbd_dev);
636 static int rbd_dev_header_info(struct rbd_device *rbd_dev);
637 static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev);
638 static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev,
639 					u64 snap_id);
640 static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
641 				u8 *order, u64 *snap_size);
642 static int rbd_dev_v2_get_flags(struct rbd_device *rbd_dev);
643 
644 static void rbd_obj_handle_request(struct rbd_obj_request *obj_req, int result);
645 static void rbd_img_handle_request(struct rbd_img_request *img_req, int result);
646 
647 /*
648  * Return true if nothing else is pending.
649  */
650 static bool pending_result_dec(struct pending_result *pending, int *result)
651 {
652 	rbd_assert(pending->num_pending > 0);
653 
654 	if (*result && !pending->result)
655 		pending->result = *result;
656 	if (--pending->num_pending)
657 		return false;
658 
659 	*result = pending->result;
660 	return true;
661 }
662 
663 static int rbd_open(struct gendisk *disk, blk_mode_t mode)
664 {
665 	struct rbd_device *rbd_dev = disk->private_data;
666 	bool removing = false;
667 
668 	spin_lock_irq(&rbd_dev->lock);
669 	if (test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags))
670 		removing = true;
671 	else
672 		rbd_dev->open_count++;
673 	spin_unlock_irq(&rbd_dev->lock);
674 	if (removing)
675 		return -ENOENT;
676 
677 	(void) get_device(&rbd_dev->dev);
678 
679 	return 0;
680 }
681 
682 static void rbd_release(struct gendisk *disk)
683 {
684 	struct rbd_device *rbd_dev = disk->private_data;
685 	unsigned long open_count_before;
686 
687 	spin_lock_irq(&rbd_dev->lock);
688 	open_count_before = rbd_dev->open_count--;
689 	spin_unlock_irq(&rbd_dev->lock);
690 	rbd_assert(open_count_before > 0);
691 
692 	put_device(&rbd_dev->dev);
693 }
694 
695 static const struct block_device_operations rbd_bd_ops = {
696 	.owner			= THIS_MODULE,
697 	.open			= rbd_open,
698 	.release		= rbd_release,
699 };
700 
701 /*
702  * Initialize an rbd client instance.  Success or not, this function
703  * consumes ceph_opts.  Caller holds client_mutex.
704  */
705 static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
706 {
707 	struct rbd_client *rbdc;
708 	int ret = -ENOMEM;
709 
710 	dout("%s:\n", __func__);
711 	rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL);
712 	if (!rbdc)
713 		goto out_opt;
714 
715 	kref_init(&rbdc->kref);
716 	INIT_LIST_HEAD(&rbdc->node);
717 
718 	rbdc->client = ceph_create_client(ceph_opts, rbdc);
719 	if (IS_ERR(rbdc->client))
720 		goto out_rbdc;
721 	ceph_opts = NULL; /* Now rbdc->client is responsible for ceph_opts */
722 
723 	ret = ceph_open_session(rbdc->client);
724 	if (ret < 0)
725 		goto out_client;
726 
727 	spin_lock(&rbd_client_list_lock);
728 	list_add_tail(&rbdc->node, &rbd_client_list);
729 	spin_unlock(&rbd_client_list_lock);
730 
731 	dout("%s: rbdc %p\n", __func__, rbdc);
732 
733 	return rbdc;
734 out_client:
735 	ceph_destroy_client(rbdc->client);
736 out_rbdc:
737 	kfree(rbdc);
738 out_opt:
739 	if (ceph_opts)
740 		ceph_destroy_options(ceph_opts);
741 	dout("%s: error %d\n", __func__, ret);
742 
743 	return ERR_PTR(ret);
744 }
745 
746 static struct rbd_client *__rbd_get_client(struct rbd_client *rbdc)
747 {
748 	kref_get(&rbdc->kref);
749 
750 	return rbdc;
751 }
752 
753 /*
754  * Find a ceph client with specific addr and configuration.  If
755  * found, bump its reference count.
756  */
757 static struct rbd_client *rbd_client_find(struct ceph_options *ceph_opts)
758 {
759 	struct rbd_client *rbdc = NULL, *iter;
760 
761 	if (ceph_opts->flags & CEPH_OPT_NOSHARE)
762 		return NULL;
763 
764 	spin_lock(&rbd_client_list_lock);
765 	list_for_each_entry(iter, &rbd_client_list, node) {
766 		if (!ceph_compare_options(ceph_opts, iter->client)) {
767 			__rbd_get_client(iter);
768 
769 			rbdc = iter;
770 			break;
771 		}
772 	}
773 	spin_unlock(&rbd_client_list_lock);
774 
775 	return rbdc;
776 }
777 
778 /*
779  * (Per device) rbd map options
780  */
781 enum {
782 	Opt_queue_depth,
783 	Opt_alloc_size,
784 	Opt_lock_timeout,
785 	/* int args above */
786 	Opt_pool_ns,
787 	Opt_compression_hint,
788 	/* string args above */
789 	Opt_read_only,
790 	Opt_read_write,
791 	Opt_lock_on_read,
792 	Opt_exclusive,
793 	Opt_notrim,
794 };
795 
796 enum {
797 	Opt_compression_hint_none,
798 	Opt_compression_hint_compressible,
799 	Opt_compression_hint_incompressible,
800 };
801 
802 static const struct constant_table rbd_param_compression_hint[] = {
803 	{"none",		Opt_compression_hint_none},
804 	{"compressible",	Opt_compression_hint_compressible},
805 	{"incompressible",	Opt_compression_hint_incompressible},
806 	{}
807 };
808 
809 static const struct fs_parameter_spec rbd_parameters[] = {
810 	fsparam_u32	("alloc_size",			Opt_alloc_size),
811 	fsparam_enum	("compression_hint",		Opt_compression_hint,
812 			 rbd_param_compression_hint),
813 	fsparam_flag	("exclusive",			Opt_exclusive),
814 	fsparam_flag	("lock_on_read",		Opt_lock_on_read),
815 	fsparam_u32	("lock_timeout",		Opt_lock_timeout),
816 	fsparam_flag	("notrim",			Opt_notrim),
817 	fsparam_string	("_pool_ns",			Opt_pool_ns),
818 	fsparam_u32	("queue_depth",			Opt_queue_depth),
819 	fsparam_flag	("read_only",			Opt_read_only),
820 	fsparam_flag	("read_write",			Opt_read_write),
821 	fsparam_flag	("ro",				Opt_read_only),
822 	fsparam_flag	("rw",				Opt_read_write),
823 	{}
824 };
825 
826 struct rbd_options {
827 	int	queue_depth;
828 	int	alloc_size;
829 	unsigned long	lock_timeout;
830 	bool	read_only;
831 	bool	lock_on_read;
832 	bool	exclusive;
833 	bool	trim;
834 
835 	u32 alloc_hint_flags;  /* CEPH_OSD_OP_ALLOC_HINT_FLAG_* */
836 };
837 
838 #define RBD_QUEUE_DEPTH_DEFAULT	BLKDEV_DEFAULT_RQ
839 #define RBD_ALLOC_SIZE_DEFAULT	(64 * 1024)
840 #define RBD_LOCK_TIMEOUT_DEFAULT 0  /* no timeout */
841 #define RBD_READ_ONLY_DEFAULT	false
842 #define RBD_LOCK_ON_READ_DEFAULT false
843 #define RBD_EXCLUSIVE_DEFAULT	false
844 #define RBD_TRIM_DEFAULT	true
845 
846 struct rbd_parse_opts_ctx {
847 	struct rbd_spec		*spec;
848 	struct ceph_options	*copts;
849 	struct rbd_options	*opts;
850 };
851 
852 static char* obj_op_name(enum obj_operation_type op_type)
853 {
854 	switch (op_type) {
855 	case OBJ_OP_READ:
856 		return "read";
857 	case OBJ_OP_WRITE:
858 		return "write";
859 	case OBJ_OP_DISCARD:
860 		return "discard";
861 	case OBJ_OP_ZEROOUT:
862 		return "zeroout";
863 	default:
864 		return "???";
865 	}
866 }
867 
868 /*
869  * Destroy ceph client
870  *
871  * Caller must hold rbd_client_list_lock.
872  */
873 static void rbd_client_release(struct kref *kref)
874 {
875 	struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref);
876 
877 	dout("%s: rbdc %p\n", __func__, rbdc);
878 	spin_lock(&rbd_client_list_lock);
879 	list_del(&rbdc->node);
880 	spin_unlock(&rbd_client_list_lock);
881 
882 	ceph_destroy_client(rbdc->client);
883 	kfree(rbdc);
884 }
885 
886 /*
887  * Drop reference to ceph client node. If it's not referenced anymore, release
888  * it.
889  */
890 static void rbd_put_client(struct rbd_client *rbdc)
891 {
892 	if (rbdc)
893 		kref_put(&rbdc->kref, rbd_client_release);
894 }
895 
896 /*
897  * Get a ceph client with specific addr and configuration, if one does
898  * not exist create it.  Either way, ceph_opts is consumed by this
899  * function.
900  */
901 static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts)
902 {
903 	struct rbd_client *rbdc;
904 	int ret;
905 
906 	mutex_lock(&client_mutex);
907 	rbdc = rbd_client_find(ceph_opts);
908 	if (rbdc) {
909 		ceph_destroy_options(ceph_opts);
910 
911 		/*
912 		 * Using an existing client.  Make sure ->pg_pools is up to
913 		 * date before we look up the pool id in do_rbd_add().
914 		 */
915 		ret = ceph_wait_for_latest_osdmap(rbdc->client,
916 					rbdc->client->options->mount_timeout);
917 		if (ret) {
918 			rbd_warn(NULL, "failed to get latest osdmap: %d", ret);
919 			rbd_put_client(rbdc);
920 			rbdc = ERR_PTR(ret);
921 		}
922 	} else {
923 		rbdc = rbd_client_create(ceph_opts);
924 	}
925 	mutex_unlock(&client_mutex);
926 
927 	return rbdc;
928 }
929 
930 static bool rbd_image_format_valid(u32 image_format)
931 {
932 	return image_format == 1 || image_format == 2;
933 }
934 
935 static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
936 {
937 	size_t size;
938 	u32 snap_count;
939 
940 	/* The header has to start with the magic rbd header text */
941 	if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT)))
942 		return false;
943 
944 	/* The bio layer requires at least sector-sized I/O */
945 
946 	if (ondisk->options.order < SECTOR_SHIFT)
947 		return false;
948 
949 	/* If we use u64 in a few spots we may be able to loosen this */
950 
951 	if (ondisk->options.order > 8 * sizeof (int) - 1)
952 		return false;
953 
954 	/*
955 	 * The size of a snapshot header has to fit in a size_t, and
956 	 * that limits the number of snapshots.
957 	 */
958 	snap_count = le32_to_cpu(ondisk->snap_count);
959 	size = SIZE_MAX - sizeof (struct ceph_snap_context);
960 	if (snap_count > size / sizeof (__le64))
961 		return false;
962 
963 	/*
964 	 * Not only that, but the size of the entire the snapshot
965 	 * header must also be representable in a size_t.
966 	 */
967 	size -= snap_count * sizeof (__le64);
968 	if ((u64) size < le64_to_cpu(ondisk->snap_names_len))
969 		return false;
970 
971 	return true;
972 }
973 
974 /*
975  * returns the size of an object in the image
976  */
977 static u32 rbd_obj_bytes(struct rbd_image_header *header)
978 {
979 	return 1U << header->obj_order;
980 }
981 
982 static void rbd_init_layout(struct rbd_device *rbd_dev)
983 {
984 	if (rbd_dev->header.stripe_unit == 0 ||
985 	    rbd_dev->header.stripe_count == 0) {
986 		rbd_dev->header.stripe_unit = rbd_obj_bytes(&rbd_dev->header);
987 		rbd_dev->header.stripe_count = 1;
988 	}
989 
990 	rbd_dev->layout.stripe_unit = rbd_dev->header.stripe_unit;
991 	rbd_dev->layout.stripe_count = rbd_dev->header.stripe_count;
992 	rbd_dev->layout.object_size = rbd_obj_bytes(&rbd_dev->header);
993 	rbd_dev->layout.pool_id = rbd_dev->header.data_pool_id == CEPH_NOPOOL ?
994 			  rbd_dev->spec->pool_id : rbd_dev->header.data_pool_id;
995 	RCU_INIT_POINTER(rbd_dev->layout.pool_ns, NULL);
996 }
997 
998 /*
999  * Fill an rbd image header with information from the given format 1
1000  * on-disk header.
1001  */
1002 static int rbd_header_from_disk(struct rbd_device *rbd_dev,
1003 				 struct rbd_image_header_ondisk *ondisk)
1004 {
1005 	struct rbd_image_header *header = &rbd_dev->header;
1006 	bool first_time = header->object_prefix == NULL;
1007 	struct ceph_snap_context *snapc;
1008 	char *object_prefix = NULL;
1009 	char *snap_names = NULL;
1010 	u64 *snap_sizes = NULL;
1011 	u32 snap_count;
1012 	int ret = -ENOMEM;
1013 	u32 i;
1014 
1015 	/* Allocate this now to avoid having to handle failure below */
1016 
1017 	if (first_time) {
1018 		object_prefix = kstrndup(ondisk->object_prefix,
1019 					 sizeof(ondisk->object_prefix),
1020 					 GFP_KERNEL);
1021 		if (!object_prefix)
1022 			return -ENOMEM;
1023 	}
1024 
1025 	/* Allocate the snapshot context and fill it in */
1026 
1027 	snap_count = le32_to_cpu(ondisk->snap_count);
1028 	snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
1029 	if (!snapc)
1030 		goto out_err;
1031 	snapc->seq = le64_to_cpu(ondisk->snap_seq);
1032 	if (snap_count) {
1033 		struct rbd_image_snap_ondisk *snaps;
1034 		u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len);
1035 
1036 		/* We'll keep a copy of the snapshot names... */
1037 
1038 		if (snap_names_len > (u64)SIZE_MAX)
1039 			goto out_2big;
1040 		snap_names = kmalloc(snap_names_len, GFP_KERNEL);
1041 		if (!snap_names)
1042 			goto out_err;
1043 
1044 		/* ...as well as the array of their sizes. */
1045 		snap_sizes = kmalloc_array(snap_count,
1046 					   sizeof(*header->snap_sizes),
1047 					   GFP_KERNEL);
1048 		if (!snap_sizes)
1049 			goto out_err;
1050 
1051 		/*
1052 		 * Copy the names, and fill in each snapshot's id
1053 		 * and size.
1054 		 *
1055 		 * Note that rbd_dev_v1_header_info() guarantees the
1056 		 * ondisk buffer we're working with has
1057 		 * snap_names_len bytes beyond the end of the
1058 		 * snapshot id array, this memcpy() is safe.
1059 		 */
1060 		memcpy(snap_names, &ondisk->snaps[snap_count], snap_names_len);
1061 		snaps = ondisk->snaps;
1062 		for (i = 0; i < snap_count; i++) {
1063 			snapc->snaps[i] = le64_to_cpu(snaps[i].id);
1064 			snap_sizes[i] = le64_to_cpu(snaps[i].image_size);
1065 		}
1066 	}
1067 
1068 	/* We won't fail any more, fill in the header */
1069 
1070 	if (first_time) {
1071 		header->object_prefix = object_prefix;
1072 		header->obj_order = ondisk->options.order;
1073 		rbd_init_layout(rbd_dev);
1074 	} else {
1075 		ceph_put_snap_context(header->snapc);
1076 		kfree(header->snap_names);
1077 		kfree(header->snap_sizes);
1078 	}
1079 
1080 	/* The remaining fields always get updated (when we refresh) */
1081 
1082 	header->image_size = le64_to_cpu(ondisk->image_size);
1083 	header->snapc = snapc;
1084 	header->snap_names = snap_names;
1085 	header->snap_sizes = snap_sizes;
1086 
1087 	return 0;
1088 out_2big:
1089 	ret = -EIO;
1090 out_err:
1091 	kfree(snap_sizes);
1092 	kfree(snap_names);
1093 	ceph_put_snap_context(snapc);
1094 	kfree(object_prefix);
1095 
1096 	return ret;
1097 }
1098 
1099 static const char *_rbd_dev_v1_snap_name(struct rbd_device *rbd_dev, u32 which)
1100 {
1101 	const char *snap_name;
1102 
1103 	rbd_assert(which < rbd_dev->header.snapc->num_snaps);
1104 
1105 	/* Skip over names until we find the one we are looking for */
1106 
1107 	snap_name = rbd_dev->header.snap_names;
1108 	while (which--)
1109 		snap_name += strlen(snap_name) + 1;
1110 
1111 	return kstrdup(snap_name, GFP_KERNEL);
1112 }
1113 
1114 /*
1115  * Snapshot id comparison function for use with qsort()/bsearch().
1116  * Note that result is for snapshots in *descending* order.
1117  */
1118 static int snapid_compare_reverse(const void *s1, const void *s2)
1119 {
1120 	u64 snap_id1 = *(u64 *)s1;
1121 	u64 snap_id2 = *(u64 *)s2;
1122 
1123 	if (snap_id1 < snap_id2)
1124 		return 1;
1125 	return snap_id1 == snap_id2 ? 0 : -1;
1126 }
1127 
1128 /*
1129  * Search a snapshot context to see if the given snapshot id is
1130  * present.
1131  *
1132  * Returns the position of the snapshot id in the array if it's found,
1133  * or BAD_SNAP_INDEX otherwise.
1134  *
1135  * Note: The snapshot array is in kept sorted (by the osd) in
1136  * reverse order, highest snapshot id first.
1137  */
1138 static u32 rbd_dev_snap_index(struct rbd_device *rbd_dev, u64 snap_id)
1139 {
1140 	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
1141 	u64 *found;
1142 
1143 	found = bsearch(&snap_id, &snapc->snaps, snapc->num_snaps,
1144 				sizeof (snap_id), snapid_compare_reverse);
1145 
1146 	return found ? (u32)(found - &snapc->snaps[0]) : BAD_SNAP_INDEX;
1147 }
1148 
1149 static const char *rbd_dev_v1_snap_name(struct rbd_device *rbd_dev,
1150 					u64 snap_id)
1151 {
1152 	u32 which;
1153 	const char *snap_name;
1154 
1155 	which = rbd_dev_snap_index(rbd_dev, snap_id);
1156 	if (which == BAD_SNAP_INDEX)
1157 		return ERR_PTR(-ENOENT);
1158 
1159 	snap_name = _rbd_dev_v1_snap_name(rbd_dev, which);
1160 	return snap_name ? snap_name : ERR_PTR(-ENOMEM);
1161 }
1162 
1163 static const char *rbd_snap_name(struct rbd_device *rbd_dev, u64 snap_id)
1164 {
1165 	if (snap_id == CEPH_NOSNAP)
1166 		return RBD_SNAP_HEAD_NAME;
1167 
1168 	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
1169 	if (rbd_dev->image_format == 1)
1170 		return rbd_dev_v1_snap_name(rbd_dev, snap_id);
1171 
1172 	return rbd_dev_v2_snap_name(rbd_dev, snap_id);
1173 }
1174 
1175 static int rbd_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
1176 				u64 *snap_size)
1177 {
1178 	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
1179 	if (snap_id == CEPH_NOSNAP) {
1180 		*snap_size = rbd_dev->header.image_size;
1181 	} else if (rbd_dev->image_format == 1) {
1182 		u32 which;
1183 
1184 		which = rbd_dev_snap_index(rbd_dev, snap_id);
1185 		if (which == BAD_SNAP_INDEX)
1186 			return -ENOENT;
1187 
1188 		*snap_size = rbd_dev->header.snap_sizes[which];
1189 	} else {
1190 		u64 size = 0;
1191 		int ret;
1192 
1193 		ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size);
1194 		if (ret)
1195 			return ret;
1196 
1197 		*snap_size = size;
1198 	}
1199 	return 0;
1200 }
1201 
1202 static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
1203 {
1204 	u64 snap_id = rbd_dev->spec->snap_id;
1205 	u64 size = 0;
1206 	int ret;
1207 
1208 	ret = rbd_snap_size(rbd_dev, snap_id, &size);
1209 	if (ret)
1210 		return ret;
1211 
1212 	rbd_dev->mapping.size = size;
1213 	return 0;
1214 }
1215 
1216 static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
1217 {
1218 	rbd_dev->mapping.size = 0;
1219 }
1220 
1221 static void zero_bios(struct ceph_bio_iter *bio_pos, u32 off, u32 bytes)
1222 {
1223 	struct ceph_bio_iter it = *bio_pos;
1224 
1225 	ceph_bio_iter_advance(&it, off);
1226 	ceph_bio_iter_advance_step(&it, bytes, ({
1227 		memzero_bvec(&bv);
1228 	}));
1229 }
1230 
1231 static void zero_bvecs(struct ceph_bvec_iter *bvec_pos, u32 off, u32 bytes)
1232 {
1233 	struct ceph_bvec_iter it = *bvec_pos;
1234 
1235 	ceph_bvec_iter_advance(&it, off);
1236 	ceph_bvec_iter_advance_step(&it, bytes, ({
1237 		memzero_bvec(&bv);
1238 	}));
1239 }
1240 
1241 /*
1242  * Zero a range in @obj_req data buffer defined by a bio (list) or
1243  * (private) bio_vec array.
1244  *
1245  * @off is relative to the start of the data buffer.
1246  */
1247 static void rbd_obj_zero_range(struct rbd_obj_request *obj_req, u32 off,
1248 			       u32 bytes)
1249 {
1250 	dout("%s %p data buf %u~%u\n", __func__, obj_req, off, bytes);
1251 
1252 	switch (obj_req->img_request->data_type) {
1253 	case OBJ_REQUEST_BIO:
1254 		zero_bios(&obj_req->bio_pos, off, bytes);
1255 		break;
1256 	case OBJ_REQUEST_BVECS:
1257 	case OBJ_REQUEST_OWN_BVECS:
1258 		zero_bvecs(&obj_req->bvec_pos, off, bytes);
1259 		break;
1260 	default:
1261 		BUG();
1262 	}
1263 }
1264 
1265 static void rbd_obj_request_destroy(struct kref *kref);
1266 static void rbd_obj_request_put(struct rbd_obj_request *obj_request)
1267 {
1268 	rbd_assert(obj_request != NULL);
1269 	dout("%s: obj %p (was %d)\n", __func__, obj_request,
1270 		kref_read(&obj_request->kref));
1271 	kref_put(&obj_request->kref, rbd_obj_request_destroy);
1272 }
1273 
1274 static inline void rbd_img_obj_request_add(struct rbd_img_request *img_request,
1275 					struct rbd_obj_request *obj_request)
1276 {
1277 	rbd_assert(obj_request->img_request == NULL);
1278 
1279 	/* Image request now owns object's original reference */
1280 	obj_request->img_request = img_request;
1281 	dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
1282 }
1283 
1284 static inline void rbd_img_obj_request_del(struct rbd_img_request *img_request,
1285 					struct rbd_obj_request *obj_request)
1286 {
1287 	dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
1288 	list_del(&obj_request->ex.oe_item);
1289 	rbd_assert(obj_request->img_request == img_request);
1290 	rbd_obj_request_put(obj_request);
1291 }
1292 
1293 static void rbd_osd_submit(struct ceph_osd_request *osd_req)
1294 {
1295 	struct rbd_obj_request *obj_req = osd_req->r_priv;
1296 
1297 	dout("%s osd_req %p for obj_req %p objno %llu %llu~%llu\n",
1298 	     __func__, osd_req, obj_req, obj_req->ex.oe_objno,
1299 	     obj_req->ex.oe_off, obj_req->ex.oe_len);
1300 	ceph_osdc_start_request(osd_req->r_osdc, osd_req);
1301 }
1302 
1303 /*
1304  * The default/initial value for all image request flags is 0.  Each
1305  * is conditionally set to 1 at image request initialization time
1306  * and currently never change thereafter.
1307  */
1308 static void img_request_layered_set(struct rbd_img_request *img_request)
1309 {
1310 	set_bit(IMG_REQ_LAYERED, &img_request->flags);
1311 }
1312 
1313 static bool img_request_layered_test(struct rbd_img_request *img_request)
1314 {
1315 	return test_bit(IMG_REQ_LAYERED, &img_request->flags) != 0;
1316 }
1317 
1318 static bool rbd_obj_is_entire(struct rbd_obj_request *obj_req)
1319 {
1320 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
1321 
1322 	return !obj_req->ex.oe_off &&
1323 	       obj_req->ex.oe_len == rbd_dev->layout.object_size;
1324 }
1325 
1326 static bool rbd_obj_is_tail(struct rbd_obj_request *obj_req)
1327 {
1328 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
1329 
1330 	return obj_req->ex.oe_off + obj_req->ex.oe_len ==
1331 					rbd_dev->layout.object_size;
1332 }
1333 
1334 /*
1335  * Must be called after rbd_obj_calc_img_extents().
1336  */
1337 static void rbd_obj_set_copyup_enabled(struct rbd_obj_request *obj_req)
1338 {
1339 	rbd_assert(obj_req->img_request->snapc);
1340 
1341 	if (obj_req->img_request->op_type == OBJ_OP_DISCARD) {
1342 		dout("%s %p objno %llu discard\n", __func__, obj_req,
1343 		     obj_req->ex.oe_objno);
1344 		return;
1345 	}
1346 
1347 	if (!obj_req->num_img_extents) {
1348 		dout("%s %p objno %llu not overlapping\n", __func__, obj_req,
1349 		     obj_req->ex.oe_objno);
1350 		return;
1351 	}
1352 
1353 	if (rbd_obj_is_entire(obj_req) &&
1354 	    !obj_req->img_request->snapc->num_snaps) {
1355 		dout("%s %p objno %llu entire\n", __func__, obj_req,
1356 		     obj_req->ex.oe_objno);
1357 		return;
1358 	}
1359 
1360 	obj_req->flags |= RBD_OBJ_FLAG_COPYUP_ENABLED;
1361 }
1362 
1363 static u64 rbd_obj_img_extents_bytes(struct rbd_obj_request *obj_req)
1364 {
1365 	return ceph_file_extents_bytes(obj_req->img_extents,
1366 				       obj_req->num_img_extents);
1367 }
1368 
1369 static bool rbd_img_is_write(struct rbd_img_request *img_req)
1370 {
1371 	switch (img_req->op_type) {
1372 	case OBJ_OP_READ:
1373 		return false;
1374 	case OBJ_OP_WRITE:
1375 	case OBJ_OP_DISCARD:
1376 	case OBJ_OP_ZEROOUT:
1377 		return true;
1378 	default:
1379 		BUG();
1380 	}
1381 }
1382 
1383 static void rbd_osd_req_callback(struct ceph_osd_request *osd_req)
1384 {
1385 	struct rbd_obj_request *obj_req = osd_req->r_priv;
1386 	int result;
1387 
1388 	dout("%s osd_req %p result %d for obj_req %p\n", __func__, osd_req,
1389 	     osd_req->r_result, obj_req);
1390 
1391 	/*
1392 	 * Writes aren't allowed to return a data payload.  In some
1393 	 * guarded write cases (e.g. stat + zero on an empty object)
1394 	 * a stat response makes it through, but we don't care.
1395 	 */
1396 	if (osd_req->r_result > 0 && rbd_img_is_write(obj_req->img_request))
1397 		result = 0;
1398 	else
1399 		result = osd_req->r_result;
1400 
1401 	rbd_obj_handle_request(obj_req, result);
1402 }
1403 
1404 static void rbd_osd_format_read(struct ceph_osd_request *osd_req)
1405 {
1406 	struct rbd_obj_request *obj_request = osd_req->r_priv;
1407 	struct rbd_device *rbd_dev = obj_request->img_request->rbd_dev;
1408 	struct ceph_options *opt = rbd_dev->rbd_client->client->options;
1409 
1410 	osd_req->r_flags = CEPH_OSD_FLAG_READ | opt->read_from_replica;
1411 	osd_req->r_snapid = obj_request->img_request->snap_id;
1412 }
1413 
1414 static void rbd_osd_format_write(struct ceph_osd_request *osd_req)
1415 {
1416 	struct rbd_obj_request *obj_request = osd_req->r_priv;
1417 
1418 	osd_req->r_flags = CEPH_OSD_FLAG_WRITE;
1419 	ktime_get_real_ts64(&osd_req->r_mtime);
1420 	osd_req->r_data_offset = obj_request->ex.oe_off;
1421 }
1422 
1423 static struct ceph_osd_request *
1424 __rbd_obj_add_osd_request(struct rbd_obj_request *obj_req,
1425 			  struct ceph_snap_context *snapc, int num_ops)
1426 {
1427 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
1428 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
1429 	struct ceph_osd_request *req;
1430 	const char *name_format = rbd_dev->image_format == 1 ?
1431 				      RBD_V1_DATA_FORMAT : RBD_V2_DATA_FORMAT;
1432 	int ret;
1433 
1434 	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, false, GFP_NOIO);
1435 	if (!req)
1436 		return ERR_PTR(-ENOMEM);
1437 
1438 	list_add_tail(&req->r_private_item, &obj_req->osd_reqs);
1439 	req->r_callback = rbd_osd_req_callback;
1440 	req->r_priv = obj_req;
1441 
1442 	/*
1443 	 * Data objects may be stored in a separate pool, but always in
1444 	 * the same namespace in that pool as the header in its pool.
1445 	 */
1446 	ceph_oloc_copy(&req->r_base_oloc, &rbd_dev->header_oloc);
1447 	req->r_base_oloc.pool = rbd_dev->layout.pool_id;
1448 
1449 	ret = ceph_oid_aprintf(&req->r_base_oid, GFP_NOIO, name_format,
1450 			       rbd_dev->header.object_prefix,
1451 			       obj_req->ex.oe_objno);
1452 	if (ret)
1453 		return ERR_PTR(ret);
1454 
1455 	return req;
1456 }
1457 
1458 static struct ceph_osd_request *
1459 rbd_obj_add_osd_request(struct rbd_obj_request *obj_req, int num_ops)
1460 {
1461 	rbd_assert(obj_req->img_request->snapc);
1462 	return __rbd_obj_add_osd_request(obj_req, obj_req->img_request->snapc,
1463 					 num_ops);
1464 }
1465 
1466 static struct rbd_obj_request *rbd_obj_request_create(void)
1467 {
1468 	struct rbd_obj_request *obj_request;
1469 
1470 	obj_request = kmem_cache_zalloc(rbd_obj_request_cache, GFP_NOIO);
1471 	if (!obj_request)
1472 		return NULL;
1473 
1474 	ceph_object_extent_init(&obj_request->ex);
1475 	INIT_LIST_HEAD(&obj_request->osd_reqs);
1476 	mutex_init(&obj_request->state_mutex);
1477 	kref_init(&obj_request->kref);
1478 
1479 	dout("%s %p\n", __func__, obj_request);
1480 	return obj_request;
1481 }
1482 
1483 static void rbd_obj_request_destroy(struct kref *kref)
1484 {
1485 	struct rbd_obj_request *obj_request;
1486 	struct ceph_osd_request *osd_req;
1487 	u32 i;
1488 
1489 	obj_request = container_of(kref, struct rbd_obj_request, kref);
1490 
1491 	dout("%s: obj %p\n", __func__, obj_request);
1492 
1493 	while (!list_empty(&obj_request->osd_reqs)) {
1494 		osd_req = list_first_entry(&obj_request->osd_reqs,
1495 				    struct ceph_osd_request, r_private_item);
1496 		list_del_init(&osd_req->r_private_item);
1497 		ceph_osdc_put_request(osd_req);
1498 	}
1499 
1500 	switch (obj_request->img_request->data_type) {
1501 	case OBJ_REQUEST_NODATA:
1502 	case OBJ_REQUEST_BIO:
1503 	case OBJ_REQUEST_BVECS:
1504 		break;		/* Nothing to do */
1505 	case OBJ_REQUEST_OWN_BVECS:
1506 		kfree(obj_request->bvec_pos.bvecs);
1507 		break;
1508 	default:
1509 		BUG();
1510 	}
1511 
1512 	kfree(obj_request->img_extents);
1513 	if (obj_request->copyup_bvecs) {
1514 		for (i = 0; i < obj_request->copyup_bvec_count; i++) {
1515 			if (obj_request->copyup_bvecs[i].bv_page)
1516 				__free_page(obj_request->copyup_bvecs[i].bv_page);
1517 		}
1518 		kfree(obj_request->copyup_bvecs);
1519 	}
1520 
1521 	kmem_cache_free(rbd_obj_request_cache, obj_request);
1522 }
1523 
1524 /* It's OK to call this for a device with no parent */
1525 
1526 static void rbd_spec_put(struct rbd_spec *spec);
1527 static void rbd_dev_unparent(struct rbd_device *rbd_dev)
1528 {
1529 	rbd_dev_remove_parent(rbd_dev);
1530 	rbd_spec_put(rbd_dev->parent_spec);
1531 	rbd_dev->parent_spec = NULL;
1532 	rbd_dev->parent_overlap = 0;
1533 }
1534 
1535 /*
1536  * Parent image reference counting is used to determine when an
1537  * image's parent fields can be safely torn down--after there are no
1538  * more in-flight requests to the parent image.  When the last
1539  * reference is dropped, cleaning them up is safe.
1540  */
1541 static void rbd_dev_parent_put(struct rbd_device *rbd_dev)
1542 {
1543 	int counter;
1544 
1545 	if (!rbd_dev->parent_spec)
1546 		return;
1547 
1548 	counter = atomic_dec_return_safe(&rbd_dev->parent_ref);
1549 	if (counter > 0)
1550 		return;
1551 
1552 	/* Last reference; clean up parent data structures */
1553 
1554 	if (!counter)
1555 		rbd_dev_unparent(rbd_dev);
1556 	else
1557 		rbd_warn(rbd_dev, "parent reference underflow");
1558 }
1559 
1560 /*
1561  * If an image has a non-zero parent overlap, get a reference to its
1562  * parent.
1563  *
1564  * Returns true if the rbd device has a parent with a non-zero
1565  * overlap and a reference for it was successfully taken, or
1566  * false otherwise.
1567  */
1568 static bool rbd_dev_parent_get(struct rbd_device *rbd_dev)
1569 {
1570 	int counter = 0;
1571 
1572 	if (!rbd_dev->parent_spec)
1573 		return false;
1574 
1575 	if (rbd_dev->parent_overlap)
1576 		counter = atomic_inc_return_safe(&rbd_dev->parent_ref);
1577 
1578 	if (counter < 0)
1579 		rbd_warn(rbd_dev, "parent reference overflow");
1580 
1581 	return counter > 0;
1582 }
1583 
1584 static void rbd_img_request_init(struct rbd_img_request *img_request,
1585 				 struct rbd_device *rbd_dev,
1586 				 enum obj_operation_type op_type)
1587 {
1588 	memset(img_request, 0, sizeof(*img_request));
1589 
1590 	img_request->rbd_dev = rbd_dev;
1591 	img_request->op_type = op_type;
1592 
1593 	INIT_LIST_HEAD(&img_request->lock_item);
1594 	INIT_LIST_HEAD(&img_request->object_extents);
1595 	mutex_init(&img_request->state_mutex);
1596 }
1597 
1598 /*
1599  * Only snap_id is captured here, for reads.  For writes, snapshot
1600  * context is captured in rbd_img_object_requests() after exclusive
1601  * lock is ensured to be held.
1602  */
1603 static void rbd_img_capture_header(struct rbd_img_request *img_req)
1604 {
1605 	struct rbd_device *rbd_dev = img_req->rbd_dev;
1606 
1607 	lockdep_assert_held(&rbd_dev->header_rwsem);
1608 
1609 	if (!rbd_img_is_write(img_req))
1610 		img_req->snap_id = rbd_dev->spec->snap_id;
1611 
1612 	if (rbd_dev_parent_get(rbd_dev))
1613 		img_request_layered_set(img_req);
1614 }
1615 
1616 static void rbd_img_request_destroy(struct rbd_img_request *img_request)
1617 {
1618 	struct rbd_obj_request *obj_request;
1619 	struct rbd_obj_request *next_obj_request;
1620 
1621 	dout("%s: img %p\n", __func__, img_request);
1622 
1623 	WARN_ON(!list_empty(&img_request->lock_item));
1624 	for_each_obj_request_safe(img_request, obj_request, next_obj_request)
1625 		rbd_img_obj_request_del(img_request, obj_request);
1626 
1627 	if (img_request_layered_test(img_request))
1628 		rbd_dev_parent_put(img_request->rbd_dev);
1629 
1630 	if (rbd_img_is_write(img_request))
1631 		ceph_put_snap_context(img_request->snapc);
1632 
1633 	if (test_bit(IMG_REQ_CHILD, &img_request->flags))
1634 		kmem_cache_free(rbd_img_request_cache, img_request);
1635 }
1636 
1637 #define BITS_PER_OBJ	2
1638 #define OBJS_PER_BYTE	(BITS_PER_BYTE / BITS_PER_OBJ)
1639 #define OBJ_MASK	((1 << BITS_PER_OBJ) - 1)
1640 
1641 static void __rbd_object_map_index(struct rbd_device *rbd_dev, u64 objno,
1642 				   u64 *index, u8 *shift)
1643 {
1644 	u32 off;
1645 
1646 	rbd_assert(objno < rbd_dev->object_map_size);
1647 	*index = div_u64_rem(objno, OBJS_PER_BYTE, &off);
1648 	*shift = (OBJS_PER_BYTE - off - 1) * BITS_PER_OBJ;
1649 }
1650 
1651 static u8 __rbd_object_map_get(struct rbd_device *rbd_dev, u64 objno)
1652 {
1653 	u64 index;
1654 	u8 shift;
1655 
1656 	lockdep_assert_held(&rbd_dev->object_map_lock);
1657 	__rbd_object_map_index(rbd_dev, objno, &index, &shift);
1658 	return (rbd_dev->object_map[index] >> shift) & OBJ_MASK;
1659 }
1660 
1661 static void __rbd_object_map_set(struct rbd_device *rbd_dev, u64 objno, u8 val)
1662 {
1663 	u64 index;
1664 	u8 shift;
1665 	u8 *p;
1666 
1667 	lockdep_assert_held(&rbd_dev->object_map_lock);
1668 	rbd_assert(!(val & ~OBJ_MASK));
1669 
1670 	__rbd_object_map_index(rbd_dev, objno, &index, &shift);
1671 	p = &rbd_dev->object_map[index];
1672 	*p = (*p & ~(OBJ_MASK << shift)) | (val << shift);
1673 }
1674 
1675 static u8 rbd_object_map_get(struct rbd_device *rbd_dev, u64 objno)
1676 {
1677 	u8 state;
1678 
1679 	spin_lock(&rbd_dev->object_map_lock);
1680 	state = __rbd_object_map_get(rbd_dev, objno);
1681 	spin_unlock(&rbd_dev->object_map_lock);
1682 	return state;
1683 }
1684 
1685 static bool use_object_map(struct rbd_device *rbd_dev)
1686 {
1687 	/*
1688 	 * An image mapped read-only can't use the object map -- it isn't
1689 	 * loaded because the header lock isn't acquired.  Someone else can
1690 	 * write to the image and update the object map behind our back.
1691 	 *
1692 	 * A snapshot can't be written to, so using the object map is always
1693 	 * safe.
1694 	 */
1695 	if (!rbd_is_snap(rbd_dev) && rbd_is_ro(rbd_dev))
1696 		return false;
1697 
1698 	return ((rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP) &&
1699 		!(rbd_dev->object_map_flags & RBD_FLAG_OBJECT_MAP_INVALID));
1700 }
1701 
1702 static bool rbd_object_map_may_exist(struct rbd_device *rbd_dev, u64 objno)
1703 {
1704 	u8 state;
1705 
1706 	/* fall back to default logic if object map is disabled or invalid */
1707 	if (!use_object_map(rbd_dev))
1708 		return true;
1709 
1710 	state = rbd_object_map_get(rbd_dev, objno);
1711 	return state != OBJECT_NONEXISTENT;
1712 }
1713 
1714 static void rbd_object_map_name(struct rbd_device *rbd_dev, u64 snap_id,
1715 				struct ceph_object_id *oid)
1716 {
1717 	if (snap_id == CEPH_NOSNAP)
1718 		ceph_oid_printf(oid, "%s%s", RBD_OBJECT_MAP_PREFIX,
1719 				rbd_dev->spec->image_id);
1720 	else
1721 		ceph_oid_printf(oid, "%s%s.%016llx", RBD_OBJECT_MAP_PREFIX,
1722 				rbd_dev->spec->image_id, snap_id);
1723 }
1724 
1725 static int rbd_object_map_lock(struct rbd_device *rbd_dev)
1726 {
1727 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
1728 	CEPH_DEFINE_OID_ONSTACK(oid);
1729 	u8 lock_type;
1730 	char *lock_tag;
1731 	struct ceph_locker *lockers;
1732 	u32 num_lockers;
1733 	bool broke_lock = false;
1734 	int ret;
1735 
1736 	rbd_object_map_name(rbd_dev, CEPH_NOSNAP, &oid);
1737 
1738 again:
1739 	ret = ceph_cls_lock(osdc, &oid, &rbd_dev->header_oloc, RBD_LOCK_NAME,
1740 			    CEPH_CLS_LOCK_EXCLUSIVE, "", "", "", 0);
1741 	if (ret != -EBUSY || broke_lock) {
1742 		if (ret == -EEXIST)
1743 			ret = 0; /* already locked by myself */
1744 		if (ret)
1745 			rbd_warn(rbd_dev, "failed to lock object map: %d", ret);
1746 		return ret;
1747 	}
1748 
1749 	ret = ceph_cls_lock_info(osdc, &oid, &rbd_dev->header_oloc,
1750 				 RBD_LOCK_NAME, &lock_type, &lock_tag,
1751 				 &lockers, &num_lockers);
1752 	if (ret) {
1753 		if (ret == -ENOENT)
1754 			goto again;
1755 
1756 		rbd_warn(rbd_dev, "failed to get object map lockers: %d", ret);
1757 		return ret;
1758 	}
1759 
1760 	kfree(lock_tag);
1761 	if (num_lockers == 0)
1762 		goto again;
1763 
1764 	rbd_warn(rbd_dev, "breaking object map lock owned by %s%llu",
1765 		 ENTITY_NAME(lockers[0].id.name));
1766 
1767 	ret = ceph_cls_break_lock(osdc, &oid, &rbd_dev->header_oloc,
1768 				  RBD_LOCK_NAME, lockers[0].id.cookie,
1769 				  &lockers[0].id.name);
1770 	ceph_free_lockers(lockers, num_lockers);
1771 	if (ret) {
1772 		if (ret == -ENOENT)
1773 			goto again;
1774 
1775 		rbd_warn(rbd_dev, "failed to break object map lock: %d", ret);
1776 		return ret;
1777 	}
1778 
1779 	broke_lock = true;
1780 	goto again;
1781 }
1782 
1783 static void rbd_object_map_unlock(struct rbd_device *rbd_dev)
1784 {
1785 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
1786 	CEPH_DEFINE_OID_ONSTACK(oid);
1787 	int ret;
1788 
1789 	rbd_object_map_name(rbd_dev, CEPH_NOSNAP, &oid);
1790 
1791 	ret = ceph_cls_unlock(osdc, &oid, &rbd_dev->header_oloc, RBD_LOCK_NAME,
1792 			      "");
1793 	if (ret && ret != -ENOENT)
1794 		rbd_warn(rbd_dev, "failed to unlock object map: %d", ret);
1795 }
1796 
1797 static int decode_object_map_header(void **p, void *end, u64 *object_map_size)
1798 {
1799 	u8 struct_v;
1800 	u32 struct_len;
1801 	u32 header_len;
1802 	void *header_end;
1803 	int ret;
1804 
1805 	ceph_decode_32_safe(p, end, header_len, e_inval);
1806 	header_end = *p + header_len;
1807 
1808 	ret = ceph_start_decoding(p, end, 1, "BitVector header", &struct_v,
1809 				  &struct_len);
1810 	if (ret)
1811 		return ret;
1812 
1813 	ceph_decode_64_safe(p, end, *object_map_size, e_inval);
1814 
1815 	*p = header_end;
1816 	return 0;
1817 
1818 e_inval:
1819 	return -EINVAL;
1820 }
1821 
1822 static int __rbd_object_map_load(struct rbd_device *rbd_dev)
1823 {
1824 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
1825 	CEPH_DEFINE_OID_ONSTACK(oid);
1826 	struct page **pages;
1827 	void *p, *end;
1828 	size_t reply_len;
1829 	u64 num_objects;
1830 	u64 object_map_bytes;
1831 	u64 object_map_size;
1832 	int num_pages;
1833 	int ret;
1834 
1835 	rbd_assert(!rbd_dev->object_map && !rbd_dev->object_map_size);
1836 
1837 	num_objects = ceph_get_num_objects(&rbd_dev->layout,
1838 					   rbd_dev->mapping.size);
1839 	object_map_bytes = DIV_ROUND_UP_ULL(num_objects * BITS_PER_OBJ,
1840 					    BITS_PER_BYTE);
1841 	num_pages = calc_pages_for(0, object_map_bytes) + 1;
1842 	pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
1843 	if (IS_ERR(pages))
1844 		return PTR_ERR(pages);
1845 
1846 	reply_len = num_pages * PAGE_SIZE;
1847 	rbd_object_map_name(rbd_dev, rbd_dev->spec->snap_id, &oid);
1848 	ret = ceph_osdc_call(osdc, &oid, &rbd_dev->header_oloc,
1849 			     "rbd", "object_map_load", CEPH_OSD_FLAG_READ,
1850 			     NULL, 0, pages, &reply_len);
1851 	if (ret)
1852 		goto out;
1853 
1854 	p = page_address(pages[0]);
1855 	end = p + min(reply_len, (size_t)PAGE_SIZE);
1856 	ret = decode_object_map_header(&p, end, &object_map_size);
1857 	if (ret)
1858 		goto out;
1859 
1860 	if (object_map_size != num_objects) {
1861 		rbd_warn(rbd_dev, "object map size mismatch: %llu vs %llu",
1862 			 object_map_size, num_objects);
1863 		ret = -EINVAL;
1864 		goto out;
1865 	}
1866 
1867 	if (offset_in_page(p) + object_map_bytes > reply_len) {
1868 		ret = -EINVAL;
1869 		goto out;
1870 	}
1871 
1872 	rbd_dev->object_map = kvmalloc(object_map_bytes, GFP_KERNEL);
1873 	if (!rbd_dev->object_map) {
1874 		ret = -ENOMEM;
1875 		goto out;
1876 	}
1877 
1878 	rbd_dev->object_map_size = object_map_size;
1879 	ceph_copy_from_page_vector(pages, rbd_dev->object_map,
1880 				   offset_in_page(p), object_map_bytes);
1881 
1882 out:
1883 	ceph_release_page_vector(pages, num_pages);
1884 	return ret;
1885 }
1886 
1887 static void rbd_object_map_free(struct rbd_device *rbd_dev)
1888 {
1889 	kvfree(rbd_dev->object_map);
1890 	rbd_dev->object_map = NULL;
1891 	rbd_dev->object_map_size = 0;
1892 }
1893 
1894 static int rbd_object_map_load(struct rbd_device *rbd_dev)
1895 {
1896 	int ret;
1897 
1898 	ret = __rbd_object_map_load(rbd_dev);
1899 	if (ret)
1900 		return ret;
1901 
1902 	ret = rbd_dev_v2_get_flags(rbd_dev);
1903 	if (ret) {
1904 		rbd_object_map_free(rbd_dev);
1905 		return ret;
1906 	}
1907 
1908 	if (rbd_dev->object_map_flags & RBD_FLAG_OBJECT_MAP_INVALID)
1909 		rbd_warn(rbd_dev, "object map is invalid");
1910 
1911 	return 0;
1912 }
1913 
1914 static int rbd_object_map_open(struct rbd_device *rbd_dev)
1915 {
1916 	int ret;
1917 
1918 	ret = rbd_object_map_lock(rbd_dev);
1919 	if (ret)
1920 		return ret;
1921 
1922 	ret = rbd_object_map_load(rbd_dev);
1923 	if (ret) {
1924 		rbd_object_map_unlock(rbd_dev);
1925 		return ret;
1926 	}
1927 
1928 	return 0;
1929 }
1930 
1931 static void rbd_object_map_close(struct rbd_device *rbd_dev)
1932 {
1933 	rbd_object_map_free(rbd_dev);
1934 	rbd_object_map_unlock(rbd_dev);
1935 }
1936 
1937 /*
1938  * This function needs snap_id (or more precisely just something to
1939  * distinguish between HEAD and snapshot object maps), new_state and
1940  * current_state that were passed to rbd_object_map_update().
1941  *
1942  * To avoid allocating and stashing a context we piggyback on the OSD
1943  * request.  A HEAD update has two ops (assert_locked).  For new_state
1944  * and current_state we decode our own object_map_update op, encoded in
1945  * rbd_cls_object_map_update().
1946  */
1947 static int rbd_object_map_update_finish(struct rbd_obj_request *obj_req,
1948 					struct ceph_osd_request *osd_req)
1949 {
1950 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
1951 	struct ceph_osd_data *osd_data;
1952 	u64 objno;
1953 	u8 state, new_state, current_state;
1954 	bool has_current_state;
1955 	void *p;
1956 
1957 	if (osd_req->r_result)
1958 		return osd_req->r_result;
1959 
1960 	/*
1961 	 * Nothing to do for a snapshot object map.
1962 	 */
1963 	if (osd_req->r_num_ops == 1)
1964 		return 0;
1965 
1966 	/*
1967 	 * Update in-memory HEAD object map.
1968 	 */
1969 	rbd_assert(osd_req->r_num_ops == 2);
1970 	osd_data = osd_req_op_data(osd_req, 1, cls, request_data);
1971 	rbd_assert(osd_data->type == CEPH_OSD_DATA_TYPE_PAGES);
1972 
1973 	p = page_address(osd_data->pages[0]);
1974 	objno = ceph_decode_64(&p);
1975 	rbd_assert(objno == obj_req->ex.oe_objno);
1976 	rbd_assert(ceph_decode_64(&p) == objno + 1);
1977 	new_state = ceph_decode_8(&p);
1978 	has_current_state = ceph_decode_8(&p);
1979 	if (has_current_state)
1980 		current_state = ceph_decode_8(&p);
1981 
1982 	spin_lock(&rbd_dev->object_map_lock);
1983 	state = __rbd_object_map_get(rbd_dev, objno);
1984 	if (!has_current_state || current_state == state ||
1985 	    (current_state == OBJECT_EXISTS && state == OBJECT_EXISTS_CLEAN))
1986 		__rbd_object_map_set(rbd_dev, objno, new_state);
1987 	spin_unlock(&rbd_dev->object_map_lock);
1988 
1989 	return 0;
1990 }
1991 
1992 static void rbd_object_map_callback(struct ceph_osd_request *osd_req)
1993 {
1994 	struct rbd_obj_request *obj_req = osd_req->r_priv;
1995 	int result;
1996 
1997 	dout("%s osd_req %p result %d for obj_req %p\n", __func__, osd_req,
1998 	     osd_req->r_result, obj_req);
1999 
2000 	result = rbd_object_map_update_finish(obj_req, osd_req);
2001 	rbd_obj_handle_request(obj_req, result);
2002 }
2003 
2004 static bool update_needed(struct rbd_device *rbd_dev, u64 objno, u8 new_state)
2005 {
2006 	u8 state = rbd_object_map_get(rbd_dev, objno);
2007 
2008 	if (state == new_state ||
2009 	    (new_state == OBJECT_PENDING && state == OBJECT_NONEXISTENT) ||
2010 	    (new_state == OBJECT_NONEXISTENT && state != OBJECT_PENDING))
2011 		return false;
2012 
2013 	return true;
2014 }
2015 
2016 static int rbd_cls_object_map_update(struct ceph_osd_request *req,
2017 				     int which, u64 objno, u8 new_state,
2018 				     const u8 *current_state)
2019 {
2020 	struct page **pages;
2021 	void *p, *start;
2022 	int ret;
2023 
2024 	ret = osd_req_op_cls_init(req, which, "rbd", "object_map_update");
2025 	if (ret)
2026 		return ret;
2027 
2028 	pages = ceph_alloc_page_vector(1, GFP_NOIO);
2029 	if (IS_ERR(pages))
2030 		return PTR_ERR(pages);
2031 
2032 	p = start = page_address(pages[0]);
2033 	ceph_encode_64(&p, objno);
2034 	ceph_encode_64(&p, objno + 1);
2035 	ceph_encode_8(&p, new_state);
2036 	if (current_state) {
2037 		ceph_encode_8(&p, 1);
2038 		ceph_encode_8(&p, *current_state);
2039 	} else {
2040 		ceph_encode_8(&p, 0);
2041 	}
2042 
2043 	osd_req_op_cls_request_data_pages(req, which, pages, p - start, 0,
2044 					  false, true);
2045 	return 0;
2046 }
2047 
2048 /*
2049  * Return:
2050  *   0 - object map update sent
2051  *   1 - object map update isn't needed
2052  *  <0 - error
2053  */
2054 static int rbd_object_map_update(struct rbd_obj_request *obj_req, u64 snap_id,
2055 				 u8 new_state, const u8 *current_state)
2056 {
2057 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2058 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
2059 	struct ceph_osd_request *req;
2060 	int num_ops = 1;
2061 	int which = 0;
2062 	int ret;
2063 
2064 	if (snap_id == CEPH_NOSNAP) {
2065 		if (!update_needed(rbd_dev, obj_req->ex.oe_objno, new_state))
2066 			return 1;
2067 
2068 		num_ops++; /* assert_locked */
2069 	}
2070 
2071 	req = ceph_osdc_alloc_request(osdc, NULL, num_ops, false, GFP_NOIO);
2072 	if (!req)
2073 		return -ENOMEM;
2074 
2075 	list_add_tail(&req->r_private_item, &obj_req->osd_reqs);
2076 	req->r_callback = rbd_object_map_callback;
2077 	req->r_priv = obj_req;
2078 
2079 	rbd_object_map_name(rbd_dev, snap_id, &req->r_base_oid);
2080 	ceph_oloc_copy(&req->r_base_oloc, &rbd_dev->header_oloc);
2081 	req->r_flags = CEPH_OSD_FLAG_WRITE;
2082 	ktime_get_real_ts64(&req->r_mtime);
2083 
2084 	if (snap_id == CEPH_NOSNAP) {
2085 		/*
2086 		 * Protect against possible race conditions during lock
2087 		 * ownership transitions.
2088 		 */
2089 		ret = ceph_cls_assert_locked(req, which++, RBD_LOCK_NAME,
2090 					     CEPH_CLS_LOCK_EXCLUSIVE, "", "");
2091 		if (ret)
2092 			return ret;
2093 	}
2094 
2095 	ret = rbd_cls_object_map_update(req, which, obj_req->ex.oe_objno,
2096 					new_state, current_state);
2097 	if (ret)
2098 		return ret;
2099 
2100 	ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
2101 	if (ret)
2102 		return ret;
2103 
2104 	ceph_osdc_start_request(osdc, req);
2105 	return 0;
2106 }
2107 
2108 static void prune_extents(struct ceph_file_extent *img_extents,
2109 			  u32 *num_img_extents, u64 overlap)
2110 {
2111 	u32 cnt = *num_img_extents;
2112 
2113 	/* drop extents completely beyond the overlap */
2114 	while (cnt && img_extents[cnt - 1].fe_off >= overlap)
2115 		cnt--;
2116 
2117 	if (cnt) {
2118 		struct ceph_file_extent *ex = &img_extents[cnt - 1];
2119 
2120 		/* trim final overlapping extent */
2121 		if (ex->fe_off + ex->fe_len > overlap)
2122 			ex->fe_len = overlap - ex->fe_off;
2123 	}
2124 
2125 	*num_img_extents = cnt;
2126 }
2127 
2128 /*
2129  * Determine the byte range(s) covered by either just the object extent
2130  * or the entire object in the parent image.
2131  */
2132 static int rbd_obj_calc_img_extents(struct rbd_obj_request *obj_req,
2133 				    bool entire)
2134 {
2135 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2136 	int ret;
2137 
2138 	if (!rbd_dev->parent_overlap)
2139 		return 0;
2140 
2141 	ret = ceph_extent_to_file(&rbd_dev->layout, obj_req->ex.oe_objno,
2142 				  entire ? 0 : obj_req->ex.oe_off,
2143 				  entire ? rbd_dev->layout.object_size :
2144 							obj_req->ex.oe_len,
2145 				  &obj_req->img_extents,
2146 				  &obj_req->num_img_extents);
2147 	if (ret)
2148 		return ret;
2149 
2150 	prune_extents(obj_req->img_extents, &obj_req->num_img_extents,
2151 		      rbd_dev->parent_overlap);
2152 	return 0;
2153 }
2154 
2155 static void rbd_osd_setup_data(struct ceph_osd_request *osd_req, int which)
2156 {
2157 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2158 
2159 	switch (obj_req->img_request->data_type) {
2160 	case OBJ_REQUEST_BIO:
2161 		osd_req_op_extent_osd_data_bio(osd_req, which,
2162 					       &obj_req->bio_pos,
2163 					       obj_req->ex.oe_len);
2164 		break;
2165 	case OBJ_REQUEST_BVECS:
2166 	case OBJ_REQUEST_OWN_BVECS:
2167 		rbd_assert(obj_req->bvec_pos.iter.bi_size ==
2168 							obj_req->ex.oe_len);
2169 		rbd_assert(obj_req->bvec_idx == obj_req->bvec_count);
2170 		osd_req_op_extent_osd_data_bvec_pos(osd_req, which,
2171 						    &obj_req->bvec_pos);
2172 		break;
2173 	default:
2174 		BUG();
2175 	}
2176 }
2177 
2178 static int rbd_osd_setup_stat(struct ceph_osd_request *osd_req, int which)
2179 {
2180 	struct page **pages;
2181 
2182 	/*
2183 	 * The response data for a STAT call consists of:
2184 	 *     le64 length;
2185 	 *     struct {
2186 	 *         le32 tv_sec;
2187 	 *         le32 tv_nsec;
2188 	 *     } mtime;
2189 	 */
2190 	pages = ceph_alloc_page_vector(1, GFP_NOIO);
2191 	if (IS_ERR(pages))
2192 		return PTR_ERR(pages);
2193 
2194 	osd_req_op_init(osd_req, which, CEPH_OSD_OP_STAT, 0);
2195 	osd_req_op_raw_data_in_pages(osd_req, which, pages,
2196 				     8 + sizeof(struct ceph_timespec),
2197 				     0, false, true);
2198 	return 0;
2199 }
2200 
2201 static int rbd_osd_setup_copyup(struct ceph_osd_request *osd_req, int which,
2202 				u32 bytes)
2203 {
2204 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2205 	int ret;
2206 
2207 	ret = osd_req_op_cls_init(osd_req, which, "rbd", "copyup");
2208 	if (ret)
2209 		return ret;
2210 
2211 	osd_req_op_cls_request_data_bvecs(osd_req, which, obj_req->copyup_bvecs,
2212 					  obj_req->copyup_bvec_count, bytes);
2213 	return 0;
2214 }
2215 
2216 static int rbd_obj_init_read(struct rbd_obj_request *obj_req)
2217 {
2218 	obj_req->read_state = RBD_OBJ_READ_START;
2219 	return 0;
2220 }
2221 
2222 static void __rbd_osd_setup_write_ops(struct ceph_osd_request *osd_req,
2223 				      int which)
2224 {
2225 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2226 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2227 	u16 opcode;
2228 
2229 	if (!use_object_map(rbd_dev) ||
2230 	    !(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST)) {
2231 		osd_req_op_alloc_hint_init(osd_req, which++,
2232 					   rbd_dev->layout.object_size,
2233 					   rbd_dev->layout.object_size,
2234 					   rbd_dev->opts->alloc_hint_flags);
2235 	}
2236 
2237 	if (rbd_obj_is_entire(obj_req))
2238 		opcode = CEPH_OSD_OP_WRITEFULL;
2239 	else
2240 		opcode = CEPH_OSD_OP_WRITE;
2241 
2242 	osd_req_op_extent_init(osd_req, which, opcode,
2243 			       obj_req->ex.oe_off, obj_req->ex.oe_len, 0, 0);
2244 	rbd_osd_setup_data(osd_req, which);
2245 }
2246 
2247 static int rbd_obj_init_write(struct rbd_obj_request *obj_req)
2248 {
2249 	int ret;
2250 
2251 	/* reverse map the entire object onto the parent */
2252 	ret = rbd_obj_calc_img_extents(obj_req, true);
2253 	if (ret)
2254 		return ret;
2255 
2256 	obj_req->write_state = RBD_OBJ_WRITE_START;
2257 	return 0;
2258 }
2259 
2260 static u16 truncate_or_zero_opcode(struct rbd_obj_request *obj_req)
2261 {
2262 	return rbd_obj_is_tail(obj_req) ? CEPH_OSD_OP_TRUNCATE :
2263 					  CEPH_OSD_OP_ZERO;
2264 }
2265 
2266 static void __rbd_osd_setup_discard_ops(struct ceph_osd_request *osd_req,
2267 					int which)
2268 {
2269 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2270 
2271 	if (rbd_obj_is_entire(obj_req) && !obj_req->num_img_extents) {
2272 		rbd_assert(obj_req->flags & RBD_OBJ_FLAG_DELETION);
2273 		osd_req_op_init(osd_req, which, CEPH_OSD_OP_DELETE, 0);
2274 	} else {
2275 		osd_req_op_extent_init(osd_req, which,
2276 				       truncate_or_zero_opcode(obj_req),
2277 				       obj_req->ex.oe_off, obj_req->ex.oe_len,
2278 				       0, 0);
2279 	}
2280 }
2281 
2282 static int rbd_obj_init_discard(struct rbd_obj_request *obj_req)
2283 {
2284 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2285 	u64 off, next_off;
2286 	int ret;
2287 
2288 	/*
2289 	 * Align the range to alloc_size boundary and punt on discards
2290 	 * that are too small to free up any space.
2291 	 *
2292 	 * alloc_size == object_size && is_tail() is a special case for
2293 	 * filestore with filestore_punch_hole = false, needed to allow
2294 	 * truncate (in addition to delete).
2295 	 */
2296 	if (rbd_dev->opts->alloc_size != rbd_dev->layout.object_size ||
2297 	    !rbd_obj_is_tail(obj_req)) {
2298 		off = round_up(obj_req->ex.oe_off, rbd_dev->opts->alloc_size);
2299 		next_off = round_down(obj_req->ex.oe_off + obj_req->ex.oe_len,
2300 				      rbd_dev->opts->alloc_size);
2301 		if (off >= next_off)
2302 			return 1;
2303 
2304 		dout("%s %p %llu~%llu -> %llu~%llu\n", __func__,
2305 		     obj_req, obj_req->ex.oe_off, obj_req->ex.oe_len,
2306 		     off, next_off - off);
2307 		obj_req->ex.oe_off = off;
2308 		obj_req->ex.oe_len = next_off - off;
2309 	}
2310 
2311 	/* reverse map the entire object onto the parent */
2312 	ret = rbd_obj_calc_img_extents(obj_req, true);
2313 	if (ret)
2314 		return ret;
2315 
2316 	obj_req->flags |= RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT;
2317 	if (rbd_obj_is_entire(obj_req) && !obj_req->num_img_extents)
2318 		obj_req->flags |= RBD_OBJ_FLAG_DELETION;
2319 
2320 	obj_req->write_state = RBD_OBJ_WRITE_START;
2321 	return 0;
2322 }
2323 
2324 static void __rbd_osd_setup_zeroout_ops(struct ceph_osd_request *osd_req,
2325 					int which)
2326 {
2327 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2328 	u16 opcode;
2329 
2330 	if (rbd_obj_is_entire(obj_req)) {
2331 		if (obj_req->num_img_extents) {
2332 			if (!(obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED))
2333 				osd_req_op_init(osd_req, which++,
2334 						CEPH_OSD_OP_CREATE, 0);
2335 			opcode = CEPH_OSD_OP_TRUNCATE;
2336 		} else {
2337 			rbd_assert(obj_req->flags & RBD_OBJ_FLAG_DELETION);
2338 			osd_req_op_init(osd_req, which++,
2339 					CEPH_OSD_OP_DELETE, 0);
2340 			opcode = 0;
2341 		}
2342 	} else {
2343 		opcode = truncate_or_zero_opcode(obj_req);
2344 	}
2345 
2346 	if (opcode)
2347 		osd_req_op_extent_init(osd_req, which, opcode,
2348 				       obj_req->ex.oe_off, obj_req->ex.oe_len,
2349 				       0, 0);
2350 }
2351 
2352 static int rbd_obj_init_zeroout(struct rbd_obj_request *obj_req)
2353 {
2354 	int ret;
2355 
2356 	/* reverse map the entire object onto the parent */
2357 	ret = rbd_obj_calc_img_extents(obj_req, true);
2358 	if (ret)
2359 		return ret;
2360 
2361 	if (!obj_req->num_img_extents) {
2362 		obj_req->flags |= RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT;
2363 		if (rbd_obj_is_entire(obj_req))
2364 			obj_req->flags |= RBD_OBJ_FLAG_DELETION;
2365 	}
2366 
2367 	obj_req->write_state = RBD_OBJ_WRITE_START;
2368 	return 0;
2369 }
2370 
2371 static int count_write_ops(struct rbd_obj_request *obj_req)
2372 {
2373 	struct rbd_img_request *img_req = obj_req->img_request;
2374 
2375 	switch (img_req->op_type) {
2376 	case OBJ_OP_WRITE:
2377 		if (!use_object_map(img_req->rbd_dev) ||
2378 		    !(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST))
2379 			return 2; /* setallochint + write/writefull */
2380 
2381 		return 1; /* write/writefull */
2382 	case OBJ_OP_DISCARD:
2383 		return 1; /* delete/truncate/zero */
2384 	case OBJ_OP_ZEROOUT:
2385 		if (rbd_obj_is_entire(obj_req) && obj_req->num_img_extents &&
2386 		    !(obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED))
2387 			return 2; /* create + truncate */
2388 
2389 		return 1; /* delete/truncate/zero */
2390 	default:
2391 		BUG();
2392 	}
2393 }
2394 
2395 static void rbd_osd_setup_write_ops(struct ceph_osd_request *osd_req,
2396 				    int which)
2397 {
2398 	struct rbd_obj_request *obj_req = osd_req->r_priv;
2399 
2400 	switch (obj_req->img_request->op_type) {
2401 	case OBJ_OP_WRITE:
2402 		__rbd_osd_setup_write_ops(osd_req, which);
2403 		break;
2404 	case OBJ_OP_DISCARD:
2405 		__rbd_osd_setup_discard_ops(osd_req, which);
2406 		break;
2407 	case OBJ_OP_ZEROOUT:
2408 		__rbd_osd_setup_zeroout_ops(osd_req, which);
2409 		break;
2410 	default:
2411 		BUG();
2412 	}
2413 }
2414 
2415 /*
2416  * Prune the list of object requests (adjust offset and/or length, drop
2417  * redundant requests).  Prepare object request state machines and image
2418  * request state machine for execution.
2419  */
2420 static int __rbd_img_fill_request(struct rbd_img_request *img_req)
2421 {
2422 	struct rbd_obj_request *obj_req, *next_obj_req;
2423 	int ret;
2424 
2425 	for_each_obj_request_safe(img_req, obj_req, next_obj_req) {
2426 		switch (img_req->op_type) {
2427 		case OBJ_OP_READ:
2428 			ret = rbd_obj_init_read(obj_req);
2429 			break;
2430 		case OBJ_OP_WRITE:
2431 			ret = rbd_obj_init_write(obj_req);
2432 			break;
2433 		case OBJ_OP_DISCARD:
2434 			ret = rbd_obj_init_discard(obj_req);
2435 			break;
2436 		case OBJ_OP_ZEROOUT:
2437 			ret = rbd_obj_init_zeroout(obj_req);
2438 			break;
2439 		default:
2440 			BUG();
2441 		}
2442 		if (ret < 0)
2443 			return ret;
2444 		if (ret > 0) {
2445 			rbd_img_obj_request_del(img_req, obj_req);
2446 			continue;
2447 		}
2448 	}
2449 
2450 	img_req->state = RBD_IMG_START;
2451 	return 0;
2452 }
2453 
2454 union rbd_img_fill_iter {
2455 	struct ceph_bio_iter	bio_iter;
2456 	struct ceph_bvec_iter	bvec_iter;
2457 };
2458 
2459 struct rbd_img_fill_ctx {
2460 	enum obj_request_type	pos_type;
2461 	union rbd_img_fill_iter	*pos;
2462 	union rbd_img_fill_iter	iter;
2463 	ceph_object_extent_fn_t	set_pos_fn;
2464 	ceph_object_extent_fn_t	count_fn;
2465 	ceph_object_extent_fn_t	copy_fn;
2466 };
2467 
2468 static struct ceph_object_extent *alloc_object_extent(void *arg)
2469 {
2470 	struct rbd_img_request *img_req = arg;
2471 	struct rbd_obj_request *obj_req;
2472 
2473 	obj_req = rbd_obj_request_create();
2474 	if (!obj_req)
2475 		return NULL;
2476 
2477 	rbd_img_obj_request_add(img_req, obj_req);
2478 	return &obj_req->ex;
2479 }
2480 
2481 /*
2482  * While su != os && sc == 1 is technically not fancy (it's the same
2483  * layout as su == os && sc == 1), we can't use the nocopy path for it
2484  * because ->set_pos_fn() should be called only once per object.
2485  * ceph_file_to_extents() invokes action_fn once per stripe unit, so
2486  * treat su != os && sc == 1 as fancy.
2487  */
2488 static bool rbd_layout_is_fancy(struct ceph_file_layout *l)
2489 {
2490 	return l->stripe_unit != l->object_size;
2491 }
2492 
2493 static int rbd_img_fill_request_nocopy(struct rbd_img_request *img_req,
2494 				       struct ceph_file_extent *img_extents,
2495 				       u32 num_img_extents,
2496 				       struct rbd_img_fill_ctx *fctx)
2497 {
2498 	u32 i;
2499 	int ret;
2500 
2501 	img_req->data_type = fctx->pos_type;
2502 
2503 	/*
2504 	 * Create object requests and set each object request's starting
2505 	 * position in the provided bio (list) or bio_vec array.
2506 	 */
2507 	fctx->iter = *fctx->pos;
2508 	for (i = 0; i < num_img_extents; i++) {
2509 		ret = ceph_file_to_extents(&img_req->rbd_dev->layout,
2510 					   img_extents[i].fe_off,
2511 					   img_extents[i].fe_len,
2512 					   &img_req->object_extents,
2513 					   alloc_object_extent, img_req,
2514 					   fctx->set_pos_fn, &fctx->iter);
2515 		if (ret)
2516 			return ret;
2517 	}
2518 
2519 	return __rbd_img_fill_request(img_req);
2520 }
2521 
2522 /*
2523  * Map a list of image extents to a list of object extents, create the
2524  * corresponding object requests (normally each to a different object,
2525  * but not always) and add them to @img_req.  For each object request,
2526  * set up its data descriptor to point to the corresponding chunk(s) of
2527  * @fctx->pos data buffer.
2528  *
2529  * Because ceph_file_to_extents() will merge adjacent object extents
2530  * together, each object request's data descriptor may point to multiple
2531  * different chunks of @fctx->pos data buffer.
2532  *
2533  * @fctx->pos data buffer is assumed to be large enough.
2534  */
2535 static int rbd_img_fill_request(struct rbd_img_request *img_req,
2536 				struct ceph_file_extent *img_extents,
2537 				u32 num_img_extents,
2538 				struct rbd_img_fill_ctx *fctx)
2539 {
2540 	struct rbd_device *rbd_dev = img_req->rbd_dev;
2541 	struct rbd_obj_request *obj_req;
2542 	u32 i;
2543 	int ret;
2544 
2545 	if (fctx->pos_type == OBJ_REQUEST_NODATA ||
2546 	    !rbd_layout_is_fancy(&rbd_dev->layout))
2547 		return rbd_img_fill_request_nocopy(img_req, img_extents,
2548 						   num_img_extents, fctx);
2549 
2550 	img_req->data_type = OBJ_REQUEST_OWN_BVECS;
2551 
2552 	/*
2553 	 * Create object requests and determine ->bvec_count for each object
2554 	 * request.  Note that ->bvec_count sum over all object requests may
2555 	 * be greater than the number of bio_vecs in the provided bio (list)
2556 	 * or bio_vec array because when mapped, those bio_vecs can straddle
2557 	 * stripe unit boundaries.
2558 	 */
2559 	fctx->iter = *fctx->pos;
2560 	for (i = 0; i < num_img_extents; i++) {
2561 		ret = ceph_file_to_extents(&rbd_dev->layout,
2562 					   img_extents[i].fe_off,
2563 					   img_extents[i].fe_len,
2564 					   &img_req->object_extents,
2565 					   alloc_object_extent, img_req,
2566 					   fctx->count_fn, &fctx->iter);
2567 		if (ret)
2568 			return ret;
2569 	}
2570 
2571 	for_each_obj_request(img_req, obj_req) {
2572 		obj_req->bvec_pos.bvecs = kmalloc_array(obj_req->bvec_count,
2573 					      sizeof(*obj_req->bvec_pos.bvecs),
2574 					      GFP_NOIO);
2575 		if (!obj_req->bvec_pos.bvecs)
2576 			return -ENOMEM;
2577 	}
2578 
2579 	/*
2580 	 * Fill in each object request's private bio_vec array, splitting and
2581 	 * rearranging the provided bio_vecs in stripe unit chunks as needed.
2582 	 */
2583 	fctx->iter = *fctx->pos;
2584 	for (i = 0; i < num_img_extents; i++) {
2585 		ret = ceph_iterate_extents(&rbd_dev->layout,
2586 					   img_extents[i].fe_off,
2587 					   img_extents[i].fe_len,
2588 					   &img_req->object_extents,
2589 					   fctx->copy_fn, &fctx->iter);
2590 		if (ret)
2591 			return ret;
2592 	}
2593 
2594 	return __rbd_img_fill_request(img_req);
2595 }
2596 
2597 static int rbd_img_fill_nodata(struct rbd_img_request *img_req,
2598 			       u64 off, u64 len)
2599 {
2600 	struct ceph_file_extent ex = { off, len };
2601 	union rbd_img_fill_iter dummy = {};
2602 	struct rbd_img_fill_ctx fctx = {
2603 		.pos_type = OBJ_REQUEST_NODATA,
2604 		.pos = &dummy,
2605 	};
2606 
2607 	return rbd_img_fill_request(img_req, &ex, 1, &fctx);
2608 }
2609 
2610 static void set_bio_pos(struct ceph_object_extent *ex, u32 bytes, void *arg)
2611 {
2612 	struct rbd_obj_request *obj_req =
2613 	    container_of(ex, struct rbd_obj_request, ex);
2614 	struct ceph_bio_iter *it = arg;
2615 
2616 	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
2617 	obj_req->bio_pos = *it;
2618 	ceph_bio_iter_advance(it, bytes);
2619 }
2620 
2621 static void count_bio_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
2622 {
2623 	struct rbd_obj_request *obj_req =
2624 	    container_of(ex, struct rbd_obj_request, ex);
2625 	struct ceph_bio_iter *it = arg;
2626 
2627 	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
2628 	ceph_bio_iter_advance_step(it, bytes, ({
2629 		obj_req->bvec_count++;
2630 	}));
2631 
2632 }
2633 
2634 static void copy_bio_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
2635 {
2636 	struct rbd_obj_request *obj_req =
2637 	    container_of(ex, struct rbd_obj_request, ex);
2638 	struct ceph_bio_iter *it = arg;
2639 
2640 	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
2641 	ceph_bio_iter_advance_step(it, bytes, ({
2642 		obj_req->bvec_pos.bvecs[obj_req->bvec_idx++] = bv;
2643 		obj_req->bvec_pos.iter.bi_size += bv.bv_len;
2644 	}));
2645 }
2646 
2647 static int __rbd_img_fill_from_bio(struct rbd_img_request *img_req,
2648 				   struct ceph_file_extent *img_extents,
2649 				   u32 num_img_extents,
2650 				   struct ceph_bio_iter *bio_pos)
2651 {
2652 	struct rbd_img_fill_ctx fctx = {
2653 		.pos_type = OBJ_REQUEST_BIO,
2654 		.pos = (union rbd_img_fill_iter *)bio_pos,
2655 		.set_pos_fn = set_bio_pos,
2656 		.count_fn = count_bio_bvecs,
2657 		.copy_fn = copy_bio_bvecs,
2658 	};
2659 
2660 	return rbd_img_fill_request(img_req, img_extents, num_img_extents,
2661 				    &fctx);
2662 }
2663 
2664 static int rbd_img_fill_from_bio(struct rbd_img_request *img_req,
2665 				 u64 off, u64 len, struct bio *bio)
2666 {
2667 	struct ceph_file_extent ex = { off, len };
2668 	struct ceph_bio_iter it = { .bio = bio, .iter = bio->bi_iter };
2669 
2670 	return __rbd_img_fill_from_bio(img_req, &ex, 1, &it);
2671 }
2672 
2673 static void set_bvec_pos(struct ceph_object_extent *ex, u32 bytes, void *arg)
2674 {
2675 	struct rbd_obj_request *obj_req =
2676 	    container_of(ex, struct rbd_obj_request, ex);
2677 	struct ceph_bvec_iter *it = arg;
2678 
2679 	obj_req->bvec_pos = *it;
2680 	ceph_bvec_iter_shorten(&obj_req->bvec_pos, bytes);
2681 	ceph_bvec_iter_advance(it, bytes);
2682 }
2683 
2684 static void count_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
2685 {
2686 	struct rbd_obj_request *obj_req =
2687 	    container_of(ex, struct rbd_obj_request, ex);
2688 	struct ceph_bvec_iter *it = arg;
2689 
2690 	ceph_bvec_iter_advance_step(it, bytes, ({
2691 		obj_req->bvec_count++;
2692 	}));
2693 }
2694 
2695 static void copy_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
2696 {
2697 	struct rbd_obj_request *obj_req =
2698 	    container_of(ex, struct rbd_obj_request, ex);
2699 	struct ceph_bvec_iter *it = arg;
2700 
2701 	ceph_bvec_iter_advance_step(it, bytes, ({
2702 		obj_req->bvec_pos.bvecs[obj_req->bvec_idx++] = bv;
2703 		obj_req->bvec_pos.iter.bi_size += bv.bv_len;
2704 	}));
2705 }
2706 
2707 static int __rbd_img_fill_from_bvecs(struct rbd_img_request *img_req,
2708 				     struct ceph_file_extent *img_extents,
2709 				     u32 num_img_extents,
2710 				     struct ceph_bvec_iter *bvec_pos)
2711 {
2712 	struct rbd_img_fill_ctx fctx = {
2713 		.pos_type = OBJ_REQUEST_BVECS,
2714 		.pos = (union rbd_img_fill_iter *)bvec_pos,
2715 		.set_pos_fn = set_bvec_pos,
2716 		.count_fn = count_bvecs,
2717 		.copy_fn = copy_bvecs,
2718 	};
2719 
2720 	return rbd_img_fill_request(img_req, img_extents, num_img_extents,
2721 				    &fctx);
2722 }
2723 
2724 static int rbd_img_fill_from_bvecs(struct rbd_img_request *img_req,
2725 				   struct ceph_file_extent *img_extents,
2726 				   u32 num_img_extents,
2727 				   struct bio_vec *bvecs)
2728 {
2729 	struct ceph_bvec_iter it = {
2730 		.bvecs = bvecs,
2731 		.iter = { .bi_size = ceph_file_extents_bytes(img_extents,
2732 							     num_img_extents) },
2733 	};
2734 
2735 	return __rbd_img_fill_from_bvecs(img_req, img_extents, num_img_extents,
2736 					 &it);
2737 }
2738 
2739 static void rbd_img_handle_request_work(struct work_struct *work)
2740 {
2741 	struct rbd_img_request *img_req =
2742 	    container_of(work, struct rbd_img_request, work);
2743 
2744 	rbd_img_handle_request(img_req, img_req->work_result);
2745 }
2746 
2747 static void rbd_img_schedule(struct rbd_img_request *img_req, int result)
2748 {
2749 	INIT_WORK(&img_req->work, rbd_img_handle_request_work);
2750 	img_req->work_result = result;
2751 	queue_work(rbd_wq, &img_req->work);
2752 }
2753 
2754 static bool rbd_obj_may_exist(struct rbd_obj_request *obj_req)
2755 {
2756 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2757 
2758 	if (rbd_object_map_may_exist(rbd_dev, obj_req->ex.oe_objno)) {
2759 		obj_req->flags |= RBD_OBJ_FLAG_MAY_EXIST;
2760 		return true;
2761 	}
2762 
2763 	dout("%s %p objno %llu assuming dne\n", __func__, obj_req,
2764 	     obj_req->ex.oe_objno);
2765 	return false;
2766 }
2767 
2768 static int rbd_obj_read_object(struct rbd_obj_request *obj_req)
2769 {
2770 	struct ceph_osd_request *osd_req;
2771 	int ret;
2772 
2773 	osd_req = __rbd_obj_add_osd_request(obj_req, NULL, 1);
2774 	if (IS_ERR(osd_req))
2775 		return PTR_ERR(osd_req);
2776 
2777 	osd_req_op_extent_init(osd_req, 0, CEPH_OSD_OP_READ,
2778 			       obj_req->ex.oe_off, obj_req->ex.oe_len, 0, 0);
2779 	rbd_osd_setup_data(osd_req, 0);
2780 	rbd_osd_format_read(osd_req);
2781 
2782 	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
2783 	if (ret)
2784 		return ret;
2785 
2786 	rbd_osd_submit(osd_req);
2787 	return 0;
2788 }
2789 
2790 static int rbd_obj_read_from_parent(struct rbd_obj_request *obj_req)
2791 {
2792 	struct rbd_img_request *img_req = obj_req->img_request;
2793 	struct rbd_device *parent = img_req->rbd_dev->parent;
2794 	struct rbd_img_request *child_img_req;
2795 	int ret;
2796 
2797 	child_img_req = kmem_cache_alloc(rbd_img_request_cache, GFP_NOIO);
2798 	if (!child_img_req)
2799 		return -ENOMEM;
2800 
2801 	rbd_img_request_init(child_img_req, parent, OBJ_OP_READ);
2802 	__set_bit(IMG_REQ_CHILD, &child_img_req->flags);
2803 	child_img_req->obj_request = obj_req;
2804 
2805 	down_read(&parent->header_rwsem);
2806 	rbd_img_capture_header(child_img_req);
2807 	up_read(&parent->header_rwsem);
2808 
2809 	dout("%s child_img_req %p for obj_req %p\n", __func__, child_img_req,
2810 	     obj_req);
2811 
2812 	if (!rbd_img_is_write(img_req)) {
2813 		switch (img_req->data_type) {
2814 		case OBJ_REQUEST_BIO:
2815 			ret = __rbd_img_fill_from_bio(child_img_req,
2816 						      obj_req->img_extents,
2817 						      obj_req->num_img_extents,
2818 						      &obj_req->bio_pos);
2819 			break;
2820 		case OBJ_REQUEST_BVECS:
2821 		case OBJ_REQUEST_OWN_BVECS:
2822 			ret = __rbd_img_fill_from_bvecs(child_img_req,
2823 						      obj_req->img_extents,
2824 						      obj_req->num_img_extents,
2825 						      &obj_req->bvec_pos);
2826 			break;
2827 		default:
2828 			BUG();
2829 		}
2830 	} else {
2831 		ret = rbd_img_fill_from_bvecs(child_img_req,
2832 					      obj_req->img_extents,
2833 					      obj_req->num_img_extents,
2834 					      obj_req->copyup_bvecs);
2835 	}
2836 	if (ret) {
2837 		rbd_img_request_destroy(child_img_req);
2838 		return ret;
2839 	}
2840 
2841 	/* avoid parent chain recursion */
2842 	rbd_img_schedule(child_img_req, 0);
2843 	return 0;
2844 }
2845 
2846 static bool rbd_obj_advance_read(struct rbd_obj_request *obj_req, int *result)
2847 {
2848 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2849 	int ret;
2850 
2851 again:
2852 	switch (obj_req->read_state) {
2853 	case RBD_OBJ_READ_START:
2854 		rbd_assert(!*result);
2855 
2856 		if (!rbd_obj_may_exist(obj_req)) {
2857 			*result = -ENOENT;
2858 			obj_req->read_state = RBD_OBJ_READ_OBJECT;
2859 			goto again;
2860 		}
2861 
2862 		ret = rbd_obj_read_object(obj_req);
2863 		if (ret) {
2864 			*result = ret;
2865 			return true;
2866 		}
2867 		obj_req->read_state = RBD_OBJ_READ_OBJECT;
2868 		return false;
2869 	case RBD_OBJ_READ_OBJECT:
2870 		if (*result == -ENOENT && rbd_dev->parent_overlap) {
2871 			/* reverse map this object extent onto the parent */
2872 			ret = rbd_obj_calc_img_extents(obj_req, false);
2873 			if (ret) {
2874 				*result = ret;
2875 				return true;
2876 			}
2877 			if (obj_req->num_img_extents) {
2878 				ret = rbd_obj_read_from_parent(obj_req);
2879 				if (ret) {
2880 					*result = ret;
2881 					return true;
2882 				}
2883 				obj_req->read_state = RBD_OBJ_READ_PARENT;
2884 				return false;
2885 			}
2886 		}
2887 
2888 		/*
2889 		 * -ENOENT means a hole in the image -- zero-fill the entire
2890 		 * length of the request.  A short read also implies zero-fill
2891 		 * to the end of the request.
2892 		 */
2893 		if (*result == -ENOENT) {
2894 			rbd_obj_zero_range(obj_req, 0, obj_req->ex.oe_len);
2895 			*result = 0;
2896 		} else if (*result >= 0) {
2897 			if (*result < obj_req->ex.oe_len)
2898 				rbd_obj_zero_range(obj_req, *result,
2899 						obj_req->ex.oe_len - *result);
2900 			else
2901 				rbd_assert(*result == obj_req->ex.oe_len);
2902 			*result = 0;
2903 		}
2904 		return true;
2905 	case RBD_OBJ_READ_PARENT:
2906 		/*
2907 		 * The parent image is read only up to the overlap -- zero-fill
2908 		 * from the overlap to the end of the request.
2909 		 */
2910 		if (!*result) {
2911 			u32 obj_overlap = rbd_obj_img_extents_bytes(obj_req);
2912 
2913 			if (obj_overlap < obj_req->ex.oe_len)
2914 				rbd_obj_zero_range(obj_req, obj_overlap,
2915 					    obj_req->ex.oe_len - obj_overlap);
2916 		}
2917 		return true;
2918 	default:
2919 		BUG();
2920 	}
2921 }
2922 
2923 static bool rbd_obj_write_is_noop(struct rbd_obj_request *obj_req)
2924 {
2925 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2926 
2927 	if (rbd_object_map_may_exist(rbd_dev, obj_req->ex.oe_objno))
2928 		obj_req->flags |= RBD_OBJ_FLAG_MAY_EXIST;
2929 
2930 	if (!(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST) &&
2931 	    (obj_req->flags & RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT)) {
2932 		dout("%s %p noop for nonexistent\n", __func__, obj_req);
2933 		return true;
2934 	}
2935 
2936 	return false;
2937 }
2938 
2939 /*
2940  * Return:
2941  *   0 - object map update sent
2942  *   1 - object map update isn't needed
2943  *  <0 - error
2944  */
2945 static int rbd_obj_write_pre_object_map(struct rbd_obj_request *obj_req)
2946 {
2947 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
2948 	u8 new_state;
2949 
2950 	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
2951 		return 1;
2952 
2953 	if (obj_req->flags & RBD_OBJ_FLAG_DELETION)
2954 		new_state = OBJECT_PENDING;
2955 	else
2956 		new_state = OBJECT_EXISTS;
2957 
2958 	return rbd_object_map_update(obj_req, CEPH_NOSNAP, new_state, NULL);
2959 }
2960 
2961 static int rbd_obj_write_object(struct rbd_obj_request *obj_req)
2962 {
2963 	struct ceph_osd_request *osd_req;
2964 	int num_ops = count_write_ops(obj_req);
2965 	int which = 0;
2966 	int ret;
2967 
2968 	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED)
2969 		num_ops++; /* stat */
2970 
2971 	osd_req = rbd_obj_add_osd_request(obj_req, num_ops);
2972 	if (IS_ERR(osd_req))
2973 		return PTR_ERR(osd_req);
2974 
2975 	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED) {
2976 		ret = rbd_osd_setup_stat(osd_req, which++);
2977 		if (ret)
2978 			return ret;
2979 	}
2980 
2981 	rbd_osd_setup_write_ops(osd_req, which);
2982 	rbd_osd_format_write(osd_req);
2983 
2984 	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
2985 	if (ret)
2986 		return ret;
2987 
2988 	rbd_osd_submit(osd_req);
2989 	return 0;
2990 }
2991 
2992 /*
2993  * copyup_bvecs pages are never highmem pages
2994  */
2995 static bool is_zero_bvecs(struct bio_vec *bvecs, u32 bytes)
2996 {
2997 	struct ceph_bvec_iter it = {
2998 		.bvecs = bvecs,
2999 		.iter = { .bi_size = bytes },
3000 	};
3001 
3002 	ceph_bvec_iter_advance_step(&it, bytes, ({
3003 		if (memchr_inv(bvec_virt(&bv), 0, bv.bv_len))
3004 			return false;
3005 	}));
3006 	return true;
3007 }
3008 
3009 #define MODS_ONLY	U32_MAX
3010 
3011 static int rbd_obj_copyup_empty_snapc(struct rbd_obj_request *obj_req,
3012 				      u32 bytes)
3013 {
3014 	struct ceph_osd_request *osd_req;
3015 	int ret;
3016 
3017 	dout("%s obj_req %p bytes %u\n", __func__, obj_req, bytes);
3018 	rbd_assert(bytes > 0 && bytes != MODS_ONLY);
3019 
3020 	osd_req = __rbd_obj_add_osd_request(obj_req, &rbd_empty_snapc, 1);
3021 	if (IS_ERR(osd_req))
3022 		return PTR_ERR(osd_req);
3023 
3024 	ret = rbd_osd_setup_copyup(osd_req, 0, bytes);
3025 	if (ret)
3026 		return ret;
3027 
3028 	rbd_osd_format_write(osd_req);
3029 
3030 	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
3031 	if (ret)
3032 		return ret;
3033 
3034 	rbd_osd_submit(osd_req);
3035 	return 0;
3036 }
3037 
3038 static int rbd_obj_copyup_current_snapc(struct rbd_obj_request *obj_req,
3039 					u32 bytes)
3040 {
3041 	struct ceph_osd_request *osd_req;
3042 	int num_ops = count_write_ops(obj_req);
3043 	int which = 0;
3044 	int ret;
3045 
3046 	dout("%s obj_req %p bytes %u\n", __func__, obj_req, bytes);
3047 
3048 	if (bytes != MODS_ONLY)
3049 		num_ops++; /* copyup */
3050 
3051 	osd_req = rbd_obj_add_osd_request(obj_req, num_ops);
3052 	if (IS_ERR(osd_req))
3053 		return PTR_ERR(osd_req);
3054 
3055 	if (bytes != MODS_ONLY) {
3056 		ret = rbd_osd_setup_copyup(osd_req, which++, bytes);
3057 		if (ret)
3058 			return ret;
3059 	}
3060 
3061 	rbd_osd_setup_write_ops(osd_req, which);
3062 	rbd_osd_format_write(osd_req);
3063 
3064 	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
3065 	if (ret)
3066 		return ret;
3067 
3068 	rbd_osd_submit(osd_req);
3069 	return 0;
3070 }
3071 
3072 static int setup_copyup_bvecs(struct rbd_obj_request *obj_req, u64 obj_overlap)
3073 {
3074 	u32 i;
3075 
3076 	rbd_assert(!obj_req->copyup_bvecs);
3077 	obj_req->copyup_bvec_count = calc_pages_for(0, obj_overlap);
3078 	obj_req->copyup_bvecs = kcalloc(obj_req->copyup_bvec_count,
3079 					sizeof(*obj_req->copyup_bvecs),
3080 					GFP_NOIO);
3081 	if (!obj_req->copyup_bvecs)
3082 		return -ENOMEM;
3083 
3084 	for (i = 0; i < obj_req->copyup_bvec_count; i++) {
3085 		unsigned int len = min(obj_overlap, (u64)PAGE_SIZE);
3086 		struct page *page = alloc_page(GFP_NOIO);
3087 
3088 		if (!page)
3089 			return -ENOMEM;
3090 
3091 		bvec_set_page(&obj_req->copyup_bvecs[i], page, len, 0);
3092 		obj_overlap -= len;
3093 	}
3094 
3095 	rbd_assert(!obj_overlap);
3096 	return 0;
3097 }
3098 
3099 /*
3100  * The target object doesn't exist.  Read the data for the entire
3101  * target object up to the overlap point (if any) from the parent,
3102  * so we can use it for a copyup.
3103  */
3104 static int rbd_obj_copyup_read_parent(struct rbd_obj_request *obj_req)
3105 {
3106 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
3107 	int ret;
3108 
3109 	rbd_assert(obj_req->num_img_extents);
3110 	prune_extents(obj_req->img_extents, &obj_req->num_img_extents,
3111 		      rbd_dev->parent_overlap);
3112 	if (!obj_req->num_img_extents) {
3113 		/*
3114 		 * The overlap has become 0 (most likely because the
3115 		 * image has been flattened).  Re-submit the original write
3116 		 * request -- pass MODS_ONLY since the copyup isn't needed
3117 		 * anymore.
3118 		 */
3119 		return rbd_obj_copyup_current_snapc(obj_req, MODS_ONLY);
3120 	}
3121 
3122 	ret = setup_copyup_bvecs(obj_req, rbd_obj_img_extents_bytes(obj_req));
3123 	if (ret)
3124 		return ret;
3125 
3126 	return rbd_obj_read_from_parent(obj_req);
3127 }
3128 
3129 static void rbd_obj_copyup_object_maps(struct rbd_obj_request *obj_req)
3130 {
3131 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
3132 	struct ceph_snap_context *snapc = obj_req->img_request->snapc;
3133 	u8 new_state;
3134 	u32 i;
3135 	int ret;
3136 
3137 	rbd_assert(!obj_req->pending.result && !obj_req->pending.num_pending);
3138 
3139 	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
3140 		return;
3141 
3142 	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ZEROS)
3143 		return;
3144 
3145 	for (i = 0; i < snapc->num_snaps; i++) {
3146 		if ((rbd_dev->header.features & RBD_FEATURE_FAST_DIFF) &&
3147 		    i + 1 < snapc->num_snaps)
3148 			new_state = OBJECT_EXISTS_CLEAN;
3149 		else
3150 			new_state = OBJECT_EXISTS;
3151 
3152 		ret = rbd_object_map_update(obj_req, snapc->snaps[i],
3153 					    new_state, NULL);
3154 		if (ret < 0) {
3155 			obj_req->pending.result = ret;
3156 			return;
3157 		}
3158 
3159 		rbd_assert(!ret);
3160 		obj_req->pending.num_pending++;
3161 	}
3162 }
3163 
3164 static void rbd_obj_copyup_write_object(struct rbd_obj_request *obj_req)
3165 {
3166 	u32 bytes = rbd_obj_img_extents_bytes(obj_req);
3167 	int ret;
3168 
3169 	rbd_assert(!obj_req->pending.result && !obj_req->pending.num_pending);
3170 
3171 	/*
3172 	 * Only send non-zero copyup data to save some I/O and network
3173 	 * bandwidth -- zero copyup data is equivalent to the object not
3174 	 * existing.
3175 	 */
3176 	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ZEROS)
3177 		bytes = 0;
3178 
3179 	if (obj_req->img_request->snapc->num_snaps && bytes > 0) {
3180 		/*
3181 		 * Send a copyup request with an empty snapshot context to
3182 		 * deep-copyup the object through all existing snapshots.
3183 		 * A second request with the current snapshot context will be
3184 		 * sent for the actual modification.
3185 		 */
3186 		ret = rbd_obj_copyup_empty_snapc(obj_req, bytes);
3187 		if (ret) {
3188 			obj_req->pending.result = ret;
3189 			return;
3190 		}
3191 
3192 		obj_req->pending.num_pending++;
3193 		bytes = MODS_ONLY;
3194 	}
3195 
3196 	ret = rbd_obj_copyup_current_snapc(obj_req, bytes);
3197 	if (ret) {
3198 		obj_req->pending.result = ret;
3199 		return;
3200 	}
3201 
3202 	obj_req->pending.num_pending++;
3203 }
3204 
3205 static bool rbd_obj_advance_copyup(struct rbd_obj_request *obj_req, int *result)
3206 {
3207 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
3208 	int ret;
3209 
3210 again:
3211 	switch (obj_req->copyup_state) {
3212 	case RBD_OBJ_COPYUP_START:
3213 		rbd_assert(!*result);
3214 
3215 		ret = rbd_obj_copyup_read_parent(obj_req);
3216 		if (ret) {
3217 			*result = ret;
3218 			return true;
3219 		}
3220 		if (obj_req->num_img_extents)
3221 			obj_req->copyup_state = RBD_OBJ_COPYUP_READ_PARENT;
3222 		else
3223 			obj_req->copyup_state = RBD_OBJ_COPYUP_WRITE_OBJECT;
3224 		return false;
3225 	case RBD_OBJ_COPYUP_READ_PARENT:
3226 		if (*result)
3227 			return true;
3228 
3229 		if (is_zero_bvecs(obj_req->copyup_bvecs,
3230 				  rbd_obj_img_extents_bytes(obj_req))) {
3231 			dout("%s %p detected zeros\n", __func__, obj_req);
3232 			obj_req->flags |= RBD_OBJ_FLAG_COPYUP_ZEROS;
3233 		}
3234 
3235 		rbd_obj_copyup_object_maps(obj_req);
3236 		if (!obj_req->pending.num_pending) {
3237 			*result = obj_req->pending.result;
3238 			obj_req->copyup_state = RBD_OBJ_COPYUP_OBJECT_MAPS;
3239 			goto again;
3240 		}
3241 		obj_req->copyup_state = __RBD_OBJ_COPYUP_OBJECT_MAPS;
3242 		return false;
3243 	case __RBD_OBJ_COPYUP_OBJECT_MAPS:
3244 		if (!pending_result_dec(&obj_req->pending, result))
3245 			return false;
3246 		fallthrough;
3247 	case RBD_OBJ_COPYUP_OBJECT_MAPS:
3248 		if (*result) {
3249 			rbd_warn(rbd_dev, "snap object map update failed: %d",
3250 				 *result);
3251 			return true;
3252 		}
3253 
3254 		rbd_obj_copyup_write_object(obj_req);
3255 		if (!obj_req->pending.num_pending) {
3256 			*result = obj_req->pending.result;
3257 			obj_req->copyup_state = RBD_OBJ_COPYUP_WRITE_OBJECT;
3258 			goto again;
3259 		}
3260 		obj_req->copyup_state = __RBD_OBJ_COPYUP_WRITE_OBJECT;
3261 		return false;
3262 	case __RBD_OBJ_COPYUP_WRITE_OBJECT:
3263 		if (!pending_result_dec(&obj_req->pending, result))
3264 			return false;
3265 		fallthrough;
3266 	case RBD_OBJ_COPYUP_WRITE_OBJECT:
3267 		return true;
3268 	default:
3269 		BUG();
3270 	}
3271 }
3272 
3273 /*
3274  * Return:
3275  *   0 - object map update sent
3276  *   1 - object map update isn't needed
3277  *  <0 - error
3278  */
3279 static int rbd_obj_write_post_object_map(struct rbd_obj_request *obj_req)
3280 {
3281 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
3282 	u8 current_state = OBJECT_PENDING;
3283 
3284 	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
3285 		return 1;
3286 
3287 	if (!(obj_req->flags & RBD_OBJ_FLAG_DELETION))
3288 		return 1;
3289 
3290 	return rbd_object_map_update(obj_req, CEPH_NOSNAP, OBJECT_NONEXISTENT,
3291 				     &current_state);
3292 }
3293 
3294 static bool rbd_obj_advance_write(struct rbd_obj_request *obj_req, int *result)
3295 {
3296 	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
3297 	int ret;
3298 
3299 again:
3300 	switch (obj_req->write_state) {
3301 	case RBD_OBJ_WRITE_START:
3302 		rbd_assert(!*result);
3303 
3304 		rbd_obj_set_copyup_enabled(obj_req);
3305 		if (rbd_obj_write_is_noop(obj_req))
3306 			return true;
3307 
3308 		ret = rbd_obj_write_pre_object_map(obj_req);
3309 		if (ret < 0) {
3310 			*result = ret;
3311 			return true;
3312 		}
3313 		obj_req->write_state = RBD_OBJ_WRITE_PRE_OBJECT_MAP;
3314 		if (ret > 0)
3315 			goto again;
3316 		return false;
3317 	case RBD_OBJ_WRITE_PRE_OBJECT_MAP:
3318 		if (*result) {
3319 			rbd_warn(rbd_dev, "pre object map update failed: %d",
3320 				 *result);
3321 			return true;
3322 		}
3323 		ret = rbd_obj_write_object(obj_req);
3324 		if (ret) {
3325 			*result = ret;
3326 			return true;
3327 		}
3328 		obj_req->write_state = RBD_OBJ_WRITE_OBJECT;
3329 		return false;
3330 	case RBD_OBJ_WRITE_OBJECT:
3331 		if (*result == -ENOENT) {
3332 			if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED) {
3333 				*result = 0;
3334 				obj_req->copyup_state = RBD_OBJ_COPYUP_START;
3335 				obj_req->write_state = __RBD_OBJ_WRITE_COPYUP;
3336 				goto again;
3337 			}
3338 			/*
3339 			 * On a non-existent object:
3340 			 *   delete - -ENOENT, truncate/zero - 0
3341 			 */
3342 			if (obj_req->flags & RBD_OBJ_FLAG_DELETION)
3343 				*result = 0;
3344 		}
3345 		if (*result)
3346 			return true;
3347 
3348 		obj_req->write_state = RBD_OBJ_WRITE_COPYUP;
3349 		goto again;
3350 	case __RBD_OBJ_WRITE_COPYUP:
3351 		if (!rbd_obj_advance_copyup(obj_req, result))
3352 			return false;
3353 		fallthrough;
3354 	case RBD_OBJ_WRITE_COPYUP:
3355 		if (*result) {
3356 			rbd_warn(rbd_dev, "copyup failed: %d", *result);
3357 			return true;
3358 		}
3359 		ret = rbd_obj_write_post_object_map(obj_req);
3360 		if (ret < 0) {
3361 			*result = ret;
3362 			return true;
3363 		}
3364 		obj_req->write_state = RBD_OBJ_WRITE_POST_OBJECT_MAP;
3365 		if (ret > 0)
3366 			goto again;
3367 		return false;
3368 	case RBD_OBJ_WRITE_POST_OBJECT_MAP:
3369 		if (*result)
3370 			rbd_warn(rbd_dev, "post object map update failed: %d",
3371 				 *result);
3372 		return true;
3373 	default:
3374 		BUG();
3375 	}
3376 }
3377 
3378 /*
3379  * Return true if @obj_req is completed.
3380  */
3381 static bool __rbd_obj_handle_request(struct rbd_obj_request *obj_req,
3382 				     int *result)
3383 {
3384 	struct rbd_img_request *img_req = obj_req->img_request;
3385 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3386 	bool done;
3387 
3388 	mutex_lock(&obj_req->state_mutex);
3389 	if (!rbd_img_is_write(img_req))
3390 		done = rbd_obj_advance_read(obj_req, result);
3391 	else
3392 		done = rbd_obj_advance_write(obj_req, result);
3393 	mutex_unlock(&obj_req->state_mutex);
3394 
3395 	if (done && *result) {
3396 		rbd_assert(*result < 0);
3397 		rbd_warn(rbd_dev, "%s at objno %llu %llu~%llu result %d",
3398 			 obj_op_name(img_req->op_type), obj_req->ex.oe_objno,
3399 			 obj_req->ex.oe_off, obj_req->ex.oe_len, *result);
3400 	}
3401 	return done;
3402 }
3403 
3404 /*
3405  * This is open-coded in rbd_img_handle_request() to avoid parent chain
3406  * recursion.
3407  */
3408 static void rbd_obj_handle_request(struct rbd_obj_request *obj_req, int result)
3409 {
3410 	if (__rbd_obj_handle_request(obj_req, &result))
3411 		rbd_img_handle_request(obj_req->img_request, result);
3412 }
3413 
3414 static bool need_exclusive_lock(struct rbd_img_request *img_req)
3415 {
3416 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3417 
3418 	if (!(rbd_dev->header.features & RBD_FEATURE_EXCLUSIVE_LOCK))
3419 		return false;
3420 
3421 	if (rbd_is_ro(rbd_dev))
3422 		return false;
3423 
3424 	rbd_assert(!test_bit(IMG_REQ_CHILD, &img_req->flags));
3425 	if (rbd_dev->opts->lock_on_read ||
3426 	    (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
3427 		return true;
3428 
3429 	return rbd_img_is_write(img_req);
3430 }
3431 
3432 static bool rbd_lock_add_request(struct rbd_img_request *img_req)
3433 {
3434 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3435 	bool locked;
3436 
3437 	lockdep_assert_held(&rbd_dev->lock_rwsem);
3438 	locked = rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED;
3439 	spin_lock(&rbd_dev->lock_lists_lock);
3440 	rbd_assert(list_empty(&img_req->lock_item));
3441 	if (!locked)
3442 		list_add_tail(&img_req->lock_item, &rbd_dev->acquiring_list);
3443 	else
3444 		list_add_tail(&img_req->lock_item, &rbd_dev->running_list);
3445 	spin_unlock(&rbd_dev->lock_lists_lock);
3446 	return locked;
3447 }
3448 
3449 static void rbd_lock_del_request(struct rbd_img_request *img_req)
3450 {
3451 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3452 	bool need_wakeup;
3453 
3454 	lockdep_assert_held(&rbd_dev->lock_rwsem);
3455 	spin_lock(&rbd_dev->lock_lists_lock);
3456 	rbd_assert(!list_empty(&img_req->lock_item));
3457 	list_del_init(&img_req->lock_item);
3458 	need_wakeup = (rbd_dev->lock_state == RBD_LOCK_STATE_RELEASING &&
3459 		       list_empty(&rbd_dev->running_list));
3460 	spin_unlock(&rbd_dev->lock_lists_lock);
3461 	if (need_wakeup)
3462 		complete(&rbd_dev->releasing_wait);
3463 }
3464 
3465 static int rbd_img_exclusive_lock(struct rbd_img_request *img_req)
3466 {
3467 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3468 
3469 	if (!need_exclusive_lock(img_req))
3470 		return 1;
3471 
3472 	if (rbd_lock_add_request(img_req))
3473 		return 1;
3474 
3475 	if (rbd_dev->opts->exclusive) {
3476 		WARN_ON(1); /* lock got released? */
3477 		return -EROFS;
3478 	}
3479 
3480 	/*
3481 	 * Note the use of mod_delayed_work() in rbd_acquire_lock()
3482 	 * and cancel_delayed_work() in wake_lock_waiters().
3483 	 */
3484 	dout("%s rbd_dev %p queueing lock_dwork\n", __func__, rbd_dev);
3485 	queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
3486 	return 0;
3487 }
3488 
3489 static void rbd_img_object_requests(struct rbd_img_request *img_req)
3490 {
3491 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3492 	struct rbd_obj_request *obj_req;
3493 
3494 	rbd_assert(!img_req->pending.result && !img_req->pending.num_pending);
3495 	rbd_assert(!need_exclusive_lock(img_req) ||
3496 		   __rbd_is_lock_owner(rbd_dev));
3497 
3498 	if (rbd_img_is_write(img_req)) {
3499 		rbd_assert(!img_req->snapc);
3500 		down_read(&rbd_dev->header_rwsem);
3501 		img_req->snapc = ceph_get_snap_context(rbd_dev->header.snapc);
3502 		up_read(&rbd_dev->header_rwsem);
3503 	}
3504 
3505 	for_each_obj_request(img_req, obj_req) {
3506 		int result = 0;
3507 
3508 		if (__rbd_obj_handle_request(obj_req, &result)) {
3509 			if (result) {
3510 				img_req->pending.result = result;
3511 				return;
3512 			}
3513 		} else {
3514 			img_req->pending.num_pending++;
3515 		}
3516 	}
3517 }
3518 
3519 static bool rbd_img_advance(struct rbd_img_request *img_req, int *result)
3520 {
3521 	int ret;
3522 
3523 again:
3524 	switch (img_req->state) {
3525 	case RBD_IMG_START:
3526 		rbd_assert(!*result);
3527 
3528 		ret = rbd_img_exclusive_lock(img_req);
3529 		if (ret < 0) {
3530 			*result = ret;
3531 			return true;
3532 		}
3533 		img_req->state = RBD_IMG_EXCLUSIVE_LOCK;
3534 		if (ret > 0)
3535 			goto again;
3536 		return false;
3537 	case RBD_IMG_EXCLUSIVE_LOCK:
3538 		if (*result)
3539 			return true;
3540 
3541 		rbd_img_object_requests(img_req);
3542 		if (!img_req->pending.num_pending) {
3543 			*result = img_req->pending.result;
3544 			img_req->state = RBD_IMG_OBJECT_REQUESTS;
3545 			goto again;
3546 		}
3547 		img_req->state = __RBD_IMG_OBJECT_REQUESTS;
3548 		return false;
3549 	case __RBD_IMG_OBJECT_REQUESTS:
3550 		if (!pending_result_dec(&img_req->pending, result))
3551 			return false;
3552 		fallthrough;
3553 	case RBD_IMG_OBJECT_REQUESTS:
3554 		return true;
3555 	default:
3556 		BUG();
3557 	}
3558 }
3559 
3560 /*
3561  * Return true if @img_req is completed.
3562  */
3563 static bool __rbd_img_handle_request(struct rbd_img_request *img_req,
3564 				     int *result)
3565 {
3566 	struct rbd_device *rbd_dev = img_req->rbd_dev;
3567 	bool done;
3568 
3569 	if (need_exclusive_lock(img_req)) {
3570 		down_read(&rbd_dev->lock_rwsem);
3571 		mutex_lock(&img_req->state_mutex);
3572 		done = rbd_img_advance(img_req, result);
3573 		if (done)
3574 			rbd_lock_del_request(img_req);
3575 		mutex_unlock(&img_req->state_mutex);
3576 		up_read(&rbd_dev->lock_rwsem);
3577 	} else {
3578 		mutex_lock(&img_req->state_mutex);
3579 		done = rbd_img_advance(img_req, result);
3580 		mutex_unlock(&img_req->state_mutex);
3581 	}
3582 
3583 	if (done && *result) {
3584 		rbd_assert(*result < 0);
3585 		rbd_warn(rbd_dev, "%s%s result %d",
3586 		      test_bit(IMG_REQ_CHILD, &img_req->flags) ? "child " : "",
3587 		      obj_op_name(img_req->op_type), *result);
3588 	}
3589 	return done;
3590 }
3591 
3592 static void rbd_img_handle_request(struct rbd_img_request *img_req, int result)
3593 {
3594 again:
3595 	if (!__rbd_img_handle_request(img_req, &result))
3596 		return;
3597 
3598 	if (test_bit(IMG_REQ_CHILD, &img_req->flags)) {
3599 		struct rbd_obj_request *obj_req = img_req->obj_request;
3600 
3601 		rbd_img_request_destroy(img_req);
3602 		if (__rbd_obj_handle_request(obj_req, &result)) {
3603 			img_req = obj_req->img_request;
3604 			goto again;
3605 		}
3606 	} else {
3607 		struct request *rq = blk_mq_rq_from_pdu(img_req);
3608 
3609 		rbd_img_request_destroy(img_req);
3610 		blk_mq_end_request(rq, errno_to_blk_status(result));
3611 	}
3612 }
3613 
3614 static const struct rbd_client_id rbd_empty_cid;
3615 
3616 static bool rbd_cid_equal(const struct rbd_client_id *lhs,
3617 			  const struct rbd_client_id *rhs)
3618 {
3619 	return lhs->gid == rhs->gid && lhs->handle == rhs->handle;
3620 }
3621 
3622 static struct rbd_client_id rbd_get_cid(struct rbd_device *rbd_dev)
3623 {
3624 	struct rbd_client_id cid;
3625 
3626 	mutex_lock(&rbd_dev->watch_mutex);
3627 	cid.gid = ceph_client_gid(rbd_dev->rbd_client->client);
3628 	cid.handle = rbd_dev->watch_cookie;
3629 	mutex_unlock(&rbd_dev->watch_mutex);
3630 	return cid;
3631 }
3632 
3633 /*
3634  * lock_rwsem must be held for write
3635  */
3636 static void rbd_set_owner_cid(struct rbd_device *rbd_dev,
3637 			      const struct rbd_client_id *cid)
3638 {
3639 	dout("%s rbd_dev %p %llu-%llu -> %llu-%llu\n", __func__, rbd_dev,
3640 	     rbd_dev->owner_cid.gid, rbd_dev->owner_cid.handle,
3641 	     cid->gid, cid->handle);
3642 	rbd_dev->owner_cid = *cid; /* struct */
3643 }
3644 
3645 static void format_lock_cookie(struct rbd_device *rbd_dev, char *buf)
3646 {
3647 	mutex_lock(&rbd_dev->watch_mutex);
3648 	sprintf(buf, "%s %llu", RBD_LOCK_COOKIE_PREFIX, rbd_dev->watch_cookie);
3649 	mutex_unlock(&rbd_dev->watch_mutex);
3650 }
3651 
3652 static void __rbd_lock(struct rbd_device *rbd_dev, const char *cookie)
3653 {
3654 	struct rbd_client_id cid = rbd_get_cid(rbd_dev);
3655 
3656 	rbd_dev->lock_state = RBD_LOCK_STATE_LOCKED;
3657 	strcpy(rbd_dev->lock_cookie, cookie);
3658 	rbd_set_owner_cid(rbd_dev, &cid);
3659 	queue_work(rbd_dev->task_wq, &rbd_dev->acquired_lock_work);
3660 }
3661 
3662 /*
3663  * lock_rwsem must be held for write
3664  */
3665 static int rbd_lock(struct rbd_device *rbd_dev)
3666 {
3667 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3668 	char cookie[32];
3669 	int ret;
3670 
3671 	WARN_ON(__rbd_is_lock_owner(rbd_dev) ||
3672 		rbd_dev->lock_cookie[0] != '\0');
3673 
3674 	format_lock_cookie(rbd_dev, cookie);
3675 	ret = ceph_cls_lock(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
3676 			    RBD_LOCK_NAME, CEPH_CLS_LOCK_EXCLUSIVE, cookie,
3677 			    RBD_LOCK_TAG, "", 0);
3678 	if (ret)
3679 		return ret;
3680 
3681 	__rbd_lock(rbd_dev, cookie);
3682 	return 0;
3683 }
3684 
3685 /*
3686  * lock_rwsem must be held for write
3687  */
3688 static void rbd_unlock(struct rbd_device *rbd_dev)
3689 {
3690 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3691 	int ret;
3692 
3693 	WARN_ON(!__rbd_is_lock_owner(rbd_dev) ||
3694 		rbd_dev->lock_cookie[0] == '\0');
3695 
3696 	ret = ceph_cls_unlock(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
3697 			      RBD_LOCK_NAME, rbd_dev->lock_cookie);
3698 	if (ret && ret != -ENOENT)
3699 		rbd_warn(rbd_dev, "failed to unlock header: %d", ret);
3700 
3701 	/* treat errors as the image is unlocked */
3702 	rbd_dev->lock_state = RBD_LOCK_STATE_UNLOCKED;
3703 	rbd_dev->lock_cookie[0] = '\0';
3704 	rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
3705 	queue_work(rbd_dev->task_wq, &rbd_dev->released_lock_work);
3706 }
3707 
3708 static int __rbd_notify_op_lock(struct rbd_device *rbd_dev,
3709 				enum rbd_notify_op notify_op,
3710 				struct page ***preply_pages,
3711 				size_t *preply_len)
3712 {
3713 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3714 	struct rbd_client_id cid = rbd_get_cid(rbd_dev);
3715 	char buf[4 + 8 + 8 + CEPH_ENCODING_START_BLK_LEN];
3716 	int buf_size = sizeof(buf);
3717 	void *p = buf;
3718 
3719 	dout("%s rbd_dev %p notify_op %d\n", __func__, rbd_dev, notify_op);
3720 
3721 	/* encode *LockPayload NotifyMessage (op + ClientId) */
3722 	ceph_start_encoding(&p, 2, 1, buf_size - CEPH_ENCODING_START_BLK_LEN);
3723 	ceph_encode_32(&p, notify_op);
3724 	ceph_encode_64(&p, cid.gid);
3725 	ceph_encode_64(&p, cid.handle);
3726 
3727 	return ceph_osdc_notify(osdc, &rbd_dev->header_oid,
3728 				&rbd_dev->header_oloc, buf, buf_size,
3729 				RBD_NOTIFY_TIMEOUT, preply_pages, preply_len);
3730 }
3731 
3732 static void rbd_notify_op_lock(struct rbd_device *rbd_dev,
3733 			       enum rbd_notify_op notify_op)
3734 {
3735 	__rbd_notify_op_lock(rbd_dev, notify_op, NULL, NULL);
3736 }
3737 
3738 static void rbd_notify_acquired_lock(struct work_struct *work)
3739 {
3740 	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
3741 						  acquired_lock_work);
3742 
3743 	rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_ACQUIRED_LOCK);
3744 }
3745 
3746 static void rbd_notify_released_lock(struct work_struct *work)
3747 {
3748 	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
3749 						  released_lock_work);
3750 
3751 	rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_RELEASED_LOCK);
3752 }
3753 
3754 static int rbd_request_lock(struct rbd_device *rbd_dev)
3755 {
3756 	struct page **reply_pages;
3757 	size_t reply_len;
3758 	bool lock_owner_responded = false;
3759 	int ret;
3760 
3761 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
3762 
3763 	ret = __rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_REQUEST_LOCK,
3764 				   &reply_pages, &reply_len);
3765 	if (ret && ret != -ETIMEDOUT) {
3766 		rbd_warn(rbd_dev, "failed to request lock: %d", ret);
3767 		goto out;
3768 	}
3769 
3770 	if (reply_len > 0 && reply_len <= PAGE_SIZE) {
3771 		void *p = page_address(reply_pages[0]);
3772 		void *const end = p + reply_len;
3773 		u32 n;
3774 
3775 		ceph_decode_32_safe(&p, end, n, e_inval); /* num_acks */
3776 		while (n--) {
3777 			u8 struct_v;
3778 			u32 len;
3779 
3780 			ceph_decode_need(&p, end, 8 + 8, e_inval);
3781 			p += 8 + 8; /* skip gid and cookie */
3782 
3783 			ceph_decode_32_safe(&p, end, len, e_inval);
3784 			if (!len)
3785 				continue;
3786 
3787 			if (lock_owner_responded) {
3788 				rbd_warn(rbd_dev,
3789 					 "duplicate lock owners detected");
3790 				ret = -EIO;
3791 				goto out;
3792 			}
3793 
3794 			lock_owner_responded = true;
3795 			ret = ceph_start_decoding(&p, end, 1, "ResponseMessage",
3796 						  &struct_v, &len);
3797 			if (ret) {
3798 				rbd_warn(rbd_dev,
3799 					 "failed to decode ResponseMessage: %d",
3800 					 ret);
3801 				goto e_inval;
3802 			}
3803 
3804 			ret = ceph_decode_32(&p);
3805 		}
3806 	}
3807 
3808 	if (!lock_owner_responded) {
3809 		rbd_warn(rbd_dev, "no lock owners detected");
3810 		ret = -ETIMEDOUT;
3811 	}
3812 
3813 out:
3814 	ceph_release_page_vector(reply_pages, calc_pages_for(0, reply_len));
3815 	return ret;
3816 
3817 e_inval:
3818 	ret = -EINVAL;
3819 	goto out;
3820 }
3821 
3822 /*
3823  * Either image request state machine(s) or rbd_add_acquire_lock()
3824  * (i.e. "rbd map").
3825  */
3826 static void wake_lock_waiters(struct rbd_device *rbd_dev, int result)
3827 {
3828 	struct rbd_img_request *img_req;
3829 
3830 	dout("%s rbd_dev %p result %d\n", __func__, rbd_dev, result);
3831 	lockdep_assert_held_write(&rbd_dev->lock_rwsem);
3832 
3833 	cancel_delayed_work(&rbd_dev->lock_dwork);
3834 	if (!completion_done(&rbd_dev->acquire_wait)) {
3835 		rbd_assert(list_empty(&rbd_dev->acquiring_list) &&
3836 			   list_empty(&rbd_dev->running_list));
3837 		rbd_dev->acquire_err = result;
3838 		complete_all(&rbd_dev->acquire_wait);
3839 		return;
3840 	}
3841 
3842 	list_for_each_entry(img_req, &rbd_dev->acquiring_list, lock_item) {
3843 		mutex_lock(&img_req->state_mutex);
3844 		rbd_assert(img_req->state == RBD_IMG_EXCLUSIVE_LOCK);
3845 		rbd_img_schedule(img_req, result);
3846 		mutex_unlock(&img_req->state_mutex);
3847 	}
3848 
3849 	list_splice_tail_init(&rbd_dev->acquiring_list, &rbd_dev->running_list);
3850 }
3851 
3852 static int get_lock_owner_info(struct rbd_device *rbd_dev,
3853 			       struct ceph_locker **lockers, u32 *num_lockers)
3854 {
3855 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3856 	u8 lock_type;
3857 	char *lock_tag;
3858 	int ret;
3859 
3860 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
3861 
3862 	ret = ceph_cls_lock_info(osdc, &rbd_dev->header_oid,
3863 				 &rbd_dev->header_oloc, RBD_LOCK_NAME,
3864 				 &lock_type, &lock_tag, lockers, num_lockers);
3865 	if (ret)
3866 		return ret;
3867 
3868 	if (*num_lockers == 0) {
3869 		dout("%s rbd_dev %p no lockers detected\n", __func__, rbd_dev);
3870 		goto out;
3871 	}
3872 
3873 	if (strcmp(lock_tag, RBD_LOCK_TAG)) {
3874 		rbd_warn(rbd_dev, "locked by external mechanism, tag %s",
3875 			 lock_tag);
3876 		ret = -EBUSY;
3877 		goto out;
3878 	}
3879 
3880 	if (lock_type == CEPH_CLS_LOCK_SHARED) {
3881 		rbd_warn(rbd_dev, "shared lock type detected");
3882 		ret = -EBUSY;
3883 		goto out;
3884 	}
3885 
3886 	if (strncmp((*lockers)[0].id.cookie, RBD_LOCK_COOKIE_PREFIX,
3887 		    strlen(RBD_LOCK_COOKIE_PREFIX))) {
3888 		rbd_warn(rbd_dev, "locked by external mechanism, cookie %s",
3889 			 (*lockers)[0].id.cookie);
3890 		ret = -EBUSY;
3891 		goto out;
3892 	}
3893 
3894 out:
3895 	kfree(lock_tag);
3896 	return ret;
3897 }
3898 
3899 static int find_watcher(struct rbd_device *rbd_dev,
3900 			const struct ceph_locker *locker)
3901 {
3902 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3903 	struct ceph_watch_item *watchers;
3904 	u32 num_watchers;
3905 	u64 cookie;
3906 	int i;
3907 	int ret;
3908 
3909 	ret = ceph_osdc_list_watchers(osdc, &rbd_dev->header_oid,
3910 				      &rbd_dev->header_oloc, &watchers,
3911 				      &num_watchers);
3912 	if (ret)
3913 		return ret;
3914 
3915 	sscanf(locker->id.cookie, RBD_LOCK_COOKIE_PREFIX " %llu", &cookie);
3916 	for (i = 0; i < num_watchers; i++) {
3917 		/*
3918 		 * Ignore addr->type while comparing.  This mimics
3919 		 * entity_addr_t::get_legacy_str() + strcmp().
3920 		 */
3921 		if (ceph_addr_equal_no_type(&watchers[i].addr,
3922 					    &locker->info.addr) &&
3923 		    watchers[i].cookie == cookie) {
3924 			struct rbd_client_id cid = {
3925 				.gid = le64_to_cpu(watchers[i].name.num),
3926 				.handle = cookie,
3927 			};
3928 
3929 			dout("%s rbd_dev %p found cid %llu-%llu\n", __func__,
3930 			     rbd_dev, cid.gid, cid.handle);
3931 			rbd_set_owner_cid(rbd_dev, &cid);
3932 			ret = 1;
3933 			goto out;
3934 		}
3935 	}
3936 
3937 	dout("%s rbd_dev %p no watchers\n", __func__, rbd_dev);
3938 	ret = 0;
3939 out:
3940 	kfree(watchers);
3941 	return ret;
3942 }
3943 
3944 /*
3945  * lock_rwsem must be held for write
3946  */
3947 static int rbd_try_lock(struct rbd_device *rbd_dev)
3948 {
3949 	struct ceph_client *client = rbd_dev->rbd_client->client;
3950 	struct ceph_locker *lockers;
3951 	u32 num_lockers;
3952 	int ret;
3953 
3954 	for (;;) {
3955 		ret = rbd_lock(rbd_dev);
3956 		if (ret != -EBUSY)
3957 			return ret;
3958 
3959 		/* determine if the current lock holder is still alive */
3960 		ret = get_lock_owner_info(rbd_dev, &lockers, &num_lockers);
3961 		if (ret)
3962 			return ret;
3963 
3964 		if (num_lockers == 0)
3965 			goto again;
3966 
3967 		ret = find_watcher(rbd_dev, lockers);
3968 		if (ret)
3969 			goto out; /* request lock or error */
3970 
3971 		rbd_warn(rbd_dev, "breaking header lock owned by %s%llu",
3972 			 ENTITY_NAME(lockers[0].id.name));
3973 
3974 		ret = ceph_monc_blocklist_add(&client->monc,
3975 					      &lockers[0].info.addr);
3976 		if (ret) {
3977 			rbd_warn(rbd_dev, "blocklist of %s%llu failed: %d",
3978 				 ENTITY_NAME(lockers[0].id.name), ret);
3979 			goto out;
3980 		}
3981 
3982 		ret = ceph_cls_break_lock(&client->osdc, &rbd_dev->header_oid,
3983 					  &rbd_dev->header_oloc, RBD_LOCK_NAME,
3984 					  lockers[0].id.cookie,
3985 					  &lockers[0].id.name);
3986 		if (ret && ret != -ENOENT)
3987 			goto out;
3988 
3989 again:
3990 		ceph_free_lockers(lockers, num_lockers);
3991 	}
3992 
3993 out:
3994 	ceph_free_lockers(lockers, num_lockers);
3995 	return ret;
3996 }
3997 
3998 static int rbd_post_acquire_action(struct rbd_device *rbd_dev)
3999 {
4000 	int ret;
4001 
4002 	ret = rbd_dev_refresh(rbd_dev);
4003 	if (ret)
4004 		return ret;
4005 
4006 	if (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP) {
4007 		ret = rbd_object_map_open(rbd_dev);
4008 		if (ret)
4009 			return ret;
4010 	}
4011 
4012 	return 0;
4013 }
4014 
4015 /*
4016  * Return:
4017  *   0 - lock acquired
4018  *   1 - caller should call rbd_request_lock()
4019  *  <0 - error
4020  */
4021 static int rbd_try_acquire_lock(struct rbd_device *rbd_dev)
4022 {
4023 	int ret;
4024 
4025 	down_read(&rbd_dev->lock_rwsem);
4026 	dout("%s rbd_dev %p read lock_state %d\n", __func__, rbd_dev,
4027 	     rbd_dev->lock_state);
4028 	if (__rbd_is_lock_owner(rbd_dev)) {
4029 		up_read(&rbd_dev->lock_rwsem);
4030 		return 0;
4031 	}
4032 
4033 	up_read(&rbd_dev->lock_rwsem);
4034 	down_write(&rbd_dev->lock_rwsem);
4035 	dout("%s rbd_dev %p write lock_state %d\n", __func__, rbd_dev,
4036 	     rbd_dev->lock_state);
4037 	if (__rbd_is_lock_owner(rbd_dev)) {
4038 		up_write(&rbd_dev->lock_rwsem);
4039 		return 0;
4040 	}
4041 
4042 	ret = rbd_try_lock(rbd_dev);
4043 	if (ret < 0) {
4044 		rbd_warn(rbd_dev, "failed to lock header: %d", ret);
4045 		if (ret == -EBLOCKLISTED)
4046 			goto out;
4047 
4048 		ret = 1; /* request lock anyway */
4049 	}
4050 	if (ret > 0) {
4051 		up_write(&rbd_dev->lock_rwsem);
4052 		return ret;
4053 	}
4054 
4055 	rbd_assert(rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED);
4056 	rbd_assert(list_empty(&rbd_dev->running_list));
4057 
4058 	ret = rbd_post_acquire_action(rbd_dev);
4059 	if (ret) {
4060 		rbd_warn(rbd_dev, "post-acquire action failed: %d", ret);
4061 		/*
4062 		 * Can't stay in RBD_LOCK_STATE_LOCKED because
4063 		 * rbd_lock_add_request() would let the request through,
4064 		 * assuming that e.g. object map is locked and loaded.
4065 		 */
4066 		rbd_unlock(rbd_dev);
4067 	}
4068 
4069 out:
4070 	wake_lock_waiters(rbd_dev, ret);
4071 	up_write(&rbd_dev->lock_rwsem);
4072 	return ret;
4073 }
4074 
4075 static void rbd_acquire_lock(struct work_struct *work)
4076 {
4077 	struct rbd_device *rbd_dev = container_of(to_delayed_work(work),
4078 					    struct rbd_device, lock_dwork);
4079 	int ret;
4080 
4081 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4082 again:
4083 	ret = rbd_try_acquire_lock(rbd_dev);
4084 	if (ret <= 0) {
4085 		dout("%s rbd_dev %p ret %d - done\n", __func__, rbd_dev, ret);
4086 		return;
4087 	}
4088 
4089 	ret = rbd_request_lock(rbd_dev);
4090 	if (ret == -ETIMEDOUT) {
4091 		goto again; /* treat this as a dead client */
4092 	} else if (ret == -EROFS) {
4093 		rbd_warn(rbd_dev, "peer will not release lock");
4094 		down_write(&rbd_dev->lock_rwsem);
4095 		wake_lock_waiters(rbd_dev, ret);
4096 		up_write(&rbd_dev->lock_rwsem);
4097 	} else if (ret < 0) {
4098 		rbd_warn(rbd_dev, "error requesting lock: %d", ret);
4099 		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork,
4100 				 RBD_RETRY_DELAY);
4101 	} else {
4102 		/*
4103 		 * lock owner acked, but resend if we don't see them
4104 		 * release the lock
4105 		 */
4106 		dout("%s rbd_dev %p requeuing lock_dwork\n", __func__,
4107 		     rbd_dev);
4108 		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork,
4109 		    msecs_to_jiffies(2 * RBD_NOTIFY_TIMEOUT * MSEC_PER_SEC));
4110 	}
4111 }
4112 
4113 static bool rbd_quiesce_lock(struct rbd_device *rbd_dev)
4114 {
4115 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4116 	lockdep_assert_held_write(&rbd_dev->lock_rwsem);
4117 
4118 	if (rbd_dev->lock_state != RBD_LOCK_STATE_LOCKED)
4119 		return false;
4120 
4121 	/*
4122 	 * Ensure that all in-flight IO is flushed.
4123 	 */
4124 	rbd_dev->lock_state = RBD_LOCK_STATE_RELEASING;
4125 	rbd_assert(!completion_done(&rbd_dev->releasing_wait));
4126 	if (list_empty(&rbd_dev->running_list))
4127 		return true;
4128 
4129 	up_write(&rbd_dev->lock_rwsem);
4130 	wait_for_completion(&rbd_dev->releasing_wait);
4131 
4132 	down_write(&rbd_dev->lock_rwsem);
4133 	if (rbd_dev->lock_state != RBD_LOCK_STATE_RELEASING)
4134 		return false;
4135 
4136 	rbd_assert(list_empty(&rbd_dev->running_list));
4137 	return true;
4138 }
4139 
4140 static void rbd_pre_release_action(struct rbd_device *rbd_dev)
4141 {
4142 	if (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP)
4143 		rbd_object_map_close(rbd_dev);
4144 }
4145 
4146 static void __rbd_release_lock(struct rbd_device *rbd_dev)
4147 {
4148 	rbd_assert(list_empty(&rbd_dev->running_list));
4149 
4150 	rbd_pre_release_action(rbd_dev);
4151 	rbd_unlock(rbd_dev);
4152 }
4153 
4154 /*
4155  * lock_rwsem must be held for write
4156  */
4157 static void rbd_release_lock(struct rbd_device *rbd_dev)
4158 {
4159 	if (!rbd_quiesce_lock(rbd_dev))
4160 		return;
4161 
4162 	__rbd_release_lock(rbd_dev);
4163 
4164 	/*
4165 	 * Give others a chance to grab the lock - we would re-acquire
4166 	 * almost immediately if we got new IO while draining the running
4167 	 * list otherwise.  We need to ack our own notifications, so this
4168 	 * lock_dwork will be requeued from rbd_handle_released_lock() by
4169 	 * way of maybe_kick_acquire().
4170 	 */
4171 	cancel_delayed_work(&rbd_dev->lock_dwork);
4172 }
4173 
4174 static void rbd_release_lock_work(struct work_struct *work)
4175 {
4176 	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
4177 						  unlock_work);
4178 
4179 	down_write(&rbd_dev->lock_rwsem);
4180 	rbd_release_lock(rbd_dev);
4181 	up_write(&rbd_dev->lock_rwsem);
4182 }
4183 
4184 static void maybe_kick_acquire(struct rbd_device *rbd_dev)
4185 {
4186 	bool have_requests;
4187 
4188 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4189 	if (__rbd_is_lock_owner(rbd_dev))
4190 		return;
4191 
4192 	spin_lock(&rbd_dev->lock_lists_lock);
4193 	have_requests = !list_empty(&rbd_dev->acquiring_list);
4194 	spin_unlock(&rbd_dev->lock_lists_lock);
4195 	if (have_requests || delayed_work_pending(&rbd_dev->lock_dwork)) {
4196 		dout("%s rbd_dev %p kicking lock_dwork\n", __func__, rbd_dev);
4197 		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
4198 	}
4199 }
4200 
4201 static void rbd_handle_acquired_lock(struct rbd_device *rbd_dev, u8 struct_v,
4202 				     void **p)
4203 {
4204 	struct rbd_client_id cid = { 0 };
4205 
4206 	if (struct_v >= 2) {
4207 		cid.gid = ceph_decode_64(p);
4208 		cid.handle = ceph_decode_64(p);
4209 	}
4210 
4211 	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
4212 	     cid.handle);
4213 	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
4214 		down_write(&rbd_dev->lock_rwsem);
4215 		if (rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
4216 			dout("%s rbd_dev %p cid %llu-%llu == owner_cid\n",
4217 			     __func__, rbd_dev, cid.gid, cid.handle);
4218 		} else {
4219 			rbd_set_owner_cid(rbd_dev, &cid);
4220 		}
4221 		downgrade_write(&rbd_dev->lock_rwsem);
4222 	} else {
4223 		down_read(&rbd_dev->lock_rwsem);
4224 	}
4225 
4226 	maybe_kick_acquire(rbd_dev);
4227 	up_read(&rbd_dev->lock_rwsem);
4228 }
4229 
4230 static void rbd_handle_released_lock(struct rbd_device *rbd_dev, u8 struct_v,
4231 				     void **p)
4232 {
4233 	struct rbd_client_id cid = { 0 };
4234 
4235 	if (struct_v >= 2) {
4236 		cid.gid = ceph_decode_64(p);
4237 		cid.handle = ceph_decode_64(p);
4238 	}
4239 
4240 	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
4241 	     cid.handle);
4242 	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
4243 		down_write(&rbd_dev->lock_rwsem);
4244 		if (!rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
4245 			dout("%s rbd_dev %p cid %llu-%llu != owner_cid %llu-%llu\n",
4246 			     __func__, rbd_dev, cid.gid, cid.handle,
4247 			     rbd_dev->owner_cid.gid, rbd_dev->owner_cid.handle);
4248 		} else {
4249 			rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
4250 		}
4251 		downgrade_write(&rbd_dev->lock_rwsem);
4252 	} else {
4253 		down_read(&rbd_dev->lock_rwsem);
4254 	}
4255 
4256 	maybe_kick_acquire(rbd_dev);
4257 	up_read(&rbd_dev->lock_rwsem);
4258 }
4259 
4260 /*
4261  * Returns result for ResponseMessage to be encoded (<= 0), or 1 if no
4262  * ResponseMessage is needed.
4263  */
4264 static int rbd_handle_request_lock(struct rbd_device *rbd_dev, u8 struct_v,
4265 				   void **p)
4266 {
4267 	struct rbd_client_id my_cid = rbd_get_cid(rbd_dev);
4268 	struct rbd_client_id cid = { 0 };
4269 	int result = 1;
4270 
4271 	if (struct_v >= 2) {
4272 		cid.gid = ceph_decode_64(p);
4273 		cid.handle = ceph_decode_64(p);
4274 	}
4275 
4276 	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
4277 	     cid.handle);
4278 	if (rbd_cid_equal(&cid, &my_cid))
4279 		return result;
4280 
4281 	down_read(&rbd_dev->lock_rwsem);
4282 	if (__rbd_is_lock_owner(rbd_dev)) {
4283 		if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED &&
4284 		    rbd_cid_equal(&rbd_dev->owner_cid, &rbd_empty_cid))
4285 			goto out_unlock;
4286 
4287 		/*
4288 		 * encode ResponseMessage(0) so the peer can detect
4289 		 * a missing owner
4290 		 */
4291 		result = 0;
4292 
4293 		if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED) {
4294 			if (!rbd_dev->opts->exclusive) {
4295 				dout("%s rbd_dev %p queueing unlock_work\n",
4296 				     __func__, rbd_dev);
4297 				queue_work(rbd_dev->task_wq,
4298 					   &rbd_dev->unlock_work);
4299 			} else {
4300 				/* refuse to release the lock */
4301 				result = -EROFS;
4302 			}
4303 		}
4304 	}
4305 
4306 out_unlock:
4307 	up_read(&rbd_dev->lock_rwsem);
4308 	return result;
4309 }
4310 
4311 static void __rbd_acknowledge_notify(struct rbd_device *rbd_dev,
4312 				     u64 notify_id, u64 cookie, s32 *result)
4313 {
4314 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4315 	char buf[4 + CEPH_ENCODING_START_BLK_LEN];
4316 	int buf_size = sizeof(buf);
4317 	int ret;
4318 
4319 	if (result) {
4320 		void *p = buf;
4321 
4322 		/* encode ResponseMessage */
4323 		ceph_start_encoding(&p, 1, 1,
4324 				    buf_size - CEPH_ENCODING_START_BLK_LEN);
4325 		ceph_encode_32(&p, *result);
4326 	} else {
4327 		buf_size = 0;
4328 	}
4329 
4330 	ret = ceph_osdc_notify_ack(osdc, &rbd_dev->header_oid,
4331 				   &rbd_dev->header_oloc, notify_id, cookie,
4332 				   buf, buf_size);
4333 	if (ret)
4334 		rbd_warn(rbd_dev, "acknowledge_notify failed: %d", ret);
4335 }
4336 
4337 static void rbd_acknowledge_notify(struct rbd_device *rbd_dev, u64 notify_id,
4338 				   u64 cookie)
4339 {
4340 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4341 	__rbd_acknowledge_notify(rbd_dev, notify_id, cookie, NULL);
4342 }
4343 
4344 static void rbd_acknowledge_notify_result(struct rbd_device *rbd_dev,
4345 					  u64 notify_id, u64 cookie, s32 result)
4346 {
4347 	dout("%s rbd_dev %p result %d\n", __func__, rbd_dev, result);
4348 	__rbd_acknowledge_notify(rbd_dev, notify_id, cookie, &result);
4349 }
4350 
4351 static void rbd_watch_cb(void *arg, u64 notify_id, u64 cookie,
4352 			 u64 notifier_id, void *data, size_t data_len)
4353 {
4354 	struct rbd_device *rbd_dev = arg;
4355 	void *p = data;
4356 	void *const end = p + data_len;
4357 	u8 struct_v = 0;
4358 	u32 len;
4359 	u32 notify_op;
4360 	int ret;
4361 
4362 	dout("%s rbd_dev %p cookie %llu notify_id %llu data_len %zu\n",
4363 	     __func__, rbd_dev, cookie, notify_id, data_len);
4364 	if (data_len) {
4365 		ret = ceph_start_decoding(&p, end, 1, "NotifyMessage",
4366 					  &struct_v, &len);
4367 		if (ret) {
4368 			rbd_warn(rbd_dev, "failed to decode NotifyMessage: %d",
4369 				 ret);
4370 			return;
4371 		}
4372 
4373 		notify_op = ceph_decode_32(&p);
4374 	} else {
4375 		/* legacy notification for header updates */
4376 		notify_op = RBD_NOTIFY_OP_HEADER_UPDATE;
4377 		len = 0;
4378 	}
4379 
4380 	dout("%s rbd_dev %p notify_op %u\n", __func__, rbd_dev, notify_op);
4381 	switch (notify_op) {
4382 	case RBD_NOTIFY_OP_ACQUIRED_LOCK:
4383 		rbd_handle_acquired_lock(rbd_dev, struct_v, &p);
4384 		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
4385 		break;
4386 	case RBD_NOTIFY_OP_RELEASED_LOCK:
4387 		rbd_handle_released_lock(rbd_dev, struct_v, &p);
4388 		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
4389 		break;
4390 	case RBD_NOTIFY_OP_REQUEST_LOCK:
4391 		ret = rbd_handle_request_lock(rbd_dev, struct_v, &p);
4392 		if (ret <= 0)
4393 			rbd_acknowledge_notify_result(rbd_dev, notify_id,
4394 						      cookie, ret);
4395 		else
4396 			rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
4397 		break;
4398 	case RBD_NOTIFY_OP_HEADER_UPDATE:
4399 		ret = rbd_dev_refresh(rbd_dev);
4400 		if (ret)
4401 			rbd_warn(rbd_dev, "refresh failed: %d", ret);
4402 
4403 		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
4404 		break;
4405 	default:
4406 		if (rbd_is_lock_owner(rbd_dev))
4407 			rbd_acknowledge_notify_result(rbd_dev, notify_id,
4408 						      cookie, -EOPNOTSUPP);
4409 		else
4410 			rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
4411 		break;
4412 	}
4413 }
4414 
4415 static void __rbd_unregister_watch(struct rbd_device *rbd_dev);
4416 
4417 static void rbd_watch_errcb(void *arg, u64 cookie, int err)
4418 {
4419 	struct rbd_device *rbd_dev = arg;
4420 
4421 	rbd_warn(rbd_dev, "encountered watch error: %d", err);
4422 
4423 	down_write(&rbd_dev->lock_rwsem);
4424 	rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
4425 	up_write(&rbd_dev->lock_rwsem);
4426 
4427 	mutex_lock(&rbd_dev->watch_mutex);
4428 	if (rbd_dev->watch_state == RBD_WATCH_STATE_REGISTERED) {
4429 		__rbd_unregister_watch(rbd_dev);
4430 		rbd_dev->watch_state = RBD_WATCH_STATE_ERROR;
4431 
4432 		queue_delayed_work(rbd_dev->task_wq, &rbd_dev->watch_dwork, 0);
4433 	}
4434 	mutex_unlock(&rbd_dev->watch_mutex);
4435 }
4436 
4437 /*
4438  * watch_mutex must be locked
4439  */
4440 static int __rbd_register_watch(struct rbd_device *rbd_dev)
4441 {
4442 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4443 	struct ceph_osd_linger_request *handle;
4444 
4445 	rbd_assert(!rbd_dev->watch_handle);
4446 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4447 
4448 	handle = ceph_osdc_watch(osdc, &rbd_dev->header_oid,
4449 				 &rbd_dev->header_oloc, rbd_watch_cb,
4450 				 rbd_watch_errcb, rbd_dev);
4451 	if (IS_ERR(handle))
4452 		return PTR_ERR(handle);
4453 
4454 	rbd_dev->watch_handle = handle;
4455 	return 0;
4456 }
4457 
4458 /*
4459  * watch_mutex must be locked
4460  */
4461 static void __rbd_unregister_watch(struct rbd_device *rbd_dev)
4462 {
4463 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4464 	int ret;
4465 
4466 	rbd_assert(rbd_dev->watch_handle);
4467 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4468 
4469 	ret = ceph_osdc_unwatch(osdc, rbd_dev->watch_handle);
4470 	if (ret)
4471 		rbd_warn(rbd_dev, "failed to unwatch: %d", ret);
4472 
4473 	rbd_dev->watch_handle = NULL;
4474 }
4475 
4476 static int rbd_register_watch(struct rbd_device *rbd_dev)
4477 {
4478 	int ret;
4479 
4480 	mutex_lock(&rbd_dev->watch_mutex);
4481 	rbd_assert(rbd_dev->watch_state == RBD_WATCH_STATE_UNREGISTERED);
4482 	ret = __rbd_register_watch(rbd_dev);
4483 	if (ret)
4484 		goto out;
4485 
4486 	rbd_dev->watch_state = RBD_WATCH_STATE_REGISTERED;
4487 	rbd_dev->watch_cookie = rbd_dev->watch_handle->linger_id;
4488 
4489 out:
4490 	mutex_unlock(&rbd_dev->watch_mutex);
4491 	return ret;
4492 }
4493 
4494 static void cancel_tasks_sync(struct rbd_device *rbd_dev)
4495 {
4496 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4497 
4498 	cancel_work_sync(&rbd_dev->acquired_lock_work);
4499 	cancel_work_sync(&rbd_dev->released_lock_work);
4500 	cancel_delayed_work_sync(&rbd_dev->lock_dwork);
4501 	cancel_work_sync(&rbd_dev->unlock_work);
4502 }
4503 
4504 /*
4505  * header_rwsem must not be held to avoid a deadlock with
4506  * rbd_dev_refresh() when flushing notifies.
4507  */
4508 static void rbd_unregister_watch(struct rbd_device *rbd_dev)
4509 {
4510 	cancel_tasks_sync(rbd_dev);
4511 
4512 	mutex_lock(&rbd_dev->watch_mutex);
4513 	if (rbd_dev->watch_state == RBD_WATCH_STATE_REGISTERED)
4514 		__rbd_unregister_watch(rbd_dev);
4515 	rbd_dev->watch_state = RBD_WATCH_STATE_UNREGISTERED;
4516 	mutex_unlock(&rbd_dev->watch_mutex);
4517 
4518 	cancel_delayed_work_sync(&rbd_dev->watch_dwork);
4519 	ceph_osdc_flush_notifies(&rbd_dev->rbd_client->client->osdc);
4520 }
4521 
4522 /*
4523  * lock_rwsem must be held for write
4524  */
4525 static void rbd_reacquire_lock(struct rbd_device *rbd_dev)
4526 {
4527 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4528 	char cookie[32];
4529 	int ret;
4530 
4531 	if (!rbd_quiesce_lock(rbd_dev))
4532 		return;
4533 
4534 	format_lock_cookie(rbd_dev, cookie);
4535 	ret = ceph_cls_set_cookie(osdc, &rbd_dev->header_oid,
4536 				  &rbd_dev->header_oloc, RBD_LOCK_NAME,
4537 				  CEPH_CLS_LOCK_EXCLUSIVE, rbd_dev->lock_cookie,
4538 				  RBD_LOCK_TAG, cookie);
4539 	if (ret) {
4540 		if (ret != -EOPNOTSUPP)
4541 			rbd_warn(rbd_dev, "failed to update lock cookie: %d",
4542 				 ret);
4543 
4544 		/*
4545 		 * Lock cookie cannot be updated on older OSDs, so do
4546 		 * a manual release and queue an acquire.
4547 		 */
4548 		__rbd_release_lock(rbd_dev);
4549 		queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
4550 	} else {
4551 		__rbd_lock(rbd_dev, cookie);
4552 		wake_lock_waiters(rbd_dev, 0);
4553 	}
4554 }
4555 
4556 static void rbd_reregister_watch(struct work_struct *work)
4557 {
4558 	struct rbd_device *rbd_dev = container_of(to_delayed_work(work),
4559 					    struct rbd_device, watch_dwork);
4560 	int ret;
4561 
4562 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
4563 
4564 	mutex_lock(&rbd_dev->watch_mutex);
4565 	if (rbd_dev->watch_state != RBD_WATCH_STATE_ERROR) {
4566 		mutex_unlock(&rbd_dev->watch_mutex);
4567 		return;
4568 	}
4569 
4570 	ret = __rbd_register_watch(rbd_dev);
4571 	if (ret) {
4572 		rbd_warn(rbd_dev, "failed to reregister watch: %d", ret);
4573 		if (ret != -EBLOCKLISTED && ret != -ENOENT) {
4574 			queue_delayed_work(rbd_dev->task_wq,
4575 					   &rbd_dev->watch_dwork,
4576 					   RBD_RETRY_DELAY);
4577 			mutex_unlock(&rbd_dev->watch_mutex);
4578 			return;
4579 		}
4580 
4581 		mutex_unlock(&rbd_dev->watch_mutex);
4582 		down_write(&rbd_dev->lock_rwsem);
4583 		wake_lock_waiters(rbd_dev, ret);
4584 		up_write(&rbd_dev->lock_rwsem);
4585 		return;
4586 	}
4587 
4588 	rbd_dev->watch_state = RBD_WATCH_STATE_REGISTERED;
4589 	rbd_dev->watch_cookie = rbd_dev->watch_handle->linger_id;
4590 	mutex_unlock(&rbd_dev->watch_mutex);
4591 
4592 	down_write(&rbd_dev->lock_rwsem);
4593 	if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED)
4594 		rbd_reacquire_lock(rbd_dev);
4595 	up_write(&rbd_dev->lock_rwsem);
4596 
4597 	ret = rbd_dev_refresh(rbd_dev);
4598 	if (ret)
4599 		rbd_warn(rbd_dev, "reregistration refresh failed: %d", ret);
4600 }
4601 
4602 /*
4603  * Synchronous osd object method call.  Returns the number of bytes
4604  * returned in the outbound buffer, or a negative error code.
4605  */
4606 static int rbd_obj_method_sync(struct rbd_device *rbd_dev,
4607 			     struct ceph_object_id *oid,
4608 			     struct ceph_object_locator *oloc,
4609 			     const char *method_name,
4610 			     const void *outbound,
4611 			     size_t outbound_size,
4612 			     void *inbound,
4613 			     size_t inbound_size)
4614 {
4615 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4616 	struct page *req_page = NULL;
4617 	struct page *reply_page;
4618 	int ret;
4619 
4620 	/*
4621 	 * Method calls are ultimately read operations.  The result
4622 	 * should placed into the inbound buffer provided.  They
4623 	 * also supply outbound data--parameters for the object
4624 	 * method.  Currently if this is present it will be a
4625 	 * snapshot id.
4626 	 */
4627 	if (outbound) {
4628 		if (outbound_size > PAGE_SIZE)
4629 			return -E2BIG;
4630 
4631 		req_page = alloc_page(GFP_KERNEL);
4632 		if (!req_page)
4633 			return -ENOMEM;
4634 
4635 		memcpy(page_address(req_page), outbound, outbound_size);
4636 	}
4637 
4638 	reply_page = alloc_page(GFP_KERNEL);
4639 	if (!reply_page) {
4640 		if (req_page)
4641 			__free_page(req_page);
4642 		return -ENOMEM;
4643 	}
4644 
4645 	ret = ceph_osdc_call(osdc, oid, oloc, RBD_DRV_NAME, method_name,
4646 			     CEPH_OSD_FLAG_READ, req_page, outbound_size,
4647 			     &reply_page, &inbound_size);
4648 	if (!ret) {
4649 		memcpy(inbound, page_address(reply_page), inbound_size);
4650 		ret = inbound_size;
4651 	}
4652 
4653 	if (req_page)
4654 		__free_page(req_page);
4655 	__free_page(reply_page);
4656 	return ret;
4657 }
4658 
4659 static void rbd_queue_workfn(struct work_struct *work)
4660 {
4661 	struct rbd_img_request *img_request =
4662 	    container_of(work, struct rbd_img_request, work);
4663 	struct rbd_device *rbd_dev = img_request->rbd_dev;
4664 	enum obj_operation_type op_type = img_request->op_type;
4665 	struct request *rq = blk_mq_rq_from_pdu(img_request);
4666 	u64 offset = (u64)blk_rq_pos(rq) << SECTOR_SHIFT;
4667 	u64 length = blk_rq_bytes(rq);
4668 	u64 mapping_size;
4669 	int result;
4670 
4671 	/* Ignore/skip any zero-length requests */
4672 	if (!length) {
4673 		dout("%s: zero-length request\n", __func__);
4674 		result = 0;
4675 		goto err_img_request;
4676 	}
4677 
4678 	blk_mq_start_request(rq);
4679 
4680 	down_read(&rbd_dev->header_rwsem);
4681 	mapping_size = rbd_dev->mapping.size;
4682 	rbd_img_capture_header(img_request);
4683 	up_read(&rbd_dev->header_rwsem);
4684 
4685 	if (offset + length > mapping_size) {
4686 		rbd_warn(rbd_dev, "beyond EOD (%llu~%llu > %llu)", offset,
4687 			 length, mapping_size);
4688 		result = -EIO;
4689 		goto err_img_request;
4690 	}
4691 
4692 	dout("%s rbd_dev %p img_req %p %s %llu~%llu\n", __func__, rbd_dev,
4693 	     img_request, obj_op_name(op_type), offset, length);
4694 
4695 	if (op_type == OBJ_OP_DISCARD || op_type == OBJ_OP_ZEROOUT)
4696 		result = rbd_img_fill_nodata(img_request, offset, length);
4697 	else
4698 		result = rbd_img_fill_from_bio(img_request, offset, length,
4699 					       rq->bio);
4700 	if (result)
4701 		goto err_img_request;
4702 
4703 	rbd_img_handle_request(img_request, 0);
4704 	return;
4705 
4706 err_img_request:
4707 	rbd_img_request_destroy(img_request);
4708 	if (result)
4709 		rbd_warn(rbd_dev, "%s %llx at %llx result %d",
4710 			 obj_op_name(op_type), length, offset, result);
4711 	blk_mq_end_request(rq, errno_to_blk_status(result));
4712 }
4713 
4714 static blk_status_t rbd_queue_rq(struct blk_mq_hw_ctx *hctx,
4715 		const struct blk_mq_queue_data *bd)
4716 {
4717 	struct rbd_device *rbd_dev = hctx->queue->queuedata;
4718 	struct rbd_img_request *img_req = blk_mq_rq_to_pdu(bd->rq);
4719 	enum obj_operation_type op_type;
4720 
4721 	switch (req_op(bd->rq)) {
4722 	case REQ_OP_DISCARD:
4723 		op_type = OBJ_OP_DISCARD;
4724 		break;
4725 	case REQ_OP_WRITE_ZEROES:
4726 		op_type = OBJ_OP_ZEROOUT;
4727 		break;
4728 	case REQ_OP_WRITE:
4729 		op_type = OBJ_OP_WRITE;
4730 		break;
4731 	case REQ_OP_READ:
4732 		op_type = OBJ_OP_READ;
4733 		break;
4734 	default:
4735 		rbd_warn(rbd_dev, "unknown req_op %d", req_op(bd->rq));
4736 		return BLK_STS_IOERR;
4737 	}
4738 
4739 	rbd_img_request_init(img_req, rbd_dev, op_type);
4740 
4741 	if (rbd_img_is_write(img_req)) {
4742 		if (rbd_is_ro(rbd_dev)) {
4743 			rbd_warn(rbd_dev, "%s on read-only mapping",
4744 				 obj_op_name(img_req->op_type));
4745 			return BLK_STS_IOERR;
4746 		}
4747 		rbd_assert(!rbd_is_snap(rbd_dev));
4748 	}
4749 
4750 	INIT_WORK(&img_req->work, rbd_queue_workfn);
4751 	queue_work(rbd_wq, &img_req->work);
4752 	return BLK_STS_OK;
4753 }
4754 
4755 static void rbd_free_disk(struct rbd_device *rbd_dev)
4756 {
4757 	put_disk(rbd_dev->disk);
4758 	blk_mq_free_tag_set(&rbd_dev->tag_set);
4759 	rbd_dev->disk = NULL;
4760 }
4761 
4762 static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
4763 			     struct ceph_object_id *oid,
4764 			     struct ceph_object_locator *oloc,
4765 			     void *buf, int buf_len)
4766 
4767 {
4768 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
4769 	struct ceph_osd_request *req;
4770 	struct page **pages;
4771 	int num_pages = calc_pages_for(0, buf_len);
4772 	int ret;
4773 
4774 	req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_KERNEL);
4775 	if (!req)
4776 		return -ENOMEM;
4777 
4778 	ceph_oid_copy(&req->r_base_oid, oid);
4779 	ceph_oloc_copy(&req->r_base_oloc, oloc);
4780 	req->r_flags = CEPH_OSD_FLAG_READ;
4781 
4782 	pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
4783 	if (IS_ERR(pages)) {
4784 		ret = PTR_ERR(pages);
4785 		goto out_req;
4786 	}
4787 
4788 	osd_req_op_extent_init(req, 0, CEPH_OSD_OP_READ, 0, buf_len, 0, 0);
4789 	osd_req_op_extent_osd_data_pages(req, 0, pages, buf_len, 0, false,
4790 					 true);
4791 
4792 	ret = ceph_osdc_alloc_messages(req, GFP_KERNEL);
4793 	if (ret)
4794 		goto out_req;
4795 
4796 	ceph_osdc_start_request(osdc, req);
4797 	ret = ceph_osdc_wait_request(osdc, req);
4798 	if (ret >= 0)
4799 		ceph_copy_from_page_vector(pages, buf, 0, ret);
4800 
4801 out_req:
4802 	ceph_osdc_put_request(req);
4803 	return ret;
4804 }
4805 
4806 /*
4807  * Read the complete header for the given rbd device.  On successful
4808  * return, the rbd_dev->header field will contain up-to-date
4809  * information about the image.
4810  */
4811 static int rbd_dev_v1_header_info(struct rbd_device *rbd_dev)
4812 {
4813 	struct rbd_image_header_ondisk *ondisk = NULL;
4814 	u32 snap_count = 0;
4815 	u64 names_size = 0;
4816 	u32 want_count;
4817 	int ret;
4818 
4819 	/*
4820 	 * The complete header will include an array of its 64-bit
4821 	 * snapshot ids, followed by the names of those snapshots as
4822 	 * a contiguous block of NUL-terminated strings.  Note that
4823 	 * the number of snapshots could change by the time we read
4824 	 * it in, in which case we re-read it.
4825 	 */
4826 	do {
4827 		size_t size;
4828 
4829 		kfree(ondisk);
4830 
4831 		size = sizeof (*ondisk);
4832 		size += snap_count * sizeof (struct rbd_image_snap_ondisk);
4833 		size += names_size;
4834 		ondisk = kmalloc(size, GFP_KERNEL);
4835 		if (!ondisk)
4836 			return -ENOMEM;
4837 
4838 		ret = rbd_obj_read_sync(rbd_dev, &rbd_dev->header_oid,
4839 					&rbd_dev->header_oloc, ondisk, size);
4840 		if (ret < 0)
4841 			goto out;
4842 		if ((size_t)ret < size) {
4843 			ret = -ENXIO;
4844 			rbd_warn(rbd_dev, "short header read (want %zd got %d)",
4845 				size, ret);
4846 			goto out;
4847 		}
4848 		if (!rbd_dev_ondisk_valid(ondisk)) {
4849 			ret = -ENXIO;
4850 			rbd_warn(rbd_dev, "invalid header");
4851 			goto out;
4852 		}
4853 
4854 		names_size = le64_to_cpu(ondisk->snap_names_len);
4855 		want_count = snap_count;
4856 		snap_count = le32_to_cpu(ondisk->snap_count);
4857 	} while (snap_count != want_count);
4858 
4859 	ret = rbd_header_from_disk(rbd_dev, ondisk);
4860 out:
4861 	kfree(ondisk);
4862 
4863 	return ret;
4864 }
4865 
4866 static void rbd_dev_update_size(struct rbd_device *rbd_dev)
4867 {
4868 	sector_t size;
4869 
4870 	/*
4871 	 * If EXISTS is not set, rbd_dev->disk may be NULL, so don't
4872 	 * try to update its size.  If REMOVING is set, updating size
4873 	 * is just useless work since the device can't be opened.
4874 	 */
4875 	if (test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags) &&
4876 	    !test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags)) {
4877 		size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE;
4878 		dout("setting size to %llu sectors", (unsigned long long)size);
4879 		set_capacity_and_notify(rbd_dev->disk, size);
4880 	}
4881 }
4882 
4883 static int rbd_dev_refresh(struct rbd_device *rbd_dev)
4884 {
4885 	u64 mapping_size;
4886 	int ret;
4887 
4888 	down_write(&rbd_dev->header_rwsem);
4889 	mapping_size = rbd_dev->mapping.size;
4890 
4891 	ret = rbd_dev_header_info(rbd_dev);
4892 	if (ret)
4893 		goto out;
4894 
4895 	/*
4896 	 * If there is a parent, see if it has disappeared due to the
4897 	 * mapped image getting flattened.
4898 	 */
4899 	if (rbd_dev->parent) {
4900 		ret = rbd_dev_v2_parent_info(rbd_dev);
4901 		if (ret)
4902 			goto out;
4903 	}
4904 
4905 	rbd_assert(!rbd_is_snap(rbd_dev));
4906 	rbd_dev->mapping.size = rbd_dev->header.image_size;
4907 
4908 out:
4909 	up_write(&rbd_dev->header_rwsem);
4910 	if (!ret && mapping_size != rbd_dev->mapping.size)
4911 		rbd_dev_update_size(rbd_dev);
4912 
4913 	return ret;
4914 }
4915 
4916 static const struct blk_mq_ops rbd_mq_ops = {
4917 	.queue_rq	= rbd_queue_rq,
4918 };
4919 
4920 static int rbd_init_disk(struct rbd_device *rbd_dev)
4921 {
4922 	struct gendisk *disk;
4923 	struct request_queue *q;
4924 	unsigned int objset_bytes =
4925 	    rbd_dev->layout.object_size * rbd_dev->layout.stripe_count;
4926 	int err;
4927 
4928 	memset(&rbd_dev->tag_set, 0, sizeof(rbd_dev->tag_set));
4929 	rbd_dev->tag_set.ops = &rbd_mq_ops;
4930 	rbd_dev->tag_set.queue_depth = rbd_dev->opts->queue_depth;
4931 	rbd_dev->tag_set.numa_node = NUMA_NO_NODE;
4932 	rbd_dev->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
4933 	rbd_dev->tag_set.nr_hw_queues = num_present_cpus();
4934 	rbd_dev->tag_set.cmd_size = sizeof(struct rbd_img_request);
4935 
4936 	err = blk_mq_alloc_tag_set(&rbd_dev->tag_set);
4937 	if (err)
4938 		return err;
4939 
4940 	disk = blk_mq_alloc_disk(&rbd_dev->tag_set, rbd_dev);
4941 	if (IS_ERR(disk)) {
4942 		err = PTR_ERR(disk);
4943 		goto out_tag_set;
4944 	}
4945 	q = disk->queue;
4946 
4947 	snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
4948 		 rbd_dev->dev_id);
4949 	disk->major = rbd_dev->major;
4950 	disk->first_minor = rbd_dev->minor;
4951 	if (single_major)
4952 		disk->minors = (1 << RBD_SINGLE_MAJOR_PART_SHIFT);
4953 	else
4954 		disk->minors = RBD_MINORS_PER_MAJOR;
4955 	disk->fops = &rbd_bd_ops;
4956 	disk->private_data = rbd_dev;
4957 
4958 	blk_queue_flag_set(QUEUE_FLAG_NONROT, q);
4959 	/* QUEUE_FLAG_ADD_RANDOM is off by default for blk-mq */
4960 
4961 	blk_queue_max_hw_sectors(q, objset_bytes >> SECTOR_SHIFT);
4962 	q->limits.max_sectors = queue_max_hw_sectors(q);
4963 	blk_queue_max_segments(q, USHRT_MAX);
4964 	blk_queue_max_segment_size(q, UINT_MAX);
4965 	blk_queue_io_min(q, rbd_dev->opts->alloc_size);
4966 	blk_queue_io_opt(q, rbd_dev->opts->alloc_size);
4967 
4968 	if (rbd_dev->opts->trim) {
4969 		q->limits.discard_granularity = rbd_dev->opts->alloc_size;
4970 		blk_queue_max_discard_sectors(q, objset_bytes >> SECTOR_SHIFT);
4971 		blk_queue_max_write_zeroes_sectors(q, objset_bytes >> SECTOR_SHIFT);
4972 	}
4973 
4974 	if (!ceph_test_opt(rbd_dev->rbd_client->client, NOCRC))
4975 		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, q);
4976 
4977 	rbd_dev->disk = disk;
4978 
4979 	return 0;
4980 out_tag_set:
4981 	blk_mq_free_tag_set(&rbd_dev->tag_set);
4982 	return err;
4983 }
4984 
4985 /*
4986   sysfs
4987 */
4988 
4989 static struct rbd_device *dev_to_rbd_dev(struct device *dev)
4990 {
4991 	return container_of(dev, struct rbd_device, dev);
4992 }
4993 
4994 static ssize_t rbd_size_show(struct device *dev,
4995 			     struct device_attribute *attr, char *buf)
4996 {
4997 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
4998 
4999 	return sprintf(buf, "%llu\n",
5000 		(unsigned long long)rbd_dev->mapping.size);
5001 }
5002 
5003 static ssize_t rbd_features_show(struct device *dev,
5004 			     struct device_attribute *attr, char *buf)
5005 {
5006 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5007 
5008 	return sprintf(buf, "0x%016llx\n", rbd_dev->header.features);
5009 }
5010 
5011 static ssize_t rbd_major_show(struct device *dev,
5012 			      struct device_attribute *attr, char *buf)
5013 {
5014 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5015 
5016 	if (rbd_dev->major)
5017 		return sprintf(buf, "%d\n", rbd_dev->major);
5018 
5019 	return sprintf(buf, "(none)\n");
5020 }
5021 
5022 static ssize_t rbd_minor_show(struct device *dev,
5023 			      struct device_attribute *attr, char *buf)
5024 {
5025 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5026 
5027 	return sprintf(buf, "%d\n", rbd_dev->minor);
5028 }
5029 
5030 static ssize_t rbd_client_addr_show(struct device *dev,
5031 				    struct device_attribute *attr, char *buf)
5032 {
5033 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5034 	struct ceph_entity_addr *client_addr =
5035 	    ceph_client_addr(rbd_dev->rbd_client->client);
5036 
5037 	return sprintf(buf, "%pISpc/%u\n", &client_addr->in_addr,
5038 		       le32_to_cpu(client_addr->nonce));
5039 }
5040 
5041 static ssize_t rbd_client_id_show(struct device *dev,
5042 				  struct device_attribute *attr, char *buf)
5043 {
5044 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5045 
5046 	return sprintf(buf, "client%lld\n",
5047 		       ceph_client_gid(rbd_dev->rbd_client->client));
5048 }
5049 
5050 static ssize_t rbd_cluster_fsid_show(struct device *dev,
5051 				     struct device_attribute *attr, char *buf)
5052 {
5053 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5054 
5055 	return sprintf(buf, "%pU\n", &rbd_dev->rbd_client->client->fsid);
5056 }
5057 
5058 static ssize_t rbd_config_info_show(struct device *dev,
5059 				    struct device_attribute *attr, char *buf)
5060 {
5061 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5062 
5063 	if (!capable(CAP_SYS_ADMIN))
5064 		return -EPERM;
5065 
5066 	return sprintf(buf, "%s\n", rbd_dev->config_info);
5067 }
5068 
5069 static ssize_t rbd_pool_show(struct device *dev,
5070 			     struct device_attribute *attr, char *buf)
5071 {
5072 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5073 
5074 	return sprintf(buf, "%s\n", rbd_dev->spec->pool_name);
5075 }
5076 
5077 static ssize_t rbd_pool_id_show(struct device *dev,
5078 			     struct device_attribute *attr, char *buf)
5079 {
5080 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5081 
5082 	return sprintf(buf, "%llu\n",
5083 			(unsigned long long) rbd_dev->spec->pool_id);
5084 }
5085 
5086 static ssize_t rbd_pool_ns_show(struct device *dev,
5087 				struct device_attribute *attr, char *buf)
5088 {
5089 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5090 
5091 	return sprintf(buf, "%s\n", rbd_dev->spec->pool_ns ?: "");
5092 }
5093 
5094 static ssize_t rbd_name_show(struct device *dev,
5095 			     struct device_attribute *attr, char *buf)
5096 {
5097 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5098 
5099 	if (rbd_dev->spec->image_name)
5100 		return sprintf(buf, "%s\n", rbd_dev->spec->image_name);
5101 
5102 	return sprintf(buf, "(unknown)\n");
5103 }
5104 
5105 static ssize_t rbd_image_id_show(struct device *dev,
5106 			     struct device_attribute *attr, char *buf)
5107 {
5108 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5109 
5110 	return sprintf(buf, "%s\n", rbd_dev->spec->image_id);
5111 }
5112 
5113 /*
5114  * Shows the name of the currently-mapped snapshot (or
5115  * RBD_SNAP_HEAD_NAME for the base image).
5116  */
5117 static ssize_t rbd_snap_show(struct device *dev,
5118 			     struct device_attribute *attr,
5119 			     char *buf)
5120 {
5121 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5122 
5123 	return sprintf(buf, "%s\n", rbd_dev->spec->snap_name);
5124 }
5125 
5126 static ssize_t rbd_snap_id_show(struct device *dev,
5127 				struct device_attribute *attr, char *buf)
5128 {
5129 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5130 
5131 	return sprintf(buf, "%llu\n", rbd_dev->spec->snap_id);
5132 }
5133 
5134 /*
5135  * For a v2 image, shows the chain of parent images, separated by empty
5136  * lines.  For v1 images or if there is no parent, shows "(no parent
5137  * image)".
5138  */
5139 static ssize_t rbd_parent_show(struct device *dev,
5140 			       struct device_attribute *attr,
5141 			       char *buf)
5142 {
5143 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5144 	ssize_t count = 0;
5145 
5146 	if (!rbd_dev->parent)
5147 		return sprintf(buf, "(no parent image)\n");
5148 
5149 	for ( ; rbd_dev->parent; rbd_dev = rbd_dev->parent) {
5150 		struct rbd_spec *spec = rbd_dev->parent_spec;
5151 
5152 		count += sprintf(&buf[count], "%s"
5153 			    "pool_id %llu\npool_name %s\n"
5154 			    "pool_ns %s\n"
5155 			    "image_id %s\nimage_name %s\n"
5156 			    "snap_id %llu\nsnap_name %s\n"
5157 			    "overlap %llu\n",
5158 			    !count ? "" : "\n", /* first? */
5159 			    spec->pool_id, spec->pool_name,
5160 			    spec->pool_ns ?: "",
5161 			    spec->image_id, spec->image_name ?: "(unknown)",
5162 			    spec->snap_id, spec->snap_name,
5163 			    rbd_dev->parent_overlap);
5164 	}
5165 
5166 	return count;
5167 }
5168 
5169 static ssize_t rbd_image_refresh(struct device *dev,
5170 				 struct device_attribute *attr,
5171 				 const char *buf,
5172 				 size_t size)
5173 {
5174 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5175 	int ret;
5176 
5177 	if (!capable(CAP_SYS_ADMIN))
5178 		return -EPERM;
5179 
5180 	ret = rbd_dev_refresh(rbd_dev);
5181 	if (ret)
5182 		return ret;
5183 
5184 	return size;
5185 }
5186 
5187 static DEVICE_ATTR(size, 0444, rbd_size_show, NULL);
5188 static DEVICE_ATTR(features, 0444, rbd_features_show, NULL);
5189 static DEVICE_ATTR(major, 0444, rbd_major_show, NULL);
5190 static DEVICE_ATTR(minor, 0444, rbd_minor_show, NULL);
5191 static DEVICE_ATTR(client_addr, 0444, rbd_client_addr_show, NULL);
5192 static DEVICE_ATTR(client_id, 0444, rbd_client_id_show, NULL);
5193 static DEVICE_ATTR(cluster_fsid, 0444, rbd_cluster_fsid_show, NULL);
5194 static DEVICE_ATTR(config_info, 0400, rbd_config_info_show, NULL);
5195 static DEVICE_ATTR(pool, 0444, rbd_pool_show, NULL);
5196 static DEVICE_ATTR(pool_id, 0444, rbd_pool_id_show, NULL);
5197 static DEVICE_ATTR(pool_ns, 0444, rbd_pool_ns_show, NULL);
5198 static DEVICE_ATTR(name, 0444, rbd_name_show, NULL);
5199 static DEVICE_ATTR(image_id, 0444, rbd_image_id_show, NULL);
5200 static DEVICE_ATTR(refresh, 0200, NULL, rbd_image_refresh);
5201 static DEVICE_ATTR(current_snap, 0444, rbd_snap_show, NULL);
5202 static DEVICE_ATTR(snap_id, 0444, rbd_snap_id_show, NULL);
5203 static DEVICE_ATTR(parent, 0444, rbd_parent_show, NULL);
5204 
5205 static struct attribute *rbd_attrs[] = {
5206 	&dev_attr_size.attr,
5207 	&dev_attr_features.attr,
5208 	&dev_attr_major.attr,
5209 	&dev_attr_minor.attr,
5210 	&dev_attr_client_addr.attr,
5211 	&dev_attr_client_id.attr,
5212 	&dev_attr_cluster_fsid.attr,
5213 	&dev_attr_config_info.attr,
5214 	&dev_attr_pool.attr,
5215 	&dev_attr_pool_id.attr,
5216 	&dev_attr_pool_ns.attr,
5217 	&dev_attr_name.attr,
5218 	&dev_attr_image_id.attr,
5219 	&dev_attr_current_snap.attr,
5220 	&dev_attr_snap_id.attr,
5221 	&dev_attr_parent.attr,
5222 	&dev_attr_refresh.attr,
5223 	NULL
5224 };
5225 
5226 static struct attribute_group rbd_attr_group = {
5227 	.attrs = rbd_attrs,
5228 };
5229 
5230 static const struct attribute_group *rbd_attr_groups[] = {
5231 	&rbd_attr_group,
5232 	NULL
5233 };
5234 
5235 static void rbd_dev_release(struct device *dev);
5236 
5237 static const struct device_type rbd_device_type = {
5238 	.name		= "rbd",
5239 	.groups		= rbd_attr_groups,
5240 	.release	= rbd_dev_release,
5241 };
5242 
5243 static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec)
5244 {
5245 	kref_get(&spec->kref);
5246 
5247 	return spec;
5248 }
5249 
5250 static void rbd_spec_free(struct kref *kref);
5251 static void rbd_spec_put(struct rbd_spec *spec)
5252 {
5253 	if (spec)
5254 		kref_put(&spec->kref, rbd_spec_free);
5255 }
5256 
5257 static struct rbd_spec *rbd_spec_alloc(void)
5258 {
5259 	struct rbd_spec *spec;
5260 
5261 	spec = kzalloc(sizeof (*spec), GFP_KERNEL);
5262 	if (!spec)
5263 		return NULL;
5264 
5265 	spec->pool_id = CEPH_NOPOOL;
5266 	spec->snap_id = CEPH_NOSNAP;
5267 	kref_init(&spec->kref);
5268 
5269 	return spec;
5270 }
5271 
5272 static void rbd_spec_free(struct kref *kref)
5273 {
5274 	struct rbd_spec *spec = container_of(kref, struct rbd_spec, kref);
5275 
5276 	kfree(spec->pool_name);
5277 	kfree(spec->pool_ns);
5278 	kfree(spec->image_id);
5279 	kfree(spec->image_name);
5280 	kfree(spec->snap_name);
5281 	kfree(spec);
5282 }
5283 
5284 static void rbd_dev_free(struct rbd_device *rbd_dev)
5285 {
5286 	WARN_ON(rbd_dev->watch_state != RBD_WATCH_STATE_UNREGISTERED);
5287 	WARN_ON(rbd_dev->lock_state != RBD_LOCK_STATE_UNLOCKED);
5288 
5289 	ceph_oid_destroy(&rbd_dev->header_oid);
5290 	ceph_oloc_destroy(&rbd_dev->header_oloc);
5291 	kfree(rbd_dev->config_info);
5292 
5293 	rbd_put_client(rbd_dev->rbd_client);
5294 	rbd_spec_put(rbd_dev->spec);
5295 	kfree(rbd_dev->opts);
5296 	kfree(rbd_dev);
5297 }
5298 
5299 static void rbd_dev_release(struct device *dev)
5300 {
5301 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
5302 	bool need_put = !!rbd_dev->opts;
5303 
5304 	if (need_put) {
5305 		destroy_workqueue(rbd_dev->task_wq);
5306 		ida_simple_remove(&rbd_dev_id_ida, rbd_dev->dev_id);
5307 	}
5308 
5309 	rbd_dev_free(rbd_dev);
5310 
5311 	/*
5312 	 * This is racy, but way better than putting module outside of
5313 	 * the release callback.  The race window is pretty small, so
5314 	 * doing something similar to dm (dm-builtin.c) is overkill.
5315 	 */
5316 	if (need_put)
5317 		module_put(THIS_MODULE);
5318 }
5319 
5320 static struct rbd_device *__rbd_dev_create(struct rbd_spec *spec)
5321 {
5322 	struct rbd_device *rbd_dev;
5323 
5324 	rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL);
5325 	if (!rbd_dev)
5326 		return NULL;
5327 
5328 	spin_lock_init(&rbd_dev->lock);
5329 	INIT_LIST_HEAD(&rbd_dev->node);
5330 	init_rwsem(&rbd_dev->header_rwsem);
5331 
5332 	rbd_dev->header.data_pool_id = CEPH_NOPOOL;
5333 	ceph_oid_init(&rbd_dev->header_oid);
5334 	rbd_dev->header_oloc.pool = spec->pool_id;
5335 	if (spec->pool_ns) {
5336 		WARN_ON(!*spec->pool_ns);
5337 		rbd_dev->header_oloc.pool_ns =
5338 		    ceph_find_or_create_string(spec->pool_ns,
5339 					       strlen(spec->pool_ns));
5340 	}
5341 
5342 	mutex_init(&rbd_dev->watch_mutex);
5343 	rbd_dev->watch_state = RBD_WATCH_STATE_UNREGISTERED;
5344 	INIT_DELAYED_WORK(&rbd_dev->watch_dwork, rbd_reregister_watch);
5345 
5346 	init_rwsem(&rbd_dev->lock_rwsem);
5347 	rbd_dev->lock_state = RBD_LOCK_STATE_UNLOCKED;
5348 	INIT_WORK(&rbd_dev->acquired_lock_work, rbd_notify_acquired_lock);
5349 	INIT_WORK(&rbd_dev->released_lock_work, rbd_notify_released_lock);
5350 	INIT_DELAYED_WORK(&rbd_dev->lock_dwork, rbd_acquire_lock);
5351 	INIT_WORK(&rbd_dev->unlock_work, rbd_release_lock_work);
5352 	spin_lock_init(&rbd_dev->lock_lists_lock);
5353 	INIT_LIST_HEAD(&rbd_dev->acquiring_list);
5354 	INIT_LIST_HEAD(&rbd_dev->running_list);
5355 	init_completion(&rbd_dev->acquire_wait);
5356 	init_completion(&rbd_dev->releasing_wait);
5357 
5358 	spin_lock_init(&rbd_dev->object_map_lock);
5359 
5360 	rbd_dev->dev.bus = &rbd_bus_type;
5361 	rbd_dev->dev.type = &rbd_device_type;
5362 	rbd_dev->dev.parent = &rbd_root_dev;
5363 	device_initialize(&rbd_dev->dev);
5364 
5365 	return rbd_dev;
5366 }
5367 
5368 /*
5369  * Create a mapping rbd_dev.
5370  */
5371 static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc,
5372 					 struct rbd_spec *spec,
5373 					 struct rbd_options *opts)
5374 {
5375 	struct rbd_device *rbd_dev;
5376 
5377 	rbd_dev = __rbd_dev_create(spec);
5378 	if (!rbd_dev)
5379 		return NULL;
5380 
5381 	/* get an id and fill in device name */
5382 	rbd_dev->dev_id = ida_simple_get(&rbd_dev_id_ida, 0,
5383 					 minor_to_rbd_dev_id(1 << MINORBITS),
5384 					 GFP_KERNEL);
5385 	if (rbd_dev->dev_id < 0)
5386 		goto fail_rbd_dev;
5387 
5388 	sprintf(rbd_dev->name, RBD_DRV_NAME "%d", rbd_dev->dev_id);
5389 	rbd_dev->task_wq = alloc_ordered_workqueue("%s-tasks", WQ_MEM_RECLAIM,
5390 						   rbd_dev->name);
5391 	if (!rbd_dev->task_wq)
5392 		goto fail_dev_id;
5393 
5394 	/* we have a ref from do_rbd_add() */
5395 	__module_get(THIS_MODULE);
5396 
5397 	rbd_dev->rbd_client = rbdc;
5398 	rbd_dev->spec = spec;
5399 	rbd_dev->opts = opts;
5400 
5401 	dout("%s rbd_dev %p dev_id %d\n", __func__, rbd_dev, rbd_dev->dev_id);
5402 	return rbd_dev;
5403 
5404 fail_dev_id:
5405 	ida_simple_remove(&rbd_dev_id_ida, rbd_dev->dev_id);
5406 fail_rbd_dev:
5407 	rbd_dev_free(rbd_dev);
5408 	return NULL;
5409 }
5410 
5411 static void rbd_dev_destroy(struct rbd_device *rbd_dev)
5412 {
5413 	if (rbd_dev)
5414 		put_device(&rbd_dev->dev);
5415 }
5416 
5417 /*
5418  * Get the size and object order for an image snapshot, or if
5419  * snap_id is CEPH_NOSNAP, gets this information for the base
5420  * image.
5421  */
5422 static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
5423 				u8 *order, u64 *snap_size)
5424 {
5425 	__le64 snapid = cpu_to_le64(snap_id);
5426 	int ret;
5427 	struct {
5428 		u8 order;
5429 		__le64 size;
5430 	} __attribute__ ((packed)) size_buf = { 0 };
5431 
5432 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5433 				  &rbd_dev->header_oloc, "get_size",
5434 				  &snapid, sizeof(snapid),
5435 				  &size_buf, sizeof(size_buf));
5436 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
5437 	if (ret < 0)
5438 		return ret;
5439 	if (ret < sizeof (size_buf))
5440 		return -ERANGE;
5441 
5442 	if (order) {
5443 		*order = size_buf.order;
5444 		dout("  order %u", (unsigned int)*order);
5445 	}
5446 	*snap_size = le64_to_cpu(size_buf.size);
5447 
5448 	dout("  snap_id 0x%016llx snap_size = %llu\n",
5449 		(unsigned long long)snap_id,
5450 		(unsigned long long)*snap_size);
5451 
5452 	return 0;
5453 }
5454 
5455 static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
5456 {
5457 	return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
5458 					&rbd_dev->header.obj_order,
5459 					&rbd_dev->header.image_size);
5460 }
5461 
5462 static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
5463 {
5464 	size_t size;
5465 	void *reply_buf;
5466 	int ret;
5467 	void *p;
5468 
5469 	/* Response will be an encoded string, which includes a length */
5470 	size = sizeof(__le32) + RBD_OBJ_PREFIX_LEN_MAX;
5471 	reply_buf = kzalloc(size, GFP_KERNEL);
5472 	if (!reply_buf)
5473 		return -ENOMEM;
5474 
5475 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5476 				  &rbd_dev->header_oloc, "get_object_prefix",
5477 				  NULL, 0, reply_buf, size);
5478 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
5479 	if (ret < 0)
5480 		goto out;
5481 
5482 	p = reply_buf;
5483 	rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
5484 						p + ret, NULL, GFP_NOIO);
5485 	ret = 0;
5486 
5487 	if (IS_ERR(rbd_dev->header.object_prefix)) {
5488 		ret = PTR_ERR(rbd_dev->header.object_prefix);
5489 		rbd_dev->header.object_prefix = NULL;
5490 	} else {
5491 		dout("  object_prefix = %s\n", rbd_dev->header.object_prefix);
5492 	}
5493 out:
5494 	kfree(reply_buf);
5495 
5496 	return ret;
5497 }
5498 
5499 static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
5500 				     bool read_only, u64 *snap_features)
5501 {
5502 	struct {
5503 		__le64 snap_id;
5504 		u8 read_only;
5505 	} features_in;
5506 	struct {
5507 		__le64 features;
5508 		__le64 incompat;
5509 	} __attribute__ ((packed)) features_buf = { 0 };
5510 	u64 unsup;
5511 	int ret;
5512 
5513 	features_in.snap_id = cpu_to_le64(snap_id);
5514 	features_in.read_only = read_only;
5515 
5516 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5517 				  &rbd_dev->header_oloc, "get_features",
5518 				  &features_in, sizeof(features_in),
5519 				  &features_buf, sizeof(features_buf));
5520 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
5521 	if (ret < 0)
5522 		return ret;
5523 	if (ret < sizeof (features_buf))
5524 		return -ERANGE;
5525 
5526 	unsup = le64_to_cpu(features_buf.incompat) & ~RBD_FEATURES_SUPPORTED;
5527 	if (unsup) {
5528 		rbd_warn(rbd_dev, "image uses unsupported features: 0x%llx",
5529 			 unsup);
5530 		return -ENXIO;
5531 	}
5532 
5533 	*snap_features = le64_to_cpu(features_buf.features);
5534 
5535 	dout("  snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
5536 		(unsigned long long)snap_id,
5537 		(unsigned long long)*snap_features,
5538 		(unsigned long long)le64_to_cpu(features_buf.incompat));
5539 
5540 	return 0;
5541 }
5542 
5543 static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
5544 {
5545 	return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
5546 					 rbd_is_ro(rbd_dev),
5547 					 &rbd_dev->header.features);
5548 }
5549 
5550 /*
5551  * These are generic image flags, but since they are used only for
5552  * object map, store them in rbd_dev->object_map_flags.
5553  *
5554  * For the same reason, this function is called only on object map
5555  * (re)load and not on header refresh.
5556  */
5557 static int rbd_dev_v2_get_flags(struct rbd_device *rbd_dev)
5558 {
5559 	__le64 snapid = cpu_to_le64(rbd_dev->spec->snap_id);
5560 	__le64 flags;
5561 	int ret;
5562 
5563 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5564 				  &rbd_dev->header_oloc, "get_flags",
5565 				  &snapid, sizeof(snapid),
5566 				  &flags, sizeof(flags));
5567 	if (ret < 0)
5568 		return ret;
5569 	if (ret < sizeof(flags))
5570 		return -EBADMSG;
5571 
5572 	rbd_dev->object_map_flags = le64_to_cpu(flags);
5573 	return 0;
5574 }
5575 
5576 struct parent_image_info {
5577 	u64		pool_id;
5578 	const char	*pool_ns;
5579 	const char	*image_id;
5580 	u64		snap_id;
5581 
5582 	bool		has_overlap;
5583 	u64		overlap;
5584 };
5585 
5586 /*
5587  * The caller is responsible for @pii.
5588  */
5589 static int decode_parent_image_spec(void **p, void *end,
5590 				    struct parent_image_info *pii)
5591 {
5592 	u8 struct_v;
5593 	u32 struct_len;
5594 	int ret;
5595 
5596 	ret = ceph_start_decoding(p, end, 1, "ParentImageSpec",
5597 				  &struct_v, &struct_len);
5598 	if (ret)
5599 		return ret;
5600 
5601 	ceph_decode_64_safe(p, end, pii->pool_id, e_inval);
5602 	pii->pool_ns = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL);
5603 	if (IS_ERR(pii->pool_ns)) {
5604 		ret = PTR_ERR(pii->pool_ns);
5605 		pii->pool_ns = NULL;
5606 		return ret;
5607 	}
5608 	pii->image_id = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL);
5609 	if (IS_ERR(pii->image_id)) {
5610 		ret = PTR_ERR(pii->image_id);
5611 		pii->image_id = NULL;
5612 		return ret;
5613 	}
5614 	ceph_decode_64_safe(p, end, pii->snap_id, e_inval);
5615 	return 0;
5616 
5617 e_inval:
5618 	return -EINVAL;
5619 }
5620 
5621 static int __get_parent_info(struct rbd_device *rbd_dev,
5622 			     struct page *req_page,
5623 			     struct page *reply_page,
5624 			     struct parent_image_info *pii)
5625 {
5626 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
5627 	size_t reply_len = PAGE_SIZE;
5628 	void *p, *end;
5629 	int ret;
5630 
5631 	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
5632 			     "rbd", "parent_get", CEPH_OSD_FLAG_READ,
5633 			     req_page, sizeof(u64), &reply_page, &reply_len);
5634 	if (ret)
5635 		return ret == -EOPNOTSUPP ? 1 : ret;
5636 
5637 	p = page_address(reply_page);
5638 	end = p + reply_len;
5639 	ret = decode_parent_image_spec(&p, end, pii);
5640 	if (ret)
5641 		return ret;
5642 
5643 	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
5644 			     "rbd", "parent_overlap_get", CEPH_OSD_FLAG_READ,
5645 			     req_page, sizeof(u64), &reply_page, &reply_len);
5646 	if (ret)
5647 		return ret;
5648 
5649 	p = page_address(reply_page);
5650 	end = p + reply_len;
5651 	ceph_decode_8_safe(&p, end, pii->has_overlap, e_inval);
5652 	if (pii->has_overlap)
5653 		ceph_decode_64_safe(&p, end, pii->overlap, e_inval);
5654 
5655 	return 0;
5656 
5657 e_inval:
5658 	return -EINVAL;
5659 }
5660 
5661 /*
5662  * The caller is responsible for @pii.
5663  */
5664 static int __get_parent_info_legacy(struct rbd_device *rbd_dev,
5665 				    struct page *req_page,
5666 				    struct page *reply_page,
5667 				    struct parent_image_info *pii)
5668 {
5669 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
5670 	size_t reply_len = PAGE_SIZE;
5671 	void *p, *end;
5672 	int ret;
5673 
5674 	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
5675 			     "rbd", "get_parent", CEPH_OSD_FLAG_READ,
5676 			     req_page, sizeof(u64), &reply_page, &reply_len);
5677 	if (ret)
5678 		return ret;
5679 
5680 	p = page_address(reply_page);
5681 	end = p + reply_len;
5682 	ceph_decode_64_safe(&p, end, pii->pool_id, e_inval);
5683 	pii->image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
5684 	if (IS_ERR(pii->image_id)) {
5685 		ret = PTR_ERR(pii->image_id);
5686 		pii->image_id = NULL;
5687 		return ret;
5688 	}
5689 	ceph_decode_64_safe(&p, end, pii->snap_id, e_inval);
5690 	pii->has_overlap = true;
5691 	ceph_decode_64_safe(&p, end, pii->overlap, e_inval);
5692 
5693 	return 0;
5694 
5695 e_inval:
5696 	return -EINVAL;
5697 }
5698 
5699 static int get_parent_info(struct rbd_device *rbd_dev,
5700 			   struct parent_image_info *pii)
5701 {
5702 	struct page *req_page, *reply_page;
5703 	void *p;
5704 	int ret;
5705 
5706 	req_page = alloc_page(GFP_KERNEL);
5707 	if (!req_page)
5708 		return -ENOMEM;
5709 
5710 	reply_page = alloc_page(GFP_KERNEL);
5711 	if (!reply_page) {
5712 		__free_page(req_page);
5713 		return -ENOMEM;
5714 	}
5715 
5716 	p = page_address(req_page);
5717 	ceph_encode_64(&p, rbd_dev->spec->snap_id);
5718 	ret = __get_parent_info(rbd_dev, req_page, reply_page, pii);
5719 	if (ret > 0)
5720 		ret = __get_parent_info_legacy(rbd_dev, req_page, reply_page,
5721 					       pii);
5722 
5723 	__free_page(req_page);
5724 	__free_page(reply_page);
5725 	return ret;
5726 }
5727 
5728 static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
5729 {
5730 	struct rbd_spec *parent_spec;
5731 	struct parent_image_info pii = { 0 };
5732 	int ret;
5733 
5734 	parent_spec = rbd_spec_alloc();
5735 	if (!parent_spec)
5736 		return -ENOMEM;
5737 
5738 	ret = get_parent_info(rbd_dev, &pii);
5739 	if (ret)
5740 		goto out_err;
5741 
5742 	dout("%s pool_id %llu pool_ns %s image_id %s snap_id %llu has_overlap %d overlap %llu\n",
5743 	     __func__, pii.pool_id, pii.pool_ns, pii.image_id, pii.snap_id,
5744 	     pii.has_overlap, pii.overlap);
5745 
5746 	if (pii.pool_id == CEPH_NOPOOL || !pii.has_overlap) {
5747 		/*
5748 		 * Either the parent never existed, or we have
5749 		 * record of it but the image got flattened so it no
5750 		 * longer has a parent.  When the parent of a
5751 		 * layered image disappears we immediately set the
5752 		 * overlap to 0.  The effect of this is that all new
5753 		 * requests will be treated as if the image had no
5754 		 * parent.
5755 		 *
5756 		 * If !pii.has_overlap, the parent image spec is not
5757 		 * applicable.  It's there to avoid duplication in each
5758 		 * snapshot record.
5759 		 */
5760 		if (rbd_dev->parent_overlap) {
5761 			rbd_dev->parent_overlap = 0;
5762 			rbd_dev_parent_put(rbd_dev);
5763 			pr_info("%s: clone image has been flattened\n",
5764 				rbd_dev->disk->disk_name);
5765 		}
5766 
5767 		goto out;	/* No parent?  No problem. */
5768 	}
5769 
5770 	/* The ceph file layout needs to fit pool id in 32 bits */
5771 
5772 	ret = -EIO;
5773 	if (pii.pool_id > (u64)U32_MAX) {
5774 		rbd_warn(NULL, "parent pool id too large (%llu > %u)",
5775 			(unsigned long long)pii.pool_id, U32_MAX);
5776 		goto out_err;
5777 	}
5778 
5779 	/*
5780 	 * The parent won't change (except when the clone is
5781 	 * flattened, already handled that).  So we only need to
5782 	 * record the parent spec we have not already done so.
5783 	 */
5784 	if (!rbd_dev->parent_spec) {
5785 		parent_spec->pool_id = pii.pool_id;
5786 		if (pii.pool_ns && *pii.pool_ns) {
5787 			parent_spec->pool_ns = pii.pool_ns;
5788 			pii.pool_ns = NULL;
5789 		}
5790 		parent_spec->image_id = pii.image_id;
5791 		pii.image_id = NULL;
5792 		parent_spec->snap_id = pii.snap_id;
5793 
5794 		rbd_dev->parent_spec = parent_spec;
5795 		parent_spec = NULL;	/* rbd_dev now owns this */
5796 	}
5797 
5798 	/*
5799 	 * We always update the parent overlap.  If it's zero we issue
5800 	 * a warning, as we will proceed as if there was no parent.
5801 	 */
5802 	if (!pii.overlap) {
5803 		if (parent_spec) {
5804 			/* refresh, careful to warn just once */
5805 			if (rbd_dev->parent_overlap)
5806 				rbd_warn(rbd_dev,
5807 				    "clone now standalone (overlap became 0)");
5808 		} else {
5809 			/* initial probe */
5810 			rbd_warn(rbd_dev, "clone is standalone (overlap 0)");
5811 		}
5812 	}
5813 	rbd_dev->parent_overlap = pii.overlap;
5814 
5815 out:
5816 	ret = 0;
5817 out_err:
5818 	kfree(pii.pool_ns);
5819 	kfree(pii.image_id);
5820 	rbd_spec_put(parent_spec);
5821 	return ret;
5822 }
5823 
5824 static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev)
5825 {
5826 	struct {
5827 		__le64 stripe_unit;
5828 		__le64 stripe_count;
5829 	} __attribute__ ((packed)) striping_info_buf = { 0 };
5830 	size_t size = sizeof (striping_info_buf);
5831 	void *p;
5832 	int ret;
5833 
5834 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5835 				&rbd_dev->header_oloc, "get_stripe_unit_count",
5836 				NULL, 0, &striping_info_buf, size);
5837 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
5838 	if (ret < 0)
5839 		return ret;
5840 	if (ret < size)
5841 		return -ERANGE;
5842 
5843 	p = &striping_info_buf;
5844 	rbd_dev->header.stripe_unit = ceph_decode_64(&p);
5845 	rbd_dev->header.stripe_count = ceph_decode_64(&p);
5846 	return 0;
5847 }
5848 
5849 static int rbd_dev_v2_data_pool(struct rbd_device *rbd_dev)
5850 {
5851 	__le64 data_pool_id;
5852 	int ret;
5853 
5854 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
5855 				  &rbd_dev->header_oloc, "get_data_pool",
5856 				  NULL, 0, &data_pool_id, sizeof(data_pool_id));
5857 	if (ret < 0)
5858 		return ret;
5859 	if (ret < sizeof(data_pool_id))
5860 		return -EBADMSG;
5861 
5862 	rbd_dev->header.data_pool_id = le64_to_cpu(data_pool_id);
5863 	WARN_ON(rbd_dev->header.data_pool_id == CEPH_NOPOOL);
5864 	return 0;
5865 }
5866 
5867 static char *rbd_dev_image_name(struct rbd_device *rbd_dev)
5868 {
5869 	CEPH_DEFINE_OID_ONSTACK(oid);
5870 	size_t image_id_size;
5871 	char *image_id;
5872 	void *p;
5873 	void *end;
5874 	size_t size;
5875 	void *reply_buf = NULL;
5876 	size_t len = 0;
5877 	char *image_name = NULL;
5878 	int ret;
5879 
5880 	rbd_assert(!rbd_dev->spec->image_name);
5881 
5882 	len = strlen(rbd_dev->spec->image_id);
5883 	image_id_size = sizeof (__le32) + len;
5884 	image_id = kmalloc(image_id_size, GFP_KERNEL);
5885 	if (!image_id)
5886 		return NULL;
5887 
5888 	p = image_id;
5889 	end = image_id + image_id_size;
5890 	ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len);
5891 
5892 	size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX;
5893 	reply_buf = kmalloc(size, GFP_KERNEL);
5894 	if (!reply_buf)
5895 		goto out;
5896 
5897 	ceph_oid_printf(&oid, "%s", RBD_DIRECTORY);
5898 	ret = rbd_obj_method_sync(rbd_dev, &oid, &rbd_dev->header_oloc,
5899 				  "dir_get_name", image_id, image_id_size,
5900 				  reply_buf, size);
5901 	if (ret < 0)
5902 		goto out;
5903 	p = reply_buf;
5904 	end = reply_buf + ret;
5905 
5906 	image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL);
5907 	if (IS_ERR(image_name))
5908 		image_name = NULL;
5909 	else
5910 		dout("%s: name is %s len is %zd\n", __func__, image_name, len);
5911 out:
5912 	kfree(reply_buf);
5913 	kfree(image_id);
5914 
5915 	return image_name;
5916 }
5917 
5918 static u64 rbd_v1_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
5919 {
5920 	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
5921 	const char *snap_name;
5922 	u32 which = 0;
5923 
5924 	/* Skip over names until we find the one we are looking for */
5925 
5926 	snap_name = rbd_dev->header.snap_names;
5927 	while (which < snapc->num_snaps) {
5928 		if (!strcmp(name, snap_name))
5929 			return snapc->snaps[which];
5930 		snap_name += strlen(snap_name) + 1;
5931 		which++;
5932 	}
5933 	return CEPH_NOSNAP;
5934 }
5935 
5936 static u64 rbd_v2_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
5937 {
5938 	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
5939 	u32 which;
5940 	bool found = false;
5941 	u64 snap_id;
5942 
5943 	for (which = 0; !found && which < snapc->num_snaps; which++) {
5944 		const char *snap_name;
5945 
5946 		snap_id = snapc->snaps[which];
5947 		snap_name = rbd_dev_v2_snap_name(rbd_dev, snap_id);
5948 		if (IS_ERR(snap_name)) {
5949 			/* ignore no-longer existing snapshots */
5950 			if (PTR_ERR(snap_name) == -ENOENT)
5951 				continue;
5952 			else
5953 				break;
5954 		}
5955 		found = !strcmp(name, snap_name);
5956 		kfree(snap_name);
5957 	}
5958 	return found ? snap_id : CEPH_NOSNAP;
5959 }
5960 
5961 /*
5962  * Assumes name is never RBD_SNAP_HEAD_NAME; returns CEPH_NOSNAP if
5963  * no snapshot by that name is found, or if an error occurs.
5964  */
5965 static u64 rbd_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
5966 {
5967 	if (rbd_dev->image_format == 1)
5968 		return rbd_v1_snap_id_by_name(rbd_dev, name);
5969 
5970 	return rbd_v2_snap_id_by_name(rbd_dev, name);
5971 }
5972 
5973 /*
5974  * An image being mapped will have everything but the snap id.
5975  */
5976 static int rbd_spec_fill_snap_id(struct rbd_device *rbd_dev)
5977 {
5978 	struct rbd_spec *spec = rbd_dev->spec;
5979 
5980 	rbd_assert(spec->pool_id != CEPH_NOPOOL && spec->pool_name);
5981 	rbd_assert(spec->image_id && spec->image_name);
5982 	rbd_assert(spec->snap_name);
5983 
5984 	if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
5985 		u64 snap_id;
5986 
5987 		snap_id = rbd_snap_id_by_name(rbd_dev, spec->snap_name);
5988 		if (snap_id == CEPH_NOSNAP)
5989 			return -ENOENT;
5990 
5991 		spec->snap_id = snap_id;
5992 	} else {
5993 		spec->snap_id = CEPH_NOSNAP;
5994 	}
5995 
5996 	return 0;
5997 }
5998 
5999 /*
6000  * A parent image will have all ids but none of the names.
6001  *
6002  * All names in an rbd spec are dynamically allocated.  It's OK if we
6003  * can't figure out the name for an image id.
6004  */
6005 static int rbd_spec_fill_names(struct rbd_device *rbd_dev)
6006 {
6007 	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
6008 	struct rbd_spec *spec = rbd_dev->spec;
6009 	const char *pool_name;
6010 	const char *image_name;
6011 	const char *snap_name;
6012 	int ret;
6013 
6014 	rbd_assert(spec->pool_id != CEPH_NOPOOL);
6015 	rbd_assert(spec->image_id);
6016 	rbd_assert(spec->snap_id != CEPH_NOSNAP);
6017 
6018 	/* Get the pool name; we have to make our own copy of this */
6019 
6020 	pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id);
6021 	if (!pool_name) {
6022 		rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id);
6023 		return -EIO;
6024 	}
6025 	pool_name = kstrdup(pool_name, GFP_KERNEL);
6026 	if (!pool_name)
6027 		return -ENOMEM;
6028 
6029 	/* Fetch the image name; tolerate failure here */
6030 
6031 	image_name = rbd_dev_image_name(rbd_dev);
6032 	if (!image_name)
6033 		rbd_warn(rbd_dev, "unable to get image name");
6034 
6035 	/* Fetch the snapshot name */
6036 
6037 	snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
6038 	if (IS_ERR(snap_name)) {
6039 		ret = PTR_ERR(snap_name);
6040 		goto out_err;
6041 	}
6042 
6043 	spec->pool_name = pool_name;
6044 	spec->image_name = image_name;
6045 	spec->snap_name = snap_name;
6046 
6047 	return 0;
6048 
6049 out_err:
6050 	kfree(image_name);
6051 	kfree(pool_name);
6052 	return ret;
6053 }
6054 
6055 static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev)
6056 {
6057 	size_t size;
6058 	int ret;
6059 	void *reply_buf;
6060 	void *p;
6061 	void *end;
6062 	u64 seq;
6063 	u32 snap_count;
6064 	struct ceph_snap_context *snapc;
6065 	u32 i;
6066 
6067 	/*
6068 	 * We'll need room for the seq value (maximum snapshot id),
6069 	 * snapshot count, and array of that many snapshot ids.
6070 	 * For now we have a fixed upper limit on the number we're
6071 	 * prepared to receive.
6072 	 */
6073 	size = sizeof (__le64) + sizeof (__le32) +
6074 			RBD_MAX_SNAP_COUNT * sizeof (__le64);
6075 	reply_buf = kzalloc(size, GFP_KERNEL);
6076 	if (!reply_buf)
6077 		return -ENOMEM;
6078 
6079 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
6080 				  &rbd_dev->header_oloc, "get_snapcontext",
6081 				  NULL, 0, reply_buf, size);
6082 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
6083 	if (ret < 0)
6084 		goto out;
6085 
6086 	p = reply_buf;
6087 	end = reply_buf + ret;
6088 	ret = -ERANGE;
6089 	ceph_decode_64_safe(&p, end, seq, out);
6090 	ceph_decode_32_safe(&p, end, snap_count, out);
6091 
6092 	/*
6093 	 * Make sure the reported number of snapshot ids wouldn't go
6094 	 * beyond the end of our buffer.  But before checking that,
6095 	 * make sure the computed size of the snapshot context we
6096 	 * allocate is representable in a size_t.
6097 	 */
6098 	if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
6099 				 / sizeof (u64)) {
6100 		ret = -EINVAL;
6101 		goto out;
6102 	}
6103 	if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
6104 		goto out;
6105 	ret = 0;
6106 
6107 	snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
6108 	if (!snapc) {
6109 		ret = -ENOMEM;
6110 		goto out;
6111 	}
6112 	snapc->seq = seq;
6113 	for (i = 0; i < snap_count; i++)
6114 		snapc->snaps[i] = ceph_decode_64(&p);
6115 
6116 	ceph_put_snap_context(rbd_dev->header.snapc);
6117 	rbd_dev->header.snapc = snapc;
6118 
6119 	dout("  snap context seq = %llu, snap_count = %u\n",
6120 		(unsigned long long)seq, (unsigned int)snap_count);
6121 out:
6122 	kfree(reply_buf);
6123 
6124 	return ret;
6125 }
6126 
6127 static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev,
6128 					u64 snap_id)
6129 {
6130 	size_t size;
6131 	void *reply_buf;
6132 	__le64 snapid;
6133 	int ret;
6134 	void *p;
6135 	void *end;
6136 	char *snap_name;
6137 
6138 	size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
6139 	reply_buf = kmalloc(size, GFP_KERNEL);
6140 	if (!reply_buf)
6141 		return ERR_PTR(-ENOMEM);
6142 
6143 	snapid = cpu_to_le64(snap_id);
6144 	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
6145 				  &rbd_dev->header_oloc, "get_snapshot_name",
6146 				  &snapid, sizeof(snapid), reply_buf, size);
6147 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
6148 	if (ret < 0) {
6149 		snap_name = ERR_PTR(ret);
6150 		goto out;
6151 	}
6152 
6153 	p = reply_buf;
6154 	end = reply_buf + ret;
6155 	snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
6156 	if (IS_ERR(snap_name))
6157 		goto out;
6158 
6159 	dout("  snap_id 0x%016llx snap_name = %s\n",
6160 		(unsigned long long)snap_id, snap_name);
6161 out:
6162 	kfree(reply_buf);
6163 
6164 	return snap_name;
6165 }
6166 
6167 static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev)
6168 {
6169 	bool first_time = rbd_dev->header.object_prefix == NULL;
6170 	int ret;
6171 
6172 	ret = rbd_dev_v2_image_size(rbd_dev);
6173 	if (ret)
6174 		return ret;
6175 
6176 	if (first_time) {
6177 		ret = rbd_dev_v2_header_onetime(rbd_dev);
6178 		if (ret)
6179 			return ret;
6180 	}
6181 
6182 	ret = rbd_dev_v2_snap_context(rbd_dev);
6183 	if (ret && first_time) {
6184 		kfree(rbd_dev->header.object_prefix);
6185 		rbd_dev->header.object_prefix = NULL;
6186 	}
6187 
6188 	return ret;
6189 }
6190 
6191 static int rbd_dev_header_info(struct rbd_device *rbd_dev)
6192 {
6193 	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
6194 
6195 	if (rbd_dev->image_format == 1)
6196 		return rbd_dev_v1_header_info(rbd_dev);
6197 
6198 	return rbd_dev_v2_header_info(rbd_dev);
6199 }
6200 
6201 /*
6202  * Skips over white space at *buf, and updates *buf to point to the
6203  * first found non-space character (if any). Returns the length of
6204  * the token (string of non-white space characters) found.  Note
6205  * that *buf must be terminated with '\0'.
6206  */
6207 static inline size_t next_token(const char **buf)
6208 {
6209         /*
6210         * These are the characters that produce nonzero for
6211         * isspace() in the "C" and "POSIX" locales.
6212         */
6213 	static const char spaces[] = " \f\n\r\t\v";
6214 
6215         *buf += strspn(*buf, spaces);	/* Find start of token */
6216 
6217 	return strcspn(*buf, spaces);   /* Return token length */
6218 }
6219 
6220 /*
6221  * Finds the next token in *buf, dynamically allocates a buffer big
6222  * enough to hold a copy of it, and copies the token into the new
6223  * buffer.  The copy is guaranteed to be terminated with '\0'.  Note
6224  * that a duplicate buffer is created even for a zero-length token.
6225  *
6226  * Returns a pointer to the newly-allocated duplicate, or a null
6227  * pointer if memory for the duplicate was not available.  If
6228  * the lenp argument is a non-null pointer, the length of the token
6229  * (not including the '\0') is returned in *lenp.
6230  *
6231  * If successful, the *buf pointer will be updated to point beyond
6232  * the end of the found token.
6233  *
6234  * Note: uses GFP_KERNEL for allocation.
6235  */
6236 static inline char *dup_token(const char **buf, size_t *lenp)
6237 {
6238 	char *dup;
6239 	size_t len;
6240 
6241 	len = next_token(buf);
6242 	dup = kmemdup(*buf, len + 1, GFP_KERNEL);
6243 	if (!dup)
6244 		return NULL;
6245 	*(dup + len) = '\0';
6246 	*buf += len;
6247 
6248 	if (lenp)
6249 		*lenp = len;
6250 
6251 	return dup;
6252 }
6253 
6254 static int rbd_parse_param(struct fs_parameter *param,
6255 			    struct rbd_parse_opts_ctx *pctx)
6256 {
6257 	struct rbd_options *opt = pctx->opts;
6258 	struct fs_parse_result result;
6259 	struct p_log log = {.prefix = "rbd"};
6260 	int token, ret;
6261 
6262 	ret = ceph_parse_param(param, pctx->copts, NULL);
6263 	if (ret != -ENOPARAM)
6264 		return ret;
6265 
6266 	token = __fs_parse(&log, rbd_parameters, param, &result);
6267 	dout("%s fs_parse '%s' token %d\n", __func__, param->key, token);
6268 	if (token < 0) {
6269 		if (token == -ENOPARAM)
6270 			return inval_plog(&log, "Unknown parameter '%s'",
6271 					  param->key);
6272 		return token;
6273 	}
6274 
6275 	switch (token) {
6276 	case Opt_queue_depth:
6277 		if (result.uint_32 < 1)
6278 			goto out_of_range;
6279 		opt->queue_depth = result.uint_32;
6280 		break;
6281 	case Opt_alloc_size:
6282 		if (result.uint_32 < SECTOR_SIZE)
6283 			goto out_of_range;
6284 		if (!is_power_of_2(result.uint_32))
6285 			return inval_plog(&log, "alloc_size must be a power of 2");
6286 		opt->alloc_size = result.uint_32;
6287 		break;
6288 	case Opt_lock_timeout:
6289 		/* 0 is "wait forever" (i.e. infinite timeout) */
6290 		if (result.uint_32 > INT_MAX / 1000)
6291 			goto out_of_range;
6292 		opt->lock_timeout = msecs_to_jiffies(result.uint_32 * 1000);
6293 		break;
6294 	case Opt_pool_ns:
6295 		kfree(pctx->spec->pool_ns);
6296 		pctx->spec->pool_ns = param->string;
6297 		param->string = NULL;
6298 		break;
6299 	case Opt_compression_hint:
6300 		switch (result.uint_32) {
6301 		case Opt_compression_hint_none:
6302 			opt->alloc_hint_flags &=
6303 			    ~(CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE |
6304 			      CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE);
6305 			break;
6306 		case Opt_compression_hint_compressible:
6307 			opt->alloc_hint_flags |=
6308 			    CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE;
6309 			opt->alloc_hint_flags &=
6310 			    ~CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE;
6311 			break;
6312 		case Opt_compression_hint_incompressible:
6313 			opt->alloc_hint_flags |=
6314 			    CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE;
6315 			opt->alloc_hint_flags &=
6316 			    ~CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE;
6317 			break;
6318 		default:
6319 			BUG();
6320 		}
6321 		break;
6322 	case Opt_read_only:
6323 		opt->read_only = true;
6324 		break;
6325 	case Opt_read_write:
6326 		opt->read_only = false;
6327 		break;
6328 	case Opt_lock_on_read:
6329 		opt->lock_on_read = true;
6330 		break;
6331 	case Opt_exclusive:
6332 		opt->exclusive = true;
6333 		break;
6334 	case Opt_notrim:
6335 		opt->trim = false;
6336 		break;
6337 	default:
6338 		BUG();
6339 	}
6340 
6341 	return 0;
6342 
6343 out_of_range:
6344 	return inval_plog(&log, "%s out of range", param->key);
6345 }
6346 
6347 /*
6348  * This duplicates most of generic_parse_monolithic(), untying it from
6349  * fs_context and skipping standard superblock and security options.
6350  */
6351 static int rbd_parse_options(char *options, struct rbd_parse_opts_ctx *pctx)
6352 {
6353 	char *key;
6354 	int ret = 0;
6355 
6356 	dout("%s '%s'\n", __func__, options);
6357 	while ((key = strsep(&options, ",")) != NULL) {
6358 		if (*key) {
6359 			struct fs_parameter param = {
6360 				.key	= key,
6361 				.type	= fs_value_is_flag,
6362 			};
6363 			char *value = strchr(key, '=');
6364 			size_t v_len = 0;
6365 
6366 			if (value) {
6367 				if (value == key)
6368 					continue;
6369 				*value++ = 0;
6370 				v_len = strlen(value);
6371 				param.string = kmemdup_nul(value, v_len,
6372 							   GFP_KERNEL);
6373 				if (!param.string)
6374 					return -ENOMEM;
6375 				param.type = fs_value_is_string;
6376 			}
6377 			param.size = v_len;
6378 
6379 			ret = rbd_parse_param(&param, pctx);
6380 			kfree(param.string);
6381 			if (ret)
6382 				break;
6383 		}
6384 	}
6385 
6386 	return ret;
6387 }
6388 
6389 /*
6390  * Parse the options provided for an "rbd add" (i.e., rbd image
6391  * mapping) request.  These arrive via a write to /sys/bus/rbd/add,
6392  * and the data written is passed here via a NUL-terminated buffer.
6393  * Returns 0 if successful or an error code otherwise.
6394  *
6395  * The information extracted from these options is recorded in
6396  * the other parameters which return dynamically-allocated
6397  * structures:
6398  *  ceph_opts
6399  *      The address of a pointer that will refer to a ceph options
6400  *      structure.  Caller must release the returned pointer using
6401  *      ceph_destroy_options() when it is no longer needed.
6402  *  rbd_opts
6403  *	Address of an rbd options pointer.  Fully initialized by
6404  *	this function; caller must release with kfree().
6405  *  spec
6406  *	Address of an rbd image specification pointer.  Fully
6407  *	initialized by this function based on parsed options.
6408  *	Caller must release with rbd_spec_put().
6409  *
6410  * The options passed take this form:
6411  *  <mon_addrs> <options> <pool_name> <image_name> [<snap_id>]
6412  * where:
6413  *  <mon_addrs>
6414  *      A comma-separated list of one or more monitor addresses.
6415  *      A monitor address is an ip address, optionally followed
6416  *      by a port number (separated by a colon).
6417  *        I.e.:  ip1[:port1][,ip2[:port2]...]
6418  *  <options>
6419  *      A comma-separated list of ceph and/or rbd options.
6420  *  <pool_name>
6421  *      The name of the rados pool containing the rbd image.
6422  *  <image_name>
6423  *      The name of the image in that pool to map.
6424  *  <snap_id>
6425  *      An optional snapshot id.  If provided, the mapping will
6426  *      present data from the image at the time that snapshot was
6427  *      created.  The image head is used if no snapshot id is
6428  *      provided.  Snapshot mappings are always read-only.
6429  */
6430 static int rbd_add_parse_args(const char *buf,
6431 				struct ceph_options **ceph_opts,
6432 				struct rbd_options **opts,
6433 				struct rbd_spec **rbd_spec)
6434 {
6435 	size_t len;
6436 	char *options;
6437 	const char *mon_addrs;
6438 	char *snap_name;
6439 	size_t mon_addrs_size;
6440 	struct rbd_parse_opts_ctx pctx = { 0 };
6441 	int ret;
6442 
6443 	/* The first four tokens are required */
6444 
6445 	len = next_token(&buf);
6446 	if (!len) {
6447 		rbd_warn(NULL, "no monitor address(es) provided");
6448 		return -EINVAL;
6449 	}
6450 	mon_addrs = buf;
6451 	mon_addrs_size = len;
6452 	buf += len;
6453 
6454 	ret = -EINVAL;
6455 	options = dup_token(&buf, NULL);
6456 	if (!options)
6457 		return -ENOMEM;
6458 	if (!*options) {
6459 		rbd_warn(NULL, "no options provided");
6460 		goto out_err;
6461 	}
6462 
6463 	pctx.spec = rbd_spec_alloc();
6464 	if (!pctx.spec)
6465 		goto out_mem;
6466 
6467 	pctx.spec->pool_name = dup_token(&buf, NULL);
6468 	if (!pctx.spec->pool_name)
6469 		goto out_mem;
6470 	if (!*pctx.spec->pool_name) {
6471 		rbd_warn(NULL, "no pool name provided");
6472 		goto out_err;
6473 	}
6474 
6475 	pctx.spec->image_name = dup_token(&buf, NULL);
6476 	if (!pctx.spec->image_name)
6477 		goto out_mem;
6478 	if (!*pctx.spec->image_name) {
6479 		rbd_warn(NULL, "no image name provided");
6480 		goto out_err;
6481 	}
6482 
6483 	/*
6484 	 * Snapshot name is optional; default is to use "-"
6485 	 * (indicating the head/no snapshot).
6486 	 */
6487 	len = next_token(&buf);
6488 	if (!len) {
6489 		buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */
6490 		len = sizeof (RBD_SNAP_HEAD_NAME) - 1;
6491 	} else if (len > RBD_MAX_SNAP_NAME_LEN) {
6492 		ret = -ENAMETOOLONG;
6493 		goto out_err;
6494 	}
6495 	snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
6496 	if (!snap_name)
6497 		goto out_mem;
6498 	*(snap_name + len) = '\0';
6499 	pctx.spec->snap_name = snap_name;
6500 
6501 	pctx.copts = ceph_alloc_options();
6502 	if (!pctx.copts)
6503 		goto out_mem;
6504 
6505 	/* Initialize all rbd options to the defaults */
6506 
6507 	pctx.opts = kzalloc(sizeof(*pctx.opts), GFP_KERNEL);
6508 	if (!pctx.opts)
6509 		goto out_mem;
6510 
6511 	pctx.opts->read_only = RBD_READ_ONLY_DEFAULT;
6512 	pctx.opts->queue_depth = RBD_QUEUE_DEPTH_DEFAULT;
6513 	pctx.opts->alloc_size = RBD_ALLOC_SIZE_DEFAULT;
6514 	pctx.opts->lock_timeout = RBD_LOCK_TIMEOUT_DEFAULT;
6515 	pctx.opts->lock_on_read = RBD_LOCK_ON_READ_DEFAULT;
6516 	pctx.opts->exclusive = RBD_EXCLUSIVE_DEFAULT;
6517 	pctx.opts->trim = RBD_TRIM_DEFAULT;
6518 
6519 	ret = ceph_parse_mon_ips(mon_addrs, mon_addrs_size, pctx.copts, NULL,
6520 				 ',');
6521 	if (ret)
6522 		goto out_err;
6523 
6524 	ret = rbd_parse_options(options, &pctx);
6525 	if (ret)
6526 		goto out_err;
6527 
6528 	*ceph_opts = pctx.copts;
6529 	*opts = pctx.opts;
6530 	*rbd_spec = pctx.spec;
6531 	kfree(options);
6532 	return 0;
6533 
6534 out_mem:
6535 	ret = -ENOMEM;
6536 out_err:
6537 	kfree(pctx.opts);
6538 	ceph_destroy_options(pctx.copts);
6539 	rbd_spec_put(pctx.spec);
6540 	kfree(options);
6541 	return ret;
6542 }
6543 
6544 static void rbd_dev_image_unlock(struct rbd_device *rbd_dev)
6545 {
6546 	down_write(&rbd_dev->lock_rwsem);
6547 	if (__rbd_is_lock_owner(rbd_dev))
6548 		__rbd_release_lock(rbd_dev);
6549 	up_write(&rbd_dev->lock_rwsem);
6550 }
6551 
6552 /*
6553  * If the wait is interrupted, an error is returned even if the lock
6554  * was successfully acquired.  rbd_dev_image_unlock() will release it
6555  * if needed.
6556  */
6557 static int rbd_add_acquire_lock(struct rbd_device *rbd_dev)
6558 {
6559 	long ret;
6560 
6561 	if (!(rbd_dev->header.features & RBD_FEATURE_EXCLUSIVE_LOCK)) {
6562 		if (!rbd_dev->opts->exclusive && !rbd_dev->opts->lock_on_read)
6563 			return 0;
6564 
6565 		rbd_warn(rbd_dev, "exclusive-lock feature is not enabled");
6566 		return -EINVAL;
6567 	}
6568 
6569 	if (rbd_is_ro(rbd_dev))
6570 		return 0;
6571 
6572 	rbd_assert(!rbd_is_lock_owner(rbd_dev));
6573 	queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
6574 	ret = wait_for_completion_killable_timeout(&rbd_dev->acquire_wait,
6575 			    ceph_timeout_jiffies(rbd_dev->opts->lock_timeout));
6576 	if (ret > 0) {
6577 		ret = rbd_dev->acquire_err;
6578 	} else {
6579 		cancel_delayed_work_sync(&rbd_dev->lock_dwork);
6580 		if (!ret)
6581 			ret = -ETIMEDOUT;
6582 	}
6583 
6584 	if (ret) {
6585 		rbd_warn(rbd_dev, "failed to acquire exclusive lock: %ld", ret);
6586 		return ret;
6587 	}
6588 
6589 	/*
6590 	 * The lock may have been released by now, unless automatic lock
6591 	 * transitions are disabled.
6592 	 */
6593 	rbd_assert(!rbd_dev->opts->exclusive || rbd_is_lock_owner(rbd_dev));
6594 	return 0;
6595 }
6596 
6597 /*
6598  * An rbd format 2 image has a unique identifier, distinct from the
6599  * name given to it by the user.  Internally, that identifier is
6600  * what's used to specify the names of objects related to the image.
6601  *
6602  * A special "rbd id" object is used to map an rbd image name to its
6603  * id.  If that object doesn't exist, then there is no v2 rbd image
6604  * with the supplied name.
6605  *
6606  * This function will record the given rbd_dev's image_id field if
6607  * it can be determined, and in that case will return 0.  If any
6608  * errors occur a negative errno will be returned and the rbd_dev's
6609  * image_id field will be unchanged (and should be NULL).
6610  */
6611 static int rbd_dev_image_id(struct rbd_device *rbd_dev)
6612 {
6613 	int ret;
6614 	size_t size;
6615 	CEPH_DEFINE_OID_ONSTACK(oid);
6616 	void *response;
6617 	char *image_id;
6618 
6619 	/*
6620 	 * When probing a parent image, the image id is already
6621 	 * known (and the image name likely is not).  There's no
6622 	 * need to fetch the image id again in this case.  We
6623 	 * do still need to set the image format though.
6624 	 */
6625 	if (rbd_dev->spec->image_id) {
6626 		rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1;
6627 
6628 		return 0;
6629 	}
6630 
6631 	/*
6632 	 * First, see if the format 2 image id file exists, and if
6633 	 * so, get the image's persistent id from it.
6634 	 */
6635 	ret = ceph_oid_aprintf(&oid, GFP_KERNEL, "%s%s", RBD_ID_PREFIX,
6636 			       rbd_dev->spec->image_name);
6637 	if (ret)
6638 		return ret;
6639 
6640 	dout("rbd id object name is %s\n", oid.name);
6641 
6642 	/* Response will be an encoded string, which includes a length */
6643 	size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
6644 	response = kzalloc(size, GFP_NOIO);
6645 	if (!response) {
6646 		ret = -ENOMEM;
6647 		goto out;
6648 	}
6649 
6650 	/* If it doesn't exist we'll assume it's a format 1 image */
6651 
6652 	ret = rbd_obj_method_sync(rbd_dev, &oid, &rbd_dev->header_oloc,
6653 				  "get_id", NULL, 0,
6654 				  response, size);
6655 	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
6656 	if (ret == -ENOENT) {
6657 		image_id = kstrdup("", GFP_KERNEL);
6658 		ret = image_id ? 0 : -ENOMEM;
6659 		if (!ret)
6660 			rbd_dev->image_format = 1;
6661 	} else if (ret >= 0) {
6662 		void *p = response;
6663 
6664 		image_id = ceph_extract_encoded_string(&p, p + ret,
6665 						NULL, GFP_NOIO);
6666 		ret = PTR_ERR_OR_ZERO(image_id);
6667 		if (!ret)
6668 			rbd_dev->image_format = 2;
6669 	}
6670 
6671 	if (!ret) {
6672 		rbd_dev->spec->image_id = image_id;
6673 		dout("image_id is %s\n", image_id);
6674 	}
6675 out:
6676 	kfree(response);
6677 	ceph_oid_destroy(&oid);
6678 	return ret;
6679 }
6680 
6681 /*
6682  * Undo whatever state changes are made by v1 or v2 header info
6683  * call.
6684  */
6685 static void rbd_dev_unprobe(struct rbd_device *rbd_dev)
6686 {
6687 	struct rbd_image_header	*header;
6688 
6689 	rbd_dev_parent_put(rbd_dev);
6690 	rbd_object_map_free(rbd_dev);
6691 	rbd_dev_mapping_clear(rbd_dev);
6692 
6693 	/* Free dynamic fields from the header, then zero it out */
6694 
6695 	header = &rbd_dev->header;
6696 	ceph_put_snap_context(header->snapc);
6697 	kfree(header->snap_sizes);
6698 	kfree(header->snap_names);
6699 	kfree(header->object_prefix);
6700 	memset(header, 0, sizeof (*header));
6701 }
6702 
6703 static int rbd_dev_v2_header_onetime(struct rbd_device *rbd_dev)
6704 {
6705 	int ret;
6706 
6707 	ret = rbd_dev_v2_object_prefix(rbd_dev);
6708 	if (ret)
6709 		goto out_err;
6710 
6711 	/*
6712 	 * Get the and check features for the image.  Currently the
6713 	 * features are assumed to never change.
6714 	 */
6715 	ret = rbd_dev_v2_features(rbd_dev);
6716 	if (ret)
6717 		goto out_err;
6718 
6719 	/* If the image supports fancy striping, get its parameters */
6720 
6721 	if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) {
6722 		ret = rbd_dev_v2_striping_info(rbd_dev);
6723 		if (ret < 0)
6724 			goto out_err;
6725 	}
6726 
6727 	if (rbd_dev->header.features & RBD_FEATURE_DATA_POOL) {
6728 		ret = rbd_dev_v2_data_pool(rbd_dev);
6729 		if (ret)
6730 			goto out_err;
6731 	}
6732 
6733 	rbd_init_layout(rbd_dev);
6734 	return 0;
6735 
6736 out_err:
6737 	rbd_dev->header.features = 0;
6738 	kfree(rbd_dev->header.object_prefix);
6739 	rbd_dev->header.object_prefix = NULL;
6740 	return ret;
6741 }
6742 
6743 /*
6744  * @depth is rbd_dev_image_probe() -> rbd_dev_probe_parent() ->
6745  * rbd_dev_image_probe() recursion depth, which means it's also the
6746  * length of the already discovered part of the parent chain.
6747  */
6748 static int rbd_dev_probe_parent(struct rbd_device *rbd_dev, int depth)
6749 {
6750 	struct rbd_device *parent = NULL;
6751 	int ret;
6752 
6753 	if (!rbd_dev->parent_spec)
6754 		return 0;
6755 
6756 	if (++depth > RBD_MAX_PARENT_CHAIN_LEN) {
6757 		pr_info("parent chain is too long (%d)\n", depth);
6758 		ret = -EINVAL;
6759 		goto out_err;
6760 	}
6761 
6762 	parent = __rbd_dev_create(rbd_dev->parent_spec);
6763 	if (!parent) {
6764 		ret = -ENOMEM;
6765 		goto out_err;
6766 	}
6767 
6768 	/*
6769 	 * Images related by parent/child relationships always share
6770 	 * rbd_client and spec/parent_spec, so bump their refcounts.
6771 	 */
6772 	parent->rbd_client = __rbd_get_client(rbd_dev->rbd_client);
6773 	parent->spec = rbd_spec_get(rbd_dev->parent_spec);
6774 
6775 	__set_bit(RBD_DEV_FLAG_READONLY, &parent->flags);
6776 
6777 	ret = rbd_dev_image_probe(parent, depth);
6778 	if (ret < 0)
6779 		goto out_err;
6780 
6781 	rbd_dev->parent = parent;
6782 	atomic_set(&rbd_dev->parent_ref, 1);
6783 	return 0;
6784 
6785 out_err:
6786 	rbd_dev_unparent(rbd_dev);
6787 	rbd_dev_destroy(parent);
6788 	return ret;
6789 }
6790 
6791 static void rbd_dev_device_release(struct rbd_device *rbd_dev)
6792 {
6793 	clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
6794 	rbd_free_disk(rbd_dev);
6795 	if (!single_major)
6796 		unregister_blkdev(rbd_dev->major, rbd_dev->name);
6797 }
6798 
6799 /*
6800  * rbd_dev->header_rwsem must be locked for write and will be unlocked
6801  * upon return.
6802  */
6803 static int rbd_dev_device_setup(struct rbd_device *rbd_dev)
6804 {
6805 	int ret;
6806 
6807 	/* Record our major and minor device numbers. */
6808 
6809 	if (!single_major) {
6810 		ret = register_blkdev(0, rbd_dev->name);
6811 		if (ret < 0)
6812 			goto err_out_unlock;
6813 
6814 		rbd_dev->major = ret;
6815 		rbd_dev->minor = 0;
6816 	} else {
6817 		rbd_dev->major = rbd_major;
6818 		rbd_dev->minor = rbd_dev_id_to_minor(rbd_dev->dev_id);
6819 	}
6820 
6821 	/* Set up the blkdev mapping. */
6822 
6823 	ret = rbd_init_disk(rbd_dev);
6824 	if (ret)
6825 		goto err_out_blkdev;
6826 
6827 	set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
6828 	set_disk_ro(rbd_dev->disk, rbd_is_ro(rbd_dev));
6829 
6830 	ret = dev_set_name(&rbd_dev->dev, "%d", rbd_dev->dev_id);
6831 	if (ret)
6832 		goto err_out_disk;
6833 
6834 	set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
6835 	up_write(&rbd_dev->header_rwsem);
6836 	return 0;
6837 
6838 err_out_disk:
6839 	rbd_free_disk(rbd_dev);
6840 err_out_blkdev:
6841 	if (!single_major)
6842 		unregister_blkdev(rbd_dev->major, rbd_dev->name);
6843 err_out_unlock:
6844 	up_write(&rbd_dev->header_rwsem);
6845 	return ret;
6846 }
6847 
6848 static int rbd_dev_header_name(struct rbd_device *rbd_dev)
6849 {
6850 	struct rbd_spec *spec = rbd_dev->spec;
6851 	int ret;
6852 
6853 	/* Record the header object name for this rbd image. */
6854 
6855 	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
6856 	if (rbd_dev->image_format == 1)
6857 		ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s",
6858 				       spec->image_name, RBD_SUFFIX);
6859 	else
6860 		ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s",
6861 				       RBD_HEADER_PREFIX, spec->image_id);
6862 
6863 	return ret;
6864 }
6865 
6866 static void rbd_print_dne(struct rbd_device *rbd_dev, bool is_snap)
6867 {
6868 	if (!is_snap) {
6869 		pr_info("image %s/%s%s%s does not exist\n",
6870 			rbd_dev->spec->pool_name,
6871 			rbd_dev->spec->pool_ns ?: "",
6872 			rbd_dev->spec->pool_ns ? "/" : "",
6873 			rbd_dev->spec->image_name);
6874 	} else {
6875 		pr_info("snap %s/%s%s%s@%s does not exist\n",
6876 			rbd_dev->spec->pool_name,
6877 			rbd_dev->spec->pool_ns ?: "",
6878 			rbd_dev->spec->pool_ns ? "/" : "",
6879 			rbd_dev->spec->image_name,
6880 			rbd_dev->spec->snap_name);
6881 	}
6882 }
6883 
6884 static void rbd_dev_image_release(struct rbd_device *rbd_dev)
6885 {
6886 	if (!rbd_is_ro(rbd_dev))
6887 		rbd_unregister_watch(rbd_dev);
6888 
6889 	rbd_dev_unprobe(rbd_dev);
6890 	rbd_dev->image_format = 0;
6891 	kfree(rbd_dev->spec->image_id);
6892 	rbd_dev->spec->image_id = NULL;
6893 }
6894 
6895 /*
6896  * Probe for the existence of the header object for the given rbd
6897  * device.  If this image is the one being mapped (i.e., not a
6898  * parent), initiate a watch on its header object before using that
6899  * object to get detailed information about the rbd image.
6900  *
6901  * On success, returns with header_rwsem held for write if called
6902  * with @depth == 0.
6903  */
6904 static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth)
6905 {
6906 	bool need_watch = !rbd_is_ro(rbd_dev);
6907 	int ret;
6908 
6909 	/*
6910 	 * Get the id from the image id object.  Unless there's an
6911 	 * error, rbd_dev->spec->image_id will be filled in with
6912 	 * a dynamically-allocated string, and rbd_dev->image_format
6913 	 * will be set to either 1 or 2.
6914 	 */
6915 	ret = rbd_dev_image_id(rbd_dev);
6916 	if (ret)
6917 		return ret;
6918 
6919 	ret = rbd_dev_header_name(rbd_dev);
6920 	if (ret)
6921 		goto err_out_format;
6922 
6923 	if (need_watch) {
6924 		ret = rbd_register_watch(rbd_dev);
6925 		if (ret) {
6926 			if (ret == -ENOENT)
6927 				rbd_print_dne(rbd_dev, false);
6928 			goto err_out_format;
6929 		}
6930 	}
6931 
6932 	if (!depth)
6933 		down_write(&rbd_dev->header_rwsem);
6934 
6935 	ret = rbd_dev_header_info(rbd_dev);
6936 	if (ret) {
6937 		if (ret == -ENOENT && !need_watch)
6938 			rbd_print_dne(rbd_dev, false);
6939 		goto err_out_probe;
6940 	}
6941 
6942 	/*
6943 	 * If this image is the one being mapped, we have pool name and
6944 	 * id, image name and id, and snap name - need to fill snap id.
6945 	 * Otherwise this is a parent image, identified by pool, image
6946 	 * and snap ids - need to fill in names for those ids.
6947 	 */
6948 	if (!depth)
6949 		ret = rbd_spec_fill_snap_id(rbd_dev);
6950 	else
6951 		ret = rbd_spec_fill_names(rbd_dev);
6952 	if (ret) {
6953 		if (ret == -ENOENT)
6954 			rbd_print_dne(rbd_dev, true);
6955 		goto err_out_probe;
6956 	}
6957 
6958 	ret = rbd_dev_mapping_set(rbd_dev);
6959 	if (ret)
6960 		goto err_out_probe;
6961 
6962 	if (rbd_is_snap(rbd_dev) &&
6963 	    (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP)) {
6964 		ret = rbd_object_map_load(rbd_dev);
6965 		if (ret)
6966 			goto err_out_probe;
6967 	}
6968 
6969 	if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
6970 		ret = rbd_dev_v2_parent_info(rbd_dev);
6971 		if (ret)
6972 			goto err_out_probe;
6973 	}
6974 
6975 	ret = rbd_dev_probe_parent(rbd_dev, depth);
6976 	if (ret)
6977 		goto err_out_probe;
6978 
6979 	dout("discovered format %u image, header name is %s\n",
6980 		rbd_dev->image_format, rbd_dev->header_oid.name);
6981 	return 0;
6982 
6983 err_out_probe:
6984 	if (!depth)
6985 		up_write(&rbd_dev->header_rwsem);
6986 	if (need_watch)
6987 		rbd_unregister_watch(rbd_dev);
6988 	rbd_dev_unprobe(rbd_dev);
6989 err_out_format:
6990 	rbd_dev->image_format = 0;
6991 	kfree(rbd_dev->spec->image_id);
6992 	rbd_dev->spec->image_id = NULL;
6993 	return ret;
6994 }
6995 
6996 static ssize_t do_rbd_add(const char *buf, size_t count)
6997 {
6998 	struct rbd_device *rbd_dev = NULL;
6999 	struct ceph_options *ceph_opts = NULL;
7000 	struct rbd_options *rbd_opts = NULL;
7001 	struct rbd_spec *spec = NULL;
7002 	struct rbd_client *rbdc;
7003 	int rc;
7004 
7005 	if (!capable(CAP_SYS_ADMIN))
7006 		return -EPERM;
7007 
7008 	if (!try_module_get(THIS_MODULE))
7009 		return -ENODEV;
7010 
7011 	/* parse add command */
7012 	rc = rbd_add_parse_args(buf, &ceph_opts, &rbd_opts, &spec);
7013 	if (rc < 0)
7014 		goto out;
7015 
7016 	rbdc = rbd_get_client(ceph_opts);
7017 	if (IS_ERR(rbdc)) {
7018 		rc = PTR_ERR(rbdc);
7019 		goto err_out_args;
7020 	}
7021 
7022 	/* pick the pool */
7023 	rc = ceph_pg_poolid_by_name(rbdc->client->osdc.osdmap, spec->pool_name);
7024 	if (rc < 0) {
7025 		if (rc == -ENOENT)
7026 			pr_info("pool %s does not exist\n", spec->pool_name);
7027 		goto err_out_client;
7028 	}
7029 	spec->pool_id = (u64)rc;
7030 
7031 	rbd_dev = rbd_dev_create(rbdc, spec, rbd_opts);
7032 	if (!rbd_dev) {
7033 		rc = -ENOMEM;
7034 		goto err_out_client;
7035 	}
7036 	rbdc = NULL;		/* rbd_dev now owns this */
7037 	spec = NULL;		/* rbd_dev now owns this */
7038 	rbd_opts = NULL;	/* rbd_dev now owns this */
7039 
7040 	/* if we are mapping a snapshot it will be a read-only mapping */
7041 	if (rbd_dev->opts->read_only ||
7042 	    strcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME))
7043 		__set_bit(RBD_DEV_FLAG_READONLY, &rbd_dev->flags);
7044 
7045 	rbd_dev->config_info = kstrdup(buf, GFP_KERNEL);
7046 	if (!rbd_dev->config_info) {
7047 		rc = -ENOMEM;
7048 		goto err_out_rbd_dev;
7049 	}
7050 
7051 	rc = rbd_dev_image_probe(rbd_dev, 0);
7052 	if (rc < 0)
7053 		goto err_out_rbd_dev;
7054 
7055 	if (rbd_dev->opts->alloc_size > rbd_dev->layout.object_size) {
7056 		rbd_warn(rbd_dev, "alloc_size adjusted to %u",
7057 			 rbd_dev->layout.object_size);
7058 		rbd_dev->opts->alloc_size = rbd_dev->layout.object_size;
7059 	}
7060 
7061 	rc = rbd_dev_device_setup(rbd_dev);
7062 	if (rc)
7063 		goto err_out_image_probe;
7064 
7065 	rc = rbd_add_acquire_lock(rbd_dev);
7066 	if (rc)
7067 		goto err_out_image_lock;
7068 
7069 	/* Everything's ready.  Announce the disk to the world. */
7070 
7071 	rc = device_add(&rbd_dev->dev);
7072 	if (rc)
7073 		goto err_out_image_lock;
7074 
7075 	rc = device_add_disk(&rbd_dev->dev, rbd_dev->disk, NULL);
7076 	if (rc)
7077 		goto err_out_cleanup_disk;
7078 
7079 	spin_lock(&rbd_dev_list_lock);
7080 	list_add_tail(&rbd_dev->node, &rbd_dev_list);
7081 	spin_unlock(&rbd_dev_list_lock);
7082 
7083 	pr_info("%s: capacity %llu features 0x%llx\n", rbd_dev->disk->disk_name,
7084 		(unsigned long long)get_capacity(rbd_dev->disk) << SECTOR_SHIFT,
7085 		rbd_dev->header.features);
7086 	rc = count;
7087 out:
7088 	module_put(THIS_MODULE);
7089 	return rc;
7090 
7091 err_out_cleanup_disk:
7092 	rbd_free_disk(rbd_dev);
7093 err_out_image_lock:
7094 	rbd_dev_image_unlock(rbd_dev);
7095 	rbd_dev_device_release(rbd_dev);
7096 err_out_image_probe:
7097 	rbd_dev_image_release(rbd_dev);
7098 err_out_rbd_dev:
7099 	rbd_dev_destroy(rbd_dev);
7100 err_out_client:
7101 	rbd_put_client(rbdc);
7102 err_out_args:
7103 	rbd_spec_put(spec);
7104 	kfree(rbd_opts);
7105 	goto out;
7106 }
7107 
7108 static ssize_t add_store(const struct bus_type *bus, const char *buf, size_t count)
7109 {
7110 	if (single_major)
7111 		return -EINVAL;
7112 
7113 	return do_rbd_add(buf, count);
7114 }
7115 
7116 static ssize_t add_single_major_store(const struct bus_type *bus, const char *buf,
7117 				      size_t count)
7118 {
7119 	return do_rbd_add(buf, count);
7120 }
7121 
7122 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev)
7123 {
7124 	while (rbd_dev->parent) {
7125 		struct rbd_device *first = rbd_dev;
7126 		struct rbd_device *second = first->parent;
7127 		struct rbd_device *third;
7128 
7129 		/*
7130 		 * Follow to the parent with no grandparent and
7131 		 * remove it.
7132 		 */
7133 		while (second && (third = second->parent)) {
7134 			first = second;
7135 			second = third;
7136 		}
7137 		rbd_assert(second);
7138 		rbd_dev_image_release(second);
7139 		rbd_dev_destroy(second);
7140 		first->parent = NULL;
7141 		first->parent_overlap = 0;
7142 
7143 		rbd_assert(first->parent_spec);
7144 		rbd_spec_put(first->parent_spec);
7145 		first->parent_spec = NULL;
7146 	}
7147 }
7148 
7149 static ssize_t do_rbd_remove(const char *buf, size_t count)
7150 {
7151 	struct rbd_device *rbd_dev = NULL;
7152 	struct list_head *tmp;
7153 	int dev_id;
7154 	char opt_buf[6];
7155 	bool force = false;
7156 	int ret;
7157 
7158 	if (!capable(CAP_SYS_ADMIN))
7159 		return -EPERM;
7160 
7161 	dev_id = -1;
7162 	opt_buf[0] = '\0';
7163 	sscanf(buf, "%d %5s", &dev_id, opt_buf);
7164 	if (dev_id < 0) {
7165 		pr_err("dev_id out of range\n");
7166 		return -EINVAL;
7167 	}
7168 	if (opt_buf[0] != '\0') {
7169 		if (!strcmp(opt_buf, "force")) {
7170 			force = true;
7171 		} else {
7172 			pr_err("bad remove option at '%s'\n", opt_buf);
7173 			return -EINVAL;
7174 		}
7175 	}
7176 
7177 	ret = -ENOENT;
7178 	spin_lock(&rbd_dev_list_lock);
7179 	list_for_each(tmp, &rbd_dev_list) {
7180 		rbd_dev = list_entry(tmp, struct rbd_device, node);
7181 		if (rbd_dev->dev_id == dev_id) {
7182 			ret = 0;
7183 			break;
7184 		}
7185 	}
7186 	if (!ret) {
7187 		spin_lock_irq(&rbd_dev->lock);
7188 		if (rbd_dev->open_count && !force)
7189 			ret = -EBUSY;
7190 		else if (test_and_set_bit(RBD_DEV_FLAG_REMOVING,
7191 					  &rbd_dev->flags))
7192 			ret = -EINPROGRESS;
7193 		spin_unlock_irq(&rbd_dev->lock);
7194 	}
7195 	spin_unlock(&rbd_dev_list_lock);
7196 	if (ret)
7197 		return ret;
7198 
7199 	if (force) {
7200 		/*
7201 		 * Prevent new IO from being queued and wait for existing
7202 		 * IO to complete/fail.
7203 		 */
7204 		blk_mq_freeze_queue(rbd_dev->disk->queue);
7205 		blk_mark_disk_dead(rbd_dev->disk);
7206 	}
7207 
7208 	del_gendisk(rbd_dev->disk);
7209 	spin_lock(&rbd_dev_list_lock);
7210 	list_del_init(&rbd_dev->node);
7211 	spin_unlock(&rbd_dev_list_lock);
7212 	device_del(&rbd_dev->dev);
7213 
7214 	rbd_dev_image_unlock(rbd_dev);
7215 	rbd_dev_device_release(rbd_dev);
7216 	rbd_dev_image_release(rbd_dev);
7217 	rbd_dev_destroy(rbd_dev);
7218 	return count;
7219 }
7220 
7221 static ssize_t remove_store(const struct bus_type *bus, const char *buf, size_t count)
7222 {
7223 	if (single_major)
7224 		return -EINVAL;
7225 
7226 	return do_rbd_remove(buf, count);
7227 }
7228 
7229 static ssize_t remove_single_major_store(const struct bus_type *bus, const char *buf,
7230 					 size_t count)
7231 {
7232 	return do_rbd_remove(buf, count);
7233 }
7234 
7235 /*
7236  * create control files in sysfs
7237  * /sys/bus/rbd/...
7238  */
7239 static int __init rbd_sysfs_init(void)
7240 {
7241 	int ret;
7242 
7243 	ret = device_register(&rbd_root_dev);
7244 	if (ret < 0) {
7245 		put_device(&rbd_root_dev);
7246 		return ret;
7247 	}
7248 
7249 	ret = bus_register(&rbd_bus_type);
7250 	if (ret < 0)
7251 		device_unregister(&rbd_root_dev);
7252 
7253 	return ret;
7254 }
7255 
7256 static void __exit rbd_sysfs_cleanup(void)
7257 {
7258 	bus_unregister(&rbd_bus_type);
7259 	device_unregister(&rbd_root_dev);
7260 }
7261 
7262 static int __init rbd_slab_init(void)
7263 {
7264 	rbd_assert(!rbd_img_request_cache);
7265 	rbd_img_request_cache = KMEM_CACHE(rbd_img_request, 0);
7266 	if (!rbd_img_request_cache)
7267 		return -ENOMEM;
7268 
7269 	rbd_assert(!rbd_obj_request_cache);
7270 	rbd_obj_request_cache = KMEM_CACHE(rbd_obj_request, 0);
7271 	if (!rbd_obj_request_cache)
7272 		goto out_err;
7273 
7274 	return 0;
7275 
7276 out_err:
7277 	kmem_cache_destroy(rbd_img_request_cache);
7278 	rbd_img_request_cache = NULL;
7279 	return -ENOMEM;
7280 }
7281 
7282 static void rbd_slab_exit(void)
7283 {
7284 	rbd_assert(rbd_obj_request_cache);
7285 	kmem_cache_destroy(rbd_obj_request_cache);
7286 	rbd_obj_request_cache = NULL;
7287 
7288 	rbd_assert(rbd_img_request_cache);
7289 	kmem_cache_destroy(rbd_img_request_cache);
7290 	rbd_img_request_cache = NULL;
7291 }
7292 
7293 static int __init rbd_init(void)
7294 {
7295 	int rc;
7296 
7297 	if (!libceph_compatible(NULL)) {
7298 		rbd_warn(NULL, "libceph incompatibility (quitting)");
7299 		return -EINVAL;
7300 	}
7301 
7302 	rc = rbd_slab_init();
7303 	if (rc)
7304 		return rc;
7305 
7306 	/*
7307 	 * The number of active work items is limited by the number of
7308 	 * rbd devices * queue depth, so leave @max_active at default.
7309 	 */
7310 	rbd_wq = alloc_workqueue(RBD_DRV_NAME, WQ_MEM_RECLAIM, 0);
7311 	if (!rbd_wq) {
7312 		rc = -ENOMEM;
7313 		goto err_out_slab;
7314 	}
7315 
7316 	if (single_major) {
7317 		rbd_major = register_blkdev(0, RBD_DRV_NAME);
7318 		if (rbd_major < 0) {
7319 			rc = rbd_major;
7320 			goto err_out_wq;
7321 		}
7322 	}
7323 
7324 	rc = rbd_sysfs_init();
7325 	if (rc)
7326 		goto err_out_blkdev;
7327 
7328 	if (single_major)
7329 		pr_info("loaded (major %d)\n", rbd_major);
7330 	else
7331 		pr_info("loaded\n");
7332 
7333 	return 0;
7334 
7335 err_out_blkdev:
7336 	if (single_major)
7337 		unregister_blkdev(rbd_major, RBD_DRV_NAME);
7338 err_out_wq:
7339 	destroy_workqueue(rbd_wq);
7340 err_out_slab:
7341 	rbd_slab_exit();
7342 	return rc;
7343 }
7344 
7345 static void __exit rbd_exit(void)
7346 {
7347 	ida_destroy(&rbd_dev_id_ida);
7348 	rbd_sysfs_cleanup();
7349 	if (single_major)
7350 		unregister_blkdev(rbd_major, RBD_DRV_NAME);
7351 	destroy_workqueue(rbd_wq);
7352 	rbd_slab_exit();
7353 }
7354 
7355 module_init(rbd_init);
7356 module_exit(rbd_exit);
7357 
7358 MODULE_AUTHOR("Alex Elder <elder@inktank.com>");
7359 MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
7360 MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
7361 /* following authorship retained from original osdblk.c */
7362 MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");
7363 
7364 MODULE_DESCRIPTION("RADOS Block Device (RBD) driver");
7365 MODULE_LICENSE("GPL");
7366