osd_client.c 83.6 KB
Newer Older
1

2
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
3

4
#include <linux/module.h>
S
Sage Weil 已提交
5 6 7 8 9 10
#include <linux/err.h>
#include <linux/highmem.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
11 12 13
#ifdef CONFIG_BLOCK
#include <linux/bio.h>
#endif
S
Sage Weil 已提交
14

15 16 17 18 19 20
#include <linux/ceph/libceph.h>
#include <linux/ceph/osd_client.h>
#include <linux/ceph/messenger.h>
#include <linux/ceph/decode.h>
#include <linux/ceph/auth.h>
#include <linux/ceph/pagelist.h>
S
Sage Weil 已提交
21

S
Sage Weil 已提交
22
#define OSD_OPREPLY_FRONT_LEN	512
23

24 25
static struct kmem_cache	*ceph_osd_request_cache;

26
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
27

A
Alex Elder 已提交
28
static void __send_queued(struct ceph_osd_client *osdc);
29
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
30 31
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req);
32 33
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req);
34 35
static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req);
36
static void __enqueue_request(struct ceph_osd_request *req);
S
Sage Weil 已提交
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

/*
 * Implement client access to distributed object storage cluster.
 *
 * All data objects are stored within a cluster/cloud of OSDs, or
 * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
 * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
 * remote daemons serving up and coordinating consistent and safe
 * access to storage.
 *
 * Cluster membership and the mapping of data objects onto storage devices
 * are described by the osd map.
 *
 * We keep track of pending OSD requests (read, write), resubmit
 * requests to different OSDs when the cluster topology/data layout
 * change, or retry the affected requests when the communications
 * channel with an OSD is reset.
 */

/*
 * calculate the mapping of a file extent onto an object, and fill out the
 * request accordingly.  shorten extent as necessary if it crosses an
 * object boundary.
 *
 * fill osd op in request message.
 */
63
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
64
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
65
{
66
	u64 orig_len = *plen;
S
Sage Weil 已提交
67
	int r;
S
Sage Weil 已提交
68

69
	/* object extent? */
70 71
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
72 73
	if (r < 0)
		return r;
74 75
	if (*objlen < orig_len) {
		*plen = *objlen;
76 77 78 79
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

80
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
81

82
	return 0;
S
Sage Weil 已提交
83 84
}

A
Alex Elder 已提交
85 86 87 88 89 90
static void ceph_osd_data_init(struct ceph_osd_data *osd_data)
{
	memset(osd_data, 0, sizeof (*osd_data));
	osd_data->type = CEPH_OSD_DATA_TYPE_NONE;
}

91
static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
92 93 94 95 96 97 98 99 100 101 102
			struct page **pages, u64 length, u32 alignment,
			bool pages_from_pool, bool own_pages)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
	osd_data->length = length;
	osd_data->alignment = alignment;
	osd_data->pages_from_pool = pages_from_pool;
	osd_data->own_pages = own_pages;
}

103
static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
104 105 106 107 108 109 110
			struct ceph_pagelist *pagelist)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
	osd_data->pagelist = pagelist;
}

#ifdef CONFIG_BLOCK
111
static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
112 113 114 115 116 117 118 119
			struct bio *bio, size_t bio_length)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_BIO;
	osd_data->bio = bio;
	osd_data->bio_length = bio_length;
}
#endif /* CONFIG_BLOCK */

120 121 122 123 124 125 126
#define osd_req_op_data(oreq, whch, typ, fld)				\
({									\
	struct ceph_osd_request *__oreq = (oreq);			\
	unsigned int __whch = (whch);					\
	BUG_ON(__whch >= __oreq->r_num_ops);				\
	&__oreq->r_ops[__whch].typ.fld;					\
})
127

A
Alex Elder 已提交
128 129 130 131 132 133 134 135
static struct ceph_osd_data *
osd_req_op_raw_data_in(struct ceph_osd_request *osd_req, unsigned int which)
{
	BUG_ON(which >= osd_req->r_num_ops);

	return &osd_req->r_ops[which].raw_data_in;
}

136 137
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
138
			unsigned int which)
139
{
140
	return osd_req_op_data(osd_req, which, extent, osd_data);
141 142 143
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data);

A
Alex Elder 已提交
144 145 146 147 148 149 150 151 152 153 154 155 156
void osd_req_op_raw_data_in_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

	osd_data = osd_req_op_raw_data_in(osd_req, which);
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_raw_data_in_pages);

157
void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
158 159
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
160 161 162 163
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

164
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
165 166 167 168 169 170
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);

void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
171
			unsigned int which, struct ceph_pagelist *pagelist)
172 173 174
{
	struct ceph_osd_data *osd_data;

175
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
176 177 178 179 180 181
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);

#ifdef CONFIG_BLOCK
void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
182
			unsigned int which, struct bio *bio, size_t bio_length)
183 184
{
	struct ceph_osd_data *osd_data;
185 186

	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
187 188 189 190 191 192 193 194 195 196 197
	ceph_osd_data_bio_init(osd_data, bio, bio_length);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio);
#endif /* CONFIG_BLOCK */

static void osd_req_op_cls_request_info_pagelist(
			struct ceph_osd_request *osd_req,
			unsigned int which, struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

198
	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
199 200 201
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

202 203 204 205 206 207
void osd_req_op_cls_request_data_pagelist(
			struct ceph_osd_request *osd_req,
			unsigned int which, struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

208
	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
209
	ceph_osd_data_pagelist_init(osd_data, pagelist);
210 211
	osd_req->r_ops[which].cls.indata_len += pagelist->length;
	osd_req->r_ops[which].indata_len += pagelist->length;
212 213 214
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);

215 216 217 218 219 220 221 222 223
void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages, u64 length,
			u32 alignment, bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
224 225
	osd_req->r_ops[which].cls.indata_len += length;
	osd_req->r_ops[which].indata_len += length;
226 227 228
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pages);

229 230 231 232 233 234
void osd_req_op_cls_response_data_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages, u64 length,
			u32 alignment, bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

235
	osd_data = osd_req_op_data(osd_req, which, cls, response_data);
236 237 238 239 240
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
static u64 ceph_osd_data_length(struct ceph_osd_data *osd_data)
{
	switch (osd_data->type) {
	case CEPH_OSD_DATA_TYPE_NONE:
		return 0;
	case CEPH_OSD_DATA_TYPE_PAGES:
		return osd_data->length;
	case CEPH_OSD_DATA_TYPE_PAGELIST:
		return (u64)osd_data->pagelist->length;
#ifdef CONFIG_BLOCK
	case CEPH_OSD_DATA_TYPE_BIO:
		return (u64)osd_data->bio_length;
#endif /* CONFIG_BLOCK */
	default:
		WARN(true, "unrecognized data type %d\n", (int)osd_data->type);
		return 0;
	}
}

A
Alex Elder 已提交
260 261
static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
{
262
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
A
Alex Elder 已提交
263 264 265 266 267 268
		int num_pages;

		num_pages = calc_pages_for((u64)osd_data->alignment,
						(u64)osd_data->length);
		ceph_release_page_vector(osd_data->pages, num_pages);
	}
269 270 271 272 273 274 275 276 277 278 279 280 281 282
	ceph_osd_data_init(osd_data);
}

static void osd_req_op_data_release(struct ceph_osd_request *osd_req,
			unsigned int which)
{
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
	op = &osd_req->r_ops[which];

	switch (op->op) {
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
283
	case CEPH_OSD_OP_WRITEFULL:
284 285 286 287
		ceph_osd_data_release(&op->extent.osd_data);
		break;
	case CEPH_OSD_OP_CALL:
		ceph_osd_data_release(&op->cls.request_info);
288
		ceph_osd_data_release(&op->cls.request_data);
289 290
		ceph_osd_data_release(&op->cls.response_data);
		break;
291 292 293 294
	case CEPH_OSD_OP_SETXATTR:
	case CEPH_OSD_OP_CMPXATTR:
		ceph_osd_data_release(&op->xattr.osd_data);
		break;
295 296 297
	case CEPH_OSD_OP_STAT:
		ceph_osd_data_release(&op->raw_data_in);
		break;
298 299 300
	default:
		break;
	}
A
Alex Elder 已提交
301 302
}

303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326
/*
 * Assumes @t is zero-initialized.
 */
static void target_init(struct ceph_osd_request_target *t)
{
	ceph_oid_init(&t->base_oid);
	ceph_oloc_init(&t->base_oloc);
	ceph_oid_init(&t->target_oid);
	ceph_oloc_init(&t->target_oloc);

	ceph_osds_init(&t->acting);
	ceph_osds_init(&t->up);
	t->size = -1;
	t->min_size = -1;

	t->osd = CEPH_HOMELESS_OSD;
}

static void target_destroy(struct ceph_osd_request_target *t)
{
	ceph_oid_destroy(&t->base_oid);
	ceph_oid_destroy(&t->target_oid);
}

S
Sage Weil 已提交
327 328 329
/*
 * requests
 */
330
static void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
331
{
332 333
	struct ceph_osd_request *req = container_of(kref,
					    struct ceph_osd_request, r_kref);
334
	unsigned int which;
S
Sage Weil 已提交
335

336 337
	dout("%s %p (r_request %p r_reply %p)\n", __func__, req,
	     req->r_request, req->r_reply);
338 339 340 341 342 343
	WARN_ON(!RB_EMPTY_NODE(&req->r_node));
	WARN_ON(!list_empty(&req->r_req_lru_item));
	WARN_ON(!list_empty(&req->r_osd_item));
	WARN_ON(!list_empty(&req->r_linger_item));
	WARN_ON(!list_empty(&req->r_linger_osd_item));
	WARN_ON(req->r_osd);
344

S
Sage Weil 已提交
345 346
	if (req->r_request)
		ceph_msg_put(req->r_request);
347
	if (req->r_reply) {
348
		ceph_msg_revoke_incoming(req->r_reply);
349
		ceph_msg_put(req->r_reply);
350
	}
351

352 353
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
354

355
	target_destroy(&req->r_t);
S
Sage Weil 已提交
356
	ceph_put_snap_context(req->r_snapc);
357

S
Sage Weil 已提交
358 359
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
360
	else if (req->r_num_ops <= CEPH_OSD_SLAB_OPS)
361
		kmem_cache_free(ceph_osd_request_cache, req);
362 363
	else
		kfree(req);
S
Sage Weil 已提交
364
}
365 366 367 368 369 370 371 372 373 374 375

void ceph_osdc_get_request(struct ceph_osd_request *req)
{
	dout("%s %p (was %d)\n", __func__, req,
	     atomic_read(&req->r_kref.refcount));
	kref_get(&req->r_kref);
}
EXPORT_SYMBOL(ceph_osdc_get_request);

void ceph_osdc_put_request(struct ceph_osd_request *req)
{
376 377 378 379 380
	if (req) {
		dout("%s %p (was %d)\n", __func__, req,
		     atomic_read(&req->r_kref.refcount));
		kref_put(&req->r_kref, ceph_osdc_release_request);
	}
381 382
}
EXPORT_SYMBOL(ceph_osdc_put_request);
383

384
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
385
					       struct ceph_snap_context *snapc,
386
					       unsigned int num_ops,
387
					       bool use_mempool,
388
					       gfp_t gfp_flags)
S
Sage Weil 已提交
389 390
{
	struct ceph_osd_request *req;
391

S
Sage Weil 已提交
392
	if (use_mempool) {
393
		BUG_ON(num_ops > CEPH_OSD_SLAB_OPS);
394
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
395 396
	} else if (num_ops <= CEPH_OSD_SLAB_OPS) {
		req = kmem_cache_alloc(ceph_osd_request_cache, gfp_flags);
S
Sage Weil 已提交
397
	} else {
398 399 400
		BUG_ON(num_ops > CEPH_OSD_MAX_OPS);
		req = kmalloc(sizeof(*req) + num_ops * sizeof(req->r_ops[0]),
			      gfp_flags);
S
Sage Weil 已提交
401
	}
402
	if (unlikely(!req))
403
		return NULL;
S
Sage Weil 已提交
404

405 406 407
	/* req only, each op is zeroed in _osd_req_op_init() */
	memset(req, 0, sizeof(*req));

S
Sage Weil 已提交
408 409
	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
410
	req->r_num_ops = num_ops;
411 412
	req->r_snapid = CEPH_NOSNAP;
	req->r_snapc = ceph_get_snap_context(snapc);
413

S
Sage Weil 已提交
414
	kref_init(&req->r_kref);
S
Sage Weil 已提交
415 416
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
417
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
418
	INIT_LIST_HEAD(&req->r_unsafe_item);
419
	INIT_LIST_HEAD(&req->r_linger_item);
420
	INIT_LIST_HEAD(&req->r_linger_osd_item);
421
	INIT_LIST_HEAD(&req->r_req_lru_item);
422 423
	INIT_LIST_HEAD(&req->r_osd_item);

424
	target_init(&req->r_t);
425

426 427 428 429
	dout("%s req %p\n", __func__, req);
	return req;
}
EXPORT_SYMBOL(ceph_osdc_alloc_request);
430

431 432 433 434 435
int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp)
{
	struct ceph_osd_client *osdc = req->r_osdc;
	struct ceph_msg *msg;
	int msg_size;
S
Sage Weil 已提交
436

437 438
	WARN_ON(ceph_oid_empty(&req->r_base_oid));

439
	/* create request message */
440 441 442 443
	msg_size = 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */
	msg_size += 4 + 4 + 4 + 8; /* mtime, reassert_version */
	msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
	msg_size += 1 + 8 + 4 + 4; /* pgid */
444 445
	msg_size += 4 + req->r_base_oid.name_len; /* oid */
	msg_size += 2 + req->r_num_ops * sizeof(struct ceph_osd_op);
446 447
	msg_size += 8; /* snapid */
	msg_size += 8; /* snap_seq */
448
	msg_size += 4 + 8 * (req->r_snapc ? req->r_snapc->num_snaps : 0);
449 450
	msg_size += 4; /* retry_attempt */

451
	if (req->r_mempool)
452
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
453
	else
454 455 456
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp, true);
	if (!msg)
		return -ENOMEM;
457

S
Sage Weil 已提交
458
	memset(msg->front.iov_base, 0, msg->front.iov_len);
459 460
	req->r_request = msg;

461 462
	/* create reply message */
	msg_size = OSD_OPREPLY_FRONT_LEN;
463 464
	msg_size += req->r_base_oid.name_len;
	msg_size += req->r_num_ops * sizeof(struct ceph_osd_op);
465 466 467 468 469 470 471 472 473 474 475

	if (req->r_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, msg_size, gfp, true);
	if (!msg)
		return -ENOMEM;

	req->r_reply = msg;

	return 0;
476
}
477
EXPORT_SYMBOL(ceph_osdc_alloc_messages);
478

479
static bool osd_req_opcode_valid(u16 opcode)
480
{
481
	switch (opcode) {
482 483 484
#define GENERATE_CASE(op, opcode, str)	case CEPH_OSD_OP_##op: return true;
__CEPH_FORALL_OSD_OPS(GENERATE_CASE)
#undef GENERATE_CASE
485 486 487 488 489
	default:
		return false;
	}
}

490 491 492 493 494
/*
 * This is an osd op init function for opcodes that have no data or
 * other information associated with them.  It also serves as a
 * common init routine for all the other init functions, below.
 */
495
static struct ceph_osd_req_op *
A
Alex Elder 已提交
496
_osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
497
		 u16 opcode, u32 flags)
498
{
499 500 501
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
502 503
	BUG_ON(!osd_req_opcode_valid(opcode));

504
	op = &osd_req->r_ops[which];
505 506
	memset(op, 0, sizeof (*op));
	op->op = opcode;
507
	op->flags = flags;
508 509

	return op;
510 511
}

A
Alex Elder 已提交
512
void osd_req_op_init(struct ceph_osd_request *osd_req,
513
		     unsigned int which, u16 opcode, u32 flags)
A
Alex Elder 已提交
514
{
515
	(void)_osd_req_op_init(osd_req, which, opcode, flags);
A
Alex Elder 已提交
516 517 518
}
EXPORT_SYMBOL(osd_req_op_init);

519 520
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
521 522 523
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
524 525
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
526 527
	size_t payload_len = 0;

L
Li Wang 已提交
528
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
529 530
	       opcode != CEPH_OSD_OP_WRITEFULL && opcode != CEPH_OSD_OP_ZERO &&
	       opcode != CEPH_OSD_OP_TRUNCATE);
531 532 533 534 535

	op->extent.offset = offset;
	op->extent.length = length;
	op->extent.truncate_size = truncate_size;
	op->extent.truncate_seq = truncate_seq;
536
	if (opcode == CEPH_OSD_OP_WRITE || opcode == CEPH_OSD_OP_WRITEFULL)
537 538
		payload_len += length;

539
	op->indata_len = payload_len;
540 541 542
}
EXPORT_SYMBOL(osd_req_op_extent_init);

543 544
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
545
{
546 547 548 549 550 551
	struct ceph_osd_req_op *op;
	u64 previous;

	BUG_ON(which >= osd_req->r_num_ops);
	op = &osd_req->r_ops[which];
	previous = op->extent.length;
552 553 554 555 556 557

	if (length == previous)
		return;		/* Nothing to do */
	BUG_ON(length > previous);

	op->extent.length = length;
558
	op->indata_len -= previous - length;
559 560 561
}
EXPORT_SYMBOL(osd_req_op_extent_update);

562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583
void osd_req_op_extent_dup_last(struct ceph_osd_request *osd_req,
				unsigned int which, u64 offset_inc)
{
	struct ceph_osd_req_op *op, *prev_op;

	BUG_ON(which + 1 >= osd_req->r_num_ops);

	prev_op = &osd_req->r_ops[which];
	op = _osd_req_op_init(osd_req, which + 1, prev_op->op, prev_op->flags);
	/* dup previous one */
	op->indata_len = prev_op->indata_len;
	op->outdata_len = prev_op->outdata_len;
	op->extent = prev_op->extent;
	/* adjust offset */
	op->extent.offset += offset_inc;
	op->extent.length -= offset_inc;

	if (op->op == CEPH_OSD_OP_WRITE || op->op == CEPH_OSD_OP_WRITEFULL)
		op->indata_len -= offset_inc;
}
EXPORT_SYMBOL(osd_req_op_extent_dup_last);

584
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
585
			u16 opcode, const char *class, const char *method)
586
{
587 588
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
589
	struct ceph_pagelist *pagelist;
590 591 592 593 594
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

595 596 597 598
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

599 600 601 602
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
603
	ceph_pagelist_append(pagelist, class, size);
604 605 606 607 608 609
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
610
	ceph_pagelist_append(pagelist, method, size);
611 612
	payload_len += size;

613
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
614

615
	op->indata_len = payload_len;
616 617
}
EXPORT_SYMBOL(osd_req_op_cls_init);
618

619 620 621 622
int osd_req_op_xattr_init(struct ceph_osd_request *osd_req, unsigned int which,
			  u16 opcode, const char *name, const void *value,
			  size_t size, u8 cmp_op, u8 cmp_mode)
{
623 624
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647
	struct ceph_pagelist *pagelist;
	size_t payload_len;

	BUG_ON(opcode != CEPH_OSD_OP_SETXATTR && opcode != CEPH_OSD_OP_CMPXATTR);

	pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS);
	if (!pagelist)
		return -ENOMEM;

	ceph_pagelist_init(pagelist);

	payload_len = strlen(name);
	op->xattr.name_len = payload_len;
	ceph_pagelist_append(pagelist, name, payload_len);

	op->xattr.value_len = size;
	ceph_pagelist_append(pagelist, value, size);
	payload_len += size;

	op->xattr.cmp_op = cmp_op;
	op->xattr.cmp_mode = cmp_mode;

	ceph_osd_data_pagelist_init(&op->xattr.osd_data, pagelist);
648
	op->indata_len = payload_len;
649 650 651 652
	return 0;
}
EXPORT_SYMBOL(osd_req_op_xattr_init);

653 654
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
655 656
				u64 cookie, u64 version, int flag)
{
657 658
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
659

660
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
661 662

	op->watch.cookie = cookie;
A
Alex Elder 已提交
663
	op->watch.ver = version;
664
	if (opcode == CEPH_OSD_OP_WATCH && flag)
665
		op->watch.flag = (u8)1;
666 667 668
}
EXPORT_SYMBOL(osd_req_op_watch_init);

669 670 671 672 673 674
void osd_req_op_alloc_hint_init(struct ceph_osd_request *osd_req,
				unsigned int which,
				u64 expected_object_size,
				u64 expected_write_size)
{
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
675 676
						      CEPH_OSD_OP_SETALLOCHINT,
						      0);
677 678 679 680 681 682 683 684 685 686 687 688 689

	op->alloc_hint.expected_object_size = expected_object_size;
	op->alloc_hint.expected_write_size = expected_write_size;

	/*
	 * CEPH_OSD_OP_SETALLOCHINT op is advisory and therefore deemed
	 * not worth a feature bit.  Set FAILOK per-op flag to make
	 * sure older osds don't trip over an unsupported opcode.
	 */
	op->flags |= CEPH_OSD_OP_FLAG_FAILOK;
}
EXPORT_SYMBOL(osd_req_op_alloc_hint_init);

690
static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
691 692 693 694 695 696 697
				struct ceph_osd_data *osd_data)
{
	u64 length = ceph_osd_data_length(osd_data);

	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
		BUG_ON(length > (u64) SIZE_MAX);
		if (length)
698
			ceph_msg_data_add_pages(msg, osd_data->pages,
699 700 701
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
702
		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
703 704
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
705
		ceph_msg_data_add_bio(msg, osd_data->bio, length);
706 707 708 709 710 711
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

712 713
static u32 osd_req_encode_op(struct ceph_osd_op *dst,
			     const struct ceph_osd_req_op *src)
714 715 716 717 718 719 720 721 722 723 724 725
{
	if (WARN_ON(!osd_req_opcode_valid(src->op))) {
		pr_err("unrecognized osd opcode %d\n", src->op);

		return 0;
	}

	switch (src->op) {
	case CEPH_OSD_OP_STAT:
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
726
	case CEPH_OSD_OP_WRITEFULL:
L
Li Wang 已提交
727 728
	case CEPH_OSD_OP_ZERO:
	case CEPH_OSD_OP_TRUNCATE:
729 730 731 732 733 734 735 736 737 738
		dst->extent.offset = cpu_to_le64(src->extent.offset);
		dst->extent.length = cpu_to_le64(src->extent.length);
		dst->extent.truncate_size =
			cpu_to_le64(src->extent.truncate_size);
		dst->extent.truncate_seq =
			cpu_to_le32(src->extent.truncate_seq);
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
739
		dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
740 741 742 743 744 745 746 747 748
		break;
	case CEPH_OSD_OP_STARTSYNC:
		break;
	case CEPH_OSD_OP_NOTIFY_ACK:
	case CEPH_OSD_OP_WATCH:
		dst->watch.cookie = cpu_to_le64(src->watch.cookie);
		dst->watch.ver = cpu_to_le64(src->watch.ver);
		dst->watch.flag = src->watch.flag;
		break;
749 750 751 752 753 754
	case CEPH_OSD_OP_SETALLOCHINT:
		dst->alloc_hint.expected_object_size =
		    cpu_to_le64(src->alloc_hint.expected_object_size);
		dst->alloc_hint.expected_write_size =
		    cpu_to_le64(src->alloc_hint.expected_write_size);
		break;
755 756 757 758 759 760 761
	case CEPH_OSD_OP_SETXATTR:
	case CEPH_OSD_OP_CMPXATTR:
		dst->xattr.name_len = cpu_to_le32(src->xattr.name_len);
		dst->xattr.value_len = cpu_to_le32(src->xattr.value_len);
		dst->xattr.cmp_op = src->xattr.cmp_op;
		dst->xattr.cmp_mode = src->xattr.cmp_mode;
		break;
762 763 764
	case CEPH_OSD_OP_CREATE:
	case CEPH_OSD_OP_DELETE:
		break;
765
	default:
766
		pr_err("unsupported osd opcode %s\n",
767
			ceph_osd_op_name(src->op));
768
		WARN_ON(1);
769 770

		return 0;
771
	}
772

773
	dst->op = cpu_to_le16(src->op);
774
	dst->flags = cpu_to_le32(src->flags);
775
	dst->payload_len = cpu_to_le32(src->indata_len);
776

777
	return src->indata_len;
778 779
}

780 781 782 783 784 785 786 787 788 789 790 791 792 793
/*
 * build new request AND message, calculate layout, and adjust file
 * extent as needed.
 *
 * if the file was recently truncated, we include information about its
 * old and new size so that the object can be updated appropriately.  (we
 * avoid synchronously deleting truncated objects because it's slow.)
 *
 * if @do_sync, include a 'startsync' command so that the osd will flush
 * data quickly.
 */
struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
					       struct ceph_file_layout *layout,
					       struct ceph_vino vino,
794 795
					       u64 off, u64 *plen,
					       unsigned int which, int num_ops,
796 797 798 799
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
800
					       bool use_mempool)
801
{
802
	struct ceph_osd_request *req;
803 804 805
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
806
	int r;
807

L
Li Wang 已提交
808
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
809 810
	       opcode != CEPH_OSD_OP_ZERO && opcode != CEPH_OSD_OP_TRUNCATE &&
	       opcode != CEPH_OSD_OP_CREATE && opcode != CEPH_OSD_OP_DELETE);
811

812
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
813
					GFP_NOFS);
814 815 816 817
	if (!req) {
		r = -ENOMEM;
		goto fail;
	}
818

819
	/* calculate max write size */
820
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
821 822
	if (r)
		goto fail;
823

824
	if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) {
825
		osd_req_op_init(req, which, opcode, 0);
826 827 828 829 830 831 832 833 834 835 836
	} else {
		u32 object_size = le32_to_cpu(layout->fl_object_size);
		u32 object_base = off - objoff;
		if (!(truncate_seq == 1 && truncate_size == -1ULL)) {
			if (truncate_size <= object_base) {
				truncate_size = 0;
			} else {
				truncate_size -= object_base;
				if (truncate_size > object_size)
					truncate_size = object_size;
			}
837
		}
838
		osd_req_op_extent_init(req, which, opcode, objoff, objlen,
839 840
				       truncate_size, truncate_seq);
	}
841

842
	req->r_flags = flags;
843
	req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout);
844
	ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum);
845

846 847 848 849
	req->r_snapid = vino.snap;
	if (flags & CEPH_OSD_FLAG_WRITE)
		req->r_data_offset = off;

850 851 852 853
	r = ceph_osdc_alloc_messages(req, GFP_NOFS);
	if (r)
		goto fail;

S
Sage Weil 已提交
854
	return req;
855 856 857 858

fail:
	ceph_osdc_put_request(req);
	return ERR_PTR(r);
S
Sage Weil 已提交
859
}
860
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
861 862 863 864

/*
 * We keep osd requests in an rbtree, sorted by ->r_tid.
 */
I
Ilya Dryomov 已提交
865
DEFINE_RB_FUNCS(request, struct ceph_osd_request, r_tid, r_node)
S
Sage Weil 已提交
866

867 868 869 870 871
static bool osd_homeless(struct ceph_osd *osd)
{
	return osd->o_osd == CEPH_HOMELESS_OSD;
}

S
Sage Weil 已提交
872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893
static struct ceph_osd_request *
__lookup_request_ge(struct ceph_osd_client *osdc,
		    u64 tid)
{
	struct ceph_osd_request *req;
	struct rb_node *n = osdc->requests.rb_node;

	while (n) {
		req = rb_entry(n, struct ceph_osd_request, r_node);
		if (tid < req->r_tid) {
			if (!n->rb_left)
				return req;
			n = n->rb_left;
		} else if (tid > req->r_tid) {
			n = n->rb_right;
		} else {
			return req;
		}
	}
	return NULL;
}

894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924
static void __kick_linger_request(struct ceph_osd_request *req)
{
	struct ceph_osd_client *osdc = req->r_osdc;
	struct ceph_osd *osd = req->r_osd;

	/*
	 * Linger requests need to be resent with a new tid to avoid
	 * the dup op detection logic on the OSDs.  Achieve this with
	 * a re-register dance instead of open-coding.
	 */
	ceph_osdc_get_request(req);
	if (!list_empty(&req->r_linger_item))
		__unregister_linger_request(osdc, req);
	else
		__unregister_request(osdc, req);
	__register_request(osdc, req);
	ceph_osdc_put_request(req);

	/*
	 * Unless request has been registered as both normal and
	 * lingering, __unregister{,_linger}_request clears r_osd.
	 * However, here we need to preserve r_osd to make sure we
	 * requeue on the same OSD.
	 */
	WARN_ON(req->r_osd || !osd);
	req->r_osd = osd;

	dout("%s requeueing %p tid %llu\n", __func__, req, req->r_tid);
	__enqueue_request(req);
}

925 926 927 928 929 930
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
931
	struct ceph_osd_request *req, *nreq;
932
	LIST_HEAD(resend);
933
	LIST_HEAD(resend_linger);
934 935
	int err;

936
	dout("%s osd%d\n", __func__, osd->o_osd);
937
	err = __reset_osd(osdc, osd);
938
	if (err)
939
		return;
940

941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957
	/*
	 * Build up a list of requests to resend by traversing the
	 * osd's list of requests.  Requests for a given object are
	 * sent in tid order, and that is also the order they're
	 * kept on this list.  Therefore all requests that are in
	 * flight will be found first, followed by all requests that
	 * have not yet been sent.  And to resend requests while
	 * preserving this order we will want to put any sent
	 * requests back on the front of the osd client's unsent
	 * list.
	 *
	 * So we build a separate ordered list of already-sent
	 * requests for the affected osd and splice it onto the
	 * front of the osd client's unsent list.  Once we've seen a
	 * request that has not yet been sent we're done.  Those
	 * requests are already sitting right where they belong.
	 */
958
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
959 960
		if (!req->r_sent)
			break;
961 962 963 964 965

		if (!req->r_linger) {
			dout("%s requeueing %p tid %llu\n", __func__, req,
			     req->r_tid);
			list_move_tail(&req->r_req_lru_item, &resend);
966
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
967 968 969
		} else {
			list_move_tail(&req->r_req_lru_item, &resend_linger);
		}
970
	}
971
	list_splice(&resend, &osdc->req_unsent);
972

973
	/*
974 975 976 977
	 * Both registered and not yet registered linger requests are
	 * enqueued with a new tid on the same OSD.  We add/move them
	 * to req_unsent/o_requests at the end to keep things in tid
	 * order.
978
	 */
979
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
980
				 r_linger_osd_item) {
981 982
		WARN_ON(!list_empty(&req->r_req_lru_item));
		__kick_linger_request(req);
983
	}
984 985 986

	list_for_each_entry_safe(req, nreq, &resend_linger, r_req_lru_item)
		__kick_linger_request(req);
987 988
}

S
Sage Weil 已提交
989
/*
990
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
991 992 993 994 995 996 997 998 999 1000 1001
 */
static void osd_reset(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc;

	if (!osd)
		return;
	dout("osd_reset osd%d\n", osd->o_osd);
	osdc = osd->o_osdc;
	down_read(&osdc->map_sem);
S
Sage Weil 已提交
1002 1003
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
1004
	__send_queued(osdc);
S
Sage Weil 已提交
1005
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1006 1007 1008
	up_read(&osdc->map_sem);
}

1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
/*
 * Assumes @osd is zero-initialized.
 */
static void osd_init(struct ceph_osd *osd)
{
	atomic_set(&osd->o_ref, 1);
	RB_CLEAR_NODE(&osd->o_node);
	INIT_LIST_HEAD(&osd->o_requests);
	INIT_LIST_HEAD(&osd->o_linger_requests);
	INIT_LIST_HEAD(&osd->o_osd_lru);
	INIT_LIST_HEAD(&osd->o_keepalive_item);
	osd->o_incarnation = 1;
}

static void osd_cleanup(struct ceph_osd *osd)
{
	WARN_ON(!RB_EMPTY_NODE(&osd->o_node));
	WARN_ON(!list_empty(&osd->o_requests));
	WARN_ON(!list_empty(&osd->o_linger_requests));
	WARN_ON(!list_empty(&osd->o_osd_lru));
	WARN_ON(!list_empty(&osd->o_keepalive_item));

	if (osd->o_auth.authorizer) {
		WARN_ON(osd_homeless(osd));
		ceph_auth_destroy_authorizer(osd->o_auth.authorizer);
	}
}

S
Sage Weil 已提交
1037 1038 1039
/*
 * Track open sessions with osds.
 */
1040
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
1041 1042 1043
{
	struct ceph_osd *osd;

1044 1045
	WARN_ON(onum == CEPH_HOMELESS_OSD);

1046
	osd = kzalloc(sizeof(*osd), GFP_NOIO | __GFP_NOFAIL);
1047
	osd_init(osd);
S
Sage Weil 已提交
1048
	osd->o_osdc = osdc;
1049
	osd->o_osd = onum;
S
Sage Weil 已提交
1050

1051
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
1052

S
Sage Weil 已提交
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
	return osd;
}

static struct ceph_osd *get_osd(struct ceph_osd *osd)
{
	if (atomic_inc_not_zero(&osd->o_ref)) {
		dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
		     atomic_read(&osd->o_ref));
		return osd;
	} else {
		dout("get_osd %p FAIL\n", osd);
		return NULL;
	}
}

static void put_osd(struct ceph_osd *osd)
{
	dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
	     atomic_read(&osd->o_ref) - 1);
1072
	if (atomic_dec_and_test(&osd->o_ref)) {
1073
		osd_cleanup(osd);
S
Sage Weil 已提交
1074
		kfree(osd);
S
Sage Weil 已提交
1075
	}
S
Sage Weil 已提交
1076 1077
}

I
Ilya Dryomov 已提交
1078 1079
DEFINE_RB_FUNCS(osd, struct ceph_osd, o_osd, o_node)

S
Sage Weil 已提交
1080 1081 1082
/*
 * remove an osd from our map
 */
1083
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1084
{
1085
	dout("%s %p osd%d\n", __func__, osd, osd->o_osd);
1086 1087
	WARN_ON(!list_empty(&osd->o_requests));
	WARN_ON(!list_empty(&osd->o_linger_requests));
1088

1089
	list_del_init(&osd->o_osd_lru);
I
Ilya Dryomov 已提交
1090
	erase_osd(&osdc->osds, osd);
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
}

static void remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
{
	dout("%s %p osd%d\n", __func__, osd, osd->o_osd);

	if (!RB_EMPTY_NODE(&osd->o_node)) {
		ceph_con_close(&osd->o_con);
		__remove_osd(osdc, osd);
		put_osd(osd);
	}
S
Sage Weil 已提交
1102 1103
}

1104
static void __move_osd_to_lru(struct ceph_osd *osd)
1105
{
1106 1107 1108
	struct ceph_osd_client *osdc = osd->o_osdc;

	dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
1109
	BUG_ON(!list_empty(&osd->o_osd_lru));
1110

1111
	spin_lock(&osdc->osd_lru_lock);
1112
	list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
1113 1114
	spin_unlock(&osdc->osd_lru_lock);

1115
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl;
1116 1117
}

1118
static void maybe_move_osd_to_lru(struct ceph_osd *osd)
1119 1120 1121
{
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests))
1122
		__move_osd_to_lru(osd);
1123 1124
}

1125 1126
static void __remove_osd_from_lru(struct ceph_osd *osd)
{
1127 1128 1129 1130 1131
	struct ceph_osd_client *osdc = osd->o_osdc;

	dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);

	spin_lock(&osdc->osd_lru_lock);
1132 1133
	if (!list_empty(&osd->o_osd_lru))
		list_del_init(&osd->o_osd_lru);
1134
	spin_unlock(&osdc->osd_lru_lock);
1135 1136
}

S
Sage Weil 已提交
1137 1138 1139
/*
 * reset osd connect
 */
1140
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1141
{
A
Alex Elder 已提交
1142
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
1143

1144
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
1145 1146
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
1147
		remove_osd(osdc, osd);
A
Alex Elder 已提交
1148 1149 1150 1151 1152 1153 1154 1155
		return -ENODEV;
	}

	peer_addr = &osdc->osdmap->osd_addr[osd->o_osd];
	if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) &&
			!ceph_con_opened(&osd->o_con)) {
		struct ceph_osd_request *req;

1156 1157
		dout("osd addr hasn't changed and connection never opened, "
		     "letting msgr retry\n");
1158 1159 1160
		/* touch each r_stamp for handle_timeout()'s benfit */
		list_for_each_entry(req, &osd->o_requests, r_osd_item)
			req->r_stamp = jiffies;
A
Alex Elder 已提交
1161 1162

		return -EAGAIN;
S
Sage Weil 已提交
1163
	}
A
Alex Elder 已提交
1164 1165 1166 1167 1168 1169

	ceph_con_close(&osd->o_con);
	ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, peer_addr);
	osd->o_incarnation++;

	return 0;
S
Sage Weil 已提交
1170 1171 1172 1173 1174 1175
}

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1176 1177
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1178 1179
{
	req->r_tid = ++osdc->last_tid;
1180
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1181
	dout("__register_request %p tid %lld\n", req, req->r_tid);
I
Ilya Dryomov 已提交
1182
	insert_request(&osdc->requests, req);
S
Sage Weil 已提交
1183 1184
	ceph_osdc_get_request(req);
	osdc->num_requests++;
1185 1186
}

S
Sage Weil 已提交
1187 1188 1189 1190 1191 1192
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1193 1194 1195 1196 1197 1198
	if (RB_EMPTY_NODE(&req->r_node)) {
		dout("__unregister_request %p tid %lld not registered\n",
			req, req->r_tid);
		return;
	}

S
Sage Weil 已提交
1199
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
I
Ilya Dryomov 已提交
1200
	erase_request(&osdc->requests, req);
S
Sage Weil 已提交
1201 1202
	osdc->num_requests--;

1203 1204
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1205
		ceph_msg_revoke(req->r_request);
1206 1207

		list_del_init(&req->r_osd_item);
1208
		maybe_move_osd_to_lru(req->r_osd);
1209
		if (list_empty(&req->r_linger_osd_item))
1210
			req->r_osd = NULL;
1211
	}
S
Sage Weil 已提交
1212

A
Alex Elder 已提交
1213
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1214 1215 1216 1217 1218 1219 1220 1221
	ceph_osdc_put_request(req);
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1222
	if (req->r_sent && req->r_osd) {
1223
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1224 1225 1226 1227
		req->r_sent = 0;
	}
}

1228 1229 1230
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
1231 1232 1233
	dout("%s %p tid %llu\n", __func__, req, req->r_tid);
	WARN_ON(!req->r_linger);

1234
	ceph_osdc_get_request(req);
1235
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1236
	if (req->r_osd)
1237
		list_add_tail(&req->r_linger_osd_item,
1238
			      &req->r_osd->o_linger_requests);
1239 1240 1241 1242 1243
}

static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req)
{
1244 1245 1246 1247 1248 1249 1250 1251 1252
	WARN_ON(!req->r_linger);

	if (list_empty(&req->r_linger_item)) {
		dout("%s %p tid %llu not registered\n", __func__, req,
		     req->r_tid);
		return;
	}

	dout("%s %p tid %llu\n", __func__, req, req->r_tid);
A
Alex Elder 已提交
1253
	list_del_init(&req->r_linger_item);
1254

1255
	if (req->r_osd) {
1256
		list_del_init(&req->r_linger_osd_item);
1257
		maybe_move_osd_to_lru(req->r_osd);
1258 1259
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1260
	}
1261
	ceph_osdc_put_request(req);
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
}

void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
				  struct ceph_osd_request *req)
{
	if (!req->r_linger) {
		dout("set_request_linger %p\n", req);
		req->r_linger = 1;
	}
}
EXPORT_SYMBOL(ceph_osdc_set_request_linger);

1274 1275 1276 1277 1278
static bool __pool_full(struct ceph_pg_pool_info *pi)
{
	return pi->flags & CEPH_POOL_FLAG_FULL;
}

I
Ilya Dryomov 已提交
1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293
static bool have_pool_full(struct ceph_osd_client *osdc)
{
	struct rb_node *n;

	for (n = rb_first(&osdc->osdmap->pg_pools); n; n = rb_next(n)) {
		struct ceph_pg_pool_info *pi =
		    rb_entry(n, struct ceph_pg_pool_info, node);

		if (__pool_full(pi))
			return true;
	}

	return false;
}

1294 1295 1296 1297 1298 1299
/*
 * Returns whether a request should be blocked from being sent
 * based on the current osdmap and osd_client settings.
 *
 * Caller should hold map_sem for read.
 */
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
static bool target_should_be_paused(struct ceph_osd_client *osdc,
				    const struct ceph_osd_request_target *t,
				    struct ceph_pg_pool_info *pi)
{
	bool pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD);
	bool pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) ||
		       ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
		       __pool_full(pi);

	WARN_ON(pi->id != t->base_oloc.pool);
	return (t->flags & CEPH_OSD_FLAG_READ && pauserd) ||
	       (t->flags & CEPH_OSD_FLAG_WRITE && pausewr);
}

enum calc_target_result {
	CALC_TARGET_NO_ACTION = 0,
	CALC_TARGET_NEED_RESEND,
	CALC_TARGET_POOL_DNE,
};

static enum calc_target_result calc_target(struct ceph_osd_client *osdc,
					   struct ceph_osd_request_target *t,
					   u32 *last_force_resend,
					   bool any_change)
{
	struct ceph_pg_pool_info *pi;
	struct ceph_pg pgid, last_pgid;
	struct ceph_osds up, acting;
	bool force_resend = false;
	bool need_check_tiering = false;
	bool need_resend = false;
	bool sort_bitwise = ceph_osdmap_flag(osdc->osdmap,
					     CEPH_OSDMAP_SORTBITWISE);
	enum calc_target_result ct_res;
	int ret;

	pi = ceph_pg_pool_by_id(osdc->osdmap, t->base_oloc.pool);
	if (!pi) {
		t->osd = CEPH_HOMELESS_OSD;
		ct_res = CALC_TARGET_POOL_DNE;
		goto out;
	}

	if (osdc->osdmap->epoch == pi->last_force_request_resend) {
		if (last_force_resend &&
		    *last_force_resend < pi->last_force_request_resend) {
			*last_force_resend = pi->last_force_request_resend;
			force_resend = true;
		} else if (!last_force_resend) {
			force_resend = true;
		}
	}
	if (ceph_oid_empty(&t->target_oid) || force_resend) {
		ceph_oid_copy(&t->target_oid, &t->base_oid);
		need_check_tiering = true;
	}
	if (ceph_oloc_empty(&t->target_oloc) || force_resend) {
		ceph_oloc_copy(&t->target_oloc, &t->base_oloc);
		need_check_tiering = true;
	}

	if (need_check_tiering &&
	    (t->flags & CEPH_OSD_FLAG_IGNORE_OVERLAY) == 0) {
		if (t->flags & CEPH_OSD_FLAG_READ && pi->read_tier >= 0)
			t->target_oloc.pool = pi->read_tier;
		if (t->flags & CEPH_OSD_FLAG_WRITE && pi->write_tier >= 0)
			t->target_oloc.pool = pi->write_tier;
	}

	ret = ceph_object_locator_to_pg(osdc->osdmap, &t->target_oid,
					&t->target_oloc, &pgid);
	if (ret) {
		WARN_ON(ret != -ENOENT);
		t->osd = CEPH_HOMELESS_OSD;
		ct_res = CALC_TARGET_POOL_DNE;
		goto out;
	}
	last_pgid.pool = pgid.pool;
	last_pgid.seed = ceph_stable_mod(pgid.seed, t->pg_num, t->pg_num_mask);

	ceph_pg_to_up_acting_osds(osdc->osdmap, &pgid, &up, &acting);
	if (any_change &&
	    ceph_is_new_interval(&t->acting,
				 &acting,
				 &t->up,
				 &up,
				 t->size,
				 pi->size,
				 t->min_size,
				 pi->min_size,
				 t->pg_num,
				 pi->pg_num,
				 t->sort_bitwise,
				 sort_bitwise,
				 &last_pgid))
		force_resend = true;

	if (t->paused && !target_should_be_paused(osdc, t, pi)) {
		t->paused = false;
		need_resend = true;
	}

	if (ceph_pg_compare(&t->pgid, &pgid) ||
	    ceph_osds_changed(&t->acting, &acting, any_change) ||
	    force_resend) {
		t->pgid = pgid; /* struct */
		ceph_osds_copy(&t->acting, &acting);
		ceph_osds_copy(&t->up, &up);
		t->size = pi->size;
		t->min_size = pi->min_size;
		t->pg_num = pi->pg_num;
		t->pg_num_mask = pi->pg_num_mask;
		t->sort_bitwise = sort_bitwise;

		t->osd = acting.primary;
		need_resend = true;
	}

	ct_res = need_resend ? CALC_TARGET_NEED_RESEND : CALC_TARGET_NO_ACTION;
out:
	dout("%s t %p -> ct_res %d osd %d\n", __func__, t, ct_res, t->osd);
	return ct_res;
}

1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439
static void __enqueue_request(struct ceph_osd_request *req)
{
	struct ceph_osd_client *osdc = req->r_osdc;

	dout("%s %p tid %llu to osd%d\n", __func__, req, req->r_tid,
	     req->r_osd ? req->r_osd->o_osd : -1);

	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
	} else {
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
	}
}

S
Sage Weil 已提交
1440 1441 1442
/*
 * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
 * (as needed), and set the request r_osd appropriately.  If there is
L
Lucas De Marchi 已提交
1443
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1444
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1445 1446 1447 1448 1449
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1450
static int __map_request(struct ceph_osd_client *osdc,
1451
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1452
{
1453
	enum calc_target_result ct_res;
S
Sage Weil 已提交
1454 1455
	int err;

1456
	dout("map_request %p tid %lld\n", req, req->r_tid);
1457

1458 1459 1460
	ct_res = calc_target(osdc, &req->r_t, NULL, force_resend);
	switch (ct_res) {
	case CALC_TARGET_POOL_DNE:
1461
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
1462 1463
		return -EIO;
	case CALC_TARGET_NO_ACTION:
S
Sage Weil 已提交
1464
		return 0;  /* no change */
1465 1466 1467
	default:
		BUG_ON(ct_res != CALC_TARGET_NEED_RESEND);
	}
S
Sage Weil 已提交
1468

1469
	dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
1470
	     req->r_tid, req->r_t.pgid.pool, req->r_t.pgid.seed, req->r_t.osd,
S
Sage Weil 已提交
1471 1472 1473 1474 1475
	     req->r_osd ? req->r_osd->o_osd : -1);

	if (req->r_osd) {
		__cancel_request(req);
		list_del_init(&req->r_osd_item);
1476
		list_del_init(&req->r_linger_osd_item);
S
Sage Weil 已提交
1477 1478 1479
		req->r_osd = NULL;
	}

1480 1481
	req->r_osd = lookup_osd(&osdc->osds, req->r_t.osd);
	if (!req->r_osd && req->r_t.osd >= 0) {
1482
		err = -ENOMEM;
1483
		req->r_osd = create_osd(osdc, req->r_t.osd);
1484 1485
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1486
			goto out;
1487
		}
S
Sage Weil 已提交
1488

1489
		dout("map_request osd %p is osd%d\n", req->r_osd,
1490
		     req->r_osd->o_osd);
I
Ilya Dryomov 已提交
1491
		insert_osd(&osdc->osds, req->r_osd);
S
Sage Weil 已提交
1492

1493
		ceph_con_open(&req->r_osd->o_con,
1494 1495
			      CEPH_ENTITY_TYPE_OSD, req->r_osd->o_osd,
			      &osdc->osdmap->osd_addr[req->r_osd->o_osd]);
S
Sage Weil 已提交
1496 1497
	}

1498
	__enqueue_request(req);
1499
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1500 1501 1502 1503 1504

out:
	return err;
}

1505 1506
static void setup_request_data(struct ceph_osd_request *req,
			       struct ceph_msg *msg)
S
Sage Weil 已提交
1507
{
1508 1509 1510 1511 1512
	u32 data_len = 0;
	int i;

	if (!list_empty(&msg->data))
		return;
S
Sage Weil 已提交
1513

1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
	WARN_ON(msg->data_length);
	for (i = 0; i < req->r_num_ops; i++) {
		struct ceph_osd_req_op *op = &req->r_ops[i];

		switch (op->op) {
		/* request */
		case CEPH_OSD_OP_WRITE:
		case CEPH_OSD_OP_WRITEFULL:
			WARN_ON(op->indata_len != op->extent.length);
			ceph_osdc_msg_data_add(msg, &op->extent.osd_data);
			break;
		case CEPH_OSD_OP_SETXATTR:
		case CEPH_OSD_OP_CMPXATTR:
			WARN_ON(op->indata_len != op->xattr.name_len +
						  op->xattr.value_len);
			ceph_osdc_msg_data_add(msg, &op->xattr.osd_data);
			break;

		/* reply */
		case CEPH_OSD_OP_STAT:
			ceph_osdc_msg_data_add(req->r_reply,
					       &op->raw_data_in);
			break;
		case CEPH_OSD_OP_READ:
			ceph_osdc_msg_data_add(req->r_reply,
					       &op->extent.osd_data);
			break;

		/* both */
		case CEPH_OSD_OP_CALL:
			WARN_ON(op->indata_len != op->cls.class_len +
						  op->cls.method_len +
						  op->cls.indata_len);
			ceph_osdc_msg_data_add(msg, &op->cls.request_info);
			/* optional, can be NONE */
			ceph_osdc_msg_data_add(msg, &op->cls.request_data);
			/* optional, can be NONE */
			ceph_osdc_msg_data_add(req->r_reply,
					       &op->cls.response_data);
			break;
		}

		data_len += op->indata_len;
	}
1558

1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597
	WARN_ON(data_len != msg->data_length);
}

static void encode_request(struct ceph_osd_request *req, struct ceph_msg *msg)
{
	void *p = msg->front.iov_base;
	void *const end = p + msg->front_alloc_len;
	u32 data_len = 0;
	int i;

	if (req->r_flags & CEPH_OSD_FLAG_WRITE) {
		/* snapshots aren't writeable */
		WARN_ON(req->r_snapid != CEPH_NOSNAP);
	} else {
		WARN_ON(req->r_mtime.tv_sec || req->r_mtime.tv_nsec ||
			req->r_data_offset || req->r_snapc);
	}

	setup_request_data(req, msg);

	ceph_encode_32(&p, 1); /* client_inc, always 1 */
	ceph_encode_32(&p, req->r_osdc->osdmap->epoch);
	ceph_encode_32(&p, req->r_flags);
	ceph_encode_timespec(p, &req->r_mtime);
	p += sizeof(struct ceph_timespec);
	/* aka reassert_version */
	memcpy(p, &req->r_replay_version, sizeof(req->r_replay_version));
	p += sizeof(req->r_replay_version);

	/* oloc */
	ceph_encode_8(&p, 4);
	ceph_encode_8(&p, 4);
	ceph_encode_32(&p, 8 + 4 + 4);
	ceph_encode_64(&p, req->r_t.target_oloc.pool);
	ceph_encode_32(&p, -1); /* preferred */
	ceph_encode_32(&p, 0); /* key len */

	/* pgid */
	ceph_encode_8(&p, 1);
1598 1599
	ceph_encode_64(&p, req->r_t.pgid.pool);
	ceph_encode_32(&p, req->r_t.pgid.seed);
1600
	ceph_encode_32(&p, -1); /* preferred */
1601

1602 1603 1604 1605
	/* oid */
	ceph_encode_32(&p, req->r_t.target_oid.name_len);
	memcpy(p, req->r_t.target_oid.name, req->r_t.target_oid.name_len);
	p += req->r_t.target_oid.name_len;
S
Sage Weil 已提交
1606

1607 1608 1609 1610 1611 1612
	/* ops, can imply data */
	ceph_encode_16(&p, req->r_num_ops);
	for (i = 0; i < req->r_num_ops; i++) {
		data_len += osd_req_encode_op(p, &req->r_ops[i]);
		p += sizeof(struct ceph_osd_op);
	}
1613

1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637
	ceph_encode_64(&p, req->r_snapid); /* snapid */
	if (req->r_snapc) {
		ceph_encode_64(&p, req->r_snapc->seq);
		ceph_encode_32(&p, req->r_snapc->num_snaps);
		for (i = 0; i < req->r_snapc->num_snaps; i++)
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
	} else {
		ceph_encode_64(&p, 0); /* snap_seq */
		ceph_encode_32(&p, 0); /* snaps len */
	}

	ceph_encode_32(&p, req->r_attempts); /* retry_attempt */

	BUG_ON(p > end);
	msg->front.iov_len = p - msg->front.iov_base;
	msg->hdr.version = cpu_to_le16(4); /* MOSDOp v4 */
	msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
	msg->hdr.data_len = cpu_to_le32(data_len);
	/*
	 * The header "data_off" is a hint to the receiver allowing it
	 * to align received data into its buffers such that there's no
	 * need to re-copy it before writing it to disk (direct I/O).
	 */
	msg->hdr.data_off = cpu_to_le16(req->r_data_offset);
1638

1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671
	dout("%s req %p oid %*pE oid_len %d front %zu data %u\n", __func__,
	     req, req->r_t.target_oid.name_len, req->r_t.target_oid.name,
	     req->r_t.target_oid.name_len, msg->front.iov_len, data_len);
}

/*
 * @req has to be assigned a tid and registered.
 */
static void send_request(struct ceph_osd_request *req)
{
	struct ceph_osd *osd = req->r_osd;

	WARN_ON(osd->o_osd != req->r_t.osd);

	req->r_flags |= CEPH_OSD_FLAG_KNOWN_REDIR;
	if (req->r_attempts)
		req->r_flags |= CEPH_OSD_FLAG_RETRY;
	else
		WARN_ON(req->r_flags & CEPH_OSD_FLAG_RETRY);

	encode_request(req, req->r_request);

	dout("%s req %p tid %llu to pg %llu.%x osd%d flags 0x%x attempt %d\n",
	     __func__, req, req->r_tid, req->r_t.pgid.pool, req->r_t.pgid.seed,
	     req->r_t.osd, req->r_flags, req->r_attempts);

	req->r_t.paused = false;
	req->r_stamp = jiffies;
	req->r_attempts++;

	req->r_sent = osd->o_incarnation;
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
	ceph_con_send(&osd->o_con, ceph_msg_get(req->r_request));
S
Sage Weil 已提交
1672 1673
}

1674 1675 1676
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1677
static void __send_queued(struct ceph_osd_client *osdc)
1678 1679 1680
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1681
	dout("__send_queued\n");
1682 1683 1684 1685
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) {
		list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
		send_request(req);
	}
1686 1687
}

I
Ilya Dryomov 已提交
1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707
static void maybe_request_map(struct ceph_osd_client *osdc)
{
	bool continuous = false;

	WARN_ON(!osdc->osdmap->epoch);

	if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
	    ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD) ||
	    ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR)) {
		dout("%s osdc %p continuous\n", __func__, osdc);
		continuous = true;
	} else {
		dout("%s osdc %p onetime\n", __func__, osdc);
	}

	if (ceph_monc_want_map(&osdc->client->monc, CEPH_SUB_OSDMAP,
			       osdc->osdmap->epoch + 1, continuous))
		ceph_monc_renew_subs(&osdc->client->monc);
}

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
/*
 * Caller should hold map_sem for read and request_mutex.
 */
static int __ceph_osdc_start_request(struct ceph_osd_client *osdc,
				     struct ceph_osd_request *req,
				     bool nofail)
{
	int rc;

	__register_request(osdc, req);
	req->r_sent = 0;
	req->r_got_reply = 0;
	rc = __map_request(osdc, req, 0);
	if (rc < 0) {
		if (nofail) {
			dout("osdc_start_request failed map, "
				" will retry %lld\n", req->r_tid);
			rc = 0;
		} else {
			__unregister_request(osdc, req);
		}
		return rc;
	}

	if (req->r_osd == NULL) {
		dout("send_request %p no up osds in pg\n", req);
		ceph_monc_request_next_osdmap(&osdc->client->monc);
	} else {
		__send_queued(osdc);
	}

	return 0;
}

1742 1743 1744 1745 1746 1747 1748 1749
static void __complete_request(struct ceph_osd_request *req)
{
	if (req->r_callback)
		req->r_callback(req);
	else
		complete_all(&req->r_completion);
}

S
Sage Weil 已提交
1750
/*
1751 1752 1753 1754
 * Timeout callback, called every N seconds.  When 1 or more OSD
 * requests has been active for more than N seconds, we send a keepalive
 * (tag + timestamp) to its OSD to ensure any communications channel
 * reset is detected.
S
Sage Weil 已提交
1755 1756 1757 1758 1759
 */
static void handle_timeout(struct work_struct *work)
{
	struct ceph_osd_client *osdc =
		container_of(work, struct ceph_osd_client, timeout_work.work);
1760
	struct ceph_options *opts = osdc->client->options;
S
Sage Weil 已提交
1761
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1762
	struct ceph_osd *osd;
1763
	struct list_head slow_osds;
S
Sage Weil 已提交
1764 1765 1766 1767 1768 1769 1770
	dout("timeout\n");
	down_read(&osdc->map_sem);

	ceph_monc_request_next_osdmap(&osdc->client->monc);

	mutex_lock(&osdc->request_mutex);

1771 1772 1773 1774 1775 1776 1777
	/*
	 * ping osds that are a bit slow.  this ensures that if there
	 * is a break in the TCP connection we will notice, and reopen
	 * a connection with that osd (from the fault callback).
	 */
	INIT_LIST_HEAD(&slow_osds);
	list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
1778 1779
		if (time_before(jiffies,
				req->r_stamp + opts->osd_keepalive_timeout))
1780 1781 1782 1783 1784
			break;

		osd = req->r_osd;
		BUG_ON(!osd);
		dout(" tid %llu is slow, will send keepalive on osd%d\n",
S
Sage Weil 已提交
1785
		     req->r_tid, osd->o_osd);
1786 1787 1788 1789 1790 1791
		list_move_tail(&osd->o_keepalive_item, &slow_osds);
	}
	while (!list_empty(&slow_osds)) {
		osd = list_entry(slow_osds.next, struct ceph_osd,
				 o_keepalive_item);
		list_del_init(&osd->o_keepalive_item);
S
Sage Weil 已提交
1792 1793 1794
		ceph_con_keepalive(&osd->o_con);
	}

A
Alex Elder 已提交
1795
	__send_queued(osdc);
S
Sage Weil 已提交
1796 1797
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
1798 1799 1800

	schedule_delayed_work(&osdc->timeout_work,
			      osdc->client->options->osd_keepalive_timeout);
S
Sage Weil 已提交
1801 1802
}

1803 1804 1805 1806 1807
static void handle_osds_timeout(struct work_struct *work)
{
	struct ceph_osd_client *osdc =
		container_of(work, struct ceph_osd_client,
			     osds_timeout_work.work);
1808
	unsigned long delay = osdc->client->options->osd_idle_ttl / 4;
1809
	struct ceph_osd *osd, *nosd;
1810

1811
	dout("%s osdc %p\n", __func__, osdc);
1812
	down_read(&osdc->map_sem);
1813 1814 1815 1816 1817 1818 1819 1820
	mutex_lock(&osdc->request_mutex);

	list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
		if (time_before(jiffies, osd->lru_ttl))
			break;

		remove_osd(osdc, osd);
	}
1821

1822 1823
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
1824 1825 1826 1827
	schedule_delayed_work(&osdc->osds_timeout_work,
			      round_jiffies_relative(delay));
}

1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930
static int ceph_oloc_decode(void **p, void *end,
			    struct ceph_object_locator *oloc)
{
	u8 struct_v, struct_cv;
	u32 len;
	void *struct_end;
	int ret = 0;

	ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
	struct_v = ceph_decode_8(p);
	struct_cv = ceph_decode_8(p);
	if (struct_v < 3) {
		pr_warn("got v %d < 3 cv %d of ceph_object_locator\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	if (struct_cv > 6) {
		pr_warn("got v %d cv %d > 6 of ceph_object_locator\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	len = ceph_decode_32(p);
	ceph_decode_need(p, end, len, e_inval);
	struct_end = *p + len;

	oloc->pool = ceph_decode_64(p);
	*p += 4; /* skip preferred */

	len = ceph_decode_32(p);
	if (len > 0) {
		pr_warn("ceph_object_locator::key is set\n");
		goto e_inval;
	}

	if (struct_v >= 5) {
		len = ceph_decode_32(p);
		if (len > 0) {
			pr_warn("ceph_object_locator::nspace is set\n");
			goto e_inval;
		}
	}

	if (struct_v >= 6) {
		s64 hash = ceph_decode_64(p);
		if (hash != -1) {
			pr_warn("ceph_object_locator::hash is set\n");
			goto e_inval;
		}
	}

	/* skip the rest */
	*p = struct_end;
out:
	return ret;

e_inval:
	ret = -EINVAL;
	goto out;
}

static int ceph_redirect_decode(void **p, void *end,
				struct ceph_request_redirect *redir)
{
	u8 struct_v, struct_cv;
	u32 len;
	void *struct_end;
	int ret;

	ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
	struct_v = ceph_decode_8(p);
	struct_cv = ceph_decode_8(p);
	if (struct_cv > 1) {
		pr_warn("got v %d cv %d > 1 of ceph_request_redirect\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	len = ceph_decode_32(p);
	ceph_decode_need(p, end, len, e_inval);
	struct_end = *p + len;

	ret = ceph_oloc_decode(p, end, &redir->oloc);
	if (ret)
		goto out;

	len = ceph_decode_32(p);
	if (len > 0) {
		pr_warn("ceph_request_redirect::object_name is set\n");
		goto e_inval;
	}

	len = ceph_decode_32(p);
	*p += len; /* skip osd_instructions */

	/* skip the rest */
	*p = struct_end;
out:
	return ret;

e_inval:
	ret = -EINVAL;
	goto out;
}

1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
struct MOSDOpReply {
	struct ceph_pg pgid;
	u64 flags;
	int result;
	u32 epoch;
	int num_ops;
	u32 outdata_len[CEPH_OSD_MAX_OPS];
	s32 rval[CEPH_OSD_MAX_OPS];
	int retry_attempt;
	struct ceph_eversion replay_version;
	u64 user_version;
	struct ceph_request_redirect redirect;
};
S
Sage Weil 已提交
1944

1945
static int decode_MOSDOpReply(const struct ceph_msg *msg, struct MOSDOpReply *m)
S
Sage Weil 已提交
1946
{
1947 1948 1949 1950
	void *p = msg->front.iov_base;
	void *const end = p + msg->front.iov_len;
	u16 version = le16_to_cpu(msg->hdr.version);
	struct ceph_eversion bad_replay_version;
I
Ilya Dryomov 已提交
1951
	u8 decode_redir;
1952 1953 1954
	u32 len;
	int ret;
	int i;
1955

1956 1957 1958
	ceph_decode_32_safe(&p, end, len, e_inval);
	ceph_decode_need(&p, end, len, e_inval);
	p += len; /* skip oid */
1959

1960 1961 1962
	ret = ceph_decode_pgid(&p, end, &m->pgid);
	if (ret)
		return ret;
1963

1964 1965 1966 1967 1968 1969
	ceph_decode_64_safe(&p, end, m->flags, e_inval);
	ceph_decode_32_safe(&p, end, m->result, e_inval);
	ceph_decode_need(&p, end, sizeof(bad_replay_version), e_inval);
	memcpy(&bad_replay_version, p, sizeof(bad_replay_version));
	p += sizeof(bad_replay_version);
	ceph_decode_32_safe(&p, end, m->epoch, e_inval);
1970

1971 1972 1973
	ceph_decode_32_safe(&p, end, m->num_ops, e_inval);
	if (m->num_ops > ARRAY_SIZE(m->outdata_len))
		goto e_inval;
1974

1975 1976 1977
	ceph_decode_need(&p, end, m->num_ops * sizeof(struct ceph_osd_op),
			 e_inval);
	for (i = 0; i < m->num_ops; i++) {
1978 1979
		struct ceph_osd_op *op = p;

1980
		m->outdata_len[i] = le32_to_cpu(op->payload_len);
1981 1982 1983
		p += sizeof(*op);
	}

1984 1985 1986
	ceph_decode_32_safe(&p, end, m->retry_attempt, e_inval);
	for (i = 0; i < m->num_ops; i++)
		ceph_decode_32_safe(&p, end, m->rval[i], e_inval);
S
Sage Weil 已提交
1987

1988 1989 1990 1991 1992 1993 1994 1995 1996
	if (version >= 5) {
		ceph_decode_need(&p, end, sizeof(m->replay_version), e_inval);
		memcpy(&m->replay_version, p, sizeof(m->replay_version));
		p += sizeof(m->replay_version);
		ceph_decode_64_safe(&p, end, m->user_version, e_inval);
	} else {
		m->replay_version = bad_replay_version; /* struct */
		m->user_version = le64_to_cpu(m->replay_version.version);
	}
Y
Yan, Zheng 已提交
1997

1998 1999 2000
	if (version >= 6) {
		if (version >= 7)
			ceph_decode_8_safe(&p, end, decode_redir, e_inval);
I
Ilya Dryomov 已提交
2001 2002 2003 2004 2005 2006 2007
		else
			decode_redir = 1;
	} else {
		decode_redir = 0;
	}

	if (decode_redir) {
2008 2009 2010
		ret = ceph_redirect_decode(&p, end, &m->redirect);
		if (ret)
			return ret;
2011
	} else {
2012
		ceph_oloc_init(&m->redirect.oloc);
2013
	}
S
Sage Weil 已提交
2014

2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
	return 0;

e_inval:
	return -EINVAL;
}

/*
 * We are done with @req if
 *   - @m is a safe reply, or
 *   - @m is an unsafe reply and we didn't want a safe one
 */
static bool done_request(const struct ceph_osd_request *req,
			 const struct MOSDOpReply *m)
{
	return (m->result < 0 ||
		(m->flags & CEPH_OSD_FLAG_ONDISK) ||
		!(req->r_flags & CEPH_OSD_FLAG_ONDISK));
}
2033

2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 *
 * ->r_unsafe_callback is set?	yes			no
 *
 * first reply is OK (needed	r_cb/r_completion,	r_cb/r_completion,
 * any or needed/got safe)	r_safe_completion	r_safe_completion
 *
 * first reply is unsafe	r_unsafe_cb(true)	(nothing)
 *
 * when we get the safe reply	r_unsafe_cb(false),	r_cb/r_completion,
 *				r_safe_completion	r_safe_completion
 */
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg)
{
	struct ceph_osd_request *req;
	struct MOSDOpReply m;
	u64 tid = le64_to_cpu(msg->hdr.tid);
	u32 data_len = 0;
	bool already_acked;
	int ret;
	int i;

	dout("%s msg %p tid %llu\n", __func__, msg, tid);

	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
	req = lookup_request(&osdc->requests, tid);
	if (!req) {
		dout("%s no tid %llu\n", __func__, tid);
		goto out_unlock;
	}
	ceph_osdc_get_request(req);

	ret = decode_MOSDOpReply(msg, &m);
	if (ret) {
		pr_err("failed to decode MOSDOpReply for tid %llu: %d\n",
		       req->r_tid, ret);
		ceph_msg_dump(msg);
		goto fail_request;
	}
	dout("%s req %p tid %llu flags 0x%llx pgid %llu.%x epoch %u attempt %d v %u'%llu uv %llu\n",
	     __func__, req, req->r_tid, m.flags, m.pgid.pool, m.pgid.seed,
	     m.epoch, m.retry_attempt, le32_to_cpu(m.replay_version.epoch),
	     le64_to_cpu(m.replay_version.version), m.user_version);

	if (m.retry_attempt >= 0) {
		if (m.retry_attempt != req->r_attempts - 1) {
			dout("req %p tid %llu retry_attempt %d != %d, ignoring\n",
			     req, req->r_tid, m.retry_attempt,
			     req->r_attempts - 1);
			goto out_put;
		}
	} else {
		WARN_ON(1); /* MOSDOpReply v4 is assumed */
	}

	if (!ceph_oloc_empty(&m.redirect.oloc)) {
		dout("req %p tid %llu redirect pool %lld\n", req, req->r_tid,
		     m.redirect.oloc.pool);
2095 2096
		__unregister_request(osdc, req);

2097
		ceph_oloc_copy(&req->r_t.target_oloc, &m.redirect.oloc);
2098 2099 2100 2101 2102 2103 2104 2105 2106

		/*
		 * Start redirect requests with nofail=true.  If
		 * mapping fails, request will end up on the notarget
		 * list, waiting for the new osdmap (which can take
		 * a while), even though the original request mapped
		 * successfully.  In the future we might want to follow
		 * original request's nofail setting here.
		 */
2107 2108
		ret = __ceph_osdc_start_request(osdc, req, true);
		BUG_ON(ret);
2109

2110
		goto out_put;
2111 2112
	}

2113 2114 2115 2116
	if (m.num_ops != req->r_num_ops) {
		pr_err("num_ops %d != %d for tid %llu\n", m.num_ops,
		       req->r_num_ops, req->r_tid);
		goto fail_request;
S
Sage Weil 已提交
2117
	}
2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
	for (i = 0; i < req->r_num_ops; i++) {
		dout(" req %p tid %llu op %d rval %d len %u\n", req,
		     req->r_tid, i, m.rval[i], m.outdata_len[i]);
		req->r_ops[i].rval = m.rval[i];
		req->r_ops[i].outdata_len = m.outdata_len[i];
		data_len += m.outdata_len[i];
	}
	if (data_len != le32_to_cpu(msg->hdr.data_len)) {
		pr_err("sum of lens %u != %u for tid %llu\n", data_len,
		       le32_to_cpu(msg->hdr.data_len), req->r_tid);
		goto fail_request;
	}
	dout("%s req %p tid %llu acked %d result %d data_len %u\n", __func__,
	     req, req->r_tid, req->r_got_reply, m.result, data_len);

	already_acked = req->r_got_reply;
	if (!already_acked) {
		req->r_result = m.result ?: data_len;
		req->r_replay_version = m.replay_version; /* struct */
		req->r_got_reply = true;
	} else if (!(m.flags & CEPH_OSD_FLAG_ONDISK)) {
		dout("req %p tid %llu dup ack\n", req, req->r_tid);
		goto out_put;
	}

	if (done_request(req, &m)) {
S
Sage Weil 已提交
2144
		__unregister_request(osdc, req);
2145 2146 2147 2148 2149
		if (req->r_linger) {
			WARN_ON(req->r_unsafe_callback);
			__register_linger_request(osdc, req);
		}
	}
S
Sage Weil 已提交
2150 2151

	mutex_unlock(&osdc->request_mutex);
2152
	up_read(&osdc->map_sem);
S
Sage Weil 已提交
2153

2154 2155 2156
	if (done_request(req, &m)) {
		if (already_acked && req->r_unsafe_callback) {
			dout("req %p tid %llu safe-cb\n", req, req->r_tid);
2157
			req->r_unsafe_callback(req, false);
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
		} else {
			dout("req %p tid %llu cb\n", req, req->r_tid);
			__complete_request(req);
		}
	} else {
		if (req->r_unsafe_callback) {
			dout("req %p tid %llu unsafe-cb\n", req, req->r_tid);
			req->r_unsafe_callback(req, true);
		} else {
			WARN_ON(1);
		}
2169
	}
2170 2171
	if (m.flags & CEPH_OSD_FLAG_ONDISK)
		complete_all(&req->r_safe_completion);
S
Sage Weil 已提交
2172 2173 2174 2175

	ceph_osdc_put_request(req);
	return;

2176
fail_request:
2177 2178
	req->r_result = -EIO;
	__unregister_request(osdc, req);
2179 2180 2181
	__complete_request(req);
	complete_all(&req->r_safe_completion);
out_put:
2182
	ceph_osdc_put_request(req);
2183
out_unlock:
2184
	mutex_unlock(&osdc->request_mutex);
2185
	up_read(&osdc->map_sem);
S
Sage Weil 已提交
2186 2187
}

I
Ilya Dryomov 已提交
2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
static void set_pool_was_full(struct ceph_osd_client *osdc)
{
	struct rb_node *n;

	for (n = rb_first(&osdc->osdmap->pg_pools); n; n = rb_next(n)) {
		struct ceph_pg_pool_info *pi =
		    rb_entry(n, struct ceph_pg_pool_info, node);

		pi->was_full = __pool_full(pi);
	}
}

2200
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
2201 2202 2203
{
	struct rb_node *p, *n;

2204
	dout("%s %p\n", __func__, osdc);
2205 2206
	for (p = rb_first(&osdc->osds); p; p = n) {
		struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
S
Sage Weil 已提交
2207

2208 2209 2210 2211 2212 2213 2214
		n = rb_next(p);
		if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
		    memcmp(&osd->o_con.peer_addr,
			   ceph_osd_addr(osdc->osdmap,
					 osd->o_osd),
			   sizeof(struct ceph_entity_addr)) != 0)
			__reset_osd(osdc, osd);
S
Sage Weil 已提交
2215
	}
2216 2217 2218
}

/*
2219 2220
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
2221
 *
2222
 * Caller should hold map_sem for read.
2223
 */
2224 2225
static void kick_requests(struct ceph_osd_client *osdc, bool force_resend,
			  bool force_resend_writes)
2226
{
2227
	struct ceph_osd_request *req, *nreq;
2228 2229 2230
	struct rb_node *p;
	int needmap = 0;
	int err;
2231
	bool force_resend_req;
2232

2233 2234
	dout("kick_requests %s %s\n", force_resend ? " (force resend)" : "",
		force_resend_writes ? " (force resend writes)" : "");
2235
	mutex_lock(&osdc->request_mutex);
2236
	for (p = rb_first(&osdc->requests); p; ) {
2237
		req = rb_entry(p, struct ceph_osd_request, r_node);
2238
		p = rb_next(p);
2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251

		/*
		 * For linger requests that have not yet been
		 * registered, move them to the linger list; they'll
		 * be sent to the osd in the loop below.  Unregister
		 * the request before re-registering it as a linger
		 * request to ensure the __map_request() below
		 * will decide it needs to be sent.
		 */
		if (req->r_linger && list_empty(&req->r_linger_item)) {
			dout("%p tid %llu restart on osd%d\n",
			     req, req->r_tid,
			     req->r_osd ? req->r_osd->o_osd : -1);
2252
			ceph_osdc_get_request(req);
2253 2254
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
2255
			ceph_osdc_put_request(req);
2256 2257 2258
			continue;
		}

2259 2260 2261 2262
		force_resend_req = force_resend ||
			(force_resend_writes &&
				req->r_flags & CEPH_OSD_FLAG_WRITE);
		err = __map_request(osdc, req, force_resend_req);
2263 2264 2265 2266 2267 2268
		if (err < 0)
			continue;  /* error */
		if (req->r_osd == NULL) {
			dout("%p tid %llu maps to no osd\n", req, req->r_tid);
			needmap++;  /* request a newer map */
		} else if (err > 0) {
2269 2270 2271 2272
			if (!req->r_linger) {
				dout("%p tid %llu requeued on osd%d\n", req,
				     req->r_tid,
				     req->r_osd ? req->r_osd->o_osd : -1);
2273
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
2274 2275
			}
		}
2276 2277 2278 2279 2280 2281
	}

	list_for_each_entry_safe(req, nreq, &osdc->req_linger,
				 r_linger_item) {
		dout("linger req=%p req->r_osd=%p\n", req, req->r_osd);

2282 2283
		err = __map_request(osdc, req,
				    force_resend || force_resend_writes);
2284
		dout("__map_request returned %d\n", err);
2285 2286
		if (err < 0)
			continue;  /* hrm! */
2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301
		if (req->r_osd == NULL || err > 0) {
			if (req->r_osd == NULL) {
				dout("lingering %p tid %llu maps to no osd\n",
				     req, req->r_tid);
				/*
				 * A homeless lingering request makes
				 * no sense, as it's job is to keep
				 * a particular OSD connection open.
				 * Request a newer map and kick the
				 * request, knowing that it won't be
				 * resent until we actually get a map
				 * that can tell us where to send it.
				 */
				needmap++;
			}
2302

2303 2304 2305 2306 2307
			dout("kicking lingering %p tid %llu osd%d\n", req,
			     req->r_tid, req->r_osd ? req->r_osd->o_osd : -1);
			__register_request(osdc, req);
			__unregister_linger_request(osdc, req);
		}
2308
	}
2309
	reset_changed_osds(osdc);
S
Sage Weil 已提交
2310 2311 2312 2313 2314 2315
	mutex_unlock(&osdc->request_mutex);

	if (needmap) {
		dout("%d requests for down osds, need new map\n", needmap);
		ceph_monc_request_next_osdmap(&osdc->client->monc);
	}
2316
}
2317

I
Ilya Dryomov 已提交
2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368
static int handle_one_map(struct ceph_osd_client *osdc,
			  void *p, void *end, bool incremental)
{
	struct ceph_osdmap *newmap;
	struct rb_node *n;
	bool skipped_map = false;
	bool was_full;

	was_full = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);
	set_pool_was_full(osdc);

	if (incremental)
		newmap = osdmap_apply_incremental(&p, end, osdc->osdmap);
	else
		newmap = ceph_osdmap_decode(&p, end);
	if (IS_ERR(newmap))
		return PTR_ERR(newmap);

	if (newmap != osdc->osdmap) {
		/*
		 * Preserve ->was_full before destroying the old map.
		 * For pools that weren't in the old map, ->was_full
		 * should be false.
		 */
		for (n = rb_first(&newmap->pg_pools); n; n = rb_next(n)) {
			struct ceph_pg_pool_info *pi =
			    rb_entry(n, struct ceph_pg_pool_info, node);
			struct ceph_pg_pool_info *old_pi;

			old_pi = ceph_pg_pool_by_id(osdc->osdmap, pi->id);
			if (old_pi)
				pi->was_full = old_pi->was_full;
			else
				WARN_ON(pi->was_full);
		}

		if (osdc->osdmap->epoch &&
		    osdc->osdmap->epoch + 1 < newmap->epoch) {
			WARN_ON(incremental);
			skipped_map = true;
		}

		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = newmap;
	}

	was_full &= !ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);
	kick_requests(osdc, skipped_map, was_full);

	return 0;
}
2369

S
Sage Weil 已提交
2370 2371 2372 2373 2374 2375 2376 2377 2378
/*
 * Process updated osd map.
 *
 * The message contains any number of incremental and full maps, normally
 * indicating some sort of topology change in the cluster.  Kick requests
 * off to different OSDs as needed.
 */
void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
{
I
Ilya Dryomov 已提交
2379 2380
	void *p = msg->front.iov_base;
	void *const end = p + msg->front.iov_len;
S
Sage Weil 已提交
2381 2382 2383
	u32 nr_maps, maplen;
	u32 epoch;
	struct ceph_fsid fsid;
I
Ilya Dryomov 已提交
2384 2385 2386 2387
	bool handled_incremental = false;
	bool was_pauserd, was_pausewr;
	bool pauserd, pausewr;
	int err;
S
Sage Weil 已提交
2388

I
Ilya Dryomov 已提交
2389 2390
	dout("%s have %u\n", __func__, osdc->osdmap->epoch);
	down_write(&osdc->map_sem);
S
Sage Weil 已提交
2391 2392 2393 2394

	/* verify fsid */
	ceph_decode_need(&p, end, sizeof(fsid), bad);
	ceph_decode_copy(&p, &fsid, sizeof(fsid));
2395
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
I
Ilya Dryomov 已提交
2396
		goto bad;
S
Sage Weil 已提交
2397

I
Ilya Dryomov 已提交
2398 2399 2400 2401
	was_pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD);
	was_pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) ||
		      ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
		      have_pool_full(osdc);
2402

S
Sage Weil 已提交
2403 2404 2405 2406 2407
	/* incremental maps */
	ceph_decode_32_safe(&p, end, nr_maps, bad);
	dout(" %d inc maps\n", nr_maps);
	while (nr_maps > 0) {
		ceph_decode_need(&p, end, 2*sizeof(u32), bad);
2408 2409
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
2410
		ceph_decode_need(&p, end, maplen, bad);
I
Ilya Dryomov 已提交
2411 2412
		if (osdc->osdmap->epoch &&
		    osdc->osdmap->epoch + 1 == epoch) {
S
Sage Weil 已提交
2413 2414
			dout("applying incremental map %u len %d\n",
			     epoch, maplen);
I
Ilya Dryomov 已提交
2415 2416
			err = handle_one_map(osdc, p, p + maplen, true);
			if (err)
S
Sage Weil 已提交
2417
				goto bad;
I
Ilya Dryomov 已提交
2418
			handled_incremental = true;
S
Sage Weil 已提交
2419 2420 2421 2422
		} else {
			dout("ignoring incremental map %u len %d\n",
			     epoch, maplen);
		}
I
Ilya Dryomov 已提交
2423
		p += maplen;
S
Sage Weil 已提交
2424 2425
		nr_maps--;
	}
I
Ilya Dryomov 已提交
2426
	if (handled_incremental)
S
Sage Weil 已提交
2427 2428 2429 2430 2431 2432 2433
		goto done;

	/* full maps */
	ceph_decode_32_safe(&p, end, nr_maps, bad);
	dout(" %d full maps\n", nr_maps);
	while (nr_maps) {
		ceph_decode_need(&p, end, 2*sizeof(u32), bad);
2434 2435
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
2436 2437 2438 2439
		ceph_decode_need(&p, end, maplen, bad);
		if (nr_maps > 1) {
			dout("skipping non-latest full map %u len %d\n",
			     epoch, maplen);
2440
		} else if (osdc->osdmap->epoch >= epoch) {
S
Sage Weil 已提交
2441 2442 2443 2444 2445
			dout("skipping full map %u len %d, "
			     "older than our %u\n", epoch, maplen,
			     osdc->osdmap->epoch);
		} else {
			dout("taking full map %u len %d\n", epoch, maplen);
I
Ilya Dryomov 已提交
2446 2447
			err = handle_one_map(osdc, p, p + maplen, false);
			if (err)
S
Sage Weil 已提交
2448 2449 2450 2451 2452 2453 2454
				goto bad;
		}
		p += maplen;
		nr_maps--;
	}

done:
2455 2456 2457 2458 2459
	/*
	 * subscribe to subsequent osdmap updates if full to ensure
	 * we find out when we are no longer full and stop returning
	 * ENOSPC.
	 */
I
Ilya Dryomov 已提交
2460 2461 2462 2463 2464 2465
	pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD);
	pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) ||
		  ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
		  have_pool_full(osdc);
	if (was_pauserd || was_pausewr || pauserd || pausewr)
		maybe_request_map(osdc);
2466

A
Alex Elder 已提交
2467 2468 2469
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
I
Ilya Dryomov 已提交
2470 2471 2472 2473

	ceph_monc_got_map(&osdc->client->monc, CEPH_SUB_OSDMAP,
			  osdc->osdmap->epoch);
	up_write(&osdc->map_sem);
2474
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
2475 2476 2477 2478
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
2479
	ceph_msg_dump(msg);
S
Sage Weil 已提交
2480 2481 2482
	up_write(&osdc->map_sem);
}

2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
/*
 * watch/notify callback event infrastructure
 *
 * These callbacks are used both for watch and notify operations.
 */
static void __release_event(struct kref *kref)
{
	struct ceph_osd_event *event =
		container_of(kref, struct ceph_osd_event, kref);

	dout("__release_event %p\n", event);
	kfree(event);
}

static void get_event(struct ceph_osd_event *event)
{
	kref_get(&event->kref);
}

void ceph_osdc_put_event(struct ceph_osd_event *event)
{
	kref_put(&event->kref, __release_event);
}
EXPORT_SYMBOL(ceph_osdc_put_event);

static void __insert_event(struct ceph_osd_client *osdc,
			     struct ceph_osd_event *new)
{
	struct rb_node **p = &osdc->event_tree.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_event *event = NULL;

	while (*p) {
		parent = *p;
		event = rb_entry(parent, struct ceph_osd_event, node);
		if (new->cookie < event->cookie)
			p = &(*p)->rb_left;
		else if (new->cookie > event->cookie)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->node, parent, p);
	rb_insert_color(&new->node, &osdc->event_tree);
}

static struct ceph_osd_event *__find_event(struct ceph_osd_client *osdc,
					        u64 cookie)
{
	struct rb_node **p = &osdc->event_tree.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_event *event = NULL;

	while (*p) {
		parent = *p;
		event = rb_entry(parent, struct ceph_osd_event, node);
		if (cookie < event->cookie)
			p = &(*p)->rb_left;
		else if (cookie > event->cookie)
			p = &(*p)->rb_right;
		else
			return event;
	}
	return NULL;
}

static void __remove_event(struct ceph_osd_event *event)
{
	struct ceph_osd_client *osdc = event->osdc;

	if (!RB_EMPTY_NODE(&event->node)) {
		dout("__remove_event removed %p\n", event);
		rb_erase(&event->node, &osdc->event_tree);
		ceph_osdc_put_event(event);
	} else {
		dout("__remove_event didn't remove %p\n", event);
	}
}

int ceph_osdc_create_event(struct ceph_osd_client *osdc,
			   void (*event_cb)(u64, u64, u8, void *),
2565
			   void *data, struct ceph_osd_event **pevent)
2566 2567 2568 2569 2570 2571 2572 2573 2574
{
	struct ceph_osd_event *event;

	event = kmalloc(sizeof(*event), GFP_NOIO);
	if (!event)
		return -ENOMEM;

	dout("create_event %p\n", event);
	event->cb = event_cb;
2575
	event->one_shot = 0;
2576 2577 2578
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
2579
	RB_CLEAR_NODE(&event->node);
2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625
	kref_init(&event->kref);   /* one ref for us */
	kref_get(&event->kref);    /* one ref for the caller */

	spin_lock(&osdc->event_lock);
	event->cookie = ++osdc->event_count;
	__insert_event(osdc, event);
	spin_unlock(&osdc->event_lock);

	*pevent = event;
	return 0;
}
EXPORT_SYMBOL(ceph_osdc_create_event);

void ceph_osdc_cancel_event(struct ceph_osd_event *event)
{
	struct ceph_osd_client *osdc = event->osdc;

	dout("cancel_event %p\n", event);
	spin_lock(&osdc->event_lock);
	__remove_event(event);
	spin_unlock(&osdc->event_lock);
	ceph_osdc_put_event(event); /* caller's */
}
EXPORT_SYMBOL(ceph_osdc_cancel_event);


static void do_event_work(struct work_struct *work)
{
	struct ceph_osd_event_work *event_work =
		container_of(work, struct ceph_osd_event_work, work);
	struct ceph_osd_event *event = event_work->event;
	u64 ver = event_work->ver;
	u64 notify_id = event_work->notify_id;
	u8 opcode = event_work->opcode;

	dout("do_event_work completing %p\n", event);
	event->cb(ver, notify_id, opcode, event->data);
	dout("do_event_work completed %p\n", event);
	ceph_osdc_put_event(event);
	kfree(event_work);
}


/*
 * Process osd watch notifications
 */
2626 2627
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647
{
	void *p, *end;
	u8 proto_ver;
	u64 cookie, ver, notify_id;
	u8 opcode;
	struct ceph_osd_event *event;
	struct ceph_osd_event_work *event_work;

	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	ceph_decode_8_safe(&p, end, proto_ver, bad);
	ceph_decode_8_safe(&p, end, opcode, bad);
	ceph_decode_64_safe(&p, end, cookie, bad);
	ceph_decode_64_safe(&p, end, ver, bad);
	ceph_decode_64_safe(&p, end, notify_id, bad);

	spin_lock(&osdc->event_lock);
	event = __find_event(osdc, cookie);
	if (event) {
2648
		BUG_ON(event->one_shot);
2649 2650 2651 2652 2653 2654 2655 2656
		get_event(event);
	}
	spin_unlock(&osdc->event_lock);
	dout("handle_watch_notify cookie %lld ver %lld event %p\n",
	     cookie, ver, event);
	if (event) {
		event_work = kmalloc(sizeof(*event_work), GFP_NOIO);
		if (!event_work) {
2657 2658 2659
			pr_err("couldn't allocate event_work\n");
			ceph_osdc_put_event(event);
			return;
2660
		}
2661
		INIT_WORK(&event_work->work, do_event_work);
2662 2663 2664 2665 2666
		event_work->event = event;
		event_work->ver = ver;
		event_work->notify_id = notify_id;
		event_work->opcode = opcode;

2667 2668
		queue_work(osdc->notify_wq, &event_work->work);
	}
2669 2670 2671 2672 2673 2674 2675

	return;

bad:
	pr_err("osdc handle_watch_notify corrupt msg\n");
}

2676 2677 2678 2679 2680 2681 2682
/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
2683
	int rc;
2684

S
Sage Weil 已提交
2685 2686
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2687 2688 2689

	rc = __ceph_osdc_start_request(osdc, req, nofail);

S
Sage Weil 已提交
2690 2691
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
2692

S
Sage Weil 已提交
2693 2694
	return rc;
}
2695
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2696

2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715
/*
 * Unregister a registered request.  The request is not completed (i.e.
 * no callbacks or wakeups) - higher layers are supposed to know what
 * they are canceling.
 */
void ceph_osdc_cancel_request(struct ceph_osd_request *req)
{
	struct ceph_osd_client *osdc = req->r_osdc;

	mutex_lock(&osdc->request_mutex);
	if (req->r_linger)
		__unregister_linger_request(osdc, req);
	__unregister_request(osdc, req);
	mutex_unlock(&osdc->request_mutex);

	dout("%s %p tid %llu canceled\n", __func__, req, req->r_tid);
}
EXPORT_SYMBOL(ceph_osdc_cancel_request);

S
Sage Weil 已提交
2716 2717 2718 2719 2720 2721 2722 2723
/*
 * wait for a request to complete
 */
int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
{
	int rc;

2724 2725
	dout("%s %p tid %llu\n", __func__, req, req->r_tid);

S
Sage Weil 已提交
2726 2727
	rc = wait_for_completion_interruptible(&req->r_completion);
	if (rc < 0) {
2728 2729
		dout("%s %p tid %llu interrupted\n", __func__, req, req->r_tid);
		ceph_osdc_cancel_request(req);
2730 2731 2732

		/* kludge - need to to wake ceph_osdc_sync() */
		complete_all(&req->r_safe_completion);
S
Sage Weil 已提交
2733 2734 2735
		return rc;
	}

2736 2737
	dout("%s %p tid %llu result %d\n", __func__, req, req->r_tid,
	     req->r_result);
S
Sage Weil 已提交
2738 2739
	return req->r_result;
}
2740
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773

/*
 * sync - wait for all in-flight requests to flush.  avoid starvation.
 */
void ceph_osdc_sync(struct ceph_osd_client *osdc)
{
	struct ceph_osd_request *req;
	u64 last_tid, next_tid = 0;

	mutex_lock(&osdc->request_mutex);
	last_tid = osdc->last_tid;
	while (1) {
		req = __lookup_request_ge(osdc, next_tid);
		if (!req)
			break;
		if (req->r_tid > last_tid)
			break;

		next_tid = req->r_tid + 1;
		if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
			continue;

		ceph_osdc_get_request(req);
		mutex_unlock(&osdc->request_mutex);
		dout("sync waiting on tid %llu (last is %llu)\n",
		     req->r_tid, last_tid);
		wait_for_completion(&req->r_safe_completion);
		mutex_lock(&osdc->request_mutex);
		ceph_osdc_put_request(req);
	}
	mutex_unlock(&osdc->request_mutex);
	dout("sync done (thru tid %llu)\n", last_tid);
}
2774
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2775

2776 2777 2778 2779
/*
 * Call all pending notify callbacks - for use after a watch is
 * unregistered, to make sure no more callbacks for it will be invoked
 */
S
stephen hemminger 已提交
2780
void ceph_osdc_flush_notifies(struct ceph_osd_client *osdc)
2781 2782 2783 2784 2785 2786
{
	flush_workqueue(osdc->notify_wq);
}
EXPORT_SYMBOL(ceph_osdc_flush_notifies);


S
Sage Weil 已提交
2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799
/*
 * init, shutdown
 */
int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
{
	int err;

	dout("init\n");
	osdc->client = client;
	init_rwsem(&osdc->map_sem);
	mutex_init(&osdc->request_mutex);
	osdc->last_tid = 0;
	osdc->osds = RB_ROOT;
2800
	INIT_LIST_HEAD(&osdc->osd_lru);
2801
	spin_lock_init(&osdc->osd_lru_lock);
S
Sage Weil 已提交
2802
	osdc->requests = RB_ROOT;
2803
	INIT_LIST_HEAD(&osdc->req_lru);
2804 2805
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2806
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2807 2808
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2809
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2810 2811 2812
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2813

2814
	err = -ENOMEM;
2815 2816 2817 2818
	osdc->osdmap = ceph_osdmap_alloc();
	if (!osdc->osdmap)
		goto out;

2819 2820
	osdc->req_mempool = mempool_create_slab_pool(10,
						     ceph_osd_request_cache);
S
Sage Weil 已提交
2821
	if (!osdc->req_mempool)
2822
		goto out_map;
S
Sage Weil 已提交
2823

2824
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
2825
				PAGE_SIZE, 10, true, "osd_op");
S
Sage Weil 已提交
2826
	if (err < 0)
2827
		goto out_mempool;
2828
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2829
				PAGE_SIZE, 10, true, "osd_op_reply");
S
Sage Weil 已提交
2830 2831
	if (err < 0)
		goto out_msgpool;
2832

2833
	err = -ENOMEM;
2834
	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
2835
	if (!osdc->notify_wq)
2836 2837
		goto out_msgpool_reply;

2838 2839
	schedule_delayed_work(&osdc->timeout_work,
			      osdc->client->options->osd_keepalive_timeout);
2840 2841 2842
	schedule_delayed_work(&osdc->osds_timeout_work,
	    round_jiffies_relative(osdc->client->options->osd_idle_ttl));

S
Sage Weil 已提交
2843
	return 0;
2844

2845 2846
out_msgpool_reply:
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2847 2848
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2849 2850
out_mempool:
	mempool_destroy(osdc->req_mempool);
2851 2852
out_map:
	ceph_osdmap_destroy(osdc->osdmap);
2853 2854
out:
	return err;
S
Sage Weil 已提交
2855 2856 2857 2858
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2859 2860
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2861
	cancel_delayed_work_sync(&osdc->timeout_work);
2862
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
2863 2864 2865 2866 2867 2868 2869 2870 2871

	mutex_lock(&osdc->request_mutex);
	while (!RB_EMPTY_ROOT(&osdc->osds)) {
		struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
						struct ceph_osd, o_node);
		remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);

2872
	ceph_osdmap_destroy(osdc->osdmap);
S
Sage Weil 已提交
2873 2874
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2875
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
}

/*
 * Read some contiguous pages.  If we cross a stripe boundary, shorten
 * *plen.  Return number of bytes read, or error.
 */
int ceph_osdc_readpages(struct ceph_osd_client *osdc,
			struct ceph_vino vino, struct ceph_file_layout *layout,
			u64 off, u64 *plen,
			u32 truncate_seq, u64 truncate_size,
2886
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2887 2888 2889 2890 2891 2892
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2893
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 0, 1,
S
Sage Weil 已提交
2894
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2895
				    NULL, truncate_seq, truncate_size,
2896
				    false);
2897 2898
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2899 2900

	/* it may be a short read due to an object boundary */
2901
	osd_req_op_extent_osd_data_pages(req, 0,
2902
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2903

2904
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2905
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2906 2907 2908 2909 2910 2911 2912 2913 2914

	rc = ceph_osdc_start_request(osdc, req, false);
	if (!rc)
		rc = ceph_osdc_wait_request(osdc, req);

	ceph_osdc_put_request(req);
	dout("readpages result %d\n", rc);
	return rc;
}
2915
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2916 2917 2918 2919 2920 2921 2922 2923 2924 2925

/*
 * do a synchronous write on N pages
 */
int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
			 struct ceph_file_layout *layout,
			 struct ceph_snap_context *snapc,
			 u64 off, u64 len,
			 u32 truncate_seq, u64 truncate_size,
			 struct timespec *mtime,
2926
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2927 2928 2929
{
	struct ceph_osd_request *req;
	int rc = 0;
2930
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2931

2932
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 0, 1,
S
Sage Weil 已提交
2933
				    CEPH_OSD_OP_WRITE,
2934
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2935
				    snapc, truncate_seq, truncate_size,
2936
				    true);
2937 2938
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2939 2940

	/* it may be a short write due to an object boundary */
2941
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2942 2943
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2944

2945
	req->r_mtime = *mtime;
2946
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2947 2948 2949 2950 2951 2952 2953 2954 2955
	if (!rc)
		rc = ceph_osdc_wait_request(osdc, req);

	ceph_osdc_put_request(req);
	if (rc == 0)
		rc = len;
	dout("writepages result %d\n", rc);
	return rc;
}
2956
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2957

2958 2959
int ceph_osdc_setup(void)
{
2960 2961 2962
	size_t size = sizeof(struct ceph_osd_request) +
	    CEPH_OSD_SLAB_OPS * sizeof(struct ceph_osd_req_op);

2963
	BUG_ON(ceph_osd_request_cache);
2964 2965
	ceph_osd_request_cache = kmem_cache_create("ceph_osd_request", size,
						   0, 0, NULL);
2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978

	return ceph_osd_request_cache ? 0 : -ENOMEM;
}
EXPORT_SYMBOL(ceph_osdc_setup);

void ceph_osdc_cleanup(void)
{
	BUG_ON(!ceph_osd_request_cache);
	kmem_cache_destroy(ceph_osd_request_cache);
	ceph_osd_request_cache = NULL;
}
EXPORT_SYMBOL(ceph_osdc_cleanup);

S
Sage Weil 已提交
2979 2980 2981 2982 2983 2984
/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2985
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2986 2987 2988
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2989
		goto out;
2990
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2991 2992 2993 2994 2995 2996

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2997
		handle_reply(osdc, msg);
S
Sage Weil 已提交
2998
		break;
2999 3000 3001
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
3002 3003 3004 3005 3006

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
3007
out:
S
Sage Weil 已提交
3008 3009 3010
	ceph_msg_put(msg);
}

3011
/*
3012 3013 3014
 * Lookup and return message for incoming reply.  Don't try to do
 * anything about a larger than preallocated data portion of the
 * message at the moment - for now, just skip the message.
3015 3016
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
3017 3018
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
3019 3020 3021
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
3022
	struct ceph_msg *m;
3023
	struct ceph_osd_request *req;
3024
	int front_len = le32_to_cpu(hdr->front_len);
3025
	int data_len = le32_to_cpu(hdr->data_len);
3026
	u64 tid;
S
Sage Weil 已提交
3027

3028 3029
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
I
Ilya Dryomov 已提交
3030
	req = lookup_request(&osdc->requests, tid);
3031
	if (!req) {
3032 3033
		dout("%s osd%d tid %llu unknown, skipping\n", __func__,
		     osd->o_osd, tid);
3034
		m = NULL;
3035
		*skip = 1;
3036 3037
		goto out;
	}
S
Sage Weil 已提交
3038

3039
	ceph_msg_revoke_incoming(req->r_reply);
3040

3041
	if (front_len > req->r_reply->front_alloc_len) {
3042 3043 3044
		pr_warn("%s osd%d tid %llu front %d > preallocated %d\n",
			__func__, osd->o_osd, req->r_tid, front_len,
			req->r_reply->front_alloc_len);
3045 3046
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front_len, GFP_NOFS,
				 false);
3047
		if (!m)
S
Sage Weil 已提交
3048 3049 3050 3051
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
3052

3053 3054 3055 3056 3057 3058 3059
	if (data_len > req->r_reply->data_length) {
		pr_warn("%s osd%d tid %llu data %d > preallocated %zu, skipping\n",
			__func__, osd->o_osd, req->r_tid, data_len,
			req->r_reply->data_length);
		m = NULL;
		*skip = 1;
		goto out;
3060
	}
3061 3062

	m = ceph_msg_get(req->r_reply);
S
Sage Weil 已提交
3063
	dout("get_reply tid %lld %p\n", tid, m);
3064 3065 3066

out:
	mutex_unlock(&osdc->request_mutex);
3067
	return m;
3068 3069 3070 3071 3072 3073 3074 3075 3076 3077
}

static struct ceph_msg *alloc_msg(struct ceph_connection *con,
				  struct ceph_msg_header *hdr,
				  int *skip)
{
	struct ceph_osd *osd = con->private;
	int type = le16_to_cpu(hdr->type);
	int front = le32_to_cpu(hdr->front_len);

A
Alex Elder 已提交
3078
	*skip = 0;
3079 3080
	switch (type) {
	case CEPH_MSG_OSD_MAP:
3081
	case CEPH_MSG_WATCH_NOTIFY:
3082
		return ceph_msg_new(type, front, GFP_NOFS, false);
3083 3084 3085 3086 3087 3088 3089 3090
	case CEPH_MSG_OSD_OPREPLY:
		return get_reply(con, hdr, skip);
	default:
		pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
			osd->o_osd);
		*skip = 1;
		return NULL;
	}
S
Sage Weil 已提交
3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109
}

/*
 * Wrappers to refcount containing ceph_osd struct
 */
static struct ceph_connection *get_osd_con(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	if (get_osd(osd))
		return con;
	return NULL;
}

static void put_osd_con(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	put_osd(osd);
}

3110 3111 3112
/*
 * authentication
 */
3113 3114 3115 3116 3117
/*
 * Note: returned pointer is the address of a structure that's
 * managed separately.  Caller must *not* attempt to free it.
 */
static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
3118
					int *proto, int force_new)
3119 3120 3121 3122
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
3123
	struct ceph_auth_handshake *auth = &o->o_auth;
3124

3125
	if (force_new && auth->authorizer) {
3126
		ceph_auth_destroy_authorizer(auth->authorizer);
3127 3128
		auth->authorizer = NULL;
	}
3129 3130 3131
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
3132
		if (ret)
3133
			return ERR_PTR(ret);
3134 3135
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
3136 3137 3138
						     auth);
		if (ret)
			return ERR_PTR(ret);
3139 3140
	}
	*proto = ac->protocol;
3141

3142
	return auth;
3143 3144 3145 3146 3147 3148 3149 3150 3151
}


static int verify_authorizer_reply(struct ceph_connection *con, int len)
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;

3152
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
3153 3154
}

3155 3156 3157 3158 3159 3160
static int invalidate_authorizer(struct ceph_connection *con)
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;

3161
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
3162 3163
	return ceph_monc_validate_auth(&osdc->client->monc);
}
3164

3165
static int osd_sign_message(struct ceph_msg *msg)
Y
Yan, Zheng 已提交
3166
{
3167
	struct ceph_osd *o = msg->con->private;
Y
Yan, Zheng 已提交
3168
	struct ceph_auth_handshake *auth = &o->o_auth;
3169

Y
Yan, Zheng 已提交
3170 3171 3172
	return ceph_auth_sign_message(auth, msg);
}

3173
static int osd_check_message_signature(struct ceph_msg *msg)
Y
Yan, Zheng 已提交
3174
{
3175
	struct ceph_osd *o = msg->con->private;
Y
Yan, Zheng 已提交
3176
	struct ceph_auth_handshake *auth = &o->o_auth;
3177

Y
Yan, Zheng 已提交
3178 3179 3180
	return ceph_auth_check_message_signature(auth, msg);
}

3181
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
3182 3183 3184
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
3185 3186
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
3187
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
3188
	.alloc_msg = alloc_msg,
3189 3190
	.sign_message = osd_sign_message,
	.check_message_signature = osd_check_message_signature,
3191
	.fault = osd_reset,
S
Sage Weil 已提交
3192
};