osd_client.c 79.5 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);
37 38
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

/*
 * 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.
 */
65
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
66
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
67
{
68
	u64 orig_len = *plen;
S
Sage Weil 已提交
69
	int r;
S
Sage Weil 已提交
70

71
	/* object extent? */
72 73
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
74 75
	if (r < 0)
		return r;
76 77
	if (*objlen < orig_len) {
		*plen = *objlen;
78 79 80 81
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

82
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
83

84
	return 0;
S
Sage Weil 已提交
85 86
}

A
Alex Elder 已提交
87 88 89 90 91 92
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;
}

93
static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
94 95 96 97 98 99 100 101 102 103 104
			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;
}

105
static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
106 107 108 109 110 111 112
			struct ceph_pagelist *pagelist)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
	osd_data->pagelist = pagelist;
}

#ifdef CONFIG_BLOCK
113
static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
114 115 116 117 118 119 120 121
			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 */

122 123 124 125 126 127 128
#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;					\
})
129

A
Alex Elder 已提交
130 131 132 133 134 135 136 137
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;
}

138 139
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
140
			unsigned int which)
141
{
142
	return osd_req_op_data(osd_req, which, extent, osd_data);
143 144 145
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data);

A
Alex Elder 已提交
146 147 148 149 150 151 152 153 154 155 156 157 158
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);

159
void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
160 161
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
162 163 164 165
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

166
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
167 168 169 170 171 172
	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,
173
			unsigned int which, struct ceph_pagelist *pagelist)
174 175 176
{
	struct ceph_osd_data *osd_data;

177
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
178 179 180 181 182 183
	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,
184
			unsigned int which, struct bio *bio, size_t bio_length)
185 186
{
	struct ceph_osd_data *osd_data;
187 188

	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
189 190 191 192 193 194 195 196 197 198 199
	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;

200
	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
201 202 203
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

204 205 206 207 208 209
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;

210
	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
211 212 213 214
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);

215 216 217 218 219 220 221 222 223 224 225 226
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);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pages);

227 228 229 230 231 232
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;

233
	osd_data = osd_req_op_data(osd_req, which, cls, response_data);
234 235 236 237 238
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);

239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
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
258 259
static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
{
260
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
A
Alex Elder 已提交
261 262 263 264 265 266
		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);
	}
267 268 269 270 271 272 273 274 275 276 277 278 279 280
	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:
281
	case CEPH_OSD_OP_WRITEFULL:
282 283 284 285
		ceph_osd_data_release(&op->extent.osd_data);
		break;
	case CEPH_OSD_OP_CALL:
		ceph_osd_data_release(&op->cls.request_info);
286
		ceph_osd_data_release(&op->cls.request_data);
287 288
		ceph_osd_data_release(&op->cls.response_data);
		break;
289 290 291 292
	case CEPH_OSD_OP_SETXATTR:
	case CEPH_OSD_OP_CMPXATTR:
		ceph_osd_data_release(&op->xattr.osd_data);
		break;
293 294 295
	case CEPH_OSD_OP_STAT:
		ceph_osd_data_release(&op->raw_data_in);
		break;
296 297 298
	default:
		break;
	}
A
Alex Elder 已提交
299 300
}

S
Sage Weil 已提交
301 302 303
/*
 * requests
 */
304
static void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
305
{
306 307
	struct ceph_osd_request *req = container_of(kref,
					    struct ceph_osd_request, r_kref);
308
	unsigned int which;
S
Sage Weil 已提交
309

310 311
	dout("%s %p (r_request %p r_reply %p)\n", __func__, req,
	     req->r_request, req->r_reply);
312 313 314 315 316 317
	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);
318

S
Sage Weil 已提交
319 320
	if (req->r_request)
		ceph_msg_put(req->r_request);
321
	if (req->r_reply) {
322
		ceph_msg_revoke_incoming(req->r_reply);
323
		ceph_msg_put(req->r_reply);
324
	}
325

326 327
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
328

329 330
	ceph_oid_destroy(&req->r_base_oid);
	ceph_oid_destroy(&req->r_target_oid);
S
Sage Weil 已提交
331
	ceph_put_snap_context(req->r_snapc);
332

S
Sage Weil 已提交
333 334
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
335
	else if (req->r_num_ops <= CEPH_OSD_SLAB_OPS)
336
		kmem_cache_free(ceph_osd_request_cache, req);
337 338
	else
		kfree(req);
S
Sage Weil 已提交
339
}
340 341 342 343 344 345 346 347 348 349 350

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)
{
351 352 353 354 355
	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);
	}
356 357
}
EXPORT_SYMBOL(ceph_osdc_put_request);
358

359
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
360
					       struct ceph_snap_context *snapc,
361
					       unsigned int num_ops,
362
					       bool use_mempool,
363
					       gfp_t gfp_flags)
S
Sage Weil 已提交
364 365
{
	struct ceph_osd_request *req;
366

S
Sage Weil 已提交
367
	if (use_mempool) {
368
		BUG_ON(num_ops > CEPH_OSD_SLAB_OPS);
369
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
370 371
	} else if (num_ops <= CEPH_OSD_SLAB_OPS) {
		req = kmem_cache_alloc(ceph_osd_request_cache, gfp_flags);
S
Sage Weil 已提交
372
	} else {
373 374 375
		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
376
	}
377
	if (unlikely(!req))
378
		return NULL;
S
Sage Weil 已提交
379

380 381 382
	/* req only, each op is zeroed in _osd_req_op_init() */
	memset(req, 0, sizeof(*req));

S
Sage Weil 已提交
383 384
	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
385
	req->r_num_ops = num_ops;
386 387
	req->r_snapid = CEPH_NOSNAP;
	req->r_snapc = ceph_get_snap_context(snapc);
388

S
Sage Weil 已提交
389
	kref_init(&req->r_kref);
S
Sage Weil 已提交
390 391
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
392
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
393
	INIT_LIST_HEAD(&req->r_unsafe_item);
394
	INIT_LIST_HEAD(&req->r_linger_item);
395
	INIT_LIST_HEAD(&req->r_linger_osd_item);
396
	INIT_LIST_HEAD(&req->r_req_lru_item);
397 398
	INIT_LIST_HEAD(&req->r_osd_item);

399
	ceph_oid_init(&req->r_base_oid);
400
	req->r_base_oloc.pool = -1;
401
	ceph_oid_init(&req->r_target_oid);
402
	req->r_target_oloc.pool = -1;
403

404 405 406 407
	dout("%s req %p\n", __func__, req);
	return req;
}
EXPORT_SYMBOL(ceph_osdc_alloc_request);
408

409 410 411 412 413
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
414

415 416
	WARN_ON(ceph_oid_empty(&req->r_base_oid));

417
	/* create request message */
418 419 420 421
	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 */
422 423
	msg_size += 4 + req->r_base_oid.name_len; /* oid */
	msg_size += 2 + req->r_num_ops * sizeof(struct ceph_osd_op);
424 425
	msg_size += 8; /* snapid */
	msg_size += 8; /* snap_seq */
426
	msg_size += 4 + 8 * (req->r_snapc ? req->r_snapc->num_snaps : 0);
427 428
	msg_size += 4; /* retry_attempt */

429
	if (req->r_mempool)
430
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
431
	else
432 433 434
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp, true);
	if (!msg)
		return -ENOMEM;
435

S
Sage Weil 已提交
436
	memset(msg->front.iov_base, 0, msg->front.iov_len);
437 438
	req->r_request = msg;

439 440
	/* create reply message */
	msg_size = OSD_OPREPLY_FRONT_LEN;
441 442
	msg_size += req->r_base_oid.name_len;
	msg_size += req->r_num_ops * sizeof(struct ceph_osd_op);
443 444 445 446 447 448 449 450 451 452 453

	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;
454
}
455
EXPORT_SYMBOL(ceph_osdc_alloc_messages);
456

457
static bool osd_req_opcode_valid(u16 opcode)
458
{
459
	switch (opcode) {
460 461 462
#define GENERATE_CASE(op, opcode, str)	case CEPH_OSD_OP_##op: return true;
__CEPH_FORALL_OSD_OPS(GENERATE_CASE)
#undef GENERATE_CASE
463 464 465 466 467
	default:
		return false;
	}
}

468 469 470 471 472
/*
 * 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.
 */
473
static struct ceph_osd_req_op *
A
Alex Elder 已提交
474
_osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
475
		 u16 opcode, u32 flags)
476
{
477 478 479
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
480 481
	BUG_ON(!osd_req_opcode_valid(opcode));

482
	op = &osd_req->r_ops[which];
483 484
	memset(op, 0, sizeof (*op));
	op->op = opcode;
485
	op->flags = flags;
486 487

	return op;
488 489
}

A
Alex Elder 已提交
490
void osd_req_op_init(struct ceph_osd_request *osd_req,
491
		     unsigned int which, u16 opcode, u32 flags)
A
Alex Elder 已提交
492
{
493
	(void)_osd_req_op_init(osd_req, which, opcode, flags);
A
Alex Elder 已提交
494 495 496
}
EXPORT_SYMBOL(osd_req_op_init);

497 498
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
499 500 501
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
502 503
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
504 505
	size_t payload_len = 0;

L
Li Wang 已提交
506
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
507 508
	       opcode != CEPH_OSD_OP_WRITEFULL && opcode != CEPH_OSD_OP_ZERO &&
	       opcode != CEPH_OSD_OP_TRUNCATE);
509 510 511 512 513

	op->extent.offset = offset;
	op->extent.length = length;
	op->extent.truncate_size = truncate_size;
	op->extent.truncate_seq = truncate_seq;
514
	if (opcode == CEPH_OSD_OP_WRITE || opcode == CEPH_OSD_OP_WRITEFULL)
515 516
		payload_len += length;

517
	op->indata_len = payload_len;
518 519 520
}
EXPORT_SYMBOL(osd_req_op_extent_init);

521 522
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
523
{
524 525 526 527 528 529
	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;
530 531 532 533 534 535

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

	op->extent.length = length;
536
	op->indata_len -= previous - length;
537 538 539
}
EXPORT_SYMBOL(osd_req_op_extent_update);

540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
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);

562
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
563
			u16 opcode, const char *class, const char *method)
564
{
565 566
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
567
	struct ceph_pagelist *pagelist;
568 569 570 571 572
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

573 574 575 576
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

577 578 579 580
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
581
	ceph_pagelist_append(pagelist, class, size);
582 583 584 585 586 587
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
588
	ceph_pagelist_append(pagelist, method, size);
589 590
	payload_len += size;

591
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
592

593 594
	op->cls.argc = 0;	/* currently unused */

595
	op->indata_len = payload_len;
596 597
}
EXPORT_SYMBOL(osd_req_op_cls_init);
598

599 600 601 602
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)
{
603 604
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627
	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);
628
	op->indata_len = payload_len;
629 630 631 632
	return 0;
}
EXPORT_SYMBOL(osd_req_op_xattr_init);

633 634
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
635 636
				u64 cookie, u64 version, int flag)
{
637 638
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
						      opcode, 0);
639

640
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
641 642

	op->watch.cookie = cookie;
A
Alex Elder 已提交
643
	op->watch.ver = version;
644
	if (opcode == CEPH_OSD_OP_WATCH && flag)
645
		op->watch.flag = (u8)1;
646 647 648
}
EXPORT_SYMBOL(osd_req_op_watch_init);

649 650 651 652 653 654
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,
655 656
						      CEPH_OSD_OP_SETALLOCHINT,
						      0);
657 658 659 660 661 662 663 664 665 666 667 668 669

	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);

670
static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
671 672 673 674 675 676 677
				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)
678
			ceph_msg_data_add_pages(msg, osd_data->pages,
679 680 681
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
682
		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
683 684
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
685
		ceph_msg_data_add_bio(msg, osd_data->bio, length);
686 687 688 689 690 691
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

692
static u64 osd_req_encode_op(struct ceph_osd_request *req,
693
			      struct ceph_osd_op *dst, unsigned int which)
694
{
695
	struct ceph_osd_req_op *src;
696
	struct ceph_osd_data *osd_data;
697
	u64 request_data_len = 0;
698
	u64 data_length;
699

700 701
	BUG_ON(which >= req->r_num_ops);
	src = &req->r_ops[which];
702 703 704 705 706 707 708 709
	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:
A
Alex Elder 已提交
710 711
		osd_data = &src->raw_data_in;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
712 713 714
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
715
	case CEPH_OSD_OP_WRITEFULL:
L
Li Wang 已提交
716 717
	case CEPH_OSD_OP_ZERO:
	case CEPH_OSD_OP_TRUNCATE:
718 719
		if (src->op == CEPH_OSD_OP_WRITE ||
		    src->op == CEPH_OSD_OP_WRITEFULL)
720
			request_data_len = src->extent.length;
721 722 723 724 725 726
		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);
727
		osd_data = &src->extent.osd_data;
728 729
		if (src->op == CEPH_OSD_OP_WRITE ||
		    src->op == CEPH_OSD_OP_WRITEFULL)
730
			ceph_osdc_msg_data_add(req->r_request, osd_data);
731
		else
732
			ceph_osdc_msg_data_add(req->r_reply, osd_data);
733 734 735 736
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
737 738 739 740 741 742 743 744 745 746 747
		osd_data = &src->cls.request_info;
		ceph_osdc_msg_data_add(req->r_request, osd_data);
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGELIST);
		request_data_len = osd_data->pagelist->length;

		osd_data = &src->cls.request_data;
		data_length = ceph_osd_data_length(osd_data);
		if (data_length) {
			BUG_ON(osd_data->type == CEPH_OSD_DATA_TYPE_NONE);
			dst->cls.indata_len = cpu_to_le32(data_length);
			ceph_osdc_msg_data_add(req->r_request, osd_data);
748
			src->indata_len += data_length;
749 750 751 752
			request_data_len += data_length;
		}
		osd_data = &src->cls.response_data;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
753 754 755 756 757 758 759 760 761
		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;
762 763 764 765 766 767
	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;
768 769 770 771 772 773 774 775 776 777
	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;
		osd_data = &src->xattr.osd_data;
		ceph_osdc_msg_data_add(req->r_request, osd_data);
		request_data_len = osd_data->pagelist->length;
		break;
778 779 780
	case CEPH_OSD_OP_CREATE:
	case CEPH_OSD_OP_DELETE:
		break;
781
	default:
782
		pr_err("unsupported osd opcode %s\n",
783
			ceph_osd_op_name(src->op));
784
		WARN_ON(1);
785 786

		return 0;
787
	}
788

789
	dst->op = cpu_to_le16(src->op);
790
	dst->flags = cpu_to_le32(src->flags);
791
	dst->payload_len = cpu_to_le32(src->indata_len);
792

793
	return request_data_len;
794 795
}

796 797 798 799 800 801 802 803 804 805 806 807 808 809
/*
 * 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,
810 811
					       u64 off, u64 *plen,
					       unsigned int which, int num_ops,
812 813 814 815
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
816
					       bool use_mempool)
817
{
818
	struct ceph_osd_request *req;
819 820 821
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
822
	int r;
823

L
Li Wang 已提交
824
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
825 826
	       opcode != CEPH_OSD_OP_ZERO && opcode != CEPH_OSD_OP_TRUNCATE &&
	       opcode != CEPH_OSD_OP_CREATE && opcode != CEPH_OSD_OP_DELETE);
827

828
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
829
					GFP_NOFS);
830 831 832 833
	if (!req) {
		r = -ENOMEM;
		goto fail;
	}
834

835
	req->r_flags = flags;
836 837

	/* calculate max write size */
838
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
839 840
	if (r)
		goto fail;
841

842
	if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) {
843
		osd_req_op_init(req, which, opcode, 0);
844 845 846 847 848 849 850 851 852 853 854
	} 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;
			}
855
		}
856
		osd_req_op_extent_init(req, which, opcode, objoff, objlen,
857 858
				       truncate_size, truncate_seq);
	}
859

860
	req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout);
861
	ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum);
862

863 864 865 866
	r = ceph_osdc_alloc_messages(req, GFP_NOFS);
	if (r)
		goto fail;

S
Sage Weil 已提交
867
	return req;
868 869 870 871

fail:
	ceph_osdc_put_request(req);
	return ERR_PTR(r);
S
Sage Weil 已提交
872
}
873
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939

/*
 * We keep osd requests in an rbtree, sorted by ->r_tid.
 */
static void __insert_request(struct ceph_osd_client *osdc,
			     struct ceph_osd_request *new)
{
	struct rb_node **p = &osdc->requests.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_request *req = NULL;

	while (*p) {
		parent = *p;
		req = rb_entry(parent, struct ceph_osd_request, r_node);
		if (new->r_tid < req->r_tid)
			p = &(*p)->rb_left;
		else if (new->r_tid > req->r_tid)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->r_node, parent, p);
	rb_insert_color(&new->r_node, &osdc->requests);
}

static struct ceph_osd_request *__lookup_request(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)
			n = n->rb_left;
		else if (tid > req->r_tid)
			n = n->rb_right;
		else
			return req;
	}
	return NULL;
}

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;
}

940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970
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);
}

971 972 973 974 975 976
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
977
	struct ceph_osd_request *req, *nreq;
978
	LIST_HEAD(resend);
979
	LIST_HEAD(resend_linger);
980 981
	int err;

982
	dout("%s osd%d\n", __func__, osd->o_osd);
983
	err = __reset_osd(osdc, osd);
984
	if (err)
985
		return;
986

987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003
	/*
	 * 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.
	 */
1004
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
1005 1006
		if (!req->r_sent)
			break;
1007 1008 1009 1010 1011

		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);
1012
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
1013 1014 1015
		} else {
			list_move_tail(&req->r_req_lru_item, &resend_linger);
		}
1016
	}
1017
	list_splice(&resend, &osdc->req_unsent);
1018

1019
	/*
1020 1021 1022 1023
	 * 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.
1024
	 */
1025
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
1026
				 r_linger_osd_item) {
1027 1028
		WARN_ON(!list_empty(&req->r_req_lru_item));
		__kick_linger_request(req);
1029
	}
1030 1031 1032

	list_for_each_entry_safe(req, nreq, &resend_linger, r_req_lru_item)
		__kick_linger_request(req);
1033 1034
}

S
Sage Weil 已提交
1035
/*
1036
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
 */
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
1048 1049
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
1050
	__send_queued(osdc);
S
Sage Weil 已提交
1051
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1052 1053 1054 1055 1056 1057
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
1058
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067
{
	struct ceph_osd *osd;

	osd = kzalloc(sizeof(*osd), GFP_NOFS);
	if (!osd)
		return NULL;

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
1068
	osd->o_osd = onum;
1069
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
1070
	INIT_LIST_HEAD(&osd->o_requests);
1071
	INIT_LIST_HEAD(&osd->o_linger_requests);
1072
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
1073 1074
	osd->o_incarnation = 1;

1075
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
1076

1077
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096
	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);
1097 1098
	if (atomic_dec_and_test(&osd->o_ref)) {
		if (osd->o_auth.authorizer)
1099
			ceph_auth_destroy_authorizer(osd->o_auth.authorizer);
S
Sage Weil 已提交
1100
		kfree(osd);
S
Sage Weil 已提交
1101
	}
S
Sage Weil 已提交
1102 1103 1104 1105 1106
}

/*
 * remove an osd from our map
 */
1107
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1108
{
1109
	dout("%s %p osd%d\n", __func__, osd, osd->o_osd);
1110 1111
	WARN_ON(!list_empty(&osd->o_requests));
	WARN_ON(!list_empty(&osd->o_linger_requests));
1112

1113
	list_del_init(&osd->o_osd_lru);
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
	rb_erase(&osd->o_node, &osdc->osds);
	RB_CLEAR_NODE(&osd->o_node);
}

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
1127 1128
}

1129 1130 1131
static void __move_osd_to_lru(struct ceph_osd_client *osdc,
			      struct ceph_osd *osd)
{
1132
	dout("%s %p\n", __func__, osd);
1133
	BUG_ON(!list_empty(&osd->o_osd_lru));
1134

1135
	list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
1136
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl;
1137 1138
}

1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
static void maybe_move_osd_to_lru(struct ceph_osd_client *osdc,
				  struct ceph_osd *osd)
{
	dout("%s %p\n", __func__, osd);

	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests))
		__move_osd_to_lru(osdc, osd);
}

1149 1150 1151 1152 1153 1154 1155
static void __remove_osd_from_lru(struct ceph_osd *osd)
{
	dout("__remove_osd_from_lru %p\n", osd);
	if (!list_empty(&osd->o_osd_lru))
		list_del_init(&osd->o_osd_lru);
}

S
Sage Weil 已提交
1156 1157 1158
/*
 * reset osd connect
 */
1159
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1160
{
A
Alex Elder 已提交
1161
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
1162

1163
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
1164 1165
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
1166
		remove_osd(osdc, osd);
A
Alex Elder 已提交
1167 1168 1169 1170 1171 1172 1173 1174
		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;

1175 1176
		dout("osd addr hasn't changed and connection never opened, "
		     "letting msgr retry\n");
1177 1178 1179
		/* 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
1180 1181

		return -EAGAIN;
S
Sage Weil 已提交
1182
	}
A
Alex Elder 已提交
1183 1184 1185 1186 1187 1188

	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
1189 1190 1191 1192 1193 1194 1195 1196
}

static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
{
	struct rb_node **p = &osdc->osds.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd *osd = NULL;

1197
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
	while (*p) {
		parent = *p;
		osd = rb_entry(parent, struct ceph_osd, o_node);
		if (new->o_osd < osd->o_osd)
			p = &(*p)->rb_left;
		else if (new->o_osd > osd->o_osd)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->o_node, parent, p);
	rb_insert_color(&new->o_node, &osdc->osds);
}

static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
{
	struct ceph_osd *osd;
	struct rb_node *n = osdc->osds.rb_node;

	while (n) {
		osd = rb_entry(n, struct ceph_osd, o_node);
		if (o < osd->o_osd)
			n = n->rb_left;
		else if (o > osd->o_osd)
			n = n->rb_right;
		else
			return osd;
	}
	return NULL;
}

1230 1231 1232
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
1233
			      osdc->client->options->osd_keepalive_timeout);
1234 1235 1236 1237 1238 1239
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
1240 1241 1242 1243 1244

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1245 1246
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1247 1248
{
	req->r_tid = ++osdc->last_tid;
1249
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1250
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
1251 1252 1253 1254
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
1255 1256
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
1257
	}
1258 1259
}

S
Sage Weil 已提交
1260 1261 1262 1263 1264 1265
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1266 1267 1268 1269 1270 1271
	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
1272 1273
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
1274
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
1275 1276
	osdc->num_requests--;

1277 1278
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1279
		ceph_msg_revoke(req->r_request);
1280 1281

		list_del_init(&req->r_osd_item);
1282
		maybe_move_osd_to_lru(osdc, req->r_osd);
1283
		if (list_empty(&req->r_linger_osd_item))
1284
			req->r_osd = NULL;
1285
	}
S
Sage Weil 已提交
1286

A
Alex Elder 已提交
1287
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1288 1289
	ceph_osdc_put_request(req);

1290 1291 1292
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1293 1294 1295 1296 1297 1298 1299 1300
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1301
	if (req->r_sent && req->r_osd) {
1302
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1303 1304 1305 1306
		req->r_sent = 0;
	}
}

1307 1308 1309
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
1310 1311 1312
	dout("%s %p tid %llu\n", __func__, req, req->r_tid);
	WARN_ON(!req->r_linger);

1313
	ceph_osdc_get_request(req);
1314
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1315
	if (req->r_osd)
1316
		list_add_tail(&req->r_linger_osd_item,
1317
			      &req->r_osd->o_linger_requests);
1318 1319 1320 1321 1322
}

static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req)
{
1323 1324 1325 1326 1327 1328 1329 1330 1331
	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
1332
	list_del_init(&req->r_linger_item);
1333

1334
	if (req->r_osd) {
1335
		list_del_init(&req->r_linger_osd_item);
1336
		maybe_move_osd_to_lru(osdc, req->r_osd);
1337 1338
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1339
	}
1340
	ceph_osdc_put_request(req);
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
}

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);

1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
/*
 * 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.
 */
static bool __req_should_be_paused(struct ceph_osd_client *osdc,
				   struct ceph_osd_request *req)
{
	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);
	return (req->r_flags & CEPH_OSD_FLAG_READ && pauserd) ||
		(req->r_flags & CEPH_OSD_FLAG_WRITE && pausewr);
}

1369 1370 1371 1372 1373 1374 1375
/*
 * Calculate mapping of a request to a PG.  Takes tiering into account.
 */
static int __calc_request_pg(struct ceph_osdmap *osdmap,
			     struct ceph_osd_request *req,
			     struct ceph_pg *pg_out)
{
1376 1377 1378 1379 1380 1381 1382
	bool need_check_tiering;

	need_check_tiering = false;
	if (req->r_target_oloc.pool == -1) {
		req->r_target_oloc = req->r_base_oloc; /* struct */
		need_check_tiering = true;
	}
1383
	if (ceph_oid_empty(&req->r_target_oid)) {
1384 1385 1386 1387 1388 1389
		ceph_oid_copy(&req->r_target_oid, &req->r_base_oid);
		need_check_tiering = true;
	}

	if (need_check_tiering &&
	    (req->r_flags & CEPH_OSD_FLAG_IGNORE_OVERLAY) == 0) {
1390 1391
		struct ceph_pg_pool_info *pi;

1392
		pi = ceph_pg_pool_by_id(osdmap, req->r_target_oloc.pool);
1393 1394 1395
		if (pi) {
			if ((req->r_flags & CEPH_OSD_FLAG_READ) &&
			    pi->read_tier >= 0)
1396
				req->r_target_oloc.pool = pi->read_tier;
1397 1398
			if ((req->r_flags & CEPH_OSD_FLAG_WRITE) &&
			    pi->write_tier >= 0)
1399
				req->r_target_oloc.pool = pi->write_tier;
1400 1401 1402 1403
		}
		/* !pi is caught in ceph_oloc_oid_to_pg() */
	}

1404 1405
	return ceph_oloc_oid_to_pg(osdmap, &req->r_target_oloc,
				   &req->r_target_oid, pg_out);
1406 1407
}

1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
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
1424 1425 1426
/*
 * 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
1427
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1428
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1429 1430 1431 1432 1433
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1434
static int __map_request(struct ceph_osd_client *osdc,
1435
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1436
{
1437
	struct ceph_pg pgid;
1438
	int acting[CEPH_PG_MAX_SIZE];
1439
	int num, o;
S
Sage Weil 已提交
1440
	int err;
1441
	bool was_paused;
S
Sage Weil 已提交
1442

1443
	dout("map_request %p tid %lld\n", req, req->r_tid);
1444 1445

	err = __calc_request_pg(osdc->osdmap, req, &pgid);
1446 1447
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1448
		return err;
1449
	}
1450 1451
	req->r_pgid = pgid;

1452 1453 1454
	num = ceph_calc_pg_acting(osdc->osdmap, pgid, acting, &o);
	if (num < 0)
		num = 0;
S
Sage Weil 已提交
1455

1456 1457 1458 1459 1460
	was_paused = req->r_paused;
	req->r_paused = __req_should_be_paused(osdc, req);
	if (was_paused && !req->r_paused)
		force_resend = 1;

1461 1462
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1463 1464 1465
	     req->r_sent >= req->r_osd->o_incarnation &&
	     req->r_num_pg_osds == num &&
	     memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
1466 1467
	    (req->r_osd == NULL && o == -1) ||
	    req->r_paused)
S
Sage Weil 已提交
1468 1469
		return 0;  /* no change */

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

1474 1475 1476 1477
	/* record full pg acting set */
	memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
	req->r_num_pg_osds = num;

S
Sage Weil 已提交
1478 1479 1480
	if (req->r_osd) {
		__cancel_request(req);
		list_del_init(&req->r_osd_item);
1481
		list_del_init(&req->r_linger_osd_item);
S
Sage Weil 已提交
1482 1483 1484 1485 1486
		req->r_osd = NULL;
	}

	req->r_osd = __lookup_osd(osdc, o);
	if (!req->r_osd && o >= 0) {
1487
		err = -ENOMEM;
1488
		req->r_osd = create_osd(osdc, o);
1489 1490
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1491
			goto out;
1492
		}
S
Sage Weil 已提交
1493

1494
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1495 1496
		__insert_osd(osdc, req->r_osd);

1497 1498 1499
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1500 1501
	}

1502
	__enqueue_request(req);
1503
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1504 1505 1506 1507 1508 1509 1510 1511

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1512 1513
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1514
{
1515
	void *p;
S
Sage Weil 已提交
1516

1517 1518 1519 1520 1521 1522 1523
	dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
	     req, req->r_tid, req->r_osd->o_osd, req->r_flags,
	     (unsigned long long)req->r_pgid.pool, req->r_pgid.seed);

	/* fill in message content that changes each time we send it */
	put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch);
	put_unaligned_le32(req->r_flags, req->r_request_flags);
1524
	put_unaligned_le64(req->r_target_oloc.pool, req->r_request_pool);
1525 1526 1527 1528 1529 1530
	p = req->r_request_pgid;
	ceph_encode_64(&p, req->r_pgid.pool);
	ceph_encode_32(&p, req->r_pgid.seed);
	put_unaligned_le64(1, req->r_request_attempts);  /* FIXME */
	memcpy(req->r_request_reassert_version, &req->r_reassert_version,
	       sizeof(req->r_reassert_version));
1531

S
Sage Weil 已提交
1532
	req->r_stamp = jiffies;
1533
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1534 1535

	ceph_msg_get(req->r_request); /* send consumes a ref */
1536

S
Sage Weil 已提交
1537
	req->r_sent = req->r_osd->o_incarnation;
1538 1539

	ceph_con_send(&req->r_osd->o_con, req->r_request);
S
Sage Weil 已提交
1540 1541
}

1542 1543 1544
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1545
static void __send_queued(struct ceph_osd_client *osdc)
1546 1547 1548
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1549 1550
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1551 1552 1553
		__send_request(osdc, req);
}

1554 1555 1556 1557 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
/*
 * 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;
}

S
Sage Weil 已提交
1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
/*
 * Timeout callback, called every N seconds when 1 or more osd
 * requests has been active for more than N seconds.  When this
 * happens, we ping all OSDs with requests who have timed out to
 * ensure any communications channel reset is detected.  Reset the
 * request timeouts another N seconds in the future as we go.
 * Reschedule the timeout event another N seconds in future (unless
 * there are no open requests).
 */
static void handle_timeout(struct work_struct *work)
{
	struct ceph_osd_client *osdc =
		container_of(work, struct ceph_osd_client, timeout_work.work);
1601
	struct ceph_options *opts = osdc->client->options;
S
Sage Weil 已提交
1602
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1603
	struct ceph_osd *osd;
1604
	struct list_head slow_osds;
S
Sage Weil 已提交
1605 1606 1607 1608 1609 1610 1611
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1612 1613 1614 1615 1616 1617 1618
	/*
	 * 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) {
1619 1620
		if (time_before(jiffies,
				req->r_stamp + opts->osd_keepalive_timeout))
1621 1622 1623 1624 1625
			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
1626
		     req->r_tid, osd->o_osd);
1627 1628 1629 1630 1631 1632
		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
1633 1634 1635
		ceph_con_keepalive(&osd->o_con);
	}

1636
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1637
	__send_queued(osdc);
S
Sage Weil 已提交
1638 1639 1640 1641
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1642 1643 1644 1645 1646
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);
1647
	unsigned long delay = osdc->client->options->osd_idle_ttl / 4;
1648
	struct ceph_osd *osd, *nosd;
1649

1650
	dout("%s osdc %p\n", __func__, osdc);
1651
	down_read(&osdc->map_sem);
1652 1653 1654 1655 1656 1657 1658 1659
	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);
	}
1660

1661 1662
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
1663 1664 1665 1666
	schedule_delayed_work(&osdc->osds_timeout_work,
			      round_jiffies_relative(delay));
}

1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 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 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769
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;
}

S
Sage Weil 已提交
1770 1771 1772 1773 1774
static void complete_request(struct ceph_osd_request *req)
{
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1775 1776 1777 1778
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1779
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg)
S
Sage Weil 已提交
1780
{
1781
	void *p, *end;
S
Sage Weil 已提交
1782
	struct ceph_osd_request *req;
1783
	struct ceph_request_redirect redir;
S
Sage Weil 已提交
1784
	u64 tid;
1785
	int object_len;
1786 1787
	unsigned int numops;
	int payload_len, flags;
1788
	s32 result;
1789 1790 1791 1792 1793 1794
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1795
	int already_completed;
1796
	u32 bytes;
I
Ilya Dryomov 已提交
1797
	u8 decode_redir;
1798
	unsigned int i;
S
Sage Weil 已提交
1799

1800
	tid = le64_to_cpu(msg->hdr.tid);
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
	dout("handle_reply %p tid %llu\n", msg, tid);

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

	ceph_decode_need(&p, end, 4, bad);
	object_len = ceph_decode_32(&p);
	ceph_decode_need(&p, end, object_len, bad);
	p += object_len;

1811
	err = ceph_decode_pgid(&p, end, &pg);
1812
	if (err)
S
Sage Weil 已提交
1813
		goto bad;
1814 1815 1816 1817 1818 1819 1820 1821

	ceph_decode_need(&p, end, 8 + 4 + 4 + 8 + 4, bad);
	flags = ceph_decode_64(&p);
	result = ceph_decode_32(&p);
	reassert_epoch = ceph_decode_32(&p);
	reassert_version = ceph_decode_64(&p);
	osdmap_epoch = ceph_decode_32(&p);

S
Sage Weil 已提交
1822
	/* lookup */
1823
	down_read(&osdc->map_sem);
S
Sage Weil 已提交
1824 1825 1826 1827
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1828
		goto bad_mutex;
S
Sage Weil 已提交
1829 1830
	}
	ceph_osdc_get_request(req);
1831 1832 1833 1834

	dout("handle_reply %p tid %llu req %p result %d\n", msg, tid,
	     req, result);

1835
	ceph_decode_need(&p, end, 4, bad_put);
1836
	numops = ceph_decode_32(&p);
1837
	if (numops > CEPH_OSD_MAX_OPS)
1838 1839 1840 1841
		goto bad_put;
	if (numops != req->r_num_ops)
		goto bad_put;
	payload_len = 0;
1842
	ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad_put);
1843 1844 1845 1846 1847
	for (i = 0; i < numops; i++) {
		struct ceph_osd_op *op = p;
		int len;

		len = le32_to_cpu(op->payload_len);
1848
		req->r_ops[i].outdata_len = len;
1849 1850 1851 1852
		dout(" op %d has %d bytes\n", i, len);
		payload_len += len;
		p += sizeof(*op);
	}
1853 1854
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1855 1856
		pr_warn("sum of op payload lens %d != data_len %d\n",
			payload_len, bytes);
1857 1858 1859
		goto bad_put;
	}

1860
	ceph_decode_need(&p, end, 4 + numops * 4, bad_put);
1861 1862
	retry_attempt = ceph_decode_32(&p);
	for (i = 0; i < numops; i++)
1863
		req->r_ops[i].rval = ceph_decode_32(&p);
S
Sage Weil 已提交
1864

1865 1866 1867
	if (le16_to_cpu(msg->hdr.version) >= 6) {
		p += 8 + 4; /* skip replay_version */
		p += 8; /* skip user_version */
Y
Yan, Zheng 已提交
1868

I
Ilya Dryomov 已提交
1869 1870 1871 1872 1873 1874 1875 1876 1877
		if (le16_to_cpu(msg->hdr.version) >= 7)
			ceph_decode_8_safe(&p, end, decode_redir, bad_put);
		else
			decode_redir = 1;
	} else {
		decode_redir = 0;
	}

	if (decode_redir) {
1878 1879 1880 1881 1882 1883
		err = ceph_redirect_decode(&p, end, &redir);
		if (err)
			goto bad_put;
	} else {
		redir.oloc.pool = -1;
	}
S
Sage Weil 已提交
1884

1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899
	if (redir.oloc.pool != -1) {
		dout("redirect pool %lld\n", redir.oloc.pool);

		__unregister_request(osdc, req);

		req->r_target_oloc = redir.oloc; /* struct */

		/*
		 * 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.
		 */
1900
		err = __ceph_osdc_start_request(osdc, req, true);
1901 1902
		BUG_ON(err);

1903
		goto out_unlock;
1904 1905 1906 1907
	}

	already_completed = req->r_got_reply;
	if (!req->r_got_reply) {
1908
		req->r_result = result;
S
Sage Weil 已提交
1909 1910 1911 1912 1913 1914
		dout("handle_reply result %d bytes %d\n", req->r_result,
		     bytes);
		if (req->r_result == 0)
			req->r_result = bytes;

		/* in case this is a write and we need to replay, */
1915 1916
		req->r_reassert_version.epoch = cpu_to_le32(reassert_epoch);
		req->r_reassert_version.version = cpu_to_le64(reassert_version);
S
Sage Weil 已提交
1917 1918 1919 1920

		req->r_got_reply = 1;
	} else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
		dout("handle_reply tid %llu dup ack\n", tid);
1921
		goto out_unlock;
S
Sage Weil 已提交
1922 1923 1924 1925
	}

	dout("handle_reply tid %llu flags %d\n", tid, flags);

1926 1927 1928
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1929
	/* either this is a read, or we got the safe response */
1930 1931
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1932 1933 1934 1935
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

	mutex_unlock(&osdc->request_mutex);
1936
	up_read(&osdc->map_sem);
S
Sage Weil 已提交
1937

Y
Yan, Zheng 已提交
1938
	if (!already_completed) {
1939 1940 1941
		if (req->r_unsafe_callback &&
		    result >= 0 && !(flags & CEPH_OSD_FLAG_ONDISK))
			req->r_unsafe_callback(req, true);
Y
Yan, Zheng 已提交
1942 1943 1944 1945 1946
		if (req->r_callback)
			req->r_callback(req, msg);
		else
			complete_all(&req->r_completion);
	}
S
Sage Weil 已提交
1947

1948 1949 1950
	if (flags & CEPH_OSD_FLAG_ONDISK) {
		if (req->r_unsafe_callback && already_completed)
			req->r_unsafe_callback(req, false);
S
Sage Weil 已提交
1951
		complete_request(req);
1952
	}
S
Sage Weil 已提交
1953

1954
out:
1955
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1956 1957
	ceph_osdc_put_request(req);
	return;
1958 1959 1960 1961
out_unlock:
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	goto out;
S
Sage Weil 已提交
1962

1963
bad_put:
1964 1965 1966 1967 1968 1969 1970
	req->r_result = -EIO;
	__unregister_request(osdc, req);
	if (req->r_callback)
		req->r_callback(req, msg);
	else
		complete_all(&req->r_completion);
	complete_request(req);
1971
	ceph_osdc_put_request(req);
1972 1973
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
1974
	up_read(&osdc->map_sem);
S
Sage Weil 已提交
1975
bad:
1976 1977
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1978
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1979 1980
}

1981
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1982 1983 1984
{
	struct rb_node *p, *n;

1985
	dout("%s %p\n", __func__, osdc);
1986 1987
	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
1988

1989 1990 1991 1992 1993 1994 1995
		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
1996
	}
1997 1998 1999
}

/*
2000 2001
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
2002
 *
2003
 * Caller should hold map_sem for read.
2004
 */
2005 2006
static void kick_requests(struct ceph_osd_client *osdc, bool force_resend,
			  bool force_resend_writes)
2007
{
2008
	struct ceph_osd_request *req, *nreq;
2009 2010 2011
	struct rb_node *p;
	int needmap = 0;
	int err;
2012
	bool force_resend_req;
2013

2014 2015
	dout("kick_requests %s %s\n", force_resend ? " (force resend)" : "",
		force_resend_writes ? " (force resend writes)" : "");
2016
	mutex_lock(&osdc->request_mutex);
2017
	for (p = rb_first(&osdc->requests); p; ) {
2018
		req = rb_entry(p, struct ceph_osd_request, r_node);
2019
		p = rb_next(p);
2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032

		/*
		 * 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);
2033
			ceph_osdc_get_request(req);
2034 2035
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
2036
			ceph_osdc_put_request(req);
2037 2038 2039
			continue;
		}

2040 2041 2042 2043
		force_resend_req = force_resend ||
			(force_resend_writes &&
				req->r_flags & CEPH_OSD_FLAG_WRITE);
		err = __map_request(osdc, req, force_resend_req);
2044 2045 2046 2047 2048 2049
		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) {
2050 2051 2052 2053
			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);
2054
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
2055 2056
			}
		}
2057 2058 2059 2060 2061 2062
	}

	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);

2063 2064
		err = __map_request(osdc, req,
				    force_resend || force_resend_writes);
2065
		dout("__map_request returned %d\n", err);
2066 2067
		if (err < 0)
			continue;  /* hrm! */
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082
		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++;
			}
2083

2084 2085 2086 2087 2088
			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);
		}
2089
	}
2090
	reset_changed_osds(osdc);
S
Sage Weil 已提交
2091 2092 2093 2094 2095 2096
	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);
	}
2097
}
2098 2099


S
Sage Weil 已提交
2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114
/*
 * 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)
{
	void *p, *end, *next;
	u32 nr_maps, maplen;
	u32 epoch;
	struct ceph_osdmap *newmap = NULL, *oldmap;
	int err;
	struct ceph_fsid fsid;
2115
	bool was_full;
S
Sage Weil 已提交
2116 2117 2118 2119 2120 2121 2122 2123

	dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	/* verify fsid */
	ceph_decode_need(&p, end, sizeof(fsid), bad);
	ceph_decode_copy(&p, &fsid, sizeof(fsid));
2124 2125
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
2126 2127 2128

	down_write(&osdc->map_sem);

2129 2130
	was_full = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);

S
Sage Weil 已提交
2131 2132 2133 2134 2135
	/* 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);
2136 2137
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
2138 2139 2140 2141 2142 2143
		ceph_decode_need(&p, end, maplen, bad);
		next = p + maplen;
		if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
			dout("applying incremental map %u len %d\n",
			     epoch, maplen);
			newmap = osdmap_apply_incremental(&p, next,
2144
							  osdc->osdmap);
S
Sage Weil 已提交
2145 2146 2147 2148
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
2149
			BUG_ON(!newmap);
S
Sage Weil 已提交
2150 2151 2152 2153
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
2154 2155 2156 2157
			was_full = was_full ||
				ceph_osdmap_flag(osdc->osdmap,
						 CEPH_OSDMAP_FULL);
			kick_requests(osdc, 0, was_full);
S
Sage Weil 已提交
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172
		} else {
			dout("ignoring incremental map %u len %d\n",
			     epoch, maplen);
		}
		p = next;
		nr_maps--;
	}
	if (newmap)
		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);
2173 2174
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
2175 2176 2177 2178 2179 2180 2181 2182 2183
		ceph_decode_need(&p, end, maplen, bad);
		if (nr_maps > 1) {
			dout("skipping non-latest full map %u len %d\n",
			     epoch, maplen);
		} else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
			dout("skipping full map %u len %d, "
			     "older than our %u\n", epoch, maplen,
			     osdc->osdmap->epoch);
		} else {
2184 2185
			int skipped_map = 0;

S
Sage Weil 已提交
2186
			dout("taking full map %u len %d\n", epoch, maplen);
2187
			newmap = ceph_osdmap_decode(&p, p+maplen);
S
Sage Weil 已提交
2188 2189 2190 2191
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
2192
			BUG_ON(!newmap);
S
Sage Weil 已提交
2193 2194
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
2195 2196 2197
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
2198
				ceph_osdmap_destroy(oldmap);
2199
			}
2200 2201 2202 2203
			was_full = was_full ||
				ceph_osdmap_flag(osdc->osdmap,
						 CEPH_OSDMAP_FULL);
			kick_requests(osdc, skipped_map, was_full);
S
Sage Weil 已提交
2204 2205 2206 2207 2208
		}
		p += maplen;
		nr_maps--;
	}

2209 2210
	if (!osdc->osdmap)
		goto bad;
S
Sage Weil 已提交
2211 2212
done:
	downgrade_write(&osdc->map_sem);
2213 2214
	ceph_monc_got_map(&osdc->client->monc, CEPH_SUB_OSDMAP,
			  osdc->osdmap->epoch);
2215 2216 2217 2218 2219 2220

	/*
	 * subscribe to subsequent osdmap updates if full to ensure
	 * we find out when we are no longer full and stop returning
	 * ENOSPC.
	 */
2221 2222 2223
	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))
2224 2225
		ceph_monc_request_next_osdmap(&osdc->client->monc);

A
Alex Elder 已提交
2226 2227 2228
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2229
	up_read(&osdc->map_sem);
2230
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
2231 2232 2233 2234
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
2235
	ceph_msg_dump(msg);
S
Sage Weil 已提交
2236 2237 2238
	up_write(&osdc->map_sem);
}

2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320
/*
 * 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 *),
2321
			   void *data, struct ceph_osd_event **pevent)
2322 2323 2324 2325 2326 2327 2328 2329 2330
{
	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;
2331
	event->one_shot = 0;
2332 2333 2334
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
2335
	RB_CLEAR_NODE(&event->node);
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 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381
	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
 */
2382 2383
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403
{
	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) {
2404
		BUG_ON(event->one_shot);
2405 2406 2407 2408 2409 2410 2411 2412
		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) {
2413 2414 2415
			pr_err("couldn't allocate event_work\n");
			ceph_osdc_put_event(event);
			return;
2416
		}
2417
		INIT_WORK(&event_work->work, do_event_work);
2418 2419 2420 2421 2422
		event_work->event = event;
		event_work->ver = ver;
		event_work->notify_id = notify_id;
		event_work->opcode = opcode;

2423 2424
		queue_work(osdc->notify_wq, &event_work->work);
	}
2425 2426 2427 2428 2429 2430 2431

	return;

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

2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
				struct ceph_snap_context *snapc, u64 snap_id,
				struct timespec *mtime)
{
	struct ceph_msg *msg = req->r_request;
	void *p;
	size_t msg_size;
	int flags = req->r_flags;
	u64 data_len;
	unsigned int i;

	req->r_snapid = snap_id;
2448
	WARN_ON(snapc != req->r_snapc);
2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479

	/* encode request */
	msg->hdr.version = cpu_to_le16(4);

	p = msg->front.iov_base;
	ceph_encode_32(&p, 1);   /* client_inc  is always 1 */
	req->r_request_osdmap_epoch = p;
	p += 4;
	req->r_request_flags = p;
	p += 4;
	if (req->r_flags & CEPH_OSD_FLAG_WRITE)
		ceph_encode_timespec(p, mtime);
	p += sizeof(struct ceph_timespec);
	req->r_request_reassert_version = p;
	p += sizeof(struct ceph_eversion); /* will get filled in */

	/* oloc */
	ceph_encode_8(&p, 4);
	ceph_encode_8(&p, 4);
	ceph_encode_32(&p, 8 + 4 + 4);
	req->r_request_pool = p;
	p += 8;
	ceph_encode_32(&p, -1);  /* preferred */
	ceph_encode_32(&p, 0);   /* key len */

	ceph_encode_8(&p, 1);
	req->r_request_pgid = p;
	p += 8 + 4;
	ceph_encode_32(&p, -1);  /* preferred */

	/* oid */
2480 2481
	ceph_encode_32(&p, req->r_base_oid.name_len);
	memcpy(p, req->r_base_oid.name, req->r_base_oid.name_len);
2482
	dout("oid %*pE len %d\n", req->r_base_oid.name_len,
2483 2484
	     req->r_base_oid.name, req->r_base_oid.name_len);
	p += req->r_base_oid.name_len;
2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498

	/* ops--can imply data */
	ceph_encode_16(&p, (u16)req->r_num_ops);
	data_len = 0;
	for (i = 0; i < req->r_num_ops; i++) {
		data_len += osd_req_encode_op(req, p, i);
		p += sizeof(struct ceph_osd_op);
	}

	/* snaps */
	ceph_encode_64(&p, req->r_snapid);
	ceph_encode_64(&p, req->r_snapc ? req->r_snapc->seq : 0);
	ceph_encode_32(&p, req->r_snapc ? req->r_snapc->num_snaps : 0);
	if (req->r_snapc) {
2499
		for (i = 0; i < req->r_snapc->num_snaps; i++) {
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
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
		}
	}

	req->r_request_attempts = p;
	p += 4;

	/* data */
	if (flags & CEPH_OSD_FLAG_WRITE) {
		u16 data_off;

		/*
		 * 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).
		 */
		data_off = (u16) (off & 0xffff);
		req->r_request->hdr.data_off = cpu_to_le16(data_off);
	}
	req->r_request->hdr.data_len = cpu_to_le32(data_len);

	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);

	dout("build_request msg_size was %d\n", (int)msg_size);
}
EXPORT_SYMBOL(ceph_osdc_build_request);

2531 2532 2533 2534 2535 2536 2537
/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
2538
	int rc;
2539

S
Sage Weil 已提交
2540 2541
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2542 2543 2544

	rc = __ceph_osdc_start_request(osdc, req, nofail);

S
Sage Weil 已提交
2545 2546
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
2547

S
Sage Weil 已提交
2548 2549
	return rc;
}
2550
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2551

2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570
/*
 * 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
2571 2572 2573 2574 2575 2576 2577 2578
/*
 * wait for a request to complete
 */
int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
{
	int rc;

2579 2580
	dout("%s %p tid %llu\n", __func__, req, req->r_tid);

S
Sage Weil 已提交
2581 2582
	rc = wait_for_completion_interruptible(&req->r_completion);
	if (rc < 0) {
2583 2584
		dout("%s %p tid %llu interrupted\n", __func__, req, req->r_tid);
		ceph_osdc_cancel_request(req);
S
Sage Weil 已提交
2585
		complete_request(req);
S
Sage Weil 已提交
2586 2587 2588
		return rc;
	}

2589 2590
	dout("%s %p tid %llu result %d\n", __func__, req, req->r_tid,
	     req->r_result);
S
Sage Weil 已提交
2591 2592
	return req->r_result;
}
2593
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
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 2626

/*
 * 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);
}
2627
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2628

2629 2630 2631 2632
/*
 * 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
2633
void ceph_osdc_flush_notifies(struct ceph_osd_client *osdc)
2634 2635 2636 2637 2638 2639
{
	flush_workqueue(osdc->notify_wq);
}
EXPORT_SYMBOL(ceph_osdc_flush_notifies);


S
Sage Weil 已提交
2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653
/*
 * init, shutdown
 */
int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
{
	int err;

	dout("init\n");
	osdc->client = client;
	osdc->osdmap = NULL;
	init_rwsem(&osdc->map_sem);
	mutex_init(&osdc->request_mutex);
	osdc->last_tid = 0;
	osdc->osds = RB_ROOT;
2654
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
2655
	osdc->requests = RB_ROOT;
2656
	INIT_LIST_HEAD(&osdc->req_lru);
2657 2658
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2659
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2660 2661
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2662
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2663 2664 2665
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2666 2667

	schedule_delayed_work(&osdc->osds_timeout_work,
2668
	    round_jiffies_relative(osdc->client->options->osd_idle_ttl));
S
Sage Weil 已提交
2669

2670
	err = -ENOMEM;
2671 2672
	osdc->req_mempool = mempool_create_slab_pool(10,
						     ceph_osd_request_cache);
S
Sage Weil 已提交
2673
	if (!osdc->req_mempool)
2674
		goto out;
S
Sage Weil 已提交
2675

2676
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
2677
				PAGE_SIZE, 10, true, "osd_op");
S
Sage Weil 已提交
2678
	if (err < 0)
2679
		goto out_mempool;
2680
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2681
				PAGE_SIZE, 10, true, "osd_op_reply");
S
Sage Weil 已提交
2682 2683
	if (err < 0)
		goto out_msgpool;
2684

2685
	err = -ENOMEM;
2686
	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
2687
	if (!osdc->notify_wq)
2688 2689
		goto out_msgpool_reply;

S
Sage Weil 已提交
2690
	return 0;
2691

2692 2693
out_msgpool_reply:
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2694 2695
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2696 2697 2698 2699
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2700 2701 2702 2703
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2704 2705
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2706
	cancel_delayed_work_sync(&osdc->timeout_work);
2707
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
2708 2709 2710 2711 2712 2713 2714 2715 2716

	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);

S
Sage Weil 已提交
2717 2718 2719 2720 2721 2722
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2723
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2724 2725 2726 2727 2728 2729 2730 2731 2732 2733
}

/*
 * 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,
2734
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2735 2736 2737 2738 2739 2740
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2741
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 0, 1,
S
Sage Weil 已提交
2742
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2743
				    NULL, truncate_seq, truncate_size,
2744
				    false);
2745 2746
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2747 2748

	/* it may be a short read due to an object boundary */
2749

2750
	osd_req_op_extent_osd_data_pages(req, 0,
2751
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2752

2753
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2754
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2755

2756
	ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2757

S
Sage Weil 已提交
2758 2759 2760 2761 2762 2763 2764 2765
	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;
}
2766
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2767 2768 2769 2770 2771 2772 2773 2774 2775 2776

/*
 * 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,
2777
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2778 2779 2780
{
	struct ceph_osd_request *req;
	int rc = 0;
2781
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2782

2783
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
2784
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 0, 1,
S
Sage Weil 已提交
2785
				    CEPH_OSD_OP_WRITE,
2786
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2787
				    snapc, truncate_seq, truncate_size,
2788
				    true);
2789 2790
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2791 2792

	/* it may be a short write due to an object boundary */
2793
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2794 2795
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2796

2797
	ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2798

2799
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2800 2801 2802 2803 2804 2805 2806 2807 2808
	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;
}
2809
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2810

2811 2812
int ceph_osdc_setup(void)
{
2813 2814 2815
	size_t size = sizeof(struct ceph_osd_request) +
	    CEPH_OSD_SLAB_OPS * sizeof(struct ceph_osd_req_op);

2816
	BUG_ON(ceph_osd_request_cache);
2817 2818
	ceph_osd_request_cache = kmem_cache_create("ceph_osd_request", size,
						   0, 0, NULL);
2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831

	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
2832 2833 2834 2835 2836 2837
/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2838
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2839 2840 2841
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2842
		goto out;
2843
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2844 2845 2846 2847 2848 2849

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2850
		handle_reply(osdc, msg);
S
Sage Weil 已提交
2851
		break;
2852 2853 2854
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2855 2856 2857 2858 2859

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2860
out:
S
Sage Weil 已提交
2861 2862 2863
	ceph_msg_put(msg);
}

2864
/*
2865 2866 2867
 * 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.
2868 2869
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2870 2871
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2872 2873 2874
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2875
	struct ceph_msg *m;
2876
	struct ceph_osd_request *req;
2877
	int front_len = le32_to_cpu(hdr->front_len);
2878
	int data_len = le32_to_cpu(hdr->data_len);
2879
	u64 tid;
S
Sage Weil 已提交
2880

2881 2882 2883 2884
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
2885 2886
		dout("%s osd%d tid %llu unknown, skipping\n", __func__,
		     osd->o_osd, tid);
2887
		m = NULL;
2888
		*skip = 1;
2889 2890
		goto out;
	}
S
Sage Weil 已提交
2891

2892
	ceph_msg_revoke_incoming(req->r_reply);
2893

2894
	if (front_len > req->r_reply->front_alloc_len) {
2895 2896 2897
		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);
2898 2899
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front_len, GFP_NOFS,
				 false);
2900
		if (!m)
S
Sage Weil 已提交
2901 2902 2903 2904
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
2905

2906 2907 2908 2909 2910 2911 2912
	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;
2913
	}
2914 2915

	m = ceph_msg_get(req->r_reply);
S
Sage Weil 已提交
2916
	dout("get_reply tid %lld %p\n", tid, m);
2917 2918 2919

out:
	mutex_unlock(&osdc->request_mutex);
2920
	return m;
2921 2922 2923 2924 2925 2926 2927 2928 2929 2930
}

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
2931
	*skip = 0;
2932 2933
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2934
	case CEPH_MSG_WATCH_NOTIFY:
2935
		return ceph_msg_new(type, front, GFP_NOFS, false);
2936 2937 2938 2939 2940 2941 2942 2943
	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
2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962
}

/*
 * 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);
}

2963 2964 2965
/*
 * authentication
 */
2966 2967 2968 2969 2970
/*
 * 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,
2971
					int *proto, int force_new)
2972 2973 2974 2975
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2976
	struct ceph_auth_handshake *auth = &o->o_auth;
2977

2978
	if (force_new && auth->authorizer) {
2979
		ceph_auth_destroy_authorizer(auth->authorizer);
2980 2981
		auth->authorizer = NULL;
	}
2982 2983 2984
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2985
		if (ret)
2986
			return ERR_PTR(ret);
2987 2988
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2989 2990 2991
						     auth);
		if (ret)
			return ERR_PTR(ret);
2992 2993
	}
	*proto = ac->protocol;
2994

2995
	return auth;
2996 2997 2998 2999 3000 3001 3002 3003 3004
}


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;

3005
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
3006 3007
}

3008 3009 3010 3011 3012 3013
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;

3014
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
3015 3016
	return ceph_monc_validate_auth(&osdc->client->monc);
}
3017

3018
static int osd_sign_message(struct ceph_msg *msg)
Y
Yan, Zheng 已提交
3019
{
3020
	struct ceph_osd *o = msg->con->private;
Y
Yan, Zheng 已提交
3021
	struct ceph_auth_handshake *auth = &o->o_auth;
3022

Y
Yan, Zheng 已提交
3023 3024 3025
	return ceph_auth_sign_message(auth, msg);
}

3026
static int osd_check_message_signature(struct ceph_msg *msg)
Y
Yan, Zheng 已提交
3027
{
3028
	struct ceph_osd *o = msg->con->private;
Y
Yan, Zheng 已提交
3029
	struct ceph_auth_handshake *auth = &o->o_auth;
3030

Y
Yan, Zheng 已提交
3031 3032 3033
	return ceph_auth_check_message_signature(auth, msg);
}

3034
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
3035 3036 3037
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
3038 3039
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
3040
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
3041
	.alloc_msg = alloc_msg,
3042 3043
	.sign_message = osd_sign_message,
	.check_message_signature = osd_check_message_signature,
3044
	.fault = osd_reset,
S
Sage Weil 已提交
3045
};