osd_client.c 60.8 KB
Newer Older
1
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
2

3
#include <linux/module.h>
S
Sage Weil 已提交
4 5 6 7 8 9
#include <linux/err.h>
#include <linux/highmem.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
10 11 12
#ifdef CONFIG_BLOCK
#include <linux/bio.h>
#endif
S
Sage Weil 已提交
13

14 15 16 17 18 19
#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
20

S
Sage Weil 已提交
21 22
#define OSD_OP_FRONT_LEN	4096
#define OSD_OPREPLY_FRONT_LEN	512
23

24
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
25

A
Alex Elder 已提交
26
static void __send_queued(struct ceph_osd_client *osdc);
27
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
28 29 30 31
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req);
static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req);
32 33
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

/*
 * 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.
 */
60
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
61
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
62
{
63
	u64 orig_len = *plen;
S
Sage Weil 已提交
64
	int r;
S
Sage Weil 已提交
65

66
	/* object extent? */
67 68
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
69 70
	if (r < 0)
		return r;
71 72
	if (*objlen < orig_len) {
		*plen = *objlen;
73 74 75 76
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

77
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
78

79
	return 0;
S
Sage Weil 已提交
80 81 82 83 84
}

/*
 * requests
 */
S
Sage Weil 已提交
85
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
86
{
87
	int num_pages;
S
Sage Weil 已提交
88 89 90 91 92 93
	struct ceph_osd_request *req = container_of(kref,
						    struct ceph_osd_request,
						    r_kref);

	if (req->r_request)
		ceph_msg_put(req->r_request);
94
	if (req->r_reply) {
95
		ceph_msg_revoke_incoming(req->r_reply);
96
		ceph_msg_put(req->r_reply);
97
	}
98 99

	if (req->r_data_in.type == CEPH_OSD_DATA_TYPE_PAGES &&
100 101 102 103 104
			req->r_data_in.own_pages) {
		num_pages = calc_pages_for((u64)req->r_data_in.alignment,
						(u64)req->r_data_in.length);
		ceph_release_page_vector(req->r_data_in.pages, num_pages);
	}
105
	if (req->r_data_out.type == CEPH_OSD_DATA_TYPE_PAGES &&
106 107 108 109 110
			req->r_data_out.own_pages) {
		num_pages = calc_pages_for((u64)req->r_data_out.alignment,
						(u64)req->r_data_out.length);
		ceph_release_page_vector(req->r_data_out.pages, num_pages);
	}
111

S
Sage Weil 已提交
112 113 114 115 116
	ceph_put_snap_context(req->r_snapc);
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
		kfree(req);
S
Sage Weil 已提交
117
}
118
EXPORT_SYMBOL(ceph_osdc_release_request);
119

120
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
121
					       struct ceph_snap_context *snapc,
122
					       unsigned int num_ops,
123
					       bool use_mempool,
124
					       gfp_t gfp_flags)
S
Sage Weil 已提交
125 126 127
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
128 129 130 131 132 133 134 135 136 137 138
	size_t msg_size;

	msg_size = 4 + 4 + 8 + 8 + 4+8;
	msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
	msg_size += 1 + 8 + 4 + 4;     /* pg_t */
	msg_size += 4 + MAX_OBJ_NAME_SIZE;
	msg_size += 2 + num_ops*sizeof(struct ceph_osd_op);
	msg_size += 8;  /* snapid */
	msg_size += 8;  /* snap_seq */
	msg_size += 8 * (snapc ? snapc->num_snaps : 0);  /* snaps */
	msg_size += 4;
S
Sage Weil 已提交
139 140

	if (use_mempool) {
141
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
142 143
		memset(req, 0, sizeof(*req));
	} else {
144
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
145 146
	}
	if (req == NULL)
147
		return NULL;
S
Sage Weil 已提交
148 149 150

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
151

S
Sage Weil 已提交
152
	kref_init(&req->r_kref);
S
Sage Weil 已提交
153 154
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
155
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
156
	INIT_LIST_HEAD(&req->r_unsafe_item);
157 158
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
159
	INIT_LIST_HEAD(&req->r_req_lru_item);
160 161
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
162 163 164 165 166
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
167
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
168
	if (!msg) {
S
Sage Weil 已提交
169
		ceph_osdc_put_request(req);
170
		return NULL;
S
Sage Weil 已提交
171 172 173
	}
	req->r_reply = msg;

174 175
	req->r_data_in.type = CEPH_OSD_DATA_TYPE_NONE;
	req->r_data_out.type = CEPH_OSD_DATA_TYPE_NONE;
176

S
Sage Weil 已提交
177
	/* create request message; allow space for oid */
S
Sage Weil 已提交
178
	if (use_mempool)
179
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
180
	else
181
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
182
	if (!msg) {
S
Sage Weil 已提交
183
		ceph_osdc_put_request(req);
184
		return NULL;
S
Sage Weil 已提交
185
	}
186

S
Sage Weil 已提交
187
	memset(msg->front.iov_base, 0, msg->front.iov_len);
188 189 190 191 192

	req->r_request = msg;

	return req;
}
193
EXPORT_SYMBOL(ceph_osdc_alloc_request);
194

195
static bool osd_req_opcode_valid(u16 opcode)
196
{
197
	switch (opcode) {
198
	case CEPH_OSD_OP_READ:
199
	case CEPH_OSD_OP_STAT:
200 201 202
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
203
	case CEPH_OSD_OP_NOTIFY:
204
	case CEPH_OSD_OP_NOTIFY_ACK:
205
	case CEPH_OSD_OP_ASSERT_VER:
206
	case CEPH_OSD_OP_WRITE:
207 208 209 210 211
	case CEPH_OSD_OP_WRITEFULL:
	case CEPH_OSD_OP_TRUNCATE:
	case CEPH_OSD_OP_ZERO:
	case CEPH_OSD_OP_DELETE:
	case CEPH_OSD_OP_APPEND:
212
	case CEPH_OSD_OP_STARTSYNC:
213 214 215 216 217 218
	case CEPH_OSD_OP_SETTRUNC:
	case CEPH_OSD_OP_TRIMTRUNC:
	case CEPH_OSD_OP_TMAPUP:
	case CEPH_OSD_OP_TMAPPUT:
	case CEPH_OSD_OP_TMAPGET:
	case CEPH_OSD_OP_CREATE:
219
	case CEPH_OSD_OP_ROLLBACK:
220
	case CEPH_OSD_OP_WATCH:
221 222 223 224 225 226 227 228 229 230 231 232
	case CEPH_OSD_OP_OMAPGETKEYS:
	case CEPH_OSD_OP_OMAPGETVALS:
	case CEPH_OSD_OP_OMAPGETHEADER:
	case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
	case CEPH_OSD_OP_OMAPSETVALS:
	case CEPH_OSD_OP_OMAPSETHEADER:
	case CEPH_OSD_OP_OMAPCLEAR:
	case CEPH_OSD_OP_OMAPRMKEYS:
	case CEPH_OSD_OP_OMAP_CMP:
	case CEPH_OSD_OP_CLONERANGE:
	case CEPH_OSD_OP_ASSERT_SRC_VERSION:
	case CEPH_OSD_OP_SRC_CMPXATTR:
233
	case CEPH_OSD_OP_GETXATTR:
234
	case CEPH_OSD_OP_GETXATTRS:
235 236
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
	case CEPH_OSD_OP_SETXATTRS:
	case CEPH_OSD_OP_RESETXATTRS:
	case CEPH_OSD_OP_RMXATTR:
	case CEPH_OSD_OP_PULL:
	case CEPH_OSD_OP_PUSH:
	case CEPH_OSD_OP_BALANCEREADS:
	case CEPH_OSD_OP_UNBALANCEREADS:
	case CEPH_OSD_OP_SCRUB:
	case CEPH_OSD_OP_SCRUB_RESERVE:
	case CEPH_OSD_OP_SCRUB_UNRESERVE:
	case CEPH_OSD_OP_SCRUB_STOP:
	case CEPH_OSD_OP_SCRUB_MAP:
	case CEPH_OSD_OP_WRLOCK:
	case CEPH_OSD_OP_WRUNLOCK:
	case CEPH_OSD_OP_RDLOCK:
	case CEPH_OSD_OP_RDUNLOCK:
	case CEPH_OSD_OP_UPLOCK:
	case CEPH_OSD_OP_DNLOCK:
255
	case CEPH_OSD_OP_CALL:
256 257
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
258 259 260 261 262 263
		return true;
	default:
		return false;
	}
}

264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298
/*
 * 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.
 */
void osd_req_op_init(struct ceph_osd_req_op *op, u16 opcode)
{
	BUG_ON(!osd_req_opcode_valid(opcode));

	memset(op, 0, sizeof (*op));

	op->op = opcode;
}

void osd_req_op_extent_init(struct ceph_osd_req_op *op, u16 opcode,
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
	size_t payload_len = 0;

	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);

	osd_req_op_init(op, opcode);

	op->extent.offset = offset;
	op->extent.length = length;
	op->extent.truncate_size = truncate_size;
	op->extent.truncate_seq = truncate_seq;
	if (opcode == CEPH_OSD_OP_WRITE)
		payload_len += length;

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_extent_init);

299 300 301 302 303 304 305 306 307 308 309 310 311
void osd_req_op_extent_update(struct ceph_osd_req_op *op, u64 length)
{
	u64 previous = op->extent.length;

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

	op->extent.length = length;
	op->payload_len -= previous - length;
}
EXPORT_SYMBOL(osd_req_op_extent_update);

312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
void osd_req_op_cls_init(struct ceph_osd_req_op *op, u16 opcode,
			const char *class, const char *method,
			const void *request_data, size_t request_data_size)
{
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

	osd_req_op_init(op, opcode);

	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
	payload_len += size;

	op->cls.indata = request_data;
	BUG_ON(request_data_size > (size_t) U32_MAX);
	op->cls.indata_len = (u32) request_data_size;
	payload_len += request_data_size;

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

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_cls_init);

void osd_req_op_watch_init(struct ceph_osd_req_op *op, u16 opcode,
				u64 cookie, u64 version, int flag)
{
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);

	osd_req_op_init(op, opcode);

	op->watch.cookie = cookie;
	/* op->watch.ver = version; */	/* XXX 3847 */
	op->watch.ver = cpu_to_le64(version);
	if (opcode == CEPH_OSD_OP_WATCH && flag)
		op->watch.flag = (u8) 1;
}
EXPORT_SYMBOL(osd_req_op_watch_init);

361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415
static u64 osd_req_encode_op(struct ceph_osd_request *req,
			      struct ceph_osd_op *dst,
			      struct ceph_osd_req_op *src)
{
	u64 out_data_len = 0;
	struct ceph_pagelist *pagelist;

	if (WARN_ON(!osd_req_opcode_valid(src->op))) {
		pr_err("unrecognized osd opcode %d\n", src->op);

		return 0;
	}

	switch (src->op) {
	case CEPH_OSD_OP_STAT:
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		if (src->op == CEPH_OSD_OP_WRITE)
			out_data_len = src->extent.length;
		dst->extent.offset = cpu_to_le64(src->extent.offset);
		dst->extent.length = cpu_to_le64(src->extent.length);
		dst->extent.truncate_size =
			cpu_to_le64(src->extent.truncate_size);
		dst->extent.truncate_seq =
			cpu_to_le32(src->extent.truncate_seq);
		break;
	case CEPH_OSD_OP_CALL:
		pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
		BUG_ON(!pagelist);
		ceph_pagelist_init(pagelist);

		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
		dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
		ceph_pagelist_append(pagelist, src->cls.class_name,
				     src->cls.class_len);
		ceph_pagelist_append(pagelist, src->cls.method_name,
				     src->cls.method_len);
		ceph_pagelist_append(pagelist, src->cls.indata,
				     src->cls.indata_len);

		req->r_data_out.type = CEPH_OSD_DATA_TYPE_PAGELIST;
		req->r_data_out.pagelist = pagelist;
		out_data_len = pagelist->length;
		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;
	default:
416
		pr_err("unsupported osd opcode %s\n",
417
			ceph_osd_op_name(src->op));
418
		WARN_ON(1);
419 420

		return 0;
421
	}
422
	dst->op = cpu_to_le16(src->op);
423
	dst->payload_len = cpu_to_le32(src->payload_len);
424 425

	return out_data_len;
426 427
}

428 429 430 431 432
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req,
433
			     u64 off, unsigned int num_ops,
434
			     struct ceph_osd_req_op *src_ops,
435
			     struct ceph_snap_context *snapc, u64 snap_id,
436
			     struct timespec *mtime)
437 438
{
	struct ceph_msg *msg = req->r_request;
439
	struct ceph_osd_req_op *src_op;
440
	void *p;
441
	size_t msg_size;
442
	int flags = req->r_flags;
443
	u64 data_len;
444
	int i;
445

446 447
	req->r_num_ops = num_ops;
	req->r_snapid = snap_id;
S
Sage Weil 已提交
448 449
	req->r_snapc = ceph_get_snap_context(snapc);

450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
	/* 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 */
	ceph_encode_32(&p, req->r_oid_len);
481
	memcpy(p, req->r_oid, req->r_oid_len);
482
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
483
	p += req->r_oid_len;
S
Sage Weil 已提交
484

485
	/* ops--can imply data */
486
	ceph_encode_16(&p, num_ops);
487
	src_op = src_ops;
488
	req->r_request_ops = p;
489
	data_len = 0;
490
	for (i = 0; i < num_ops; i++, src_op++) {
491
		data_len += osd_req_encode_op(req, p, src_op);
492 493
		p += sizeof(struct ceph_osd_op);
	}
494

495 496 497 498 499
	/* 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) {
S
Sage Weil 已提交
500
		for (i = 0; i < snapc->num_snaps; i++) {
501
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
S
Sage Weil 已提交
502 503 504
		}
	}

505 506 507
	req->r_request_attempts = p;
	p += 4;

508
	/* data */
509 510 511 512 513 514 515 516 517 518 519 520
	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);
	}
521
	req->r_request->hdr.data_len = cpu_to_le32(data_len);
522

S
Sage Weil 已提交
523
	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
524 525 526
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);
527

528
	dout("build_request msg_size was %d\n", (int)msg_size);
529
}
530
EXPORT_SYMBOL(ceph_osdc_build_request);
531 532 533 534 535 536 537 538 539 540 541 542 543 544 545

/*
 * 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,
546 547
					       u64 off, u64 *plen, int num_ops,
					       struct ceph_osd_req_op *ops,
548 549 550 551
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
552
					       bool use_mempool)
553
{
554
	struct ceph_osd_request *req;
555 556 557
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
558 559
	u32 object_size;
	u64 object_base;
560
	int r;
561

562
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
563

564
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
565
					GFP_NOFS);
566
	if (!req)
567
		return ERR_PTR(-ENOMEM);
568
	req->r_flags = flags;
569 570

	/* calculate max write size */
571
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
572 573
	if (r < 0) {
		ceph_osdc_put_request(req);
574
		return ERR_PTR(r);
575
	}
576

577 578 579 580 581 582 583 584
	object_size = le32_to_cpu(layout->fl_object_size);
	object_base = off - objoff;
	if (truncate_size <= object_base) {
		truncate_size = 0;
	} else {
		truncate_size -= object_base;
		if (truncate_size > object_size)
			truncate_size = object_size;
585
	}
586

587 588
	osd_req_op_extent_init(&ops[0], opcode, objoff, objlen,
				truncate_size, truncate_seq);
589 590 591 592 593 594
	/*
	 * A second op in the ops array means the caller wants to
	 * also issue a include a 'startsync' command so that the
	 * osd will flush data quickly.
	 */
	if (num_ops > 1)
595
		osd_req_op_init(&ops[1], CEPH_OSD_OP_STARTSYNC);
596

597 598
	req->r_file_layout = *layout;  /* keep a copy */

599 600
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx",
		vino.ino, objnum);
601 602
	req->r_oid_len = strlen(req->r_oid);

S
Sage Weil 已提交
603 604
	return req;
}
605
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671

/*
 * 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;
}

672 673 674 675 676 677
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
678
	struct ceph_osd_request *req, *nreq;
679
	LIST_HEAD(resend);
680 681 682 683
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
684
	if (err)
685
		return;
686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702
	/*
	 * 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.
	 */
703
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
704 705 706 707
		if (!req->r_sent)
			break;
		list_move_tail(&req->r_req_lru_item, &resend);
		dout("requeueing %p tid %llu osd%d\n", req, req->r_tid,
708
		     osd->o_osd);
709 710 711
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}
712
	list_splice(&resend, &osdc->req_unsent);
713

714 715 716 717 718
	/*
	 * Linger requests are re-registered before sending, which
	 * sets up a new tid for each.  We add them to the unsent
	 * list at the end to keep things in tid order.
	 */
719 720
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
721 722 723 724 725
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
726
		__register_request(osdc, req);
727
		list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
728
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
729
		__unregister_linger_request(osdc, req);
730 731
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
732 733 734
	}
}

S
Sage Weil 已提交
735
/*
736
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
737 738 739 740 741 742 743 744 745 746 747
 */
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
748 749
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
750
	__send_queued(osdc);
S
Sage Weil 已提交
751
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
752 753 754 755 756 757
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
758
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
759 760 761 762 763 764 765 766 767
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
768
	osd->o_osd = onum;
769
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
770
	INIT_LIST_HEAD(&osd->o_requests);
771
	INIT_LIST_HEAD(&osd->o_linger_requests);
772
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
773 774
	osd->o_incarnation = 1;

775
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
776

777
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796
	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);
797
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
798 799
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

800
		ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
801
		kfree(osd);
S
Sage Weil 已提交
802
	}
S
Sage Weil 已提交
803 804 805 806 807
}

/*
 * remove an osd from our map
 */
808
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
809
{
810
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
811 812
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
813
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
814 815 816 817
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

818 819
static void remove_all_osds(struct ceph_osd_client *osdc)
{
820
	dout("%s %p\n", __func__, osdc);
821 822 823 824 825 826 827 828 829
	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);
}

830 831 832 833 834 835
static void __move_osd_to_lru(struct ceph_osd_client *osdc,
			      struct ceph_osd *osd)
{
	dout("__move_osd_to_lru %p\n", osd);
	BUG_ON(!list_empty(&osd->o_osd_lru));
	list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
836
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
837 838 839 840 841 842 843 844 845
}

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);
}

846
static void remove_old_osds(struct ceph_osd_client *osdc)
847 848 849 850 851 852
{
	struct ceph_osd *osd, *nosd;

	dout("__remove_old_osds %p\n", osdc);
	mutex_lock(&osdc->request_mutex);
	list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
853
		if (time_before(jiffies, osd->lru_ttl))
854 855 856 857 858 859
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
860 861 862
/*
 * reset osd connect
 */
863
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
864
{
A
Alex Elder 已提交
865
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
866

867
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
868 869
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
870
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
871 872 873 874 875 876 877 878 879

		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;

880 881 882 883 884
		dout(" osd addr hasn't changed and connection never opened,"
		     " letting msgr retry");
		/* 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
885 886

		return -EAGAIN;
S
Sage Weil 已提交
887
	}
A
Alex Elder 已提交
888 889 890 891 892 893

	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
894 895 896 897 898 899 900 901
}

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;

902
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
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
	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;
}

935 936 937
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
938
			osdc->client->options->osd_keepalive_timeout * HZ);
939 940 941 942 943 944
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
945 946 947 948 949

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
950 951
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
952 953
{
	req->r_tid = ++osdc->last_tid;
954
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
955
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
956 957 958 959
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
960 961
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
962
	}
963 964
}

S
Sage Weil 已提交
965 966 967 968 969 970
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
971 972 973 974 975 976
	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
977 978 979 980
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

981 982
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
983
		ceph_msg_revoke(req->r_request);
984 985

		list_del_init(&req->r_osd_item);
986 987 988
		if (list_empty(&req->r_osd->o_requests) &&
		    list_empty(&req->r_osd->o_linger_requests)) {
			dout("moving osd to %p lru\n", req->r_osd);
989
			__move_osd_to_lru(osdc, req->r_osd);
990
		}
991
		if (list_empty(&req->r_linger_item))
992
			req->r_osd = NULL;
993
	}
S
Sage Weil 已提交
994

A
Alex Elder 已提交
995
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
996 997
	ceph_osdc_put_request(req);

998 999 1000
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1001 1002 1003 1004 1005 1006 1007 1008
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1009
	if (req->r_sent && req->r_osd) {
1010
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1011 1012 1013 1014
		req->r_sent = 0;
	}
}

1015 1016 1017 1018 1019
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
	dout("__register_linger_request %p\n", req);
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1020 1021 1022
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
1023 1024 1025 1026 1027 1028
}

static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req)
{
	dout("__unregister_linger_request %p\n", req);
A
Alex Elder 已提交
1029
	list_del_init(&req->r_linger_item);
1030 1031 1032 1033 1034 1035 1036 1037
	if (req->r_osd) {
		list_del_init(&req->r_linger_osd);

		if (list_empty(&req->r_osd->o_requests) &&
		    list_empty(&req->r_osd->o_linger_requests)) {
			dout("moving osd to %p lru\n", req->r_osd);
			__move_osd_to_lru(osdc, req->r_osd);
		}
1038 1039
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
	}
}

void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
					 struct ceph_osd_request *req)
{
	mutex_lock(&osdc->request_mutex);
	if (req->r_linger) {
		__unregister_linger_request(osdc, req);
		ceph_osdc_put_request(req);
	}
	mutex_unlock(&osdc->request_mutex);
}
EXPORT_SYMBOL(ceph_osdc_unregister_linger_request);

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;
		/*
		 * caller is now responsible for calling
		 * unregister_linger_request
		 */
		ceph_osdc_get_request(req);
	}
}
EXPORT_SYMBOL(ceph_osdc_set_request_linger);

S
Sage Weil 已提交
1070 1071 1072
/*
 * 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
1073
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1074
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1075 1076 1077 1078 1079
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1080
static int __map_request(struct ceph_osd_client *osdc,
1081
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1082
{
1083
	struct ceph_pg pgid;
1084 1085
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
1086 1087
	int err;

1088
	dout("map_request %p tid %lld\n", req, req->r_tid);
1089 1090
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
1091 1092
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1093
		return err;
1094
	}
1095 1096
	req->r_pgid = pgid;

1097 1098 1099 1100 1101
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
1102

1103 1104
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1105 1106 1107
	     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) ||
S
Sage Weil 已提交
1108 1109 1110
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

1111 1112
	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
1113 1114
	     req->r_osd ? req->r_osd->o_osd : -1);

1115 1116 1117 1118
	/* 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
1119 1120 1121 1122 1123 1124 1125 1126
	if (req->r_osd) {
		__cancel_request(req);
		list_del_init(&req->r_osd_item);
		req->r_osd = NULL;
	}

	req->r_osd = __lookup_osd(osdc, o);
	if (!req->r_osd && o >= 0) {
1127
		err = -ENOMEM;
1128
		req->r_osd = create_osd(osdc, o);
1129 1130
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1131
			goto out;
1132
		}
S
Sage Weil 已提交
1133

1134
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1135 1136
		__insert_osd(osdc, req->r_osd);

1137 1138 1139
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1140 1141
	}

1142 1143
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
1144 1145
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1146
	} else {
1147
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1148
	}
1149
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1150 1151 1152 1153 1154 1155 1156 1157

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1158 1159
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1160
{
1161
	void *p;
S
Sage Weil 已提交
1162

1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
	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);
	put_unaligned_le64(req->r_pgid.pool, req->r_request_pool);
	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));
1177

S
Sage Weil 已提交
1178
	req->r_stamp = jiffies;
1179
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1180 1181 1182 1183 1184 1185

	ceph_msg_get(req->r_request); /* send consumes a ref */
	ceph_con_send(&req->r_osd->o_con, req->r_request);
	req->r_sent = req->r_osd->o_incarnation;
}

1186 1187 1188
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1189
static void __send_queued(struct ceph_osd_client *osdc)
1190 1191 1192
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1193 1194
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1195 1196 1197
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
/*
 * 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);
S
Sage Weil 已提交
1211
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1212
	struct ceph_osd *osd;
1213
	unsigned long keepalive =
1214
		osdc->client->options->osd_keepalive_timeout * HZ;
1215
	struct list_head slow_osds;
S
Sage Weil 已提交
1216 1217 1218 1219 1220 1221 1222
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1223 1224 1225 1226 1227 1228 1229
	/*
	 * 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) {
S
Sage Weil 已提交
1230
		if (time_before(jiffies, req->r_stamp + keepalive))
1231 1232 1233 1234 1235
			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
1236
		     req->r_tid, osd->o_osd);
1237 1238 1239 1240 1241 1242
		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
1243 1244 1245
		ceph_con_keepalive(&osd->o_con);
	}

1246
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1247
	__send_queued(osdc);
S
Sage Weil 已提交
1248 1249 1250 1251
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1252 1253 1254 1255 1256 1257
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);
	unsigned long delay =
1258
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1259 1260 1261

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1262
	remove_old_osds(osdc);
1263 1264 1265 1266 1267 1268
	up_read(&osdc->map_sem);

	schedule_delayed_work(&osdc->osds_timeout_work,
			      round_jiffies_relative(delay));
}

S
Sage Weil 已提交
1269 1270 1271 1272 1273 1274 1275
static void complete_request(struct ceph_osd_request *req)
{
	if (req->r_safe_callback)
		req->r_safe_callback(req, NULL);
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1276 1277 1278 1279
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1280 1281
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1282
{
1283
	void *p, *end;
S
Sage Weil 已提交
1284 1285
	struct ceph_osd_request *req;
	u64 tid;
1286 1287
	int object_len;
	int numops, payload_len, flags;
1288
	s32 result;
1289 1290 1291 1292 1293 1294
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1295
	int already_completed;
1296
	u32 bytes;
1297
	int i;
S
Sage Weil 已提交
1298

1299
	tid = le64_to_cpu(msg->hdr.tid);
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
	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;

1310
	err = ceph_decode_pgid(&p, end, &pg);
1311
	if (err)
S
Sage Weil 已提交
1312
		goto bad;
1313 1314 1315 1316 1317 1318 1319 1320

	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
1321 1322 1323 1324 1325
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1326
		goto bad_mutex;
S
Sage Weil 已提交
1327 1328
	}
	ceph_osdc_get_request(req);
1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350

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

	ceph_decode_need(&p, end, 4, bad);
	numops = ceph_decode_32(&p);
	if (numops > CEPH_OSD_MAX_OP)
		goto bad_put;
	if (numops != req->r_num_ops)
		goto bad_put;
	payload_len = 0;
	ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad);
	for (i = 0; i < numops; i++) {
		struct ceph_osd_op *op = p;
		int len;

		len = le32_to_cpu(op->payload_len);
		req->r_reply_op_len[i] = len;
		dout(" op %d has %d bytes\n", i, len);
		payload_len += len;
		p += sizeof(*op);
	}
1351 1352
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1353
		pr_warning("sum of op payload lens %d != data_len %d",
1354
			   payload_len, bytes);
1355 1356 1357 1358 1359 1360 1361
		goto bad_put;
	}

	ceph_decode_need(&p, end, 4 + numops * 4, bad);
	retry_attempt = ceph_decode_32(&p);
	for (i = 0; i < numops; i++)
		req->r_reply_op_result[i] = ceph_decode_32(&p);
S
Sage Weil 已提交
1362 1363 1364

	if (!req->r_got_reply) {

1365
		req->r_result = result;
S
Sage Weil 已提交
1366 1367 1368 1369 1370 1371
		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, */
1372 1373
		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
1374 1375 1376 1377

		req->r_got_reply = 1;
	} else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
		dout("handle_reply tid %llu dup ack\n", tid);
S
Sage Weil 已提交
1378
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1379 1380 1381 1382 1383
		goto done;
	}

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

1384 1385 1386
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1387
	/* either this is a read, or we got the safe response */
1388 1389
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1390 1391 1392
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1393 1394
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1395
	mutex_unlock(&osdc->request_mutex);
1396 1397
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1398 1399 1400 1401

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1402
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1403

S
Sage Weil 已提交
1404 1405
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1406 1407

done:
1408
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1409 1410 1411
	ceph_osdc_put_request(req);
	return;

1412 1413
bad_put:
	ceph_osdc_put_request(req);
1414 1415
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1416
bad:
1417 1418
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1419
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1420 1421
}

1422
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1423 1424 1425
{
	struct rb_node *p, *n;

1426 1427
	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
1428

1429 1430 1431 1432 1433 1434 1435
		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
1436
	}
1437 1438 1439
}

/*
1440 1441
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1442
 *
1443
 * Caller should hold map_sem for read.
1444
 */
1445
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1446
{
1447
	struct ceph_osd_request *req, *nreq;
1448 1449 1450
	struct rb_node *p;
	int needmap = 0;
	int err;
1451

1452
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1453
	mutex_lock(&osdc->request_mutex);
1454
	for (p = rb_first(&osdc->requests); p; ) {
1455
		req = rb_entry(p, struct ceph_osd_request, r_node);
1456
		p = rb_next(p);
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474

		/*
		 * 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);
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
			continue;
		}

1475
		err = __map_request(osdc, req, force_resend);
1476 1477 1478 1479 1480 1481
		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) {
1482 1483 1484 1485
			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);
1486
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1487 1488
			}
		}
1489 1490 1491 1492 1493 1494
	}

	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);

1495
		err = __map_request(osdc, req, force_resend);
1496
		dout("__map_request returned %d\n", err);
1497 1498 1499 1500 1501 1502 1503 1504
		if (err == 0)
			continue;  /* no change and no osd was specified */
		if (err < 0)
			continue;  /* hrm! */
		if (req->r_osd == NULL) {
			dout("tid %llu maps to no valid osd\n", req->r_tid);
			needmap++;  /* request a newer map */
			continue;
1505
		}
1506 1507 1508 1509

		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);
1510
		__unregister_linger_request(osdc, req);
1511
	}
S
Sage Weil 已提交
1512 1513 1514 1515 1516 1517
	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);
	}
1518
	reset_changed_osds(osdc);
1519
}
1520 1521


S
Sage Weil 已提交
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544
/*
 * 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;

	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));
1545 1546
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1547 1548 1549 1550 1551 1552 1553 1554

	down_write(&osdc->map_sem);

	/* 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);
1555 1556
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1557 1558 1559 1560 1561 1562 1563
		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,
							  osdc->osdmap,
1564
							  &osdc->client->msgr);
S
Sage Weil 已提交
1565 1566 1567 1568
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1569
			BUG_ON(!newmap);
S
Sage Weil 已提交
1570 1571 1572 1573
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1574
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
		} 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);
1590 1591
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1592 1593 1594 1595 1596 1597 1598 1599 1600
		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 {
1601 1602
			int skipped_map = 0;

S
Sage Weil 已提交
1603 1604 1605 1606 1607 1608
			dout("taking full map %u len %d\n", epoch, maplen);
			newmap = osdmap_decode(&p, p+maplen);
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1609
			BUG_ON(!newmap);
S
Sage Weil 已提交
1610 1611
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1612 1613 1614
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1615
				ceph_osdmap_destroy(oldmap);
1616 1617
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1618 1619 1620 1621 1622 1623 1624 1625
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1626 1627 1628 1629 1630 1631 1632 1633 1634

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

A
Alex Elder 已提交
1635 1636 1637
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1638
	up_read(&osdc->map_sem);
1639
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1640 1641 1642 1643
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1644
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1645 1646 1647 1648
	up_write(&osdc->map_sem);
	return;
}

1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 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
/*
 * 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 *),
1731
			   void *data, struct ceph_osd_event **pevent)
1732 1733 1734 1735 1736 1737 1738 1739 1740
{
	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;
1741
	event->one_shot = 0;
1742 1743 1744
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1745
	RB_CLEAR_NODE(&event->node);
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
	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
 */
1792 1793
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
{
	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) {
1814
		BUG_ON(event->one_shot);
1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
		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) {
			dout("ERROR: could not allocate event_work\n");
			goto done_err;
		}
1826
		INIT_WORK(&event_work->work, do_event_work);
1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847
		event_work->event = event;
		event_work->ver = ver;
		event_work->notify_id = notify_id;
		event_work->opcode = opcode;
		if (!queue_work(osdc->notify_wq, &event_work->work)) {
			dout("WARNING: failed to queue notify event work\n");
			goto done_err;
		}
	}

	return;

done_err:
	ceph_osdc_put_event(event);
	return;

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

1848 1849
static void ceph_osdc_msg_data_set(struct ceph_msg *msg,
				struct ceph_osd_data *osd_data)
S
Sage Weil 已提交
1850
{
1851
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
1852
		BUG_ON(osd_data->length > (u64) SIZE_MAX);
1853
		if (osd_data->length)
1854 1855
			ceph_msg_data_set_pages(msg, osd_data->pages,
				osd_data->length, osd_data->alignment);
1856 1857 1858
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!osd_data->pagelist->length);
		ceph_msg_data_set_pagelist(msg, osd_data->pagelist);
1859
#ifdef CONFIG_BLOCK
1860
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
A
Alex Elder 已提交
1861
		ceph_msg_data_set_bio(msg, osd_data->bio, osd_data->bio_length);
1862
#endif
1863
	} else {
1864
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
1865
	}
1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
}

/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
	int rc = 0;

	/* Set up response incoming data and request outgoing data fields */

	ceph_osdc_msg_data_set(req->r_reply, &req->r_data_in);
	ceph_osdc_msg_data_set(req->r_request, &req->r_data_out);
S
Sage Weil 已提交
1881 1882 1883

	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
1884
	__register_request(osdc, req);
1885 1886 1887 1888 1889 1890 1891
	WARN_ON(req->r_sent);
	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;
S
Sage Weil 已提交
1892
		}
1893
		goto out_unlock;
S
Sage Weil 已提交
1894
	}
1895 1896 1897 1898
	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 {
1899
		__send_queued(osdc);
1900 1901
	}
	rc = 0;
1902
out_unlock:
S
Sage Weil 已提交
1903 1904 1905 1906
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
1907
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920

/*
 * wait for a request to complete
 */
int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
{
	int rc;

	rc = wait_for_completion_interruptible(&req->r_completion);
	if (rc < 0) {
		mutex_lock(&osdc->request_mutex);
		__cancel_request(req);
1921
		__unregister_request(osdc, req);
S
Sage Weil 已提交
1922
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1923
		complete_request(req);
1924
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
1925 1926 1927 1928 1929 1930
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
1931
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964

/*
 * 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);
}
1965
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982

/*
 * 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);
	init_completion(&osdc->map_waiters);
	osdc->last_requested_map = 0;
	mutex_init(&osdc->request_mutex);
	osdc->last_tid = 0;
	osdc->osds = RB_ROOT;
1983
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
1984
	osdc->requests = RB_ROOT;
1985
	INIT_LIST_HEAD(&osdc->req_lru);
1986 1987
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
1988
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
1989 1990
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
1991
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1992 1993 1994
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
1995 1996

	schedule_delayed_work(&osdc->osds_timeout_work,
1997
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
1998

1999
	err = -ENOMEM;
S
Sage Weil 已提交
2000 2001 2002
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
2003
		goto out;
S
Sage Weil 已提交
2004

2005 2006
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
2007
				"osd_op");
S
Sage Weil 已提交
2008
	if (err < 0)
2009
		goto out_mempool;
2010
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2011 2012
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
2013 2014
	if (err < 0)
		goto out_msgpool;
2015 2016 2017 2018 2019 2020 2021

	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
	if (IS_ERR(osdc->notify_wq)) {
		err = PTR_ERR(osdc->notify_wq);
		osdc->notify_wq = NULL;
		goto out_msgpool;
	}
S
Sage Weil 已提交
2022
	return 0;
2023

S
Sage Weil 已提交
2024 2025
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2026 2027 2028 2029
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2030 2031 2032 2033
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2034 2035
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2036
	cancel_delayed_work_sync(&osdc->timeout_work);
2037
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
2038 2039 2040 2041
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
2042
	remove_all_osds(osdc);
S
Sage Weil 已提交
2043 2044
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2045
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
}

/*
 * 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,
2056
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2057 2058
{
	struct ceph_osd_request *req;
2059
	struct ceph_osd_data *osd_data;
2060
	struct ceph_osd_req_op op;
S
Sage Weil 已提交
2061 2062 2063 2064
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2065
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1, &op,
S
Sage Weil 已提交
2066
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2067
				    NULL, truncate_seq, truncate_size,
2068
				    false);
2069 2070
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2071 2072

	/* it may be a short read due to an object boundary */
2073 2074 2075 2076

	osd_data = &req->r_data_in;
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
2077
	osd_data->length = *plen;
2078
	osd_data->alignment = page_align;
S
Sage Weil 已提交
2079

2080 2081
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
	     off, *plen, osd_data->length, page_align);
S
Sage Weil 已提交
2082

2083 2084
	ceph_osdc_build_request(req, off, 1, &op, NULL, vino.snap, NULL);

S
Sage Weil 已提交
2085 2086 2087 2088 2089 2090 2091 2092
	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;
}
2093
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103

/*
 * 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,
2104
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2105 2106
{
	struct ceph_osd_request *req;
2107
	struct ceph_osd_data *osd_data;
2108
	struct ceph_osd_req_op op;
S
Sage Weil 已提交
2109
	int rc = 0;
2110
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2111

2112 2113
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1, &op,
S
Sage Weil 已提交
2114
				    CEPH_OSD_OP_WRITE,
2115
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2116
				    snapc, truncate_seq, truncate_size,
2117
				    true);
2118 2119
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2120 2121

	/* it may be a short write due to an object boundary */
2122 2123 2124
	osd_data = &req->r_data_out;
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
2125
	osd_data->length = len;
2126
	osd_data->alignment = page_align;
2127
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, osd_data->length);
S
Sage Weil 已提交
2128

2129 2130
	ceph_osdc_build_request(req, off, 1, &op, snapc, CEPH_NOSNAP, mtime);

2131
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2132 2133 2134 2135 2136 2137 2138 2139 2140
	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;
}
2141
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2142 2143 2144 2145 2146 2147 2148

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2149
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2150 2151 2152
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2153
		goto out;
2154
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2155 2156 2157 2158 2159 2160

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2161
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2162
		break;
2163 2164 2165
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2166 2167 2168 2169 2170

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2171
out:
S
Sage Weil 已提交
2172 2173 2174
	ceph_msg_put(msg);
}

2175
/*
2176 2177
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2178 2179
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2180 2181
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2182 2183 2184
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2185
	struct ceph_msg *m;
2186
	struct ceph_osd_request *req;
2187 2188
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2189
	u64 tid;
S
Sage Weil 已提交
2190

2191 2192 2193 2194 2195 2196
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2197 2198
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2199 2200
		goto out;
	}
S
Sage Weil 已提交
2201

2202
	if (req->r_reply->con)
2203
		dout("%s revoking msg %p from old con %p\n", __func__,
2204 2205
		     req->r_reply, req->r_reply->con);
	ceph_msg_revoke_incoming(req->r_reply);
2206

S
Sage Weil 已提交
2207 2208 2209
	if (front > req->r_reply->front.iov_len) {
		pr_warning("get_reply front %d > preallocated %d\n",
			   front, (int)req->r_reply->front.iov_len);
2210
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2211
		if (!m)
S
Sage Weil 已提交
2212 2213 2214 2215 2216 2217
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2218
	if (data_len > 0) {
2219 2220 2221 2222
		struct ceph_osd_data *osd_data = &req->r_data_in;

		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2223
				unlikely(osd_data->length < data_len)) {
2224

2225 2226 2227
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2228 2229 2230 2231 2232 2233
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2234
	}
2235
	*skip = 0;
S
Sage Weil 已提交
2236
	dout("get_reply tid %lld %p\n", tid, m);
2237 2238 2239

out:
	mutex_unlock(&osdc->request_mutex);
2240
	return m;
2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251

}

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
2252
	*skip = 0;
2253 2254
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2255
	case CEPH_MSG_WATCH_NOTIFY:
2256
		return ceph_msg_new(type, front, GFP_NOFS, false);
2257 2258 2259 2260 2261 2262 2263 2264
	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
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
}

/*
 * 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);
}

2284 2285 2286
/*
 * authentication
 */
2287 2288 2289 2290 2291
/*
 * 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,
2292
					int *proto, int force_new)
2293 2294 2295 2296
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2297
	struct ceph_auth_handshake *auth = &o->o_auth;
2298

2299
	if (force_new && auth->authorizer) {
2300
		ceph_auth_destroy_authorizer(ac, auth->authorizer);
2301 2302
		auth->authorizer = NULL;
	}
2303 2304 2305
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2306
		if (ret)
2307
			return ERR_PTR(ret);
2308 2309
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2310 2311 2312
						     auth);
		if (ret)
			return ERR_PTR(ret);
2313 2314
	}
	*proto = ac->protocol;
2315

2316
	return auth;
2317 2318 2319 2320 2321 2322 2323 2324 2325
}


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;

2326
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2327 2328
}

2329 2330 2331 2332 2333 2334
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;

2335
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2336 2337
	return ceph_monc_validate_auth(&osdc->client->monc);
}
2338

2339
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2340 2341 2342
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2343 2344
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2345
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2346
	.alloc_msg = alloc_msg,
2347
	.fault = osd_reset,
S
Sage Weil 已提交
2348
};