osd_client.c 58.4 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
static int op_has_extent(int op)
{
	return (op == CEPH_OSD_OP_READ ||
		op == CEPH_OSD_OP_WRITE);
}

S
Sage Weil 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
/*
 * 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.
 */
66
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
67
		       struct ceph_osd_req_op *op, u64 *bno)
S
Sage Weil 已提交
68
{
69 70 71
	u64 orig_len = *plen;
	u64 objoff = 0;
	u64 objlen = 0;
S
Sage Weil 已提交
72
	int r;
S
Sage Weil 已提交
73

74
	/* object extent? */
75
	r = ceph_calc_file_object_mapping(layout, off, orig_len, bno,
76
					  &objoff, &objlen);
S
Sage Weil 已提交
77 78
	if (r < 0)
		return r;
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
	if (objlen < orig_len) {
		*plen = objlen;
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

	if (op_has_extent(op->op)) {
		u32 osize = le32_to_cpu(layout->fl_object_size);
		op->extent.offset = objoff;
		op->extent.length = objlen;
		if (op->extent.truncate_size <= off - objoff) {
			op->extent.truncate_size = 0;
		} else {
			op->extent.truncate_size -= off - objoff;
			if (op->extent.truncate_size > osize)
				op->extent.truncate_size = osize;
		}
	}
	if (op->op == CEPH_OSD_OP_WRITE)
		op->payload_len = *plen;

100
	dout("calc_layout bno=%llx %llu~%llu\n", *bno, objoff, objlen);
S
Sage Weil 已提交
101

102
	return 0;
S
Sage Weil 已提交
103 104 105 106 107
}

/*
 * requests
 */
S
Sage Weil 已提交
108
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
109
{
110
	int num_pages;
S
Sage Weil 已提交
111 112 113 114 115 116
	struct ceph_osd_request *req = container_of(kref,
						    struct ceph_osd_request,
						    r_kref);

	if (req->r_request)
		ceph_msg_put(req->r_request);
117
	if (req->r_con_filling_msg) {
118 119
		dout("%s revoking msg %p from con %p\n", __func__,
		     req->r_reply, req->r_con_filling_msg);
120
		ceph_msg_revoke_incoming(req->r_reply);
121
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
122
		req->r_con_filling_msg = NULL;
123
	}
124 125
	if (req->r_reply)
		ceph_msg_put(req->r_reply);
126 127

	if (req->r_data_in.type == CEPH_OSD_DATA_TYPE_PAGES &&
128 129 130 131 132
			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);
	}
133
	if (req->r_data_out.type == CEPH_OSD_DATA_TYPE_PAGES &&
134 135 136 137 138
			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);
	}
139

S
Sage Weil 已提交
140 141 142 143 144
	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
145
}
146
EXPORT_SYMBOL(ceph_osdc_release_request);
147

148
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
149
					       struct ceph_snap_context *snapc,
150
					       unsigned int num_ops,
151
					       bool use_mempool,
152
					       gfp_t gfp_flags)
S
Sage Weil 已提交
153 154 155
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
156 157 158 159 160 161 162 163 164 165 166
	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
167 168

	if (use_mempool) {
169
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
170 171
		memset(req, 0, sizeof(*req));
	} else {
172
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
173 174
	}
	if (req == NULL)
175
		return NULL;
S
Sage Weil 已提交
176 177 178

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
179

S
Sage Weil 已提交
180
	kref_init(&req->r_kref);
S
Sage Weil 已提交
181 182
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
183
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
184
	INIT_LIST_HEAD(&req->r_unsafe_item);
185 186
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
187
	INIT_LIST_HEAD(&req->r_req_lru_item);
188 189
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
190 191 192 193 194
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
195
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
196
	if (!msg) {
S
Sage Weil 已提交
197
		ceph_osdc_put_request(req);
198
		return NULL;
S
Sage Weil 已提交
199 200 201
	}
	req->r_reply = msg;

202 203
	req->r_data_in.type = CEPH_OSD_DATA_TYPE_NONE;
	req->r_data_out.type = CEPH_OSD_DATA_TYPE_NONE;
204

S
Sage Weil 已提交
205
	/* create request message; allow space for oid */
S
Sage Weil 已提交
206
	if (use_mempool)
207
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
208
	else
209
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
210
	if (!msg) {
S
Sage Weil 已提交
211
		ceph_osdc_put_request(req);
212
		return NULL;
S
Sage Weil 已提交
213
	}
214

S
Sage Weil 已提交
215
	memset(msg->front.iov_base, 0, msg->front.iov_len);
216 217 218 219 220

	req->r_request = msg;

	return req;
}
221
EXPORT_SYMBOL(ceph_osdc_alloc_request);
222

223
static u64 osd_req_encode_op(struct ceph_osd_request *req,
224 225 226
			      struct ceph_osd_op *dst,
			      struct ceph_osd_req_op *src)
{
227
	u64 out_data_len = 0;
A
Alex Elder 已提交
228
	struct ceph_pagelist *pagelist;
229

230 231
	dst->op = cpu_to_le16(src->op);

232
	switch (src->op) {
A
Alex Elder 已提交
233 234
	case CEPH_OSD_OP_STAT:
		break;
235 236
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
237 238 239 240
		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);
241 242 243 244 245
		dst->extent.truncate_size =
			cpu_to_le64(src->extent.truncate_size);
		dst->extent.truncate_seq =
			cpu_to_le32(src->extent.truncate_seq);
		break;
246
	case CEPH_OSD_OP_CALL:
A
Alex Elder 已提交
247 248 249 250
		pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
		BUG_ON(!pagelist);
		ceph_pagelist_init(pagelist);

251 252 253
		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);
A
Alex Elder 已提交
254
		ceph_pagelist_append(pagelist, src->cls.class_name,
255
				     src->cls.class_len);
A
Alex Elder 已提交
256
		ceph_pagelist_append(pagelist, src->cls.method_name,
257
				     src->cls.method_len);
A
Alex Elder 已提交
258
		ceph_pagelist_append(pagelist, src->cls.indata,
259
				     src->cls.indata_len);
A
Alex Elder 已提交
260 261 262 263

		req->r_data_out.type = CEPH_OSD_DATA_TYPE_PAGELIST;
		req->r_data_out.pagelist = pagelist;
		out_data_len = pagelist->length;
264
		break;
265 266
	case CEPH_OSD_OP_STARTSYNC:
		break;
267 268 269 270 271 272
	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;
273
	default:
274
		pr_err("unrecognized osd opcode %d\n", src->op);
275 276
		WARN_ON(1);
		break;
277 278 279
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
280
	case CEPH_OSD_OP_NOTIFY:
281 282 283 284 285 286 287 288 289 290 291 292
	case CEPH_OSD_OP_ASSERT_VER:
	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:
	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:
293
	case CEPH_OSD_OP_ROLLBACK:
294 295 296 297 298 299 300 301 302 303 304 305 306
	case CEPH_OSD_OP_OMAPGETKEYS:
	case CEPH_OSD_OP_OMAPGETVALS:
	case CEPH_OSD_OP_OMAPGETHEADER:
	case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
	case CEPH_OSD_OP_MODE_RD:
	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:
307
	case CEPH_OSD_OP_GETXATTR:
308
	case CEPH_OSD_OP_GETXATTRS:
309 310
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
	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:
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
		pr_err("unsupported osd opcode %s\n",
332
			ceph_osd_op_name(src->op));
333 334
		WARN_ON(1);
		break;
335 336
	}
	dst->payload_len = cpu_to_le32(src->payload_len);
337 338

	return out_data_len;
339 340
}

341 342 343 344 345
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req,
346
			     u64 off, unsigned int num_ops,
347
			     struct ceph_osd_req_op *src_ops,
348
			     struct ceph_snap_context *snapc, u64 snap_id,
349
			     struct timespec *mtime)
350 351
{
	struct ceph_msg *msg = req->r_request;
352
	struct ceph_osd_req_op *src_op;
353
	void *p;
354
	size_t msg_size;
355
	int flags = req->r_flags;
356
	u64 data_len;
357
	int i;
358

359 360
	req->r_num_ops = num_ops;
	req->r_snapid = snap_id;
S
Sage Weil 已提交
361 362
	req->r_snapc = ceph_get_snap_context(snapc);

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
	/* 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);
394
	memcpy(p, req->r_oid, req->r_oid_len);
395
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
396
	p += req->r_oid_len;
S
Sage Weil 已提交
397

398
	/* ops--can imply data */
399
	ceph_encode_16(&p, num_ops);
400
	src_op = src_ops;
401
	req->r_request_ops = p;
402
	data_len = 0;
403
	for (i = 0; i < num_ops; i++, src_op++) {
404
		data_len += osd_req_encode_op(req, p, src_op);
405 406
		p += sizeof(struct ceph_osd_op);
	}
407

408 409 410 411 412
	/* 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
413
		for (i = 0; i < snapc->num_snaps; i++) {
414
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
S
Sage Weil 已提交
415 416 417
		}
	}

418 419 420
	req->r_request_attempts = p;
	p += 4;

421 422
	/* data */
	if (flags & CEPH_OSD_FLAG_WRITE)
423
		req->r_request->hdr.data_off = cpu_to_le16(off);
424
	req->r_request->hdr.data_len = cpu_to_le32(data_len);
425

S
Sage Weil 已提交
426
	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
427 428 429
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);
430 431 432

	dout("build_request msg_size was %d num_ops %d\n", (int)msg_size,
	     num_ops);
433 434
	return;
}
435
EXPORT_SYMBOL(ceph_osdc_build_request);
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457

/*
 * 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,
					       u64 off, u64 *plen,
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       int do_sync,
					       u32 truncate_seq,
					       u64 truncate_size,
					       struct timespec *mtime,
458
					       bool use_mempool)
459
{
A
Alex Elder 已提交
460
	struct ceph_osd_req_op ops[2];
461
	struct ceph_osd_request *req;
A
Alex Elder 已提交
462
	unsigned int num_op = 1;
463
	u64 bno = 0;
464
	int r;
465

A
Alex Elder 已提交
466 467
	memset(&ops, 0, sizeof ops);

468 469 470 471 472 473
	ops[0].op = opcode;
	ops[0].extent.truncate_seq = truncate_seq;
	ops[0].extent.truncate_size = truncate_size;

	if (do_sync) {
		ops[1].op = CEPH_OSD_OP_STARTSYNC;
A
Alex Elder 已提交
474 475
		num_op++;
	}
476

A
Alex Elder 已提交
477 478
	req = ceph_osdc_alloc_request(osdc, snapc, num_op, use_mempool,
					GFP_NOFS);
479
	if (!req)
480
		return ERR_PTR(-ENOMEM);
481
	req->r_flags = flags;
482 483

	/* calculate max write size */
484
	r = calc_layout(layout, off, plen, ops, &bno);
485 486
	if (r < 0) {
		ceph_osdc_put_request(req);
487
		return ERR_PTR(r);
488
	}
489 490
	req->r_file_layout = *layout;  /* keep a copy */

491 492 493
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx", vino.ino, bno);
	req->r_oid_len = strlen(req->r_oid);

494
	ceph_osdc_build_request(req, off, num_op, ops,
A
Alex Elder 已提交
495
				snapc, vino.snap, mtime);
496

S
Sage Weil 已提交
497 498
	return req;
}
499
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565

/*
 * 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;
}

566 567 568 569 570 571
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
572
	struct ceph_osd_request *req, *nreq;
573 574 575 576
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
577
	if (err)
578 579 580 581 582 583
		return;

	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
		list_move(&req->r_req_lru_item, &osdc->req_unsent);
		dout("requeued %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
584 585 586 587 588 589
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}

	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
590 591 592 593 594
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
595
		__register_request(osdc, req);
596 597 598
		list_add(&req->r_req_lru_item, &osdc->req_unsent);
		list_add(&req->r_osd_item, &req->r_osd->o_requests);
		__unregister_linger_request(osdc, req);
599 600
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
601 602 603
	}
}

S
Sage Weil 已提交
604
/*
605
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
606 607 608 609 610 611 612 613 614 615 616
 */
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
617 618
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
619
	__send_queued(osdc);
S
Sage Weil 已提交
620
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
621 622 623 624 625 626
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
627
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
628 629 630 631 632 633 634 635 636
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
637
	osd->o_osd = onum;
638
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
639
	INIT_LIST_HEAD(&osd->o_requests);
640
	INIT_LIST_HEAD(&osd->o_linger_requests);
641
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
642 643
	osd->o_incarnation = 1;

644
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
645

646
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665
	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);
666
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
667 668
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

669
		if (ac->ops && ac->ops->destroy_authorizer)
670
			ac->ops->destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
671
		kfree(osd);
S
Sage Weil 已提交
672
	}
S
Sage Weil 已提交
673 674 675 676 677
}

/*
 * remove an osd from our map
 */
678
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
679
{
680
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
681 682
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
683
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
684 685 686 687
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

688 689
static void remove_all_osds(struct ceph_osd_client *osdc)
{
690
	dout("%s %p\n", __func__, osdc);
691 692 693 694 695 696 697 698 699
	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);
}

700 701 702 703 704 705
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);
706
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
707 708 709 710 711 712 713 714 715
}

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);
}

716
static void remove_old_osds(struct ceph_osd_client *osdc)
717 718 719 720 721 722
{
	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) {
723
		if (time_before(jiffies, osd->lru_ttl))
724 725 726 727 728 729
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
730 731 732
/*
 * reset osd connect
 */
733
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
734
{
A
Alex Elder 已提交
735
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
736

737
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
738 739
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
740
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
741 742 743 744 745 746 747 748 749

		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;

750 751 752 753 754
		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
755 756

		return -EAGAIN;
S
Sage Weil 已提交
757
	}
A
Alex Elder 已提交
758 759 760 761 762 763

	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
764 765 766 767 768 769 770 771
}

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;

772
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804
	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;
}

805 806 807
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
808
			osdc->client->options->osd_keepalive_timeout * HZ);
809 810 811 812 813 814
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
815 816 817 818 819

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
820 821
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
822 823
{
	req->r_tid = ++osdc->last_tid;
824
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
825
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
826 827 828 829
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
830 831
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
832
	}
833 834 835 836 837 838 839
}

static void register_request(struct ceph_osd_client *osdc,
			     struct ceph_osd_request *req)
{
	mutex_lock(&osdc->request_mutex);
	__register_request(osdc, req);
S
Sage Weil 已提交
840 841 842 843 844 845 846 847 848
	mutex_unlock(&osdc->request_mutex);
}

/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
849 850 851 852 853 854
	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
855 856 857 858
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

859 860
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
861
		ceph_msg_revoke(req->r_request);
862 863

		list_del_init(&req->r_osd_item);
864 865 866
		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);
867
			__move_osd_to_lru(osdc, req->r_osd);
868
		}
869
		if (list_empty(&req->r_linger_item))
870
			req->r_osd = NULL;
871
	}
S
Sage Weil 已提交
872

A
Alex Elder 已提交
873
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
874 875
	ceph_osdc_put_request(req);

876 877 878
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
879 880 881 882 883 884 885 886
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
887
	if (req->r_sent && req->r_osd) {
888
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
889 890 891 892
		req->r_sent = 0;
	}
}

893 894 895 896 897
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);
898 899 900
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
901 902 903 904 905 906
}

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
907
	list_del_init(&req->r_linger_item);
908 909 910 911 912 913 914 915
	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);
		}
916 917
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947
	}
}

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
948 949 950
/*
 * 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
951
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
952
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
953 954 955 956 957
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
958
static int __map_request(struct ceph_osd_client *osdc,
959
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
960
{
961
	struct ceph_pg pgid;
962 963
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
964 965
	int err;

966
	dout("map_request %p tid %lld\n", req, req->r_tid);
967 968
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
969 970
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
971
		return err;
972
	}
973 974
	req->r_pgid = pgid;

975 976 977 978 979
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
980

981 982
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
983 984 985
	     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
986 987 988
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

989 990
	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
991 992
	     req->r_osd ? req->r_osd->o_osd : -1);

993 994 995 996
	/* 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
997 998 999 1000 1001 1002 1003 1004
	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) {
1005
		err = -ENOMEM;
1006
		req->r_osd = create_osd(osdc, o);
1007 1008
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1009
			goto out;
1010
		}
S
Sage Weil 已提交
1011

1012
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1013 1014
		__insert_osd(osdc, req->r_osd);

1015 1016 1017
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1018 1019
	}

1020 1021
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
S
Sage Weil 已提交
1022
		list_add(&req->r_osd_item, &req->r_osd->o_requests);
1023 1024 1025
		list_move(&req->r_req_lru_item, &osdc->req_unsent);
	} else {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
1026
	}
1027
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1028 1029 1030 1031 1032 1033 1034 1035

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1036 1037
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1038
{
1039
	void *p;
S
Sage Weil 已提交
1040

1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
	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));
1055

S
Sage Weil 已提交
1056
	req->r_stamp = jiffies;
1057
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1058 1059 1060 1061 1062 1063

	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;
}

1064 1065 1066
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1067
static void __send_queued(struct ceph_osd_client *osdc)
1068 1069 1070
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1071 1072
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1073 1074 1075
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
/*
 * 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
1089
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1090
	struct ceph_osd *osd;
1091
	unsigned long keepalive =
1092
		osdc->client->options->osd_keepalive_timeout * HZ;
1093
	struct list_head slow_osds;
S
Sage Weil 已提交
1094 1095 1096 1097 1098 1099 1100
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1101 1102 1103 1104 1105 1106 1107
	/*
	 * 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
1108
		if (time_before(jiffies, req->r_stamp + keepalive))
1109 1110 1111 1112 1113
			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
1114
		     req->r_tid, osd->o_osd);
1115 1116 1117 1118 1119 1120
		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
1121 1122 1123
		ceph_con_keepalive(&osd->o_con);
	}

1124
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1125
	__send_queued(osdc);
S
Sage Weil 已提交
1126 1127 1128 1129
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1130 1131 1132 1133 1134 1135
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 =
1136
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1137 1138 1139

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1140
	remove_old_osds(osdc);
1141 1142 1143 1144 1145 1146
	up_read(&osdc->map_sem);

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

S
Sage Weil 已提交
1147 1148 1149 1150 1151 1152 1153
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 */
}

1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173
static int __decode_pgid(void **p, void *end, struct ceph_pg *pgid)
{
	__u8 v;

	ceph_decode_need(p, end, 1 + 8 + 4 + 4, bad);
	v = ceph_decode_8(p);
	if (v > 1) {
		pr_warning("do not understand pg encoding %d > 1", v);
		return -EINVAL;
	}
	pgid->pool = ceph_decode_64(p);
	pgid->seed = ceph_decode_32(p);
	*p += 4;
	return 0;

bad:
	pr_warning("incomplete pg encoding");
	return -EINVAL;
}

S
Sage Weil 已提交
1174 1175 1176 1177
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1178 1179
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1180
{
1181
	void *p, *end;
S
Sage Weil 已提交
1182 1183
	struct ceph_osd_request *req;
	u64 tid;
1184 1185
	int object_len;
	int numops, payload_len, flags;
1186
	s32 result;
1187 1188 1189 1190 1191 1192
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1193
	int already_completed;
1194
	int i;
S
Sage Weil 已提交
1195

1196
	tid = le64_to_cpu(msg->hdr.tid);
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
	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;

	err = __decode_pgid(&p, end, &pg);
	if (err)
S
Sage Weil 已提交
1209
		goto bad;
1210 1211 1212 1213 1214 1215 1216 1217

	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
1218 1219 1220 1221 1222 1223 1224 1225 1226
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
		mutex_unlock(&osdc->request_mutex);
		return;
	}
	ceph_osdc_get_request(req);
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258

	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);
	}
	if (payload_len != le32_to_cpu(msg->hdr.data_len)) {
		pr_warning("sum of op payload lens %d != data_len %d",
			   payload_len, le32_to_cpu(msg->hdr.data_len));
		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
1259

1260
	/*
1261
	 * if this connection filled our message, drop our reference now, to
1262 1263
	 * avoid a (safe but slower) revoke later.
	 */
1264
	if (req->r_con_filling_msg == con && req->r_reply == msg) {
S
Sage Weil 已提交
1265
		dout(" dropping con_filling_msg ref %p\n", con);
1266
		req->r_con_filling_msg = NULL;
1267
		con->ops->put(con);
1268 1269
	}

S
Sage Weil 已提交
1270
	if (!req->r_got_reply) {
1271
		unsigned int bytes;
S
Sage Weil 已提交
1272

1273
		req->r_result = result;
S
Sage Weil 已提交
1274 1275 1276 1277 1278 1279 1280
		bytes = le32_to_cpu(msg->hdr.data_len);
		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, */
1281 1282
		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
1283 1284 1285 1286

		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
1287
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1288 1289 1290 1291 1292
		goto done;
	}

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

1293 1294 1295
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1296
	/* either this is a read, or we got the safe response */
1297 1298
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1299 1300 1301
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1302 1303
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1304
	mutex_unlock(&osdc->request_mutex);
1305 1306
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1307 1308 1309 1310

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1311
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1312

S
Sage Weil 已提交
1313 1314
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1315 1316

done:
1317
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1318 1319 1320
	ceph_osdc_put_request(req);
	return;

1321 1322
bad_put:
	ceph_osdc_put_request(req);
S
Sage Weil 已提交
1323
bad:
1324 1325
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1326
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1327 1328
}

1329
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1330 1331 1332
{
	struct rb_node *p, *n;

1333 1334
	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
1335

1336 1337 1338 1339 1340 1341 1342
		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
1343
	}
1344 1345 1346
}

/*
1347 1348
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1349
 *
1350
 * Caller should hold map_sem for read.
1351
 */
1352
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1353
{
1354
	struct ceph_osd_request *req, *nreq;
1355 1356 1357
	struct rb_node *p;
	int needmap = 0;
	int err;
1358

1359
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1360
	mutex_lock(&osdc->request_mutex);
1361
	for (p = rb_first(&osdc->requests); p; ) {
1362
		req = rb_entry(p, struct ceph_osd_request, r_node);
1363
		p = rb_next(p);
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381

		/*
		 * 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;
		}

1382
		err = __map_request(osdc, req, force_resend);
1383 1384 1385 1386 1387 1388
		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) {
1389 1390 1391 1392
			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);
1393
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1394 1395
			}
		}
1396 1397 1398 1399 1400 1401
	}

	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);

1402
		err = __map_request(osdc, req, force_resend);
1403
		dout("__map_request returned %d\n", err);
1404 1405 1406 1407 1408 1409 1410 1411
		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;
1412
		}
1413 1414 1415 1416

		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);
1417
		__unregister_linger_request(osdc, req);
1418
	}
S
Sage Weil 已提交
1419 1420 1421 1422 1423 1424
	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);
	}
1425
	reset_changed_osds(osdc);
1426
}
1427 1428


S
Sage Weil 已提交
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
/*
 * 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));
1452 1453
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1454 1455 1456 1457 1458 1459 1460 1461

	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);
1462 1463
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1464 1465 1466 1467 1468 1469 1470
		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,
1471
							  &osdc->client->msgr);
S
Sage Weil 已提交
1472 1473 1474 1475
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1476
			BUG_ON(!newmap);
S
Sage Weil 已提交
1477 1478 1479 1480
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1481
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
		} 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);
1497 1498
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1499 1500 1501 1502 1503 1504 1505 1506 1507
		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 {
1508 1509
			int skipped_map = 0;

S
Sage Weil 已提交
1510 1511 1512 1513 1514 1515
			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;
			}
1516
			BUG_ON(!newmap);
S
Sage Weil 已提交
1517 1518
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1519 1520 1521
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1522
				ceph_osdmap_destroy(oldmap);
1523 1524
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1525 1526 1527 1528 1529 1530 1531 1532
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1533 1534 1535 1536 1537 1538 1539 1540 1541

	/*
	 * 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
1542 1543 1544
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1545
	up_read(&osdc->map_sem);
1546
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1547 1548 1549 1550
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1551
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1552 1553 1554 1555
	up_write(&osdc->map_sem);
	return;
}

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 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637
/*
 * 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 *),
1638
			   void *data, struct ceph_osd_event **pevent)
1639 1640 1641 1642 1643 1644 1645 1646 1647
{
	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;
1648
	event->one_shot = 0;
1649 1650 1651
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1652
	RB_CLEAR_NODE(&event->node);
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
	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
 */
1699 1700
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
{
	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) {
1721
		BUG_ON(event->one_shot);
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732
		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;
		}
1733
		INIT_WORK(&event_work->work, do_event_work);
1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
		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;
}

1755 1756
static void ceph_osdc_msg_data_set(struct ceph_msg *msg,
				struct ceph_osd_data *osd_data)
S
Sage Weil 已提交
1757
{
1758
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
1759
		BUG_ON(osd_data->length > (u64) SIZE_MAX);
1760
		if (osd_data->length)
1761 1762
			ceph_msg_data_set_pages(msg, osd_data->pages,
				osd_data->length, osd_data->alignment);
1763 1764 1765
	} 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);
1766
#ifdef CONFIG_BLOCK
1767
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
1768
		ceph_msg_data_set_bio(msg, osd_data->bio);
1769
#endif
1770
	} else {
1771
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
1772
	}
1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787
}

/*
 * 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
1788 1789 1790 1791 1792

	register_request(osdc, req);

	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
S
Sage Weil 已提交
1793 1794 1795 1796 1797 1798
	/*
	 * a racing kick_requests() may have sent the message for us
	 * while we dropped request_mutex above, so only send now if
	 * the request still han't been touched yet.
	 */
	if (req->r_sent == 0) {
1799
		rc = __map_request(osdc, req, 0);
1800 1801 1802 1803 1804 1805
		if (rc < 0) {
			if (nofail) {
				dout("osdc_start_request failed map, "
				     " will retry %lld\n", req->r_tid);
				rc = 0;
			}
1806
			goto out_unlock;
1807
		}
1808 1809 1810 1811
		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 {
1812
			__send_request(osdc, req);
S
Sage Weil 已提交
1813
		}
1814
		rc = 0;
S
Sage Weil 已提交
1815
	}
1816 1817

out_unlock:
S
Sage Weil 已提交
1818 1819 1820 1821
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
1822
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835

/*
 * 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);
1836
		__unregister_request(osdc, req);
S
Sage Weil 已提交
1837
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1838
		complete_request(req);
1839
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
1840 1841 1842 1843 1844 1845
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
1846
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879

/*
 * 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);
}
1880
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897

/*
 * 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;
1898
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
1899
	osdc->requests = RB_ROOT;
1900
	INIT_LIST_HEAD(&osdc->req_lru);
1901 1902
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
1903
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
1904 1905
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
1906
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1907 1908 1909
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
1910 1911

	schedule_delayed_work(&osdc->osds_timeout_work,
1912
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
1913

1914
	err = -ENOMEM;
S
Sage Weil 已提交
1915 1916 1917
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
1918
		goto out;
S
Sage Weil 已提交
1919

1920 1921
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
1922
				"osd_op");
S
Sage Weil 已提交
1923
	if (err < 0)
1924
		goto out_mempool;
1925
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
1926 1927
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
1928 1929
	if (err < 0)
		goto out_msgpool;
1930 1931 1932 1933 1934 1935 1936

	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
1937
	return 0;
1938

S
Sage Weil 已提交
1939 1940
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
1941 1942 1943 1944
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
1945 1946 1947 1948
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
1949 1950
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
1951
	cancel_delayed_work_sync(&osdc->timeout_work);
1952
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
1953 1954 1955 1956
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
1957
	remove_all_osds(osdc);
S
Sage Weil 已提交
1958 1959
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
1960
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970
}

/*
 * 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,
1971
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
1972 1973
{
	struct ceph_osd_request *req;
1974
	struct ceph_osd_data *osd_data;
S
Sage Weil 已提交
1975 1976 1977 1978 1979 1980 1981
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
				    NULL, 0, truncate_seq, truncate_size, NULL,
1982
				    false);
1983 1984
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
1985 1986

	/* it may be a short read due to an object boundary */
1987 1988 1989 1990

	osd_data = &req->r_data_in;
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
1991
	osd_data->length = *plen;
1992
	osd_data->alignment = page_align;
S
Sage Weil 已提交
1993

1994 1995
	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
1996 1997 1998 1999 2000 2001 2002 2003 2004

	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;
}
2005
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

/*
 * 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,
2016
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2017 2018
{
	struct ceph_osd_request *req;
2019
	struct ceph_osd_data *osd_data;
S
Sage Weil 已提交
2020
	int rc = 0;
2021
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2022 2023 2024 2025

	BUG_ON(vino.snap != CEPH_NOSNAP);
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
				    CEPH_OSD_OP_WRITE,
2026
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2027
				    snapc, 0,
S
Sage Weil 已提交
2028
				    truncate_seq, truncate_size, mtime,
2029
				    true);
2030 2031
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2032 2033

	/* it may be a short write due to an object boundary */
2034 2035 2036
	osd_data = &req->r_data_out;
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
2037
	osd_data->length = len;
2038
	osd_data->alignment = page_align;
2039
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, osd_data->length);
S
Sage Weil 已提交
2040

2041
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2042 2043 2044 2045 2046 2047 2048 2049 2050
	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;
}
2051
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2052 2053 2054 2055 2056 2057 2058

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2059
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2060 2061 2062
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2063
		goto out;
2064
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2065 2066 2067 2068 2069 2070

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2071
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2072
		break;
2073 2074 2075
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2076 2077 2078 2079 2080

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2081
out:
S
Sage Weil 已提交
2082 2083 2084
	ceph_msg_put(msg);
}

2085
/*
2086 2087
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2088 2089
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2090 2091
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2092 2093 2094
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2095
	struct ceph_msg *m;
2096
	struct ceph_osd_request *req;
2097 2098
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2099
	u64 tid;
S
Sage Weil 已提交
2100

2101 2102 2103 2104 2105 2106
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2107 2108
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2109 2110
		goto out;
	}
S
Sage Weil 已提交
2111 2112

	if (req->r_con_filling_msg) {
2113
		dout("%s revoking msg %p from old con %p\n", __func__,
S
Sage Weil 已提交
2114
		     req->r_reply, req->r_con_filling_msg);
2115
		ceph_msg_revoke_incoming(req->r_reply);
2116
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
2117
		req->r_con_filling_msg = NULL;
2118 2119
	}

S
Sage Weil 已提交
2120 2121 2122
	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);
2123
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2124
		if (!m)
S
Sage Weil 已提交
2125 2126 2127 2128 2129 2130
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2131
	if (data_len > 0) {
2132 2133 2134 2135
		struct ceph_osd_data *osd_data = &req->r_data_in;

		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2136
				unlikely(osd_data->length < data_len)) {
2137

2138 2139 2140
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2141 2142 2143 2144 2145 2146
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2147
	}
2148
	*skip = 0;
2149
	req->r_con_filling_msg = con->ops->get(con);
S
Sage Weil 已提交
2150
	dout("get_reply tid %lld %p\n", tid, m);
2151 2152 2153

out:
	mutex_unlock(&osdc->request_mutex);
2154
	return m;
2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165

}

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
2166
	*skip = 0;
2167 2168
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2169
	case CEPH_MSG_WATCH_NOTIFY:
2170
		return ceph_msg_new(type, front, GFP_NOFS, false);
2171 2172 2173 2174 2175 2176 2177 2178
	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
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197
}

/*
 * 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);
}

2198 2199 2200
/*
 * authentication
 */
2201 2202 2203 2204 2205
/*
 * 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,
2206
					int *proto, int force_new)
2207 2208 2209 2210
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2211
	struct ceph_auth_handshake *auth = &o->o_auth;
2212

2213
	if (force_new && auth->authorizer) {
2214 2215
		if (ac->ops && ac->ops->destroy_authorizer)
			ac->ops->destroy_authorizer(ac, auth->authorizer);
2216 2217
		auth->authorizer = NULL;
	}
2218
	if (!auth->authorizer && ac->ops && ac->ops->create_authorizer) {
2219 2220
		int ret = ac->ops->create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
							auth);
2221
		if (ret)
2222
			return ERR_PTR(ret);
2223 2224
	}
	*proto = ac->protocol;
2225

2226
	return auth;
2227 2228 2229 2230 2231 2232 2233 2234 2235
}


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;

2236 2237 2238 2239
	/*
	 * XXX If ac->ops or ac->ops->verify_authorizer_reply is null,
	 * XXX which do we do:  succeed or fail?
	 */
2240
	return ac->ops->verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2241 2242
}

2243 2244 2245 2246 2247 2248
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;

2249
	if (ac->ops && ac->ops->invalidate_authorizer)
2250 2251 2252 2253
		ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);

	return ceph_monc_validate_auth(&osdc->client->monc);
}
2254

2255
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2256 2257 2258
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2259 2260
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2261
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2262
	.alloc_msg = alloc_msg,
2263
	.fault = osd_reset,
S
Sage Weil 已提交
2264
};