osd_client.c 57.1 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
{
S
Sage Weil 已提交
110 111 112 113 114 115
	struct ceph_osd_request *req = container_of(kref,
						    struct ceph_osd_request,
						    r_kref);

	if (req->r_request)
		ceph_msg_put(req->r_request);
116
	if (req->r_con_filling_msg) {
117 118
		dout("%s revoking msg %p from con %p\n", __func__,
		     req->r_reply, req->r_con_filling_msg);
119
		ceph_msg_revoke_incoming(req->r_reply);
120
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
121
		req->r_con_filling_msg = NULL;
122
	}
123 124
	if (req->r_reply)
		ceph_msg_put(req->r_reply);
125 126 127
	if (req->r_data.own_pages)
		ceph_release_page_vector(req->r_data.pages,
					 req->r_data.num_pages);
S
Sage Weil 已提交
128
	ceph_put_snap_context(req->r_snapc);
129
	ceph_pagelist_release(&req->r_trail);
S
Sage Weil 已提交
130 131 132 133
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
		kfree(req);
S
Sage Weil 已提交
134
}
135
EXPORT_SYMBOL(ceph_osdc_release_request);
136

137
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
138
					       struct ceph_snap_context *snapc,
139
					       unsigned int num_ops,
140
					       bool use_mempool,
141
					       gfp_t gfp_flags)
S
Sage Weil 已提交
142 143 144
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
145 146 147 148 149 150 151 152 153 154 155
	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
156 157

	if (use_mempool) {
158
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
159 160
		memset(req, 0, sizeof(*req));
	} else {
161
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
162 163
	}
	if (req == NULL)
164
		return NULL;
S
Sage Weil 已提交
165 166 167

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
168

S
Sage Weil 已提交
169
	kref_init(&req->r_kref);
S
Sage Weil 已提交
170 171
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
172
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
173
	INIT_LIST_HEAD(&req->r_unsafe_item);
174 175
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
176
	INIT_LIST_HEAD(&req->r_req_lru_item);
177 178
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
179 180 181 182 183
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
184
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
185
	if (!msg) {
S
Sage Weil 已提交
186
		ceph_osdc_put_request(req);
187
		return NULL;
S
Sage Weil 已提交
188 189 190
	}
	req->r_reply = msg;

191
	ceph_pagelist_init(&req->r_trail);
192

S
Sage Weil 已提交
193
	/* create request message; allow space for oid */
S
Sage Weil 已提交
194
	if (use_mempool)
195
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
196
	else
197
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
198
	if (!msg) {
S
Sage Weil 已提交
199
		ceph_osdc_put_request(req);
200
		return NULL;
S
Sage Weil 已提交
201
	}
202

S
Sage Weil 已提交
203
	memset(msg->front.iov_base, 0, msg->front.iov_len);
204 205 206 207 208

	req->r_request = msg;

	return req;
}
209
EXPORT_SYMBOL(ceph_osdc_alloc_request);
210

211 212 213 214 215 216
static void osd_req_encode_op(struct ceph_osd_request *req,
			      struct ceph_osd_op *dst,
			      struct ceph_osd_req_op *src)
{
	dst->op = cpu_to_le16(src->op);

217
	switch (src->op) {
A
Alex Elder 已提交
218 219
	case CEPH_OSD_OP_STAT:
		break;
220 221 222 223 224 225 226 227 228 229 230
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		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;
231 232 233 234 235
	case CEPH_OSD_OP_CALL:
		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);

236
		ceph_pagelist_append(&req->r_trail, src->cls.class_name,
237
				     src->cls.class_len);
238
		ceph_pagelist_append(&req->r_trail, src->cls.method_name,
239
				     src->cls.method_len);
240
		ceph_pagelist_append(&req->r_trail, src->cls.indata,
241 242
				     src->cls.indata_len);
		break;
243 244
	case CEPH_OSD_OP_STARTSYNC:
		break;
245 246 247 248 249 250
	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;
251
	default:
252
		pr_err("unrecognized osd opcode %d\n", src->op);
253 254
		WARN_ON(1);
		break;
255 256 257
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
258
	case CEPH_OSD_OP_NOTIFY:
259 260 261 262 263 264 265 266 267 268 269 270
	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:
271
	case CEPH_OSD_OP_ROLLBACK:
272 273 274 275 276 277 278 279 280 281 282 283 284
	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:
285
	case CEPH_OSD_OP_GETXATTR:
286
	case CEPH_OSD_OP_GETXATTRS:
287 288
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
	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",
310
			ceph_osd_op_name(src->op));
311 312
		WARN_ON(1);
		break;
313 314 315 316
	}
	dst->payload_len = cpu_to_le32(src->payload_len);
}

317 318 319 320 321
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req,
322
			     u64 off, u64 len, unsigned int num_ops,
323
			     struct ceph_osd_req_op *src_ops,
324
			     struct ceph_snap_context *snapc, u64 snap_id,
325
			     struct timespec *mtime)
326 327
{
	struct ceph_msg *msg = req->r_request;
328
	struct ceph_osd_req_op *src_op;
329
	void *p;
330
	size_t msg_size;
331
	int flags = req->r_flags;
332
	u64 data_len;
333
	int i;
334

335 336
	req->r_num_ops = num_ops;
	req->r_snapid = snap_id;
S
Sage Weil 已提交
337 338
	req->r_snapc = ceph_get_snap_context(snapc);

339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
	/* 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);
370
	memcpy(p, req->r_oid, req->r_oid_len);
371
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
372
	p += req->r_oid_len;
S
Sage Weil 已提交
373

374 375
	/* ops */
	ceph_encode_16(&p, num_ops);
376
	src_op = src_ops;
377 378 379 380 381
	req->r_request_ops = p;
	for (i = 0; i < num_ops; i++, src_op++) {
		osd_req_encode_op(req, p, src_op);
		p += sizeof(struct ceph_osd_op);
	}
382

383 384 385 386 387
	/* 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
388
		for (i = 0; i < snapc->num_snaps; i++) {
389
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
S
Sage Weil 已提交
390 391 392
		}
	}

393 394 395
	req->r_request_attempts = p;
	p += 4;

396
	data_len = req->r_trail.length;
397 398
	if (flags & CEPH_OSD_FLAG_WRITE) {
		req->r_request->hdr.data_off = cpu_to_le16(off);
399
		data_len += len;
400
	}
401
	req->r_request->hdr.data_len = cpu_to_le32(data_len);
402

S
Sage Weil 已提交
403
	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
404 405 406
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);
407 408 409

	dout("build_request msg_size was %d num_ops %d\n", (int)msg_size,
	     num_ops);
410 411
	return;
}
412
EXPORT_SYMBOL(ceph_osdc_build_request);
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434

/*
 * 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,
435
					       bool use_mempool)
436
{
A
Alex Elder 已提交
437
	struct ceph_osd_req_op ops[2];
438
	struct ceph_osd_request *req;
A
Alex Elder 已提交
439
	unsigned int num_op = 1;
440
	u64 bno = 0;
441
	int r;
442

A
Alex Elder 已提交
443 444
	memset(&ops, 0, sizeof ops);

445 446 447 448 449 450
	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
451 452
		num_op++;
	}
453

A
Alex Elder 已提交
454 455
	req = ceph_osdc_alloc_request(osdc, snapc, num_op, use_mempool,
					GFP_NOFS);
456
	if (!req)
457
		return ERR_PTR(-ENOMEM);
458
	req->r_flags = flags;
459 460

	/* calculate max write size */
461
	r = calc_layout(layout, off, plen, ops, &bno);
462 463
	if (r < 0) {
		ceph_osdc_put_request(req);
464
		return ERR_PTR(r);
465
	}
466

467 468
	req->r_file_layout = *layout;  /* keep a copy */

469 470 471
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx", vino.ino, bno);
	req->r_oid_len = strlen(req->r_oid);

A
Alex Elder 已提交
472 473
	ceph_osdc_build_request(req, off, *plen, num_op, ops,
				snapc, vino.snap, mtime);
474

S
Sage Weil 已提交
475 476
	return req;
}
477
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543

/*
 * 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;
}

544 545 546 547 548 549
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
550
	struct ceph_osd_request *req, *nreq;
551 552 553 554
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
555
	if (err)
556 557 558 559 560 561
		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);
562 563 564 565 566 567
		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) {
568 569 570 571 572
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
573
		__register_request(osdc, req);
574 575 576
		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);
577 578
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
579 580 581
	}
}

S
Sage Weil 已提交
582
/*
583
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
584 585 586 587 588 589 590 591 592 593 594
 */
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
595 596
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
597
	__send_queued(osdc);
S
Sage Weil 已提交
598
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
599 600 601 602 603 604
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
605
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
606 607 608 609 610 611 612 613 614
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
615
	osd->o_osd = onum;
616
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
617
	INIT_LIST_HEAD(&osd->o_requests);
618
	INIT_LIST_HEAD(&osd->o_linger_requests);
619
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
620 621
	osd->o_incarnation = 1;

622
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
623

624
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
	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);
644
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
645 646
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

647
		if (ac->ops && ac->ops->destroy_authorizer)
648
			ac->ops->destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
649
		kfree(osd);
S
Sage Weil 已提交
650
	}
S
Sage Weil 已提交
651 652 653 654 655
}

/*
 * remove an osd from our map
 */
656
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
657
{
658
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
659 660
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
661
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
662 663 664 665
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

666 667
static void remove_all_osds(struct ceph_osd_client *osdc)
{
668
	dout("%s %p\n", __func__, osdc);
669 670 671 672 673 674 675 676 677
	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);
}

678 679 680 681 682 683
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);
684
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
685 686 687 688 689 690 691 692 693
}

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);
}

694
static void remove_old_osds(struct ceph_osd_client *osdc)
695 696 697 698 699 700
{
	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) {
701
		if (time_before(jiffies, osd->lru_ttl))
702 703 704 705 706 707
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
708 709 710
/*
 * reset osd connect
 */
711
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
712
{
A
Alex Elder 已提交
713
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
714

715
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
716 717
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
718
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
719 720 721 722 723 724 725 726 727

		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;

728 729 730 731 732
		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
733 734

		return -EAGAIN;
S
Sage Weil 已提交
735
	}
A
Alex Elder 已提交
736 737 738 739 740 741

	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
742 743 744 745 746 747 748 749
}

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;

750
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
	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;
}

783 784 785
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
786
			osdc->client->options->osd_keepalive_timeout * HZ);
787 788 789 790 791 792
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
793 794 795 796 797

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
798 799
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
800 801
{
	req->r_tid = ++osdc->last_tid;
802
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
803
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
804 805 806 807
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
808 809
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
810
	}
811 812 813 814 815 816 817
}

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
818 819 820 821 822 823 824 825 826
	mutex_unlock(&osdc->request_mutex);
}

/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
827 828 829 830 831 832
	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
833 834 835 836
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

837 838
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
839
		ceph_msg_revoke(req->r_request);
840 841

		list_del_init(&req->r_osd_item);
842 843 844
		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);
845
			__move_osd_to_lru(osdc, req->r_osd);
846
		}
847
		if (list_empty(&req->r_linger_item))
848
			req->r_osd = NULL;
849
	}
S
Sage Weil 已提交
850

A
Alex Elder 已提交
851
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
852 853
	ceph_osdc_put_request(req);

854 855 856
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
857 858 859 860 861 862 863 864
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
865
	if (req->r_sent && req->r_osd) {
866
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
867 868 869 870
		req->r_sent = 0;
	}
}

871 872 873 874 875
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);
876 877 878
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
879 880 881 882 883 884
}

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
885
	list_del_init(&req->r_linger_item);
886 887 888 889 890 891 892 893
	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);
		}
894 895
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925
	}
}

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
926 927 928
/*
 * 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
929
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
930
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
931 932 933 934 935
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
936
static int __map_request(struct ceph_osd_client *osdc,
937
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
938
{
939
	struct ceph_pg pgid;
940 941
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
942 943
	int err;

944
	dout("map_request %p tid %lld\n", req, req->r_tid);
945 946
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
947 948
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
949
		return err;
950
	}
951 952
	req->r_pgid = pgid;

953 954 955 956 957
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
958

959 960
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
961 962 963
	     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
964 965 966
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

967 968
	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
969 970
	     req->r_osd ? req->r_osd->o_osd : -1);

971 972 973 974
	/* 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
975 976 977 978 979 980 981 982
	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) {
983
		err = -ENOMEM;
984
		req->r_osd = create_osd(osdc, o);
985 986
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
987
			goto out;
988
		}
S
Sage Weil 已提交
989

990
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
991 992
		__insert_osd(osdc, req->r_osd);

993 994 995
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
996 997
	}

998 999
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
S
Sage Weil 已提交
1000
		list_add(&req->r_osd_item, &req->r_osd->o_requests);
1001 1002 1003
		list_move(&req->r_req_lru_item, &osdc->req_unsent);
	} else {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
1004
	}
1005
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1006 1007 1008 1009 1010 1011 1012 1013

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1014 1015
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1016
{
1017
	void *p;
S
Sage Weil 已提交
1018

1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
	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));
1033

S
Sage Weil 已提交
1034
	req->r_stamp = jiffies;
1035
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1036 1037 1038 1039 1040 1041

	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;
}

1042 1043 1044
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1045
static void __send_queued(struct ceph_osd_client *osdc)
1046 1047 1048
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1049 1050
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1051 1052 1053
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
/*
 * 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
1067
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1068
	struct ceph_osd *osd;
1069
	unsigned long keepalive =
1070
		osdc->client->options->osd_keepalive_timeout * HZ;
1071
	struct list_head slow_osds;
S
Sage Weil 已提交
1072 1073 1074 1075 1076 1077 1078
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1079 1080 1081 1082 1083 1084 1085
	/*
	 * 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
1086
		if (time_before(jiffies, req->r_stamp + keepalive))
1087 1088 1089 1090 1091
			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
1092
		     req->r_tid, osd->o_osd);
1093 1094 1095 1096 1097 1098
		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
1099 1100 1101
		ceph_con_keepalive(&osd->o_con);
	}

1102
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1103
	__send_queued(osdc);
S
Sage Weil 已提交
1104 1105 1106 1107
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1108 1109 1110 1111 1112 1113
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 =
1114
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1115 1116 1117

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1118
	remove_old_osds(osdc);
1119 1120 1121 1122 1123 1124
	up_read(&osdc->map_sem);

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

S
Sage Weil 已提交
1125 1126 1127 1128 1129 1130 1131
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 */
}

1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
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
1152 1153 1154 1155
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1156 1157
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1158
{
1159
	void *p, *end;
S
Sage Weil 已提交
1160 1161
	struct ceph_osd_request *req;
	u64 tid;
1162 1163
	int object_len;
	int numops, payload_len, flags;
1164
	s32 result;
1165 1166 1167 1168 1169 1170
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1171
	int already_completed;
1172
	int i;
S
Sage Weil 已提交
1173

1174
	tid = le64_to_cpu(msg->hdr.tid);
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
	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
1187
		goto bad;
1188 1189 1190 1191 1192 1193 1194 1195

	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
1196 1197 1198 1199 1200 1201 1202 1203 1204
	/* 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);
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236

	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
1237

1238
	/*
1239
	 * if this connection filled our message, drop our reference now, to
1240 1241
	 * avoid a (safe but slower) revoke later.
	 */
1242
	if (req->r_con_filling_msg == con && req->r_reply == msg) {
S
Sage Weil 已提交
1243
		dout(" dropping con_filling_msg ref %p\n", con);
1244
		req->r_con_filling_msg = NULL;
1245
		con->ops->put(con);
1246 1247
	}

S
Sage Weil 已提交
1248
	if (!req->r_got_reply) {
1249
		unsigned int bytes;
S
Sage Weil 已提交
1250

1251
		req->r_result = result;
S
Sage Weil 已提交
1252 1253 1254 1255 1256 1257 1258
		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, */
1259 1260
		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
1261 1262 1263 1264

		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
1265
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1266 1267 1268 1269 1270
		goto done;
	}

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

1271 1272 1273
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1274
	/* either this is a read, or we got the safe response */
1275 1276
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1277 1278 1279
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1280 1281
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1282
	mutex_unlock(&osdc->request_mutex);
1283 1284
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1285 1286 1287 1288

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1289
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1290

S
Sage Weil 已提交
1291 1292
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1293 1294

done:
1295
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1296 1297 1298
	ceph_osdc_put_request(req);
	return;

1299 1300
bad_put:
	ceph_osdc_put_request(req);
S
Sage Weil 已提交
1301
bad:
1302 1303
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1304
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1305 1306
}

1307
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1308 1309 1310
{
	struct rb_node *p, *n;

1311 1312
	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
1313

1314 1315 1316 1317 1318 1319 1320
		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
1321
	}
1322 1323 1324
}

/*
1325 1326
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1327
 *
1328
 * Caller should hold map_sem for read.
1329
 */
1330
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1331
{
1332
	struct ceph_osd_request *req, *nreq;
1333 1334 1335
	struct rb_node *p;
	int needmap = 0;
	int err;
1336

1337
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1338
	mutex_lock(&osdc->request_mutex);
1339
	for (p = rb_first(&osdc->requests); p; ) {
1340
		req = rb_entry(p, struct ceph_osd_request, r_node);
1341
		p = rb_next(p);
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359

		/*
		 * 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;
		}

1360
		err = __map_request(osdc, req, force_resend);
1361 1362 1363 1364 1365 1366
		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) {
1367 1368 1369 1370
			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);
1371
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1372 1373
			}
		}
1374 1375 1376 1377 1378 1379
	}

	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);

1380
		err = __map_request(osdc, req, force_resend);
1381
		dout("__map_request returned %d\n", err);
1382 1383 1384 1385 1386 1387 1388 1389
		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;
1390
		}
1391 1392 1393 1394

		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);
1395
		__unregister_linger_request(osdc, req);
1396
	}
S
Sage Weil 已提交
1397 1398 1399 1400 1401 1402
	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);
	}
1403
	reset_changed_osds(osdc);
1404
}
1405 1406


S
Sage Weil 已提交
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
/*
 * 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));
1430 1431
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1432 1433 1434 1435 1436 1437 1438 1439

	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);
1440 1441
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1442 1443 1444 1445 1446 1447 1448
		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,
1449
							  &osdc->client->msgr);
S
Sage Weil 已提交
1450 1451 1452 1453
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1454
			BUG_ON(!newmap);
S
Sage Weil 已提交
1455 1456 1457 1458
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1459
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
		} 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);
1475 1476
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1477 1478 1479 1480 1481 1482 1483 1484 1485
		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 {
1486 1487
			int skipped_map = 0;

S
Sage Weil 已提交
1488 1489 1490 1491 1492 1493
			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;
			}
1494
			BUG_ON(!newmap);
S
Sage Weil 已提交
1495 1496
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1497 1498 1499
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1500
				ceph_osdmap_destroy(oldmap);
1501 1502
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1503 1504 1505 1506 1507 1508 1509 1510
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1511 1512 1513 1514 1515 1516 1517 1518 1519

	/*
	 * 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
1520 1521 1522
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1523
	up_read(&osdc->map_sem);
1524
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1525 1526 1527 1528
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1529
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1530 1531 1532 1533
	up_write(&osdc->map_sem);
	return;
}

1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 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
/*
 * 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 *),
1616
			   void *data, struct ceph_osd_event **pevent)
1617 1618 1619 1620 1621 1622 1623 1624 1625
{
	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;
1626
	event->one_shot = 0;
1627 1628 1629
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1630
	RB_CLEAR_NODE(&event->node);
1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
	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
 */
1677 1678
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698
{
	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) {
1699
		BUG_ON(event->one_shot);
1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		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;
		}
1711
		INIT_WORK(&event_work->work, do_event_work);
1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732
		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;
}

S
Sage Weil 已提交
1733 1734 1735 1736 1737 1738 1739
/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
S
Sage Weil 已提交
1740
	int rc = 0;
S
Sage Weil 已提交
1741

1742 1743 1744
	req->r_request->pages = req->r_data.pages;
	req->r_request->page_count = req->r_data.num_pages;
	req->r_request->page_alignment = req->r_data.alignment;
1745
#ifdef CONFIG_BLOCK
1746
	req->r_request->bio = req->r_data.bio;
1747
#endif
1748
	req->r_request->trail = &req->r_trail;
S
Sage Weil 已提交
1749 1750 1751 1752 1753

	register_request(osdc, req);

	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
S
Sage Weil 已提交
1754 1755 1756 1757 1758 1759
	/*
	 * 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) {
1760
		rc = __map_request(osdc, req, 0);
1761 1762 1763 1764 1765 1766
		if (rc < 0) {
			if (nofail) {
				dout("osdc_start_request failed map, "
				     " will retry %lld\n", req->r_tid);
				rc = 0;
			}
1767
			goto out_unlock;
1768
		}
1769 1770 1771 1772
		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 {
1773
			__send_request(osdc, req);
S
Sage Weil 已提交
1774
		}
1775
		rc = 0;
S
Sage Weil 已提交
1776
	}
1777 1778

out_unlock:
S
Sage Weil 已提交
1779 1780 1781 1782
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
1783
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796

/*
 * 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);
1797
		__unregister_request(osdc, req);
S
Sage Weil 已提交
1798
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1799
		complete_request(req);
1800
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
1801 1802 1803 1804 1805 1806
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
1807
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840

/*
 * 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);
}
1841
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858

/*
 * 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;
1859
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
1860
	osdc->requests = RB_ROOT;
1861
	INIT_LIST_HEAD(&osdc->req_lru);
1862 1863
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
1864
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
1865 1866
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
1867
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1868 1869 1870
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
1871 1872

	schedule_delayed_work(&osdc->osds_timeout_work,
1873
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
1874

1875
	err = -ENOMEM;
S
Sage Weil 已提交
1876 1877 1878
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
1879
		goto out;
S
Sage Weil 已提交
1880

1881 1882
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
1883
				"osd_op");
S
Sage Weil 已提交
1884
	if (err < 0)
1885
		goto out_mempool;
1886
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
1887 1888
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
1889 1890
	if (err < 0)
		goto out_msgpool;
1891 1892 1893 1894 1895 1896 1897

	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
1898
	return 0;
1899

S
Sage Weil 已提交
1900 1901
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
1902 1903 1904 1905
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
1906 1907 1908 1909
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
1910 1911
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
1912
	cancel_delayed_work_sync(&osdc->timeout_work);
1913
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
1914 1915 1916 1917
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
1918
	remove_all_osds(osdc);
S
Sage Weil 已提交
1919 1920
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
1921
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931
}

/*
 * 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,
1932
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
1933 1934 1935 1936 1937 1938 1939 1940 1941
{
	struct ceph_osd_request *req;
	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,
1942
				    false);
1943 1944
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
1945 1946

	/* it may be a short read due to an object boundary */
1947 1948 1949
	req->r_data.pages = pages;
	req->r_data.num_pages = calc_pages_for(page_align, *plen);
	req->r_data.alignment = page_align;
S
Sage Weil 已提交
1950

1951
	dout("readpages  final extent is %llu~%llu (%d pages align %d)\n",
1952
	     off, *plen, req->r_data.num_pages, page_align);
S
Sage Weil 已提交
1953 1954 1955 1956 1957 1958 1959 1960 1961

	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;
}
1962
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
1963 1964 1965 1966 1967 1968 1969 1970 1971 1972

/*
 * 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,
1973
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
1974 1975 1976
{
	struct ceph_osd_request *req;
	int rc = 0;
1977
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
1978 1979 1980 1981

	BUG_ON(vino.snap != CEPH_NOSNAP);
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
				    CEPH_OSD_OP_WRITE,
1982
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
1983
				    snapc, 0,
S
Sage Weil 已提交
1984
				    truncate_seq, truncate_size, mtime,
1985
				    true);
1986 1987
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
1988 1989

	/* it may be a short write due to an object boundary */
1990 1991 1992 1993
	req->r_data.pages = pages;
	req->r_data.num_pages = calc_pages_for(page_align, len);
	req->r_data.alignment = page_align;
	dout("writepages %llu~%llu (%d pages)\n", off, len, req->r_data.num_pages);
S
Sage Weil 已提交
1994

1995
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
1996 1997 1998 1999 2000 2001 2002 2003 2004
	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;
}
2005
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2006 2007 2008 2009 2010 2011 2012

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2013
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2014 2015 2016
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2017
		goto out;
2018
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2019 2020 2021 2022 2023 2024

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2025
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2026
		break;
2027 2028 2029
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2030 2031 2032 2033 2034

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2035
out:
S
Sage Weil 已提交
2036 2037 2038
	ceph_msg_put(msg);
}

2039
/*
2040 2041
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2042 2043
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2044 2045
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2046 2047 2048
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2049
	struct ceph_msg *m;
2050
	struct ceph_osd_request *req;
2051 2052
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2053
	u64 tid;
S
Sage Weil 已提交
2054

2055 2056 2057 2058 2059 2060
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2061 2062
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2063 2064
		goto out;
	}
S
Sage Weil 已提交
2065 2066

	if (req->r_con_filling_msg) {
2067
		dout("%s revoking msg %p from old con %p\n", __func__,
S
Sage Weil 已提交
2068
		     req->r_reply, req->r_con_filling_msg);
2069
		ceph_msg_revoke_incoming(req->r_reply);
2070
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
2071
		req->r_con_filling_msg = NULL;
2072 2073
	}

S
Sage Weil 已提交
2074 2075 2076
	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);
2077
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2078
		if (!m)
S
Sage Weil 已提交
2079 2080 2081 2082 2083 2084
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2085
	if (data_len > 0) {
2086
		int want = calc_pages_for(req->r_data.alignment, data_len);
2087

2088
		if (req->r_data.pages && unlikely(req->r_data.num_pages < want)) {
2089 2090
			pr_warning("tid %lld reply has %d bytes %d pages, we"
				   " had only %d pages ready\n", tid, data_len,
2091
				   want, req->r_data.num_pages);
2092 2093
			*skip = 1;
			ceph_msg_put(m);
2094
			m = NULL;
2095
			goto out;
2096
		}
2097 2098 2099
		m->pages = req->r_data.pages;
		m->page_count = req->r_data.num_pages;
		m->page_alignment = req->r_data.alignment;
2100
#ifdef CONFIG_BLOCK
2101
		m->bio = req->r_data.bio;
2102
#endif
2103
	}
2104
	*skip = 0;
2105
	req->r_con_filling_msg = con->ops->get(con);
S
Sage Weil 已提交
2106
	dout("get_reply tid %lld %p\n", tid, m);
2107 2108 2109

out:
	mutex_unlock(&osdc->request_mutex);
2110
	return m;
2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121

}

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
2122
	*skip = 0;
2123 2124
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2125
	case CEPH_MSG_WATCH_NOTIFY:
2126
		return ceph_msg_new(type, front, GFP_NOFS, false);
2127 2128 2129 2130 2131 2132 2133 2134
	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
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153
}

/*
 * 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);
}

2154 2155 2156
/*
 * authentication
 */
2157 2158 2159 2160 2161
/*
 * 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,
2162
					int *proto, int force_new)
2163 2164 2165 2166
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2167
	struct ceph_auth_handshake *auth = &o->o_auth;
2168

2169
	if (force_new && auth->authorizer) {
2170 2171
		if (ac->ops && ac->ops->destroy_authorizer)
			ac->ops->destroy_authorizer(ac, auth->authorizer);
2172 2173
		auth->authorizer = NULL;
	}
2174
	if (!auth->authorizer && ac->ops && ac->ops->create_authorizer) {
2175 2176
		int ret = ac->ops->create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
							auth);
2177
		if (ret)
2178
			return ERR_PTR(ret);
2179 2180
	}
	*proto = ac->protocol;
2181

2182
	return auth;
2183 2184 2185 2186 2187 2188 2189 2190 2191
}


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;

2192 2193 2194 2195
	/*
	 * XXX If ac->ops or ac->ops->verify_authorizer_reply is null,
	 * XXX which do we do:  succeed or fail?
	 */
2196
	return ac->ops->verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2197 2198
}

2199 2200 2201 2202 2203 2204
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;

2205
	if (ac->ops && ac->ops->invalidate_authorizer)
2206 2207 2208 2209
		ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);

	return ceph_monc_validate_auth(&osdc->client->monc);
}
2210

2211
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2212 2213 2214
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2215 2216
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2217
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2218
	.alloc_msg = alloc_msg,
2219
	.fault = osd_reset,
S
Sage Weil 已提交
2220
};