osd_client.c 67.2 KB
Newer Older
1

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

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

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

S
Sage Weil 已提交
22 23
#define OSD_OP_FRONT_LEN	4096
#define OSD_OPREPLY_FRONT_LEN	512
24

25
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
26

A
Alex Elder 已提交
27
static void __send_queued(struct ceph_osd_client *osdc);
28
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
29 30 31 32
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);
33 34
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

/*
 * 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.
 */
61
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
62
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
63
{
64
	u64 orig_len = *plen;
S
Sage Weil 已提交
65
	int r;
S
Sage Weil 已提交
66

67
	/* object extent? */
68 69
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
70 71
	if (r < 0)
		return r;
72 73
	if (*objlen < orig_len) {
		*plen = *objlen;
74 75 76 77
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

78
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
79

80
	return 0;
S
Sage Weil 已提交
81 82
}

A
Alex Elder 已提交
83 84 85 86 87 88
static void ceph_osd_data_init(struct ceph_osd_data *osd_data)
{
	memset(osd_data, 0, sizeof (*osd_data));
	osd_data->type = CEPH_OSD_DATA_TYPE_NONE;
}

89
static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
90 91 92 93 94 95 96 97 98 99 100
			struct page **pages, u64 length, u32 alignment,
			bool pages_from_pool, bool own_pages)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
	osd_data->length = length;
	osd_data->alignment = alignment;
	osd_data->pages_from_pool = pages_from_pool;
	osd_data->own_pages = own_pages;
}

101
static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
102 103 104 105 106 107 108
			struct ceph_pagelist *pagelist)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
	osd_data->pagelist = pagelist;
}

#ifdef CONFIG_BLOCK
109
static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
110 111 112 113 114 115 116 117
			struct bio *bio, size_t bio_length)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_BIO;
	osd_data->bio = bio;
	osd_data->bio_length = bio_length;
}
#endif /* CONFIG_BLOCK */

118 119 120 121 122 123
#define osd_req_op_data(oreq, whch, typ, fld)	\
	({						\
		BUG_ON(whch >= (oreq)->r_num_ops);	\
		&(oreq)->r_ops[whch].typ.fld;		\
	})

A
Alex Elder 已提交
124 125 126 127 128 129 130 131
static struct ceph_osd_data *
osd_req_op_raw_data_in(struct ceph_osd_request *osd_req, unsigned int which)
{
	BUG_ON(which >= osd_req->r_num_ops);

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

132 133
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
134
			unsigned int which)
135
{
136
	return osd_req_op_data(osd_req, which, extent, osd_data);
137 138 139 140 141 142 143
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data);

struct ceph_osd_data *
osd_req_op_cls_response_data(struct ceph_osd_request *osd_req,
			unsigned int which)
{
144
	return osd_req_op_data(osd_req, which, cls, response_data);
145 146 147
}
EXPORT_SYMBOL(osd_req_op_cls_response_data);	/* ??? */

A
Alex Elder 已提交
148 149 150 151 152 153 154 155 156 157 158 159 160
void osd_req_op_raw_data_in_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

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

161
void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
162 163
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
164 165 166 167
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

168
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
169 170 171 172 173 174
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);

void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
175
			unsigned int which, struct ceph_pagelist *pagelist)
176 177 178
{
	struct ceph_osd_data *osd_data;

179
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
180 181 182 183 184 185
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);

#ifdef CONFIG_BLOCK
void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
186
			unsigned int which, struct bio *bio, size_t bio_length)
187 188
{
	struct ceph_osd_data *osd_data;
189 190

	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
191 192 193 194 195 196 197 198 199 200 201
	ceph_osd_data_bio_init(osd_data, bio, bio_length);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio);
#endif /* CONFIG_BLOCK */

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

202
	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
203 204 205
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

206 207 208 209 210 211
void osd_req_op_cls_request_data_pagelist(
			struct ceph_osd_request *osd_req,
			unsigned int which, struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

212
	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
213 214 215 216
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);

217 218 219 220 221 222 223 224 225 226 227 228
void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages, u64 length,
			u32 alignment, bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pages);

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

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

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

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

		num_pages = calc_pages_for((u64)osd_data->alignment,
						(u64)osd_data->length);
		ceph_release_page_vector(osd_data->pages, num_pages);
	}
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
	ceph_osd_data_init(osd_data);
}

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

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

	switch (op->op) {
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		ceph_osd_data_release(&op->extent.osd_data);
		break;
	case CEPH_OSD_OP_CALL:
		ceph_osd_data_release(&op->cls.request_info);
287
		ceph_osd_data_release(&op->cls.request_data);
288 289 290 291 292
		ceph_osd_data_release(&op->cls.response_data);
		break;
	default:
		break;
	}
A
Alex Elder 已提交
293 294
}

S
Sage Weil 已提交
295 296 297
/*
 * requests
 */
S
Sage Weil 已提交
298
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
299
{
A
Alex Elder 已提交
300
	struct ceph_osd_request *req;
301
	unsigned int which;
S
Sage Weil 已提交
302

A
Alex Elder 已提交
303
	req = container_of(kref, struct ceph_osd_request, r_kref);
S
Sage Weil 已提交
304 305
	if (req->r_request)
		ceph_msg_put(req->r_request);
306
	if (req->r_reply) {
307
		ceph_msg_revoke_incoming(req->r_reply);
308
		ceph_msg_put(req->r_reply);
309
	}
310

311 312
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
313

S
Sage Weil 已提交
314 315 316 317 318
	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
319
}
320
EXPORT_SYMBOL(ceph_osdc_release_request);
321

322
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
323
					       struct ceph_snap_context *snapc,
324
					       unsigned int num_ops,
325
					       bool use_mempool,
326
					       gfp_t gfp_flags)
S
Sage Weil 已提交
327 328 329
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
330 331
	size_t msg_size;

332 333 334
	BUILD_BUG_ON(CEPH_OSD_MAX_OP > U16_MAX);
	BUG_ON(num_ops > CEPH_OSD_MAX_OP);

335 336 337 338 339 340 341 342 343
	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
344 345

	if (use_mempool) {
346
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
347 348
		memset(req, 0, sizeof(*req));
	} else {
349
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
350 351
	}
	if (req == NULL)
352
		return NULL;
S
Sage Weil 已提交
353 354 355

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
356
	req->r_num_ops = num_ops;
357

S
Sage Weil 已提交
358
	kref_init(&req->r_kref);
S
Sage Weil 已提交
359 360
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
361
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
362
	INIT_LIST_HEAD(&req->r_unsafe_item);
363 364
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
365
	INIT_LIST_HEAD(&req->r_req_lru_item);
366 367
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
368 369 370 371 372
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
373
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
374
	if (!msg) {
S
Sage Weil 已提交
375
		ceph_osdc_put_request(req);
376
		return NULL;
S
Sage Weil 已提交
377 378 379 380
	}
	req->r_reply = msg;

	/* create request message; allow space for oid */
S
Sage Weil 已提交
381
	if (use_mempool)
382
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
383
	else
384
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
385
	if (!msg) {
S
Sage Weil 已提交
386
		ceph_osdc_put_request(req);
387
		return NULL;
S
Sage Weil 已提交
388
	}
389

S
Sage Weil 已提交
390
	memset(msg->front.iov_base, 0, msg->front.iov_len);
391 392 393 394 395

	req->r_request = msg;

	return req;
}
396
EXPORT_SYMBOL(ceph_osdc_alloc_request);
397

398
static bool osd_req_opcode_valid(u16 opcode)
399
{
400
	switch (opcode) {
401
	case CEPH_OSD_OP_READ:
402
	case CEPH_OSD_OP_STAT:
403 404 405
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
406
	case CEPH_OSD_OP_NOTIFY:
407
	case CEPH_OSD_OP_NOTIFY_ACK:
408
	case CEPH_OSD_OP_ASSERT_VER:
409
	case CEPH_OSD_OP_WRITE:
410 411 412 413 414
	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:
415
	case CEPH_OSD_OP_STARTSYNC:
416 417 418 419 420 421
	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:
422
	case CEPH_OSD_OP_ROLLBACK:
423
	case CEPH_OSD_OP_WATCH:
424 425 426 427 428 429 430 431 432 433 434 435
	case CEPH_OSD_OP_OMAPGETKEYS:
	case CEPH_OSD_OP_OMAPGETVALS:
	case CEPH_OSD_OP_OMAPGETHEADER:
	case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
	case CEPH_OSD_OP_OMAPSETVALS:
	case CEPH_OSD_OP_OMAPSETHEADER:
	case CEPH_OSD_OP_OMAPCLEAR:
	case CEPH_OSD_OP_OMAPRMKEYS:
	case CEPH_OSD_OP_OMAP_CMP:
	case CEPH_OSD_OP_CLONERANGE:
	case CEPH_OSD_OP_ASSERT_SRC_VERSION:
	case CEPH_OSD_OP_SRC_CMPXATTR:
436
	case CEPH_OSD_OP_GETXATTR:
437
	case CEPH_OSD_OP_GETXATTRS:
438 439
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
	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:
458
	case CEPH_OSD_OP_CALL:
459 460
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
461 462 463 464 465 466
		return true;
	default:
		return false;
	}
}

467 468 469 470 471
/*
 * This is an osd op init function for opcodes that have no data or
 * other information associated with them.  It also serves as a
 * common init routine for all the other init functions, below.
 */
472
static struct ceph_osd_req_op *
A
Alex Elder 已提交
473
_osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
474
				u16 opcode)
475
{
476 477 478
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
479 480
	BUG_ON(!osd_req_opcode_valid(opcode));

481
	op = &osd_req->r_ops[which];
482 483
	memset(op, 0, sizeof (*op));
	op->op = opcode;
484 485

	return op;
486 487
}

A
Alex Elder 已提交
488 489 490 491 492 493 494
void osd_req_op_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode)
{
	(void)_osd_req_op_init(osd_req, which, opcode);
}
EXPORT_SYMBOL(osd_req_op_init);

495 496
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
497 498 499
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
A
Alex Elder 已提交
500
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
	size_t payload_len = 0;

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

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

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

516 517
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
518
{
519 520 521 522 523 524
	struct ceph_osd_req_op *op;
	u64 previous;

	BUG_ON(which >= osd_req->r_num_ops);
	op = &osd_req->r_ops[which];
	previous = op->extent.length;
525 526 527 528 529 530 531 532 533 534

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

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

535
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
536
			u16 opcode, const char *class, const char *method)
537
{
A
Alex Elder 已提交
538
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
539
	struct ceph_pagelist *pagelist;
540 541 542 543 544
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

545 546 547 548
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

549 550 551 552
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
553
	ceph_pagelist_append(pagelist, class, size);
554 555 556 557 558 559
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
560
	ceph_pagelist_append(pagelist, method, size);
561 562
	payload_len += size;

563
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
564

565 566 567 568 569
	op->cls.argc = 0;	/* currently unused */

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_cls_init);
570

571 572
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
573 574
				u64 cookie, u64 version, int flag)
{
A
Alex Elder 已提交
575
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
576

577
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
578 579

	op->watch.cookie = cookie;
A
Alex Elder 已提交
580
	op->watch.ver = version;
581
	if (opcode == CEPH_OSD_OP_WATCH && flag)
582
		op->watch.flag = (u8)1;
583 584 585
}
EXPORT_SYMBOL(osd_req_op_watch_init);

586
static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
587 588 589 590 591 592 593
				struct ceph_osd_data *osd_data)
{
	u64 length = ceph_osd_data_length(osd_data);

	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
		BUG_ON(length > (u64) SIZE_MAX);
		if (length)
594
			ceph_msg_data_add_pages(msg, osd_data->pages,
595 596 597
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
598
		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
599 600
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
601
		ceph_msg_data_add_bio(msg, osd_data->bio, length);
602 603 604 605 606 607
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

608
static u64 osd_req_encode_op(struct ceph_osd_request *req,
609
			      struct ceph_osd_op *dst, unsigned int which)
610
{
611
	struct ceph_osd_req_op *src;
612
	struct ceph_osd_data *osd_data;
613
	u64 request_data_len = 0;
614
	u64 data_length;
615

616 617
	BUG_ON(which >= req->r_num_ops);
	src = &req->r_ops[which];
618 619 620 621 622 623 624 625
	if (WARN_ON(!osd_req_opcode_valid(src->op))) {
		pr_err("unrecognized osd opcode %d\n", src->op);

		return 0;
	}

	switch (src->op) {
	case CEPH_OSD_OP_STAT:
A
Alex Elder 已提交
626 627
		osd_data = &src->raw_data_in;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
628 629 630 631
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		if (src->op == CEPH_OSD_OP_WRITE)
632
			request_data_len = src->extent.length;
633 634 635 636 637 638
		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);
639
		osd_data = &src->extent.osd_data;
640
		if (src->op == CEPH_OSD_OP_WRITE)
641
			ceph_osdc_msg_data_add(req->r_request, osd_data);
642
		else
643
			ceph_osdc_msg_data_add(req->r_reply, osd_data);
644 645 646 647
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
		osd_data = &src->cls.request_info;
		ceph_osdc_msg_data_add(req->r_request, osd_data);
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGELIST);
		request_data_len = osd_data->pagelist->length;

		osd_data = &src->cls.request_data;
		data_length = ceph_osd_data_length(osd_data);
		if (data_length) {
			BUG_ON(osd_data->type == CEPH_OSD_DATA_TYPE_NONE);
			dst->cls.indata_len = cpu_to_le32(data_length);
			ceph_osdc_msg_data_add(req->r_request, osd_data);
			src->payload_len += data_length;
			request_data_len += data_length;
		}
		osd_data = &src->cls.response_data;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
664 665 666 667 668 669 670 671 672 673
		break;
	case CEPH_OSD_OP_STARTSYNC:
		break;
	case CEPH_OSD_OP_NOTIFY_ACK:
	case CEPH_OSD_OP_WATCH:
		dst->watch.cookie = cpu_to_le64(src->watch.cookie);
		dst->watch.ver = cpu_to_le64(src->watch.ver);
		dst->watch.flag = src->watch.flag;
		break;
	default:
674
		pr_err("unsupported osd opcode %s\n",
675
			ceph_osd_op_name(src->op));
676
		WARN_ON(1);
677 678

		return 0;
679
	}
680
	dst->op = cpu_to_le16(src->op);
681
	dst->payload_len = cpu_to_le32(src->payload_len);
682

683
	return request_data_len;
684 685
}

686 687 688 689 690 691 692 693 694 695 696 697 698 699
/*
 * 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,
700
					       u64 off, u64 *plen, int num_ops,
701 702 703 704
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
705
					       bool use_mempool)
706
{
707
	struct ceph_osd_request *req;
708 709 710
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
711 712
	u32 object_size;
	u64 object_base;
713
	int r;
714

715
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
716

717
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
718
					GFP_NOFS);
719
	if (!req)
720
		return ERR_PTR(-ENOMEM);
721

722
	req->r_flags = flags;
723 724

	/* calculate max write size */
725
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
726 727
	if (r < 0) {
		ceph_osdc_put_request(req);
728
		return ERR_PTR(r);
729
	}
730

731 732 733 734 735 736 737 738
	object_size = le32_to_cpu(layout->fl_object_size);
	object_base = off - objoff;
	if (truncate_size <= object_base) {
		truncate_size = 0;
	} else {
		truncate_size -= object_base;
		if (truncate_size > object_size)
			truncate_size = object_size;
739
	}
740

741
	osd_req_op_extent_init(req, 0, opcode, objoff, objlen,
742
				truncate_size, truncate_seq);
743

744 745 746 747 748 749
	/*
	 * A second op in the ops array means the caller wants to
	 * also issue a include a 'startsync' command so that the
	 * osd will flush data quickly.
	 */
	if (num_ops > 1)
750
		osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC);
751

752 753
	req->r_file_layout = *layout;  /* keep a copy */

754 755
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx",
		vino.ino, objnum);
756 757
	req->r_oid_len = strlen(req->r_oid);

S
Sage Weil 已提交
758 759
	return req;
}
760
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826

/*
 * 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;
}

827 828 829 830 831 832
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
833
	struct ceph_osd_request *req, *nreq;
834
	LIST_HEAD(resend);
835 836 837 838
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
839
	if (err)
840
		return;
841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
	/*
	 * Build up a list of requests to resend by traversing the
	 * osd's list of requests.  Requests for a given object are
	 * sent in tid order, and that is also the order they're
	 * kept on this list.  Therefore all requests that are in
	 * flight will be found first, followed by all requests that
	 * have not yet been sent.  And to resend requests while
	 * preserving this order we will want to put any sent
	 * requests back on the front of the osd client's unsent
	 * list.
	 *
	 * So we build a separate ordered list of already-sent
	 * requests for the affected osd and splice it onto the
	 * front of the osd client's unsent list.  Once we've seen a
	 * request that has not yet been sent we're done.  Those
	 * requests are already sitting right where they belong.
	 */
858
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
859 860 861 862
		if (!req->r_sent)
			break;
		list_move_tail(&req->r_req_lru_item, &resend);
		dout("requeueing %p tid %llu osd%d\n", req, req->r_tid,
863
		     osd->o_osd);
864 865 866
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}
867
	list_splice(&resend, &osdc->req_unsent);
868

869 870 871 872 873
	/*
	 * Linger requests are re-registered before sending, which
	 * sets up a new tid for each.  We add them to the unsent
	 * list at the end to keep things in tid order.
	 */
874 875
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
876 877 878 879 880
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
881
		__register_request(osdc, req);
882
		list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
883
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
884
		__unregister_linger_request(osdc, req);
885 886
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
887 888 889
	}
}

S
Sage Weil 已提交
890
/*
891
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
892 893 894 895 896 897 898 899 900 901 902
 */
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
903 904
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
905
	__send_queued(osdc);
S
Sage Weil 已提交
906
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
907 908 909 910 911 912
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
913
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
914 915 916 917 918 919 920 921 922
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
923
	osd->o_osd = onum;
924
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
925
	INIT_LIST_HEAD(&osd->o_requests);
926
	INIT_LIST_HEAD(&osd->o_linger_requests);
927
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
928 929
	osd->o_incarnation = 1;

930
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
931

932
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951
	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);
952
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
953 954
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

955
		ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
956
		kfree(osd);
S
Sage Weil 已提交
957
	}
S
Sage Weil 已提交
958 959 960 961 962
}

/*
 * remove an osd from our map
 */
963
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
964
{
965
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
966 967
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
968
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
969 970 971 972
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

973 974
static void remove_all_osds(struct ceph_osd_client *osdc)
{
975
	dout("%s %p\n", __func__, osdc);
976 977 978 979 980 981 982 983 984
	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);
}

985 986 987 988 989 990
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);
991
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
992 993 994 995 996 997 998 999 1000
}

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);
}

1001
static void remove_old_osds(struct ceph_osd_client *osdc)
1002 1003 1004 1005 1006 1007
{
	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) {
1008
		if (time_before(jiffies, osd->lru_ttl))
1009 1010 1011 1012 1013 1014
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
1015 1016 1017
/*
 * reset osd connect
 */
1018
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1019
{
A
Alex Elder 已提交
1020
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
1021

1022
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
1023 1024
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
1025
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
1026 1027 1028 1029 1030 1031 1032 1033 1034

		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;

1035 1036 1037 1038 1039
		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
1040 1041

		return -EAGAIN;
S
Sage Weil 已提交
1042
	}
A
Alex Elder 已提交
1043 1044 1045 1046 1047 1048

	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
1049 1050 1051 1052 1053 1054 1055 1056
}

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;

1057
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
	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;
}

1090 1091 1092
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
1093
			osdc->client->options->osd_keepalive_timeout * HZ);
1094 1095 1096 1097 1098 1099
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
1100 1101 1102 1103 1104

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1105 1106
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1107 1108
{
	req->r_tid = ++osdc->last_tid;
1109
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1110
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
1111 1112 1113 1114
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
1115 1116
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
1117
	}
1118 1119
}

S
Sage Weil 已提交
1120 1121 1122 1123 1124 1125
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1126 1127 1128 1129 1130 1131
	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
1132 1133 1134 1135
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

1136 1137
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1138
		ceph_msg_revoke(req->r_request);
1139 1140

		list_del_init(&req->r_osd_item);
1141 1142 1143
		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);
1144
			__move_osd_to_lru(osdc, req->r_osd);
1145
		}
1146
		if (list_empty(&req->r_linger_item))
1147
			req->r_osd = NULL;
1148
	}
S
Sage Weil 已提交
1149

A
Alex Elder 已提交
1150
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1151 1152
	ceph_osdc_put_request(req);

1153 1154 1155
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1156 1157 1158 1159 1160 1161 1162 1163
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1164
	if (req->r_sent && req->r_osd) {
1165
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1166 1167 1168 1169
		req->r_sent = 0;
	}
}

1170 1171 1172 1173 1174
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);
1175 1176 1177
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
1178 1179 1180 1181 1182 1183
}

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
1184
	list_del_init(&req->r_linger_item);
1185 1186 1187 1188 1189 1190 1191 1192
	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);
		}
1193 1194
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
	}
}

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
1225 1226 1227
/*
 * 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
1228
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1229
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1230 1231 1232 1233 1234
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1235
static int __map_request(struct ceph_osd_client *osdc,
1236
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1237
{
1238
	struct ceph_pg pgid;
1239 1240
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
1241 1242
	int err;

1243
	dout("map_request %p tid %lld\n", req, req->r_tid);
1244 1245
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
1246 1247
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1248
		return err;
1249
	}
1250 1251
	req->r_pgid = pgid;

1252 1253 1254 1255 1256
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
1257

1258 1259
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1260 1261 1262
	     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
1263 1264 1265
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

1266 1267
	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
1268 1269
	     req->r_osd ? req->r_osd->o_osd : -1);

1270 1271 1272 1273
	/* 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
1274 1275 1276 1277 1278 1279 1280 1281
	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) {
1282
		err = -ENOMEM;
1283
		req->r_osd = create_osd(osdc, o);
1284 1285
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1286
			goto out;
1287
		}
S
Sage Weil 已提交
1288

1289
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1290 1291
		__insert_osd(osdc, req->r_osd);

1292 1293 1294
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1295 1296
	}

1297 1298
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
1299 1300
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1301
	} else {
1302
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1303
	}
1304
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1305 1306 1307 1308 1309 1310 1311 1312

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1313 1314
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1315
{
1316
	void *p;
S
Sage Weil 已提交
1317

1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
	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));
1332

S
Sage Weil 已提交
1333
	req->r_stamp = jiffies;
1334
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1335 1336

	ceph_msg_get(req->r_request); /* send consumes a ref */
1337 1338 1339 1340 1341

	/* Mark the request unsafe if this is the first timet's being sent. */

	if (!req->r_sent && req->r_unsafe_callback)
		req->r_unsafe_callback(req, true);
S
Sage Weil 已提交
1342
	req->r_sent = req->r_osd->o_incarnation;
1343 1344

	ceph_con_send(&req->r_osd->o_con, req->r_request);
S
Sage Weil 已提交
1345 1346
}

1347 1348 1349
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1350
static void __send_queued(struct ceph_osd_client *osdc)
1351 1352 1353
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1354 1355
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1356 1357 1358
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
/*
 * 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
1372
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1373
	struct ceph_osd *osd;
1374
	unsigned long keepalive =
1375
		osdc->client->options->osd_keepalive_timeout * HZ;
1376
	struct list_head slow_osds;
S
Sage Weil 已提交
1377 1378 1379 1380 1381 1382 1383
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1384 1385 1386 1387 1388 1389 1390
	/*
	 * 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
1391
		if (time_before(jiffies, req->r_stamp + keepalive))
1392 1393 1394 1395 1396
			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
1397
		     req->r_tid, osd->o_osd);
1398 1399 1400 1401 1402 1403
		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
1404 1405 1406
		ceph_con_keepalive(&osd->o_con);
	}

1407
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1408
	__send_queued(osdc);
S
Sage Weil 已提交
1409 1410 1411 1412
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1413 1414 1415 1416 1417 1418
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 =
1419
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1420 1421 1422

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1423
	remove_old_osds(osdc);
1424 1425 1426 1427 1428 1429
	up_read(&osdc->map_sem);

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

S
Sage Weil 已提交
1430 1431
static void complete_request(struct ceph_osd_request *req)
{
1432 1433
	if (req->r_unsafe_callback)
		req->r_unsafe_callback(req, false);
S
Sage Weil 已提交
1434 1435 1436
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1437 1438 1439 1440
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1441 1442
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1443
{
1444
	void *p, *end;
S
Sage Weil 已提交
1445 1446
	struct ceph_osd_request *req;
	u64 tid;
1447
	int object_len;
1448 1449
	unsigned int numops;
	int payload_len, flags;
1450
	s32 result;
1451 1452 1453 1454 1455 1456
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1457
	int already_completed;
1458
	u32 bytes;
1459
	unsigned int i;
S
Sage Weil 已提交
1460

1461
	tid = le64_to_cpu(msg->hdr.tid);
1462 1463 1464 1465 1466 1467 1468 1469 1470 1471
	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;

1472
	err = ceph_decode_pgid(&p, end, &pg);
1473
	if (err)
S
Sage Weil 已提交
1474
		goto bad;
1475 1476 1477 1478 1479 1480 1481 1482

	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
1483 1484 1485 1486 1487
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1488
		goto bad_mutex;
S
Sage Weil 已提交
1489 1490
	}
	ceph_osdc_get_request(req);
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512

	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);
	}
1513 1514
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1515
		pr_warning("sum of op payload lens %d != data_len %d",
1516
			   payload_len, bytes);
1517 1518 1519 1520 1521 1522 1523
		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
1524 1525 1526

	if (!req->r_got_reply) {

1527
		req->r_result = result;
S
Sage Weil 已提交
1528 1529 1530 1531 1532 1533
		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, */
1534 1535
		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
1536 1537 1538 1539

		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
1540
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1541 1542 1543 1544 1545
		goto done;
	}

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

1546 1547 1548
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1549
	/* either this is a read, or we got the safe response */
1550 1551
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1552 1553 1554
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1555 1556
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1557
	mutex_unlock(&osdc->request_mutex);
1558 1559
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1560 1561 1562 1563

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1564
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1565

S
Sage Weil 已提交
1566 1567
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1568 1569

done:
1570
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1571 1572 1573
	ceph_osdc_put_request(req);
	return;

1574 1575
bad_put:
	ceph_osdc_put_request(req);
1576 1577
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1578
bad:
1579 1580
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1581
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1582 1583
}

1584
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1585 1586 1587
{
	struct rb_node *p, *n;

1588 1589
	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
1590

1591 1592 1593 1594 1595 1596 1597
		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
1598
	}
1599 1600 1601
}

/*
1602 1603
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1604
 *
1605
 * Caller should hold map_sem for read.
1606
 */
1607
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1608
{
1609
	struct ceph_osd_request *req, *nreq;
1610 1611 1612
	struct rb_node *p;
	int needmap = 0;
	int err;
1613

1614
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1615
	mutex_lock(&osdc->request_mutex);
1616
	for (p = rb_first(&osdc->requests); p; ) {
1617
		req = rb_entry(p, struct ceph_osd_request, r_node);
1618
		p = rb_next(p);
1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636

		/*
		 * 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;
		}

1637
		err = __map_request(osdc, req, force_resend);
1638 1639 1640 1641 1642 1643
		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) {
1644 1645 1646 1647
			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);
1648
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1649 1650
			}
		}
1651 1652 1653 1654 1655 1656
	}

	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);

1657
		err = __map_request(osdc, req, force_resend);
1658
		dout("__map_request returned %d\n", err);
1659 1660 1661 1662 1663 1664 1665 1666
		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;
1667
		}
1668 1669 1670 1671

		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);
1672
		__unregister_linger_request(osdc, req);
1673
	}
S
Sage Weil 已提交
1674 1675 1676 1677 1678 1679
	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);
	}
1680
	reset_changed_osds(osdc);
1681
}
1682 1683


S
Sage Weil 已提交
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706
/*
 * 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));
1707 1708
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1709 1710 1711 1712 1713 1714 1715 1716

	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);
1717 1718
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1719 1720 1721 1722 1723 1724 1725
		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,
1726
							  &osdc->client->msgr);
S
Sage Weil 已提交
1727 1728 1729 1730
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1731
			BUG_ON(!newmap);
S
Sage Weil 已提交
1732 1733 1734 1735
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1736
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
		} 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);
1752 1753
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1754 1755 1756 1757 1758 1759 1760 1761 1762
		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 {
1763 1764
			int skipped_map = 0;

S
Sage Weil 已提交
1765 1766 1767 1768 1769 1770
			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;
			}
1771
			BUG_ON(!newmap);
S
Sage Weil 已提交
1772 1773
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1774 1775 1776
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1777
				ceph_osdmap_destroy(oldmap);
1778 1779
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1780 1781 1782 1783 1784 1785 1786 1787
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1788 1789 1790 1791 1792 1793 1794 1795 1796

	/*
	 * 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
1797 1798 1799
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1800
	up_read(&osdc->map_sem);
1801
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1802 1803 1804 1805
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1806
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1807 1808 1809 1810
	up_write(&osdc->map_sem);
	return;
}

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 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892
/*
 * 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 *),
1893
			   void *data, struct ceph_osd_event **pevent)
1894 1895 1896 1897 1898 1899 1900 1901 1902
{
	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;
1903
	event->one_shot = 0;
1904 1905 1906
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1907
	RB_CLEAR_NODE(&event->node);
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
	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
 */
1954 1955
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
{
	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) {
1976
		BUG_ON(event->one_shot);
1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
		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;
		}
1988
		INIT_WORK(&event_work->work, do_event_work);
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
		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;
}

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
				struct ceph_snap_context *snapc, u64 snap_id,
				struct timespec *mtime)
{
	struct ceph_msg *msg = req->r_request;
	void *p;
	size_t msg_size;
	int flags = req->r_flags;
	u64 data_len;
	unsigned int i;

	req->r_snapid = snap_id;
	req->r_snapc = ceph_get_snap_context(snapc);

	/* 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);
	memcpy(p, req->r_oid, req->r_oid_len);
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
	p += req->r_oid_len;

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

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

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

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

		/*
		 * The header "data_off" is a hint to the receiver
		 * allowing it to align received data into its
		 * buffers such that there's no need to re-copy
		 * it before writing it to disk (direct I/O).
		 */
		data_off = (u16) (off & 0xffff);
		req->r_request->hdr.data_off = cpu_to_le16(data_off);
	}
	req->r_request->hdr.data_len = cpu_to_le32(data_len);

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

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

2108 2109 2110 2111 2112 2113 2114 2115 2116
/*
 * 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;

S
Sage Weil 已提交
2117 2118
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2119
	__register_request(osdc, req);
2120 2121 2122 2123 2124 2125 2126
	WARN_ON(req->r_sent);
	rc = __map_request(osdc, req, 0);
	if (rc < 0) {
		if (nofail) {
			dout("osdc_start_request failed map, "
				" will retry %lld\n", req->r_tid);
			rc = 0;
S
Sage Weil 已提交
2127
		}
2128
		goto out_unlock;
S
Sage Weil 已提交
2129
	}
2130 2131 2132 2133
	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 {
2134
		__send_queued(osdc);
2135 2136
	}
	rc = 0;
2137
out_unlock:
S
Sage Weil 已提交
2138 2139 2140 2141
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
2142
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155

/*
 * 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);
2156
		__unregister_request(osdc, req);
S
Sage Weil 已提交
2157
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2158
		complete_request(req);
2159
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
2160 2161 2162 2163 2164 2165
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
2166
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199

/*
 * 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);
}
2200
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217

/*
 * 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;
2218
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
2219
	osdc->requests = RB_ROOT;
2220
	INIT_LIST_HEAD(&osdc->req_lru);
2221 2222
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2223
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2224 2225
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2226
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2227 2228 2229
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2230 2231

	schedule_delayed_work(&osdc->osds_timeout_work,
2232
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
2233

2234
	err = -ENOMEM;
S
Sage Weil 已提交
2235 2236 2237
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
2238
		goto out;
S
Sage Weil 已提交
2239

2240 2241
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
2242
				"osd_op");
S
Sage Weil 已提交
2243
	if (err < 0)
2244
		goto out_mempool;
2245
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2246 2247
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
2248 2249
	if (err < 0)
		goto out_msgpool;
2250 2251 2252 2253 2254 2255 2256

	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
2257
	return 0;
2258

S
Sage Weil 已提交
2259 2260
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2261 2262 2263 2264
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2265 2266 2267 2268
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2269 2270
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2271
	cancel_delayed_work_sync(&osdc->timeout_work);
2272
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
2273 2274 2275 2276
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
2277
	remove_all_osds(osdc);
S
Sage Weil 已提交
2278 2279
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2280
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2281 2282 2283 2284 2285 2286 2287 2288 2289 2290
}

/*
 * 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,
2291
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2292 2293 2294 2295 2296 2297
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2298
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1,
S
Sage Weil 已提交
2299
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2300
				    NULL, truncate_seq, truncate_size,
2301
				    false);
2302 2303
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2304 2305

	/* it may be a short read due to an object boundary */
2306

2307
	osd_req_op_extent_osd_data_pages(req, 0,
2308
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2309

2310
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2311
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2312

2313
	ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2314

S
Sage Weil 已提交
2315 2316 2317 2318 2319 2320 2321 2322
	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;
}
2323
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2324 2325 2326 2327 2328 2329 2330 2331 2332 2333

/*
 * 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,
2334
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2335 2336 2337
{
	struct ceph_osd_request *req;
	int rc = 0;
2338
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2339

2340
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
2341
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1,
S
Sage Weil 已提交
2342
				    CEPH_OSD_OP_WRITE,
2343
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2344
				    snapc, truncate_seq, truncate_size,
2345
				    true);
2346 2347
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2348 2349

	/* it may be a short write due to an object boundary */
2350
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2351 2352
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2353

2354
	ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2355

2356
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2357 2358 2359 2360 2361 2362 2363 2364 2365
	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;
}
2366
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2367 2368 2369 2370 2371 2372 2373

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2374
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2375 2376 2377
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2378
		goto out;
2379
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2380 2381 2382 2383 2384 2385

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2386
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2387
		break;
2388 2389 2390
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2391 2392 2393 2394 2395

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2396
out:
S
Sage Weil 已提交
2397 2398 2399
	ceph_msg_put(msg);
}

2400
/*
2401 2402
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2403 2404
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2405 2406
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2407 2408 2409
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2410
	struct ceph_msg *m;
2411
	struct ceph_osd_request *req;
2412 2413
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2414
	u64 tid;
S
Sage Weil 已提交
2415

2416 2417 2418 2419 2420 2421
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2422 2423
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2424 2425
		goto out;
	}
S
Sage Weil 已提交
2426

2427
	if (req->r_reply->con)
2428
		dout("%s revoking msg %p from old con %p\n", __func__,
2429 2430
		     req->r_reply, req->r_reply->con);
	ceph_msg_revoke_incoming(req->r_reply);
2431

S
Sage Weil 已提交
2432 2433 2434
	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);
2435
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2436
		if (!m)
S
Sage Weil 已提交
2437 2438 2439 2440 2441 2442
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2443
	if (data_len > 0) {
2444
		struct ceph_osd_data *osd_data;
2445

2446 2447 2448 2449 2450
		/*
		 * XXX This is assuming there is only one op containing
		 * XXX page data.  Probably OK for reads, but this
		 * XXX ought to be done more generally.
		 */
2451
		osd_data = osd_req_op_extent_osd_data(req, 0);
2452 2453
		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2454
				unlikely(osd_data->length < data_len)) {
2455

2456 2457 2458
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2459 2460 2461 2462 2463 2464
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2465
	}
2466
	*skip = 0;
S
Sage Weil 已提交
2467
	dout("get_reply tid %lld %p\n", tid, m);
2468 2469 2470

out:
	mutex_unlock(&osdc->request_mutex);
2471
	return m;
2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482

}

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
2483
	*skip = 0;
2484 2485
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2486
	case CEPH_MSG_WATCH_NOTIFY:
2487
		return ceph_msg_new(type, front, GFP_NOFS, false);
2488 2489 2490 2491 2492 2493 2494 2495
	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
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514
}

/*
 * 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);
}

2515 2516 2517
/*
 * authentication
 */
2518 2519 2520 2521 2522
/*
 * 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,
2523
					int *proto, int force_new)
2524 2525 2526 2527
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2528
	struct ceph_auth_handshake *auth = &o->o_auth;
2529

2530
	if (force_new && auth->authorizer) {
2531
		ceph_auth_destroy_authorizer(ac, auth->authorizer);
2532 2533
		auth->authorizer = NULL;
	}
2534 2535 2536
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2537
		if (ret)
2538
			return ERR_PTR(ret);
2539 2540
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2541 2542 2543
						     auth);
		if (ret)
			return ERR_PTR(ret);
2544 2545
	}
	*proto = ac->protocol;
2546

2547
	return auth;
2548 2549 2550 2551 2552 2553 2554 2555 2556
}


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;

2557
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2558 2559
}

2560 2561 2562 2563 2564 2565
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;

2566
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2567 2568
	return ceph_monc_validate_auth(&osdc->client->monc);
}
2569

2570
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2571 2572 2573
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2574 2575
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2576
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2577
	.alloc_msg = alloc_msg,
2578
	.fault = osd_reset,
S
Sage Weil 已提交
2579
};