osd_client.c 74.1 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 26
static struct kmem_cache	*ceph_osd_request_cache;

27
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
28

A
Alex Elder 已提交
29
static void __send_queued(struct ceph_osd_client *osdc);
30
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
31 32 33 34
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);
35 36
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

/*
 * 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.
 */
63
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
64
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
65
{
66
	u64 orig_len = *plen;
S
Sage Weil 已提交
67
	int r;
S
Sage Weil 已提交
68

69
	/* object extent? */
70 71
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
72 73
	if (r < 0)
		return r;
74 75
	if (*objlen < orig_len) {
		*plen = *objlen;
76 77 78 79
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

80
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
81

82
	return 0;
S
Sage Weil 已提交
83 84
}

A
Alex Elder 已提交
85 86 87 88 89 90
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;
}

91
static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
92 93 94 95 96 97 98 99 100 101 102
			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;
}

103
static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
104 105 106 107 108 109 110
			struct ceph_pagelist *pagelist)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
	osd_data->pagelist = pagelist;
}

#ifdef CONFIG_BLOCK
111
static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
112 113 114 115 116 117 118 119
			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 */

120 121 122 123 124 125
#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
126 127 128 129 130 131 132 133
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;
}

134 135
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
136
			unsigned int which)
137
{
138
	return osd_req_op_data(osd_req, which, extent, osd_data);
139 140 141 142 143 144 145
}
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)
{
146
	return osd_req_op_data(osd_req, which, cls, response_data);
147 148 149
}
EXPORT_SYMBOL(osd_req_op_cls_response_data);	/* ??? */

A
Alex Elder 已提交
150 151 152 153 154 155 156 157 158 159 160 161 162
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);

163
void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
164 165
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
166 167 168 169
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

170
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
171 172 173 174 175 176
	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,
177
			unsigned int which, struct ceph_pagelist *pagelist)
178 179 180
{
	struct ceph_osd_data *osd_data;

181
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
182 183 184 185 186 187
	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,
188
			unsigned int which, struct bio *bio, size_t bio_length)
189 190
{
	struct ceph_osd_data *osd_data;
191 192

	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
193 194 195 196 197 198 199 200 201 202 203
	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;

204
	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
205 206 207
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

208 209 210 211 212 213
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;

214
	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
215 216 217 218
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);

219 220 221 222 223 224 225 226 227 228 229 230
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);

231 232 233 234 235 236
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;

237
	osd_data = osd_req_op_data(osd_req, which, cls, response_data);
238 239 240 241 242
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);

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

S
Sage Weil 已提交
297 298 299
/*
 * requests
 */
S
Sage Weil 已提交
300
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
301
{
A
Alex Elder 已提交
302
	struct ceph_osd_request *req;
303
	unsigned int which;
S
Sage Weil 已提交
304

A
Alex Elder 已提交
305
	req = container_of(kref, struct ceph_osd_request, r_kref);
S
Sage Weil 已提交
306 307
	if (req->r_request)
		ceph_msg_put(req->r_request);
308
	if (req->r_reply) {
309
		ceph_msg_revoke_incoming(req->r_reply);
310
		ceph_msg_put(req->r_reply);
311
	}
312

313 314
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
315

S
Sage Weil 已提交
316 317 318 319
	ceph_put_snap_context(req->r_snapc);
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
320 321
		kmem_cache_free(ceph_osd_request_cache, req);

S
Sage Weil 已提交
322
}
323
EXPORT_SYMBOL(ceph_osdc_release_request);
324

325
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
326
					       struct ceph_snap_context *snapc,
327
					       unsigned int num_ops,
328
					       bool use_mempool,
329
					       gfp_t gfp_flags)
S
Sage Weil 已提交
330 331 332
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
333 334
	size_t msg_size;

335 336 337
	BUILD_BUG_ON(CEPH_OSD_MAX_OP > U16_MAX);
	BUG_ON(num_ops > CEPH_OSD_MAX_OP);

338 339 340
	msg_size = 4 + 4 + 8 + 8 + 4+8;
	msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
	msg_size += 1 + 8 + 4 + 4;     /* pg_t */
341
	msg_size += 4 + CEPH_MAX_OID_NAME_LEN; /* oid */
342 343 344 345 346
	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
347 348

	if (use_mempool) {
349
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
350 351
		memset(req, 0, sizeof(*req));
	} else {
352
		req = kmem_cache_zalloc(ceph_osd_request_cache, gfp_flags);
S
Sage Weil 已提交
353 354
	}
	if (req == NULL)
355
		return NULL;
S
Sage Weil 已提交
356 357 358

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
359
	req->r_num_ops = num_ops;
360

S
Sage Weil 已提交
361
	kref_init(&req->r_kref);
S
Sage Weil 已提交
362 363
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
364
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
365
	INIT_LIST_HEAD(&req->r_unsafe_item);
366 367
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
368
	INIT_LIST_HEAD(&req->r_req_lru_item);
369 370
	INIT_LIST_HEAD(&req->r_osd_item);

371
	req->r_base_oloc.pool = -1;
372
	req->r_target_oloc.pool = -1;
373

S
Sage Weil 已提交
374 375 376 377 378
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
379
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
380
	if (!msg) {
S
Sage Weil 已提交
381
		ceph_osdc_put_request(req);
382
		return NULL;
S
Sage Weil 已提交
383 384 385 386
	}
	req->r_reply = msg;

	/* create request message; allow space for oid */
S
Sage Weil 已提交
387
	if (use_mempool)
388
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
389
	else
390
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
391
	if (!msg) {
S
Sage Weil 已提交
392
		ceph_osdc_put_request(req);
393
		return NULL;
S
Sage Weil 已提交
394
	}
395

S
Sage Weil 已提交
396
	memset(msg->front.iov_base, 0, msg->front.iov_len);
397 398 399 400 401

	req->r_request = msg;

	return req;
}
402
EXPORT_SYMBOL(ceph_osdc_alloc_request);
403

404
static bool osd_req_opcode_valid(u16 opcode)
405
{
406
	switch (opcode) {
407
	case CEPH_OSD_OP_READ:
408
	case CEPH_OSD_OP_STAT:
409 410 411
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
412
	case CEPH_OSD_OP_NOTIFY:
413
	case CEPH_OSD_OP_NOTIFY_ACK:
414
	case CEPH_OSD_OP_ASSERT_VER:
415
	case CEPH_OSD_OP_WRITE:
416 417 418 419 420
	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:
421
	case CEPH_OSD_OP_STARTSYNC:
422 423 424 425 426 427
	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:
428
	case CEPH_OSD_OP_ROLLBACK:
429
	case CEPH_OSD_OP_WATCH:
430 431 432 433 434 435 436 437 438 439 440 441
	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:
442
	case CEPH_OSD_OP_GETXATTR:
443
	case CEPH_OSD_OP_GETXATTRS:
444 445
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
	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:
464
	case CEPH_OSD_OP_CALL:
465 466
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
467 468 469 470 471 472
		return true;
	default:
		return false;
	}
}

473 474 475 476 477
/*
 * 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.
 */
478
static struct ceph_osd_req_op *
A
Alex Elder 已提交
479
_osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
480
				u16 opcode)
481
{
482 483 484
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
485 486
	BUG_ON(!osd_req_opcode_valid(opcode));

487
	op = &osd_req->r_ops[which];
488 489
	memset(op, 0, sizeof (*op));
	op->op = opcode;
490 491

	return op;
492 493
}

A
Alex Elder 已提交
494 495 496 497 498 499 500
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);

501 502
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
503 504 505
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
A
Alex Elder 已提交
506
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
507 508
	size_t payload_len = 0;

L
Li Wang 已提交
509 510 511
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
	       opcode != CEPH_OSD_OP_DELETE && opcode != CEPH_OSD_OP_ZERO &&
	       opcode != CEPH_OSD_OP_TRUNCATE);
512 513 514 515 516 517 518 519 520 521 522 523

	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);

524 525
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
526
{
527 528 529 530 531 532
	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;
533 534 535 536 537 538 539 540 541 542

	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);

543
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
544
			u16 opcode, const char *class, const char *method)
545
{
A
Alex Elder 已提交
546
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
547
	struct ceph_pagelist *pagelist;
548 549 550 551 552
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

553 554 555 556
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

557 558 559 560
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
561
	ceph_pagelist_append(pagelist, class, size);
562 563 564 565 566 567
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
568
	ceph_pagelist_append(pagelist, method, size);
569 570
	payload_len += size;

571
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
572

573 574 575 576 577
	op->cls.argc = 0;	/* currently unused */

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_cls_init);
578

579 580
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
581 582
				u64 cookie, u64 version, int flag)
{
A
Alex Elder 已提交
583
	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
584

585
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
586 587

	op->watch.cookie = cookie;
A
Alex Elder 已提交
588
	op->watch.ver = version;
589
	if (opcode == CEPH_OSD_OP_WATCH && flag)
590
		op->watch.flag = (u8)1;
591 592 593
}
EXPORT_SYMBOL(osd_req_op_watch_init);

594
static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
595 596 597 598 599 600 601
				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)
602
			ceph_msg_data_add_pages(msg, osd_data->pages,
603 604 605
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
606
		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
607 608
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
609
		ceph_msg_data_add_bio(msg, osd_data->bio, length);
610 611 612 613 614 615
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

616
static u64 osd_req_encode_op(struct ceph_osd_request *req,
617
			      struct ceph_osd_op *dst, unsigned int which)
618
{
619
	struct ceph_osd_req_op *src;
620
	struct ceph_osd_data *osd_data;
621
	u64 request_data_len = 0;
622
	u64 data_length;
623

624 625
	BUG_ON(which >= req->r_num_ops);
	src = &req->r_ops[which];
626 627 628 629 630 631 632 633
	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
634 635
		osd_data = &src->raw_data_in;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
636 637 638
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
L
Li Wang 已提交
639 640 641
	case CEPH_OSD_OP_ZERO:
	case CEPH_OSD_OP_DELETE:
	case CEPH_OSD_OP_TRUNCATE:
642
		if (src->op == CEPH_OSD_OP_WRITE)
643
			request_data_len = src->extent.length;
644 645 646 647 648 649
		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);
650
		osd_data = &src->extent.osd_data;
651
		if (src->op == CEPH_OSD_OP_WRITE)
652
			ceph_osdc_msg_data_add(req->r_request, osd_data);
653
		else
654
			ceph_osdc_msg_data_add(req->r_reply, osd_data);
655 656 657 658
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674
		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);
675 676 677 678 679 680 681 682 683 684
		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:
685
		pr_err("unsupported osd opcode %s\n",
686
			ceph_osd_op_name(src->op));
687
		WARN_ON(1);
688 689

		return 0;
690
	}
691
	dst->op = cpu_to_le16(src->op);
692
	dst->payload_len = cpu_to_le32(src->payload_len);
693

694
	return request_data_len;
695 696
}

697 698 699 700 701 702 703 704 705 706 707 708 709 710
/*
 * 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,
711
					       u64 off, u64 *plen, int num_ops,
712 713 714 715
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
716
					       bool use_mempool)
717
{
718
	struct ceph_osd_request *req;
719 720 721
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
722 723
	u32 object_size;
	u64 object_base;
724
	int r;
725

L
Li Wang 已提交
726 727 728
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
	       opcode != CEPH_OSD_OP_DELETE && opcode != CEPH_OSD_OP_ZERO &&
	       opcode != CEPH_OSD_OP_TRUNCATE);
729

730
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
731
					GFP_NOFS);
732
	if (!req)
733
		return ERR_PTR(-ENOMEM);
734

735
	req->r_flags = flags;
736 737

	/* calculate max write size */
738
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
739 740
	if (r < 0) {
		ceph_osdc_put_request(req);
741
		return ERR_PTR(r);
742
	}
743

744 745
	object_size = le32_to_cpu(layout->fl_object_size);
	object_base = off - objoff;
746 747 748 749 750 751 752 753
	if (!(truncate_seq == 1 && truncate_size == -1ULL)) {
		if (truncate_size <= object_base) {
			truncate_size = 0;
		} else {
			truncate_size -= object_base;
			if (truncate_size > object_size)
				truncate_size = object_size;
		}
754
	}
755

756
	osd_req_op_extent_init(req, 0, opcode, objoff, objlen,
757
				truncate_size, truncate_seq);
758

759 760 761 762 763 764
	/*
	 * 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)
765
		osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC);
766

767
	req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout);
768

769
	snprintf(req->r_base_oid.name, sizeof(req->r_base_oid.name),
770
		 "%llx.%08llx", vino.ino, objnum);
771
	req->r_base_oid.name_len = strlen(req->r_base_oid.name);
772

S
Sage Weil 已提交
773 774
	return req;
}
775
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841

/*
 * 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;
}

842 843 844 845 846 847
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
848
	struct ceph_osd_request *req, *nreq;
849
	LIST_HEAD(resend);
850 851 852 853
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
854
	if (err)
855
		return;
856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
	/*
	 * 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.
	 */
873
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
874 875 876 877
		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,
878
		     osd->o_osd);
879 880 881
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}
882
	list_splice(&resend, &osdc->req_unsent);
883

884 885 886 887 888
	/*
	 * 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.
	 */
889 890
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
891 892 893 894 895
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
896
		__register_request(osdc, req);
897
		list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
898
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
899
		__unregister_linger_request(osdc, req);
900 901
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
902 903 904
	}
}

S
Sage Weil 已提交
905
/*
906
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
907 908 909 910 911 912 913 914 915 916 917
 */
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
918 919
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
920
	__send_queued(osdc);
S
Sage Weil 已提交
921
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
922 923 924 925 926 927
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
928
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
929 930 931 932 933 934 935 936 937
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
938
	osd->o_osd = onum;
939
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
940
	INIT_LIST_HEAD(&osd->o_requests);
941
	INIT_LIST_HEAD(&osd->o_linger_requests);
942
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
943 944
	osd->o_incarnation = 1;

945
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
946

947
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966
	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);
967
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
968 969
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

970
		ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
971
		kfree(osd);
S
Sage Weil 已提交
972
	}
S
Sage Weil 已提交
973 974 975 976 977
}

/*
 * remove an osd from our map
 */
978
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
979
{
980
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
981 982
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
983
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
984 985 986 987
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

988 989
static void remove_all_osds(struct ceph_osd_client *osdc)
{
990
	dout("%s %p\n", __func__, osdc);
991 992 993 994 995 996 997 998 999
	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);
}

1000 1001 1002 1003 1004 1005
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);
1006
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
1007 1008 1009 1010 1011 1012 1013 1014 1015
}

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);
}

1016
static void remove_old_osds(struct ceph_osd_client *osdc)
1017 1018 1019 1020 1021 1022
{
	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) {
1023
		if (time_before(jiffies, osd->lru_ttl))
1024 1025 1026 1027 1028 1029
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
1030 1031 1032
/*
 * reset osd connect
 */
1033
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
1034
{
A
Alex Elder 已提交
1035
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
1036

1037
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
1038 1039
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
1040
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
1041 1042 1043 1044 1045 1046 1047 1048 1049

		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;

1050 1051
		dout("osd addr hasn't changed and connection never opened, "
		     "letting msgr retry\n");
1052 1053 1054
		/* 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
1055 1056

		return -EAGAIN;
S
Sage Weil 已提交
1057
	}
A
Alex Elder 已提交
1058 1059 1060 1061 1062 1063

	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
1064 1065 1066 1067 1068 1069 1070 1071
}

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;

1072
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
	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;
}

1105 1106 1107
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
1108
			osdc->client->options->osd_keepalive_timeout * HZ);
1109 1110 1111 1112 1113 1114
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
1115 1116 1117 1118 1119

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1120 1121
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1122 1123
{
	req->r_tid = ++osdc->last_tid;
1124
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1125
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
1126 1127 1128 1129
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
1130 1131
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
1132
	}
1133 1134
}

S
Sage Weil 已提交
1135 1136 1137 1138 1139 1140
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1141 1142 1143 1144 1145 1146
	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
1147 1148 1149 1150
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

1151 1152
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1153
		ceph_msg_revoke(req->r_request);
1154 1155

		list_del_init(&req->r_osd_item);
1156 1157 1158
		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);
1159
			__move_osd_to_lru(osdc, req->r_osd);
1160
		}
1161
		if (list_empty(&req->r_linger_item))
1162
			req->r_osd = NULL;
1163
	}
S
Sage Weil 已提交
1164

A
Alex Elder 已提交
1165
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1166 1167
	ceph_osdc_put_request(req);

1168 1169 1170
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1171 1172 1173 1174 1175 1176 1177 1178
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1179
	if (req->r_sent && req->r_osd) {
1180
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1181 1182 1183 1184
		req->r_sent = 0;
	}
}

1185 1186 1187 1188
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
	dout("__register_linger_request %p\n", req);
1189
	ceph_osdc_get_request(req);
1190
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1191 1192 1193
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
1194 1195 1196 1197 1198 1199
}

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
1200
	list_del_init(&req->r_linger_item);
1201 1202 1203 1204 1205 1206 1207 1208
	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);
		}
1209 1210
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1211
	}
1212
	ceph_osdc_put_request(req);
1213 1214 1215 1216 1217 1218 1219
}

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) {
1220
		req->r_linger = 0;
1221
		__unregister_linger_request(osdc, req);
1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
	}
	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;
	}
}
EXPORT_SYMBOL(ceph_osdc_set_request_linger);

1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
/*
 * Returns whether a request should be blocked from being sent
 * based on the current osdmap and osd_client settings.
 *
 * Caller should hold map_sem for read.
 */
static bool __req_should_be_paused(struct ceph_osd_client *osdc,
				   struct ceph_osd_request *req)
{
	bool pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD);
	bool pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) ||
		ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);
	return (req->r_flags & CEPH_OSD_FLAG_READ && pauserd) ||
		(req->r_flags & CEPH_OSD_FLAG_WRITE && pausewr);
}

1253 1254 1255 1256 1257 1258 1259
/*
 * Calculate mapping of a request to a PG.  Takes tiering into account.
 */
static int __calc_request_pg(struct ceph_osdmap *osdmap,
			     struct ceph_osd_request *req,
			     struct ceph_pg *pg_out)
{
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
	bool need_check_tiering;

	need_check_tiering = false;
	if (req->r_target_oloc.pool == -1) {
		req->r_target_oloc = req->r_base_oloc; /* struct */
		need_check_tiering = true;
	}
	if (req->r_target_oid.name_len == 0) {
		ceph_oid_copy(&req->r_target_oid, &req->r_base_oid);
		need_check_tiering = true;
	}

	if (need_check_tiering &&
	    (req->r_flags & CEPH_OSD_FLAG_IGNORE_OVERLAY) == 0) {
1274 1275
		struct ceph_pg_pool_info *pi;

1276
		pi = ceph_pg_pool_by_id(osdmap, req->r_target_oloc.pool);
1277 1278 1279
		if (pi) {
			if ((req->r_flags & CEPH_OSD_FLAG_READ) &&
			    pi->read_tier >= 0)
1280
				req->r_target_oloc.pool = pi->read_tier;
1281 1282
			if ((req->r_flags & CEPH_OSD_FLAG_WRITE) &&
			    pi->write_tier >= 0)
1283
				req->r_target_oloc.pool = pi->write_tier;
1284 1285 1286 1287
		}
		/* !pi is caught in ceph_oloc_oid_to_pg() */
	}

1288 1289
	return ceph_oloc_oid_to_pg(osdmap, &req->r_target_oloc,
				   &req->r_target_oid, pg_out);
1290 1291
}

S
Sage Weil 已提交
1292 1293 1294
/*
 * 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
1295
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1296
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1297 1298 1299 1300 1301
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1302
static int __map_request(struct ceph_osd_client *osdc,
1303
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1304
{
1305
	struct ceph_pg pgid;
1306 1307
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
1308
	int err;
1309
	bool was_paused;
S
Sage Weil 已提交
1310

1311
	dout("map_request %p tid %lld\n", req, req->r_tid);
1312 1313

	err = __calc_request_pg(osdc->osdmap, req, &pgid);
1314 1315
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1316
		return err;
1317
	}
1318 1319
	req->r_pgid = pgid;

1320 1321 1322 1323 1324
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
1325

1326 1327 1328 1329 1330
	was_paused = req->r_paused;
	req->r_paused = __req_should_be_paused(osdc, req);
	if (was_paused && !req->r_paused)
		force_resend = 1;

1331 1332
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1333 1334 1335
	     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) ||
1336 1337
	    (req->r_osd == NULL && o == -1) ||
	    req->r_paused)
S
Sage Weil 已提交
1338 1339
		return 0;  /* no change */

1340 1341
	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
1342 1343
	     req->r_osd ? req->r_osd->o_osd : -1);

1344 1345 1346 1347
	/* 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
1348 1349 1350 1351 1352 1353 1354 1355
	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) {
1356
		err = -ENOMEM;
1357
		req->r_osd = create_osd(osdc, o);
1358 1359
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1360
			goto out;
1361
		}
S
Sage Weil 已提交
1362

1363
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1364 1365
		__insert_osd(osdc, req->r_osd);

1366 1367 1368
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1369 1370
	}

1371 1372
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
1373 1374
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1375
	} else {
1376
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1377
	}
1378
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1379 1380 1381 1382 1383 1384 1385 1386

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1387 1388
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1389
{
1390
	void *p;
S
Sage Weil 已提交
1391

1392 1393 1394 1395 1396 1397 1398
	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);
1399
	put_unaligned_le64(req->r_target_oloc.pool, req->r_request_pool);
1400 1401 1402 1403 1404 1405
	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));
1406

S
Sage Weil 已提交
1407
	req->r_stamp = jiffies;
1408
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1409 1410

	ceph_msg_get(req->r_request); /* send consumes a ref */
1411

S
Sage Weil 已提交
1412
	req->r_sent = req->r_osd->o_incarnation;
1413 1414

	ceph_con_send(&req->r_osd->o_con, req->r_request);
S
Sage Weil 已提交
1415 1416
}

1417 1418 1419
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1420
static void __send_queued(struct ceph_osd_client *osdc)
1421 1422 1423
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1424 1425
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1426 1427 1428
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
/*
 * 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
1442
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1443
	struct ceph_osd *osd;
1444
	unsigned long keepalive =
1445
		osdc->client->options->osd_keepalive_timeout * HZ;
1446
	struct list_head slow_osds;
S
Sage Weil 已提交
1447 1448 1449 1450 1451 1452 1453
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1454 1455 1456 1457 1458 1459 1460
	/*
	 * 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
1461
		if (time_before(jiffies, req->r_stamp + keepalive))
1462 1463 1464 1465 1466
			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
1467
		     req->r_tid, osd->o_osd);
1468 1469 1470 1471 1472 1473
		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
1474 1475 1476
		ceph_con_keepalive(&osd->o_con);
	}

1477
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1478
	__send_queued(osdc);
S
Sage Weil 已提交
1479 1480 1481 1482
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1483 1484 1485 1486 1487 1488
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 =
1489
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1490 1491 1492

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1493
	remove_old_osds(osdc);
1494 1495 1496 1497 1498 1499
	up_read(&osdc->map_sem);

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

1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 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
static int ceph_oloc_decode(void **p, void *end,
			    struct ceph_object_locator *oloc)
{
	u8 struct_v, struct_cv;
	u32 len;
	void *struct_end;
	int ret = 0;

	ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
	struct_v = ceph_decode_8(p);
	struct_cv = ceph_decode_8(p);
	if (struct_v < 3) {
		pr_warn("got v %d < 3 cv %d of ceph_object_locator\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	if (struct_cv > 6) {
		pr_warn("got v %d cv %d > 6 of ceph_object_locator\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	len = ceph_decode_32(p);
	ceph_decode_need(p, end, len, e_inval);
	struct_end = *p + len;

	oloc->pool = ceph_decode_64(p);
	*p += 4; /* skip preferred */

	len = ceph_decode_32(p);
	if (len > 0) {
		pr_warn("ceph_object_locator::key is set\n");
		goto e_inval;
	}

	if (struct_v >= 5) {
		len = ceph_decode_32(p);
		if (len > 0) {
			pr_warn("ceph_object_locator::nspace is set\n");
			goto e_inval;
		}
	}

	if (struct_v >= 6) {
		s64 hash = ceph_decode_64(p);
		if (hash != -1) {
			pr_warn("ceph_object_locator::hash is set\n");
			goto e_inval;
		}
	}

	/* skip the rest */
	*p = struct_end;
out:
	return ret;

e_inval:
	ret = -EINVAL;
	goto out;
}

static int ceph_redirect_decode(void **p, void *end,
				struct ceph_request_redirect *redir)
{
	u8 struct_v, struct_cv;
	u32 len;
	void *struct_end;
	int ret;

	ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
	struct_v = ceph_decode_8(p);
	struct_cv = ceph_decode_8(p);
	if (struct_cv > 1) {
		pr_warn("got v %d cv %d > 1 of ceph_request_redirect\n",
			struct_v, struct_cv);
		goto e_inval;
	}
	len = ceph_decode_32(p);
	ceph_decode_need(p, end, len, e_inval);
	struct_end = *p + len;

	ret = ceph_oloc_decode(p, end, &redir->oloc);
	if (ret)
		goto out;

	len = ceph_decode_32(p);
	if (len > 0) {
		pr_warn("ceph_request_redirect::object_name is set\n");
		goto e_inval;
	}

	len = ceph_decode_32(p);
	*p += len; /* skip osd_instructions */

	/* skip the rest */
	*p = struct_end;
out:
	return ret;

e_inval:
	ret = -EINVAL;
	goto out;
}

S
Sage Weil 已提交
1603 1604 1605 1606 1607
static void complete_request(struct ceph_osd_request *req)
{
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1608 1609 1610 1611
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1612 1613
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1614
{
1615
	void *p, *end;
S
Sage Weil 已提交
1616
	struct ceph_osd_request *req;
1617
	struct ceph_request_redirect redir;
S
Sage Weil 已提交
1618
	u64 tid;
1619
	int object_len;
1620 1621
	unsigned int numops;
	int payload_len, flags;
1622
	s32 result;
1623 1624 1625 1626 1627 1628
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1629
	int already_completed;
1630
	u32 bytes;
1631
	unsigned int i;
S
Sage Weil 已提交
1632

1633
	tid = le64_to_cpu(msg->hdr.tid);
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
	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;

1644
	err = ceph_decode_pgid(&p, end, &pg);
1645
	if (err)
S
Sage Weil 已提交
1646
		goto bad;
1647 1648 1649 1650 1651 1652 1653 1654

	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
1655 1656 1657 1658 1659
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1660
		goto bad_mutex;
S
Sage Weil 已提交
1661 1662
	}
	ceph_osdc_get_request(req);
1663 1664 1665 1666

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

1667
	ceph_decode_need(&p, end, 4, bad_put);
1668 1669 1670 1671 1672 1673
	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;
1674
	ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad_put);
1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
	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);
	}
1685 1686
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1687
		pr_warning("sum of op payload lens %d != data_len %d",
1688
			   payload_len, bytes);
1689 1690 1691
		goto bad_put;
	}

1692
	ceph_decode_need(&p, end, 4 + numops * 4, bad_put);
1693 1694 1695
	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
1696

1697 1698 1699
	if (le16_to_cpu(msg->hdr.version) >= 6) {
		p += 8 + 4; /* skip replay_version */
		p += 8; /* skip user_version */
Y
Yan, Zheng 已提交
1700

1701 1702 1703 1704 1705 1706
		err = ceph_redirect_decode(&p, end, &redir);
		if (err)
			goto bad_put;
	} else {
		redir.oloc.pool = -1;
	}
S
Sage Weil 已提交
1707

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731
	if (redir.oloc.pool != -1) {
		dout("redirect pool %lld\n", redir.oloc.pool);

		__unregister_request(osdc, req);
		mutex_unlock(&osdc->request_mutex);

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

		/*
		 * Start redirect requests with nofail=true.  If
		 * mapping fails, request will end up on the notarget
		 * list, waiting for the new osdmap (which can take
		 * a while), even though the original request mapped
		 * successfully.  In the future we might want to follow
		 * original request's nofail setting here.
		 */
		err = ceph_osdc_start_request(osdc, req, true);
		BUG_ON(err);

		goto done;
	}

	already_completed = req->r_got_reply;
	if (!req->r_got_reply) {
1732
		req->r_result = result;
S
Sage Weil 已提交
1733 1734 1735 1736 1737 1738
		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, */
1739 1740
		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
1741 1742 1743 1744

		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
1745
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1746 1747 1748 1749 1750
		goto done;
	}

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

1751 1752 1753
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1754
	/* either this is a read, or we got the safe response */
1755 1756
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1757 1758 1759 1760 1761
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

	mutex_unlock(&osdc->request_mutex);

Y
Yan, Zheng 已提交
1762
	if (!already_completed) {
1763 1764 1765
		if (req->r_unsafe_callback &&
		    result >= 0 && !(flags & CEPH_OSD_FLAG_ONDISK))
			req->r_unsafe_callback(req, true);
Y
Yan, Zheng 已提交
1766 1767 1768 1769 1770
		if (req->r_callback)
			req->r_callback(req, msg);
		else
			complete_all(&req->r_completion);
	}
S
Sage Weil 已提交
1771

1772 1773 1774
	if (flags & CEPH_OSD_FLAG_ONDISK) {
		if (req->r_unsafe_callback && already_completed)
			req->r_unsafe_callback(req, false);
S
Sage Weil 已提交
1775
		complete_request(req);
1776
	}
S
Sage Weil 已提交
1777 1778

done:
1779
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1780 1781 1782
	ceph_osdc_put_request(req);
	return;

1783
bad_put:
1784 1785 1786 1787 1788 1789 1790
	req->r_result = -EIO;
	__unregister_request(osdc, req);
	if (req->r_callback)
		req->r_callback(req, msg);
	else
		complete_all(&req->r_completion);
	complete_request(req);
1791
	ceph_osdc_put_request(req);
1792 1793
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1794
bad:
1795 1796
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1797
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1798 1799
}

1800
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1801 1802 1803
{
	struct rb_node *p, *n;

1804 1805
	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
1806

1807 1808 1809 1810 1811 1812 1813
		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
1814
	}
1815 1816 1817
}

/*
1818 1819
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1820
 *
1821
 * Caller should hold map_sem for read.
1822
 */
1823 1824
static void kick_requests(struct ceph_osd_client *osdc, bool force_resend,
			  bool force_resend_writes)
1825
{
1826
	struct ceph_osd_request *req, *nreq;
1827 1828 1829
	struct rb_node *p;
	int needmap = 0;
	int err;
1830
	bool force_resend_req;
1831

1832 1833
	dout("kick_requests %s %s\n", force_resend ? " (force resend)" : "",
		force_resend_writes ? " (force resend writes)" : "");
1834
	mutex_lock(&osdc->request_mutex);
1835
	for (p = rb_first(&osdc->requests); p; ) {
1836
		req = rb_entry(p, struct ceph_osd_request, r_node);
1837
		p = rb_next(p);
1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850

		/*
		 * 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);
1851
			ceph_osdc_get_request(req);
1852 1853
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
1854
			ceph_osdc_put_request(req);
1855 1856 1857
			continue;
		}

1858 1859 1860 1861
		force_resend_req = force_resend ||
			(force_resend_writes &&
				req->r_flags & CEPH_OSD_FLAG_WRITE);
		err = __map_request(osdc, req, force_resend_req);
1862 1863 1864 1865 1866 1867
		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) {
1868 1869 1870 1871
			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);
1872
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1873 1874
			}
		}
1875 1876 1877 1878 1879 1880
	}

	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);

1881 1882
		err = __map_request(osdc, req,
				    force_resend || force_resend_writes);
1883
		dout("__map_request returned %d\n", err);
1884 1885 1886 1887 1888 1889 1890 1891
		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;
1892
		}
1893 1894 1895 1896

		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);
1897
		__unregister_linger_request(osdc, req);
1898
	}
1899
	reset_changed_osds(osdc);
S
Sage Weil 已提交
1900 1901 1902 1903 1904 1905
	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);
	}
1906
}
1907 1908


S
Sage Weil 已提交
1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
/*
 * 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;
1924
	bool was_full;
S
Sage Weil 已提交
1925 1926 1927 1928 1929 1930 1931 1932

	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));
1933 1934
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1935 1936 1937

	down_write(&osdc->map_sem);

1938 1939
	was_full = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);

S
Sage Weil 已提交
1940 1941 1942 1943 1944
	/* 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);
1945 1946
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1947 1948 1949 1950 1951 1952 1953
		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,
1954
							  &osdc->client->msgr);
S
Sage Weil 已提交
1955 1956 1957 1958
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1959
			BUG_ON(!newmap);
S
Sage Weil 已提交
1960 1961 1962 1963
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1964 1965 1966 1967
			was_full = was_full ||
				ceph_osdmap_flag(osdc->osdmap,
						 CEPH_OSDMAP_FULL);
			kick_requests(osdc, 0, was_full);
S
Sage Weil 已提交
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
		} 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);
1983 1984
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1985 1986 1987 1988 1989 1990 1991 1992 1993
		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 {
1994 1995
			int skipped_map = 0;

S
Sage Weil 已提交
1996 1997 1998 1999 2000 2001
			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;
			}
2002
			BUG_ON(!newmap);
S
Sage Weil 已提交
2003 2004
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
2005 2006 2007
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
2008
				ceph_osdmap_destroy(oldmap);
2009
			}
2010 2011 2012 2013
			was_full = was_full ||
				ceph_osdmap_flag(osdc->osdmap,
						 CEPH_OSDMAP_FULL);
			kick_requests(osdc, skipped_map, was_full);
S
Sage Weil 已提交
2014 2015 2016 2017 2018
		}
		p += maplen;
		nr_maps--;
	}

2019 2020
	if (!osdc->osdmap)
		goto bad;
S
Sage Weil 已提交
2021 2022 2023
done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
2024 2025 2026 2027 2028 2029

	/*
	 * subscribe to subsequent osdmap updates if full to ensure
	 * we find out when we are no longer full and stop returning
	 * ENOSPC.
	 */
2030 2031 2032
	if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
		ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD) ||
		ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR))
2033 2034
		ceph_monc_request_next_osdmap(&osdc->client->monc);

A
Alex Elder 已提交
2035 2036 2037
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2038
	up_read(&osdc->map_sem);
2039
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
2040 2041 2042 2043
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
2044
	ceph_msg_dump(msg);
S
Sage Weil 已提交
2045 2046 2047 2048
	up_write(&osdc->map_sem);
	return;
}

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 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
/*
 * 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 *),
2131
			   void *data, struct ceph_osd_event **pevent)
2132 2133 2134 2135 2136 2137 2138 2139 2140
{
	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;
2141
	event->one_shot = 0;
2142 2143 2144
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
2145
	RB_CLEAR_NODE(&event->node);
2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 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
	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
 */
2192 2193
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213
{
	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) {
2214
		BUG_ON(event->one_shot);
2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225
		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;
		}
2226
		INIT_WORK(&event_work->work, do_event_work);
2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247
		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;
}

2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
/*
 * 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 */
2296 2297 2298 2299 2300
	ceph_encode_32(&p, req->r_base_oid.name_len);
	memcpy(p, req->r_base_oid.name, req->r_base_oid.name_len);
	dout("oid '%.*s' len %d\n", req->r_base_oid.name_len,
	     req->r_base_oid.name, req->r_base_oid.name_len);
	p += req->r_base_oid.name_len;
2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346

	/* 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);

2347 2348 2349 2350 2351 2352 2353 2354 2355
/*
 * 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
2356 2357
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2358
	__register_request(osdc, req);
2359 2360
	req->r_sent = 0;
	req->r_got_reply = 0;
2361 2362 2363 2364 2365 2366
	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;
2367 2368
		} else {
			__unregister_request(osdc, req);
S
Sage Weil 已提交
2369
		}
2370
		goto out_unlock;
S
Sage Weil 已提交
2371
	}
2372 2373 2374 2375
	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 {
2376
		__send_queued(osdc);
2377 2378
	}
	rc = 0;
2379
out_unlock:
S
Sage Weil 已提交
2380 2381 2382 2383
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
2384
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397

/*
 * 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);
2398
		__unregister_request(osdc, req);
S
Sage Weil 已提交
2399
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2400
		complete_request(req);
2401
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
2402 2403 2404 2405 2406 2407
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
2408
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441

/*
 * 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);
}
2442
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2443

2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
/*
 * Call all pending notify callbacks - for use after a watch is
 * unregistered, to make sure no more callbacks for it will be invoked
 */
extern void ceph_osdc_flush_notifies(struct ceph_osd_client *osdc)
{
	flush_workqueue(osdc->notify_wq);
}
EXPORT_SYMBOL(ceph_osdc_flush_notifies);


S
Sage Weil 已提交
2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470
/*
 * 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;
2471
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
2472
	osdc->requests = RB_ROOT;
2473
	INIT_LIST_HEAD(&osdc->req_lru);
2474 2475
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2476
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2477 2478
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2479
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2480 2481 2482
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2483 2484

	schedule_delayed_work(&osdc->osds_timeout_work,
2485
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
2486

2487
	err = -ENOMEM;
S
Sage Weil 已提交
2488 2489 2490
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
2491
		goto out;
S
Sage Weil 已提交
2492

2493 2494
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
2495
				"osd_op");
S
Sage Weil 已提交
2496
	if (err < 0)
2497
		goto out_mempool;
2498
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2499 2500
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
2501 2502
	if (err < 0)
		goto out_msgpool;
2503

2504
	err = -ENOMEM;
2505
	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
2506
	if (!osdc->notify_wq)
2507
		goto out_msgpool;
S
Sage Weil 已提交
2508
	return 0;
2509

S
Sage Weil 已提交
2510 2511
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2512 2513 2514 2515
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2516 2517 2518 2519
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2520 2521
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2522
	cancel_delayed_work_sync(&osdc->timeout_work);
2523
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
2524 2525 2526 2527
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
2528
	remove_all_osds(osdc);
S
Sage Weil 已提交
2529 2530
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2531
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2532 2533 2534 2535 2536 2537 2538 2539 2540 2541
}

/*
 * 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,
2542
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2543 2544 2545 2546 2547 2548
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2549
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1,
S
Sage Weil 已提交
2550
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2551
				    NULL, truncate_seq, truncate_size,
2552
				    false);
2553 2554
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2555 2556

	/* it may be a short read due to an object boundary */
2557

2558
	osd_req_op_extent_osd_data_pages(req, 0,
2559
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2560

2561
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2562
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2563

2564
	ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2565

S
Sage Weil 已提交
2566 2567 2568 2569 2570 2571 2572 2573
	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;
}
2574
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2575 2576 2577 2578 2579 2580 2581 2582 2583 2584

/*
 * 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,
2585
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2586 2587 2588
{
	struct ceph_osd_request *req;
	int rc = 0;
2589
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2590

2591
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
2592
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1,
S
Sage Weil 已提交
2593
				    CEPH_OSD_OP_WRITE,
2594
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2595
				    snapc, truncate_seq, truncate_size,
2596
				    true);
2597 2598
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2599 2600

	/* it may be a short write due to an object boundary */
2601
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2602 2603
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2604

2605
	ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2606

2607
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2608 2609 2610 2611 2612 2613 2614 2615 2616
	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;
}
2617
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2618

2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638
int ceph_osdc_setup(void)
{
	BUG_ON(ceph_osd_request_cache);
	ceph_osd_request_cache = kmem_cache_create("ceph_osd_request",
					sizeof (struct ceph_osd_request),
					__alignof__(struct ceph_osd_request),
					0, NULL);

	return ceph_osd_request_cache ? 0 : -ENOMEM;
}
EXPORT_SYMBOL(ceph_osdc_setup);

void ceph_osdc_cleanup(void)
{
	BUG_ON(!ceph_osd_request_cache);
	kmem_cache_destroy(ceph_osd_request_cache);
	ceph_osd_request_cache = NULL;
}
EXPORT_SYMBOL(ceph_osdc_cleanup);

S
Sage Weil 已提交
2639 2640 2641 2642 2643 2644
/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2645
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2646 2647 2648
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2649
		goto out;
2650
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2651 2652 2653 2654 2655 2656

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2657
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2658
		break;
2659 2660 2661
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2662 2663 2664 2665 2666

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2667
out:
S
Sage Weil 已提交
2668 2669 2670
	ceph_msg_put(msg);
}

2671
/*
2672 2673
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2674 2675
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2676 2677
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2678 2679 2680
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2681
	struct ceph_msg *m;
2682
	struct ceph_osd_request *req;
2683
	int front_len = le32_to_cpu(hdr->front_len);
2684
	int data_len = le32_to_cpu(hdr->data_len);
2685
	u64 tid;
S
Sage Weil 已提交
2686

2687 2688 2689 2690 2691 2692
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2693 2694
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2695 2696
		goto out;
	}
S
Sage Weil 已提交
2697

2698
	if (req->r_reply->con)
2699
		dout("%s revoking msg %p from old con %p\n", __func__,
2700 2701
		     req->r_reply, req->r_reply->con);
	ceph_msg_revoke_incoming(req->r_reply);
2702

2703
	if (front_len > req->r_reply->front_alloc_len) {
2704
		pr_warning("get_reply front %d > preallocated %d (%u#%llu)\n",
2705
			   front_len, req->r_reply->front_alloc_len,
2706 2707
			   (unsigned int)con->peer_name.type,
			   le64_to_cpu(con->peer_name.num));
2708 2709
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front_len, GFP_NOFS,
				 false);
2710
		if (!m)
S
Sage Weil 已提交
2711 2712 2713 2714 2715 2716
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2717
	if (data_len > 0) {
2718
		struct ceph_osd_data *osd_data;
2719

2720 2721 2722 2723 2724
		/*
		 * 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.
		 */
2725
		osd_data = osd_req_op_extent_osd_data(req, 0);
2726 2727
		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2728
				unlikely(osd_data->length < data_len)) {
2729

2730 2731 2732
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2733 2734 2735 2736 2737 2738
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2739
	}
2740
	*skip = 0;
S
Sage Weil 已提交
2741
	dout("get_reply tid %lld %p\n", tid, m);
2742 2743 2744

out:
	mutex_unlock(&osdc->request_mutex);
2745
	return m;
2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756

}

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
2757
	*skip = 0;
2758 2759
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2760
	case CEPH_MSG_WATCH_NOTIFY:
2761
		return ceph_msg_new(type, front, GFP_NOFS, false);
2762 2763 2764 2765 2766 2767 2768 2769
	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
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788
}

/*
 * 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);
}

2789 2790 2791
/*
 * authentication
 */
2792 2793 2794 2795 2796
/*
 * 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,
2797
					int *proto, int force_new)
2798 2799 2800 2801
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2802
	struct ceph_auth_handshake *auth = &o->o_auth;
2803

2804
	if (force_new && auth->authorizer) {
2805
		ceph_auth_destroy_authorizer(ac, auth->authorizer);
2806 2807
		auth->authorizer = NULL;
	}
2808 2809 2810
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2811
		if (ret)
2812
			return ERR_PTR(ret);
2813 2814
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2815 2816 2817
						     auth);
		if (ret)
			return ERR_PTR(ret);
2818 2819
	}
	*proto = ac->protocol;
2820

2821
	return auth;
2822 2823 2824 2825 2826 2827 2828 2829 2830
}


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;

2831
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2832 2833
}

2834 2835 2836 2837 2838 2839
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;

2840
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2841 2842
	return ceph_monc_validate_auth(&osdc->client->monc);
}
2843

2844
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2845 2846 2847
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2848 2849
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2850
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2851
	.alloc_msg = alloc_msg,
2852
	.fault = osd_reset,
S
Sage Weil 已提交
2853
};