osd_client.c 65.7 KB
Newer Older
1

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

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

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

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

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

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

/*
 * Implement client access to distributed object storage cluster.
 *
 * All data objects are stored within a cluster/cloud of OSDs, or
 * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
 * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
 * remote daemons serving up and coordinating consistent and safe
 * access to storage.
 *
 * Cluster membership and the mapping of data objects onto storage devices
 * are described by the osd map.
 *
 * We keep track of pending OSD requests (read, write), resubmit
 * requests to different OSDs when the cluster topology/data layout
 * change, or retry the affected requests when the communications
 * channel with an OSD is reset.
 */

/*
 * calculate the mapping of a file extent onto an object, and fill out the
 * request accordingly.  shorten extent as necessary if it crosses an
 * object boundary.
 *
 * fill osd op in request message.
 */
61
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
62
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
63
{
64
	u64 orig_len = *plen;
S
Sage Weil 已提交
65
	int r;
S
Sage Weil 已提交
66

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

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

80
	return 0;
S
Sage Weil 已提交
81 82
}

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

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

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

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

118 119 120 121 122 123
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
			unsigned int which, bool write_request)
{
	BUG_ON(which >= osd_req->r_num_ops);

124
	return &osd_req->r_ops[which].extent.osd_data;
125 126 127 128 129 130 131 132 133
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data);

struct ceph_osd_data *
osd_req_op_cls_request_info(struct ceph_osd_request *osd_req,
			unsigned int which)
{
	BUG_ON(which >= osd_req->r_num_ops);

134
	return &osd_req->r_ops[which].cls.request_info;
135 136 137 138 139 140 141 142 143
}
EXPORT_SYMBOL(osd_req_op_cls_request_info);	/* ??? */

struct ceph_osd_data *
osd_req_op_cls_response_data(struct ceph_osd_request *osd_req,
			unsigned int which)
{
	BUG_ON(which >= osd_req->r_num_ops);

144
	return &osd_req->r_ops[which].cls.response_data;
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206
}
EXPORT_SYMBOL(osd_req_op_cls_response_data);	/* ??? */

void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
			unsigned int which, bool write_request,
			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_extent_osd_data(osd_req, which, write_request);
	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,
			unsigned int which, bool write_request,
			struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

	osd_data = osd_req_op_extent_osd_data(osd_req, which, write_request);
	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,
			unsigned int which, bool write_request,
			struct bio *bio, size_t bio_length)
{
	struct ceph_osd_data *osd_data;

	osd_data = osd_req_op_extent_osd_data(osd_req, which, write_request);
	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;

	osd_data = osd_req_op_cls_request_info(osd_req, which);
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

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;

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

207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
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;
	}
}

226

A
Alex Elder 已提交
227 228
static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
{
229
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
A
Alex Elder 已提交
230 231 232 233 234 235
		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);
	}
236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
	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);
		ceph_osd_data_release(&op->cls.response_data);
		break;
	default:
		break;
	}
A
Alex Elder 已提交
259 260
}

S
Sage Weil 已提交
261 262 263
/*
 * requests
 */
S
Sage Weil 已提交
264
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
265
{
A
Alex Elder 已提交
266
	struct ceph_osd_request *req;
267
	unsigned int which;
S
Sage Weil 已提交
268

A
Alex Elder 已提交
269
	req = container_of(kref, struct ceph_osd_request, r_kref);
S
Sage Weil 已提交
270 271
	if (req->r_request)
		ceph_msg_put(req->r_request);
272
	if (req->r_reply) {
273
		ceph_msg_revoke_incoming(req->r_reply);
274
		ceph_msg_put(req->r_reply);
275
	}
276

277 278
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
279

S
Sage Weil 已提交
280 281 282 283 284
	ceph_put_snap_context(req->r_snapc);
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
		kfree(req);
S
Sage Weil 已提交
285
}
286
EXPORT_SYMBOL(ceph_osdc_release_request);
287

288
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
289
					       struct ceph_snap_context *snapc,
290
					       unsigned int num_ops,
291
					       bool use_mempool,
292
					       gfp_t gfp_flags)
S
Sage Weil 已提交
293 294 295
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
296 297
	size_t msg_size;

298 299 300
	BUILD_BUG_ON(CEPH_OSD_MAX_OP > U16_MAX);
	BUG_ON(num_ops > CEPH_OSD_MAX_OP);

301 302 303 304 305 306 307 308 309
	msg_size = 4 + 4 + 8 + 8 + 4+8;
	msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
	msg_size += 1 + 8 + 4 + 4;     /* pg_t */
	msg_size += 4 + MAX_OBJ_NAME_SIZE;
	msg_size += 2 + num_ops*sizeof(struct ceph_osd_op);
	msg_size += 8;  /* snapid */
	msg_size += 8;  /* snap_seq */
	msg_size += 8 * (snapc ? snapc->num_snaps : 0);  /* snaps */
	msg_size += 4;
S
Sage Weil 已提交
310 311

	if (use_mempool) {
312
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
313 314
		memset(req, 0, sizeof(*req));
	} else {
315
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
316 317
	}
	if (req == NULL)
318
		return NULL;
S
Sage Weil 已提交
319 320 321

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
322
	req->r_num_ops = num_ops;
323

S
Sage Weil 已提交
324
	kref_init(&req->r_kref);
S
Sage Weil 已提交
325 326
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
327
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
328
	INIT_LIST_HEAD(&req->r_unsafe_item);
329 330
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
331
	INIT_LIST_HEAD(&req->r_req_lru_item);
332 333
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
334 335 336 337 338
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
339
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
340
	if (!msg) {
S
Sage Weil 已提交
341
		ceph_osdc_put_request(req);
342
		return NULL;
S
Sage Weil 已提交
343 344 345 346
	}
	req->r_reply = msg;

	/* create request message; allow space for oid */
S
Sage Weil 已提交
347
	if (use_mempool)
348
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
349
	else
350
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
351
	if (!msg) {
S
Sage Weil 已提交
352
		ceph_osdc_put_request(req);
353
		return NULL;
S
Sage Weil 已提交
354
	}
355

S
Sage Weil 已提交
356
	memset(msg->front.iov_base, 0, msg->front.iov_len);
357 358 359 360 361

	req->r_request = msg;

	return req;
}
362
EXPORT_SYMBOL(ceph_osdc_alloc_request);
363

364
static bool osd_req_opcode_valid(u16 opcode)
365
{
366
	switch (opcode) {
367
	case CEPH_OSD_OP_READ:
368
	case CEPH_OSD_OP_STAT:
369 370 371
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
372
	case CEPH_OSD_OP_NOTIFY:
373
	case CEPH_OSD_OP_NOTIFY_ACK:
374
	case CEPH_OSD_OP_ASSERT_VER:
375
	case CEPH_OSD_OP_WRITE:
376 377 378 379 380
	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:
381
	case CEPH_OSD_OP_STARTSYNC:
382 383 384 385 386 387
	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:
388
	case CEPH_OSD_OP_ROLLBACK:
389
	case CEPH_OSD_OP_WATCH:
390 391 392 393 394 395 396 397 398 399 400 401
	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:
402
	case CEPH_OSD_OP_GETXATTR:
403
	case CEPH_OSD_OP_GETXATTRS:
404 405
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
	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:
424
	case CEPH_OSD_OP_CALL:
425 426
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
427 428 429 430 431 432
		return true;
	default:
		return false;
	}
}

433 434 435 436 437
/*
 * 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.
 */
438 439 440
static struct ceph_osd_req_op *
osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
				u16 opcode)
441
{
442 443 444
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
445 446
	BUG_ON(!osd_req_opcode_valid(opcode));

447
	op = &osd_req->r_ops[which];
448 449
	memset(op, 0, sizeof (*op));
	op->op = opcode;
450 451

	return op;
452 453
}

454 455
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
456 457 458
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
459
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
	size_t payload_len = 0;

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

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

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

475 476
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
477
{
478 479 480 481 482 483
	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;
484 485 486 487 488 489 490 491 492 493

	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);

494 495
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
			u16 opcode, const char *class, const char *method,
496 497
			const void *request_data, size_t request_data_size)
{
498
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
499
	struct ceph_pagelist *pagelist;
500 501 502 503 504
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

505 506 507 508
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

509 510 511 512
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
513
	ceph_pagelist_append(pagelist, class, size);
514 515 516 517 518 519
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
520
	ceph_pagelist_append(pagelist, method, size);
521 522
	payload_len += size;

523
	op->cls.request_data = request_data;
524
	BUG_ON(request_data_size > (size_t) U32_MAX);
525
	op->cls.request_data_len = (u32) request_data_size;
526
	ceph_pagelist_append(pagelist, request_data, request_data_size);
527 528
	payload_len += request_data_size;

529
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
530

531 532 533 534 535
	op->cls.argc = 0;	/* currently unused */

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_cls_init);
536

537 538
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
539 540
				u64 cookie, u64 version, int flag)
{
541
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
542

543
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
544 545 546 547 548

	op->watch.cookie = cookie;
	/* op->watch.ver = version; */	/* XXX 3847 */
	op->watch.ver = cpu_to_le64(version);
	if (opcode == CEPH_OSD_OP_WATCH && flag)
549
		op->watch.flag = (u8)1;
550 551 552
}
EXPORT_SYMBOL(osd_req_op_watch_init);

553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
static void ceph_osdc_msg_data_set(struct ceph_msg *msg,
				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)
			ceph_msg_data_set_pages(msg, osd_data->pages,
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
		ceph_msg_data_set_pagelist(msg, osd_data->pagelist);
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
		ceph_msg_data_set_bio(msg, osd_data->bio, length);
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

575
static u64 osd_req_encode_op(struct ceph_osd_request *req,
576
			      struct ceph_osd_op *dst, unsigned int which)
577
{
578
	struct ceph_osd_req_op *src;
579
	u64 request_data_len = 0;
580

581 582
	BUG_ON(which >= req->r_num_ops);
	src = &req->r_ops[which];
583 584 585 586 587 588 589 590 591 592 593 594
	if (WARN_ON(!osd_req_opcode_valid(src->op))) {
		pr_err("unrecognized osd opcode %d\n", src->op);

		return 0;
	}

	switch (src->op) {
	case CEPH_OSD_OP_STAT:
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		if (src->op == CEPH_OSD_OP_WRITE)
595
			request_data_len = src->extent.length;
596 597 598 599 600 601
		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);
602
		if (src->op == CEPH_OSD_OP_WRITE)
603
			ceph_osdc_msg_data_set(req->r_request,
604 605
						&src->extent.osd_data);
		else
606
			ceph_osdc_msg_data_set(req->r_reply,
607
						&src->extent.osd_data);
608 609 610 611
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
612
		dst->cls.indata_len = cpu_to_le32(src->cls.request_data_len);
613 614 615
		ceph_osdc_msg_data_set(req->r_reply, &src->cls.response_data);
		ceph_osdc_msg_data_set(req->r_request, &src->cls.request_info);
		BUG_ON(src->cls.request_info.type !=
616
					CEPH_OSD_DATA_TYPE_PAGELIST);
617
		request_data_len = src->cls.request_info.pagelist->length;
618 619 620 621 622 623 624 625 626 627
		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:
628
		pr_err("unsupported osd opcode %s\n",
629
			ceph_osd_op_name(src->op));
630
		WARN_ON(1);
631 632

		return 0;
633
	}
634
	dst->op = cpu_to_le16(src->op);
635
	dst->payload_len = cpu_to_le32(src->payload_len);
636

637
	return request_data_len;
638 639
}

640 641 642 643 644 645 646 647 648 649 650 651 652 653
/*
 * 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,
654
					       u64 off, u64 *plen, int num_ops,
655 656 657 658
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
659
					       bool use_mempool)
660
{
661
	struct ceph_osd_request *req;
662 663 664
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
665 666
	u32 object_size;
	u64 object_base;
667
	int r;
668

669
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
670

671
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
672
					GFP_NOFS);
673
	if (!req)
674
		return ERR_PTR(-ENOMEM);
675

676
	req->r_flags = flags;
677 678

	/* calculate max write size */
679
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
680 681
	if (r < 0) {
		ceph_osdc_put_request(req);
682
		return ERR_PTR(r);
683
	}
684

685 686 687 688 689 690 691 692
	object_size = le32_to_cpu(layout->fl_object_size);
	object_base = off - objoff;
	if (truncate_size <= object_base) {
		truncate_size = 0;
	} else {
		truncate_size -= object_base;
		if (truncate_size > object_size)
			truncate_size = object_size;
693
	}
694

695
	osd_req_op_extent_init(req, 0, opcode, objoff, objlen,
696
				truncate_size, truncate_seq);
697

698 699 700 701 702 703
	/*
	 * 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)
704
		osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC);
705

706 707
	req->r_file_layout = *layout;  /* keep a copy */

708 709
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx",
		vino.ino, objnum);
710 711
	req->r_oid_len = strlen(req->r_oid);

S
Sage Weil 已提交
712 713
	return req;
}
714
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780

/*
 * 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;
}

781 782 783 784 785 786
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
787
	struct ceph_osd_request *req, *nreq;
788
	LIST_HEAD(resend);
789 790 791 792
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
793
	if (err)
794
		return;
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811
	/*
	 * 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.
	 */
812
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
813 814 815 816
		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,
817
		     osd->o_osd);
818 819 820
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}
821
	list_splice(&resend, &osdc->req_unsent);
822

823 824 825 826 827
	/*
	 * 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.
	 */
828 829
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
830 831 832 833 834
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
835
		__register_request(osdc, req);
836
		list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
837
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
838
		__unregister_linger_request(osdc, req);
839 840
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
841 842 843
	}
}

S
Sage Weil 已提交
844
/*
845
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
846 847 848 849 850 851 852 853 854 855 856
 */
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
857 858
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
859
	__send_queued(osdc);
S
Sage Weil 已提交
860
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
861 862 863 864 865 866
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
867
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
868 869 870 871 872 873 874 875 876
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
877
	osd->o_osd = onum;
878
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
879
	INIT_LIST_HEAD(&osd->o_requests);
880
	INIT_LIST_HEAD(&osd->o_linger_requests);
881
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
882 883
	osd->o_incarnation = 1;

884
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
885

886
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905
	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);
906
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
907 908
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

909
		ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
910
		kfree(osd);
S
Sage Weil 已提交
911
	}
S
Sage Weil 已提交
912 913 914 915 916
}

/*
 * remove an osd from our map
 */
917
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
918
{
919
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
920 921
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
922
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
923 924 925 926
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

927 928
static void remove_all_osds(struct ceph_osd_client *osdc)
{
929
	dout("%s %p\n", __func__, osdc);
930 931 932 933 934 935 936 937 938
	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);
}

939 940 941 942 943 944
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);
945
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
946 947 948 949 950 951 952 953 954
}

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);
}

955
static void remove_old_osds(struct ceph_osd_client *osdc)
956 957 958 959 960 961
{
	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) {
962
		if (time_before(jiffies, osd->lru_ttl))
963 964 965 966 967 968
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
969 970 971
/*
 * reset osd connect
 */
972
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
973
{
A
Alex Elder 已提交
974
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
975

976
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
977 978
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
979
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
980 981 982 983 984 985 986 987 988

		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;

989 990 991 992 993
		dout(" osd addr hasn't changed and connection never opened,"
		     " letting msgr retry");
		/* touch each r_stamp for handle_timeout()'s benfit */
		list_for_each_entry(req, &osd->o_requests, r_osd_item)
			req->r_stamp = jiffies;
A
Alex Elder 已提交
994 995

		return -EAGAIN;
S
Sage Weil 已提交
996
	}
A
Alex Elder 已提交
997 998 999 1000 1001 1002

	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
1003 1004 1005 1006 1007 1008 1009 1010
}

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;

1011
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
	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;
}

1044 1045 1046
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
1047
			osdc->client->options->osd_keepalive_timeout * HZ);
1048 1049 1050 1051 1052 1053
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
1054 1055 1056 1057 1058

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1059 1060
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1061 1062
{
	req->r_tid = ++osdc->last_tid;
1063
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1064
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
1065 1066 1067 1068
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
1069 1070
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
1071
	}
1072 1073
}

S
Sage Weil 已提交
1074 1075 1076 1077 1078 1079
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1080 1081 1082 1083 1084 1085
	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
1086 1087 1088 1089
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

1090 1091
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1092
		ceph_msg_revoke(req->r_request);
1093 1094

		list_del_init(&req->r_osd_item);
1095 1096 1097
		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);
1098
			__move_osd_to_lru(osdc, req->r_osd);
1099
		}
1100
		if (list_empty(&req->r_linger_item))
1101
			req->r_osd = NULL;
1102
	}
S
Sage Weil 已提交
1103

A
Alex Elder 已提交
1104
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1105 1106
	ceph_osdc_put_request(req);

1107 1108 1109
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1110 1111 1112 1113 1114 1115 1116 1117
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1118
	if (req->r_sent && req->r_osd) {
1119
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1120 1121 1122 1123
		req->r_sent = 0;
	}
}

1124 1125 1126 1127 1128
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
	dout("__register_linger_request %p\n", req);
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1129 1130 1131
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
1132 1133 1134 1135 1136 1137
}

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
1138
	list_del_init(&req->r_linger_item);
1139 1140 1141 1142 1143 1144 1145 1146
	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);
		}
1147 1148
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
	}
}

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

void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
				  struct ceph_osd_request *req)
{
	if (!req->r_linger) {
		dout("set_request_linger %p\n", req);
		req->r_linger = 1;
		/*
		 * caller is now responsible for calling
		 * unregister_linger_request
		 */
		ceph_osdc_get_request(req);
	}
}
EXPORT_SYMBOL(ceph_osdc_set_request_linger);

S
Sage Weil 已提交
1179 1180 1181
/*
 * 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
1182
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1183
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1184 1185 1186 1187 1188
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1189
static int __map_request(struct ceph_osd_client *osdc,
1190
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1191
{
1192
	struct ceph_pg pgid;
1193 1194
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
1195 1196
	int err;

1197
	dout("map_request %p tid %lld\n", req, req->r_tid);
1198 1199
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
1200 1201
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1202
		return err;
1203
	}
1204 1205
	req->r_pgid = pgid;

1206 1207 1208 1209 1210
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
1211

1212 1213
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1214 1215 1216
	     req->r_sent >= req->r_osd->o_incarnation &&
	     req->r_num_pg_osds == num &&
	     memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
S
Sage Weil 已提交
1217 1218 1219
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

1220 1221
	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
1222 1223
	     req->r_osd ? req->r_osd->o_osd : -1);

1224 1225 1226 1227
	/* 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
1228 1229 1230 1231 1232 1233 1234 1235
	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) {
1236
		err = -ENOMEM;
1237
		req->r_osd = create_osd(osdc, o);
1238 1239
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1240
			goto out;
1241
		}
S
Sage Weil 已提交
1242

1243
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1244 1245
		__insert_osd(osdc, req->r_osd);

1246 1247 1248
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1249 1250
	}

1251 1252
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
1253 1254
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1255
	} else {
1256
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1257
	}
1258
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1259 1260 1261 1262 1263 1264 1265 1266

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1267 1268
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1269
{
1270
	void *p;
S
Sage Weil 已提交
1271

1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
	dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
	     req, req->r_tid, req->r_osd->o_osd, req->r_flags,
	     (unsigned long long)req->r_pgid.pool, req->r_pgid.seed);

	/* fill in message content that changes each time we send it */
	put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch);
	put_unaligned_le32(req->r_flags, req->r_request_flags);
	put_unaligned_le64(req->r_pgid.pool, req->r_request_pool);
	p = req->r_request_pgid;
	ceph_encode_64(&p, req->r_pgid.pool);
	ceph_encode_32(&p, req->r_pgid.seed);
	put_unaligned_le64(1, req->r_request_attempts);  /* FIXME */
	memcpy(req->r_request_reassert_version, &req->r_reassert_version,
	       sizeof(req->r_reassert_version));
1286

S
Sage Weil 已提交
1287
	req->r_stamp = jiffies;
1288
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1289 1290 1291 1292 1293 1294

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

1295 1296 1297
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1298
static void __send_queued(struct ceph_osd_client *osdc)
1299 1300 1301
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1302 1303
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1304 1305 1306
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319
/*
 * 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
1320
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1321
	struct ceph_osd *osd;
1322
	unsigned long keepalive =
1323
		osdc->client->options->osd_keepalive_timeout * HZ;
1324
	struct list_head slow_osds;
S
Sage Weil 已提交
1325 1326 1327 1328 1329 1330 1331
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1332 1333 1334 1335 1336 1337 1338
	/*
	 * 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
1339
		if (time_before(jiffies, req->r_stamp + keepalive))
1340 1341 1342 1343 1344
			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
1345
		     req->r_tid, osd->o_osd);
1346 1347 1348 1349 1350 1351
		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
1352 1353 1354
		ceph_con_keepalive(&osd->o_con);
	}

1355
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1356
	__send_queued(osdc);
S
Sage Weil 已提交
1357 1358 1359 1360
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1361 1362 1363 1364 1365 1366
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 =
1367
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1368 1369 1370

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1371
	remove_old_osds(osdc);
1372 1373 1374 1375 1376 1377
	up_read(&osdc->map_sem);

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

S
Sage Weil 已提交
1378 1379 1380 1381 1382 1383 1384
static void complete_request(struct ceph_osd_request *req)
{
	if (req->r_safe_callback)
		req->r_safe_callback(req, NULL);
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1385 1386 1387 1388
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1389 1390
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1391
{
1392
	void *p, *end;
S
Sage Weil 已提交
1393 1394
	struct ceph_osd_request *req;
	u64 tid;
1395
	int object_len;
1396 1397
	unsigned int numops;
	int payload_len, flags;
1398
	s32 result;
1399 1400 1401 1402 1403 1404
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1405
	int already_completed;
1406
	u32 bytes;
1407
	unsigned int i;
S
Sage Weil 已提交
1408

1409
	tid = le64_to_cpu(msg->hdr.tid);
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
	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;

1420
	err = ceph_decode_pgid(&p, end, &pg);
1421
	if (err)
S
Sage Weil 已提交
1422
		goto bad;
1423 1424 1425 1426 1427 1428 1429 1430

	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
1431 1432 1433 1434 1435
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1436
		goto bad_mutex;
S
Sage Weil 已提交
1437 1438
	}
	ceph_osdc_get_request(req);
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460

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

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

		len = le32_to_cpu(op->payload_len);
		req->r_reply_op_len[i] = len;
		dout(" op %d has %d bytes\n", i, len);
		payload_len += len;
		p += sizeof(*op);
	}
1461 1462
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1463
		pr_warning("sum of op payload lens %d != data_len %d",
1464
			   payload_len, bytes);
1465 1466 1467 1468 1469 1470 1471
		goto bad_put;
	}

	ceph_decode_need(&p, end, 4 + numops * 4, bad);
	retry_attempt = ceph_decode_32(&p);
	for (i = 0; i < numops; i++)
		req->r_reply_op_result[i] = ceph_decode_32(&p);
S
Sage Weil 已提交
1472 1473 1474

	if (!req->r_got_reply) {

1475
		req->r_result = result;
S
Sage Weil 已提交
1476 1477 1478 1479 1480 1481
		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, */
1482 1483
		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
1484 1485 1486 1487

		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
1488
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1489 1490 1491 1492 1493
		goto done;
	}

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

1494 1495 1496
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1497
	/* either this is a read, or we got the safe response */
1498 1499
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1500 1501 1502
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1503 1504
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1505
	mutex_unlock(&osdc->request_mutex);
1506 1507
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1508 1509 1510 1511

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1512
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1513

S
Sage Weil 已提交
1514 1515
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1516 1517

done:
1518
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1519 1520 1521
	ceph_osdc_put_request(req);
	return;

1522 1523
bad_put:
	ceph_osdc_put_request(req);
1524 1525
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1526
bad:
1527 1528
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1529
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1530 1531
}

1532
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1533 1534 1535
{
	struct rb_node *p, *n;

1536 1537
	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
1538

1539 1540 1541 1542 1543 1544 1545
		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
1546
	}
1547 1548 1549
}

/*
1550 1551
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1552
 *
1553
 * Caller should hold map_sem for read.
1554
 */
1555
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1556
{
1557
	struct ceph_osd_request *req, *nreq;
1558 1559 1560
	struct rb_node *p;
	int needmap = 0;
	int err;
1561

1562
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1563
	mutex_lock(&osdc->request_mutex);
1564
	for (p = rb_first(&osdc->requests); p; ) {
1565
		req = rb_entry(p, struct ceph_osd_request, r_node);
1566
		p = rb_next(p);
1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584

		/*
		 * For linger requests that have not yet been
		 * registered, move them to the linger list; they'll
		 * be sent to the osd in the loop below.  Unregister
		 * the request before re-registering it as a linger
		 * request to ensure the __map_request() below
		 * will decide it needs to be sent.
		 */
		if (req->r_linger && list_empty(&req->r_linger_item)) {
			dout("%p tid %llu restart on osd%d\n",
			     req, req->r_tid,
			     req->r_osd ? req->r_osd->o_osd : -1);
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
			continue;
		}

1585
		err = __map_request(osdc, req, force_resend);
1586 1587 1588 1589 1590 1591
		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) {
1592 1593 1594 1595
			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);
1596
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1597 1598
			}
		}
1599 1600 1601 1602 1603 1604
	}

	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);

1605
		err = __map_request(osdc, req, force_resend);
1606
		dout("__map_request returned %d\n", err);
1607 1608 1609 1610 1611 1612 1613 1614
		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;
1615
		}
1616 1617 1618 1619

		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);
1620
		__unregister_linger_request(osdc, req);
1621
	}
S
Sage Weil 已提交
1622 1623 1624 1625 1626 1627
	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);
	}
1628
	reset_changed_osds(osdc);
1629
}
1630 1631


S
Sage Weil 已提交
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654
/*
 * Process updated osd map.
 *
 * The message contains any number of incremental and full maps, normally
 * indicating some sort of topology change in the cluster.  Kick requests
 * off to different OSDs as needed.
 */
void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
{
	void *p, *end, *next;
	u32 nr_maps, maplen;
	u32 epoch;
	struct ceph_osdmap *newmap = NULL, *oldmap;
	int err;
	struct ceph_fsid fsid;

	dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	/* verify fsid */
	ceph_decode_need(&p, end, sizeof(fsid), bad);
	ceph_decode_copy(&p, &fsid, sizeof(fsid));
1655 1656
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1657 1658 1659 1660 1661 1662 1663 1664

	down_write(&osdc->map_sem);

	/* incremental maps */
	ceph_decode_32_safe(&p, end, nr_maps, bad);
	dout(" %d inc maps\n", nr_maps);
	while (nr_maps > 0) {
		ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1665 1666
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1667 1668 1669 1670 1671 1672 1673
		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,
1674
							  &osdc->client->msgr);
S
Sage Weil 已提交
1675 1676 1677 1678
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1679
			BUG_ON(!newmap);
S
Sage Weil 已提交
1680 1681 1682 1683
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1684
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
		} 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);
1700 1701
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1702 1703 1704 1705 1706 1707 1708 1709 1710
		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 {
1711 1712
			int skipped_map = 0;

S
Sage Weil 已提交
1713 1714 1715 1716 1717 1718
			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;
			}
1719
			BUG_ON(!newmap);
S
Sage Weil 已提交
1720 1721
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1722 1723 1724
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1725
				ceph_osdmap_destroy(oldmap);
1726 1727
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1728 1729 1730 1731 1732 1733 1734 1735
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1736 1737 1738 1739 1740 1741 1742 1743 1744

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

A
Alex Elder 已提交
1745 1746 1747
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1748
	up_read(&osdc->map_sem);
1749
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1750 1751 1752 1753
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1754
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1755 1756 1757 1758
	up_write(&osdc->map_sem);
	return;
}

1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840
/*
 * 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 *),
1841
			   void *data, struct ceph_osd_event **pevent)
1842 1843 1844 1845 1846 1847 1848 1849 1850
{
	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;
1851
	event->one_shot = 0;
1852 1853 1854
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1855
	RB_CLEAR_NODE(&event->node);
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
	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
 */
1902 1903
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
{
	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) {
1924
		BUG_ON(event->one_shot);
1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
		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;
		}
1936
		INIT_WORK(&event_work->work, do_event_work);
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
		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;
}

1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
				struct ceph_snap_context *snapc, u64 snap_id,
				struct timespec *mtime)
{
	struct ceph_msg *msg = req->r_request;
	void *p;
	size_t msg_size;
	int flags = req->r_flags;
	u64 data_len;
	unsigned int i;

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

	/* encode request */
	msg->hdr.version = cpu_to_le16(4);

	p = msg->front.iov_base;
	ceph_encode_32(&p, 1);   /* client_inc  is always 1 */
	req->r_request_osdmap_epoch = p;
	p += 4;
	req->r_request_flags = p;
	p += 4;
	if (req->r_flags & CEPH_OSD_FLAG_WRITE)
		ceph_encode_timespec(p, mtime);
	p += sizeof(struct ceph_timespec);
	req->r_request_reassert_version = p;
	p += sizeof(struct ceph_eversion); /* will get filled in */

	/* oloc */
	ceph_encode_8(&p, 4);
	ceph_encode_8(&p, 4);
	ceph_encode_32(&p, 8 + 4 + 4);
	req->r_request_pool = p;
	p += 8;
	ceph_encode_32(&p, -1);  /* preferred */
	ceph_encode_32(&p, 0);   /* key len */

	ceph_encode_8(&p, 1);
	req->r_request_pgid = p;
	p += 8 + 4;
	ceph_encode_32(&p, -1);  /* preferred */

	/* oid */
	ceph_encode_32(&p, req->r_oid_len);
	memcpy(p, req->r_oid, req->r_oid_len);
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
	p += req->r_oid_len;

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

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

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

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

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

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

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

2056 2057 2058 2059 2060 2061 2062 2063 2064
/*
 * 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
2065 2066
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2067
	__register_request(osdc, req);
2068 2069 2070 2071 2072 2073 2074
	WARN_ON(req->r_sent);
	rc = __map_request(osdc, req, 0);
	if (rc < 0) {
		if (nofail) {
			dout("osdc_start_request failed map, "
				" will retry %lld\n", req->r_tid);
			rc = 0;
S
Sage Weil 已提交
2075
		}
2076
		goto out_unlock;
S
Sage Weil 已提交
2077
	}
2078 2079 2080 2081
	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 {
2082
		__send_queued(osdc);
2083 2084
	}
	rc = 0;
2085
out_unlock:
S
Sage Weil 已提交
2086 2087 2088 2089
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
2090
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103

/*
 * 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);
2104
		__unregister_request(osdc, req);
S
Sage Weil 已提交
2105
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2106
		complete_request(req);
2107
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
2108 2109 2110 2111 2112 2113
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
2114
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147

/*
 * 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);
}
2148
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165

/*
 * 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;
2166
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
2167
	osdc->requests = RB_ROOT;
2168
	INIT_LIST_HEAD(&osdc->req_lru);
2169 2170
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2171
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2172 2173
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2174
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2175 2176 2177
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2178 2179

	schedule_delayed_work(&osdc->osds_timeout_work,
2180
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
2181

2182
	err = -ENOMEM;
S
Sage Weil 已提交
2183 2184 2185
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
2186
		goto out;
S
Sage Weil 已提交
2187

2188 2189
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
2190
				"osd_op");
S
Sage Weil 已提交
2191
	if (err < 0)
2192
		goto out_mempool;
2193
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2194 2195
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
2196 2197
	if (err < 0)
		goto out_msgpool;
2198 2199 2200 2201 2202 2203 2204

	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
	if (IS_ERR(osdc->notify_wq)) {
		err = PTR_ERR(osdc->notify_wq);
		osdc->notify_wq = NULL;
		goto out_msgpool;
	}
S
Sage Weil 已提交
2205
	return 0;
2206

S
Sage Weil 已提交
2207 2208
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2209 2210 2211 2212
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2213 2214 2215 2216
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2217 2218
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2219
	cancel_delayed_work_sync(&osdc->timeout_work);
2220
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
2221 2222 2223 2224
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
2225
	remove_all_osds(osdc);
S
Sage Weil 已提交
2226 2227
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2228
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
}

/*
 * 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,
2239
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2240 2241 2242 2243 2244 2245
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2246
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1,
S
Sage Weil 已提交
2247
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2248
				    NULL, truncate_seq, truncate_size,
2249
				    false);
2250 2251
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2252 2253

	/* it may be a short read due to an object boundary */
2254

2255 2256
	osd_req_op_extent_osd_data_pages(req, 0, false,
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2257

2258
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2259
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2260

2261
	ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2262

S
Sage Weil 已提交
2263 2264 2265 2266 2267 2268 2269 2270
	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;
}
2271
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281

/*
 * 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,
2282
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2283 2284 2285
{
	struct ceph_osd_request *req;
	int rc = 0;
2286
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2287

2288
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
2289
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1,
S
Sage Weil 已提交
2290
				    CEPH_OSD_OP_WRITE,
2291
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2292
				    snapc, truncate_seq, truncate_size,
2293
				    true);
2294 2295
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2296 2297

	/* it may be a short write due to an object boundary */
2298
	osd_req_op_extent_osd_data_pages(req, 0, true, pages, len, page_align,
2299 2300
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2301

2302
	ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2303

2304
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313
	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;
}
2314
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2315 2316 2317 2318 2319 2320 2321

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2322
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2323 2324 2325
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2326
		goto out;
2327
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2328 2329 2330 2331 2332 2333

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2334
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2335
		break;
2336 2337 2338
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2339 2340 2341 2342 2343

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2344
out:
S
Sage Weil 已提交
2345 2346 2347
	ceph_msg_put(msg);
}

2348
/*
2349 2350
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2351 2352
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2353 2354
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2355 2356 2357
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2358
	struct ceph_msg *m;
2359
	struct ceph_osd_request *req;
2360 2361
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2362
	u64 tid;
S
Sage Weil 已提交
2363

2364 2365 2366 2367 2368 2369
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2370 2371
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2372 2373
		goto out;
	}
S
Sage Weil 已提交
2374

2375
	if (req->r_reply->con)
2376
		dout("%s revoking msg %p from old con %p\n", __func__,
2377 2378
		     req->r_reply, req->r_reply->con);
	ceph_msg_revoke_incoming(req->r_reply);
2379

S
Sage Weil 已提交
2380 2381 2382
	if (front > req->r_reply->front.iov_len) {
		pr_warning("get_reply front %d > preallocated %d\n",
			   front, (int)req->r_reply->front.iov_len);
2383
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2384
		if (!m)
S
Sage Weil 已提交
2385 2386 2387 2388 2389 2390
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2391
	if (data_len > 0) {
2392
		struct ceph_osd_data *osd_data;
2393

2394 2395 2396 2397 2398 2399
		/*
		 * 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.
		 */
		osd_data = osd_req_op_extent_osd_data(req, 0, false);
2400 2401
		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2402
				unlikely(osd_data->length < data_len)) {
2403

2404 2405 2406
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2407 2408 2409 2410 2411 2412
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2413
	}
2414
	*skip = 0;
S
Sage Weil 已提交
2415
	dout("get_reply tid %lld %p\n", tid, m);
2416 2417 2418

out:
	mutex_unlock(&osdc->request_mutex);
2419
	return m;
2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430

}

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
2431
	*skip = 0;
2432 2433
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2434
	case CEPH_MSG_WATCH_NOTIFY:
2435
		return ceph_msg_new(type, front, GFP_NOFS, false);
2436 2437 2438 2439 2440 2441 2442 2443
	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
2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462
}

/*
 * 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);
}

2463 2464 2465
/*
 * authentication
 */
2466 2467 2468 2469 2470
/*
 * 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,
2471
					int *proto, int force_new)
2472 2473 2474 2475
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2476
	struct ceph_auth_handshake *auth = &o->o_auth;
2477

2478
	if (force_new && auth->authorizer) {
2479
		ceph_auth_destroy_authorizer(ac, auth->authorizer);
2480 2481
		auth->authorizer = NULL;
	}
2482 2483 2484
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2485
		if (ret)
2486
			return ERR_PTR(ret);
2487 2488
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2489 2490 2491
						     auth);
		if (ret)
			return ERR_PTR(ret);
2492 2493
	}
	*proto = ac->protocol;
2494

2495
	return auth;
2496 2497 2498 2499 2500 2501 2502 2503 2504
}


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;

2505
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2506 2507
}

2508 2509 2510 2511 2512 2513
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;

2514
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2515 2516
	return ceph_monc_validate_auth(&osdc->client->monc);
}
2517

2518
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2519 2520 2521
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2522 2523
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2524
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2525
	.alloc_msg = alloc_msg,
2526
	.fault = osd_reset,
S
Sage Weil 已提交
2527
};