osd_client.c 54.3 KB
Newer Older
1
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
2

3
#include <linux/module.h>
S
Sage Weil 已提交
4 5 6 7 8 9
#include <linux/err.h>
#include <linux/highmem.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
10 11 12
#ifdef CONFIG_BLOCK
#include <linux/bio.h>
#endif
S
Sage Weil 已提交
13

14 15 16 17 18 19
#include <linux/ceph/libceph.h>
#include <linux/ceph/osd_client.h>
#include <linux/ceph/messenger.h>
#include <linux/ceph/decode.h>
#include <linux/ceph/auth.h>
#include <linux/ceph/pagelist.h>
S
Sage Weil 已提交
20

S
Sage Weil 已提交
21 22
#define OSD_OP_FRONT_LEN	4096
#define OSD_OPREPLY_FRONT_LEN	512
23

24
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
25

A
Alex Elder 已提交
26
static void __send_queued(struct ceph_osd_client *osdc);
27
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
28 29 30 31
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req);
static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req);
32 33
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
34

35 36 37 38 39 40
static int op_has_extent(int op)
{
	return (op == CEPH_OSD_OP_READ ||
		op == CEPH_OSD_OP_WRITE);
}

S
Sage Weil 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
/*
 * Implement client access to distributed object storage cluster.
 *
 * All data objects are stored within a cluster/cloud of OSDs, or
 * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
 * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
 * remote daemons serving up and coordinating consistent and safe
 * access to storage.
 *
 * Cluster membership and the mapping of data objects onto storage devices
 * are described by the osd map.
 *
 * We keep track of pending OSD requests (read, write), resubmit
 * requests to different OSDs when the cluster topology/data layout
 * change, or retry the affected requests when the communications
 * channel with an OSD is reset.
 */

/*
 * calculate the mapping of a file extent onto an object, and fill out the
 * request accordingly.  shorten extent as necessary if it crosses an
 * object boundary.
 *
 * fill osd op in request message.
 */
66
static int calc_layout(struct ceph_vino vino,
S
Sage Weil 已提交
67 68 69 70
		       struct ceph_file_layout *layout,
		       u64 off, u64 *plen,
		       struct ceph_osd_request *req,
		       struct ceph_osd_req_op *op)
S
Sage Weil 已提交
71
{
72 73 74 75
	u64 orig_len = *plen;
	u64 bno = 0;
	u64 objoff = 0;
	u64 objlen = 0;
S
Sage Weil 已提交
76
	int r;
S
Sage Weil 已提交
77

78 79 80
	/* object extent? */
	r = ceph_calc_file_object_mapping(layout, off, orig_len, &bno,
					  &objoff, &objlen);
S
Sage Weil 已提交
81 82
	if (r < 0)
		return r;
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
	if (objlen < orig_len) {
		*plen = objlen;
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

	if (op_has_extent(op->op)) {
		u32 osize = le32_to_cpu(layout->fl_object_size);
		op->extent.offset = objoff;
		op->extent.length = objlen;
		if (op->extent.truncate_size <= off - objoff) {
			op->extent.truncate_size = 0;
		} else {
			op->extent.truncate_size -= off - objoff;
			if (op->extent.truncate_size > osize)
				op->extent.truncate_size = osize;
		}
	}
	req->r_num_pages = calc_pages_for(off, *plen);
	req->r_page_alignment = off & ~PAGE_MASK;
	if (op->op == CEPH_OSD_OP_WRITE)
		op->payload_len = *plen;

	dout("calc_layout bno=%llx %llu~%llu (%d pages)\n",
	     bno, objoff, objlen, req->r_num_pages);
S
Sage Weil 已提交
108

109
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx", vino.ino, bno);
S
Sage Weil 已提交
110
	req->r_oid_len = strlen(req->r_oid);
S
Sage Weil 已提交
111 112

	return r;
S
Sage Weil 已提交
113 114 115 116 117
}

/*
 * requests
 */
S
Sage Weil 已提交
118
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
119
{
S
Sage Weil 已提交
120 121 122 123 124 125
	struct ceph_osd_request *req = container_of(kref,
						    struct ceph_osd_request,
						    r_kref);

	if (req->r_request)
		ceph_msg_put(req->r_request);
126
	if (req->r_con_filling_msg) {
127 128
		dout("%s revoking msg %p from con %p\n", __func__,
		     req->r_reply, req->r_con_filling_msg);
129
		ceph_msg_revoke_incoming(req->r_reply);
130
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
131
		req->r_con_filling_msg = NULL;
132
	}
133 134
	if (req->r_reply)
		ceph_msg_put(req->r_reply);
S
Sage Weil 已提交
135 136 137 138
	if (req->r_own_pages)
		ceph_release_page_vector(req->r_pages,
					 req->r_num_pages);
	ceph_put_snap_context(req->r_snapc);
139
	ceph_pagelist_release(&req->r_trail);
S
Sage Weil 已提交
140 141 142 143
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
		kfree(req);
S
Sage Weil 已提交
144
}
145
EXPORT_SYMBOL(ceph_osdc_release_request);
146

147
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
148
					       struct ceph_snap_context *snapc,
A
Alex Elder 已提交
149
					       unsigned int num_op,
150
					       bool use_mempool,
151
					       gfp_t gfp_flags)
S
Sage Weil 已提交
152 153 154
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
155
	size_t msg_size = sizeof(struct ceph_osd_request_head);
156

157
	msg_size += num_op*sizeof(struct ceph_osd_op);
S
Sage Weil 已提交
158 159

	if (use_mempool) {
160
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
161 162
		memset(req, 0, sizeof(*req));
	} else {
163
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
164 165
	}
	if (req == NULL)
166
		return NULL;
S
Sage Weil 已提交
167 168 169

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
170

S
Sage Weil 已提交
171
	kref_init(&req->r_kref);
S
Sage Weil 已提交
172 173
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
174
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
175
	INIT_LIST_HEAD(&req->r_unsafe_item);
176 177
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
178
	INIT_LIST_HEAD(&req->r_req_lru_item);
179 180
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
181 182 183 184 185
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
186
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
187
	if (!msg) {
S
Sage Weil 已提交
188
		ceph_osdc_put_request(req);
189
		return NULL;
S
Sage Weil 已提交
190 191 192
	}
	req->r_reply = msg;

193
	ceph_pagelist_init(&req->r_trail);
194

S
Sage Weil 已提交
195
	/* create request message; allow space for oid */
196
	msg_size += MAX_OBJ_NAME_SIZE;
S
Sage Weil 已提交
197 198 199
	if (snapc)
		msg_size += sizeof(u64) * snapc->num_snaps;
	if (use_mempool)
200
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
201
	else
202
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
203
	if (!msg) {
S
Sage Weil 已提交
204
		ceph_osdc_put_request(req);
205
		return NULL;
S
Sage Weil 已提交
206
	}
207

S
Sage Weil 已提交
208
	memset(msg->front.iov_base, 0, msg->front.iov_len);
209 210 211 212 213

	req->r_request = msg;

	return req;
}
214
EXPORT_SYMBOL(ceph_osdc_alloc_request);
215

216 217 218 219 220 221
static void osd_req_encode_op(struct ceph_osd_request *req,
			      struct ceph_osd_op *dst,
			      struct ceph_osd_req_op *src)
{
	dst->op = cpu_to_le16(src->op);

222
	switch (src->op) {
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		dst->extent.offset =
			cpu_to_le64(src->extent.offset);
		dst->extent.length =
			cpu_to_le64(src->extent.length);
		dst->extent.truncate_size =
			cpu_to_le64(src->extent.truncate_size);
		dst->extent.truncate_seq =
			cpu_to_le32(src->extent.truncate_seq);
		break;

	case CEPH_OSD_OP_GETXATTR:
	case CEPH_OSD_OP_SETXATTR:
	case CEPH_OSD_OP_CMPXATTR:
		dst->xattr.name_len = cpu_to_le32(src->xattr.name_len);
		dst->xattr.value_len = cpu_to_le32(src->xattr.value_len);
		dst->xattr.cmp_op = src->xattr.cmp_op;
		dst->xattr.cmp_mode = src->xattr.cmp_mode;
242
		ceph_pagelist_append(&req->r_trail, src->xattr.name,
243
				     src->xattr.name_len);
244
		ceph_pagelist_append(&req->r_trail, src->xattr.val,
245 246
				     src->xattr.value_len);
		break;
247 248 249 250 251
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
		dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);

252
		ceph_pagelist_append(&req->r_trail, src->cls.class_name,
253
				     src->cls.class_len);
254
		ceph_pagelist_append(&req->r_trail, src->cls.method_name,
255
				     src->cls.method_len);
256
		ceph_pagelist_append(&req->r_trail, src->cls.indata,
257 258 259 260 261
				     src->cls.indata_len);
		break;
	case CEPH_OSD_OP_ROLLBACK:
		dst->snap.snapid = cpu_to_le64(src->snap.snapid);
		break;
262 263
	case CEPH_OSD_OP_STARTSYNC:
		break;
264 265 266 267 268
	case CEPH_OSD_OP_NOTIFY:
		{
			__le32 prot_ver = cpu_to_le32(src->watch.prot_ver);
			__le32 timeout = cpu_to_le32(src->watch.timeout);

269
			ceph_pagelist_append(&req->r_trail,
270
						&prot_ver, sizeof(prot_ver));
271
			ceph_pagelist_append(&req->r_trail,
272 273 274 275 276 277 278 279
						&timeout, sizeof(timeout));
		}
	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;
280 281 282 283 284 285 286 287
	default:
		pr_err("unrecognized osd opcode %d\n", dst->op);
		WARN_ON(1);
		break;
	}
	dst->payload_len = cpu_to_le32(src->payload_len);
}

288 289 290 291 292
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req,
A
Alex Elder 已提交
293
			     u64 off, u64 len, unsigned int num_op,
294
			     struct ceph_osd_req_op *src_ops,
295
			     struct ceph_snap_context *snapc, u64 snap_id,
296
			     struct timespec *mtime)
297 298 299
{
	struct ceph_msg *msg = req->r_request;
	struct ceph_osd_request_head *head;
300
	struct ceph_osd_req_op *src_op;
301 302 303 304
	struct ceph_osd_op *op;
	void *p;
	size_t msg_size = sizeof(*head) + num_op*sizeof(*op);
	int flags = req->r_flags;
305 306
	u64 data_len = 0;
	int i;
307

308 309
	WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0);

S
Sage Weil 已提交
310
	head = msg->front.iov_base;
311
	head->snapid = cpu_to_le64(snap_id);
S
Sage Weil 已提交
312 313 314 315 316 317 318 319 320
	op = (void *)(head + 1);
	p = (void *)(op + num_op);

	req->r_snapc = ceph_get_snap_context(snapc);

	head->client_inc = cpu_to_le32(1); /* always, for now. */
	head->flags = cpu_to_le32(flags);
	if (flags & CEPH_OSD_FLAG_WRITE)
		ceph_encode_timespec(&head->mtime, mtime);
A
Alex Elder 已提交
321
	BUG_ON(num_op > (unsigned int) ((u16) -1));
S
Sage Weil 已提交
322 323 324
	head->num_ops = cpu_to_le16(num_op);

	/* fill in oid */
325 326 327
	head->object_len = cpu_to_le32(req->r_oid_len);
	memcpy(p, req->r_oid, req->r_oid_len);
	p += req->r_oid_len;
S
Sage Weil 已提交
328

329
	src_op = src_ops;
A
Alex Elder 已提交
330 331
	while (num_op--)
		osd_req_encode_op(req, op++, src_op++);
332

333
	data_len += req->r_trail.length;
334

S
Sage Weil 已提交
335 336 337 338 339 340 341 342 343
	if (snapc) {
		head->snap_seq = cpu_to_le64(snapc->seq);
		head->num_snaps = cpu_to_le32(snapc->num_snaps);
		for (i = 0; i < snapc->num_snaps; i++) {
			put_unaligned_le64(snapc->snaps[i], p);
			p += sizeof(u64);
		}
	}

344 345
	if (flags & CEPH_OSD_FLAG_WRITE) {
		req->r_request->hdr.data_off = cpu_to_le16(off);
346
		req->r_request->hdr.data_len = cpu_to_le32(len + data_len);
347 348 349 350 351
	} else if (data_len) {
		req->r_request->hdr.data_off = 0;
		req->r_request->hdr.data_len = cpu_to_le32(data_len);
	}

352 353
	req->r_request->page_alignment = req->r_page_alignment;

S
Sage Weil 已提交
354
	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
355 356 357
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);
358 359
	return;
}
360
EXPORT_SYMBOL(ceph_osdc_build_request);
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382

/*
 * build new request AND message, calculate layout, and adjust file
 * extent as needed.
 *
 * if the file was recently truncated, we include information about its
 * old and new size so that the object can be updated appropriately.  (we
 * avoid synchronously deleting truncated objects because it's slow.)
 *
 * if @do_sync, include a 'startsync' command so that the osd will flush
 * data quickly.
 */
struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
					       struct ceph_file_layout *layout,
					       struct ceph_vino vino,
					       u64 off, u64 *plen,
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       int do_sync,
					       u32 truncate_seq,
					       u64 truncate_size,
					       struct timespec *mtime,
383
					       bool use_mempool,
384
					       int page_align)
385
{
A
Alex Elder 已提交
386
	struct ceph_osd_req_op ops[2];
387
	struct ceph_osd_request *req;
A
Alex Elder 已提交
388
	unsigned int num_op = 1;
389
	int r;
390

A
Alex Elder 已提交
391 392
	memset(&ops, 0, sizeof ops);

393 394 395 396 397 398
	ops[0].op = opcode;
	ops[0].extent.truncate_seq = truncate_seq;
	ops[0].extent.truncate_size = truncate_size;

	if (do_sync) {
		ops[1].op = CEPH_OSD_OP_STARTSYNC;
A
Alex Elder 已提交
399 400
		num_op++;
	}
401

A
Alex Elder 已提交
402 403
	req = ceph_osdc_alloc_request(osdc, snapc, num_op, use_mempool,
					GFP_NOFS);
404
	if (!req)
405
		return ERR_PTR(-ENOMEM);
406
	req->r_flags = flags;
407 408

	/* calculate max write size */
409
	r = calc_layout(vino, layout, off, plen, req, ops);
410 411
	if (r < 0)
		return ERR_PTR(r);
412 413
	req->r_file_layout = *layout;  /* keep a copy */

414 415 416
	/* in case it differs from natural (file) alignment that
	   calc_layout filled in for us */
	req->r_num_pages = calc_pages_for(page_align, *plen);
417 418
	req->r_page_alignment = page_align;

A
Alex Elder 已提交
419 420
	ceph_osdc_build_request(req, off, *plen, num_op, ops,
				snapc, vino.snap, mtime);
421

S
Sage Weil 已提交
422 423
	return req;
}
424
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490

/*
 * 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;
}

491 492 493 494 495 496
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
497
	struct ceph_osd_request *req, *nreq;
498 499 500 501
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
502
	if (err)
503 504 505 506 507 508
		return;

	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
		list_move(&req->r_req_lru_item, &osdc->req_unsent);
		dout("requeued %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
509 510 511 512 513 514
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}

	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
515 516 517 518 519
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
520
		__register_request(osdc, req);
521 522 523
		list_add(&req->r_req_lru_item, &osdc->req_unsent);
		list_add(&req->r_osd_item, &req->r_osd->o_requests);
		__unregister_linger_request(osdc, req);
524 525
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
526 527 528
	}
}

S
Sage Weil 已提交
529
/*
530
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
531 532 533 534 535 536 537 538 539 540 541
 */
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
542 543
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
544
	__send_queued(osdc);
S
Sage Weil 已提交
545
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
546 547 548 549 550 551
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
552
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
553 554 555 556 557 558 559 560 561
{
	struct ceph_osd *osd;

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

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
562
	osd->o_osd = onum;
563
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
564
	INIT_LIST_HEAD(&osd->o_requests);
565
	INIT_LIST_HEAD(&osd->o_linger_requests);
566
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
567 568
	osd->o_incarnation = 1;

569
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
570

571
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590
	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);
591
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
592 593
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

594
		if (ac->ops && ac->ops->destroy_authorizer)
595
			ac->ops->destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
596
		kfree(osd);
S
Sage Weil 已提交
597
	}
S
Sage Weil 已提交
598 599 600 601 602
}

/*
 * remove an osd from our map
 */
603
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
604
{
605
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
606 607
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
608
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
609 610 611 612
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

613 614
static void remove_all_osds(struct ceph_osd_client *osdc)
{
615
	dout("%s %p\n", __func__, osdc);
616 617 618 619 620 621 622 623 624
	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);
}

625 626 627 628 629 630
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);
631
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
632 633 634 635 636 637 638 639 640
}

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);
}

641
static void remove_old_osds(struct ceph_osd_client *osdc)
642 643 644 645 646 647
{
	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) {
648
		if (time_before(jiffies, osd->lru_ttl))
649 650 651 652 653 654
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
655 656 657
/*
 * reset osd connect
 */
658
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
659
{
A
Alex Elder 已提交
660
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
661

662
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
663 664
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
665
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
666 667 668 669 670 671 672 673 674

		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;

675 676 677 678 679
		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
680 681

		return -EAGAIN;
S
Sage Weil 已提交
682
	}
A
Alex Elder 已提交
683 684 685 686 687 688

	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
689 690 691 692 693 694 695 696
}

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;

697
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729
	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;
}

730 731 732
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
733
			osdc->client->options->osd_keepalive_timeout * HZ);
734 735 736 737 738 739
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
740 741 742 743 744

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
745 746
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
747 748
{
	req->r_tid = ++osdc->last_tid;
749
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
750
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
751 752 753 754
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
755 756
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
757
	}
758 759 760 761 762 763 764
}

static void register_request(struct ceph_osd_client *osdc,
			     struct ceph_osd_request *req)
{
	mutex_lock(&osdc->request_mutex);
	__register_request(osdc, req);
S
Sage Weil 已提交
765 766 767 768 769 770 771 772 773
	mutex_unlock(&osdc->request_mutex);
}

/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
774 775 776 777 778 779
	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
780 781 782 783
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

784 785
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
786
		ceph_msg_revoke(req->r_request);
787 788

		list_del_init(&req->r_osd_item);
789 790 791
		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);
792
			__move_osd_to_lru(osdc, req->r_osd);
793
		}
794
		if (list_empty(&req->r_linger_item))
795
			req->r_osd = NULL;
796
	}
S
Sage Weil 已提交
797

A
Alex Elder 已提交
798
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
799 800
	ceph_osdc_put_request(req);

801 802 803
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
804 805 806 807 808 809 810 811
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
812
	if (req->r_sent && req->r_osd) {
813
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
814 815 816 817
		req->r_sent = 0;
	}
}

818 819 820 821 822
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);
823 824 825
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
826 827 828 829 830 831
}

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
832
	list_del_init(&req->r_linger_item);
833 834 835 836 837 838 839 840
	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);
		}
841 842
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
	}
}

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
873 874 875
/*
 * 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
876
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
877
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
878 879 880 881 882
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
883
static int __map_request(struct ceph_osd_client *osdc,
884
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
885 886
{
	struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
887
	struct ceph_pg pgid;
888 889
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
890 891
	int err;

892
	dout("map_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
893 894
	err = ceph_calc_object_layout(&reqhead->layout, req->r_oid,
				      &req->r_file_layout, osdc->osdmap);
895 896
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
897
		return err;
898
	}
899
	pgid = reqhead->layout.ol_pgid;
900 901
	req->r_pgid = pgid;

902 903 904 905 906
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
907

908 909
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
910 911 912
	     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
913 914 915
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

916
	dout("map_request tid %llu pgid %d.%x osd%d (was osd%d)\n",
917
	     req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o,
S
Sage Weil 已提交
918 919
	     req->r_osd ? req->r_osd->o_osd : -1);

920 921 922 923
	/* 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
924 925 926 927 928 929 930 931
	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) {
932
		err = -ENOMEM;
933
		req->r_osd = create_osd(osdc, o);
934 935
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
936
			goto out;
937
		}
S
Sage Weil 已提交
938

939
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
940 941
		__insert_osd(osdc, req->r_osd);

942 943 944
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
945 946
	}

947 948
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
S
Sage Weil 已提交
949
		list_add(&req->r_osd_item, &req->r_osd->o_requests);
950 951 952
		list_move(&req->r_req_lru_item, &osdc->req_unsent);
	} else {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
953
	}
954
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
955 956 957 958 959 960 961 962

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
963 964
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
965 966 967 968 969 970 971 972 973 974 975
{
	struct ceph_osd_request_head *reqhead;

	dout("send_request %p tid %llu to osd%d flags %d\n",
	     req, req->r_tid, req->r_osd->o_osd, req->r_flags);

	reqhead = req->r_request->front.iov_base;
	reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch);
	reqhead->flags |= cpu_to_le32(req->r_flags);  /* e.g., RETRY */
	reqhead->reassert_version = req->r_reassert_version;

S
Sage Weil 已提交
976
	req->r_stamp = jiffies;
977
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
978 979 980 981 982 983

	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;
}

984 985 986
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
987
static void __send_queued(struct ceph_osd_client *osdc)
988 989 990
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
991 992
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
993 994 995
		__send_request(osdc, req);
}

S
Sage Weil 已提交
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
/*
 * 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
1009
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1010
	struct ceph_osd *osd;
1011
	unsigned long keepalive =
1012
		osdc->client->options->osd_keepalive_timeout * HZ;
1013
	struct list_head slow_osds;
S
Sage Weil 已提交
1014 1015 1016 1017 1018 1019 1020
	dout("timeout\n");
	down_read(&osdc->map_sem);

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

	mutex_lock(&osdc->request_mutex);

1021 1022 1023 1024 1025 1026 1027
	/*
	 * 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
1028
		if (time_before(jiffies, req->r_stamp + keepalive))
1029 1030 1031 1032 1033
			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
1034
		     req->r_tid, osd->o_osd);
1035 1036 1037 1038 1039 1040
		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
1041 1042 1043
		ceph_con_keepalive(&osd->o_con);
	}

1044
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1045
	__send_queued(osdc);
S
Sage Weil 已提交
1046 1047 1048 1049
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1050 1051 1052 1053 1054 1055
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 =
1056
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1057 1058 1059

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1060
	remove_old_osds(osdc);
1061 1062 1063 1064 1065 1066
	up_read(&osdc->map_sem);

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

S
Sage Weil 已提交
1067 1068 1069 1070 1071 1072 1073
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
1074 1075 1076 1077
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1078 1079
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1080 1081 1082 1083 1084
{
	struct ceph_osd_reply_head *rhead = msg->front.iov_base;
	struct ceph_osd_request *req;
	u64 tid;
	int numops, object_len, flags;
1085
	s32 result;
S
Sage Weil 已提交
1086

1087
	tid = le64_to_cpu(msg->hdr.tid);
S
Sage Weil 已提交
1088 1089 1090 1091
	if (msg->front.iov_len < sizeof(*rhead))
		goto bad;
	numops = le32_to_cpu(rhead->num_ops);
	object_len = le32_to_cpu(rhead->object_len);
1092
	result = le32_to_cpu(rhead->result);
S
Sage Weil 已提交
1093 1094 1095
	if (msg->front.iov_len != sizeof(*rhead) + object_len +
	    numops * sizeof(struct ceph_osd_op))
		goto bad;
1096
	dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result);
S
Sage Weil 已提交
1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
		mutex_unlock(&osdc->request_mutex);
		return;
	}
	ceph_osdc_get_request(req);
	flags = le32_to_cpu(rhead->flags);

1108
	/*
1109
	 * if this connection filled our message, drop our reference now, to
1110 1111
	 * avoid a (safe but slower) revoke later.
	 */
1112
	if (req->r_con_filling_msg == con && req->r_reply == msg) {
S
Sage Weil 已提交
1113
		dout(" dropping con_filling_msg ref %p\n", con);
1114
		req->r_con_filling_msg = NULL;
1115
		con->ops->put(con);
1116 1117
	}

S
Sage Weil 已提交
1118
	if (!req->r_got_reply) {
1119
		unsigned int bytes;
S
Sage Weil 已提交
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133

		req->r_result = le32_to_cpu(rhead->result);
		bytes = le32_to_cpu(msg->hdr.data_len);
		dout("handle_reply result %d bytes %d\n", req->r_result,
		     bytes);
		if (req->r_result == 0)
			req->r_result = bytes;

		/* in case this is a write and we need to replay, */
		req->r_reassert_version = rhead->reassert_version;

		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
1134
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1135 1136 1137 1138 1139
		goto done;
	}

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

1140 1141 1142
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1143
	/* either this is a read, or we got the safe response */
1144 1145
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1146 1147 1148 1149 1150 1151 1152 1153
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

	mutex_unlock(&osdc->request_mutex);

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1154
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1155

S
Sage Weil 已提交
1156 1157
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1158 1159

done:
1160
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1161 1162 1163 1164 1165 1166 1167
	ceph_osdc_put_request(req);
	return;

bad:
	pr_err("corrupt osd_op_reply got %d %d expected %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len),
	       (int)sizeof(*rhead));
1168
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1169 1170
}

1171
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1172 1173 1174
{
	struct rb_node *p, *n;

1175 1176
	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
1177

1178 1179 1180 1181 1182 1183 1184
		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
1185
	}
1186 1187 1188
}

/*
1189 1190
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1191
 *
1192
 * Caller should hold map_sem for read.
1193
 */
1194
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1195
{
1196
	struct ceph_osd_request *req, *nreq;
1197 1198 1199
	struct rb_node *p;
	int needmap = 0;
	int err;
1200

1201
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1202
	mutex_lock(&osdc->request_mutex);
1203
	for (p = rb_first(&osdc->requests); p; ) {
1204
		req = rb_entry(p, struct ceph_osd_request, r_node);
1205
		p = rb_next(p);
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223

		/*
		 * 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;
		}

1224
		err = __map_request(osdc, req, force_resend);
1225 1226 1227 1228 1229 1230
		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) {
1231 1232 1233 1234
			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);
1235
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1236 1237
			}
		}
1238 1239 1240 1241 1242 1243
	}

	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);

1244
		err = __map_request(osdc, req, force_resend);
1245
		dout("__map_request returned %d\n", err);
1246 1247 1248 1249 1250 1251 1252 1253
		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;
1254
		}
1255 1256 1257 1258

		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);
1259
		__unregister_linger_request(osdc, req);
1260
	}
S
Sage Weil 已提交
1261 1262 1263 1264 1265 1266
	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);
	}
1267
	reset_changed_osds(osdc);
1268
}
1269 1270


S
Sage Weil 已提交
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293
/*
 * 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));
1294 1295
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1296 1297 1298 1299 1300 1301 1302 1303

	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);
1304 1305
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1306 1307 1308 1309 1310 1311 1312
		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,
1313
							  &osdc->client->msgr);
S
Sage Weil 已提交
1314 1315 1316 1317
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1318
			BUG_ON(!newmap);
S
Sage Weil 已提交
1319 1320 1321 1322
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1323
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338
		} 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);
1339 1340
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1341 1342 1343 1344 1345 1346 1347 1348 1349
		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 {
1350 1351
			int skipped_map = 0;

S
Sage Weil 已提交
1352 1353 1354 1355 1356 1357
			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;
			}
1358
			BUG_ON(!newmap);
S
Sage Weil 已提交
1359 1360
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1361 1362 1363
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1364
				ceph_osdmap_destroy(oldmap);
1365 1366
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1367 1368 1369 1370 1371 1372 1373 1374
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1375 1376 1377 1378 1379 1380 1381 1382 1383

	/*
	 * 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
1384 1385 1386
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1387
	up_read(&osdc->map_sem);
1388
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1389 1390 1391 1392
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1393
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1394 1395 1396 1397
	up_write(&osdc->map_sem);
	return;
}

1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494
/*
 * 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 *),
			   int one_shot, void *data,
			   struct ceph_osd_event **pevent)
{
	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;
	event->one_shot = one_shot;
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1495
	RB_CLEAR_NODE(&event->node);
1496 1497 1498 1499 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
	kref_init(&event->kref);   /* one ref for us */
	kref_get(&event->kref);    /* one ref for the caller */
	init_completion(&event->completion);

	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);
	complete(&event->completion);
	dout("do_event_work completed %p\n", event);
	ceph_osdc_put_event(event);
	kfree(event_work);
}


/*
 * Process osd watch notifications
 */
void handle_watch_notify(struct ceph_osd_client *osdc, struct ceph_msg *msg)
{
	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) {
		get_event(event);
		if (event->one_shot)
			__remove_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;
		}
1578
		INIT_WORK(&event_work->work, do_event_work);
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
		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:
	complete(&event->completion);
	ceph_osdc_put_event(event);
	return;

bad:
	pr_err("osdc handle_watch_notify corrupt msg\n");
	return;
}

int ceph_osdc_wait_event(struct ceph_osd_event *event, unsigned long timeout)
{
	int err;

	dout("wait_event %p\n", event);
	err = wait_for_completion_interruptible_timeout(&event->completion,
							timeout * HZ);
	ceph_osdc_put_event(event);
	if (err > 0)
		err = 0;
	dout("wait_event %p returns %d\n", event, err);
	return err;
}
EXPORT_SYMBOL(ceph_osdc_wait_event);

S
Sage Weil 已提交
1616 1617 1618 1619 1620 1621 1622
/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
S
Sage Weil 已提交
1623
	int rc = 0;
S
Sage Weil 已提交
1624 1625 1626

	req->r_request->pages = req->r_pages;
	req->r_request->nr_pages = req->r_num_pages;
1627 1628 1629
#ifdef CONFIG_BLOCK
	req->r_request->bio = req->r_bio;
#endif
1630
	req->r_request->trail = &req->r_trail;
S
Sage Weil 已提交
1631 1632 1633 1634 1635

	register_request(osdc, req);

	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
S
Sage Weil 已提交
1636 1637 1638 1639 1640 1641
	/*
	 * a racing kick_requests() may have sent the message for us
	 * while we dropped request_mutex above, so only send now if
	 * the request still han't been touched yet.
	 */
	if (req->r_sent == 0) {
1642
		rc = __map_request(osdc, req, 0);
1643 1644 1645 1646 1647 1648
		if (rc < 0) {
			if (nofail) {
				dout("osdc_start_request failed map, "
				     " will retry %lld\n", req->r_tid);
				rc = 0;
			}
1649
			goto out_unlock;
1650
		}
1651 1652 1653 1654
		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 {
1655
			__send_request(osdc, req);
S
Sage Weil 已提交
1656
		}
1657
		rc = 0;
S
Sage Weil 已提交
1658
	}
1659 1660

out_unlock:
S
Sage Weil 已提交
1661 1662 1663 1664
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
1665
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678

/*
 * 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);
1679
		__unregister_request(osdc, req);
S
Sage Weil 已提交
1680
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1681
		complete_request(req);
1682
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
1683 1684 1685 1686 1687 1688
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
1689
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722

/*
 * 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);
}
1723
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740

/*
 * 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;
1741
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
1742
	osdc->requests = RB_ROOT;
1743
	INIT_LIST_HEAD(&osdc->req_lru);
1744 1745
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
1746
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
1747 1748
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
1749
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1750 1751 1752
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
1753 1754

	schedule_delayed_work(&osdc->osds_timeout_work,
1755
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
1756

1757
	err = -ENOMEM;
S
Sage Weil 已提交
1758 1759 1760
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
1761
		goto out;
S
Sage Weil 已提交
1762

1763 1764
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
1765
				"osd_op");
S
Sage Weil 已提交
1766
	if (err < 0)
1767
		goto out_mempool;
1768
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
1769 1770
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
1771 1772
	if (err < 0)
		goto out_msgpool;
1773 1774 1775 1776 1777 1778 1779

	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
1780
	return 0;
1781

S
Sage Weil 已提交
1782 1783
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
1784 1785 1786 1787
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
1788 1789 1790 1791
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
1792 1793
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
1794
	cancel_delayed_work_sync(&osdc->timeout_work);
1795
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
1796 1797 1798 1799
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
1800
	remove_all_osds(osdc);
S
Sage Weil 已提交
1801 1802
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
1803
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
}

/*
 * 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,
1814
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
1815 1816 1817 1818 1819 1820 1821 1822 1823
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
				    NULL, 0, truncate_seq, truncate_size, NULL,
1824
				    false, page_align);
1825 1826
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
1827 1828 1829 1830

	/* it may be a short read due to an object boundary */
	req->r_pages = pages;

1831 1832
	dout("readpages  final extent is %llu~%llu (%d pages align %d)\n",
	     off, *plen, req->r_num_pages, page_align);
S
Sage Weil 已提交
1833 1834 1835 1836 1837 1838 1839 1840 1841

	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;
}
1842
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
1843 1844 1845 1846 1847 1848 1849 1850 1851 1852

/*
 * 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,
1853
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
1854 1855 1856
{
	struct ceph_osd_request *req;
	int rc = 0;
1857
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
1858 1859 1860 1861

	BUG_ON(vino.snap != CEPH_NOSNAP);
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
				    CEPH_OSD_OP_WRITE,
1862
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
1863
				    snapc, 0,
S
Sage Weil 已提交
1864
				    truncate_seq, truncate_size, mtime,
1865
				    true, page_align);
1866 1867
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
1868 1869 1870 1871 1872 1873

	/* it may be a short write due to an object boundary */
	req->r_pages = pages;
	dout("writepages %llu~%llu (%d pages)\n", off, len,
	     req->r_num_pages);

1874
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
1875 1876 1877 1878 1879 1880 1881 1882 1883
	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;
}
1884
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
1885 1886 1887 1888 1889 1890 1891

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
1892
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
1893 1894 1895
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
1896
		goto out;
1897
	osdc = osd->o_osdc;
S
Sage Weil 已提交
1898 1899 1900 1901 1902 1903

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
1904
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
1905
		break;
1906 1907 1908
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
1909 1910 1911 1912 1913

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
1914
out:
S
Sage Weil 已提交
1915 1916 1917
	ceph_msg_put(msg);
}

1918
/*
1919 1920
 * lookup and return message for incoming reply.  set up reply message
 * pages.
1921 1922
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
1923 1924
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
1925 1926 1927
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
1928
	struct ceph_msg *m;
1929
	struct ceph_osd_request *req;
1930 1931
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
1932
	u64 tid;
S
Sage Weil 已提交
1933

1934 1935 1936 1937 1938 1939
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
1940 1941
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
1942 1943
		goto out;
	}
S
Sage Weil 已提交
1944 1945

	if (req->r_con_filling_msg) {
1946
		dout("%s revoking msg %p from old con %p\n", __func__,
S
Sage Weil 已提交
1947
		     req->r_reply, req->r_con_filling_msg);
1948
		ceph_msg_revoke_incoming(req->r_reply);
1949
		req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
1950
		req->r_con_filling_msg = NULL;
1951 1952
	}

S
Sage Weil 已提交
1953 1954 1955
	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);
1956
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
1957
		if (!m)
S
Sage Weil 已提交
1958 1959 1960 1961 1962 1963
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

1964
	if (data_len > 0) {
1965
		int want = calc_pages_for(req->r_page_alignment, data_len);
1966

1967
		if (req->r_pages && unlikely(req->r_num_pages < want)) {
1968 1969 1970
			pr_warning("tid %lld reply has %d bytes %d pages, we"
				   " had only %d pages ready\n", tid, data_len,
				   want, req->r_num_pages);
1971 1972
			*skip = 1;
			ceph_msg_put(m);
1973
			m = NULL;
1974
			goto out;
1975
		}
1976 1977
		m->pages = req->r_pages;
		m->nr_pages = req->r_num_pages;
1978
		m->page_alignment = req->r_page_alignment;
1979 1980 1981
#ifdef CONFIG_BLOCK
		m->bio = req->r_bio;
#endif
1982
	}
1983
	*skip = 0;
1984
	req->r_con_filling_msg = con->ops->get(con);
S
Sage Weil 已提交
1985
	dout("get_reply tid %lld %p\n", tid, m);
1986 1987 1988

out:
	mutex_unlock(&osdc->request_mutex);
1989
	return m;
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

}

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
2001
	*skip = 0;
2002 2003
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2004
	case CEPH_MSG_WATCH_NOTIFY:
2005
		return ceph_msg_new(type, front, GFP_NOFS, false);
2006 2007 2008 2009 2010 2011 2012 2013
	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
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
}

/*
 * 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);
}

2033 2034 2035
/*
 * authentication
 */
2036 2037 2038 2039 2040
/*
 * 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,
2041
					int *proto, int force_new)
2042 2043 2044 2045
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2046
	struct ceph_auth_handshake *auth = &o->o_auth;
2047

2048
	if (force_new && auth->authorizer) {
2049 2050
		if (ac->ops && ac->ops->destroy_authorizer)
			ac->ops->destroy_authorizer(ac, auth->authorizer);
2051 2052
		auth->authorizer = NULL;
	}
2053
	if (!auth->authorizer && ac->ops && ac->ops->create_authorizer) {
2054 2055
		int ret = ac->ops->create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
							auth);
2056
		if (ret)
2057
			return ERR_PTR(ret);
2058 2059
	}
	*proto = ac->protocol;
2060

2061
	return auth;
2062 2063 2064 2065 2066 2067 2068 2069 2070
}


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;

2071 2072 2073 2074
	/*
	 * XXX If ac->ops or ac->ops->verify_authorizer_reply is null,
	 * XXX which do we do:  succeed or fail?
	 */
2075
	return ac->ops->verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2076 2077
}

2078 2079 2080 2081 2082 2083
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;

2084
	if (ac->ops && ac->ops->invalidate_authorizer)
2085 2086 2087 2088
		ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);

	return ceph_monc_validate_auth(&osdc->client->monc);
}
2089

2090
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2091 2092 2093
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2094 2095
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2096
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2097
	.alloc_msg = alloc_msg,
2098
	.fault = osd_reset,
S
Sage Weil 已提交
2099
};