net/ceph/osd_client.c

   1
   2 #include <linux/ceph/ceph_debug.h>
   3
   4 #include <linux/module.h>
   5 #include <linux/err.h>
   6 #include <linux/highmem.h>
   7 #include <linux/mm.h>
   8 #include <linux/pagemap.h>
   9 #include <linux/slab.h>
  10 #include <linux/uaccess.h>
  11 #ifdef CONFIG_BLOCK
  12 #include <linux/bio.h>
  13 #endif
  14
  15 #include <linux/ceph/libceph.h>
  16 #include <linux/ceph/osd_client.h>
  17 #include <linux/ceph/messenger.h>
  18 #include <linux/ceph/decode.h>
  19 #include <linux/ceph/auth.h>
  20 #include <linux/ceph/pagelist.h>
  21
  22 #define OSD_OP_FRONT_LEN        4096
  23 #define OSD_OPREPLY_FRONT_LEN   512
  24
  25 static struct kmem_cache        *ceph_osd_request_cache;
  26
  27 static const struct ceph_connection_operations osd_con_ops;
  28
  29 static void __send_queued(struct ceph_osd_client *osdc);
  30 static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
  31 static void __register_request(struct ceph_osd_client *osdc,
  32                                struct ceph_osd_request *req);
  33 static void __unregister_linger_request(struct ceph_osd_client *osdc,
  34                                         struct ceph_osd_request *req);
  35 static void __send_request(struct ceph_osd_client *osdc,
  36                            struct ceph_osd_request *req);
  37
  38 /*
  39  * Implement client access to distributed object storage cluster.
  40  *
  41  * All data objects are stored within a cluster/cloud of OSDs, or
  42  * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
  43  * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
  44  * remote daemons serving up and coordinating consistent and safe
  45  * access to storage.
  46  *
  47  * Cluster membership and the mapping of data objects onto storage devices
  48  * are described by the osd map.
  49  *
  50  * We keep track of pending OSD requests (read, write), resubmit
  51  * requests to different OSDs when the cluster topology/data layout
  52  * change, or retry the affected requests when the communications
  53  * channel with an OSD is reset.
  54  */
  55
  56 /*
  57  * calculate the mapping of a file extent onto an object, and fill out the
  58  * request accordingly.  shorten extent as necessary if it crosses an
  59  * object boundary.
  60  *
  61  * fill osd op in request message.
  62  */
  63 static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
  64                         u64 *objnum, u64 *objoff, u64 *objlen)
  65 {
  66         u64 orig_len = *plen;
  67         int r;
  68
  69         /* object extent? */
  70         r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
  71                                           objoff, objlen);
  72         if (r < 0)
  73                 return r;
  74         if (*objlen < orig_len) {
  75                 *plen = *objlen;
  76                 dout(" skipping last %llu, final file extent %llu~%llu\n",
  77                      orig_len - *plen, off, *plen);
  78         }
  79
  80         dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
  81
  82         return 0;
  83 }
  84
  85 static void ceph_osd_data_init(struct ceph_osd_data *osd_data)
  86 {
  87         memset(osd_data, 0, sizeof (*osd_data));
  88         osd_data->type = CEPH_OSD_DATA_TYPE_NONE;
  89 }
  90
  91 static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
  92                         struct page **pages, u64 length, u32 alignment,
  93                         bool pages_from_pool, bool own_pages)
  94 {
  95         osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
  96         osd_data->pages = pages;
  97         osd_data->length = length;
  98         osd_data->alignment = alignment;
  99         osd_data->pages_from_pool = pages_from_pool;
 100         osd_data->own_pages = own_pages;
 101 }
 102
 103 static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
 104                         struct ceph_pagelist *pagelist)
 105 {
 106         osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
 107         osd_data->pagelist = pagelist;
 108 }
 109
 110 #ifdef CONFIG_BLOCK
 111 static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
 112                         struct bio *bio, size_t bio_length)
 113 {
 114         osd_data->type = CEPH_OSD_DATA_TYPE_BIO;
 115         osd_data->bio = bio;
 116         osd_data->bio_length = bio_length;
 117 }
 118 #endif /* CONFIG_BLOCK */
 119
 120 #define osd_req_op_data(oreq, whch, typ, fld)   \
 121         ({                                              \
 122                 BUG_ON(whch >= (oreq)->r_num_ops);      \
 123                 &(oreq)->r_ops[whch].typ.fld;           \
 124         })
 125
 126 static struct ceph_osd_data *
 127 osd_req_op_raw_data_in(struct ceph_osd_request *osd_req, unsigned int which)
 128 {
 129         BUG_ON(which >= osd_req->r_num_ops);
 130
 131         return &osd_req->r_ops[which].raw_data_in;
 132 }
 133
 134 struct ceph_osd_data *
 135 osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
 136                         unsigned int which)
 137 {
 138         return osd_req_op_data(osd_req, which, extent, osd_data);
 139 }
 140 EXPORT_SYMBOL(osd_req_op_extent_osd_data);
 141
 142 struct ceph_osd_data *
 143 osd_req_op_cls_response_data(struct ceph_osd_request *osd_req,
 144                         unsigned int which)
 145 {
 146         return osd_req_op_data(osd_req, which, cls, response_data);
 147 }
 148 EXPORT_SYMBOL(osd_req_op_cls_response_data);    /* ??? */
 149
 150 void osd_req_op_raw_data_in_pages(struct ceph_osd_request *osd_req,
 151                         unsigned int which, struct page **pages,
 152                         u64 length, u32 alignment,
 153                         bool pages_from_pool, bool own_pages)
 154 {
 155         struct ceph_osd_data *osd_data;
 156
 157         osd_data = osd_req_op_raw_data_in(osd_req, which);
 158         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 159                                 pages_from_pool, own_pages);
 160 }
 161 EXPORT_SYMBOL(osd_req_op_raw_data_in_pages);
 162
 163 void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
 164                         unsigned int which, struct page **pages,
 165                         u64 length, u32 alignment,
 166                         bool pages_from_pool, bool own_pages)
 167 {
 168         struct ceph_osd_data *osd_data;
 169
 170         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 171         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 172                                 pages_from_pool, own_pages);
 173 }
 174 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);
 175
 176 void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
 177                         unsigned int which, struct ceph_pagelist *pagelist)
 178 {
 179         struct ceph_osd_data *osd_data;
 180
 181         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 182         ceph_osd_data_pagelist_init(osd_data, pagelist);
 183 }
 184 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);
 185
 186 #ifdef CONFIG_BLOCK
 187 void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
 188                         unsigned int which, struct bio *bio, size_t bio_length)
 189 {
 190         struct ceph_osd_data *osd_data;
 191
 192         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 193         ceph_osd_data_bio_init(osd_data, bio, bio_length);
 194 }
 195 EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio);
 196 #endif /* CONFIG_BLOCK */
 197
 198 static void osd_req_op_cls_request_info_pagelist(
 199                         struct ceph_osd_request *osd_req,
 200                         unsigned int which, struct ceph_pagelist *pagelist)
 201 {
 202         struct ceph_osd_data *osd_data;
 203
 204         osd_data = osd_req_op_data(osd_req, which, cls, request_info);
 205         ceph_osd_data_pagelist_init(osd_data, pagelist);
 206 }
 207
 208 void osd_req_op_cls_request_data_pagelist(
 209                         struct ceph_osd_request *osd_req,
 210                         unsigned int which, struct ceph_pagelist *pagelist)
 211 {
 212         struct ceph_osd_data *osd_data;
 213
 214         osd_data = osd_req_op_data(osd_req, which, cls, request_data);
 215         ceph_osd_data_pagelist_init(osd_data, pagelist);
 216 }
 217 EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);
 218
 219 void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req,
 220                         unsigned int which, struct page **pages, u64 length,
 221                         u32 alignment, bool pages_from_pool, bool own_pages)
 222 {
 223         struct ceph_osd_data *osd_data;
 224
 225         osd_data = osd_req_op_data(osd_req, which, cls, request_data);
 226         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 227                                 pages_from_pool, own_pages);
 228 }
 229 EXPORT_SYMBOL(osd_req_op_cls_request_data_pages);
 230
 231 void osd_req_op_cls_response_data_pages(struct ceph_osd_request *osd_req,
 232                         unsigned int which, struct page **pages, u64 length,
 233                         u32 alignment, bool pages_from_pool, bool own_pages)
 234 {
 235         struct ceph_osd_data *osd_data;
 236
 237         osd_data = osd_req_op_data(osd_req, which, cls, response_data);
 238         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 239                                 pages_from_pool, own_pages);
 240 }
 241 EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);
 242
 243 static u64 ceph_osd_data_length(struct ceph_osd_data *osd_data)
 244 {
 245         switch (osd_data->type) {
 246         case CEPH_OSD_DATA_TYPE_NONE:
 247                 return 0;
 248         case CEPH_OSD_DATA_TYPE_PAGES:
 249                 return osd_data->length;
 250         case CEPH_OSD_DATA_TYPE_PAGELIST:
 251                 return (u64)osd_data->pagelist->length;
 252 #ifdef CONFIG_BLOCK
 253         case CEPH_OSD_DATA_TYPE_BIO:
 254                 return (u64)osd_data->bio_length;
 255 #endif /* CONFIG_BLOCK */
 256         default:
 257                 WARN(true, "unrecognized data type %d\n", (int)osd_data->type);
 258                 return 0;
 259         }
 260 }
 261
 262 static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
 263 {
 264         if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
 265                 int num_pages;
 266
 267                 num_pages = calc_pages_for((u64)osd_data->alignment,
 268                                                 (u64)osd_data->length);
 269                 ceph_release_page_vector(osd_data->pages, num_pages);
 270         }
 271         ceph_osd_data_init(osd_data);
 272 }
 273
 274 static void osd_req_op_data_release(struct ceph_osd_request *osd_req,
 275                         unsigned int which)
 276 {
 277         struct ceph_osd_req_op *op;
 278
 279         BUG_ON(which >= osd_req->r_num_ops);
 280         op = &osd_req->r_ops[which];
 281
 282         switch (op->op) {
 283         case CEPH_OSD_OP_READ:
 284         case CEPH_OSD_OP_WRITE:
 285                 ceph_osd_data_release(&op->extent.osd_data);
 286                 break;
 287         case CEPH_OSD_OP_CALL:
 288                 ceph_osd_data_release(&op->cls.request_info);
 289                 ceph_osd_data_release(&op->cls.request_data);
 290                 ceph_osd_data_release(&op->cls.response_data);
 291                 break;
 292         default:
 293                 break;
 294         }
 295 }
 296
 297 /*
 298  * requests
 299  */
 300 void ceph_osdc_release_request(struct kref *kref)
 301 {
 302         struct ceph_osd_request *req;
 303         unsigned int which;
 304
 305         req = container_of(kref, struct ceph_osd_request, r_kref);
 306         if (req->r_request)
 307                 ceph_msg_put(req->r_request);
 308         if (req->r_reply) {
 309                 ceph_msg_revoke_incoming(req->r_reply);
 310                 ceph_msg_put(req->r_reply);
 311         }
 312
 313         for (which = 0; which < req->r_num_ops; which++)
 314                 osd_req_op_data_release(req, which);
 315
 316         ceph_put_snap_context(req->r_snapc);
 317         if (req->r_mempool)
 318                 mempool_free(req, req->r_osdc->req_mempool);
 319         else
 320                 kmem_cache_free(ceph_osd_request_cache, req);
 321
 322 }
 323 EXPORT_SYMBOL(ceph_osdc_release_request);
 324
 325 struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
 326                                                struct ceph_snap_context *snapc,
 327                                                unsigned int num_ops,
 328                                                bool use_mempool,
 329                                                gfp_t gfp_flags)
 330 {
 331         struct ceph_osd_request *req;
 332         struct ceph_msg *msg;
 333         size_t msg_size;
 334
 335         BUILD_BUG_ON(CEPH_OSD_MAX_OP > U16_MAX);
 336         BUG_ON(num_ops > CEPH_OSD_MAX_OP);
 337
 338         msg_size = 4 + 4 + 8 + 8 + 4+8;
 339         msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
 340         msg_size += 1 + 8 + 4 + 4;     /* pg_t */
 341         msg_size += 4 + MAX_OBJ_NAME_SIZE;
 342         msg_size += 2 + num_ops*sizeof(struct ceph_osd_op);
 343         msg_size += 8;  /* snapid */
 344         msg_size += 8;  /* snap_seq */
 345         msg_size += 8 * (snapc ? snapc->num_snaps : 0);  /* snaps */
 346         msg_size += 4;
 347
 348         if (use_mempool) {
 349                 req = mempool_alloc(osdc->req_mempool, gfp_flags);
 350                 memset(req, 0, sizeof(*req));
 351         } else {
 352                 req = kmem_cache_zalloc(ceph_osd_request_cache, gfp_flags);
 353         }
 354         if (req == NULL)
 355                 return NULL;
 356
 357         req->r_osdc = osdc;
 358         req->r_mempool = use_mempool;
 359         req->r_num_ops = num_ops;
 360
 361         kref_init(&req->r_kref);
 362         init_completion(&req->r_completion);
 363         init_completion(&req->r_safe_completion);
 364         RB_CLEAR_NODE(&req->r_node);
 365         INIT_LIST_HEAD(&req->r_unsafe_item);
 366         INIT_LIST_HEAD(&req->r_linger_item);
 367         INIT_LIST_HEAD(&req->r_linger_osd);
 368         INIT_LIST_HEAD(&req->r_req_lru_item);
 369         INIT_LIST_HEAD(&req->r_osd_item);
 370
 371         /* create reply message */
 372         if (use_mempool)
 373                 msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
 374         else
 375                 msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
 376                                    OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
 377         if (!msg) {
 378                 ceph_osdc_put_request(req);
 379                 return NULL;
 380         }
 381         req->r_reply = msg;
 382
 383         /* create request message; allow space for oid */
 384         if (use_mempool)
 385                 msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
 386         else
 387                 msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
 388         if (!msg) {
 389                 ceph_osdc_put_request(req);
 390                 return NULL;
 391         }
 392
 393         memset(msg->front.iov_base, 0, msg->front.iov_len);
 394
 395         req->r_request = msg;
 396
 397         return req;
 398 }
 399 EXPORT_SYMBOL(ceph_osdc_alloc_request);
 400
 401 static bool osd_req_opcode_valid(u16 opcode)
 402 {
 403         switch (opcode) {
 404         case CEPH_OSD_OP_READ:
 405         case CEPH_OSD_OP_STAT:
 406         case CEPH_OSD_OP_MAPEXT:
 407         case CEPH_OSD_OP_MASKTRUNC:
 408         case CEPH_OSD_OP_SPARSE_READ:
 409         case CEPH_OSD_OP_NOTIFY:
 410         case CEPH_OSD_OP_NOTIFY_ACK:
 411         case CEPH_OSD_OP_ASSERT_VER:
 412         case CEPH_OSD_OP_WRITE:
 413         case CEPH_OSD_OP_WRITEFULL:
 414         case CEPH_OSD_OP_TRUNCATE:
 415         case CEPH_OSD_OP_ZERO:
 416         case CEPH_OSD_OP_DELETE:
 417         case CEPH_OSD_OP_APPEND:
 418         case CEPH_OSD_OP_STARTSYNC:
 419         case CEPH_OSD_OP_SETTRUNC:
 420         case CEPH_OSD_OP_TRIMTRUNC:
 421         case CEPH_OSD_OP_TMAPUP:
 422         case CEPH_OSD_OP_TMAPPUT:
 423         case CEPH_OSD_OP_TMAPGET:
 424         case CEPH_OSD_OP_CREATE:
 425         case CEPH_OSD_OP_ROLLBACK:
 426         case CEPH_OSD_OP_WATCH:
 427         case CEPH_OSD_OP_OMAPGETKEYS:
 428         case CEPH_OSD_OP_OMAPGETVALS:
 429         case CEPH_OSD_OP_OMAPGETHEADER:
 430         case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
 431         case CEPH_OSD_OP_OMAPSETVALS:
 432         case CEPH_OSD_OP_OMAPSETHEADER:
 433         case CEPH_OSD_OP_OMAPCLEAR:
 434         case CEPH_OSD_OP_OMAPRMKEYS:
 435         case CEPH_OSD_OP_OMAP_CMP:
 436         case CEPH_OSD_OP_CLONERANGE:
 437         case CEPH_OSD_OP_ASSERT_SRC_VERSION:
 438         case CEPH_OSD_OP_SRC_CMPXATTR:
 439         case CEPH_OSD_OP_GETXATTR:
 440         case CEPH_OSD_OP_GETXATTRS:
 441         case CEPH_OSD_OP_CMPXATTR:
 442         case CEPH_OSD_OP_SETXATTR:
 443         case CEPH_OSD_OP_SETXATTRS:
 444         case CEPH_OSD_OP_RESETXATTRS:
 445         case CEPH_OSD_OP_RMXATTR:
 446         case CEPH_OSD_OP_PULL:
 447         case CEPH_OSD_OP_PUSH:
 448         case CEPH_OSD_OP_BALANCEREADS:
 449         case CEPH_OSD_OP_UNBALANCEREADS:
 450         case CEPH_OSD_OP_SCRUB:
 451         case CEPH_OSD_OP_SCRUB_RESERVE:
 452         case CEPH_OSD_OP_SCRUB_UNRESERVE:
 453         case CEPH_OSD_OP_SCRUB_STOP:
 454         case CEPH_OSD_OP_SCRUB_MAP:
 455         case CEPH_OSD_OP_WRLOCK:
 456         case CEPH_OSD_OP_WRUNLOCK:
 457         case CEPH_OSD_OP_RDLOCK:
 458         case CEPH_OSD_OP_RDUNLOCK:
 459         case CEPH_OSD_OP_UPLOCK:
 460         case CEPH_OSD_OP_DNLOCK:
 461         case CEPH_OSD_OP_CALL:
 462         case CEPH_OSD_OP_PGLS:
 463         case CEPH_OSD_OP_PGLS_FILTER:
 464                 return true;
 465         default:
 466                 return false;
 467         }
 468 }
 469
 470 /*
 471  * This is an osd op init function for opcodes that have no data or
 472  * other information associated with them.  It also serves as a
 473  * common init routine for all the other init functions, below.
 474  */
 475 static struct ceph_osd_req_op *
 476 _osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
 477                                 u16 opcode)
 478 {
 479         struct ceph_osd_req_op *op;
 480
 481         BUG_ON(which >= osd_req->r_num_ops);
 482         BUG_ON(!osd_req_opcode_valid(opcode));
 483
 484         op = &osd_req->r_ops[which];
 485         memset(op, 0, sizeof (*op));
 486         op->op = opcode;
 487
 488         return op;
 489 }
 490
 491 void osd_req_op_init(struct ceph_osd_request *osd_req,
 492                                 unsigned int which, u16 opcode)
 493 {
 494         (void)_osd_req_op_init(osd_req, which, opcode);
 495 }
 496 EXPORT_SYMBOL(osd_req_op_init);
 497
 498 void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
 499                                 unsigned int which, u16 opcode,
 500                                 u64 offset, u64 length,
 501                                 u64 truncate_size, u32 truncate_seq)
 502 {
 503         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
 504         size_t payload_len = 0;
 505
 506         BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
 507
 508         op->extent.offset = offset;
 509         op->extent.length = length;
 510         op->extent.truncate_size = truncate_size;
 511         op->extent.truncate_seq = truncate_seq;
 512         if (opcode == CEPH_OSD_OP_WRITE)
 513                 payload_len += length;
 514
 515         op->payload_len = payload_len;
 516 }
 517 EXPORT_SYMBOL(osd_req_op_extent_init);
 518
 519 void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
 520                                 unsigned int which, u64 length)
 521 {
 522         struct ceph_osd_req_op *op;
 523         u64 previous;
 524
 525         BUG_ON(which >= osd_req->r_num_ops);
 526         op = &osd_req->r_ops[which];
 527         previous = op->extent.length;
 528
 529         if (length == previous)
 530                 return;         /* Nothing to do */
 531         BUG_ON(length > previous);
 532
 533         op->extent.length = length;
 534         op->payload_len -= previous - length;
 535 }
 536 EXPORT_SYMBOL(osd_req_op_extent_update);
 537
 538 void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
 539                         u16 opcode, const char *class, const char *method)
 540 {
 541         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
 542         struct ceph_pagelist *pagelist;
 543         size_t payload_len = 0;
 544         size_t size;
 545
 546         BUG_ON(opcode != CEPH_OSD_OP_CALL);
 547
 548         pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
 549         BUG_ON(!pagelist);
 550         ceph_pagelist_init(pagelist);
 551
 552         op->cls.class_name = class;
 553         size = strlen(class);
 554         BUG_ON(size > (size_t) U8_MAX);
 555         op->cls.class_len = size;
 556         ceph_pagelist_append(pagelist, class, size);
 557         payload_len += size;
 558
 559         op->cls.method_name = method;
 560         size = strlen(method);
 561         BUG_ON(size > (size_t) U8_MAX);
 562         op->cls.method_len = size;
 563         ceph_pagelist_append(pagelist, method, size);
 564         payload_len += size;
 565
 566         osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
 567
 568         op->cls.argc = 0;       /* currently unused */
 569
 570         op->payload_len = payload_len;
 571 }
 572 EXPORT_SYMBOL(osd_req_op_cls_init);
 573
 574 void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
 575                                 unsigned int which, u16 opcode,
 576                                 u64 cookie, u64 version, int flag)
 577 {
 578         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode);
 579
 580         BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
 581
 582         op->watch.cookie = cookie;
 583         op->watch.ver = version;
 584         if (opcode == CEPH_OSD_OP_WATCH && flag)
 585                 op->watch.flag = (u8)1;
 586 }
 587 EXPORT_SYMBOL(osd_req_op_watch_init);
 588
 589 static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
 590                                 struct ceph_osd_data *osd_data)
 591 {
 592         u64 length = ceph_osd_data_length(osd_data);
 593
 594         if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
 595                 BUG_ON(length > (u64) SIZE_MAX);
 596                 if (length)
 597                         ceph_msg_data_add_pages(msg, osd_data->pages,
 598                                         length, osd_data->alignment);
 599         } else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
 600                 BUG_ON(!length);
 601                 ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
 602 #ifdef CONFIG_BLOCK
 603         } else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
 604                 ceph_msg_data_add_bio(msg, osd_data->bio, length);
 605 #endif
 606         } else {
 607                 BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
 608         }
 609 }
 610
 611 static u64 osd_req_encode_op(struct ceph_osd_request *req,
 612                               struct ceph_osd_op *dst, unsigned int which)
 613 {
 614         struct ceph_osd_req_op *src;
 615         struct ceph_osd_data *osd_data;
 616         u64 request_data_len = 0;
 617         u64 data_length;
 618
 619         BUG_ON(which >= req->r_num_ops);
 620         src = &req->r_ops[which];
 621         if (WARN_ON(!osd_req_opcode_valid(src->op))) {
 622                 pr_err("unrecognized osd opcode %d\n", src->op);
 623
 624                 return 0;
 625         }
 626
 627         switch (src->op) {
 628         case CEPH_OSD_OP_STAT:
 629                 osd_data = &src->raw_data_in;
 630                 ceph_osdc_msg_data_add(req->r_reply, osd_data);
 631                 break;
 632         case CEPH_OSD_OP_READ:
 633         case CEPH_OSD_OP_WRITE:
 634                 if (src->op == CEPH_OSD_OP_WRITE)
 635                         request_data_len = src->extent.length;
 636                 dst->extent.offset = cpu_to_le64(src->extent.offset);
 637                 dst->extent.length = cpu_to_le64(src->extent.length);
 638                 dst->extent.truncate_size =
 639                         cpu_to_le64(src->extent.truncate_size);
 640                 dst->extent.truncate_seq =
 641                         cpu_to_le32(src->extent.truncate_seq);
 642                 osd_data = &src->extent.osd_data;
 643                 if (src->op == CEPH_OSD_OP_WRITE)
 644                         ceph_osdc_msg_data_add(req->r_request, osd_data);
 645                 else
 646                         ceph_osdc_msg_data_add(req->r_reply, osd_data);
 647                 break;
 648         case CEPH_OSD_OP_CALL:
 649                 dst->cls.class_len = src->cls.class_len;
 650                 dst->cls.method_len = src->cls.method_len;
 651                 osd_data = &src->cls.request_info;
 652                 ceph_osdc_msg_data_add(req->r_request, osd_data);
 653                 BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGELIST);
 654                 request_data_len = osd_data->pagelist->length;
 655
 656                 osd_data = &src->cls.request_data;
 657                 data_length = ceph_osd_data_length(osd_data);
 658                 if (data_length) {
 659                         BUG_ON(osd_data->type == CEPH_OSD_DATA_TYPE_NONE);
 660                         dst->cls.indata_len = cpu_to_le32(data_length);
 661                         ceph_osdc_msg_data_add(req->r_request, osd_data);
 662                         src->payload_len += data_length;
 663                         request_data_len += data_length;
 664                 }
 665                 osd_data = &src->cls.response_data;
 666                 ceph_osdc_msg_data_add(req->r_reply, osd_data);
 667                 break;
 668         case CEPH_OSD_OP_STARTSYNC:
 669                 break;
 670         case CEPH_OSD_OP_NOTIFY_ACK:
 671         case CEPH_OSD_OP_WATCH:
 672                 dst->watch.cookie = cpu_to_le64(src->watch.cookie);
 673                 dst->watch.ver = cpu_to_le64(src->watch.ver);
 674                 dst->watch.flag = src->watch.flag;
 675                 break;
 676         default:
 677                 pr_err("unsupported osd opcode %s\n",
 678                         ceph_osd_op_name(src->op));
 679                 WARN_ON(1);
 680
 681                 return 0;
 682         }
 683         dst->op = cpu_to_le16(src->op);
 684         dst->payload_len = cpu_to_le32(src->payload_len);
 685
 686         return request_data_len;
 687 }
 688
 689 /*
 690  * build new request AND message, calculate layout, and adjust file
 691  * extent as needed.
 692  *
 693  * if the file was recently truncated, we include information about its
 694  * old and new size so that the object can be updated appropriately.  (we
 695  * avoid synchronously deleting truncated objects because it's slow.)
 696  *
 697  * if @do_sync, include a 'startsync' command so that the osd will flush
 698  * data quickly.
 699  */
 700 struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 701                                                struct ceph_file_layout *layout,
 702                                                struct ceph_vino vino,
 703                                                u64 off, u64 *plen, int num_ops,
 704                                                int opcode, int flags,
 705                                                struct ceph_snap_context *snapc,
 706                                                u32 truncate_seq,
 707                                                u64 truncate_size,
 708                                                bool use_mempool)
 709 {
 710         struct ceph_osd_request *req;
 711         u64 objnum = 0;
 712         u64 objoff = 0;
 713         u64 objlen = 0;
 714         u32 object_size;
 715         u64 object_base;
 716         int r;
 717
 718         BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
 719
 720         req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
 721                                         GFP_NOFS);
 722         if (!req)
 723                 return ERR_PTR(-ENOMEM);
 724
 725         req->r_flags = flags;
 726
 727         /* calculate max write size */
 728         r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
 729         if (r < 0) {
 730                 ceph_osdc_put_request(req);
 731                 return ERR_PTR(r);
 732         }
 733
 734         object_size = le32_to_cpu(layout->fl_object_size);
 735         object_base = off - objoff;
 736         if (!(truncate_seq == 1 && truncate_size == -1ULL)) {
 737                 if (truncate_size <= object_base) {
 738                         truncate_size = 0;
 739                 } else {
 740                         truncate_size -= object_base;
 741                         if (truncate_size > object_size)
 742                                 truncate_size = object_size;
 743                 }
 744         }
 745
 746         osd_req_op_extent_init(req, 0, opcode, objoff, objlen,
 747                                 truncate_size, truncate_seq);
 748
 749         /*
 750          * A second op in the ops array means the caller wants to
 751          * also issue a include a 'startsync' command so that the
 752          * osd will flush data quickly.
 753          */
 754         if (num_ops > 1)
 755                 osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC);
 756
 757         req->r_file_layout = *layout;  /* keep a copy */
 758
 759         snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx",
 760                 vino.ino, objnum);
 761         req->r_oid_len = strlen(req->r_oid);
 762
 763         return req;
 764 }
 765 EXPORT_SYMBOL(ceph_osdc_new_request);
 766
 767 /*
 768  * We keep osd requests in an rbtree, sorted by ->r_tid.
 769  */
 770 static void __insert_request(struct ceph_osd_client *osdc,
 771                              struct ceph_osd_request *new)
 772 {
 773         struct rb_node **p = &osdc->requests.rb_node;
 774         struct rb_node *parent = NULL;
 775         struct ceph_osd_request *req = NULL;
 776
 777         while (*p) {
 778                 parent = *p;
 779                 req = rb_entry(parent, struct ceph_osd_request, r_node);
 780                 if (new->r_tid < req->r_tid)
 781                         p = &(*p)->rb_left;
 782                 else if (new->r_tid > req->r_tid)
 783                         p = &(*p)->rb_right;
 784                 else
 785                         BUG();
 786         }
 787
 788         rb_link_node(&new->r_node, parent, p);
 789         rb_insert_color(&new->r_node, &osdc->requests);
 790 }
 791
 792 static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
 793                                                  u64 tid)
 794 {
 795         struct ceph_osd_request *req;
 796         struct rb_node *n = osdc->requests.rb_node;
 797
 798         while (n) {
 799                 req = rb_entry(n, struct ceph_osd_request, r_node);
 800                 if (tid < req->r_tid)
 801                         n = n->rb_left;
 802                 else if (tid > req->r_tid)
 803                         n = n->rb_right;
 804                 else
 805                         return req;
 806         }
 807         return NULL;
 808 }
 809
 810 static struct ceph_osd_request *
 811 __lookup_request_ge(struct ceph_osd_client *osdc,
 812                     u64 tid)
 813 {
 814         struct ceph_osd_request *req;
 815         struct rb_node *n = osdc->requests.rb_node;
 816
 817         while (n) {
 818                 req = rb_entry(n, struct ceph_osd_request, r_node);
 819                 if (tid < req->r_tid) {
 820                         if (!n->rb_left)
 821                                 return req;
 822                         n = n->rb_left;
 823                 } else if (tid > req->r_tid) {
 824                         n = n->rb_right;
 825                 } else {
 826                         return req;
 827                 }
 828         }
 829         return NULL;
 830 }
 831
 832 /*
 833  * Resubmit requests pending on the given osd.
 834  */
 835 static void __kick_osd_requests(struct ceph_osd_client *osdc,
 836                                 struct ceph_osd *osd)
 837 {
 838         struct ceph_osd_request *req, *nreq;
 839         LIST_HEAD(resend);
 840         int err;
 841
 842         dout("__kick_osd_requests osd%d\n", osd->o_osd);
 843         err = __reset_osd(osdc, osd);
 844         if (err)
 845                 return;
 846         /*
 847          * Build up a list of requests to resend by traversing the
 848          * osd's list of requests.  Requests for a given object are
 849          * sent in tid order, and that is also the order they're
 850          * kept on this list.  Therefore all requests that are in
 851          * flight will be found first, followed by all requests that
 852          * have not yet been sent.  And to resend requests while
 853          * preserving this order we will want to put any sent
 854          * requests back on the front of the osd client's unsent
 855          * list.
 856          *
 857          * So we build a separate ordered list of already-sent
 858          * requests for the affected osd and splice it onto the
 859          * front of the osd client's unsent list.  Once we've seen a
 860          * request that has not yet been sent we're done.  Those
 861          * requests are already sitting right where they belong.
 862          */
 863         list_for_each_entry(req, &osd->o_requests, r_osd_item) {
 864                 if (!req->r_sent)
 865                         break;
 866                 list_move_tail(&req->r_req_lru_item, &resend);
 867                 dout("requeueing %p tid %llu osd%d\n", req, req->r_tid,
 868                      osd->o_osd);
 869                 if (!req->r_linger)
 870                         req->r_flags |= CEPH_OSD_FLAG_RETRY;
 871         }
 872         list_splice(&resend, &osdc->req_unsent);
 873
 874         /*
 875          * Linger requests are re-registered before sending, which
 876          * sets up a new tid for each.  We add them to the unsent
 877          * list at the end to keep things in tid order.
 878          */
 879         list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
 880                                  r_linger_osd) {
 881                 /*
 882                  * reregister request prior to unregistering linger so
 883                  * that r_osd is preserved.
 884                  */
 885                 BUG_ON(!list_empty(&req->r_req_lru_item));
 886                 __register_request(osdc, req);
 887                 list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
 888                 list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
 889                 __unregister_linger_request(osdc, req);
 890                 dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
 891                      osd->o_osd);
 892         }
 893 }
 894
 895 /*
 896  * If the osd connection drops, we need to resubmit all requests.
 897  */
 898 static void osd_reset(struct ceph_connection *con)
 899 {
 900         struct ceph_osd *osd = con->private;
 901         struct ceph_osd_client *osdc;
 902
 903         if (!osd)
 904                 return;
 905         dout("osd_reset osd%d\n", osd->o_osd);
 906         osdc = osd->o_osdc;
 907         down_read(&osdc->map_sem);
 908         mutex_lock(&osdc->request_mutex);
 909         __kick_osd_requests(osdc, osd);
 910         __send_queued(osdc);
 911         mutex_unlock(&osdc->request_mutex);
 912         up_read(&osdc->map_sem);
 913 }
 914
 915 /*
 916  * Track open sessions with osds.
 917  */
 918 static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
 919 {
 920         struct ceph_osd *osd;
 921
 922         osd = kzalloc(sizeof(*osd), GFP_NOFS);
 923         if (!osd)
 924                 return NULL;
 925
 926         atomic_set(&osd->o_ref, 1);
 927         osd->o_osdc = osdc;
 928         osd->o_osd = onum;
 929         RB_CLEAR_NODE(&osd->o_node);
 930         INIT_LIST_HEAD(&osd->o_requests);
 931         INIT_LIST_HEAD(&osd->o_linger_requests);
 932         INIT_LIST_HEAD(&osd->o_osd_lru);
 933         osd->o_incarnation = 1;
 934
 935         ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
 936
 937         INIT_LIST_HEAD(&osd->o_keepalive_item);
 938         return osd;
 939 }
 940
 941 static struct ceph_osd *get_osd(struct ceph_osd *osd)
 942 {
 943         if (atomic_inc_not_zero(&osd->o_ref)) {
 944                 dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
 945                      atomic_read(&osd->o_ref));
 946                 return osd;
 947         } else {
 948                 dout("get_osd %p FAIL\n", osd);
 949                 return NULL;
 950         }
 951 }
 952
 953 static void put_osd(struct ceph_osd *osd)
 954 {
 955         dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
 956              atomic_read(&osd->o_ref) - 1);
 957         if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
 958                 struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;
 959
 960                 ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
 961                 kfree(osd);
 962         }
 963 }
 964
 965 /*
 966  * remove an osd from our map
 967  */
 968 static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
 969 {
 970         dout("%s %p osd%d\n", __func__, osd, osd->o_osd);
 971         WARN_ON(!list_empty(&osd->o_requests));
 972         WARN_ON(!list_empty(&osd->o_linger_requests));
 973
 974         list_del_init(&osd->o_osd_lru);
 975         rb_erase(&osd->o_node, &osdc->osds);
 976         RB_CLEAR_NODE(&osd->o_node);
 977 }
 978
 979 static void remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
 980 {
 981         dout("%s %p osd%d\n", __func__, osd, osd->o_osd);
 982
 983         if (!RB_EMPTY_NODE(&osd->o_node)) {
 984                 ceph_con_close(&osd->o_con);
 985                 __remove_osd(osdc, osd);
 986                 put_osd(osd);
 987         }
 988 }
 989
 990 static void remove_all_osds(struct ceph_osd_client *osdc)
 991 {
 992         dout("%s %p\n", __func__, osdc);
 993         mutex_lock(&osdc->request_mutex);
 994         while (!RB_EMPTY_ROOT(&osdc->osds)) {
 995                 struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
 996                                                 struct ceph_osd, o_node);
 997                 remove_osd(osdc, osd);
 998         }
 999         mutex_unlock(&osdc->request_mutex);
1000 }
1001
1002 static void __move_osd_to_lru(struct ceph_osd_client *osdc,
1003                               struct ceph_osd *osd)
1004 {
1005         dout("__move_osd_to_lru %p\n", osd);
1006         BUG_ON(!list_empty(&osd->o_osd_lru));
1007         list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
1008         osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
1009 }
1010
1011 static void __remove_osd_from_lru(struct ceph_osd *osd)
1012 {
1013         dout("__remove_osd_from_lru %p\n", osd);
1014         if (!list_empty(&osd->o_osd_lru))
1015                 list_del_init(&osd->o_osd_lru);
1016 }
1017
1018 static void remove_old_osds(struct ceph_osd_client *osdc)
1019 {
1020         struct ceph_osd *osd, *nosd;
1021
1022         dout("__remove_old_osds %p\n", osdc);
1023         mutex_lock(&osdc->request_mutex);
1024         list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
1025                 if (time_before(jiffies, osd->lru_ttl))
1026                         break;
1027                 remove_osd(osdc, osd);
1028         }
1029         mutex_unlock(&osdc->request_mutex);
1030 }
1031
1032 /*
1033  * reset osd connect
1034  */
1035 static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
1036 {
1037         struct ceph_entity_addr *peer_addr;
1038
1039         dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
1040         if (list_empty(&osd->o_requests) &&
1041             list_empty(&osd->o_linger_requests)) {
1042                 remove_osd(osdc, osd);
1043                 return -ENODEV;
1044         }
1045
1046         peer_addr = &osdc->osdmap->osd_addr[osd->o_osd];
1047         if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) &&
1048                         !ceph_con_opened(&osd->o_con)) {
1049                 struct ceph_osd_request *req;
1050
1051                 dout(" osd addr hasn't changed and connection never opened,"
1052                      " letting msgr retry");
1053                 /* touch each r_stamp for handle_timeout()'s benfit */
1054                 list_for_each_entry(req, &osd->o_requests, r_osd_item)
1055                         req->r_stamp = jiffies;
1056
1057                 return -EAGAIN;
1058         }
1059
1060         ceph_con_close(&osd->o_con);
1061         ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, peer_addr);
1062         osd->o_incarnation++;
1063
1064         return 0;
1065 }
1066
1067 static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
1068 {
1069         struct rb_node **p = &osdc->osds.rb_node;
1070         struct rb_node *parent = NULL;
1071         struct ceph_osd *osd = NULL;
1072
1073         dout("__insert_osd %p osd%d\n", new, new->o_osd);
1074         while (*p) {
1075                 parent = *p;
1076                 osd = rb_entry(parent, struct ceph_osd, o_node);
1077                 if (new->o_osd < osd->o_osd)
1078                         p = &(*p)->rb_left;
1079                 else if (new->o_osd > osd->o_osd)
1080                         p = &(*p)->rb_right;
1081                 else
1082                         BUG();
1083         }
1084
1085         rb_link_node(&new->o_node, parent, p);
1086         rb_insert_color(&new->o_node, &osdc->osds);
1087 }
1088
1089 static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
1090 {
1091         struct ceph_osd *osd;
1092         struct rb_node *n = osdc->osds.rb_node;
1093
1094         while (n) {
1095                 osd = rb_entry(n, struct ceph_osd, o_node);
1096                 if (o < osd->o_osd)
1097                         n = n->rb_left;
1098                 else if (o > osd->o_osd)
1099                         n = n->rb_right;
1100                 else
1101                         return osd;
1102         }
1103         return NULL;
1104 }
1105
1106 static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
1107 {
1108         schedule_delayed_work(&osdc->timeout_work,
1109                         osdc->client->options->osd_keepalive_timeout * HZ);
1110 }
1111
1112 static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
1113 {
1114         cancel_delayed_work(&osdc->timeout_work);
1115 }
1116
1117 /*
1118  * Register request, assign tid.  If this is the first request, set up
1119  * the timeout event.
1120  */
1121 static void __register_request(struct ceph_osd_client *osdc,
1122                                struct ceph_osd_request *req)
1123 {
1124         req->r_tid = ++osdc->last_tid;
1125         req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1126         dout("__register_request %p tid %lld\n", req, req->r_tid);
1127         __insert_request(osdc, req);
1128         ceph_osdc_get_request(req);
1129         osdc->num_requests++;
1130         if (osdc->num_requests == 1) {
1131                 dout(" first request, scheduling timeout\n");
1132                 __schedule_osd_timeout(osdc);
1133         }
1134 }
1135
1136 /*
1137  * called under osdc->request_mutex
1138  */
1139 static void __unregister_request(struct ceph_osd_client *osdc,
1140                                  struct ceph_osd_request *req)
1141 {
1142         if (RB_EMPTY_NODE(&req->r_node)) {
1143                 dout("__unregister_request %p tid %lld not registered\n",
1144                         req, req->r_tid);
1145                 return;
1146         }
1147
1148         dout("__unregister_request %p tid %lld\n", req, req->r_tid);
1149         rb_erase(&req->r_node, &osdc->requests);
1150         osdc->num_requests--;
1151
1152         if (req->r_osd) {
1153                 /* make sure the original request isn't in flight. */
1154                 ceph_msg_revoke(req->r_request);
1155
1156                 list_del_init(&req->r_osd_item);
1157                 if (list_empty(&req->r_osd->o_requests) &&
1158                     list_empty(&req->r_osd->o_linger_requests)) {
1159                         dout("moving osd to %p lru\n", req->r_osd);
1160                         __move_osd_to_lru(osdc, req->r_osd);
1161                 }
1162                 if (list_empty(&req->r_linger_item))
1163                         req->r_osd = NULL;
1164         }
1165
1166         list_del_init(&req->r_req_lru_item);
1167         ceph_osdc_put_request(req);
1168
1169         if (osdc->num_requests == 0) {
1170                 dout(" no requests, canceling timeout\n");
1171                 __cancel_osd_timeout(osdc);
1172         }
1173 }
1174
1175 /*
1176  * Cancel a previously queued request message
1177  */
1178 static void __cancel_request(struct ceph_osd_request *req)
1179 {
1180         if (req->r_sent && req->r_osd) {
1181                 ceph_msg_revoke(req->r_request);
1182                 req->r_sent = 0;
1183         }
1184 }
1185
1186 static void __register_linger_request(struct ceph_osd_client *osdc,
1187                                     struct ceph_osd_request *req)
1188 {
1189         dout("__register_linger_request %p\n", req);
1190         ceph_osdc_get_request(req);
1191         list_add_tail(&req->r_linger_item, &osdc->req_linger);
1192         if (req->r_osd)
1193                 list_add_tail(&req->r_linger_osd,
1194                               &req->r_osd->o_linger_requests);
1195 }
1196
1197 static void __unregister_linger_request(struct ceph_osd_client *osdc,
1198                                         struct ceph_osd_request *req)
1199 {
1200         dout("__unregister_linger_request %p\n", req);
1201         list_del_init(&req->r_linger_item);
1202         if (req->r_osd) {
1203                 list_del_init(&req->r_linger_osd);
1204
1205                 if (list_empty(&req->r_osd->o_requests) &&
1206                     list_empty(&req->r_osd->o_linger_requests)) {
1207                         dout("moving osd to %p lru\n", req->r_osd);
1208                         __move_osd_to_lru(osdc, req->r_osd);
1209                 }
1210                 if (list_empty(&req->r_osd_item))
1211                         req->r_osd = NULL;
1212         }
1213         ceph_osdc_put_request(req);
1214 }
1215
1216 void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
1217                                          struct ceph_osd_request *req)
1218 {
1219         mutex_lock(&osdc->request_mutex);
1220         if (req->r_linger) {
1221                 req->r_linger = 0;
1222                 __unregister_linger_request(osdc, req);
1223         }
1224         mutex_unlock(&osdc->request_mutex);
1225 }
1226 EXPORT_SYMBOL(ceph_osdc_unregister_linger_request);
1227
1228 void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
1229                                   struct ceph_osd_request *req)
1230 {
1231         if (!req->r_linger) {
1232                 dout("set_request_linger %p\n", req);
1233                 req->r_linger = 1;
1234         }
1235 }
1236 EXPORT_SYMBOL(ceph_osdc_set_request_linger);
1237
1238 /*
1239  * Returns whether a request should be blocked from being sent
1240  * based on the current osdmap and osd_client settings.
1241  *
1242  * Caller should hold map_sem for read.
1243  */
1244 static bool __req_should_be_paused(struct ceph_osd_client *osdc,
1245                                    struct ceph_osd_request *req)
1246 {
1247         bool pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD);
1248         bool pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) ||
1249                 ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);
1250         return (req->r_flags & CEPH_OSD_FLAG_READ && pauserd) ||
1251                 (req->r_flags & CEPH_OSD_FLAG_WRITE && pausewr);
1252 }
1253
1254 /*
1255  * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
1256  * (as needed), and set the request r_osd appropriately.  If there is
1257  * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1258  * (unsent, homeless) or leave on in-flight lru.
1259  *
1260  * Return 0 if unchanged, 1 if changed, or negative on error.
1261  *
1262  * Caller should hold map_sem for read and request_mutex.
1263  */
1264 static int __map_request(struct ceph_osd_client *osdc,
1265                          struct ceph_osd_request *req, int force_resend)
1266 {
1267         struct ceph_pg pgid;
1268         int acting[CEPH_PG_MAX_SIZE];
1269         int o = -1, num = 0;
1270         int err;
1271         bool was_paused;
1272
1273         dout("map_request %p tid %lld\n", req, req->r_tid);
1274         err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
1275                                 ceph_file_layout_pg_pool(req->r_file_layout));
1276         if (err) {
1277                 list_move(&req->r_req_lru_item, &osdc->req_notarget);
1278                 return err;
1279         }
1280         req->r_pgid = pgid;
1281
1282         err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
1283         if (err > 0) {
1284                 o = acting[0];
1285                 num = err;
1286         }
1287
1288         was_paused = req->r_paused;
1289         req->r_paused = __req_should_be_paused(osdc, req);
1290         if (was_paused && !req->r_paused)
1291                 force_resend = 1;
1292
1293         if ((!force_resend &&
1294              req->r_osd && req->r_osd->o_osd == o &&
1295              req->r_sent >= req->r_osd->o_incarnation &&
1296              req->r_num_pg_osds == num &&
1297              memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
1298             (req->r_osd == NULL && o == -1) ||
1299             req->r_paused)
1300                 return 0;  /* no change */
1301
1302         dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
1303              req->r_tid, pgid.pool, pgid.seed, o,
1304              req->r_osd ? req->r_osd->o_osd : -1);
1305
1306         /* record full pg acting set */
1307         memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
1308         req->r_num_pg_osds = num;
1309
1310         if (req->r_osd) {
1311                 __cancel_request(req);
1312                 list_del_init(&req->r_osd_item);
1313                 req->r_osd = NULL;
1314         }
1315
1316         req->r_osd = __lookup_osd(osdc, o);
1317         if (!req->r_osd && o >= 0) {
1318                 err = -ENOMEM;
1319                 req->r_osd = create_osd(osdc, o);
1320                 if (!req->r_osd) {
1321                         list_move(&req->r_req_lru_item, &osdc->req_notarget);
1322                         goto out;
1323                 }
1324
1325                 dout("map_request osd %p is osd%d\n", req->r_osd, o);
1326                 __insert_osd(osdc, req->r_osd);
1327
1328                 ceph_con_open(&req->r_osd->o_con,
1329                               CEPH_ENTITY_TYPE_OSD, o,
1330                               &osdc->osdmap->osd_addr[o]);
1331         }
1332
1333         if (req->r_osd) {
1334                 __remove_osd_from_lru(req->r_osd);
1335                 list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
1336                 list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1337         } else {
1338                 list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1339         }
1340         err = 1;   /* osd or pg changed */
1341
1342 out:
1343         return err;
1344 }
1345
1346 /*
1347  * caller should hold map_sem (for read) and request_mutex
1348  */
1349 static void __send_request(struct ceph_osd_client *osdc,
1350                            struct ceph_osd_request *req)
1351 {
1352         void *p;
1353
1354         dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
1355              req, req->r_tid, req->r_osd->o_osd, req->r_flags,
1356              (unsigned long long)req->r_pgid.pool, req->r_pgid.seed);
1357
1358         /* fill in message content that changes each time we send it */
1359         put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch);
1360         put_unaligned_le32(req->r_flags, req->r_request_flags);
1361         put_unaligned_le64(req->r_pgid.pool, req->r_request_pool);
1362         p = req->r_request_pgid;
1363         ceph_encode_64(&p, req->r_pgid.pool);
1364         ceph_encode_32(&p, req->r_pgid.seed);
1365         put_unaligned_le64(1, req->r_request_attempts);  /* FIXME */
1366         memcpy(req->r_request_reassert_version, &req->r_reassert_version,
1367                sizeof(req->r_reassert_version));
1368
1369         req->r_stamp = jiffies;
1370         list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
1371
1372         ceph_msg_get(req->r_request); /* send consumes a ref */
1373
1374         req->r_sent = req->r_osd->o_incarnation;
1375
1376         ceph_con_send(&req->r_osd->o_con, req->r_request);
1377 }
1378
1379 /*
1380  * Send any requests in the queue (req_unsent).
1381  */
1382 static void __send_queued(struct ceph_osd_client *osdc)
1383 {
1384         struct ceph_osd_request *req, *tmp;
1385
1386         dout("__send_queued\n");
1387         list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1388                 __send_request(osdc, req);
1389 }
1390
1391 /*
1392  * Timeout callback, called every N seconds when 1 or more osd
1393  * requests has been active for more than N seconds.  When this
1394  * happens, we ping all OSDs with requests who have timed out to
1395  * ensure any communications channel reset is detected.  Reset the
1396  * request timeouts another N seconds in the future as we go.
1397  * Reschedule the timeout event another N seconds in future (unless
1398  * there are no open requests).
1399  */
1400 static void handle_timeout(struct work_struct *work)
1401 {
1402         struct ceph_osd_client *osdc =
1403                 container_of(work, struct ceph_osd_client, timeout_work.work);
1404         struct ceph_osd_request *req;
1405         struct ceph_osd *osd;
1406         unsigned long keepalive =
1407                 osdc->client->options->osd_keepalive_timeout * HZ;
1408         struct list_head slow_osds;
1409         dout("timeout\n");
1410         down_read(&osdc->map_sem);
1411
1412         ceph_monc_request_next_osdmap(&osdc->client->monc);
1413
1414         mutex_lock(&osdc->request_mutex);
1415
1416         /*
1417          * ping osds that are a bit slow.  this ensures that if there
1418          * is a break in the TCP connection we will notice, and reopen
1419          * a connection with that osd (from the fault callback).
1420          */
1421         INIT_LIST_HEAD(&slow_osds);
1422         list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
1423                 if (time_before(jiffies, req->r_stamp + keepalive))
1424                         break;
1425
1426                 osd = req->r_osd;
1427                 BUG_ON(!osd);
1428                 dout(" tid %llu is slow, will send keepalive on osd%d\n",
1429                      req->r_tid, osd->o_osd);
1430                 list_move_tail(&osd->o_keepalive_item, &slow_osds);
1431         }
1432         while (!list_empty(&slow_osds)) {
1433                 osd = list_entry(slow_osds.next, struct ceph_osd,
1434                                  o_keepalive_item);
1435                 list_del_init(&osd->o_keepalive_item);
1436                 ceph_con_keepalive(&osd->o_con);
1437         }
1438
1439         __schedule_osd_timeout(osdc);
1440         __send_queued(osdc);
1441         mutex_unlock(&osdc->request_mutex);
1442         up_read(&osdc->map_sem);
1443 }
1444
1445 static void handle_osds_timeout(struct work_struct *work)
1446 {
1447         struct ceph_osd_client *osdc =
1448                 container_of(work, struct ceph_osd_client,
1449                              osds_timeout_work.work);
1450         unsigned long delay =
1451                 osdc->client->options->osd_idle_ttl * HZ >> 2;
1452
1453         dout("osds timeout\n");
1454         down_read(&osdc->map_sem);
1455         remove_old_osds(osdc);
1456         up_read(&osdc->map_sem);
1457
1458         schedule_delayed_work(&osdc->osds_timeout_work,
1459                               round_jiffies_relative(delay));
1460 }
1461
1462 static void complete_request(struct ceph_osd_request *req)
1463 {
1464         complete_all(&req->r_safe_completion);  /* fsync waiter */
1465 }
1466
1467 /*
1468  * handle osd op reply.  either call the callback if it is specified,
1469  * or do the completion to wake up the waiting thread.
1470  */
1471 static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
1472                          struct ceph_connection *con)
1473 {
1474         void *p, *end;
1475         struct ceph_osd_request *req;
1476         u64 tid;
1477         int object_len;
1478         unsigned int numops;
1479         int payload_len, flags;
1480         s32 result;
1481         s32 retry_attempt;
1482         struct ceph_pg pg;
1483         int err;
1484         u32 reassert_epoch;
1485         u64 reassert_version;
1486         u32 osdmap_epoch;
1487         int already_completed;
1488         u32 bytes;
1489         unsigned int i;
1490
1491         tid = le64_to_cpu(msg->hdr.tid);
1492         dout("handle_reply %p tid %llu\n", msg, tid);
1493
1494         p = msg->front.iov_base;
1495         end = p + msg->front.iov_len;
1496
1497         ceph_decode_need(&p, end, 4, bad);
1498         object_len = ceph_decode_32(&p);
1499         ceph_decode_need(&p, end, object_len, bad);
1500         p += object_len;
1501
1502         err = ceph_decode_pgid(&p, end, &pg);
1503         if (err)
1504                 goto bad;
1505
1506         ceph_decode_need(&p, end, 8 + 4 + 4 + 8 + 4, bad);
1507         flags = ceph_decode_64(&p);
1508         result = ceph_decode_32(&p);
1509         reassert_epoch = ceph_decode_32(&p);
1510         reassert_version = ceph_decode_64(&p);
1511         osdmap_epoch = ceph_decode_32(&p);
1512
1513         /* lookup */
1514         mutex_lock(&osdc->request_mutex);
1515         req = __lookup_request(osdc, tid);
1516         if (req == NULL) {
1517                 dout("handle_reply tid %llu dne\n", tid);
1518                 goto bad_mutex;
1519         }
1520         ceph_osdc_get_request(req);
1521
1522         dout("handle_reply %p tid %llu req %p result %d\n", msg, tid,
1523              req, result);
1524
1525         ceph_decode_need(&p, end, 4, bad_put);
1526         numops = ceph_decode_32(&p);
1527         if (numops > CEPH_OSD_MAX_OP)
1528                 goto bad_put;
1529         if (numops != req->r_num_ops)
1530                 goto bad_put;
1531         payload_len = 0;
1532         ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad_put);
1533         for (i = 0; i < numops; i++) {
1534                 struct ceph_osd_op *op = p;
1535                 int len;
1536
1537                 len = le32_to_cpu(op->payload_len);
1538                 req->r_reply_op_len[i] = len;
1539                 dout(" op %d has %d bytes\n", i, len);
1540                 payload_len += len;
1541                 p += sizeof(*op);
1542         }
1543         bytes = le32_to_cpu(msg->hdr.data_len);
1544         if (payload_len != bytes) {
1545                 pr_warning("sum of op payload lens %d != data_len %d",
1546                            payload_len, bytes);
1547                 goto bad_put;
1548         }
1549
1550         ceph_decode_need(&p, end, 4 + numops * 4, bad_put);
1551         retry_attempt = ceph_decode_32(&p);
1552         for (i = 0; i < numops; i++)
1553                 req->r_reply_op_result[i] = ceph_decode_32(&p);
1554
1555         already_completed = req->r_got_reply;
1556
1557         if (!req->r_got_reply) {
1558
1559                 req->r_result = result;
1560                 dout("handle_reply result %d bytes %d\n", req->r_result,
1561                      bytes);
1562                 if (req->r_result == 0)
1563                         req->r_result = bytes;
1564
1565                 /* in case this is a write and we need to replay, */
1566                 req->r_reassert_version.epoch = cpu_to_le32(reassert_epoch);
1567                 req->r_reassert_version.version = cpu_to_le64(reassert_version);
1568
1569                 req->r_got_reply = 1;
1570         } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
1571                 dout("handle_reply tid %llu dup ack\n", tid);
1572                 mutex_unlock(&osdc->request_mutex);
1573                 goto done;
1574         }
1575
1576         dout("handle_reply tid %llu flags %d\n", tid, flags);
1577
1578         if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
1579                 __register_linger_request(osdc, req);
1580
1581         /* either this is a read, or we got the safe response */
1582         if (result < 0 ||
1583             (flags & CEPH_OSD_FLAG_ONDISK) ||
1584             ((flags & CEPH_OSD_FLAG_WRITE) == 0))
1585                 __unregister_request(osdc, req);
1586
1587         mutex_unlock(&osdc->request_mutex);
1588
1589         if (!already_completed) {
1590                 if (req->r_unsafe_callback &&
1591                     result >= 0 && !(flags & CEPH_OSD_FLAG_ONDISK))
1592                         req->r_unsafe_callback(req, true);
1593                 if (req->r_callback)
1594                         req->r_callback(req, msg);
1595                 else
1596                         complete_all(&req->r_completion);
1597         }
1598
1599         if (flags & CEPH_OSD_FLAG_ONDISK) {
1600                 if (req->r_unsafe_callback && already_completed)
1601                         req->r_unsafe_callback(req, false);
1602                 complete_request(req);
1603         }
1604
1605 done:
1606         dout("req=%p req->r_linger=%d\n", req, req->r_linger);
1607         ceph_osdc_put_request(req);
1608         return;
1609
1610 bad_put:
1611         ceph_osdc_put_request(req);
1612 bad_mutex:
1613         mutex_unlock(&osdc->request_mutex);
1614 bad:
1615         pr_err("corrupt osd_op_reply got %d %d\n",
1616                (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1617         ceph_msg_dump(msg);
1618 }
1619
1620 static void reset_changed_osds(struct ceph_osd_client *osdc)
1621 {
1622         struct rb_node *p, *n;
1623
1624         dout("%s %p\n", __func__, osdc);
1625         for (p = rb_first(&osdc->osds); p; p = n) {
1626                 struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
1627
1628                 n = rb_next(p);
1629                 if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
1630                     memcmp(&osd->o_con.peer_addr,
1631                            ceph_osd_addr(osdc->osdmap,
1632                                          osd->o_osd),
1633                            sizeof(struct ceph_entity_addr)) != 0)
1634                         __reset_osd(osdc, osd);
1635         }
1636 }
1637
1638 /*
1639  * Requeue requests whose mapping to an OSD has changed.  If requests map to
1640  * no osd, request a new map.
1641  *
1642  * Caller should hold map_sem for read.
1643  */
1644 static void kick_requests(struct ceph_osd_client *osdc, bool force_resend,
1645                           bool force_resend_writes)
1646 {
1647         struct ceph_osd_request *req, *nreq;
1648         struct rb_node *p;
1649         int needmap = 0;
1650         int err;
1651         bool force_resend_req;
1652
1653         dout("kick_requests %s %s\n", force_resend ? " (force resend)" : "",
1654                 force_resend_writes ? " (force resend writes)" : "");
1655         mutex_lock(&osdc->request_mutex);
1656         for (p = rb_first(&osdc->requests); p; ) {
1657                 req = rb_entry(p, struct ceph_osd_request, r_node);
1658                 p = rb_next(p);
1659
1660                 /*
1661                  * For linger requests that have not yet been
1662                  * registered, move them to the linger list; they'll
1663                  * be sent to the osd in the loop below.  Unregister
1664                  * the request before re-registering it as a linger
1665                  * request to ensure the __map_request() below
1666                  * will decide it needs to be sent.
1667                  */
1668                 if (req->r_linger && list_empty(&req->r_linger_item)) {
1669                         dout("%p tid %llu restart on osd%d\n",
1670                              req, req->r_tid,
1671                              req->r_osd ? req->r_osd->o_osd : -1);
1672                         ceph_osdc_get_request(req);
1673                         __unregister_request(osdc, req);
1674                         __register_linger_request(osdc, req);
1675                         ceph_osdc_put_request(req);
1676                         continue;
1677                 }
1678
1679                 force_resend_req = force_resend ||
1680                         (force_resend_writes &&
1681                                 req->r_flags & CEPH_OSD_FLAG_WRITE);
1682                 err = __map_request(osdc, req, force_resend_req);
1683                 if (err < 0)
1684                         continue;  /* error */
1685                 if (req->r_osd == NULL) {
1686                         dout("%p tid %llu maps to no osd\n", req, req->r_tid);
1687                         needmap++;  /* request a newer map */
1688                 } else if (err > 0) {
1689                         if (!req->r_linger) {
1690                                 dout("%p tid %llu requeued on osd%d\n", req,
1691                                      req->r_tid,
1692                                      req->r_osd ? req->r_osd->o_osd : -1);
1693                                 req->r_flags |= CEPH_OSD_FLAG_RETRY;
1694                         }
1695                 }
1696         }
1697
1698         list_for_each_entry_safe(req, nreq, &osdc->req_linger,
1699                                  r_linger_item) {
1700                 dout("linger req=%p req->r_osd=%p\n", req, req->r_osd);
1701
1702                 err = __map_request(osdc, req,
1703                                     force_resend || force_resend_writes);
1704                 dout("__map_request returned %d\n", err);
1705                 if (err < 0)
1706                         continue;  /* hrm! */
1707                 if (req->r_osd == NULL || err > 0) {
1708                         if (req->r_osd == NULL) {
1709                                 dout("lingering %p tid %llu maps to no osd\n",
1710                                      req, req->r_tid);
1711                                 /*
1712                                  * A homeless lingering request makes
1713                                  * no sense, as it's job is to keep
1714                                  * a particular OSD connection open.
1715                                  * Request a newer map and kick the
1716                                  * request, knowing that it won't be
1717                                  * resent until we actually get a map
1718                                  * that can tell us where to send it.
1719                                  */
1720                                 needmap++;
1721                         }
1722
1723                         dout("kicking lingering %p tid %llu osd%d\n", req,
1724                              req->r_tid, req->r_osd ? req->r_osd->o_osd : -1);
1725                         __register_request(osdc, req);
1726                         __unregister_linger_request(osdc, req);
1727                 }
1728         }
1729         reset_changed_osds(osdc);
1730         mutex_unlock(&osdc->request_mutex);
1731
1732         if (needmap) {
1733                 dout("%d requests for down osds, need new map\n", needmap);
1734                 ceph_monc_request_next_osdmap(&osdc->client->monc);
1735         }
1736 }
1737
1738
1739 /*
1740  * Process updated osd map.
1741  *
1742  * The message contains any number of incremental and full maps, normally
1743  * indicating some sort of topology change in the cluster.  Kick requests
1744  * off to different OSDs as needed.
1745  */
1746 void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
1747 {
1748         void *p, *end, *next;
1749         u32 nr_maps, maplen;
1750         u32 epoch;
1751         struct ceph_osdmap *newmap = NULL, *oldmap;
1752         int err;
1753         struct ceph_fsid fsid;
1754         bool was_full;
1755
1756         dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
1757         p = msg->front.iov_base;
1758         end = p + msg->front.iov_len;
1759
1760         /* verify fsid */
1761         ceph_decode_need(&p, end, sizeof(fsid), bad);
1762         ceph_decode_copy(&p, &fsid, sizeof(fsid));
1763         if (ceph_check_fsid(osdc->client, &fsid) < 0)
1764                 return;
1765
1766         down_write(&osdc->map_sem);
1767
1768         was_full = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL);
1769
1770         /* incremental maps */
1771         ceph_decode_32_safe(&p, end, nr_maps, bad);
1772         dout(" %d inc maps\n", nr_maps);
1773         while (nr_maps > 0) {
1774                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1775                 epoch = ceph_decode_32(&p);
1776                 maplen = ceph_decode_32(&p);
1777                 ceph_decode_need(&p, end, maplen, bad);
1778                 next = p + maplen;
1779                 if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
1780                         dout("applying incremental map %u len %d\n",
1781                              epoch, maplen);
1782                         newmap = osdmap_apply_incremental(&p, next,
1783                                                           osdc->osdmap,
1784                                                           &osdc->client->msgr);
1785                         if (IS_ERR(newmap)) {
1786                                 err = PTR_ERR(newmap);
1787                                 goto bad;
1788                         }
1789                         BUG_ON(!newmap);
1790                         if (newmap != osdc->osdmap) {
1791                                 ceph_osdmap_destroy(osdc->osdmap);
1792                                 osdc->osdmap = newmap;
1793                         }
1794                         was_full = was_full ||
1795                                 ceph_osdmap_flag(osdc->osdmap,
1796                                                  CEPH_OSDMAP_FULL);
1797                         kick_requests(osdc, 0, was_full);
1798                 } else {
1799                         dout("ignoring incremental map %u len %d\n",
1800                              epoch, maplen);
1801                 }
1802                 p = next;
1803                 nr_maps--;
1804         }
1805         if (newmap)
1806                 goto done;
1807
1808         /* full maps */
1809         ceph_decode_32_safe(&p, end, nr_maps, bad);
1810         dout(" %d full maps\n", nr_maps);
1811         while (nr_maps) {
1812                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1813                 epoch = ceph_decode_32(&p);
1814                 maplen = ceph_decode_32(&p);
1815                 ceph_decode_need(&p, end, maplen, bad);
1816                 if (nr_maps > 1) {
1817                         dout("skipping non-latest full map %u len %d\n",
1818                              epoch, maplen);
1819                 } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
1820                         dout("skipping full map %u len %d, "
1821                              "older than our %u\n", epoch, maplen,
1822                              osdc->osdmap->epoch);
1823                 } else {
1824                         int skipped_map = 0;
1825
1826                         dout("taking full map %u len %d\n", epoch, maplen);
1827                         newmap = osdmap_decode(&p, p+maplen);
1828                         if (IS_ERR(newmap)) {
1829                                 err = PTR_ERR(newmap);
1830                                 goto bad;
1831                         }
1832                         BUG_ON(!newmap);
1833                         oldmap = osdc->osdmap;
1834                         osdc->osdmap = newmap;
1835                         if (oldmap) {
1836                                 if (oldmap->epoch + 1 < newmap->epoch)
1837                                         skipped_map = 1;
1838                                 ceph_osdmap_destroy(oldmap);
1839                         }
1840                         was_full = was_full ||
1841                                 ceph_osdmap_flag(osdc->osdmap,
1842                                                  CEPH_OSDMAP_FULL);
1843                         kick_requests(osdc, skipped_map, was_full);
1844                 }
1845                 p += maplen;
1846                 nr_maps--;
1847         }
1848
1849         if (!osdc->osdmap)
1850                 goto bad;
1851 done:
1852         downgrade_write(&osdc->map_sem);
1853         ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1854
1855         /*
1856          * subscribe to subsequent osdmap updates if full to ensure
1857          * we find out when we are no longer full and stop returning
1858          * ENOSPC.
1859          */
1860         if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) ||
1861                 ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD) ||
1862                 ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR))
1863                 ceph_monc_request_next_osdmap(&osdc->client->monc);
1864
1865         mutex_lock(&osdc->request_mutex);
1866         __send_queued(osdc);
1867         mutex_unlock(&osdc->request_mutex);
1868         up_read(&osdc->map_sem);
1869         wake_up_all(&osdc->client->auth_wq);
1870         return;
1871
1872 bad:
1873         pr_err("osdc handle_map corrupt msg\n");
1874         ceph_msg_dump(msg);
1875         up_write(&osdc->map_sem);
1876         return;
1877 }
1878
1879 /*
1880  * watch/notify callback event infrastructure
1881  *
1882  * These callbacks are used both for watch and notify operations.
1883  */
1884 static void __release_event(struct kref *kref)
1885 {
1886         struct ceph_osd_event *event =
1887                 container_of(kref, struct ceph_osd_event, kref);
1888
1889         dout("__release_event %p\n", event);
1890         kfree(event);
1891 }
1892
1893 static void get_event(struct ceph_osd_event *event)
1894 {
1895         kref_get(&event->kref);
1896 }
1897
1898 void ceph_osdc_put_event(struct ceph_osd_event *event)
1899 {
1900         kref_put(&event->kref, __release_event);
1901 }
1902 EXPORT_SYMBOL(ceph_osdc_put_event);
1903
1904 static void __insert_event(struct ceph_osd_client *osdc,
1905                              struct ceph_osd_event *new)
1906 {
1907         struct rb_node **p = &osdc->event_tree.rb_node;
1908         struct rb_node *parent = NULL;
1909         struct ceph_osd_event *event = NULL;
1910
1911         while (*p) {
1912                 parent = *p;
1913                 event = rb_entry(parent, struct ceph_osd_event, node);
1914                 if (new->cookie < event->cookie)
1915                         p = &(*p)->rb_left;
1916                 else if (new->cookie > event->cookie)
1917                         p = &(*p)->rb_right;
1918                 else
1919                         BUG();
1920         }
1921
1922         rb_link_node(&new->node, parent, p);
1923         rb_insert_color(&new->node, &osdc->event_tree);
1924 }
1925
1926 static struct ceph_osd_event *__find_event(struct ceph_osd_client *osdc,
1927                                                 u64 cookie)
1928 {
1929         struct rb_node **p = &osdc->event_tree.rb_node;
1930         struct rb_node *parent = NULL;
1931         struct ceph_osd_event *event = NULL;
1932
1933         while (*p) {
1934                 parent = *p;
1935                 event = rb_entry(parent, struct ceph_osd_event, node);
1936                 if (cookie < event->cookie)
1937                         p = &(*p)->rb_left;
1938                 else if (cookie > event->cookie)
1939                         p = &(*p)->rb_right;
1940                 else
1941                         return event;
1942         }
1943         return NULL;
1944 }
1945
1946 static void __remove_event(struct ceph_osd_event *event)
1947 {
1948         struct ceph_osd_client *osdc = event->osdc;
1949
1950         if (!RB_EMPTY_NODE(&event->node)) {
1951                 dout("__remove_event removed %p\n", event);
1952                 rb_erase(&event->node, &osdc->event_tree);
1953                 ceph_osdc_put_event(event);
1954         } else {
1955                 dout("__remove_event didn't remove %p\n", event);
1956         }
1957 }
1958
1959 int ceph_osdc_create_event(struct ceph_osd_client *osdc,
1960                            void (*event_cb)(u64, u64, u8, void *),
1961                            void *data, struct ceph_osd_event **pevent)
1962 {
1963         struct ceph_osd_event *event;
1964
1965         event = kmalloc(sizeof(*event), GFP_NOIO);
1966         if (!event)
1967                 return -ENOMEM;
1968
1969         dout("create_event %p\n", event);
1970         event->cb = event_cb;
1971         event->one_shot = 0;
1972         event->data = data;
1973         event->osdc = osdc;
1974         INIT_LIST_HEAD(&event->osd_node);
1975         RB_CLEAR_NODE(&event->node);
1976         kref_init(&event->kref);   /* one ref for us */
1977         kref_get(&event->kref);    /* one ref for the caller */
1978
1979         spin_lock(&osdc->event_lock);
1980         event->cookie = ++osdc->event_count;
1981         __insert_event(osdc, event);
1982         spin_unlock(&osdc->event_lock);
1983
1984         *pevent = event;
1985         return 0;
1986 }
1987 EXPORT_SYMBOL(ceph_osdc_create_event);
1988
1989 void ceph_osdc_cancel_event(struct ceph_osd_event *event)
1990 {
1991         struct ceph_osd_client *osdc = event->osdc;
1992
1993         dout("cancel_event %p\n", event);
1994         spin_lock(&osdc->event_lock);
1995         __remove_event(event);
1996         spin_unlock(&osdc->event_lock);
1997         ceph_osdc_put_event(event); /* caller's */
1998 }
1999 EXPORT_SYMBOL(ceph_osdc_cancel_event);
2000
2001
2002 static void do_event_work(struct work_struct *work)
2003 {
2004         struct ceph_osd_event_work *event_work =
2005                 container_of(work, struct ceph_osd_event_work, work);
2006         struct ceph_osd_event *event = event_work->event;
2007         u64 ver = event_work->ver;
2008         u64 notify_id = event_work->notify_id;
2009         u8 opcode = event_work->opcode;
2010
2011         dout("do_event_work completing %p\n", event);
2012         event->cb(ver, notify_id, opcode, event->data);
2013         dout("do_event_work completed %p\n", event);
2014         ceph_osdc_put_event(event);
2015         kfree(event_work);
2016 }
2017
2018
2019 /*
2020  * Process osd watch notifications
2021  */
2022 static void handle_watch_notify(struct ceph_osd_client *osdc,
2023                                 struct ceph_msg *msg)
2024 {
2025         void *p, *end;
2026         u8 proto_ver;
2027         u64 cookie, ver, notify_id;
2028         u8 opcode;
2029         struct ceph_osd_event *event;
2030         struct ceph_osd_event_work *event_work;
2031
2032         p = msg->front.iov_base;
2033         end = p + msg->front.iov_len;
2034
2035         ceph_decode_8_safe(&p, end, proto_ver, bad);
2036         ceph_decode_8_safe(&p, end, opcode, bad);
2037         ceph_decode_64_safe(&p, end, cookie, bad);
2038         ceph_decode_64_safe(&p, end, ver, bad);
2039         ceph_decode_64_safe(&p, end, notify_id, bad);
2040
2041         spin_lock(&osdc->event_lock);
2042         event = __find_event(osdc, cookie);
2043         if (event) {
2044                 BUG_ON(event->one_shot);
2045                 get_event(event);
2046         }
2047         spin_unlock(&osdc->event_lock);
2048         dout("handle_watch_notify cookie %lld ver %lld event %p\n",
2049              cookie, ver, event);
2050         if (event) {
2051                 event_work = kmalloc(sizeof(*event_work), GFP_NOIO);
2052                 if (!event_work) {
2053                         dout("ERROR: could not allocate event_work\n");
2054                         goto done_err;
2055                 }
2056                 INIT_WORK(&event_work->work, do_event_work);
2057                 event_work->event = event;
2058                 event_work->ver = ver;
2059                 event_work->notify_id = notify_id;
2060                 event_work->opcode = opcode;
2061                 if (!queue_work(osdc->notify_wq, &event_work->work)) {
2062                         dout("WARNING: failed to queue notify event work\n");
2063                         goto done_err;
2064                 }
2065         }
2066
2067         return;
2068
2069 done_err:
2070         ceph_osdc_put_event(event);
2071         return;
2072
2073 bad:
2074         pr_err("osdc handle_watch_notify corrupt msg\n");
2075         return;
2076 }
2077
2078 /*
2079  * build new request AND message
2080  *
2081  */
2082 void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
2083                                 struct ceph_snap_context *snapc, u64 snap_id,
2084                                 struct timespec *mtime)
2085 {
2086         struct ceph_msg *msg = req->r_request;
2087         void *p;
2088         size_t msg_size;
2089         int flags = req->r_flags;
2090         u64 data_len;
2091         unsigned int i;
2092
2093         req->r_snapid = snap_id;
2094         req->r_snapc = ceph_get_snap_context(snapc);
2095
2096         /* encode request */
2097         msg->hdr.version = cpu_to_le16(4);
2098
2099         p = msg->front.iov_base;
2100         ceph_encode_32(&p, 1);   /* client_inc  is always 1 */
2101         req->r_request_osdmap_epoch = p;
2102         p += 4;
2103         req->r_request_flags = p;
2104         p += 4;
2105         if (req->r_flags & CEPH_OSD_FLAG_WRITE)
2106                 ceph_encode_timespec(p, mtime);
2107         p += sizeof(struct ceph_timespec);
2108         req->r_request_reassert_version = p;
2109         p += sizeof(struct ceph_eversion); /* will get filled in */
2110
2111         /* oloc */
2112         ceph_encode_8(&p, 4);
2113         ceph_encode_8(&p, 4);
2114         ceph_encode_32(&p, 8 + 4 + 4);
2115         req->r_request_pool = p;
2116         p += 8;
2117         ceph_encode_32(&p, -1);  /* preferred */
2118         ceph_encode_32(&p, 0);   /* key len */
2119
2120         ceph_encode_8(&p, 1);
2121         req->r_request_pgid = p;
2122         p += 8 + 4;
2123         ceph_encode_32(&p, -1);  /* preferred */
2124
2125         /* oid */
2126         ceph_encode_32(&p, req->r_oid_len);
2127         memcpy(p, req->r_oid, req->r_oid_len);
2128         dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
2129         p += req->r_oid_len;
2130
2131         /* ops--can imply data */
2132         ceph_encode_16(&p, (u16)req->r_num_ops);
2133         data_len = 0;
2134         for (i = 0; i < req->r_num_ops; i++) {
2135                 data_len += osd_req_encode_op(req, p, i);
2136                 p += sizeof(struct ceph_osd_op);
2137         }
2138
2139         /* snaps */
2140         ceph_encode_64(&p, req->r_snapid);
2141         ceph_encode_64(&p, req->r_snapc ? req->r_snapc->seq : 0);
2142         ceph_encode_32(&p, req->r_snapc ? req->r_snapc->num_snaps : 0);
2143         if (req->r_snapc) {
2144                 for (i = 0; i < snapc->num_snaps; i++) {
2145                         ceph_encode_64(&p, req->r_snapc->snaps[i]);
2146                 }
2147         }
2148
2149         req->r_request_attempts = p;
2150         p += 4;
2151
2152         /* data */
2153         if (flags & CEPH_OSD_FLAG_WRITE) {
2154                 u16 data_off;
2155
2156                 /*
2157                  * The header "data_off" is a hint to the receiver
2158                  * allowing it to align received data into its
2159                  * buffers such that there's no need to re-copy
2160                  * it before writing it to disk (direct I/O).
2161                  */
2162                 data_off = (u16) (off & 0xffff);
2163                 req->r_request->hdr.data_off = cpu_to_le16(data_off);
2164         }
2165         req->r_request->hdr.data_len = cpu_to_le32(data_len);
2166
2167         BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
2168         msg_size = p - msg->front.iov_base;
2169         msg->front.iov_len = msg_size;
2170         msg->hdr.front_len = cpu_to_le32(msg_size);
2171
2172         dout("build_request msg_size was %d\n", (int)msg_size);
2173 }
2174 EXPORT_SYMBOL(ceph_osdc_build_request);
2175
2176 /*
2177  * Register request, send initial attempt.
2178  */
2179 int ceph_osdc_start_request(struct ceph_osd_client *osdc,
2180                             struct ceph_osd_request *req,
2181                             bool nofail)
2182 {
2183         int rc = 0;
2184
2185         down_read(&osdc->map_sem);
2186         mutex_lock(&osdc->request_mutex);
2187         __register_request(osdc, req);
2188         req->r_sent = 0;
2189         req->r_got_reply = 0;
2190         rc = __map_request(osdc, req, 0);
2191         if (rc < 0) {
2192                 if (nofail) {
2193                         dout("osdc_start_request failed map, "
2194                                 " will retry %lld\n", req->r_tid);
2195                         rc = 0;
2196                 } else {
2197                         __unregister_request(osdc, req);
2198                 }
2199                 goto out_unlock;
2200         }
2201         if (req->r_osd == NULL) {
2202                 dout("send_request %p no up osds in pg\n", req);
2203                 ceph_monc_request_next_osdmap(&osdc->client->monc);
2204         } else {
2205                 __send_queued(osdc);
2206         }
2207         rc = 0;
2208 out_unlock:
2209         mutex_unlock(&osdc->request_mutex);
2210         up_read(&osdc->map_sem);
2211         return rc;
2212 }
2213 EXPORT_SYMBOL(ceph_osdc_start_request);
2214
2215 /*
2216  * wait for a request to complete
2217  */
2218 int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
2219                            struct ceph_osd_request *req)
2220 {
2221         int rc;
2222
2223         rc = wait_for_completion_interruptible(&req->r_completion);
2224         if (rc < 0) {
2225                 mutex_lock(&osdc->request_mutex);
2226                 __cancel_request(req);
2227                 __unregister_request(osdc, req);
2228                 mutex_unlock(&osdc->request_mutex);
2229                 complete_request(req);
2230                 dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
2231                 return rc;
2232         }
2233
2234         dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
2235         return req->r_result;
2236 }
2237 EXPORT_SYMBOL(ceph_osdc_wait_request);
2238
2239 /*
2240  * sync - wait for all in-flight requests to flush.  avoid starvation.
2241  */
2242 void ceph_osdc_sync(struct ceph_osd_client *osdc)
2243 {
2244         struct ceph_osd_request *req;
2245         u64 last_tid, next_tid = 0;
2246
2247         mutex_lock(&osdc->request_mutex);
2248         last_tid = osdc->last_tid;
2249         while (1) {
2250                 req = __lookup_request_ge(osdc, next_tid);
2251                 if (!req)
2252                         break;
2253                 if (req->r_tid > last_tid)
2254                         break;
2255
2256                 next_tid = req->r_tid + 1;
2257                 if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
2258                         continue;
2259
2260                 ceph_osdc_get_request(req);
2261                 mutex_unlock(&osdc->request_mutex);
2262                 dout("sync waiting on tid %llu (last is %llu)\n",
2263                      req->r_tid, last_tid);
2264                 wait_for_completion(&req->r_safe_completion);
2265                 mutex_lock(&osdc->request_mutex);
2266                 ceph_osdc_put_request(req);
2267         }
2268         mutex_unlock(&osdc->request_mutex);
2269         dout("sync done (thru tid %llu)\n", last_tid);
2270 }
2271 EXPORT_SYMBOL(ceph_osdc_sync);
2272
2273 /*
2274  * Call all pending notify callbacks - for use after a watch is
2275  * unregistered, to make sure no more callbacks for it will be invoked
2276  */
2277 extern void ceph_osdc_flush_notifies(struct ceph_osd_client *osdc)
2278 {
2279         flush_workqueue(osdc->notify_wq);
2280 }
2281 EXPORT_SYMBOL(ceph_osdc_flush_notifies);
2282
2283
2284 /*
2285  * init, shutdown
2286  */
2287 int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
2288 {
2289         int err;
2290
2291         dout("init\n");
2292         osdc->client = client;
2293         osdc->osdmap = NULL;
2294         init_rwsem(&osdc->map_sem);
2295         init_completion(&osdc->map_waiters);
2296         osdc->last_requested_map = 0;
2297         mutex_init(&osdc->request_mutex);
2298         osdc->last_tid = 0;
2299         osdc->osds = RB_ROOT;
2300         INIT_LIST_HEAD(&osdc->osd_lru);
2301         osdc->requests = RB_ROOT;
2302         INIT_LIST_HEAD(&osdc->req_lru);
2303         INIT_LIST_HEAD(&osdc->req_unsent);
2304         INIT_LIST_HEAD(&osdc->req_notarget);
2305         INIT_LIST_HEAD(&osdc->req_linger);
2306         osdc->num_requests = 0;
2307         INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2308         INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2309         spin_lock_init(&osdc->event_lock);
2310         osdc->event_tree = RB_ROOT;
2311         osdc->event_count = 0;
2312
2313         schedule_delayed_work(&osdc->osds_timeout_work,
2314            round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
2315
2316         err = -ENOMEM;
2317         osdc->req_mempool = mempool_create_kmalloc_pool(10,
2318                                         sizeof(struct ceph_osd_request));
2319         if (!osdc->req_mempool)
2320                 goto out;
2321
2322         err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
2323                                 OSD_OP_FRONT_LEN, 10, true,
2324                                 "osd_op");
2325         if (err < 0)
2326                 goto out_mempool;
2327         err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2328                                 OSD_OPREPLY_FRONT_LEN, 10, true,
2329                                 "osd_op_reply");
2330         if (err < 0)
2331                 goto out_msgpool;
2332
2333         err = -ENOMEM;
2334         osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
2335         if (!osdc->notify_wq)
2336                 goto out_msgpool;
2337         return 0;
2338
2339 out_msgpool:
2340         ceph_msgpool_destroy(&osdc->msgpool_op);
2341 out_mempool:
2342         mempool_destroy(osdc->req_mempool);
2343 out:
2344         return err;
2345 }
2346
2347 void ceph_osdc_stop(struct ceph_osd_client *osdc)
2348 {
2349         flush_workqueue(osdc->notify_wq);
2350         destroy_workqueue(osdc->notify_wq);
2351         cancel_delayed_work_sync(&osdc->timeout_work);
2352         cancel_delayed_work_sync(&osdc->osds_timeout_work);
2353         if (osdc->osdmap) {
2354                 ceph_osdmap_destroy(osdc->osdmap);
2355                 osdc->osdmap = NULL;
2356         }
2357         remove_all_osds(osdc);
2358         mempool_destroy(osdc->req_mempool);
2359         ceph_msgpool_destroy(&osdc->msgpool_op);
2360         ceph_msgpool_destroy(&osdc->msgpool_op_reply);
2361 }
2362
2363 /*
2364  * Read some contiguous pages.  If we cross a stripe boundary, shorten
2365  * *plen.  Return number of bytes read, or error.
2366  */
2367 int ceph_osdc_readpages(struct ceph_osd_client *osdc,
2368                         struct ceph_vino vino, struct ceph_file_layout *layout,
2369                         u64 off, u64 *plen,
2370                         u32 truncate_seq, u64 truncate_size,
2371                         struct page **pages, int num_pages, int page_align)
2372 {
2373         struct ceph_osd_request *req;
2374         int rc = 0;
2375
2376         dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
2377              vino.snap, off, *plen);
2378         req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1,
2379                                     CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2380                                     NULL, truncate_seq, truncate_size,
2381                                     false);
2382         if (IS_ERR(req))
2383                 return PTR_ERR(req);
2384
2385         /* it may be a short read due to an object boundary */
2386
2387         osd_req_op_extent_osd_data_pages(req, 0,
2388                                 pages, *plen, page_align, false, false);
2389
2390         dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2391              off, *plen, *plen, page_align);
2392
2393         ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2394
2395         rc = ceph_osdc_start_request(osdc, req, false);
2396         if (!rc)
2397                 rc = ceph_osdc_wait_request(osdc, req);
2398
2399         ceph_osdc_put_request(req);
2400         dout("readpages result %d\n", rc);
2401         return rc;
2402 }
2403 EXPORT_SYMBOL(ceph_osdc_readpages);
2404
2405 /*
2406  * do a synchronous write on N pages
2407  */
2408 int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
2409                          struct ceph_file_layout *layout,
2410                          struct ceph_snap_context *snapc,
2411                          u64 off, u64 len,
2412                          u32 truncate_seq, u64 truncate_size,
2413                          struct timespec *mtime,
2414                          struct page **pages, int num_pages)
2415 {
2416         struct ceph_osd_request *req;
2417         int rc = 0;
2418         int page_align = off & ~PAGE_MASK;
2419
2420         BUG_ON(vino.snap != CEPH_NOSNAP);       /* snapshots aren't writeable */
2421         req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1,
2422                                     CEPH_OSD_OP_WRITE,
2423                                     CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2424                                     snapc, truncate_seq, truncate_size,
2425                                     true);
2426         if (IS_ERR(req))
2427                 return PTR_ERR(req);
2428
2429         /* it may be a short write due to an object boundary */
2430         osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2431                                 false, false);
2432         dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
2433
2434         ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2435
2436         rc = ceph_osdc_start_request(osdc, req, true);
2437         if (!rc)
2438                 rc = ceph_osdc_wait_request(osdc, req);
2439
2440         ceph_osdc_put_request(req);
2441         if (rc == 0)
2442                 rc = len;
2443         dout("writepages result %d\n", rc);
2444         return rc;
2445 }
2446 EXPORT_SYMBOL(ceph_osdc_writepages);
2447
2448 int ceph_osdc_setup(void)
2449 {
2450         BUG_ON(ceph_osd_request_cache);
2451         ceph_osd_request_cache = kmem_cache_create("ceph_osd_request",
2452                                         sizeof (struct ceph_osd_request),
2453                                         __alignof__(struct ceph_osd_request),
2454                                         0, NULL);
2455
2456         return ceph_osd_request_cache ? 0 : -ENOMEM;
2457 }
2458 EXPORT_SYMBOL(ceph_osdc_setup);
2459
2460 void ceph_osdc_cleanup(void)
2461 {
2462         BUG_ON(!ceph_osd_request_cache);
2463         kmem_cache_destroy(ceph_osd_request_cache);
2464         ceph_osd_request_cache = NULL;
2465 }
2466 EXPORT_SYMBOL(ceph_osdc_cleanup);
2467
2468 /*
2469  * handle incoming message
2470  */
2471 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
2472 {
2473         struct ceph_osd *osd = con->private;
2474         struct ceph_osd_client *osdc;
2475         int type = le16_to_cpu(msg->hdr.type);
2476
2477         if (!osd)
2478                 goto out;
2479         osdc = osd->o_osdc;
2480
2481         switch (type) {
2482         case CEPH_MSG_OSD_MAP:
2483                 ceph_osdc_handle_map(osdc, msg);
2484                 break;
2485         case CEPH_MSG_OSD_OPREPLY:
2486                 handle_reply(osdc, msg, con);
2487                 break;
2488         case CEPH_MSG_WATCH_NOTIFY:
2489                 handle_watch_notify(osdc, msg);
2490                 break;
2491
2492         default:
2493                 pr_err("received unknown message type %d %s\n", type,
2494                        ceph_msg_type_name(type));
2495         }
2496 out:
2497         ceph_msg_put(msg);
2498 }
2499
2500 /*
2501  * lookup and return message for incoming reply.  set up reply message
2502  * pages.
2503  */
2504 static struct ceph_msg *get_reply(struct ceph_connection *con,
2505                                   struct ceph_msg_header *hdr,
2506                                   int *skip)
2507 {
2508         struct ceph_osd *osd = con->private;
2509         struct ceph_osd_client *osdc = osd->o_osdc;
2510         struct ceph_msg *m;
2511         struct ceph_osd_request *req;
2512         int front = le32_to_cpu(hdr->front_len);
2513         int data_len = le32_to_cpu(hdr->data_len);
2514         u64 tid;
2515
2516         tid = le64_to_cpu(hdr->tid);
2517         mutex_lock(&osdc->request_mutex);
2518         req = __lookup_request(osdc, tid);
2519         if (!req) {
2520                 *skip = 1;
2521                 m = NULL;
2522                 dout("get_reply unknown tid %llu from osd%d\n", tid,
2523                      osd->o_osd);
2524                 goto out;
2525         }
2526
2527         if (req->r_reply->con)
2528                 dout("%s revoking msg %p from old con %p\n", __func__,
2529                      req->r_reply, req->r_reply->con);
2530         ceph_msg_revoke_incoming(req->r_reply);
2531
2532         if (front > req->r_reply->front.iov_len) {
2533                 pr_warning("get_reply front %d > preallocated %d\n",
2534                            front, (int)req->r_reply->front.iov_len);
2535                 m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2536                 if (!m)
2537                         goto out;
2538                 ceph_msg_put(req->r_reply);
2539                 req->r_reply = m;
2540         }
2541         m = ceph_msg_get(req->r_reply);
2542
2543         if (data_len > 0) {
2544                 struct ceph_osd_data *osd_data;
2545
2546                 /*
2547                  * XXX This is assuming there is only one op containing
2548                  * XXX page data.  Probably OK for reads, but this
2549                  * XXX ought to be done more generally.
2550                  */
2551                 osd_data = osd_req_op_extent_osd_data(req, 0);
2552                 if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
2553                         if (osd_data->pages &&
2554                                 unlikely(osd_data->length < data_len)) {
2555
2556                                 pr_warning("tid %lld reply has %d bytes "
2557                                         "we had only %llu bytes ready\n",
2558                                         tid, data_len, osd_data->length);
2559                                 *skip = 1;
2560                                 ceph_msg_put(m);
2561                                 m = NULL;
2562                                 goto out;
2563                         }
2564                 }
2565         }
2566         *skip = 0;
2567         dout("get_reply tid %lld %p\n", tid, m);
2568
2569 out:
2570         mutex_unlock(&osdc->request_mutex);
2571         return m;
2572
2573 }
2574
2575 static struct ceph_msg *alloc_msg(struct ceph_connection *con,
2576                                   struct ceph_msg_header *hdr,
2577                                   int *skip)
2578 {
2579         struct ceph_osd *osd = con->private;
2580         int type = le16_to_cpu(hdr->type);
2581         int front = le32_to_cpu(hdr->front_len);
2582
2583         *skip = 0;
2584         switch (type) {
2585         case CEPH_MSG_OSD_MAP:
2586         case CEPH_MSG_WATCH_NOTIFY:
2587                 return ceph_msg_new(type, front, GFP_NOFS, false);
2588         case CEPH_MSG_OSD_OPREPLY:
2589                 return get_reply(con, hdr, skip);
2590         default:
2591                 pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
2592                         osd->o_osd);
2593                 *skip = 1;
2594                 return NULL;
2595         }
2596 }
2597
2598 /*
2599  * Wrappers to refcount containing ceph_osd struct
2600  */
2601 static struct ceph_connection *get_osd_con(struct ceph_connection *con)
2602 {
2603         struct ceph_osd *osd = con->private;
2604         if (get_osd(osd))
2605                 return con;
2606         return NULL;
2607 }
2608
2609 static void put_osd_con(struct ceph_connection *con)
2610 {
2611         struct ceph_osd *osd = con->private;
2612         put_osd(osd);
2613 }
2614
2615 /*
2616  * authentication
2617  */
2618 /*
2619  * Note: returned pointer is the address of a structure that's
2620  * managed separately.  Caller must *not* attempt to free it.
2621  */
2622 static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
2623                                         int *proto, int force_new)
2624 {
2625         struct ceph_osd *o = con->private;
2626         struct ceph_osd_client *osdc = o->o_osdc;
2627         struct ceph_auth_client *ac = osdc->client->monc.auth;
2628         struct ceph_auth_handshake *auth = &o->o_auth;
2629
2630         if (force_new && auth->authorizer) {
2631                 ceph_auth_destroy_authorizer(ac, auth->authorizer);
2632                 auth->authorizer = NULL;
2633         }
2634         if (!auth->authorizer) {
2635                 int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2636                                                       auth);
2637                 if (ret)
2638                         return ERR_PTR(ret);
2639         } else {
2640                 int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2641                                                      auth);
2642                 if (ret)
2643                         return ERR_PTR(ret);
2644         }
2645         *proto = ac->protocol;
2646
2647         return auth;
2648 }
2649
2650
2651 static int verify_authorizer_reply(struct ceph_connection *con, int len)
2652 {
2653         struct ceph_osd *o = con->private;
2654         struct ceph_osd_client *osdc = o->o_osdc;
2655         struct ceph_auth_client *ac = osdc->client->monc.auth;
2656
2657         return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2658 }
2659
2660 static int invalidate_authorizer(struct ceph_connection *con)
2661 {
2662         struct ceph_osd *o = con->private;
2663         struct ceph_osd_client *osdc = o->o_osdc;
2664         struct ceph_auth_client *ac = osdc->client->monc.auth;
2665
2666         ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2667         return ceph_monc_validate_auth(&osdc->client->monc);
2668 }
2669
2670 static const struct ceph_connection_operations osd_con_ops = {
2671         .get = get_osd_con,
2672         .put = put_osd_con,
2673         .dispatch = dispatch,
2674         .get_authorizer = get_authorizer,
2675         .verify_authorizer_reply = verify_authorizer_reply,
2676         .invalidate_authorizer = invalidate_authorizer,
2677         .alloc_msg = alloc_msg,
2678         .fault = osd_reset,
2679 };