net/ceph/osd_client.c

   1 #include <linux/ceph/ceph_debug.h>
   2
   3 #include <linux/module.h>
   4 #include <linux/err.h>
   5 #include <linux/highmem.h>
   6 #include <linux/mm.h>
   7 #include <linux/pagemap.h>
   8 #include <linux/slab.h>
   9 #include <linux/uaccess.h>
  10 #ifdef CONFIG_BLOCK
  11 #include <linux/bio.h>
  12 #endif
  13
  14 #include <linux/ceph/libceph.h>
  15 #include <linux/ceph/osd_client.h>
  16 #include <linux/ceph/messenger.h>
  17 #include <linux/ceph/decode.h>
  18 #include <linux/ceph/auth.h>
  19 #include <linux/ceph/pagelist.h>
  20
  21 #define OSD_OP_FRONT_LEN        4096
  22 #define OSD_OPREPLY_FRONT_LEN   512
  23
  24 static const struct ceph_connection_operations osd_con_ops;
  25
  26 static void send_queued(struct ceph_osd_client *osdc);
  27 static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
  28 static void __register_request(struct ceph_osd_client *osdc,
  29                                struct ceph_osd_request *req);
  30 static void __unregister_linger_request(struct ceph_osd_client *osdc,
  31                                         struct ceph_osd_request *req);
  32 static void __send_request(struct ceph_osd_client *osdc,
  33                            struct ceph_osd_request *req);
  34
  35 static int op_needs_trail(int op)
  36 {
  37         switch (op) {
  38         case CEPH_OSD_OP_GETXATTR:
  39         case CEPH_OSD_OP_SETXATTR:
  40         case CEPH_OSD_OP_CMPXATTR:
  41         case CEPH_OSD_OP_CALL:
  42         case CEPH_OSD_OP_NOTIFY:
  43                 return 1;
  44         default:
  45                 return 0;
  46         }
  47 }
  48
  49 static int op_has_extent(int op)
  50 {
  51         return (op == CEPH_OSD_OP_READ ||
  52                 op == CEPH_OSD_OP_WRITE);
  53 }
  54
  55 void ceph_calc_raw_layout(struct ceph_osd_client *osdc,
  56                         struct ceph_file_layout *layout,
  57                         u64 snapid,
  58                         u64 off, u64 *plen, u64 *bno,
  59                         struct ceph_osd_request *req,
  60                         struct ceph_osd_req_op *op)
  61 {
  62         struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
  63         u64 orig_len = *plen;
  64         u64 objoff, objlen;    /* extent in object */
  65
  66         reqhead->snapid = cpu_to_le64(snapid);
  67
  68         /* object extent? */
  69         ceph_calc_file_object_mapping(layout, off, plen, bno,
  70                                       &objoff, &objlen);
  71         if (*plen < orig_len)
  72                 dout(" skipping last %llu, final file extent %llu~%llu\n",
  73                      orig_len - *plen, off, *plen);
  74
  75         if (op_has_extent(op->op)) {
  76                 op->extent.offset = objoff;
  77                 op->extent.length = objlen;
  78         }
  79         req->r_num_pages = calc_pages_for(off, *plen);
  80         req->r_page_alignment = off & ~PAGE_MASK;
  81         if (op->op == CEPH_OSD_OP_WRITE)
  82                 op->payload_len = *plen;
  83
  84         dout("calc_layout bno=%llx %llu~%llu (%d pages)\n",
  85              *bno, objoff, objlen, req->r_num_pages);
  86
  87 }
  88 EXPORT_SYMBOL(ceph_calc_raw_layout);
  89
  90 /*
  91  * Implement client access to distributed object storage cluster.
  92  *
  93  * All data objects are stored within a cluster/cloud of OSDs, or
  94  * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
  95  * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
  96  * remote daemons serving up and coordinating consistent and safe
  97  * access to storage.
  98  *
  99  * Cluster membership and the mapping of data objects onto storage devices
 100  * are described by the osd map.
 101  *
 102  * We keep track of pending OSD requests (read, write), resubmit
 103  * requests to different OSDs when the cluster topology/data layout
 104  * change, or retry the affected requests when the communications
 105  * channel with an OSD is reset.
 106  */
 107
 108 /*
 109  * calculate the mapping of a file extent onto an object, and fill out the
 110  * request accordingly.  shorten extent as necessary if it crosses an
 111  * object boundary.
 112  *
 113  * fill osd op in request message.
 114  */
 115 static void calc_layout(struct ceph_osd_client *osdc,
 116                         struct ceph_vino vino,
 117                         struct ceph_file_layout *layout,
 118                         u64 off, u64 *plen,
 119                         struct ceph_osd_request *req,
 120                         struct ceph_osd_req_op *op)
 121 {
 122         u64 bno;
 123
 124         ceph_calc_raw_layout(osdc, layout, vino.snap, off,
 125                              plen, &bno, req, op);
 126
 127         snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx", vino.ino, bno);
 128         req->r_oid_len = strlen(req->r_oid);
 129 }
 130
 131 /*
 132  * requests
 133  */
 134 void ceph_osdc_release_request(struct kref *kref)
 135 {
 136         struct ceph_osd_request *req = container_of(kref,
 137                                                     struct ceph_osd_request,
 138                                                     r_kref);
 139
 140         if (req->r_request)
 141                 ceph_msg_put(req->r_request);
 142         if (req->r_reply)
 143                 ceph_msg_put(req->r_reply);
 144         if (req->r_con_filling_msg) {
 145                 dout("release_request revoking pages %p from con %p\n",
 146                      req->r_pages, req->r_con_filling_msg);
 147                 ceph_con_revoke_message(req->r_con_filling_msg,
 148                                       req->r_reply);
 149                 ceph_con_put(req->r_con_filling_msg);
 150         }
 151         if (req->r_own_pages)
 152                 ceph_release_page_vector(req->r_pages,
 153                                          req->r_num_pages);
 154 #ifdef CONFIG_BLOCK
 155         if (req->r_bio)
 156                 bio_put(req->r_bio);
 157 #endif
 158         ceph_put_snap_context(req->r_snapc);
 159         if (req->r_trail) {
 160                 ceph_pagelist_release(req->r_trail);
 161                 kfree(req->r_trail);
 162         }
 163         if (req->r_mempool)
 164                 mempool_free(req, req->r_osdc->req_mempool);
 165         else
 166                 kfree(req);
 167 }
 168 EXPORT_SYMBOL(ceph_osdc_release_request);
 169
 170 static int get_num_ops(struct ceph_osd_req_op *ops, int *needs_trail)
 171 {
 172         int i = 0;
 173
 174         if (needs_trail)
 175                 *needs_trail = 0;
 176         while (ops[i].op) {
 177                 if (needs_trail && op_needs_trail(ops[i].op))
 178                         *needs_trail = 1;
 179                 i++;
 180         }
 181
 182         return i;
 183 }
 184
 185 struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
 186                                                int flags,
 187                                                struct ceph_snap_context *snapc,
 188                                                struct ceph_osd_req_op *ops,
 189                                                bool use_mempool,
 190                                                gfp_t gfp_flags,
 191                                                struct page **pages,
 192                                                struct bio *bio)
 193 {
 194         struct ceph_osd_request *req;
 195         struct ceph_msg *msg;
 196         int needs_trail;
 197         int num_op = get_num_ops(ops, &needs_trail);
 198         size_t msg_size = sizeof(struct ceph_osd_request_head);
 199
 200         msg_size += num_op*sizeof(struct ceph_osd_op);
 201
 202         if (use_mempool) {
 203                 req = mempool_alloc(osdc->req_mempool, gfp_flags);
 204                 memset(req, 0, sizeof(*req));
 205         } else {
 206                 req = kzalloc(sizeof(*req), gfp_flags);
 207         }
 208         if (req == NULL)
 209                 return NULL;
 210
 211         req->r_osdc = osdc;
 212         req->r_mempool = use_mempool;
 213
 214         kref_init(&req->r_kref);
 215         init_completion(&req->r_completion);
 216         init_completion(&req->r_safe_completion);
 217         INIT_LIST_HEAD(&req->r_unsafe_item);
 218         INIT_LIST_HEAD(&req->r_linger_item);
 219         INIT_LIST_HEAD(&req->r_linger_osd);
 220         INIT_LIST_HEAD(&req->r_req_lru_item);
 221         req->r_flags = flags;
 222
 223         WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0);
 224
 225         /* create reply message */
 226         if (use_mempool)
 227                 msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
 228         else
 229                 msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
 230                                    OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
 231         if (!msg) {
 232                 ceph_osdc_put_request(req);
 233                 return NULL;
 234         }
 235         req->r_reply = msg;
 236
 237         /* allocate space for the trailing data */
 238         if (needs_trail) {
 239                 req->r_trail = kmalloc(sizeof(struct ceph_pagelist), gfp_flags);
 240                 if (!req->r_trail) {
 241                         ceph_osdc_put_request(req);
 242                         return NULL;
 243                 }
 244                 ceph_pagelist_init(req->r_trail);
 245         }
 246         /* create request message; allow space for oid */
 247         msg_size += MAX_OBJ_NAME_SIZE;
 248         if (snapc)
 249                 msg_size += sizeof(u64) * snapc->num_snaps;
 250         if (use_mempool)
 251                 msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
 252         else
 253                 msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
 254         if (!msg) {
 255                 ceph_osdc_put_request(req);
 256                 return NULL;
 257         }
 258
 259         msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP);
 260         memset(msg->front.iov_base, 0, msg->front.iov_len);
 261
 262         req->r_request = msg;
 263         req->r_pages = pages;
 264 #ifdef CONFIG_BLOCK
 265         if (bio) {
 266                 req->r_bio = bio;
 267                 bio_get(req->r_bio);
 268         }
 269 #endif
 270
 271         return req;
 272 }
 273 EXPORT_SYMBOL(ceph_osdc_alloc_request);
 274
 275 static void osd_req_encode_op(struct ceph_osd_request *req,
 276                               struct ceph_osd_op *dst,
 277                               struct ceph_osd_req_op *src)
 278 {
 279         dst->op = cpu_to_le16(src->op);
 280
 281         switch (src->op) {
 282         case CEPH_OSD_OP_READ:
 283         case CEPH_OSD_OP_WRITE:
 284                 dst->extent.offset =
 285                         cpu_to_le64(src->extent.offset);
 286                 dst->extent.length =
 287                         cpu_to_le64(src->extent.length);
 288                 dst->extent.truncate_size =
 289                         cpu_to_le64(src->extent.truncate_size);
 290                 dst->extent.truncate_seq =
 291                         cpu_to_le32(src->extent.truncate_seq);
 292                 break;
 293
 294         case CEPH_OSD_OP_GETXATTR:
 295         case CEPH_OSD_OP_SETXATTR:
 296         case CEPH_OSD_OP_CMPXATTR:
 297                 BUG_ON(!req->r_trail);
 298
 299                 dst->xattr.name_len = cpu_to_le32(src->xattr.name_len);
 300                 dst->xattr.value_len = cpu_to_le32(src->xattr.value_len);
 301                 dst->xattr.cmp_op = src->xattr.cmp_op;
 302                 dst->xattr.cmp_mode = src->xattr.cmp_mode;
 303                 ceph_pagelist_append(req->r_trail, src->xattr.name,
 304                                      src->xattr.name_len);
 305                 ceph_pagelist_append(req->r_trail, src->xattr.val,
 306                                      src->xattr.value_len);
 307                 break;
 308         case CEPH_OSD_OP_CALL:
 309                 BUG_ON(!req->r_trail);
 310
 311                 dst->cls.class_len = src->cls.class_len;
 312                 dst->cls.method_len = src->cls.method_len;
 313                 dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
 314
 315                 ceph_pagelist_append(req->r_trail, src->cls.class_name,
 316                                      src->cls.class_len);
 317                 ceph_pagelist_append(req->r_trail, src->cls.method_name,
 318                                      src->cls.method_len);
 319                 ceph_pagelist_append(req->r_trail, src->cls.indata,
 320                                      src->cls.indata_len);
 321                 break;
 322         case CEPH_OSD_OP_ROLLBACK:
 323                 dst->snap.snapid = cpu_to_le64(src->snap.snapid);
 324                 break;
 325         case CEPH_OSD_OP_STARTSYNC:
 326                 break;
 327         case CEPH_OSD_OP_NOTIFY:
 328                 {
 329                         __le32 prot_ver = cpu_to_le32(src->watch.prot_ver);
 330                         __le32 timeout = cpu_to_le32(src->watch.timeout);
 331
 332                         BUG_ON(!req->r_trail);
 333
 334                         ceph_pagelist_append(req->r_trail,
 335                                                 &prot_ver, sizeof(prot_ver));
 336                         ceph_pagelist_append(req->r_trail,
 337                                                 &timeout, sizeof(timeout));
 338                 }
 339         case CEPH_OSD_OP_NOTIFY_ACK:
 340         case CEPH_OSD_OP_WATCH:
 341                 dst->watch.cookie = cpu_to_le64(src->watch.cookie);
 342                 dst->watch.ver = cpu_to_le64(src->watch.ver);
 343                 dst->watch.flag = src->watch.flag;
 344                 break;
 345         default:
 346                 pr_err("unrecognized osd opcode %d\n", dst->op);
 347                 WARN_ON(1);
 348                 break;
 349         }
 350         dst->payload_len = cpu_to_le32(src->payload_len);
 351 }
 352
 353 /*
 354  * build new request AND message
 355  *
 356  */
 357 void ceph_osdc_build_request(struct ceph_osd_request *req,
 358                              u64 off, u64 *plen,
 359                              struct ceph_osd_req_op *src_ops,
 360                              struct ceph_snap_context *snapc,
 361                              struct timespec *mtime,
 362                              const char *oid,
 363                              int oid_len)
 364 {
 365         struct ceph_msg *msg = req->r_request;
 366         struct ceph_osd_request_head *head;
 367         struct ceph_osd_req_op *src_op;
 368         struct ceph_osd_op *op;
 369         void *p;
 370         int num_op = get_num_ops(src_ops, NULL);
 371         size_t msg_size = sizeof(*head) + num_op*sizeof(*op);
 372         int flags = req->r_flags;
 373         u64 data_len = 0;
 374         int i;
 375
 376         head = msg->front.iov_base;
 377         op = (void *)(head + 1);
 378         p = (void *)(op + num_op);
 379
 380         req->r_snapc = ceph_get_snap_context(snapc);
 381
 382         head->client_inc = cpu_to_le32(1); /* always, for now. */
 383         head->flags = cpu_to_le32(flags);
 384         if (flags & CEPH_OSD_FLAG_WRITE)
 385                 ceph_encode_timespec(&head->mtime, mtime);
 386         head->num_ops = cpu_to_le16(num_op);
 387
 388
 389         /* fill in oid */
 390         head->object_len = cpu_to_le32(oid_len);
 391         memcpy(p, oid, oid_len);
 392         p += oid_len;
 393
 394         src_op = src_ops;
 395         while (src_op->op) {
 396                 osd_req_encode_op(req, op, src_op);
 397                 src_op++;
 398                 op++;
 399         }
 400
 401         if (req->r_trail)
 402                 data_len += req->r_trail->length;
 403
 404         if (snapc) {
 405                 head->snap_seq = cpu_to_le64(snapc->seq);
 406                 head->num_snaps = cpu_to_le32(snapc->num_snaps);
 407                 for (i = 0; i < snapc->num_snaps; i++) {
 408                         put_unaligned_le64(snapc->snaps[i], p);
 409                         p += sizeof(u64);
 410                 }
 411         }
 412
 413         if (flags & CEPH_OSD_FLAG_WRITE) {
 414                 req->r_request->hdr.data_off = cpu_to_le16(off);
 415                 req->r_request->hdr.data_len = cpu_to_le32(*plen + data_len);
 416         } else if (data_len) {
 417                 req->r_request->hdr.data_off = 0;
 418                 req->r_request->hdr.data_len = cpu_to_le32(data_len);
 419         }
 420
 421         req->r_request->page_alignment = req->r_page_alignment;
 422
 423         BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
 424         msg_size = p - msg->front.iov_base;
 425         msg->front.iov_len = msg_size;
 426         msg->hdr.front_len = cpu_to_le32(msg_size);
 427         return;
 428 }
 429 EXPORT_SYMBOL(ceph_osdc_build_request);
 430
 431 /*
 432  * build new request AND message, calculate layout, and adjust file
 433  * extent as needed.
 434  *
 435  * if the file was recently truncated, we include information about its
 436  * old and new size so that the object can be updated appropriately.  (we
 437  * avoid synchronously deleting truncated objects because it's slow.)
 438  *
 439  * if @do_sync, include a 'startsync' command so that the osd will flush
 440  * data quickly.
 441  */
 442 struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 443                                                struct ceph_file_layout *layout,
 444                                                struct ceph_vino vino,
 445                                                u64 off, u64 *plen,
 446                                                int opcode, int flags,
 447                                                struct ceph_snap_context *snapc,
 448                                                int do_sync,
 449                                                u32 truncate_seq,
 450                                                u64 truncate_size,
 451                                                struct timespec *mtime,
 452                                                bool use_mempool, int num_reply,
 453                                                int page_align)
 454 {
 455         struct ceph_osd_req_op ops[3];
 456         struct ceph_osd_request *req;
 457
 458         ops[0].op = opcode;
 459         ops[0].extent.truncate_seq = truncate_seq;
 460         ops[0].extent.truncate_size = truncate_size;
 461         ops[0].payload_len = 0;
 462
 463         if (do_sync) {
 464                 ops[1].op = CEPH_OSD_OP_STARTSYNC;
 465                 ops[1].payload_len = 0;
 466                 ops[2].op = 0;
 467         } else
 468                 ops[1].op = 0;
 469
 470         req = ceph_osdc_alloc_request(osdc, flags,
 471                                          snapc, ops,
 472                                          use_mempool,
 473                                          GFP_NOFS, NULL, NULL);
 474         if (!req)
 475                 return NULL;
 476
 477         /* calculate max write size */
 478         calc_layout(osdc, vino, layout, off, plen, req, ops);
 479         req->r_file_layout = *layout;  /* keep a copy */
 480
 481         /* in case it differs from natural (file) alignment that
 482            calc_layout filled in for us */
 483         req->r_num_pages = calc_pages_for(page_align, *plen);
 484         req->r_page_alignment = page_align;
 485
 486         ceph_osdc_build_request(req, off, plen, ops,
 487                                 snapc,
 488                                 mtime,
 489                                 req->r_oid, req->r_oid_len);
 490
 491         return req;
 492 }
 493 EXPORT_SYMBOL(ceph_osdc_new_request);
 494
 495 /*
 496  * We keep osd requests in an rbtree, sorted by ->r_tid.
 497  */
 498 static void __insert_request(struct ceph_osd_client *osdc,
 499                              struct ceph_osd_request *new)
 500 {
 501         struct rb_node **p = &osdc->requests.rb_node;
 502         struct rb_node *parent = NULL;
 503         struct ceph_osd_request *req = NULL;
 504
 505         while (*p) {
 506                 parent = *p;
 507                 req = rb_entry(parent, struct ceph_osd_request, r_node);
 508                 if (new->r_tid < req->r_tid)
 509                         p = &(*p)->rb_left;
 510                 else if (new->r_tid > req->r_tid)
 511                         p = &(*p)->rb_right;
 512                 else
 513                         BUG();
 514         }
 515
 516         rb_link_node(&new->r_node, parent, p);
 517         rb_insert_color(&new->r_node, &osdc->requests);
 518 }
 519
 520 static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
 521                                                  u64 tid)
 522 {
 523         struct ceph_osd_request *req;
 524         struct rb_node *n = osdc->requests.rb_node;
 525
 526         while (n) {
 527                 req = rb_entry(n, struct ceph_osd_request, r_node);
 528                 if (tid < req->r_tid)
 529                         n = n->rb_left;
 530                 else if (tid > req->r_tid)
 531                         n = n->rb_right;
 532                 else
 533                         return req;
 534         }
 535         return NULL;
 536 }
 537
 538 static struct ceph_osd_request *
 539 __lookup_request_ge(struct ceph_osd_client *osdc,
 540                     u64 tid)
 541 {
 542         struct ceph_osd_request *req;
 543         struct rb_node *n = osdc->requests.rb_node;
 544
 545         while (n) {
 546                 req = rb_entry(n, struct ceph_osd_request, r_node);
 547                 if (tid < req->r_tid) {
 548                         if (!n->rb_left)
 549                                 return req;
 550                         n = n->rb_left;
 551                 } else if (tid > req->r_tid) {
 552                         n = n->rb_right;
 553                 } else {
 554                         return req;
 555                 }
 556         }
 557         return NULL;
 558 }
 559
 560 /*
 561  * Resubmit requests pending on the given osd.
 562  */
 563 static void __kick_osd_requests(struct ceph_osd_client *osdc,
 564                                 struct ceph_osd *osd)
 565 {
 566         struct ceph_osd_request *req, *nreq;
 567         int err;
 568
 569         dout("__kick_osd_requests osd%d\n", osd->o_osd);
 570         err = __reset_osd(osdc, osd);
 571         if (err == -EAGAIN)
 572                 return;
 573
 574         list_for_each_entry(req, &osd->o_requests, r_osd_item) {
 575                 list_move(&req->r_req_lru_item, &osdc->req_unsent);
 576                 dout("requeued %p tid %llu osd%d\n", req, req->r_tid,
 577                      osd->o_osd);
 578                 if (!req->r_linger)
 579                         req->r_flags |= CEPH_OSD_FLAG_RETRY;
 580         }
 581
 582         list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
 583                                  r_linger_osd) {
 584                 /*
 585                  * reregister request prior to unregistering linger so
 586                  * that r_osd is preserved.
 587                  */
 588                 BUG_ON(!list_empty(&req->r_req_lru_item));
 589                 __register_request(osdc, req);
 590                 list_add(&req->r_req_lru_item, &osdc->req_unsent);
 591                 list_add(&req->r_osd_item, &req->r_osd->o_requests);
 592                 __unregister_linger_request(osdc, req);
 593                 dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
 594                      osd->o_osd);
 595         }
 596 }
 597
 598 static void kick_osd_requests(struct ceph_osd_client *osdc,
 599                               struct ceph_osd *kickosd)
 600 {
 601         mutex_lock(&osdc->request_mutex);
 602         __kick_osd_requests(osdc, kickosd);
 603         mutex_unlock(&osdc->request_mutex);
 604 }
 605
 606 /*
 607  * If the osd connection drops, we need to resubmit all requests.
 608  */
 609 static void osd_reset(struct ceph_connection *con)
 610 {
 611         struct ceph_osd *osd = con->private;
 612         struct ceph_osd_client *osdc;
 613
 614         if (!osd)
 615                 return;
 616         dout("osd_reset osd%d\n", osd->o_osd);
 617         osdc = osd->o_osdc;
 618         down_read(&osdc->map_sem);
 619         kick_osd_requests(osdc, osd);
 620         send_queued(osdc);
 621         up_read(&osdc->map_sem);
 622 }
 623
 624 /*
 625  * Track open sessions with osds.
 626  */
 627 static struct ceph_osd *create_osd(struct ceph_osd_client *osdc)
 628 {
 629         struct ceph_osd *osd;
 630
 631         osd = kzalloc(sizeof(*osd), GFP_NOFS);
 632         if (!osd)
 633                 return NULL;
 634
 635         atomic_set(&osd->o_ref, 1);
 636         osd->o_osdc = osdc;
 637         INIT_LIST_HEAD(&osd->o_requests);
 638         INIT_LIST_HEAD(&osd->o_linger_requests);
 639         INIT_LIST_HEAD(&osd->o_osd_lru);
 640         osd->o_incarnation = 1;
 641
 642         ceph_con_init(osdc->client->msgr, &osd->o_con);
 643         osd->o_con.private = osd;
 644         osd->o_con.ops = &osd_con_ops;
 645         osd->o_con.peer_name.type = CEPH_ENTITY_TYPE_OSD;
 646
 647         INIT_LIST_HEAD(&osd->o_keepalive_item);
 648         return osd;
 649 }
 650
 651 static struct ceph_osd *get_osd(struct ceph_osd *osd)
 652 {
 653         if (atomic_inc_not_zero(&osd->o_ref)) {
 654                 dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
 655                      atomic_read(&osd->o_ref));
 656                 return osd;
 657         } else {
 658                 dout("get_osd %p FAIL\n", osd);
 659                 return NULL;
 660         }
 661 }
 662
 663 static void put_osd(struct ceph_osd *osd)
 664 {
 665         dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
 666              atomic_read(&osd->o_ref) - 1);
 667         if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
 668                 struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;
 669
 670                 if (ac->ops && ac->ops->destroy_authorizer)
 671                         ac->ops->destroy_authorizer(ac, osd->o_auth.authorizer);
 672                 kfree(osd);
 673         }
 674 }
 675
 676 /*
 677  * remove an osd from our map
 678  */
 679 static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
 680 {
 681         dout("__remove_osd %p\n", osd);
 682         BUG_ON(!list_empty(&osd->o_requests));
 683         rb_erase(&osd->o_node, &osdc->osds);
 684         list_del_init(&osd->o_osd_lru);
 685         ceph_con_close(&osd->o_con);
 686         put_osd(osd);
 687 }
 688
 689 static void remove_all_osds(struct ceph_osd_client *osdc)
 690 {
 691         dout("__remove_old_osds %p\n", osdc);
 692         mutex_lock(&osdc->request_mutex);
 693         while (!RB_EMPTY_ROOT(&osdc->osds)) {
 694                 struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
 695                                                 struct ceph_osd, o_node);
 696                 __remove_osd(osdc, osd);
 697         }
 698         mutex_unlock(&osdc->request_mutex);
 699 }
 700
 701 static void __move_osd_to_lru(struct ceph_osd_client *osdc,
 702                               struct ceph_osd *osd)
 703 {
 704         dout("__move_osd_to_lru %p\n", osd);
 705         BUG_ON(!list_empty(&osd->o_osd_lru));
 706         list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
 707         osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
 708 }
 709
 710 static void __remove_osd_from_lru(struct ceph_osd *osd)
 711 {
 712         dout("__remove_osd_from_lru %p\n", osd);
 713         if (!list_empty(&osd->o_osd_lru))
 714                 list_del_init(&osd->o_osd_lru);
 715 }
 716
 717 static void remove_old_osds(struct ceph_osd_client *osdc)
 718 {
 719         struct ceph_osd *osd, *nosd;
 720
 721         dout("__remove_old_osds %p\n", osdc);
 722         mutex_lock(&osdc->request_mutex);
 723         list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
 724                 if (time_before(jiffies, osd->lru_ttl))
 725                         break;
 726                 __remove_osd(osdc, osd);
 727         }
 728         mutex_unlock(&osdc->request_mutex);
 729 }
 730
 731 /*
 732  * reset osd connect
 733  */
 734 static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
 735 {
 736         struct ceph_osd_request *req;
 737         int ret = 0;
 738
 739         dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
 740         if (list_empty(&osd->o_requests) &&
 741             list_empty(&osd->o_linger_requests)) {
 742                 __remove_osd(osdc, osd);
 743         } else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd],
 744                           &osd->o_con.peer_addr,
 745                           sizeof(osd->o_con.peer_addr)) == 0 &&
 746                    !ceph_con_opened(&osd->o_con)) {
 747                 dout(" osd addr hasn't changed and connection never opened,"
 748                      " letting msgr retry");
 749                 /* touch each r_stamp for handle_timeout()'s benfit */
 750                 list_for_each_entry(req, &osd->o_requests, r_osd_item)
 751                         req->r_stamp = jiffies;
 752                 ret = -EAGAIN;
 753         } else {
 754                 ceph_con_close(&osd->o_con);
 755                 ceph_con_open(&osd->o_con, &osdc->osdmap->osd_addr[osd->o_osd]);
 756                 osd->o_incarnation++;
 757         }
 758         return ret;
 759 }
 760
 761 static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
 762 {
 763         struct rb_node **p = &osdc->osds.rb_node;
 764         struct rb_node *parent = NULL;
 765         struct ceph_osd *osd = NULL;
 766
 767         dout("__insert_osd %p osd%d\n", new, new->o_osd);
 768         while (*p) {
 769                 parent = *p;
 770                 osd = rb_entry(parent, struct ceph_osd, o_node);
 771                 if (new->o_osd < osd->o_osd)
 772                         p = &(*p)->rb_left;
 773                 else if (new->o_osd > osd->o_osd)
 774                         p = &(*p)->rb_right;
 775                 else
 776                         BUG();
 777         }
 778
 779         rb_link_node(&new->o_node, parent, p);
 780         rb_insert_color(&new->o_node, &osdc->osds);
 781 }
 782
 783 static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
 784 {
 785         struct ceph_osd *osd;
 786         struct rb_node *n = osdc->osds.rb_node;
 787
 788         while (n) {
 789                 osd = rb_entry(n, struct ceph_osd, o_node);
 790                 if (o < osd->o_osd)
 791                         n = n->rb_left;
 792                 else if (o > osd->o_osd)
 793                         n = n->rb_right;
 794                 else
 795                         return osd;
 796         }
 797         return NULL;
 798 }
 799
 800 static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
 801 {
 802         schedule_delayed_work(&osdc->timeout_work,
 803                         osdc->client->options->osd_keepalive_timeout * HZ);
 804 }
 805
 806 static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
 807 {
 808         cancel_delayed_work(&osdc->timeout_work);
 809 }
 810
 811 /*
 812  * Register request, assign tid.  If this is the first request, set up
 813  * the timeout event.
 814  */
 815 static void __register_request(struct ceph_osd_client *osdc,
 816                                struct ceph_osd_request *req)
 817 {
 818         req->r_tid = ++osdc->last_tid;
 819         req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
 820         dout("__register_request %p tid %lld\n", req, req->r_tid);
 821         __insert_request(osdc, req);
 822         ceph_osdc_get_request(req);
 823         osdc->num_requests++;
 824         if (osdc->num_requests == 1) {
 825                 dout(" first request, scheduling timeout\n");
 826                 __schedule_osd_timeout(osdc);
 827         }
 828 }
 829
 830 static void register_request(struct ceph_osd_client *osdc,
 831                              struct ceph_osd_request *req)
 832 {
 833         mutex_lock(&osdc->request_mutex);
 834         __register_request(osdc, req);
 835         mutex_unlock(&osdc->request_mutex);
 836 }
 837
 838 /*
 839  * called under osdc->request_mutex
 840  */
 841 static void __unregister_request(struct ceph_osd_client *osdc,
 842                                  struct ceph_osd_request *req)
 843 {
 844         if (RB_EMPTY_NODE(&req->r_node)) {
 845                 dout("__unregister_request %p tid %lld not registered\n",
 846                         req, req->r_tid);
 847                 return;
 848         }
 849
 850         dout("__unregister_request %p tid %lld\n", req, req->r_tid);
 851         rb_erase(&req->r_node, &osdc->requests);
 852         osdc->num_requests--;
 853
 854         if (req->r_osd) {
 855                 /* make sure the original request isn't in flight. */
 856                 ceph_con_revoke(&req->r_osd->o_con, req->r_request);
 857
 858                 list_del_init(&req->r_osd_item);
 859                 if (list_empty(&req->r_osd->o_requests) &&
 860                     list_empty(&req->r_osd->o_linger_requests)) {
 861                         dout("moving osd to %p lru\n", req->r_osd);
 862                         __move_osd_to_lru(osdc, req->r_osd);
 863                 }
 864                 if (list_empty(&req->r_linger_item))
 865                         req->r_osd = NULL;
 866         }
 867
 868         ceph_osdc_put_request(req);
 869
 870         list_del_init(&req->r_req_lru_item);
 871         if (osdc->num_requests == 0) {
 872                 dout(" no requests, canceling timeout\n");
 873                 __cancel_osd_timeout(osdc);
 874         }
 875 }
 876
 877 /*
 878  * Cancel a previously queued request message
 879  */
 880 static void __cancel_request(struct ceph_osd_request *req)
 881 {
 882         if (req->r_sent && req->r_osd) {
 883                 ceph_con_revoke(&req->r_osd->o_con, req->r_request);
 884                 req->r_sent = 0;
 885         }
 886 }
 887
 888 static void __register_linger_request(struct ceph_osd_client *osdc,
 889                                     struct ceph_osd_request *req)
 890 {
 891         dout("__register_linger_request %p\n", req);
 892         list_add_tail(&req->r_linger_item, &osdc->req_linger);
 893         list_add_tail(&req->r_linger_osd, &req->r_osd->o_linger_requests);
 894 }
 895
 896 static void __unregister_linger_request(struct ceph_osd_client *osdc,
 897                                         struct ceph_osd_request *req)
 898 {
 899         dout("__unregister_linger_request %p\n", req);
 900         if (req->r_osd) {
 901                 list_del_init(&req->r_linger_item);
 902                 list_del_init(&req->r_linger_osd);
 903
 904                 if (list_empty(&req->r_osd->o_requests) &&
 905                     list_empty(&req->r_osd->o_linger_requests)) {
 906                         dout("moving osd to %p lru\n", req->r_osd);
 907                         __move_osd_to_lru(osdc, req->r_osd);
 908                 }
 909                 if (list_empty(&req->r_osd_item))
 910                         req->r_osd = NULL;
 911         }
 912 }
 913
 914 void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
 915                                          struct ceph_osd_request *req)
 916 {
 917         mutex_lock(&osdc->request_mutex);
 918         if (req->r_linger) {
 919                 __unregister_linger_request(osdc, req);
 920                 ceph_osdc_put_request(req);
 921         }
 922         mutex_unlock(&osdc->request_mutex);
 923 }
 924 EXPORT_SYMBOL(ceph_osdc_unregister_linger_request);
 925
 926 void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
 927                                   struct ceph_osd_request *req)
 928 {
 929         if (!req->r_linger) {
 930                 dout("set_request_linger %p\n", req);
 931                 req->r_linger = 1;
 932                 /*
 933                  * caller is now responsible for calling
 934                  * unregister_linger_request
 935                  */
 936                 ceph_osdc_get_request(req);
 937         }
 938 }
 939 EXPORT_SYMBOL(ceph_osdc_set_request_linger);
 940
 941 /*
 942  * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
 943  * (as needed), and set the request r_osd appropriately.  If there is
 944  * no up osd, set r_osd to NULL.  Move the request to the appropriate list
 945  * (unsent, homeless) or leave on in-flight lru.
 946  *
 947  * Return 0 if unchanged, 1 if changed, or negative on error.
 948  *
 949  * Caller should hold map_sem for read and request_mutex.
 950  */
 951 static int __map_request(struct ceph_osd_client *osdc,
 952                          struct ceph_osd_request *req, int force_resend)
 953 {
 954         struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
 955         struct ceph_pg pgid;
 956         int acting[CEPH_PG_MAX_SIZE];
 957         int o = -1, num = 0;
 958         int err;
 959
 960         dout("map_request %p tid %lld\n", req, req->r_tid);
 961         err = ceph_calc_object_layout(&reqhead->layout, req->r_oid,
 962                                       &req->r_file_layout, osdc->osdmap);
 963         if (err) {
 964                 list_move(&req->r_req_lru_item, &osdc->req_notarget);
 965                 return err;
 966         }
 967         pgid = reqhead->layout.ol_pgid;
 968         req->r_pgid = pgid;
 969
 970         err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
 971         if (err > 0) {
 972                 o = acting[0];
 973                 num = err;
 974         }
 975
 976         if ((!force_resend &&
 977              req->r_osd && req->r_osd->o_osd == o &&
 978              req->r_sent >= req->r_osd->o_incarnation &&
 979              req->r_num_pg_osds == num &&
 980              memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
 981             (req->r_osd == NULL && o == -1))
 982                 return 0;  /* no change */
 983
 984         dout("map_request tid %llu pgid %d.%x osd%d (was osd%d)\n",
 985              req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o,
 986              req->r_osd ? req->r_osd->o_osd : -1);
 987
 988         /* record full pg acting set */
 989         memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
 990         req->r_num_pg_osds = num;
 991
 992         if (req->r_osd) {
 993                 __cancel_request(req);
 994                 list_del_init(&req->r_osd_item);
 995                 req->r_osd = NULL;
 996         }
 997
 998         req->r_osd = __lookup_osd(osdc, o);
 999         if (!req->r_osd && o >= 0) {
1000                 err = -ENOMEM;
1001                 req->r_osd = create_osd(osdc);
1002                 if (!req->r_osd) {
1003                         list_move(&req->r_req_lru_item, &osdc->req_notarget);
1004                         goto out;
1005                 }
1006
1007                 dout("map_request osd %p is osd%d\n", req->r_osd, o);
1008                 req->r_osd->o_osd = o;
1009                 req->r_osd->o_con.peer_name.num = cpu_to_le64(o);
1010                 __insert_osd(osdc, req->r_osd);
1011
1012                 ceph_con_open(&req->r_osd->o_con, &osdc->osdmap->osd_addr[o]);
1013         }
1014
1015         if (req->r_osd) {
1016                 __remove_osd_from_lru(req->r_osd);
1017                 list_add(&req->r_osd_item, &req->r_osd->o_requests);
1018                 list_move(&req->r_req_lru_item, &osdc->req_unsent);
1019         } else {
1020                 list_move(&req->r_req_lru_item, &osdc->req_notarget);
1021         }
1022         err = 1;   /* osd or pg changed */
1023
1024 out:
1025         return err;
1026 }
1027
1028 /*
1029  * caller should hold map_sem (for read) and request_mutex
1030  */
1031 static void __send_request(struct ceph_osd_client *osdc,
1032                            struct ceph_osd_request *req)
1033 {
1034         struct ceph_osd_request_head *reqhead;
1035
1036         dout("send_request %p tid %llu to osd%d flags %d\n",
1037              req, req->r_tid, req->r_osd->o_osd, req->r_flags);
1038
1039         reqhead = req->r_request->front.iov_base;
1040         reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch);
1041         reqhead->flags |= cpu_to_le32(req->r_flags);  /* e.g., RETRY */
1042         reqhead->reassert_version = req->r_reassert_version;
1043
1044         req->r_stamp = jiffies;
1045         list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
1046
1047         ceph_msg_get(req->r_request); /* send consumes a ref */
1048         ceph_con_send(&req->r_osd->o_con, req->r_request);
1049         req->r_sent = req->r_osd->o_incarnation;
1050 }
1051
1052 /*
1053  * Send any requests in the queue (req_unsent).
1054  */
1055 static void send_queued(struct ceph_osd_client *osdc)
1056 {
1057         struct ceph_osd_request *req, *tmp;
1058
1059         dout("send_queued\n");
1060         mutex_lock(&osdc->request_mutex);
1061         list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) {
1062                 __send_request(osdc, req);
1063         }
1064         mutex_unlock(&osdc->request_mutex);
1065 }
1066
1067 /*
1068  * Timeout callback, called every N seconds when 1 or more osd
1069  * requests has been active for more than N seconds.  When this
1070  * happens, we ping all OSDs with requests who have timed out to
1071  * ensure any communications channel reset is detected.  Reset the
1072  * request timeouts another N seconds in the future as we go.
1073  * Reschedule the timeout event another N seconds in future (unless
1074  * there are no open requests).
1075  */
1076 static void handle_timeout(struct work_struct *work)
1077 {
1078         struct ceph_osd_client *osdc =
1079                 container_of(work, struct ceph_osd_client, timeout_work.work);
1080         struct ceph_osd_request *req, *last_req = NULL;
1081         struct ceph_osd *osd;
1082         unsigned long timeout = osdc->client->options->osd_timeout * HZ;
1083         unsigned long keepalive =
1084                 osdc->client->options->osd_keepalive_timeout * HZ;
1085         unsigned long last_stamp = 0;
1086         struct list_head slow_osds;
1087         dout("timeout\n");
1088         down_read(&osdc->map_sem);
1089
1090         ceph_monc_request_next_osdmap(&osdc->client->monc);
1091
1092         mutex_lock(&osdc->request_mutex);
1093
1094         /*
1095          * reset osds that appear to be _really_ unresponsive.  this
1096          * is a failsafe measure.. we really shouldn't be getting to
1097          * this point if the system is working properly.  the monitors
1098          * should mark the osd as failed and we should find out about
1099          * it from an updated osd map.
1100          */
1101         while (timeout && !list_empty(&osdc->req_lru)) {
1102                 req = list_entry(osdc->req_lru.next, struct ceph_osd_request,
1103                                  r_req_lru_item);
1104
1105                 /* hasn't been long enough since we sent it? */
1106                 if (time_before(jiffies, req->r_stamp + timeout))
1107                         break;
1108
1109                 /* hasn't been long enough since it was acked? */
1110                 if (req->r_request->ack_stamp == 0 ||
1111                     time_before(jiffies, req->r_request->ack_stamp + timeout))
1112                         break;
1113
1114                 BUG_ON(req == last_req && req->r_stamp == last_stamp);
1115                 last_req = req;
1116                 last_stamp = req->r_stamp;
1117
1118                 osd = req->r_osd;
1119                 BUG_ON(!osd);
1120                 pr_warning(" tid %llu timed out on osd%d, will reset osd\n",
1121                            req->r_tid, osd->o_osd);
1122                 __kick_osd_requests(osdc, osd);
1123         }
1124
1125         /*
1126          * ping osds that are a bit slow.  this ensures that if there
1127          * is a break in the TCP connection we will notice, and reopen
1128          * a connection with that osd (from the fault callback).
1129          */
1130         INIT_LIST_HEAD(&slow_osds);
1131         list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
1132                 if (time_before(jiffies, req->r_stamp + keepalive))
1133                         break;
1134
1135                 osd = req->r_osd;
1136                 BUG_ON(!osd);
1137                 dout(" tid %llu is slow, will send keepalive on osd%d\n",
1138                      req->r_tid, osd->o_osd);
1139                 list_move_tail(&osd->o_keepalive_item, &slow_osds);
1140         }
1141         while (!list_empty(&slow_osds)) {
1142                 osd = list_entry(slow_osds.next, struct ceph_osd,
1143                                  o_keepalive_item);
1144                 list_del_init(&osd->o_keepalive_item);
1145                 ceph_con_keepalive(&osd->o_con);
1146         }
1147
1148         __schedule_osd_timeout(osdc);
1149         mutex_unlock(&osdc->request_mutex);
1150         send_queued(osdc);
1151         up_read(&osdc->map_sem);
1152 }
1153
1154 static void handle_osds_timeout(struct work_struct *work)
1155 {
1156         struct ceph_osd_client *osdc =
1157                 container_of(work, struct ceph_osd_client,
1158                              osds_timeout_work.work);
1159         unsigned long delay =
1160                 osdc->client->options->osd_idle_ttl * HZ >> 2;
1161
1162         dout("osds timeout\n");
1163         down_read(&osdc->map_sem);
1164         remove_old_osds(osdc);
1165         up_read(&osdc->map_sem);
1166
1167         schedule_delayed_work(&osdc->osds_timeout_work,
1168                               round_jiffies_relative(delay));
1169 }
1170
1171 static void complete_request(struct ceph_osd_request *req)
1172 {
1173         if (req->r_safe_callback)
1174                 req->r_safe_callback(req, NULL);
1175         complete_all(&req->r_safe_completion);  /* fsync waiter */
1176 }
1177
1178 /*
1179  * handle osd op reply.  either call the callback if it is specified,
1180  * or do the completion to wake up the waiting thread.
1181  */
1182 static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
1183                          struct ceph_connection *con)
1184 {
1185         struct ceph_osd_reply_head *rhead = msg->front.iov_base;
1186         struct ceph_osd_request *req;
1187         u64 tid;
1188         int numops, object_len, flags;
1189         s32 result;
1190
1191         tid = le64_to_cpu(msg->hdr.tid);
1192         if (msg->front.iov_len < sizeof(*rhead))
1193                 goto bad;
1194         numops = le32_to_cpu(rhead->num_ops);
1195         object_len = le32_to_cpu(rhead->object_len);
1196         result = le32_to_cpu(rhead->result);
1197         if (msg->front.iov_len != sizeof(*rhead) + object_len +
1198             numops * sizeof(struct ceph_osd_op))
1199                 goto bad;
1200         dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result);
1201         /* lookup */
1202         mutex_lock(&osdc->request_mutex);
1203         req = __lookup_request(osdc, tid);
1204         if (req == NULL) {
1205                 dout("handle_reply tid %llu dne\n", tid);
1206                 mutex_unlock(&osdc->request_mutex);
1207                 return;
1208         }
1209         ceph_osdc_get_request(req);
1210         flags = le32_to_cpu(rhead->flags);
1211
1212         /*
1213          * if this connection filled our message, drop our reference now, to
1214          * avoid a (safe but slower) revoke later.
1215          */
1216         if (req->r_con_filling_msg == con && req->r_reply == msg) {
1217                 dout(" dropping con_filling_msg ref %p\n", con);
1218                 req->r_con_filling_msg = NULL;
1219                 ceph_con_put(con);
1220         }
1221
1222         if (!req->r_got_reply) {
1223                 unsigned bytes;
1224
1225                 req->r_result = le32_to_cpu(rhead->result);
1226                 bytes = le32_to_cpu(msg->hdr.data_len);
1227                 dout("handle_reply result %d bytes %d\n", req->r_result,
1228                      bytes);
1229                 if (req->r_result == 0)
1230                         req->r_result = bytes;
1231
1232                 /* in case this is a write and we need to replay, */
1233                 req->r_reassert_version = rhead->reassert_version;
1234
1235                 req->r_got_reply = 1;
1236         } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
1237                 dout("handle_reply tid %llu dup ack\n", tid);
1238                 mutex_unlock(&osdc->request_mutex);
1239                 goto done;
1240         }
1241
1242         dout("handle_reply tid %llu flags %d\n", tid, flags);
1243
1244         if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
1245                 __register_linger_request(osdc, req);
1246
1247         /* either this is a read, or we got the safe response */
1248         if (result < 0 ||
1249             (flags & CEPH_OSD_FLAG_ONDISK) ||
1250             ((flags & CEPH_OSD_FLAG_WRITE) == 0))
1251                 __unregister_request(osdc, req);
1252
1253         mutex_unlock(&osdc->request_mutex);
1254
1255         if (req->r_callback)
1256                 req->r_callback(req, msg);
1257         else
1258                 complete_all(&req->r_completion);
1259
1260         if (flags & CEPH_OSD_FLAG_ONDISK)
1261                 complete_request(req);
1262
1263 done:
1264         dout("req=%p req->r_linger=%d\n", req, req->r_linger);
1265         ceph_osdc_put_request(req);
1266         return;
1267
1268 bad:
1269         pr_err("corrupt osd_op_reply got %d %d expected %d\n",
1270                (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len),
1271                (int)sizeof(*rhead));
1272         ceph_msg_dump(msg);
1273 }
1274
1275 static void reset_changed_osds(struct ceph_osd_client *osdc)
1276 {
1277         struct rb_node *p, *n;
1278
1279         for (p = rb_first(&osdc->osds); p; p = n) {
1280                 struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
1281
1282                 n = rb_next(p);
1283                 if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
1284                     memcmp(&osd->o_con.peer_addr,
1285                            ceph_osd_addr(osdc->osdmap,
1286                                          osd->o_osd),
1287                            sizeof(struct ceph_entity_addr)) != 0)
1288                         __reset_osd(osdc, osd);
1289         }
1290 }
1291
1292 /*
1293  * Requeue requests whose mapping to an OSD has changed.  If requests map to
1294  * no osd, request a new map.
1295  *
1296  * Caller should hold map_sem for read and request_mutex.
1297  */
1298 static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1299 {
1300         struct ceph_osd_request *req, *nreq;
1301         struct rb_node *p;
1302         int needmap = 0;
1303         int err;
1304
1305         dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1306         mutex_lock(&osdc->request_mutex);
1307         for (p = rb_first(&osdc->requests); p; p = rb_next(p)) {
1308                 req = rb_entry(p, struct ceph_osd_request, r_node);
1309                 err = __map_request(osdc, req, force_resend);
1310                 if (err < 0)
1311                         continue;  /* error */
1312                 if (req->r_osd == NULL) {
1313                         dout("%p tid %llu maps to no osd\n", req, req->r_tid);
1314                         needmap++;  /* request a newer map */
1315                 } else if (err > 0) {
1316                         dout("%p tid %llu requeued on osd%d\n", req, req->r_tid,
1317                              req->r_osd ? req->r_osd->o_osd : -1);
1318                         if (!req->r_linger)
1319                                 req->r_flags |= CEPH_OSD_FLAG_RETRY;
1320                 }
1321         }
1322
1323         list_for_each_entry_safe(req, nreq, &osdc->req_linger,
1324                                  r_linger_item) {
1325                 dout("linger req=%p req->r_osd=%p\n", req, req->r_osd);
1326
1327                 err = __map_request(osdc, req, force_resend);
1328                 if (err == 0)
1329                         continue;  /* no change and no osd was specified */
1330                 if (err < 0)
1331                         continue;  /* hrm! */
1332                 if (req->r_osd == NULL) {
1333                         dout("tid %llu maps to no valid osd\n", req->r_tid);
1334                         needmap++;  /* request a newer map */
1335                         continue;
1336                 }
1337
1338                 dout("kicking lingering %p tid %llu osd%d\n", req, req->r_tid,
1339                      req->r_osd ? req->r_osd->o_osd : -1);
1340                 __unregister_linger_request(osdc, req);
1341                 __register_request(osdc, req);
1342         }
1343         mutex_unlock(&osdc->request_mutex);
1344
1345         if (needmap) {
1346                 dout("%d requests for down osds, need new map\n", needmap);
1347                 ceph_monc_request_next_osdmap(&osdc->client->monc);
1348         }
1349 }
1350
1351
1352 /*
1353  * Process updated osd map.
1354  *
1355  * The message contains any number of incremental and full maps, normally
1356  * indicating some sort of topology change in the cluster.  Kick requests
1357  * off to different OSDs as needed.
1358  */
1359 void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
1360 {
1361         void *p, *end, *next;
1362         u32 nr_maps, maplen;
1363         u32 epoch;
1364         struct ceph_osdmap *newmap = NULL, *oldmap;
1365         int err;
1366         struct ceph_fsid fsid;
1367
1368         dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
1369         p = msg->front.iov_base;
1370         end = p + msg->front.iov_len;
1371
1372         /* verify fsid */
1373         ceph_decode_need(&p, end, sizeof(fsid), bad);
1374         ceph_decode_copy(&p, &fsid, sizeof(fsid));
1375         if (ceph_check_fsid(osdc->client, &fsid) < 0)
1376                 return;
1377
1378         down_write(&osdc->map_sem);
1379
1380         /* incremental maps */
1381         ceph_decode_32_safe(&p, end, nr_maps, bad);
1382         dout(" %d inc maps\n", nr_maps);
1383         while (nr_maps > 0) {
1384                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1385                 epoch = ceph_decode_32(&p);
1386                 maplen = ceph_decode_32(&p);
1387                 ceph_decode_need(&p, end, maplen, bad);
1388                 next = p + maplen;
1389                 if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
1390                         dout("applying incremental map %u len %d\n",
1391                              epoch, maplen);
1392                         newmap = osdmap_apply_incremental(&p, next,
1393                                                           osdc->osdmap,
1394                                                           osdc->client->msgr);
1395                         if (IS_ERR(newmap)) {
1396                                 err = PTR_ERR(newmap);
1397                                 goto bad;
1398                         }
1399                         BUG_ON(!newmap);
1400                         if (newmap != osdc->osdmap) {
1401                                 ceph_osdmap_destroy(osdc->osdmap);
1402                                 osdc->osdmap = newmap;
1403                         }
1404                         kick_requests(osdc, 0);
1405                         reset_changed_osds(osdc);
1406                 } else {
1407                         dout("ignoring incremental map %u len %d\n",
1408                              epoch, maplen);
1409                 }
1410                 p = next;
1411                 nr_maps--;
1412         }
1413         if (newmap)
1414                 goto done;
1415
1416         /* full maps */
1417         ceph_decode_32_safe(&p, end, nr_maps, bad);
1418         dout(" %d full maps\n", nr_maps);
1419         while (nr_maps) {
1420                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1421                 epoch = ceph_decode_32(&p);
1422                 maplen = ceph_decode_32(&p);
1423                 ceph_decode_need(&p, end, maplen, bad);
1424                 if (nr_maps > 1) {
1425                         dout("skipping non-latest full map %u len %d\n",
1426                              epoch, maplen);
1427                 } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
1428                         dout("skipping full map %u len %d, "
1429                              "older than our %u\n", epoch, maplen,
1430                              osdc->osdmap->epoch);
1431                 } else {
1432                         int skipped_map = 0;
1433
1434                         dout("taking full map %u len %d\n", epoch, maplen);
1435                         newmap = osdmap_decode(&p, p+maplen);
1436                         if (IS_ERR(newmap)) {
1437                                 err = PTR_ERR(newmap);
1438                                 goto bad;
1439                         }
1440                         BUG_ON(!newmap);
1441                         oldmap = osdc->osdmap;
1442                         osdc->osdmap = newmap;
1443                         if (oldmap) {
1444                                 if (oldmap->epoch + 1 < newmap->epoch)
1445                                         skipped_map = 1;
1446                                 ceph_osdmap_destroy(oldmap);
1447                         }
1448                         kick_requests(osdc, skipped_map);
1449                 }
1450                 p += maplen;
1451                 nr_maps--;
1452         }
1453
1454 done:
1455         downgrade_write(&osdc->map_sem);
1456         ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1457
1458         /*
1459          * subscribe to subsequent osdmap updates if full to ensure
1460          * we find out when we are no longer full and stop returning
1461          * ENOSPC.
1462          */
1463         if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL))
1464                 ceph_monc_request_next_osdmap(&osdc->client->monc);
1465
1466         send_queued(osdc);
1467         up_read(&osdc->map_sem);
1468         wake_up_all(&osdc->client->auth_wq);
1469         return;
1470
1471 bad:
1472         pr_err("osdc handle_map corrupt msg\n");
1473         ceph_msg_dump(msg);
1474         up_write(&osdc->map_sem);
1475         return;
1476 }
1477
1478 /*
1479  * watch/notify callback event infrastructure
1480  *
1481  * These callbacks are used both for watch and notify operations.
1482  */
1483 static void __release_event(struct kref *kref)
1484 {
1485         struct ceph_osd_event *event =
1486                 container_of(kref, struct ceph_osd_event, kref);
1487
1488         dout("__release_event %p\n", event);
1489         kfree(event);
1490 }
1491
1492 static void get_event(struct ceph_osd_event *event)
1493 {
1494         kref_get(&event->kref);
1495 }
1496
1497 void ceph_osdc_put_event(struct ceph_osd_event *event)
1498 {
1499         kref_put(&event->kref, __release_event);
1500 }
1501 EXPORT_SYMBOL(ceph_osdc_put_event);
1502
1503 static void __insert_event(struct ceph_osd_client *osdc,
1504                              struct ceph_osd_event *new)
1505 {
1506         struct rb_node **p = &osdc->event_tree.rb_node;
1507         struct rb_node *parent = NULL;
1508         struct ceph_osd_event *event = NULL;
1509
1510         while (*p) {
1511                 parent = *p;
1512                 event = rb_entry(parent, struct ceph_osd_event, node);
1513                 if (new->cookie < event->cookie)
1514                         p = &(*p)->rb_left;
1515                 else if (new->cookie > event->cookie)
1516                         p = &(*p)->rb_right;
1517                 else
1518                         BUG();
1519         }
1520
1521         rb_link_node(&new->node, parent, p);
1522         rb_insert_color(&new->node, &osdc->event_tree);
1523 }
1524
1525 static struct ceph_osd_event *__find_event(struct ceph_osd_client *osdc,
1526                                                 u64 cookie)
1527 {
1528         struct rb_node **p = &osdc->event_tree.rb_node;
1529         struct rb_node *parent = NULL;
1530         struct ceph_osd_event *event = NULL;
1531
1532         while (*p) {
1533                 parent = *p;
1534                 event = rb_entry(parent, struct ceph_osd_event, node);
1535                 if (cookie < event->cookie)
1536                         p = &(*p)->rb_left;
1537                 else if (cookie > event->cookie)
1538                         p = &(*p)->rb_right;
1539                 else
1540                         return event;
1541         }
1542         return NULL;
1543 }
1544
1545 static void __remove_event(struct ceph_osd_event *event)
1546 {
1547         struct ceph_osd_client *osdc = event->osdc;
1548
1549         if (!RB_EMPTY_NODE(&event->node)) {
1550                 dout("__remove_event removed %p\n", event);
1551                 rb_erase(&event->node, &osdc->event_tree);
1552                 ceph_osdc_put_event(event);
1553         } else {
1554                 dout("__remove_event didn't remove %p\n", event);
1555         }
1556 }
1557
1558 int ceph_osdc_create_event(struct ceph_osd_client *osdc,
1559                            void (*event_cb)(u64, u64, u8, void *),
1560                            int one_shot, void *data,
1561                            struct ceph_osd_event **pevent)
1562 {
1563         struct ceph_osd_event *event;
1564
1565         event = kmalloc(sizeof(*event), GFP_NOIO);
1566         if (!event)
1567                 return -ENOMEM;
1568
1569         dout("create_event %p\n", event);
1570         event->cb = event_cb;
1571         event->one_shot = one_shot;
1572         event->data = data;
1573         event->osdc = osdc;
1574         INIT_LIST_HEAD(&event->osd_node);
1575         kref_init(&event->kref);   /* one ref for us */
1576         kref_get(&event->kref);    /* one ref for the caller */
1577         init_completion(&event->completion);
1578
1579         spin_lock(&osdc->event_lock);
1580         event->cookie = ++osdc->event_count;
1581         __insert_event(osdc, event);
1582         spin_unlock(&osdc->event_lock);
1583
1584         *pevent = event;
1585         return 0;
1586 }
1587 EXPORT_SYMBOL(ceph_osdc_create_event);
1588
1589 void ceph_osdc_cancel_event(struct ceph_osd_event *event)
1590 {
1591         struct ceph_osd_client *osdc = event->osdc;
1592
1593         dout("cancel_event %p\n", event);
1594         spin_lock(&osdc->event_lock);
1595         __remove_event(event);
1596         spin_unlock(&osdc->event_lock);
1597         ceph_osdc_put_event(event); /* caller's */
1598 }
1599 EXPORT_SYMBOL(ceph_osdc_cancel_event);
1600
1601
1602 static void do_event_work(struct work_struct *work)
1603 {
1604         struct ceph_osd_event_work *event_work =
1605                 container_of(work, struct ceph_osd_event_work, work);
1606         struct ceph_osd_event *event = event_work->event;
1607         u64 ver = event_work->ver;
1608         u64 notify_id = event_work->notify_id;
1609         u8 opcode = event_work->opcode;
1610
1611         dout("do_event_work completing %p\n", event);
1612         event->cb(ver, notify_id, opcode, event->data);
1613         complete(&event->completion);
1614         dout("do_event_work completed %p\n", event);
1615         ceph_osdc_put_event(event);
1616         kfree(event_work);
1617 }
1618
1619
1620 /*
1621  * Process osd watch notifications
1622  */
1623 void handle_watch_notify(struct ceph_osd_client *osdc, struct ceph_msg *msg)
1624 {
1625         void *p, *end;
1626         u8 proto_ver;
1627         u64 cookie, ver, notify_id;
1628         u8 opcode;
1629         struct ceph_osd_event *event;
1630         struct ceph_osd_event_work *event_work;
1631
1632         p = msg->front.iov_base;
1633         end = p + msg->front.iov_len;
1634
1635         ceph_decode_8_safe(&p, end, proto_ver, bad);
1636         ceph_decode_8_safe(&p, end, opcode, bad);
1637         ceph_decode_64_safe(&p, end, cookie, bad);
1638         ceph_decode_64_safe(&p, end, ver, bad);
1639         ceph_decode_64_safe(&p, end, notify_id, bad);
1640
1641         spin_lock(&osdc->event_lock);
1642         event = __find_event(osdc, cookie);
1643         if (event) {
1644                 get_event(event);
1645                 if (event->one_shot)
1646                         __remove_event(event);
1647         }
1648         spin_unlock(&osdc->event_lock);
1649         dout("handle_watch_notify cookie %lld ver %lld event %p\n",
1650              cookie, ver, event);
1651         if (event) {
1652                 event_work = kmalloc(sizeof(*event_work), GFP_NOIO);
1653                 if (!event_work) {
1654                         dout("ERROR: could not allocate event_work\n");
1655                         goto done_err;
1656                 }
1657                 INIT_WORK(&event_work->work, do_event_work);
1658                 event_work->event = event;
1659                 event_work->ver = ver;
1660                 event_work->notify_id = notify_id;
1661                 event_work->opcode = opcode;
1662                 if (!queue_work(osdc->notify_wq, &event_work->work)) {
1663                         dout("WARNING: failed to queue notify event work\n");
1664                         goto done_err;
1665                 }
1666         }
1667
1668         return;
1669
1670 done_err:
1671         complete(&event->completion);
1672         ceph_osdc_put_event(event);
1673         return;
1674
1675 bad:
1676         pr_err("osdc handle_watch_notify corrupt msg\n");
1677         return;
1678 }
1679
1680 int ceph_osdc_wait_event(struct ceph_osd_event *event, unsigned long timeout)
1681 {
1682         int err;
1683
1684         dout("wait_event %p\n", event);
1685         err = wait_for_completion_interruptible_timeout(&event->completion,
1686                                                         timeout * HZ);
1687         ceph_osdc_put_event(event);
1688         if (err > 0)
1689                 err = 0;
1690         dout("wait_event %p returns %d\n", event, err);
1691         return err;
1692 }
1693 EXPORT_SYMBOL(ceph_osdc_wait_event);
1694
1695 /*
1696  * Register request, send initial attempt.
1697  */
1698 int ceph_osdc_start_request(struct ceph_osd_client *osdc,
1699                             struct ceph_osd_request *req,
1700                             bool nofail)
1701 {
1702         int rc = 0;
1703
1704         req->r_request->pages = req->r_pages;
1705         req->r_request->nr_pages = req->r_num_pages;
1706 #ifdef CONFIG_BLOCK
1707         req->r_request->bio = req->r_bio;
1708 #endif
1709         req->r_request->trail = req->r_trail;
1710
1711         register_request(osdc, req);
1712
1713         down_read(&osdc->map_sem);
1714         mutex_lock(&osdc->request_mutex);
1715         /*
1716          * a racing kick_requests() may have sent the message for us
1717          * while we dropped request_mutex above, so only send now if
1718          * the request still han't been touched yet.
1719          */
1720         if (req->r_sent == 0) {
1721                 rc = __map_request(osdc, req, 0);
1722                 if (rc < 0) {
1723                         if (nofail) {
1724                                 dout("osdc_start_request failed map, "
1725                                      " will retry %lld\n", req->r_tid);
1726                                 rc = 0;
1727                         }
1728                         goto out_unlock;
1729                 }
1730                 if (req->r_osd == NULL) {
1731                         dout("send_request %p no up osds in pg\n", req);
1732                         ceph_monc_request_next_osdmap(&osdc->client->monc);
1733                 } else {
1734                         __send_request(osdc, req);
1735                 }
1736                 rc = 0;
1737         }
1738
1739 out_unlock:
1740         mutex_unlock(&osdc->request_mutex);
1741         up_read(&osdc->map_sem);
1742         return rc;
1743 }
1744 EXPORT_SYMBOL(ceph_osdc_start_request);
1745
1746 /*
1747  * wait for a request to complete
1748  */
1749 int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
1750                            struct ceph_osd_request *req)
1751 {
1752         int rc;
1753
1754         rc = wait_for_completion_interruptible(&req->r_completion);
1755         if (rc < 0) {
1756                 mutex_lock(&osdc->request_mutex);
1757                 __cancel_request(req);
1758                 __unregister_request(osdc, req);
1759                 mutex_unlock(&osdc->request_mutex);
1760                 complete_request(req);
1761                 dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
1762                 return rc;
1763         }
1764
1765         dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
1766         return req->r_result;
1767 }
1768 EXPORT_SYMBOL(ceph_osdc_wait_request);
1769
1770 /*
1771  * sync - wait for all in-flight requests to flush.  avoid starvation.
1772  */
1773 void ceph_osdc_sync(struct ceph_osd_client *osdc)
1774 {
1775         struct ceph_osd_request *req;
1776         u64 last_tid, next_tid = 0;
1777
1778         mutex_lock(&osdc->request_mutex);
1779         last_tid = osdc->last_tid;
1780         while (1) {
1781                 req = __lookup_request_ge(osdc, next_tid);
1782                 if (!req)
1783                         break;
1784                 if (req->r_tid > last_tid)
1785                         break;
1786
1787                 next_tid = req->r_tid + 1;
1788                 if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
1789                         continue;
1790
1791                 ceph_osdc_get_request(req);
1792                 mutex_unlock(&osdc->request_mutex);
1793                 dout("sync waiting on tid %llu (last is %llu)\n",
1794                      req->r_tid, last_tid);
1795                 wait_for_completion(&req->r_safe_completion);
1796                 mutex_lock(&osdc->request_mutex);
1797                 ceph_osdc_put_request(req);
1798         }
1799         mutex_unlock(&osdc->request_mutex);
1800         dout("sync done (thru tid %llu)\n", last_tid);
1801 }
1802 EXPORT_SYMBOL(ceph_osdc_sync);
1803
1804 /*
1805  * init, shutdown
1806  */
1807 int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
1808 {
1809         int err;
1810
1811         dout("init\n");
1812         osdc->client = client;
1813         osdc->osdmap = NULL;
1814         init_rwsem(&osdc->map_sem);
1815         init_completion(&osdc->map_waiters);
1816         osdc->last_requested_map = 0;
1817         mutex_init(&osdc->request_mutex);
1818         osdc->last_tid = 0;
1819         osdc->osds = RB_ROOT;
1820         INIT_LIST_HEAD(&osdc->osd_lru);
1821         osdc->requests = RB_ROOT;
1822         INIT_LIST_HEAD(&osdc->req_lru);
1823         INIT_LIST_HEAD(&osdc->req_unsent);
1824         INIT_LIST_HEAD(&osdc->req_notarget);
1825         INIT_LIST_HEAD(&osdc->req_linger);
1826         osdc->num_requests = 0;
1827         INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
1828         INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1829         spin_lock_init(&osdc->event_lock);
1830         osdc->event_tree = RB_ROOT;
1831         osdc->event_count = 0;
1832
1833         schedule_delayed_work(&osdc->osds_timeout_work,
1834            round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
1835
1836         err = -ENOMEM;
1837         osdc->req_mempool = mempool_create_kmalloc_pool(10,
1838                                         sizeof(struct ceph_osd_request));
1839         if (!osdc->req_mempool)
1840                 goto out;
1841
1842         err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true,
1843                                 "osd_op");
1844         if (err < 0)
1845                 goto out_mempool;
1846         err = ceph_msgpool_init(&osdc->msgpool_op_reply,
1847                                 OSD_OPREPLY_FRONT_LEN, 10, true,
1848                                 "osd_op_reply");
1849         if (err < 0)
1850                 goto out_msgpool;
1851
1852         osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
1853         if (IS_ERR(osdc->notify_wq)) {
1854                 err = PTR_ERR(osdc->notify_wq);
1855                 osdc->notify_wq = NULL;
1856                 goto out_msgpool;
1857         }
1858         return 0;
1859
1860 out_msgpool:
1861         ceph_msgpool_destroy(&osdc->msgpool_op);
1862 out_mempool:
1863         mempool_destroy(osdc->req_mempool);
1864 out:
1865         return err;
1866 }
1867 EXPORT_SYMBOL(ceph_osdc_init);
1868
1869 void ceph_osdc_stop(struct ceph_osd_client *osdc)
1870 {
1871         flush_workqueue(osdc->notify_wq);
1872         destroy_workqueue(osdc->notify_wq);
1873         cancel_delayed_work_sync(&osdc->timeout_work);
1874         cancel_delayed_work_sync(&osdc->osds_timeout_work);
1875         if (osdc->osdmap) {
1876                 ceph_osdmap_destroy(osdc->osdmap);
1877                 osdc->osdmap = NULL;
1878         }
1879         remove_all_osds(osdc);
1880         mempool_destroy(osdc->req_mempool);
1881         ceph_msgpool_destroy(&osdc->msgpool_op);
1882         ceph_msgpool_destroy(&osdc->msgpool_op_reply);
1883 }
1884 EXPORT_SYMBOL(ceph_osdc_stop);
1885
1886 /*
1887  * Read some contiguous pages.  If we cross a stripe boundary, shorten
1888  * *plen.  Return number of bytes read, or error.
1889  */
1890 int ceph_osdc_readpages(struct ceph_osd_client *osdc,
1891                         struct ceph_vino vino, struct ceph_file_layout *layout,
1892                         u64 off, u64 *plen,
1893                         u32 truncate_seq, u64 truncate_size,
1894                         struct page **pages, int num_pages, int page_align)
1895 {
1896         struct ceph_osd_request *req;
1897         int rc = 0;
1898
1899         dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
1900              vino.snap, off, *plen);
1901         req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
1902                                     CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
1903                                     NULL, 0, truncate_seq, truncate_size, NULL,
1904                                     false, 1, page_align);
1905         if (!req)
1906                 return -ENOMEM;
1907
1908         /* it may be a short read due to an object boundary */
1909         req->r_pages = pages;
1910
1911         dout("readpages  final extent is %llu~%llu (%d pages align %d)\n",
1912              off, *plen, req->r_num_pages, page_align);
1913
1914         rc = ceph_osdc_start_request(osdc, req, false);
1915         if (!rc)
1916                 rc = ceph_osdc_wait_request(osdc, req);
1917
1918         ceph_osdc_put_request(req);
1919         dout("readpages result %d\n", rc);
1920         return rc;
1921 }
1922 EXPORT_SYMBOL(ceph_osdc_readpages);
1923
1924 /*
1925  * do a synchronous write on N pages
1926  */
1927 int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
1928                          struct ceph_file_layout *layout,
1929                          struct ceph_snap_context *snapc,
1930                          u64 off, u64 len,
1931                          u32 truncate_seq, u64 truncate_size,
1932                          struct timespec *mtime,
1933                          struct page **pages, int num_pages,
1934                          int flags, int do_sync, bool nofail)
1935 {
1936         struct ceph_osd_request *req;
1937         int rc = 0;
1938         int page_align = off & ~PAGE_MASK;
1939
1940         BUG_ON(vino.snap != CEPH_NOSNAP);
1941         req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
1942                                     CEPH_OSD_OP_WRITE,
1943                                     flags | CEPH_OSD_FLAG_ONDISK |
1944                                             CEPH_OSD_FLAG_WRITE,
1945                                     snapc, do_sync,
1946                                     truncate_seq, truncate_size, mtime,
1947                                     nofail, 1, page_align);
1948         if (!req)
1949                 return -ENOMEM;
1950
1951         /* it may be a short write due to an object boundary */
1952         req->r_pages = pages;
1953         dout("writepages %llu~%llu (%d pages)\n", off, len,
1954              req->r_num_pages);
1955
1956         rc = ceph_osdc_start_request(osdc, req, nofail);
1957         if (!rc)
1958                 rc = ceph_osdc_wait_request(osdc, req);
1959
1960         ceph_osdc_put_request(req);
1961         if (rc == 0)
1962                 rc = len;
1963         dout("writepages result %d\n", rc);
1964         return rc;
1965 }
1966 EXPORT_SYMBOL(ceph_osdc_writepages);
1967
1968 /*
1969  * handle incoming message
1970  */
1971 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
1972 {
1973         struct ceph_osd *osd = con->private;
1974         struct ceph_osd_client *osdc;
1975         int type = le16_to_cpu(msg->hdr.type);
1976
1977         if (!osd)
1978                 goto out;
1979         osdc = osd->o_osdc;
1980
1981         switch (type) {
1982         case CEPH_MSG_OSD_MAP:
1983                 ceph_osdc_handle_map(osdc, msg);
1984                 break;
1985         case CEPH_MSG_OSD_OPREPLY:
1986                 handle_reply(osdc, msg, con);
1987                 break;
1988         case CEPH_MSG_WATCH_NOTIFY:
1989                 handle_watch_notify(osdc, msg);
1990                 break;
1991
1992         default:
1993                 pr_err("received unknown message type %d %s\n", type,
1994                        ceph_msg_type_name(type));
1995         }
1996 out:
1997         ceph_msg_put(msg);
1998 }
1999
2000 /*
2001  * lookup and return message for incoming reply.  set up reply message
2002  * pages.
2003  */
2004 static struct ceph_msg *get_reply(struct ceph_connection *con,
2005                                   struct ceph_msg_header *hdr,
2006                                   int *skip)
2007 {
2008         struct ceph_osd *osd = con->private;
2009         struct ceph_osd_client *osdc = osd->o_osdc;
2010         struct ceph_msg *m;
2011         struct ceph_osd_request *req;
2012         int front = le32_to_cpu(hdr->front_len);
2013         int data_len = le32_to_cpu(hdr->data_len);
2014         u64 tid;
2015
2016         tid = le64_to_cpu(hdr->tid);
2017         mutex_lock(&osdc->request_mutex);
2018         req = __lookup_request(osdc, tid);
2019         if (!req) {
2020                 *skip = 1;
2021                 m = NULL;
2022                 pr_info("get_reply unknown tid %llu from osd%d\n", tid,
2023                         osd->o_osd);
2024                 goto out;
2025         }
2026
2027         if (req->r_con_filling_msg) {
2028                 dout("get_reply revoking msg %p from old con %p\n",
2029                      req->r_reply, req->r_con_filling_msg);
2030                 ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply);
2031                 ceph_con_put(req->r_con_filling_msg);
2032                 req->r_con_filling_msg = NULL;
2033         }
2034
2035         if (front > req->r_reply->front.iov_len) {
2036                 pr_warning("get_reply front %d > preallocated %d\n",
2037                            front, (int)req->r_reply->front.iov_len);
2038                 m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2039                 if (!m)
2040                         goto out;
2041                 ceph_msg_put(req->r_reply);
2042                 req->r_reply = m;
2043         }
2044         m = ceph_msg_get(req->r_reply);
2045
2046         if (data_len > 0) {
2047                 int want = calc_pages_for(req->r_page_alignment, data_len);
2048
2049                 if (unlikely(req->r_num_pages < want)) {
2050                         pr_warning("tid %lld reply has %d bytes %d pages, we"
2051                                    " had only %d pages ready\n", tid, data_len,
2052                                    want, req->r_num_pages);
2053                         *skip = 1;
2054                         ceph_msg_put(m);
2055                         m = NULL;
2056                         goto out;
2057                 }
2058                 m->pages = req->r_pages;
2059                 m->nr_pages = req->r_num_pages;
2060                 m->page_alignment = req->r_page_alignment;
2061 #ifdef CONFIG_BLOCK
2062                 m->bio = req->r_bio;
2063 #endif
2064         }
2065         *skip = 0;
2066         req->r_con_filling_msg = ceph_con_get(con);
2067         dout("get_reply tid %lld %p\n", tid, m);
2068
2069 out:
2070         mutex_unlock(&osdc->request_mutex);
2071         return m;
2072
2073 }
2074
2075 static struct ceph_msg *alloc_msg(struct ceph_connection *con,
2076                                   struct ceph_msg_header *hdr,
2077                                   int *skip)
2078 {
2079         struct ceph_osd *osd = con->private;
2080         int type = le16_to_cpu(hdr->type);
2081         int front = le32_to_cpu(hdr->front_len);
2082
2083         switch (type) {
2084         case CEPH_MSG_OSD_MAP:
2085         case CEPH_MSG_WATCH_NOTIFY:
2086                 return ceph_msg_new(type, front, GFP_NOFS, false);
2087         case CEPH_MSG_OSD_OPREPLY:
2088                 return get_reply(con, hdr, skip);
2089         default:
2090                 pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
2091                         osd->o_osd);
2092                 *skip = 1;
2093                 return NULL;
2094         }
2095 }
2096
2097 /*
2098  * Wrappers to refcount containing ceph_osd struct
2099  */
2100 static struct ceph_connection *get_osd_con(struct ceph_connection *con)
2101 {
2102         struct ceph_osd *osd = con->private;
2103         if (get_osd(osd))
2104                 return con;
2105         return NULL;
2106 }
2107
2108 static void put_osd_con(struct ceph_connection *con)
2109 {
2110         struct ceph_osd *osd = con->private;
2111         put_osd(osd);
2112 }
2113
2114 /*
2115  * authentication
2116  */
2117 /*
2118  * Note: returned pointer is the address of a structure that's
2119  * managed separately.  Caller must *not* attempt to free it.
2120  */
2121 static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
2122                                         int *proto, int force_new)
2123 {
2124         struct ceph_osd *o = con->private;
2125         struct ceph_osd_client *osdc = o->o_osdc;
2126         struct ceph_auth_client *ac = osdc->client->monc.auth;
2127         struct ceph_auth_handshake *auth = &o->o_auth;
2128
2129         if (force_new && auth->authorizer) {
2130                 if (ac->ops && ac->ops->destroy_authorizer)
2131                         ac->ops->destroy_authorizer(ac, auth->authorizer);
2132                 auth->authorizer = NULL;
2133         }
2134         if (!auth->authorizer && ac->ops && ac->ops->create_authorizer) {
2135                 int ret = ac->ops->create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2136                                                         auth);
2137                 if (ret)
2138                         return ERR_PTR(ret);
2139         }
2140         *proto = ac->protocol;
2141
2142         return auth;
2143 }
2144
2145
2146 static int verify_authorizer_reply(struct ceph_connection *con, int len)
2147 {
2148         struct ceph_osd *o = con->private;
2149         struct ceph_osd_client *osdc = o->o_osdc;
2150         struct ceph_auth_client *ac = osdc->client->monc.auth;
2151
2152         /*
2153          * XXX If ac->ops or ac->ops->verify_authorizer_reply is null,
2154          * XXX which do we do:  succeed or fail?
2155          */
2156         return ac->ops->verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2157 }
2158
2159 static int invalidate_authorizer(struct ceph_connection *con)
2160 {
2161         struct ceph_osd *o = con->private;
2162         struct ceph_osd_client *osdc = o->o_osdc;
2163         struct ceph_auth_client *ac = osdc->client->monc.auth;
2164
2165         if (ac->ops && ac->ops->invalidate_authorizer)
2166                 ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2167
2168         return ceph_monc_validate_auth(&osdc->client->monc);
2169 }
2170
2171 static const struct ceph_connection_operations osd_con_ops = {
2172         .get = get_osd_con,
2173         .put = put_osd_con,
2174         .dispatch = dispatch,
2175         .get_authorizer = get_authorizer,
2176         .verify_authorizer_reply = verify_authorizer_reply,
2177         .invalidate_authorizer = invalidate_authorizer,
2178         .alloc_msg = alloc_msg,
2179         .fault = osd_reset,
2180 };