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