net/ceph/osd_client.c

   1
   2 #include <linux/ceph/ceph_debug.h>
   3
   4 #include <linux/module.h>
   5 #include <linux/err.h>
   6 #include <linux/highmem.h>
   7 #include <linux/mm.h>
   8 #include <linux/pagemap.h>
   9 #include <linux/slab.h>
  10 #include <linux/uaccess.h>
  11 #ifdef CONFIG_BLOCK
  12 #include <linux/bio.h>
  13 #endif
  14
  15 #include <linux/ceph/ceph_features.h>
  16 #include <linux/ceph/libceph.h>
  17 #include <linux/ceph/osd_client.h>
  18 #include <linux/ceph/messenger.h>
  19 #include <linux/ceph/decode.h>
  20 #include <linux/ceph/auth.h>
  21 #include <linux/ceph/pagelist.h>
  22
  23 #define OSD_OPREPLY_FRONT_LEN   512
  24
  25 static struct kmem_cache        *ceph_osd_request_cache;
  26
  27 static const struct ceph_connection_operations osd_con_ops;
  28
  29 /*
  30  * Implement client access to distributed object storage cluster.
  31  *
  32  * All data objects are stored within a cluster/cloud of OSDs, or
  33  * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
  34  * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
  35  * remote daemons serving up and coordinating consistent and safe
  36  * access to storage.
  37  *
  38  * Cluster membership and the mapping of data objects onto storage devices
  39  * are described by the osd map.
  40  *
  41  * We keep track of pending OSD requests (read, write), resubmit
  42  * requests to different OSDs when the cluster topology/data layout
  43  * change, or retry the affected requests when the communications
  44  * channel with an OSD is reset.
  45  */
  46
  47 static void link_request(struct ceph_osd *osd, struct ceph_osd_request *req);
  48 static void unlink_request(struct ceph_osd *osd, struct ceph_osd_request *req);
  49 static void link_linger(struct ceph_osd *osd,
  50                         struct ceph_osd_linger_request *lreq);
  51 static void unlink_linger(struct ceph_osd *osd,
  52                           struct ceph_osd_linger_request *lreq);
  53 static void clear_backoffs(struct ceph_osd *osd);
  54
  55 #if 1
  56 static inline bool rwsem_is_wrlocked(struct rw_semaphore *sem)
  57 {
  58         bool wrlocked = true;
  59
  60         if (unlikely(down_read_trylock(sem))) {
  61                 wrlocked = false;
  62                 up_read(sem);
  63         }
  64
  65         return wrlocked;
  66 }
  67 static inline void verify_osdc_locked(struct ceph_osd_client *osdc)
  68 {
  69         WARN_ON(!rwsem_is_locked(&osdc->lock));
  70 }
  71 static inline void verify_osdc_wrlocked(struct ceph_osd_client *osdc)
  72 {
  73         WARN_ON(!rwsem_is_wrlocked(&osdc->lock));
  74 }
  75 static inline void verify_osd_locked(struct ceph_osd *osd)
  76 {
  77         struct ceph_osd_client *osdc = osd->o_osdc;
  78
  79         WARN_ON(!(mutex_is_locked(&osd->lock) &&
  80                   rwsem_is_locked(&osdc->lock)) &&
  81                 !rwsem_is_wrlocked(&osdc->lock));
  82 }
  83 static inline void verify_lreq_locked(struct ceph_osd_linger_request *lreq)
  84 {
  85         WARN_ON(!mutex_is_locked(&lreq->lock));
  86 }
  87 #else
  88 static inline void verify_osdc_locked(struct ceph_osd_client *osdc) { }
  89 static inline void verify_osdc_wrlocked(struct ceph_osd_client *osdc) { }
  90 static inline void verify_osd_locked(struct ceph_osd *osd) { }
  91 static inline void verify_lreq_locked(struct ceph_osd_linger_request *lreq) { }
  92 #endif
  93
  94 /*
  95  * calculate the mapping of a file extent onto an object, and fill out the
  96  * request accordingly.  shorten extent as necessary if it crosses an
  97  * object boundary.
  98  *
  99  * fill osd op in request message.
 100  */
 101 static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
 102                         u64 *objnum, u64 *objoff, u64 *objlen)
 103 {
 104         u64 orig_len = *plen;
 105         int r;
 106
 107         /* object extent? */
 108         r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
 109                                           objoff, objlen);
 110         if (r < 0)
 111                 return r;
 112         if (*objlen < orig_len) {
 113                 *plen = *objlen;
 114                 dout(" skipping last %llu, final file extent %llu~%llu\n",
 115                      orig_len - *plen, off, *plen);
 116         }
 117
 118         dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
 119
 120         return 0;
 121 }
 122
 123 static void ceph_osd_data_init(struct ceph_osd_data *osd_data)
 124 {
 125         memset(osd_data, 0, sizeof (*osd_data));
 126         osd_data->type = CEPH_OSD_DATA_TYPE_NONE;
 127 }
 128
 129 static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
 130                         struct page **pages, u64 length, u32 alignment,
 131                         bool pages_from_pool, bool own_pages)
 132 {
 133         osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
 134         osd_data->pages = pages;
 135         osd_data->length = length;
 136         osd_data->alignment = alignment;
 137         osd_data->pages_from_pool = pages_from_pool;
 138         osd_data->own_pages = own_pages;
 139 }
 140
 141 static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
 142                         struct ceph_pagelist *pagelist)
 143 {
 144         osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
 145         osd_data->pagelist = pagelist;
 146 }
 147
 148 #ifdef CONFIG_BLOCK
 149 static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
 150                         struct bio *bio, size_t bio_length)
 151 {
 152         osd_data->type = CEPH_OSD_DATA_TYPE_BIO;
 153         osd_data->bio = bio;
 154         osd_data->bio_length = bio_length;
 155 }
 156 #endif /* CONFIG_BLOCK */
 157
 158 #define osd_req_op_data(oreq, whch, typ, fld)                           \
 159 ({                                                                      \
 160         struct ceph_osd_request *__oreq = (oreq);                       \
 161         unsigned int __whch = (whch);                                   \
 162         BUG_ON(__whch >= __oreq->r_num_ops);                            \
 163         &__oreq->r_ops[__whch].typ.fld;                                 \
 164 })
 165
 166 static struct ceph_osd_data *
 167 osd_req_op_raw_data_in(struct ceph_osd_request *osd_req, unsigned int which)
 168 {
 169         BUG_ON(which >= osd_req->r_num_ops);
 170
 171         return &osd_req->r_ops[which].raw_data_in;
 172 }
 173
 174 struct ceph_osd_data *
 175 osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
 176                         unsigned int which)
 177 {
 178         return osd_req_op_data(osd_req, which, extent, osd_data);
 179 }
 180 EXPORT_SYMBOL(osd_req_op_extent_osd_data);
 181
 182 void osd_req_op_raw_data_in_pages(struct ceph_osd_request *osd_req,
 183                         unsigned int which, struct page **pages,
 184                         u64 length, u32 alignment,
 185                         bool pages_from_pool, bool own_pages)
 186 {
 187         struct ceph_osd_data *osd_data;
 188
 189         osd_data = osd_req_op_raw_data_in(osd_req, which);
 190         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 191                                 pages_from_pool, own_pages);
 192 }
 193 EXPORT_SYMBOL(osd_req_op_raw_data_in_pages);
 194
 195 void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
 196                         unsigned int which, struct page **pages,
 197                         u64 length, u32 alignment,
 198                         bool pages_from_pool, bool own_pages)
 199 {
 200         struct ceph_osd_data *osd_data;
 201
 202         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 203         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 204                                 pages_from_pool, own_pages);
 205 }
 206 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);
 207
 208 void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
 209                         unsigned int which, struct ceph_pagelist *pagelist)
 210 {
 211         struct ceph_osd_data *osd_data;
 212
 213         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 214         ceph_osd_data_pagelist_init(osd_data, pagelist);
 215 }
 216 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);
 217
 218 #ifdef CONFIG_BLOCK
 219 void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
 220                         unsigned int which, struct bio *bio, size_t bio_length)
 221 {
 222         struct ceph_osd_data *osd_data;
 223
 224         osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
 225         ceph_osd_data_bio_init(osd_data, bio, bio_length);
 226 }
 227 EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio);
 228 #endif /* CONFIG_BLOCK */
 229
 230 static void osd_req_op_cls_request_info_pagelist(
 231                         struct ceph_osd_request *osd_req,
 232                         unsigned int which, struct ceph_pagelist *pagelist)
 233 {
 234         struct ceph_osd_data *osd_data;
 235
 236         osd_data = osd_req_op_data(osd_req, which, cls, request_info);
 237         ceph_osd_data_pagelist_init(osd_data, pagelist);
 238 }
 239
 240 void osd_req_op_cls_request_data_pagelist(
 241                         struct ceph_osd_request *osd_req,
 242                         unsigned int which, struct ceph_pagelist *pagelist)
 243 {
 244         struct ceph_osd_data *osd_data;
 245
 246         osd_data = osd_req_op_data(osd_req, which, cls, request_data);
 247         ceph_osd_data_pagelist_init(osd_data, pagelist);
 248         osd_req->r_ops[which].cls.indata_len += pagelist->length;
 249         osd_req->r_ops[which].indata_len += pagelist->length;
 250 }
 251 EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);
 252
 253 void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req,
 254                         unsigned int which, struct page **pages, u64 length,
 255                         u32 alignment, bool pages_from_pool, bool own_pages)
 256 {
 257         struct ceph_osd_data *osd_data;
 258
 259         osd_data = osd_req_op_data(osd_req, which, cls, request_data);
 260         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 261                                 pages_from_pool, own_pages);
 262         osd_req->r_ops[which].cls.indata_len += length;
 263         osd_req->r_ops[which].indata_len += length;
 264 }
 265 EXPORT_SYMBOL(osd_req_op_cls_request_data_pages);
 266
 267 void osd_req_op_cls_response_data_pages(struct ceph_osd_request *osd_req,
 268                         unsigned int which, struct page **pages, u64 length,
 269                         u32 alignment, bool pages_from_pool, bool own_pages)
 270 {
 271         struct ceph_osd_data *osd_data;
 272
 273         osd_data = osd_req_op_data(osd_req, which, cls, response_data);
 274         ceph_osd_data_pages_init(osd_data, pages, length, alignment,
 275                                 pages_from_pool, own_pages);
 276 }
 277 EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);
 278
 279 static u64 ceph_osd_data_length(struct ceph_osd_data *osd_data)
 280 {
 281         switch (osd_data->type) {
 282         case CEPH_OSD_DATA_TYPE_NONE:
 283                 return 0;
 284         case CEPH_OSD_DATA_TYPE_PAGES:
 285                 return osd_data->length;
 286         case CEPH_OSD_DATA_TYPE_PAGELIST:
 287                 return (u64)osd_data->pagelist->length;
 288 #ifdef CONFIG_BLOCK
 289         case CEPH_OSD_DATA_TYPE_BIO:
 290                 return (u64)osd_data->bio_length;
 291 #endif /* CONFIG_BLOCK */
 292         default:
 293                 WARN(true, "unrecognized data type %d\n", (int)osd_data->type);
 294                 return 0;
 295         }
 296 }
 297
 298 static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
 299 {
 300         if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
 301                 int num_pages;
 302
 303                 num_pages = calc_pages_for((u64)osd_data->alignment,
 304                                                 (u64)osd_data->length);
 305                 ceph_release_page_vector(osd_data->pages, num_pages);
 306         }
 307         ceph_osd_data_init(osd_data);
 308 }
 309
 310 static void osd_req_op_data_release(struct ceph_osd_request *osd_req,
 311                         unsigned int which)
 312 {
 313         struct ceph_osd_req_op *op;
 314
 315         BUG_ON(which >= osd_req->r_num_ops);
 316         op = &osd_req->r_ops[which];
 317
 318         switch (op->op) {
 319         case CEPH_OSD_OP_READ:
 320         case CEPH_OSD_OP_WRITE:
 321         case CEPH_OSD_OP_WRITEFULL:
 322                 ceph_osd_data_release(&op->extent.osd_data);
 323                 break;
 324         case CEPH_OSD_OP_CALL:
 325                 ceph_osd_data_release(&op->cls.request_info);
 326                 ceph_osd_data_release(&op->cls.request_data);
 327                 ceph_osd_data_release(&op->cls.response_data);
 328                 break;
 329         case CEPH_OSD_OP_SETXATTR:
 330         case CEPH_OSD_OP_CMPXATTR:
 331                 ceph_osd_data_release(&op->xattr.osd_data);
 332                 break;
 333         case CEPH_OSD_OP_STAT:
 334                 ceph_osd_data_release(&op->raw_data_in);
 335                 break;
 336         case CEPH_OSD_OP_NOTIFY_ACK:
 337                 ceph_osd_data_release(&op->notify_ack.request_data);
 338                 break;
 339         case CEPH_OSD_OP_NOTIFY:
 340                 ceph_osd_data_release(&op->notify.request_data);
 341                 ceph_osd_data_release(&op->notify.response_data);
 342                 break;
 343         case CEPH_OSD_OP_LIST_WATCHERS:
 344                 ceph_osd_data_release(&op->list_watchers.response_data);
 345                 break;
 346         default:
 347                 break;
 348         }
 349 }
 350
 351 /*
 352  * Assumes @t is zero-initialized.
 353  */
 354 static void target_init(struct ceph_osd_request_target *t)
 355 {
 356         ceph_oid_init(&t->base_oid);
 357         ceph_oloc_init(&t->base_oloc);
 358         ceph_oid_init(&t->target_oid);
 359         ceph_oloc_init(&t->target_oloc);
 360
 361         ceph_osds_init(&t->acting);
 362         ceph_osds_init(&t->up);
 363         t->size = -1;
 364         t->min_size = -1;
 365
 366         t->osd = CEPH_HOMELESS_OSD;
 367 }
 368
 369 static void target_copy(struct ceph_osd_request_target *dest,
 370                         const struct ceph_osd_request_target *src)
 371 {
 372         ceph_oid_copy(&dest->base_oid, &src->base_oid);
 373         ceph_oloc_copy(&dest->base_oloc, &src->base_oloc);
 374         ceph_oid_copy(&dest->target_oid, &src->target_oid);
 375         ceph_oloc_copy(&dest->target_oloc, &src->target_oloc);
 376
 377         dest->pgid = src->pgid; /* struct */
 378         dest->spgid = src->spgid; /* struct */
 379         dest->pg_num = src->pg_num;
 380         dest->pg_num_mask = src->pg_num_mask;
 381         ceph_osds_copy(&dest->acting, &src->acting);
 382         ceph_osds_copy(&dest->up, &src->up);
 383         dest->size = src->size;
 384         dest->min_size = src->min_size;
 385         dest->sort_bitwise = src->sort_bitwise;
 386
 387         dest->flags = src->flags;
 388         dest->paused = src->paused;
 389
 390         dest->epoch = src->epoch;
 391         dest->last_force_resend = src->last_force_resend;
 392
 393         dest->osd = src->osd;
 394 }
 395
 396 static void target_destroy(struct ceph_osd_request_target *t)
 397 {
 398         ceph_oid_destroy(&t->base_oid);
 399         ceph_oloc_destroy(&t->base_oloc);
 400         ceph_oid_destroy(&t->target_oid);
 401         ceph_oloc_destroy(&t->target_oloc);
 402 }
 403
 404 /*
 405  * requests
 406  */
 407 static void request_release_checks(struct ceph_osd_request *req)
 408 {
 409         WARN_ON(!RB_EMPTY_NODE(&req->r_node));
 410         WARN_ON(!RB_EMPTY_NODE(&req->r_mc_node));
 411         WARN_ON(!list_empty(&req->r_unsafe_item));
 412         WARN_ON(req->r_osd);
 413 }
 414
 415 static void ceph_osdc_release_request(struct kref *kref)
 416 {
 417         struct ceph_osd_request *req = container_of(kref,
 418                                             struct ceph_osd_request, r_kref);
 419         unsigned int which;
 420
 421         dout("%s %p (r_request %p r_reply %p)\n", __func__, req,
 422              req->r_request, req->r_reply);
 423         request_release_checks(req);
 424
 425         if (req->r_request)
 426                 ceph_msg_put(req->r_request);
 427         if (req->r_reply)
 428                 ceph_msg_put(req->r_reply);
 429
 430         for (which = 0; which < req->r_num_ops; which++)
 431                 osd_req_op_data_release(req, which);
 432
 433         target_destroy(&req->r_t);
 434         ceph_put_snap_context(req->r_snapc);
 435
 436         if (req->r_mempool)
 437                 mempool_free(req, req->r_osdc->req_mempool);
 438         else if (req->r_num_ops <= CEPH_OSD_SLAB_OPS)
 439                 kmem_cache_free(ceph_osd_request_cache, req);
 440         else
 441                 kfree(req);
 442 }
 443
 444 void ceph_osdc_get_request(struct ceph_osd_request *req)
 445 {
 446         dout("%s %p (was %d)\n", __func__, req,
 447              kref_read(&req->r_kref));
 448         kref_get(&req->r_kref);
 449 }
 450 EXPORT_SYMBOL(ceph_osdc_get_request);
 451
 452 void ceph_osdc_put_request(struct ceph_osd_request *req)
 453 {
 454         if (req) {
 455                 dout("%s %p (was %d)\n", __func__, req,
 456                      kref_read(&req->r_kref));
 457                 kref_put(&req->r_kref, ceph_osdc_release_request);
 458         }
 459 }
 460 EXPORT_SYMBOL(ceph_osdc_put_request);
 461
 462 static void request_init(struct ceph_osd_request *req)
 463 {
 464         /* req only, each op is zeroed in _osd_req_op_init() */
 465         memset(req, 0, sizeof(*req));
 466
 467         kref_init(&req->r_kref);
 468         init_completion(&req->r_completion);
 469         RB_CLEAR_NODE(&req->r_node);
 470         RB_CLEAR_NODE(&req->r_mc_node);
 471         INIT_LIST_HEAD(&req->r_unsafe_item);
 472
 473         target_init(&req->r_t);
 474 }
 475
 476 /*
 477  * This is ugly, but it allows us to reuse linger registration and ping
 478  * requests, keeping the structure of the code around send_linger{_ping}()
 479  * reasonable.  Setting up a min_nr=2 mempool for each linger request
 480  * and dealing with copying ops (this blasts req only, watch op remains
 481  * intact) isn't any better.
 482  */
 483 static void request_reinit(struct ceph_osd_request *req)
 484 {
 485         struct ceph_osd_client *osdc = req->r_osdc;
 486         bool mempool = req->r_mempool;
 487         unsigned int num_ops = req->r_num_ops;
 488         u64 snapid = req->r_snapid;
 489         struct ceph_snap_context *snapc = req->r_snapc;
 490         bool linger = req->r_linger;
 491         struct ceph_msg *request_msg = req->r_request;
 492         struct ceph_msg *reply_msg = req->r_reply;
 493
 494         dout("%s req %p\n", __func__, req);
 495         WARN_ON(kref_read(&req->r_kref) != 1);
 496         request_release_checks(req);
 497
 498         WARN_ON(kref_read(&request_msg->kref) != 1);
 499         WARN_ON(kref_read(&reply_msg->kref) != 1);
 500         target_destroy(&req->r_t);
 501
 502         request_init(req);
 503         req->r_osdc = osdc;
 504         req->r_mempool = mempool;
 505         req->r_num_ops = num_ops;
 506         req->r_snapid = snapid;
 507         req->r_snapc = snapc;
 508         req->r_linger = linger;
 509         req->r_request = request_msg;
 510         req->r_reply = reply_msg;
 511 }
 512
 513 struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
 514                                                struct ceph_snap_context *snapc,
 515                                                unsigned int num_ops,
 516                                                bool use_mempool,
 517                                                gfp_t gfp_flags)
 518 {
 519         struct ceph_osd_request *req;
 520
 521         if (use_mempool) {
 522                 BUG_ON(num_ops > CEPH_OSD_SLAB_OPS);
 523                 req = mempool_alloc(osdc->req_mempool, gfp_flags);
 524         } else if (num_ops <= CEPH_OSD_SLAB_OPS) {
 525                 req = kmem_cache_alloc(ceph_osd_request_cache, gfp_flags);
 526         } else {
 527                 BUG_ON(num_ops > CEPH_OSD_MAX_OPS);
 528                 req = kmalloc(sizeof(*req) + num_ops * sizeof(req->r_ops[0]),
 529                               gfp_flags);
 530         }
 531         if (unlikely(!req))
 532                 return NULL;
 533
 534         request_init(req);
 535         req->r_osdc = osdc;
 536         req->r_mempool = use_mempool;
 537         req->r_num_ops = num_ops;
 538         req->r_snapid = CEPH_NOSNAP;
 539         req->r_snapc = ceph_get_snap_context(snapc);
 540
 541         dout("%s req %p\n", __func__, req);
 542         return req;
 543 }
 544 EXPORT_SYMBOL(ceph_osdc_alloc_request);
 545
 546 static int ceph_oloc_encoding_size(const struct ceph_object_locator *oloc)
 547 {
 548         return 8 + 4 + 4 + 4 + (oloc->pool_ns ? oloc->pool_ns->len : 0);
 549 }
 550
 551 int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp)
 552 {
 553         struct ceph_osd_client *osdc = req->r_osdc;
 554         struct ceph_msg *msg;
 555         int msg_size;
 556
 557         WARN_ON(ceph_oid_empty(&req->r_base_oid));
 558         WARN_ON(ceph_oloc_empty(&req->r_base_oloc));
 559
 560         /* create request message */
 561         msg_size = CEPH_ENCODING_START_BLK_LEN +
 562                         CEPH_PGID_ENCODING_LEN + 1; /* spgid */
 563         msg_size += 4 + 4 + 4; /* hash, osdmap_epoch, flags */
 564         msg_size += CEPH_ENCODING_START_BLK_LEN +
 565                         sizeof(struct ceph_osd_reqid); /* reqid */
 566         msg_size += sizeof(struct ceph_blkin_trace_info); /* trace */
 567         msg_size += 4 + sizeof(struct ceph_timespec); /* client_inc, mtime */
 568         msg_size += CEPH_ENCODING_START_BLK_LEN +
 569                         ceph_oloc_encoding_size(&req->r_base_oloc); /* oloc */
 570         msg_size += 4 + req->r_base_oid.name_len; /* oid */
 571         msg_size += 2 + req->r_num_ops * sizeof(struct ceph_osd_op);
 572         msg_size += 8; /* snapid */
 573         msg_size += 8; /* snap_seq */
 574         msg_size += 4 + 8 * (req->r_snapc ? req->r_snapc->num_snaps : 0);
 575         msg_size += 4 + 8; /* retry_attempt, features */
 576
 577         if (req->r_mempool)
 578                 msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
 579         else
 580                 msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp, true);
 581         if (!msg)
 582                 return -ENOMEM;
 583
 584         memset(msg->front.iov_base, 0, msg->front.iov_len);
 585         req->r_request = msg;
 586
 587         /* create reply message */
 588         msg_size = OSD_OPREPLY_FRONT_LEN;
 589         msg_size += req->r_base_oid.name_len;
 590         msg_size += req->r_num_ops * sizeof(struct ceph_osd_op);
 591
 592         if (req->r_mempool)
 593                 msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
 594         else
 595                 msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, msg_size, gfp, true);
 596         if (!msg)
 597                 return -ENOMEM;
 598
 599         req->r_reply = msg;
 600
 601         return 0;
 602 }
 603 EXPORT_SYMBOL(ceph_osdc_alloc_messages);
 604
 605 static bool osd_req_opcode_valid(u16 opcode)
 606 {
 607         switch (opcode) {
 608 #define GENERATE_CASE(op, opcode, str)  case CEPH_OSD_OP_##op: return true;
 609 __CEPH_FORALL_OSD_OPS(GENERATE_CASE)
 610 #undef GENERATE_CASE
 611         default:
 612                 return false;
 613         }
 614 }
 615
 616 /*
 617  * This is an osd op init function for opcodes that have no data or
 618  * other information associated with them.  It also serves as a
 619  * common init routine for all the other init functions, below.
 620  */
 621 static struct ceph_osd_req_op *
 622 _osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
 623                  u16 opcode, u32 flags)
 624 {
 625         struct ceph_osd_req_op *op;
 626
 627         BUG_ON(which >= osd_req->r_num_ops);
 628         BUG_ON(!osd_req_opcode_valid(opcode));
 629
 630         op = &osd_req->r_ops[which];
 631         memset(op, 0, sizeof (*op));
 632         op->op = opcode;
 633         op->flags = flags;
 634
 635         return op;
 636 }
 637
 638 void osd_req_op_init(struct ceph_osd_request *osd_req,
 639                      unsigned int which, u16 opcode, u32 flags)
 640 {
 641         (void)_osd_req_op_init(osd_req, which, opcode, flags);
 642 }
 643 EXPORT_SYMBOL(osd_req_op_init);
 644
 645 void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
 646                                 unsigned int which, u16 opcode,
 647                                 u64 offset, u64 length,
 648                                 u64 truncate_size, u32 truncate_seq)
 649 {
 650         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
 651                                                       opcode, 0);
 652         size_t payload_len = 0;
 653
 654         BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
 655                opcode != CEPH_OSD_OP_WRITEFULL && opcode != CEPH_OSD_OP_ZERO &&
 656                opcode != CEPH_OSD_OP_TRUNCATE);
 657
 658         op->extent.offset = offset;
 659         op->extent.length = length;
 660         op->extent.truncate_size = truncate_size;
 661         op->extent.truncate_seq = truncate_seq;
 662         if (opcode == CEPH_OSD_OP_WRITE || opcode == CEPH_OSD_OP_WRITEFULL)
 663                 payload_len += length;
 664
 665         op->indata_len = payload_len;
 666 }
 667 EXPORT_SYMBOL(osd_req_op_extent_init);
 668
 669 void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
 670                                 unsigned int which, u64 length)
 671 {
 672         struct ceph_osd_req_op *op;
 673         u64 previous;
 674
 675         BUG_ON(which >= osd_req->r_num_ops);
 676         op = &osd_req->r_ops[which];
 677         previous = op->extent.length;
 678
 679         if (length == previous)
 680                 return;         /* Nothing to do */
 681         BUG_ON(length > previous);
 682
 683         op->extent.length = length;
 684         if (op->op == CEPH_OSD_OP_WRITE || op->op == CEPH_OSD_OP_WRITEFULL)
 685                 op->indata_len -= previous - length;
 686 }
 687 EXPORT_SYMBOL(osd_req_op_extent_update);
 688
 689 void osd_req_op_extent_dup_last(struct ceph_osd_request *osd_req,
 690                                 unsigned int which, u64 offset_inc)
 691 {
 692         struct ceph_osd_req_op *op, *prev_op;
 693
 694         BUG_ON(which + 1 >= osd_req->r_num_ops);
 695
 696         prev_op = &osd_req->r_ops[which];
 697         op = _osd_req_op_init(osd_req, which + 1, prev_op->op, prev_op->flags);
 698         /* dup previous one */
 699         op->indata_len = prev_op->indata_len;
 700         op->outdata_len = prev_op->outdata_len;
 701         op->extent = prev_op->extent;
 702         /* adjust offset */
 703         op->extent.offset += offset_inc;
 704         op->extent.length -= offset_inc;
 705
 706         if (op->op == CEPH_OSD_OP_WRITE || op->op == CEPH_OSD_OP_WRITEFULL)
 707                 op->indata_len -= offset_inc;
 708 }
 709 EXPORT_SYMBOL(osd_req_op_extent_dup_last);
 710
 711 void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
 712                         u16 opcode, const char *class, const char *method)
 713 {
 714         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
 715                                                       opcode, 0);
 716         struct ceph_pagelist *pagelist;
 717         size_t payload_len = 0;
 718         size_t size;
 719
 720         BUG_ON(opcode != CEPH_OSD_OP_CALL);
 721
 722         pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
 723         BUG_ON(!pagelist);
 724         ceph_pagelist_init(pagelist);
 725
 726         op->cls.class_name = class;
 727         size = strlen(class);
 728         BUG_ON(size > (size_t) U8_MAX);
 729         op->cls.class_len = size;
 730         ceph_pagelist_append(pagelist, class, size);
 731         payload_len += size;
 732
 733         op->cls.method_name = method;
 734         size = strlen(method);
 735         BUG_ON(size > (size_t) U8_MAX);
 736         op->cls.method_len = size;
 737         ceph_pagelist_append(pagelist, method, size);
 738         payload_len += size;
 739
 740         osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
 741
 742         op->indata_len = payload_len;
 743 }
 744 EXPORT_SYMBOL(osd_req_op_cls_init);
 745
 746 int osd_req_op_xattr_init(struct ceph_osd_request *osd_req, unsigned int which,
 747                           u16 opcode, const char *name, const void *value,
 748                           size_t size, u8 cmp_op, u8 cmp_mode)
 749 {
 750         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
 751                                                       opcode, 0);
 752         struct ceph_pagelist *pagelist;
 753         size_t payload_len;
 754
 755         BUG_ON(opcode != CEPH_OSD_OP_SETXATTR && opcode != CEPH_OSD_OP_CMPXATTR);
 756
 757         pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS);
 758         if (!pagelist)
 759                 return -ENOMEM;
 760
 761         ceph_pagelist_init(pagelist);
 762
 763         payload_len = strlen(name);
 764         op->xattr.name_len = payload_len;
 765         ceph_pagelist_append(pagelist, name, payload_len);
 766
 767         op->xattr.value_len = size;
 768         ceph_pagelist_append(pagelist, value, size);
 769         payload_len += size;
 770
 771         op->xattr.cmp_op = cmp_op;
 772         op->xattr.cmp_mode = cmp_mode;
 773
 774         ceph_osd_data_pagelist_init(&op->xattr.osd_data, pagelist);
 775         op->indata_len = payload_len;
 776         return 0;
 777 }
 778 EXPORT_SYMBOL(osd_req_op_xattr_init);
 779
 780 /*
 781  * @watch_opcode: CEPH_OSD_WATCH_OP_*
 782  */
 783 static void osd_req_op_watch_init(struct ceph_osd_request *req, int which,
 784                                   u64 cookie, u8 watch_opcode)
 785 {
 786         struct ceph_osd_req_op *op;
 787
 788         op = _osd_req_op_init(req, which, CEPH_OSD_OP_WATCH, 0);
 789         op->watch.cookie = cookie;
 790         op->watch.op = watch_opcode;
 791         op->watch.gen = 0;
 792 }
 793
 794 void osd_req_op_alloc_hint_init(struct ceph_osd_request *osd_req,
 795                                 unsigned int which,
 796                                 u64 expected_object_size,
 797                                 u64 expected_write_size)
 798 {
 799         struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
 800                                                       CEPH_OSD_OP_SETALLOCHINT,
 801                                                       0);
 802
 803         op->alloc_hint.expected_object_size = expected_object_size;
 804         op->alloc_hint.expected_write_size = expected_write_size;
 805
 806         /*
 807          * CEPH_OSD_OP_SETALLOCHINT op is advisory and therefore deemed
 808          * not worth a feature bit.  Set FAILOK per-op flag to make
 809          * sure older osds don't trip over an unsupported opcode.
 810          */
 811         op->flags |= CEPH_OSD_OP_FLAG_FAILOK;
 812 }
 813 EXPORT_SYMBOL(osd_req_op_alloc_hint_init);
 814
 815 static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
 816                                 struct ceph_osd_data *osd_data)
 817 {
 818         u64 length = ceph_osd_data_length(osd_data);
 819
 820         if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
 821                 BUG_ON(length > (u64) SIZE_MAX);
 822                 if (length)
 823                         ceph_msg_data_add_pages(msg, osd_data->pages,
 824                                         length, osd_data->alignment);
 825         } else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
 826                 BUG_ON(!length);
 827                 ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
 828 #ifdef CONFIG_BLOCK
 829         } else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
 830                 ceph_msg_data_add_bio(msg, osd_data->bio, length);
 831 #endif
 832         } else {
 833                 BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
 834         }
 835 }
 836
 837 static u32 osd_req_encode_op(struct ceph_osd_op *dst,
 838                              const struct ceph_osd_req_op *src)
 839 {
 840         if (WARN_ON(!osd_req_opcode_valid(src->op))) {
 841                 pr_err("unrecognized osd opcode %d\n", src->op);
 842
 843                 return 0;
 844         }
 845
 846         switch (src->op) {
 847         case CEPH_OSD_OP_STAT:
 848                 break;
 849         case CEPH_OSD_OP_READ:
 850         case CEPH_OSD_OP_WRITE:
 851         case CEPH_OSD_OP_WRITEFULL:
 852         case CEPH_OSD_OP_ZERO:
 853         case CEPH_OSD_OP_TRUNCATE:
 854                 dst->extent.offset = cpu_to_le64(src->extent.offset);
 855                 dst->extent.length = cpu_to_le64(src->extent.length);
 856                 dst->extent.truncate_size =
 857                         cpu_to_le64(src->extent.truncate_size);
 858                 dst->extent.truncate_seq =
 859                         cpu_to_le32(src->extent.truncate_seq);
 860                 break;
 861         case CEPH_OSD_OP_CALL:
 862                 dst->cls.class_len = src->cls.class_len;
 863                 dst->cls.method_len = src->cls.method_len;
 864                 dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
 865                 break;
 866         case CEPH_OSD_OP_STARTSYNC:
 867                 break;
 868         case CEPH_OSD_OP_WATCH:
 869                 dst->watch.cookie = cpu_to_le64(src->watch.cookie);
 870                 dst->watch.ver = cpu_to_le64(0);
 871                 dst->watch.op = src->watch.op;
 872                 dst->watch.gen = cpu_to_le32(src->watch.gen);
 873                 break;
 874         case CEPH_OSD_OP_NOTIFY_ACK:
 875                 break;
 876         case CEPH_OSD_OP_NOTIFY:
 877                 dst->notify.cookie = cpu_to_le64(src->notify.cookie);
 878                 break;
 879         case CEPH_OSD_OP_LIST_WATCHERS:
 880                 break;
 881         case CEPH_OSD_OP_SETALLOCHINT:
 882                 dst->alloc_hint.expected_object_size =
 883                     cpu_to_le64(src->alloc_hint.expected_object_size);
 884                 dst->alloc_hint.expected_write_size =
 885                     cpu_to_le64(src->alloc_hint.expected_write_size);
 886                 break;
 887         case CEPH_OSD_OP_SETXATTR:
 888         case CEPH_OSD_OP_CMPXATTR:
 889                 dst->xattr.name_len = cpu_to_le32(src->xattr.name_len);
 890                 dst->xattr.value_len = cpu_to_le32(src->xattr.value_len);
 891                 dst->xattr.cmp_op = src->xattr.cmp_op;
 892                 dst->xattr.cmp_mode = src->xattr.cmp_mode;
 893                 break;
 894         case CEPH_OSD_OP_CREATE:
 895         case CEPH_OSD_OP_DELETE:
 896                 break;
 897         default:
 898                 pr_err("unsupported osd opcode %s\n",
 899                         ceph_osd_op_name(src->op));
 900                 WARN_ON(1);
 901
 902                 return 0;
 903         }
 904
 905         dst->op = cpu_to_le16(src->op);
 906         dst->flags = cpu_to_le32(src->flags);
 907         dst->payload_len = cpu_to_le32(src->indata_len);
 908
 909         return src->indata_len;
 910 }
 911
 912 /*
 913  * build new request AND message, calculate layout, and adjust file
 914  * extent as needed.
 915  *
 916  * if the file was recently truncated, we include information about its
 917  * old and new size so that the object can be updated appropriately.  (we
 918  * avoid synchronously deleting truncated objects because it's slow.)
 919  *
 920  * if @do_sync, include a 'startsync' command so that the osd will flush
 921  * data quickly.
 922  */
 923 struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 924                                                struct ceph_file_layout *layout,
 925                                                struct ceph_vino vino,
 926                                                u64 off, u64 *plen,
 927                                                unsigned int which, int num_ops,
 928                                                int opcode, int flags,
 929                                                struct ceph_snap_context *snapc,
 930                                                u32 truncate_seq,
 931                                                u64 truncate_size,
 932                                                bool use_mempool)
 933 {
 934         struct ceph_osd_request *req;
 935         u64 objnum = 0;
 936         u64 objoff = 0;
 937         u64 objlen = 0;
 938         int r;
 939
 940         BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
 941                opcode != CEPH_OSD_OP_ZERO && opcode != CEPH_OSD_OP_TRUNCATE &&
 942                opcode != CEPH_OSD_OP_CREATE && opcode != CEPH_OSD_OP_DELETE);
 943
 944         req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
 945                                         GFP_NOFS);
 946         if (!req) {
 947                 r = -ENOMEM;
 948                 goto fail;
 949         }
 950
 951         /* calculate max write size */
 952         r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
 953         if (r)
 954                 goto fail;
 955
 956         if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) {
 957                 osd_req_op_init(req, which, opcode, 0);
 958         } else {
 959                 u32 object_size = layout->object_size;
 960                 u32 object_base = off - objoff;
 961                 if (!(truncate_seq == 1 && truncate_size == -1ULL)) {
 962                         if (truncate_size <= object_base) {
 963                                 truncate_size = 0;
 964                         } else {
 965                                 truncate_size -= object_base;
 966                                 if (truncate_size > object_size)
 967                                         truncate_size = object_size;
 968                         }
 969                 }
 970                 osd_req_op_extent_init(req, which, opcode, objoff, objlen,
 971                                        truncate_size, truncate_seq);
 972         }
 973
 974         req->r_abort_on_full = true;
 975         req->r_flags = flags;
 976         req->r_base_oloc.pool = layout->pool_id;
 977         req->r_base_oloc.pool_ns = ceph_try_get_string(layout->pool_ns);
 978         ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum);
 979
 980         req->r_snapid = vino.snap;
 981         if (flags & CEPH_OSD_FLAG_WRITE)
 982                 req->r_data_offset = off;
 983
 984         r = ceph_osdc_alloc_messages(req, GFP_NOFS);
 985         if (r)
 986                 goto fail;
 987
 988         return req;
 989
 990 fail:
 991         ceph_osdc_put_request(req);
 992         return ERR_PTR(r);
 993 }
 994 EXPORT_SYMBOL(ceph_osdc_new_request);
 995
 996 /*
 997  * We keep osd requests in an rbtree, sorted by ->r_tid.
 998  */
 999 DEFINE_RB_FUNCS(request, struct ceph_osd_request, r_tid, r_node)
1000 DEFINE_RB_FUNCS(request_mc, struct ceph_osd_request, r_tid, r_mc_node)
1001
1002 static bool osd_homeless(struct ceph_osd *osd)
1003 {
1004         return osd->o_osd == CEPH_HOMELESS_OSD;
1005 }
1006
1007 static bool osd_registered(struct ceph_osd *osd)
1008 {
1009         verify_osdc_locked(osd->o_osdc);
1010
1011         return !RB_EMPTY_NODE(&osd->o_node);
1012 }
1013
1014 /*
1015  * Assumes @osd is zero-initialized.
1016  */
1017 static void osd_init(struct ceph_osd *osd)
1018 {
1019         refcount_set(&osd->o_ref, 1);
1020         RB_CLEAR_NODE(&osd->o_node);
1021         osd->o_requests = RB_ROOT;
1022         osd->o_linger_requests = RB_ROOT;
1023         osd->o_backoff_mappings = RB_ROOT;
1024         osd->o_backoffs_by_id = RB_ROOT;
1025         INIT_LIST_HEAD(&osd->o_osd_lru);
1026         INIT_LIST_HEAD(&osd->o_keepalive_item);
1027         osd->o_incarnation = 1;
1028         mutex_init(&osd->lock);
1029 }
1030
1031 static void osd_cleanup(struct ceph_osd *osd)
1032 {
1033         WARN_ON(!RB_EMPTY_NODE(&osd->o_node));
1034         WARN_ON(!RB_EMPTY_ROOT(&osd->o_requests));
1035         WARN_ON(!RB_EMPTY_ROOT(&osd->o_linger_requests));
1036         WARN_ON(!RB_EMPTY_ROOT(&osd->o_backoff_mappings));
1037         WARN_ON(!RB_EMPTY_ROOT(&osd->o_backoffs_by_id));
1038         WARN_ON(!list_empty(&osd->o_osd_lru));
1039         WARN_ON(!list_empty(&osd->o_keepalive_item));
1040
1041         if (osd->o_auth.authorizer) {
1042                 WARN_ON(osd_homeless(osd));
1043                 ceph_auth_destroy_authorizer(osd->o_auth.authorizer);
1044         }
1045 }
1046
1047 /*
1048  * Track open sessions with osds.
1049  */
1050 static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
1051 {
1052         struct ceph_osd *osd;
1053
1054         WARN_ON(onum == CEPH_HOMELESS_OSD);
1055
1056         osd = kzalloc(sizeof(*osd), GFP_NOIO | __GFP_NOFAIL);
1057         osd_init(osd);
1058         osd->o_osdc = osdc;
1059         osd->o_osd = onum;
1060
1061         ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
1062
1063         return osd;
1064 }
1065
1066 static struct ceph_osd *get_osd(struct ceph_osd *osd)
1067 {
1068         if (refcount_inc_not_zero(&osd->o_ref)) {
1069                 dout("get_osd %p %d -> %d\n", osd, refcount_read(&osd->o_ref)-1,
1070                      refcount_read(&osd->o_ref));
1071                 return osd;
1072         } else {
1073                 dout("get_osd %p FAIL\n", osd);
1074                 return NULL;
1075         }
1076 }
1077
1078 static void put_osd(struct ceph_osd *osd)
1079 {
1080         dout("put_osd %p %d -> %d\n", osd, refcount_read(&osd->o_ref),
1081              refcount_read(&osd->o_ref) - 1);
1082         if (refcount_dec_and_test(&osd->o_ref)) {
1083                 osd_cleanup(osd);
1084                 kfree(osd);
1085         }
1086 }
1087
1088 DEFINE_RB_FUNCS(osd, struct ceph_osd, o_osd, o_node)
1089
1090 static void __move_osd_to_lru(struct ceph_osd *osd)
1091 {
1092         struct ceph_osd_client *osdc = osd->o_osdc;
1093
1094         dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
1095         BUG_ON(!list_empty(&osd->o_osd_lru));
1096
1097         spin_lock(&osdc->osd_lru_lock);
1098         list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
1099         spin_unlock(&osdc->osd_lru_lock);
1100
1101         osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl;
1102 }
1103
1104 static void maybe_move_osd_to_lru(struct ceph_osd *osd)
1105 {
1106         if (RB_EMPTY_ROOT(&osd->o_requests) &&
1107             RB_EMPTY_ROOT(&osd->o_linger_requests))
1108                 __move_osd_to_lru(osd);
1109 }
1110
1111 static void __remove_osd_from_lru(struct ceph_osd *osd)
1112 {
1113         struct ceph_osd_client *osdc = osd->o_osdc;
1114
1115         dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
1116
1117         spin_lock(&osdc->osd_lru_lock);
1118         if (!list_empty(&osd->o_osd_lru))
1119                 list_del_init(&osd->o_osd_lru);
1120         spin_unlock(&osdc->osd_lru_lock);
1121 }
1122
1123 /*
1124  * Close the connection and assign any leftover requests to the
1125  * homeless session.
1126  */
1127 static void close_osd(struct ceph_osd *osd)
1128 {
1129         struct ceph_osd_client *osdc = osd->o_osdc;
1130         struct rb_node *n;
1131
1132         verify_osdc_wrlocked(osdc);
1133         dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
1134
1135         ceph_con_close(&osd->o_con);
1136
1137         for (n = rb_first(&osd->o_requests); n; ) {
1138                 struct ceph_osd_request *req =
1139                     rb_entry(n, struct ceph_osd_request, r_node);
1140
1141                 n = rb_next(n); /* unlink_request() */
1142
1143                 dout(" reassigning req %p tid %llu\n", req, req->r_tid);
1144                 unlink_request(osd, req);
1145                 link_request(&osdc->homeless_osd, req);
1146         }
1147         for (n = rb_first(&osd->o_linger_requests); n; ) {
1148                 struct ceph_osd_linger_request *lreq =
1149                     rb_entry(n, struct ceph_osd_linger_request, node);
1150
1151                 n = rb_next(n); /* unlink_linger() */
1152
1153                 dout(" reassigning lreq %p linger_id %llu\n", lreq,
1154                      lreq->linger_id);
1155                 unlink_linger(osd, lreq);
1156                 link_linger(&osdc->homeless_osd, lreq);
1157         }
1158         clear_backoffs(osd);
1159
1160         __remove_osd_from_lru(osd);
1161         erase_osd(&osdc->osds, osd);
1162         put_osd(osd);
1163 }
1164
1165 /*
1166  * reset osd connect
1167  */
1168 static int reopen_osd(struct ceph_osd *osd)
1169 {
1170         struct ceph_entity_addr *peer_addr;
1171
1172         dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
1173
1174         if (RB_EMPTY_ROOT(&osd->o_requests) &&
1175             RB_EMPTY_ROOT(&osd->o_linger_requests)) {
1176                 close_osd(osd);
1177                 return -ENODEV;
1178         }
1179
1180         peer_addr = &osd->o_osdc->osdmap->osd_addr[osd->o_osd];
1181         if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) &&
1182                         !ceph_con_opened(&osd->o_con)) {
1183                 struct rb_node *n;
1184
1185                 dout("osd addr hasn't changed and connection never opened, "
1186                      "letting msgr retry\n");
1187                 /* touch each r_stamp for handle_timeout()'s benfit */
1188                 for (n = rb_first(&osd->o_requests); n; n = rb_next(n)) {
1189                         struct ceph_osd_request *req =
1190                             rb_entry(n, struct ceph_osd_request, r_node);
1191                         req->r_stamp = jiffies;
1192                 }
1193
1194                 return -EAGAIN;
1195         }
1196
1197         ceph_con_close(&osd->o_con);
1198         ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, peer_addr);
1199         osd->o_incarnation++;
1200
1201         return 0;
1202 }
1203
1204 static struct ceph_osd *lookup_create_osd(struct ceph_osd_client *osdc, int o,
1205                                           bool wrlocked)
1206 {
1207         struct ceph_osd *osd;
1208
1209         if (wrlocked)
1210                 verify_osdc_wrlocked(osdc);
1211         else
1212                 verify_osdc_locked(osdc);
1213
1214         if (o != CEPH_HOMELESS_OSD)
1215                 osd = lookup_osd(&osdc->osds, o);
1216         else
1217                 osd = &osdc->homeless_osd;
1218         if (!osd) {
1219                 if (!wrlocked)
1220                         return ERR_PTR(-EAGAIN);
1221
1222                 osd = create_osd(osdc, o);
1223                 insert_osd(&osdc->osds, osd);
1224                 ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd,
1225                               &osdc->osdmap->osd_addr[osd->o_osd]);
1226         }
1227
1228         dout("%s osdc %p osd%d -> osd %p\n", __func__, osdc, o, osd);
1229         return osd;
1230 }
1231
1232 /*
1233  * Create request <-> OSD session relation.
1234  *
1235  * @req has to be assigned a tid, @osd may be homeless.
1236  */
1237 static void link_request(struct ceph_osd *osd, struct ceph_osd_request *req)
1238 {
1239         verify_osd_locked(osd);
1240         WARN_ON(!req->r_tid || req->r_osd);
1241         dout("%s osd %p osd%d req %p tid %llu\n", __func__, osd, osd->o_osd,
1242              req, req->r_tid);
1243
1244         if (!osd_homeless(osd))
1245                 __remove_osd_from_lru(osd);
1246         else
1247                 atomic_inc(&osd->o_osdc->num_homeless);
1248
1249         get_osd(osd);
1250         insert_request(&osd->o_requests, req);
1251         req->r_osd = osd;
1252 }
1253
1254 static void unlink_request(struct ceph_osd *osd, struct ceph_osd_request *req)
1255 {
1256         verify_osd_locked(osd);
1257         WARN_ON(req->r_osd != osd);
1258         dout("%s osd %p osd%d req %p tid %llu\n", __func__, osd, osd->o_osd,
1259              req, req->r_tid);
1260
1261         req->r_osd = NULL;
1262         erase_request(&osd->o_requests, req);
1263         put_osd(osd);
1264
1265         if (!osd_homeless(osd))
1266                 maybe_move_osd_to_lru(osd);
1267         else
1268                 atomic_dec(&osd->o_osdc->num_homeless);
1269 }
1270
1271 static bool __pool_full(struct ceph_pg_pool_info *pi)
1272 {
1273         return pi->flags & CEPH_POOL_FLAG_FULL;
1274 }
1275
1276 static bool have_pool_full(struct ceph_osd_client *osdc)
1277 {
1278         struct rb_node *n;
1279
1280         for (n = rb_first(&osdc->osdmap->pg_pools); n; n = rb_next(n)) {
1281                 struct ceph_pg_pool_info *pi =
1282                     rb_entry(n, struct ceph_pg_pool_info, node);
1283
1284                 if (__pool_full(pi))
1285                         return true;
1286         }
1287
1288         return false;
1289 }
1290
1291 static bool pool_full(struct ceph_osd_client *osdc, s64 pool_id)
1292 {
1293         struct ceph_pg_pool_info *pi;
1294
1295         pi = ceph_pg_pool_by_id(osdc->osdmap, pool_id);
1296         if (!pi)
1297                 return false;
1298
1299         return __pool_full(pi);
1300 }
1301
1302 /*
1303  * Returns whether a request should be blocked from being sent
1304  * based on the current osdmap and osd_client settings.
1305  */
1306 static bool target_should_be_paused(struct ceph_osd_client *osdc,
1307                                     const struct ceph_osd_request_target *t,
1308                                     struct ceph_pg_pool_info *pi)
1309 {
1310         bool pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD);
1311         bool pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) ||
1312                        ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
1313                        __pool_full(pi);
1314
1315         WARN_ON(pi->id != t->target_oloc.pool);
1316         return ((t->flags & CEPH_OSD_FLAG_READ) && pauserd) ||
1317                ((t->flags & CEPH_OSD_FLAG_WRITE) && pausewr) ||
1318                (osdc->osdmap->epoch < osdc->epoch_barrier);
1319 }
1320
1321 enum calc_target_result {
1322         CALC_TARGET_NO_ACTION = 0,
1323         CALC_TARGET_NEED_RESEND,
1324         CALC_TARGET_POOL_DNE,
1325 };
1326
1327 static enum calc_target_result calc_target(struct ceph_osd_client *osdc,
1328                                            struct ceph_osd_request_target *t,
1329                                            struct ceph_connection *con,
1330                                            bool any_change)
1331 {
1332         struct ceph_pg_pool_info *pi;
1333         struct ceph_pg pgid, last_pgid;
1334         struct ceph_osds up, acting;
1335         bool force_resend = false;
1336         bool unpaused = false;
1337         bool legacy_change;
1338         bool split = false;
1339         bool sort_bitwise = ceph_osdmap_flag(osdc, CEPH_OSDMAP_SORTBITWISE);
1340         enum calc_target_result ct_res;
1341         int ret;
1342
1343         t->epoch = osdc->osdmap->epoch;
1344         pi = ceph_pg_pool_by_id(osdc->osdmap, t->base_oloc.pool);
1345         if (!pi) {
1346                 t->osd = CEPH_HOMELESS_OSD;
1347                 ct_res = CALC_TARGET_POOL_DNE;
1348                 goto out;
1349         }
1350
1351         if (osdc->osdmap->epoch == pi->last_force_request_resend) {
1352                 if (t->last_force_resend < pi->last_force_request_resend) {
1353                         t->last_force_resend = pi->last_force_request_resend;
1354                         force_resend = true;
1355                 } else if (t->last_force_resend == 0) {
1356                         force_resend = true;
1357                 }
1358         }
1359
1360         /* apply tiering */
1361         ceph_oid_copy(&t->target_oid, &t->base_oid);
1362         ceph_oloc_copy(&t->target_oloc, &t->base_oloc);
1363         if ((t->flags & CEPH_OSD_FLAG_IGNORE_OVERLAY) == 0) {
1364                 if (t->flags & CEPH_OSD_FLAG_READ && pi->read_tier >= 0)
1365                         t->target_oloc.pool = pi->read_tier;
1366                 if (t->flags & CEPH_OSD_FLAG_WRITE && pi->write_tier >= 0)
1367                         t->target_oloc.pool = pi->write_tier;
1368
1369                 pi = ceph_pg_pool_by_id(osdc->osdmap, t->target_oloc.pool);
1370                 if (!pi) {
1371                         t->osd = CEPH_HOMELESS_OSD;
1372                         ct_res = CALC_TARGET_POOL_DNE;
1373                         goto out;
1374                 }
1375         }
1376
1377         ret = __ceph_object_locator_to_pg(pi, &t->target_oid, &t->target_oloc,
1378                                           &pgid);
1379         if (ret) {
1380                 WARN_ON(ret != -ENOENT);
1381                 t->osd = CEPH_HOMELESS_OSD;
1382                 ct_res = CALC_TARGET_POOL_DNE;
1383                 goto out;
1384         }
1385         last_pgid.pool = pgid.pool;
1386         last_pgid.seed = ceph_stable_mod(pgid.seed, t->pg_num, t->pg_num_mask);
1387
1388         ceph_pg_to_up_acting_osds(osdc->osdmap, pi, &pgid, &up, &acting);
1389         if (any_change &&
1390             ceph_is_new_interval(&t->acting,
1391                                  &acting,
1392                                  &t->up,
1393                                  &up,
1394                                  t->size,
1395                                  pi->size,
1396                                  t->min_size,
1397                                  pi->min_size,
1398                                  t->pg_num,
1399                                  pi->pg_num,
1400                                  t->sort_bitwise,
1401                                  sort_bitwise,
1402                                  &last_pgid))
1403                 force_resend = true;
1404
1405         if (t->paused && !target_should_be_paused(osdc, t, pi)) {
1406                 t->paused = false;
1407                 unpaused = true;
1408         }
1409         legacy_change = ceph_pg_compare(&t->pgid, &pgid) ||
1410                         ceph_osds_changed(&t->acting, &acting, any_change);
1411         if (t->pg_num)
1412                 split = ceph_pg_is_split(&last_pgid, t->pg_num, pi->pg_num);
1413
1414         if (legacy_change || force_resend || split) {
1415                 t->pgid = pgid; /* struct */
1416                 ceph_pg_to_primary_shard(osdc->osdmap, pi, &pgid, &t->spgid);
1417                 ceph_osds_copy(&t->acting, &acting);
1418                 ceph_osds_copy(&t->up, &up);
1419                 t->size = pi->size;
1420                 t->min_size = pi->min_size;
1421                 t->pg_num = pi->pg_num;
1422                 t->pg_num_mask = pi->pg_num_mask;
1423                 t->sort_bitwise = sort_bitwise;
1424
1425                 t->osd = acting.primary;
1426         }
1427
1428         if (unpaused || legacy_change || force_resend ||
1429             (split && con && CEPH_HAVE_FEATURE(con->peer_features,
1430                                                RESEND_ON_SPLIT)))
1431                 ct_res = CALC_TARGET_NEED_RESEND;
1432         else
1433                 ct_res = CALC_TARGET_NO_ACTION;
1434
1435 out:
1436         dout("%s t %p -> ct_res %d osd %d\n", __func__, t, ct_res, t->osd);
1437         return ct_res;
1438 }
1439
1440 static struct ceph_spg_mapping *alloc_spg_mapping(void)
1441 {
1442         struct ceph_spg_mapping *spg;
1443
1444         spg = kmalloc(sizeof(*spg), GFP_NOIO);
1445         if (!spg)
1446                 return NULL;
1447
1448         RB_CLEAR_NODE(&spg->node);
1449         spg->backoffs = RB_ROOT;
1450         return spg;
1451 }
1452
1453 static void free_spg_mapping(struct ceph_spg_mapping *spg)
1454 {
1455         WARN_ON(!RB_EMPTY_NODE(&spg->node));
1456         WARN_ON(!RB_EMPTY_ROOT(&spg->backoffs));
1457
1458         kfree(spg);
1459 }
1460
1461 /*
1462  * rbtree of ceph_spg_mapping for handling map<spg_t, ...>, similar to
1463  * ceph_pg_mapping.  Used to track OSD backoffs -- a backoff [range] is
1464  * defined only within a specific spgid; it does not pass anything to
1465  * children on split, or to another primary.
1466  */
1467 DEFINE_RB_FUNCS2(spg_mapping, struct ceph_spg_mapping, spgid, ceph_spg_compare,
1468                  RB_BYPTR, const struct ceph_spg *, node)
1469
1470 static u64 hoid_get_bitwise_key(const struct ceph_hobject_id *hoid)
1471 {
1472         return hoid->is_max ? 0x100000000ull : hoid->hash_reverse_bits;
1473 }
1474
1475 static void hoid_get_effective_key(const struct ceph_hobject_id *hoid,
1476                                    void **pkey, size_t *pkey_len)
1477 {
1478         if (hoid->key_len) {
1479                 *pkey = hoid->key;
1480                 *pkey_len = hoid->key_len;
1481         } else {
1482                 *pkey = hoid->oid;
1483                 *pkey_len = hoid->oid_len;
1484         }
1485 }
1486
1487 static int compare_names(const void *name1, size_t name1_len,
1488                          const void *name2, size_t name2_len)
1489 {
1490         int ret;
1491
1492         ret = memcmp(name1, name2, min(name1_len, name2_len));
1493         if (!ret) {
1494                 if (name1_len < name2_len)
1495                         ret = -1;
1496                 else if (name1_len > name2_len)
1497                         ret = 1;
1498         }
1499         return ret;
1500 }
1501
1502 static int hoid_compare(const struct ceph_hobject_id *lhs,
1503                         const struct ceph_hobject_id *rhs)
1504 {
1505         void *effective_key1, *effective_key2;
1506         size_t effective_key1_len, effective_key2_len;
1507         int ret;
1508
1509         if (lhs->is_max < rhs->is_max)
1510                 return -1;
1511         if (lhs->is_max > rhs->is_max)
1512                 return 1;
1513
1514         if (lhs->pool < rhs->pool)
1515                 return -1;
1516         if (lhs->pool > rhs->pool)
1517                 return 1;
1518
1519         if (hoid_get_bitwise_key(lhs) < hoid_get_bitwise_key(rhs))
1520                 return -1;
1521         if (hoid_get_bitwise_key(lhs) > hoid_get_bitwise_key(rhs))
1522                 return 1;
1523
1524         ret = compare_names(lhs->nspace, lhs->nspace_len,
1525                             rhs->nspace, rhs->nspace_len);
1526         if (ret)
1527                 return ret;
1528
1529         hoid_get_effective_key(lhs, &effective_key1, &effective_key1_len);
1530         hoid_get_effective_key(rhs, &effective_key2, &effective_key2_len);
1531         ret = compare_names(effective_key1, effective_key1_len,
1532                             effective_key2, effective_key2_len);
1533         if (ret)
1534                 return ret;
1535
1536         ret = compare_names(lhs->oid, lhs->oid_len, rhs->oid, rhs->oid_len);
1537         if (ret)
1538                 return ret;
1539
1540         if (lhs->snapid < rhs->snapid)
1541                 return -1;
1542         if (lhs->snapid > rhs->snapid)
1543                 return 1;
1544
1545         return 0;
1546 }
1547
1548 /*
1549  * For decoding ->begin and ->end of MOSDBackoff only -- no MIN/MAX
1550  * compat stuff here.
1551  *
1552  * Assumes @hoid is zero-initialized.
1553  */
1554 static int decode_hoid(void **p, void *end, struct ceph_hobject_id *hoid)
1555 {
1556         u8 struct_v;
1557         u32 struct_len;
1558         int ret;
1559
1560         ret = ceph_start_decoding(p, end, 4, "hobject_t", &struct_v,
1561                                   &struct_len);
1562         if (ret)
1563                 return ret;
1564
1565         if (struct_v < 4) {
1566                 pr_err("got struct_v %d < 4 of hobject_t\n", struct_v);
1567                 goto e_inval;
1568         }
1569
1570         hoid->key = ceph_extract_encoded_string(p, end, &hoid->key_len,
1571                                                 GFP_NOIO);
1572         if (IS_ERR(hoid->key)) {
1573                 ret = PTR_ERR(hoid->key);
1574                 hoid->key = NULL;
1575                 return ret;
1576         }
1577
1578         hoid->oid = ceph_extract_encoded_string(p, end, &hoid->oid_len,
1579                                                 GFP_NOIO);
1580         if (IS_ERR(hoid->oid)) {
1581                 ret = PTR_ERR(hoid->oid);
1582                 hoid->oid = NULL;
1583                 return ret;
1584         }
1585
1586         ceph_decode_64_safe(p, end, hoid->snapid, e_inval);
1587         ceph_decode_32_safe(p, end, hoid->hash, e_inval);
1588         ceph_decode_8_safe(p, end, hoid->is_max, e_inval);
1589
1590         hoid->nspace = ceph_extract_encoded_string(p, end, &hoid->nspace_len,
1591                                                    GFP_NOIO);
1592         if (IS_ERR(hoid->nspace)) {
1593                 ret = PTR_ERR(hoid->nspace);
1594                 hoid->nspace = NULL;
1595                 return ret;
1596         }
1597
1598         ceph_decode_64_safe(p, end, hoid->pool, e_inval);
1599
1600         ceph_hoid_build_hash_cache(hoid);
1601         return 0;
1602
1603 e_inval:
1604         return -EINVAL;
1605 }
1606
1607 static int hoid_encoding_size(const struct ceph_hobject_id *hoid)
1608 {
1609         return 8 + 4 + 1 + 8 + /* snapid, hash, is_max, pool */
1610                4 + hoid->key_len + 4 + hoid->oid_len + 4 + hoid->nspace_len;
1611 }
1612
1613 static void encode_hoid(void **p, void *end, const struct ceph_hobject_id *hoid)
1614 {
1615         ceph_start_encoding(p, 4, 3, hoid_encoding_size(hoid));
1616         ceph_encode_string(p, end, hoid->key, hoid->key_len);
1617         ceph_encode_string(p, end, hoid->oid, hoid->oid_len);
1618         ceph_encode_64(p, hoid->snapid);
1619         ceph_encode_32(p, hoid->hash);
1620         ceph_encode_8(p, hoid->is_max);
1621         ceph_encode_string(p, end, hoid->nspace, hoid->nspace_len);
1622         ceph_encode_64(p, hoid->pool);
1623 }
1624
1625 static void free_hoid(struct ceph_hobject_id *hoid)
1626 {
1627         if (hoid) {
1628                 kfree(hoid->key);
1629                 kfree(hoid->oid);
1630                 kfree(hoid->nspace);
1631                 kfree(hoid);
1632         }
1633 }
1634
1635 static struct ceph_osd_backoff *alloc_backoff(void)
1636 {
1637         struct ceph_osd_backoff *backoff;
1638
1639         backoff = kzalloc(sizeof(*backoff), GFP_NOIO);
1640         if (!backoff)
1641                 return NULL;
1642
1643         RB_CLEAR_NODE(&backoff->spg_node);
1644         RB_CLEAR_NODE(&backoff->id_node);
1645         return backoff;
1646 }
1647
1648 static void free_backoff(struct ceph_osd_backoff *backoff)
1649 {
1650         WARN_ON(!RB_EMPTY_NODE(&backoff->spg_node));
1651         WARN_ON(!RB_EMPTY_NODE(&backoff->id_node));
1652
1653         free_hoid(backoff->begin);
1654         free_hoid(backoff->end);
1655         kfree(backoff);
1656 }
1657
1658 /*
1659  * Within a specific spgid, backoffs are managed by ->begin hoid.
1660  */
1661 DEFINE_RB_INSDEL_FUNCS2(backoff, struct ceph_osd_backoff, begin, hoid_compare,
1662                         RB_BYVAL, spg_node);
1663
1664 static struct ceph_osd_backoff *lookup_containing_backoff(struct rb_root *root,
1665                                             const struct ceph_hobject_id *hoid)
1666 {
1667         struct rb_node *n = root->rb_node;
1668
1669         while (n) {
1670                 struct ceph_osd_backoff *cur =
1671                     rb_entry(n, struct ceph_osd_backoff, spg_node);
1672                 int cmp;
1673
1674                 cmp = hoid_compare(hoid, cur->begin);
1675                 if (cmp < 0) {
1676                         n = n->rb_left;
1677                 } else if (cmp > 0) {
1678                         if (hoid_compare(hoid, cur->end) < 0)
1679                                 return cur;
1680
1681                         n = n->rb_right;
1682                 } else {
1683                         return cur;
1684                 }
1685         }
1686
1687         return NULL;
1688 }
1689
1690 /*
1691  * Each backoff has a unique id within its OSD session.
1692  */
1693 DEFINE_RB_FUNCS(backoff_by_id, struct ceph_osd_backoff, id, id_node)
1694
1695 static void clear_backoffs(struct ceph_osd *osd)
1696 {
1697         while (!RB_EMPTY_ROOT(&osd->o_backoff_mappings)) {
1698                 struct ceph_spg_mapping *spg =
1699                     rb_entry(rb_first(&osd->o_backoff_mappings),
1700                              struct ceph_spg_mapping, node);
1701
1702                 while (!RB_EMPTY_ROOT(&spg->backoffs)) {
1703                         struct ceph_osd_backoff *backoff =
1704                             rb_entry(rb_first(&spg->backoffs),
1705                                      struct ceph_osd_backoff, spg_node);
1706
1707                         erase_backoff(&spg->backoffs, backoff);
1708                         erase_backoff_by_id(&osd->o_backoffs_by_id, backoff);
1709                         free_backoff(backoff);
1710                 }
1711                 erase_spg_mapping(&osd->o_backoff_mappings, spg);
1712                 free_spg_mapping(spg);
1713         }
1714 }
1715
1716 /*
1717  * Set up a temporary, non-owning view into @t.
1718  */
1719 static void hoid_fill_from_target(struct ceph_hobject_id *hoid,
1720                                   const struct ceph_osd_request_target *t)
1721 {
1722         hoid->key = NULL;
1723         hoid->key_len = 0;
1724         hoid->oid = t->target_oid.name;
1725         hoid->oid_len = t->target_oid.name_len;
1726         hoid->snapid = CEPH_NOSNAP;
1727         hoid->hash = t->pgid.seed;
1728         hoid->is_max = false;
1729         if (t->target_oloc.pool_ns) {
1730                 hoid->nspace = t->target_oloc.pool_ns->str;
1731                 hoid->nspace_len = t->target_oloc.pool_ns->len;
1732         } else {
1733                 hoid->nspace = NULL;
1734                 hoid->nspace_len = 0;
1735         }
1736         hoid->pool = t->target_oloc.pool;
1737         ceph_hoid_build_hash_cache(hoid);
1738 }
1739
1740 static bool should_plug_request(struct ceph_osd_request *req)
1741 {
1742         struct ceph_osd *osd = req->r_osd;
1743         struct ceph_spg_mapping *spg;
1744         struct ceph_osd_backoff *backoff;
1745         struct ceph_hobject_id hoid;
1746
1747         spg = lookup_spg_mapping(&osd->o_backoff_mappings, &req->r_t.spgid);
1748         if (!spg)
1749                 return false;
1750
1751         hoid_fill_from_target(&hoid, &req->r_t);
1752         backoff = lookup_containing_backoff(&spg->backoffs, &hoid);
1753         if (!backoff)
1754                 return false;
1755
1756         dout("%s req %p tid %llu backoff osd%d spgid %llu.%xs%d id %llu\n",
1757              __func__, req, req->r_tid, osd->o_osd, backoff->spgid.pgid.pool,
1758              backoff->spgid.pgid.seed, backoff->spgid.shard, backoff->id);
1759         return true;
1760 }
1761
1762 static void setup_request_data(struct ceph_osd_request *req,
1763                                struct ceph_msg *msg)
1764 {
1765         u32 data_len = 0;
1766         int i;
1767
1768         if (!list_empty(&msg->data))
1769                 return;
1770
1771         WARN_ON(msg->data_length);
1772         for (i = 0; i < req->r_num_ops; i++) {
1773                 struct ceph_osd_req_op *op = &req->r_ops[i];
1774
1775                 switch (op->op) {
1776                 /* request */
1777                 case CEPH_OSD_OP_WRITE:
1778                 case CEPH_OSD_OP_WRITEFULL:
1779                         WARN_ON(op->indata_len != op->extent.length);
1780                         ceph_osdc_msg_data_add(msg, &op->extent.osd_data);
1781                         break;
1782                 case CEPH_OSD_OP_SETXATTR:
1783                 case CEPH_OSD_OP_CMPXATTR:
1784                         WARN_ON(op->indata_len != op->xattr.name_len +
1785                                                   op->xattr.value_len);
1786                         ceph_osdc_msg_data_add(msg, &op->xattr.osd_data);
1787                         break;
1788                 case CEPH_OSD_OP_NOTIFY_ACK:
1789                         ceph_osdc_msg_data_add(msg,
1790                                                &op->notify_ack.request_data);
1791                         break;
1792
1793                 /* reply */
1794                 case CEPH_OSD_OP_STAT:
1795                         ceph_osdc_msg_data_add(req->r_reply,
1796                                                &op->raw_data_in);
1797                         break;
1798                 case CEPH_OSD_OP_READ:
1799                         ceph_osdc_msg_data_add(req->r_reply,
1800                                                &op->extent.osd_data);
1801                         break;
1802                 case CEPH_OSD_OP_LIST_WATCHERS:
1803                         ceph_osdc_msg_data_add(req->r_reply,
1804                                                &op->list_watchers.response_data);
1805                         break;
1806
1807                 /* both */
1808                 case CEPH_OSD_OP_CALL:
1809                         WARN_ON(op->indata_len != op->cls.class_len +
1810                                                   op->cls.method_len +
1811                                                   op->cls.indata_len);
1812                         ceph_osdc_msg_data_add(msg, &op->cls.request_info);
1813                         /* optional, can be NONE */
1814                         ceph_osdc_msg_data_add(msg, &op->cls.request_data);
1815                         /* optional, can be NONE */
1816                         ceph_osdc_msg_data_add(req->r_reply,
1817                                                &op->cls.response_data);
1818                         break;
1819                 case CEPH_OSD_OP_NOTIFY:
1820                         ceph_osdc_msg_data_add(msg,
1821                                                &op->notify.request_data);
1822                         ceph_osdc_msg_data_add(req->r_reply,
1823                                                &op->notify.response_data);
1824                         break;
1825                 }
1826
1827                 data_len += op->indata_len;
1828         }
1829
1830         WARN_ON(data_len != msg->data_length);
1831 }
1832
1833 static void encode_pgid(void **p, const struct ceph_pg *pgid)
1834 {
1835         ceph_encode_8(p, 1);
1836         ceph_encode_64(p, pgid->pool);
1837         ceph_encode_32(p, pgid->seed);
1838         ceph_encode_32(p, -1); /* preferred */
1839 }
1840
1841 static void encode_spgid(void **p, const struct ceph_spg *spgid)
1842 {
1843         ceph_start_encoding(p, 1, 1, CEPH_PGID_ENCODING_LEN + 1);
1844         encode_pgid(p, &spgid->pgid);
1845         ceph_encode_8(p, spgid->shard);
1846 }
1847
1848 static void encode_oloc(void **p, void *end,
1849                         const struct ceph_object_locator *oloc)
1850 {
1851         ceph_start_encoding(p, 5, 4, ceph_oloc_encoding_size(oloc));
1852         ceph_encode_64(p, oloc->pool);
1853         ceph_encode_32(p, -1); /* preferred */
1854         ceph_encode_32(p, 0);  /* key len */
1855         if (oloc->pool_ns)
1856                 ceph_encode_string(p, end, oloc->pool_ns->str,
1857                                    oloc->pool_ns->len);
1858         else
1859                 ceph_encode_32(p, 0);
1860 }
1861
1862 static void encode_request_partial(struct ceph_osd_request *req,
1863                                    struct ceph_msg *msg)
1864 {
1865         void *p = msg->front.iov_base;
1866         void *const end = p + msg->front_alloc_len;
1867         u32 data_len = 0;
1868         int i;
1869
1870         if (req->r_flags & CEPH_OSD_FLAG_WRITE) {
1871                 /* snapshots aren't writeable */
1872                 WARN_ON(req->r_snapid != CEPH_NOSNAP);
1873         } else {
1874                 WARN_ON(req->r_mtime.tv_sec || req->r_mtime.tv_nsec ||
1875                         req->r_data_offset || req->r_snapc);
1876         }
1877
1878         setup_request_data(req, msg);
1879
1880         encode_spgid(&p, &req->r_t.spgid); /* actual spg */
1881         ceph_encode_32(&p, req->r_t.pgid.seed); /* raw hash */
1882         ceph_encode_32(&p, req->r_osdc->osdmap->epoch);
1883         ceph_encode_32(&p, req->r_flags);
1884
1885         /* reqid */
1886         ceph_start_encoding(&p, 2, 2, sizeof(struct ceph_osd_reqid));
1887         memset(p, 0, sizeof(struct ceph_osd_reqid));
1888         p += sizeof(struct ceph_osd_reqid);
1889
1890         /* trace */
1891         memset(p, 0, sizeof(struct ceph_blkin_trace_info));
1892         p += sizeof(struct ceph_blkin_trace_info);
1893
1894         ceph_encode_32(&p, 0); /* client_inc, always 0 */
1895         ceph_encode_timespec(p, &req->r_mtime);
1896         p += sizeof(struct ceph_timespec);
1897
1898         encode_oloc(&p, end, &req->r_t.target_oloc);
1899         ceph_encode_string(&p, end, req->r_t.target_oid.name,
1900                            req->r_t.target_oid.name_len);
1901
1902         /* ops, can imply data */
1903         ceph_encode_16(&p, req->r_num_ops);
1904         for (i = 0; i < req->r_num_ops; i++) {
1905                 data_len += osd_req_encode_op(p, &req->r_ops[i]);
1906                 p += sizeof(struct ceph_osd_op);
1907         }
1908
1909         ceph_encode_64(&p, req->r_snapid); /* snapid */
1910         if (req->r_snapc) {
1911                 ceph_encode_64(&p, req->r_snapc->seq);
1912                 ceph_encode_32(&p, req->r_snapc->num_snaps);
1913                 for (i = 0; i < req->r_snapc->num_snaps; i++)
1914                         ceph_encode_64(&p, req->r_snapc->snaps[i]);
1915         } else {
1916                 ceph_encode_64(&p, 0); /* snap_seq */
1917                 ceph_encode_32(&p, 0); /* snaps len */
1918         }
1919
1920         ceph_encode_32(&p, req->r_attempts); /* retry_attempt */
1921         BUG_ON(p > end - 8); /* space for features */
1922
1923         msg->hdr.version = cpu_to_le16(8); /* MOSDOp v8 */
1924         /* front_len is finalized in encode_request_finish() */
1925         msg->front.iov_len = p - msg->front.iov_base;
1926         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
1927         msg->hdr.data_len = cpu_to_le32(data_len);
1928         /*
1929          * The header "data_off" is a hint to the receiver allowing it
1930          * to align received data into its buffers such that there's no
1931          * need to re-copy it before writing it to disk (direct I/O).
1932          */
1933         msg->hdr.data_off = cpu_to_le16(req->r_data_offset);
1934
1935         dout("%s req %p msg %p oid %s oid_len %d\n", __func__, req, msg,
1936              req->r_t.target_oid.name, req->r_t.target_oid.name_len);
1937 }
1938
1939 static void encode_request_finish(struct ceph_msg *msg)
1940 {
1941         void *p = msg->front.iov_base;
1942         void *const partial_end = p + msg->front.iov_len;
1943         void *const end = p + msg->front_alloc_len;
1944
1945         if (CEPH_HAVE_FEATURE(msg->con->peer_features, RESEND_ON_SPLIT)) {
1946                 /* luminous OSD -- encode features and be done */
1947                 p = partial_end;
1948                 ceph_encode_64(&p, msg->con->peer_features);
1949         } else {
1950                 struct {
1951                         char spgid[CEPH_ENCODING_START_BLK_LEN +
1952                                    CEPH_PGID_ENCODING_LEN + 1];
1953                         __le32 hash;
1954                         __le32 epoch;
1955                         __le32 flags;
1956                         char reqid[CEPH_ENCODING_START_BLK_LEN +
1957                                    sizeof(struct ceph_osd_reqid)];
1958                         char trace[sizeof(struct ceph_blkin_trace_info)];
1959                         __le32 client_inc;
1960                         struct ceph_timespec mtime;
1961                 } __packed head;
1962                 struct ceph_pg pgid;
1963                 void *oloc, *oid, *tail;
1964                 int oloc_len, oid_len, tail_len;
1965                 int len;
1966
1967                 /*
1968                  * Pre-luminous OSD -- reencode v8 into v4 using @head
1969                  * as a temporary buffer.  Encode the raw PG; the rest
1970                  * is just a matter of moving oloc, oid and tail blobs
1971                  * around.
1972                  */
1973                 memcpy(&head, p, sizeof(head));
1974                 p += sizeof(head);
1975
1976                 oloc = p;
1977                 p += CEPH_ENCODING_START_BLK_LEN;
1978                 pgid.pool = ceph_decode_64(&p);
1979                 p += 4 + 4; /* preferred, key len */
1980                 len = ceph_decode_32(&p);
1981                 p += len;   /* nspace */
1982                 oloc_len = p - oloc;
1983
1984                 oid = p;
1985                 len = ceph_decode_32(&p);
1986                 p += len;
1987                 oid_len = p - oid;
1988
1989                 tail = p;
1990                 tail_len = partial_end - p;
1991
1992                 p = msg->front.iov_base;
1993                 ceph_encode_copy(&p, &head.client_inc, sizeof(head.client_inc));
1994                 ceph_encode_copy(&p, &head.epoch, sizeof(head.epoch));
1995                 ceph_encode_copy(&p, &head.flags, sizeof(head.flags));
1996                 ceph_encode_copy(&p, &head.mtime, sizeof(head.mtime));
1997
1998                 /* reassert_version */
1999                 memset(p, 0, sizeof(struct ceph_eversion));
2000                 p += sizeof(struct ceph_eversion);
2001
2002                 BUG_ON(p >= oloc);
2003                 memmove(p, oloc, oloc_len);
2004                 p += oloc_len;
2005
2006                 pgid.seed = le32_to_cpu(head.hash);
2007                 encode_pgid(&p, &pgid); /* raw pg */
2008
2009                 BUG_ON(p >= oid);
2010                 memmove(p, oid, oid_len);
2011                 p += oid_len;
2012
2013                 /* tail -- ops, snapid, snapc, retry_attempt */
2014                 BUG_ON(p >= tail);
2015                 memmove(p, tail, tail_len);
2016                 p += tail_len;
2017
2018                 msg->hdr.version = cpu_to_le16(4); /* MOSDOp v4 */
2019         }
2020
2021         BUG_ON(p > end);
2022         msg->front.iov_len = p - msg->front.iov_base;
2023         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
2024
2025         dout("%s msg %p tid %llu %u+%u+%u v%d\n", __func__, msg,
2026              le64_to_cpu(msg->hdr.tid), le32_to_cpu(msg->hdr.front_len),
2027              le32_to_cpu(msg->hdr.middle_len), le32_to_cpu(msg->hdr.data_len),
2028              le16_to_cpu(msg->hdr.version));
2029 }
2030
2031 /*
2032  * @req has to be assigned a tid and registered.
2033  */
2034 static void send_request(struct ceph_osd_request *req)
2035 {
2036         struct ceph_osd *osd = req->r_osd;
2037
2038         verify_osd_locked(osd);
2039         WARN_ON(osd->o_osd != req->r_t.osd);
2040
2041         /* backoff? */
2042         if (should_plug_request(req))
2043                 return;
2044
2045         /*
2046          * We may have a previously queued request message hanging
2047          * around.  Cancel it to avoid corrupting the msgr.
2048          */
2049         if (req->r_sent)
2050                 ceph_msg_revoke(req->r_request);
2051
2052         req->r_flags |= CEPH_OSD_FLAG_KNOWN_REDIR;
2053         if (req->r_attempts)
2054                 req->r_flags |= CEPH_OSD_FLAG_RETRY;
2055         else
2056                 WARN_ON(req->r_flags & CEPH_OSD_FLAG_RETRY);
2057
2058         encode_request_partial(req, req->r_request);
2059
2060         dout("%s req %p tid %llu to pgid %llu.%x spgid %llu.%xs%d osd%d e%u flags 0x%x attempt %d\n",
2061              __func__, req, req->r_tid, req->r_t.pgid.pool, req->r_t.pgid.seed,
2062              req->r_t.spgid.pgid.pool, req->r_t.spgid.pgid.seed,
2063              req->r_t.spgid.shard, osd->o_osd, req->r_t.epoch, req->r_flags,
2064              req->r_attempts);
2065
2066         req->r_t.paused = false;
2067         req->r_stamp = jiffies;
2068         req->r_attempts++;
2069
2070         req->r_sent = osd->o_incarnation;
2071         req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
2072         ceph_con_send(&osd->o_con, ceph_msg_get(req->r_request));
2073 }
2074
2075 static void maybe_request_map(struct ceph_osd_client *osdc)
2076 {
2077         bool continuous = false;
2078
2079         verify_osdc_locked(osdc);
2080         WARN_ON(!osdc->osdmap->epoch);
2081
2082         if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
2083             ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD) ||
2084             ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR)) {
2085                 dout("%s osdc %p continuous\n", __func__, osdc);
2086                 continuous = true;
2087         } else {
2088                 dout("%s osdc %p onetime\n", __func__, osdc);
2089         }
2090
2091         if (ceph_monc_want_map(&osdc->client->monc, CEPH_SUB_OSDMAP,
2092                                osdc->osdmap->epoch + 1, continuous))
2093                 ceph_monc_renew_subs(&osdc->client->monc);
2094 }
2095
2096 static void complete_request(struct ceph_osd_request *req, int err);
2097 static void send_map_check(struct ceph_osd_request *req);
2098
2099 static void __submit_request(struct ceph_osd_request *req, bool wrlocked)
2100 {
2101         struct ceph_osd_client *osdc = req->r_osdc;
2102         struct ceph_osd *osd;
2103         enum calc_target_result ct_res;
2104         bool need_send = false;
2105         bool promoted = false;
2106         bool need_abort = false;
2107
2108         WARN_ON(req->r_tid);
2109         dout("%s req %p wrlocked %d\n", __func__, req, wrlocked);
2110
2111 again:
2112         ct_res = calc_target(osdc, &req->r_t, NULL, false);
2113         if (ct_res == CALC_TARGET_POOL_DNE && !wrlocked)
2114                 goto promote;
2115
2116         osd = lookup_create_osd(osdc, req->r_t.osd, wrlocked);
2117         if (IS_ERR(osd)) {
2118                 WARN_ON(PTR_ERR(osd) != -EAGAIN || wrlocked);
2119                 goto promote;
2120         }
2121
2122         if (osdc->osdmap->epoch < osdc->epoch_barrier) {
2123                 dout("req %p epoch %u barrier %u\n", req, osdc->osdmap->epoch,
2124                      osdc->epoch_barrier);
2125                 req->r_t.paused = true;
2126                 maybe_request_map(osdc);
2127         } else if ((req->r_flags & CEPH_OSD_FLAG_WRITE) &&
2128                    ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR)) {
2129                 dout("req %p pausewr\n", req);
2130                 req->r_t.paused = true;
2131                 maybe_request_map(osdc);
2132         } else if ((req->r_flags & CEPH_OSD_FLAG_READ) &&
2133                    ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD)) {
2134                 dout("req %p pauserd\n", req);
2135                 req->r_t.paused = true;
2136                 maybe_request_map(osdc);
2137         } else if ((req->r_flags & CEPH_OSD_FLAG_WRITE) &&
2138                    !(req->r_flags & (CEPH_OSD_FLAG_FULL_TRY |
2139                                      CEPH_OSD_FLAG_FULL_FORCE)) &&
2140                    (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
2141                     pool_full(osdc, req->r_t.base_oloc.pool))) {
2142                 dout("req %p full/pool_full\n", req);
2143                 pr_warn_ratelimited("FULL or reached pool quota\n");
2144                 req->r_t.paused = true;
2145                 maybe_request_map(osdc);
2146                 if (req->r_abort_on_full)
2147                         need_abort = true;
2148         } else if (!osd_homeless(osd)) {
2149                 need_send = true;
2150         } else {
2151                 maybe_request_map(osdc);
2152         }
2153
2154         mutex_lock(&osd->lock);
2155         /*
2156          * Assign the tid atomically with send_request() to protect
2157          * multiple writes to the same object from racing with each
2158          * other, resulting in out of order ops on the OSDs.
2159          */
2160         req->r_tid = atomic64_inc_return(&osdc->last_tid);
2161         link_request(osd, req);
2162         if (need_send)
2163                 send_request(req);
2164         else if (need_abort)
2165                 complete_request(req, -ENOSPC);
2166         mutex_unlock(&osd->lock);
2167
2168         if (ct_res == CALC_TARGET_POOL_DNE)
2169                 send_map_check(req);
2170
2171         if (promoted)
2172                 downgrade_write(&osdc->lock);
2173         return;
2174
2175 promote:
2176         up_read(&osdc->lock);
2177         down_write(&osdc->lock);
2178         wrlocked = true;
2179         promoted = true;
2180         goto again;
2181 }
2182
2183 static void account_request(struct ceph_osd_request *req)
2184 {
2185         WARN_ON(req->r_flags & (CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK));
2186         WARN_ON(!(req->r_flags & (CEPH_OSD_FLAG_READ | CEPH_OSD_FLAG_WRITE)));
2187
2188         req->r_flags |= CEPH_OSD_FLAG_ONDISK;
2189         atomic_inc(&req->r_osdc->num_requests);
2190
2191         req->r_start_stamp = jiffies;
2192 }
2193
2194 static void submit_request(struct ceph_osd_request *req, bool wrlocked)
2195 {
2196         ceph_osdc_get_request(req);
2197         account_request(req);
2198         __submit_request(req, wrlocked);
2199 }
2200
2201 static void finish_request(struct ceph_osd_request *req)
2202 {
2203         struct ceph_osd_client *osdc = req->r_osdc;
2204
2205         WARN_ON(lookup_request_mc(&osdc->map_checks, req->r_tid));
2206         dout("%s req %p tid %llu\n", __func__, req, req->r_tid);
2207
2208         if (req->r_osd)
2209                 unlink_request(req->r_osd, req);
2210         atomic_dec(&osdc->num_requests);
2211
2212         /*
2213          * If an OSD has failed or returned and a request has been sent
2214          * twice, it's possible to get a reply and end up here while the
2215          * request message is queued for delivery.  We will ignore the
2216          * reply, so not a big deal, but better to try and catch it.
2217          */
2218         ceph_msg_revoke(req->r_request);
2219         ceph_msg_revoke_incoming(req->r_reply);
2220 }
2221
2222 static void __complete_request(struct ceph_osd_request *req)
2223 {
2224         if (req->r_callback) {
2225                 dout("%s req %p tid %llu cb %pf result %d\n", __func__, req,
2226                      req->r_tid, req->r_callback, req->r_result);
2227                 req->r_callback(req);
2228         }
2229 }
2230
2231 /*
2232  * This is open-coded in handle_reply().
2233  */
2234 static void complete_request(struct ceph_osd_request *req, int err)
2235 {
2236         dout("%s req %p tid %llu err %d\n", __func__, req, req->r_tid, err);
2237
2238         req->r_result = err;
2239         finish_request(req);
2240         __complete_request(req);
2241         complete_all(&req->r_completion);
2242         ceph_osdc_put_request(req);
2243 }
2244
2245 static void cancel_map_check(struct ceph_osd_request *req)
2246 {
2247         struct ceph_osd_client *osdc = req->r_osdc;
2248         struct ceph_osd_request *lookup_req;
2249
2250         verify_osdc_wrlocked(osdc);
2251
2252         lookup_req = lookup_request_mc(&osdc->map_checks, req->r_tid);
2253         if (!lookup_req)
2254                 return;
2255
2256         WARN_ON(lookup_req != req);
2257         erase_request_mc(&osdc->map_checks, req);
2258         ceph_osdc_put_request(req);
2259 }
2260
2261 static void cancel_request(struct ceph_osd_request *req)
2262 {
2263         dout("%s req %p tid %llu\n", __func__, req, req->r_tid);
2264
2265         cancel_map_check(req);
2266         finish_request(req);
2267         complete_all(&req->r_completion);
2268         ceph_osdc_put_request(req);
2269 }
2270
2271 static void abort_request(struct ceph_osd_request *req, int err)
2272 {
2273         dout("%s req %p tid %llu err %d\n", __func__, req, req->r_tid, err);
2274
2275         cancel_map_check(req);
2276         complete_request(req, err);
2277 }
2278
2279 static void update_epoch_barrier(struct ceph_osd_client *osdc, u32 eb)
2280 {
2281         if (likely(eb > osdc->epoch_barrier)) {
2282                 dout("updating epoch_barrier from %u to %u\n",
2283                                 osdc->epoch_barrier, eb);
2284                 osdc->epoch_barrier = eb;
2285                 /* Request map if we're not to the barrier yet */
2286                 if (eb > osdc->osdmap->epoch)
2287                         maybe_request_map(osdc);
2288         }
2289 }
2290
2291 void ceph_osdc_update_epoch_barrier(struct ceph_osd_client *osdc, u32 eb)
2292 {
2293         down_read(&osdc->lock);
2294         if (unlikely(eb > osdc->epoch_barrier)) {
2295                 up_read(&osdc->lock);
2296                 down_write(&osdc->lock);
2297                 update_epoch_barrier(osdc, eb);
2298                 up_write(&osdc->lock);
2299         } else {
2300                 up_read(&osdc->lock);
2301         }
2302 }
2303 EXPORT_SYMBOL(ceph_osdc_update_epoch_barrier);
2304
2305 /*
2306  * Drop all pending requests that are stalled waiting on a full condition to
2307  * clear, and complete them with ENOSPC as the return code. Set the
2308  * osdc->epoch_barrier to the latest map epoch that we've seen if any were
2309  * cancelled.
2310  */
2311 static void ceph_osdc_abort_on_full(struct ceph_osd_client *osdc)
2312 {
2313         struct rb_node *n;
2314         bool victims = false;
2315
2316         dout("enter abort_on_full\n");
2317
2318         if (!ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) && !have_pool_full(osdc))
2319                 goto out;
2320
2321         /* Scan list and see if there is anything to abort */
2322         for (n = rb_first(&osdc->osds); n; n = rb_next(n)) {
2323                 struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node);
2324                 struct rb_node *m;
2325
2326                 m = rb_first(&osd->o_requests);
2327                 while (m) {
2328                         struct ceph_osd_request *req = rb_entry(m,
2329                                         struct ceph_osd_request, r_node);
2330                         m = rb_next(m);
2331
2332                         if (req->r_abort_on_full) {
2333                                 victims = true;
2334                                 break;
2335                         }
2336                 }
2337                 if (victims)
2338                         break;
2339         }
2340
2341         if (!victims)
2342                 goto out;
2343
2344         /*
2345          * Update the barrier to current epoch if it's behind that point,
2346          * since we know we have some calls to be aborted in the tree.
2347          */
2348         update_epoch_barrier(osdc, osdc->osdmap->epoch);
2349
2350         for (n = rb_first(&osdc->osds); n; n = rb_next(n)) {
2351                 struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node);
2352                 struct rb_node *m;
2353
2354                 m = rb_first(&osd->o_requests);
2355                 while (m) {
2356                         struct ceph_osd_request *req = rb_entry(m,
2357                                         struct ceph_osd_request, r_node);
2358                         m = rb_next(m);
2359
2360                         if (req->r_abort_on_full &&
2361                             (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
2362                              pool_full(osdc, req->r_t.target_oloc.pool)))
2363                                 abort_request(req, -ENOSPC);
2364                 }
2365         }
2366 out:
2367         dout("return abort_on_full barrier=%u\n", osdc->epoch_barrier);
2368 }
2369
2370 static void check_pool_dne(struct ceph_osd_request *req)
2371 {
2372         struct ceph_osd_client *osdc = req->r_osdc;
2373         struct ceph_osdmap *map = osdc->osdmap;
2374
2375         verify_osdc_wrlocked(osdc);
2376         WARN_ON(!map->epoch);
2377
2378         if (req->r_attempts) {
2379                 /*
2380                  * We sent a request earlier, which means that
2381                  * previously the pool existed, and now it does not
2382                  * (i.e., it was deleted).
2383                  */
2384                 req->r_map_dne_bound = map->epoch;
2385                 dout("%s req %p tid %llu pool disappeared\n", __func__, req,
2386                      req->r_tid);
2387         } else {
2388                 dout("%s req %p tid %llu map_dne_bound %u have %u\n", __func__,
2389                      req, req->r_tid, req->r_map_dne_bound, map->epoch);
2390         }
2391
2392         if (req->r_map_dne_bound) {
2393                 if (map->epoch >= req->r_map_dne_bound) {
2394                         /* we had a new enough map */
2395                         pr_info_ratelimited("tid %llu pool does not exist\n",
2396                                             req->r_tid);
2397                         complete_request(req, -ENOENT);
2398                 }
2399         } else {
2400                 send_map_check(req);
2401         }
2402 }
2403
2404 static void map_check_cb(struct ceph_mon_generic_request *greq)
2405 {
2406         struct ceph_osd_client *osdc = &greq->monc->client->osdc;
2407         struct ceph_osd_request *req;
2408         u64 tid = greq->private_data;
2409
2410         WARN_ON(greq->result || !greq->u.newest);
2411
2412         down_write(&osdc->lock);
2413         req = lookup_request_mc(&osdc->map_checks, tid);
2414         if (!req) {
2415                 dout("%s tid %llu dne\n", __func__, tid);
2416                 goto out_unlock;
2417         }
2418
2419         dout("%s req %p tid %llu map_dne_bound %u newest %llu\n", __func__,
2420              req, req->r_tid, req->r_map_dne_bound, greq->u.newest);
2421         if (!req->r_map_dne_bound)
2422                 req->r_map_dne_bound = greq->u.newest;
2423         erase_request_mc(&osdc->map_checks, req);
2424         check_pool_dne(req);
2425
2426         ceph_osdc_put_request(req);
2427 out_unlock:
2428         up_write(&osdc->lock);
2429 }
2430
2431 static void send_map_check(struct ceph_osd_request *req)
2432 {
2433         struct ceph_osd_client *osdc = req->r_osdc;
2434         struct ceph_osd_request *lookup_req;
2435         int ret;
2436
2437         verify_osdc_wrlocked(osdc);
2438
2439         lookup_req = lookup_request_mc(&osdc->map_checks, req->r_tid);
2440         if (lookup_req) {
2441                 WARN_ON(lookup_req != req);
2442                 return;
2443         }
2444
2445         ceph_osdc_get_request(req);
2446         insert_request_mc(&osdc->map_checks, req);
2447         ret = ceph_monc_get_version_async(&osdc->client->monc, "osdmap",
2448                                           map_check_cb, req->r_tid);
2449         WARN_ON(ret);
2450 }
2451
2452 /*
2453  * lingering requests, watch/notify v2 infrastructure
2454  */
2455 static void linger_release(struct kref *kref)
2456 {
2457         struct ceph_osd_linger_request *lreq =
2458             container_of(kref, struct ceph_osd_linger_request, kref);
2459
2460         dout("%s lreq %p reg_req %p ping_req %p\n", __func__, lreq,
2461              lreq->reg_req, lreq->ping_req);
2462         WARN_ON(!RB_EMPTY_NODE(&lreq->node));
2463         WARN_ON(!RB_EMPTY_NODE(&lreq->osdc_node));
2464         WARN_ON(!RB_EMPTY_NODE(&lreq->mc_node));
2465         WARN_ON(!list_empty(&lreq->scan_item));
2466         WARN_ON(!list_empty(&lreq->pending_lworks));
2467         WARN_ON(lreq->osd);
2468
2469         if (lreq->reg_req)
2470                 ceph_osdc_put_request(lreq->reg_req);
2471         if (lreq->ping_req)
2472                 ceph_osdc_put_request(lreq->ping_req);
2473         target_destroy(&lreq->t);
2474         kfree(lreq);
2475 }
2476
2477 static void linger_put(struct ceph_osd_linger_request *lreq)
2478 {
2479         if (lreq)
2480                 kref_put(&lreq->kref, linger_release);
2481 }
2482
2483 static struct ceph_osd_linger_request *
2484 linger_get(struct ceph_osd_linger_request *lreq)
2485 {
2486         kref_get(&lreq->kref);
2487         return lreq;
2488 }
2489
2490 static struct ceph_osd_linger_request *
2491 linger_alloc(struct ceph_osd_client *osdc)
2492 {
2493         struct ceph_osd_linger_request *lreq;
2494
2495         lreq = kzalloc(sizeof(*lreq), GFP_NOIO);
2496         if (!lreq)
2497                 return NULL;
2498
2499         kref_init(&lreq->kref);
2500         mutex_init(&lreq->lock);
2501         RB_CLEAR_NODE(&lreq->node);
2502         RB_CLEAR_NODE(&lreq->osdc_node);
2503         RB_CLEAR_NODE(&lreq->mc_node);
2504         INIT_LIST_HEAD(&lreq->scan_item);
2505         INIT_LIST_HEAD(&lreq->pending_lworks);
2506         init_completion(&lreq->reg_commit_wait);
2507         init_completion(&lreq->notify_finish_wait);
2508
2509         lreq->osdc = osdc;
2510         target_init(&lreq->t);
2511
2512         dout("%s lreq %p\n", __func__, lreq);
2513         return lreq;
2514 }
2515
2516 DEFINE_RB_INSDEL_FUNCS(linger, struct ceph_osd_linger_request, linger_id, node)
2517 DEFINE_RB_FUNCS(linger_osdc, struct ceph_osd_linger_request, linger_id, osdc_node)
2518 DEFINE_RB_FUNCS(linger_mc, struct ceph_osd_linger_request, linger_id, mc_node)
2519
2520 /*
2521  * Create linger request <-> OSD session relation.
2522  *
2523  * @lreq has to be registered, @osd may be homeless.
2524  */
2525 static void link_linger(struct ceph_osd *osd,
2526                         struct ceph_osd_linger_request *lreq)
2527 {
2528         verify_osd_locked(osd);
2529         WARN_ON(!lreq->linger_id || lreq->osd);
2530         dout("%s osd %p osd%d lreq %p linger_id %llu\n", __func__, osd,
2531              osd->o_osd, lreq, lreq->linger_id);
2532
2533         if (!osd_homeless(osd))
2534                 __remove_osd_from_lru(osd);
2535         else
2536                 atomic_inc(&osd->o_osdc->num_homeless);
2537
2538         get_osd(osd);
2539         insert_linger(&osd->o_linger_requests, lreq);
2540         lreq->osd = osd;
2541 }
2542
2543 static void unlink_linger(struct ceph_osd *osd,
2544                           struct ceph_osd_linger_request *lreq)
2545 {
2546         verify_osd_locked(osd);
2547         WARN_ON(lreq->osd != osd);
2548         dout("%s osd %p osd%d lreq %p linger_id %llu\n", __func__, osd,
2549              osd->o_osd, lreq, lreq->linger_id);
2550
2551         lreq->osd = NULL;
2552         erase_linger(&osd->o_linger_requests, lreq);
2553         put_osd(osd);
2554
2555         if (!osd_homeless(osd))
2556                 maybe_move_osd_to_lru(osd);
2557         else
2558                 atomic_dec(&osd->o_osdc->num_homeless);
2559 }
2560
2561 static bool __linger_registered(struct ceph_osd_linger_request *lreq)
2562 {
2563         verify_osdc_locked(lreq->osdc);
2564
2565         return !RB_EMPTY_NODE(&lreq->osdc_node);
2566 }
2567
2568 static bool linger_registered(struct ceph_osd_linger_request *lreq)
2569 {
2570         struct ceph_osd_client *osdc = lreq->osdc;
2571         bool registered;
2572
2573         down_read(&osdc->lock);
2574         registered = __linger_registered(lreq);
2575         up_read(&osdc->lock);
2576
2577         return registered;
2578 }
2579
2580 static void linger_register(struct ceph_osd_linger_request *lreq)
2581 {
2582         struct ceph_osd_client *osdc = lreq->osdc;
2583
2584         verify_osdc_wrlocked(osdc);
2585         WARN_ON(lreq->linger_id);
2586
2587         linger_get(lreq);
2588         lreq->linger_id = ++osdc->last_linger_id;
2589         insert_linger_osdc(&osdc->linger_requests, lreq);
2590 }
2591
2592 static void linger_unregister(struct ceph_osd_linger_request *lreq)
2593 {
2594         struct ceph_osd_client *osdc = lreq->osdc;
2595
2596         verify_osdc_wrlocked(osdc);
2597
2598         erase_linger_osdc(&osdc->linger_requests, lreq);
2599         linger_put(lreq);
2600 }
2601
2602 static void cancel_linger_request(struct ceph_osd_request *req)
2603 {
2604         struct ceph_osd_linger_request *lreq = req->r_priv;
2605
2606         WARN_ON(!req->r_linger);
2607         cancel_request(req);
2608         linger_put(lreq);
2609 }
2610
2611 struct linger_work {
2612         struct work_struct work;
2613         struct ceph_osd_linger_request *lreq;
2614         struct list_head pending_item;
2615         unsigned long queued_stamp;
2616
2617         union {
2618                 struct {
2619                         u64 notify_id;
2620                         u64 notifier_id;
2621                         void *payload; /* points into @msg front */
2622                         size_t payload_len;
2623
2624                         struct ceph_msg *msg; /* for ceph_msg_put() */
2625                 } notify;
2626                 struct {
2627                         int err;
2628                 } error;
2629         };
2630 };
2631
2632 static struct linger_work *lwork_alloc(struct ceph_osd_linger_request *lreq,
2633                                        work_func_t workfn)
2634 {
2635         struct linger_work *lwork;
2636
2637         lwork = kzalloc(sizeof(*lwork), GFP_NOIO);
2638         if (!lwork)
2639                 return NULL;
2640
2641         INIT_WORK(&lwork->work, workfn);
2642         INIT_LIST_HEAD(&lwork->pending_item);
2643         lwork->lreq = linger_get(lreq);
2644
2645         return lwork;
2646 }
2647
2648 static void lwork_free(struct linger_work *lwork)
2649 {
2650         struct ceph_osd_linger_request *lreq = lwork->lreq;
2651
2652         mutex_lock(&lreq->lock);
2653         list_del(&lwork->pending_item);
2654         mutex_unlock(&lreq->lock);
2655
2656         linger_put(lreq);
2657         kfree(lwork);
2658 }
2659
2660 static void lwork_queue(struct linger_work *lwork)
2661 {
2662         struct ceph_osd_linger_request *lreq = lwork->lreq;
2663         struct ceph_osd_client *osdc = lreq->osdc;
2664
2665         verify_lreq_locked(lreq);
2666         WARN_ON(!list_empty(&lwork->pending_item));
2667
2668         lwork->queued_stamp = jiffies;
2669         list_add_tail(&lwork->pending_item, &lreq->pending_lworks);
2670         queue_work(osdc->notify_wq, &lwork->work);
2671 }
2672
2673 static void do_watch_notify(struct work_struct *w)
2674 {
2675         struct linger_work *lwork = container_of(w, struct linger_work, work);
2676         struct ceph_osd_linger_request *lreq = lwork->lreq;
2677
2678         if (!linger_registered(lreq)) {
2679                 dout("%s lreq %p not registered\n", __func__, lreq);
2680                 goto out;
2681         }
2682
2683         WARN_ON(!lreq->is_watch);
2684         dout("%s lreq %p notify_id %llu notifier_id %llu payload_len %zu\n",
2685              __func__, lreq, lwork->notify.notify_id, lwork->notify.notifier_id,
2686              lwork->notify.payload_len);
2687         lreq->wcb(lreq->data, lwork->notify.notify_id, lreq->linger_id,
2688                   lwork->notify.notifier_id, lwork->notify.payload,
2689                   lwork->notify.payload_len);
2690
2691 out:
2692         ceph_msg_put(lwork->notify.msg);
2693         lwork_free(lwork);
2694 }
2695
2696 static void do_watch_error(struct work_struct *w)
2697 {
2698         struct linger_work *lwork = container_of(w, struct linger_work, work);
2699         struct ceph_osd_linger_request *lreq = lwork->lreq;
2700
2701         if (!linger_registered(lreq)) {
2702                 dout("%s lreq %p not registered\n", __func__, lreq);
2703                 goto out;
2704         }
2705
2706         dout("%s lreq %p err %d\n", __func__, lreq, lwork->error.err);
2707         lreq->errcb(lreq->data, lreq->linger_id, lwork->error.err);
2708
2709 out:
2710         lwork_free(lwork);
2711 }
2712
2713 static void queue_watch_error(struct ceph_osd_linger_request *lreq)
2714 {
2715         struct linger_work *lwork;
2716
2717         lwork = lwork_alloc(lreq, do_watch_error);
2718         if (!lwork) {
2719                 pr_err("failed to allocate error-lwork\n");
2720                 return;
2721         }
2722
2723         lwork->error.err = lreq->last_error;
2724         lwork_queue(lwork);
2725 }
2726
2727 static void linger_reg_commit_complete(struct ceph_osd_linger_request *lreq,
2728                                        int result)
2729 {
2730         if (!completion_done(&lreq->reg_commit_wait)) {
2731                 lreq->reg_commit_error = (result <= 0 ? result : 0);
2732                 complete_all(&lreq->reg_commit_wait);
2733         }
2734 }
2735
2736 static void linger_commit_cb(struct ceph_osd_request *req)
2737 {
2738         struct ceph_osd_linger_request *lreq = req->r_priv;
2739
2740         mutex_lock(&lreq->lock);
2741         dout("%s lreq %p linger_id %llu result %d\n", __func__, lreq,
2742              lreq->linger_id, req->r_result);
2743         linger_reg_commit_complete(lreq, req->r_result);
2744         lreq->committed = true;
2745
2746         if (!lreq->is_watch) {
2747                 struct ceph_osd_data *osd_data =
2748                     osd_req_op_data(req, 0, notify, response_data);
2749                 void *p = page_address(osd_data->pages[0]);
2750
2751                 WARN_ON(req->r_ops[0].op != CEPH_OSD_OP_NOTIFY ||
2752                         osd_data->type != CEPH_OSD_DATA_TYPE_PAGES);
2753
2754                 /* make note of the notify_id */
2755                 if (req->r_ops[0].outdata_len >= sizeof(u64)) {
2756                         lreq->notify_id = ceph_decode_64(&p);
2757                         dout("lreq %p notify_id %llu\n", lreq,
2758                              lreq->notify_id);
2759                 } else {
2760                         dout("lreq %p no notify_id\n", lreq);
2761                 }
2762         }
2763
2764         mutex_unlock(&lreq->lock);
2765         linger_put(lreq);
2766 }
2767
2768 static int normalize_watch_error(int err)
2769 {
2770         /*
2771          * Translate ENOENT -> ENOTCONN so that a delete->disconnection
2772          * notification and a failure to reconnect because we raced with
2773          * the delete appear the same to the user.
2774          */
2775         if (err == -ENOENT)
2776                 err = -ENOTCONN;
2777
2778         return err;
2779 }
2780
2781 static void linger_reconnect_cb(struct ceph_osd_request *req)
2782 {
2783         struct ceph_osd_linger_request *lreq = req->r_priv;
2784
2785         mutex_lock(&lreq->lock);
2786         dout("%s lreq %p linger_id %llu result %d last_error %d\n", __func__,
2787              lreq, lreq->linger_id, req->r_result, lreq->last_error);
2788         if (req->r_result < 0) {
2789                 if (!lreq->last_error) {
2790                         lreq->last_error = normalize_watch_error(req->r_result);
2791                         queue_watch_error(lreq);
2792                 }
2793         }
2794
2795         mutex_unlock(&lreq->lock);
2796         linger_put(lreq);
2797 }
2798
2799 static void send_linger(struct ceph_osd_linger_request *lreq)
2800 {
2801         struct ceph_osd_request *req = lreq->reg_req;
2802         struct ceph_osd_req_op *op = &req->r_ops[0];
2803
2804         verify_osdc_wrlocked(req->r_osdc);
2805         dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
2806
2807         if (req->r_osd)
2808                 cancel_linger_request(req);
2809
2810         request_reinit(req);
2811         ceph_oid_copy(&req->r_base_oid, &lreq->t.base_oid);
2812         ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
2813         req->r_flags = lreq->t.flags;
2814         req->r_mtime = lreq->mtime;
2815
2816         mutex_lock(&lreq->lock);
2817         if (lreq->is_watch && lreq->committed) {
2818                 WARN_ON(op->op != CEPH_OSD_OP_WATCH ||
2819                         op->watch.cookie != lreq->linger_id);
2820                 op->watch.op = CEPH_OSD_WATCH_OP_RECONNECT;
2821                 op->watch.gen = ++lreq->register_gen;
2822                 dout("lreq %p reconnect register_gen %u\n", lreq,
2823                      op->watch.gen);
2824                 req->r_callback = linger_reconnect_cb;
2825         } else {
2826                 if (!lreq->is_watch)
2827                         lreq->notify_id = 0;
2828                 else
2829                         WARN_ON(op->watch.op != CEPH_OSD_WATCH_OP_WATCH);
2830                 dout("lreq %p register\n", lreq);
2831                 req->r_callback = linger_commit_cb;
2832         }
2833         mutex_unlock(&lreq->lock);
2834
2835         req->r_priv = linger_get(lreq);
2836         req->r_linger = true;
2837
2838         submit_request(req, true);
2839 }
2840
2841 static void linger_ping_cb(struct ceph_osd_request *req)
2842 {
2843         struct ceph_osd_linger_request *lreq = req->r_priv;
2844
2845         mutex_lock(&lreq->lock);
2846         dout("%s lreq %p linger_id %llu result %d ping_sent %lu last_error %d\n",
2847              __func__, lreq, lreq->linger_id, req->r_result, lreq->ping_sent,
2848              lreq->last_error);
2849         if (lreq->register_gen == req->r_ops[0].watch.gen) {
2850                 if (!req->r_result) {
2851                         lreq->watch_valid_thru = lreq->ping_sent;
2852                 } else if (!lreq->last_error) {
2853                         lreq->last_error = normalize_watch_error(req->r_result);
2854                         queue_watch_error(lreq);
2855                 }
2856         } else {
2857                 dout("lreq %p register_gen %u ignoring old pong %u\n", lreq,
2858                      lreq->register_gen, req->r_ops[0].watch.gen);
2859         }
2860
2861         mutex_unlock(&lreq->lock);
2862         linger_put(lreq);
2863 }
2864
2865 static void send_linger_ping(struct ceph_osd_linger_request *lreq)
2866 {
2867         struct ceph_osd_client *osdc = lreq->osdc;
2868         struct ceph_osd_request *req = lreq->ping_req;
2869         struct ceph_osd_req_op *op = &req->r_ops[0];
2870
2871         if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD)) {
2872                 dout("%s PAUSERD\n", __func__);
2873                 return;
2874         }
2875
2876         lreq->ping_sent = jiffies;
2877         dout("%s lreq %p linger_id %llu ping_sent %lu register_gen %u\n",
2878              __func__, lreq, lreq->linger_id, lreq->ping_sent,
2879              lreq->register_gen);
2880
2881         if (req->r_osd)
2882                 cancel_linger_request(req);
2883
2884         request_reinit(req);
2885         target_copy(&req->r_t, &lreq->t);
2886
2887         WARN_ON(op->op != CEPH_OSD_OP_WATCH ||
2888                 op->watch.cookie != lreq->linger_id ||
2889                 op->watch.op != CEPH_OSD_WATCH_OP_PING);
2890         op->watch.gen = lreq->register_gen;
2891         req->r_callback = linger_ping_cb;
2892         req->r_priv = linger_get(lreq);
2893         req->r_linger = true;
2894
2895         ceph_osdc_get_request(req);
2896         account_request(req);
2897         req->r_tid = atomic64_inc_return(&osdc->last_tid);
2898         link_request(lreq->osd, req);
2899         send_request(req);
2900 }
2901
2902 static void linger_submit(struct ceph_osd_linger_request *lreq)
2903 {
2904         struct ceph_osd_client *osdc = lreq->osdc;
2905         struct ceph_osd *osd;
2906
2907         calc_target(osdc, &lreq->t, NULL, false);
2908         osd = lookup_create_osd(osdc, lreq->t.osd, true);
2909         link_linger(osd, lreq);
2910
2911         send_linger(lreq);
2912 }
2913
2914 static void cancel_linger_map_check(struct ceph_osd_linger_request *lreq)
2915 {
2916         struct ceph_osd_client *osdc = lreq->osdc;
2917         struct ceph_osd_linger_request *lookup_lreq;
2918
2919         verify_osdc_wrlocked(osdc);
2920
2921         lookup_lreq = lookup_linger_mc(&osdc->linger_map_checks,
2922                                        lreq->linger_id);
2923         if (!lookup_lreq)
2924                 return;
2925
2926         WARN_ON(lookup_lreq != lreq);
2927         erase_linger_mc(&osdc->linger_map_checks, lreq);
2928         linger_put(lreq);
2929 }
2930
2931 /*
2932  * @lreq has to be both registered and linked.
2933  */
2934 static void __linger_cancel(struct ceph_osd_linger_request *lreq)
2935 {
2936         if (lreq->is_watch && lreq->ping_req->r_osd)
2937                 cancel_linger_request(lreq->ping_req);
2938         if (lreq->reg_req->r_osd)
2939                 cancel_linger_request(lreq->reg_req);
2940         cancel_linger_map_check(lreq);
2941         unlink_linger(lreq->osd, lreq);
2942         linger_unregister(lreq);
2943 }
2944
2945 static void linger_cancel(struct ceph_osd_linger_request *lreq)
2946 {
2947         struct ceph_osd_client *osdc = lreq->osdc;
2948
2949         down_write(&osdc->lock);
2950         if (__linger_registered(lreq))
2951                 __linger_cancel(lreq);
2952         up_write(&osdc->lock);
2953 }
2954
2955 static void send_linger_map_check(struct ceph_osd_linger_request *lreq);
2956
2957 static void check_linger_pool_dne(struct ceph_osd_linger_request *lreq)
2958 {
2959         struct ceph_osd_client *osdc = lreq->osdc;
2960         struct ceph_osdmap *map = osdc->osdmap;
2961
2962         verify_osdc_wrlocked(osdc);
2963         WARN_ON(!map->epoch);
2964
2965         if (lreq->register_gen) {
2966                 lreq->map_dne_bound = map->epoch;
2967                 dout("%s lreq %p linger_id %llu pool disappeared\n", __func__,
2968                      lreq, lreq->linger_id);
2969         } else {
2970                 dout("%s lreq %p linger_id %llu map_dne_bound %u have %u\n",
2971                      __func__, lreq, lreq->linger_id, lreq->map_dne_bound,
2972                      map->epoch);
2973         }
2974
2975         if (lreq->map_dne_bound) {
2976                 if (map->epoch >= lreq->map_dne_bound) {
2977                         /* we had a new enough map */
2978                         pr_info("linger_id %llu pool does not exist\n",
2979                                 lreq->linger_id);
2980                         linger_reg_commit_complete(lreq, -ENOENT);
2981                         __linger_cancel(lreq);
2982                 }
2983         } else {
2984                 send_linger_map_check(lreq);
2985         }
2986 }
2987
2988 static void linger_map_check_cb(struct ceph_mon_generic_request *greq)
2989 {
2990         struct ceph_osd_client *osdc = &greq->monc->client->osdc;
2991         struct ceph_osd_linger_request *lreq;
2992         u64 linger_id = greq->private_data;
2993
2994         WARN_ON(greq->result || !greq->u.newest);
2995
2996         down_write(&osdc->lock);
2997         lreq = lookup_linger_mc(&osdc->linger_map_checks, linger_id);
2998         if (!lreq) {
2999                 dout("%s linger_id %llu dne\n", __func__, linger_id);
3000                 goto out_unlock;
3001         }
3002
3003         dout("%s lreq %p linger_id %llu map_dne_bound %u newest %llu\n",
3004              __func__, lreq, lreq->linger_id, lreq->map_dne_bound,
3005              greq->u.newest);
3006         if (!lreq->map_dne_bound)
3007                 lreq->map_dne_bound = greq->u.newest;
3008         erase_linger_mc(&osdc->linger_map_checks, lreq);
3009         check_linger_pool_dne(lreq);
3010
3011         linger_put(lreq);
3012 out_unlock:
3013         up_write(&osdc->lock);
3014 }
3015
3016 static void send_linger_map_check(struct ceph_osd_linger_request *lreq)
3017 {
3018         struct ceph_osd_client *osdc = lreq->osdc;
3019         struct ceph_osd_linger_request *lookup_lreq;
3020         int ret;
3021
3022         verify_osdc_wrlocked(osdc);
3023
3024         lookup_lreq = lookup_linger_mc(&osdc->linger_map_checks,
3025                                        lreq->linger_id);
3026         if (lookup_lreq) {
3027                 WARN_ON(lookup_lreq != lreq);
3028                 return;
3029         }
3030
3031         linger_get(lreq);
3032         insert_linger_mc(&osdc->linger_map_checks, lreq);
3033         ret = ceph_monc_get_version_async(&osdc->client->monc, "osdmap",
3034                                           linger_map_check_cb, lreq->linger_id);
3035         WARN_ON(ret);
3036 }
3037
3038 static int linger_reg_commit_wait(struct ceph_osd_linger_request *lreq)
3039 {
3040         int ret;
3041
3042         dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
3043         ret = wait_for_completion_interruptible(&lreq->reg_commit_wait);
3044         return ret ?: lreq->reg_commit_error;
3045 }
3046
3047 static int linger_notify_finish_wait(struct ceph_osd_linger_request *lreq)
3048 {
3049         int ret;
3050
3051         dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
3052         ret = wait_for_completion_interruptible(&lreq->notify_finish_wait);
3053         return ret ?: lreq->notify_finish_error;
3054 }
3055
3056 /*
3057  * Timeout callback, called every N seconds.  When 1 or more OSD
3058  * requests has been active for more than N seconds, we send a keepalive
3059  * (tag + timestamp) to its OSD to ensure any communications channel
3060  * reset is detected.
3061  */
3062 static void handle_timeout(struct work_struct *work)
3063 {
3064         struct ceph_osd_client *osdc =
3065                 container_of(work, struct ceph_osd_client, timeout_work.work);
3066         struct ceph_options *opts = osdc->client->options;
3067         unsigned long cutoff = jiffies - opts->osd_keepalive_timeout;
3068         unsigned long expiry_cutoff = jiffies - opts->osd_request_timeout;
3069         LIST_HEAD(slow_osds);
3070         struct rb_node *n, *p;
3071
3072         dout("%s osdc %p\n", __func__, osdc);
3073         down_write(&osdc->lock);
3074
3075         /*
3076          * ping osds that are a bit slow.  this ensures that if there
3077          * is a break in the TCP connection we will notice, and reopen
3078          * a connection with that osd (from the fault callback).
3079          */
3080         for (n = rb_first(&osdc->osds); n; n = rb_next(n)) {
3081                 struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node);
3082                 bool found = false;
3083
3084                 for (p = rb_first(&osd->o_requests); p; ) {
3085                         struct ceph_osd_request *req =
3086                             rb_entry(p, struct ceph_osd_request, r_node);
3087
3088                         p = rb_next(p); /* abort_request() */
3089
3090                         if (time_before(req->r_stamp, cutoff)) {
3091                                 dout(" req %p tid %llu on osd%d is laggy\n",
3092                                      req, req->r_tid, osd->o_osd);
3093                                 found = true;
3094                         }
3095                         if (opts->osd_request_timeout &&
3096                             time_before(req->r_start_stamp, expiry_cutoff)) {
3097                                 pr_err_ratelimited("tid %llu on osd%d timeout\n",
3098                                        req->r_tid, osd->o_osd);
3099                                 abort_request(req, -ETIMEDOUT);
3100                         }
3101                 }
3102                 for (p = rb_first(&osd->o_linger_requests); p; p = rb_next(p)) {
3103                         struct ceph_osd_linger_request *lreq =
3104                             rb_entry(p, struct ceph_osd_linger_request, node);
3105
3106                         dout(" lreq %p linger_id %llu is served by osd%d\n",
3107                              lreq, lreq->linger_id, osd->o_osd);
3108                         found = true;
3109
3110                         mutex_lock(&lreq->lock);
3111                         if (lreq->is_watch && lreq->committed && !lreq->last_error)
3112                                 send_linger_ping(lreq);
3113                         mutex_unlock(&lreq->lock);
3114                 }
3115
3116                 if (found)
3117                         list_move_tail(&osd->o_keepalive_item, &slow_osds);
3118         }
3119
3120         if (opts->osd_request_timeout) {
3121                 for (p = rb_first(&osdc->homeless_osd.o_requests); p; ) {
3122                         struct ceph_osd_request *req =
3123                             rb_entry(p, struct ceph_osd_request, r_node);
3124
3125                         p = rb_next(p); /* abort_request() */
3126
3127                         if (time_before(req->r_start_stamp, expiry_cutoff)) {
3128                                 pr_err_ratelimited("tid %llu on osd%d timeout\n",
3129                                        req->r_tid, osdc->homeless_osd.o_osd);
3130                                 abort_request(req, -ETIMEDOUT);
3131                         }
3132                 }
3133         }
3134
3135         if (atomic_read(&osdc->num_homeless) || !list_empty(&slow_osds))
3136                 maybe_request_map(osdc);
3137
3138         while (!list_empty(&slow_osds)) {
3139                 struct ceph_osd *osd = list_first_entry(&slow_osds,
3140                                                         struct ceph_osd,
3141                                                         o_keepalive_item);
3142                 list_del_init(&osd->o_keepalive_item);
3143                 ceph_con_keepalive(&osd->o_con);
3144         }
3145
3146         up_write(&osdc->lock);
3147         schedule_delayed_work(&osdc->timeout_work,
3148                               osdc->client->options->osd_keepalive_timeout);
3149 }
3150
3151 static void handle_osds_timeout(struct work_struct *work)
3152 {
3153         struct ceph_osd_client *osdc =
3154                 container_of(work, struct ceph_osd_client,
3155                              osds_timeout_work.work);
3156         unsigned long delay = osdc->client->options->osd_idle_ttl / 4;
3157         struct ceph_osd *osd, *nosd;
3158
3159         dout("%s osdc %p\n", __func__, osdc);
3160         down_write(&osdc->lock);
3161         list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
3162                 if (time_before(jiffies, osd->lru_ttl))
3163                         break;
3164
3165                 WARN_ON(!RB_EMPTY_ROOT(&osd->o_requests));
3166                 WARN_ON(!RB_EMPTY_ROOT(&osd->o_linger_requests));
3167                 close_osd(osd);
3168         }
3169
3170         up_write(&osdc->lock);
3171         schedule_delayed_work(&osdc->osds_timeout_work,
3172                               round_jiffies_relative(delay));
3173 }
3174
3175 static int ceph_oloc_decode(void **p, void *end,
3176                             struct ceph_object_locator *oloc)
3177 {
3178         u8 struct_v, struct_cv;
3179         u32 len;
3180         void *struct_end;
3181         int ret = 0;
3182
3183         ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
3184         struct_v = ceph_decode_8(p);
3185         struct_cv = ceph_decode_8(p);
3186         if (struct_v < 3) {
3187                 pr_warn("got v %d < 3 cv %d of ceph_object_locator\n",
3188                         struct_v, struct_cv);
3189                 goto e_inval;
3190         }
3191         if (struct_cv > 6) {
3192                 pr_warn("got v %d cv %d > 6 of ceph_object_locator\n",
3193                         struct_v, struct_cv);
3194                 goto e_inval;
3195         }
3196         len = ceph_decode_32(p);
3197         ceph_decode_need(p, end, len, e_inval);
3198         struct_end = *p + len;
3199
3200         oloc->pool = ceph_decode_64(p);
3201         *p += 4; /* skip preferred */
3202
3203         len = ceph_decode_32(p);
3204         if (len > 0) {
3205                 pr_warn("ceph_object_locator::key is set\n");
3206                 goto e_inval;
3207         }
3208
3209         if (struct_v >= 5) {
3210                 bool changed = false;
3211
3212                 len = ceph_decode_32(p);
3213                 if (len > 0) {
3214                         ceph_decode_need(p, end, len, e_inval);
3215                         if (!oloc->pool_ns ||
3216                             ceph_compare_string(oloc->pool_ns, *p, len))
3217                                 changed = true;
3218                         *p += len;
3219                 } else {
3220                         if (oloc->pool_ns)
3221                                 changed = true;
3222                 }
3223                 if (changed) {
3224                         /* redirect changes namespace */
3225                         pr_warn("ceph_object_locator::nspace is changed\n");
3226                         goto e_inval;
3227                 }
3228         }
3229
3230         if (struct_v >= 6) {
3231                 s64 hash = ceph_decode_64(p);
3232                 if (hash != -1) {
3233                         pr_warn("ceph_object_locator::hash is set\n");
3234                         goto e_inval;
3235                 }
3236         }
3237
3238         /* skip the rest */
3239         *p = struct_end;
3240 out:
3241         return ret;
3242
3243 e_inval:
3244         ret = -EINVAL;
3245         goto out;
3246 }
3247
3248 static int ceph_redirect_decode(void **p, void *end,
3249                                 struct ceph_request_redirect *redir)
3250 {
3251         u8 struct_v, struct_cv;
3252         u32 len;
3253         void *struct_end;
3254         int ret;
3255
3256         ceph_decode_need(p, end, 1 + 1 + 4, e_inval);
3257         struct_v = ceph_decode_8(p);
3258         struct_cv = ceph_decode_8(p);
3259         if (struct_cv > 1) {
3260                 pr_warn("got v %d cv %d > 1 of ceph_request_redirect\n",
3261                         struct_v, struct_cv);
3262                 goto e_inval;
3263         }
3264         len = ceph_decode_32(p);
3265         ceph_decode_need(p, end, len, e_inval);
3266         struct_end = *p + len;
3267
3268         ret = ceph_oloc_decode(p, end, &redir->oloc);
3269         if (ret)
3270                 goto out;
3271
3272         len = ceph_decode_32(p);
3273         if (len > 0) {
3274                 pr_warn("ceph_request_redirect::object_name is set\n");
3275                 goto e_inval;
3276         }
3277
3278         len = ceph_decode_32(p);
3279         *p += len; /* skip osd_instructions */
3280
3281         /* skip the rest */
3282         *p = struct_end;
3283 out:
3284         return ret;
3285
3286 e_inval:
3287         ret = -EINVAL;
3288         goto out;
3289 }
3290
3291 struct MOSDOpReply {
3292         struct ceph_pg pgid;
3293         u64 flags;
3294         int result;
3295         u32 epoch;
3296         int num_ops;
3297         u32 outdata_len[CEPH_OSD_MAX_OPS];
3298         s32 rval[CEPH_OSD_MAX_OPS];
3299         int retry_attempt;
3300         struct ceph_eversion replay_version;
3301         u64 user_version;
3302         struct ceph_request_redirect redirect;
3303 };
3304
3305 static int decode_MOSDOpReply(const struct ceph_msg *msg, struct MOSDOpReply *m)
3306 {
3307         void *p = msg->front.iov_base;
3308         void *const end = p + msg->front.iov_len;
3309         u16 version = le16_to_cpu(msg->hdr.version);
3310         struct ceph_eversion bad_replay_version;
3311         u8 decode_redir;
3312         u32 len;
3313         int ret;
3314         int i;
3315
3316         ceph_decode_32_safe(&p, end, len, e_inval);
3317         ceph_decode_need(&p, end, len, e_inval);
3318         p += len; /* skip oid */
3319
3320         ret = ceph_decode_pgid(&p, end, &m->pgid);
3321         if (ret)
3322                 return ret;
3323
3324         ceph_decode_64_safe(&p, end, m->flags, e_inval);
3325         ceph_decode_32_safe(&p, end, m->result, e_inval);
3326         ceph_decode_need(&p, end, sizeof(bad_replay_version), e_inval);
3327         memcpy(&bad_replay_version, p, sizeof(bad_replay_version));
3328         p += sizeof(bad_replay_version);
3329         ceph_decode_32_safe(&p, end, m->epoch, e_inval);
3330
3331         ceph_decode_32_safe(&p, end, m->num_ops, e_inval);
3332         if (m->num_ops > ARRAY_SIZE(m->outdata_len))
3333                 goto e_inval;
3334
3335         ceph_decode_need(&p, end, m->num_ops * sizeof(struct ceph_osd_op),
3336                          e_inval);
3337         for (i = 0; i < m->num_ops; i++) {
3338                 struct ceph_osd_op *op = p;
3339
3340                 m->outdata_len[i] = le32_to_cpu(op->payload_len);
3341                 p += sizeof(*op);
3342         }
3343
3344         ceph_decode_32_safe(&p, end, m->retry_attempt, e_inval);
3345         for (i = 0; i < m->num_ops; i++)
3346                 ceph_decode_32_safe(&p, end, m->rval[i], e_inval);
3347
3348         if (version >= 5) {
3349                 ceph_decode_need(&p, end, sizeof(m->replay_version), e_inval);
3350                 memcpy(&m->replay_version, p, sizeof(m->replay_version));
3351                 p += sizeof(m->replay_version);
3352                 ceph_decode_64_safe(&p, end, m->user_version, e_inval);
3353         } else {
3354                 m->replay_version = bad_replay_version; /* struct */
3355                 m->user_version = le64_to_cpu(m->replay_version.version);
3356         }
3357
3358         if (version >= 6) {
3359                 if (version >= 7)
3360                         ceph_decode_8_safe(&p, end, decode_redir, e_inval);
3361                 else
3362                         decode_redir = 1;
3363         } else {
3364                 decode_redir = 0;
3365         }
3366
3367         if (decode_redir) {
3368                 ret = ceph_redirect_decode(&p, end, &m->redirect);
3369                 if (ret)
3370                         return ret;
3371         } else {
3372                 ceph_oloc_init(&m->redirect.oloc);
3373         }
3374
3375         return 0;
3376
3377 e_inval:
3378         return -EINVAL;
3379 }
3380
3381 /*
3382  * Handle MOSDOpReply.  Set ->r_result and call the callback if it is
3383  * specified.
3384  */
3385 static void handle_reply(struct ceph_osd *osd, struct ceph_msg *msg)
3386 {
3387         struct ceph_osd_client *osdc = osd->o_osdc;
3388         struct ceph_osd_request *req;
3389         struct MOSDOpReply m;
3390         u64 tid = le64_to_cpu(msg->hdr.tid);
3391         u32 data_len = 0;
3392         int ret;
3393         int i;
3394
3395         dout("%s msg %p tid %llu\n", __func__, msg, tid);
3396
3397         down_read(&osdc->lock);
3398         if (!osd_registered(osd)) {
3399                 dout("%s osd%d unknown\n", __func__, osd->o_osd);
3400                 goto out_unlock_osdc;
3401         }
3402         WARN_ON(osd->o_osd != le64_to_cpu(msg->hdr.src.num));
3403
3404         mutex_lock(&osd->lock);
3405         req = lookup_request(&osd->o_requests, tid);
3406         if (!req) {
3407                 dout("%s osd%d tid %llu unknown\n", __func__, osd->o_osd, tid);
3408                 goto out_unlock_session;
3409         }
3410
3411         m.redirect.oloc.pool_ns = req->r_t.target_oloc.pool_ns;
3412         ret = decode_MOSDOpReply(msg, &m);
3413         m.redirect.oloc.pool_ns = NULL;
3414         if (ret) {
3415                 pr_err("failed to decode MOSDOpReply for tid %llu: %d\n",
3416                        req->r_tid, ret);
3417                 ceph_msg_dump(msg);
3418                 goto fail_request;
3419         }
3420         dout("%s req %p tid %llu flags 0x%llx pgid %llu.%x epoch %u attempt %d v %u'%llu uv %llu\n",
3421              __func__, req, req->r_tid, m.flags, m.pgid.pool, m.pgid.seed,
3422              m.epoch, m.retry_attempt, le32_to_cpu(m.replay_version.epoch),
3423              le64_to_cpu(m.replay_version.version), m.user_version);
3424
3425         if (m.retry_attempt >= 0) {
3426                 if (m.retry_attempt != req->r_attempts - 1) {
3427                         dout("req %p tid %llu retry_attempt %d != %d, ignoring\n",
3428                              req, req->r_tid, m.retry_attempt,
3429                              req->r_attempts - 1);
3430                         goto out_unlock_session;
3431                 }
3432         } else {
3433                 WARN_ON(1); /* MOSDOpReply v4 is assumed */
3434         }
3435
3436         if (!ceph_oloc_empty(&m.redirect.oloc)) {
3437                 dout("req %p tid %llu redirect pool %lld\n", req, req->r_tid,
3438                      m.redirect.oloc.pool);
3439                 unlink_request(osd, req);
3440                 mutex_unlock(&osd->lock);
3441
3442                 /*
3443                  * Not ceph_oloc_copy() - changing pool_ns is not
3444                  * supported.
3445                  */
3446                 req->r_t.target_oloc.pool = m.redirect.oloc.pool;
3447                 req->r_flags |= CEPH_OSD_FLAG_REDIRECTED;
3448                 req->r_tid = 0;
3449                 __submit_request(req, false);
3450                 goto out_unlock_osdc;
3451         }
3452
3453         if (m.num_ops != req->r_num_ops) {
3454                 pr_err("num_ops %d != %d for tid %llu\n", m.num_ops,
3455                        req->r_num_ops, req->r_tid);
3456                 goto fail_request;
3457         }
3458         for (i = 0; i < req->r_num_ops; i++) {
3459                 dout(" req %p tid %llu op %d rval %d len %u\n", req,
3460                      req->r_tid, i, m.rval[i], m.outdata_len[i]);
3461                 req->r_ops[i].rval = m.rval[i];
3462                 req->r_ops[i].outdata_len = m.outdata_len[i];
3463                 data_len += m.outdata_len[i];
3464         }
3465         if (data_len != le32_to_cpu(msg->hdr.data_len)) {
3466                 pr_err("sum of lens %u != %u for tid %llu\n", data_len,
3467                        le32_to_cpu(msg->hdr.data_len), req->r_tid);
3468                 goto fail_request;
3469         }
3470         dout("%s req %p tid %llu result %d data_len %u\n", __func__,
3471              req, req->r_tid, m.result, data_len);
3472
3473         /*
3474          * Since we only ever request ONDISK, we should only ever get
3475          * one (type of) reply back.
3476          */
3477         WARN_ON(!(m.flags & CEPH_OSD_FLAG_ONDISK));
3478         req->r_result = m.result ?: data_len;
3479         finish_request(req);
3480         mutex_unlock(&osd->lock);
3481         up_read(&osdc->lock);
3482
3483         __complete_request(req);
3484         complete_all(&req->r_completion);
3485         ceph_osdc_put_request(req);
3486         return;
3487
3488 fail_request:
3489         complete_request(req, -EIO);
3490 out_unlock_session:
3491         mutex_unlock(&osd->lock);
3492 out_unlock_osdc:
3493         up_read(&osdc->lock);
3494 }
3495
3496 static void set_pool_was_full(struct ceph_osd_client *osdc)
3497 {
3498         struct rb_node *n;
3499
3500         for (n = rb_first(&osdc->osdmap->pg_pools); n; n = rb_next(n)) {
3501                 struct ceph_pg_pool_info *pi =
3502                     rb_entry(n, struct ceph_pg_pool_info, node);
3503
3504                 pi->was_full = __pool_full(pi);
3505         }
3506 }
3507
3508 static bool pool_cleared_full(struct ceph_osd_client *osdc, s64 pool_id)
3509 {
3510         struct ceph_pg_pool_info *pi;
3511
3512         pi = ceph_pg_pool_by_id(osdc->osdmap, pool_id);
3513         if (!pi)
3514                 return false;
3515
3516         return pi->was_full && !__pool_full(pi);
3517 }
3518
3519 static enum calc_target_result
3520 recalc_linger_target(struct ceph_osd_linger_request *lreq)
3521 {
3522         struct ceph_osd_client *osdc = lreq->osdc;
3523         enum calc_target_result ct_res;
3524
3525         ct_res = calc_target(osdc, &lreq->t, NULL, true);
3526         if (ct_res == CALC_TARGET_NEED_RESEND) {
3527                 struct ceph_osd *osd;
3528
3529                 osd = lookup_create_osd(osdc, lreq->t.osd, true);
3530                 if (osd != lreq->osd) {
3531                         unlink_linger(lreq->osd, lreq);
3532                         link_linger(osd, lreq);
3533                 }
3534         }
3535
3536         return ct_res;
3537 }
3538
3539 /*
3540  * Requeue requests whose mapping to an OSD has changed.
3541  */
3542 static void scan_requests(struct ceph_osd *osd,
3543                           bool force_resend,
3544                           bool cleared_full,
3545                           bool check_pool_cleared_full,
3546                           struct rb_root *need_resend,
3547                           struct list_head *need_resend_linger)
3548 {
3549         struct ceph_osd_client *osdc = osd->o_osdc;
3550         struct rb_node *n;
3551         bool force_resend_writes;
3552
3553         for (n = rb_first(&osd->o_linger_requests); n; ) {
3554                 struct ceph_osd_linger_request *lreq =
3555                     rb_entry(n, struct ceph_osd_linger_request, node);
3556                 enum calc_target_result ct_res;
3557
3558                 n = rb_next(n); /* recalc_linger_target() */
3559
3560                 dout("%s lreq %p linger_id %llu\n", __func__, lreq,
3561                      lreq->linger_id);
3562                 ct_res = recalc_linger_target(lreq);
3563                 switch (ct_res) {
3564                 case CALC_TARGET_NO_ACTION:
3565                         force_resend_writes = cleared_full ||
3566                             (check_pool_cleared_full &&
3567                              pool_cleared_full(osdc, lreq->t.base_oloc.pool));
3568                         if (!force_resend && !force_resend_writes)
3569                                 break;
3570
3571                         /* fall through */
3572                 case CALC_TARGET_NEED_RESEND:
3573                         cancel_linger_map_check(lreq);
3574                         /*
3575                          * scan_requests() for the previous epoch(s)
3576                          * may have already added it to the list, since
3577                          * it's not unlinked here.
3578                          */
3579                         if (list_empty(&lreq->scan_item))
3580                                 list_add_tail(&lreq->scan_item, need_resend_linger);
3581                         break;
3582                 case CALC_TARGET_POOL_DNE:
3583                         list_del_init(&lreq->scan_item);
3584                         check_linger_pool_dne(lreq);
3585                         break;
3586                 }
3587         }
3588
3589         for (n = rb_first(&osd->o_requests); n; ) {
3590                 struct ceph_osd_request *req =
3591                     rb_entry(n, struct ceph_osd_request, r_node);
3592                 enum calc_target_result ct_res;
3593
3594                 n = rb_next(n); /* unlink_request(), check_pool_dne() */
3595
3596                 dout("%s req %p tid %llu\n", __func__, req, req->r_tid);
3597                 ct_res = calc_target(osdc, &req->r_t, &req->r_osd->o_con,
3598                                      false);
3599                 switch (ct_res) {
3600                 case CALC_TARGET_NO_ACTION:
3601                         force_resend_writes = cleared_full ||
3602                             (check_pool_cleared_full &&
3603                              pool_cleared_full(osdc, req->r_t.base_oloc.pool));
3604                         if (!force_resend &&
3605                             (!(req->r_flags & CEPH_OSD_FLAG_WRITE) ||
3606                              !force_resend_writes))
3607                                 break;
3608
3609                         /* fall through */
3610                 case CALC_TARGET_NEED_RESEND:
3611                         cancel_map_check(req);
3612                         unlink_request(osd, req);
3613                         insert_request(need_resend, req);
3614                         break;
3615                 case CALC_TARGET_POOL_DNE:
3616                         check_pool_dne(req);
3617                         break;
3618                 }
3619         }
3620 }
3621
3622 static int handle_one_map(struct ceph_osd_client *osdc,
3623                           void *p, void *end, bool incremental,
3624                           struct rb_root *need_resend,
3625                           struct list_head *need_resend_linger)
3626 {
3627         struct ceph_osdmap *newmap;
3628         struct rb_node *n;
3629         bool skipped_map = false;
3630         bool was_full;
3631
3632         was_full = ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL);
3633         set_pool_was_full(osdc);
3634
3635         if (incremental)
3636                 newmap = osdmap_apply_incremental(&p, end, osdc->osdmap);
3637         else
3638                 newmap = ceph_osdmap_decode(&p, end);
3639         if (IS_ERR(newmap))
3640                 return PTR_ERR(newmap);
3641
3642         if (newmap != osdc->osdmap) {
3643                 /*
3644                  * Preserve ->was_full before destroying the old map.
3645                  * For pools that weren't in the old map, ->was_full
3646                  * should be false.
3647                  */
3648                 for (n = rb_first(&newmap->pg_pools); n; n = rb_next(n)) {
3649                         struct ceph_pg_pool_info *pi =
3650                             rb_entry(n, struct ceph_pg_pool_info, node);
3651                         struct ceph_pg_pool_info *old_pi;
3652
3653                         old_pi = ceph_pg_pool_by_id(osdc->osdmap, pi->id);
3654                         if (old_pi)
3655                                 pi->was_full = old_pi->was_full;
3656                         else
3657                                 WARN_ON(pi->was_full);
3658                 }
3659
3660                 if (osdc->osdmap->epoch &&
3661                     osdc->osdmap->epoch + 1 < newmap->epoch) {
3662                         WARN_ON(incremental);
3663                         skipped_map = true;
3664                 }
3665
3666                 ceph_osdmap_destroy(osdc->osdmap);
3667                 osdc->osdmap = newmap;
3668         }
3669
3670         was_full &= !ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL);
3671         scan_requests(&osdc->homeless_osd, skipped_map, was_full, true,
3672                       need_resend, need_resend_linger);
3673
3674         for (n = rb_first(&osdc->osds); n; ) {
3675                 struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node);
3676
3677                 n = rb_next(n); /* close_osd() */
3678
3679                 scan_requests(osd, skipped_map, was_full, true, need_resend,
3680                               need_resend_linger);
3681                 if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
3682                     memcmp(&osd->o_con.peer_addr,
3683                            ceph_osd_addr(osdc->osdmap, osd->o_osd),
3684                            sizeof(struct ceph_entity_addr)))
3685                         close_osd(osd);
3686         }
3687
3688         return 0;
3689 }
3690
3691 static void kick_requests(struct ceph_osd_client *osdc,
3692                           struct rb_root *need_resend,
3693                           struct list_head *need_resend_linger)
3694 {
3695         struct ceph_osd_linger_request *lreq, *nlreq;
3696         enum calc_target_result ct_res;
3697         struct rb_node *n;
3698
3699         /* make sure need_resend targets reflect latest map */
3700         for (n = rb_first(need_resend); n; ) {
3701                 struct ceph_osd_request *req =
3702                     rb_entry(n, struct ceph_osd_request, r_node);
3703
3704                 n = rb_next(n);
3705
3706                 if (req->r_t.epoch < osdc->osdmap->epoch) {
3707                         ct_res = calc_target(osdc, &req->r_t, NULL, false);
3708                         if (ct_res == CALC_TARGET_POOL_DNE) {
3709                                 erase_request(need_resend, req);
3710                                 check_pool_dne(req);
3711                         }
3712                 }
3713         }
3714
3715         for (n = rb_first(need_resend); n; ) {
3716                 struct ceph_osd_request *req =
3717                     rb_entry(n, struct ceph_osd_request, r_node);
3718                 struct ceph_osd *osd;
3719
3720                 n = rb_next(n);
3721                 erase_request(need_resend, req); /* before link_request() */
3722
3723                 osd = lookup_create_osd(osdc, req->r_t.osd, true);
3724                 link_request(osd, req);
3725                 if (!req->r_linger) {
3726                         if (!osd_homeless(osd) && !req->r_t.paused)
3727                                 send_request(req);
3728                 } else {
3729                         cancel_linger_request(req);
3730                 }
3731         }
3732
3733         list_for_each_entry_safe(lreq, nlreq, need_resend_linger, scan_item) {
3734                 if (!osd_homeless(lreq->osd))
3735                         send_linger(lreq);
3736
3737                 list_del_init(&lreq->scan_item);
3738         }
3739 }
3740
3741 /*
3742  * Process updated osd map.
3743  *
3744  * The message contains any number of incremental and full maps, normally
3745  * indicating some sort of topology change in the cluster.  Kick requests
3746  * off to different OSDs as needed.
3747  */
3748 void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
3749 {
3750         void *p = msg->front.iov_base;
3751         void *const end = p + msg->front.iov_len;
3752         u32 nr_maps, maplen;
3753         u32 epoch;
3754         struct ceph_fsid fsid;
3755         struct rb_root need_resend = RB_ROOT;
3756         LIST_HEAD(need_resend_linger);
3757         bool handled_incremental = false;
3758         bool was_pauserd, was_pausewr;
3759         bool pauserd, pausewr;
3760         int err;
3761
3762         dout("%s have %u\n", __func__, osdc->osdmap->epoch);
3763         down_write(&osdc->lock);
3764
3765         /* verify fsid */
3766         ceph_decode_need(&p, end, sizeof(fsid), bad);
3767         ceph_decode_copy(&p, &fsid, sizeof(fsid));
3768         if (ceph_check_fsid(osdc->client, &fsid) < 0)
3769                 goto bad;
3770
3771         was_pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD);
3772         was_pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) ||
3773                       ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
3774                       have_pool_full(osdc);
3775
3776         /* incremental maps */
3777         ceph_decode_32_safe(&p, end, nr_maps, bad);
3778         dout(" %d inc maps\n", nr_maps);
3779         while (nr_maps > 0) {
3780                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
3781                 epoch = ceph_decode_32(&p);
3782                 maplen = ceph_decode_32(&p);
3783                 ceph_decode_need(&p, end, maplen, bad);
3784                 if (osdc->osdmap->epoch &&
3785                     osdc->osdmap->epoch + 1 == epoch) {
3786                         dout("applying incremental map %u len %d\n",
3787                              epoch, maplen);
3788                         err = handle_one_map(osdc, p, p + maplen, true,
3789                                              &need_resend, &need_resend_linger);
3790                         if (err)
3791                                 goto bad;
3792                         handled_incremental = true;
3793                 } else {
3794                         dout("ignoring incremental map %u len %d\n",
3795                              epoch, maplen);
3796                 }
3797                 p += maplen;
3798                 nr_maps--;
3799         }
3800         if (handled_incremental)
3801                 goto done;
3802
3803         /* full maps */
3804         ceph_decode_32_safe(&p, end, nr_maps, bad);
3805         dout(" %d full maps\n", nr_maps);
3806         while (nr_maps) {
3807                 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
3808                 epoch = ceph_decode_32(&p);
3809                 maplen = ceph_decode_32(&p);
3810                 ceph_decode_need(&p, end, maplen, bad);
3811                 if (nr_maps > 1) {
3812                         dout("skipping non-latest full map %u len %d\n",
3813                              epoch, maplen);
3814                 } else if (osdc->osdmap->epoch >= epoch) {
3815                         dout("skipping full map %u len %d, "
3816                              "older than our %u\n", epoch, maplen,
3817                              osdc->osdmap->epoch);
3818                 } else {
3819                         dout("taking full map %u len %d\n", epoch, maplen);
3820                         err = handle_one_map(osdc, p, p + maplen, false,
3821                                              &need_resend, &need_resend_linger);
3822                         if (err)
3823                                 goto bad;
3824                 }
3825                 p += maplen;
3826                 nr_maps--;
3827         }
3828
3829 done:
3830         /*
3831          * subscribe to subsequent osdmap updates if full to ensure
3832          * we find out when we are no longer full and stop returning
3833          * ENOSPC.
3834          */
3835         pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD);
3836         pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) ||
3837                   ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) ||
3838                   have_pool_full(osdc);
3839         if (was_pauserd || was_pausewr || pauserd || pausewr ||
3840             osdc->osdmap->epoch < osdc->epoch_barrier)
3841                 maybe_request_map(osdc);
3842
3843         kick_requests(osdc, &need_resend, &need_resend_linger);
3844
3845         ceph_osdc_abort_on_full(osdc);
3846         ceph_monc_got_map(&osdc->client->monc, CEPH_SUB_OSDMAP,
3847                           osdc->osdmap->epoch);
3848         up_write(&osdc->lock);
3849         wake_up_all(&osdc->client->auth_wq);
3850         return;
3851
3852 bad:
3853         pr_err("osdc handle_map corrupt msg\n");
3854         ceph_msg_dump(msg);
3855         up_write(&osdc->lock);
3856 }
3857
3858 /*
3859  * Resubmit requests pending on the given osd.
3860  */
3861 static void kick_osd_requests(struct ceph_osd *osd)
3862 {
3863         struct rb_node *n;
3864
3865         clear_backoffs(osd);
3866
3867         for (n = rb_first(&osd->o_requests); n; ) {
3868                 struct ceph_osd_request *req =
3869                     rb_entry(n, struct ceph_osd_request, r_node);
3870
3871                 n = rb_next(n); /* cancel_linger_request() */
3872
3873                 if (!req->r_linger) {
3874                         if (!req->r_t.paused)
3875                                 send_request(req);
3876                 } else {
3877                         cancel_linger_request(req);
3878                 }
3879         }
3880         for (n = rb_first(&osd->o_linger_requests); n; n = rb_next(n)) {
3881                 struct ceph_osd_linger_request *lreq =
3882                     rb_entry(n, struct ceph_osd_linger_request, node);
3883
3884                 send_linger(lreq);
3885         }
3886 }
3887
3888 /*
3889  * If the osd connection drops, we need to resubmit all requests.
3890  */
3891 static void osd_fault(struct ceph_connection *con)
3892 {
3893         struct ceph_osd *osd = con->private;
3894         struct ceph_osd_client *osdc = osd->o_osdc;
3895
3896         dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd);
3897
3898         down_write(&osdc->lock);
3899         if (!osd_registered(osd)) {
3900                 dout("%s osd%d unknown\n", __func__, osd->o_osd);
3901                 goto out_unlock;
3902         }
3903
3904         if (!reopen_osd(osd))
3905                 kick_osd_requests(osd);
3906         maybe_request_map(osdc);
3907
3908 out_unlock:
3909         up_write(&osdc->lock);
3910 }
3911
3912 struct MOSDBackoff {
3913         struct ceph_spg spgid;
3914         u32 map_epoch;
3915         u8 op;
3916         u64 id;
3917         struct ceph_hobject_id *begin;
3918         struct ceph_hobject_id *end;
3919 };
3920
3921 static int decode_MOSDBackoff(const struct ceph_msg *msg, struct MOSDBackoff *m)
3922 {
3923         void *p = msg->front.iov_base;
3924         void *const end = p + msg->front.iov_len;
3925         u8 struct_v;
3926         u32 struct_len;
3927         int ret;
3928
3929         ret = ceph_start_decoding(&p, end, 1, "spg_t", &struct_v, &struct_len);
3930         if (ret)
3931                 return ret;
3932
3933         ret = ceph_decode_pgid(&p, end, &m->spgid.pgid);
3934         if (ret)
3935                 return ret;
3936
3937         ceph_decode_8_safe(&p, end, m->spgid.shard, e_inval);
3938         ceph_decode_32_safe(&p, end, m->map_epoch, e_inval);
3939         ceph_decode_8_safe(&p, end, m->op, e_inval);
3940         ceph_decode_64_safe(&p, end, m->id, e_inval);
3941
3942         m->begin = kzalloc(sizeof(*m->begin), GFP_NOIO);
3943         if (!m->begin)
3944                 return -ENOMEM;
3945
3946         ret = decode_hoid(&p, end, m->begin);
3947         if (ret) {
3948                 free_hoid(m->begin);
3949                 return ret;
3950         }
3951
3952         m->end = kzalloc(sizeof(*m->end), GFP_NOIO);
3953         if (!m->end) {
3954                 free_hoid(m->begin);
3955                 return -ENOMEM;
3956         }
3957
3958         ret = decode_hoid(&p, end, m->end);
3959         if (ret) {
3960                 free_hoid(m->begin);
3961                 free_hoid(m->end);
3962                 return ret;
3963         }
3964
3965         return 0;
3966
3967 e_inval:
3968         return -EINVAL;
3969 }
3970
3971 static struct ceph_msg *create_backoff_message(
3972                                 const struct ceph_osd_backoff *backoff,
3973                                 u32 map_epoch)
3974 {
3975         struct ceph_msg *msg;
3976         void *p, *end;
3977         int msg_size;
3978
3979         msg_size = CEPH_ENCODING_START_BLK_LEN +
3980                         CEPH_PGID_ENCODING_LEN + 1; /* spgid */
3981         msg_size += 4 + 1 + 8; /* map_epoch, op, id */
3982         msg_size += CEPH_ENCODING_START_BLK_LEN +
3983                         hoid_encoding_size(backoff->begin);
3984         msg_size += CEPH_ENCODING_START_BLK_LEN +
3985                         hoid_encoding_size(backoff->end);
3986
3987         msg = ceph_msg_new(CEPH_MSG_OSD_BACKOFF, msg_size, GFP_NOIO, true);
3988         if (!msg)
3989                 return NULL;
3990
3991         p = msg->front.iov_base;
3992         end = p + msg->front_alloc_len;
3993
3994         encode_spgid(&p, &backoff->spgid);
3995         ceph_encode_32(&p, map_epoch);
3996         ceph_encode_8(&p, CEPH_OSD_BACKOFF_OP_ACK_BLOCK);
3997         ceph_encode_64(&p, backoff->id);
3998         encode_hoid(&p, end, backoff->begin);
3999         encode_hoid(&p, end, backoff->end);
4000         BUG_ON(p != end);
4001
4002         msg->front.iov_len = p - msg->front.iov_base;
4003         msg->hdr.version = cpu_to_le16(1); /* MOSDBackoff v1 */
4004         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
4005
4006         return msg;
4007 }
4008
4009 static void handle_backoff_block(struct ceph_osd *osd, struct MOSDBackoff *m)
4010 {
4011         struct ceph_spg_mapping *spg;
4012         struct ceph_osd_backoff *backoff;
4013         struct ceph_msg *msg;
4014
4015         dout("%s osd%d spgid %llu.%xs%d id %llu\n", __func__, osd->o_osd,
4016              m->spgid.pgid.pool, m->spgid.pgid.seed, m->spgid.shard, m->id);
4017
4018         spg = lookup_spg_mapping(&osd->o_backoff_mappings, &m->spgid);
4019         if (!spg) {
4020                 spg = alloc_spg_mapping();
4021                 if (!spg) {
4022                         pr_err("%s failed to allocate spg\n", __func__);
4023                         return;
4024                 }
4025                 spg->spgid = m->spgid; /* struct */
4026                 insert_spg_mapping(&osd->o_backoff_mappings, spg);
4027         }
4028
4029         backoff = alloc_backoff();
4030         if (!backoff) {
4031                 pr_err("%s failed to allocate backoff\n", __func__);
4032                 return;
4033         }
4034         backoff->spgid = m->spgid; /* struct */
4035         backoff->id = m->id;
4036         backoff->begin = m->begin;
4037         m->begin = NULL; /* backoff now owns this */
4038         backoff->end = m->end;
4039         m->end = NULL;   /* ditto */
4040
4041         insert_backoff(&spg->backoffs, backoff);
4042         insert_backoff_by_id(&osd->o_backoffs_by_id, backoff);
4043
4044         /*
4045          * Ack with original backoff's epoch so that the OSD can
4046          * discard this if there was a PG split.
4047          */
4048         msg = create_backoff_message(backoff, m->map_epoch);
4049         if (!msg) {
4050                 pr_err("%s failed to allocate msg\n", __func__);
4051                 return;
4052         }
4053         ceph_con_send(&osd->o_con, msg);
4054 }
4055
4056 static bool target_contained_by(const struct ceph_osd_request_target *t,
4057                                 const struct ceph_hobject_id *begin,
4058                                 const struct ceph_hobject_id *end)
4059 {
4060         struct ceph_hobject_id hoid;
4061         int cmp;
4062
4063         hoid_fill_from_target(&hoid, t);
4064         cmp = hoid_compare(&hoid, begin);
4065         return !cmp || (cmp > 0 && hoid_compare(&hoid, end) < 0);
4066 }
4067
4068 static void handle_backoff_unblock(struct ceph_osd *osd,
4069                                    const struct MOSDBackoff *m)
4070 {
4071         struct ceph_spg_mapping *spg;
4072         struct ceph_osd_backoff *backoff;
4073         struct rb_node *n;
4074
4075         dout("%s osd%d spgid %llu.%xs%d id %llu\n", __func__, osd->o_osd,
4076              m->spgid.pgid.pool, m->spgid.pgid.seed, m->spgid.shard, m->id);
4077
4078         backoff = lookup_backoff_by_id(&osd->o_backoffs_by_id, m->id);
4079         if (!backoff) {
4080                 pr_err("%s osd%d spgid %llu.%xs%d id %llu backoff dne\n",
4081                        __func__, osd->o_osd, m->spgid.pgid.pool,
4082                        m->spgid.pgid.seed, m->spgid.shard, m->id);
4083                 return;
4084         }
4085
4086         if (hoid_compare(backoff->begin, m->begin) &&
4087             hoid_compare(backoff->end, m->end)) {
4088                 pr_err("%s osd%d spgid %llu.%xs%d id %llu bad range?\n",
4089                        __func__, osd->o_osd, m->spgid.pgid.pool,
4090                        m->spgid.pgid.seed, m->spgid.shard, m->id);
4091                 /* unblock it anyway... */
4092         }
4093
4094         spg = lookup_spg_mapping(&osd->o_backoff_mappings, &backoff->spgid);
4095         BUG_ON(!spg);
4096
4097         erase_backoff(&spg->backoffs, backoff);
4098         erase_backoff_by_id(&osd->o_backoffs_by_id, backoff);
4099         free_backoff(backoff);
4100
4101         if (RB_EMPTY_ROOT(&spg->backoffs)) {
4102                 erase_spg_mapping(&osd->o_backoff_mappings, spg);
4103                 free_spg_mapping(spg);
4104         }
4105
4106         for (n = rb_first(&osd->o_requests); n; n = rb_next(n)) {
4107                 struct ceph_osd_request *req =
4108                     rb_entry(n, struct ceph_osd_request, r_node);
4109
4110                 if (!ceph_spg_compare(&req->r_t.spgid, &m->spgid)) {
4111                         /*
4112                          * Match against @m, not @backoff -- the PG may
4113                          * have split on the OSD.
4114                          */
4115                         if (target_contained_by(&req->r_t, m->begin, m->end)) {
4116                                 /*
4117                                  * If no other installed backoff applies,
4118                                  * resend.
4119                                  */
4120                                 send_request(req);
4121                         }
4122                 }
4123         }
4124 }
4125
4126 static void handle_backoff(struct ceph_osd *osd, struct ceph_msg *msg)
4127 {
4128         struct ceph_osd_client *osdc = osd->o_osdc;
4129         struct MOSDBackoff m;
4130         int ret;
4131
4132         down_read(&osdc->lock);
4133         if (!osd_registered(osd)) {
4134                 dout("%s osd%d unknown\n", __func__, osd->o_osd);
4135                 up_read(&osdc->lock);
4136                 return;
4137         }
4138         WARN_ON(osd->o_osd != le64_to_cpu(msg->hdr.src.num));
4139
4140         mutex_lock(&osd->lock);
4141         ret = decode_MOSDBackoff(msg, &m);
4142         if (ret) {
4143                 pr_err("failed to decode MOSDBackoff: %d\n", ret);
4144                 ceph_msg_dump(msg);
4145                 goto out_unlock;
4146         }
4147
4148         switch (m.op) {
4149         case CEPH_OSD_BACKOFF_OP_BLOCK:
4150                 handle_backoff_block(osd, &m);
4151                 break;
4152         case CEPH_OSD_BACKOFF_OP_UNBLOCK:
4153                 handle_backoff_unblock(osd, &m);
4154                 break;
4155         default:
4156                 pr_err("%s osd%d unknown op %d\n", __func__, osd->o_osd, m.op);
4157         }
4158
4159         free_hoid(m.begin);
4160         free_hoid(m.end);
4161
4162 out_unlock:
4163         mutex_unlock(&osd->lock);
4164         up_read(&osdc->lock);
4165 }
4166
4167 /*
4168  * Process osd watch notifications
4169  */
4170 static void handle_watch_notify(struct ceph_osd_client *osdc,
4171                                 struct ceph_msg *msg)
4172 {
4173         void *p = msg->front.iov_base;
4174         void *const end = p + msg->front.iov_len;
4175         struct ceph_osd_linger_request *lreq;
4176         struct linger_work *lwork;
4177         u8 proto_ver, opcode;
4178         u64 cookie, notify_id;
4179         u64 notifier_id = 0;
4180         s32 return_code = 0;
4181         void *payload = NULL;
4182         u32 payload_len = 0;
4183
4184         ceph_decode_8_safe(&p, end, proto_ver, bad);
4185         ceph_decode_8_safe(&p, end, opcode, bad);
4186         ceph_decode_64_safe(&p, end, cookie, bad);
4187         p += 8; /* skip ver */
4188         ceph_decode_64_safe(&p, end, notify_id, bad);
4189
4190         if (proto_ver >= 1) {
4191                 ceph_decode_32_safe(&p, end, payload_len, bad);
4192                 ceph_decode_need(&p, end, payload_len, bad);
4193                 payload = p;
4194                 p += payload_len;
4195         }
4196
4197         if (le16_to_cpu(msg->hdr.version) >= 2)
4198                 ceph_decode_32_safe(&p, end, return_code, bad);
4199
4200         if (le16_to_cpu(msg->hdr.version) >= 3)
4201                 ceph_decode_64_safe(&p, end, notifier_id, bad);
4202
4203         down_read(&osdc->lock);
4204         lreq = lookup_linger_osdc(&osdc->linger_requests, cookie);
4205         if (!lreq) {
4206                 dout("%s opcode %d cookie %llu dne\n", __func__, opcode,
4207                      cookie);
4208                 goto out_unlock_osdc;
4209         }
4210
4211         mutex_lock(&lreq->lock);
4212         dout("%s opcode %d cookie %llu lreq %p is_watch %d\n", __func__,
4213              opcode, cookie, lreq, lreq->is_watch);
4214         if (opcode == CEPH_WATCH_EVENT_DISCONNECT) {
4215                 if (!lreq->last_error) {
4216                         lreq->last_error = -ENOTCONN;
4217                         queue_watch_error(lreq);
4218                 }
4219         } else if (!lreq->is_watch) {
4220                 /* CEPH_WATCH_EVENT_NOTIFY_COMPLETE */
4221                 if (lreq->notify_id && lreq->notify_id != notify_id) {
4222                         dout("lreq %p notify_id %llu != %llu, ignoring\n", lreq,
4223                              lreq->notify_id, notify_id);
4224                 } else if (!completion_done(&lreq->notify_finish_wait)) {
4225                         struct ceph_msg_data *data =
4226                             list_first_entry_or_null(&msg->data,
4227                                                      struct ceph_msg_data,
4228                                                      links);
4229
4230                         if (data) {
4231                                 if (lreq->preply_pages) {
4232                                         WARN_ON(data->type !=
4233                                                         CEPH_MSG_DATA_PAGES);
4234                                         *lreq->preply_pages = data->pages;
4235                                         *lreq->preply_len = data->length;
4236                                 } else {
4237                                         ceph_release_page_vector(data->pages,
4238                                                calc_pages_for(0, data->length));
4239                                 }
4240                         }
4241                         lreq->notify_finish_error = return_code;
4242                         complete_all(&lreq->notify_finish_wait);
4243                 }
4244         } else {
4245                 /* CEPH_WATCH_EVENT_NOTIFY */
4246                 lwork = lwork_alloc(lreq, do_watch_notify);
4247                 if (!lwork) {
4248                         pr_err("failed to allocate notify-lwork\n");
4249                         goto out_unlock_lreq;
4250                 }
4251
4252                 lwork->notify.notify_id = notify_id;
4253                 lwork->notify.notifier_id = notifier_id;
4254                 lwork->notify.payload = payload;
4255                 lwork->notify.payload_len = payload_len;
4256                 lwork->notify.msg = ceph_msg_get(msg);
4257                 lwork_queue(lwork);
4258         }
4259
4260 out_unlock_lreq:
4261         mutex_unlock(&lreq->lock);
4262 out_unlock_osdc:
4263         up_read(&osdc->lock);
4264         return;
4265
4266 bad:
4267         pr_err("osdc handle_watch_notify corrupt msg\n");
4268 }
4269
4270 /*
4271  * Register request, send initial attempt.
4272  */
4273 int ceph_osdc_start_request(struct ceph_osd_client *osdc,
4274                             struct ceph_osd_request *req,
4275                             bool nofail)
4276 {
4277         down_read(&osdc->lock);
4278         submit_request(req, false);
4279         up_read(&osdc->lock);
4280
4281         return 0;
4282 }
4283 EXPORT_SYMBOL(ceph_osdc_start_request);
4284
4285 /*
4286  * Unregister a registered request.  The request is not completed:
4287  * ->r_result isn't set and __complete_request() isn't called.
4288  */
4289 void ceph_osdc_cancel_request(struct ceph_osd_request *req)
4290 {
4291         struct ceph_osd_client *osdc = req->r_osdc;
4292
4293         down_write(&osdc->lock);
4294         if (req->r_osd)
4295                 cancel_request(req);
4296         up_write(&osdc->lock);
4297 }
4298 EXPORT_SYMBOL(ceph_osdc_cancel_request);
4299
4300 /*
4301  * @timeout: in jiffies, 0 means "wait forever"
4302  */
4303 static int wait_request_timeout(struct ceph_osd_request *req,
4304                                 unsigned long timeout)
4305 {
4306         long left;
4307
4308         dout("%s req %p tid %llu\n", __func__, req, req->r_tid);
4309         left = wait_for_completion_killable_timeout(&req->r_completion,
4310                                                 ceph_timeout_jiffies(timeout));
4311         if (left <= 0) {
4312                 left = left ?: -ETIMEDOUT;
4313                 ceph_osdc_cancel_request(req);
4314         } else {
4315                 left = req->r_result; /* completed */
4316         }
4317
4318         return left;
4319 }
4320
4321 /*
4322  * wait for a request to complete
4323  */
4324 int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
4325                            struct ceph_osd_request *req)
4326 {
4327         return wait_request_timeout(req, 0);
4328 }
4329 EXPORT_SYMBOL(ceph_osdc_wait_request);
4330
4331 /*
4332  * sync - wait for all in-flight requests to flush.  avoid starvation.
4333  */
4334 void ceph_osdc_sync(struct ceph_osd_client *osdc)
4335 {
4336         struct rb_node *n, *p;
4337         u64 last_tid = atomic64_read(&osdc->last_tid);
4338
4339 again:
4340         down_read(&osdc->lock);
4341         for (n = rb_first(&osdc->osds); n; n = rb_next(n)) {
4342                 struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node);
4343
4344                 mutex_lock(&osd->lock);
4345                 for (p = rb_first(&osd->o_requests); p; p = rb_next(p)) {
4346                         struct ceph_osd_request *req =
4347                             rb_entry(p, struct ceph_osd_request, r_node);
4348
4349                         if (req->r_tid > last_tid)
4350                                 break;
4351
4352                         if (!(req->r_flags & CEPH_OSD_FLAG_WRITE))
4353                                 continue;
4354
4355                         ceph_osdc_get_request(req);
4356                         mutex_unlock(&osd->lock);
4357                         up_read(&osdc->lock);
4358                         dout("%s waiting on req %p tid %llu last_tid %llu\n",
4359                              __func__, req, req->r_tid, last_tid);
4360                         wait_for_completion(&req->r_completion);
4361                         ceph_osdc_put_request(req);
4362                         goto again;
4363                 }
4364
4365                 mutex_unlock(&osd->lock);
4366         }
4367
4368         up_read(&osdc->lock);
4369         dout("%s done last_tid %llu\n", __func__, last_tid);
4370 }
4371 EXPORT_SYMBOL(ceph_osdc_sync);
4372
4373 static struct ceph_osd_request *
4374 alloc_linger_request(struct ceph_osd_linger_request *lreq)
4375 {
4376         struct ceph_osd_request *req;
4377
4378         req = ceph_osdc_alloc_request(lreq->osdc, NULL, 1, false, GFP_NOIO);
4379         if (!req)
4380                 return NULL;
4381
4382         ceph_oid_copy(&req->r_base_oid, &lreq->t.base_oid);
4383         ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
4384
4385         if (ceph_osdc_alloc_messages(req, GFP_NOIO)) {
4386                 ceph_osdc_put_request(req);
4387                 return NULL;
4388         }
4389
4390         return req;
4391 }
4392
4393 /*
4394  * Returns a handle, caller owns a ref.
4395  */
4396 struct ceph_osd_linger_request *
4397 ceph_osdc_watch(struct ceph_osd_client *osdc,
4398                 struct ceph_object_id *oid,
4399                 struct ceph_object_locator *oloc,
4400                 rados_watchcb2_t wcb,
4401                 rados_watcherrcb_t errcb,
4402                 void *data)
4403 {
4404         struct ceph_osd_linger_request *lreq;
4405         int ret;
4406
4407         lreq = linger_alloc(osdc);
4408         if (!lreq)
4409                 return ERR_PTR(-ENOMEM);
4410
4411         lreq->is_watch = true;
4412         lreq->wcb = wcb;
4413         lreq->errcb = errcb;
4414         lreq->data = data;
4415         lreq->watch_valid_thru = jiffies;
4416
4417         ceph_oid_copy(&lreq->t.base_oid, oid);
4418         ceph_oloc_copy(&lreq->t.base_oloc, oloc);
4419         lreq->t.flags = CEPH_OSD_FLAG_WRITE;
4420         ktime_get_real_ts(&lreq->mtime);
4421
4422         lreq->reg_req = alloc_linger_request(lreq);
4423         if (!lreq->reg_req) {
4424                 ret = -ENOMEM;
4425                 goto err_put_lreq;
4426         }
4427
4428         lreq->ping_req = alloc_linger_request(lreq);
4429         if (!lreq->ping_req) {
4430                 ret = -ENOMEM;
4431                 goto err_put_lreq;
4432         }
4433
4434         down_write(&osdc->lock);
4435         linger_register(lreq); /* before osd_req_op_* */
4436         osd_req_op_watch_init(lreq->reg_req, 0, lreq->linger_id,
4437                               CEPH_OSD_WATCH_OP_WATCH);
4438         osd_req_op_watch_init(lreq->ping_req, 0, lreq->linger_id,
4439                               CEPH_OSD_WATCH_OP_PING);
4440         linger_submit(lreq);
4441         up_write(&osdc->lock);
4442
4443         ret = linger_reg_commit_wait(lreq);
4444         if (ret) {
4445                 linger_cancel(lreq);
4446                 goto err_put_lreq;
4447         }
4448
4449         return lreq;
4450
4451 err_put_lreq:
4452         linger_put(lreq);
4453         return ERR_PTR(ret);
4454 }
4455 EXPORT_SYMBOL(ceph_osdc_watch);
4456
4457 /*
4458  * Releases a ref.
4459  *
4460  * Times out after mount_timeout to preserve rbd unmap behaviour
4461  * introduced in 2894e1d76974 ("rbd: timeout watch teardown on unmap
4462  * with mount_timeout").
4463  */
4464 int ceph_osdc_unwatch(struct ceph_osd_client *osdc,
4465                       struct ceph_osd_linger_request *lreq)
4466 {
4467         struct ceph_options *opts = osdc->client->options;
4468         struct ceph_osd_request *req;
4469         int ret;
4470
4471         req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_NOIO);
4472         if (!req)
4473                 return -ENOMEM;
4474
4475         ceph_oid_copy(&req->r_base_oid, &lreq->t.base_oid);
4476         ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
4477         req->r_flags = CEPH_OSD_FLAG_WRITE;
4478         ktime_get_real_ts(&req->r_mtime);
4479         osd_req_op_watch_init(req, 0, lreq->linger_id,
4480                               CEPH_OSD_WATCH_OP_UNWATCH);
4481
4482         ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
4483         if (ret)
4484                 goto out_put_req;
4485
4486         ceph_osdc_start_request(osdc, req, false);
4487         linger_cancel(lreq);
4488         linger_put(lreq);
4489         ret = wait_request_timeout(req, opts->mount_timeout);
4490
4491 out_put_req:
4492         ceph_osdc_put_request(req);
4493         return ret;
4494 }
4495 EXPORT_SYMBOL(ceph_osdc_unwatch);
4496
4497 static int osd_req_op_notify_ack_init(struct ceph_osd_request *req, int which,
4498                                       u64 notify_id, u64 cookie, void *payload,
4499                                       size_t payload_len)
4500 {
4501         struct ceph_osd_req_op *op;
4502         struct ceph_pagelist *pl;
4503         int ret;
4504
4505         op = _osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY_ACK, 0);
4506
4507         pl = kmalloc(sizeof(*pl), GFP_NOIO);
4508         if (!pl)
4509                 return -ENOMEM;
4510
4511         ceph_pagelist_init(pl);
4512         ret = ceph_pagelist_encode_64(pl, notify_id);
4513         ret |= ceph_pagelist_encode_64(pl, cookie);
4514         if (payload) {
4515                 ret |= ceph_pagelist_encode_32(pl, payload_len);
4516                 ret |= ceph_pagelist_append(pl, payload, payload_len);
4517         } else {
4518                 ret |= ceph_pagelist_encode_32(pl, 0);
4519         }
4520         if (ret) {
4521                 ceph_pagelist_release(pl);
4522                 return -ENOMEM;
4523         }
4524
4525         ceph_osd_data_pagelist_init(&op->notify_ack.request_data, pl);
4526         op->indata_len = pl->length;
4527         return 0;
4528 }
4529
4530 int ceph_osdc_notify_ack(struct ceph_osd_client *osdc,
4531                          struct ceph_object_id *oid,
4532                          struct ceph_object_locator *oloc,
4533                          u64 notify_id,
4534                          u64 cookie,
4535                          void *payload,
4536                          size_t payload_len)
4537 {
4538         struct ceph_osd_request *req;
4539         int ret;
4540
4541         req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_NOIO);
4542         if (!req)
4543                 return -ENOMEM;
4544
4545         ceph_oid_copy(&req->r_base_oid, oid);
4546         ceph_oloc_copy(&req->r_base_oloc, oloc);
4547         req->r_flags = CEPH_OSD_FLAG_READ;
4548
4549         ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
4550         if (ret)
4551                 goto out_put_req;
4552
4553         ret = osd_req_op_notify_ack_init(req, 0, notify_id, cookie, payload,
4554                                          payload_len);
4555         if (ret)
4556                 goto out_put_req;
4557
4558         ceph_osdc_start_request(osdc, req, false);
4559         ret = ceph_osdc_wait_request(osdc, req);
4560
4561 out_put_req:
4562         ceph_osdc_put_request(req);
4563         return ret;
4564 }
4565 EXPORT_SYMBOL(ceph_osdc_notify_ack);
4566
4567 static int osd_req_op_notify_init(struct ceph_osd_request *req, int which,
4568                                   u64 cookie, u32 prot_ver, u32 timeout,
4569                                   void *payload, size_t payload_len)
4570 {
4571         struct ceph_osd_req_op *op;
4572         struct ceph_pagelist *pl;
4573         int ret;
4574
4575         op = _osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY, 0);
4576         op->notify.cookie = cookie;
4577
4578         pl = kmalloc(sizeof(*pl), GFP_NOIO);
4579         if (!pl)
4580                 return -ENOMEM;
4581
4582         ceph_pagelist_init(pl);
4583         ret = ceph_pagelist_encode_32(pl, 1); /* prot_ver */
4584         ret |= ceph_pagelist_encode_32(pl, timeout);
4585         ret |= ceph_pagelist_encode_32(pl, payload_len);
4586         ret |= ceph_pagelist_append(pl, payload, payload_len);
4587         if (ret) {
4588                 ceph_pagelist_release(pl);
4589                 return -ENOMEM;
4590         }
4591
4592         ceph_osd_data_pagelist_init(&op->notify.request_data, pl);
4593         op->indata_len = pl->length;
4594         return 0;
4595 }
4596
4597 /*
4598  * @timeout: in seconds
4599  *
4600  * @preply_{pages,len} are initialized both on success and error.
4601  * The caller is responsible for:
4602  *
4603  *     ceph_release_page_vector(reply_pages, calc_pages_for(0, reply_len))
4604  */
4605 int ceph_osdc_notify(struct ceph_osd_client *osdc,
4606                      struct ceph_object_id *oid,
4607                      struct ceph_object_locator *oloc,
4608                      void *payload,
4609                      size_t payload_len,
4610                      u32 timeout,
4611                      struct page ***preply_pages,
4612                      size_t *preply_len)
4613 {
4614         struct ceph_osd_linger_request *lreq;
4615         struct page **pages;
4616         int ret;
4617
4618         WARN_ON(!timeout);
4619         if (preply_pages) {
4620                 *preply_pages = NULL;
4621                 *preply_len = 0;
4622         }
4623
4624         lreq = linger_alloc(osdc);
4625         if (!lreq)
4626                 return -ENOMEM;
4627
4628         lreq->preply_pages = preply_pages;
4629         lreq->preply_len = preply_len;
4630
4631         ceph_oid_copy(&lreq->t.base_oid, oid);
4632         ceph_oloc_copy(&lreq->t.base_oloc, oloc);
4633         lreq->t.flags = CEPH_OSD_FLAG_READ;
4634
4635         lreq->reg_req = alloc_linger_request(lreq);
4636         if (!lreq->reg_req) {
4637                 ret = -ENOMEM;
4638                 goto out_put_lreq;
4639         }
4640
4641         /* for notify_id */
4642         pages = ceph_alloc_page_vector(1, GFP_NOIO);
4643         if (IS_ERR(pages)) {
4644                 ret = PTR_ERR(pages);
4645                 goto out_put_lreq;
4646         }
4647
4648         down_write(&osdc->lock);
4649         linger_register(lreq); /* before osd_req_op_* */
4650         ret = osd_req_op_notify_init(lreq->reg_req, 0, lreq->linger_id, 1,
4651                                      timeout, payload, payload_len);
4652         if (ret) {
4653                 linger_unregister(lreq);
4654                 up_write(&osdc->lock);
4655                 ceph_release_page_vector(pages, 1);
4656                 goto out_put_lreq;
4657         }
4658         ceph_osd_data_pages_init(osd_req_op_data(lreq->reg_req, 0, notify,
4659                                                  response_data),
4660                                  pages, PAGE_SIZE, 0, false, true);
4661         linger_submit(lreq);
4662         up_write(&osdc->lock);
4663
4664         ret = linger_reg_commit_wait(lreq);
4665         if (!ret)
4666                 ret = linger_notify_finish_wait(lreq);
4667         else
4668                 dout("lreq %p failed to initiate notify %d\n", lreq, ret);
4669
4670         linger_cancel(lreq);
4671 out_put_lreq:
4672         linger_put(lreq);
4673         return ret;
4674 }
4675 EXPORT_SYMBOL(ceph_osdc_notify);
4676
4677 /*
4678  * Return the number of milliseconds since the watch was last
4679  * confirmed, or an error.  If there is an error, the watch is no
4680  * longer valid, and should be destroyed with ceph_osdc_unwatch().
4681  */
4682 int ceph_osdc_watch_check(struct ceph_osd_client *osdc,
4683                           struct ceph_osd_linger_request *lreq)
4684 {
4685         unsigned long stamp, age;
4686         int ret;
4687
4688         down_read(&osdc->lock);
4689         mutex_lock(&lreq->lock);
4690         stamp = lreq->watch_valid_thru;
4691         if (!list_empty(&lreq->pending_lworks)) {
4692                 struct linger_work *lwork =
4693                     list_first_entry(&lreq->pending_lworks,
4694                                      struct linger_work,
4695                                      pending_item);
4696
4697                 if (time_before(lwork->queued_stamp, stamp))
4698                         stamp = lwork->queued_stamp;
4699         }
4700         age = jiffies - stamp;
4701         dout("%s lreq %p linger_id %llu age %lu last_error %d\n", __func__,
4702              lreq, lreq->linger_id, age, lreq->last_error);
4703         /* we are truncating to msecs, so return a safe upper bound */
4704         ret = lreq->last_error ?: 1 + jiffies_to_msecs(age);
4705
4706         mutex_unlock(&lreq->lock);
4707         up_read(&osdc->lock);
4708         return ret;
4709 }
4710
4711 static int decode_watcher(void **p, void *end, struct ceph_watch_item *item)
4712 {
4713         u8 struct_v;
4714         u32 struct_len;
4715         int ret;
4716
4717         ret = ceph_start_decoding(p, end, 2, "watch_item_t",
4718                                   &struct_v, &struct_len);
4719         if (ret)
4720                 return ret;
4721
4722         ceph_decode_copy(p, &item->name, sizeof(item->name));
4723         item->cookie = ceph_decode_64(p);
4724         *p += 4; /* skip timeout_seconds */
4725         if (struct_v >= 2) {
4726                 ceph_decode_copy(p, &item->addr, sizeof(item->addr));
4727                 ceph_decode_addr(&item->addr);
4728         }
4729
4730         dout("%s %s%llu cookie %llu addr %s\n", __func__,
4731              ENTITY_NAME(item->name), item->cookie,
4732              ceph_pr_addr(&item->addr.in_addr));
4733         return 0;
4734 }
4735
4736 static int decode_watchers(void **p, void *end,
4737                            struct ceph_watch_item **watchers,
4738                            u32 *num_watchers)
4739 {
4740         u8 struct_v;
4741         u32 struct_len;
4742         int i;
4743         int ret;
4744
4745         ret = ceph_start_decoding(p, end, 1, "obj_list_watch_response_t",
4746                                   &struct_v, &struct_len);
4747         if (ret)
4748                 return ret;
4749
4750         *num_watchers = ceph_decode_32(p);
4751         *watchers = kcalloc(*num_watchers, sizeof(**watchers), GFP_NOIO);
4752         if (!*watchers)
4753                 return -ENOMEM;
4754
4755         for (i = 0; i < *num_watchers; i++) {
4756                 ret = decode_watcher(p, end, *watchers + i);
4757                 if (ret) {
4758                         kfree(*watchers);
4759                         return ret;
4760                 }
4761         }
4762
4763         return 0;
4764 }
4765
4766 /*
4767  * On success, the caller is responsible for:
4768  *
4769  *     kfree(watchers);
4770  */
4771 int ceph_osdc_list_watchers(struct ceph_osd_client *osdc,
4772                             struct ceph_object_id *oid,
4773                             struct ceph_object_locator *oloc,
4774                             struct ceph_watch_item **watchers,
4775                             u32 *num_watchers)
4776 {
4777         struct ceph_osd_request *req;
4778         struct page **pages;
4779         int ret;
4780
4781         req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_NOIO);
4782         if (!req)
4783                 return -ENOMEM;
4784
4785         ceph_oid_copy(&req->r_base_oid, oid);
4786         ceph_oloc_copy(&req->r_base_oloc, oloc);
4787         req->r_flags = CEPH_OSD_FLAG_READ;
4788
4789         ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
4790         if (ret)
4791                 goto out_put_req;
4792
4793         pages = ceph_alloc_page_vector(1, GFP_NOIO);
4794         if (IS_ERR(pages)) {
4795                 ret = PTR_ERR(pages);
4796                 goto out_put_req;
4797         }
4798
4799         osd_req_op_init(req, 0, CEPH_OSD_OP_LIST_WATCHERS, 0);
4800         ceph_osd_data_pages_init(osd_req_op_data(req, 0, list_watchers,
4801                                                  response_data),
4802                                  pages, PAGE_SIZE, 0, false, true);
4803
4804         ceph_osdc_start_request(osdc, req, false);
4805         ret = ceph_osdc_wait_request(osdc, req);
4806         if (ret >= 0) {
4807                 void *p = page_address(pages[0]);
4808                 void *const end = p + req->r_ops[0].outdata_len;
4809
4810                 ret = decode_watchers(&p, end, watchers, num_watchers);
4811         }
4812
4813 out_put_req:
4814         ceph_osdc_put_request(req);
4815         return ret;
4816 }
4817 EXPORT_SYMBOL(ceph_osdc_list_watchers);
4818
4819 /*
4820  * Call all pending notify callbacks - for use after a watch is
4821  * unregistered, to make sure no more callbacks for it will be invoked
4822  */
4823 void ceph_osdc_flush_notifies(struct ceph_osd_client *osdc)
4824 {
4825         dout("%s osdc %p\n", __func__, osdc);
4826         flush_workqueue(osdc->notify_wq);
4827 }
4828 EXPORT_SYMBOL(ceph_osdc_flush_notifies);
4829
4830 void ceph_osdc_maybe_request_map(struct ceph_osd_client *osdc)
4831 {
4832         down_read(&osdc->lock);
4833         maybe_request_map(osdc);
4834         up_read(&osdc->lock);
4835 }
4836 EXPORT_SYMBOL(ceph_osdc_maybe_request_map);
4837
4838 /*
4839  * Execute an OSD class method on an object.
4840  *
4841  * @flags: CEPH_OSD_FLAG_*
4842  * @resp_len: in/out param for reply length
4843  */
4844 int ceph_osdc_call(struct ceph_osd_client *osdc,
4845                    struct ceph_object_id *oid,
4846                    struct ceph_object_locator *oloc,
4847                    const char *class, const char *method,
4848                    unsigned int flags,
4849                    struct page *req_page, size_t req_len,
4850                    struct page *resp_page, size_t *resp_len)
4851 {
4852         struct ceph_osd_request *req;
4853         int ret;
4854
4855         if (req_len > PAGE_SIZE || (resp_page && *resp_len > PAGE_SIZE))
4856                 return -E2BIG;
4857
4858         req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_NOIO);
4859         if (!req)
4860                 return -ENOMEM;
4861
4862         ceph_oid_copy(&req->r_base_oid, oid);
4863         ceph_oloc_copy(&req->r_base_oloc, oloc);
4864         req->r_flags = flags;
4865
4866         ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
4867         if (ret)
4868                 goto out_put_req;
4869
4870         osd_req_op_cls_init(req, 0, CEPH_OSD_OP_CALL, class, method);
4871         if (req_page)
4872                 osd_req_op_cls_request_data_pages(req, 0, &req_page, req_len,
4873                                                   0, false, false);
4874         if (resp_page)
4875                 osd_req_op_cls_response_data_pages(req, 0, &resp_page,
4876                                                    *resp_len, 0, false, false);
4877
4878         ceph_osdc_start_request(osdc, req, false);
4879         ret = ceph_osdc_wait_request(osdc, req);
4880         if (ret >= 0) {
4881                 ret = req->r_ops[0].rval;
4882                 if (resp_page)
4883                         *resp_len = req->r_ops[0].outdata_len;
4884         }
4885
4886 out_put_req:
4887         ceph_osdc_put_request(req);
4888         return ret;
4889 }
4890 EXPORT_SYMBOL(ceph_osdc_call);
4891
4892 /*
4893  * init, shutdown
4894  */
4895 int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
4896 {
4897         int err;
4898
4899         dout("init\n");
4900         osdc->client = client;
4901         init_rwsem(&osdc->lock);
4902         osdc->osds = RB_ROOT;
4903         INIT_LIST_HEAD(&osdc->osd_lru);
4904         spin_lock_init(&osdc->osd_lru_lock);
4905         osd_init(&osdc->homeless_osd);
4906         osdc->homeless_osd.o_osdc = osdc;
4907         osdc->homeless_osd.o_osd = CEPH_HOMELESS_OSD;
4908         osdc->last_linger_id = CEPH_LINGER_ID_START;
4909         osdc->linger_requests = RB_ROOT;
4910         osdc->map_checks = RB_ROOT;
4911         osdc->linger_map_checks = RB_ROOT;
4912         INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
4913         INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
4914
4915         err = -ENOMEM;
4916         osdc->osdmap = ceph_osdmap_alloc();
4917         if (!osdc->osdmap)
4918                 goto out;
4919
4920         osdc->req_mempool = mempool_create_slab_pool(10,
4921                                                      ceph_osd_request_cache);
4922         if (!osdc->req_mempool)
4923                 goto out_map;
4924
4925         err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
4926                                 PAGE_SIZE, 10, true, "osd_op");
4927         if (err < 0)
4928                 goto out_mempool;
4929         err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
4930                                 PAGE_SIZE, 10, true, "osd_op_reply");
4931         if (err < 0)
4932                 goto out_msgpool;
4933
4934         err = -ENOMEM;
4935         osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
4936         if (!osdc->notify_wq)
4937                 goto out_msgpool_reply;
4938
4939         schedule_delayed_work(&osdc->timeout_work,
4940                               osdc->client->options->osd_keepalive_timeout);
4941         schedule_delayed_work(&osdc->osds_timeout_work,
4942             round_jiffies_relative(osdc->client->options->osd_idle_ttl));
4943
4944         return 0;
4945
4946 out_msgpool_reply:
4947         ceph_msgpool_destroy(&osdc->msgpool_op_reply);
4948 out_msgpool:
4949         ceph_msgpool_destroy(&osdc->msgpool_op);
4950 out_mempool:
4951         mempool_destroy(osdc->req_mempool);
4952 out_map:
4953         ceph_osdmap_destroy(osdc->osdmap);
4954 out:
4955         return err;
4956 }
4957
4958 void ceph_osdc_stop(struct ceph_osd_client *osdc)
4959 {
4960         flush_workqueue(osdc->notify_wq);
4961         destroy_workqueue(osdc->notify_wq);
4962         cancel_delayed_work_sync(&osdc->timeout_work);
4963         cancel_delayed_work_sync(&osdc->osds_timeout_work);
4964
4965         down_write(&osdc->lock);
4966         while (!RB_EMPTY_ROOT(&osdc->osds)) {
4967                 struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
4968                                                 struct ceph_osd, o_node);
4969                 close_osd(osd);
4970         }
4971         up_write(&osdc->lock);
4972         WARN_ON(refcount_read(&osdc->homeless_osd.o_ref) != 1);
4973         osd_cleanup(&osdc->homeless_osd);
4974
4975         WARN_ON(!list_empty(&osdc->osd_lru));
4976         WARN_ON(!RB_EMPTY_ROOT(&osdc->linger_requests));
4977         WARN_ON(!RB_EMPTY_ROOT(&osdc->map_checks));
4978         WARN_ON(!RB_EMPTY_ROOT(&osdc->linger_map_checks));
4979         WARN_ON(atomic_read(&osdc->num_requests));
4980         WARN_ON(atomic_read(&osdc->num_homeless));
4981
4982         ceph_osdmap_destroy(osdc->osdmap);
4983         mempool_destroy(osdc->req_mempool);
4984         ceph_msgpool_destroy(&osdc->msgpool_op);
4985         ceph_msgpool_destroy(&osdc->msgpool_op_reply);
4986 }
4987
4988 /*
4989  * Read some contiguous pages.  If we cross a stripe boundary, shorten
4990  * *plen.  Return number of bytes read, or error.
4991  */
4992 int ceph_osdc_readpages(struct ceph_osd_client *osdc,
4993                         struct ceph_vino vino, struct ceph_file_layout *layout,
4994                         u64 off, u64 *plen,
4995                         u32 truncate_seq, u64 truncate_size,
4996                         struct page **pages, int num_pages, int page_align)
4997 {
4998         struct ceph_osd_request *req;
4999         int rc = 0;
5000
5001         dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
5002              vino.snap, off, *plen);
5003         req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 0, 1,
5004                                     CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
5005                                     NULL, truncate_seq, truncate_size,
5006                                     false);
5007         if (IS_ERR(req))
5008                 return PTR_ERR(req);
5009
5010         /* it may be a short read due to an object boundary */
5011         osd_req_op_extent_osd_data_pages(req, 0,
5012                                 pages, *plen, page_align, false, false);
5013
5014         dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
5015              off, *plen, *plen, page_align);
5016
5017         rc = ceph_osdc_start_request(osdc, req, false);
5018         if (!rc)
5019                 rc = ceph_osdc_wait_request(osdc, req);
5020
5021         ceph_osdc_put_request(req);
5022         dout("readpages result %d\n", rc);
5023         return rc;
5024 }
5025 EXPORT_SYMBOL(ceph_osdc_readpages);
5026
5027 /*
5028  * do a synchronous write on N pages
5029  */
5030 int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
5031                          struct ceph_file_layout *layout,
5032                          struct ceph_snap_context *snapc,
5033                          u64 off, u64 len,
5034                          u32 truncate_seq, u64 truncate_size,
5035                          struct timespec *mtime,
5036                          struct page **pages, int num_pages)
5037 {
5038         struct ceph_osd_request *req;
5039         int rc = 0;
5040         int page_align = off & ~PAGE_MASK;
5041
5042         req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 0, 1,
5043                                     CEPH_OSD_OP_WRITE, CEPH_OSD_FLAG_WRITE,
5044                                     snapc, truncate_seq, truncate_size,
5045                                     true);
5046         if (IS_ERR(req))
5047                 return PTR_ERR(req);
5048
5049         /* it may be a short write due to an object boundary */
5050         osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
5051                                 false, false);
5052         dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
5053
5054         req->r_mtime = *mtime;
5055         rc = ceph_osdc_start_request(osdc, req, true);
5056         if (!rc)
5057                 rc = ceph_osdc_wait_request(osdc, req);
5058
5059         ceph_osdc_put_request(req);
5060         if (rc == 0)
5061                 rc = len;
5062         dout("writepages result %d\n", rc);
5063         return rc;
5064 }
5065 EXPORT_SYMBOL(ceph_osdc_writepages);
5066
5067 int ceph_osdc_setup(void)
5068 {
5069         size_t size = sizeof(struct ceph_osd_request) +
5070             CEPH_OSD_SLAB_OPS * sizeof(struct ceph_osd_req_op);
5071
5072         BUG_ON(ceph_osd_request_cache);
5073         ceph_osd_request_cache = kmem_cache_create("ceph_osd_request", size,
5074                                                    0, 0, NULL);
5075
5076         return ceph_osd_request_cache ? 0 : -ENOMEM;
5077 }
5078 EXPORT_SYMBOL(ceph_osdc_setup);
5079
5080 void ceph_osdc_cleanup(void)
5081 {
5082         BUG_ON(!ceph_osd_request_cache);
5083         kmem_cache_destroy(ceph_osd_request_cache);
5084         ceph_osd_request_cache = NULL;
5085 }
5086 EXPORT_SYMBOL(ceph_osdc_cleanup);
5087
5088 /*
5089  * handle incoming message
5090  */
5091 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
5092 {
5093         struct ceph_osd *osd = con->private;
5094         struct ceph_osd_client *osdc = osd->o_osdc;
5095         int type = le16_to_cpu(msg->hdr.type);
5096
5097         switch (type) {
5098         case CEPH_MSG_OSD_MAP:
5099                 ceph_osdc_handle_map(osdc, msg);
5100                 break;
5101         case CEPH_MSG_OSD_OPREPLY:
5102                 handle_reply(osd, msg);
5103                 break;
5104         case CEPH_MSG_OSD_BACKOFF:
5105                 handle_backoff(osd, msg);
5106                 break;
5107         case CEPH_MSG_WATCH_NOTIFY:
5108                 handle_watch_notify(osdc, msg);
5109                 break;
5110
5111         default:
5112                 pr_err("received unknown message type %d %s\n", type,
5113                        ceph_msg_type_name(type));
5114         }
5115
5116         ceph_msg_put(msg);
5117 }
5118
5119 /*
5120  * Lookup and return message for incoming reply.  Don't try to do
5121  * anything about a larger than preallocated data portion of the
5122  * message at the moment - for now, just skip the message.
5123  */
5124 static struct ceph_msg *get_reply(struct ceph_connection *con,
5125                                   struct ceph_msg_header *hdr,
5126                                   int *skip)
5127 {
5128         struct ceph_osd *osd = con->private;
5129         struct ceph_osd_client *osdc = osd->o_osdc;
5130         struct ceph_msg *m = NULL;
5131         struct ceph_osd_request *req;
5132         int front_len = le32_to_cpu(hdr->front_len);
5133         int data_len = le32_to_cpu(hdr->data_len);
5134         u64 tid = le64_to_cpu(hdr->tid);
5135
5136         down_read(&osdc->lock);
5137         if (!osd_registered(osd)) {
5138                 dout("%s osd%d unknown, skipping\n", __func__, osd->o_osd);
5139                 *skip = 1;
5140                 goto out_unlock_osdc;
5141         }
5142         WARN_ON(osd->o_osd != le64_to_cpu(hdr->src.num));
5143
5144         mutex_lock(&osd->lock);
5145         req = lookup_request(&osd->o_requests, tid);
5146         if (!req) {
5147                 dout("%s osd%d tid %llu unknown, skipping\n", __func__,
5148                      osd->o_osd, tid);
5149                 *skip = 1;
5150                 goto out_unlock_session;
5151         }
5152
5153         ceph_msg_revoke_incoming(req->r_reply);
5154
5155         if (front_len > req->r_reply->front_alloc_len) {
5156                 pr_warn("%s osd%d tid %llu front %d > preallocated %d\n",
5157                         __func__, osd->o_osd, req->r_tid, front_len,
5158                         req->r_reply->front_alloc_len);
5159                 m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front_len, GFP_NOFS,
5160                                  false);
5161                 if (!m)
5162                         goto out_unlock_session;
5163                 ceph_msg_put(req->r_reply);
5164                 req->r_reply = m;
5165         }
5166
5167         if (data_len > req->r_reply->data_length) {
5168                 pr_warn("%s osd%d tid %llu data %d > preallocated %zu, skipping\n",
5169                         __func__, osd->o_osd, req->r_tid, data_len,
5170                         req->r_reply->data_length);
5171                 m = NULL;
5172                 *skip = 1;
5173                 goto out_unlock_session;
5174         }
5175
5176         m = ceph_msg_get(req->r_reply);
5177         dout("get_reply tid %lld %p\n", tid, m);
5178
5179 out_unlock_session:
5180         mutex_unlock(&osd->lock);
5181 out_unlock_osdc:
5182         up_read(&osdc->lock);
5183         return m;
5184 }
5185
5186 /*
5187  * TODO: switch to a msg-owned pagelist
5188  */
5189 static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
5190 {
5191         struct ceph_msg *m;
5192         int type = le16_to_cpu(hdr->type);
5193         u32 front_len = le32_to_cpu(hdr->front_len);
5194         u32 data_len = le32_to_cpu(hdr->data_len);
5195
5196         m = ceph_msg_new(type, front_len, GFP_NOIO, false);
5197         if (!m)
5198                 return NULL;
5199
5200         if (data_len) {
5201                 struct page **pages;
5202                 struct ceph_osd_data osd_data;
5203
5204                 pages = ceph_alloc_page_vector(calc_pages_for(0, data_len),
5205                                                GFP_NOIO);
5206                 if (IS_ERR(pages)) {
5207                         ceph_msg_put(m);
5208                         return NULL;
5209                 }
5210
5211                 ceph_osd_data_pages_init(&osd_data, pages, data_len, 0, false,
5212                                          false);
5213                 ceph_osdc_msg_data_add(m, &osd_data);
5214         }
5215
5216         return m;
5217 }
5218
5219 static struct ceph_msg *alloc_msg(struct ceph_connection *con,
5220                                   struct ceph_msg_header *hdr,
5221                                   int *skip)
5222 {
5223         struct ceph_osd *osd = con->private;
5224         int type = le16_to_cpu(hdr->type);
5225
5226         *skip = 0;
5227         switch (type) {
5228         case CEPH_MSG_OSD_MAP:
5229         case CEPH_MSG_OSD_BACKOFF:
5230         case CEPH_MSG_WATCH_NOTIFY:
5231                 return alloc_msg_with_page_vector(hdr);
5232         case CEPH_MSG_OSD_OPREPLY:
5233                 return get_reply(con, hdr, skip);
5234         default:
5235                 pr_warn("%s osd%d unknown msg type %d, skipping\n", __func__,
5236                         osd->o_osd, type);
5237                 *skip = 1;
5238                 return NULL;
5239         }
5240 }
5241
5242 /*
5243  * Wrappers to refcount containing ceph_osd struct
5244  */
5245 static struct ceph_connection *get_osd_con(struct ceph_connection *con)
5246 {
5247         struct ceph_osd *osd = con->private;
5248         if (get_osd(osd))
5249                 return con;
5250         return NULL;
5251 }
5252
5253 static void put_osd_con(struct ceph_connection *con)
5254 {
5255         struct ceph_osd *osd = con->private;
5256         put_osd(osd);
5257 }
5258
5259 /*
5260  * authentication
5261  */
5262 /*
5263  * Note: returned pointer is the address of a structure that's
5264  * managed separately.  Caller must *not* attempt to free it.
5265  */
5266 static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
5267                                         int *proto, int force_new)
5268 {
5269         struct ceph_osd *o = con->private;
5270         struct ceph_osd_client *osdc = o->o_osdc;
5271         struct ceph_auth_client *ac = osdc->client->monc.auth;
5272         struct ceph_auth_handshake *auth = &o->o_auth;
5273
5274         if (force_new && auth->authorizer) {
5275                 ceph_auth_destroy_authorizer(auth->authorizer);
5276                 auth->authorizer = NULL;
5277         }
5278         if (!auth->authorizer) {
5279                 int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
5280                                                       auth);
5281                 if (ret)
5282                         return ERR_PTR(ret);
5283         } else {
5284                 int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
5285                                                      auth);
5286                 if (ret)
5287                         return ERR_PTR(ret);
5288         }
5289         *proto = ac->protocol;
5290
5291         return auth;
5292 }
5293
5294
5295 static int verify_authorizer_reply(struct ceph_connection *con)
5296 {
5297         struct ceph_osd *o = con->private;
5298         struct ceph_osd_client *osdc = o->o_osdc;
5299         struct ceph_auth_client *ac = osdc->client->monc.auth;
5300
5301         return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer);
5302 }
5303
5304 static int invalidate_authorizer(struct ceph_connection *con)
5305 {
5306         struct ceph_osd *o = con->private;
5307         struct ceph_osd_client *osdc = o->o_osdc;
5308         struct ceph_auth_client *ac = osdc->client->monc.auth;
5309
5310         ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
5311         return ceph_monc_validate_auth(&osdc->client->monc);
5312 }
5313
5314 static void osd_reencode_message(struct ceph_msg *msg)
5315 {
5316         int type = le16_to_cpu(msg->hdr.type);
5317
5318         if (type == CEPH_MSG_OSD_OP)
5319                 encode_request_finish(msg);
5320 }
5321
5322 static int osd_sign_message(struct ceph_msg *msg)
5323 {
5324         struct ceph_osd *o = msg->con->private;
5325         struct ceph_auth_handshake *auth = &o->o_auth;
5326
5327         return ceph_auth_sign_message(auth, msg);
5328 }
5329
5330 static int osd_check_message_signature(struct ceph_msg *msg)
5331 {
5332         struct ceph_osd *o = msg->con->private;
5333         struct ceph_auth_handshake *auth = &o->o_auth;
5334
5335         return ceph_auth_check_message_signature(auth, msg);
5336 }
5337
5338 static const struct ceph_connection_operations osd_con_ops = {
5339         .get = get_osd_con,
5340         .put = put_osd_con,
5341         .dispatch = dispatch,
5342         .get_authorizer = get_authorizer,
5343         .verify_authorizer_reply = verify_authorizer_reply,
5344         .invalidate_authorizer = invalidate_authorizer,
5345         .alloc_msg = alloc_msg,
5346         .reencode_message = osd_reencode_message,
5347         .sign_message = osd_sign_message,
5348         .check_message_signature = osd_check_message_signature,
5349         .fault = osd_fault,
5350 };