net/ceph/osdmap.c

   1 // SPDX-License-Identifier: GPL-2.0
   2
   3 #include <linux/ceph/ceph_debug.h>
   4
   5 #include <linux/module.h>
   6 #include <linux/slab.h>
   7
   8 #include <linux/ceph/libceph.h>
   9 #include <linux/ceph/osdmap.h>
  10 #include <linux/ceph/decode.h>
  11 #include <linux/crush/hash.h>
  12 #include <linux/crush/mapper.h>
  13
  14 char *ceph_osdmap_state_str(char *str, int len, u32 state)
  15 {
  16         if (!len)
  17                 return str;
  18
  19         if ((state & CEPH_OSD_EXISTS) && (state & CEPH_OSD_UP))
  20                 snprintf(str, len, "exists, up");
  21         else if (state & CEPH_OSD_EXISTS)
  22                 snprintf(str, len, "exists");
  23         else if (state & CEPH_OSD_UP)
  24                 snprintf(str, len, "up");
  25         else
  26                 snprintf(str, len, "doesn't exist");
  27
  28         return str;
  29 }
  30
  31 /* maps */
  32
  33 static int calc_bits_of(unsigned int t)
  34 {
  35         int b = 0;
  36         while (t) {
  37                 t = t >> 1;
  38                 b++;
  39         }
  40         return b;
  41 }
  42
  43 /*
  44  * the foo_mask is the smallest value 2^n-1 that is >= foo.
  45  */
  46 static void calc_pg_masks(struct ceph_pg_pool_info *pi)
  47 {
  48         pi->pg_num_mask = (1 << calc_bits_of(pi->pg_num-1)) - 1;
  49         pi->pgp_num_mask = (1 << calc_bits_of(pi->pgp_num-1)) - 1;
  50 }
  51
  52 /*
  53  * decode crush map
  54  */
  55 static int crush_decode_uniform_bucket(void **p, void *end,
  56                                        struct crush_bucket_uniform *b)
  57 {
  58         dout("crush_decode_uniform_bucket %p to %p\n", *p, end);
  59         ceph_decode_need(p, end, (1+b->h.size) * sizeof(u32), bad);
  60         b->item_weight = ceph_decode_32(p);
  61         return 0;
  62 bad:
  63         return -EINVAL;
  64 }
  65
  66 static int crush_decode_list_bucket(void **p, void *end,
  67                                     struct crush_bucket_list *b)
  68 {
  69         int j;
  70         dout("crush_decode_list_bucket %p to %p\n", *p, end);
  71         b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
  72         if (b->item_weights == NULL)
  73                 return -ENOMEM;
  74         b->sum_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
  75         if (b->sum_weights == NULL)
  76                 return -ENOMEM;
  77         ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad);
  78         for (j = 0; j < b->h.size; j++) {
  79                 b->item_weights[j] = ceph_decode_32(p);
  80                 b->sum_weights[j] = ceph_decode_32(p);
  81         }
  82         return 0;
  83 bad:
  84         return -EINVAL;
  85 }
  86
  87 static int crush_decode_tree_bucket(void **p, void *end,
  88                                     struct crush_bucket_tree *b)
  89 {
  90         int j;
  91         dout("crush_decode_tree_bucket %p to %p\n", *p, end);
  92         ceph_decode_8_safe(p, end, b->num_nodes, bad);
  93         b->node_weights = kcalloc(b->num_nodes, sizeof(u32), GFP_NOFS);
  94         if (b->node_weights == NULL)
  95                 return -ENOMEM;
  96         ceph_decode_need(p, end, b->num_nodes * sizeof(u32), bad);
  97         for (j = 0; j < b->num_nodes; j++)
  98                 b->node_weights[j] = ceph_decode_32(p);
  99         return 0;
 100 bad:
 101         return -EINVAL;
 102 }
 103
 104 static int crush_decode_straw_bucket(void **p, void *end,
 105                                      struct crush_bucket_straw *b)
 106 {
 107         int j;
 108         dout("crush_decode_straw_bucket %p to %p\n", *p, end);
 109         b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
 110         if (b->item_weights == NULL)
 111                 return -ENOMEM;
 112         b->straws = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
 113         if (b->straws == NULL)
 114                 return -ENOMEM;
 115         ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad);
 116         for (j = 0; j < b->h.size; j++) {
 117                 b->item_weights[j] = ceph_decode_32(p);
 118                 b->straws[j] = ceph_decode_32(p);
 119         }
 120         return 0;
 121 bad:
 122         return -EINVAL;
 123 }
 124
 125 static int crush_decode_straw2_bucket(void **p, void *end,
 126                                       struct crush_bucket_straw2 *b)
 127 {
 128         int j;
 129         dout("crush_decode_straw2_bucket %p to %p\n", *p, end);
 130         b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS);
 131         if (b->item_weights == NULL)
 132                 return -ENOMEM;
 133         ceph_decode_need(p, end, b->h.size * sizeof(u32), bad);
 134         for (j = 0; j < b->h.size; j++)
 135                 b->item_weights[j] = ceph_decode_32(p);
 136         return 0;
 137 bad:
 138         return -EINVAL;
 139 }
 140
 141 struct crush_name_node {
 142         struct rb_node cn_node;
 143         int cn_id;
 144         char cn_name[];
 145 };
 146
 147 static struct crush_name_node *alloc_crush_name(size_t name_len)
 148 {
 149         struct crush_name_node *cn;
 150
 151         cn = kmalloc(sizeof(*cn) + name_len + 1, GFP_NOIO);
 152         if (!cn)
 153                 return NULL;
 154
 155         RB_CLEAR_NODE(&cn->cn_node);
 156         return cn;
 157 }
 158
 159 static void free_crush_name(struct crush_name_node *cn)
 160 {
 161         WARN_ON(!RB_EMPTY_NODE(&cn->cn_node));
 162
 163         kfree(cn);
 164 }
 165
 166 DEFINE_RB_FUNCS(crush_name, struct crush_name_node, cn_id, cn_node)
 167
 168 static int decode_crush_names(void **p, void *end, struct rb_root *root)
 169 {
 170         u32 n;
 171
 172         ceph_decode_32_safe(p, end, n, e_inval);
 173         while (n--) {
 174                 struct crush_name_node *cn;
 175                 int id;
 176                 u32 name_len;
 177
 178                 ceph_decode_32_safe(p, end, id, e_inval);
 179                 ceph_decode_32_safe(p, end, name_len, e_inval);
 180                 ceph_decode_need(p, end, name_len, e_inval);
 181
 182                 cn = alloc_crush_name(name_len);
 183                 if (!cn)
 184                         return -ENOMEM;
 185
 186                 cn->cn_id = id;
 187                 memcpy(cn->cn_name, *p, name_len);
 188                 cn->cn_name[name_len] = '\0';
 189                 *p += name_len;
 190
 191                 if (!__insert_crush_name(root, cn)) {
 192                         free_crush_name(cn);
 193                         return -EEXIST;
 194                 }
 195         }
 196
 197         return 0;
 198
 199 e_inval:
 200         return -EINVAL;
 201 }
 202
 203 void clear_crush_names(struct rb_root *root)
 204 {
 205         while (!RB_EMPTY_ROOT(root)) {
 206                 struct crush_name_node *cn =
 207                     rb_entry(rb_first(root), struct crush_name_node, cn_node);
 208
 209                 erase_crush_name(root, cn);
 210                 free_crush_name(cn);
 211         }
 212 }
 213
 214 static struct crush_choose_arg_map *alloc_choose_arg_map(void)
 215 {
 216         struct crush_choose_arg_map *arg_map;
 217
 218         arg_map = kzalloc(sizeof(*arg_map), GFP_NOIO);
 219         if (!arg_map)
 220                 return NULL;
 221
 222         RB_CLEAR_NODE(&arg_map->node);
 223         return arg_map;
 224 }
 225
 226 static void free_choose_arg_map(struct crush_choose_arg_map *arg_map)
 227 {
 228         if (arg_map) {
 229                 int i, j;
 230
 231                 WARN_ON(!RB_EMPTY_NODE(&arg_map->node));
 232
 233                 for (i = 0; i < arg_map->size; i++) {
 234                         struct crush_choose_arg *arg = &arg_map->args[i];
 235
 236                         for (j = 0; j < arg->weight_set_size; j++)
 237                                 kfree(arg->weight_set[j].weights);
 238                         kfree(arg->weight_set);
 239                         kfree(arg->ids);
 240                 }
 241                 kfree(arg_map->args);
 242                 kfree(arg_map);
 243         }
 244 }
 245
 246 DEFINE_RB_FUNCS(choose_arg_map, struct crush_choose_arg_map, choose_args_index,
 247                 node);
 248
 249 void clear_choose_args(struct crush_map *c)
 250 {
 251         while (!RB_EMPTY_ROOT(&c->choose_args)) {
 252                 struct crush_choose_arg_map *arg_map =
 253                     rb_entry(rb_first(&c->choose_args),
 254                              struct crush_choose_arg_map, node);
 255
 256                 erase_choose_arg_map(&c->choose_args, arg_map);
 257                 free_choose_arg_map(arg_map);
 258         }
 259 }
 260
 261 static u32 *decode_array_32_alloc(void **p, void *end, u32 *plen)
 262 {
 263         u32 *a = NULL;
 264         u32 len;
 265         int ret;
 266
 267         ceph_decode_32_safe(p, end, len, e_inval);
 268         if (len) {
 269                 u32 i;
 270
 271                 a = kmalloc_array(len, sizeof(u32), GFP_NOIO);
 272                 if (!a) {
 273                         ret = -ENOMEM;
 274                         goto fail;
 275                 }
 276
 277                 ceph_decode_need(p, end, len * sizeof(u32), e_inval);
 278                 for (i = 0; i < len; i++)
 279                         a[i] = ceph_decode_32(p);
 280         }
 281
 282         *plen = len;
 283         return a;
 284
 285 e_inval:
 286         ret = -EINVAL;
 287 fail:
 288         kfree(a);
 289         return ERR_PTR(ret);
 290 }
 291
 292 /*
 293  * Assumes @arg is zero-initialized.
 294  */
 295 static int decode_choose_arg(void **p, void *end, struct crush_choose_arg *arg)
 296 {
 297         int ret;
 298
 299         ceph_decode_32_safe(p, end, arg->weight_set_size, e_inval);
 300         if (arg->weight_set_size) {
 301                 u32 i;
 302
 303                 arg->weight_set = kmalloc_array(arg->weight_set_size,
 304                                                 sizeof(*arg->weight_set),
 305                                                 GFP_NOIO);
 306                 if (!arg->weight_set)
 307                         return -ENOMEM;
 308
 309                 for (i = 0; i < arg->weight_set_size; i++) {
 310                         struct crush_weight_set *w = &arg->weight_set[i];
 311
 312                         w->weights = decode_array_32_alloc(p, end, &w->size);
 313                         if (IS_ERR(w->weights)) {
 314                                 ret = PTR_ERR(w->weights);
 315                                 w->weights = NULL;
 316                                 return ret;
 317                         }
 318                 }
 319         }
 320
 321         arg->ids = decode_array_32_alloc(p, end, &arg->ids_size);
 322         if (IS_ERR(arg->ids)) {
 323                 ret = PTR_ERR(arg->ids);
 324                 arg->ids = NULL;
 325                 return ret;
 326         }
 327
 328         return 0;
 329
 330 e_inval:
 331         return -EINVAL;
 332 }
 333
 334 static int decode_choose_args(void **p, void *end, struct crush_map *c)
 335 {
 336         struct crush_choose_arg_map *arg_map = NULL;
 337         u32 num_choose_arg_maps, num_buckets;
 338         int ret;
 339
 340         ceph_decode_32_safe(p, end, num_choose_arg_maps, e_inval);
 341         while (num_choose_arg_maps--) {
 342                 arg_map = alloc_choose_arg_map();
 343                 if (!arg_map) {
 344                         ret = -ENOMEM;
 345                         goto fail;
 346                 }
 347
 348                 ceph_decode_64_safe(p, end, arg_map->choose_args_index,
 349                                     e_inval);
 350                 arg_map->size = c->max_buckets;
 351                 arg_map->args = kcalloc(arg_map->size, sizeof(*arg_map->args),
 352                                         GFP_NOIO);
 353                 if (!arg_map->args) {
 354                         ret = -ENOMEM;
 355                         goto fail;
 356                 }
 357
 358                 ceph_decode_32_safe(p, end, num_buckets, e_inval);
 359                 while (num_buckets--) {
 360                         struct crush_choose_arg *arg;
 361                         u32 bucket_index;
 362
 363                         ceph_decode_32_safe(p, end, bucket_index, e_inval);
 364                         if (bucket_index >= arg_map->size)
 365                                 goto e_inval;
 366
 367                         arg = &arg_map->args[bucket_index];
 368                         ret = decode_choose_arg(p, end, arg);
 369                         if (ret)
 370                                 goto fail;
 371
 372                         if (arg->ids_size &&
 373                             arg->ids_size != c->buckets[bucket_index]->size)
 374                                 goto e_inval;
 375                 }
 376
 377                 insert_choose_arg_map(&c->choose_args, arg_map);
 378         }
 379
 380         return 0;
 381
 382 e_inval:
 383         ret = -EINVAL;
 384 fail:
 385         free_choose_arg_map(arg_map);
 386         return ret;
 387 }
 388
 389 static void crush_finalize(struct crush_map *c)
 390 {
 391         __s32 b;
 392
 393         /* Space for the array of pointers to per-bucket workspace */
 394         c->working_size = sizeof(struct crush_work) +
 395             c->max_buckets * sizeof(struct crush_work_bucket *);
 396
 397         for (b = 0; b < c->max_buckets; b++) {
 398                 if (!c->buckets[b])
 399                         continue;
 400
 401                 switch (c->buckets[b]->alg) {
 402                 default:
 403                         /*
 404                          * The base case, permutation variables and
 405                          * the pointer to the permutation array.
 406                          */
 407                         c->working_size += sizeof(struct crush_work_bucket);
 408                         break;
 409                 }
 410                 /* Every bucket has a permutation array. */
 411                 c->working_size += c->buckets[b]->size * sizeof(__u32);
 412         }
 413 }
 414
 415 static struct crush_map *crush_decode(void *pbyval, void *end)
 416 {
 417         struct crush_map *c;
 418         int err;
 419         int i, j;
 420         void **p = &pbyval;
 421         void *start = pbyval;
 422         u32 magic;
 423
 424         dout("crush_decode %p to %p len %d\n", *p, end, (int)(end - *p));
 425
 426         c = kzalloc(sizeof(*c), GFP_NOFS);
 427         if (c == NULL)
 428                 return ERR_PTR(-ENOMEM);
 429
 430         c->type_names = RB_ROOT;
 431         c->names = RB_ROOT;
 432         c->choose_args = RB_ROOT;
 433
 434         /* set tunables to default values */
 435         c->choose_local_tries = 2;
 436         c->choose_local_fallback_tries = 5;
 437         c->choose_total_tries = 19;
 438         c->chooseleaf_descend_once = 0;
 439
 440         ceph_decode_need(p, end, 4*sizeof(u32), bad);
 441         magic = ceph_decode_32(p);
 442         if (magic != CRUSH_MAGIC) {
 443                 pr_err("crush_decode magic %x != current %x\n",
 444                        (unsigned int)magic, (unsigned int)CRUSH_MAGIC);
 445                 goto bad;
 446         }
 447         c->max_buckets = ceph_decode_32(p);
 448         c->max_rules = ceph_decode_32(p);
 449         c->max_devices = ceph_decode_32(p);
 450
 451         c->buckets = kcalloc(c->max_buckets, sizeof(*c->buckets), GFP_NOFS);
 452         if (c->buckets == NULL)
 453                 goto badmem;
 454         c->rules = kcalloc(c->max_rules, sizeof(*c->rules), GFP_NOFS);
 455         if (c->rules == NULL)
 456                 goto badmem;
 457
 458         /* buckets */
 459         for (i = 0; i < c->max_buckets; i++) {
 460                 int size = 0;
 461                 u32 alg;
 462                 struct crush_bucket *b;
 463
 464                 ceph_decode_32_safe(p, end, alg, bad);
 465                 if (alg == 0) {
 466                         c->buckets[i] = NULL;
 467                         continue;
 468                 }
 469                 dout("crush_decode bucket %d off %x %p to %p\n",
 470                      i, (int)(*p-start), *p, end);
 471
 472                 switch (alg) {
 473                 case CRUSH_BUCKET_UNIFORM:
 474                         size = sizeof(struct crush_bucket_uniform);
 475                         break;
 476                 case CRUSH_BUCKET_LIST:
 477                         size = sizeof(struct crush_bucket_list);
 478                         break;
 479                 case CRUSH_BUCKET_TREE:
 480                         size = sizeof(struct crush_bucket_tree);
 481                         break;
 482                 case CRUSH_BUCKET_STRAW:
 483                         size = sizeof(struct crush_bucket_straw);
 484                         break;
 485                 case CRUSH_BUCKET_STRAW2:
 486                         size = sizeof(struct crush_bucket_straw2);
 487                         break;
 488                 default:
 489                         goto bad;
 490                 }
 491                 BUG_ON(size == 0);
 492                 b = c->buckets[i] = kzalloc(size, GFP_NOFS);
 493                 if (b == NULL)
 494                         goto badmem;
 495
 496                 ceph_decode_need(p, end, 4*sizeof(u32), bad);
 497                 b->id = ceph_decode_32(p);
 498                 b->type = ceph_decode_16(p);
 499                 b->alg = ceph_decode_8(p);
 500                 b->hash = ceph_decode_8(p);
 501                 b->weight = ceph_decode_32(p);
 502                 b->size = ceph_decode_32(p);
 503
 504                 dout("crush_decode bucket size %d off %x %p to %p\n",
 505                      b->size, (int)(*p-start), *p, end);
 506
 507                 b->items = kcalloc(b->size, sizeof(__s32), GFP_NOFS);
 508                 if (b->items == NULL)
 509                         goto badmem;
 510
 511                 ceph_decode_need(p, end, b->size*sizeof(u32), bad);
 512                 for (j = 0; j < b->size; j++)
 513                         b->items[j] = ceph_decode_32(p);
 514
 515                 switch (b->alg) {
 516                 case CRUSH_BUCKET_UNIFORM:
 517                         err = crush_decode_uniform_bucket(p, end,
 518                                   (struct crush_bucket_uniform *)b);
 519                         if (err < 0)
 520                                 goto fail;
 521                         break;
 522                 case CRUSH_BUCKET_LIST:
 523                         err = crush_decode_list_bucket(p, end,
 524                                (struct crush_bucket_list *)b);
 525                         if (err < 0)
 526                                 goto fail;
 527                         break;
 528                 case CRUSH_BUCKET_TREE:
 529                         err = crush_decode_tree_bucket(p, end,
 530                                 (struct crush_bucket_tree *)b);
 531                         if (err < 0)
 532                                 goto fail;
 533                         break;
 534                 case CRUSH_BUCKET_STRAW:
 535                         err = crush_decode_straw_bucket(p, end,
 536                                 (struct crush_bucket_straw *)b);
 537                         if (err < 0)
 538                                 goto fail;
 539                         break;
 540                 case CRUSH_BUCKET_STRAW2:
 541                         err = crush_decode_straw2_bucket(p, end,
 542                                 (struct crush_bucket_straw2 *)b);
 543                         if (err < 0)
 544                                 goto fail;
 545                         break;
 546                 }
 547         }
 548
 549         /* rules */
 550         dout("rule vec is %p\n", c->rules);
 551         for (i = 0; i < c->max_rules; i++) {
 552                 u32 yes;
 553                 struct crush_rule *r;
 554
 555                 ceph_decode_32_safe(p, end, yes, bad);
 556                 if (!yes) {
 557                         dout("crush_decode NO rule %d off %x %p to %p\n",
 558                              i, (int)(*p-start), *p, end);
 559                         c->rules[i] = NULL;
 560                         continue;
 561                 }
 562
 563                 dout("crush_decode rule %d off %x %p to %p\n",
 564                      i, (int)(*p-start), *p, end);
 565
 566                 /* len */
 567                 ceph_decode_32_safe(p, end, yes, bad);
 568 #if BITS_PER_LONG == 32
 569                 if (yes > (ULONG_MAX - sizeof(*r))
 570                           / sizeof(struct crush_rule_step))
 571                         goto bad;
 572 #endif
 573                 r = kmalloc(struct_size(r, steps, yes), GFP_NOFS);
 574                 c->rules[i] = r;
 575                 if (r == NULL)
 576                         goto badmem;
 577                 dout(" rule %d is at %p\n", i, r);
 578                 r->len = yes;
 579                 ceph_decode_copy_safe(p, end, &r->mask, 4, bad); /* 4 u8's */
 580                 ceph_decode_need(p, end, r->len*3*sizeof(u32), bad);
 581                 for (j = 0; j < r->len; j++) {
 582                         r->steps[j].op = ceph_decode_32(p);
 583                         r->steps[j].arg1 = ceph_decode_32(p);
 584                         r->steps[j].arg2 = ceph_decode_32(p);
 585                 }
 586         }
 587
 588         err = decode_crush_names(p, end, &c->type_names);
 589         if (err)
 590                 goto fail;
 591
 592         err = decode_crush_names(p, end, &c->names);
 593         if (err)
 594                 goto fail;
 595
 596         ceph_decode_skip_map(p, end, 32, string, bad); /* rule_name_map */
 597
 598         /* tunables */
 599         ceph_decode_need(p, end, 3*sizeof(u32), done);
 600         c->choose_local_tries = ceph_decode_32(p);
 601         c->choose_local_fallback_tries =  ceph_decode_32(p);
 602         c->choose_total_tries = ceph_decode_32(p);
 603         dout("crush decode tunable choose_local_tries = %d\n",
 604              c->choose_local_tries);
 605         dout("crush decode tunable choose_local_fallback_tries = %d\n",
 606              c->choose_local_fallback_tries);
 607         dout("crush decode tunable choose_total_tries = %d\n",
 608              c->choose_total_tries);
 609
 610         ceph_decode_need(p, end, sizeof(u32), done);
 611         c->chooseleaf_descend_once = ceph_decode_32(p);
 612         dout("crush decode tunable chooseleaf_descend_once = %d\n",
 613              c->chooseleaf_descend_once);
 614
 615         ceph_decode_need(p, end, sizeof(u8), done);
 616         c->chooseleaf_vary_r = ceph_decode_8(p);
 617         dout("crush decode tunable chooseleaf_vary_r = %d\n",
 618              c->chooseleaf_vary_r);
 619
 620         /* skip straw_calc_version, allowed_bucket_algs */
 621         ceph_decode_need(p, end, sizeof(u8) + sizeof(u32), done);
 622         *p += sizeof(u8) + sizeof(u32);
 623
 624         ceph_decode_need(p, end, sizeof(u8), done);
 625         c->chooseleaf_stable = ceph_decode_8(p);
 626         dout("crush decode tunable chooseleaf_stable = %d\n",
 627              c->chooseleaf_stable);
 628
 629         if (*p != end) {
 630                 /* class_map */
 631                 ceph_decode_skip_map(p, end, 32, 32, bad);
 632                 /* class_name */
 633                 ceph_decode_skip_map(p, end, 32, string, bad);
 634                 /* class_bucket */
 635                 ceph_decode_skip_map_of_map(p, end, 32, 32, 32, bad);
 636         }
 637
 638         if (*p != end) {
 639                 err = decode_choose_args(p, end, c);
 640                 if (err)
 641                         goto fail;
 642         }
 643
 644 done:
 645         crush_finalize(c);
 646         dout("crush_decode success\n");
 647         return c;
 648
 649 badmem:
 650         err = -ENOMEM;
 651 fail:
 652         dout("crush_decode fail %d\n", err);
 653         crush_destroy(c);
 654         return ERR_PTR(err);
 655
 656 bad:
 657         err = -EINVAL;
 658         goto fail;
 659 }
 660
 661 int ceph_pg_compare(const struct ceph_pg *lhs, const struct ceph_pg *rhs)
 662 {
 663         if (lhs->pool < rhs->pool)
 664                 return -1;
 665         if (lhs->pool > rhs->pool)
 666                 return 1;
 667         if (lhs->seed < rhs->seed)
 668                 return -1;
 669         if (lhs->seed > rhs->seed)
 670                 return 1;
 671
 672         return 0;
 673 }
 674
 675 int ceph_spg_compare(const struct ceph_spg *lhs, const struct ceph_spg *rhs)
 676 {
 677         int ret;
 678
 679         ret = ceph_pg_compare(&lhs->pgid, &rhs->pgid);
 680         if (ret)
 681                 return ret;
 682
 683         if (lhs->shard < rhs->shard)
 684                 return -1;
 685         if (lhs->shard > rhs->shard)
 686                 return 1;
 687
 688         return 0;
 689 }
 690
 691 static struct ceph_pg_mapping *alloc_pg_mapping(size_t payload_len)
 692 {
 693         struct ceph_pg_mapping *pg;
 694
 695         pg = kmalloc(sizeof(*pg) + payload_len, GFP_NOIO);
 696         if (!pg)
 697                 return NULL;
 698
 699         RB_CLEAR_NODE(&pg->node);
 700         return pg;
 701 }
 702
 703 static void free_pg_mapping(struct ceph_pg_mapping *pg)
 704 {
 705         WARN_ON(!RB_EMPTY_NODE(&pg->node));
 706
 707         kfree(pg);
 708 }
 709
 710 /*
 711  * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid
 712  * to a set of osds) and primary_temp (explicit primary setting)
 713  */
 714 DEFINE_RB_FUNCS2(pg_mapping, struct ceph_pg_mapping, pgid, ceph_pg_compare,
 715                  RB_BYPTR, const struct ceph_pg *, node)
 716
 717 /*
 718  * rbtree of pg pool info
 719  */
 720 DEFINE_RB_FUNCS(pg_pool, struct ceph_pg_pool_info, id, node)
 721
 722 struct ceph_pg_pool_info *ceph_pg_pool_by_id(struct ceph_osdmap *map, u64 id)
 723 {
 724         return lookup_pg_pool(&map->pg_pools, id);
 725 }
 726
 727 const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id)
 728 {
 729         struct ceph_pg_pool_info *pi;
 730
 731         if (id == CEPH_NOPOOL)
 732                 return NULL;
 733
 734         if (WARN_ON_ONCE(id > (u64) INT_MAX))
 735                 return NULL;
 736
 737         pi = lookup_pg_pool(&map->pg_pools, id);
 738         return pi ? pi->name : NULL;
 739 }
 740 EXPORT_SYMBOL(ceph_pg_pool_name_by_id);
 741
 742 int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name)
 743 {
 744         struct rb_node *rbp;
 745
 746         for (rbp = rb_first(&map->pg_pools); rbp; rbp = rb_next(rbp)) {
 747                 struct ceph_pg_pool_info *pi =
 748                         rb_entry(rbp, struct ceph_pg_pool_info, node);
 749                 if (pi->name && strcmp(pi->name, name) == 0)
 750                         return pi->id;
 751         }
 752         return -ENOENT;
 753 }
 754 EXPORT_SYMBOL(ceph_pg_poolid_by_name);
 755
 756 u64 ceph_pg_pool_flags(struct ceph_osdmap *map, u64 id)
 757 {
 758         struct ceph_pg_pool_info *pi;
 759
 760         pi = lookup_pg_pool(&map->pg_pools, id);
 761         return pi ? pi->flags : 0;
 762 }
 763 EXPORT_SYMBOL(ceph_pg_pool_flags);
 764
 765 static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi)
 766 {
 767         erase_pg_pool(root, pi);
 768         kfree(pi->name);
 769         kfree(pi);
 770 }
 771
 772 static int decode_pool(void **p, void *end, struct ceph_pg_pool_info *pi)
 773 {
 774         u8 ev, cv;
 775         unsigned len, num;
 776         void *pool_end;
 777
 778         ceph_decode_need(p, end, 2 + 4, bad);
 779         ev = ceph_decode_8(p);  /* encoding version */
 780         cv = ceph_decode_8(p); /* compat version */
 781         if (ev < 5) {
 782                 pr_warn("got v %d < 5 cv %d of ceph_pg_pool\n", ev, cv);
 783                 return -EINVAL;
 784         }
 785         if (cv > 9) {
 786                 pr_warn("got v %d cv %d > 9 of ceph_pg_pool\n", ev, cv);
 787                 return -EINVAL;
 788         }
 789         len = ceph_decode_32(p);
 790         ceph_decode_need(p, end, len, bad);
 791         pool_end = *p + len;
 792
 793         pi->type = ceph_decode_8(p);
 794         pi->size = ceph_decode_8(p);
 795         pi->crush_ruleset = ceph_decode_8(p);
 796         pi->object_hash = ceph_decode_8(p);
 797
 798         pi->pg_num = ceph_decode_32(p);
 799         pi->pgp_num = ceph_decode_32(p);
 800
 801         *p += 4 + 4;  /* skip lpg* */
 802         *p += 4;      /* skip last_change */
 803         *p += 8 + 4;  /* skip snap_seq, snap_epoch */
 804
 805         /* skip snaps */
 806         num = ceph_decode_32(p);
 807         while (num--) {
 808                 *p += 8;  /* snapid key */
 809                 *p += 1 + 1; /* versions */
 810                 len = ceph_decode_32(p);
 811                 *p += len;
 812         }
 813
 814         /* skip removed_snaps */
 815         num = ceph_decode_32(p);
 816         *p += num * (8 + 8);
 817
 818         *p += 8;  /* skip auid */
 819         pi->flags = ceph_decode_64(p);
 820         *p += 4;  /* skip crash_replay_interval */
 821
 822         if (ev >= 7)
 823                 pi->min_size = ceph_decode_8(p);
 824         else
 825                 pi->min_size = pi->size - pi->size / 2;
 826
 827         if (ev >= 8)
 828                 *p += 8 + 8;  /* skip quota_max_* */
 829
 830         if (ev >= 9) {
 831                 /* skip tiers */
 832                 num = ceph_decode_32(p);
 833                 *p += num * 8;
 834
 835                 *p += 8;  /* skip tier_of */
 836                 *p += 1;  /* skip cache_mode */
 837
 838                 pi->read_tier = ceph_decode_64(p);
 839                 pi->write_tier = ceph_decode_64(p);
 840         } else {
 841                 pi->read_tier = -1;
 842                 pi->write_tier = -1;
 843         }
 844
 845         if (ev >= 10) {
 846                 /* skip properties */
 847                 num = ceph_decode_32(p);
 848                 while (num--) {
 849                         len = ceph_decode_32(p);
 850                         *p += len; /* key */
 851                         len = ceph_decode_32(p);
 852                         *p += len; /* val */
 853                 }
 854         }
 855
 856         if (ev >= 11) {
 857                 /* skip hit_set_params */
 858                 *p += 1 + 1; /* versions */
 859                 len = ceph_decode_32(p);
 860                 *p += len;
 861
 862                 *p += 4; /* skip hit_set_period */
 863                 *p += 4; /* skip hit_set_count */
 864         }
 865
 866         if (ev >= 12)
 867                 *p += 4; /* skip stripe_width */
 868
 869         if (ev >= 13) {
 870                 *p += 8; /* skip target_max_bytes */
 871                 *p += 8; /* skip target_max_objects */
 872                 *p += 4; /* skip cache_target_dirty_ratio_micro */
 873                 *p += 4; /* skip cache_target_full_ratio_micro */
 874                 *p += 4; /* skip cache_min_flush_age */
 875                 *p += 4; /* skip cache_min_evict_age */
 876         }
 877
 878         if (ev >=  14) {
 879                 /* skip erasure_code_profile */
 880                 len = ceph_decode_32(p);
 881                 *p += len;
 882         }
 883
 884         /*
 885          * last_force_op_resend_preluminous, will be overridden if the
 886          * map was encoded with RESEND_ON_SPLIT
 887          */
 888         if (ev >= 15)
 889                 pi->last_force_request_resend = ceph_decode_32(p);
 890         else
 891                 pi->last_force_request_resend = 0;
 892
 893         if (ev >= 16)
 894                 *p += 4; /* skip min_read_recency_for_promote */
 895
 896         if (ev >= 17)
 897                 *p += 8; /* skip expected_num_objects */
 898
 899         if (ev >= 19)
 900                 *p += 4; /* skip cache_target_dirty_high_ratio_micro */
 901
 902         if (ev >= 20)
 903                 *p += 4; /* skip min_write_recency_for_promote */
 904
 905         if (ev >= 21)
 906                 *p += 1; /* skip use_gmt_hitset */
 907
 908         if (ev >= 22)
 909                 *p += 1; /* skip fast_read */
 910
 911         if (ev >= 23) {
 912                 *p += 4; /* skip hit_set_grade_decay_rate */
 913                 *p += 4; /* skip hit_set_search_last_n */
 914         }
 915
 916         if (ev >= 24) {
 917                 /* skip opts */
 918                 *p += 1 + 1; /* versions */
 919                 len = ceph_decode_32(p);
 920                 *p += len;
 921         }
 922
 923         if (ev >= 25)
 924                 pi->last_force_request_resend = ceph_decode_32(p);
 925
 926         /* ignore the rest */
 927
 928         *p = pool_end;
 929         calc_pg_masks(pi);
 930         return 0;
 931
 932 bad:
 933         return -EINVAL;
 934 }
 935
 936 static int decode_pool_names(void **p, void *end, struct ceph_osdmap *map)
 937 {
 938         struct ceph_pg_pool_info *pi;
 939         u32 num, len;
 940         u64 pool;
 941
 942         ceph_decode_32_safe(p, end, num, bad);
 943         dout(" %d pool names\n", num);
 944         while (num--) {
 945                 ceph_decode_64_safe(p, end, pool, bad);
 946                 ceph_decode_32_safe(p, end, len, bad);
 947                 dout("  pool %llu len %d\n", pool, len);
 948                 ceph_decode_need(p, end, len, bad);
 949                 pi = lookup_pg_pool(&map->pg_pools, pool);
 950                 if (pi) {
 951                         char *name = kstrndup(*p, len, GFP_NOFS);
 952
 953                         if (!name)
 954                                 return -ENOMEM;
 955                         kfree(pi->name);
 956                         pi->name = name;
 957                         dout("  name is %s\n", pi->name);
 958                 }
 959                 *p += len;
 960         }
 961         return 0;
 962
 963 bad:
 964         return -EINVAL;
 965 }
 966
 967 /*
 968  * osd map
 969  */
 970 struct ceph_osdmap *ceph_osdmap_alloc(void)
 971 {
 972         struct ceph_osdmap *map;
 973
 974         map = kzalloc(sizeof(*map), GFP_NOIO);
 975         if (!map)
 976                 return NULL;
 977
 978         map->pg_pools = RB_ROOT;
 979         map->pool_max = -1;
 980         map->pg_temp = RB_ROOT;
 981         map->primary_temp = RB_ROOT;
 982         map->pg_upmap = RB_ROOT;
 983         map->pg_upmap_items = RB_ROOT;
 984         mutex_init(&map->crush_workspace_mutex);
 985
 986         return map;
 987 }
 988
 989 void ceph_osdmap_destroy(struct ceph_osdmap *map)
 990 {
 991         dout("osdmap_destroy %p\n", map);
 992         if (map->crush)
 993                 crush_destroy(map->crush);
 994         while (!RB_EMPTY_ROOT(&map->pg_temp)) {
 995                 struct ceph_pg_mapping *pg =
 996                         rb_entry(rb_first(&map->pg_temp),
 997                                  struct ceph_pg_mapping, node);
 998                 erase_pg_mapping(&map->pg_temp, pg);
 999                 free_pg_mapping(pg);
1000         }
1001         while (!RB_EMPTY_ROOT(&map->primary_temp)) {
1002                 struct ceph_pg_mapping *pg =
1003                         rb_entry(rb_first(&map->primary_temp),
1004                                  struct ceph_pg_mapping, node);
1005                 erase_pg_mapping(&map->primary_temp, pg);
1006                 free_pg_mapping(pg);
1007         }
1008         while (!RB_EMPTY_ROOT(&map->pg_upmap)) {
1009                 struct ceph_pg_mapping *pg =
1010                         rb_entry(rb_first(&map->pg_upmap),
1011                                  struct ceph_pg_mapping, node);
1012                 rb_erase(&pg->node, &map->pg_upmap);
1013                 kfree(pg);
1014         }
1015         while (!RB_EMPTY_ROOT(&map->pg_upmap_items)) {
1016                 struct ceph_pg_mapping *pg =
1017                         rb_entry(rb_first(&map->pg_upmap_items),
1018                                  struct ceph_pg_mapping, node);
1019                 rb_erase(&pg->node, &map->pg_upmap_items);
1020                 kfree(pg);
1021         }
1022         while (!RB_EMPTY_ROOT(&map->pg_pools)) {
1023                 struct ceph_pg_pool_info *pi =
1024                         rb_entry(rb_first(&map->pg_pools),
1025                                  struct ceph_pg_pool_info, node);
1026                 __remove_pg_pool(&map->pg_pools, pi);
1027         }
1028         kvfree(map->osd_state);
1029         kvfree(map->osd_weight);
1030         kvfree(map->osd_addr);
1031         kvfree(map->osd_primary_affinity);
1032         kvfree(map->crush_workspace);
1033         kfree(map);
1034 }
1035
1036 /*
1037  * Adjust max_osd value, (re)allocate arrays.
1038  *
1039  * The new elements are properly initialized.
1040  */
1041 static int osdmap_set_max_osd(struct ceph_osdmap *map, u32 max)
1042 {
1043         u32 *state;
1044         u32 *weight;
1045         struct ceph_entity_addr *addr;
1046         u32 to_copy;
1047         int i;
1048
1049         dout("%s old %u new %u\n", __func__, map->max_osd, max);
1050         if (max == map->max_osd)
1051                 return 0;
1052
1053         state = ceph_kvmalloc(array_size(max, sizeof(*state)), GFP_NOFS);
1054         weight = ceph_kvmalloc(array_size(max, sizeof(*weight)), GFP_NOFS);
1055         addr = ceph_kvmalloc(array_size(max, sizeof(*addr)), GFP_NOFS);
1056         if (!state || !weight || !addr) {
1057                 kvfree(state);
1058                 kvfree(weight);
1059                 kvfree(addr);
1060                 return -ENOMEM;
1061         }
1062
1063         to_copy = min(map->max_osd, max);
1064         if (map->osd_state) {
1065                 memcpy(state, map->osd_state, to_copy * sizeof(*state));
1066                 memcpy(weight, map->osd_weight, to_copy * sizeof(*weight));
1067                 memcpy(addr, map->osd_addr, to_copy * sizeof(*addr));
1068                 kvfree(map->osd_state);
1069                 kvfree(map->osd_weight);
1070                 kvfree(map->osd_addr);
1071         }
1072
1073         map->osd_state = state;
1074         map->osd_weight = weight;
1075         map->osd_addr = addr;
1076         for (i = map->max_osd; i < max; i++) {
1077                 map->osd_state[i] = 0;
1078                 map->osd_weight[i] = CEPH_OSD_OUT;
1079                 memset(map->osd_addr + i, 0, sizeof(*map->osd_addr));
1080         }
1081
1082         if (map->osd_primary_affinity) {
1083                 u32 *affinity;
1084
1085                 affinity = ceph_kvmalloc(array_size(max, sizeof(*affinity)),
1086                                          GFP_NOFS);
1087                 if (!affinity)
1088                         return -ENOMEM;
1089
1090                 memcpy(affinity, map->osd_primary_affinity,
1091                        to_copy * sizeof(*affinity));
1092                 kvfree(map->osd_primary_affinity);
1093
1094                 map->osd_primary_affinity = affinity;
1095                 for (i = map->max_osd; i < max; i++)
1096                         map->osd_primary_affinity[i] =
1097                             CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1098         }
1099
1100         map->max_osd = max;
1101
1102         return 0;
1103 }
1104
1105 static int osdmap_set_crush(struct ceph_osdmap *map, struct crush_map *crush)
1106 {
1107         void *workspace;
1108         size_t work_size;
1109
1110         if (IS_ERR(crush))
1111                 return PTR_ERR(crush);
1112
1113         work_size = crush_work_size(crush, CEPH_PG_MAX_SIZE);
1114         dout("%s work_size %zu bytes\n", __func__, work_size);
1115         workspace = ceph_kvmalloc(work_size, GFP_NOIO);
1116         if (!workspace) {
1117                 crush_destroy(crush);
1118                 return -ENOMEM;
1119         }
1120         crush_init_workspace(crush, workspace);
1121
1122         if (map->crush)
1123                 crush_destroy(map->crush);
1124         kvfree(map->crush_workspace);
1125         map->crush = crush;
1126         map->crush_workspace = workspace;
1127         return 0;
1128 }
1129
1130 #define OSDMAP_WRAPPER_COMPAT_VER       7
1131 #define OSDMAP_CLIENT_DATA_COMPAT_VER   1
1132
1133 /*
1134  * Return 0 or error.  On success, *v is set to 0 for old (v6) osdmaps,
1135  * to struct_v of the client_data section for new (v7 and above)
1136  * osdmaps.
1137  */
1138 static int get_osdmap_client_data_v(void **p, void *end,
1139                                     const char *prefix, u8 *v)
1140 {
1141         u8 struct_v;
1142
1143         ceph_decode_8_safe(p, end, struct_v, e_inval);
1144         if (struct_v >= 7) {
1145                 u8 struct_compat;
1146
1147                 ceph_decode_8_safe(p, end, struct_compat, e_inval);
1148                 if (struct_compat > OSDMAP_WRAPPER_COMPAT_VER) {
1149                         pr_warn("got v %d cv %d > %d of %s ceph_osdmap\n",
1150                                 struct_v, struct_compat,
1151                                 OSDMAP_WRAPPER_COMPAT_VER, prefix);
1152                         return -EINVAL;
1153                 }
1154                 *p += 4; /* ignore wrapper struct_len */
1155
1156                 ceph_decode_8_safe(p, end, struct_v, e_inval);
1157                 ceph_decode_8_safe(p, end, struct_compat, e_inval);
1158                 if (struct_compat > OSDMAP_CLIENT_DATA_COMPAT_VER) {
1159                         pr_warn("got v %d cv %d > %d of %s ceph_osdmap client data\n",
1160                                 struct_v, struct_compat,
1161                                 OSDMAP_CLIENT_DATA_COMPAT_VER, prefix);
1162                         return -EINVAL;
1163                 }
1164                 *p += 4; /* ignore client data struct_len */
1165         } else {
1166                 u16 version;
1167
1168                 *p -= 1;
1169                 ceph_decode_16_safe(p, end, version, e_inval);
1170                 if (version < 6) {
1171                         pr_warn("got v %d < 6 of %s ceph_osdmap\n",
1172                                 version, prefix);
1173                         return -EINVAL;
1174                 }
1175
1176                 /* old osdmap enconding */
1177                 struct_v = 0;
1178         }
1179
1180         *v = struct_v;
1181         return 0;
1182
1183 e_inval:
1184         return -EINVAL;
1185 }
1186
1187 static int __decode_pools(void **p, void *end, struct ceph_osdmap *map,
1188                           bool incremental)
1189 {
1190         u32 n;
1191
1192         ceph_decode_32_safe(p, end, n, e_inval);
1193         while (n--) {
1194                 struct ceph_pg_pool_info *pi;
1195                 u64 pool;
1196                 int ret;
1197
1198                 ceph_decode_64_safe(p, end, pool, e_inval);
1199
1200                 pi = lookup_pg_pool(&map->pg_pools, pool);
1201                 if (!incremental || !pi) {
1202                         pi = kzalloc(sizeof(*pi), GFP_NOFS);
1203                         if (!pi)
1204                                 return -ENOMEM;
1205
1206                         RB_CLEAR_NODE(&pi->node);
1207                         pi->id = pool;
1208
1209                         if (!__insert_pg_pool(&map->pg_pools, pi)) {
1210                                 kfree(pi);
1211                                 return -EEXIST;
1212                         }
1213                 }
1214
1215                 ret = decode_pool(p, end, pi);
1216                 if (ret)
1217                         return ret;
1218         }
1219
1220         return 0;
1221
1222 e_inval:
1223         return -EINVAL;
1224 }
1225
1226 static int decode_pools(void **p, void *end, struct ceph_osdmap *map)
1227 {
1228         return __decode_pools(p, end, map, false);
1229 }
1230
1231 static int decode_new_pools(void **p, void *end, struct ceph_osdmap *map)
1232 {
1233         return __decode_pools(p, end, map, true);
1234 }
1235
1236 typedef struct ceph_pg_mapping *(*decode_mapping_fn_t)(void **, void *, bool);
1237
1238 static int decode_pg_mapping(void **p, void *end, struct rb_root *mapping_root,
1239                              decode_mapping_fn_t fn, bool incremental)
1240 {
1241         u32 n;
1242
1243         WARN_ON(!incremental && !fn);
1244
1245         ceph_decode_32_safe(p, end, n, e_inval);
1246         while (n--) {
1247                 struct ceph_pg_mapping *pg;
1248                 struct ceph_pg pgid;
1249                 int ret;
1250
1251                 ret = ceph_decode_pgid(p, end, &pgid);
1252                 if (ret)
1253                         return ret;
1254
1255                 pg = lookup_pg_mapping(mapping_root, &pgid);
1256                 if (pg) {
1257                         WARN_ON(!incremental);
1258                         erase_pg_mapping(mapping_root, pg);
1259                         free_pg_mapping(pg);
1260                 }
1261
1262                 if (fn) {
1263                         pg = fn(p, end, incremental);
1264                         if (IS_ERR(pg))
1265                                 return PTR_ERR(pg);
1266
1267                         if (pg) {
1268                                 pg->pgid = pgid; /* struct */
1269                                 insert_pg_mapping(mapping_root, pg);
1270                         }
1271                 }
1272         }
1273
1274         return 0;
1275
1276 e_inval:
1277         return -EINVAL;
1278 }
1279
1280 static struct ceph_pg_mapping *__decode_pg_temp(void **p, void *end,
1281                                                 bool incremental)
1282 {
1283         struct ceph_pg_mapping *pg;
1284         u32 len, i;
1285
1286         ceph_decode_32_safe(p, end, len, e_inval);
1287         if (len == 0 && incremental)
1288                 return NULL;    /* new_pg_temp: [] to remove */
1289         if (len > (SIZE_MAX - sizeof(*pg)) / sizeof(u32))
1290                 return ERR_PTR(-EINVAL);
1291
1292         ceph_decode_need(p, end, len * sizeof(u32), e_inval);
1293         pg = alloc_pg_mapping(len * sizeof(u32));
1294         if (!pg)
1295                 return ERR_PTR(-ENOMEM);
1296
1297         pg->pg_temp.len = len;
1298         for (i = 0; i < len; i++)
1299                 pg->pg_temp.osds[i] = ceph_decode_32(p);
1300
1301         return pg;
1302
1303 e_inval:
1304         return ERR_PTR(-EINVAL);
1305 }
1306
1307 static int decode_pg_temp(void **p, void *end, struct ceph_osdmap *map)
1308 {
1309         return decode_pg_mapping(p, end, &map->pg_temp, __decode_pg_temp,
1310                                  false);
1311 }
1312
1313 static int decode_new_pg_temp(void **p, void *end, struct ceph_osdmap *map)
1314 {
1315         return decode_pg_mapping(p, end, &map->pg_temp, __decode_pg_temp,
1316                                  true);
1317 }
1318
1319 static struct ceph_pg_mapping *__decode_primary_temp(void **p, void *end,
1320                                                      bool incremental)
1321 {
1322         struct ceph_pg_mapping *pg;
1323         u32 osd;
1324
1325         ceph_decode_32_safe(p, end, osd, e_inval);
1326         if (osd == (u32)-1 && incremental)
1327                 return NULL;    /* new_primary_temp: -1 to remove */
1328
1329         pg = alloc_pg_mapping(0);
1330         if (!pg)
1331                 return ERR_PTR(-ENOMEM);
1332
1333         pg->primary_temp.osd = osd;
1334         return pg;
1335
1336 e_inval:
1337         return ERR_PTR(-EINVAL);
1338 }
1339
1340 static int decode_primary_temp(void **p, void *end, struct ceph_osdmap *map)
1341 {
1342         return decode_pg_mapping(p, end, &map->primary_temp,
1343                                  __decode_primary_temp, false);
1344 }
1345
1346 static int decode_new_primary_temp(void **p, void *end,
1347                                    struct ceph_osdmap *map)
1348 {
1349         return decode_pg_mapping(p, end, &map->primary_temp,
1350                                  __decode_primary_temp, true);
1351 }
1352
1353 u32 ceph_get_primary_affinity(struct ceph_osdmap *map, int osd)
1354 {
1355         BUG_ON(osd >= map->max_osd);
1356
1357         if (!map->osd_primary_affinity)
1358                 return CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1359
1360         return map->osd_primary_affinity[osd];
1361 }
1362
1363 static int set_primary_affinity(struct ceph_osdmap *map, int osd, u32 aff)
1364 {
1365         BUG_ON(osd >= map->max_osd);
1366
1367         if (!map->osd_primary_affinity) {
1368                 int i;
1369
1370                 map->osd_primary_affinity = ceph_kvmalloc(
1371                     array_size(map->max_osd, sizeof(*map->osd_primary_affinity)),
1372                     GFP_NOFS);
1373                 if (!map->osd_primary_affinity)
1374                         return -ENOMEM;
1375
1376                 for (i = 0; i < map->max_osd; i++)
1377                         map->osd_primary_affinity[i] =
1378                             CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
1379         }
1380
1381         map->osd_primary_affinity[osd] = aff;
1382
1383         return 0;
1384 }
1385
1386 static int decode_primary_affinity(void **p, void *end,
1387                                    struct ceph_osdmap *map)
1388 {
1389         u32 len, i;
1390
1391         ceph_decode_32_safe(p, end, len, e_inval);
1392         if (len == 0) {
1393                 kvfree(map->osd_primary_affinity);
1394                 map->osd_primary_affinity = NULL;
1395                 return 0;
1396         }
1397         if (len != map->max_osd)
1398                 goto e_inval;
1399
1400         ceph_decode_need(p, end, map->max_osd*sizeof(u32), e_inval);
1401
1402         for (i = 0; i < map->max_osd; i++) {
1403                 int ret;
1404
1405                 ret = set_primary_affinity(map, i, ceph_decode_32(p));
1406                 if (ret)
1407                         return ret;
1408         }
1409
1410         return 0;
1411
1412 e_inval:
1413         return -EINVAL;
1414 }
1415
1416 static int decode_new_primary_affinity(void **p, void *end,
1417                                        struct ceph_osdmap *map)
1418 {
1419         u32 n;
1420
1421         ceph_decode_32_safe(p, end, n, e_inval);
1422         while (n--) {
1423                 u32 osd, aff;
1424                 int ret;
1425
1426                 ceph_decode_32_safe(p, end, osd, e_inval);
1427                 ceph_decode_32_safe(p, end, aff, e_inval);
1428
1429                 ret = set_primary_affinity(map, osd, aff);
1430                 if (ret)
1431                         return ret;
1432
1433                 pr_info("osd%d primary-affinity 0x%x\n", osd, aff);
1434         }
1435
1436         return 0;
1437
1438 e_inval:
1439         return -EINVAL;
1440 }
1441
1442 static struct ceph_pg_mapping *__decode_pg_upmap(void **p, void *end,
1443                                                  bool __unused)
1444 {
1445         return __decode_pg_temp(p, end, false);
1446 }
1447
1448 static int decode_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1449 {
1450         return decode_pg_mapping(p, end, &map->pg_upmap, __decode_pg_upmap,
1451                                  false);
1452 }
1453
1454 static int decode_new_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1455 {
1456         return decode_pg_mapping(p, end, &map->pg_upmap, __decode_pg_upmap,
1457                                  true);
1458 }
1459
1460 static int decode_old_pg_upmap(void **p, void *end, struct ceph_osdmap *map)
1461 {
1462         return decode_pg_mapping(p, end, &map->pg_upmap, NULL, true);
1463 }
1464
1465 static struct ceph_pg_mapping *__decode_pg_upmap_items(void **p, void *end,
1466                                                        bool __unused)
1467 {
1468         struct ceph_pg_mapping *pg;
1469         u32 len, i;
1470
1471         ceph_decode_32_safe(p, end, len, e_inval);
1472         if (len > (SIZE_MAX - sizeof(*pg)) / (2 * sizeof(u32)))
1473                 return ERR_PTR(-EINVAL);
1474
1475         ceph_decode_need(p, end, 2 * len * sizeof(u32), e_inval);
1476         pg = alloc_pg_mapping(2 * len * sizeof(u32));
1477         if (!pg)
1478                 return ERR_PTR(-ENOMEM);
1479
1480         pg->pg_upmap_items.len = len;
1481         for (i = 0; i < len; i++) {
1482                 pg->pg_upmap_items.from_to[i][0] = ceph_decode_32(p);
1483                 pg->pg_upmap_items.from_to[i][1] = ceph_decode_32(p);
1484         }
1485
1486         return pg;
1487
1488 e_inval:
1489         return ERR_PTR(-EINVAL);
1490 }
1491
1492 static int decode_pg_upmap_items(void **p, void *end, struct ceph_osdmap *map)
1493 {
1494         return decode_pg_mapping(p, end, &map->pg_upmap_items,
1495                                  __decode_pg_upmap_items, false);
1496 }
1497
1498 static int decode_new_pg_upmap_items(void **p, void *end,
1499                                      struct ceph_osdmap *map)
1500 {
1501         return decode_pg_mapping(p, end, &map->pg_upmap_items,
1502                                  __decode_pg_upmap_items, true);
1503 }
1504
1505 static int decode_old_pg_upmap_items(void **p, void *end,
1506                                      struct ceph_osdmap *map)
1507 {
1508         return decode_pg_mapping(p, end, &map->pg_upmap_items, NULL, true);
1509 }
1510
1511 /*
1512  * decode a full map.
1513  */
1514 static int osdmap_decode(void **p, void *end, struct ceph_osdmap *map)
1515 {
1516         u8 struct_v;
1517         u32 epoch = 0;
1518         void *start = *p;
1519         u32 max;
1520         u32 len, i;
1521         int err;
1522
1523         dout("%s %p to %p len %d\n", __func__, *p, end, (int)(end - *p));
1524
1525         err = get_osdmap_client_data_v(p, end, "full", &struct_v);
1526         if (err)
1527                 goto bad;
1528
1529         /* fsid, epoch, created, modified */
1530         ceph_decode_need(p, end, sizeof(map->fsid) + sizeof(u32) +
1531                          sizeof(map->created) + sizeof(map->modified), e_inval);
1532         ceph_decode_copy(p, &map->fsid, sizeof(map->fsid));
1533         epoch = map->epoch = ceph_decode_32(p);
1534         ceph_decode_copy(p, &map->created, sizeof(map->created));
1535         ceph_decode_copy(p, &map->modified, sizeof(map->modified));
1536
1537         /* pools */
1538         err = decode_pools(p, end, map);
1539         if (err)
1540                 goto bad;
1541
1542         /* pool_name */
1543         err = decode_pool_names(p, end, map);
1544         if (err)
1545                 goto bad;
1546
1547         ceph_decode_32_safe(p, end, map->pool_max, e_inval);
1548
1549         ceph_decode_32_safe(p, end, map->flags, e_inval);
1550
1551         /* max_osd */
1552         ceph_decode_32_safe(p, end, max, e_inval);
1553
1554         /* (re)alloc osd arrays */
1555         err = osdmap_set_max_osd(map, max);
1556         if (err)
1557                 goto bad;
1558
1559         /* osd_state, osd_weight, osd_addrs->client_addr */
1560         ceph_decode_need(p, end, 3*sizeof(u32) +
1561                          map->max_osd*(struct_v >= 5 ? sizeof(u32) :
1562                                                        sizeof(u8)) +
1563                                        sizeof(*map->osd_weight), e_inval);
1564         if (ceph_decode_32(p) != map->max_osd)
1565                 goto e_inval;
1566
1567         if (struct_v >= 5) {
1568                 for (i = 0; i < map->max_osd; i++)
1569                         map->osd_state[i] = ceph_decode_32(p);
1570         } else {
1571                 for (i = 0; i < map->max_osd; i++)
1572                         map->osd_state[i] = ceph_decode_8(p);
1573         }
1574
1575         if (ceph_decode_32(p) != map->max_osd)
1576                 goto e_inval;
1577
1578         for (i = 0; i < map->max_osd; i++)
1579                 map->osd_weight[i] = ceph_decode_32(p);
1580
1581         if (ceph_decode_32(p) != map->max_osd)
1582                 goto e_inval;
1583
1584         for (i = 0; i < map->max_osd; i++) {
1585                 err = ceph_decode_entity_addr(p, end, &map->osd_addr[i]);
1586                 if (err)
1587                         goto bad;
1588         }
1589
1590         /* pg_temp */
1591         err = decode_pg_temp(p, end, map);
1592         if (err)
1593                 goto bad;
1594
1595         /* primary_temp */
1596         if (struct_v >= 1) {
1597                 err = decode_primary_temp(p, end, map);
1598                 if (err)
1599                         goto bad;
1600         }
1601
1602         /* primary_affinity */
1603         if (struct_v >= 2) {
1604                 err = decode_primary_affinity(p, end, map);
1605                 if (err)
1606                         goto bad;
1607         } else {
1608                 WARN_ON(map->osd_primary_affinity);
1609         }
1610
1611         /* crush */
1612         ceph_decode_32_safe(p, end, len, e_inval);
1613         err = osdmap_set_crush(map, crush_decode(*p, min(*p + len, end)));
1614         if (err)
1615                 goto bad;
1616
1617         *p += len;
1618         if (struct_v >= 3) {
1619                 /* erasure_code_profiles */
1620                 ceph_decode_skip_map_of_map(p, end, string, string, string,
1621                                             e_inval);
1622         }
1623
1624         if (struct_v >= 4) {
1625                 err = decode_pg_upmap(p, end, map);
1626                 if (err)
1627                         goto bad;
1628
1629                 err = decode_pg_upmap_items(p, end, map);
1630                 if (err)
1631                         goto bad;
1632         } else {
1633                 WARN_ON(!RB_EMPTY_ROOT(&map->pg_upmap));
1634                 WARN_ON(!RB_EMPTY_ROOT(&map->pg_upmap_items));
1635         }
1636
1637         /* ignore the rest */
1638         *p = end;
1639
1640         dout("full osdmap epoch %d max_osd %d\n", map->epoch, map->max_osd);
1641         return 0;
1642
1643 e_inval:
1644         err = -EINVAL;
1645 bad:
1646         pr_err("corrupt full osdmap (%d) epoch %d off %d (%p of %p-%p)\n",
1647                err, epoch, (int)(*p - start), *p, start, end);
1648         print_hex_dump(KERN_DEBUG, "osdmap: ",
1649                        DUMP_PREFIX_OFFSET, 16, 1,
1650                        start, end - start, true);
1651         return err;
1652 }
1653
1654 /*
1655  * Allocate and decode a full map.
1656  */
1657 struct ceph_osdmap *ceph_osdmap_decode(void **p, void *end)
1658 {
1659         struct ceph_osdmap *map;
1660         int ret;
1661
1662         map = ceph_osdmap_alloc();
1663         if (!map)
1664                 return ERR_PTR(-ENOMEM);
1665
1666         ret = osdmap_decode(p, end, map);
1667         if (ret) {
1668                 ceph_osdmap_destroy(map);
1669                 return ERR_PTR(ret);
1670         }
1671
1672         return map;
1673 }
1674
1675 /*
1676  * Encoding order is (new_up_client, new_state, new_weight).  Need to
1677  * apply in the (new_weight, new_state, new_up_client) order, because
1678  * an incremental map may look like e.g.
1679  *
1680  *     new_up_client: { osd=6, addr=... } # set osd_state and addr
1681  *     new_state: { osd=6, xorstate=EXISTS } # clear osd_state
1682  */
1683 static int decode_new_up_state_weight(void **p, void *end, u8 struct_v,
1684                                       struct ceph_osdmap *map)
1685 {
1686         void *new_up_client;
1687         void *new_state;
1688         void *new_weight_end;
1689         u32 len;
1690         int i;
1691
1692         new_up_client = *p;
1693         ceph_decode_32_safe(p, end, len, e_inval);
1694         for (i = 0; i < len; ++i) {
1695                 struct ceph_entity_addr addr;
1696
1697                 ceph_decode_skip_32(p, end, e_inval);
1698                 if (ceph_decode_entity_addr(p, end, &addr))
1699                         goto e_inval;
1700         }
1701
1702         new_state = *p;
1703         ceph_decode_32_safe(p, end, len, e_inval);
1704         len *= sizeof(u32) + (struct_v >= 5 ? sizeof(u32) : sizeof(u8));
1705         ceph_decode_need(p, end, len, e_inval);
1706         *p += len;
1707
1708         /* new_weight */
1709         ceph_decode_32_safe(p, end, len, e_inval);
1710         while (len--) {
1711                 s32 osd;
1712                 u32 w;
1713
1714                 ceph_decode_need(p, end, 2*sizeof(u32), e_inval);
1715                 osd = ceph_decode_32(p);
1716                 w = ceph_decode_32(p);
1717                 BUG_ON(osd >= map->max_osd);
1718                 pr_info("osd%d weight 0x%x %s\n", osd, w,
1719                      w == CEPH_OSD_IN ? "(in)" :
1720                      (w == CEPH_OSD_OUT ? "(out)" : ""));
1721                 map->osd_weight[osd] = w;
1722
1723                 /*
1724                  * If we are marking in, set the EXISTS, and clear the
1725                  * AUTOOUT and NEW bits.
1726                  */
1727                 if (w) {
1728                         map->osd_state[osd] |= CEPH_OSD_EXISTS;
1729                         map->osd_state[osd] &= ~(CEPH_OSD_AUTOOUT |
1730                                                  CEPH_OSD_NEW);
1731                 }
1732         }
1733         new_weight_end = *p;
1734
1735         /* new_state (up/down) */
1736         *p = new_state;
1737         len = ceph_decode_32(p);
1738         while (len--) {
1739                 s32 osd;
1740                 u32 xorstate;
1741                 int ret;
1742
1743                 osd = ceph_decode_32(p);
1744                 if (struct_v >= 5)
1745                         xorstate = ceph_decode_32(p);
1746                 else
1747                         xorstate = ceph_decode_8(p);
1748                 if (xorstate == 0)
1749                         xorstate = CEPH_OSD_UP;
1750                 BUG_ON(osd >= map->max_osd);
1751                 if ((map->osd_state[osd] & CEPH_OSD_UP) &&
1752                     (xorstate & CEPH_OSD_UP))
1753                         pr_info("osd%d down\n", osd);
1754                 if ((map->osd_state[osd] & CEPH_OSD_EXISTS) &&
1755                     (xorstate & CEPH_OSD_EXISTS)) {
1756                         pr_info("osd%d does not exist\n", osd);
1757                         ret = set_primary_affinity(map, osd,
1758                                                    CEPH_OSD_DEFAULT_PRIMARY_AFFINITY);
1759                         if (ret)
1760                                 return ret;
1761                         memset(map->osd_addr + osd, 0, sizeof(*map->osd_addr));
1762                         map->osd_state[osd] = 0;
1763                 } else {
1764                         map->osd_state[osd] ^= xorstate;
1765                 }
1766         }
1767
1768         /* new_up_client */
1769         *p = new_up_client;
1770         len = ceph_decode_32(p);
1771         while (len--) {
1772                 s32 osd;
1773                 struct ceph_entity_addr addr;
1774
1775                 osd = ceph_decode_32(p);
1776                 BUG_ON(osd >= map->max_osd);
1777                 if (ceph_decode_entity_addr(p, end, &addr))
1778                         goto e_inval;
1779                 pr_info("osd%d up\n", osd);
1780                 map->osd_state[osd] |= CEPH_OSD_EXISTS | CEPH_OSD_UP;
1781                 map->osd_addr[osd] = addr;
1782         }
1783
1784         *p = new_weight_end;
1785         return 0;
1786
1787 e_inval:
1788         return -EINVAL;
1789 }
1790
1791 /*
1792  * decode and apply an incremental map update.
1793  */
1794 struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
1795                                              struct ceph_osdmap *map)
1796 {
1797         struct ceph_fsid fsid;
1798         u32 epoch = 0;
1799         struct ceph_timespec modified;
1800         s32 len;
1801         u64 pool;
1802         __s64 new_pool_max;
1803         __s32 new_flags, max;
1804         void *start = *p;
1805         int err;
1806         u8 struct_v;
1807
1808         dout("%s %p to %p len %d\n", __func__, *p, end, (int)(end - *p));
1809
1810         err = get_osdmap_client_data_v(p, end, "inc", &struct_v);
1811         if (err)
1812                 goto bad;
1813
1814         /* fsid, epoch, modified, new_pool_max, new_flags */
1815         ceph_decode_need(p, end, sizeof(fsid) + sizeof(u32) + sizeof(modified) +
1816                          sizeof(u64) + sizeof(u32), e_inval);
1817         ceph_decode_copy(p, &fsid, sizeof(fsid));
1818         epoch = ceph_decode_32(p);
1819         BUG_ON(epoch != map->epoch+1);
1820         ceph_decode_copy(p, &modified, sizeof(modified));
1821         new_pool_max = ceph_decode_64(p);
1822         new_flags = ceph_decode_32(p);
1823
1824         /* full map? */
1825         ceph_decode_32_safe(p, end, len, e_inval);
1826         if (len > 0) {
1827                 dout("apply_incremental full map len %d, %p to %p\n",
1828                      len, *p, end);
1829                 return ceph_osdmap_decode(p, min(*p+len, end));
1830         }
1831
1832         /* new crush? */
1833         ceph_decode_32_safe(p, end, len, e_inval);
1834         if (len > 0) {
1835                 err = osdmap_set_crush(map,
1836                                        crush_decode(*p, min(*p + len, end)));
1837                 if (err)
1838                         goto bad;
1839                 *p += len;
1840         }
1841
1842         /* new flags? */
1843         if (new_flags >= 0)
1844                 map->flags = new_flags;
1845         if (new_pool_max >= 0)
1846                 map->pool_max = new_pool_max;
1847
1848         /* new max? */
1849         ceph_decode_32_safe(p, end, max, e_inval);
1850         if (max >= 0) {
1851                 err = osdmap_set_max_osd(map, max);
1852                 if (err)
1853                         goto bad;
1854         }
1855
1856         map->epoch++;
1857         map->modified = modified;
1858
1859         /* new_pools */
1860         err = decode_new_pools(p, end, map);
1861         if (err)
1862                 goto bad;
1863
1864         /* new_pool_names */
1865         err = decode_pool_names(p, end, map);
1866         if (err)
1867                 goto bad;
1868
1869         /* old_pool */
1870         ceph_decode_32_safe(p, end, len, e_inval);
1871         while (len--) {
1872                 struct ceph_pg_pool_info *pi;
1873
1874                 ceph_decode_64_safe(p, end, pool, e_inval);
1875                 pi = lookup_pg_pool(&map->pg_pools, pool);
1876                 if (pi)
1877                         __remove_pg_pool(&map->pg_pools, pi);
1878         }
1879
1880         /* new_up_client, new_state, new_weight */
1881         err = decode_new_up_state_weight(p, end, struct_v, map);
1882         if (err)
1883                 goto bad;
1884
1885         /* new_pg_temp */
1886         err = decode_new_pg_temp(p, end, map);
1887         if (err)
1888                 goto bad;
1889
1890         /* new_primary_temp */
1891         if (struct_v >= 1) {
1892                 err = decode_new_primary_temp(p, end, map);
1893                 if (err)
1894                         goto bad;
1895         }
1896
1897         /* new_primary_affinity */
1898         if (struct_v >= 2) {
1899                 err = decode_new_primary_affinity(p, end, map);
1900                 if (err)
1901                         goto bad;
1902         }
1903
1904         if (struct_v >= 3) {
1905                 /* new_erasure_code_profiles */
1906                 ceph_decode_skip_map_of_map(p, end, string, string, string,
1907                                             e_inval);
1908                 /* old_erasure_code_profiles */
1909                 ceph_decode_skip_set(p, end, string, e_inval);
1910         }
1911
1912         if (struct_v >= 4) {
1913                 err = decode_new_pg_upmap(p, end, map);
1914                 if (err)
1915                         goto bad;
1916
1917                 err = decode_old_pg_upmap(p, end, map);
1918                 if (err)
1919                         goto bad;
1920
1921                 err = decode_new_pg_upmap_items(p, end, map);
1922                 if (err)
1923                         goto bad;
1924
1925                 err = decode_old_pg_upmap_items(p, end, map);
1926                 if (err)
1927                         goto bad;
1928         }
1929
1930         /* ignore the rest */
1931         *p = end;
1932
1933         dout("inc osdmap epoch %d max_osd %d\n", map->epoch, map->max_osd);
1934         return map;
1935
1936 e_inval:
1937         err = -EINVAL;
1938 bad:
1939         pr_err("corrupt inc osdmap (%d) epoch %d off %d (%p of %p-%p)\n",
1940                err, epoch, (int)(*p - start), *p, start, end);
1941         print_hex_dump(KERN_DEBUG, "osdmap: ",
1942                        DUMP_PREFIX_OFFSET, 16, 1,
1943                        start, end - start, true);
1944         return ERR_PTR(err);
1945 }
1946
1947 void ceph_oloc_copy(struct ceph_object_locator *dest,
1948                     const struct ceph_object_locator *src)
1949 {
1950         ceph_oloc_destroy(dest);
1951
1952         dest->pool = src->pool;
1953         if (src->pool_ns)
1954                 dest->pool_ns = ceph_get_string(src->pool_ns);
1955         else
1956                 dest->pool_ns = NULL;
1957 }
1958 EXPORT_SYMBOL(ceph_oloc_copy);
1959
1960 void ceph_oloc_destroy(struct ceph_object_locator *oloc)
1961 {
1962         ceph_put_string(oloc->pool_ns);
1963 }
1964 EXPORT_SYMBOL(ceph_oloc_destroy);
1965
1966 void ceph_oid_copy(struct ceph_object_id *dest,
1967                    const struct ceph_object_id *src)
1968 {
1969         ceph_oid_destroy(dest);
1970
1971         if (src->name != src->inline_name) {
1972                 /* very rare, see ceph_object_id definition */
1973                 dest->name = kmalloc(src->name_len + 1,
1974                                      GFP_NOIO | __GFP_NOFAIL);
1975         } else {
1976                 dest->name = dest->inline_name;
1977         }
1978         memcpy(dest->name, src->name, src->name_len + 1);
1979         dest->name_len = src->name_len;
1980 }
1981 EXPORT_SYMBOL(ceph_oid_copy);
1982
1983 static __printf(2, 0)
1984 int oid_printf_vargs(struct ceph_object_id *oid, const char *fmt, va_list ap)
1985 {
1986         int len;
1987
1988         WARN_ON(!ceph_oid_empty(oid));
1989
1990         len = vsnprintf(oid->inline_name, sizeof(oid->inline_name), fmt, ap);
1991         if (len >= sizeof(oid->inline_name))
1992                 return len;
1993
1994         oid->name_len = len;
1995         return 0;
1996 }
1997
1998 /*
1999  * If oid doesn't fit into inline buffer, BUG.
2000  */
2001 void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...)
2002 {
2003         va_list ap;
2004
2005         va_start(ap, fmt);
2006         BUG_ON(oid_printf_vargs(oid, fmt, ap));
2007         va_end(ap);
2008 }
2009 EXPORT_SYMBOL(ceph_oid_printf);
2010
2011 static __printf(3, 0)
2012 int oid_aprintf_vargs(struct ceph_object_id *oid, gfp_t gfp,
2013                       const char *fmt, va_list ap)
2014 {
2015         va_list aq;
2016         int len;
2017
2018         va_copy(aq, ap);
2019         len = oid_printf_vargs(oid, fmt, aq);
2020         va_end(aq);
2021
2022         if (len) {
2023                 char *external_name;
2024
2025                 external_name = kmalloc(len + 1, gfp);
2026                 if (!external_name)
2027                         return -ENOMEM;
2028
2029                 oid->name = external_name;
2030                 WARN_ON(vsnprintf(oid->name, len + 1, fmt, ap) != len);
2031                 oid->name_len = len;
2032         }
2033
2034         return 0;
2035 }
2036
2037 /*
2038  * If oid doesn't fit into inline buffer, allocate.
2039  */
2040 int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp,
2041                      const char *fmt, ...)
2042 {
2043         va_list ap;
2044         int ret;
2045
2046         va_start(ap, fmt);
2047         ret = oid_aprintf_vargs(oid, gfp, fmt, ap);
2048         va_end(ap);
2049
2050         return ret;
2051 }
2052 EXPORT_SYMBOL(ceph_oid_aprintf);
2053
2054 void ceph_oid_destroy(struct ceph_object_id *oid)
2055 {
2056         if (oid->name != oid->inline_name)
2057                 kfree(oid->name);
2058 }
2059 EXPORT_SYMBOL(ceph_oid_destroy);
2060
2061 /*
2062  * osds only
2063  */
2064 static bool __osds_equal(const struct ceph_osds *lhs,
2065                          const struct ceph_osds *rhs)
2066 {
2067         if (lhs->size == rhs->size &&
2068             !memcmp(lhs->osds, rhs->osds, rhs->size * sizeof(rhs->osds[0])))
2069                 return true;
2070
2071         return false;
2072 }
2073
2074 /*
2075  * osds + primary
2076  */
2077 static bool osds_equal(const struct ceph_osds *lhs,
2078                        const struct ceph_osds *rhs)
2079 {
2080         if (__osds_equal(lhs, rhs) &&
2081             lhs->primary == rhs->primary)
2082                 return true;
2083
2084         return false;
2085 }
2086
2087 static bool osds_valid(const struct ceph_osds *set)
2088 {
2089         /* non-empty set */
2090         if (set->size > 0 && set->primary >= 0)
2091                 return true;
2092
2093         /* empty can_shift_osds set */
2094         if (!set->size && set->primary == -1)
2095                 return true;
2096
2097         /* empty !can_shift_osds set - all NONE */
2098         if (set->size > 0 && set->primary == -1) {
2099                 int i;
2100
2101                 for (i = 0; i < set->size; i++) {
2102                         if (set->osds[i] != CRUSH_ITEM_NONE)
2103                                 break;
2104                 }
2105                 if (i == set->size)
2106                         return true;
2107         }
2108
2109         return false;
2110 }
2111
2112 void ceph_osds_copy(struct ceph_osds *dest, const struct ceph_osds *src)
2113 {
2114         memcpy(dest->osds, src->osds, src->size * sizeof(src->osds[0]));
2115         dest->size = src->size;
2116         dest->primary = src->primary;
2117 }
2118
2119 bool ceph_pg_is_split(const struct ceph_pg *pgid, u32 old_pg_num,
2120                       u32 new_pg_num)
2121 {
2122         int old_bits = calc_bits_of(old_pg_num);
2123         int old_mask = (1 << old_bits) - 1;
2124         int n;
2125
2126         WARN_ON(pgid->seed >= old_pg_num);
2127         if (new_pg_num <= old_pg_num)
2128                 return false;
2129
2130         for (n = 1; ; n++) {
2131                 int next_bit = n << (old_bits - 1);
2132                 u32 s = next_bit | pgid->seed;
2133
2134                 if (s < old_pg_num || s == pgid->seed)
2135                         continue;
2136                 if (s >= new_pg_num)
2137                         break;
2138
2139                 s = ceph_stable_mod(s, old_pg_num, old_mask);
2140                 if (s == pgid->seed)
2141                         return true;
2142         }
2143
2144         return false;
2145 }
2146
2147 bool ceph_is_new_interval(const struct ceph_osds *old_acting,
2148                           const struct ceph_osds *new_acting,
2149                           const struct ceph_osds *old_up,
2150                           const struct ceph_osds *new_up,
2151                           int old_size,
2152                           int new_size,
2153                           int old_min_size,
2154                           int new_min_size,
2155                           u32 old_pg_num,
2156                           u32 new_pg_num,
2157                           bool old_sort_bitwise,
2158                           bool new_sort_bitwise,
2159                           bool old_recovery_deletes,
2160                           bool new_recovery_deletes,
2161                           const struct ceph_pg *pgid)
2162 {
2163         return !osds_equal(old_acting, new_acting) ||
2164                !osds_equal(old_up, new_up) ||
2165                old_size != new_size ||
2166                old_min_size != new_min_size ||
2167                ceph_pg_is_split(pgid, old_pg_num, new_pg_num) ||
2168                old_sort_bitwise != new_sort_bitwise ||
2169                old_recovery_deletes != new_recovery_deletes;
2170 }
2171
2172 static int calc_pg_rank(int osd, const struct ceph_osds *acting)
2173 {
2174         int i;
2175
2176         for (i = 0; i < acting->size; i++) {
2177                 if (acting->osds[i] == osd)
2178                         return i;
2179         }
2180
2181         return -1;
2182 }
2183
2184 static bool primary_changed(const struct ceph_osds *old_acting,
2185                             const struct ceph_osds *new_acting)
2186 {
2187         if (!old_acting->size && !new_acting->size)
2188                 return false; /* both still empty */
2189
2190         if (!old_acting->size ^ !new_acting->size)
2191                 return true; /* was empty, now not, or vice versa */
2192
2193         if (old_acting->primary != new_acting->primary)
2194                 return true; /* primary changed */
2195
2196         if (calc_pg_rank(old_acting->primary, old_acting) !=
2197             calc_pg_rank(new_acting->primary, new_acting))
2198                 return true;
2199
2200         return false; /* same primary (tho replicas may have changed) */
2201 }
2202
2203 bool ceph_osds_changed(const struct ceph_osds *old_acting,
2204                        const struct ceph_osds *new_acting,
2205                        bool any_change)
2206 {
2207         if (primary_changed(old_acting, new_acting))
2208                 return true;
2209
2210         if (any_change && !__osds_equal(old_acting, new_acting))
2211                 return true;
2212
2213         return false;
2214 }
2215
2216 /*
2217  * Map an object into a PG.
2218  *
2219  * Should only be called with target_oid and target_oloc (as opposed to
2220  * base_oid and base_oloc), since tiering isn't taken into account.
2221  */
2222 void __ceph_object_locator_to_pg(struct ceph_pg_pool_info *pi,
2223                                  const struct ceph_object_id *oid,
2224                                  const struct ceph_object_locator *oloc,
2225                                  struct ceph_pg *raw_pgid)
2226 {
2227         WARN_ON(pi->id != oloc->pool);
2228
2229         if (!oloc->pool_ns) {
2230                 raw_pgid->pool = oloc->pool;
2231                 raw_pgid->seed = ceph_str_hash(pi->object_hash, oid->name,
2232                                              oid->name_len);
2233                 dout("%s %s -> raw_pgid %llu.%x\n", __func__, oid->name,
2234                      raw_pgid->pool, raw_pgid->seed);
2235         } else {
2236                 char stack_buf[256];
2237                 char *buf = stack_buf;
2238                 int nsl = oloc->pool_ns->len;
2239                 size_t total = nsl + 1 + oid->name_len;
2240
2241                 if (total > sizeof(stack_buf))
2242                         buf = kmalloc(total, GFP_NOIO | __GFP_NOFAIL);
2243                 memcpy(buf, oloc->pool_ns->str, nsl);
2244                 buf[nsl] = '\037';
2245                 memcpy(buf + nsl + 1, oid->name, oid->name_len);
2246                 raw_pgid->pool = oloc->pool;
2247                 raw_pgid->seed = ceph_str_hash(pi->object_hash, buf, total);
2248                 if (buf != stack_buf)
2249                         kfree(buf);
2250                 dout("%s %s ns %.*s -> raw_pgid %llu.%x\n", __func__,
2251                      oid->name, nsl, oloc->pool_ns->str,
2252                      raw_pgid->pool, raw_pgid->seed);
2253         }
2254 }
2255
2256 int ceph_object_locator_to_pg(struct ceph_osdmap *osdmap,
2257                               const struct ceph_object_id *oid,
2258                               const struct ceph_object_locator *oloc,
2259                               struct ceph_pg *raw_pgid)
2260 {
2261         struct ceph_pg_pool_info *pi;
2262
2263         pi = ceph_pg_pool_by_id(osdmap, oloc->pool);
2264         if (!pi)
2265                 return -ENOENT;
2266
2267         __ceph_object_locator_to_pg(pi, oid, oloc, raw_pgid);
2268         return 0;
2269 }
2270 EXPORT_SYMBOL(ceph_object_locator_to_pg);
2271
2272 /*
2273  * Map a raw PG (full precision ps) into an actual PG.
2274  */
2275 static void raw_pg_to_pg(struct ceph_pg_pool_info *pi,
2276                          const struct ceph_pg *raw_pgid,
2277                          struct ceph_pg *pgid)
2278 {
2279         pgid->pool = raw_pgid->pool;
2280         pgid->seed = ceph_stable_mod(raw_pgid->seed, pi->pg_num,
2281                                      pi->pg_num_mask);
2282 }
2283
2284 /*
2285  * Map a raw PG (full precision ps) into a placement ps (placement
2286  * seed).  Include pool id in that value so that different pools don't
2287  * use the same seeds.
2288  */
2289 static u32 raw_pg_to_pps(struct ceph_pg_pool_info *pi,
2290                          const struct ceph_pg *raw_pgid)
2291 {
2292         if (pi->flags & CEPH_POOL_FLAG_HASHPSPOOL) {
2293                 /* hash pool id and seed so that pool PGs do not overlap */
2294                 return crush_hash32_2(CRUSH_HASH_RJENKINS1,
2295                                       ceph_stable_mod(raw_pgid->seed,
2296                                                       pi->pgp_num,
2297                                                       pi->pgp_num_mask),
2298                                       raw_pgid->pool);
2299         } else {
2300                 /*
2301                  * legacy behavior: add ps and pool together.  this is
2302                  * not a great approach because the PGs from each pool
2303                  * will overlap on top of each other: 0.5 == 1.4 ==
2304                  * 2.3 == ...
2305                  */
2306                 return ceph_stable_mod(raw_pgid->seed, pi->pgp_num,
2307                                        pi->pgp_num_mask) +
2308                        (unsigned)raw_pgid->pool;
2309         }
2310 }
2311
2312 /*
2313  * Magic value used for a "default" fallback choose_args, used if the
2314  * crush_choose_arg_map passed to do_crush() does not exist.  If this
2315  * also doesn't exist, fall back to canonical weights.
2316  */
2317 #define CEPH_DEFAULT_CHOOSE_ARGS        -1
2318
2319 static int do_crush(struct ceph_osdmap *map, int ruleno, int x,
2320                     int *result, int result_max,
2321                     const __u32 *weight, int weight_max,
2322                     s64 choose_args_index)
2323 {
2324         struct crush_choose_arg_map *arg_map;
2325         int r;
2326
2327         BUG_ON(result_max > CEPH_PG_MAX_SIZE);
2328
2329         arg_map = lookup_choose_arg_map(&map->crush->choose_args,
2330                                         choose_args_index);
2331         if (!arg_map)
2332                 arg_map = lookup_choose_arg_map(&map->crush->choose_args,
2333                                                 CEPH_DEFAULT_CHOOSE_ARGS);
2334
2335         mutex_lock(&map->crush_workspace_mutex);
2336         r = crush_do_rule(map->crush, ruleno, x, result, result_max,
2337                           weight, weight_max, map->crush_workspace,
2338                           arg_map ? arg_map->args : NULL);
2339         mutex_unlock(&map->crush_workspace_mutex);
2340
2341         return r;
2342 }
2343
2344 static void remove_nonexistent_osds(struct ceph_osdmap *osdmap,
2345                                     struct ceph_pg_pool_info *pi,
2346                                     struct ceph_osds *set)
2347 {
2348         int i;
2349
2350         if (ceph_can_shift_osds(pi)) {
2351                 int removed = 0;
2352
2353                 /* shift left */
2354                 for (i = 0; i < set->size; i++) {
2355                         if (!ceph_osd_exists(osdmap, set->osds[i])) {
2356                                 removed++;
2357                                 continue;
2358                         }
2359                         if (removed)
2360                                 set->osds[i - removed] = set->osds[i];
2361                 }
2362                 set->size -= removed;
2363         } else {
2364                 /* set dne devices to NONE */
2365                 for (i = 0; i < set->size; i++) {
2366                         if (!ceph_osd_exists(osdmap, set->osds[i]))
2367                                 set->osds[i] = CRUSH_ITEM_NONE;
2368                 }
2369         }
2370 }
2371
2372 /*
2373  * Calculate raw set (CRUSH output) for given PG and filter out
2374  * nonexistent OSDs.  ->primary is undefined for a raw set.
2375  *
2376  * Placement seed (CRUSH input) is returned through @ppps.
2377  */
2378 static void pg_to_raw_osds(struct ceph_osdmap *osdmap,
2379                            struct ceph_pg_pool_info *pi,
2380                            const struct ceph_pg *raw_pgid,
2381                            struct ceph_osds *raw,
2382                            u32 *ppps)
2383 {
2384         u32 pps = raw_pg_to_pps(pi, raw_pgid);
2385         int ruleno;
2386         int len;
2387
2388         ceph_osds_init(raw);
2389         if (ppps)
2390                 *ppps = pps;
2391
2392         ruleno = crush_find_rule(osdmap->crush, pi->crush_ruleset, pi->type,
2393                                  pi->size);
2394         if (ruleno < 0) {
2395                 pr_err("no crush rule: pool %lld ruleset %d type %d size %d\n",
2396                        pi->id, pi->crush_ruleset, pi->type, pi->size);
2397                 return;
2398         }
2399
2400         if (pi->size > ARRAY_SIZE(raw->osds)) {
2401                 pr_err_ratelimited("pool %lld ruleset %d type %d too wide: size %d > %zu\n",
2402                        pi->id, pi->crush_ruleset, pi->type, pi->size,
2403                        ARRAY_SIZE(raw->osds));
2404                 return;
2405         }
2406
2407         len = do_crush(osdmap, ruleno, pps, raw->osds, pi->size,
2408                        osdmap->osd_weight, osdmap->max_osd, pi->id);
2409         if (len < 0) {
2410                 pr_err("error %d from crush rule %d: pool %lld ruleset %d type %d size %d\n",
2411                        len, ruleno, pi->id, pi->crush_ruleset, pi->type,
2412                        pi->size);
2413                 return;
2414         }
2415
2416         raw->size = len;
2417         remove_nonexistent_osds(osdmap, pi, raw);
2418 }
2419
2420 /* apply pg_upmap[_items] mappings */
2421 static void apply_upmap(struct ceph_osdmap *osdmap,
2422                         const struct ceph_pg *pgid,
2423                         struct ceph_osds *raw)
2424 {
2425         struct ceph_pg_mapping *pg;
2426         int i, j;
2427
2428         pg = lookup_pg_mapping(&osdmap->pg_upmap, pgid);
2429         if (pg) {
2430                 /* make sure targets aren't marked out */
2431                 for (i = 0; i < pg->pg_upmap.len; i++) {
2432                         int osd = pg->pg_upmap.osds[i];
2433
2434                         if (osd != CRUSH_ITEM_NONE &&
2435                             osd < osdmap->max_osd &&
2436                             osdmap->osd_weight[osd] == 0) {
2437                                 /* reject/ignore explicit mapping */
2438                                 return;
2439                         }
2440                 }
2441                 for (i = 0; i < pg->pg_upmap.len; i++)
2442                         raw->osds[i] = pg->pg_upmap.osds[i];
2443                 raw->size = pg->pg_upmap.len;
2444                 /* check and apply pg_upmap_items, if any */
2445         }
2446
2447         pg = lookup_pg_mapping(&osdmap->pg_upmap_items, pgid);
2448         if (pg) {
2449                 /*
2450                  * Note: this approach does not allow a bidirectional swap,
2451                  * e.g., [[1,2],[2,1]] applied to [0,1,2] -> [0,2,1].
2452                  */
2453                 for (i = 0; i < pg->pg_upmap_items.len; i++) {
2454                         int from = pg->pg_upmap_items.from_to[i][0];
2455                         int to = pg->pg_upmap_items.from_to[i][1];
2456                         int pos = -1;
2457                         bool exists = false;
2458
2459                         /* make sure replacement doesn't already appear */
2460                         for (j = 0; j < raw->size; j++) {
2461                                 int osd = raw->osds[j];
2462
2463                                 if (osd == to) {
2464                                         exists = true;
2465                                         break;
2466                                 }
2467                                 /* ignore mapping if target is marked out */
2468                                 if (osd == from && pos < 0 &&
2469                                     !(to != CRUSH_ITEM_NONE &&
2470                                       to < osdmap->max_osd &&
2471                                       osdmap->osd_weight[to] == 0)) {
2472                                         pos = j;
2473                                 }
2474                         }
2475                         if (!exists && pos >= 0)
2476                                 raw->osds[pos] = to;
2477                 }
2478         }
2479 }
2480
2481 /*
2482  * Given raw set, calculate up set and up primary.  By definition of an
2483  * up set, the result won't contain nonexistent or down OSDs.
2484  *
2485  * This is done in-place - on return @set is the up set.  If it's
2486  * empty, ->primary will remain undefined.
2487  */
2488 static void raw_to_up_osds(struct ceph_osdmap *osdmap,
2489                            struct ceph_pg_pool_info *pi,
2490                            struct ceph_osds *set)
2491 {
2492         int i;
2493
2494         /* ->primary is undefined for a raw set */
2495         BUG_ON(set->primary != -1);
2496
2497         if (ceph_can_shift_osds(pi)) {
2498                 int removed = 0;
2499
2500                 /* shift left */
2501                 for (i = 0; i < set->size; i++) {
2502                         if (ceph_osd_is_down(osdmap, set->osds[i])) {
2503                                 removed++;
2504                                 continue;
2505                         }
2506                         if (removed)
2507                                 set->osds[i - removed] = set->osds[i];
2508                 }
2509                 set->size -= removed;
2510                 if (set->size > 0)
2511                         set->primary = set->osds[0];
2512         } else {
2513                 /* set down/dne devices to NONE */
2514                 for (i = set->size - 1; i >= 0; i--) {
2515                         if (ceph_osd_is_down(osdmap, set->osds[i]))
2516                                 set->osds[i] = CRUSH_ITEM_NONE;
2517                         else
2518                                 set->primary = set->osds[i];
2519                 }
2520         }
2521 }
2522
2523 static void apply_primary_affinity(struct ceph_osdmap *osdmap,
2524                                    struct ceph_pg_pool_info *pi,
2525                                    u32 pps,
2526                                    struct ceph_osds *up)
2527 {
2528         int i;
2529         int pos = -1;
2530
2531         /*
2532          * Do we have any non-default primary_affinity values for these
2533          * osds?
2534          */
2535         if (!osdmap->osd_primary_affinity)
2536                 return;
2537
2538         for (i = 0; i < up->size; i++) {
2539                 int osd = up->osds[i];
2540
2541                 if (osd != CRUSH_ITEM_NONE &&
2542                     osdmap->osd_primary_affinity[osd] !=
2543                                         CEPH_OSD_DEFAULT_PRIMARY_AFFINITY) {
2544                         break;
2545                 }
2546         }
2547         if (i == up->size)
2548                 return;
2549
2550         /*
2551          * Pick the primary.  Feed both the seed (for the pg) and the
2552          * osd into the hash/rng so that a proportional fraction of an
2553          * osd's pgs get rejected as primary.
2554          */
2555         for (i = 0; i < up->size; i++) {
2556                 int osd = up->osds[i];
2557                 u32 aff;
2558
2559                 if (osd == CRUSH_ITEM_NONE)
2560                         continue;
2561
2562                 aff = osdmap->osd_primary_affinity[osd];
2563                 if (aff < CEPH_OSD_MAX_PRIMARY_AFFINITY &&
2564                     (crush_hash32_2(CRUSH_HASH_RJENKINS1,
2565                                     pps, osd) >> 16) >= aff) {
2566                         /*
2567                          * We chose not to use this primary.  Note it
2568                          * anyway as a fallback in case we don't pick
2569                          * anyone else, but keep looking.
2570                          */
2571                         if (pos < 0)
2572                                 pos = i;
2573                 } else {
2574                         pos = i;
2575                         break;
2576                 }
2577         }
2578         if (pos < 0)
2579                 return;
2580
2581         up->primary = up->osds[pos];
2582
2583         if (ceph_can_shift_osds(pi) && pos > 0) {
2584                 /* move the new primary to the front */
2585                 for (i = pos; i > 0; i--)
2586                         up->osds[i] = up->osds[i - 1];
2587                 up->osds[0] = up->primary;
2588         }
2589 }
2590
2591 /*
2592  * Get pg_temp and primary_temp mappings for given PG.
2593  *
2594  * Note that a PG may have none, only pg_temp, only primary_temp or
2595  * both pg_temp and primary_temp mappings.  This means @temp isn't
2596  * always a valid OSD set on return: in the "only primary_temp" case,
2597  * @temp will have its ->primary >= 0 but ->size == 0.
2598  */
2599 static void get_temp_osds(struct ceph_osdmap *osdmap,
2600                           struct ceph_pg_pool_info *pi,
2601                           const struct ceph_pg *pgid,
2602                           struct ceph_osds *temp)
2603 {
2604         struct ceph_pg_mapping *pg;
2605         int i;
2606
2607         ceph_osds_init(temp);
2608
2609         /* pg_temp? */
2610         pg = lookup_pg_mapping(&osdmap->pg_temp, pgid);
2611         if (pg) {
2612                 for (i = 0; i < pg->pg_temp.len; i++) {
2613                         if (ceph_osd_is_down(osdmap, pg->pg_temp.osds[i])) {
2614                                 if (ceph_can_shift_osds(pi))
2615                                         continue;
2616
2617                                 temp->osds[temp->size++] = CRUSH_ITEM_NONE;
2618                         } else {
2619                                 temp->osds[temp->size++] = pg->pg_temp.osds[i];
2620                         }
2621                 }
2622
2623                 /* apply pg_temp's primary */
2624                 for (i = 0; i < temp->size; i++) {
2625                         if (temp->osds[i] != CRUSH_ITEM_NONE) {
2626                                 temp->primary = temp->osds[i];
2627                                 break;
2628                         }
2629                 }
2630         }
2631
2632         /* primary_temp? */
2633         pg = lookup_pg_mapping(&osdmap->primary_temp, pgid);
2634         if (pg)
2635                 temp->primary = pg->primary_temp.osd;
2636 }
2637
2638 /*
2639  * Map a PG to its acting set as well as its up set.
2640  *
2641  * Acting set is used for data mapping purposes, while up set can be
2642  * recorded for detecting interval changes and deciding whether to
2643  * resend a request.
2644  */
2645 void ceph_pg_to_up_acting_osds(struct ceph_osdmap *osdmap,
2646                                struct ceph_pg_pool_info *pi,
2647                                const struct ceph_pg *raw_pgid,
2648                                struct ceph_osds *up,
2649                                struct ceph_osds *acting)
2650 {
2651         struct ceph_pg pgid;
2652         u32 pps;
2653
2654         WARN_ON(pi->id != raw_pgid->pool);
2655         raw_pg_to_pg(pi, raw_pgid, &pgid);
2656
2657         pg_to_raw_osds(osdmap, pi, raw_pgid, up, &pps);
2658         apply_upmap(osdmap, &pgid, up);
2659         raw_to_up_osds(osdmap, pi, up);
2660         apply_primary_affinity(osdmap, pi, pps, up);
2661         get_temp_osds(osdmap, pi, &pgid, acting);
2662         if (!acting->size) {
2663                 memcpy(acting->osds, up->osds, up->size * sizeof(up->osds[0]));
2664                 acting->size = up->size;
2665                 if (acting->primary == -1)
2666                         acting->primary = up->primary;
2667         }
2668         WARN_ON(!osds_valid(up) || !osds_valid(acting));
2669 }
2670
2671 bool ceph_pg_to_primary_shard(struct ceph_osdmap *osdmap,
2672                               struct ceph_pg_pool_info *pi,
2673                               const struct ceph_pg *raw_pgid,
2674                               struct ceph_spg *spgid)
2675 {
2676         struct ceph_pg pgid;
2677         struct ceph_osds up, acting;
2678         int i;
2679
2680         WARN_ON(pi->id != raw_pgid->pool);
2681         raw_pg_to_pg(pi, raw_pgid, &pgid);
2682
2683         if (ceph_can_shift_osds(pi)) {
2684                 spgid->pgid = pgid; /* struct */
2685                 spgid->shard = CEPH_SPG_NOSHARD;
2686                 return true;
2687         }
2688
2689         ceph_pg_to_up_acting_osds(osdmap, pi, &pgid, &up, &acting);
2690         for (i = 0; i < acting.size; i++) {
2691                 if (acting.osds[i] == acting.primary) {
2692                         spgid->pgid = pgid; /* struct */
2693                         spgid->shard = i;
2694                         return true;
2695                 }
2696         }
2697
2698         return false;
2699 }
2700
2701 /*
2702  * Return acting primary for given PG, or -1 if none.
2703  */
2704 int ceph_pg_to_acting_primary(struct ceph_osdmap *osdmap,
2705                               const struct ceph_pg *raw_pgid)
2706 {
2707         struct ceph_pg_pool_info *pi;
2708         struct ceph_osds up, acting;
2709
2710         pi = ceph_pg_pool_by_id(osdmap, raw_pgid->pool);
2711         if (!pi)
2712                 return -1;
2713
2714         ceph_pg_to_up_acting_osds(osdmap, pi, raw_pgid, &up, &acting);
2715         return acting.primary;
2716 }
2717 EXPORT_SYMBOL(ceph_pg_to_acting_primary);
2718
2719 static struct crush_loc_node *alloc_crush_loc(size_t type_name_len,
2720                                               size_t name_len)
2721 {
2722         struct crush_loc_node *loc;
2723
2724         loc = kmalloc(sizeof(*loc) + type_name_len + name_len + 2, GFP_NOIO);
2725         if (!loc)
2726                 return NULL;
2727
2728         RB_CLEAR_NODE(&loc->cl_node);
2729         return loc;
2730 }
2731
2732 static void free_crush_loc(struct crush_loc_node *loc)
2733 {
2734         WARN_ON(!RB_EMPTY_NODE(&loc->cl_node));
2735
2736         kfree(loc);
2737 }
2738
2739 static int crush_loc_compare(const struct crush_loc *loc1,
2740                              const struct crush_loc *loc2)
2741 {
2742         return strcmp(loc1->cl_type_name, loc2->cl_type_name) ?:
2743                strcmp(loc1->cl_name, loc2->cl_name);
2744 }
2745
2746 DEFINE_RB_FUNCS2(crush_loc, struct crush_loc_node, cl_loc, crush_loc_compare,
2747                  RB_BYPTR, const struct crush_loc *, cl_node)
2748
2749 /*
2750  * Parses a set of <bucket type name>':'<bucket name> pairs separated
2751  * by '|', e.g. "rack:foo1|rack:foo2|datacenter:bar".
2752  *
2753  * Note that @crush_location is modified by strsep().
2754  */
2755 int ceph_parse_crush_location(char *crush_location, struct rb_root *locs)
2756 {
2757         struct crush_loc_node *loc;
2758         const char *type_name, *name, *colon;
2759         size_t type_name_len, name_len;
2760
2761         dout("%s '%s'\n", __func__, crush_location);
2762         while ((type_name = strsep(&crush_location, "|"))) {
2763                 colon = strchr(type_name, ':');
2764                 if (!colon)
2765                         return -EINVAL;
2766
2767                 type_name_len = colon - type_name;
2768                 if (type_name_len == 0)
2769                         return -EINVAL;
2770
2771                 name = colon + 1;
2772                 name_len = strlen(name);
2773                 if (name_len == 0)
2774                         return -EINVAL;
2775
2776                 loc = alloc_crush_loc(type_name_len, name_len);
2777                 if (!loc)
2778                         return -ENOMEM;
2779
2780                 loc->cl_loc.cl_type_name = loc->cl_data;
2781                 memcpy(loc->cl_loc.cl_type_name, type_name, type_name_len);
2782                 loc->cl_loc.cl_type_name[type_name_len] = '\0';
2783
2784                 loc->cl_loc.cl_name = loc->cl_data + type_name_len + 1;
2785                 memcpy(loc->cl_loc.cl_name, name, name_len);
2786                 loc->cl_loc.cl_name[name_len] = '\0';
2787
2788                 if (!__insert_crush_loc(locs, loc)) {
2789                         free_crush_loc(loc);
2790                         return -EEXIST;
2791                 }
2792
2793                 dout("%s type_name '%s' name '%s'\n", __func__,
2794                      loc->cl_loc.cl_type_name, loc->cl_loc.cl_name);
2795         }
2796
2797         return 0;
2798 }
2799
2800 int ceph_compare_crush_locs(struct rb_root *locs1, struct rb_root *locs2)
2801 {
2802         struct rb_node *n1 = rb_first(locs1);
2803         struct rb_node *n2 = rb_first(locs2);
2804         int ret;
2805
2806         for ( ; n1 && n2; n1 = rb_next(n1), n2 = rb_next(n2)) {
2807                 struct crush_loc_node *loc1 =
2808                     rb_entry(n1, struct crush_loc_node, cl_node);
2809                 struct crush_loc_node *loc2 =
2810                     rb_entry(n2, struct crush_loc_node, cl_node);
2811
2812                 ret = crush_loc_compare(&loc1->cl_loc, &loc2->cl_loc);
2813                 if (ret)
2814                         return ret;
2815         }
2816
2817         if (!n1 && n2)
2818                 return -1;
2819         if (n1 && !n2)
2820                 return 1;
2821         return 0;
2822 }
2823
2824 void ceph_clear_crush_locs(struct rb_root *locs)
2825 {
2826         while (!RB_EMPTY_ROOT(locs)) {
2827                 struct crush_loc_node *loc =
2828                     rb_entry(rb_first(locs), struct crush_loc_node, cl_node);
2829
2830                 erase_crush_loc(locs, loc);
2831                 free_crush_loc(loc);
2832         }
2833 }
2834
2835 /*
2836  * [a-zA-Z0-9-_.]+
2837  */
2838 static bool is_valid_crush_name(const char *name)
2839 {
2840         do {
2841                 if (!('a' <= *name && *name <= 'z') &&
2842                     !('A' <= *name && *name <= 'Z') &&
2843                     !('0' <= *name && *name <= '9') &&
2844                     *name != '-' && *name != '_' && *name != '.')
2845                         return false;
2846         } while (*++name != '\0');
2847
2848         return true;
2849 }
2850
2851 /*
2852  * Gets the parent of an item.  Returns its id (<0 because the
2853  * parent is always a bucket), type id (>0 for the same reason,
2854  * via @parent_type_id) and location (via @parent_loc).  If no
2855  * parent, returns 0.
2856  *
2857  * Does a linear search, as there are no parent pointers of any
2858  * kind.  Note that the result is ambigous for items that occur
2859  * multiple times in the map.
2860  */
2861 static int get_immediate_parent(struct crush_map *c, int id,
2862                                 u16 *parent_type_id,
2863                                 struct crush_loc *parent_loc)
2864 {
2865         struct crush_bucket *b;
2866         struct crush_name_node *type_cn, *cn;
2867         int i, j;
2868
2869         for (i = 0; i < c->max_buckets; i++) {
2870                 b = c->buckets[i];
2871                 if (!b)
2872                         continue;
2873
2874                 /* ignore per-class shadow hierarchy */
2875                 cn = lookup_crush_name(&c->names, b->id);
2876                 if (!cn || !is_valid_crush_name(cn->cn_name))
2877                         continue;
2878
2879                 for (j = 0; j < b->size; j++) {
2880                         if (b->items[j] != id)
2881                                 continue;
2882
2883                         *parent_type_id = b->type;
2884                         type_cn = lookup_crush_name(&c->type_names, b->type);
2885                         parent_loc->cl_type_name = type_cn->cn_name;
2886                         parent_loc->cl_name = cn->cn_name;
2887                         return b->id;
2888                 }
2889         }
2890
2891         return 0;  /* no parent */
2892 }
2893
2894 /*
2895  * Calculates the locality/distance from an item to a client
2896  * location expressed in terms of CRUSH hierarchy as a set of
2897  * (bucket type name, bucket name) pairs.  Specifically, looks
2898  * for the lowest-valued bucket type for which the location of
2899  * @id matches one of the locations in @locs, so for standard
2900  * bucket types (host = 1, rack = 3, datacenter = 8, zone = 9)
2901  * a matching host is closer than a matching rack and a matching
2902  * data center is closer than a matching zone.
2903  *
2904  * Specifying multiple locations (a "multipath" location) such
2905  * as "rack=foo1 rack=foo2 datacenter=bar" is allowed -- @locs
2906  * is a multimap.  The locality will be:
2907  *
2908  * - 3 for OSDs in racks foo1 and foo2
2909  * - 8 for OSDs in data center bar
2910  * - -1 for all other OSDs
2911  *
2912  * The lowest possible bucket type is 1, so the best locality
2913  * for an OSD is 1 (i.e. a matching host).  Locality 0 would be
2914  * the OSD itself.
2915  */
2916 int ceph_get_crush_locality(struct ceph_osdmap *osdmap, int id,
2917                             struct rb_root *locs)
2918 {
2919         struct crush_loc loc;
2920         u16 type_id;
2921
2922         /*
2923          * Instead of repeated get_immediate_parent() calls,
2924          * the location of @id could be obtained with a single
2925          * depth-first traversal.
2926          */
2927         for (;;) {
2928                 id = get_immediate_parent(osdmap->crush, id, &type_id, &loc);
2929                 if (id >= 0)
2930                         return -1;  /* not local */
2931
2932                 if (lookup_crush_loc(locs, &loc))
2933                         return type_id;
2934         }
2935 }