1 // SPDX-License-Identifier: GPL-2.0-only 2 /* 3 * linux/net/sunrpc/svc.c 4 * 5 * High-level RPC service routines 6 * 7 * Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de> 8 * 9 * Multiple threads pools and NUMAisation 10 * Copyright (c) 2006 Silicon Graphics, Inc. 11 * by Greg Banks <gnb@melbourne.sgi.com> 12 */ 13 14 #include <linux/linkage.h> 15 #include <linux/sched/signal.h> 16 #include <linux/errno.h> 17 #include <linux/net.h> 18 #include <linux/in.h> 19 #include <linux/mm.h> 20 #include <linux/interrupt.h> 21 #include <linux/module.h> 22 #include <linux/kthread.h> 23 #include <linux/slab.h> 24 25 #include <linux/sunrpc/types.h> 26 #include <linux/sunrpc/xdr.h> 27 #include <linux/sunrpc/stats.h> 28 #include <linux/sunrpc/svcsock.h> 29 #include <linux/sunrpc/clnt.h> 30 #include <linux/sunrpc/bc_xprt.h> 31 32 #include <trace/events/sunrpc.h> 33 34 #include "fail.h" 35 36 #define RPCDBG_FACILITY RPCDBG_SVCDSP 37 38 static void svc_unregister(const struct svc_serv *serv, struct net *net); 39 40 #define svc_serv_is_pooled(serv) ((serv)->sv_ops->svo_function) 41 42 #define SVC_POOL_DEFAULT SVC_POOL_GLOBAL 43 44 /* 45 * Structure for mapping cpus to pools and vice versa. 46 * Setup once during sunrpc initialisation. 47 */ 48 struct svc_pool_map svc_pool_map = { 49 .mode = SVC_POOL_DEFAULT 50 }; 51 EXPORT_SYMBOL_GPL(svc_pool_map); 52 53 static DEFINE_MUTEX(svc_pool_map_mutex);/* protects svc_pool_map.count only */ 54 55 static int 56 param_set_pool_mode(const char *val, const struct kernel_param *kp) 57 { 58 int *ip = (int *)kp->arg; 59 struct svc_pool_map *m = &svc_pool_map; 60 int err; 61 62 mutex_lock(&svc_pool_map_mutex); 63 64 err = -EBUSY; 65 if (m->count) 66 goto out; 67 68 err = 0; 69 if (!strncmp(val, "auto", 4)) 70 *ip = SVC_POOL_AUTO; 71 else if (!strncmp(val, "global", 6)) 72 *ip = SVC_POOL_GLOBAL; 73 else if (!strncmp(val, "percpu", 6)) 74 *ip = SVC_POOL_PERCPU; 75 else if (!strncmp(val, "pernode", 7)) 76 *ip = SVC_POOL_PERNODE; 77 else 78 err = -EINVAL; 79 80 out: 81 mutex_unlock(&svc_pool_map_mutex); 82 return err; 83 } 84 85 static int 86 param_get_pool_mode(char *buf, const struct kernel_param *kp) 87 { 88 int *ip = (int *)kp->arg; 89 90 switch (*ip) 91 { 92 case SVC_POOL_AUTO: 93 return strlcpy(buf, "auto\n", 20); 94 case SVC_POOL_GLOBAL: 95 return strlcpy(buf, "global\n", 20); 96 case SVC_POOL_PERCPU: 97 return strlcpy(buf, "percpu\n", 20); 98 case SVC_POOL_PERNODE: 99 return strlcpy(buf, "pernode\n", 20); 100 default: 101 return sprintf(buf, "%d\n", *ip); 102 } 103 } 104 105 module_param_call(pool_mode, param_set_pool_mode, param_get_pool_mode, 106 &svc_pool_map.mode, 0644); 107 108 /* 109 * Detect best pool mapping mode heuristically, 110 * according to the machine's topology. 111 */ 112 static int 113 svc_pool_map_choose_mode(void) 114 { 115 unsigned int node; 116 117 if (nr_online_nodes > 1) { 118 /* 119 * Actually have multiple NUMA nodes, 120 * so split pools on NUMA node boundaries 121 */ 122 return SVC_POOL_PERNODE; 123 } 124 125 node = first_online_node; 126 if (nr_cpus_node(node) > 2) { 127 /* 128 * Non-trivial SMP, or CONFIG_NUMA on 129 * non-NUMA hardware, e.g. with a generic 130 * x86_64 kernel on Xeons. In this case we 131 * want to divide the pools on cpu boundaries. 132 */ 133 return SVC_POOL_PERCPU; 134 } 135 136 /* default: one global pool */ 137 return SVC_POOL_GLOBAL; 138 } 139 140 /* 141 * Allocate the to_pool[] and pool_to[] arrays. 142 * Returns 0 on success or an errno. 143 */ 144 static int 145 svc_pool_map_alloc_arrays(struct svc_pool_map *m, unsigned int maxpools) 146 { 147 m->to_pool = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL); 148 if (!m->to_pool) 149 goto fail; 150 m->pool_to = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL); 151 if (!m->pool_to) 152 goto fail_free; 153 154 return 0; 155 156 fail_free: 157 kfree(m->to_pool); 158 m->to_pool = NULL; 159 fail: 160 return -ENOMEM; 161 } 162 163 /* 164 * Initialise the pool map for SVC_POOL_PERCPU mode. 165 * Returns number of pools or <0 on error. 166 */ 167 static int 168 svc_pool_map_init_percpu(struct svc_pool_map *m) 169 { 170 unsigned int maxpools = nr_cpu_ids; 171 unsigned int pidx = 0; 172 unsigned int cpu; 173 int err; 174 175 err = svc_pool_map_alloc_arrays(m, maxpools); 176 if (err) 177 return err; 178 179 for_each_online_cpu(cpu) { 180 BUG_ON(pidx >= maxpools); 181 m->to_pool[cpu] = pidx; 182 m->pool_to[pidx] = cpu; 183 pidx++; 184 } 185 /* cpus brought online later all get mapped to pool0, sorry */ 186 187 return pidx; 188 }; 189 190 191 /* 192 * Initialise the pool map for SVC_POOL_PERNODE mode. 193 * Returns number of pools or <0 on error. 194 */ 195 static int 196 svc_pool_map_init_pernode(struct svc_pool_map *m) 197 { 198 unsigned int maxpools = nr_node_ids; 199 unsigned int pidx = 0; 200 unsigned int node; 201 int err; 202 203 err = svc_pool_map_alloc_arrays(m, maxpools); 204 if (err) 205 return err; 206 207 for_each_node_with_cpus(node) { 208 /* some architectures (e.g. SN2) have cpuless nodes */ 209 BUG_ON(pidx > maxpools); 210 m->to_pool[node] = pidx; 211 m->pool_to[pidx] = node; 212 pidx++; 213 } 214 /* nodes brought online later all get mapped to pool0, sorry */ 215 216 return pidx; 217 } 218 219 220 /* 221 * Add a reference to the global map of cpus to pools (and 222 * vice versa). Initialise the map if we're the first user. 223 * Returns the number of pools. 224 */ 225 unsigned int 226 svc_pool_map_get(void) 227 { 228 struct svc_pool_map *m = &svc_pool_map; 229 int npools = -1; 230 231 mutex_lock(&svc_pool_map_mutex); 232 233 if (m->count++) { 234 mutex_unlock(&svc_pool_map_mutex); 235 return m->npools; 236 } 237 238 if (m->mode == SVC_POOL_AUTO) 239 m->mode = svc_pool_map_choose_mode(); 240 241 switch (m->mode) { 242 case SVC_POOL_PERCPU: 243 npools = svc_pool_map_init_percpu(m); 244 break; 245 case SVC_POOL_PERNODE: 246 npools = svc_pool_map_init_pernode(m); 247 break; 248 } 249 250 if (npools < 0) { 251 /* default, or memory allocation failure */ 252 npools = 1; 253 m->mode = SVC_POOL_GLOBAL; 254 } 255 m->npools = npools; 256 257 mutex_unlock(&svc_pool_map_mutex); 258 return m->npools; 259 } 260 EXPORT_SYMBOL_GPL(svc_pool_map_get); 261 262 /* 263 * Drop a reference to the global map of cpus to pools. 264 * When the last reference is dropped, the map data is 265 * freed; this allows the sysadmin to change the pool 266 * mode using the pool_mode module option without 267 * rebooting or re-loading sunrpc.ko. 268 */ 269 void 270 svc_pool_map_put(void) 271 { 272 struct svc_pool_map *m = &svc_pool_map; 273 274 mutex_lock(&svc_pool_map_mutex); 275 276 if (!--m->count) { 277 kfree(m->to_pool); 278 m->to_pool = NULL; 279 kfree(m->pool_to); 280 m->pool_to = NULL; 281 m->npools = 0; 282 } 283 284 mutex_unlock(&svc_pool_map_mutex); 285 } 286 EXPORT_SYMBOL_GPL(svc_pool_map_put); 287 288 static int svc_pool_map_get_node(unsigned int pidx) 289 { 290 const struct svc_pool_map *m = &svc_pool_map; 291 292 if (m->count) { 293 if (m->mode == SVC_POOL_PERCPU) 294 return cpu_to_node(m->pool_to[pidx]); 295 if (m->mode == SVC_POOL_PERNODE) 296 return m->pool_to[pidx]; 297 } 298 return NUMA_NO_NODE; 299 } 300 /* 301 * Set the given thread's cpus_allowed mask so that it 302 * will only run on cpus in the given pool. 303 */ 304 static inline void 305 svc_pool_map_set_cpumask(struct task_struct *task, unsigned int pidx) 306 { 307 struct svc_pool_map *m = &svc_pool_map; 308 unsigned int node = m->pool_to[pidx]; 309 310 /* 311 * The caller checks for sv_nrpools > 1, which 312 * implies that we've been initialized. 313 */ 314 WARN_ON_ONCE(m->count == 0); 315 if (m->count == 0) 316 return; 317 318 switch (m->mode) { 319 case SVC_POOL_PERCPU: 320 { 321 set_cpus_allowed_ptr(task, cpumask_of(node)); 322 break; 323 } 324 case SVC_POOL_PERNODE: 325 { 326 set_cpus_allowed_ptr(task, cpumask_of_node(node)); 327 break; 328 } 329 } 330 } 331 332 /* 333 * Use the mapping mode to choose a pool for a given CPU. 334 * Used when enqueueing an incoming RPC. Always returns 335 * a non-NULL pool pointer. 336 */ 337 struct svc_pool * 338 svc_pool_for_cpu(struct svc_serv *serv, int cpu) 339 { 340 struct svc_pool_map *m = &svc_pool_map; 341 unsigned int pidx = 0; 342 343 /* 344 * An uninitialised map happens in a pure client when 345 * lockd is brought up, so silently treat it the 346 * same as SVC_POOL_GLOBAL. 347 */ 348 if (svc_serv_is_pooled(serv)) { 349 switch (m->mode) { 350 case SVC_POOL_PERCPU: 351 pidx = m->to_pool[cpu]; 352 break; 353 case SVC_POOL_PERNODE: 354 pidx = m->to_pool[cpu_to_node(cpu)]; 355 break; 356 } 357 } 358 return &serv->sv_pools[pidx % serv->sv_nrpools]; 359 } 360 361 int svc_rpcb_setup(struct svc_serv *serv, struct net *net) 362 { 363 int err; 364 365 err = rpcb_create_local(net); 366 if (err) 367 return err; 368 369 /* Remove any stale portmap registrations */ 370 svc_unregister(serv, net); 371 return 0; 372 } 373 EXPORT_SYMBOL_GPL(svc_rpcb_setup); 374 375 void svc_rpcb_cleanup(struct svc_serv *serv, struct net *net) 376 { 377 svc_unregister(serv, net); 378 rpcb_put_local(net); 379 } 380 EXPORT_SYMBOL_GPL(svc_rpcb_cleanup); 381 382 static int svc_uses_rpcbind(struct svc_serv *serv) 383 { 384 struct svc_program *progp; 385 unsigned int i; 386 387 for (progp = serv->sv_program; progp; progp = progp->pg_next) { 388 for (i = 0; i < progp->pg_nvers; i++) { 389 if (progp->pg_vers[i] == NULL) 390 continue; 391 if (!progp->pg_vers[i]->vs_hidden) 392 return 1; 393 } 394 } 395 396 return 0; 397 } 398 399 int svc_bind(struct svc_serv *serv, struct net *net) 400 { 401 if (!svc_uses_rpcbind(serv)) 402 return 0; 403 return svc_rpcb_setup(serv, net); 404 } 405 EXPORT_SYMBOL_GPL(svc_bind); 406 407 #if defined(CONFIG_SUNRPC_BACKCHANNEL) 408 static void 409 __svc_init_bc(struct svc_serv *serv) 410 { 411 INIT_LIST_HEAD(&serv->sv_cb_list); 412 spin_lock_init(&serv->sv_cb_lock); 413 init_waitqueue_head(&serv->sv_cb_waitq); 414 } 415 #else 416 static void 417 __svc_init_bc(struct svc_serv *serv) 418 { 419 } 420 #endif 421 422 /* 423 * Create an RPC service 424 */ 425 static struct svc_serv * 426 __svc_create(struct svc_program *prog, unsigned int bufsize, int npools, 427 const struct svc_serv_ops *ops) 428 { 429 struct svc_serv *serv; 430 unsigned int vers; 431 unsigned int xdrsize; 432 unsigned int i; 433 434 if (!(serv = kzalloc(sizeof(*serv), GFP_KERNEL))) 435 return NULL; 436 serv->sv_name = prog->pg_name; 437 serv->sv_program = prog; 438 serv->sv_nrthreads = 1; 439 serv->sv_stats = prog->pg_stats; 440 if (bufsize > RPCSVC_MAXPAYLOAD) 441 bufsize = RPCSVC_MAXPAYLOAD; 442 serv->sv_max_payload = bufsize? bufsize : 4096; 443 serv->sv_max_mesg = roundup(serv->sv_max_payload + PAGE_SIZE, PAGE_SIZE); 444 serv->sv_ops = ops; 445 xdrsize = 0; 446 while (prog) { 447 prog->pg_lovers = prog->pg_nvers-1; 448 for (vers=0; vers<prog->pg_nvers ; vers++) 449 if (prog->pg_vers[vers]) { 450 prog->pg_hivers = vers; 451 if (prog->pg_lovers > vers) 452 prog->pg_lovers = vers; 453 if (prog->pg_vers[vers]->vs_xdrsize > xdrsize) 454 xdrsize = prog->pg_vers[vers]->vs_xdrsize; 455 } 456 prog = prog->pg_next; 457 } 458 serv->sv_xdrsize = xdrsize; 459 INIT_LIST_HEAD(&serv->sv_tempsocks); 460 INIT_LIST_HEAD(&serv->sv_permsocks); 461 timer_setup(&serv->sv_temptimer, NULL, 0); 462 spin_lock_init(&serv->sv_lock); 463 464 __svc_init_bc(serv); 465 466 serv->sv_nrpools = npools; 467 serv->sv_pools = 468 kcalloc(serv->sv_nrpools, sizeof(struct svc_pool), 469 GFP_KERNEL); 470 if (!serv->sv_pools) { 471 kfree(serv); 472 return NULL; 473 } 474 475 for (i = 0; i < serv->sv_nrpools; i++) { 476 struct svc_pool *pool = &serv->sv_pools[i]; 477 478 dprintk("svc: initialising pool %u for %s\n", 479 i, serv->sv_name); 480 481 pool->sp_id = i; 482 INIT_LIST_HEAD(&pool->sp_sockets); 483 INIT_LIST_HEAD(&pool->sp_all_threads); 484 spin_lock_init(&pool->sp_lock); 485 } 486 487 return serv; 488 } 489 490 struct svc_serv * 491 svc_create(struct svc_program *prog, unsigned int bufsize, 492 const struct svc_serv_ops *ops) 493 { 494 return __svc_create(prog, bufsize, /*npools*/1, ops); 495 } 496 EXPORT_SYMBOL_GPL(svc_create); 497 498 struct svc_serv * 499 svc_create_pooled(struct svc_program *prog, unsigned int bufsize, 500 const struct svc_serv_ops *ops) 501 { 502 struct svc_serv *serv; 503 unsigned int npools = svc_pool_map_get(); 504 505 serv = __svc_create(prog, bufsize, npools, ops); 506 if (!serv) 507 goto out_err; 508 return serv; 509 out_err: 510 svc_pool_map_put(); 511 return NULL; 512 } 513 EXPORT_SYMBOL_GPL(svc_create_pooled); 514 515 void svc_shutdown_net(struct svc_serv *serv, struct net *net) 516 { 517 svc_close_net(serv, net); 518 519 if (serv->sv_ops->svo_shutdown) 520 serv->sv_ops->svo_shutdown(serv, net); 521 } 522 EXPORT_SYMBOL_GPL(svc_shutdown_net); 523 524 /* 525 * Destroy an RPC service. Should be called with appropriate locking to 526 * protect the sv_nrthreads, sv_permsocks and sv_tempsocks. 527 */ 528 void 529 svc_destroy(struct svc_serv *serv) 530 { 531 dprintk("svc: svc_destroy(%s, %d)\n", 532 serv->sv_program->pg_name, 533 serv->sv_nrthreads); 534 535 if (serv->sv_nrthreads) { 536 if (--(serv->sv_nrthreads) != 0) { 537 svc_sock_update_bufs(serv); 538 return; 539 } 540 } else 541 printk("svc_destroy: no threads for serv=%p!\n", serv); 542 543 del_timer_sync(&serv->sv_temptimer); 544 545 /* 546 * The last user is gone and thus all sockets have to be destroyed to 547 * the point. Check this. 548 */ 549 BUG_ON(!list_empty(&serv->sv_permsocks)); 550 BUG_ON(!list_empty(&serv->sv_tempsocks)); 551 552 cache_clean_deferred(serv); 553 554 if (svc_serv_is_pooled(serv)) 555 svc_pool_map_put(); 556 557 kfree(serv->sv_pools); 558 kfree(serv); 559 } 560 EXPORT_SYMBOL_GPL(svc_destroy); 561 562 /* 563 * Allocate an RPC server's buffer space. 564 * We allocate pages and place them in rq_pages. 565 */ 566 static int 567 svc_init_buffer(struct svc_rqst *rqstp, unsigned int size, int node) 568 { 569 unsigned int pages, arghi; 570 571 /* bc_xprt uses fore channel allocated buffers */ 572 if (svc_is_backchannel(rqstp)) 573 return 1; 574 575 pages = size / PAGE_SIZE + 1; /* extra page as we hold both request and reply. 576 * We assume one is at most one page 577 */ 578 arghi = 0; 579 WARN_ON_ONCE(pages > RPCSVC_MAXPAGES); 580 if (pages > RPCSVC_MAXPAGES) 581 pages = RPCSVC_MAXPAGES; 582 while (pages) { 583 struct page *p = alloc_pages_node(node, GFP_KERNEL, 0); 584 if (!p) 585 break; 586 rqstp->rq_pages[arghi++] = p; 587 pages--; 588 } 589 return pages == 0; 590 } 591 592 /* 593 * Release an RPC server buffer 594 */ 595 static void 596 svc_release_buffer(struct svc_rqst *rqstp) 597 { 598 unsigned int i; 599 600 for (i = 0; i < ARRAY_SIZE(rqstp->rq_pages); i++) 601 if (rqstp->rq_pages[i]) 602 put_page(rqstp->rq_pages[i]); 603 } 604 605 struct svc_rqst * 606 svc_rqst_alloc(struct svc_serv *serv, struct svc_pool *pool, int node) 607 { 608 struct svc_rqst *rqstp; 609 610 rqstp = kzalloc_node(sizeof(*rqstp), GFP_KERNEL, node); 611 if (!rqstp) 612 return rqstp; 613 614 __set_bit(RQ_BUSY, &rqstp->rq_flags); 615 spin_lock_init(&rqstp->rq_lock); 616 rqstp->rq_server = serv; 617 rqstp->rq_pool = pool; 618 619 rqstp->rq_scratch_page = alloc_pages_node(node, GFP_KERNEL, 0); 620 if (!rqstp->rq_scratch_page) 621 goto out_enomem; 622 623 rqstp->rq_argp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node); 624 if (!rqstp->rq_argp) 625 goto out_enomem; 626 627 rqstp->rq_resp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node); 628 if (!rqstp->rq_resp) 629 goto out_enomem; 630 631 if (!svc_init_buffer(rqstp, serv->sv_max_mesg, node)) 632 goto out_enomem; 633 634 return rqstp; 635 out_enomem: 636 svc_rqst_free(rqstp); 637 return NULL; 638 } 639 EXPORT_SYMBOL_GPL(svc_rqst_alloc); 640 641 struct svc_rqst * 642 svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node) 643 { 644 struct svc_rqst *rqstp; 645 646 rqstp = svc_rqst_alloc(serv, pool, node); 647 if (!rqstp) 648 return ERR_PTR(-ENOMEM); 649 650 serv->sv_nrthreads++; 651 spin_lock_bh(&pool->sp_lock); 652 pool->sp_nrthreads++; 653 list_add_rcu(&rqstp->rq_all, &pool->sp_all_threads); 654 spin_unlock_bh(&pool->sp_lock); 655 return rqstp; 656 } 657 EXPORT_SYMBOL_GPL(svc_prepare_thread); 658 659 /* 660 * Choose a pool in which to create a new thread, for svc_set_num_threads 661 */ 662 static inline struct svc_pool * 663 choose_pool(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state) 664 { 665 if (pool != NULL) 666 return pool; 667 668 return &serv->sv_pools[(*state)++ % serv->sv_nrpools]; 669 } 670 671 /* 672 * Choose a thread to kill, for svc_set_num_threads 673 */ 674 static inline struct task_struct * 675 choose_victim(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state) 676 { 677 unsigned int i; 678 struct task_struct *task = NULL; 679 680 if (pool != NULL) { 681 spin_lock_bh(&pool->sp_lock); 682 } else { 683 /* choose a pool in round-robin fashion */ 684 for (i = 0; i < serv->sv_nrpools; i++) { 685 pool = &serv->sv_pools[--(*state) % serv->sv_nrpools]; 686 spin_lock_bh(&pool->sp_lock); 687 if (!list_empty(&pool->sp_all_threads)) 688 goto found_pool; 689 spin_unlock_bh(&pool->sp_lock); 690 } 691 return NULL; 692 } 693 694 found_pool: 695 if (!list_empty(&pool->sp_all_threads)) { 696 struct svc_rqst *rqstp; 697 698 /* 699 * Remove from the pool->sp_all_threads list 700 * so we don't try to kill it again. 701 */ 702 rqstp = list_entry(pool->sp_all_threads.next, struct svc_rqst, rq_all); 703 set_bit(RQ_VICTIM, &rqstp->rq_flags); 704 list_del_rcu(&rqstp->rq_all); 705 task = rqstp->rq_task; 706 } 707 spin_unlock_bh(&pool->sp_lock); 708 709 return task; 710 } 711 712 /* create new threads */ 713 static int 714 svc_start_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) 715 { 716 struct svc_rqst *rqstp; 717 struct task_struct *task; 718 struct svc_pool *chosen_pool; 719 unsigned int state = serv->sv_nrthreads-1; 720 int node; 721 722 do { 723 nrservs--; 724 chosen_pool = choose_pool(serv, pool, &state); 725 726 node = svc_pool_map_get_node(chosen_pool->sp_id); 727 rqstp = svc_prepare_thread(serv, chosen_pool, node); 728 if (IS_ERR(rqstp)) 729 return PTR_ERR(rqstp); 730 731 __module_get(serv->sv_ops->svo_module); 732 task = kthread_create_on_node(serv->sv_ops->svo_function, rqstp, 733 node, "%s", serv->sv_name); 734 if (IS_ERR(task)) { 735 module_put(serv->sv_ops->svo_module); 736 svc_exit_thread(rqstp); 737 return PTR_ERR(task); 738 } 739 740 rqstp->rq_task = task; 741 if (serv->sv_nrpools > 1) 742 svc_pool_map_set_cpumask(task, chosen_pool->sp_id); 743 744 svc_sock_update_bufs(serv); 745 wake_up_process(task); 746 } while (nrservs > 0); 747 748 return 0; 749 } 750 751 752 /* destroy old threads */ 753 static int 754 svc_signal_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) 755 { 756 struct task_struct *task; 757 unsigned int state = serv->sv_nrthreads-1; 758 759 /* destroy old threads */ 760 do { 761 task = choose_victim(serv, pool, &state); 762 if (task == NULL) 763 break; 764 send_sig(SIGINT, task, 1); 765 nrservs++; 766 } while (nrservs < 0); 767 768 return 0; 769 } 770 771 /* 772 * Create or destroy enough new threads to make the number 773 * of threads the given number. If `pool' is non-NULL, applies 774 * only to threads in that pool, otherwise round-robins between 775 * all pools. Caller must ensure that mutual exclusion between this and 776 * server startup or shutdown. 777 * 778 * Destroying threads relies on the service threads filling in 779 * rqstp->rq_task, which only the nfs ones do. Assumes the serv 780 * has been created using svc_create_pooled(). 781 * 782 * Based on code that used to be in nfsd_svc() but tweaked 783 * to be pool-aware. 784 */ 785 int 786 svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) 787 { 788 if (pool == NULL) { 789 /* The -1 assumes caller has done a svc_get() */ 790 nrservs -= (serv->sv_nrthreads-1); 791 } else { 792 spin_lock_bh(&pool->sp_lock); 793 nrservs -= pool->sp_nrthreads; 794 spin_unlock_bh(&pool->sp_lock); 795 } 796 797 if (nrservs > 0) 798 return svc_start_kthreads(serv, pool, nrservs); 799 if (nrservs < 0) 800 return svc_signal_kthreads(serv, pool, nrservs); 801 return 0; 802 } 803 EXPORT_SYMBOL_GPL(svc_set_num_threads); 804 805 /* destroy old threads */ 806 static int 807 svc_stop_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) 808 { 809 struct task_struct *task; 810 unsigned int state = serv->sv_nrthreads-1; 811 812 /* destroy old threads */ 813 do { 814 task = choose_victim(serv, pool, &state); 815 if (task == NULL) 816 break; 817 kthread_stop(task); 818 nrservs++; 819 } while (nrservs < 0); 820 return 0; 821 } 822 823 int 824 svc_set_num_threads_sync(struct svc_serv *serv, struct svc_pool *pool, int nrservs) 825 { 826 if (pool == NULL) { 827 /* The -1 assumes caller has done a svc_get() */ 828 nrservs -= (serv->sv_nrthreads-1); 829 } else { 830 spin_lock_bh(&pool->sp_lock); 831 nrservs -= pool->sp_nrthreads; 832 spin_unlock_bh(&pool->sp_lock); 833 } 834 835 if (nrservs > 0) 836 return svc_start_kthreads(serv, pool, nrservs); 837 if (nrservs < 0) 838 return svc_stop_kthreads(serv, pool, nrservs); 839 return 0; 840 } 841 EXPORT_SYMBOL_GPL(svc_set_num_threads_sync); 842 843 /** 844 * svc_rqst_replace_page - Replace one page in rq_pages[] 845 * @rqstp: svc_rqst with pages to replace 846 * @page: replacement page 847 * 848 * When replacing a page in rq_pages, batch the release of the 849 * replaced pages to avoid hammering the page allocator. 850 */ 851 void svc_rqst_replace_page(struct svc_rqst *rqstp, struct page *page) 852 { 853 if (*rqstp->rq_next_page) { 854 if (!pagevec_space(&rqstp->rq_pvec)) 855 __pagevec_release(&rqstp->rq_pvec); 856 pagevec_add(&rqstp->rq_pvec, *rqstp->rq_next_page); 857 } 858 859 get_page(page); 860 *(rqstp->rq_next_page++) = page; 861 } 862 EXPORT_SYMBOL_GPL(svc_rqst_replace_page); 863 864 /* 865 * Called from a server thread as it's exiting. Caller must hold the "service 866 * mutex" for the service. 867 */ 868 void 869 svc_rqst_free(struct svc_rqst *rqstp) 870 { 871 svc_release_buffer(rqstp); 872 if (rqstp->rq_scratch_page) 873 put_page(rqstp->rq_scratch_page); 874 kfree(rqstp->rq_resp); 875 kfree(rqstp->rq_argp); 876 kfree(rqstp->rq_auth_data); 877 kfree_rcu(rqstp, rq_rcu_head); 878 } 879 EXPORT_SYMBOL_GPL(svc_rqst_free); 880 881 void 882 svc_exit_thread(struct svc_rqst *rqstp) 883 { 884 struct svc_serv *serv = rqstp->rq_server; 885 struct svc_pool *pool = rqstp->rq_pool; 886 887 spin_lock_bh(&pool->sp_lock); 888 pool->sp_nrthreads--; 889 if (!test_and_set_bit(RQ_VICTIM, &rqstp->rq_flags)) 890 list_del_rcu(&rqstp->rq_all); 891 spin_unlock_bh(&pool->sp_lock); 892 893 svc_rqst_free(rqstp); 894 895 /* Release the server */ 896 if (serv) 897 svc_destroy(serv); 898 } 899 EXPORT_SYMBOL_GPL(svc_exit_thread); 900 901 /* 902 * Register an "inet" protocol family netid with the local 903 * rpcbind daemon via an rpcbind v4 SET request. 904 * 905 * No netconfig infrastructure is available in the kernel, so 906 * we map IP_ protocol numbers to netids by hand. 907 * 908 * Returns zero on success; a negative errno value is returned 909 * if any error occurs. 910 */ 911 static int __svc_rpcb_register4(struct net *net, const u32 program, 912 const u32 version, 913 const unsigned short protocol, 914 const unsigned short port) 915 { 916 const struct sockaddr_in sin = { 917 .sin_family = AF_INET, 918 .sin_addr.s_addr = htonl(INADDR_ANY), 919 .sin_port = htons(port), 920 }; 921 const char *netid; 922 int error; 923 924 switch (protocol) { 925 case IPPROTO_UDP: 926 netid = RPCBIND_NETID_UDP; 927 break; 928 case IPPROTO_TCP: 929 netid = RPCBIND_NETID_TCP; 930 break; 931 default: 932 return -ENOPROTOOPT; 933 } 934 935 error = rpcb_v4_register(net, program, version, 936 (const struct sockaddr *)&sin, netid); 937 938 /* 939 * User space didn't support rpcbind v4, so retry this 940 * registration request with the legacy rpcbind v2 protocol. 941 */ 942 if (error == -EPROTONOSUPPORT) 943 error = rpcb_register(net, program, version, protocol, port); 944 945 return error; 946 } 947 948 #if IS_ENABLED(CONFIG_IPV6) 949 /* 950 * Register an "inet6" protocol family netid with the local 951 * rpcbind daemon via an rpcbind v4 SET request. 952 * 953 * No netconfig infrastructure is available in the kernel, so 954 * we map IP_ protocol numbers to netids by hand. 955 * 956 * Returns zero on success; a negative errno value is returned 957 * if any error occurs. 958 */ 959 static int __svc_rpcb_register6(struct net *net, const u32 program, 960 const u32 version, 961 const unsigned short protocol, 962 const unsigned short port) 963 { 964 const struct sockaddr_in6 sin6 = { 965 .sin6_family = AF_INET6, 966 .sin6_addr = IN6ADDR_ANY_INIT, 967 .sin6_port = htons(port), 968 }; 969 const char *netid; 970 int error; 971 972 switch (protocol) { 973 case IPPROTO_UDP: 974 netid = RPCBIND_NETID_UDP6; 975 break; 976 case IPPROTO_TCP: 977 netid = RPCBIND_NETID_TCP6; 978 break; 979 default: 980 return -ENOPROTOOPT; 981 } 982 983 error = rpcb_v4_register(net, program, version, 984 (const struct sockaddr *)&sin6, netid); 985 986 /* 987 * User space didn't support rpcbind version 4, so we won't 988 * use a PF_INET6 listener. 989 */ 990 if (error == -EPROTONOSUPPORT) 991 error = -EAFNOSUPPORT; 992 993 return error; 994 } 995 #endif /* IS_ENABLED(CONFIG_IPV6) */ 996 997 /* 998 * Register a kernel RPC service via rpcbind version 4. 999 * 1000 * Returns zero on success; a negative errno value is returned 1001 * if any error occurs. 1002 */ 1003 static int __svc_register(struct net *net, const char *progname, 1004 const u32 program, const u32 version, 1005 const int family, 1006 const unsigned short protocol, 1007 const unsigned short port) 1008 { 1009 int error = -EAFNOSUPPORT; 1010 1011 switch (family) { 1012 case PF_INET: 1013 error = __svc_rpcb_register4(net, program, version, 1014 protocol, port); 1015 break; 1016 #if IS_ENABLED(CONFIG_IPV6) 1017 case PF_INET6: 1018 error = __svc_rpcb_register6(net, program, version, 1019 protocol, port); 1020 #endif 1021 } 1022 1023 trace_svc_register(progname, version, protocol, port, family, error); 1024 return error; 1025 } 1026 1027 int svc_rpcbind_set_version(struct net *net, 1028 const struct svc_program *progp, 1029 u32 version, int family, 1030 unsigned short proto, 1031 unsigned short port) 1032 { 1033 return __svc_register(net, progp->pg_name, progp->pg_prog, 1034 version, family, proto, port); 1035 1036 } 1037 EXPORT_SYMBOL_GPL(svc_rpcbind_set_version); 1038 1039 int svc_generic_rpcbind_set(struct net *net, 1040 const struct svc_program *progp, 1041 u32 version, int family, 1042 unsigned short proto, 1043 unsigned short port) 1044 { 1045 const struct svc_version *vers = progp->pg_vers[version]; 1046 int error; 1047 1048 if (vers == NULL) 1049 return 0; 1050 1051 if (vers->vs_hidden) { 1052 trace_svc_noregister(progp->pg_name, version, proto, 1053 port, family, 0); 1054 return 0; 1055 } 1056 1057 /* 1058 * Don't register a UDP port if we need congestion 1059 * control. 1060 */ 1061 if (vers->vs_need_cong_ctrl && proto == IPPROTO_UDP) 1062 return 0; 1063 1064 error = svc_rpcbind_set_version(net, progp, version, 1065 family, proto, port); 1066 1067 return (vers->vs_rpcb_optnl) ? 0 : error; 1068 } 1069 EXPORT_SYMBOL_GPL(svc_generic_rpcbind_set); 1070 1071 /** 1072 * svc_register - register an RPC service with the local portmapper 1073 * @serv: svc_serv struct for the service to register 1074 * @net: net namespace for the service to register 1075 * @family: protocol family of service's listener socket 1076 * @proto: transport protocol number to advertise 1077 * @port: port to advertise 1078 * 1079 * Service is registered for any address in the passed-in protocol family 1080 */ 1081 int svc_register(const struct svc_serv *serv, struct net *net, 1082 const int family, const unsigned short proto, 1083 const unsigned short port) 1084 { 1085 struct svc_program *progp; 1086 unsigned int i; 1087 int error = 0; 1088 1089 WARN_ON_ONCE(proto == 0 && port == 0); 1090 if (proto == 0 && port == 0) 1091 return -EINVAL; 1092 1093 for (progp = serv->sv_program; progp; progp = progp->pg_next) { 1094 for (i = 0; i < progp->pg_nvers; i++) { 1095 1096 error = progp->pg_rpcbind_set(net, progp, i, 1097 family, proto, port); 1098 if (error < 0) { 1099 printk(KERN_WARNING "svc: failed to register " 1100 "%sv%u RPC service (errno %d).\n", 1101 progp->pg_name, i, -error); 1102 break; 1103 } 1104 } 1105 } 1106 1107 return error; 1108 } 1109 1110 /* 1111 * If user space is running rpcbind, it should take the v4 UNSET 1112 * and clear everything for this [program, version]. If user space 1113 * is running portmap, it will reject the v4 UNSET, but won't have 1114 * any "inet6" entries anyway. So a PMAP_UNSET should be sufficient 1115 * in this case to clear all existing entries for [program, version]. 1116 */ 1117 static void __svc_unregister(struct net *net, const u32 program, const u32 version, 1118 const char *progname) 1119 { 1120 int error; 1121 1122 error = rpcb_v4_register(net, program, version, NULL, ""); 1123 1124 /* 1125 * User space didn't support rpcbind v4, so retry this 1126 * request with the legacy rpcbind v2 protocol. 1127 */ 1128 if (error == -EPROTONOSUPPORT) 1129 error = rpcb_register(net, program, version, 0, 0); 1130 1131 trace_svc_unregister(progname, version, error); 1132 } 1133 1134 /* 1135 * All netids, bind addresses and ports registered for [program, version] 1136 * are removed from the local rpcbind database (if the service is not 1137 * hidden) to make way for a new instance of the service. 1138 * 1139 * The result of unregistration is reported via dprintk for those who want 1140 * verification of the result, but is otherwise not important. 1141 */ 1142 static void svc_unregister(const struct svc_serv *serv, struct net *net) 1143 { 1144 struct svc_program *progp; 1145 unsigned long flags; 1146 unsigned int i; 1147 1148 clear_thread_flag(TIF_SIGPENDING); 1149 1150 for (progp = serv->sv_program; progp; progp = progp->pg_next) { 1151 for (i = 0; i < progp->pg_nvers; i++) { 1152 if (progp->pg_vers[i] == NULL) 1153 continue; 1154 if (progp->pg_vers[i]->vs_hidden) 1155 continue; 1156 __svc_unregister(net, progp->pg_prog, i, progp->pg_name); 1157 } 1158 } 1159 1160 spin_lock_irqsave(¤t->sighand->siglock, flags); 1161 recalc_sigpending(); 1162 spin_unlock_irqrestore(¤t->sighand->siglock, flags); 1163 } 1164 1165 /* 1166 * dprintk the given error with the address of the client that caused it. 1167 */ 1168 #if IS_ENABLED(CONFIG_SUNRPC_DEBUG) 1169 static __printf(2, 3) 1170 void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...) 1171 { 1172 struct va_format vaf; 1173 va_list args; 1174 char buf[RPC_MAX_ADDRBUFLEN]; 1175 1176 va_start(args, fmt); 1177 1178 vaf.fmt = fmt; 1179 vaf.va = &args; 1180 1181 dprintk("svc: %s: %pV", svc_print_addr(rqstp, buf, sizeof(buf)), &vaf); 1182 1183 va_end(args); 1184 } 1185 #else 1186 static __printf(2,3) void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...) {} 1187 #endif 1188 1189 static int 1190 svc_generic_dispatch(struct svc_rqst *rqstp, __be32 *statp) 1191 { 1192 struct kvec *argv = &rqstp->rq_arg.head[0]; 1193 struct kvec *resv = &rqstp->rq_res.head[0]; 1194 const struct svc_procedure *procp = rqstp->rq_procinfo; 1195 1196 /* 1197 * Decode arguments 1198 * XXX: why do we ignore the return value? 1199 */ 1200 if (procp->pc_decode && 1201 !procp->pc_decode(rqstp, argv->iov_base)) { 1202 *statp = rpc_garbage_args; 1203 return 1; 1204 } 1205 1206 *statp = procp->pc_func(rqstp); 1207 1208 if (*statp == rpc_drop_reply || 1209 test_bit(RQ_DROPME, &rqstp->rq_flags)) 1210 return 0; 1211 1212 if (rqstp->rq_auth_stat != rpc_auth_ok) 1213 return 1; 1214 1215 if (*statp != rpc_success) 1216 return 1; 1217 1218 /* Encode reply */ 1219 if (procp->pc_encode && 1220 !procp->pc_encode(rqstp, resv->iov_base + resv->iov_len)) { 1221 dprintk("svc: failed to encode reply\n"); 1222 /* serv->sv_stats->rpcsystemerr++; */ 1223 *statp = rpc_system_err; 1224 } 1225 return 1; 1226 } 1227 1228 __be32 1229 svc_generic_init_request(struct svc_rqst *rqstp, 1230 const struct svc_program *progp, 1231 struct svc_process_info *ret) 1232 { 1233 const struct svc_version *versp = NULL; /* compiler food */ 1234 const struct svc_procedure *procp = NULL; 1235 1236 if (rqstp->rq_vers >= progp->pg_nvers ) 1237 goto err_bad_vers; 1238 versp = progp->pg_vers[rqstp->rq_vers]; 1239 if (!versp) 1240 goto err_bad_vers; 1241 1242 /* 1243 * Some protocol versions (namely NFSv4) require some form of 1244 * congestion control. (See RFC 7530 section 3.1 paragraph 2) 1245 * In other words, UDP is not allowed. We mark those when setting 1246 * up the svc_xprt, and verify that here. 1247 * 1248 * The spec is not very clear about what error should be returned 1249 * when someone tries to access a server that is listening on UDP 1250 * for lower versions. RPC_PROG_MISMATCH seems to be the closest 1251 * fit. 1252 */ 1253 if (versp->vs_need_cong_ctrl && rqstp->rq_xprt && 1254 !test_bit(XPT_CONG_CTRL, &rqstp->rq_xprt->xpt_flags)) 1255 goto err_bad_vers; 1256 1257 if (rqstp->rq_proc >= versp->vs_nproc) 1258 goto err_bad_proc; 1259 rqstp->rq_procinfo = procp = &versp->vs_proc[rqstp->rq_proc]; 1260 if (!procp) 1261 goto err_bad_proc; 1262 1263 /* Initialize storage for argp and resp */ 1264 memset(rqstp->rq_argp, 0, procp->pc_argsize); 1265 memset(rqstp->rq_resp, 0, procp->pc_ressize); 1266 1267 /* Bump per-procedure stats counter */ 1268 versp->vs_count[rqstp->rq_proc]++; 1269 1270 ret->dispatch = versp->vs_dispatch; 1271 return rpc_success; 1272 err_bad_vers: 1273 ret->mismatch.lovers = progp->pg_lovers; 1274 ret->mismatch.hivers = progp->pg_hivers; 1275 return rpc_prog_mismatch; 1276 err_bad_proc: 1277 return rpc_proc_unavail; 1278 } 1279 EXPORT_SYMBOL_GPL(svc_generic_init_request); 1280 1281 /* 1282 * Common routine for processing the RPC request. 1283 */ 1284 static int 1285 svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) 1286 { 1287 struct svc_program *progp; 1288 const struct svc_procedure *procp = NULL; 1289 struct svc_serv *serv = rqstp->rq_server; 1290 struct svc_process_info process; 1291 __be32 *statp; 1292 u32 prog, vers; 1293 __be32 rpc_stat; 1294 int auth_res; 1295 __be32 *reply_statp; 1296 1297 rpc_stat = rpc_success; 1298 1299 if (argv->iov_len < 6*4) 1300 goto err_short_len; 1301 1302 /* Will be turned off by GSS integrity and privacy services */ 1303 set_bit(RQ_SPLICE_OK, &rqstp->rq_flags); 1304 /* Will be turned off only when NFSv4 Sessions are used */ 1305 set_bit(RQ_USEDEFERRAL, &rqstp->rq_flags); 1306 clear_bit(RQ_DROPME, &rqstp->rq_flags); 1307 1308 svc_putu32(resv, rqstp->rq_xid); 1309 1310 vers = svc_getnl(argv); 1311 1312 /* First words of reply: */ 1313 svc_putnl(resv, 1); /* REPLY */ 1314 1315 if (vers != 2) /* RPC version number */ 1316 goto err_bad_rpc; 1317 1318 /* Save position in case we later decide to reject: */ 1319 reply_statp = resv->iov_base + resv->iov_len; 1320 1321 svc_putnl(resv, 0); /* ACCEPT */ 1322 1323 rqstp->rq_prog = prog = svc_getnl(argv); /* program number */ 1324 rqstp->rq_vers = svc_getnl(argv); /* version number */ 1325 rqstp->rq_proc = svc_getnl(argv); /* procedure number */ 1326 1327 for (progp = serv->sv_program; progp; progp = progp->pg_next) 1328 if (prog == progp->pg_prog) 1329 break; 1330 1331 /* 1332 * Decode auth data, and add verifier to reply buffer. 1333 * We do this before anything else in order to get a decent 1334 * auth verifier. 1335 */ 1336 auth_res = svc_authenticate(rqstp); 1337 /* Also give the program a chance to reject this call: */ 1338 if (auth_res == SVC_OK && progp) 1339 auth_res = progp->pg_authenticate(rqstp); 1340 if (auth_res != SVC_OK) 1341 trace_svc_authenticate(rqstp, auth_res); 1342 switch (auth_res) { 1343 case SVC_OK: 1344 break; 1345 case SVC_GARBAGE: 1346 goto err_garbage; 1347 case SVC_SYSERR: 1348 rpc_stat = rpc_system_err; 1349 goto err_bad; 1350 case SVC_DENIED: 1351 goto err_bad_auth; 1352 case SVC_CLOSE: 1353 goto close; 1354 case SVC_DROP: 1355 goto dropit; 1356 case SVC_COMPLETE: 1357 goto sendit; 1358 } 1359 1360 if (progp == NULL) 1361 goto err_bad_prog; 1362 1363 rpc_stat = progp->pg_init_request(rqstp, progp, &process); 1364 switch (rpc_stat) { 1365 case rpc_success: 1366 break; 1367 case rpc_prog_unavail: 1368 goto err_bad_prog; 1369 case rpc_prog_mismatch: 1370 goto err_bad_vers; 1371 case rpc_proc_unavail: 1372 goto err_bad_proc; 1373 } 1374 1375 procp = rqstp->rq_procinfo; 1376 /* Should this check go into the dispatcher? */ 1377 if (!procp || !procp->pc_func) 1378 goto err_bad_proc; 1379 1380 /* Syntactic check complete */ 1381 serv->sv_stats->rpccnt++; 1382 trace_svc_process(rqstp, progp->pg_name); 1383 1384 /* Build the reply header. */ 1385 statp = resv->iov_base +resv->iov_len; 1386 svc_putnl(resv, RPC_SUCCESS); 1387 1388 /* un-reserve some of the out-queue now that we have a 1389 * better idea of reply size 1390 */ 1391 if (procp->pc_xdrressize) 1392 svc_reserve_auth(rqstp, procp->pc_xdrressize<<2); 1393 1394 /* Call the function that processes the request. */ 1395 if (!process.dispatch) { 1396 if (!svc_generic_dispatch(rqstp, statp)) 1397 goto release_dropit; 1398 if (*statp == rpc_garbage_args) 1399 goto err_garbage; 1400 } else { 1401 dprintk("svc: calling dispatcher\n"); 1402 if (!process.dispatch(rqstp, statp)) 1403 goto release_dropit; /* Release reply info */ 1404 } 1405 1406 if (rqstp->rq_auth_stat != rpc_auth_ok) 1407 goto err_release_bad_auth; 1408 1409 /* Check RPC status result */ 1410 if (*statp != rpc_success) 1411 resv->iov_len = ((void*)statp) - resv->iov_base + 4; 1412 1413 /* Release reply info */ 1414 if (procp->pc_release) 1415 procp->pc_release(rqstp); 1416 1417 if (procp->pc_encode == NULL) 1418 goto dropit; 1419 1420 sendit: 1421 if (svc_authorise(rqstp)) 1422 goto close_xprt; 1423 return 1; /* Caller can now send it */ 1424 1425 release_dropit: 1426 if (procp->pc_release) 1427 procp->pc_release(rqstp); 1428 dropit: 1429 svc_authorise(rqstp); /* doesn't hurt to call this twice */ 1430 dprintk("svc: svc_process dropit\n"); 1431 return 0; 1432 1433 close: 1434 svc_authorise(rqstp); 1435 close_xprt: 1436 if (rqstp->rq_xprt && test_bit(XPT_TEMP, &rqstp->rq_xprt->xpt_flags)) 1437 svc_close_xprt(rqstp->rq_xprt); 1438 dprintk("svc: svc_process close\n"); 1439 return 0; 1440 1441 err_short_len: 1442 svc_printk(rqstp, "short len %zd, dropping request\n", 1443 argv->iov_len); 1444 goto close_xprt; 1445 1446 err_bad_rpc: 1447 serv->sv_stats->rpcbadfmt++; 1448 svc_putnl(resv, 1); /* REJECT */ 1449 svc_putnl(resv, 0); /* RPC_MISMATCH */ 1450 svc_putnl(resv, 2); /* Only RPCv2 supported */ 1451 svc_putnl(resv, 2); 1452 goto sendit; 1453 1454 err_release_bad_auth: 1455 if (procp->pc_release) 1456 procp->pc_release(rqstp); 1457 err_bad_auth: 1458 dprintk("svc: authentication failed (%d)\n", 1459 be32_to_cpu(rqstp->rq_auth_stat)); 1460 serv->sv_stats->rpcbadauth++; 1461 /* Restore write pointer to location of accept status: */ 1462 xdr_ressize_check(rqstp, reply_statp); 1463 svc_putnl(resv, 1); /* REJECT */ 1464 svc_putnl(resv, 1); /* AUTH_ERROR */ 1465 svc_putu32(resv, rqstp->rq_auth_stat); /* status */ 1466 goto sendit; 1467 1468 err_bad_prog: 1469 dprintk("svc: unknown program %d\n", prog); 1470 serv->sv_stats->rpcbadfmt++; 1471 svc_putnl(resv, RPC_PROG_UNAVAIL); 1472 goto sendit; 1473 1474 err_bad_vers: 1475 svc_printk(rqstp, "unknown version (%d for prog %d, %s)\n", 1476 rqstp->rq_vers, rqstp->rq_prog, progp->pg_name); 1477 1478 serv->sv_stats->rpcbadfmt++; 1479 svc_putnl(resv, RPC_PROG_MISMATCH); 1480 svc_putnl(resv, process.mismatch.lovers); 1481 svc_putnl(resv, process.mismatch.hivers); 1482 goto sendit; 1483 1484 err_bad_proc: 1485 svc_printk(rqstp, "unknown procedure (%d)\n", rqstp->rq_proc); 1486 1487 serv->sv_stats->rpcbadfmt++; 1488 svc_putnl(resv, RPC_PROC_UNAVAIL); 1489 goto sendit; 1490 1491 err_garbage: 1492 svc_printk(rqstp, "failed to decode args\n"); 1493 1494 rpc_stat = rpc_garbage_args; 1495 err_bad: 1496 serv->sv_stats->rpcbadfmt++; 1497 svc_putnl(resv, ntohl(rpc_stat)); 1498 goto sendit; 1499 } 1500 1501 /* 1502 * Process the RPC request. 1503 */ 1504 int 1505 svc_process(struct svc_rqst *rqstp) 1506 { 1507 struct kvec *argv = &rqstp->rq_arg.head[0]; 1508 struct kvec *resv = &rqstp->rq_res.head[0]; 1509 struct svc_serv *serv = rqstp->rq_server; 1510 u32 dir; 1511 1512 #if IS_ENABLED(CONFIG_FAIL_SUNRPC) 1513 if (!fail_sunrpc.ignore_server_disconnect && 1514 should_fail(&fail_sunrpc.attr, 1)) 1515 svc_xprt_deferred_close(rqstp->rq_xprt); 1516 #endif 1517 1518 /* 1519 * Setup response xdr_buf. 1520 * Initially it has just one page 1521 */ 1522 rqstp->rq_next_page = &rqstp->rq_respages[1]; 1523 resv->iov_base = page_address(rqstp->rq_respages[0]); 1524 resv->iov_len = 0; 1525 rqstp->rq_res.pages = rqstp->rq_respages + 1; 1526 rqstp->rq_res.len = 0; 1527 rqstp->rq_res.page_base = 0; 1528 rqstp->rq_res.page_len = 0; 1529 rqstp->rq_res.buflen = PAGE_SIZE; 1530 rqstp->rq_res.tail[0].iov_base = NULL; 1531 rqstp->rq_res.tail[0].iov_len = 0; 1532 1533 dir = svc_getnl(argv); 1534 if (dir != 0) { 1535 /* direction != CALL */ 1536 svc_printk(rqstp, "bad direction %d, dropping request\n", dir); 1537 serv->sv_stats->rpcbadfmt++; 1538 goto out_drop; 1539 } 1540 1541 /* Returns 1 for send, 0 for drop */ 1542 if (likely(svc_process_common(rqstp, argv, resv))) 1543 return svc_send(rqstp); 1544 1545 out_drop: 1546 svc_drop(rqstp); 1547 return 0; 1548 } 1549 EXPORT_SYMBOL_GPL(svc_process); 1550 1551 #if defined(CONFIG_SUNRPC_BACKCHANNEL) 1552 /* 1553 * Process a backchannel RPC request that arrived over an existing 1554 * outbound connection 1555 */ 1556 int 1557 bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, 1558 struct svc_rqst *rqstp) 1559 { 1560 struct kvec *argv = &rqstp->rq_arg.head[0]; 1561 struct kvec *resv = &rqstp->rq_res.head[0]; 1562 struct rpc_task *task; 1563 int proc_error; 1564 int error; 1565 1566 dprintk("svc: %s(%p)\n", __func__, req); 1567 1568 /* Build the svc_rqst used by the common processing routine */ 1569 rqstp->rq_xid = req->rq_xid; 1570 rqstp->rq_prot = req->rq_xprt->prot; 1571 rqstp->rq_server = serv; 1572 rqstp->rq_bc_net = req->rq_xprt->xprt_net; 1573 1574 rqstp->rq_addrlen = sizeof(req->rq_xprt->addr); 1575 memcpy(&rqstp->rq_addr, &req->rq_xprt->addr, rqstp->rq_addrlen); 1576 memcpy(&rqstp->rq_arg, &req->rq_rcv_buf, sizeof(rqstp->rq_arg)); 1577 memcpy(&rqstp->rq_res, &req->rq_snd_buf, sizeof(rqstp->rq_res)); 1578 1579 /* Adjust the argument buffer length */ 1580 rqstp->rq_arg.len = req->rq_private_buf.len; 1581 if (rqstp->rq_arg.len <= rqstp->rq_arg.head[0].iov_len) { 1582 rqstp->rq_arg.head[0].iov_len = rqstp->rq_arg.len; 1583 rqstp->rq_arg.page_len = 0; 1584 } else if (rqstp->rq_arg.len <= rqstp->rq_arg.head[0].iov_len + 1585 rqstp->rq_arg.page_len) 1586 rqstp->rq_arg.page_len = rqstp->rq_arg.len - 1587 rqstp->rq_arg.head[0].iov_len; 1588 else 1589 rqstp->rq_arg.len = rqstp->rq_arg.head[0].iov_len + 1590 rqstp->rq_arg.page_len; 1591 1592 /* reset result send buffer "put" position */ 1593 resv->iov_len = 0; 1594 1595 /* 1596 * Skip the next two words because they've already been 1597 * processed in the transport 1598 */ 1599 svc_getu32(argv); /* XID */ 1600 svc_getnl(argv); /* CALLDIR */ 1601 1602 /* Parse and execute the bc call */ 1603 proc_error = svc_process_common(rqstp, argv, resv); 1604 1605 atomic_dec(&req->rq_xprt->bc_slot_count); 1606 if (!proc_error) { 1607 /* Processing error: drop the request */ 1608 xprt_free_bc_request(req); 1609 error = -EINVAL; 1610 goto out; 1611 } 1612 /* Finally, send the reply synchronously */ 1613 memcpy(&req->rq_snd_buf, &rqstp->rq_res, sizeof(req->rq_snd_buf)); 1614 task = rpc_run_bc_task(req); 1615 if (IS_ERR(task)) { 1616 error = PTR_ERR(task); 1617 goto out; 1618 } 1619 1620 WARN_ON_ONCE(atomic_read(&task->tk_count) != 1); 1621 error = task->tk_status; 1622 rpc_put_task(task); 1623 1624 out: 1625 dprintk("svc: %s(), error=%d\n", __func__, error); 1626 return error; 1627 } 1628 EXPORT_SYMBOL_GPL(bc_svc_process); 1629 #endif /* CONFIG_SUNRPC_BACKCHANNEL */ 1630 1631 /* 1632 * Return (transport-specific) limit on the rpc payload. 1633 */ 1634 u32 svc_max_payload(const struct svc_rqst *rqstp) 1635 { 1636 u32 max = rqstp->rq_xprt->xpt_class->xcl_max_payload; 1637 1638 if (rqstp->rq_server->sv_max_payload < max) 1639 max = rqstp->rq_server->sv_max_payload; 1640 return max; 1641 } 1642 EXPORT_SYMBOL_GPL(svc_max_payload); 1643 1644 /** 1645 * svc_proc_name - Return RPC procedure name in string form 1646 * @rqstp: svc_rqst to operate on 1647 * 1648 * Return value: 1649 * Pointer to a NUL-terminated string 1650 */ 1651 const char *svc_proc_name(const struct svc_rqst *rqstp) 1652 { 1653 if (rqstp && rqstp->rq_procinfo) 1654 return rqstp->rq_procinfo->pc_name; 1655 return "unknown"; 1656 } 1657 1658 1659 /** 1660 * svc_encode_result_payload - mark a range of bytes as a result payload 1661 * @rqstp: svc_rqst to operate on 1662 * @offset: payload's byte offset in rqstp->rq_res 1663 * @length: size of payload, in bytes 1664 * 1665 * Returns zero on success, or a negative errno if a permanent 1666 * error occurred. 1667 */ 1668 int svc_encode_result_payload(struct svc_rqst *rqstp, unsigned int offset, 1669 unsigned int length) 1670 { 1671 return rqstp->rq_xprt->xpt_ops->xpo_result_payload(rqstp, offset, 1672 length); 1673 } 1674 EXPORT_SYMBOL_GPL(svc_encode_result_payload); 1675 1676 /** 1677 * svc_fill_write_vector - Construct data argument for VFS write call 1678 * @rqstp: svc_rqst to operate on 1679 * @pages: list of pages containing data payload 1680 * @first: buffer containing first section of write payload 1681 * @total: total number of bytes of write payload 1682 * 1683 * Fills in rqstp::rq_vec, and returns the number of elements. 1684 */ 1685 unsigned int svc_fill_write_vector(struct svc_rqst *rqstp, struct page **pages, 1686 struct kvec *first, size_t total) 1687 { 1688 struct kvec *vec = rqstp->rq_vec; 1689 unsigned int i; 1690 1691 /* Some types of transport can present the write payload 1692 * entirely in rq_arg.pages. In this case, @first is empty. 1693 */ 1694 i = 0; 1695 if (first->iov_len) { 1696 vec[i].iov_base = first->iov_base; 1697 vec[i].iov_len = min_t(size_t, total, first->iov_len); 1698 total -= vec[i].iov_len; 1699 ++i; 1700 } 1701 1702 while (total) { 1703 vec[i].iov_base = page_address(*pages); 1704 vec[i].iov_len = min_t(size_t, total, PAGE_SIZE); 1705 total -= vec[i].iov_len; 1706 ++i; 1707 ++pages; 1708 } 1709 1710 WARN_ON_ONCE(i > ARRAY_SIZE(rqstp->rq_vec)); 1711 return i; 1712 } 1713 EXPORT_SYMBOL_GPL(svc_fill_write_vector); 1714 1715 /** 1716 * svc_fill_symlink_pathname - Construct pathname argument for VFS symlink call 1717 * @rqstp: svc_rqst to operate on 1718 * @first: buffer containing first section of pathname 1719 * @p: buffer containing remaining section of pathname 1720 * @total: total length of the pathname argument 1721 * 1722 * The VFS symlink API demands a NUL-terminated pathname in mapped memory. 1723 * Returns pointer to a NUL-terminated string, or an ERR_PTR. Caller must free 1724 * the returned string. 1725 */ 1726 char *svc_fill_symlink_pathname(struct svc_rqst *rqstp, struct kvec *first, 1727 void *p, size_t total) 1728 { 1729 size_t len, remaining; 1730 char *result, *dst; 1731 1732 result = kmalloc(total + 1, GFP_KERNEL); 1733 if (!result) 1734 return ERR_PTR(-ESERVERFAULT); 1735 1736 dst = result; 1737 remaining = total; 1738 1739 len = min_t(size_t, total, first->iov_len); 1740 if (len) { 1741 memcpy(dst, first->iov_base, len); 1742 dst += len; 1743 remaining -= len; 1744 } 1745 1746 if (remaining) { 1747 len = min_t(size_t, remaining, PAGE_SIZE); 1748 memcpy(dst, p, len); 1749 dst += len; 1750 } 1751 1752 *dst = '\0'; 1753 1754 /* Sanity check: Linux doesn't allow the pathname argument to 1755 * contain a NUL byte. 1756 */ 1757 if (strlen(result) != total) { 1758 kfree(result); 1759 return ERR_PTR(-EINVAL); 1760 } 1761 return result; 1762 } 1763 EXPORT_SYMBOL_GPL(svc_fill_symlink_pathname); 1764