1 // SPDX-License-Identifier: GPL-2.0 2 /* 3 * Block device elevator/IO-scheduler. 4 * 5 * Copyright (C) 2000 Andrea Arcangeli <andrea@suse.de> SuSE 6 * 7 * 30042000 Jens Axboe <axboe@kernel.dk> : 8 * 9 * Split the elevator a bit so that it is possible to choose a different 10 * one or even write a new "plug in". There are three pieces: 11 * - elevator_fn, inserts a new request in the queue list 12 * - elevator_merge_fn, decides whether a new buffer can be merged with 13 * an existing request 14 * - elevator_dequeue_fn, called when a request is taken off the active list 15 * 16 * 20082000 Dave Jones <davej@suse.de> : 17 * Removed tests for max-bomb-segments, which was breaking elvtune 18 * when run without -bN 19 * 20 * Jens: 21 * - Rework again to work with bio instead of buffer_heads 22 * - loose bi_dev comparisons, partition handling is right now 23 * - completely modularize elevator setup and teardown 24 * 25 */ 26 #include <linux/kernel.h> 27 #include <linux/fs.h> 28 #include <linux/blkdev.h> 29 #include <linux/elevator.h> 30 #include <linux/bio.h> 31 #include <linux/module.h> 32 #include <linux/slab.h> 33 #include <linux/init.h> 34 #include <linux/compiler.h> 35 #include <linux/blktrace_api.h> 36 #include <linux/hash.h> 37 #include <linux/uaccess.h> 38 #include <linux/pm_runtime.h> 39 #include <linux/blk-cgroup.h> 40 41 #include <trace/events/block.h> 42 43 #include "blk.h" 44 #include "blk-mq-sched.h" 45 #include "blk-pm.h" 46 #include "blk-wbt.h" 47 48 static DEFINE_SPINLOCK(elv_list_lock); 49 static LIST_HEAD(elv_list); 50 51 /* 52 * Merge hash stuff. 53 */ 54 #define rq_hash_key(rq) (blk_rq_pos(rq) + blk_rq_sectors(rq)) 55 56 /* 57 * Query io scheduler to see if the current process issuing bio may be 58 * merged with rq. 59 */ 60 static int elv_iosched_allow_bio_merge(struct request *rq, struct bio *bio) 61 { 62 struct request_queue *q = rq->q; 63 struct elevator_queue *e = q->elevator; 64 65 if (e->type->ops.allow_merge) 66 return e->type->ops.allow_merge(q, rq, bio); 67 68 return 1; 69 } 70 71 /* 72 * can we safely merge with this request? 73 */ 74 bool elv_bio_merge_ok(struct request *rq, struct bio *bio) 75 { 76 if (!blk_rq_merge_ok(rq, bio)) 77 return false; 78 79 if (!elv_iosched_allow_bio_merge(rq, bio)) 80 return false; 81 82 return true; 83 } 84 EXPORT_SYMBOL(elv_bio_merge_ok); 85 86 static inline bool elv_support_features(unsigned int elv_features, 87 unsigned int required_features) 88 { 89 return (required_features & elv_features) == required_features; 90 } 91 92 /** 93 * elevator_match - Test an elevator name and features 94 * @e: Scheduler to test 95 * @name: Elevator name to test 96 * @required_features: Features that the elevator must provide 97 * 98 * Return true if the elevator @e name matches @name and if @e provides all 99 * the features specified by @required_features. 100 */ 101 static bool elevator_match(const struct elevator_type *e, const char *name, 102 unsigned int required_features) 103 { 104 if (!elv_support_features(e->elevator_features, required_features)) 105 return false; 106 if (!strcmp(e->elevator_name, name)) 107 return true; 108 if (e->elevator_alias && !strcmp(e->elevator_alias, name)) 109 return true; 110 111 return false; 112 } 113 114 /** 115 * elevator_find - Find an elevator 116 * @name: Name of the elevator to find 117 * @required_features: Features that the elevator must provide 118 * 119 * Return the first registered scheduler with name @name and supporting the 120 * features @required_features and NULL otherwise. 121 */ 122 static struct elevator_type *elevator_find(const char *name, 123 unsigned int required_features) 124 { 125 struct elevator_type *e; 126 127 list_for_each_entry(e, &elv_list, list) { 128 if (elevator_match(e, name, required_features)) 129 return e; 130 } 131 132 return NULL; 133 } 134 135 static void elevator_put(struct elevator_type *e) 136 { 137 module_put(e->elevator_owner); 138 } 139 140 static struct elevator_type *elevator_get(struct request_queue *q, 141 const char *name, bool try_loading) 142 { 143 struct elevator_type *e; 144 145 spin_lock(&elv_list_lock); 146 147 e = elevator_find(name, q->required_elevator_features); 148 if (!e && try_loading) { 149 spin_unlock(&elv_list_lock); 150 request_module("%s-iosched", name); 151 spin_lock(&elv_list_lock); 152 e = elevator_find(name, q->required_elevator_features); 153 } 154 155 if (e && !try_module_get(e->elevator_owner)) 156 e = NULL; 157 158 spin_unlock(&elv_list_lock); 159 return e; 160 } 161 162 static struct kobj_type elv_ktype; 163 164 struct elevator_queue *elevator_alloc(struct request_queue *q, 165 struct elevator_type *e) 166 { 167 struct elevator_queue *eq; 168 169 eq = kzalloc_node(sizeof(*eq), GFP_KERNEL, q->node); 170 if (unlikely(!eq)) 171 return NULL; 172 173 eq->type = e; 174 kobject_init(&eq->kobj, &elv_ktype); 175 mutex_init(&eq->sysfs_lock); 176 hash_init(eq->hash); 177 178 return eq; 179 } 180 EXPORT_SYMBOL(elevator_alloc); 181 182 static void elevator_release(struct kobject *kobj) 183 { 184 struct elevator_queue *e; 185 186 e = container_of(kobj, struct elevator_queue, kobj); 187 elevator_put(e->type); 188 kfree(e); 189 } 190 191 void __elevator_exit(struct request_queue *q, struct elevator_queue *e) 192 { 193 mutex_lock(&e->sysfs_lock); 194 blk_mq_exit_sched(q, e); 195 mutex_unlock(&e->sysfs_lock); 196 197 kobject_put(&e->kobj); 198 } 199 200 static inline void __elv_rqhash_del(struct request *rq) 201 { 202 hash_del(&rq->hash); 203 rq->rq_flags &= ~RQF_HASHED; 204 } 205 206 void elv_rqhash_del(struct request_queue *q, struct request *rq) 207 { 208 if (ELV_ON_HASH(rq)) 209 __elv_rqhash_del(rq); 210 } 211 EXPORT_SYMBOL_GPL(elv_rqhash_del); 212 213 void elv_rqhash_add(struct request_queue *q, struct request *rq) 214 { 215 struct elevator_queue *e = q->elevator; 216 217 BUG_ON(ELV_ON_HASH(rq)); 218 hash_add(e->hash, &rq->hash, rq_hash_key(rq)); 219 rq->rq_flags |= RQF_HASHED; 220 } 221 EXPORT_SYMBOL_GPL(elv_rqhash_add); 222 223 void elv_rqhash_reposition(struct request_queue *q, struct request *rq) 224 { 225 __elv_rqhash_del(rq); 226 elv_rqhash_add(q, rq); 227 } 228 229 struct request *elv_rqhash_find(struct request_queue *q, sector_t offset) 230 { 231 struct elevator_queue *e = q->elevator; 232 struct hlist_node *next; 233 struct request *rq; 234 235 hash_for_each_possible_safe(e->hash, rq, next, hash, offset) { 236 BUG_ON(!ELV_ON_HASH(rq)); 237 238 if (unlikely(!rq_mergeable(rq))) { 239 __elv_rqhash_del(rq); 240 continue; 241 } 242 243 if (rq_hash_key(rq) == offset) 244 return rq; 245 } 246 247 return NULL; 248 } 249 250 /* 251 * RB-tree support functions for inserting/lookup/removal of requests 252 * in a sorted RB tree. 253 */ 254 void elv_rb_add(struct rb_root *root, struct request *rq) 255 { 256 struct rb_node **p = &root->rb_node; 257 struct rb_node *parent = NULL; 258 struct request *__rq; 259 260 while (*p) { 261 parent = *p; 262 __rq = rb_entry(parent, struct request, rb_node); 263 264 if (blk_rq_pos(rq) < blk_rq_pos(__rq)) 265 p = &(*p)->rb_left; 266 else if (blk_rq_pos(rq) >= blk_rq_pos(__rq)) 267 p = &(*p)->rb_right; 268 } 269 270 rb_link_node(&rq->rb_node, parent, p); 271 rb_insert_color(&rq->rb_node, root); 272 } 273 EXPORT_SYMBOL(elv_rb_add); 274 275 void elv_rb_del(struct rb_root *root, struct request *rq) 276 { 277 BUG_ON(RB_EMPTY_NODE(&rq->rb_node)); 278 rb_erase(&rq->rb_node, root); 279 RB_CLEAR_NODE(&rq->rb_node); 280 } 281 EXPORT_SYMBOL(elv_rb_del); 282 283 struct request *elv_rb_find(struct rb_root *root, sector_t sector) 284 { 285 struct rb_node *n = root->rb_node; 286 struct request *rq; 287 288 while (n) { 289 rq = rb_entry(n, struct request, rb_node); 290 291 if (sector < blk_rq_pos(rq)) 292 n = n->rb_left; 293 else if (sector > blk_rq_pos(rq)) 294 n = n->rb_right; 295 else 296 return rq; 297 } 298 299 return NULL; 300 } 301 EXPORT_SYMBOL(elv_rb_find); 302 303 enum elv_merge elv_merge(struct request_queue *q, struct request **req, 304 struct bio *bio) 305 { 306 struct elevator_queue *e = q->elevator; 307 struct request *__rq; 308 309 /* 310 * Levels of merges: 311 * nomerges: No merges at all attempted 312 * noxmerges: Only simple one-hit cache try 313 * merges: All merge tries attempted 314 */ 315 if (blk_queue_nomerges(q) || !bio_mergeable(bio)) 316 return ELEVATOR_NO_MERGE; 317 318 /* 319 * First try one-hit cache. 320 */ 321 if (q->last_merge && elv_bio_merge_ok(q->last_merge, bio)) { 322 enum elv_merge ret = blk_try_merge(q->last_merge, bio); 323 324 if (ret != ELEVATOR_NO_MERGE) { 325 *req = q->last_merge; 326 return ret; 327 } 328 } 329 330 if (blk_queue_noxmerges(q)) 331 return ELEVATOR_NO_MERGE; 332 333 /* 334 * See if our hash lookup can find a potential backmerge. 335 */ 336 __rq = elv_rqhash_find(q, bio->bi_iter.bi_sector); 337 if (__rq && elv_bio_merge_ok(__rq, bio)) { 338 *req = __rq; 339 return ELEVATOR_BACK_MERGE; 340 } 341 342 if (e->type->ops.request_merge) 343 return e->type->ops.request_merge(q, req, bio); 344 345 return ELEVATOR_NO_MERGE; 346 } 347 348 /* 349 * Attempt to do an insertion back merge. Only check for the case where 350 * we can append 'rq' to an existing request, so we can throw 'rq' away 351 * afterwards. 352 * 353 * Returns true if we merged, false otherwise 354 */ 355 bool elv_attempt_insert_merge(struct request_queue *q, struct request *rq) 356 { 357 struct request *__rq; 358 bool ret; 359 360 if (blk_queue_nomerges(q)) 361 return false; 362 363 /* 364 * First try one-hit cache. 365 */ 366 if (q->last_merge && blk_attempt_req_merge(q, q->last_merge, rq)) 367 return true; 368 369 if (blk_queue_noxmerges(q)) 370 return false; 371 372 ret = false; 373 /* 374 * See if our hash lookup can find a potential backmerge. 375 */ 376 while (1) { 377 __rq = elv_rqhash_find(q, blk_rq_pos(rq)); 378 if (!__rq || !blk_attempt_req_merge(q, __rq, rq)) 379 break; 380 381 /* The merged request could be merged with others, try again */ 382 ret = true; 383 rq = __rq; 384 } 385 386 return ret; 387 } 388 389 void elv_merged_request(struct request_queue *q, struct request *rq, 390 enum elv_merge type) 391 { 392 struct elevator_queue *e = q->elevator; 393 394 if (e->type->ops.request_merged) 395 e->type->ops.request_merged(q, rq, type); 396 397 if (type == ELEVATOR_BACK_MERGE) 398 elv_rqhash_reposition(q, rq); 399 400 q->last_merge = rq; 401 } 402 403 void elv_merge_requests(struct request_queue *q, struct request *rq, 404 struct request *next) 405 { 406 struct elevator_queue *e = q->elevator; 407 408 if (e->type->ops.requests_merged) 409 e->type->ops.requests_merged(q, rq, next); 410 411 elv_rqhash_reposition(q, rq); 412 q->last_merge = rq; 413 } 414 415 struct request *elv_latter_request(struct request_queue *q, struct request *rq) 416 { 417 struct elevator_queue *e = q->elevator; 418 419 if (e->type->ops.next_request) 420 return e->type->ops.next_request(q, rq); 421 422 return NULL; 423 } 424 425 struct request *elv_former_request(struct request_queue *q, struct request *rq) 426 { 427 struct elevator_queue *e = q->elevator; 428 429 if (e->type->ops.former_request) 430 return e->type->ops.former_request(q, rq); 431 432 return NULL; 433 } 434 435 #define to_elv(atr) container_of((atr), struct elv_fs_entry, attr) 436 437 static ssize_t 438 elv_attr_show(struct kobject *kobj, struct attribute *attr, char *page) 439 { 440 struct elv_fs_entry *entry = to_elv(attr); 441 struct elevator_queue *e; 442 ssize_t error; 443 444 if (!entry->show) 445 return -EIO; 446 447 e = container_of(kobj, struct elevator_queue, kobj); 448 mutex_lock(&e->sysfs_lock); 449 error = e->type ? entry->show(e, page) : -ENOENT; 450 mutex_unlock(&e->sysfs_lock); 451 return error; 452 } 453 454 static ssize_t 455 elv_attr_store(struct kobject *kobj, struct attribute *attr, 456 const char *page, size_t length) 457 { 458 struct elv_fs_entry *entry = to_elv(attr); 459 struct elevator_queue *e; 460 ssize_t error; 461 462 if (!entry->store) 463 return -EIO; 464 465 e = container_of(kobj, struct elevator_queue, kobj); 466 mutex_lock(&e->sysfs_lock); 467 error = e->type ? entry->store(e, page, length) : -ENOENT; 468 mutex_unlock(&e->sysfs_lock); 469 return error; 470 } 471 472 static const struct sysfs_ops elv_sysfs_ops = { 473 .show = elv_attr_show, 474 .store = elv_attr_store, 475 }; 476 477 static struct kobj_type elv_ktype = { 478 .sysfs_ops = &elv_sysfs_ops, 479 .release = elevator_release, 480 }; 481 482 int elv_register_queue(struct request_queue *q, bool uevent) 483 { 484 struct elevator_queue *e = q->elevator; 485 int error; 486 487 lockdep_assert_held(&q->sysfs_lock); 488 489 error = kobject_add(&e->kobj, &q->kobj, "%s", "iosched"); 490 if (!error) { 491 struct elv_fs_entry *attr = e->type->elevator_attrs; 492 if (attr) { 493 while (attr->attr.name) { 494 if (sysfs_create_file(&e->kobj, &attr->attr)) 495 break; 496 attr++; 497 } 498 } 499 if (uevent) 500 kobject_uevent(&e->kobj, KOBJ_ADD); 501 502 e->registered = 1; 503 } 504 return error; 505 } 506 507 void elv_unregister_queue(struct request_queue *q) 508 { 509 lockdep_assert_held(&q->sysfs_lock); 510 511 if (q) { 512 struct elevator_queue *e = q->elevator; 513 514 kobject_uevent(&e->kobj, KOBJ_REMOVE); 515 kobject_del(&e->kobj); 516 517 e->registered = 0; 518 /* Re-enable throttling in case elevator disabled it */ 519 wbt_enable_default(q); 520 } 521 } 522 523 int elv_register(struct elevator_type *e) 524 { 525 /* create icq_cache if requested */ 526 if (e->icq_size) { 527 if (WARN_ON(e->icq_size < sizeof(struct io_cq)) || 528 WARN_ON(e->icq_align < __alignof__(struct io_cq))) 529 return -EINVAL; 530 531 snprintf(e->icq_cache_name, sizeof(e->icq_cache_name), 532 "%s_io_cq", e->elevator_name); 533 e->icq_cache = kmem_cache_create(e->icq_cache_name, e->icq_size, 534 e->icq_align, 0, NULL); 535 if (!e->icq_cache) 536 return -ENOMEM; 537 } 538 539 /* register, don't allow duplicate names */ 540 spin_lock(&elv_list_lock); 541 if (elevator_find(e->elevator_name, 0)) { 542 spin_unlock(&elv_list_lock); 543 kmem_cache_destroy(e->icq_cache); 544 return -EBUSY; 545 } 546 list_add_tail(&e->list, &elv_list); 547 spin_unlock(&elv_list_lock); 548 549 printk(KERN_INFO "io scheduler %s registered\n", e->elevator_name); 550 551 return 0; 552 } 553 EXPORT_SYMBOL_GPL(elv_register); 554 555 void elv_unregister(struct elevator_type *e) 556 { 557 /* unregister */ 558 spin_lock(&elv_list_lock); 559 list_del_init(&e->list); 560 spin_unlock(&elv_list_lock); 561 562 /* 563 * Destroy icq_cache if it exists. icq's are RCU managed. Make 564 * sure all RCU operations are complete before proceeding. 565 */ 566 if (e->icq_cache) { 567 rcu_barrier(); 568 kmem_cache_destroy(e->icq_cache); 569 e->icq_cache = NULL; 570 } 571 } 572 EXPORT_SYMBOL_GPL(elv_unregister); 573 574 int elevator_switch_mq(struct request_queue *q, 575 struct elevator_type *new_e) 576 { 577 int ret; 578 579 lockdep_assert_held(&q->sysfs_lock); 580 581 if (q->elevator) { 582 if (q->elevator->registered) 583 elv_unregister_queue(q); 584 585 ioc_clear_queue(q); 586 elevator_exit(q, q->elevator); 587 } 588 589 ret = blk_mq_init_sched(q, new_e); 590 if (ret) 591 goto out; 592 593 if (new_e) { 594 ret = elv_register_queue(q, true); 595 if (ret) { 596 elevator_exit(q, q->elevator); 597 goto out; 598 } 599 } 600 601 if (new_e) 602 blk_add_trace_msg(q, "elv switch: %s", new_e->elevator_name); 603 else 604 blk_add_trace_msg(q, "elv switch: none"); 605 606 out: 607 return ret; 608 } 609 610 static inline bool elv_support_iosched(struct request_queue *q) 611 { 612 if (!queue_is_mq(q) || 613 (q->tag_set && (q->tag_set->flags & BLK_MQ_F_NO_SCHED))) 614 return false; 615 return true; 616 } 617 618 /* 619 * For single queue devices, default to using mq-deadline. If we have multiple 620 * queues or mq-deadline is not available, default to "none". 621 */ 622 static struct elevator_type *elevator_get_default(struct request_queue *q) 623 { 624 if (q->nr_hw_queues != 1) 625 return NULL; 626 627 return elevator_get(q, "mq-deadline", false); 628 } 629 630 /* 631 * Get the first elevator providing the features required by the request queue. 632 * Default to "none" if no matching elevator is found. 633 */ 634 static struct elevator_type *elevator_get_by_features(struct request_queue *q) 635 { 636 struct elevator_type *e, *found = NULL; 637 638 spin_lock(&elv_list_lock); 639 640 list_for_each_entry(e, &elv_list, list) { 641 if (elv_support_features(e->elevator_features, 642 q->required_elevator_features)) { 643 found = e; 644 break; 645 } 646 } 647 648 if (found && !try_module_get(found->elevator_owner)) 649 found = NULL; 650 651 spin_unlock(&elv_list_lock); 652 return found; 653 } 654 655 /* 656 * For a device queue that has no required features, use the default elevator 657 * settings. Otherwise, use the first elevator available matching the required 658 * features. If no suitable elevator is find or if the chosen elevator 659 * initialization fails, fall back to the "none" elevator (no elevator). 660 */ 661 void elevator_init_mq(struct request_queue *q) 662 { 663 struct elevator_type *e; 664 int err; 665 666 if (!elv_support_iosched(q)) 667 return; 668 669 WARN_ON_ONCE(blk_queue_registered(q)); 670 671 if (unlikely(q->elevator)) 672 return; 673 674 if (!q->required_elevator_features) 675 e = elevator_get_default(q); 676 else 677 e = elevator_get_by_features(q); 678 if (!e) 679 return; 680 681 blk_mq_freeze_queue(q); 682 blk_mq_quiesce_queue(q); 683 684 err = blk_mq_init_sched(q, e); 685 686 blk_mq_unquiesce_queue(q); 687 blk_mq_unfreeze_queue(q); 688 689 if (err) { 690 pr_warn("\"%s\" elevator initialization failed, " 691 "falling back to \"none\"\n", e->elevator_name); 692 elevator_put(e); 693 } 694 } 695 696 697 /* 698 * switch to new_e io scheduler. be careful not to introduce deadlocks - 699 * we don't free the old io scheduler, before we have allocated what we 700 * need for the new one. this way we have a chance of going back to the old 701 * one, if the new one fails init for some reason. 702 */ 703 static int elevator_switch(struct request_queue *q, struct elevator_type *new_e) 704 { 705 int err; 706 707 lockdep_assert_held(&q->sysfs_lock); 708 709 blk_mq_freeze_queue(q); 710 blk_mq_quiesce_queue(q); 711 712 err = elevator_switch_mq(q, new_e); 713 714 blk_mq_unquiesce_queue(q); 715 blk_mq_unfreeze_queue(q); 716 717 return err; 718 } 719 720 /* 721 * Switch this queue to the given IO scheduler. 722 */ 723 static int __elevator_change(struct request_queue *q, const char *name) 724 { 725 char elevator_name[ELV_NAME_MAX]; 726 struct elevator_type *e; 727 728 /* Make sure queue is not in the middle of being removed */ 729 if (!blk_queue_registered(q)) 730 return -ENOENT; 731 732 /* 733 * Special case for mq, turn off scheduling 734 */ 735 if (!strncmp(name, "none", 4)) { 736 if (!q->elevator) 737 return 0; 738 return elevator_switch(q, NULL); 739 } 740 741 strlcpy(elevator_name, name, sizeof(elevator_name)); 742 e = elevator_get(q, strstrip(elevator_name), true); 743 if (!e) 744 return -EINVAL; 745 746 if (q->elevator && 747 elevator_match(q->elevator->type, elevator_name, 0)) { 748 elevator_put(e); 749 return 0; 750 } 751 752 return elevator_switch(q, e); 753 } 754 755 ssize_t elv_iosched_store(struct request_queue *q, const char *name, 756 size_t count) 757 { 758 int ret; 759 760 if (!elv_support_iosched(q)) 761 return count; 762 763 ret = __elevator_change(q, name); 764 if (!ret) 765 return count; 766 767 return ret; 768 } 769 770 ssize_t elv_iosched_show(struct request_queue *q, char *name) 771 { 772 struct elevator_queue *e = q->elevator; 773 struct elevator_type *elv = NULL; 774 struct elevator_type *__e; 775 int len = 0; 776 777 if (!queue_is_mq(q)) 778 return sprintf(name, "none\n"); 779 780 if (!q->elevator) 781 len += sprintf(name+len, "[none] "); 782 else 783 elv = e->type; 784 785 spin_lock(&elv_list_lock); 786 list_for_each_entry(__e, &elv_list, list) { 787 if (elv && elevator_match(elv, __e->elevator_name, 0)) { 788 len += sprintf(name+len, "[%s] ", elv->elevator_name); 789 continue; 790 } 791 if (elv_support_iosched(q) && 792 elevator_match(__e, __e->elevator_name, 793 q->required_elevator_features)) 794 len += sprintf(name+len, "%s ", __e->elevator_name); 795 } 796 spin_unlock(&elv_list_lock); 797 798 if (q->elevator) 799 len += sprintf(name+len, "none"); 800 801 len += sprintf(len+name, "\n"); 802 return len; 803 } 804 805 struct request *elv_rb_former_request(struct request_queue *q, 806 struct request *rq) 807 { 808 struct rb_node *rbprev = rb_prev(&rq->rb_node); 809 810 if (rbprev) 811 return rb_entry_rq(rbprev); 812 813 return NULL; 814 } 815 EXPORT_SYMBOL(elv_rb_former_request); 816 817 struct request *elv_rb_latter_request(struct request_queue *q, 818 struct request *rq) 819 { 820 struct rb_node *rbnext = rb_next(&rq->rb_node); 821 822 if (rbnext) 823 return rb_entry_rq(rbnext); 824 825 return NULL; 826 } 827 EXPORT_SYMBOL(elv_rb_latter_request); 828 829 static int __init elevator_setup(char *str) 830 { 831 pr_warn("Kernel parameter elevator= does not have any effect anymore.\n" 832 "Please use sysfs to set IO scheduler for individual devices.\n"); 833 return 1; 834 } 835 836 __setup("elevator=", elevator_setup); 837