Blame view
block/blk-mq-sched.c
14.5 KB
bd166ef18
|
1 2 3 4 5 6 7 8 9 10 11 12 13 |
/* * blk-mq scheduling framework * * Copyright (C) 2016 Jens Axboe */ #include <linux/kernel.h> #include <linux/module.h> #include <linux/blk-mq.h> #include <trace/events/block.h> #include "blk.h" #include "blk-mq.h" |
d332ce091
|
14 |
#include "blk-mq-debugfs.h" |
bd166ef18
|
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 |
#include "blk-mq-sched.h" #include "blk-mq-tag.h" #include "blk-wbt.h" void blk_mq_sched_free_hctx_data(struct request_queue *q, void (*exit)(struct blk_mq_hw_ctx *)) { struct blk_mq_hw_ctx *hctx; int i; queue_for_each_hw_ctx(q, hctx, i) { if (exit && hctx->sched_data) exit(hctx); kfree(hctx->sched_data); hctx->sched_data = NULL; } } EXPORT_SYMBOL_GPL(blk_mq_sched_free_hctx_data); |
44e8c2bff
|
33 |
void blk_mq_sched_assign_ioc(struct request *rq, struct bio *bio) |
bd166ef18
|
34 |
{ |
44e8c2bff
|
35 36 |
struct request_queue *q = rq->q; struct io_context *ioc = rq_ioc(bio); |
bd166ef18
|
37 38 39 40 41 42 43 44 45 46 47 |
struct io_cq *icq; spin_lock_irq(q->queue_lock); icq = ioc_lookup_icq(ioc, q); spin_unlock_irq(q->queue_lock); if (!icq) { icq = ioc_create_icq(ioc, q, GFP_ATOMIC); if (!icq) return; } |
ea511e3c2
|
48 |
get_io_context(icq->ioc); |
44e8c2bff
|
49 |
rq->elv.icq = icq; |
bd166ef18
|
50 |
} |
8e8320c93
|
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 |
/* * Mark a hardware queue as needing a restart. For shared queues, maintain * a count of how many hardware queues are marked for restart. */ static void blk_mq_sched_mark_restart_hctx(struct blk_mq_hw_ctx *hctx) { if (test_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state)) return; if (hctx->flags & BLK_MQ_F_TAG_SHARED) { struct request_queue *q = hctx->queue; if (!test_and_set_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state)) atomic_inc(&q->shared_hctx_restart); } else set_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state); } static bool blk_mq_sched_restart_hctx(struct blk_mq_hw_ctx *hctx) { if (!test_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state)) return false; if (hctx->flags & BLK_MQ_F_TAG_SHARED) { struct request_queue *q = hctx->queue; if (test_and_clear_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state)) atomic_dec(&q->shared_hctx_restart); } else clear_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state); if (blk_mq_hctx_has_pending(hctx)) { blk_mq_run_hw_queue(hctx, true); return true; } return false; } |
bd166ef18
|
89 90 |
void blk_mq_sched_dispatch_requests(struct blk_mq_hw_ctx *hctx) { |
81380ca10
|
91 92 |
struct request_queue *q = hctx->queue; struct elevator_queue *e = q->elevator; |
64765a75e
|
93 |
const bool has_sched_dispatch = e && e->type->ops.mq.dispatch_request; |
85dcb3c85
|
94 |
bool do_sched_dispatch = true; |
bd166ef18
|
95 |
LIST_HEAD(rq_list); |
f4560ffe8
|
96 97 |
/* RCU or SRCU read lock is needed before checking quiesced flag */ if (unlikely(blk_mq_hctx_stopped(hctx) || blk_queue_quiesced(q))) |
bd166ef18
|
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 |
return; hctx->run++; /* * If we have previous entries on our dispatch list, grab them first for * more fair dispatch. */ if (!list_empty_careful(&hctx->dispatch)) { spin_lock(&hctx->lock); if (!list_empty(&hctx->dispatch)) list_splice_init(&hctx->dispatch, &rq_list); spin_unlock(&hctx->lock); } /* * Only ask the scheduler for requests, if we didn't have residual * requests from the dispatch list. This is to avoid the case where * we only ever dispatch a fraction of the requests available because * of low device queue depth. Once we pull requests out of the IO * scheduler, we can no longer merge or sort them. So it's best to * leave them there for as long as we can. Mark the hw queue as * needing a restart in that case. */ |
c13660a08
|
122 |
if (!list_empty(&rq_list)) { |
d38d35155
|
123 |
blk_mq_sched_mark_restart_hctx(hctx); |
85dcb3c85
|
124 |
do_sched_dispatch = blk_mq_dispatch_rq_list(q, &rq_list); |
64765a75e
|
125 |
} else if (!has_sched_dispatch) { |
c13660a08
|
126 |
blk_mq_flush_busy_ctxs(hctx, &rq_list); |
81380ca10
|
127 |
blk_mq_dispatch_rq_list(q, &rq_list); |
64765a75e
|
128 129 130 |
} /* |
85dcb3c85
|
131 132 133 |
* We want to dispatch from the scheduler if there was nothing * on the dispatch list or we were able to dispatch from the * dispatch list. |
64765a75e
|
134 |
*/ |
85dcb3c85
|
135 |
if (do_sched_dispatch && has_sched_dispatch) { |
c13660a08
|
136 137 138 139 140 141 142 |
do { struct request *rq; rq = e->type->ops.mq.dispatch_request(hctx); if (!rq) break; list_add(&rq->queuelist, &rq_list); |
81380ca10
|
143 |
} while (blk_mq_dispatch_rq_list(q, &rq_list)); |
c13660a08
|
144 |
} |
bd166ef18
|
145 |
} |
e4d750c97
|
146 147 |
bool blk_mq_sched_try_merge(struct request_queue *q, struct bio *bio, struct request **merged_request) |
bd166ef18
|
148 149 |
{ struct request *rq; |
bd166ef18
|
150 |
|
34fe7c054
|
151 152 |
switch (elv_merge(q, &rq, bio)) { case ELEVATOR_BACK_MERGE: |
bd166ef18
|
153 154 |
if (!blk_mq_sched_allow_merge(q, rq, bio)) return false; |
34fe7c054
|
155 156 157 158 159 160 161 |
if (!bio_attempt_back_merge(q, rq, bio)) return false; *merged_request = attempt_back_merge(q, rq); if (!*merged_request) elv_merged_request(q, rq, ELEVATOR_BACK_MERGE); return true; case ELEVATOR_FRONT_MERGE: |
bd166ef18
|
162 163 |
if (!blk_mq_sched_allow_merge(q, rq, bio)) return false; |
34fe7c054
|
164 165 166 167 168 169 170 171 |
if (!bio_attempt_front_merge(q, rq, bio)) return false; *merged_request = attempt_front_merge(q, rq); if (!*merged_request) elv_merged_request(q, rq, ELEVATOR_FRONT_MERGE); return true; default: return false; |
bd166ef18
|
172 |
} |
bd166ef18
|
173 174 |
} EXPORT_SYMBOL_GPL(blk_mq_sched_try_merge); |
9bddeb2a5
|
175 176 177 178 179 180 181 182 183 184 |
/* * Reverse check our software queue for entries that we could potentially * merge with. Currently includes a hand-wavy stop count of 8, to not spend * too much time checking for merges. */ static bool blk_mq_attempt_merge(struct request_queue *q, struct blk_mq_ctx *ctx, struct bio *bio) { struct request *rq; int checked = 8; |
7b6078146
|
185 |
lockdep_assert_held(&ctx->lock); |
9bddeb2a5
|
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 |
list_for_each_entry_reverse(rq, &ctx->rq_list, queuelist) { bool merged = false; if (!checked--) break; if (!blk_rq_merge_ok(rq, bio)) continue; switch (blk_try_merge(rq, bio)) { case ELEVATOR_BACK_MERGE: if (blk_mq_sched_allow_merge(q, rq, bio)) merged = bio_attempt_back_merge(q, rq, bio); break; case ELEVATOR_FRONT_MERGE: if (blk_mq_sched_allow_merge(q, rq, bio)) merged = bio_attempt_front_merge(q, rq, bio); break; case ELEVATOR_DISCARD_MERGE: merged = bio_attempt_discard_merge(q, rq, bio); break; default: continue; } if (merged) ctx->rq_merged++; return merged; } return false; } |
bd166ef18
|
218 219 220 |
bool __blk_mq_sched_bio_merge(struct request_queue *q, struct bio *bio) { struct elevator_queue *e = q->elevator; |
9bddeb2a5
|
221 222 223 |
struct blk_mq_ctx *ctx = blk_mq_get_ctx(q); struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, ctx->cpu); bool ret = false; |
bd166ef18
|
224 |
|
9bddeb2a5
|
225 |
if (e && e->type->ops.mq.bio_merge) { |
bd166ef18
|
226 227 228 |
blk_mq_put_ctx(ctx); return e->type->ops.mq.bio_merge(hctx, bio); } |
07a252b47
|
229 230 |
if ((hctx->flags & BLK_MQ_F_SHOULD_MERGE) && !list_empty_careful(&ctx->rq_list)) { |
9bddeb2a5
|
231 232 233 234 235 236 237 238 |
/* default per sw-queue merge */ spin_lock(&ctx->lock); ret = blk_mq_attempt_merge(q, ctx, bio); spin_unlock(&ctx->lock); } blk_mq_put_ctx(ctx); return ret; |
bd166ef18
|
239 240 241 242 243 244 245 246 247 248 249 250 251 |
} bool blk_mq_sched_try_insert_merge(struct request_queue *q, struct request *rq) { return rq_mergeable(rq) && elv_attempt_insert_merge(q, rq); } EXPORT_SYMBOL_GPL(blk_mq_sched_try_insert_merge); void blk_mq_sched_request_inserted(struct request *rq) { trace_block_rq_insert(rq->q, rq); } EXPORT_SYMBOL_GPL(blk_mq_sched_request_inserted); |
0cacba6cf
|
252 253 |
static bool blk_mq_sched_bypass_insert(struct blk_mq_hw_ctx *hctx, struct request *rq) |
bd166ef18
|
254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 |
{ if (rq->tag == -1) { rq->rq_flags |= RQF_SORTED; return false; } /* * If we already have a real request tag, send directly to * the dispatch list. */ spin_lock(&hctx->lock); list_add(&rq->queuelist, &hctx->dispatch); spin_unlock(&hctx->lock); return true; } |
bd166ef18
|
269 |
|
6d8c6c0f9
|
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 |
/** * list_for_each_entry_rcu_rr - iterate in a round-robin fashion over rcu list * @pos: loop cursor. * @skip: the list element that will not be examined. Iteration starts at * @skip->next. * @head: head of the list to examine. This list must have at least one * element, namely @skip. * @member: name of the list_head structure within typeof(*pos). */ #define list_for_each_entry_rcu_rr(pos, skip, head, member) \ for ((pos) = (skip); \ (pos = (pos)->member.next != (head) ? list_entry_rcu( \ (pos)->member.next, typeof(*pos), member) : \ list_entry_rcu((pos)->member.next->next, typeof(*pos), member)), \ (pos) != (skip); ) |
50e1dab86
|
285 |
|
6d8c6c0f9
|
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 |
/* * Called after a driver tag has been freed to check whether a hctx needs to * be restarted. Restarts @hctx if its tag set is not shared. Restarts hardware * queues in a round-robin fashion if the tag set of @hctx is shared with other * hardware queues. */ void blk_mq_sched_restart(struct blk_mq_hw_ctx *const hctx) { struct blk_mq_tags *const tags = hctx->tags; struct blk_mq_tag_set *const set = hctx->queue->tag_set; struct request_queue *const queue = hctx->queue, *q; struct blk_mq_hw_ctx *hctx2; unsigned int i, j; if (set->flags & BLK_MQ_F_TAG_SHARED) { |
8e8320c93
|
301 302 303 304 305 306 |
/* * If this is 0, then we know that no hardware queues * have RESTART marked. We're done. */ if (!atomic_read(&queue->shared_hctx_restart)) return; |
6d8c6c0f9
|
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 |
rcu_read_lock(); list_for_each_entry_rcu_rr(q, queue, &set->tag_list, tag_set_list) { queue_for_each_hw_ctx(q, hctx2, i) if (hctx2->tags == tags && blk_mq_sched_restart_hctx(hctx2)) goto done; } j = hctx->queue_num + 1; for (i = 0; i < queue->nr_hw_queues; i++, j++) { if (j == queue->nr_hw_queues) j = 0; hctx2 = queue->queue_hw_ctx[j]; if (hctx2->tags == tags && blk_mq_sched_restart_hctx(hctx2)) break; |
d38d35155
|
323 |
} |
6d8c6c0f9
|
324 325 |
done: rcu_read_unlock(); |
d38d35155
|
326 |
} else { |
50e1dab86
|
327 |
blk_mq_sched_restart_hctx(hctx); |
50e1dab86
|
328 329 |
} } |
bd6737f1a
|
330 331 332 333 334 335 336 337 338 339 340 341 |
/* * Add flush/fua to the queue. If we fail getting a driver tag, then * punt to the requeue list. Requeue will re-invoke us from a context * that's safe to block from. */ static void blk_mq_sched_insert_flush(struct blk_mq_hw_ctx *hctx, struct request *rq, bool can_block) { if (blk_mq_get_driver_tag(rq, &hctx, can_block)) { blk_insert_flush(rq); blk_mq_run_hw_queue(hctx, true); } else |
c7a571b45
|
342 |
blk_mq_add_to_requeue_list(rq, false, true); |
bd6737f1a
|
343 344 345 346 347 348 349 350 351 |
} void blk_mq_sched_insert_request(struct request *rq, bool at_head, bool run_queue, bool async, bool can_block) { struct request_queue *q = rq->q; struct elevator_queue *e = q->elevator; struct blk_mq_ctx *ctx = rq->mq_ctx; struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, ctx->cpu); |
f3a8ab7d5
|
352 |
if (rq->tag == -1 && op_is_flush(rq->cmd_flags)) { |
bd6737f1a
|
353 354 355 |
blk_mq_sched_insert_flush(hctx, rq, can_block); return; } |
0cacba6cf
|
356 357 |
if (e && blk_mq_sched_bypass_insert(hctx, rq)) goto run; |
bd6737f1a
|
358 359 360 361 362 363 364 365 366 367 |
if (e && e->type->ops.mq.insert_requests) { LIST_HEAD(list); list_add(&rq->queuelist, &list); e->type->ops.mq.insert_requests(hctx, &list, at_head); } else { spin_lock(&ctx->lock); __blk_mq_insert_request(hctx, rq, at_head); spin_unlock(&ctx->lock); } |
0cacba6cf
|
368 |
run: |
bd6737f1a
|
369 370 371 372 373 374 375 376 377 378 |
if (run_queue) blk_mq_run_hw_queue(hctx, async); } void blk_mq_sched_insert_requests(struct request_queue *q, struct blk_mq_ctx *ctx, struct list_head *list, bool run_queue_async) { struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, ctx->cpu); struct elevator_queue *e = hctx->queue->elevator; |
0cacba6cf
|
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 |
if (e) { struct request *rq, *next; /* * We bypass requests that already have a driver tag assigned, * which should only be flushes. Flushes are only ever inserted * as single requests, so we shouldn't ever hit the * WARN_ON_ONCE() below (but let's handle it just in case). */ list_for_each_entry_safe(rq, next, list, queuelist) { if (WARN_ON_ONCE(rq->tag != -1)) { list_del_init(&rq->queuelist); blk_mq_sched_bypass_insert(hctx, rq); } } } |
bd6737f1a
|
395 396 397 398 399 400 401 |
if (e && e->type->ops.mq.insert_requests) e->type->ops.mq.insert_requests(hctx, list, false); else blk_mq_insert_requests(hctx, ctx, list); blk_mq_run_hw_queue(hctx, run_queue_async); } |
bd166ef18
|
402 403 404 405 406 407 408 409 410 411 |
static void blk_mq_sched_free_tags(struct blk_mq_tag_set *set, struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx) { if (hctx->sched_tags) { blk_mq_free_rqs(set, hctx->sched_tags, hctx_idx); blk_mq_free_rq_map(hctx->sched_tags); hctx->sched_tags = NULL; } } |
6917ff0b5
|
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 |
static int blk_mq_sched_alloc_tags(struct request_queue *q, struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx) { struct blk_mq_tag_set *set = q->tag_set; int ret; hctx->sched_tags = blk_mq_alloc_rq_map(set, hctx_idx, q->nr_requests, set->reserved_tags); if (!hctx->sched_tags) return -ENOMEM; ret = blk_mq_alloc_rqs(set, hctx->sched_tags, hctx_idx, q->nr_requests); if (ret) blk_mq_sched_free_tags(set, hctx, hctx_idx); return ret; } |
54d5329d4
|
430 |
static void blk_mq_sched_tags_teardown(struct request_queue *q) |
bd166ef18
|
431 432 433 |
{ struct blk_mq_tag_set *set = q->tag_set; struct blk_mq_hw_ctx *hctx; |
6917ff0b5
|
434 435 436 437 438 |
int i; queue_for_each_hw_ctx(q, hctx, i) blk_mq_sched_free_tags(set, hctx, i); } |
93252632e
|
439 440 441 442 |
int blk_mq_sched_init_hctx(struct request_queue *q, struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx) { struct elevator_queue *e = q->elevator; |
ee056f981
|
443 |
int ret; |
93252632e
|
444 445 446 |
if (!e) return 0; |
ee056f981
|
447 448 449 450 451 452 453 454 455 456 457 |
ret = blk_mq_sched_alloc_tags(q, hctx, hctx_idx); if (ret) return ret; if (e->type->ops.mq.init_hctx) { ret = e->type->ops.mq.init_hctx(hctx, hctx_idx); if (ret) { blk_mq_sched_free_tags(q->tag_set, hctx, hctx_idx); return ret; } } |
d332ce091
|
458 |
blk_mq_debugfs_register_sched_hctx(q, hctx); |
ee056f981
|
459 |
return 0; |
93252632e
|
460 461 462 463 464 465 466 467 468 |
} void blk_mq_sched_exit_hctx(struct request_queue *q, struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx) { struct elevator_queue *e = q->elevator; if (!e) return; |
d332ce091
|
469 |
blk_mq_debugfs_unregister_sched_hctx(hctx); |
ee056f981
|
470 471 472 473 |
if (e->type->ops.mq.exit_hctx && hctx->sched_data) { e->type->ops.mq.exit_hctx(hctx, hctx_idx); hctx->sched_data = NULL; } |
93252632e
|
474 475 |
blk_mq_sched_free_tags(q->tag_set, hctx, hctx_idx); } |
6917ff0b5
|
476 477 478 |
int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e) { struct blk_mq_hw_ctx *hctx; |
ee056f981
|
479 |
struct elevator_queue *eq; |
6917ff0b5
|
480 481 482 483 484 485 486 |
unsigned int i; int ret; if (!e) { q->elevator = NULL; return 0; } |
bd166ef18
|
487 488 |
/* |
32825c45f
|
489 490 491 |
* Default to double of smaller one between hw queue_depth and 128, * since we don't split into sync/async like the old code did. * Additionally, this is a per-hw queue depth. |
bd166ef18
|
492 |
*/ |
32825c45f
|
493 494 |
q->nr_requests = 2 * min_t(unsigned int, q->tag_set->queue_depth, BLKDEV_MAX_RQ); |
bd166ef18
|
495 |
|
bd166ef18
|
496 |
queue_for_each_hw_ctx(q, hctx, i) { |
6917ff0b5
|
497 |
ret = blk_mq_sched_alloc_tags(q, hctx, i); |
bd166ef18
|
498 |
if (ret) |
6917ff0b5
|
499 |
goto err; |
bd166ef18
|
500 |
} |
6917ff0b5
|
501 502 503 |
ret = e->ops.mq.init_sched(q, e); if (ret) goto err; |
bd166ef18
|
504 |
|
d332ce091
|
505 506 507 508 |
blk_mq_debugfs_register_sched(q); queue_for_each_hw_ctx(q, hctx, i) { if (e->ops.mq.init_hctx) { |
ee056f981
|
509 510 511 512 513 514 515 516 |
ret = e->ops.mq.init_hctx(hctx, i); if (ret) { eq = q->elevator; blk_mq_exit_sched(q, eq); kobject_put(&eq->kobj); return ret; } } |
d332ce091
|
517 |
blk_mq_debugfs_register_sched_hctx(q, hctx); |
ee056f981
|
518 |
} |
bd166ef18
|
519 |
return 0; |
bd166ef18
|
520 |
|
6917ff0b5
|
521 |
err: |
54d5329d4
|
522 523 |
blk_mq_sched_tags_teardown(q); q->elevator = NULL; |
6917ff0b5
|
524 |
return ret; |
bd166ef18
|
525 |
} |
d34849913
|
526 |
|
54d5329d4
|
527 528 |
void blk_mq_exit_sched(struct request_queue *q, struct elevator_queue *e) { |
ee056f981
|
529 530 |
struct blk_mq_hw_ctx *hctx; unsigned int i; |
d332ce091
|
531 532 533 534 535 |
queue_for_each_hw_ctx(q, hctx, i) { blk_mq_debugfs_unregister_sched_hctx(hctx); if (e->type->ops.mq.exit_hctx && hctx->sched_data) { e->type->ops.mq.exit_hctx(hctx, i); hctx->sched_data = NULL; |
ee056f981
|
536 537 |
} } |
d332ce091
|
538 |
blk_mq_debugfs_unregister_sched(q); |
54d5329d4
|
539 540 541 542 543 |
if (e->type->ops.mq.exit_sched) e->type->ops.mq.exit_sched(e); blk_mq_sched_tags_teardown(q); q->elevator = NULL; } |
d34849913
|
544 545 546 |
int blk_mq_sched_init(struct request_queue *q) { int ret; |
d34849913
|
547 548 549 550 551 552 |
mutex_lock(&q->sysfs_lock); ret = elevator_init(q, NULL); mutex_unlock(&q->sysfs_lock); return ret; } |