Line data Source code
1 : // SPDX-License-Identifier: GPL-2.0
2 : /*
3 : * The Kyber I/O scheduler. Controls latency by throttling queue depths using
4 : * scalable techniques.
5 : *
6 : * Copyright (C) 2017 Facebook
7 : */
8 :
9 : #include <linux/kernel.h>
10 : #include <linux/blkdev.h>
11 : #include <linux/module.h>
12 : #include <linux/sbitmap.h>
13 :
14 : #include <trace/events/block.h>
15 :
16 : #include "elevator.h"
17 : #include "blk.h"
18 : #include "blk-mq.h"
19 : #include "blk-mq-debugfs.h"
20 : #include "blk-mq-sched.h"
21 :
22 : #define CREATE_TRACE_POINTS
23 : #include <trace/events/kyber.h>
24 :
25 : /*
26 : * Scheduling domains: the device is divided into multiple domains based on the
27 : * request type.
28 : */
29 : enum {
30 : KYBER_READ,
31 : KYBER_WRITE,
32 : KYBER_DISCARD,
33 : KYBER_OTHER,
34 : KYBER_NUM_DOMAINS,
35 : };
36 :
37 : static const char *kyber_domain_names[] = {
38 : [KYBER_READ] = "READ",
39 : [KYBER_WRITE] = "WRITE",
40 : [KYBER_DISCARD] = "DISCARD",
41 : [KYBER_OTHER] = "OTHER",
42 : };
43 :
44 : enum {
45 : /*
46 : * In order to prevent starvation of synchronous requests by a flood of
47 : * asynchronous requests, we reserve 25% of requests for synchronous
48 : * operations.
49 : */
50 : KYBER_ASYNC_PERCENT = 75,
51 : };
52 :
53 : /*
54 : * Maximum device-wide depth for each scheduling domain.
55 : *
56 : * Even for fast devices with lots of tags like NVMe, you can saturate the
57 : * device with only a fraction of the maximum possible queue depth. So, we cap
58 : * these to a reasonable value.
59 : */
60 : static const unsigned int kyber_depth[] = {
61 : [KYBER_READ] = 256,
62 : [KYBER_WRITE] = 128,
63 : [KYBER_DISCARD] = 64,
64 : [KYBER_OTHER] = 16,
65 : };
66 :
67 : /*
68 : * Default latency targets for each scheduling domain.
69 : */
70 : static const u64 kyber_latency_targets[] = {
71 : [KYBER_READ] = 2ULL * NSEC_PER_MSEC,
72 : [KYBER_WRITE] = 10ULL * NSEC_PER_MSEC,
73 : [KYBER_DISCARD] = 5ULL * NSEC_PER_SEC,
74 : };
75 :
76 : /*
77 : * Batch size (number of requests we'll dispatch in a row) for each scheduling
78 : * domain.
79 : */
80 : static const unsigned int kyber_batch_size[] = {
81 : [KYBER_READ] = 16,
82 : [KYBER_WRITE] = 8,
83 : [KYBER_DISCARD] = 1,
84 : [KYBER_OTHER] = 1,
85 : };
86 :
87 : /*
88 : * Requests latencies are recorded in a histogram with buckets defined relative
89 : * to the target latency:
90 : *
91 : * <= 1/4 * target latency
92 : * <= 1/2 * target latency
93 : * <= 3/4 * target latency
94 : * <= target latency
95 : * <= 1 1/4 * target latency
96 : * <= 1 1/2 * target latency
97 : * <= 1 3/4 * target latency
98 : * > 1 3/4 * target latency
99 : */
100 : enum {
101 : /*
102 : * The width of the latency histogram buckets is
103 : * 1 / (1 << KYBER_LATENCY_SHIFT) * target latency.
104 : */
105 : KYBER_LATENCY_SHIFT = 2,
106 : /*
107 : * The first (1 << KYBER_LATENCY_SHIFT) buckets are <= target latency,
108 : * thus, "good".
109 : */
110 : KYBER_GOOD_BUCKETS = 1 << KYBER_LATENCY_SHIFT,
111 : /* There are also (1 << KYBER_LATENCY_SHIFT) "bad" buckets. */
112 : KYBER_LATENCY_BUCKETS = 2 << KYBER_LATENCY_SHIFT,
113 : };
114 :
115 : /*
116 : * We measure both the total latency and the I/O latency (i.e., latency after
117 : * submitting to the device).
118 : */
119 : enum {
120 : KYBER_TOTAL_LATENCY,
121 : KYBER_IO_LATENCY,
122 : };
123 :
124 : static const char *kyber_latency_type_names[] = {
125 : [KYBER_TOTAL_LATENCY] = "total",
126 : [KYBER_IO_LATENCY] = "I/O",
127 : };
128 :
129 : /*
130 : * Per-cpu latency histograms: total latency and I/O latency for each scheduling
131 : * domain except for KYBER_OTHER.
132 : */
133 : struct kyber_cpu_latency {
134 : atomic_t buckets[KYBER_OTHER][2][KYBER_LATENCY_BUCKETS];
135 : };
136 :
137 : /*
138 : * There is a same mapping between ctx & hctx and kcq & khd,
139 : * we use request->mq_ctx->index_hw to index the kcq in khd.
140 : */
141 : struct kyber_ctx_queue {
142 : /*
143 : * Used to ensure operations on rq_list and kcq_map to be an atmoic one.
144 : * Also protect the rqs on rq_list when merge.
145 : */
146 : spinlock_t lock;
147 : struct list_head rq_list[KYBER_NUM_DOMAINS];
148 : } ____cacheline_aligned_in_smp;
149 :
150 : struct kyber_queue_data {
151 : struct request_queue *q;
152 : dev_t dev;
153 :
154 : /*
155 : * Each scheduling domain has a limited number of in-flight requests
156 : * device-wide, limited by these tokens.
157 : */
158 : struct sbitmap_queue domain_tokens[KYBER_NUM_DOMAINS];
159 :
160 : /*
161 : * Async request percentage, converted to per-word depth for
162 : * sbitmap_get_shallow().
163 : */
164 : unsigned int async_depth;
165 :
166 : struct kyber_cpu_latency __percpu *cpu_latency;
167 :
168 : /* Timer for stats aggregation and adjusting domain tokens. */
169 : struct timer_list timer;
170 :
171 : unsigned int latency_buckets[KYBER_OTHER][2][KYBER_LATENCY_BUCKETS];
172 :
173 : unsigned long latency_timeout[KYBER_OTHER];
174 :
175 : int domain_p99[KYBER_OTHER];
176 :
177 : /* Target latencies in nanoseconds. */
178 : u64 latency_targets[KYBER_OTHER];
179 : };
180 :
181 : struct kyber_hctx_data {
182 : spinlock_t lock;
183 : struct list_head rqs[KYBER_NUM_DOMAINS];
184 : unsigned int cur_domain;
185 : unsigned int batching;
186 : struct kyber_ctx_queue *kcqs;
187 : struct sbitmap kcq_map[KYBER_NUM_DOMAINS];
188 : struct sbq_wait domain_wait[KYBER_NUM_DOMAINS];
189 : struct sbq_wait_state *domain_ws[KYBER_NUM_DOMAINS];
190 : atomic_t wait_index[KYBER_NUM_DOMAINS];
191 : };
192 :
193 : static int kyber_domain_wake(wait_queue_entry_t *wait, unsigned mode, int flags,
194 : void *key);
195 :
196 : static unsigned int kyber_sched_domain(blk_opf_t opf)
197 : {
198 0 : switch (opf & REQ_OP_MASK) {
199 : case REQ_OP_READ:
200 : return KYBER_READ;
201 : case REQ_OP_WRITE:
202 : return KYBER_WRITE;
203 : case REQ_OP_DISCARD:
204 : return KYBER_DISCARD;
205 : default:
206 : return KYBER_OTHER;
207 : }
208 : }
209 :
210 : static void flush_latency_buckets(struct kyber_queue_data *kqd,
211 : struct kyber_cpu_latency *cpu_latency,
212 : unsigned int sched_domain, unsigned int type)
213 : {
214 0 : unsigned int *buckets = kqd->latency_buckets[sched_domain][type];
215 0 : atomic_t *cpu_buckets = cpu_latency->buckets[sched_domain][type];
216 : unsigned int bucket;
217 :
218 0 : for (bucket = 0; bucket < KYBER_LATENCY_BUCKETS; bucket++)
219 0 : buckets[bucket] += atomic_xchg(&cpu_buckets[bucket], 0);
220 : }
221 :
222 : /*
223 : * Calculate the histogram bucket with the given percentile rank, or -1 if there
224 : * aren't enough samples yet.
225 : */
226 0 : static int calculate_percentile(struct kyber_queue_data *kqd,
227 : unsigned int sched_domain, unsigned int type,
228 : unsigned int percentile)
229 : {
230 0 : unsigned int *buckets = kqd->latency_buckets[sched_domain][type];
231 0 : unsigned int bucket, samples = 0, percentile_samples;
232 :
233 0 : for (bucket = 0; bucket < KYBER_LATENCY_BUCKETS; bucket++)
234 0 : samples += buckets[bucket];
235 :
236 0 : if (!samples)
237 : return -1;
238 :
239 : /*
240 : * We do the calculation once we have 500 samples or one second passes
241 : * since the first sample was recorded, whichever comes first.
242 : */
243 0 : if (!kqd->latency_timeout[sched_domain])
244 0 : kqd->latency_timeout[sched_domain] = max(jiffies + HZ, 1UL);
245 0 : if (samples < 500 &&
246 0 : time_is_after_jiffies(kqd->latency_timeout[sched_domain])) {
247 : return -1;
248 : }
249 0 : kqd->latency_timeout[sched_domain] = 0;
250 :
251 0 : percentile_samples = DIV_ROUND_UP(samples * percentile, 100);
252 0 : for (bucket = 0; bucket < KYBER_LATENCY_BUCKETS - 1; bucket++) {
253 0 : if (buckets[bucket] >= percentile_samples)
254 : break;
255 0 : percentile_samples -= buckets[bucket];
256 : }
257 0 : memset(buckets, 0, sizeof(kqd->latency_buckets[sched_domain][type]));
258 :
259 0 : trace_kyber_latency(kqd->dev, kyber_domain_names[sched_domain],
260 : kyber_latency_type_names[type], percentile,
261 : bucket + 1, 1 << KYBER_LATENCY_SHIFT, samples);
262 :
263 0 : return bucket;
264 : }
265 :
266 : static void kyber_resize_domain(struct kyber_queue_data *kqd,
267 : unsigned int sched_domain, unsigned int depth)
268 : {
269 0 : depth = clamp(depth, 1U, kyber_depth[sched_domain]);
270 0 : if (depth != kqd->domain_tokens[sched_domain].sb.depth) {
271 0 : sbitmap_queue_resize(&kqd->domain_tokens[sched_domain], depth);
272 0 : trace_kyber_adjust(kqd->dev, kyber_domain_names[sched_domain],
273 : depth);
274 : }
275 : }
276 :
277 0 : static void kyber_timer_fn(struct timer_list *t)
278 : {
279 0 : struct kyber_queue_data *kqd = from_timer(kqd, t, timer);
280 : unsigned int sched_domain;
281 : int cpu;
282 0 : bool bad = false;
283 :
284 : /* Sum all of the per-cpu latency histograms. */
285 0 : for_each_online_cpu(cpu) {
286 : struct kyber_cpu_latency *cpu_latency;
287 :
288 0 : cpu_latency = per_cpu_ptr(kqd->cpu_latency, cpu);
289 0 : for (sched_domain = 0; sched_domain < KYBER_OTHER; sched_domain++) {
290 0 : flush_latency_buckets(kqd, cpu_latency, sched_domain,
291 : KYBER_TOTAL_LATENCY);
292 0 : flush_latency_buckets(kqd, cpu_latency, sched_domain,
293 : KYBER_IO_LATENCY);
294 : }
295 : }
296 :
297 : /*
298 : * Check if any domains have a high I/O latency, which might indicate
299 : * congestion in the device. Note that we use the p90; we don't want to
300 : * be too sensitive to outliers here.
301 : */
302 0 : for (sched_domain = 0; sched_domain < KYBER_OTHER; sched_domain++) {
303 : int p90;
304 :
305 0 : p90 = calculate_percentile(kqd, sched_domain, KYBER_IO_LATENCY,
306 : 90);
307 0 : if (p90 >= KYBER_GOOD_BUCKETS)
308 0 : bad = true;
309 : }
310 :
311 : /*
312 : * Adjust the scheduling domain depths. If we determined that there was
313 : * congestion, we throttle all domains with good latencies. Either way,
314 : * we ease up on throttling domains with bad latencies.
315 : */
316 0 : for (sched_domain = 0; sched_domain < KYBER_OTHER; sched_domain++) {
317 : unsigned int orig_depth, depth;
318 : int p99;
319 :
320 0 : p99 = calculate_percentile(kqd, sched_domain,
321 : KYBER_TOTAL_LATENCY, 99);
322 : /*
323 : * This is kind of subtle: different domains will not
324 : * necessarily have enough samples to calculate the latency
325 : * percentiles during the same window, so we have to remember
326 : * the p99 for the next time we observe congestion; once we do,
327 : * we don't want to throttle again until we get more data, so we
328 : * reset it to -1.
329 : */
330 0 : if (bad) {
331 0 : if (p99 < 0)
332 0 : p99 = kqd->domain_p99[sched_domain];
333 0 : kqd->domain_p99[sched_domain] = -1;
334 0 : } else if (p99 >= 0) {
335 0 : kqd->domain_p99[sched_domain] = p99;
336 : }
337 0 : if (p99 < 0)
338 0 : continue;
339 :
340 : /*
341 : * If this domain has bad latency, throttle less. Otherwise,
342 : * throttle more iff we determined that there is congestion.
343 : *
344 : * The new depth is scaled linearly with the p99 latency vs the
345 : * latency target. E.g., if the p99 is 3/4 of the target, then
346 : * we throttle down to 3/4 of the current depth, and if the p99
347 : * is 2x the target, then we double the depth.
348 : */
349 0 : if (bad || p99 >= KYBER_GOOD_BUCKETS) {
350 0 : orig_depth = kqd->domain_tokens[sched_domain].sb.depth;
351 0 : depth = (orig_depth * (p99 + 1)) >> KYBER_LATENCY_SHIFT;
352 : kyber_resize_domain(kqd, sched_domain, depth);
353 : }
354 : }
355 0 : }
356 :
357 0 : static struct kyber_queue_data *kyber_queue_data_alloc(struct request_queue *q)
358 : {
359 : struct kyber_queue_data *kqd;
360 0 : int ret = -ENOMEM;
361 : int i;
362 :
363 0 : kqd = kzalloc_node(sizeof(*kqd), GFP_KERNEL, q->node);
364 0 : if (!kqd)
365 : goto err;
366 :
367 0 : kqd->q = q;
368 0 : kqd->dev = disk_devt(q->disk);
369 :
370 0 : kqd->cpu_latency = alloc_percpu_gfp(struct kyber_cpu_latency,
371 : GFP_KERNEL | __GFP_ZERO);
372 0 : if (!kqd->cpu_latency)
373 : goto err_kqd;
374 :
375 0 : timer_setup(&kqd->timer, kyber_timer_fn, 0);
376 :
377 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++) {
378 0 : WARN_ON(!kyber_depth[i]);
379 0 : WARN_ON(!kyber_batch_size[i]);
380 0 : ret = sbitmap_queue_init_node(&kqd->domain_tokens[i],
381 : kyber_depth[i], -1, false,
382 : GFP_KERNEL, q->node);
383 0 : if (ret) {
384 0 : while (--i >= 0)
385 0 : sbitmap_queue_free(&kqd->domain_tokens[i]);
386 : goto err_buckets;
387 : }
388 : }
389 :
390 0 : for (i = 0; i < KYBER_OTHER; i++) {
391 0 : kqd->domain_p99[i] = -1;
392 0 : kqd->latency_targets[i] = kyber_latency_targets[i];
393 : }
394 :
395 : return kqd;
396 :
397 : err_buckets:
398 0 : free_percpu(kqd->cpu_latency);
399 : err_kqd:
400 0 : kfree(kqd);
401 : err:
402 0 : return ERR_PTR(ret);
403 : }
404 :
405 0 : static int kyber_init_sched(struct request_queue *q, struct elevator_type *e)
406 : {
407 : struct kyber_queue_data *kqd;
408 : struct elevator_queue *eq;
409 :
410 0 : eq = elevator_alloc(q, e);
411 0 : if (!eq)
412 : return -ENOMEM;
413 :
414 0 : kqd = kyber_queue_data_alloc(q);
415 0 : if (IS_ERR(kqd)) {
416 0 : kobject_put(&eq->kobj);
417 0 : return PTR_ERR(kqd);
418 : }
419 :
420 0 : blk_stat_enable_accounting(q);
421 :
422 0 : blk_queue_flag_clear(QUEUE_FLAG_SQ_SCHED, q);
423 :
424 0 : eq->elevator_data = kqd;
425 0 : q->elevator = eq;
426 :
427 0 : return 0;
428 : }
429 :
430 0 : static void kyber_exit_sched(struct elevator_queue *e)
431 : {
432 0 : struct kyber_queue_data *kqd = e->elevator_data;
433 : int i;
434 :
435 0 : timer_shutdown_sync(&kqd->timer);
436 0 : blk_stat_disable_accounting(kqd->q);
437 :
438 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++)
439 0 : sbitmap_queue_free(&kqd->domain_tokens[i]);
440 0 : free_percpu(kqd->cpu_latency);
441 0 : kfree(kqd);
442 0 : }
443 :
444 : static void kyber_ctx_queue_init(struct kyber_ctx_queue *kcq)
445 : {
446 : unsigned int i;
447 :
448 0 : spin_lock_init(&kcq->lock);
449 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++)
450 0 : INIT_LIST_HEAD(&kcq->rq_list[i]);
451 : }
452 :
453 0 : static void kyber_depth_updated(struct blk_mq_hw_ctx *hctx)
454 : {
455 0 : struct kyber_queue_data *kqd = hctx->queue->elevator->elevator_data;
456 0 : struct blk_mq_tags *tags = hctx->sched_tags;
457 0 : unsigned int shift = tags->bitmap_tags.sb.shift;
458 :
459 0 : kqd->async_depth = (1U << shift) * KYBER_ASYNC_PERCENT / 100U;
460 :
461 0 : sbitmap_queue_min_shallow_depth(&tags->bitmap_tags, kqd->async_depth);
462 0 : }
463 :
464 0 : static int kyber_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
465 : {
466 : struct kyber_hctx_data *khd;
467 : int i;
468 :
469 0 : khd = kmalloc_node(sizeof(*khd), GFP_KERNEL, hctx->numa_node);
470 0 : if (!khd)
471 : return -ENOMEM;
472 :
473 0 : khd->kcqs = kmalloc_array_node(hctx->nr_ctx,
474 : sizeof(struct kyber_ctx_queue),
475 0 : GFP_KERNEL, hctx->numa_node);
476 0 : if (!khd->kcqs)
477 : goto err_khd;
478 :
479 0 : for (i = 0; i < hctx->nr_ctx; i++)
480 0 : kyber_ctx_queue_init(&khd->kcqs[i]);
481 :
482 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++) {
483 0 : if (sbitmap_init_node(&khd->kcq_map[i], hctx->nr_ctx,
484 0 : ilog2(8), GFP_KERNEL, hctx->numa_node,
485 : false, false)) {
486 0 : while (--i >= 0)
487 0 : sbitmap_free(&khd->kcq_map[i]);
488 : goto err_kcqs;
489 : }
490 : }
491 :
492 : spin_lock_init(&khd->lock);
493 :
494 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++) {
495 0 : INIT_LIST_HEAD(&khd->rqs[i]);
496 0 : khd->domain_wait[i].sbq = NULL;
497 0 : init_waitqueue_func_entry(&khd->domain_wait[i].wait,
498 : kyber_domain_wake);
499 0 : khd->domain_wait[i].wait.private = hctx;
500 0 : INIT_LIST_HEAD(&khd->domain_wait[i].wait.entry);
501 0 : atomic_set(&khd->wait_index[i], 0);
502 : }
503 :
504 0 : khd->cur_domain = 0;
505 0 : khd->batching = 0;
506 :
507 0 : hctx->sched_data = khd;
508 0 : kyber_depth_updated(hctx);
509 :
510 0 : return 0;
511 :
512 : err_kcqs:
513 0 : kfree(khd->kcqs);
514 : err_khd:
515 0 : kfree(khd);
516 0 : return -ENOMEM;
517 : }
518 :
519 0 : static void kyber_exit_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
520 : {
521 0 : struct kyber_hctx_data *khd = hctx->sched_data;
522 : int i;
523 :
524 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++)
525 0 : sbitmap_free(&khd->kcq_map[i]);
526 0 : kfree(khd->kcqs);
527 0 : kfree(hctx->sched_data);
528 0 : }
529 :
530 : static int rq_get_domain_token(struct request *rq)
531 : {
532 0 : return (long)rq->elv.priv[0];
533 : }
534 :
535 : static void rq_set_domain_token(struct request *rq, int token)
536 : {
537 0 : rq->elv.priv[0] = (void *)(long)token;
538 : }
539 :
540 0 : static void rq_clear_domain_token(struct kyber_queue_data *kqd,
541 : struct request *rq)
542 : {
543 : unsigned int sched_domain;
544 : int nr;
545 :
546 0 : nr = rq_get_domain_token(rq);
547 0 : if (nr != -1) {
548 0 : sched_domain = kyber_sched_domain(rq->cmd_flags);
549 0 : sbitmap_queue_clear(&kqd->domain_tokens[sched_domain], nr,
550 0 : rq->mq_ctx->cpu);
551 : }
552 0 : }
553 :
554 0 : static void kyber_limit_depth(blk_opf_t opf, struct blk_mq_alloc_data *data)
555 : {
556 : /*
557 : * We use the scheduler tags as per-hardware queue queueing tokens.
558 : * Async requests can be limited at this stage.
559 : */
560 0 : if (!op_is_sync(opf)) {
561 0 : struct kyber_queue_data *kqd = data->q->elevator->elevator_data;
562 :
563 0 : data->shallow_depth = kqd->async_depth;
564 : }
565 0 : }
566 :
567 0 : static bool kyber_bio_merge(struct request_queue *q, struct bio *bio,
568 : unsigned int nr_segs)
569 : {
570 0 : struct blk_mq_ctx *ctx = blk_mq_get_ctx(q);
571 0 : struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, bio->bi_opf, ctx);
572 0 : struct kyber_hctx_data *khd = hctx->sched_data;
573 0 : struct kyber_ctx_queue *kcq = &khd->kcqs[ctx->index_hw[hctx->type]];
574 0 : unsigned int sched_domain = kyber_sched_domain(bio->bi_opf);
575 0 : struct list_head *rq_list = &kcq->rq_list[sched_domain];
576 : bool merged;
577 :
578 0 : spin_lock(&kcq->lock);
579 0 : merged = blk_bio_list_merge(hctx->queue, rq_list, bio, nr_segs);
580 0 : spin_unlock(&kcq->lock);
581 :
582 0 : return merged;
583 : }
584 :
585 0 : static void kyber_prepare_request(struct request *rq)
586 : {
587 0 : rq_set_domain_token(rq, -1);
588 0 : }
589 :
590 0 : static void kyber_insert_requests(struct blk_mq_hw_ctx *hctx,
591 : struct list_head *rq_list,
592 : blk_insert_t flags)
593 : {
594 0 : struct kyber_hctx_data *khd = hctx->sched_data;
595 : struct request *rq, *next;
596 :
597 0 : list_for_each_entry_safe(rq, next, rq_list, queuelist) {
598 0 : unsigned int sched_domain = kyber_sched_domain(rq->cmd_flags);
599 0 : struct kyber_ctx_queue *kcq = &khd->kcqs[rq->mq_ctx->index_hw[hctx->type]];
600 0 : struct list_head *head = &kcq->rq_list[sched_domain];
601 :
602 0 : spin_lock(&kcq->lock);
603 0 : trace_block_rq_insert(rq);
604 0 : if (flags & BLK_MQ_INSERT_AT_HEAD)
605 0 : list_move(&rq->queuelist, head);
606 : else
607 0 : list_move_tail(&rq->queuelist, head);
608 0 : sbitmap_set_bit(&khd->kcq_map[sched_domain],
609 0 : rq->mq_ctx->index_hw[hctx->type]);
610 0 : spin_unlock(&kcq->lock);
611 : }
612 0 : }
613 :
614 0 : static void kyber_finish_request(struct request *rq)
615 : {
616 0 : struct kyber_queue_data *kqd = rq->q->elevator->elevator_data;
617 :
618 0 : rq_clear_domain_token(kqd, rq);
619 0 : }
620 :
621 : static void add_latency_sample(struct kyber_cpu_latency *cpu_latency,
622 : unsigned int sched_domain, unsigned int type,
623 : u64 target, u64 latency)
624 : {
625 : unsigned int bucket;
626 : u64 divisor;
627 :
628 0 : if (latency > 0) {
629 0 : divisor = max_t(u64, target >> KYBER_LATENCY_SHIFT, 1);
630 0 : bucket = min_t(unsigned int, div64_u64(latency - 1, divisor),
631 : KYBER_LATENCY_BUCKETS - 1);
632 : } else {
633 : bucket = 0;
634 : }
635 :
636 0 : atomic_inc(&cpu_latency->buckets[sched_domain][type][bucket]);
637 : }
638 :
639 0 : static void kyber_completed_request(struct request *rq, u64 now)
640 : {
641 0 : struct kyber_queue_data *kqd = rq->q->elevator->elevator_data;
642 : struct kyber_cpu_latency *cpu_latency;
643 : unsigned int sched_domain;
644 : u64 target;
645 :
646 0 : sched_domain = kyber_sched_domain(rq->cmd_flags);
647 0 : if (sched_domain == KYBER_OTHER)
648 : return;
649 :
650 0 : cpu_latency = get_cpu_ptr(kqd->cpu_latency);
651 0 : target = kqd->latency_targets[sched_domain];
652 0 : add_latency_sample(cpu_latency, sched_domain, KYBER_TOTAL_LATENCY,
653 0 : target, now - rq->start_time_ns);
654 0 : add_latency_sample(cpu_latency, sched_domain, KYBER_IO_LATENCY, target,
655 0 : now - rq->io_start_time_ns);
656 0 : put_cpu_ptr(kqd->cpu_latency);
657 :
658 0 : timer_reduce(&kqd->timer, jiffies + HZ / 10);
659 : }
660 :
661 : struct flush_kcq_data {
662 : struct kyber_hctx_data *khd;
663 : unsigned int sched_domain;
664 : struct list_head *list;
665 : };
666 :
667 0 : static bool flush_busy_kcq(struct sbitmap *sb, unsigned int bitnr, void *data)
668 : {
669 0 : struct flush_kcq_data *flush_data = data;
670 0 : struct kyber_ctx_queue *kcq = &flush_data->khd->kcqs[bitnr];
671 :
672 0 : spin_lock(&kcq->lock);
673 0 : list_splice_tail_init(&kcq->rq_list[flush_data->sched_domain],
674 : flush_data->list);
675 0 : sbitmap_clear_bit(sb, bitnr);
676 0 : spin_unlock(&kcq->lock);
677 :
678 0 : return true;
679 : }
680 :
681 : static void kyber_flush_busy_kcqs(struct kyber_hctx_data *khd,
682 : unsigned int sched_domain,
683 : struct list_head *list)
684 : {
685 0 : struct flush_kcq_data data = {
686 : .khd = khd,
687 : .sched_domain = sched_domain,
688 : .list = list,
689 : };
690 :
691 0 : sbitmap_for_each_set(&khd->kcq_map[sched_domain],
692 : flush_busy_kcq, &data);
693 : }
694 :
695 0 : static int kyber_domain_wake(wait_queue_entry_t *wqe, unsigned mode, int flags,
696 : void *key)
697 : {
698 0 : struct blk_mq_hw_ctx *hctx = READ_ONCE(wqe->private);
699 0 : struct sbq_wait *wait = container_of(wqe, struct sbq_wait, wait);
700 :
701 0 : sbitmap_del_wait_queue(wait);
702 0 : blk_mq_run_hw_queue(hctx, true);
703 0 : return 1;
704 : }
705 :
706 0 : static int kyber_get_domain_token(struct kyber_queue_data *kqd,
707 : struct kyber_hctx_data *khd,
708 : struct blk_mq_hw_ctx *hctx)
709 : {
710 0 : unsigned int sched_domain = khd->cur_domain;
711 0 : struct sbitmap_queue *domain_tokens = &kqd->domain_tokens[sched_domain];
712 0 : struct sbq_wait *wait = &khd->domain_wait[sched_domain];
713 : struct sbq_wait_state *ws;
714 : int nr;
715 :
716 0 : nr = __sbitmap_queue_get(domain_tokens);
717 :
718 : /*
719 : * If we failed to get a domain token, make sure the hardware queue is
720 : * run when one becomes available. Note that this is serialized on
721 : * khd->lock, but we still need to be careful about the waker.
722 : */
723 0 : if (nr < 0 && list_empty_careful(&wait->wait.entry)) {
724 0 : ws = sbq_wait_ptr(domain_tokens,
725 : &khd->wait_index[sched_domain]);
726 0 : khd->domain_ws[sched_domain] = ws;
727 0 : sbitmap_add_wait_queue(domain_tokens, ws, wait);
728 :
729 : /*
730 : * Try again in case a token was freed before we got on the wait
731 : * queue.
732 : */
733 0 : nr = __sbitmap_queue_get(domain_tokens);
734 : }
735 :
736 : /*
737 : * If we got a token while we were on the wait queue, remove ourselves
738 : * from the wait queue to ensure that all wake ups make forward
739 : * progress. It's possible that the waker already deleted the entry
740 : * between the !list_empty_careful() check and us grabbing the lock, but
741 : * list_del_init() is okay with that.
742 : */
743 0 : if (nr >= 0 && !list_empty_careful(&wait->wait.entry)) {
744 0 : ws = khd->domain_ws[sched_domain];
745 0 : spin_lock_irq(&ws->wait.lock);
746 0 : sbitmap_del_wait_queue(wait);
747 0 : spin_unlock_irq(&ws->wait.lock);
748 : }
749 :
750 0 : return nr;
751 : }
752 :
753 : static struct request *
754 0 : kyber_dispatch_cur_domain(struct kyber_queue_data *kqd,
755 : struct kyber_hctx_data *khd,
756 : struct blk_mq_hw_ctx *hctx)
757 : {
758 : struct list_head *rqs;
759 : struct request *rq;
760 : int nr;
761 :
762 0 : rqs = &khd->rqs[khd->cur_domain];
763 :
764 : /*
765 : * If we already have a flushed request, then we just need to get a
766 : * token for it. Otherwise, if there are pending requests in the kcqs,
767 : * flush the kcqs, but only if we can get a token. If not, we should
768 : * leave the requests in the kcqs so that they can be merged. Note that
769 : * khd->lock serializes the flushes, so if we observed any bit set in
770 : * the kcq_map, we will always get a request.
771 : */
772 0 : rq = list_first_entry_or_null(rqs, struct request, queuelist);
773 0 : if (rq) {
774 0 : nr = kyber_get_domain_token(kqd, khd, hctx);
775 0 : if (nr >= 0) {
776 0 : khd->batching++;
777 0 : rq_set_domain_token(rq, nr);
778 0 : list_del_init(&rq->queuelist);
779 : return rq;
780 : } else {
781 : trace_kyber_throttled(kqd->dev,
782 : kyber_domain_names[khd->cur_domain]);
783 : }
784 0 : } else if (sbitmap_any_bit_set(&khd->kcq_map[khd->cur_domain])) {
785 0 : nr = kyber_get_domain_token(kqd, khd, hctx);
786 0 : if (nr >= 0) {
787 0 : kyber_flush_busy_kcqs(khd, khd->cur_domain, rqs);
788 0 : rq = list_first_entry(rqs, struct request, queuelist);
789 0 : khd->batching++;
790 0 : rq_set_domain_token(rq, nr);
791 0 : list_del_init(&rq->queuelist);
792 : return rq;
793 : } else {
794 : trace_kyber_throttled(kqd->dev,
795 : kyber_domain_names[khd->cur_domain]);
796 : }
797 : }
798 :
799 : /* There were either no pending requests or no tokens. */
800 : return NULL;
801 : }
802 :
803 0 : static struct request *kyber_dispatch_request(struct blk_mq_hw_ctx *hctx)
804 : {
805 0 : struct kyber_queue_data *kqd = hctx->queue->elevator->elevator_data;
806 0 : struct kyber_hctx_data *khd = hctx->sched_data;
807 : struct request *rq;
808 : int i;
809 :
810 0 : spin_lock(&khd->lock);
811 :
812 : /*
813 : * First, if we are still entitled to batch, try to dispatch a request
814 : * from the batch.
815 : */
816 0 : if (khd->batching < kyber_batch_size[khd->cur_domain]) {
817 0 : rq = kyber_dispatch_cur_domain(kqd, khd, hctx);
818 0 : if (rq)
819 : goto out;
820 : }
821 :
822 : /*
823 : * Either,
824 : * 1. We were no longer entitled to a batch.
825 : * 2. The domain we were batching didn't have any requests.
826 : * 3. The domain we were batching was out of tokens.
827 : *
828 : * Start another batch. Note that this wraps back around to the original
829 : * domain if no other domains have requests or tokens.
830 : */
831 0 : khd->batching = 0;
832 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++) {
833 0 : if (khd->cur_domain == KYBER_NUM_DOMAINS - 1)
834 0 : khd->cur_domain = 0;
835 : else
836 0 : khd->cur_domain++;
837 :
838 0 : rq = kyber_dispatch_cur_domain(kqd, khd, hctx);
839 0 : if (rq)
840 : goto out;
841 : }
842 :
843 : rq = NULL;
844 : out:
845 0 : spin_unlock(&khd->lock);
846 0 : return rq;
847 : }
848 :
849 0 : static bool kyber_has_work(struct blk_mq_hw_ctx *hctx)
850 : {
851 0 : struct kyber_hctx_data *khd = hctx->sched_data;
852 : int i;
853 :
854 0 : for (i = 0; i < KYBER_NUM_DOMAINS; i++) {
855 0 : if (!list_empty_careful(&khd->rqs[i]) ||
856 0 : sbitmap_any_bit_set(&khd->kcq_map[i]))
857 : return true;
858 : }
859 :
860 : return false;
861 : }
862 :
863 : #define KYBER_LAT_SHOW_STORE(domain, name) \
864 : static ssize_t kyber_##name##_lat_show(struct elevator_queue *e, \
865 : char *page) \
866 : { \
867 : struct kyber_queue_data *kqd = e->elevator_data; \
868 : \
869 : return sprintf(page, "%llu\n", kqd->latency_targets[domain]); \
870 : } \
871 : \
872 : static ssize_t kyber_##name##_lat_store(struct elevator_queue *e, \
873 : const char *page, size_t count) \
874 : { \
875 : struct kyber_queue_data *kqd = e->elevator_data; \
876 : unsigned long long nsec; \
877 : int ret; \
878 : \
879 : ret = kstrtoull(page, 10, &nsec); \
880 : if (ret) \
881 : return ret; \
882 : \
883 : kqd->latency_targets[domain] = nsec; \
884 : \
885 : return count; \
886 : }
887 0 : KYBER_LAT_SHOW_STORE(KYBER_READ, read);
888 0 : KYBER_LAT_SHOW_STORE(KYBER_WRITE, write);
889 : #undef KYBER_LAT_SHOW_STORE
890 :
891 : #define KYBER_LAT_ATTR(op) __ATTR(op##_lat_nsec, 0644, kyber_##op##_lat_show, kyber_##op##_lat_store)
892 : static struct elv_fs_entry kyber_sched_attrs[] = {
893 : KYBER_LAT_ATTR(read),
894 : KYBER_LAT_ATTR(write),
895 : __ATTR_NULL
896 : };
897 : #undef KYBER_LAT_ATTR
898 :
899 : #ifdef CONFIG_BLK_DEBUG_FS
900 : #define KYBER_DEBUGFS_DOMAIN_ATTRS(domain, name) \
901 : static int kyber_##name##_tokens_show(void *data, struct seq_file *m) \
902 : { \
903 : struct request_queue *q = data; \
904 : struct kyber_queue_data *kqd = q->elevator->elevator_data; \
905 : \
906 : sbitmap_queue_show(&kqd->domain_tokens[domain], m); \
907 : return 0; \
908 : } \
909 : \
910 : static void *kyber_##name##_rqs_start(struct seq_file *m, loff_t *pos) \
911 : __acquires(&khd->lock) \
912 : { \
913 : struct blk_mq_hw_ctx *hctx = m->private; \
914 : struct kyber_hctx_data *khd = hctx->sched_data; \
915 : \
916 : spin_lock(&khd->lock); \
917 : return seq_list_start(&khd->rqs[domain], *pos); \
918 : } \
919 : \
920 : static void *kyber_##name##_rqs_next(struct seq_file *m, void *v, \
921 : loff_t *pos) \
922 : { \
923 : struct blk_mq_hw_ctx *hctx = m->private; \
924 : struct kyber_hctx_data *khd = hctx->sched_data; \
925 : \
926 : return seq_list_next(v, &khd->rqs[domain], pos); \
927 : } \
928 : \
929 : static void kyber_##name##_rqs_stop(struct seq_file *m, void *v) \
930 : __releases(&khd->lock) \
931 : { \
932 : struct blk_mq_hw_ctx *hctx = m->private; \
933 : struct kyber_hctx_data *khd = hctx->sched_data; \
934 : \
935 : spin_unlock(&khd->lock); \
936 : } \
937 : \
938 : static const struct seq_operations kyber_##name##_rqs_seq_ops = { \
939 : .start = kyber_##name##_rqs_start, \
940 : .next = kyber_##name##_rqs_next, \
941 : .stop = kyber_##name##_rqs_stop, \
942 : .show = blk_mq_debugfs_rq_show, \
943 : }; \
944 : \
945 : static int kyber_##name##_waiting_show(void *data, struct seq_file *m) \
946 : { \
947 : struct blk_mq_hw_ctx *hctx = data; \
948 : struct kyber_hctx_data *khd = hctx->sched_data; \
949 : wait_queue_entry_t *wait = &khd->domain_wait[domain].wait; \
950 : \
951 : seq_printf(m, "%d\n", !list_empty_careful(&wait->entry)); \
952 : return 0; \
953 : }
954 : KYBER_DEBUGFS_DOMAIN_ATTRS(KYBER_READ, read)
955 : KYBER_DEBUGFS_DOMAIN_ATTRS(KYBER_WRITE, write)
956 : KYBER_DEBUGFS_DOMAIN_ATTRS(KYBER_DISCARD, discard)
957 : KYBER_DEBUGFS_DOMAIN_ATTRS(KYBER_OTHER, other)
958 : #undef KYBER_DEBUGFS_DOMAIN_ATTRS
959 :
960 : static int kyber_async_depth_show(void *data, struct seq_file *m)
961 : {
962 : struct request_queue *q = data;
963 : struct kyber_queue_data *kqd = q->elevator->elevator_data;
964 :
965 : seq_printf(m, "%u\n", kqd->async_depth);
966 : return 0;
967 : }
968 :
969 : static int kyber_cur_domain_show(void *data, struct seq_file *m)
970 : {
971 : struct blk_mq_hw_ctx *hctx = data;
972 : struct kyber_hctx_data *khd = hctx->sched_data;
973 :
974 : seq_printf(m, "%s\n", kyber_domain_names[khd->cur_domain]);
975 : return 0;
976 : }
977 :
978 : static int kyber_batching_show(void *data, struct seq_file *m)
979 : {
980 : struct blk_mq_hw_ctx *hctx = data;
981 : struct kyber_hctx_data *khd = hctx->sched_data;
982 :
983 : seq_printf(m, "%u\n", khd->batching);
984 : return 0;
985 : }
986 :
987 : #define KYBER_QUEUE_DOMAIN_ATTRS(name) \
988 : {#name "_tokens", 0400, kyber_##name##_tokens_show}
989 : static const struct blk_mq_debugfs_attr kyber_queue_debugfs_attrs[] = {
990 : KYBER_QUEUE_DOMAIN_ATTRS(read),
991 : KYBER_QUEUE_DOMAIN_ATTRS(write),
992 : KYBER_QUEUE_DOMAIN_ATTRS(discard),
993 : KYBER_QUEUE_DOMAIN_ATTRS(other),
994 : {"async_depth", 0400, kyber_async_depth_show},
995 : {},
996 : };
997 : #undef KYBER_QUEUE_DOMAIN_ATTRS
998 :
999 : #define KYBER_HCTX_DOMAIN_ATTRS(name) \
1000 : {#name "_rqs", 0400, .seq_ops = &kyber_##name##_rqs_seq_ops}, \
1001 : {#name "_waiting", 0400, kyber_##name##_waiting_show}
1002 : static const struct blk_mq_debugfs_attr kyber_hctx_debugfs_attrs[] = {
1003 : KYBER_HCTX_DOMAIN_ATTRS(read),
1004 : KYBER_HCTX_DOMAIN_ATTRS(write),
1005 : KYBER_HCTX_DOMAIN_ATTRS(discard),
1006 : KYBER_HCTX_DOMAIN_ATTRS(other),
1007 : {"cur_domain", 0400, kyber_cur_domain_show},
1008 : {"batching", 0400, kyber_batching_show},
1009 : {},
1010 : };
1011 : #undef KYBER_HCTX_DOMAIN_ATTRS
1012 : #endif
1013 :
1014 : static struct elevator_type kyber_sched = {
1015 : .ops = {
1016 : .init_sched = kyber_init_sched,
1017 : .exit_sched = kyber_exit_sched,
1018 : .init_hctx = kyber_init_hctx,
1019 : .exit_hctx = kyber_exit_hctx,
1020 : .limit_depth = kyber_limit_depth,
1021 : .bio_merge = kyber_bio_merge,
1022 : .prepare_request = kyber_prepare_request,
1023 : .insert_requests = kyber_insert_requests,
1024 : .finish_request = kyber_finish_request,
1025 : .requeue_request = kyber_finish_request,
1026 : .completed_request = kyber_completed_request,
1027 : .dispatch_request = kyber_dispatch_request,
1028 : .has_work = kyber_has_work,
1029 : .depth_updated = kyber_depth_updated,
1030 : },
1031 : #ifdef CONFIG_BLK_DEBUG_FS
1032 : .queue_debugfs_attrs = kyber_queue_debugfs_attrs,
1033 : .hctx_debugfs_attrs = kyber_hctx_debugfs_attrs,
1034 : #endif
1035 : .elevator_attrs = kyber_sched_attrs,
1036 : .elevator_name = "kyber",
1037 : .elevator_owner = THIS_MODULE,
1038 : };
1039 :
1040 1 : static int __init kyber_init(void)
1041 : {
1042 1 : return elv_register(&kyber_sched);
1043 : }
1044 :
1045 0 : static void __exit kyber_exit(void)
1046 : {
1047 0 : elv_unregister(&kyber_sched);
1048 0 : }
1049 :
1050 : module_init(kyber_init);
1051 : module_exit(kyber_exit);
1052 :
1053 : MODULE_AUTHOR("Omar Sandoval");
1054 : MODULE_LICENSE("GPL");
1055 : MODULE_DESCRIPTION("Kyber I/O scheduler");
|
