1 // SPDX-License-Identifier: GPL-2.0-or-later
3 * Budget Fair Queueing (BFQ) I/O scheduler.
5 * Based on ideas and code from CFQ:
6 * Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
8 * Copyright (C) 2008 Fabio Checconi <fabio@gandalf.sssup.it>
9 * Paolo Valente <paolo.valente@unimore.it>
11 * Copyright (C) 2010 Paolo Valente <paolo.valente@unimore.it>
12 * Arianna Avanzini <avanzini@google.com>
14 * Copyright (C) 2017 Paolo Valente <paolo.valente@linaro.org>
16 * BFQ is a proportional-share I/O scheduler, with some extra
17 * low-latency capabilities. BFQ also supports full hierarchical
18 * scheduling through cgroups. Next paragraphs provide an introduction
19 * on BFQ inner workings. Details on BFQ benefits, usage and
20 * limitations can be found in Documentation/block/bfq-iosched.rst.
22 * BFQ is a proportional-share storage-I/O scheduling algorithm based
23 * on the slice-by-slice service scheme of CFQ. But BFQ assigns
24 * budgets, measured in number of sectors, to processes instead of
25 * time slices. The device is not granted to the in-service process
26 * for a given time slice, but until it has exhausted its assigned
27 * budget. This change from the time to the service domain enables BFQ
28 * to distribute the device throughput among processes as desired,
29 * without any distortion due to throughput fluctuations, or to device
30 * internal queueing. BFQ uses an ad hoc internal scheduler, called
31 * B-WF2Q+, to schedule processes according to their budgets. More
32 * precisely, BFQ schedules queues associated with processes. Each
33 * process/queue is assigned a user-configurable weight, and B-WF2Q+
34 * guarantees that each queue receives a fraction of the throughput
35 * proportional to its weight. Thanks to the accurate policy of
36 * B-WF2Q+, BFQ can afford to assign high budgets to I/O-bound
37 * processes issuing sequential requests (to boost the throughput),
38 * and yet guarantee a low latency to interactive and soft real-time
41 * In particular, to provide these low-latency guarantees, BFQ
42 * explicitly privileges the I/O of two classes of time-sensitive
43 * applications: interactive and soft real-time. In more detail, BFQ
44 * behaves this way if the low_latency parameter is set (default
45 * configuration). This feature enables BFQ to provide applications in
46 * these classes with a very low latency.
48 * To implement this feature, BFQ constantly tries to detect whether
49 * the I/O requests in a bfq_queue come from an interactive or a soft
50 * real-time application. For brevity, in these cases, the queue is
51 * said to be interactive or soft real-time. In both cases, BFQ
52 * privileges the service of the queue, over that of non-interactive
53 * and non-soft-real-time queues. This privileging is performed,
54 * mainly, by raising the weight of the queue. So, for brevity, we
55 * call just weight-raising periods the time periods during which a
56 * queue is privileged, because deemed interactive or soft real-time.
58 * The detection of soft real-time queues/applications is described in
59 * detail in the comments on the function
60 * bfq_bfqq_softrt_next_start. On the other hand, the detection of an
61 * interactive queue works as follows: a queue is deemed interactive
62 * if it is constantly non empty only for a limited time interval,
63 * after which it does become empty. The queue may be deemed
64 * interactive again (for a limited time), if it restarts being
65 * constantly non empty, provided that this happens only after the
66 * queue has remained empty for a given minimum idle time.
68 * By default, BFQ computes automatically the above maximum time
69 * interval, i.e., the time interval after which a constantly
70 * non-empty queue stops being deemed interactive. Since a queue is
71 * weight-raised while it is deemed interactive, this maximum time
72 * interval happens to coincide with the (maximum) duration of the
73 * weight-raising for interactive queues.
75 * Finally, BFQ also features additional heuristics for
76 * preserving both a low latency and a high throughput on NCQ-capable,
77 * rotational or flash-based devices, and to get the job done quickly
78 * for applications consisting in many I/O-bound processes.
80 * NOTE: if the main or only goal, with a given device, is to achieve
81 * the maximum-possible throughput at all times, then do switch off
82 * all low-latency heuristics for that device, by setting low_latency
85 * BFQ is described in [1], where also a reference to the initial,
86 * more theoretical paper on BFQ can be found. The interested reader
87 * can find in the latter paper full details on the main algorithm, as
88 * well as formulas of the guarantees and formal proofs of all the
89 * properties. With respect to the version of BFQ presented in these
90 * papers, this implementation adds a few more heuristics, such as the
91 * ones that guarantee a low latency to interactive and soft real-time
92 * applications, and a hierarchical extension based on H-WF2Q+.
94 * B-WF2Q+ is based on WF2Q+, which is described in [2], together with
95 * H-WF2Q+, while the augmented tree used here to implement B-WF2Q+
96 * with O(log N) complexity derives from the one introduced with EEVDF
99 * [1] P. Valente, A. Avanzini, "Evolution of the BFQ Storage I/O
100 * Scheduler", Proceedings of the First Workshop on Mobile System
101 * Technologies (MST-2015), May 2015.
102 * http://algogroup.unimore.it/people/paolo/disk_sched/mst-2015.pdf
104 * [2] Jon C.R. Bennett and H. Zhang, "Hierarchical Packet Fair Queueing
105 * Algorithms", IEEE/ACM Transactions on Networking, 5(5):675-689,
108 * http://www.cs.cmu.edu/~hzhang/papers/TON-97-Oct.ps.gz
110 * [3] I. Stoica and H. Abdel-Wahab, "Earliest Eligible Virtual Deadline
111 * First: A Flexible and Accurate Mechanism for Proportional Share
112 * Resource Allocation", technical report.
114 * http://www.cs.berkeley.edu/~istoica/papers/eevdf-tr-95.pdf
116 #include <linux/module.h>
117 #include <linux/slab.h>
118 #include <linux/blkdev.h>
119 #include <linux/cgroup.h>
120 #include <linux/elevator.h>
121 #include <linux/ktime.h>
122 #include <linux/rbtree.h>
123 #include <linux/ioprio.h>
124 #include <linux/sbitmap.h>
125 #include <linux/delay.h>
126 #include <linux/backing-dev.h>
128 #include <trace/events/block.h>
132 #include "blk-mq-tag.h"
133 #include "blk-mq-sched.h"
134 #include "bfq-iosched.h"
137 #define BFQ_BFQQ_FNS(name) \
138 void bfq_mark_bfqq_##name(struct bfq_queue *bfqq) \
140 __set_bit(BFQQF_##name, &(bfqq)->flags); \
142 void bfq_clear_bfqq_##name(struct bfq_queue *bfqq) \
144 __clear_bit(BFQQF_##name, &(bfqq)->flags); \
146 int bfq_bfqq_##name(const struct bfq_queue *bfqq) \
148 return test_bit(BFQQF_##name, &(bfqq)->flags); \
151 BFQ_BFQQ_FNS(just_created);
153 BFQ_BFQQ_FNS(wait_request);
154 BFQ_BFQQ_FNS(non_blocking_wait_rq);
155 BFQ_BFQQ_FNS(fifo_expire);
156 BFQ_BFQQ_FNS(has_short_ttime);
158 BFQ_BFQQ_FNS(IO_bound);
159 BFQ_BFQQ_FNS(in_large_burst);
161 BFQ_BFQQ_FNS(split_coop);
162 BFQ_BFQQ_FNS(softrt_update);
163 #undef BFQ_BFQQ_FNS \
165 /* Expiration time of async (0) and sync (1) requests, in ns. */
166 static const u64 bfq_fifo_expire[2] = { NSEC_PER_SEC / 4, NSEC_PER_SEC / 8 };
168 /* Maximum backwards seek (magic number lifted from CFQ), in KiB. */
169 static const int bfq_back_max = 16 * 1024;
171 /* Penalty of a backwards seek, in number of sectors. */
172 static const int bfq_back_penalty = 2;
174 /* Idling period duration, in ns. */
175 static u64 bfq_slice_idle = NSEC_PER_SEC / 125;
177 /* Minimum number of assigned budgets for which stats are safe to compute. */
178 static const int bfq_stats_min_budgets = 194;
180 /* Default maximum budget values, in sectors and number of requests. */
181 static const int bfq_default_max_budget = 16 * 1024;
184 * When a sync request is dispatched, the queue that contains that
185 * request, and all the ancestor entities of that queue, are charged
186 * with the number of sectors of the request. In contrast, if the
187 * request is async, then the queue and its ancestor entities are
188 * charged with the number of sectors of the request, multiplied by
189 * the factor below. This throttles the bandwidth for async I/O,
190 * w.r.t. to sync I/O, and it is done to counter the tendency of async
191 * writes to steal I/O throughput to reads.
193 * The current value of this parameter is the result of a tuning with
194 * several hardware and software configurations. We tried to find the
195 * lowest value for which writes do not cause noticeable problems to
196 * reads. In fact, the lower this parameter, the stabler I/O control,
197 * in the following respect. The lower this parameter is, the less
198 * the bandwidth enjoyed by a group decreases
199 * - when the group does writes, w.r.t. to when it does reads;
200 * - when other groups do reads, w.r.t. to when they do writes.
202 static const int bfq_async_charge_factor = 3;
204 /* Default timeout values, in jiffies, approximating CFQ defaults. */
205 const int bfq_timeout = HZ / 8;
208 * Time limit for merging (see comments in bfq_setup_cooperator). Set
209 * to the slowest value that, in our tests, proved to be effective in
210 * removing false positives, while not causing true positives to miss
213 * As can be deduced from the low time limit below, queue merging, if
214 * successful, happens at the very beginning of the I/O of the involved
215 * cooperating processes, as a consequence of the arrival of the very
216 * first requests from each cooperator. After that, there is very
217 * little chance to find cooperators.
219 static const unsigned long bfq_merge_time_limit = HZ/10;
221 static struct kmem_cache *bfq_pool;
223 /* Below this threshold (in ns), we consider thinktime immediate. */
224 #define BFQ_MIN_TT (2 * NSEC_PER_MSEC)
226 /* hw_tag detection: parallel requests threshold and min samples needed. */
227 #define BFQ_HW_QUEUE_THRESHOLD 3
228 #define BFQ_HW_QUEUE_SAMPLES 32
230 #define BFQQ_SEEK_THR (sector_t)(8 * 100)
231 #define BFQQ_SECT_THR_NONROT (sector_t)(2 * 32)
232 #define BFQ_RQ_SEEKY(bfqd, last_pos, rq) \
233 (get_sdist(last_pos, rq) > \
235 (!blk_queue_nonrot(bfqd->queue) || \
236 blk_rq_sectors(rq) < BFQQ_SECT_THR_NONROT))
237 #define BFQQ_CLOSE_THR (sector_t)(8 * 1024)
238 #define BFQQ_SEEKY(bfqq) (hweight32(bfqq->seek_history) > 19)
240 * Sync random I/O is likely to be confused with soft real-time I/O,
241 * because it is characterized by limited throughput and apparently
242 * isochronous arrival pattern. To avoid false positives, queues
243 * containing only random (seeky) I/O are prevented from being tagged
246 #define BFQQ_TOTALLY_SEEKY(bfqq) (bfqq->seek_history == -1)
248 /* Min number of samples required to perform peak-rate update */
249 #define BFQ_RATE_MIN_SAMPLES 32
250 /* Min observation time interval required to perform a peak-rate update (ns) */
251 #define BFQ_RATE_MIN_INTERVAL (300*NSEC_PER_MSEC)
252 /* Target observation time interval for a peak-rate update (ns) */
253 #define BFQ_RATE_REF_INTERVAL NSEC_PER_SEC
256 * Shift used for peak-rate fixed precision calculations.
258 * - the current shift: 16 positions
259 * - the current type used to store rate: u32
260 * - the current unit of measure for rate: [sectors/usec], or, more precisely,
261 * [(sectors/usec) / 2^BFQ_RATE_SHIFT] to take into account the shift,
262 * the range of rates that can be stored is
263 * [1 / 2^BFQ_RATE_SHIFT, 2^(32 - BFQ_RATE_SHIFT)] sectors/usec =
264 * [1 / 2^16, 2^16] sectors/usec = [15e-6, 65536] sectors/usec =
265 * [15, 65G] sectors/sec
266 * Which, assuming a sector size of 512B, corresponds to a range of
269 #define BFQ_RATE_SHIFT 16
272 * When configured for computing the duration of the weight-raising
273 * for interactive queues automatically (see the comments at the
274 * beginning of this file), BFQ does it using the following formula:
275 * duration = (ref_rate / r) * ref_wr_duration,
276 * where r is the peak rate of the device, and ref_rate and
277 * ref_wr_duration are two reference parameters. In particular,
278 * ref_rate is the peak rate of the reference storage device (see
279 * below), and ref_wr_duration is about the maximum time needed, with
280 * BFQ and while reading two files in parallel, to load typical large
281 * applications on the reference device (see the comments on
282 * max_service_from_wr below, for more details on how ref_wr_duration
283 * is obtained). In practice, the slower/faster the device at hand
284 * is, the more/less it takes to load applications with respect to the
285 * reference device. Accordingly, the longer/shorter BFQ grants
286 * weight raising to interactive applications.
288 * BFQ uses two different reference pairs (ref_rate, ref_wr_duration),
289 * depending on whether the device is rotational or non-rotational.
291 * In the following definitions, ref_rate[0] and ref_wr_duration[0]
292 * are the reference values for a rotational device, whereas
293 * ref_rate[1] and ref_wr_duration[1] are the reference values for a
294 * non-rotational device. The reference rates are not the actual peak
295 * rates of the devices used as a reference, but slightly lower
296 * values. The reason for using slightly lower values is that the
297 * peak-rate estimator tends to yield slightly lower values than the
298 * actual peak rate (it can yield the actual peak rate only if there
299 * is only one process doing I/O, and the process does sequential
302 * The reference peak rates are measured in sectors/usec, left-shifted
305 static int ref_rate[2] = {14000, 33000};
307 * To improve readability, a conversion function is used to initialize
308 * the following array, which entails that the array can be
309 * initialized only in a function.
311 static int ref_wr_duration[2];
314 * BFQ uses the above-detailed, time-based weight-raising mechanism to
315 * privilege interactive tasks. This mechanism is vulnerable to the
316 * following false positives: I/O-bound applications that will go on
317 * doing I/O for much longer than the duration of weight
318 * raising. These applications have basically no benefit from being
319 * weight-raised at the beginning of their I/O. On the opposite end,
320 * while being weight-raised, these applications
321 * a) unjustly steal throughput to applications that may actually need
323 * b) make BFQ uselessly perform device idling; device idling results
324 * in loss of device throughput with most flash-based storage, and may
325 * increase latencies when used purposelessly.
327 * BFQ tries to reduce these problems, by adopting the following
328 * countermeasure. To introduce this countermeasure, we need first to
329 * finish explaining how the duration of weight-raising for
330 * interactive tasks is computed.
332 * For a bfq_queue deemed as interactive, the duration of weight
333 * raising is dynamically adjusted, as a function of the estimated
334 * peak rate of the device, so as to be equal to the time needed to
335 * execute the 'largest' interactive task we benchmarked so far. By
336 * largest task, we mean the task for which each involved process has
337 * to do more I/O than for any of the other tasks we benchmarked. This
338 * reference interactive task is the start-up of LibreOffice Writer,
339 * and in this task each process/bfq_queue needs to have at most ~110K
340 * sectors transferred.
342 * This last piece of information enables BFQ to reduce the actual
343 * duration of weight-raising for at least one class of I/O-bound
344 * applications: those doing sequential or quasi-sequential I/O. An
345 * example is file copy. In fact, once started, the main I/O-bound
346 * processes of these applications usually consume the above 110K
347 * sectors in much less time than the processes of an application that
348 * is starting, because these I/O-bound processes will greedily devote
349 * almost all their CPU cycles only to their target,
350 * throughput-friendly I/O operations. This is even more true if BFQ
351 * happens to be underestimating the device peak rate, and thus
352 * overestimating the duration of weight raising. But, according to
353 * our measurements, once transferred 110K sectors, these processes
354 * have no right to be weight-raised any longer.
356 * Basing on the last consideration, BFQ ends weight-raising for a
357 * bfq_queue if the latter happens to have received an amount of
358 * service at least equal to the following constant. The constant is
359 * set to slightly more than 110K, to have a minimum safety margin.
361 * This early ending of weight-raising reduces the amount of time
362 * during which interactive false positives cause the two problems
363 * described at the beginning of these comments.
365 static const unsigned long max_service_from_wr = 120000;
367 #define RQ_BIC(rq) icq_to_bic((rq)->elv.priv[0])
368 #define RQ_BFQQ(rq) ((rq)->elv.priv[1])
370 struct bfq_queue *bic_to_bfqq(struct bfq_io_cq *bic, bool is_sync)
372 return bic->bfqq[is_sync];
375 static void bfq_put_stable_ref(struct bfq_queue *bfqq);
377 void bic_set_bfqq(struct bfq_io_cq *bic, struct bfq_queue *bfqq, bool is_sync)
380 * If bfqq != NULL, then a non-stable queue merge between
381 * bic->bfqq and bfqq is happening here. This causes troubles
382 * in the following case: bic->bfqq has also been scheduled
383 * for a possible stable merge with bic->stable_merge_bfqq,
384 * and bic->stable_merge_bfqq == bfqq happens to
385 * hold. Troubles occur because bfqq may then undergo a split,
386 * thereby becoming eligible for a stable merge. Yet, if
387 * bic->stable_merge_bfqq points exactly to bfqq, then bfqq
388 * would be stably merged with itself. To avoid this anomaly,
389 * we cancel the stable merge if
390 * bic->stable_merge_bfqq == bfqq.
392 bic->bfqq[is_sync] = bfqq;
394 if (bfqq && bic->stable_merge_bfqq == bfqq) {
396 * Actually, these same instructions are executed also
397 * in bfq_setup_cooperator, in case of abort or actual
398 * execution of a stable merge. We could avoid
399 * repeating these instructions there too, but if we
400 * did so, we would nest even more complexity in this
403 bfq_put_stable_ref(bic->stable_merge_bfqq);
405 bic->stable_merge_bfqq = NULL;
409 struct bfq_data *bic_to_bfqd(struct bfq_io_cq *bic)
411 return bic->icq.q->elevator->elevator_data;
415 * icq_to_bic - convert iocontext queue structure to bfq_io_cq.
416 * @icq: the iocontext queue.
418 static struct bfq_io_cq *icq_to_bic(struct io_cq *icq)
420 /* bic->icq is the first member, %NULL will convert to %NULL */
421 return container_of(icq, struct bfq_io_cq, icq);
425 * bfq_bic_lookup - search into @ioc a bic associated to @bfqd.
426 * @bfqd: the lookup key.
427 * @ioc: the io_context of the process doing I/O.
428 * @q: the request queue.
430 static struct bfq_io_cq *bfq_bic_lookup(struct bfq_data *bfqd,
431 struct io_context *ioc,
432 struct request_queue *q)
436 struct bfq_io_cq *icq;
438 spin_lock_irqsave(&q->queue_lock, flags);
439 icq = icq_to_bic(ioc_lookup_icq(ioc, q));
440 spin_unlock_irqrestore(&q->queue_lock, flags);
449 * Scheduler run of queue, if there are requests pending and no one in the
450 * driver that will restart queueing.
452 void bfq_schedule_dispatch(struct bfq_data *bfqd)
454 if (bfqd->queued != 0) {
455 bfq_log(bfqd, "schedule dispatch");
456 blk_mq_run_hw_queues(bfqd->queue, true);
460 #define bfq_class_idle(bfqq) ((bfqq)->ioprio_class == IOPRIO_CLASS_IDLE)
462 #define bfq_sample_valid(samples) ((samples) > 80)
465 * Lifted from AS - choose which of rq1 and rq2 that is best served now.
466 * We choose the request that is closer to the head right now. Distance
467 * behind the head is penalized and only allowed to a certain extent.
469 static struct request *bfq_choose_req(struct bfq_data *bfqd,
474 sector_t s1, s2, d1 = 0, d2 = 0;
475 unsigned long back_max;
476 #define BFQ_RQ1_WRAP 0x01 /* request 1 wraps */
477 #define BFQ_RQ2_WRAP 0x02 /* request 2 wraps */
478 unsigned int wrap = 0; /* bit mask: requests behind the disk head? */
480 if (!rq1 || rq1 == rq2)
485 if (rq_is_sync(rq1) && !rq_is_sync(rq2))
487 else if (rq_is_sync(rq2) && !rq_is_sync(rq1))
489 if ((rq1->cmd_flags & REQ_META) && !(rq2->cmd_flags & REQ_META))
491 else if ((rq2->cmd_flags & REQ_META) && !(rq1->cmd_flags & REQ_META))
494 s1 = blk_rq_pos(rq1);
495 s2 = blk_rq_pos(rq2);
498 * By definition, 1KiB is 2 sectors.
500 back_max = bfqd->bfq_back_max * 2;
503 * Strict one way elevator _except_ in the case where we allow
504 * short backward seeks which are biased as twice the cost of a
505 * similar forward seek.
509 else if (s1 + back_max >= last)
510 d1 = (last - s1) * bfqd->bfq_back_penalty;
512 wrap |= BFQ_RQ1_WRAP;
516 else if (s2 + back_max >= last)
517 d2 = (last - s2) * bfqd->bfq_back_penalty;
519 wrap |= BFQ_RQ2_WRAP;
521 /* Found required data */
524 * By doing switch() on the bit mask "wrap" we avoid having to
525 * check two variables for all permutations: --> faster!
528 case 0: /* common case for CFQ: rq1 and rq2 not wrapped */
543 case BFQ_RQ1_WRAP|BFQ_RQ2_WRAP: /* both rqs wrapped */
546 * Since both rqs are wrapped,
547 * start with the one that's further behind head
548 * (--> only *one* back seek required),
549 * since back seek takes more time than forward.
559 * Async I/O can easily starve sync I/O (both sync reads and sync
560 * writes), by consuming all tags. Similarly, storms of sync writes,
561 * such as those that sync(2) may trigger, can starve sync reads.
562 * Limit depths of async I/O and sync writes so as to counter both
565 static void bfq_limit_depth(unsigned int op, struct blk_mq_alloc_data *data)
567 struct bfq_data *bfqd = data->q->elevator->elevator_data;
569 if (op_is_sync(op) && !op_is_write(op))
572 data->shallow_depth =
573 bfqd->word_depths[!!bfqd->wr_busy_queues][op_is_sync(op)];
575 bfq_log(bfqd, "[%s] wr_busy %d sync %d depth %u",
576 __func__, bfqd->wr_busy_queues, op_is_sync(op),
577 data->shallow_depth);
580 static struct bfq_queue *
581 bfq_rq_pos_tree_lookup(struct bfq_data *bfqd, struct rb_root *root,
582 sector_t sector, struct rb_node **ret_parent,
583 struct rb_node ***rb_link)
585 struct rb_node **p, *parent;
586 struct bfq_queue *bfqq = NULL;
594 bfqq = rb_entry(parent, struct bfq_queue, pos_node);
597 * Sort strictly based on sector. Smallest to the left,
598 * largest to the right.
600 if (sector > blk_rq_pos(bfqq->next_rq))
602 else if (sector < blk_rq_pos(bfqq->next_rq))
610 *ret_parent = parent;
614 bfq_log(bfqd, "rq_pos_tree_lookup %llu: returning %d",
615 (unsigned long long)sector,
616 bfqq ? bfqq->pid : 0);
621 static bool bfq_too_late_for_merging(struct bfq_queue *bfqq)
623 return bfqq->service_from_backlogged > 0 &&
624 time_is_before_jiffies(bfqq->first_IO_time +
625 bfq_merge_time_limit);
629 * The following function is not marked as __cold because it is
630 * actually cold, but for the same performance goal described in the
631 * comments on the likely() at the beginning of
632 * bfq_setup_cooperator(). Unexpectedly, to reach an even lower
633 * execution time for the case where this function is not invoked, we
634 * had to add an unlikely() in each involved if().
637 bfq_pos_tree_add_move(struct bfq_data *bfqd, struct bfq_queue *bfqq)
639 struct rb_node **p, *parent;
640 struct bfq_queue *__bfqq;
642 if (bfqq->pos_root) {
643 rb_erase(&bfqq->pos_node, bfqq->pos_root);
644 bfqq->pos_root = NULL;
647 /* oom_bfqq does not participate in queue merging */
648 if (bfqq == &bfqd->oom_bfqq)
652 * bfqq cannot be merged any longer (see comments in
653 * bfq_setup_cooperator): no point in adding bfqq into the
656 if (bfq_too_late_for_merging(bfqq))
659 if (bfq_class_idle(bfqq))
664 bfqq->pos_root = &bfq_bfqq_to_bfqg(bfqq)->rq_pos_tree;
665 __bfqq = bfq_rq_pos_tree_lookup(bfqd, bfqq->pos_root,
666 blk_rq_pos(bfqq->next_rq), &parent, &p);
668 rb_link_node(&bfqq->pos_node, parent, p);
669 rb_insert_color(&bfqq->pos_node, bfqq->pos_root);
671 bfqq->pos_root = NULL;
675 * The following function returns false either if every active queue
676 * must receive the same share of the throughput (symmetric scenario),
677 * or, as a special case, if bfqq must receive a share of the
678 * throughput lower than or equal to the share that every other active
679 * queue must receive. If bfqq does sync I/O, then these are the only
680 * two cases where bfqq happens to be guaranteed its share of the
681 * throughput even if I/O dispatching is not plugged when bfqq remains
682 * temporarily empty (for more details, see the comments in the
683 * function bfq_better_to_idle()). For this reason, the return value
684 * of this function is used to check whether I/O-dispatch plugging can
687 * The above first case (symmetric scenario) occurs when:
688 * 1) all active queues have the same weight,
689 * 2) all active queues belong to the same I/O-priority class,
690 * 3) all active groups at the same level in the groups tree have the same
692 * 4) all active groups at the same level in the groups tree have the same
693 * number of children.
695 * Unfortunately, keeping the necessary state for evaluating exactly
696 * the last two symmetry sub-conditions above would be quite complex
697 * and time consuming. Therefore this function evaluates, instead,
698 * only the following stronger three sub-conditions, for which it is
699 * much easier to maintain the needed state:
700 * 1) all active queues have the same weight,
701 * 2) all active queues belong to the same I/O-priority class,
702 * 3) there are no active groups.
703 * In particular, the last condition is always true if hierarchical
704 * support or the cgroups interface are not enabled, thus no state
705 * needs to be maintained in this case.
707 static bool bfq_asymmetric_scenario(struct bfq_data *bfqd,
708 struct bfq_queue *bfqq)
710 bool smallest_weight = bfqq &&
711 bfqq->weight_counter &&
712 bfqq->weight_counter ==
714 rb_first_cached(&bfqd->queue_weights_tree),
715 struct bfq_weight_counter,
719 * For queue weights to differ, queue_weights_tree must contain
720 * at least two nodes.
722 bool varied_queue_weights = !smallest_weight &&
723 !RB_EMPTY_ROOT(&bfqd->queue_weights_tree.rb_root) &&
724 (bfqd->queue_weights_tree.rb_root.rb_node->rb_left ||
725 bfqd->queue_weights_tree.rb_root.rb_node->rb_right);
727 bool multiple_classes_busy =
728 (bfqd->busy_queues[0] && bfqd->busy_queues[1]) ||
729 (bfqd->busy_queues[0] && bfqd->busy_queues[2]) ||
730 (bfqd->busy_queues[1] && bfqd->busy_queues[2]);
732 return varied_queue_weights || multiple_classes_busy
733 #ifdef CONFIG_BFQ_GROUP_IOSCHED
734 || bfqd->num_groups_with_pending_reqs > 0
740 * If the weight-counter tree passed as input contains no counter for
741 * the weight of the input queue, then add that counter; otherwise just
742 * increment the existing counter.
744 * Note that weight-counter trees contain few nodes in mostly symmetric
745 * scenarios. For example, if all queues have the same weight, then the
746 * weight-counter tree for the queues may contain at most one node.
747 * This holds even if low_latency is on, because weight-raised queues
748 * are not inserted in the tree.
749 * In most scenarios, the rate at which nodes are created/destroyed
752 void bfq_weights_tree_add(struct bfq_data *bfqd, struct bfq_queue *bfqq,
753 struct rb_root_cached *root)
755 struct bfq_entity *entity = &bfqq->entity;
756 struct rb_node **new = &(root->rb_root.rb_node), *parent = NULL;
757 bool leftmost = true;
760 * Do not insert if the queue is already associated with a
761 * counter, which happens if:
762 * 1) a request arrival has caused the queue to become both
763 * non-weight-raised, and hence change its weight, and
764 * backlogged; in this respect, each of the two events
765 * causes an invocation of this function,
766 * 2) this is the invocation of this function caused by the
767 * second event. This second invocation is actually useless,
768 * and we handle this fact by exiting immediately. More
769 * efficient or clearer solutions might possibly be adopted.
771 if (bfqq->weight_counter)
775 struct bfq_weight_counter *__counter = container_of(*new,
776 struct bfq_weight_counter,
780 if (entity->weight == __counter->weight) {
781 bfqq->weight_counter = __counter;
784 if (entity->weight < __counter->weight)
785 new = &((*new)->rb_left);
787 new = &((*new)->rb_right);
792 bfqq->weight_counter = kzalloc(sizeof(struct bfq_weight_counter),
796 * In the unlucky event of an allocation failure, we just
797 * exit. This will cause the weight of queue to not be
798 * considered in bfq_asymmetric_scenario, which, in its turn,
799 * causes the scenario to be deemed wrongly symmetric in case
800 * bfqq's weight would have been the only weight making the
801 * scenario asymmetric. On the bright side, no unbalance will
802 * however occur when bfqq becomes inactive again (the
803 * invocation of this function is triggered by an activation
804 * of queue). In fact, bfq_weights_tree_remove does nothing
805 * if !bfqq->weight_counter.
807 if (unlikely(!bfqq->weight_counter))
810 bfqq->weight_counter->weight = entity->weight;
811 rb_link_node(&bfqq->weight_counter->weights_node, parent, new);
812 rb_insert_color_cached(&bfqq->weight_counter->weights_node, root,
816 bfqq->weight_counter->num_active++;
821 * Decrement the weight counter associated with the queue, and, if the
822 * counter reaches 0, remove the counter from the tree.
823 * See the comments to the function bfq_weights_tree_add() for considerations
826 void __bfq_weights_tree_remove(struct bfq_data *bfqd,
827 struct bfq_queue *bfqq,
828 struct rb_root_cached *root)
830 if (!bfqq->weight_counter)
833 bfqq->weight_counter->num_active--;
834 if (bfqq->weight_counter->num_active > 0)
835 goto reset_entity_pointer;
837 rb_erase_cached(&bfqq->weight_counter->weights_node, root);
838 kfree(bfqq->weight_counter);
840 reset_entity_pointer:
841 bfqq->weight_counter = NULL;
846 * Invoke __bfq_weights_tree_remove on bfqq and decrement the number
847 * of active groups for each queue's inactive parent entity.
849 void bfq_weights_tree_remove(struct bfq_data *bfqd,
850 struct bfq_queue *bfqq)
852 struct bfq_entity *entity = bfqq->entity.parent;
854 for_each_entity(entity) {
855 struct bfq_sched_data *sd = entity->my_sched_data;
857 if (sd->next_in_service || sd->in_service_entity) {
859 * entity is still active, because either
860 * next_in_service or in_service_entity is not
861 * NULL (see the comments on the definition of
862 * next_in_service for details on why
863 * in_service_entity must be checked too).
865 * As a consequence, its parent entities are
866 * active as well, and thus this loop must
873 * The decrement of num_groups_with_pending_reqs is
874 * not performed immediately upon the deactivation of
875 * entity, but it is delayed to when it also happens
876 * that the first leaf descendant bfqq of entity gets
877 * all its pending requests completed. The following
878 * instructions perform this delayed decrement, if
879 * needed. See the comments on
880 * num_groups_with_pending_reqs for details.
882 if (entity->in_groups_with_pending_reqs) {
883 entity->in_groups_with_pending_reqs = false;
884 bfqd->num_groups_with_pending_reqs--;
889 * Next function is invoked last, because it causes bfqq to be
890 * freed if the following holds: bfqq is not in service and
891 * has no dispatched request. DO NOT use bfqq after the next
892 * function invocation.
894 __bfq_weights_tree_remove(bfqd, bfqq,
895 &bfqd->queue_weights_tree);
899 * Return expired entry, or NULL to just start from scratch in rbtree.
901 static struct request *bfq_check_fifo(struct bfq_queue *bfqq,
902 struct request *last)
906 if (bfq_bfqq_fifo_expire(bfqq))
909 bfq_mark_bfqq_fifo_expire(bfqq);
911 rq = rq_entry_fifo(bfqq->fifo.next);
913 if (rq == last || ktime_get_ns() < rq->fifo_time)
916 bfq_log_bfqq(bfqq->bfqd, bfqq, "check_fifo: returned %p", rq);
920 static struct request *bfq_find_next_rq(struct bfq_data *bfqd,
921 struct bfq_queue *bfqq,
922 struct request *last)
924 struct rb_node *rbnext = rb_next(&last->rb_node);
925 struct rb_node *rbprev = rb_prev(&last->rb_node);
926 struct request *next, *prev = NULL;
928 /* Follow expired path, else get first next available. */
929 next = bfq_check_fifo(bfqq, last);
934 prev = rb_entry_rq(rbprev);
937 next = rb_entry_rq(rbnext);
939 rbnext = rb_first(&bfqq->sort_list);
940 if (rbnext && rbnext != &last->rb_node)
941 next = rb_entry_rq(rbnext);
944 return bfq_choose_req(bfqd, next, prev, blk_rq_pos(last));
947 /* see the definition of bfq_async_charge_factor for details */
948 static unsigned long bfq_serv_to_charge(struct request *rq,
949 struct bfq_queue *bfqq)
951 if (bfq_bfqq_sync(bfqq) || bfqq->wr_coeff > 1 ||
952 bfq_asymmetric_scenario(bfqq->bfqd, bfqq))
953 return blk_rq_sectors(rq);
955 return blk_rq_sectors(rq) * bfq_async_charge_factor;
959 * bfq_updated_next_req - update the queue after a new next_rq selection.
960 * @bfqd: the device data the queue belongs to.
961 * @bfqq: the queue to update.
963 * If the first request of a queue changes we make sure that the queue
964 * has enough budget to serve at least its first request (if the
965 * request has grown). We do this because if the queue has not enough
966 * budget for its first request, it has to go through two dispatch
967 * rounds to actually get it dispatched.
969 static void bfq_updated_next_req(struct bfq_data *bfqd,
970 struct bfq_queue *bfqq)
972 struct bfq_entity *entity = &bfqq->entity;
973 struct request *next_rq = bfqq->next_rq;
974 unsigned long new_budget;
979 if (bfqq == bfqd->in_service_queue)
981 * In order not to break guarantees, budgets cannot be
982 * changed after an entity has been selected.
986 new_budget = max_t(unsigned long,
987 max_t(unsigned long, bfqq->max_budget,
988 bfq_serv_to_charge(next_rq, bfqq)),
990 if (entity->budget != new_budget) {
991 entity->budget = new_budget;
992 bfq_log_bfqq(bfqd, bfqq, "updated next rq: new budget %lu",
994 bfq_requeue_bfqq(bfqd, bfqq, false);
998 static unsigned int bfq_wr_duration(struct bfq_data *bfqd)
1002 if (bfqd->bfq_wr_max_time > 0)
1003 return bfqd->bfq_wr_max_time;
1005 dur = bfqd->rate_dur_prod;
1006 do_div(dur, bfqd->peak_rate);
1009 * Limit duration between 3 and 25 seconds. The upper limit
1010 * has been conservatively set after the following worst case:
1011 * on a QEMU/KVM virtual machine
1012 * - running in a slow PC
1013 * - with a virtual disk stacked on a slow low-end 5400rpm HDD
1014 * - serving a heavy I/O workload, such as the sequential reading
1016 * mplayer took 23 seconds to start, if constantly weight-raised.
1018 * As for higher values than that accommodating the above bad
1019 * scenario, tests show that higher values would often yield
1020 * the opposite of the desired result, i.e., would worsen
1021 * responsiveness by allowing non-interactive applications to
1022 * preserve weight raising for too long.
1024 * On the other end, lower values than 3 seconds make it
1025 * difficult for most interactive tasks to complete their jobs
1026 * before weight-raising finishes.
1028 return clamp_val(dur, msecs_to_jiffies(3000), msecs_to_jiffies(25000));
1031 /* switch back from soft real-time to interactive weight raising */
1032 static void switch_back_to_interactive_wr(struct bfq_queue *bfqq,
1033 struct bfq_data *bfqd)
1035 bfqq->wr_coeff = bfqd->bfq_wr_coeff;
1036 bfqq->wr_cur_max_time = bfq_wr_duration(bfqd);
1037 bfqq->last_wr_start_finish = bfqq->wr_start_at_switch_to_srt;
1041 bfq_bfqq_resume_state(struct bfq_queue *bfqq, struct bfq_data *bfqd,
1042 struct bfq_io_cq *bic, bool bfq_already_existing)
1044 unsigned int old_wr_coeff = 1;
1045 bool busy = bfq_already_existing && bfq_bfqq_busy(bfqq);
1047 if (bic->saved_has_short_ttime)
1048 bfq_mark_bfqq_has_short_ttime(bfqq);
1050 bfq_clear_bfqq_has_short_ttime(bfqq);
1052 if (bic->saved_IO_bound)
1053 bfq_mark_bfqq_IO_bound(bfqq);
1055 bfq_clear_bfqq_IO_bound(bfqq);
1057 bfqq->last_serv_time_ns = bic->saved_last_serv_time_ns;
1058 bfqq->inject_limit = bic->saved_inject_limit;
1059 bfqq->decrease_time_jif = bic->saved_decrease_time_jif;
1061 bfqq->entity.new_weight = bic->saved_weight;
1062 bfqq->ttime = bic->saved_ttime;
1063 bfqq->io_start_time = bic->saved_io_start_time;
1064 bfqq->tot_idle_time = bic->saved_tot_idle_time;
1066 * Restore weight coefficient only if low_latency is on
1068 if (bfqd->low_latency) {
1069 old_wr_coeff = bfqq->wr_coeff;
1070 bfqq->wr_coeff = bic->saved_wr_coeff;
1072 bfqq->service_from_wr = bic->saved_service_from_wr;
1073 bfqq->wr_start_at_switch_to_srt = bic->saved_wr_start_at_switch_to_srt;
1074 bfqq->last_wr_start_finish = bic->saved_last_wr_start_finish;
1075 bfqq->wr_cur_max_time = bic->saved_wr_cur_max_time;
1077 if (bfqq->wr_coeff > 1 && (bfq_bfqq_in_large_burst(bfqq) ||
1078 time_is_before_jiffies(bfqq->last_wr_start_finish +
1079 bfqq->wr_cur_max_time))) {
1080 if (bfqq->wr_cur_max_time == bfqd->bfq_wr_rt_max_time &&
1081 !bfq_bfqq_in_large_burst(bfqq) &&
1082 time_is_after_eq_jiffies(bfqq->wr_start_at_switch_to_srt +
1083 bfq_wr_duration(bfqd))) {
1084 switch_back_to_interactive_wr(bfqq, bfqd);
1087 bfq_log_bfqq(bfqq->bfqd, bfqq,
1088 "resume state: switching off wr");
1092 /* make sure weight will be updated, however we got here */
1093 bfqq->entity.prio_changed = 1;
1098 if (old_wr_coeff == 1 && bfqq->wr_coeff > 1)
1099 bfqd->wr_busy_queues++;
1100 else if (old_wr_coeff > 1 && bfqq->wr_coeff == 1)
1101 bfqd->wr_busy_queues--;
1104 static int bfqq_process_refs(struct bfq_queue *bfqq)
1106 return bfqq->ref - bfqq->allocated - bfqq->entity.on_st_or_in_serv -
1107 (bfqq->weight_counter != NULL) - bfqq->stable_ref;
1110 /* Empty burst list and add just bfqq (see comments on bfq_handle_burst) */
1111 static void bfq_reset_burst_list(struct bfq_data *bfqd, struct bfq_queue *bfqq)
1113 struct bfq_queue *item;
1114 struct hlist_node *n;
1116 hlist_for_each_entry_safe(item, n, &bfqd->burst_list, burst_list_node)
1117 hlist_del_init(&item->burst_list_node);
1120 * Start the creation of a new burst list only if there is no
1121 * active queue. See comments on the conditional invocation of
1122 * bfq_handle_burst().
1124 if (bfq_tot_busy_queues(bfqd) == 0) {
1125 hlist_add_head(&bfqq->burst_list_node, &bfqd->burst_list);
1126 bfqd->burst_size = 1;
1128 bfqd->burst_size = 0;
1130 bfqd->burst_parent_entity = bfqq->entity.parent;
1133 /* Add bfqq to the list of queues in current burst (see bfq_handle_burst) */
1134 static void bfq_add_to_burst(struct bfq_data *bfqd, struct bfq_queue *bfqq)
1136 /* Increment burst size to take into account also bfqq */
1139 if (bfqd->burst_size == bfqd->bfq_large_burst_thresh) {
1140 struct bfq_queue *pos, *bfqq_item;
1141 struct hlist_node *n;
1144 * Enough queues have been activated shortly after each
1145 * other to consider this burst as large.
1147 bfqd->large_burst = true;
1150 * We can now mark all queues in the burst list as
1151 * belonging to a large burst.
1153 hlist_for_each_entry(bfqq_item, &bfqd->burst_list,
1155 bfq_mark_bfqq_in_large_burst(bfqq_item);
1156 bfq_mark_bfqq_in_large_burst(bfqq);
1159 * From now on, and until the current burst finishes, any
1160 * new queue being activated shortly after the last queue
1161 * was inserted in the burst can be immediately marked as
1162 * belonging to a large burst. So the burst list is not
1163 * needed any more. Remove it.
1165 hlist_for_each_entry_safe(pos, n, &bfqd->burst_list,
1167 hlist_del_init(&pos->burst_list_node);
1169 * Burst not yet large: add bfqq to the burst list. Do
1170 * not increment the ref counter for bfqq, because bfqq
1171 * is removed from the burst list before freeing bfqq
1174 hlist_add_head(&bfqq->burst_list_node, &bfqd->burst_list);
1178 * If many queues belonging to the same group happen to be created
1179 * shortly after each other, then the processes associated with these
1180 * queues have typically a common goal. In particular, bursts of queue
1181 * creations are usually caused by services or applications that spawn
1182 * many parallel threads/processes. Examples are systemd during boot,
1183 * or git grep. To help these processes get their job done as soon as
1184 * possible, it is usually better to not grant either weight-raising
1185 * or device idling to their queues, unless these queues must be
1186 * protected from the I/O flowing through other active queues.
1188 * In this comment we describe, firstly, the reasons why this fact
1189 * holds, and, secondly, the next function, which implements the main
1190 * steps needed to properly mark these queues so that they can then be
1191 * treated in a different way.
1193 * The above services or applications benefit mostly from a high
1194 * throughput: the quicker the requests of the activated queues are
1195 * cumulatively served, the sooner the target job of these queues gets
1196 * completed. As a consequence, weight-raising any of these queues,
1197 * which also implies idling the device for it, is almost always
1198 * counterproductive, unless there are other active queues to isolate
1199 * these new queues from. If there no other active queues, then
1200 * weight-raising these new queues just lowers throughput in most
1203 * On the other hand, a burst of queue creations may be caused also by
1204 * the start of an application that does not consist of a lot of
1205 * parallel I/O-bound threads. In fact, with a complex application,
1206 * several short processes may need to be executed to start-up the
1207 * application. In this respect, to start an application as quickly as
1208 * possible, the best thing to do is in any case to privilege the I/O
1209 * related to the application with respect to all other
1210 * I/O. Therefore, the best strategy to start as quickly as possible
1211 * an application that causes a burst of queue creations is to
1212 * weight-raise all the queues created during the burst. This is the
1213 * exact opposite of the best strategy for the other type of bursts.
1215 * In the end, to take the best action for each of the two cases, the
1216 * two types of bursts need to be distinguished. Fortunately, this
1217 * seems relatively easy, by looking at the sizes of the bursts. In
1218 * particular, we found a threshold such that only bursts with a
1219 * larger size than that threshold are apparently caused by
1220 * services or commands such as systemd or git grep. For brevity,
1221 * hereafter we call just 'large' these bursts. BFQ *does not*
1222 * weight-raise queues whose creation occurs in a large burst. In
1223 * addition, for each of these queues BFQ performs or does not perform
1224 * idling depending on which choice boosts the throughput more. The
1225 * exact choice depends on the device and request pattern at
1228 * Unfortunately, false positives may occur while an interactive task
1229 * is starting (e.g., an application is being started). The
1230 * consequence is that the queues associated with the task do not
1231 * enjoy weight raising as expected. Fortunately these false positives
1232 * are very rare. They typically occur if some service happens to
1233 * start doing I/O exactly when the interactive task starts.
1235 * Turning back to the next function, it is invoked only if there are
1236 * no active queues (apart from active queues that would belong to the
1237 * same, possible burst bfqq would belong to), and it implements all
1238 * the steps needed to detect the occurrence of a large burst and to
1239 * properly mark all the queues belonging to it (so that they can then
1240 * be treated in a different way). This goal is achieved by
1241 * maintaining a "burst list" that holds, temporarily, the queues that
1242 * belong to the burst in progress. The list is then used to mark
1243 * these queues as belonging to a large burst if the burst does become
1244 * large. The main steps are the following.
1246 * . when the very first queue is created, the queue is inserted into the
1247 * list (as it could be the first queue in a possible burst)
1249 * . if the current burst has not yet become large, and a queue Q that does
1250 * not yet belong to the burst is activated shortly after the last time
1251 * at which a new queue entered the burst list, then the function appends
1252 * Q to the burst list
1254 * . if, as a consequence of the previous step, the burst size reaches
1255 * the large-burst threshold, then
1257 * . all the queues in the burst list are marked as belonging to a
1260 * . the burst list is deleted; in fact, the burst list already served
1261 * its purpose (keeping temporarily track of the queues in a burst,
1262 * so as to be able to mark them as belonging to a large burst in the
1263 * previous sub-step), and now is not needed any more
1265 * . the device enters a large-burst mode
1267 * . if a queue Q that does not belong to the burst is created while
1268 * the device is in large-burst mode and shortly after the last time
1269 * at which a queue either entered the burst list or was marked as
1270 * belonging to the current large burst, then Q is immediately marked
1271 * as belonging to a large burst.
1273 * . if a queue Q that does not belong to the burst is created a while
1274 * later, i.e., not shortly after, than the last time at which a queue
1275 * either entered the burst list or was marked as belonging to the
1276 * current large burst, then the current burst is deemed as finished and:
1278 * . the large-burst mode is reset if set
1280 * . the burst list is emptied
1282 * . Q is inserted in the burst list, as Q may be the first queue
1283 * in a possible new burst (then the burst list contains just Q
1286 static void bfq_handle_burst(struct bfq_data *bfqd, struct bfq_queue *bfqq)
1289 * If bfqq is already in the burst list or is part of a large
1290 * burst, or finally has just been split, then there is
1291 * nothing else to do.
1293 if (!hlist_unhashed(&bfqq->burst_list_node) ||
1294 bfq_bfqq_in_large_burst(bfqq) ||
1295 time_is_after_eq_jiffies(bfqq->split_time +
1296 msecs_to_jiffies(10)))
1300 * If bfqq's creation happens late enough, or bfqq belongs to
1301 * a different group than the burst group, then the current
1302 * burst is finished, and related data structures must be
1305 * In this respect, consider the special case where bfqq is
1306 * the very first queue created after BFQ is selected for this
1307 * device. In this case, last_ins_in_burst and
1308 * burst_parent_entity are not yet significant when we get
1309 * here. But it is easy to verify that, whether or not the
1310 * following condition is true, bfqq will end up being
1311 * inserted into the burst list. In particular the list will
1312 * happen to contain only bfqq. And this is exactly what has
1313 * to happen, as bfqq may be the first queue of the first
1316 if (time_is_before_jiffies(bfqd->last_ins_in_burst +
1317 bfqd->bfq_burst_interval) ||
1318 bfqq->entity.parent != bfqd->burst_parent_entity) {
1319 bfqd->large_burst = false;
1320 bfq_reset_burst_list(bfqd, bfqq);
1325 * If we get here, then bfqq is being activated shortly after the
1326 * last queue. So, if the current burst is also large, we can mark
1327 * bfqq as belonging to this large burst immediately.
1329 if (bfqd->large_burst) {
1330 bfq_mark_bfqq_in_large_burst(bfqq);
1335 * If we get here, then a large-burst state has not yet been
1336 * reached, but bfqq is being activated shortly after the last
1337 * queue. Then we add bfqq to the burst.
1339 bfq_add_to_burst(bfqd, bfqq);
1342 * At this point, bfqq either has been added to the current
1343 * burst or has caused the current burst to terminate and a
1344 * possible new burst to start. In particular, in the second
1345 * case, bfqq has become the first queue in the possible new
1346 * burst. In both cases last_ins_in_burst needs to be moved
1349 bfqd->last_ins_in_burst = jiffies;
1352 static int bfq_bfqq_budget_left(struct bfq_queue *bfqq)
1354 struct bfq_entity *entity = &bfqq->entity;
1356 return entity->budget - entity->service;
1360 * If enough samples have been computed, return the current max budget
1361 * stored in bfqd, which is dynamically updated according to the
1362 * estimated disk peak rate; otherwise return the default max budget
1364 static int bfq_max_budget(struct bfq_data *bfqd)
1366 if (bfqd->budgets_assigned < bfq_stats_min_budgets)
1367 return bfq_default_max_budget;
1369 return bfqd->bfq_max_budget;
1373 * Return min budget, which is a fraction of the current or default
1374 * max budget (trying with 1/32)
1376 static int bfq_min_budget(struct bfq_data *bfqd)
1378 if (bfqd->budgets_assigned < bfq_stats_min_budgets)
1379 return bfq_default_max_budget / 32;
1381 return bfqd->bfq_max_budget / 32;
1385 * The next function, invoked after the input queue bfqq switches from
1386 * idle to busy, updates the budget of bfqq. The function also tells
1387 * whether the in-service queue should be expired, by returning
1388 * true. The purpose of expiring the in-service queue is to give bfqq
1389 * the chance to possibly preempt the in-service queue, and the reason
1390 * for preempting the in-service queue is to achieve one of the two
1393 * 1. Guarantee to bfqq its reserved bandwidth even if bfqq has
1394 * expired because it has remained idle. In particular, bfqq may have
1395 * expired for one of the following two reasons:
1397 * - BFQQE_NO_MORE_REQUESTS bfqq did not enjoy any device idling
1398 * and did not make it to issue a new request before its last
1399 * request was served;
1401 * - BFQQE_TOO_IDLE bfqq did enjoy device idling, but did not issue
1402 * a new request before the expiration of the idling-time.
1404 * Even if bfqq has expired for one of the above reasons, the process
1405 * associated with the queue may be however issuing requests greedily,
1406 * and thus be sensitive to the bandwidth it receives (bfqq may have
1407 * remained idle for other reasons: CPU high load, bfqq not enjoying
1408 * idling, I/O throttling somewhere in the path from the process to
1409 * the I/O scheduler, ...). But if, after every expiration for one of
1410 * the above two reasons, bfqq has to wait for the service of at least
1411 * one full budget of another queue before being served again, then
1412 * bfqq is likely to get a much lower bandwidth or resource time than
1413 * its reserved ones. To address this issue, two countermeasures need
1416 * First, the budget and the timestamps of bfqq need to be updated in
1417 * a special way on bfqq reactivation: they need to be updated as if
1418 * bfqq did not remain idle and did not expire. In fact, if they are
1419 * computed as if bfqq expired and remained idle until reactivation,
1420 * then the process associated with bfqq is treated as if, instead of
1421 * being greedy, it stopped issuing requests when bfqq remained idle,
1422 * and restarts issuing requests only on this reactivation. In other
1423 * words, the scheduler does not help the process recover the "service
1424 * hole" between bfqq expiration and reactivation. As a consequence,
1425 * the process receives a lower bandwidth than its reserved one. In
1426 * contrast, to recover this hole, the budget must be updated as if
1427 * bfqq was not expired at all before this reactivation, i.e., it must
1428 * be set to the value of the remaining budget when bfqq was
1429 * expired. Along the same line, timestamps need to be assigned the
1430 * value they had the last time bfqq was selected for service, i.e.,
1431 * before last expiration. Thus timestamps need to be back-shifted
1432 * with respect to their normal computation (see [1] for more details
1433 * on this tricky aspect).
1435 * Secondly, to allow the process to recover the hole, the in-service
1436 * queue must be expired too, to give bfqq the chance to preempt it
1437 * immediately. In fact, if bfqq has to wait for a full budget of the
1438 * in-service queue to be completed, then it may become impossible to
1439 * let the process recover the hole, even if the back-shifted
1440 * timestamps of bfqq are lower than those of the in-service queue. If
1441 * this happens for most or all of the holes, then the process may not
1442 * receive its reserved bandwidth. In this respect, it is worth noting
1443 * that, being the service of outstanding requests unpreemptible, a
1444 * little fraction of the holes may however be unrecoverable, thereby
1445 * causing a little loss of bandwidth.
1447 * The last important point is detecting whether bfqq does need this
1448 * bandwidth recovery. In this respect, the next function deems the
1449 * process associated with bfqq greedy, and thus allows it to recover
1450 * the hole, if: 1) the process is waiting for the arrival of a new
1451 * request (which implies that bfqq expired for one of the above two
1452 * reasons), and 2) such a request has arrived soon. The first
1453 * condition is controlled through the flag non_blocking_wait_rq,
1454 * while the second through the flag arrived_in_time. If both
1455 * conditions hold, then the function computes the budget in the
1456 * above-described special way, and signals that the in-service queue
1457 * should be expired. Timestamp back-shifting is done later in
1458 * __bfq_activate_entity.
1460 * 2. Reduce latency. Even if timestamps are not backshifted to let
1461 * the process associated with bfqq recover a service hole, bfqq may
1462 * however happen to have, after being (re)activated, a lower finish
1463 * timestamp than the in-service queue. That is, the next budget of
1464 * bfqq may have to be completed before the one of the in-service
1465 * queue. If this is the case, then preempting the in-service queue
1466 * allows this goal to be achieved, apart from the unpreemptible,
1467 * outstanding requests mentioned above.
1469 * Unfortunately, regardless of which of the above two goals one wants
1470 * to achieve, service trees need first to be updated to know whether
1471 * the in-service queue must be preempted. To have service trees
1472 * correctly updated, the in-service queue must be expired and
1473 * rescheduled, and bfqq must be scheduled too. This is one of the
1474 * most costly operations (in future versions, the scheduling
1475 * mechanism may be re-designed in such a way to make it possible to
1476 * know whether preemption is needed without needing to update service
1477 * trees). In addition, queue preemptions almost always cause random
1478 * I/O, which may in turn cause loss of throughput. Finally, there may
1479 * even be no in-service queue when the next function is invoked (so,
1480 * no queue to compare timestamps with). Because of these facts, the
1481 * next function adopts the following simple scheme to avoid costly
1482 * operations, too frequent preemptions and too many dependencies on
1483 * the state of the scheduler: it requests the expiration of the
1484 * in-service queue (unconditionally) only for queues that need to
1485 * recover a hole. Then it delegates to other parts of the code the
1486 * responsibility of handling the above case 2.
1488 static bool bfq_bfqq_update_budg_for_activation(struct bfq_data *bfqd,
1489 struct bfq_queue *bfqq,
1490 bool arrived_in_time)
1492 struct bfq_entity *entity = &bfqq->entity;
1495 * In the next compound condition, we check also whether there
1496 * is some budget left, because otherwise there is no point in
1497 * trying to go on serving bfqq with this same budget: bfqq
1498 * would be expired immediately after being selected for
1499 * service. This would only cause useless overhead.
1501 if (bfq_bfqq_non_blocking_wait_rq(bfqq) && arrived_in_time &&
1502 bfq_bfqq_budget_left(bfqq) > 0) {
1504 * We do not clear the flag non_blocking_wait_rq here, as
1505 * the latter is used in bfq_activate_bfqq to signal
1506 * that timestamps need to be back-shifted (and is
1507 * cleared right after).
1511 * In next assignment we rely on that either
1512 * entity->service or entity->budget are not updated
1513 * on expiration if bfqq is empty (see
1514 * __bfq_bfqq_recalc_budget). Thus both quantities
1515 * remain unchanged after such an expiration, and the
1516 * following statement therefore assigns to
1517 * entity->budget the remaining budget on such an
1520 entity->budget = min_t(unsigned long,
1521 bfq_bfqq_budget_left(bfqq),
1525 * At this point, we have used entity->service to get
1526 * the budget left (needed for updating
1527 * entity->budget). Thus we finally can, and have to,
1528 * reset entity->service. The latter must be reset
1529 * because bfqq would otherwise be charged again for
1530 * the service it has received during its previous
1533 entity->service = 0;
1539 * We can finally complete expiration, by setting service to 0.
1541 entity->service = 0;
1542 entity->budget = max_t(unsigned long, bfqq->max_budget,
1543 bfq_serv_to_charge(bfqq->next_rq, bfqq));
1544 bfq_clear_bfqq_non_blocking_wait_rq(bfqq);
1549 * Return the farthest past time instant according to jiffies
1552 static unsigned long bfq_smallest_from_now(void)
1554 return jiffies - MAX_JIFFY_OFFSET;
1557 static void bfq_update_bfqq_wr_on_rq_arrival(struct bfq_data *bfqd,
1558 struct bfq_queue *bfqq,
1559 unsigned int old_wr_coeff,
1560 bool wr_or_deserves_wr,
1565 if (old_wr_coeff == 1 && wr_or_deserves_wr) {
1566 /* start a weight-raising period */
1568 bfqq->service_from_wr = 0;
1569 bfqq->wr_coeff = bfqd->bfq_wr_coeff;
1570 bfqq->wr_cur_max_time = bfq_wr_duration(bfqd);
1573 * No interactive weight raising in progress
1574 * here: assign minus infinity to
1575 * wr_start_at_switch_to_srt, to make sure
1576 * that, at the end of the soft-real-time
1577 * weight raising periods that is starting
1578 * now, no interactive weight-raising period
1579 * may be wrongly considered as still in
1580 * progress (and thus actually started by
1583 bfqq->wr_start_at_switch_to_srt =
1584 bfq_smallest_from_now();
1585 bfqq->wr_coeff = bfqd->bfq_wr_coeff *
1586 BFQ_SOFTRT_WEIGHT_FACTOR;
1587 bfqq->wr_cur_max_time =
1588 bfqd->bfq_wr_rt_max_time;
1592 * If needed, further reduce budget to make sure it is
1593 * close to bfqq's backlog, so as to reduce the
1594 * scheduling-error component due to a too large
1595 * budget. Do not care about throughput consequences,
1596 * but only about latency. Finally, do not assign a
1597 * too small budget either, to avoid increasing
1598 * latency by causing too frequent expirations.
1600 bfqq->entity.budget = min_t(unsigned long,
1601 bfqq->entity.budget,
1602 2 * bfq_min_budget(bfqd));
1603 } else if (old_wr_coeff > 1) {
1604 if (interactive) { /* update wr coeff and duration */
1605 bfqq->wr_coeff = bfqd->bfq_wr_coeff;
1606 bfqq->wr_cur_max_time = bfq_wr_duration(bfqd);
1607 } else if (in_burst)
1611 * The application is now or still meeting the
1612 * requirements for being deemed soft rt. We
1613 * can then correctly and safely (re)charge
1614 * the weight-raising duration for the
1615 * application with the weight-raising
1616 * duration for soft rt applications.
1618 * In particular, doing this recharge now, i.e.,
1619 * before the weight-raising period for the
1620 * application finishes, reduces the probability
1621 * of the following negative scenario:
1622 * 1) the weight of a soft rt application is
1623 * raised at startup (as for any newly
1624 * created application),
1625 * 2) since the application is not interactive,
1626 * at a certain time weight-raising is
1627 * stopped for the application,
1628 * 3) at that time the application happens to
1629 * still have pending requests, and hence
1630 * is destined to not have a chance to be
1631 * deemed soft rt before these requests are
1632 * completed (see the comments to the
1633 * function bfq_bfqq_softrt_next_start()
1634 * for details on soft rt detection),
1635 * 4) these pending requests experience a high
1636 * latency because the application is not
1637 * weight-raised while they are pending.
1639 if (bfqq->wr_cur_max_time !=
1640 bfqd->bfq_wr_rt_max_time) {
1641 bfqq->wr_start_at_switch_to_srt =
1642 bfqq->last_wr_start_finish;
1644 bfqq->wr_cur_max_time =
1645 bfqd->bfq_wr_rt_max_time;
1646 bfqq->wr_coeff = bfqd->bfq_wr_coeff *
1647 BFQ_SOFTRT_WEIGHT_FACTOR;
1649 bfqq->last_wr_start_finish = jiffies;
1654 static bool bfq_bfqq_idle_for_long_time(struct bfq_data *bfqd,
1655 struct bfq_queue *bfqq)
1657 return bfqq->dispatched == 0 &&
1658 time_is_before_jiffies(
1659 bfqq->budget_timeout +
1660 bfqd->bfq_wr_min_idle_time);
1665 * Return true if bfqq is in a higher priority class, or has a higher
1666 * weight than the in-service queue.
1668 static bool bfq_bfqq_higher_class_or_weight(struct bfq_queue *bfqq,
1669 struct bfq_queue *in_serv_bfqq)
1671 int bfqq_weight, in_serv_weight;
1673 if (bfqq->ioprio_class < in_serv_bfqq->ioprio_class)
1676 if (in_serv_bfqq->entity.parent == bfqq->entity.parent) {
1677 bfqq_weight = bfqq->entity.weight;
1678 in_serv_weight = in_serv_bfqq->entity.weight;
1680 if (bfqq->entity.parent)
1681 bfqq_weight = bfqq->entity.parent->weight;
1683 bfqq_weight = bfqq->entity.weight;
1684 if (in_serv_bfqq->entity.parent)
1685 in_serv_weight = in_serv_bfqq->entity.parent->weight;
1687 in_serv_weight = in_serv_bfqq->entity.weight;
1690 return bfqq_weight > in_serv_weight;
1693 static bool bfq_better_to_idle(struct bfq_queue *bfqq);
1695 static void bfq_bfqq_handle_idle_busy_switch(struct bfq_data *bfqd,
1696 struct bfq_queue *bfqq,
1701 bool soft_rt, in_burst, wr_or_deserves_wr,
1702 bfqq_wants_to_preempt,
1703 idle_for_long_time = bfq_bfqq_idle_for_long_time(bfqd, bfqq),
1705 * See the comments on
1706 * bfq_bfqq_update_budg_for_activation for
1707 * details on the usage of the next variable.
1709 arrived_in_time = ktime_get_ns() <=
1710 bfqq->ttime.last_end_request +
1711 bfqd->bfq_slice_idle * 3;
1715 * bfqq deserves to be weight-raised if:
1717 * - it does not belong to a large burst,
1718 * - it has been idle for enough time or is soft real-time,
1719 * - is linked to a bfq_io_cq (it is not shared in any sense),
1720 * - has a default weight (otherwise we assume the user wanted
1721 * to control its weight explicitly)
1723 in_burst = bfq_bfqq_in_large_burst(bfqq);
1724 soft_rt = bfqd->bfq_wr_max_softrt_rate > 0 &&
1725 !BFQQ_TOTALLY_SEEKY(bfqq) &&
1727 time_is_before_jiffies(bfqq->soft_rt_next_start) &&
1728 bfqq->dispatched == 0 &&
1729 bfqq->entity.new_weight == 40;
1730 *interactive = !in_burst && idle_for_long_time &&
1731 bfqq->entity.new_weight == 40;
1732 wr_or_deserves_wr = bfqd->low_latency &&
1733 (bfqq->wr_coeff > 1 ||
1734 (bfq_bfqq_sync(bfqq) &&
1735 bfqq->bic && (*interactive || soft_rt)));
1738 * Using the last flag, update budget and check whether bfqq
1739 * may want to preempt the in-service queue.
1741 bfqq_wants_to_preempt =
1742 bfq_bfqq_update_budg_for_activation(bfqd, bfqq,
1746 * If bfqq happened to be activated in a burst, but has been
1747 * idle for much more than an interactive queue, then we
1748 * assume that, in the overall I/O initiated in the burst, the
1749 * I/O associated with bfqq is finished. So bfqq does not need
1750 * to be treated as a queue belonging to a burst
1751 * anymore. Accordingly, we reset bfqq's in_large_burst flag
1752 * if set, and remove bfqq from the burst list if it's
1753 * there. We do not decrement burst_size, because the fact
1754 * that bfqq does not need to belong to the burst list any
1755 * more does not invalidate the fact that bfqq was created in
1758 if (likely(!bfq_bfqq_just_created(bfqq)) &&
1759 idle_for_long_time &&
1760 time_is_before_jiffies(
1761 bfqq->budget_timeout +
1762 msecs_to_jiffies(10000))) {
1763 hlist_del_init(&bfqq->burst_list_node);
1764 bfq_clear_bfqq_in_large_burst(bfqq);
1767 bfq_clear_bfqq_just_created(bfqq);
1769 if (bfqd->low_latency) {
1770 if (unlikely(time_is_after_jiffies(bfqq->split_time)))
1773 jiffies - bfqd->bfq_wr_min_idle_time - 1;
1775 if (time_is_before_jiffies(bfqq->split_time +
1776 bfqd->bfq_wr_min_idle_time)) {
1777 bfq_update_bfqq_wr_on_rq_arrival(bfqd, bfqq,
1784 if (old_wr_coeff != bfqq->wr_coeff)
1785 bfqq->entity.prio_changed = 1;
1789 bfqq->last_idle_bklogged = jiffies;
1790 bfqq->service_from_backlogged = 0;
1791 bfq_clear_bfqq_softrt_update(bfqq);
1793 bfq_add_bfqq_busy(bfqd, bfqq);
1796 * Expire in-service queue if preemption may be needed for
1797 * guarantees or throughput. As for guarantees, we care
1798 * explicitly about two cases. The first is that bfqq has to
1799 * recover a service hole, as explained in the comments on
1800 * bfq_bfqq_update_budg_for_activation(), i.e., that
1801 * bfqq_wants_to_preempt is true. However, if bfqq does not
1802 * carry time-critical I/O, then bfqq's bandwidth is less
1803 * important than that of queues that carry time-critical I/O.
1804 * So, as a further constraint, we consider this case only if
1805 * bfqq is at least as weight-raised, i.e., at least as time
1806 * critical, as the in-service queue.
1808 * The second case is that bfqq is in a higher priority class,
1809 * or has a higher weight than the in-service queue. If this
1810 * condition does not hold, we don't care because, even if
1811 * bfqq does not start to be served immediately, the resulting
1812 * delay for bfqq's I/O is however lower or much lower than
1813 * the ideal completion time to be guaranteed to bfqq's I/O.
1815 * In both cases, preemption is needed only if, according to
1816 * the timestamps of both bfqq and of the in-service queue,
1817 * bfqq actually is the next queue to serve. So, to reduce
1818 * useless preemptions, the return value of
1819 * next_queue_may_preempt() is considered in the next compound
1820 * condition too. Yet next_queue_may_preempt() just checks a
1821 * simple, necessary condition for bfqq to be the next queue
1822 * to serve. In fact, to evaluate a sufficient condition, the
1823 * timestamps of the in-service queue would need to be
1824 * updated, and this operation is quite costly (see the
1825 * comments on bfq_bfqq_update_budg_for_activation()).
1827 * As for throughput, we ask bfq_better_to_idle() whether we
1828 * still need to plug I/O dispatching. If bfq_better_to_idle()
1829 * says no, then plugging is not needed any longer, either to
1830 * boost throughput or to perserve service guarantees. Then
1831 * the best option is to stop plugging I/O, as not doing so
1832 * would certainly lower throughput. We may end up in this
1833 * case if: (1) upon a dispatch attempt, we detected that it
1834 * was better to plug I/O dispatch, and to wait for a new
1835 * request to arrive for the currently in-service queue, but
1836 * (2) this switch of bfqq to busy changes the scenario.
1838 if (bfqd->in_service_queue &&
1839 ((bfqq_wants_to_preempt &&
1840 bfqq->wr_coeff >= bfqd->in_service_queue->wr_coeff) ||
1841 bfq_bfqq_higher_class_or_weight(bfqq, bfqd->in_service_queue) ||
1842 !bfq_better_to_idle(bfqd->in_service_queue)) &&
1843 next_queue_may_preempt(bfqd))
1844 bfq_bfqq_expire(bfqd, bfqd->in_service_queue,
1845 false, BFQQE_PREEMPTED);
1848 static void bfq_reset_inject_limit(struct bfq_data *bfqd,
1849 struct bfq_queue *bfqq)
1851 /* invalidate baseline total service time */
1852 bfqq->last_serv_time_ns = 0;
1855 * Reset pointer in case we are waiting for
1856 * some request completion.
1858 bfqd->waited_rq = NULL;
1861 * If bfqq has a short think time, then start by setting the
1862 * inject limit to 0 prudentially, because the service time of
1863 * an injected I/O request may be higher than the think time
1864 * of bfqq, and therefore, if one request was injected when
1865 * bfqq remains empty, this injected request might delay the
1866 * service of the next I/O request for bfqq significantly. In
1867 * case bfqq can actually tolerate some injection, then the
1868 * adaptive update will however raise the limit soon. This
1869 * lucky circumstance holds exactly because bfqq has a short
1870 * think time, and thus, after remaining empty, is likely to
1871 * get new I/O enqueued---and then completed---before being
1872 * expired. This is the very pattern that gives the
1873 * limit-update algorithm the chance to measure the effect of
1874 * injection on request service times, and then to update the
1875 * limit accordingly.
1877 * However, in the following special case, the inject limit is
1878 * left to 1 even if the think time is short: bfqq's I/O is
1879 * synchronized with that of some other queue, i.e., bfqq may
1880 * receive new I/O only after the I/O of the other queue is
1881 * completed. Keeping the inject limit to 1 allows the
1882 * blocking I/O to be served while bfqq is in service. And
1883 * this is very convenient both for bfqq and for overall
1884 * throughput, as explained in detail in the comments in
1885 * bfq_update_has_short_ttime().
1887 * On the opposite end, if bfqq has a long think time, then
1888 * start directly by 1, because:
1889 * a) on the bright side, keeping at most one request in
1890 * service in the drive is unlikely to cause any harm to the
1891 * latency of bfqq's requests, as the service time of a single
1892 * request is likely to be lower than the think time of bfqq;
1893 * b) on the downside, after becoming empty, bfqq is likely to
1894 * expire before getting its next request. With this request
1895 * arrival pattern, it is very hard to sample total service
1896 * times and update the inject limit accordingly (see comments
1897 * on bfq_update_inject_limit()). So the limit is likely to be
1898 * never, or at least seldom, updated. As a consequence, by
1899 * setting the limit to 1, we avoid that no injection ever
1900 * occurs with bfqq. On the downside, this proactive step
1901 * further reduces chances to actually compute the baseline
1902 * total service time. Thus it reduces chances to execute the
1903 * limit-update algorithm and possibly raise the limit to more
1906 if (bfq_bfqq_has_short_ttime(bfqq))
1907 bfqq->inject_limit = 0;
1909 bfqq->inject_limit = 1;
1911 bfqq->decrease_time_jif = jiffies;
1914 static void bfq_update_io_intensity(struct bfq_queue *bfqq, u64 now_ns)
1916 u64 tot_io_time = now_ns - bfqq->io_start_time;
1918 if (RB_EMPTY_ROOT(&bfqq->sort_list) && bfqq->dispatched == 0)
1919 bfqq->tot_idle_time +=
1920 now_ns - bfqq->ttime.last_end_request;
1922 if (unlikely(bfq_bfqq_just_created(bfqq)))
1926 * Must be busy for at least about 80% of the time to be
1927 * considered I/O bound.
1929 if (bfqq->tot_idle_time * 5 > tot_io_time)
1930 bfq_clear_bfqq_IO_bound(bfqq);
1932 bfq_mark_bfqq_IO_bound(bfqq);
1935 * Keep an observation window of at most 200 ms in the past
1938 if (tot_io_time > 200 * NSEC_PER_MSEC) {
1939 bfqq->io_start_time = now_ns - (tot_io_time>>1);
1940 bfqq->tot_idle_time >>= 1;
1945 * Detect whether bfqq's I/O seems synchronized with that of some
1946 * other queue, i.e., whether bfqq, after remaining empty, happens to
1947 * receive new I/O only right after some I/O request of the other
1948 * queue has been completed. We call waker queue the other queue, and
1949 * we assume, for simplicity, that bfqq may have at most one waker
1952 * A remarkable throughput boost can be reached by unconditionally
1953 * injecting the I/O of the waker queue, every time a new
1954 * bfq_dispatch_request happens to be invoked while I/O is being
1955 * plugged for bfqq. In addition to boosting throughput, this
1956 * unblocks bfqq's I/O, thereby improving bandwidth and latency for
1957 * bfqq. Note that these same results may be achieved with the general
1958 * injection mechanism, but less effectively. For details on this
1959 * aspect, see the comments on the choice of the queue for injection
1960 * in bfq_select_queue().
1962 * Turning back to the detection of a waker queue, a queue Q is deemed
1963 * as a waker queue for bfqq if, for three consecutive times, bfqq
1964 * happens to become non empty right after a request of Q has been
1965 * completed. In particular, on the first time, Q is tentatively set
1966 * as a candidate waker queue, while on the third consecutive time
1967 * that Q is detected, the field waker_bfqq is set to Q, to confirm
1968 * that Q is a waker queue for bfqq. These detection steps are
1969 * performed only if bfqq has a long think time, so as to make it more
1970 * likely that bfqq's I/O is actually being blocked by a
1971 * synchronization. This last filter, plus the above three-times
1972 * requirement, make false positives less likely.
1976 * The sooner a waker queue is detected, the sooner throughput can be
1977 * boosted by injecting I/O from the waker queue. Fortunately,
1978 * detection is likely to be actually fast, for the following
1979 * reasons. While blocked by synchronization, bfqq has a long think
1980 * time. This implies that bfqq's inject limit is at least equal to 1
1981 * (see the comments in bfq_update_inject_limit()). So, thanks to
1982 * injection, the waker queue is likely to be served during the very
1983 * first I/O-plugging time interval for bfqq. This triggers the first
1984 * step of the detection mechanism. Thanks again to injection, the
1985 * candidate waker queue is then likely to be confirmed no later than
1986 * during the next I/O-plugging interval for bfqq.
1990 * On queue merging all waker information is lost.
1992 static void bfq_check_waker(struct bfq_data *bfqd, struct bfq_queue *bfqq,
1995 if (!bfqd->last_completed_rq_bfqq ||
1996 bfqd->last_completed_rq_bfqq == bfqq ||
1997 bfq_bfqq_has_short_ttime(bfqq) ||
1998 now_ns - bfqd->last_completion >= 4 * NSEC_PER_MSEC ||
1999 bfqd->last_completed_rq_bfqq == bfqq->waker_bfqq)
2002 if (bfqd->last_completed_rq_bfqq !=
2003 bfqq->tentative_waker_bfqq) {
2005 * First synchronization detected with a
2006 * candidate waker queue, or with a different
2007 * candidate waker queue from the current one.
2009 bfqq->tentative_waker_bfqq =
2010 bfqd->last_completed_rq_bfqq;
2011 bfqq->num_waker_detections = 1;
2012 } else /* Same tentative waker queue detected again */
2013 bfqq->num_waker_detections++;
2015 if (bfqq->num_waker_detections == 3) {
2016 bfqq->waker_bfqq = bfqd->last_completed_rq_bfqq;
2017 bfqq->tentative_waker_bfqq = NULL;
2020 * If the waker queue disappears, then
2021 * bfqq->waker_bfqq must be reset. To
2022 * this goal, we maintain in each
2023 * waker queue a list, woken_list, of
2024 * all the queues that reference the
2025 * waker queue through their
2026 * waker_bfqq pointer. When the waker
2027 * queue exits, the waker_bfqq pointer
2028 * of all the queues in the woken_list
2031 * In addition, if bfqq is already in
2032 * the woken_list of a waker queue,
2033 * then, before being inserted into
2034 * the woken_list of a new waker
2035 * queue, bfqq must be removed from
2036 * the woken_list of the old waker
2039 if (!hlist_unhashed(&bfqq->woken_list_node))
2040 hlist_del_init(&bfqq->woken_list_node);
2041 hlist_add_head(&bfqq->woken_list_node,
2042 &bfqd->last_completed_rq_bfqq->woken_list);
2046 static void bfq_add_request(struct request *rq)
2048 struct bfq_queue *bfqq = RQ_BFQQ(rq);
2049 struct bfq_data *bfqd = bfqq->bfqd;
2050 struct request *next_rq, *prev;
2051 unsigned int old_wr_coeff = bfqq->wr_coeff;
2052 bool interactive = false;
2053 u64 now_ns = ktime_get_ns();
2055 bfq_log_bfqq(bfqd, bfqq, "add_request %d", rq_is_sync(rq));
2056 bfqq->queued[rq_is_sync(rq)]++;
2059 if (RB_EMPTY_ROOT(&bfqq->sort_list) && bfq_bfqq_sync(bfqq)) {
2060 bfq_check_waker(bfqd, bfqq, now_ns);
2063 * Periodically reset inject limit, to make sure that
2064 * the latter eventually drops in case workload
2065 * changes, see step (3) in the comments on
2066 * bfq_update_inject_limit().
2068 if (time_is_before_eq_jiffies(bfqq->decrease_time_jif +
2069 msecs_to_jiffies(1000)))
2070 bfq_reset_inject_limit(bfqd, bfqq);
2073 * The following conditions must hold to setup a new
2074 * sampling of total service time, and then a new
2075 * update of the inject limit:
2076 * - bfqq is in service, because the total service
2077 * time is evaluated only for the I/O requests of
2078 * the queues in service;
2079 * - this is the right occasion to compute or to
2080 * lower the baseline total service time, because
2081 * there are actually no requests in the drive,
2083 * the baseline total service time is available, and
2084 * this is the right occasion to compute the other
2085 * quantity needed to update the inject limit, i.e.,
2086 * the total service time caused by the amount of
2087 * injection allowed by the current value of the
2088 * limit. It is the right occasion because injection
2089 * has actually been performed during the service
2090 * hole, and there are still in-flight requests,
2091 * which are very likely to be exactly the injected
2092 * requests, or part of them;
2093 * - the minimum interval for sampling the total
2094 * service time and updating the inject limit has
2097 if (bfqq == bfqd->in_service_queue &&
2098 (bfqd->rq_in_driver == 0 ||
2099 (bfqq->last_serv_time_ns > 0 &&
2100 bfqd->rqs_injected && bfqd->rq_in_driver > 0)) &&
2101 time_is_before_eq_jiffies(bfqq->decrease_time_jif +
2102 msecs_to_jiffies(10))) {
2103 bfqd->last_empty_occupied_ns = ktime_get_ns();
2105 * Start the state machine for measuring the
2106 * total service time of rq: setting
2107 * wait_dispatch will cause bfqd->waited_rq to
2108 * be set when rq will be dispatched.
2110 bfqd->wait_dispatch = true;
2112 * If there is no I/O in service in the drive,
2113 * then possible injection occurred before the
2114 * arrival of rq will not affect the total
2115 * service time of rq. So the injection limit
2116 * must not be updated as a function of such
2117 * total service time, unless new injection
2118 * occurs before rq is completed. To have the
2119 * injection limit updated only in the latter
2120 * case, reset rqs_injected here (rqs_injected
2121 * will be set in case injection is performed
2122 * on bfqq before rq is completed).
2124 if (bfqd->rq_in_driver == 0)
2125 bfqd->rqs_injected = false;
2129 if (bfq_bfqq_sync(bfqq))
2130 bfq_update_io_intensity(bfqq, now_ns);
2132 elv_rb_add(&bfqq->sort_list, rq);
2135 * Check if this request is a better next-serve candidate.
2137 prev = bfqq->next_rq;
2138 next_rq = bfq_choose_req(bfqd, bfqq->next_rq, rq, bfqd->last_position);
2139 bfqq->next_rq = next_rq;
2142 * Adjust priority tree position, if next_rq changes.
2143 * See comments on bfq_pos_tree_add_move() for the unlikely().
2145 if (unlikely(!bfqd->nonrot_with_queueing && prev != bfqq->next_rq))
2146 bfq_pos_tree_add_move(bfqd, bfqq);
2148 if (!bfq_bfqq_busy(bfqq)) /* switching to busy ... */
2149 bfq_bfqq_handle_idle_busy_switch(bfqd, bfqq, old_wr_coeff,
2152 if (bfqd->low_latency && old_wr_coeff == 1 && !rq_is_sync(rq) &&
2153 time_is_before_jiffies(
2154 bfqq->last_wr_start_finish +
2155 bfqd->bfq_wr_min_inter_arr_async)) {
2156 bfqq->wr_coeff = bfqd->bfq_wr_coeff;
2157 bfqq->wr_cur_max_time = bfq_wr_duration(bfqd);
2159 bfqd->wr_busy_queues++;
2160 bfqq->entity.prio_changed = 1;
2162 if (prev != bfqq->next_rq)
2163 bfq_updated_next_req(bfqd, bfqq);
2167 * Assign jiffies to last_wr_start_finish in the following
2170 * . if bfqq is not going to be weight-raised, because, for
2171 * non weight-raised queues, last_wr_start_finish stores the
2172 * arrival time of the last request; as of now, this piece
2173 * of information is used only for deciding whether to
2174 * weight-raise async queues
2176 * . if bfqq is not weight-raised, because, if bfqq is now
2177 * switching to weight-raised, then last_wr_start_finish
2178 * stores the time when weight-raising starts
2180 * . if bfqq is interactive, because, regardless of whether
2181 * bfqq is currently weight-raised, the weight-raising
2182 * period must start or restart (this case is considered
2183 * separately because it is not detected by the above
2184 * conditions, if bfqq is already weight-raised)
2186 * last_wr_start_finish has to be updated also if bfqq is soft
2187 * real-time, because the weight-raising period is constantly
2188 * restarted on idle-to-busy transitions for these queues, but
2189 * this is already done in bfq_bfqq_handle_idle_busy_switch if
2192 if (bfqd->low_latency &&
2193 (old_wr_coeff == 1 || bfqq->wr_coeff == 1 || interactive))
2194 bfqq->last_wr_start_finish = jiffies;
2197 static struct request *bfq_find_rq_fmerge(struct bfq_data *bfqd,
2199 struct request_queue *q)
2201 struct bfq_queue *bfqq = bfqd->bio_bfqq;
2205 return elv_rb_find(&bfqq->sort_list, bio_end_sector(bio));
2210 static sector_t get_sdist(sector_t last_pos, struct request *rq)
2213 return abs(blk_rq_pos(rq) - last_pos);
2218 #if 0 /* Still not clear if we can do without next two functions */
2219 static void bfq_activate_request(struct request_queue *q, struct request *rq)
2221 struct bfq_data *bfqd = q->elevator->elevator_data;
2223 bfqd->rq_in_driver++;
2226 static void bfq_deactivate_request(struct request_queue *q, struct request *rq)
2228 struct bfq_data *bfqd = q->elevator->elevator_data;
2230 bfqd->rq_in_driver--;
2234 static void bfq_remove_request(struct request_queue *q,
2237 struct bfq_queue *bfqq = RQ_BFQQ(rq);
2238 struct bfq_data *bfqd = bfqq->bfqd;
2239 const int sync = rq_is_sync(rq);
2241 if (bfqq->next_rq == rq) {
2242 bfqq->next_rq = bfq_find_next_rq(bfqd, bfqq, rq);
2243 bfq_updated_next_req(bfqd, bfqq);
2246 if (rq->queuelist.prev != &rq->queuelist)
2247 list_del_init(&rq->queuelist);
2248 bfqq->queued[sync]--;
2250 elv_rb_del(&bfqq->sort_list, rq);
2252 elv_rqhash_del(q, rq);
2253 if (q->last_merge == rq)
2254 q->last_merge = NULL;
2256 if (RB_EMPTY_ROOT(&bfqq->sort_list)) {
2257 bfqq->next_rq = NULL;
2259 if (bfq_bfqq_busy(bfqq) && bfqq != bfqd->in_service_queue) {
2260 bfq_del_bfqq_busy(bfqd, bfqq, false);
2262 * bfqq emptied. In normal operation, when
2263 * bfqq is empty, bfqq->entity.service and
2264 * bfqq->entity.budget must contain,
2265 * respectively, the service received and the
2266 * budget used last time bfqq emptied. These
2267 * facts do not hold in this case, as at least
2268 * this last removal occurred while bfqq is
2269 * not in service. To avoid inconsistencies,
2270 * reset both bfqq->entity.service and
2271 * bfqq->entity.budget, if bfqq has still a
2272 * process that may issue I/O requests to it.
2274 bfqq->entity.budget = bfqq->entity.service = 0;
2278 * Remove queue from request-position tree as it is empty.
2280 if (bfqq->pos_root) {
2281 rb_erase(&bfqq->pos_node, bfqq->pos_root);
2282 bfqq->pos_root = NULL;
2285 /* see comments on bfq_pos_tree_add_move() for the unlikely() */
2286 if (unlikely(!bfqd->nonrot_with_queueing))
2287 bfq_pos_tree_add_move(bfqd, bfqq);
2290 if (rq->cmd_flags & REQ_META)
2291 bfqq->meta_pending--;
2295 static bool bfq_bio_merge(struct request_queue *q, struct bio *bio,
2296 unsigned int nr_segs)
2298 struct bfq_data *bfqd = q->elevator->elevator_data;
2299 struct request *free = NULL;
2301 * bfq_bic_lookup grabs the queue_lock: invoke it now and
2302 * store its return value for later use, to avoid nesting
2303 * queue_lock inside the bfqd->lock. We assume that the bic
2304 * returned by bfq_bic_lookup does not go away before
2305 * bfqd->lock is taken.
2307 struct bfq_io_cq *bic = bfq_bic_lookup(bfqd, current->io_context, q);
2310 spin_lock_irq(&bfqd->lock);
2313 bfqd->bio_bfqq = bic_to_bfqq(bic, op_is_sync(bio->bi_opf));
2315 bfqd->bio_bfqq = NULL;
2316 bfqd->bio_bic = bic;
2318 ret = blk_mq_sched_try_merge(q, bio, nr_segs, &free);
2321 blk_mq_free_request(free);
2322 spin_unlock_irq(&bfqd->lock);
2327 static int bfq_request_merge(struct request_queue *q, struct request **req,
2330 struct bfq_data *bfqd = q->elevator->elevator_data;
2331 struct request *__rq;
2333 __rq = bfq_find_rq_fmerge(bfqd, bio, q);
2334 if (__rq && elv_bio_merge_ok(__rq, bio)) {
2336 return ELEVATOR_FRONT_MERGE;
2339 return ELEVATOR_NO_MERGE;
2342 static struct bfq_queue *bfq_init_rq(struct request *rq);
2344 static void bfq_request_merged(struct request_queue *q, struct request *req,
2345 enum elv_merge type)
2347 if (type == ELEVATOR_FRONT_MERGE &&
2348 rb_prev(&req->rb_node) &&
2350 blk_rq_pos(container_of(rb_prev(&req->rb_node),
2351 struct request, rb_node))) {
2352 struct bfq_queue *bfqq = bfq_init_rq(req);
2353 struct bfq_data *bfqd;
2354 struct request *prev, *next_rq;
2361 /* Reposition request in its sort_list */
2362 elv_rb_del(&bfqq->sort_list, req);
2363 elv_rb_add(&bfqq->sort_list, req);
2365 /* Choose next request to be served for bfqq */
2366 prev = bfqq->next_rq;
2367 next_rq = bfq_choose_req(bfqd, bfqq->next_rq, req,
2368 bfqd->last_position);
2369 bfqq->next_rq = next_rq;
2371 * If next_rq changes, update both the queue's budget to
2372 * fit the new request and the queue's position in its
2375 if (prev != bfqq->next_rq) {
2376 bfq_updated_next_req(bfqd, bfqq);
2378 * See comments on bfq_pos_tree_add_move() for
2381 if (unlikely(!bfqd->nonrot_with_queueing))
2382 bfq_pos_tree_add_move(bfqd, bfqq);
2388 * This function is called to notify the scheduler that the requests
2389 * rq and 'next' have been merged, with 'next' going away. BFQ
2390 * exploits this hook to address the following issue: if 'next' has a
2391 * fifo_time lower that rq, then the fifo_time of rq must be set to
2392 * the value of 'next', to not forget the greater age of 'next'.
2394 * NOTE: in this function we assume that rq is in a bfq_queue, basing
2395 * on that rq is picked from the hash table q->elevator->hash, which,
2396 * in its turn, is filled only with I/O requests present in
2397 * bfq_queues, while BFQ is in use for the request queue q. In fact,
2398 * the function that fills this hash table (elv_rqhash_add) is called
2399 * only by bfq_insert_request.
2401 static void bfq_requests_merged(struct request_queue *q, struct request *rq,
2402 struct request *next)
2404 struct bfq_queue *bfqq = bfq_init_rq(rq),
2405 *next_bfqq = bfq_init_rq(next);
2411 * If next and rq belong to the same bfq_queue and next is older
2412 * than rq, then reposition rq in the fifo (by substituting next
2413 * with rq). Otherwise, if next and rq belong to different
2414 * bfq_queues, never reposition rq: in fact, we would have to
2415 * reposition it with respect to next's position in its own fifo,
2416 * which would most certainly be too expensive with respect to
2419 if (bfqq == next_bfqq &&
2420 !list_empty(&rq->queuelist) && !list_empty(&next->queuelist) &&
2421 next->fifo_time < rq->fifo_time) {
2422 list_del_init(&rq->queuelist);
2423 list_replace_init(&next->queuelist, &rq->queuelist);
2424 rq->fifo_time = next->fifo_time;
2427 if (bfqq->next_rq == next)
2430 bfqg_stats_update_io_merged(bfqq_group(bfqq), next->cmd_flags);
2433 /* Must be called with bfqq != NULL */
2434 static void bfq_bfqq_end_wr(struct bfq_queue *bfqq)
2437 * If bfqq has been enjoying interactive weight-raising, then
2438 * reset soft_rt_next_start. We do it for the following
2439 * reason. bfqq may have been conveying the I/O needed to load
2440 * a soft real-time application. Such an application actually
2441 * exhibits a soft real-time I/O pattern after it finishes
2442 * loading, and finally starts doing its job. But, if bfqq has
2443 * been receiving a lot of bandwidth so far (likely to happen
2444 * on a fast device), then soft_rt_next_start now contains a
2445 * high value that. So, without this reset, bfqq would be
2446 * prevented from being possibly considered as soft_rt for a
2450 if (bfqq->wr_cur_max_time !=
2451 bfqq->bfqd->bfq_wr_rt_max_time)
2452 bfqq->soft_rt_next_start = jiffies;
2454 if (bfq_bfqq_busy(bfqq))
2455 bfqq->bfqd->wr_busy_queues--;
2457 bfqq->wr_cur_max_time = 0;
2458 bfqq->last_wr_start_finish = jiffies;
2460 * Trigger a weight change on the next invocation of
2461 * __bfq_entity_update_weight_prio.
2463 bfqq->entity.prio_changed = 1;
2466 void bfq_end_wr_async_queues(struct bfq_data *bfqd,
2467 struct bfq_group *bfqg)
2471 for (i = 0; i < 2; i++)
2472 for (j = 0; j < IOPRIO_BE_NR; j++)
2473 if (bfqg->async_bfqq[i][j])
2474 bfq_bfqq_end_wr(bfqg->async_bfqq[i][j]);
2475 if (bfqg->async_idle_bfqq)
2476 bfq_bfqq_end_wr(bfqg->async_idle_bfqq);
2479 static void bfq_end_wr(struct bfq_data *bfqd)
2481 struct bfq_queue *bfqq;
2483 spin_lock_irq(&bfqd->lock);
2485 list_for_each_entry(bfqq, &bfqd->active_list, bfqq_list)
2486 bfq_bfqq_end_wr(bfqq);
2487 list_for_each_entry(bfqq, &bfqd->idle_list, bfqq_list)
2488 bfq_bfqq_end_wr(bfqq);
2489 bfq_end_wr_async(bfqd);
2491 spin_unlock_irq(&bfqd->lock);
2494 static sector_t bfq_io_struct_pos(void *io_struct, bool request)
2497 return blk_rq_pos(io_struct);
2499 return ((struct bio *)io_struct)->bi_iter.bi_sector;
2502 static int bfq_rq_close_to_sector(void *io_struct, bool request,
2505 return abs(bfq_io_struct_pos(io_struct, request) - sector) <=
2509 static struct bfq_queue *bfqq_find_close(struct bfq_data *bfqd,
2510 struct bfq_queue *bfqq,
2513 struct rb_root *root = &bfq_bfqq_to_bfqg(bfqq)->rq_pos_tree;
2514 struct rb_node *parent, *node;
2515 struct bfq_queue *__bfqq;
2517 if (RB_EMPTY_ROOT(root))
2521 * First, if we find a request starting at the end of the last
2522 * request, choose it.
2524 __bfqq = bfq_rq_pos_tree_lookup(bfqd, root, sector, &parent, NULL);
2529 * If the exact sector wasn't found, the parent of the NULL leaf
2530 * will contain the closest sector (rq_pos_tree sorted by
2531 * next_request position).
2533 __bfqq = rb_entry(parent, struct bfq_queue, pos_node);
2534 if (bfq_rq_close_to_sector(__bfqq->next_rq, true, sector))
2537 if (blk_rq_pos(__bfqq->next_rq) < sector)
2538 node = rb_next(&__bfqq->pos_node);
2540 node = rb_prev(&__bfqq->pos_node);
2544 __bfqq = rb_entry(node, struct bfq_queue, pos_node);
2545 if (bfq_rq_close_to_sector(__bfqq->next_rq, true, sector))
2551 static struct bfq_queue *bfq_find_close_cooperator(struct bfq_data *bfqd,
2552 struct bfq_queue *cur_bfqq,
2555 struct bfq_queue *bfqq;
2558 * We shall notice if some of the queues are cooperating,
2559 * e.g., working closely on the same area of the device. In
2560 * that case, we can group them together and: 1) don't waste
2561 * time idling, and 2) serve the union of their requests in
2562 * the best possible order for throughput.
2564 bfqq = bfqq_find_close(bfqd, cur_bfqq, sector);
2565 if (!bfqq || bfqq == cur_bfqq)
2571 static struct bfq_queue *
2572 bfq_setup_merge(struct bfq_queue *bfqq, struct bfq_queue *new_bfqq)
2574 int process_refs, new_process_refs;
2575 struct bfq_queue *__bfqq;
2578 * If there are no process references on the new_bfqq, then it is
2579 * unsafe to follow the ->new_bfqq chain as other bfqq's in the chain
2580 * may have dropped their last reference (not just their last process
2583 if (!bfqq_process_refs(new_bfqq))
2586 /* Avoid a circular list and skip interim queue merges. */
2587 while ((__bfqq = new_bfqq->new_bfqq)) {
2593 process_refs = bfqq_process_refs(bfqq);
2594 new_process_refs = bfqq_process_refs(new_bfqq);
2596 * If the process for the bfqq has gone away, there is no
2597 * sense in merging the queues.
2599 if (process_refs == 0 || new_process_refs == 0)
2602 bfq_log_bfqq(bfqq->bfqd, bfqq, "scheduling merge with queue %d",
2606 * Merging is just a redirection: the requests of the process
2607 * owning one of the two queues are redirected to the other queue.
2608 * The latter queue, in its turn, is set as shared if this is the
2609 * first time that the requests of some process are redirected to
2612 * We redirect bfqq to new_bfqq and not the opposite, because
2613 * we are in the context of the process owning bfqq, thus we
2614 * have the io_cq of this process. So we can immediately
2615 * configure this io_cq to redirect the requests of the
2616 * process to new_bfqq. In contrast, the io_cq of new_bfqq is
2617 * not available any more (new_bfqq->bic == NULL).
2619 * Anyway, even in case new_bfqq coincides with the in-service
2620 * queue, redirecting requests the in-service queue is the
2621 * best option, as we feed the in-service queue with new
2622 * requests close to the last request served and, by doing so,
2623 * are likely to increase the throughput.
2625 bfqq->new_bfqq = new_bfqq;
2626 new_bfqq->ref += process_refs;
2630 static bool bfq_may_be_close_cooperator(struct bfq_queue *bfqq,
2631 struct bfq_queue *new_bfqq)
2633 if (bfq_too_late_for_merging(new_bfqq))
2636 if (bfq_class_idle(bfqq) || bfq_class_idle(new_bfqq) ||
2637 (bfqq->ioprio_class != new_bfqq->ioprio_class))
2641 * If either of the queues has already been detected as seeky,
2642 * then merging it with the other queue is unlikely to lead to
2645 if (BFQQ_SEEKY(bfqq) || BFQQ_SEEKY(new_bfqq))
2649 * Interleaved I/O is known to be done by (some) applications
2650 * only for reads, so it does not make sense to merge async
2653 if (!bfq_bfqq_sync(bfqq) || !bfq_bfqq_sync(new_bfqq))
2659 static bool idling_boosts_thr_without_issues(struct bfq_data *bfqd,
2660 struct bfq_queue *bfqq);
2663 * Attempt to schedule a merge of bfqq with the currently in-service
2664 * queue or with a close queue among the scheduled queues. Return
2665 * NULL if no merge was scheduled, a pointer to the shared bfq_queue
2666 * structure otherwise.
2668 * The OOM queue is not allowed to participate to cooperation: in fact, since
2669 * the requests temporarily redirected to the OOM queue could be redirected
2670 * again to dedicated queues at any time, the state needed to correctly
2671 * handle merging with the OOM queue would be quite complex and expensive
2672 * to maintain. Besides, in such a critical condition as an out of memory,
2673 * the benefits of queue merging may be little relevant, or even negligible.
2675 * WARNING: queue merging may impair fairness among non-weight raised
2676 * queues, for at least two reasons: 1) the original weight of a
2677 * merged queue may change during the merged state, 2) even being the
2678 * weight the same, a merged queue may be bloated with many more
2679 * requests than the ones produced by its originally-associated
2682 static struct bfq_queue *
2683 bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
2684 void *io_struct, bool request, struct bfq_io_cq *bic)
2686 struct bfq_queue *in_service_bfqq, *new_bfqq;
2689 * Check delayed stable merge for rotational or non-queueing
2690 * devs. For this branch to be executed, bfqq must not be
2691 * currently merged with some other queue (i.e., bfqq->bic
2692 * must be non null). If we considered also merged queues,
2693 * then we should also check whether bfqq has already been
2694 * merged with bic->stable_merge_bfqq. But this would be
2695 * costly and complicated.
2697 if (unlikely(!bfqd->nonrot_with_queueing)) {
2699 * Make sure also that bfqq is sync, because
2700 * bic->stable_merge_bfqq may point to some queue (for
2701 * stable merging) also if bic is associated with a
2702 * sync queue, but this bfqq is async
2704 if (bfq_bfqq_sync(bfqq) && bic->stable_merge_bfqq &&
2705 !bfq_bfqq_just_created(bfqq) &&
2706 time_is_before_jiffies(bfqq->split_time +
2707 msecs_to_jiffies(200))) {
2708 struct bfq_queue *stable_merge_bfqq =
2709 bic->stable_merge_bfqq;
2710 int proc_ref = min(bfqq_process_refs(bfqq),
2711 bfqq_process_refs(stable_merge_bfqq));
2713 /* deschedule stable merge, because done or aborted here */
2714 bfq_put_stable_ref(stable_merge_bfqq);
2716 bic->stable_merge_bfqq = NULL;
2718 if (!idling_boosts_thr_without_issues(bfqd, bfqq) &&
2720 /* next function will take at least one ref */
2721 struct bfq_queue *new_bfqq =
2722 bfq_setup_merge(bfqq, stable_merge_bfqq);
2724 bic->stably_merged = true;
2725 if (new_bfqq && new_bfqq->bic)
2726 new_bfqq->bic->stably_merged = true;
2734 * Do not perform queue merging if the device is non
2735 * rotational and performs internal queueing. In fact, such a
2736 * device reaches a high speed through internal parallelism
2737 * and pipelining. This means that, to reach a high
2738 * throughput, it must have many requests enqueued at the same
2739 * time. But, in this configuration, the internal scheduling
2740 * algorithm of the device does exactly the job of queue
2741 * merging: it reorders requests so as to obtain as much as
2742 * possible a sequential I/O pattern. As a consequence, with
2743 * the workload generated by processes doing interleaved I/O,
2744 * the throughput reached by the device is likely to be the
2745 * same, with and without queue merging.
2747 * Disabling merging also provides a remarkable benefit in
2748 * terms of throughput. Merging tends to make many workloads
2749 * artificially more uneven, because of shared queues
2750 * remaining non empty for incomparably more time than
2751 * non-merged queues. This may accentuate workload
2752 * asymmetries. For example, if one of the queues in a set of
2753 * merged queues has a higher weight than a normal queue, then
2754 * the shared queue may inherit such a high weight and, by
2755 * staying almost always active, may force BFQ to perform I/O
2756 * plugging most of the time. This evidently makes it harder
2757 * for BFQ to let the device reach a high throughput.
2759 * Finally, the likely() macro below is not used because one
2760 * of the two branches is more likely than the other, but to
2761 * have the code path after the following if() executed as
2762 * fast as possible for the case of a non rotational device
2763 * with queueing. We want it because this is the fastest kind
2764 * of device. On the opposite end, the likely() may lengthen
2765 * the execution time of BFQ for the case of slower devices
2766 * (rotational or at least without queueing). But in this case
2767 * the execution time of BFQ matters very little, if not at
2770 if (likely(bfqd->nonrot_with_queueing))
2774 * Prevent bfqq from being merged if it has been created too
2775 * long ago. The idea is that true cooperating processes, and
2776 * thus their associated bfq_queues, are supposed to be
2777 * created shortly after each other. This is the case, e.g.,
2778 * for KVM/QEMU and dump I/O threads. Basing on this
2779 * assumption, the following filtering greatly reduces the
2780 * probability that two non-cooperating processes, which just
2781 * happen to do close I/O for some short time interval, have
2782 * their queues merged by mistake.
2784 if (bfq_too_late_for_merging(bfqq))
2788 return bfqq->new_bfqq;
2790 if (!io_struct || unlikely(bfqq == &bfqd->oom_bfqq))
2793 /* If there is only one backlogged queue, don't search. */
2794 if (bfq_tot_busy_queues(bfqd) == 1)
2797 in_service_bfqq = bfqd->in_service_queue;
2799 if (in_service_bfqq && in_service_bfqq != bfqq &&
2800 likely(in_service_bfqq != &bfqd->oom_bfqq) &&
2801 bfq_rq_close_to_sector(io_struct, request,
2802 bfqd->in_serv_last_pos) &&
2803 bfqq->entity.parent == in_service_bfqq->entity.parent &&
2804 bfq_may_be_close_cooperator(bfqq, in_service_bfqq)) {
2805 new_bfqq = bfq_setup_merge(bfqq, in_service_bfqq);
2810 * Check whether there is a cooperator among currently scheduled
2811 * queues. The only thing we need is that the bio/request is not
2812 * NULL, as we need it to establish whether a cooperator exists.
2814 new_bfqq = bfq_find_close_cooperator(bfqd, bfqq,
2815 bfq_io_struct_pos(io_struct, request));
2817 if (new_bfqq && likely(new_bfqq != &bfqd->oom_bfqq) &&
2818 bfq_may_be_close_cooperator(bfqq, new_bfqq))
2819 return bfq_setup_merge(bfqq, new_bfqq);
2824 static void bfq_bfqq_save_state(struct bfq_queue *bfqq)
2826 struct bfq_io_cq *bic = bfqq->bic;
2829 * If !bfqq->bic, the queue is already shared or its requests
2830 * have already been redirected to a shared queue; both idle window
2831 * and weight raising state have already been saved. Do nothing.
2836 bic->saved_last_serv_time_ns = bfqq->last_serv_time_ns;
2837 bic->saved_inject_limit = bfqq->inject_limit;
2838 bic->saved_decrease_time_jif = bfqq->decrease_time_jif;
2840 bic->saved_weight = bfqq->entity.orig_weight;
2841 bic->saved_ttime = bfqq->ttime;
2842 bic->saved_has_short_ttime = bfq_bfqq_has_short_ttime(bfqq);
2843 bic->saved_IO_bound = bfq_bfqq_IO_bound(bfqq);
2844 bic->saved_io_start_time = bfqq->io_start_time;
2845 bic->saved_tot_idle_time = bfqq->tot_idle_time;
2846 bic->saved_in_large_burst = bfq_bfqq_in_large_burst(bfqq);
2847 bic->was_in_burst_list = !hlist_unhashed(&bfqq->burst_list_node);
2848 if (unlikely(bfq_bfqq_just_created(bfqq) &&
2849 !bfq_bfqq_in_large_burst(bfqq) &&
2850 bfqq->bfqd->low_latency)) {
2852 * bfqq being merged right after being created: bfqq
2853 * would have deserved interactive weight raising, but
2854 * did not make it to be set in a weight-raised state,
2855 * because of this early merge. Store directly the
2856 * weight-raising state that would have been assigned
2857 * to bfqq, so that to avoid that bfqq unjustly fails
2858 * to enjoy weight raising if split soon.
2860 bic->saved_wr_coeff = bfqq->bfqd->bfq_wr_coeff;
2861 bic->saved_wr_start_at_switch_to_srt = bfq_smallest_from_now();
2862 bic->saved_wr_cur_max_time = bfq_wr_duration(bfqq->bfqd);
2863 bic->saved_last_wr_start_finish = jiffies;
2865 bic->saved_wr_coeff = bfqq->wr_coeff;
2866 bic->saved_wr_start_at_switch_to_srt =
2867 bfqq->wr_start_at_switch_to_srt;
2868 bic->saved_service_from_wr = bfqq->service_from_wr;
2869 bic->saved_last_wr_start_finish = bfqq->last_wr_start_finish;
2870 bic->saved_wr_cur_max_time = bfqq->wr_cur_max_time;
2876 bfq_reassign_last_bfqq(struct bfq_queue *cur_bfqq, struct bfq_queue *new_bfqq)
2878 if (cur_bfqq->entity.parent &&
2879 cur_bfqq->entity.parent->last_bfqq_created == cur_bfqq)
2880 cur_bfqq->entity.parent->last_bfqq_created = new_bfqq;
2881 else if (cur_bfqq->bfqd && cur_bfqq->bfqd->last_bfqq_created == cur_bfqq)
2882 cur_bfqq->bfqd->last_bfqq_created = new_bfqq;
2885 void bfq_release_process_ref(struct bfq_data *bfqd, struct bfq_queue *bfqq)
2888 * To prevent bfqq's service guarantees from being violated,
2889 * bfqq may be left busy, i.e., queued for service, even if
2890 * empty (see comments in __bfq_bfqq_expire() for
2891 * details). But, if no process will send requests to bfqq any
2892 * longer, then there is no point in keeping bfqq queued for
2893 * service. In addition, keeping bfqq queued for service, but
2894 * with no process ref any longer, may have caused bfqq to be
2895 * freed when dequeued from service. But this is assumed to
2898 if (bfq_bfqq_busy(bfqq) && RB_EMPTY_ROOT(&bfqq->sort_list) &&
2899 bfqq != bfqd->in_service_queue)
2900 bfq_del_bfqq_busy(bfqd, bfqq, false);
2902 bfq_reassign_last_bfqq(bfqq, NULL);
2904 bfq_put_queue(bfqq);
2908 bfq_merge_bfqqs(struct bfq_data *bfqd, struct bfq_io_cq *bic,
2909 struct bfq_queue *bfqq, struct bfq_queue *new_bfqq)
2911 bfq_log_bfqq(bfqd, bfqq, "merging with queue %lu",
2912 (unsigned long)new_bfqq->pid);
2913 /* Save weight raising and idle window of the merged queues */
2914 bfq_bfqq_save_state(bfqq);
2915 bfq_bfqq_save_state(new_bfqq);
2916 if (bfq_bfqq_IO_bound(bfqq))
2917 bfq_mark_bfqq_IO_bound(new_bfqq);
2918 bfq_clear_bfqq_IO_bound(bfqq);
2921 * The processes associated with bfqq are cooperators of the
2922 * processes associated with new_bfqq. So, if bfqq has a
2923 * waker, then assume that all these processes will be happy
2924 * to let bfqq's waker freely inject I/O when they have no
2927 if (bfqq->waker_bfqq && !new_bfqq->waker_bfqq &&
2928 bfqq->waker_bfqq != new_bfqq) {
2929 new_bfqq->waker_bfqq = bfqq->waker_bfqq;
2930 new_bfqq->tentative_waker_bfqq = NULL;
2933 * If the waker queue disappears, then
2934 * new_bfqq->waker_bfqq must be reset. So insert
2935 * new_bfqq into the woken_list of the waker. See
2936 * bfq_check_waker for details.
2938 hlist_add_head(&new_bfqq->woken_list_node,
2939 &new_bfqq->waker_bfqq->woken_list);
2944 * If bfqq is weight-raised, then let new_bfqq inherit
2945 * weight-raising. To reduce false positives, neglect the case
2946 * where bfqq has just been created, but has not yet made it
2947 * to be weight-raised (which may happen because EQM may merge
2948 * bfqq even before bfq_add_request is executed for the first
2949 * time for bfqq). Handling this case would however be very
2950 * easy, thanks to the flag just_created.
2952 if (new_bfqq->wr_coeff == 1 && bfqq->wr_coeff > 1) {
2953 new_bfqq->wr_coeff = bfqq->wr_coeff;
2954 new_bfqq->wr_cur_max_time = bfqq->wr_cur_max_time;
2955 new_bfqq->last_wr_start_finish = bfqq->last_wr_start_finish;
2956 new_bfqq->wr_start_at_switch_to_srt =
2957 bfqq->wr_start_at_switch_to_srt;
2958 if (bfq_bfqq_busy(new_bfqq))
2959 bfqd->wr_busy_queues++;
2960 new_bfqq->entity.prio_changed = 1;
2963 if (bfqq->wr_coeff > 1) { /* bfqq has given its wr to new_bfqq */
2965 bfqq->entity.prio_changed = 1;
2966 if (bfq_bfqq_busy(bfqq))
2967 bfqd->wr_busy_queues--;
2970 bfq_log_bfqq(bfqd, new_bfqq, "merge_bfqqs: wr_busy %d",
2971 bfqd->wr_busy_queues);
2974 * Merge queues (that is, let bic redirect its requests to new_bfqq)
2976 bic_set_bfqq(bic, new_bfqq, 1);
2977 bfq_mark_bfqq_coop(new_bfqq);
2979 * new_bfqq now belongs to at least two bics (it is a shared queue):
2980 * set new_bfqq->bic to NULL. bfqq either:
2981 * - does not belong to any bic any more, and hence bfqq->bic must
2982 * be set to NULL, or
2983 * - is a queue whose owning bics have already been redirected to a
2984 * different queue, hence the queue is destined to not belong to
2985 * any bic soon and bfqq->bic is already NULL (therefore the next
2986 * assignment causes no harm).
2988 new_bfqq->bic = NULL;
2990 * If the queue is shared, the pid is the pid of one of the associated
2991 * processes. Which pid depends on the exact sequence of merge events
2992 * the queue underwent. So printing such a pid is useless and confusing
2993 * because it reports a random pid between those of the associated
2995 * We mark such a queue with a pid -1, and then print SHARED instead of
2996 * a pid in logging messages.
3001 bfq_reassign_last_bfqq(bfqq, new_bfqq);
3003 bfq_release_process_ref(bfqd, bfqq);
3006 static bool bfq_allow_bio_merge(struct request_queue *q, struct request *rq,
3009 struct bfq_data *bfqd = q->elevator->elevator_data;
3010 bool is_sync = op_is_sync(bio->bi_opf);
3011 struct bfq_queue *bfqq = bfqd->bio_bfqq, *new_bfqq;
3014 * Disallow merge of a sync bio into an async request.
3016 if (is_sync && !rq_is_sync(rq))
3020 * Lookup the bfqq that this bio will be queued with. Allow
3021 * merge only if rq is queued there.
3027 * We take advantage of this function to perform an early merge
3028 * of the queues of possible cooperating processes.
3030 new_bfqq = bfq_setup_cooperator(bfqd, bfqq, bio, false, bfqd->bio_bic);
3033 * bic still points to bfqq, then it has not yet been
3034 * redirected to some other bfq_queue, and a queue
3035 * merge between bfqq and new_bfqq can be safely
3036 * fulfilled, i.e., bic can be redirected to new_bfqq
3037 * and bfqq can be put.
3039 bfq_merge_bfqqs(bfqd, bfqd->bio_bic, bfqq,
3042 * If we get here, bio will be queued into new_queue,
3043 * so use new_bfqq to decide whether bio and rq can be
3049 * Change also bqfd->bio_bfqq, as
3050 * bfqd->bio_bic now points to new_bfqq, and
3051 * this function may be invoked again (and then may
3052 * use again bqfd->bio_bfqq).
3054 bfqd->bio_bfqq = bfqq;
3057 return bfqq == RQ_BFQQ(rq);
3061 * Set the maximum time for the in-service queue to consume its
3062 * budget. This prevents seeky processes from lowering the throughput.
3063 * In practice, a time-slice service scheme is used with seeky
3066 static void bfq_set_budget_timeout(struct bfq_data *bfqd,
3067 struct bfq_queue *bfqq)
3069 unsigned int timeout_coeff;
3071 if (bfqq->wr_cur_max_time == bfqd->bfq_wr_rt_max_time)
3074 timeout_coeff = bfqq->entity.weight / bfqq->entity.orig_weight;
3076 bfqd->last_budget_start = ktime_get();
3078 bfqq->budget_timeout = jiffies +
3079 bfqd->bfq_timeout * timeout_coeff;
3082 static void __bfq_set_in_service_queue(struct bfq_data *bfqd,
3083 struct bfq_queue *bfqq)
3086 bfq_clear_bfqq_fifo_expire(bfqq);
3088 bfqd->budgets_assigned = (bfqd->budgets_assigned * 7 + 256) / 8;
3090 if (time_is_before_jiffies(bfqq->last_wr_start_finish) &&
3091 bfqq->wr_coeff > 1 &&
3092 bfqq->wr_cur_max_time == bfqd->bfq_wr_rt_max_time &&
3093 time_is_before_jiffies(bfqq->budget_timeout)) {
3095 * For soft real-time queues, move the start
3096 * of the weight-raising period forward by the
3097 * time the queue has not received any
3098 * service. Otherwise, a relatively long
3099 * service delay is likely to cause the
3100 * weight-raising period of the queue to end,
3101 * because of the short duration of the
3102 * weight-raising period of a soft real-time
3103 * queue. It is worth noting that this move
3104 * is not so dangerous for the other queues,
3105 * because soft real-time queues are not
3108 * To not add a further variable, we use the
3109 * overloaded field budget_timeout to
3110 * determine for how long the queue has not
3111 * received service, i.e., how much time has
3112 * elapsed since the queue expired. However,
3113 * this is a little imprecise, because
3114 * budget_timeout is set to jiffies if bfqq
3115 * not only expires, but also remains with no
3118 if (time_after(bfqq->budget_timeout,
3119 bfqq->last_wr_start_finish))
3120 bfqq->last_wr_start_finish +=
3121 jiffies - bfqq->budget_timeout;
3123 bfqq->last_wr_start_finish = jiffies;
3126 bfq_set_budget_timeout(bfqd, bfqq);
3127 bfq_log_bfqq(bfqd, bfqq,
3128 "set_in_service_queue, cur-budget = %d",
3129 bfqq->entity.budget);
3132 bfqd->in_service_queue = bfqq;
3133 bfqd->in_serv_last_pos = 0;
3137 * Get and set a new queue for service.
3139 static struct bfq_queue *bfq_set_in_service_queue(struct bfq_data *bfqd)
3141 struct bfq_queue *bfqq = bfq_get_next_queue(bfqd);
3143 __bfq_set_in_service_queue(bfqd, bfqq);
3147 static void bfq_arm_slice_timer(struct bfq_data *bfqd)
3149 struct bfq_queue *bfqq = bfqd->in_service_queue;
3152 bfq_mark_bfqq_wait_request(bfqq);
3155 * We don't want to idle for seeks, but we do want to allow
3156 * fair distribution of slice time for a process doing back-to-back
3157 * seeks. So allow a little bit of time for him to submit a new rq.
3159 sl = bfqd->bfq_slice_idle;
3161 * Unless the queue is being weight-raised or the scenario is
3162 * asymmetric, grant only minimum idle time if the queue
3163 * is seeky. A long idling is preserved for a weight-raised
3164 * queue, or, more in general, in an asymmetric scenario,
3165 * because a long idling is needed for guaranteeing to a queue
3166 * its reserved share of the throughput (in particular, it is
3167 * needed if the queue has a higher weight than some other
3170 if (BFQQ_SEEKY(bfqq) && bfqq->wr_coeff == 1 &&
3171 !bfq_asymmetric_scenario(bfqd, bfqq))
3172 sl = min_t(u64, sl, BFQ_MIN_TT);
3173 else if (bfqq->wr_coeff > 1)
3174 sl = max_t(u32, sl, 20ULL * NSEC_PER_MSEC);
3176 bfqd->last_idling_start = ktime_get();
3177 bfqd->last_idling_start_jiffies = jiffies;
3179 hrtimer_start(&bfqd->idle_slice_timer, ns_to_ktime(sl),
3181 bfqg_stats_set_start_idle_time(bfqq_group(bfqq));
3185 * In autotuning mode, max_budget is dynamically recomputed as the
3186 * amount of sectors transferred in timeout at the estimated peak
3187 * rate. This enables BFQ to utilize a full timeslice with a full
3188 * budget, even if the in-service queue is served at peak rate. And
3189 * this maximises throughput with sequential workloads.
3191 static unsigned long bfq_calc_max_budget(struct bfq_data *bfqd)
3193 return (u64)bfqd->peak_rate * USEC_PER_MSEC *
3194 jiffies_to_msecs(bfqd->bfq_timeout)>>BFQ_RATE_SHIFT;
3198 * Update parameters related to throughput and responsiveness, as a
3199 * function of the estimated peak rate. See comments on
3200 * bfq_calc_max_budget(), and on the ref_wr_duration array.
3202 static void update_thr_responsiveness_params(struct bfq_data *bfqd)
3204 if (bfqd->bfq_user_max_budget == 0) {
3205 bfqd->bfq_max_budget =
3206 bfq_calc_max_budget(bfqd);
3207 bfq_log(bfqd, "new max_budget = %d", bfqd->bfq_max_budget);
3211 static void bfq_reset_rate_computation(struct bfq_data *bfqd,
3214 if (rq != NULL) { /* new rq dispatch now, reset accordingly */
3215 bfqd->last_dispatch = bfqd->first_dispatch = ktime_get_ns();
3216 bfqd->peak_rate_samples = 1;
3217 bfqd->sequential_samples = 0;
3218 bfqd->tot_sectors_dispatched = bfqd->last_rq_max_size =
3220 } else /* no new rq dispatched, just reset the number of samples */
3221 bfqd->peak_rate_samples = 0; /* full re-init on next disp. */
3224 "reset_rate_computation at end, sample %u/%u tot_sects %llu",
3225 bfqd->peak_rate_samples, bfqd->sequential_samples,
3226 bfqd->tot_sectors_dispatched);
3229 static void bfq_update_rate_reset(struct bfq_data *bfqd, struct request *rq)
3231 u32 rate, weight, divisor;
3234 * For the convergence property to hold (see comments on
3235 * bfq_update_peak_rate()) and for the assessment to be
3236 * reliable, a minimum number of samples must be present, and
3237 * a minimum amount of time must have elapsed. If not so, do
3238 * not compute new rate. Just reset parameters, to get ready
3239 * for a new evaluation attempt.
3241 if (bfqd->peak_rate_samples < BFQ_RATE_MIN_SAMPLES ||
3242 bfqd->delta_from_first < BFQ_RATE_MIN_INTERVAL)
3243 goto reset_computation;
3246 * If a new request completion has occurred after last
3247 * dispatch, then, to approximate the rate at which requests
3248 * have been served by the device, it is more precise to
3249 * extend the observation interval to the last completion.
3251 bfqd->delta_from_first =
3252 max_t(u64, bfqd->delta_from_first,
3253 bfqd->last_completion - bfqd->first_dispatch);
3256 * Rate computed in sects/usec, and not sects/nsec, for
3259 rate = div64_ul(bfqd->tot_sectors_dispatched<<BFQ_RATE_SHIFT,
3260 div_u64(bfqd->delta_from_first, NSEC_PER_USEC));
3263 * Peak rate not updated if:
3264 * - the percentage of sequential dispatches is below 3/4 of the
3265 * total, and rate is below the current estimated peak rate
3266 * - rate is unreasonably high (> 20M sectors/sec)
3268 if ((bfqd->sequential_samples < (3 * bfqd->peak_rate_samples)>>2 &&
3269 rate <= bfqd->peak_rate) ||
3270 rate > 20<<BFQ_RATE_SHIFT)
3271 goto reset_computation;
3274 * We have to update the peak rate, at last! To this purpose,
3275 * we use a low-pass filter. We compute the smoothing constant
3276 * of the filter as a function of the 'weight' of the new
3279 * As can be seen in next formulas, we define this weight as a
3280 * quantity proportional to how sequential the workload is,
3281 * and to how long the observation time interval is.
3283 * The weight runs from 0 to 8. The maximum value of the
3284 * weight, 8, yields the minimum value for the smoothing
3285 * constant. At this minimum value for the smoothing constant,
3286 * the measured rate contributes for half of the next value of
3287 * the estimated peak rate.
3289 * So, the first step is to compute the weight as a function
3290 * of how sequential the workload is. Note that the weight
3291 * cannot reach 9, because bfqd->sequential_samples cannot
3292 * become equal to bfqd->peak_rate_samples, which, in its
3293 * turn, holds true because bfqd->sequential_samples is not
3294 * incremented for the first sample.
3296 weight = (9 * bfqd->sequential_samples) / bfqd->peak_rate_samples;
3299 * Second step: further refine the weight as a function of the
3300 * duration of the observation interval.
3302 weight = min_t(u32, 8,
3303 div_u64(weight * bfqd->delta_from_first,
3304 BFQ_RATE_REF_INTERVAL));
3307 * Divisor ranging from 10, for minimum weight, to 2, for
3310 divisor = 10 - weight;
3313 * Finally, update peak rate:
3315 * peak_rate = peak_rate * (divisor-1) / divisor + rate / divisor
3317 bfqd->peak_rate *= divisor-1;
3318 bfqd->peak_rate /= divisor;
3319 rate /= divisor; /* smoothing constant alpha = 1/divisor */
3321 bfqd->peak_rate += rate;
3324 * For a very slow device, bfqd->peak_rate can reach 0 (see
3325 * the minimum representable values reported in the comments
3326 * on BFQ_RATE_SHIFT). Push to 1 if this happens, to avoid
3327 * divisions by zero where bfqd->peak_rate is used as a
3330 bfqd->peak_rate = max_t(u32, 1, bfqd->peak_rate);
3332 update_thr_responsiveness_params(bfqd);
3335 bfq_reset_rate_computation(bfqd, rq);
3339 * Update the read/write peak rate (the main quantity used for
3340 * auto-tuning, see update_thr_responsiveness_params()).
3342 * It is not trivial to estimate the peak rate (correctly): because of
3343 * the presence of sw and hw queues between the scheduler and the
3344 * device components that finally serve I/O requests, it is hard to
3345 * say exactly when a given dispatched request is served inside the
3346 * device, and for how long. As a consequence, it is hard to know
3347 * precisely at what rate a given set of requests is actually served
3350 * On the opposite end, the dispatch time of any request is trivially
3351 * available, and, from this piece of information, the "dispatch rate"
3352 * of requests can be immediately computed. So, the idea in the next
3353 * function is to use what is known, namely request dispatch times
3354 * (plus, when useful, request completion times), to estimate what is
3355 * unknown, namely in-device request service rate.
3357 * The main issue is that, because of the above facts, the rate at
3358 * which a certain set of requests is dispatched over a certain time
3359 * interval can vary greatly with respect to the rate at which the
3360 * same requests are then served. But, since the size of any
3361 * intermediate queue is limited, and the service scheme is lossless
3362 * (no request is silently dropped), the following obvious convergence
3363 * property holds: the number of requests dispatched MUST become
3364 * closer and closer to the number of requests completed as the
3365 * observation interval grows. This is the key property used in
3366 * the next function to estimate the peak service rate as a function
3367 * of the observed dispatch rate. The function assumes to be invoked
3368 * on every request dispatch.
3370 static void bfq_update_peak_rate(struct bfq_data *bfqd, struct request *rq)
3372 u64 now_ns = ktime_get_ns();
3374 if (bfqd->peak_rate_samples == 0) { /* first dispatch */
3375 bfq_log(bfqd, "update_peak_rate: goto reset, samples %d",
3376 bfqd->peak_rate_samples);
3377 bfq_reset_rate_computation(bfqd, rq);
3378 goto update_last_values; /* will add one sample */
3382 * Device idle for very long: the observation interval lasting
3383 * up to this dispatch cannot be a valid observation interval
3384 * for computing a new peak rate (similarly to the late-
3385 * completion event in bfq_completed_request()). Go to
3386 * update_rate_and_reset to have the following three steps
3388 * - close the observation interval at the last (previous)
3389 * request dispatch or completion
3390 * - compute rate, if possible, for that observation interval
3391 * - start a new observation interval with this dispatch
3393 if (now_ns - bfqd->last_dispatch > 100*NSEC_PER_MSEC &&
3394 bfqd->rq_in_driver == 0)
3395 goto update_rate_and_reset;
3397 /* Update sampling information */
3398 bfqd->peak_rate_samples++;
3400 if ((bfqd->rq_in_driver > 0 ||
3401 now_ns - bfqd->last_completion < BFQ_MIN_TT)
3402 && !BFQ_RQ_SEEKY(bfqd, bfqd->last_position, rq))
3403 bfqd->sequential_samples++;
3405 bfqd->tot_sectors_dispatched += blk_rq_sectors(rq);
3407 /* Reset max observed rq size every 32 dispatches */
3408 if (likely(bfqd->peak_rate_samples % 32))
3409 bfqd->last_rq_max_size =
3410 max_t(u32, blk_rq_sectors(rq), bfqd->last_rq_max_size);
3412 bfqd->last_rq_max_size = blk_rq_sectors(rq);
3414 bfqd->delta_from_first = now_ns - bfqd->first_dispatch;
3416 /* Target observation interval not yet reached, go on sampling */
3417 if (bfqd->delta_from_first < BFQ_RATE_REF_INTERVAL)
3418 goto update_last_values;
3420 update_rate_and_reset:
3421 bfq_update_rate_reset(bfqd, rq);
3423 bfqd->last_position = blk_rq_pos(rq) + blk_rq_sectors(rq);
3424 if (RQ_BFQQ(rq) == bfqd->in_service_queue)
3425 bfqd->in_serv_last_pos = bfqd->last_position;
3426 bfqd->last_dispatch = now_ns;
3430 * Remove request from internal lists.
3432 static void bfq_dispatch_remove(struct request_queue *q, struct request *rq)
3434 struct bfq_queue *bfqq = RQ_BFQQ(rq);
3437 * For consistency, the next instruction should have been
3438 * executed after removing the request from the queue and
3439 * dispatching it. We execute instead this instruction before
3440 * bfq_remove_request() (and hence introduce a temporary
3441 * inconsistency), for efficiency. In fact, should this
3442 * dispatch occur for a non in-service bfqq, this anticipated
3443 * increment prevents two counters related to bfqq->dispatched
3444 * from risking to be, first, uselessly decremented, and then
3445 * incremented again when the (new) value of bfqq->dispatched
3446 * happens to be taken into account.
3449 bfq_update_peak_rate(q->elevator->elevator_data, rq);
3451 bfq_remove_request(q, rq);
3455 * There is a case where idling does not have to be performed for
3456 * throughput concerns, but to preserve the throughput share of
3457 * the process associated with bfqq.
3459 * To introduce this case, we can note that allowing the drive
3460 * to enqueue more than one request at a time, and hence
3461 * delegating de facto final scheduling decisions to the
3462 * drive's internal scheduler, entails loss of control on the
3463 * actual request service order. In particular, the critical
3464 * situation is when requests from different processes happen
3465 * to be present, at the same time, in the internal queue(s)
3466 * of the drive. In such a situation, the drive, by deciding
3467 * the service order of the internally-queued requests, does
3468 * determine also the actual throughput distribution among
3469 * these processes. But the drive typically has no notion or
3470 * concern about per-process throughput distribution, and
3471 * makes its decisions only on a per-request basis. Therefore,
3472 * the service distribution enforced by the drive's internal
3473 * scheduler is likely to coincide with the desired throughput
3474 * distribution only in a completely symmetric, or favorably
3475 * skewed scenario where:
3476 * (i-a) each of these processes must get the same throughput as
3478 * (i-b) in case (i-a) does not hold, it holds that the process
3479 * associated with bfqq must receive a lower or equal
3480 * throughput than any of the other processes;
3481 * (ii) the I/O of each process has the same properties, in
3482 * terms of locality (sequential or random), direction
3483 * (reads or writes), request sizes, greediness
3484 * (from I/O-bound to sporadic), and so on;
3486 * In fact, in such a scenario, the drive tends to treat the requests
3487 * of each process in about the same way as the requests of the
3488 * others, and thus to provide each of these processes with about the
3489 * same throughput. This is exactly the desired throughput
3490 * distribution if (i-a) holds, or, if (i-b) holds instead, this is an
3491 * even more convenient distribution for (the process associated with)
3494 * In contrast, in any asymmetric or unfavorable scenario, device
3495 * idling (I/O-dispatch plugging) is certainly needed to guarantee
3496 * that bfqq receives its assigned fraction of the device throughput
3497 * (see [1] for details).
3499 * The problem is that idling may significantly reduce throughput with
3500 * certain combinations of types of I/O and devices. An important
3501 * example is sync random I/O on flash storage with command
3502 * queueing. So, unless bfqq falls in cases where idling also boosts
3503 * throughput, it is important to check conditions (i-a), i(-b) and
3504 * (ii) accurately, so as to avoid idling when not strictly needed for
3505 * service guarantees.
3507 * Unfortunately, it is extremely difficult to thoroughly check
3508 * condition (ii). And, in case there are active groups, it becomes
3509 * very difficult to check conditions (i-a) and (i-b) too. In fact,
3510 * if there are active groups, then, for conditions (i-a) or (i-b) to
3511 * become false 'indirectly', it is enough that an active group
3512 * contains more active processes or sub-groups than some other active
3513 * group. More precisely, for conditions (i-a) or (i-b) to become
3514 * false because of such a group, it is not even necessary that the
3515 * group is (still) active: it is sufficient that, even if the group
3516 * has become inactive, some of its descendant processes still have
3517 * some request already dispatched but still waiting for
3518 * completion. In fact, requests have still to be guaranteed their
3519 * share of the throughput even after being dispatched. In this
3520 * respect, it is easy to show that, if a group frequently becomes
3521 * inactive while still having in-flight requests, and if, when this
3522 * happens, the group is not considered in the calculation of whether
3523 * the scenario is asymmetric, then the group may fail to be
3524 * guaranteed its fair share of the throughput (basically because
3525 * idling may not be performed for the descendant processes of the
3526 * group, but it had to be). We address this issue with the following
3527 * bi-modal behavior, implemented in the function
3528 * bfq_asymmetric_scenario().
3530 * If there are groups with requests waiting for completion
3531 * (as commented above, some of these groups may even be
3532 * already inactive), then the scenario is tagged as
3533 * asymmetric, conservatively, without checking any of the
3534 * conditions (i-a), (i-b) or (ii). So the device is idled for bfqq.
3535 * This behavior matches also the fact that groups are created
3536 * exactly if controlling I/O is a primary concern (to
3537 * preserve bandwidth and latency guarantees).
3539 * On the opposite end, if there are no groups with requests waiting
3540 * for completion, then only conditions (i-a) and (i-b) are actually
3541 * controlled, i.e., provided that conditions (i-a) or (i-b) holds,
3542 * idling is not performed, regardless of whether condition (ii)
3543 * holds. In other words, only if conditions (i-a) and (i-b) do not
3544 * hold, then idling is allowed, and the device tends to be prevented
3545 * from queueing many requests, possibly of several processes. Since
3546 * there are no groups with requests waiting for completion, then, to
3547 * control conditions (i-a) and (i-b) it is enough to check just
3548 * whether all the queues with requests waiting for completion also
3549 * have the same weight.
3551 * Not checking condition (ii) evidently exposes bfqq to the
3552 * risk of getting less throughput than its fair share.
3553 * However, for queues with the same weight, a further
3554 * mechanism, preemption, mitigates or even eliminates this
3555 * problem. And it does so without consequences on overall
3556 * throughput. This mechanism and its benefits are explained
3557 * in the next three paragraphs.
3559 * Even if a queue, say Q, is expired when it remains idle, Q
3560 * can still preempt the new in-service queue if the next
3561 * request of Q arrives soon (see the comments on
3562 * bfq_bfqq_update_budg_for_activation). If all queues and
3563 * groups have the same weight, this form of preemption,
3564 * combined with the hole-recovery heuristic described in the
3565 * comments on function bfq_bfqq_update_budg_for_activation,
3566 * are enough to preserve a correct bandwidth distribution in
3567 * the mid term, even without idling. In fact, even if not
3568 * idling allows the internal queues of the device to contain
3569 * many requests, and thus to reorder requests, we can rather
3570 * safely assume that the internal scheduler still preserves a
3571 * minimum of mid-term fairness.
3573 * More precisely, this preemption-based, idleless approach
3574 * provides fairness in terms of IOPS, and not sectors per
3575 * second. This can be seen with a simple example. Suppose
3576 * that there are two queues with the same weight, but that
3577 * the first queue receives requests of 8 sectors, while the
3578 * second queue receives requests of 1024 sectors. In
3579 * addition, suppose that each of the two queues contains at
3580 * most one request at a time, which implies that each queue
3581 * always remains idle after it is served. Finally, after
3582 * remaining idle, each queue receives very quickly a new
3583 * request. It follows that the two queues are served
3584 * alternatively, preempting each other if needed. This
3585 * implies that, although both queues have the same weight,
3586 * the queue with large requests receives a service that is
3587 * 1024/8 times as high as the service received by the other
3590 * The motivation for using preemption instead of idling (for
3591 * queues with the same weight) is that, by not idling,
3592 * service guarantees are preserved (completely or at least in
3593 * part) without minimally sacrificing throughput. And, if
3594 * there is no active group, then the primary expectation for
3595 * this device is probably a high throughput.
3597 * We are now left only with explaining the two sub-conditions in the
3598 * additional compound condition that is checked below for deciding
3599 * whether the scenario is asymmetric. To explain the first
3600 * sub-condition, we need to add that the function
3601 * bfq_asymmetric_scenario checks the weights of only
3602 * non-weight-raised queues, for efficiency reasons (see comments on
3603 * bfq_weights_tree_add()). Then the fact that bfqq is weight-raised
3604 * is checked explicitly here. More precisely, the compound condition
3605 * below takes into account also the fact that, even if bfqq is being
3606 * weight-raised, the scenario is still symmetric if all queues with
3607 * requests waiting for completion happen to be
3608 * weight-raised. Actually, we should be even more precise here, and
3609 * differentiate between interactive weight raising and soft real-time
3612 * The second sub-condition checked in the compound condition is
3613 * whether there is a fair amount of already in-flight I/O not
3614 * belonging to bfqq. If so, I/O dispatching is to be plugged, for the
3615 * following reason. The drive may decide to serve in-flight
3616 * non-bfqq's I/O requests before bfqq's ones, thereby delaying the
3617 * arrival of new I/O requests for bfqq (recall that bfqq is sync). If
3618 * I/O-dispatching is not plugged, then, while bfqq remains empty, a
3619 * basically uncontrolled amount of I/O from other queues may be
3620 * dispatched too, possibly causing the service of bfqq's I/O to be
3621 * delayed even longer in the drive. This problem gets more and more
3622 * serious as the speed and the queue depth of the drive grow,
3623 * because, as these two quantities grow, the probability to find no
3624 * queue busy but many requests in flight grows too. By contrast,
3625 * plugging I/O dispatching minimizes the delay induced by already
3626 * in-flight I/O, and enables bfqq to recover the bandwidth it may
3627 * lose because of this delay.
3629 * As a side note, it is worth considering that the above
3630 * device-idling countermeasures may however fail in the following
3631 * unlucky scenario: if I/O-dispatch plugging is (correctly) disabled
3632 * in a time period during which all symmetry sub-conditions hold, and
3633 * therefore the device is allowed to enqueue many requests, but at
3634 * some later point in time some sub-condition stops to hold, then it
3635 * may become impossible to make requests be served in the desired
3636 * order until all the requests already queued in the device have been
3637 * served. The last sub-condition commented above somewhat mitigates
3638 * this problem for weight-raised queues.
3640 * However, as an additional mitigation for this problem, we preserve
3641 * plugging for a special symmetric case that may suddenly turn into
3642 * asymmetric: the case where only bfqq is busy. In this case, not
3643 * expiring bfqq does not cause any harm to any other queues in terms
3644 * of service guarantees. In contrast, it avoids the following unlucky
3645 * sequence of events: (1) bfqq is expired, (2) a new queue with a
3646 * lower weight than bfqq becomes busy (or more queues), (3) the new
3647 * queue is served until a new request arrives for bfqq, (4) when bfqq
3648 * is finally served, there are so many requests of the new queue in
3649 * the drive that the pending requests for bfqq take a lot of time to
3650 * be served. In particular, event (2) may case even already
3651 * dispatched requests of bfqq to be delayed, inside the drive. So, to
3652 * avoid this series of events, the scenario is preventively declared
3653 * as asymmetric also if bfqq is the only busy queues
3655 static bool idling_needed_for_service_guarantees(struct bfq_data *bfqd,
3656 struct bfq_queue *bfqq)
3658 int tot_busy_queues = bfq_tot_busy_queues(bfqd);
3660 /* No point in idling for bfqq if it won't get requests any longer */
3661 if (unlikely(!bfqq_process_refs(bfqq)))
3664 return (bfqq->wr_coeff > 1 &&
3665 (bfqd->wr_busy_queues <
3667 bfqd->rq_in_driver >=
3668 bfqq->dispatched + 4)) ||
3669 bfq_asymmetric_scenario(bfqd, bfqq) ||
3670 tot_busy_queues == 1;
3673 static bool __bfq_bfqq_expire(struct bfq_data *bfqd, struct bfq_queue *bfqq,
3674 enum bfqq_expiration reason)
3677 * If this bfqq is shared between multiple processes, check
3678 * to make sure that those processes are still issuing I/Os
3679 * within the mean seek distance. If not, it may be time to
3680 * break the queues apart again.
3682 if (bfq_bfqq_coop(bfqq) && BFQQ_SEEKY(bfqq))
3683 bfq_mark_bfqq_split_coop(bfqq);
3686 * Consider queues with a higher finish virtual time than
3687 * bfqq. If idling_needed_for_service_guarantees(bfqq) returns
3688 * true, then bfqq's bandwidth would be violated if an
3689 * uncontrolled amount of I/O from these queues were
3690 * dispatched while bfqq is waiting for its new I/O to
3691 * arrive. This is exactly what may happen if this is a forced
3692 * expiration caused by a preemption attempt, and if bfqq is
3693 * not re-scheduled. To prevent this from happening, re-queue
3694 * bfqq if it needs I/O-dispatch plugging, even if it is
3695 * empty. By doing so, bfqq is granted to be served before the
3696 * above queues (provided that bfqq is of course eligible).
3698 if (RB_EMPTY_ROOT(&bfqq->sort_list) &&
3699 !(reason == BFQQE_PREEMPTED &&
3700 idling_needed_for_service_guarantees(bfqd, bfqq))) {
3701 if (bfqq->dispatched == 0)
3703 * Overloading budget_timeout field to store
3704 * the time at which the queue remains with no
3705 * backlog and no outstanding request; used by
3706 * the weight-raising mechanism.
3708 bfqq->budget_timeout = jiffies;
3710 bfq_del_bfqq_busy(bfqd, bfqq, true);
3712 bfq_requeue_bfqq(bfqd, bfqq, true);
3714 * Resort priority tree of potential close cooperators.
3715 * See comments on bfq_pos_tree_add_move() for the unlikely().
3717 if (unlikely(!bfqd->nonrot_with_queueing &&
3718 !RB_EMPTY_ROOT(&bfqq->sort_list)))
3719 bfq_pos_tree_add_move(bfqd, bfqq);
3723 * All in-service entities must have been properly deactivated
3724 * or requeued before executing the next function, which
3725 * resets all in-service entities as no more in service. This
3726 * may cause bfqq to be freed. If this happens, the next
3727 * function returns true.
3729 return __bfq_bfqd_reset_in_service(bfqd);
3733 * __bfq_bfqq_recalc_budget - try to adapt the budget to the @bfqq behavior.
3734 * @bfqd: device data.
3735 * @bfqq: queue to update.
3736 * @reason: reason for expiration.
3738 * Handle the feedback on @bfqq budget at queue expiration.
3739 * See the body for detailed comments.
3741 static void __bfq_bfqq_recalc_budget(struct bfq_data *bfqd,
3742 struct bfq_queue *bfqq,
3743 enum bfqq_expiration reason)
3745 struct request *next_rq;
3746 int budget, min_budget;
3748 min_budget = bfq_min_budget(bfqd);
3750 if (bfqq->wr_coeff == 1)
3751 budget = bfqq->max_budget;
3753 * Use a constant, low budget for weight-raised queues,
3754 * to help achieve a low latency. Keep it slightly higher
3755 * than the minimum possible budget, to cause a little
3756 * bit fewer expirations.
3758 budget = 2 * min_budget;
3760 bfq_log_bfqq(bfqd, bfqq, "recalc_budg: last budg %d, budg left %d",
3761 bfqq->entity.budget, bfq_bfqq_budget_left(bfqq));
3762 bfq_log_bfqq(bfqd, bfqq, "recalc_budg: last max_budg %d, min budg %d",
3763 budget, bfq_min_budget(bfqd));
3764 bfq_log_bfqq(bfqd, bfqq, "recalc_budg: sync %d, seeky %d",
3765 bfq_bfqq_sync(bfqq), BFQQ_SEEKY(bfqd->in_service_queue));
3767 if (bfq_bfqq_sync(bfqq) && bfqq->wr_coeff == 1) {
3770 * Caveat: in all the following cases we trade latency
3773 case BFQQE_TOO_IDLE:
3775 * This is the only case where we may reduce
3776 * the budget: if there is no request of the
3777 * process still waiting for completion, then
3778 * we assume (tentatively) that the timer has
3779 * expired because the batch of requests of
3780 * the process could have been served with a
3781 * smaller budget. Hence, betting that
3782 * process will behave in the same way when it
3783 * becomes backlogged again, we reduce its
3784 * next budget. As long as we guess right,
3785 * this budget cut reduces the latency
3786 * experienced by the process.
3788 * However, if there are still outstanding
3789 * requests, then the process may have not yet
3790 * issued its next request just because it is
3791 * still waiting for the completion of some of
3792 * the still outstanding ones. So in this
3793 * subcase we do not reduce its budget, on the
3794 * contrary we increase it to possibly boost
3795 * the throughput, as discussed in the
3796 * comments to the BUDGET_TIMEOUT case.
3798 if (bfqq->dispatched > 0) /* still outstanding reqs */
3799 budget = min(budget * 2, bfqd->bfq_max_budget);
3801 if (budget > 5 * min_budget)
3802 budget -= 4 * min_budget;
3804 budget = min_budget;
3807 case BFQQE_BUDGET_TIMEOUT:
3809 * We double the budget here because it gives
3810 * the chance to boost the throughput if this
3811 * is not a seeky process (and has bumped into
3812 * this timeout because of, e.g., ZBR).
3814 budget = min(budget * 2, bfqd->bfq_max_budget);
3816 case BFQQE_BUDGET_EXHAUSTED:
3818 * The process still has backlog, and did not
3819 * let either the budget timeout or the disk
3820 * idling timeout expire. Hence it is not
3821 * seeky, has a short thinktime and may be
3822 * happy with a higher budget too. So
3823 * definitely increase the budget of this good
3824 * candidate to boost the disk throughput.
3826 budget = min(budget * 4, bfqd->bfq_max_budget);
3828 case BFQQE_NO_MORE_REQUESTS:
3830 * For queues that expire for this reason, it
3831 * is particularly important to keep the
3832 * budget close to the actual service they
3833 * need. Doing so reduces the timestamp
3834 * misalignment problem described in the
3835 * comments in the body of
3836 * __bfq_activate_entity. In fact, suppose
3837 * that a queue systematically expires for
3838 * BFQQE_NO_MORE_REQUESTS and presents a
3839 * new request in time to enjoy timestamp
3840 * back-shifting. The larger the budget of the
3841 * queue is with respect to the service the
3842 * queue actually requests in each service
3843 * slot, the more times the queue can be
3844 * reactivated with the same virtual finish
3845 * time. It follows that, even if this finish
3846 * time is pushed to the system virtual time
3847 * to reduce the consequent timestamp
3848 * misalignment, the queue unjustly enjoys for
3849 * many re-activations a lower finish time
3850 * than all newly activated queues.
3852 * The service needed by bfqq is measured
3853 * quite precisely by bfqq->entity.service.
3854 * Since bfqq does not enjoy device idling,
3855 * bfqq->entity.service is equal to the number
3856 * of sectors that the process associated with
3857 * bfqq requested to read/write before waiting
3858 * for request completions, or blocking for
3861 budget = max_t(int, bfqq->entity.service, min_budget);
3866 } else if (!bfq_bfqq_sync(bfqq)) {
3868 * Async queues get always the maximum possible
3869 * budget, as for them we do not care about latency
3870 * (in addition, their ability to dispatch is limited
3871 * by the charging factor).
3873 budget = bfqd->bfq_max_budget;
3876 bfqq->max_budget = budget;
3878 if (bfqd->budgets_assigned >= bfq_stats_min_budgets &&
3879 !bfqd->bfq_user_max_budget)
3880 bfqq->max_budget = min(bfqq->max_budget, bfqd->bfq_max_budget);
3883 * If there is still backlog, then assign a new budget, making
3884 * sure that it is large enough for the next request. Since
3885 * the finish time of bfqq must be kept in sync with the
3886 * budget, be sure to call __bfq_bfqq_expire() *after* this
3889 * If there is no backlog, then no need to update the budget;
3890 * it will be updated on the arrival of a new request.
3892 next_rq = bfqq->next_rq;
3894 bfqq->entity.budget = max_t(unsigned long, bfqq->max_budget,
3895 bfq_serv_to_charge(next_rq, bfqq));
3897 bfq_log_bfqq(bfqd, bfqq, "head sect: %u, new budget %d",
3898 next_rq ? blk_rq_sectors(next_rq) : 0,
3899 bfqq->entity.budget);
3903 * Return true if the process associated with bfqq is "slow". The slow
3904 * flag is used, in addition to the budget timeout, to reduce the
3905 * amount of service provided to seeky processes, and thus reduce
3906 * their chances to lower the throughput. More details in the comments
3907 * on the function bfq_bfqq_expire().
3909 * An important observation is in order: as discussed in the comments
3910 * on the function bfq_update_peak_rate(), with devices with internal
3911 * queues, it is hard if ever possible to know when and for how long
3912 * an I/O request is processed by the device (apart from the trivial
3913 * I/O pattern where a new request is dispatched only after the
3914 * previous one has been completed). This makes it hard to evaluate
3915 * the real rate at which the I/O requests of each bfq_queue are
3916 * served. In fact, for an I/O scheduler like BFQ, serving a
3917 * bfq_queue means just dispatching its requests during its service
3918 * slot (i.e., until the budget of the queue is exhausted, or the
3919 * queue remains idle, or, finally, a timeout fires). But, during the
3920 * service slot of a bfq_queue, around 100 ms at most, the device may
3921 * be even still processing requests of bfq_queues served in previous
3922 * service slots. On the opposite end, the requests of the in-service
3923 * bfq_queue may be completed after the service slot of the queue
3926 * Anyway, unless more sophisticated solutions are used
3927 * (where possible), the sum of the sizes of the requests dispatched
3928 * during the service slot of a bfq_queue is probably the only
3929 * approximation available for the service received by the bfq_queue
3930 * during its service slot. And this sum is the quantity used in this
3931 * function to evaluate the I/O speed of a process.
3933 static bool bfq_bfqq_is_slow(struct bfq_data *bfqd, struct bfq_queue *bfqq,
3934 bool compensate, enum bfqq_expiration reason,
3935 unsigned long *delta_ms)
3937 ktime_t delta_ktime;
3939 bool slow = BFQQ_SEEKY(bfqq); /* if delta too short, use seekyness */
3941 if (!bfq_bfqq_sync(bfqq))
3945 delta_ktime = bfqd->last_idling_start;
3947 delta_ktime = ktime_get();
3948 delta_ktime = ktime_sub(delta_ktime, bfqd->last_budget_start);
3949 delta_usecs = ktime_to_us(delta_ktime);
3951 /* don't use too short time intervals */
3952 if (delta_usecs < 1000) {
3953 if (blk_queue_nonrot(bfqd->queue))
3955 * give same worst-case guarantees as idling
3958 *delta_ms = BFQ_MIN_TT / NSEC_PER_MSEC;
3959 else /* charge at least one seek */
3960 *delta_ms = bfq_slice_idle / NSEC_PER_MSEC;
3965 *delta_ms = delta_usecs / USEC_PER_MSEC;
3968 * Use only long (> 20ms) intervals to filter out excessive
3969 * spikes in service rate estimation.
3971 if (delta_usecs > 20000) {
3973 * Caveat for rotational devices: processes doing I/O
3974 * in the slower disk zones tend to be slow(er) even
3975 * if not seeky. In this respect, the estimated peak
3976 * rate is likely to be an average over the disk
3977 * surface. Accordingly, to not be too harsh with
3978 * unlucky processes, a process is deemed slow only if
3979 * its rate has been lower than half of the estimated
3982 slow = bfqq->entity.service < bfqd->bfq_max_budget / 2;
3985 bfq_log_bfqq(bfqd, bfqq, "bfq_bfqq_is_slow: slow %d", slow);
3991 * To be deemed as soft real-time, an application must meet two
3992 * requirements. First, the application must not require an average
3993 * bandwidth higher than the approximate bandwidth required to playback or
3994 * record a compressed high-definition video.
3995 * The next function is invoked on the completion of the last request of a
3996 * batch, to compute the next-start time instant, soft_rt_next_start, such
3997 * that, if the next request of the application does not arrive before
3998 * soft_rt_next_start, then the above requirement on the bandwidth is met.
4000 * The second requirement is that the request pattern of the application is
4001 * isochronous, i.e., that, after issuing a request or a batch of requests,
4002 * the application stops issuing new requests until all its pending requests
4003 * have been completed. After that, the application may issue a new batch,
4005 * For this reason the next function is invoked to compute
4006 * soft_rt_next_start only for applications that meet this requirement,
4007 * whereas soft_rt_next_start is set to infinity for applications that do
4010 * Unfortunately, even a greedy (i.e., I/O-bound) application may
4011 * happen to meet, occasionally or systematically, both the above
4012 * bandwidth and isochrony requirements. This may happen at least in
4013 * the following circumstances. First, if the CPU load is high. The
4014 * application may stop issuing requests while the CPUs are busy
4015 * serving other processes, then restart, then stop again for a while,
4016 * and so on. The other circumstances are related to the storage
4017 * device: the storage device is highly loaded or reaches a low-enough
4018 * throughput with the I/O of the application (e.g., because the I/O
4019 * is random and/or the device is slow). In all these cases, the
4020 * I/O of the application may be simply slowed down enough to meet
4021 * the bandwidth and isochrony requirements. To reduce the probability
4022 * that greedy applications are deemed as soft real-time in these
4023 * corner cases, a further rule is used in the computation of
4024 * soft_rt_next_start: the return value of this function is forced to
4025 * be higher than the maximum between the following two quantities.
4027 * (a) Current time plus: (1) the maximum time for which the arrival
4028 * of a request is waited for when a sync queue becomes idle,
4029 * namely bfqd->bfq_slice_idle, and (2) a few extra jiffies. We
4030 * postpone for a moment the reason for adding a few extra
4031 * jiffies; we get back to it after next item (b). Lower-bounding
4032 * the return value of this function with the current time plus
4033 * bfqd->bfq_slice_idle tends to filter out greedy applications,
4034 * because the latter issue their next request as soon as possible
4035 * after the last one has been completed. In contrast, a soft
4036 * real-time application spends some time processing data, after a
4037 * batch of its requests has been completed.
4039 * (b) Current value of bfqq->soft_rt_next_start. As pointed out
4040 * above, greedy applications may happen to meet both the
4041 * bandwidth and isochrony requirements under heavy CPU or
4042 * storage-device load. In more detail, in these scenarios, these
4043 * applications happen, only for limited time periods, to do I/O
4044 * slowly enough to meet all the requirements described so far,
4045 * including the filtering in above item (a). These slow-speed
4046 * time intervals are usually interspersed between other time
4047 * intervals during which these applications do I/O at a very high
4048 * speed. Fortunately, exactly because of the high speed of the
4049 * I/O in the high-speed intervals, the values returned by this
4050 * function happen to be so high, near the end of any such
4051 * high-speed interval, to be likely to fall *after* the end of
4052 * the low-speed time interval that follows. These high values are
4053 * stored in bfqq->soft_rt_next_start after each invocation of
4054 * this function. As a consequence, if the last value of
4055 * bfqq->soft_rt_next_start is constantly used to lower-bound the
4056 * next value that this function may return, then, from the very
4057 * beginning of a low-speed interval, bfqq->soft_rt_next_start is
4058 * likely to be constantly kept so high that any I/O request
4059 * issued during the low-speed interval is considered as arriving
4060 * to soon for the application to be deemed as soft
4061 * real-time. Then, in the high-speed interval that follows, the
4062 * application will not be deemed as soft real-time, just because
4063 * it will do I/O at a high speed. And so on.
4065 * Getting back to the filtering in item (a), in the following two
4066 * cases this filtering might be easily passed by a greedy
4067 * application, if the reference quantity was just
4068 * bfqd->bfq_slice_idle:
4069 * 1) HZ is so low that the duration of a jiffy is comparable to or
4070 * higher than bfqd->bfq_slice_idle. This happens, e.g., on slow
4071 * devices with HZ=100. The time granularity may be so coarse
4072 * that the approximation, in jiffies, of bfqd->bfq_slice_idle
4073 * is rather lower than the exact value.
4074 * 2) jiffies, instead of increasing at a constant rate, may stop increasing
4075 * for a while, then suddenly 'jump' by several units to recover the lost
4076 * increments. This seems to happen, e.g., inside virtual machines.
4077 * To address this issue, in the filtering in (a) we do not use as a
4078 * reference time interval just bfqd->bfq_slice_idle, but
4079 * bfqd->bfq_slice_idle plus a few jiffies. In particular, we add the
4080 * minimum number of jiffies for which the filter seems to be quite
4081 * precise also in embedded systems and KVM/QEMU virtual machines.
4083 static unsigned long bfq_bfqq_softrt_next_start(struct bfq_data *bfqd,
4084 struct bfq_queue *bfqq)
4086 return max3(bfqq->soft_rt_next_start,
4087 bfqq->last_idle_bklogged +
4088 HZ * bfqq->service_from_backlogged /
4089 bfqd->bfq_wr_max_softrt_rate,
4090 jiffies + nsecs_to_jiffies(bfqq->bfqd->bfq_slice_idle) + 4);
4094 * bfq_bfqq_expire - expire a queue.
4095 * @bfqd: device owning the queue.
4096 * @bfqq: the queue to expire.
4097 * @compensate: if true, compensate for the time spent idling.
4098 * @reason: the reason causing the expiration.
4100 * If the process associated with bfqq does slow I/O (e.g., because it
4101 * issues random requests), we charge bfqq with the time it has been
4102 * in service instead of the service it has received (see
4103 * bfq_bfqq_charge_time for details on how this goal is achieved). As
4104 * a consequence, bfqq will typically get higher timestamps upon
4105 * reactivation, and hence it will be rescheduled as if it had
4106 * received more service than what it has actually received. In the
4107 * end, bfqq receives less service in proportion to how slowly its
4108 * associated process consumes its budgets (and hence how seriously it
4109 * tends to lower the throughput). In addition, this time-charging
4110 * strategy guarantees time fairness among slow processes. In
4111 * contrast, if the process associated with bfqq is not slow, we
4112 * charge bfqq exactly with the service it has received.
4114 * Charging time to the first type of queues and the exact service to
4115 * the other has the effect of using the WF2Q+ policy to schedule the
4116 * former on a timeslice basis, without violating service domain
4117 * guarantees among the latter.
4119 void bfq_bfqq_expire(struct bfq_data *bfqd,
4120 struct bfq_queue *bfqq,
4122 enum bfqq_expiration reason)
4125 unsigned long delta = 0;
4126 struct bfq_entity *entity = &bfqq->entity;
4129 * Check whether the process is slow (see bfq_bfqq_is_slow).
4131 slow = bfq_bfqq_is_slow(bfqd, bfqq, compensate, reason, &delta);
4134 * As above explained, charge slow (typically seeky) and
4135 * timed-out queues with the time and not the service
4136 * received, to favor sequential workloads.
4138 * Processes doing I/O in the slower disk zones will tend to
4139 * be slow(er) even if not seeky. Therefore, since the
4140 * estimated peak rate is actually an average over the disk
4141 * surface, these processes may timeout just for bad luck. To
4142 * avoid punishing them, do not charge time to processes that
4143 * succeeded in consuming at least 2/3 of their budget. This
4144 * allows BFQ to preserve enough elasticity to still perform
4145 * bandwidth, and not time, distribution with little unlucky
4146 * or quasi-sequential processes.
4148 if (bfqq->wr_coeff == 1 &&
4150 (reason == BFQQE_BUDGET_TIMEOUT &&
4151 bfq_bfqq_budget_left(bfqq) >= entity->budget / 3)))
4152 bfq_bfqq_charge_time(bfqd, bfqq, delta);
4154 if (bfqd->low_latency && bfqq->wr_coeff == 1)
4155 bfqq->last_wr_start_finish = jiffies;
4157 if (bfqd->low_latency && bfqd->bfq_wr_max_softrt_rate > 0 &&
4158 RB_EMPTY_ROOT(&bfqq->sort_list)) {
4160 * If we get here, and there are no outstanding
4161 * requests, then the request pattern is isochronous
4162 * (see the comments on the function
4163 * bfq_bfqq_softrt_next_start()). Therefore we can
4164 * compute soft_rt_next_start.
4166 * If, instead, the queue still has outstanding
4167 * requests, then we have to wait for the completion
4168 * of all the outstanding requests to discover whether
4169 * the request pattern is actually isochronous.
4171 if (bfqq->dispatched == 0)
4172 bfqq->soft_rt_next_start =
4173 bfq_bfqq_softrt_next_start(bfqd, bfqq);
4174 else if (bfqq->dispatched > 0) {
4176 * Schedule an update of soft_rt_next_start to when
4177 * the task may be discovered to be isochronous.
4179 bfq_mark_bfqq_softrt_update(bfqq);
4183 bfq_log_bfqq(bfqd, bfqq,
4184 "expire (%d, slow %d, num_disp %d, short_ttime %d)", reason,
4185 slow, bfqq->dispatched, bfq_bfqq_has_short_ttime(bfqq));
4188 * bfqq expired, so no total service time needs to be computed
4189 * any longer: reset state machine for measuring total service
4192 bfqd->rqs_injected = bfqd->wait_dispatch = false;
4193 bfqd->waited_rq = NULL;
4196 * Increase, decrease or leave budget unchanged according to
4199 __bfq_bfqq_recalc_budget(bfqd, bfqq, reason);
4200 if (__bfq_bfqq_expire(bfqd, bfqq, reason))
4201 /* bfqq is gone, no more actions on it */
4204 /* mark bfqq as waiting a request only if a bic still points to it */
4205 if (!bfq_bfqq_busy(bfqq) &&
4206 reason != BFQQE_BUDGET_TIMEOUT &&
4207 reason != BFQQE_BUDGET_EXHAUSTED) {
4208 bfq_mark_bfqq_non_blocking_wait_rq(bfqq);
4210 * Not setting service to 0, because, if the next rq
4211 * arrives in time, the queue will go on receiving
4212 * service with this same budget (as if it never expired)
4215 entity->service = 0;
4218 * Reset the received-service counter for every parent entity.
4219 * Differently from what happens with bfqq->entity.service,
4220 * the resetting of this counter never needs to be postponed
4221 * for parent entities. In fact, in case bfqq may have a
4222 * chance to go on being served using the last, partially
4223 * consumed budget, bfqq->entity.service needs to be kept,
4224 * because if bfqq then actually goes on being served using
4225 * the same budget, the last value of bfqq->entity.service is
4226 * needed to properly decrement bfqq->entity.budget by the
4227 * portion already consumed. In contrast, it is not necessary
4228 * to keep entity->service for parent entities too, because
4229 * the bubble up of the new value of bfqq->entity.budget will
4230 * make sure that the budgets of parent entities are correct,
4231 * even in case bfqq and thus parent entities go on receiving
4232 * service with the same budget.
4234 entity = entity->parent;
4235 for_each_entity(entity)
4236 entity->service = 0;
4240 * Budget timeout is not implemented through a dedicated timer, but
4241 * just checked on request arrivals and completions, as well as on
4242 * idle timer expirations.
4244 static bool bfq_bfqq_budget_timeout(struct bfq_queue *bfqq)
4246 return time_is_before_eq_jiffies(bfqq->budget_timeout);
4250 * If we expire a queue that is actively waiting (i.e., with the
4251 * device idled) for the arrival of a new request, then we may incur
4252 * the timestamp misalignment problem described in the body of the
4253 * function __bfq_activate_entity. Hence we return true only if this
4254 * condition does not hold, or if the queue is slow enough to deserve
4255 * only to be kicked off for preserving a high throughput.
4257 static bool bfq_may_expire_for_budg_timeout(struct bfq_queue *bfqq)
4259 bfq_log_bfqq(bfqq->bfqd, bfqq,
4260 "may_budget_timeout: wait_request %d left %d timeout %d",
4261 bfq_bfqq_wait_request(bfqq),
4262 bfq_bfqq_budget_left(bfqq) >= bfqq->entity.budget / 3,
4263 bfq_bfqq_budget_timeout(bfqq));
4265 return (!bfq_bfqq_wait_request(bfqq) ||
4266 bfq_bfqq_budget_left(bfqq) >= bfqq->entity.budget / 3)
4268 bfq_bfqq_budget_timeout(bfqq);
4271 static bool idling_boosts_thr_without_issues(struct bfq_data *bfqd,
4272 struct bfq_queue *bfqq)
4274 bool rot_without_queueing =
4275 !blk_queue_nonrot(bfqd->queue) && !bfqd->hw_tag,
4276 bfqq_sequential_and_IO_bound,
4279 /* No point in idling for bfqq if it won't get requests any longer */
4280 if (unlikely(!bfqq_process_refs(bfqq)))
4283 bfqq_sequential_and_IO_bound = !BFQQ_SEEKY(bfqq) &&
4284 bfq_bfqq_IO_bound(bfqq) && bfq_bfqq_has_short_ttime(bfqq);
4287 * The next variable takes into account the cases where idling
4288 * boosts the throughput.
4290 * The value of the variable is computed considering, first, that
4291 * idling is virtually always beneficial for the throughput if:
4292 * (a) the device is not NCQ-capable and rotational, or
4293 * (b) regardless of the presence of NCQ, the device is rotational and
4294 * the request pattern for bfqq is I/O-bound and sequential, or
4295 * (c) regardless of whether it is rotational, the device is
4296 * not NCQ-capable and the request pattern for bfqq is
4297 * I/O-bound and sequential.
4299 * Secondly, and in contrast to the above item (b), idling an
4300 * NCQ-capable flash-based device would not boost the
4301 * throughput even with sequential I/O; rather it would lower
4302 * the throughput in proportion to how fast the device
4303 * is. Accordingly, the next variable is true if any of the
4304 * above conditions (a), (b) or (c) is true, and, in
4305 * particular, happens to be false if bfqd is an NCQ-capable
4306 * flash-based device.
4308 idling_boosts_thr = rot_without_queueing ||
4309 ((!blk_queue_nonrot(bfqd->queue) || !bfqd->hw_tag) &&
4310 bfqq_sequential_and_IO_bound);
4313 * The return value of this function is equal to that of
4314 * idling_boosts_thr, unless a special case holds. In this
4315 * special case, described below, idling may cause problems to
4316 * weight-raised queues.
4318 * When the request pool is saturated (e.g., in the presence
4319 * of write hogs), if the processes associated with
4320 * non-weight-raised queues ask for requests at a lower rate,
4321 * then processes associated with weight-raised queues have a
4322 * higher probability to get a request from the pool
4323 * immediately (or at least soon) when they need one. Thus
4324 * they have a higher probability to actually get a fraction
4325 * of the device throughput proportional to their high
4326 * weight. This is especially true with NCQ-capable drives,
4327 * which enqueue several requests in advance, and further
4328 * reorder internally-queued requests.
4330 * For this reason, we force to false the return value if
4331 * there are weight-raised busy queues. In this case, and if
4332 * bfqq is not weight-raised, this guarantees that the device
4333 * is not idled for bfqq (if, instead, bfqq is weight-raised,
4334 * then idling will be guaranteed by another variable, see
4335 * below). Combined with the timestamping rules of BFQ (see
4336 * [1] for details), this behavior causes bfqq, and hence any
4337 * sync non-weight-raised queue, to get a lower number of
4338 * requests served, and thus to ask for a lower number of
4339 * requests from the request pool, before the busy
4340 * weight-raised queues get served again. This often mitigates
4341 * starvation problems in the presence of heavy write
4342 * workloads and NCQ, thereby guaranteeing a higher
4343 * application and system responsiveness in these hostile
4346 return idling_boosts_thr &&
4347 bfqd->wr_busy_queues == 0;
4351 * For a queue that becomes empty, device idling is allowed only if
4352 * this function returns true for that queue. As a consequence, since
4353 * device idling plays a critical role for both throughput boosting
4354 * and service guarantees, the return value of this function plays a
4355 * critical role as well.
4357 * In a nutshell, this function returns true only if idling is
4358 * beneficial for throughput or, even if detrimental for throughput,
4359 * idling is however necessary to preserve service guarantees (low
4360 * latency, desired throughput distribution, ...). In particular, on
4361 * NCQ-capable devices, this function tries to return false, so as to
4362 * help keep the drives' internal queues full, whenever this helps the
4363 * device boost the throughput without causing any service-guarantee
4366 * Most of the issues taken into account to get the return value of
4367 * this function are not trivial. We discuss these issues in the two
4368 * functions providing the main pieces of information needed by this
4371 static bool bfq_better_to_idle(struct bfq_queue *bfqq)
4373 struct bfq_data *bfqd = bfqq->bfqd;
4374 bool idling_boosts_thr_with_no_issue, idling_needed_for_service_guar;
4376 /* No point in idling for bfqq if it won't get requests any longer */
4377 if (unlikely(!bfqq_process_refs(bfqq)))
4380 if (unlikely(bfqd->strict_guarantees))
4384 * Idling is performed only if slice_idle > 0. In addition, we
4387 * (b) bfqq is in the idle io prio class: in this case we do
4388 * not idle because we want to minimize the bandwidth that
4389 * queues in this class can steal to higher-priority queues
4391 if (bfqd->bfq_slice_idle == 0 || !bfq_bfqq_sync(bfqq) ||
4392 bfq_class_idle(bfqq))
4395 idling_boosts_thr_with_no_issue =
4396 idling_boosts_thr_without_issues(bfqd, bfqq);
4398 idling_needed_for_service_guar =
4399 idling_needed_for_service_guarantees(bfqd, bfqq);
4402 * We have now the two components we need to compute the
4403 * return value of the function, which is true only if idling
4404 * either boosts the throughput (without issues), or is
4405 * necessary to preserve service guarantees.
4407 return idling_boosts_thr_with_no_issue ||
4408 idling_needed_for_service_guar;
4412 * If the in-service queue is empty but the function bfq_better_to_idle
4413 * returns true, then:
4414 * 1) the queue must remain in service and cannot be expired, and
4415 * 2) the device must be idled to wait for the possible arrival of a new
4416 * request for the queue.
4417 * See the comments on the function bfq_better_to_idle for the reasons
4418 * why performing device idling is the best choice to boost the throughput
4419 * and preserve service guarantees when bfq_better_to_idle itself
4422 static bool bfq_bfqq_must_idle(struct bfq_queue *bfqq)
4424 return RB_EMPTY_ROOT(&bfqq->sort_list) && bfq_better_to_idle(bfqq);
4428 * This function chooses the queue from which to pick the next extra
4429 * I/O request to inject, if it finds a compatible queue. See the
4430 * comments on bfq_update_inject_limit() for details on the injection
4431 * mechanism, and for the definitions of the quantities mentioned
4434 static struct bfq_queue *
4435 bfq_choose_bfqq_for_injection(struct bfq_data *bfqd)
4437 struct bfq_queue *bfqq, *in_serv_bfqq = bfqd->in_service_queue;
4438 unsigned int limit = in_serv_bfqq->inject_limit;
4441 * - bfqq is not weight-raised and therefore does not carry
4442 * time-critical I/O,
4444 * - regardless of whether bfqq is weight-raised, bfqq has
4445 * however a long think time, during which it can absorb the
4446 * effect of an appropriate number of extra I/O requests
4447 * from other queues (see bfq_update_inject_limit for
4448 * details on the computation of this number);
4449 * then injection can be performed without restrictions.
4451 bool in_serv_always_inject = in_serv_bfqq->wr_coeff == 1 ||
4452 !bfq_bfqq_has_short_ttime(in_serv_bfqq);
4456 * - the baseline total service time could not be sampled yet,
4457 * so the inject limit happens to be still 0, and
4458 * - a lot of time has elapsed since the plugging of I/O
4459 * dispatching started, so drive speed is being wasted
4461 * then temporarily raise inject limit to one request.
4463 if (limit == 0 && in_serv_bfqq->last_serv_time_ns == 0 &&
4464 bfq_bfqq_wait_request(in_serv_bfqq) &&
4465 time_is_before_eq_jiffies(bfqd->last_idling_start_jiffies +
4466 bfqd->bfq_slice_idle)
4470 if (bfqd->rq_in_driver >= limit)
4474 * Linear search of the source queue for injection; but, with
4475 * a high probability, very few steps are needed to find a
4476 * candidate queue, i.e., a queue with enough budget left for
4477 * its next request. In fact:
4478 * - BFQ dynamically updates the budget of every queue so as
4479 * to accommodate the expected backlog of the queue;
4480 * - if a queue gets all its requests dispatched as injected
4481 * service, then the queue is removed from the active list
4482 * (and re-added only if it gets new requests, but then it
4483 * is assigned again enough budget for its new backlog).
4485 list_for_each_entry(bfqq, &bfqd->active_list, bfqq_list)
4486 if (!RB_EMPTY_ROOT(&bfqq->sort_list) &&
4487 (in_serv_always_inject || bfqq->wr_coeff > 1) &&
4488 bfq_serv_to_charge(bfqq->next_rq, bfqq) <=
4489 bfq_bfqq_budget_left(bfqq)) {
4491 * Allow for only one large in-flight request
4492 * on non-rotational devices, for the
4493 * following reason. On non-rotationl drives,
4494 * large requests take much longer than
4495 * smaller requests to be served. In addition,
4496 * the drive prefers to serve large requests
4497 * w.r.t. to small ones, if it can choose. So,
4498 * having more than one large requests queued
4499 * in the drive may easily make the next first
4500 * request of the in-service queue wait for so
4501 * long to break bfqq's service guarantees. On
4502 * the bright side, large requests let the
4503 * drive reach a very high throughput, even if
4504 * there is only one in-flight large request
4507 if (blk_queue_nonrot(bfqd->queue) &&
4508 blk_rq_sectors(bfqq->next_rq) >=
4509 BFQQ_SECT_THR_NONROT)
4510 limit = min_t(unsigned int, 1, limit);
4512 limit = in_serv_bfqq->inject_limit;
4514 if (bfqd->rq_in_driver < limit) {
4515 bfqd->rqs_injected = true;
4524 * Select a queue for service. If we have a current queue in service,
4525 * check whether to continue servicing it, or retrieve and set a new one.
4527 static struct bfq_queue *bfq_select_queue(struct bfq_data *bfqd)
4529 struct bfq_queue *bfqq;
4530 struct request *next_rq;
4531 enum bfqq_expiration reason = BFQQE_BUDGET_TIMEOUT;
4533 bfqq = bfqd->in_service_queue;
4537 bfq_log_bfqq(bfqd, bfqq, "select_queue: already in-service queue");
4540 * Do not expire bfqq for budget timeout if bfqq may be about
4541 * to enjoy device idling. The reason why, in this case, we
4542 * prevent bfqq from expiring is the same as in the comments
4543 * on the case where bfq_bfqq_must_idle() returns true, in
4544 * bfq_completed_request().
4546 if (bfq_may_expire_for_budg_timeout(bfqq) &&
4547 !bfq_bfqq_must_idle(bfqq))
4552 * This loop is rarely executed more than once. Even when it
4553 * happens, it is much more convenient to re-execute this loop
4554 * than to return NULL and trigger a new dispatch to get a
4557 next_rq = bfqq->next_rq;
4559 * If bfqq has requests queued and it has enough budget left to
4560 * serve them, keep the queue, otherwise expire it.
4563 if (bfq_serv_to_charge(next_rq, bfqq) >
4564 bfq_bfqq_budget_left(bfqq)) {
4566 * Expire the queue for budget exhaustion,
4567 * which makes sure that the next budget is
4568 * enough to serve the next request, even if
4569 * it comes from the fifo expired path.
4571 reason = BFQQE_BUDGET_EXHAUSTED;
4575 * The idle timer may be pending because we may
4576 * not disable disk idling even when a new request
4579 if (bfq_bfqq_wait_request(bfqq)) {
4581 * If we get here: 1) at least a new request
4582 * has arrived but we have not disabled the
4583 * timer because the request was too small,
4584 * 2) then the block layer has unplugged
4585 * the device, causing the dispatch to be
4588 * Since the device is unplugged, now the
4589 * requests are probably large enough to
4590 * provide a reasonable throughput.
4591 * So we disable idling.
4593 bfq_clear_bfqq_wait_request(bfqq);
4594 hrtimer_try_to_cancel(&bfqd->idle_slice_timer);
4601 * No requests pending. However, if the in-service queue is idling
4602 * for a new request, or has requests waiting for a completion and
4603 * may idle after their completion, then keep it anyway.
4605 * Yet, inject service from other queues if it boosts
4606 * throughput and is possible.
4608 if (bfq_bfqq_wait_request(bfqq) ||
4609 (bfqq->dispatched != 0 && bfq_better_to_idle(bfqq))) {
4610 struct bfq_queue *async_bfqq =
4611 bfqq->bic && bfqq->bic->bfqq[0] &&
4612 bfq_bfqq_busy(bfqq->bic->bfqq[0]) &&
4613 bfqq->bic->bfqq[0]->next_rq ?
4614 bfqq->bic->bfqq[0] : NULL;
4615 struct bfq_queue *blocked_bfqq =
4616 !hlist_empty(&bfqq->woken_list) ?
4617 container_of(bfqq->woken_list.first,
4623 * The next four mutually-exclusive ifs decide
4624 * whether to try injection, and choose the queue to
4625 * pick an I/O request from.
4627 * The first if checks whether the process associated
4628 * with bfqq has also async I/O pending. If so, it
4629 * injects such I/O unconditionally. Injecting async
4630 * I/O from the same process can cause no harm to the
4631 * process. On the contrary, it can only increase
4632 * bandwidth and reduce latency for the process.
4634 * The second if checks whether there happens to be a
4635 * non-empty waker queue for bfqq, i.e., a queue whose
4636 * I/O needs to be completed for bfqq to receive new
4637 * I/O. This happens, e.g., if bfqq is associated with
4638 * a process that does some sync. A sync generates
4639 * extra blocking I/O, which must be completed before
4640 * the process associated with bfqq can go on with its
4641 * I/O. If the I/O of the waker queue is not served,
4642 * then bfqq remains empty, and no I/O is dispatched,
4643 * until the idle timeout fires for bfqq. This is
4644 * likely to result in lower bandwidth and higher
4645 * latencies for bfqq, and in a severe loss of total
4646 * throughput. The best action to take is therefore to
4647 * serve the waker queue as soon as possible. So do it
4648 * (without relying on the third alternative below for
4649 * eventually serving waker_bfqq's I/O; see the last
4650 * paragraph for further details). This systematic
4651 * injection of I/O from the waker queue does not
4652 * cause any delay to bfqq's I/O. On the contrary,
4653 * next bfqq's I/O is brought forward dramatically,
4654 * for it is not blocked for milliseconds.
4656 * The third if checks whether there is a queue woken
4657 * by bfqq, and currently with pending I/O. Such a
4658 * woken queue does not steal bandwidth from bfqq,
4659 * because it remains soon without I/O if bfqq is not
4660 * served. So there is virtually no risk of loss of
4661 * bandwidth for bfqq if this woken queue has I/O
4662 * dispatched while bfqq is waiting for new I/O.
4664 * The fourth if checks whether bfqq is a queue for
4665 * which it is better to avoid injection. It is so if
4666 * bfqq delivers more throughput when served without
4667 * any further I/O from other queues in the middle, or
4668 * if the service times of bfqq's I/O requests both
4669 * count more than overall throughput, and may be
4670 * easily increased by injection (this happens if bfqq
4671 * has a short think time). If none of these
4672 * conditions holds, then a candidate queue for
4673 * injection is looked for through
4674 * bfq_choose_bfqq_for_injection(). Note that the
4675 * latter may return NULL (for example if the inject
4676 * limit for bfqq is currently 0).
4678 * NOTE: motivation for the second alternative
4680 * Thanks to the way the inject limit is updated in
4681 * bfq_update_has_short_ttime(), it is rather likely
4682 * that, if I/O is being plugged for bfqq and the
4683 * waker queue has pending I/O requests that are
4684 * blocking bfqq's I/O, then the fourth alternative
4685 * above lets the waker queue get served before the
4686 * I/O-plugging timeout fires. So one may deem the
4687 * second alternative superfluous. It is not, because
4688 * the fourth alternative may be way less effective in
4689 * case of a synchronization. For two main
4690 * reasons. First, throughput may be low because the
4691 * inject limit may be too low to guarantee the same
4692 * amount of injected I/O, from the waker queue or
4693 * other queues, that the second alternative
4694 * guarantees (the second alternative unconditionally
4695 * injects a pending I/O request of the waker queue
4696 * for each bfq_dispatch_request()). Second, with the
4697 * fourth alternative, the duration of the plugging,
4698 * i.e., the time before bfqq finally receives new I/O,
4699 * may not be minimized, because the waker queue may
4700 * happen to be served only after other queues.
4703 icq_to_bic(async_bfqq->next_rq->elv.icq) == bfqq->bic &&
4704 bfq_serv_to_charge(async_bfqq->next_rq, async_bfqq) <=
4705 bfq_bfqq_budget_left(async_bfqq))
4706 bfqq = bfqq->bic->bfqq[0];
4707 else if (bfqq->waker_bfqq &&
4708 bfq_bfqq_busy(bfqq->waker_bfqq) &&
4709 bfqq->waker_bfqq->next_rq &&
4710 bfq_serv_to_charge(bfqq->waker_bfqq->next_rq,
4711 bfqq->waker_bfqq) <=
4712 bfq_bfqq_budget_left(bfqq->waker_bfqq)
4714 bfqq = bfqq->waker_bfqq;
4715 else if (blocked_bfqq &&
4716 bfq_bfqq_busy(blocked_bfqq) &&
4717 blocked_bfqq->next_rq &&
4718 bfq_serv_to_charge(blocked_bfqq->next_rq,
4720 bfq_bfqq_budget_left(blocked_bfqq)
4722 bfqq = blocked_bfqq;
4723 else if (!idling_boosts_thr_without_issues(bfqd, bfqq) &&
4724 (bfqq->wr_coeff == 1 || bfqd->wr_busy_queues > 1 ||
4725 !bfq_bfqq_has_short_ttime(bfqq)))
4726 bfqq = bfq_choose_bfqq_for_injection(bfqd);
4733 reason = BFQQE_NO_MORE_REQUESTS;
4735 bfq_bfqq_expire(bfqd, bfqq, false, reason);
4737 bfqq = bfq_set_in_service_queue(bfqd);
4739 bfq_log_bfqq(bfqd, bfqq, "select_queue: checking new queue");
4744 bfq_log_bfqq(bfqd, bfqq, "select_queue: returned this queue");
4746 bfq_log(bfqd, "select_queue: no queue returned");
4751 static void bfq_update_wr_data(struct bfq_data *bfqd, struct bfq_queue *bfqq)
4753 struct bfq_entity *entity = &bfqq->entity;
4755 if (bfqq->wr_coeff > 1) { /* queue is being weight-raised */
4756 bfq_log_bfqq(bfqd, bfqq,
4757 "raising period dur %u/%u msec, old coeff %u, w %d(%d)",
4758 jiffies_to_msecs(jiffies - bfqq->last_wr_start_finish),
4759 jiffies_to_msecs(bfqq->wr_cur_max_time),
4761 bfqq->entity.weight, bfqq->entity.orig_weight);
4763 if (entity->prio_changed)
4764 bfq_log_bfqq(bfqd, bfqq, "WARN: pending prio change");
4767 * If the queue was activated in a burst, or too much
4768 * time has elapsed from the beginning of this
4769 * weight-raising period, then end weight raising.
4771 if (bfq_bfqq_in_large_burst(bfqq))
4772 bfq_bfqq_end_wr(bfqq);
4773 else if (time_is_before_jiffies(bfqq->last_wr_start_finish +
4774 bfqq->wr_cur_max_time)) {
4775 if (bfqq->wr_cur_max_time != bfqd->bfq_wr_rt_max_time ||
4776 time_is_before_jiffies(bfqq->wr_start_at_switch_to_srt +
4777 bfq_wr_duration(bfqd))) {
4779 * Either in interactive weight
4780 * raising, or in soft_rt weight
4782 * interactive-weight-raising period
4783 * elapsed (so no switch back to
4784 * interactive weight raising).
4786 bfq_bfqq_end_wr(bfqq);
4788 * soft_rt finishing while still in
4789 * interactive period, switch back to
4790 * interactive weight raising
4792 switch_back_to_interactive_wr(bfqq, bfqd);
4793 bfqq->entity.prio_changed = 1;
4796 if (bfqq->wr_coeff > 1 &&
4797 bfqq->wr_cur_max_time != bfqd->bfq_wr_rt_max_time &&
4798 bfqq->service_from_wr > max_service_from_wr) {
4799 /* see comments on max_service_from_wr */
4800 bfq_bfqq_end_wr(bfqq);
4804 * To improve latency (for this or other queues), immediately
4805 * update weight both if it must be raised and if it must be
4806 * lowered. Since, entity may be on some active tree here, and
4807 * might have a pending change of its ioprio class, invoke
4808 * next function with the last parameter unset (see the
4809 * comments on the function).
4811 if ((entity->weight > entity->orig_weight) != (bfqq->wr_coeff > 1))
4812 __bfq_entity_update_weight_prio(bfq_entity_service_tree(entity),
4817 * Dispatch next request from bfqq.
4819 static struct request *bfq_dispatch_rq_from_bfqq(struct bfq_data *bfqd,
4820 struct bfq_queue *bfqq)
4822 struct request *rq = bfqq->next_rq;
4823 unsigned long service_to_charge;
4825 service_to_charge = bfq_serv_to_charge(rq, bfqq);
4827 bfq_bfqq_served(bfqq, service_to_charge);
4829 if (bfqq == bfqd->in_service_queue && bfqd->wait_dispatch) {
4830 bfqd->wait_dispatch = false;
4831 bfqd->waited_rq = rq;
4834 bfq_dispatch_remove(bfqd->queue, rq);
4836 if (bfqq != bfqd->in_service_queue)
4840 * If weight raising has to terminate for bfqq, then next
4841 * function causes an immediate update of bfqq's weight,
4842 * without waiting for next activation. As a consequence, on
4843 * expiration, bfqq will be timestamped as if has never been
4844 * weight-raised during this service slot, even if it has
4845 * received part or even most of the service as a
4846 * weight-raised queue. This inflates bfqq's timestamps, which
4847 * is beneficial, as bfqq is then more willing to leave the
4848 * device immediately to possible other weight-raised queues.
4850 bfq_update_wr_data(bfqd, bfqq);
4853 * Expire bfqq, pretending that its budget expired, if bfqq
4854 * belongs to CLASS_IDLE and other queues are waiting for
4857 if (!(bfq_tot_busy_queues(bfqd) > 1 && bfq_class_idle(bfqq)))
4860 bfq_bfqq_expire(bfqd, bfqq, false, BFQQE_BUDGET_EXHAUSTED);
4866 static bool bfq_has_work(struct blk_mq_hw_ctx *hctx)
4868 struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
4871 * Avoiding lock: a race on bfqd->busy_queues should cause at
4872 * most a call to dispatch for nothing
4874 return !list_empty_careful(&bfqd->dispatch) ||
4875 bfq_tot_busy_queues(bfqd) > 0;
4878 static struct request *__bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
4880 struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
4881 struct request *rq = NULL;
4882 struct bfq_queue *bfqq = NULL;
4884 if (!list_empty(&bfqd->dispatch)) {
4885 rq = list_first_entry(&bfqd->dispatch, struct request,
4887 list_del_init(&rq->queuelist);
4893 * Increment counters here, because this
4894 * dispatch does not follow the standard
4895 * dispatch flow (where counters are
4900 goto inc_in_driver_start_rq;
4904 * We exploit the bfq_finish_requeue_request hook to
4905 * decrement rq_in_driver, but
4906 * bfq_finish_requeue_request will not be invoked on
4907 * this request. So, to avoid unbalance, just start
4908 * this request, without incrementing rq_in_driver. As
4909 * a negative consequence, rq_in_driver is deceptively
4910 * lower than it should be while this request is in
4911 * service. This may cause bfq_schedule_dispatch to be
4912 * invoked uselessly.
4914 * As for implementing an exact solution, the
4915 * bfq_finish_requeue_request hook, if defined, is
4916 * probably invoked also on this request. So, by
4917 * exploiting this hook, we could 1) increment
4918 * rq_in_driver here, and 2) decrement it in
4919 * bfq_finish_requeue_request. Such a solution would
4920 * let the value of the counter be always accurate,
4921 * but it would entail using an extra interface
4922 * function. This cost seems higher than the benefit,
4923 * being the frequency of non-elevator-private
4924 * requests very low.
4929 bfq_log(bfqd, "dispatch requests: %d busy queues",
4930 bfq_tot_busy_queues(bfqd));
4932 if (bfq_tot_busy_queues(bfqd) == 0)
4936 * Force device to serve one request at a time if
4937 * strict_guarantees is true. Forcing this service scheme is
4938 * currently the ONLY way to guarantee that the request
4939 * service order enforced by the scheduler is respected by a
4940 * queueing device. Otherwise the device is free even to make
4941 * some unlucky request wait for as long as the device
4944 * Of course, serving one request at a time may cause loss of
4947 if (bfqd->strict_guarantees && bfqd->rq_in_driver > 0)
4950 bfqq = bfq_select_queue(bfqd);
4954 rq = bfq_dispatch_rq_from_bfqq(bfqd, bfqq);
4957 inc_in_driver_start_rq:
4958 bfqd->rq_in_driver++;
4960 rq->rq_flags |= RQF_STARTED;
4966 #ifdef CONFIG_BFQ_CGROUP_DEBUG
4967 static void bfq_update_dispatch_stats(struct request_queue *q,
4969 struct bfq_queue *in_serv_queue,
4970 bool idle_timer_disabled)
4972 struct bfq_queue *bfqq = rq ? RQ_BFQQ(rq) : NULL;
4974 if (!idle_timer_disabled && !bfqq)
4978 * rq and bfqq are guaranteed to exist until this function
4979 * ends, for the following reasons. First, rq can be
4980 * dispatched to the device, and then can be completed and
4981 * freed, only after this function ends. Second, rq cannot be
4982 * merged (and thus freed because of a merge) any longer,
4983 * because it has already started. Thus rq cannot be freed
4984 * before this function ends, and, since rq has a reference to
4985 * bfqq, the same guarantee holds for bfqq too.
4987 * In addition, the following queue lock guarantees that
4988 * bfqq_group(bfqq) exists as well.
4990 spin_lock_irq(&q->queue_lock);
4991 if (idle_timer_disabled)
4993 * Since the idle timer has been disabled,
4994 * in_serv_queue contained some request when
4995 * __bfq_dispatch_request was invoked above, which
4996 * implies that rq was picked exactly from
4997 * in_serv_queue. Thus in_serv_queue == bfqq, and is
4998 * therefore guaranteed to exist because of the above
5001 bfqg_stats_update_idle_time(bfqq_group(in_serv_queue));
5003 struct bfq_group *bfqg = bfqq_group(bfqq);
5005 bfqg_stats_update_avg_queue_size(bfqg);
5006 bfqg_stats_set_start_empty_time(bfqg);
5007 bfqg_stats_update_io_remove(bfqg, rq->cmd_flags);
5009 spin_unlock_irq(&q->queue_lock);
5012 static inline void bfq_update_dispatch_stats(struct request_queue *q,
5014 struct bfq_queue *in_serv_queue,
5015 bool idle_timer_disabled) {}
5016 #endif /* CONFIG_BFQ_CGROUP_DEBUG */
5018 static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
5020 struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
5022 struct bfq_queue *in_serv_queue;
5023 bool waiting_rq, idle_timer_disabled;
5025 spin_lock_irq(&bfqd->lock);
5027 in_serv_queue = bfqd->in_service_queue;
5028 waiting_rq = in_serv_queue && bfq_bfqq_wait_request(in_serv_queue);
5030 rq = __bfq_dispatch_request(hctx);
5032 idle_timer_disabled =
5033 waiting_rq && !bfq_bfqq_wait_request(in_serv_queue);
5035 spin_unlock_irq(&bfqd->lock);
5037 bfq_update_dispatch_stats(hctx->queue, rq, in_serv_queue,
5038 idle_timer_disabled);
5044 * Task holds one reference to the queue, dropped when task exits. Each rq
5045 * in-flight on this queue also holds a reference, dropped when rq is freed.
5047 * Scheduler lock must be held here. Recall not to use bfqq after calling
5048 * this function on it.
5050 void bfq_put_queue(struct bfq_queue *bfqq)
5052 struct bfq_queue *item;
5053 struct hlist_node *n;
5054 struct bfq_group *bfqg = bfqq_group(bfqq);
5057 bfq_log_bfqq(bfqq->bfqd, bfqq, "put_queue: %p %d",
5064 if (!hlist_unhashed(&bfqq->burst_list_node)) {
5065 hlist_del_init(&bfqq->burst_list_node);
5067 * Decrement also burst size after the removal, if the
5068 * process associated with bfqq is exiting, and thus
5069 * does not contribute to the burst any longer. This
5070 * decrement helps filter out false positives of large
5071 * bursts, when some short-lived process (often due to
5072 * the execution of commands by some service) happens
5073 * to start and exit while a complex application is
5074 * starting, and thus spawning several processes that
5075 * do I/O (and that *must not* be treated as a large
5076 * burst, see comments on bfq_handle_burst).
5078 * In particular, the decrement is performed only if:
5079 * 1) bfqq is not a merged queue, because, if it is,
5080 * then this free of bfqq is not triggered by the exit
5081 * of the process bfqq is associated with, but exactly
5082 * by the fact that bfqq has just been merged.
5083 * 2) burst_size is greater than 0, to handle
5084 * unbalanced decrements. Unbalanced decrements may
5085 * happen in te following case: bfqq is inserted into
5086 * the current burst list--without incrementing
5087 * bust_size--because of a split, but the current
5088 * burst list is not the burst list bfqq belonged to
5089 * (see comments on the case of a split in
5092 if (bfqq->bic && bfqq->bfqd->burst_size > 0)
5093 bfqq->bfqd->burst_size--;
5097 * bfqq does not exist any longer, so it cannot be woken by
5098 * any other queue, and cannot wake any other queue. Then bfqq
5099 * must be removed from the woken list of its possible waker
5100 * queue, and all queues in the woken list of bfqq must stop
5101 * having a waker queue. Strictly speaking, these updates
5102 * should be performed when bfqq remains with no I/O source
5103 * attached to it, which happens before bfqq gets freed. In
5104 * particular, this happens when the last process associated
5105 * with bfqq exits or gets associated with a different
5106 * queue. However, both events lead to bfqq being freed soon,
5107 * and dangling references would come out only after bfqq gets
5108 * freed. So these updates are done here, as a simple and safe
5109 * way to handle all cases.
5111 /* remove bfqq from woken list */
5112 if (!hlist_unhashed(&bfqq->woken_list_node))
5113 hlist_del_init(&bfqq->woken_list_node);
5115 /* reset waker for all queues in woken list */
5116 hlist_for_each_entry_safe(item, n, &bfqq->woken_list,
5118 item->waker_bfqq = NULL;
5119 hlist_del_init(&item->woken_list_node);
5122 if (bfqq->bfqd && bfqq->bfqd->last_completed_rq_bfqq == bfqq)
5123 bfqq->bfqd->last_completed_rq_bfqq = NULL;
5125 kmem_cache_free(bfq_pool, bfqq);
5126 bfqg_and_blkg_put(bfqg);
5129 static void bfq_put_stable_ref(struct bfq_queue *bfqq)
5132 bfq_put_queue(bfqq);
5135 static void bfq_put_cooperator(struct bfq_queue *bfqq)
5137 struct bfq_queue *__bfqq, *next;
5140 * If this queue was scheduled to merge with another queue, be
5141 * sure to drop the reference taken on that queue (and others in
5142 * the merge chain). See bfq_setup_merge and bfq_merge_bfqqs.
5144 __bfqq = bfqq->new_bfqq;
5148 next = __bfqq->new_bfqq;
5149 bfq_put_queue(__bfqq);
5154 static void bfq_exit_bfqq(struct bfq_data *bfqd, struct bfq_queue *bfqq)
5156 if (bfqq == bfqd->in_service_queue) {
5157 __bfq_bfqq_expire(bfqd, bfqq, BFQQE_BUDGET_TIMEOUT);
5158 bfq_schedule_dispatch(bfqd);
5161 bfq_log_bfqq(bfqd, bfqq, "exit_bfqq: %p, %d", bfqq, bfqq->ref);
5163 bfq_put_cooperator(bfqq);
5165 bfq_release_process_ref(bfqd, bfqq);
5168 static void bfq_exit_icq_bfqq(struct bfq_io_cq *bic, bool is_sync)
5170 struct bfq_queue *bfqq = bic_to_bfqq(bic, is_sync);
5171 struct bfq_data *bfqd;
5174 bfqd = bfqq->bfqd; /* NULL if scheduler already exited */
5177 unsigned long flags;
5179 spin_lock_irqsave(&bfqd->lock, flags);
5181 bfq_exit_bfqq(bfqd, bfqq);
5182 bic_set_bfqq(bic, NULL, is_sync);
5183 spin_unlock_irqrestore(&bfqd->lock, flags);
5187 static void bfq_exit_icq(struct io_cq *icq)
5189 struct bfq_io_cq *bic = icq_to_bic(icq);
5191 if (bic->stable_merge_bfqq) {
5192 struct bfq_data *bfqd = bic->stable_merge_bfqq->bfqd;
5195 * bfqd is NULL if scheduler already exited, and in
5196 * that case this is the last time bfqq is accessed.
5199 unsigned long flags;
5201 spin_lock_irqsave(&bfqd->lock, flags);
5202 bfq_put_stable_ref(bic->stable_merge_bfqq);
5203 spin_unlock_irqrestore(&bfqd->lock, flags);
5205 bfq_put_stable_ref(bic->stable_merge_bfqq);
5209 bfq_exit_icq_bfqq(bic, true);
5210 bfq_exit_icq_bfqq(bic, false);
5214 * Update the entity prio values; note that the new values will not
5215 * be used until the next (re)activation.
5218 bfq_set_next_ioprio_data(struct bfq_queue *bfqq, struct bfq_io_cq *bic)
5220 struct task_struct *tsk = current;
5222 struct bfq_data *bfqd = bfqq->bfqd;
5227 ioprio_class = IOPRIO_PRIO_CLASS(bic->ioprio);
5228 switch (ioprio_class) {
5230 pr_err("bdi %s: bfq: bad prio class %d\n",
5231 bdi_dev_name(bfqq->bfqd->queue->backing_dev_info),
5234 case IOPRIO_CLASS_NONE:
5236 * No prio set, inherit CPU scheduling settings.
5238 bfqq->new_ioprio = task_nice_ioprio(tsk);
5239 bfqq->new_ioprio_class = task_nice_ioclass(tsk);
5241 case IOPRIO_CLASS_RT:
5242 bfqq->new_ioprio = IOPRIO_PRIO_DATA(bic->ioprio);
5243 bfqq->new_ioprio_class = IOPRIO_CLASS_RT;
5245 case IOPRIO_CLASS_BE:
5246 bfqq->new_ioprio = IOPRIO_PRIO_DATA(bic->ioprio);
5247 bfqq->new_ioprio_class = IOPRIO_CLASS_BE;
5249 case IOPRIO_CLASS_IDLE:
5250 bfqq->new_ioprio_class = IOPRIO_CLASS_IDLE;
5251 bfqq->new_ioprio = 7;
5255 if (bfqq->new_ioprio >= IOPRIO_BE_NR) {
5256 pr_crit("bfq_set_next_ioprio_data: new_ioprio %d\n",
5258 bfqq->new_ioprio = IOPRIO_BE_NR;
5261 bfqq->entity.new_weight = bfq_ioprio_to_weight(bfqq->new_ioprio);
5262 bfq_log_bfqq(bfqd, bfqq, "new_ioprio %d new_weight %d",
5263 bfqq->new_ioprio, bfqq->entity.new_weight);
5264 bfqq->entity.prio_changed = 1;
5267 static struct bfq_queue *bfq_get_queue(struct bfq_data *bfqd,
5268 struct bio *bio, bool is_sync,
5269 struct bfq_io_cq *bic,
5272 static void bfq_check_ioprio_change(struct bfq_io_cq *bic, struct bio *bio)
5274 struct bfq_data *bfqd = bic_to_bfqd(bic);
5275 struct bfq_queue *bfqq;
5276 int ioprio = bic->icq.ioc->ioprio;
5279 * This condition may trigger on a newly created bic, be sure to
5280 * drop the lock before returning.
5282 if (unlikely(!bfqd) || likely(bic->ioprio == ioprio))
5285 bic->ioprio = ioprio;
5287 bfqq = bic_to_bfqq(bic, false);
5289 bfq_release_process_ref(bfqd, bfqq);
5290 bfqq = bfq_get_queue(bfqd, bio, BLK_RW_ASYNC, bic, true);
5291 bic_set_bfqq(bic, bfqq, false);
5294 bfqq = bic_to_bfqq(bic, true);
5296 bfq_set_next_ioprio_data(bfqq, bic);
5299 static void bfq_init_bfqq(struct bfq_data *bfqd, struct bfq_queue *bfqq,
5300 struct bfq_io_cq *bic, pid_t pid, int is_sync)
5302 u64 now_ns = ktime_get_ns();
5304 RB_CLEAR_NODE(&bfqq->entity.rb_node);
5305 INIT_LIST_HEAD(&bfqq->fifo);
5306 INIT_HLIST_NODE(&bfqq->burst_list_node);
5307 INIT_HLIST_NODE(&bfqq->woken_list_node);
5308 INIT_HLIST_HEAD(&bfqq->woken_list);
5314 bfq_set_next_ioprio_data(bfqq, bic);
5318 * No need to mark as has_short_ttime if in
5319 * idle_class, because no device idling is performed
5320 * for queues in idle class
5322 if (!bfq_class_idle(bfqq))
5323 /* tentatively mark as has_short_ttime */
5324 bfq_mark_bfqq_has_short_ttime(bfqq);
5325 bfq_mark_bfqq_sync(bfqq);
5326 bfq_mark_bfqq_just_created(bfqq);
5328 bfq_clear_bfqq_sync(bfqq);
5330 /* set end request to minus infinity from now */
5331 bfqq->ttime.last_end_request = now_ns + 1;
5333 bfqq->creation_time = jiffies;
5335 bfqq->io_start_time = now_ns;
5337 bfq_mark_bfqq_IO_bound(bfqq);
5341 /* Tentative initial value to trade off between thr and lat */
5342 bfqq->max_budget = (2 * bfq_max_budget(bfqd)) / 3;
5343 bfqq->budget_timeout = bfq_smallest_from_now();
5346 bfqq->last_wr_start_finish = jiffies;
5347 bfqq->wr_start_at_switch_to_srt = bfq_smallest_from_now();
5348 bfqq->split_time = bfq_smallest_from_now();
5351 * To not forget the possibly high bandwidth consumed by a
5352 * process/queue in the recent past,
5353 * bfq_bfqq_softrt_next_start() returns a value at least equal
5354 * to the current value of bfqq->soft_rt_next_start (see
5355 * comments on bfq_bfqq_softrt_next_start). Set
5356 * soft_rt_next_start to now, to mean that bfqq has consumed
5357 * no bandwidth so far.
5359 bfqq->soft_rt_next_start = jiffies;
5361 /* first request is almost certainly seeky */
5362 bfqq->seek_history = 1;
5365 static struct bfq_queue **bfq_async_queue_prio(struct bfq_data *bfqd,
5366 struct bfq_group *bfqg,
5367 int ioprio_class, int ioprio)
5369 switch (ioprio_class) {
5370 case IOPRIO_CLASS_RT:
5371 return &bfqg->async_bfqq[0][ioprio];
5372 case IOPRIO_CLASS_NONE:
5373 ioprio = IOPRIO_NORM;
5375 case IOPRIO_CLASS_BE:
5376 return &bfqg->async_bfqq[1][ioprio];
5377 case IOPRIO_CLASS_IDLE:
5378 return &bfqg->async_idle_bfqq;
5384 static struct bfq_queue *
5385 bfq_do_early_stable_merge(struct bfq_data *bfqd, struct bfq_queue *bfqq,
5386 struct bfq_io_cq *bic,
5387 struct bfq_queue *last_bfqq_created)
5389 struct bfq_queue *new_bfqq =
5390 bfq_setup_merge(bfqq, last_bfqq_created);
5396 new_bfqq->bic->stably_merged = true;
5397 bic->stably_merged = true;
5400 * Reusing merge functions. This implies that
5401 * bfqq->bic must be set too, for
5402 * bfq_merge_bfqqs to correctly save bfqq's
5403 * state before killing it.
5406 bfq_merge_bfqqs(bfqd, bic, bfqq, new_bfqq);
5412 * Many throughput-sensitive workloads are made of several parallel
5413 * I/O flows, with all flows generated by the same application, or
5414 * more generically by the same task (e.g., system boot). The most
5415 * counterproductive action with these workloads is plugging I/O
5416 * dispatch when one of the bfq_queues associated with these flows
5417 * remains temporarily empty.
5419 * To avoid this plugging, BFQ has been using a burst-handling
5420 * mechanism for years now. This mechanism has proven effective for
5421 * throughput, and not detrimental for service guarantees. The
5422 * following function pushes this mechanism a little bit further,
5423 * basing on the following two facts.
5425 * First, all the I/O flows of a the same application or task
5426 * contribute to the execution/completion of that common application
5427 * or task. So the performance figures that matter are total
5428 * throughput of the flows and task-wide I/O latency. In particular,
5429 * these flows do not need to be protected from each other, in terms
5430 * of individual bandwidth or latency.
5432 * Second, the above fact holds regardless of the number of flows.
5434 * Putting these two facts together, this commits merges stably the
5435 * bfq_queues associated with these I/O flows, i.e., with the
5436 * processes that generate these IO/ flows, regardless of how many the
5437 * involved processes are.
5439 * To decide whether a set of bfq_queues is actually associated with
5440 * the I/O flows of a common application or task, and to merge these
5441 * queues stably, this function operates as follows: given a bfq_queue,
5442 * say Q2, currently being created, and the last bfq_queue, say Q1,
5443 * created before Q2, Q2 is merged stably with Q1 if
5444 * - very little time has elapsed since when Q1 was created
5445 * - Q2 has the same ioprio as Q1
5446 * - Q2 belongs to the same group as Q1
5448 * Merging bfq_queues also reduces scheduling overhead. A fio test
5449 * with ten random readers on /dev/nullb shows a throughput boost of
5450 * 40%, with a quadcore. Since BFQ's execution time amounts to ~50% of
5451 * the total per-request processing time, the above throughput boost
5452 * implies that BFQ's overhead is reduced by more than 50%.
5454 * This new mechanism most certainly obsoletes the current
5455 * burst-handling heuristics. We keep those heuristics for the moment.
5457 static struct bfq_queue *bfq_do_or_sched_stable_merge(struct bfq_data *bfqd,
5458 struct bfq_queue *bfqq,
5459 struct bfq_io_cq *bic)
5461 struct bfq_queue **source_bfqq = bfqq->entity.parent ?
5462 &bfqq->entity.parent->last_bfqq_created :
5463 &bfqd->last_bfqq_created;
5465 struct bfq_queue *last_bfqq_created = *source_bfqq;
5468 * If last_bfqq_created has not been set yet, then init it. If
5469 * it has been set already, but too long ago, then move it
5470 * forward to bfqq. Finally, move also if bfqq belongs to a
5471 * different group than last_bfqq_created, or if bfqq has a
5472 * different ioprio or ioprio_class. If none of these
5473 * conditions holds true, then try an early stable merge or
5474 * schedule a delayed stable merge.
5476 * A delayed merge is scheduled (instead of performing an
5477 * early merge), in case bfqq might soon prove to be more
5478 * throughput-beneficial if not merged. Currently this is
5479 * possible only if bfqd is rotational with no queueing. For
5480 * such a drive, not merging bfqq is better for throughput if
5481 * bfqq happens to contain sequential I/O. So, we wait a
5482 * little bit for enough I/O to flow through bfqq. After that,
5483 * if such an I/O is sequential, then the merge is
5484 * canceled. Otherwise the merge is finally performed.
5486 if (!last_bfqq_created ||
5487 time_before(last_bfqq_created->creation_time +
5488 bfqd->bfq_burst_interval,
5489 bfqq->creation_time) ||
5490 bfqq->entity.parent != last_bfqq_created->entity.parent ||
5491 bfqq->ioprio != last_bfqq_created->ioprio ||
5492 bfqq->ioprio_class != last_bfqq_created->ioprio_class)
5493 *source_bfqq = bfqq;
5494 else if (time_after_eq(last_bfqq_created->creation_time +
5495 bfqd->bfq_burst_interval,
5496 bfqq->creation_time)) {
5497 if (likely(bfqd->nonrot_with_queueing))
5499 * With this type of drive, leaving
5500 * bfqq alone may provide no
5501 * throughput benefits compared with
5502 * merging bfqq. So merge bfqq now.
5504 bfqq = bfq_do_early_stable_merge(bfqd, bfqq,
5507 else { /* schedule tentative stable merge */
5509 * get reference on last_bfqq_created,
5510 * to prevent it from being freed,
5511 * until we decide whether to merge
5513 last_bfqq_created->ref++;
5515 * need to keep track of stable refs, to
5516 * compute process refs correctly
5518 last_bfqq_created->stable_ref++;
5520 * Record the bfqq to merge to.
5522 bic->stable_merge_bfqq = last_bfqq_created;
5530 static struct bfq_queue *bfq_get_queue(struct bfq_data *bfqd,
5531 struct bio *bio, bool is_sync,
5532 struct bfq_io_cq *bic,
5535 const int ioprio = IOPRIO_PRIO_DATA(bic->ioprio);
5536 const int ioprio_class = IOPRIO_PRIO_CLASS(bic->ioprio);
5537 struct bfq_queue **async_bfqq = NULL;
5538 struct bfq_queue *bfqq;
5539 struct bfq_group *bfqg;
5543 bfqg = bfq_find_set_group(bfqd, __bio_blkcg(bio));
5545 bfqq = &bfqd->oom_bfqq;
5550 async_bfqq = bfq_async_queue_prio(bfqd, bfqg, ioprio_class,
5557 bfqq = kmem_cache_alloc_node(bfq_pool,
5558 GFP_NOWAIT | __GFP_ZERO | __GFP_NOWARN,
5562 bfq_init_bfqq(bfqd, bfqq, bic, current->pid,
5564 bfq_init_entity(&bfqq->entity, bfqg);
5565 bfq_log_bfqq(bfqd, bfqq, "allocated");
5567 bfqq = &bfqd->oom_bfqq;
5568 bfq_log_bfqq(bfqd, bfqq, "using oom bfqq");
5573 * Pin the queue now that it's allocated, scheduler exit will
5578 * Extra group reference, w.r.t. sync
5579 * queue. This extra reference is removed
5580 * only if bfqq->bfqg disappears, to
5581 * guarantee that this queue is not freed
5582 * until its group goes away.
5584 bfq_log_bfqq(bfqd, bfqq, "get_queue, bfqq not in async: %p, %d",
5590 bfqq->ref++; /* get a process reference to this queue */
5592 if (bfqq != &bfqd->oom_bfqq && is_sync && !respawn)
5593 bfqq = bfq_do_or_sched_stable_merge(bfqd, bfqq, bic);
5599 static void bfq_update_io_thinktime(struct bfq_data *bfqd,
5600 struct bfq_queue *bfqq)
5602 struct bfq_ttime *ttime = &bfqq->ttime;
5606 * We are really interested in how long it takes for the queue to
5607 * become busy when there is no outstanding IO for this queue. So
5608 * ignore cases when the bfq queue has already IO queued.
5610 if (bfqq->dispatched || bfq_bfqq_busy(bfqq))
5612 elapsed = ktime_get_ns() - bfqq->ttime.last_end_request;
5613 elapsed = min_t(u64, elapsed, 2ULL * bfqd->bfq_slice_idle);
5615 ttime->ttime_samples = (7*ttime->ttime_samples + 256) / 8;
5616 ttime->ttime_total = div_u64(7*ttime->ttime_total + 256*elapsed, 8);
5617 ttime->ttime_mean = div64_ul(ttime->ttime_total + 128,
5618 ttime->ttime_samples);
5622 bfq_update_io_seektime(struct bfq_data *bfqd, struct bfq_queue *bfqq,
5625 bfqq->seek_history <<= 1;
5626 bfqq->seek_history |= BFQ_RQ_SEEKY(bfqd, bfqq->last_request_pos, rq);
5628 if (bfqq->wr_coeff > 1 &&
5629 bfqq->wr_cur_max_time == bfqd->bfq_wr_rt_max_time &&
5630 BFQQ_TOTALLY_SEEKY(bfqq)) {
5631 if (time_is_before_jiffies(bfqq->wr_start_at_switch_to_srt +
5632 bfq_wr_duration(bfqd))) {
5634 * In soft_rt weight raising with the
5635 * interactive-weight-raising period
5636 * elapsed (so no switch back to
5637 * interactive weight raising).
5639 bfq_bfqq_end_wr(bfqq);
5641 * stopping soft_rt weight raising
5642 * while still in interactive period,
5643 * switch back to interactive weight
5646 switch_back_to_interactive_wr(bfqq, bfqd);
5647 bfqq->entity.prio_changed = 1;
5652 static void bfq_update_has_short_ttime(struct bfq_data *bfqd,
5653 struct bfq_queue *bfqq,
5654 struct bfq_io_cq *bic)
5656 bool has_short_ttime = true, state_changed;
5659 * No need to update has_short_ttime if bfqq is async or in
5660 * idle io prio class, or if bfq_slice_idle is zero, because
5661 * no device idling is performed for bfqq in this case.
5663 if (!bfq_bfqq_sync(bfqq) || bfq_class_idle(bfqq) ||
5664 bfqd->bfq_slice_idle == 0)
5667 /* Idle window just restored, statistics are meaningless. */
5668 if (time_is_after_eq_jiffies(bfqq->split_time +
5669 bfqd->bfq_wr_min_idle_time))
5672 /* Think time is infinite if no process is linked to
5673 * bfqq. Otherwise check average think time to decide whether
5674 * to mark as has_short_ttime. To this goal, compare average
5675 * think time with half the I/O-plugging timeout.
5677 if (atomic_read(&bic->icq.ioc->active_ref) == 0 ||
5678 (bfq_sample_valid(bfqq->ttime.ttime_samples) &&
5679 bfqq->ttime.ttime_mean > bfqd->bfq_slice_idle>>1))
5680 has_short_ttime = false;
5682 state_changed = has_short_ttime != bfq_bfqq_has_short_ttime(bfqq);
5684 if (has_short_ttime)
5685 bfq_mark_bfqq_has_short_ttime(bfqq);
5687 bfq_clear_bfqq_has_short_ttime(bfqq);
5690 * Until the base value for the total service time gets
5691 * finally computed for bfqq, the inject limit does depend on
5692 * the think-time state (short|long). In particular, the limit
5693 * is 0 or 1 if the think time is deemed, respectively, as
5694 * short or long (details in the comments in
5695 * bfq_update_inject_limit()). Accordingly, the next
5696 * instructions reset the inject limit if the think-time state
5697 * has changed and the above base value is still to be
5700 * However, the reset is performed only if more than 100 ms
5701 * have elapsed since the last update of the inject limit, or
5702 * (inclusive) if the change is from short to long think
5703 * time. The reason for this waiting is as follows.
5705 * bfqq may have a long think time because of a
5706 * synchronization with some other queue, i.e., because the
5707 * I/O of some other queue may need to be completed for bfqq
5708 * to receive new I/O. Details in the comments on the choice
5709 * of the queue for injection in bfq_select_queue().
5711 * As stressed in those comments, if such a synchronization is
5712 * actually in place, then, without injection on bfqq, the
5713 * blocking I/O cannot happen to served while bfqq is in
5714 * service. As a consequence, if bfqq is granted
5715 * I/O-dispatch-plugging, then bfqq remains empty, and no I/O
5716 * is dispatched, until the idle timeout fires. This is likely
5717 * to result in lower bandwidth and higher latencies for bfqq,
5718 * and in a severe loss of total throughput.
5720 * On the opposite end, a non-zero inject limit may allow the
5721 * I/O that blocks bfqq to be executed soon, and therefore
5722 * bfqq to receive new I/O soon.
5724 * But, if the blocking gets actually eliminated, then the
5725 * next think-time sample for bfqq may be very low. This in
5726 * turn may cause bfqq's think time to be deemed
5727 * short. Without the 100 ms barrier, this new state change
5728 * would cause the body of the next if to be executed
5729 * immediately. But this would set to 0 the inject
5730 * limit. Without injection, the blocking I/O would cause the
5731 * think time of bfqq to become long again, and therefore the
5732 * inject limit to be raised again, and so on. The only effect
5733 * of such a steady oscillation between the two think-time
5734 * states would be to prevent effective injection on bfqq.
5736 * In contrast, if the inject limit is not reset during such a
5737 * long time interval as 100 ms, then the number of short
5738 * think time samples can grow significantly before the reset
5739 * is performed. As a consequence, the think time state can
5740 * become stable before the reset. Therefore there will be no
5741 * state change when the 100 ms elapse, and no reset of the
5742 * inject limit. The inject limit remains steadily equal to 1
5743 * both during and after the 100 ms. So injection can be
5744 * performed at all times, and throughput gets boosted.
5746 * An inject limit equal to 1 is however in conflict, in
5747 * general, with the fact that the think time of bfqq is
5748 * short, because injection may be likely to delay bfqq's I/O
5749 * (as explained in the comments in
5750 * bfq_update_inject_limit()). But this does not happen in
5751 * this special case, because bfqq's low think time is due to
5752 * an effective handling of a synchronization, through
5753 * injection. In this special case, bfqq's I/O does not get
5754 * delayed by injection; on the contrary, bfqq's I/O is
5755 * brought forward, because it is not blocked for
5758 * In addition, serving the blocking I/O much sooner, and much
5759 * more frequently than once per I/O-plugging timeout, makes
5760 * it much quicker to detect a waker queue (the concept of
5761 * waker queue is defined in the comments in
5762 * bfq_add_request()). This makes it possible to start sooner
5763 * to boost throughput more effectively, by injecting the I/O
5764 * of the waker queue unconditionally on every
5765 * bfq_dispatch_request().
5767 * One last, important benefit of not resetting the inject
5768 * limit before 100 ms is that, during this time interval, the
5769 * base value for the total service time is likely to get
5770 * finally computed for bfqq, freeing the inject limit from
5771 * its relation with the think time.
5773 if (state_changed && bfqq->last_serv_time_ns == 0 &&
5774 (time_is_before_eq_jiffies(bfqq->decrease_time_jif +
5775 msecs_to_jiffies(100)) ||
5777 bfq_reset_inject_limit(bfqd, bfqq);
5781 * Called when a new fs request (rq) is added to bfqq. Check if there's
5782 * something we should do about it.
5784 static void bfq_rq_enqueued(struct bfq_data *bfqd, struct bfq_queue *bfqq,
5787 if (rq->cmd_flags & REQ_META)
5788 bfqq->meta_pending++;
5790 bfqq->last_request_pos = blk_rq_pos(rq) + blk_rq_sectors(rq);
5792 if (bfqq == bfqd->in_service_queue && bfq_bfqq_wait_request(bfqq)) {
5793 bool small_req = bfqq->queued[rq_is_sync(rq)] == 1 &&
5794 blk_rq_sectors(rq) < 32;
5795 bool budget_timeout = bfq_bfqq_budget_timeout(bfqq);
5798 * There is just this request queued: if
5799 * - the request is small, and
5800 * - we are idling to boost throughput, and
5801 * - the queue is not to be expired,
5804 * In this way, if the device is being idled to wait
5805 * for a new request from the in-service queue, we
5806 * avoid unplugging the device and committing the
5807 * device to serve just a small request. In contrast
5808 * we wait for the block layer to decide when to
5809 * unplug the device: hopefully, new requests will be
5810 * merged to this one quickly, then the device will be
5811 * unplugged and larger requests will be dispatched.
5813 if (small_req && idling_boosts_thr_without_issues(bfqd, bfqq) &&
5818 * A large enough request arrived, or idling is being
5819 * performed to preserve service guarantees, or
5820 * finally the queue is to be expired: in all these
5821 * cases disk idling is to be stopped, so clear
5822 * wait_request flag and reset timer.
5824 bfq_clear_bfqq_wait_request(bfqq);
5825 hrtimer_try_to_cancel(&bfqd->idle_slice_timer);
5828 * The queue is not empty, because a new request just
5829 * arrived. Hence we can safely expire the queue, in
5830 * case of budget timeout, without risking that the
5831 * timestamps of the queue are not updated correctly.
5832 * See [1] for more details.
5835 bfq_bfqq_expire(bfqd, bfqq, false,
5836 BFQQE_BUDGET_TIMEOUT);
5840 /* returns true if it causes the idle timer to be disabled */
5841 static bool __bfq_insert_request(struct bfq_data *bfqd, struct request *rq)
5843 struct bfq_queue *bfqq = RQ_BFQQ(rq),
5844 *new_bfqq = bfq_setup_cooperator(bfqd, bfqq, rq, true,
5846 bool waiting, idle_timer_disabled = false;
5850 * Release the request's reference to the old bfqq
5851 * and make sure one is taken to the shared queue.
5853 new_bfqq->allocated++;
5857 * If the bic associated with the process
5858 * issuing this request still points to bfqq
5859 * (and thus has not been already redirected
5860 * to new_bfqq or even some other bfq_queue),
5861 * then complete the merge and redirect it to
5864 if (bic_to_bfqq(RQ_BIC(rq), 1) == bfqq)
5865 bfq_merge_bfqqs(bfqd, RQ_BIC(rq),
5868 bfq_clear_bfqq_just_created(bfqq);
5870 * rq is about to be enqueued into new_bfqq,
5871 * release rq reference on bfqq
5873 bfq_put_queue(bfqq);
5874 rq->elv.priv[1] = new_bfqq;
5878 bfq_update_io_thinktime(bfqd, bfqq);
5879 bfq_update_has_short_ttime(bfqd, bfqq, RQ_BIC(rq));
5880 bfq_update_io_seektime(bfqd, bfqq, rq);
5882 waiting = bfqq && bfq_bfqq_wait_request(bfqq);
5883 bfq_add_request(rq);
5884 idle_timer_disabled = waiting && !bfq_bfqq_wait_request(bfqq);
5886 rq->fifo_time = ktime_get_ns() + bfqd->bfq_fifo_expire[rq_is_sync(rq)];
5887 list_add_tail(&rq->queuelist, &bfqq->fifo);
5889 bfq_rq_enqueued(bfqd, bfqq, rq);
5891 return idle_timer_disabled;
5894 #ifdef CONFIG_BFQ_CGROUP_DEBUG
5895 static void bfq_update_insert_stats(struct request_queue *q,
5896 struct bfq_queue *bfqq,
5897 bool idle_timer_disabled,
5898 unsigned int cmd_flags)
5904 * bfqq still exists, because it can disappear only after
5905 * either it is merged with another queue, or the process it
5906 * is associated with exits. But both actions must be taken by
5907 * the same process currently executing this flow of
5910 * In addition, the following queue lock guarantees that
5911 * bfqq_group(bfqq) exists as well.
5913 spin_lock_irq(&q->queue_lock);
5914 bfqg_stats_update_io_add(bfqq_group(bfqq), bfqq, cmd_flags);
5915 if (idle_timer_disabled)
5916 bfqg_stats_update_idle_time(bfqq_group(bfqq));
5917 spin_unlock_irq(&q->queue_lock);
5920 static inline void bfq_update_insert_stats(struct request_queue *q,
5921 struct bfq_queue *bfqq,
5922 bool idle_timer_disabled,
5923 unsigned int cmd_flags) {}
5924 #endif /* CONFIG_BFQ_CGROUP_DEBUG */
5926 static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
5929 struct request_queue *q = hctx->queue;
5930 struct bfq_data *bfqd = q->elevator->elevator_data;
5931 struct bfq_queue *bfqq;
5932 bool idle_timer_disabled = false;
5933 unsigned int cmd_flags;
5935 #ifdef CONFIG_BFQ_GROUP_IOSCHED
5936 if (!cgroup_subsys_on_dfl(io_cgrp_subsys) && rq->bio)
5937 bfqg_stats_update_legacy_io(q, rq);
5939 spin_lock_irq(&bfqd->lock);
5940 if (blk_mq_sched_try_insert_merge(q, rq)) {
5941 spin_unlock_irq(&bfqd->lock);
5945 spin_unlock_irq(&bfqd->lock);
5947 trace_block_rq_insert(rq);
5949 spin_lock_irq(&bfqd->lock);
5950 bfqq = bfq_init_rq(rq);
5953 * Reqs with at_head or passthrough flags set are to be put
5954 * directly into dispatch list. Additional case for putting rq
5955 * directly into the dispatch queue: the only active
5956 * bfq_queues are bfqq and either its waker bfq_queue or one
5957 * of its woken bfq_queues. The rationale behind this
5958 * additional condition is as follows:
5959 * - consider a bfq_queue, say Q1, detected as a waker of
5960 * another bfq_queue, say Q2
5961 * - by definition of a waker, Q1 blocks the I/O of Q2, i.e.,
5962 * some I/O of Q1 needs to be completed for new I/O of Q2
5963 * to arrive. A notable example of waker is journald
5964 * - so, Q1 and Q2 are in any respect the queues of two
5965 * cooperating processes (or of two cooperating sets of
5966 * processes): the goal of Q1's I/O is doing what needs to
5967 * be done so that new Q2's I/O can finally be
5968 * issued. Therefore, if the service of Q1's I/O is delayed,
5969 * then Q2's I/O is delayed too. Conversely, if Q2's I/O is
5970 * delayed, the goal of Q1's I/O is hindered.
5971 * - as a consequence, if some I/O of Q1/Q2 arrives while
5972 * Q2/Q1 is the only queue in service, there is absolutely
5973 * no point in delaying the service of such an I/O. The
5974 * only possible result is a throughput loss
5975 * - so, when the above condition holds, the best option is to
5976 * have the new I/O dispatched as soon as possible
5977 * - the most effective and efficient way to attain the above
5978 * goal is to put the new I/O directly in the dispatch
5980 * - as an additional restriction, Q1 and Q2 must be the only
5981 * busy queues for this commit to put the I/O of Q2/Q1 in
5982 * the dispatch list. This is necessary, because, if also
5983 * other queues are waiting for service, then putting new
5984 * I/O directly in the dispatch list may evidently cause a
5985 * violation of service guarantees for the other queues
5988 (bfqq != bfqd->in_service_queue &&
5989 bfqd->in_service_queue != NULL &&
5990 bfq_tot_busy_queues(bfqd) == 1 + bfq_bfqq_busy(bfqq) &&
5991 (bfqq->waker_bfqq == bfqd->in_service_queue ||
5992 bfqd->in_service_queue->waker_bfqq == bfqq)) || at_head) {
5994 list_add(&rq->queuelist, &bfqd->dispatch);
5996 list_add_tail(&rq->queuelist, &bfqd->dispatch);
5998 idle_timer_disabled = __bfq_insert_request(bfqd, rq);
6000 * Update bfqq, because, if a queue merge has occurred
6001 * in __bfq_insert_request, then rq has been
6002 * redirected into a new queue.
6006 if (rq_mergeable(rq)) {
6007 elv_rqhash_add(q, rq);
6014 * Cache cmd_flags before releasing scheduler lock, because rq
6015 * may disappear afterwards (for example, because of a request
6018 cmd_flags = rq->cmd_flags;
6020 spin_unlock_irq(&bfqd->lock);
6022 bfq_update_insert_stats(q, bfqq, idle_timer_disabled,
6026 static void bfq_insert_requests(struct blk_mq_hw_ctx *hctx,
6027 struct list_head *list, bool at_head)
6029 while (!list_empty(list)) {
6032 rq = list_first_entry(list, struct request, queuelist);
6033 list_del_init(&rq->queuelist);
6034 bfq_insert_request(hctx, rq, at_head);
6038 static void bfq_update_hw_tag(struct bfq_data *bfqd)
6040 struct bfq_queue *bfqq = bfqd->in_service_queue;
6042 bfqd->max_rq_in_driver = max_t(int, bfqd->max_rq_in_driver,
6043 bfqd->rq_in_driver);
6045 if (bfqd->hw_tag == 1)
6049 * This sample is valid if the number of outstanding requests
6050 * is large enough to allow a queueing behavior. Note that the
6051 * sum is not exact, as it's not taking into account deactivated
6054 if (bfqd->rq_in_driver + bfqd->queued <= BFQ_HW_QUEUE_THRESHOLD)
6058 * If active queue hasn't enough requests and can idle, bfq might not
6059 * dispatch sufficient requests to hardware. Don't zero hw_tag in this
6062 if (bfqq && bfq_bfqq_has_short_ttime(bfqq) &&
6063 bfqq->dispatched + bfqq->queued[0] + bfqq->queued[1] <
6064 BFQ_HW_QUEUE_THRESHOLD &&
6065 bfqd->rq_in_driver < BFQ_HW_QUEUE_THRESHOLD)
6068 if (bfqd->hw_tag_samples++ < BFQ_HW_QUEUE_SAMPLES)
6071 bfqd->hw_tag = bfqd->max_rq_in_driver > BFQ_HW_QUEUE_THRESHOLD;
6072 bfqd->max_rq_in_driver = 0;
6073 bfqd->hw_tag_samples = 0;
6075 bfqd->nonrot_with_queueing =
6076 blk_queue_nonrot(bfqd->queue) && bfqd->hw_tag;
6079 static void bfq_completed_request(struct bfq_queue *bfqq, struct bfq_data *bfqd)
6084 bfq_update_hw_tag(bfqd);
6086 bfqd->rq_in_driver--;
6089 if (!bfqq->dispatched && !bfq_bfqq_busy(bfqq)) {
6091 * Set budget_timeout (which we overload to store the
6092 * time at which the queue remains with no backlog and
6093 * no outstanding request; used by the weight-raising
6096 bfqq->budget_timeout = jiffies;
6098 bfq_weights_tree_remove(bfqd, bfqq);
6101 now_ns = ktime_get_ns();
6103 bfqq->ttime.last_end_request = now_ns;
6106 * Using us instead of ns, to get a reasonable precision in
6107 * computing rate in next check.
6109 delta_us = div_u64(now_ns - bfqd->last_completion, NSEC_PER_USEC);
6112 * If the request took rather long to complete, and, according
6113 * to the maximum request size recorded, this completion latency
6114 * implies that the request was certainly served at a very low
6115 * rate (less than 1M sectors/sec), then the whole observation
6116 * interval that lasts up to this time instant cannot be a
6117 * valid time interval for computing a new peak rate. Invoke
6118 * bfq_update_rate_reset to have the following three steps
6120 * - close the observation interval at the last (previous)
6121 * request dispatch or completion
6122 * - compute rate, if possible, for that observation interval
6123 * - reset to zero samples, which will trigger a proper
6124 * re-initialization of the observation interval on next
6127 if (delta_us > BFQ_MIN_TT/NSEC_PER_USEC &&
6128 (bfqd->last_rq_max_size<<BFQ_RATE_SHIFT)/delta_us <
6129 1UL<<(BFQ_RATE_SHIFT - 10))
6130 bfq_update_rate_reset(bfqd, NULL);
6131 bfqd->last_completion = now_ns;
6133 * Shared queues are likely to receive I/O at a high
6134 * rate. This may deceptively let them be considered as wakers
6135 * of other queues. But a false waker will unjustly steal
6136 * bandwidth to its supposedly woken queue. So considering
6137 * also shared queues in the waking mechanism may cause more
6138 * control troubles than throughput benefits. Then reset
6139 * last_completed_rq_bfqq if bfqq is a shared queue.
6141 if (!bfq_bfqq_coop(bfqq))
6142 bfqd->last_completed_rq_bfqq = bfqq;
6144 bfqd->last_completed_rq_bfqq = NULL;
6147 * If we are waiting to discover whether the request pattern
6148 * of the task associated with the queue is actually
6149 * isochronous, and both requisites for this condition to hold
6150 * are now satisfied, then compute soft_rt_next_start (see the
6151 * comments on the function bfq_bfqq_softrt_next_start()). We
6152 * do not compute soft_rt_next_start if bfqq is in interactive
6153 * weight raising (see the comments in bfq_bfqq_expire() for
6154 * an explanation). We schedule this delayed update when bfqq
6155 * expires, if it still has in-flight requests.
6157 if (bfq_bfqq_softrt_update(bfqq) && bfqq->dispatched == 0 &&
6158 RB_EMPTY_ROOT(&bfqq->sort_list) &&
6159 bfqq->wr_coeff != bfqd->bfq_wr_coeff)
6160 bfqq->soft_rt_next_start =
6161 bfq_bfqq_softrt_next_start(bfqd, bfqq);
6164 * If this is the in-service queue, check if it needs to be expired,
6165 * or if we want to idle in case it has no pending requests.
6167 if (bfqd->in_service_queue == bfqq) {
6168 if (bfq_bfqq_must_idle(bfqq)) {
6169 if (bfqq->dispatched == 0)
6170 bfq_arm_slice_timer(bfqd);
6172 * If we get here, we do not expire bfqq, even
6173 * if bfqq was in budget timeout or had no
6174 * more requests (as controlled in the next
6175 * conditional instructions). The reason for
6176 * not expiring bfqq is as follows.
6178 * Here bfqq->dispatched > 0 holds, but
6179 * bfq_bfqq_must_idle() returned true. This
6180 * implies that, even if no request arrives
6181 * for bfqq before bfqq->dispatched reaches 0,
6182 * bfqq will, however, not be expired on the
6183 * completion event that causes bfqq->dispatch
6184 * to reach zero. In contrast, on this event,
6185 * bfqq will start enjoying device idling
6186 * (I/O-dispatch plugging).
6188 * But, if we expired bfqq here, bfqq would
6189 * not have the chance to enjoy device idling
6190 * when bfqq->dispatched finally reaches
6191 * zero. This would expose bfqq to violation
6192 * of its reserved service guarantees.
6195 } else if (bfq_may_expire_for_budg_timeout(bfqq))
6196 bfq_bfqq_expire(bfqd, bfqq, false,
6197 BFQQE_BUDGET_TIMEOUT);
6198 else if (RB_EMPTY_ROOT(&bfqq->sort_list) &&
6199 (bfqq->dispatched == 0 ||
6200 !bfq_better_to_idle(bfqq)))
6201 bfq_bfqq_expire(bfqd, bfqq, false,
6202 BFQQE_NO_MORE_REQUESTS);
6205 if (!bfqd->rq_in_driver)
6206 bfq_schedule_dispatch(bfqd);
6209 static void bfq_finish_requeue_request_body(struct bfq_queue *bfqq)
6213 bfq_put_queue(bfqq);
6217 * The processes associated with bfqq may happen to generate their
6218 * cumulative I/O at a lower rate than the rate at which the device
6219 * could serve the same I/O. This is rather probable, e.g., if only
6220 * one process is associated with bfqq and the device is an SSD. It
6221 * results in bfqq becoming often empty while in service. In this
6222 * respect, if BFQ is allowed to switch to another queue when bfqq
6223 * remains empty, then the device goes on being fed with I/O requests,
6224 * and the throughput is not affected. In contrast, if BFQ is not
6225 * allowed to switch to another queue---because bfqq is sync and
6226 * I/O-dispatch needs to be plugged while bfqq is temporarily
6227 * empty---then, during the service of bfqq, there will be frequent
6228 * "service holes", i.e., time intervals during which bfqq gets empty
6229 * and the device can only consume the I/O already queued in its
6230 * hardware queues. During service holes, the device may even get to
6231 * remaining idle. In the end, during the service of bfqq, the device
6232 * is driven at a lower speed than the one it can reach with the kind
6233 * of I/O flowing through bfqq.
6235 * To counter this loss of throughput, BFQ implements a "request
6236 * injection mechanism", which tries to fill the above service holes
6237 * with I/O requests taken from other queues. The hard part in this
6238 * mechanism is finding the right amount of I/O to inject, so as to
6239 * both boost throughput and not break bfqq's bandwidth and latency
6240 * guarantees. In this respect, the mechanism maintains a per-queue
6241 * inject limit, computed as below. While bfqq is empty, the injection
6242 * mechanism dispatches extra I/O requests only until the total number
6243 * of I/O requests in flight---i.e., already dispatched but not yet
6244 * completed---remains lower than this limit.
6246 * A first definition comes in handy to introduce the algorithm by
6247 * which the inject limit is computed. We define as first request for
6248 * bfqq, an I/O request for bfqq that arrives while bfqq is in
6249 * service, and causes bfqq to switch from empty to non-empty. The
6250 * algorithm updates the limit as a function of the effect of
6251 * injection on the service times of only the first requests of
6252 * bfqq. The reason for this restriction is that these are the
6253 * requests whose service time is affected most, because they are the
6254 * first to arrive after injection possibly occurred.
6256 * To evaluate the effect of injection, the algorithm measures the
6257 * "total service time" of first requests. We define as total service
6258 * time of an I/O request, the time that elapses since when the
6259 * request is enqueued into bfqq, to when it is completed. This
6260 * quantity allows the whole effect of injection to be measured. It is
6261 * easy to see why. Suppose that some requests of other queues are
6262 * actually injected while bfqq is empty, and that a new request R
6263 * then arrives for bfqq. If the device does start to serve all or
6264 * part of the injected requests during the service hole, then,
6265 * because of this extra service, it may delay the next invocation of
6266 * the dispatch hook of BFQ. Then, even after R gets eventually
6267 * dispatched, the device may delay the actual service of R if it is
6268 * still busy serving the extra requests, or if it decides to serve,
6269 * before R, some extra request still present in its queues. As a
6270 * conclusion, the cumulative extra delay caused by injection can be
6271 * easily evaluated by just comparing the total service time of first
6272 * requests with and without injection.
6274 * The limit-update algorithm works as follows. On the arrival of a
6275 * first request of bfqq, the algorithm measures the total time of the
6276 * request only if one of the three cases below holds, and, for each
6277 * case, it updates the limit as described below:
6279 * (1) If there is no in-flight request. This gives a baseline for the
6280 * total service time of the requests of bfqq. If the baseline has
6281 * not been computed yet, then, after computing it, the limit is
6282 * set to 1, to start boosting throughput, and to prepare the
6283 * ground for the next case. If the baseline has already been
6284 * computed, then it is updated, in case it results to be lower
6285 * than the previous value.
6287 * (2) If the limit is higher than 0 and there are in-flight
6288 * requests. By comparing the total service time in this case with
6289 * the above baseline, it is possible to know at which extent the
6290 * current value of the limit is inflating the total service
6291 * time. If the inflation is below a certain threshold, then bfqq
6292 * is assumed to be suffering from no perceivable loss of its
6293 * service guarantees, and the limit is even tentatively
6294 * increased. If the inflation is above the threshold, then the
6295 * limit is decreased. Due to the lack of any hysteresis, this
6296 * logic makes the limit oscillate even in steady workload
6297 * conditions. Yet we opted for it, because it is fast in reaching
6298 * the best value for the limit, as a function of the current I/O
6299 * workload. To reduce oscillations, this step is disabled for a
6300 * short time interval after the limit happens to be decreased.
6302 * (3) Periodically, after resetting the limit, to make sure that the
6303 * limit eventually drops in case the workload changes. This is
6304 * needed because, after the limit has gone safely up for a
6305 * certain workload, it is impossible to guess whether the
6306 * baseline total service time may have changed, without measuring
6307 * it again without injection. A more effective version of this
6308 * step might be to just sample the baseline, by interrupting
6309 * injection only once, and then to reset/lower the limit only if
6310 * the total service time with the current limit does happen to be
6313 * More details on each step are provided in the comments on the
6314 * pieces of code that implement these steps: the branch handling the
6315 * transition from empty to non empty in bfq_add_request(), the branch
6316 * handling injection in bfq_select_queue(), and the function
6317 * bfq_choose_bfqq_for_injection(). These comments also explain some
6318 * exceptions, made by the injection mechanism in some special cases.
6320 static void bfq_update_inject_limit(struct bfq_data *bfqd,
6321 struct bfq_queue *bfqq)
6323 u64 tot_time_ns = ktime_get_ns() - bfqd->last_empty_occupied_ns;
6324 unsigned int old_limit = bfqq->inject_limit;
6326 if (bfqq->last_serv_time_ns > 0 && bfqd->rqs_injected) {
6327 u64 threshold = (bfqq->last_serv_time_ns * 3)>>1;
6329 if (tot_time_ns >= threshold && old_limit > 0) {
6330 bfqq->inject_limit--;
6331 bfqq->decrease_time_jif = jiffies;
6332 } else if (tot_time_ns < threshold &&
6333 old_limit <= bfqd->max_rq_in_driver)
6334 bfqq->inject_limit++;
6338 * Either we still have to compute the base value for the
6339 * total service time, and there seem to be the right
6340 * conditions to do it, or we can lower the last base value
6343 * NOTE: (bfqd->rq_in_driver == 1) means that there is no I/O
6344 * request in flight, because this function is in the code
6345 * path that handles the completion of a request of bfqq, and,
6346 * in particular, this function is executed before
6347 * bfqd->rq_in_driver is decremented in such a code path.
6349 if ((bfqq->last_serv_time_ns == 0 && bfqd->rq_in_driver == 1) ||
6350 tot_time_ns < bfqq->last_serv_time_ns) {
6351 if (bfqq->last_serv_time_ns == 0) {
6353 * Now we certainly have a base value: make sure we
6354 * start trying injection.
6356 bfqq->inject_limit = max_t(unsigned int, 1, old_limit);
6358 bfqq->last_serv_time_ns = tot_time_ns;
6359 } else if (!bfqd->rqs_injected && bfqd->rq_in_driver == 1)
6361 * No I/O injected and no request still in service in
6362 * the drive: these are the exact conditions for
6363 * computing the base value of the total service time
6364 * for bfqq. So let's update this value, because it is
6365 * rather variable. For example, it varies if the size
6366 * or the spatial locality of the I/O requests in bfqq
6369 bfqq->last_serv_time_ns = tot_time_ns;
6372 /* update complete, not waiting for any request completion any longer */
6373 bfqd->waited_rq = NULL;
6374 bfqd->rqs_injected = false;
6378 * Handle either a requeue or a finish for rq. The things to do are
6379 * the same in both cases: all references to rq are to be dropped. In
6380 * particular, rq is considered completed from the point of view of
6383 static void bfq_finish_requeue_request(struct request *rq)
6385 struct bfq_queue *bfqq = RQ_BFQQ(rq);
6386 struct bfq_data *bfqd;
6389 * rq either is not associated with any icq, or is an already
6390 * requeued request that has not (yet) been re-inserted into
6393 if (!rq->elv.icq || !bfqq)
6398 if (rq->rq_flags & RQF_STARTED)
6399 bfqg_stats_update_completion(bfqq_group(bfqq),
6401 rq->io_start_time_ns,
6404 if (likely(rq->rq_flags & RQF_STARTED)) {
6405 unsigned long flags;
6407 spin_lock_irqsave(&bfqd->lock, flags);
6409 if (rq == bfqd->waited_rq)
6410 bfq_update_inject_limit(bfqd, bfqq);
6412 bfq_completed_request(bfqq, bfqd);
6413 bfq_finish_requeue_request_body(bfqq);
6415 spin_unlock_irqrestore(&bfqd->lock, flags);
6418 * Request rq may be still/already in the scheduler,
6419 * in which case we need to remove it (this should
6420 * never happen in case of requeue). And we cannot
6421 * defer such a check and removal, to avoid
6422 * inconsistencies in the time interval from the end
6423 * of this function to the start of the deferred work.
6424 * This situation seems to occur only in process
6425 * context, as a consequence of a merge. In the
6426 * current version of the code, this implies that the
6430 if (!RB_EMPTY_NODE(&rq->rb_node)) {
6431 bfq_remove_request(rq->q, rq);
6432 bfqg_stats_update_io_remove(bfqq_group(bfqq),
6435 bfq_finish_requeue_request_body(bfqq);
6439 * Reset private fields. In case of a requeue, this allows
6440 * this function to correctly do nothing if it is spuriously
6441 * invoked again on this same request (see the check at the
6442 * beginning of the function). Probably, a better general
6443 * design would be to prevent blk-mq from invoking the requeue
6444 * or finish hooks of an elevator, for a request that is not
6445 * referred by that elevator.
6447 * Resetting the following fields would break the
6448 * request-insertion logic if rq is re-inserted into a bfq
6449 * internal queue, without a re-preparation. Here we assume
6450 * that re-insertions of requeued requests, without
6451 * re-preparation, can happen only for pass_through or at_head
6452 * requests (which are not re-inserted into bfq internal
6455 rq->elv.priv[0] = NULL;
6456 rq->elv.priv[1] = NULL;
6460 * Removes the association between the current task and bfqq, assuming
6461 * that bic points to the bfq iocontext of the task.
6462 * Returns NULL if a new bfqq should be allocated, or the old bfqq if this
6463 * was the last process referring to that bfqq.
6465 static struct bfq_queue *
6466 bfq_split_bfqq(struct bfq_io_cq *bic, struct bfq_queue *bfqq)
6468 bfq_log_bfqq(bfqq->bfqd, bfqq, "splitting queue");
6470 if (bfqq_process_refs(bfqq) == 1) {
6471 bfqq->pid = current->pid;
6472 bfq_clear_bfqq_coop(bfqq);
6473 bfq_clear_bfqq_split_coop(bfqq);
6477 bic_set_bfqq(bic, NULL, 1);
6479 bfq_put_cooperator(bfqq);
6481 bfq_release_process_ref(bfqq->bfqd, bfqq);
6485 static struct bfq_queue *bfq_get_bfqq_handle_split(struct bfq_data *bfqd,
6486 struct bfq_io_cq *bic,
6488 bool split, bool is_sync,
6491 struct bfq_queue *bfqq = bic_to_bfqq(bic, is_sync);
6493 if (likely(bfqq && bfqq != &bfqd->oom_bfqq))
6500 bfq_put_queue(bfqq);
6501 bfqq = bfq_get_queue(bfqd, bio, is_sync, bic, split);
6503 bic_set_bfqq(bic, bfqq, is_sync);
6504 if (split && is_sync) {
6505 if ((bic->was_in_burst_list && bfqd->large_burst) ||
6506 bic->saved_in_large_burst)
6507 bfq_mark_bfqq_in_large_burst(bfqq);
6509 bfq_clear_bfqq_in_large_burst(bfqq);
6510 if (bic->was_in_burst_list)
6512 * If bfqq was in the current
6513 * burst list before being
6514 * merged, then we have to add
6515 * it back. And we do not need
6516 * to increase burst_size, as
6517 * we did not decrement
6518 * burst_size when we removed
6519 * bfqq from the burst list as
6520 * a consequence of a merge
6522 * bfq_put_queue). In this
6523 * respect, it would be rather
6524 * costly to know whether the
6525 * current burst list is still
6526 * the same burst list from
6527 * which bfqq was removed on
6528 * the merge. To avoid this
6529 * cost, if bfqq was in a
6530 * burst list, then we add
6531 * bfqq to the current burst
6532 * list without any further
6533 * check. This can cause
6534 * inappropriate insertions,
6535 * but rarely enough to not
6536 * harm the detection of large
6537 * bursts significantly.
6539 hlist_add_head(&bfqq->burst_list_node,
6542 bfqq->split_time = jiffies;
6549 * Only reset private fields. The actual request preparation will be
6550 * performed by bfq_init_rq, when rq is either inserted or merged. See
6551 * comments on bfq_init_rq for the reason behind this delayed
6554 static void bfq_prepare_request(struct request *rq)
6557 * Regardless of whether we have an icq attached, we have to
6558 * clear the scheduler pointers, as they might point to
6559 * previously allocated bic/bfqq structs.
6561 rq->elv.priv[0] = rq->elv.priv[1] = NULL;
6565 * If needed, init rq, allocate bfq data structures associated with
6566 * rq, and increment reference counters in the destination bfq_queue
6567 * for rq. Return the destination bfq_queue for rq, or NULL is rq is
6568 * not associated with any bfq_queue.
6570 * This function is invoked by the functions that perform rq insertion
6571 * or merging. One may have expected the above preparation operations
6572 * to be performed in bfq_prepare_request, and not delayed to when rq
6573 * is inserted or merged. The rationale behind this delayed
6574 * preparation is that, after the prepare_request hook is invoked for
6575 * rq, rq may still be transformed into a request with no icq, i.e., a
6576 * request not associated with any queue. No bfq hook is invoked to
6577 * signal this transformation. As a consequence, should these
6578 * preparation operations be performed when the prepare_request hook
6579 * is invoked, and should rq be transformed one moment later, bfq
6580 * would end up in an inconsistent state, because it would have
6581 * incremented some queue counters for an rq destined to
6582 * transformation, without any chance to correctly lower these
6583 * counters back. In contrast, no transformation can still happen for
6584 * rq after rq has been inserted or merged. So, it is safe to execute
6585 * these preparation operations when rq is finally inserted or merged.
6587 static struct bfq_queue *bfq_init_rq(struct request *rq)
6589 struct request_queue *q = rq->q;
6590 struct bio *bio = rq->bio;
6591 struct bfq_data *bfqd = q->elevator->elevator_data;
6592 struct bfq_io_cq *bic;
6593 const int is_sync = rq_is_sync(rq);
6594 struct bfq_queue *bfqq;
6595 bool new_queue = false;
6596 bool bfqq_already_existing = false, split = false;
6598 if (unlikely(!rq->elv.icq))
6602 * Assuming that elv.priv[1] is set only if everything is set
6603 * for this rq. This holds true, because this function is
6604 * invoked only for insertion or merging, and, after such
6605 * events, a request cannot be manipulated any longer before
6606 * being removed from bfq.
6608 if (rq->elv.priv[1])
6609 return rq->elv.priv[1];
6611 bic = icq_to_bic(rq->elv.icq);
6613 bfq_check_ioprio_change(bic, bio);
6615 bfq_bic_update_cgroup(bic, bio);
6617 bfqq = bfq_get_bfqq_handle_split(bfqd, bic, bio, false, is_sync,
6620 if (likely(!new_queue)) {
6621 /* If the queue was seeky for too long, break it apart. */
6622 if (bfq_bfqq_coop(bfqq) && bfq_bfqq_split_coop(bfqq) &&
6623 !bic->stably_merged) {
6624 struct bfq_queue *old_bfqq = bfqq;
6626 /* Update bic before losing reference to bfqq */
6627 if (bfq_bfqq_in_large_burst(bfqq))
6628 bic->saved_in_large_burst = true;
6630 bfqq = bfq_split_bfqq(bic, bfqq);
6634 bfqq = bfq_get_bfqq_handle_split(bfqd, bic, bio,
6637 bfqq->waker_bfqq = old_bfqq->waker_bfqq;
6638 bfqq->tentative_waker_bfqq = NULL;
6641 * If the waker queue disappears, then
6642 * new_bfqq->waker_bfqq must be
6643 * reset. So insert new_bfqq into the
6644 * woken_list of the waker. See
6645 * bfq_check_waker for details.
6647 if (bfqq->waker_bfqq)
6648 hlist_add_head(&bfqq->woken_list_node,
6649 &bfqq->waker_bfqq->woken_list);
6651 bfqq_already_existing = true;
6657 bfq_log_bfqq(bfqd, bfqq, "get_request %p: bfqq %p, %d",
6658 rq, bfqq, bfqq->ref);
6660 rq->elv.priv[0] = bic;
6661 rq->elv.priv[1] = bfqq;
6664 * If a bfq_queue has only one process reference, it is owned
6665 * by only this bic: we can then set bfqq->bic = bic. in
6666 * addition, if the queue has also just been split, we have to
6669 if (likely(bfqq != &bfqd->oom_bfqq) && bfqq_process_refs(bfqq) == 1) {
6673 * The queue has just been split from a shared
6674 * queue: restore the idle window and the
6675 * possible weight raising period.
6677 bfq_bfqq_resume_state(bfqq, bfqd, bic,
6678 bfqq_already_existing);
6683 * Consider bfqq as possibly belonging to a burst of newly
6684 * created queues only if:
6685 * 1) A burst is actually happening (bfqd->burst_size > 0)
6687 * 2) There is no other active queue. In fact, if, in
6688 * contrast, there are active queues not belonging to the
6689 * possible burst bfqq may belong to, then there is no gain
6690 * in considering bfqq as belonging to a burst, and
6691 * therefore in not weight-raising bfqq. See comments on
6692 * bfq_handle_burst().
6694 * This filtering also helps eliminating false positives,
6695 * occurring when bfqq does not belong to an actual large
6696 * burst, but some background task (e.g., a service) happens
6697 * to trigger the creation of new queues very close to when
6698 * bfqq and its possible companion queues are created. See
6699 * comments on bfq_handle_burst() for further details also on
6702 if (unlikely(bfq_bfqq_just_created(bfqq) &&
6703 (bfqd->burst_size > 0 ||
6704 bfq_tot_busy_queues(bfqd) == 0)))
6705 bfq_handle_burst(bfqd, bfqq);
6711 bfq_idle_slice_timer_body(struct bfq_data *bfqd, struct bfq_queue *bfqq)
6713 enum bfqq_expiration reason;
6714 unsigned long flags;
6716 spin_lock_irqsave(&bfqd->lock, flags);
6719 * Considering that bfqq may be in race, we should firstly check
6720 * whether bfqq is in service before doing something on it. If
6721 * the bfqq in race is not in service, it has already been expired
6722 * through __bfq_bfqq_expire func and its wait_request flags has
6723 * been cleared in __bfq_bfqd_reset_in_service func.
6725 if (bfqq != bfqd->in_service_queue) {
6726 spin_unlock_irqrestore(&bfqd->lock, flags);
6730 bfq_clear_bfqq_wait_request(bfqq);
6732 if (bfq_bfqq_budget_timeout(bfqq))
6734 * Also here the queue can be safely expired
6735 * for budget timeout without wasting
6738 reason = BFQQE_BUDGET_TIMEOUT;
6739 else if (bfqq->queued[0] == 0 && bfqq->queued[1] == 0)
6741 * The queue may not be empty upon timer expiration,
6742 * because we may not disable the timer when the
6743 * first request of the in-service queue arrives
6744 * during disk idling.
6746 reason = BFQQE_TOO_IDLE;
6748 goto schedule_dispatch;
6750 bfq_bfqq_expire(bfqd, bfqq, true, reason);
6753 spin_unlock_irqrestore(&bfqd->lock, flags);
6754 bfq_schedule_dispatch(bfqd);
6758 * Handler of the expiration of the timer running if the in-service queue
6759 * is idling inside its time slice.
6761 static enum hrtimer_restart bfq_idle_slice_timer(struct hrtimer *timer)
6763 struct bfq_data *bfqd = container_of(timer, struct bfq_data,
6765 struct bfq_queue *bfqq = bfqd->in_service_queue;
6768 * Theoretical race here: the in-service queue can be NULL or
6769 * different from the queue that was idling if a new request
6770 * arrives for the current queue and there is a full dispatch
6771 * cycle that changes the in-service queue. This can hardly
6772 * happen, but in the worst case we just expire a queue too
6776 bfq_idle_slice_timer_body(bfqd, bfqq);
6778 return HRTIMER_NORESTART;
6781 static void __bfq_put_async_bfqq(struct bfq_data *bfqd,
6782 struct bfq_queue **bfqq_ptr)
6784 struct bfq_queue *bfqq = *bfqq_ptr;
6786 bfq_log(bfqd, "put_async_bfqq: %p", bfqq);
6788 bfq_bfqq_move(bfqd, bfqq, bfqd->root_group);
6790 bfq_log_bfqq(bfqd, bfqq, "put_async_bfqq: putting %p, %d",
6792 bfq_put_queue(bfqq);
6798 * Release all the bfqg references to its async queues. If we are
6799 * deallocating the group these queues may still contain requests, so
6800 * we reparent them to the root cgroup (i.e., the only one that will
6801 * exist for sure until all the requests on a device are gone).
6803 void bfq_put_async_queues(struct bfq_data *bfqd, struct bfq_group *bfqg)
6807 for (i = 0; i < 2; i++)
6808 for (j = 0; j < IOPRIO_BE_NR; j++)
6809 __bfq_put_async_bfqq(bfqd, &bfqg->async_bfqq[i][j]);
6811 __bfq_put_async_bfqq(bfqd, &bfqg->async_idle_bfqq);
6815 * See the comments on bfq_limit_depth for the purpose of
6816 * the depths set in the function. Return minimum shallow depth we'll use.
6818 static unsigned int bfq_update_depths(struct bfq_data *bfqd,
6819 struct sbitmap_queue *bt)
6821 unsigned int i, j, min_shallow = UINT_MAX;
6824 * In-word depths if no bfq_queue is being weight-raised:
6825 * leaving 25% of tags only for sync reads.
6827 * In next formulas, right-shift the value
6828 * (1U<<bt->sb.shift), instead of computing directly
6829 * (1U<<(bt->sb.shift - something)), to be robust against
6830 * any possible value of bt->sb.shift, without having to
6831 * limit 'something'.
6833 /* no more than 50% of tags for async I/O */
6834 bfqd->word_depths[0][0] = max((1U << bt->sb.shift) >> 1, 1U);
6836 * no more than 75% of tags for sync writes (25% extra tags
6837 * w.r.t. async I/O, to prevent async I/O from starving sync
6840 bfqd->word_depths[0][1] = max(((1U << bt->sb.shift) * 3) >> 2, 1U);
6843 * In-word depths in case some bfq_queue is being weight-
6844 * raised: leaving ~63% of tags for sync reads. This is the
6845 * highest percentage for which, in our tests, application
6846 * start-up times didn't suffer from any regression due to tag
6849 /* no more than ~18% of tags for async I/O */
6850 bfqd->word_depths[1][0] = max(((1U << bt->sb.shift) * 3) >> 4, 1U);
6851 /* no more than ~37% of tags for sync writes (~20% extra tags) */
6852 bfqd->word_depths[1][1] = max(((1U << bt->sb.shift) * 6) >> 4, 1U);
6854 for (i = 0; i < 2; i++)
6855 for (j = 0; j < 2; j++)
6856 min_shallow = min(min_shallow, bfqd->word_depths[i][j]);
6861 static void bfq_depth_updated(struct blk_mq_hw_ctx *hctx)
6863 struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
6864 struct blk_mq_tags *tags = hctx->sched_tags;
6865 unsigned int min_shallow;
6867 min_shallow = bfq_update_depths(bfqd, tags->bitmap_tags);
6868 sbitmap_queue_min_shallow_depth(tags->bitmap_tags, min_shallow);
6871 static int bfq_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int index)
6873 bfq_depth_updated(hctx);
6877 static void bfq_exit_queue(struct elevator_queue *e)
6879 struct bfq_data *bfqd = e->elevator_data;
6880 struct bfq_queue *bfqq, *n;
6882 hrtimer_cancel(&bfqd->idle_slice_timer);
6884 spin_lock_irq(&bfqd->lock);
6885 list_for_each_entry_safe(bfqq, n, &bfqd->idle_list, bfqq_list)
6886 bfq_deactivate_bfqq(bfqd, bfqq, false, false);
6887 spin_unlock_irq(&bfqd->lock);
6889 hrtimer_cancel(&bfqd->idle_slice_timer);
6891 /* release oom-queue reference to root group */
6892 bfqg_and_blkg_put(bfqd->root_group);
6894 #ifdef CONFIG_BFQ_GROUP_IOSCHED
6895 blkcg_deactivate_policy(bfqd->queue, &blkcg_policy_bfq);
6897 spin_lock_irq(&bfqd->lock);
6898 bfq_put_async_queues(bfqd, bfqd->root_group);
6899 kfree(bfqd->root_group);
6900 spin_unlock_irq(&bfqd->lock);
6906 static void bfq_init_root_group(struct bfq_group *root_group,
6907 struct bfq_data *bfqd)
6911 #ifdef CONFIG_BFQ_GROUP_IOSCHED
6912 root_group->entity.parent = NULL;
6913 root_group->my_entity = NULL;
6914 root_group->bfqd = bfqd;
6916 root_group->rq_pos_tree = RB_ROOT;
6917 for (i = 0; i < BFQ_IOPRIO_CLASSES; i++)
6918 root_group->sched_data.service_tree[i] = BFQ_SERVICE_TREE_INIT;
6919 root_group->sched_data.bfq_class_idle_last_service = jiffies;
6922 static int bfq_init_queue(struct request_queue *q, struct elevator_type *e)
6924 struct bfq_data *bfqd;
6925 struct elevator_queue *eq;
6927 eq = elevator_alloc(q, e);
6931 bfqd = kzalloc_node(sizeof(*bfqd), GFP_KERNEL, q->node);
6933 kobject_put(&eq->kobj);
6936 eq->elevator_data = bfqd;
6938 spin_lock_irq(&q->queue_lock);
6940 spin_unlock_irq(&q->queue_lock);
6943 * Our fallback bfqq if bfq_find_alloc_queue() runs into OOM issues.
6944 * Grab a permanent reference to it, so that the normal code flow
6945 * will not attempt to free it.
6947 bfq_init_bfqq(bfqd, &bfqd->oom_bfqq, NULL, 1, 0);
6948 bfqd->oom_bfqq.ref++;
6949 bfqd->oom_bfqq.new_ioprio = BFQ_DEFAULT_QUEUE_IOPRIO;
6950 bfqd->oom_bfqq.new_ioprio_class = IOPRIO_CLASS_BE;
6951 bfqd->oom_bfqq.entity.new_weight =
6952 bfq_ioprio_to_weight(bfqd->oom_bfqq.new_ioprio);
6954 /* oom_bfqq does not participate to bursts */
6955 bfq_clear_bfqq_just_created(&bfqd->oom_bfqq);
6958 * Trigger weight initialization, according to ioprio, at the
6959 * oom_bfqq's first activation. The oom_bfqq's ioprio and ioprio
6960 * class won't be changed any more.
6962 bfqd->oom_bfqq.entity.prio_changed = 1;
6966 INIT_LIST_HEAD(&bfqd->dispatch);
6968 hrtimer_init(&bfqd->idle_slice_timer, CLOCK_MONOTONIC,
6970 bfqd->idle_slice_timer.function = bfq_idle_slice_timer;
6972 bfqd->queue_weights_tree = RB_ROOT_CACHED;
6973 bfqd->num_groups_with_pending_reqs = 0;
6975 INIT_LIST_HEAD(&bfqd->active_list);
6976 INIT_LIST_HEAD(&bfqd->idle_list);
6977 INIT_HLIST_HEAD(&bfqd->burst_list);
6980 bfqd->nonrot_with_queueing = blk_queue_nonrot(bfqd->queue);
6982 bfqd->bfq_max_budget = bfq_default_max_budget;
6984 bfqd->bfq_fifo_expire[0] = bfq_fifo_expire[0];
6985 bfqd->bfq_fifo_expire[1] = bfq_fifo_expire[1];
6986 bfqd->bfq_back_max = bfq_back_max;
6987 bfqd->bfq_back_penalty = bfq_back_penalty;
6988 bfqd->bfq_slice_idle = bfq_slice_idle;
6989 bfqd->bfq_timeout = bfq_timeout;
6991 bfqd->bfq_large_burst_thresh = 8;
6992 bfqd->bfq_burst_interval = msecs_to_jiffies(180);
6994 bfqd->low_latency = true;
6997 * Trade-off between responsiveness and fairness.
6999 bfqd->bfq_wr_coeff = 30;
7000 bfqd->bfq_wr_rt_max_time = msecs_to_jiffies(300);
7001 bfqd->bfq_wr_max_time = 0;
7002 bfqd->bfq_wr_min_idle_time = msecs_to_jiffies(2000);
7003 bfqd->bfq_wr_min_inter_arr_async = msecs_to_jiffies(500);
7004 bfqd->bfq_wr_max_softrt_rate = 7000; /*
7005 * Approximate rate required
7006 * to playback or record a
7007 * high-definition compressed
7010 bfqd->wr_busy_queues = 0;
7013 * Begin by assuming, optimistically, that the device peak
7014 * rate is equal to 2/3 of the highest reference rate.
7016 bfqd->rate_dur_prod = ref_rate[blk_queue_nonrot(bfqd->queue)] *
7017 ref_wr_duration[blk_queue_nonrot(bfqd->queue)];
7018 bfqd->peak_rate = ref_rate[blk_queue_nonrot(bfqd->queue)] * 2 / 3;
7020 spin_lock_init(&bfqd->lock);
7023 * The invocation of the next bfq_create_group_hierarchy
7024 * function is the head of a chain of function calls
7025 * (bfq_create_group_hierarchy->blkcg_activate_policy->
7026 * blk_mq_freeze_queue) that may lead to the invocation of the
7027 * has_work hook function. For this reason,
7028 * bfq_create_group_hierarchy is invoked only after all
7029 * scheduler data has been initialized, apart from the fields
7030 * that can be initialized only after invoking
7031 * bfq_create_group_hierarchy. This, in particular, enables
7032 * has_work to correctly return false. Of course, to avoid
7033 * other inconsistencies, the blk-mq stack must then refrain
7034 * from invoking further scheduler hooks before this init
7035 * function is finished.
7037 bfqd->root_group = bfq_create_group_hierarchy(bfqd, q->node);
7038 if (!bfqd->root_group)
7040 bfq_init_root_group(bfqd->root_group, bfqd);
7041 bfq_init_entity(&bfqd->oom_bfqq.entity, bfqd->root_group);
7043 wbt_disable_default(q);
7048 kobject_put(&eq->kobj);
7052 static void bfq_slab_kill(void)
7054 kmem_cache_destroy(bfq_pool);
7057 static int __init bfq_slab_setup(void)
7059 bfq_pool = KMEM_CACHE(bfq_queue, 0);
7065 static ssize_t bfq_var_show(unsigned int var, char *page)
7067 return sprintf(page, "%u\n", var);
7070 static int bfq_var_store(unsigned long *var, const char *page)
7072 unsigned long new_val;
7073 int ret = kstrtoul(page, 10, &new_val);
7081 #define SHOW_FUNCTION(__FUNC, __VAR, __CONV) \
7082 static ssize_t __FUNC(struct elevator_queue *e, char *page) \
7084 struct bfq_data *bfqd = e->elevator_data; \
7085 u64 __data = __VAR; \
7087 __data = jiffies_to_msecs(__data); \
7088 else if (__CONV == 2) \
7089 __data = div_u64(__data, NSEC_PER_MSEC); \
7090 return bfq_var_show(__data, (page)); \
7092 SHOW_FUNCTION(bfq_fifo_expire_sync_show, bfqd->bfq_fifo_expire[1], 2);
7093 SHOW_FUNCTION(bfq_fifo_expire_async_show, bfqd->bfq_fifo_expire[0], 2);
7094 SHOW_FUNCTION(bfq_back_seek_max_show, bfqd->bfq_back_max, 0);
7095 SHOW_FUNCTION(bfq_back_seek_penalty_show, bfqd->bfq_back_penalty, 0);
7096 SHOW_FUNCTION(bfq_slice_idle_show, bfqd->bfq_slice_idle, 2);
7097 SHOW_FUNCTION(bfq_max_budget_show, bfqd->bfq_user_max_budget, 0);
7098 SHOW_FUNCTION(bfq_timeout_sync_show, bfqd->bfq_timeout, 1);
7099 SHOW_FUNCTION(bfq_strict_guarantees_show, bfqd->strict_guarantees, 0);
7100 SHOW_FUNCTION(bfq_low_latency_show, bfqd->low_latency, 0);
7101 #undef SHOW_FUNCTION
7103 #define USEC_SHOW_FUNCTION(__FUNC, __VAR) \
7104 static ssize_t __FUNC(struct elevator_queue *e, char *page) \
7106 struct bfq_data *bfqd = e->elevator_data; \
7107 u64 __data = __VAR; \
7108 __data = div_u64(__data, NSEC_PER_USEC); \
7109 return bfq_var_show(__data, (page)); \
7111 USEC_SHOW_FUNCTION(bfq_slice_idle_us_show, bfqd->bfq_slice_idle);
7112 #undef USEC_SHOW_FUNCTION
7114 #define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX, __CONV) \
7116 __FUNC(struct elevator_queue *e, const char *page, size_t count) \
7118 struct bfq_data *bfqd = e->elevator_data; \
7119 unsigned long __data, __min = (MIN), __max = (MAX); \
7122 ret = bfq_var_store(&__data, (page)); \
7125 if (__data < __min) \
7127 else if (__data > __max) \
7130 *(__PTR) = msecs_to_jiffies(__data); \
7131 else if (__CONV == 2) \
7132 *(__PTR) = (u64)__data * NSEC_PER_MSEC; \
7134 *(__PTR) = __data; \
7137 STORE_FUNCTION(bfq_fifo_expire_sync_store, &bfqd->bfq_fifo_expire[1], 1,
7139 STORE_FUNCTION(bfq_fifo_expire_async_store, &bfqd->bfq_fifo_expire[0], 1,
7141 STORE_FUNCTION(bfq_back_seek_max_store, &bfqd->bfq_back_max, 0, INT_MAX, 0);
7142 STORE_FUNCTION(bfq_back_seek_penalty_store, &bfqd->bfq_back_penalty, 1,
7144 STORE_FUNCTION(bfq_slice_idle_store, &bfqd->bfq_slice_idle, 0, INT_MAX, 2);
7145 #undef STORE_FUNCTION
7147 #define USEC_STORE_FUNCTION(__FUNC, __PTR, MIN, MAX) \
7148 static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count)\
7150 struct bfq_data *bfqd = e->elevator_data; \
7151 unsigned long __data, __min = (MIN), __max = (MAX); \
7154 ret = bfq_var_store(&__data, (page)); \
7157 if (__data < __min) \
7159 else if (__data > __max) \
7161 *(__PTR) = (u64)__data * NSEC_PER_USEC; \
7164 USEC_STORE_FUNCTION(bfq_slice_idle_us_store, &bfqd->bfq_slice_idle, 0,
7166 #undef USEC_STORE_FUNCTION
7168 static ssize_t bfq_max_budget_store(struct elevator_queue *e,
7169 const char *page, size_t count)
7171 struct bfq_data *bfqd = e->elevator_data;
7172 unsigned long __data;
7175 ret = bfq_var_store(&__data, (page));
7180 bfqd->bfq_max_budget = bfq_calc_max_budget(bfqd);
7182 if (__data > INT_MAX)
7184 bfqd->bfq_max_budget = __data;
7187 bfqd->bfq_user_max_budget = __data;
7193 * Leaving this name to preserve name compatibility with cfq
7194 * parameters, but this timeout is used for both sync and async.
7196 static ssize_t bfq_timeout_sync_store(struct elevator_queue *e,
7197 const char *page, size_t count)
7199 struct bfq_data *bfqd = e->elevator_data;
7200 unsigned long __data;
7203 ret = bfq_var_store(&__data, (page));
7209 else if (__data > INT_MAX)
7212 bfqd->bfq_timeout = msecs_to_jiffies(__data);
7213 if (bfqd->bfq_user_max_budget == 0)
7214 bfqd->bfq_max_budget = bfq_calc_max_budget(bfqd);
7219 static ssize_t bfq_strict_guarantees_store(struct elevator_queue *e,
7220 const char *page, size_t count)
7222 struct bfq_data *bfqd = e->elevator_data;
7223 unsigned long __data;
7226 ret = bfq_var_store(&__data, (page));
7232 if (!bfqd->strict_guarantees && __data == 1
7233 && bfqd->bfq_slice_idle < 8 * NSEC_PER_MSEC)
7234 bfqd->bfq_slice_idle = 8 * NSEC_PER_MSEC;
7236 bfqd->strict_guarantees = __data;
7241 static ssize_t bfq_low_latency_store(struct elevator_queue *e,
7242 const char *page, size_t count)
7244 struct bfq_data *bfqd = e->elevator_data;
7245 unsigned long __data;
7248 ret = bfq_var_store(&__data, (page));
7254 if (__data == 0 && bfqd->low_latency != 0)
7256 bfqd->low_latency = __data;
7261 #define BFQ_ATTR(name) \
7262 __ATTR(name, 0644, bfq_##name##_show, bfq_##name##_store)
7264 static struct elv_fs_entry bfq_attrs[] = {
7265 BFQ_ATTR(fifo_expire_sync),
7266 BFQ_ATTR(fifo_expire_async),
7267 BFQ_ATTR(back_seek_max),
7268 BFQ_ATTR(back_seek_penalty),
7269 BFQ_ATTR(slice_idle),
7270 BFQ_ATTR(slice_idle_us),
7271 BFQ_ATTR(max_budget),
7272 BFQ_ATTR(timeout_sync),
7273 BFQ_ATTR(strict_guarantees),
7274 BFQ_ATTR(low_latency),
7278 static struct elevator_type iosched_bfq_mq = {
7280 .limit_depth = bfq_limit_depth,
7281 .prepare_request = bfq_prepare_request,
7282 .requeue_request = bfq_finish_requeue_request,
7283 .finish_request = bfq_finish_requeue_request,
7284 .exit_icq = bfq_exit_icq,
7285 .insert_requests = bfq_insert_requests,
7286 .dispatch_request = bfq_dispatch_request,
7287 .next_request = elv_rb_latter_request,
7288 .former_request = elv_rb_former_request,
7289 .allow_merge = bfq_allow_bio_merge,
7290 .bio_merge = bfq_bio_merge,
7291 .request_merge = bfq_request_merge,
7292 .requests_merged = bfq_requests_merged,
7293 .request_merged = bfq_request_merged,
7294 .has_work = bfq_has_work,
7295 .depth_updated = bfq_depth_updated,
7296 .init_hctx = bfq_init_hctx,
7297 .init_sched = bfq_init_queue,
7298 .exit_sched = bfq_exit_queue,
7301 .icq_size = sizeof(struct bfq_io_cq),
7302 .icq_align = __alignof__(struct bfq_io_cq),
7303 .elevator_attrs = bfq_attrs,
7304 .elevator_name = "bfq",
7305 .elevator_owner = THIS_MODULE,
7307 MODULE_ALIAS("bfq-iosched");
7309 static int __init bfq_init(void)
7313 #ifdef CONFIG_BFQ_GROUP_IOSCHED
7314 ret = blkcg_policy_register(&blkcg_policy_bfq);
7320 if (bfq_slab_setup())
7324 * Times to load large popular applications for the typical
7325 * systems installed on the reference devices (see the
7326 * comments before the definition of the next
7327 * array). Actually, we use slightly lower values, as the
7328 * estimated peak rate tends to be smaller than the actual
7329 * peak rate. The reason for this last fact is that estimates
7330 * are computed over much shorter time intervals than the long
7331 * intervals typically used for benchmarking. Why? First, to
7332 * adapt more quickly to variations. Second, because an I/O
7333 * scheduler cannot rely on a peak-rate-evaluation workload to
7334 * be run for a long time.
7336 ref_wr_duration[0] = msecs_to_jiffies(7000); /* actually 8 sec */
7337 ref_wr_duration[1] = msecs_to_jiffies(2500); /* actually 3 sec */
7339 ret = elv_register(&iosched_bfq_mq);
7348 #ifdef CONFIG_BFQ_GROUP_IOSCHED
7349 blkcg_policy_unregister(&blkcg_policy_bfq);
7354 static void __exit bfq_exit(void)
7356 elv_unregister(&iosched_bfq_mq);
7357 #ifdef CONFIG_BFQ_GROUP_IOSCHED
7358 blkcg_policy_unregister(&blkcg_policy_bfq);
7363 module_init(bfq_init);
7364 module_exit(bfq_exit);
7366 MODULE_AUTHOR("Paolo Valente");
7367 MODULE_LICENSE("GPL");
7368 MODULE_DESCRIPTION("MQ Budget Fair Queueing I/O Scheduler");