1 /* SPDX-License-Identifier: GPL-2.0
3 * IO cost model based controller.
5 * Copyright (C) 2019 Tejun Heo <tj@kernel.org>
6 * Copyright (C) 2019 Andy Newell <newella@fb.com>
7 * Copyright (C) 2019 Facebook
9 * One challenge of controlling IO resources is the lack of trivially
10 * observable cost metric. This is distinguished from CPU and memory where
11 * wallclock time and the number of bytes can serve as accurate enough
14 * Bandwidth and iops are the most commonly used metrics for IO devices but
15 * depending on the type and specifics of the device, different IO patterns
16 * easily lead to multiple orders of magnitude variations rendering them
17 * useless for the purpose of IO capacity distribution. While on-device
18 * time, with a lot of clutches, could serve as a useful approximation for
19 * non-queued rotational devices, this is no longer viable with modern
20 * devices, even the rotational ones.
22 * While there is no cost metric we can trivially observe, it isn't a
23 * complete mystery. For example, on a rotational device, seek cost
24 * dominates while a contiguous transfer contributes a smaller amount
25 * proportional to the size. If we can characterize at least the relative
26 * costs of these different types of IOs, it should be possible to
27 * implement a reasonable work-conserving proportional IO resource
32 * IO cost model estimates the cost of an IO given its basic parameters and
33 * history (e.g. the end sector of the last IO). The cost is measured in
34 * device time. If a given IO is estimated to cost 10ms, the device should
35 * be able to process ~100 of those IOs in a second.
37 * Currently, there's only one builtin cost model - linear. Each IO is
38 * classified as sequential or random and given a base cost accordingly.
39 * On top of that, a size cost proportional to the length of the IO is
40 * added. While simple, this model captures the operational
41 * characteristics of a wide varienty of devices well enough. Default
42 * parameters for several different classes of devices are provided and the
43 * parameters can be configured from userspace via
44 * /sys/fs/cgroup/io.cost.model.
46 * If needed, tools/cgroup/iocost_coef_gen.py can be used to generate
47 * device-specific coefficients.
51 * The device virtual time (vtime) is used as the primary control metric.
52 * The control strategy is composed of the following three parts.
54 * 2-1. Vtime Distribution
56 * When a cgroup becomes active in terms of IOs, its hierarchical share is
57 * calculated. Please consider the following hierarchy where the numbers
58 * inside parentheses denote the configured weights.
64 * A0 (w:100) A1 (w:100)
66 * If B is idle and only A0 and A1 are actively issuing IOs, as the two are
67 * of equal weight, each gets 50% share. If then B starts issuing IOs, B
68 * gets 300/(100+300) or 75% share, and A0 and A1 equally splits the rest,
69 * 12.5% each. The distribution mechanism only cares about these flattened
70 * shares. They're called hweights (hierarchical weights) and always add
71 * upto 1 (WEIGHT_ONE).
73 * A given cgroup's vtime runs slower in inverse proportion to its hweight.
74 * For example, with 12.5% weight, A0's time runs 8 times slower (100/12.5)
75 * against the device vtime - an IO which takes 10ms on the underlying
76 * device is considered to take 80ms on A0.
78 * This constitutes the basis of IO capacity distribution. Each cgroup's
79 * vtime is running at a rate determined by its hweight. A cgroup tracks
80 * the vtime consumed by past IOs and can issue a new IO if doing so
81 * wouldn't outrun the current device vtime. Otherwise, the IO is
82 * suspended until the vtime has progressed enough to cover it.
84 * 2-2. Vrate Adjustment
86 * It's unrealistic to expect the cost model to be perfect. There are too
87 * many devices and even on the same device the overall performance
88 * fluctuates depending on numerous factors such as IO mixture and device
89 * internal garbage collection. The controller needs to adapt dynamically.
91 * This is achieved by adjusting the overall IO rate according to how busy
92 * the device is. If the device becomes overloaded, we're sending down too
93 * many IOs and should generally slow down. If there are waiting issuers
94 * but the device isn't saturated, we're issuing too few and should
97 * To slow down, we lower the vrate - the rate at which the device vtime
98 * passes compared to the wall clock. For example, if the vtime is running
99 * at the vrate of 75%, all cgroups added up would only be able to issue
100 * 750ms worth of IOs per second, and vice-versa for speeding up.
102 * Device business is determined using two criteria - rq wait and
103 * completion latencies.
105 * When a device gets saturated, the on-device and then the request queues
106 * fill up and a bio which is ready to be issued has to wait for a request
107 * to become available. When this delay becomes noticeable, it's a clear
108 * indication that the device is saturated and we lower the vrate. This
109 * saturation signal is fairly conservative as it only triggers when both
110 * hardware and software queues are filled up, and is used as the default
113 * As devices can have deep queues and be unfair in how the queued commands
114 * are executed, soley depending on rq wait may not result in satisfactory
115 * control quality. For a better control quality, completion latency QoS
116 * parameters can be configured so that the device is considered saturated
117 * if N'th percentile completion latency rises above the set point.
119 * The completion latency requirements are a function of both the
120 * underlying device characteristics and the desired IO latency quality of
121 * service. There is an inherent trade-off - the tighter the latency QoS,
122 * the higher the bandwidth lossage. Latency QoS is disabled by default
123 * and can be set through /sys/fs/cgroup/io.cost.qos.
125 * 2-3. Work Conservation
127 * Imagine two cgroups A and B with equal weights. A is issuing a small IO
128 * periodically while B is sending out enough parallel IOs to saturate the
129 * device on its own. Let's say A's usage amounts to 100ms worth of IO
130 * cost per second, i.e., 10% of the device capacity. The naive
131 * distribution of half and half would lead to 60% utilization of the
132 * device, a significant reduction in the total amount of work done
133 * compared to free-for-all competition. This is too high a cost to pay
136 * To conserve the total amount of work done, we keep track of how much
137 * each active cgroup is actually using and yield part of its weight if
138 * there are other cgroups which can make use of it. In the above case,
139 * A's weight will be lowered so that it hovers above the actual usage and
140 * B would be able to use the rest.
142 * As we don't want to penalize a cgroup for donating its weight, the
143 * surplus weight adjustment factors in a margin and has an immediate
144 * snapback mechanism in case the cgroup needs more IO vtime for itself.
146 * Note that adjusting down surplus weights has the same effects as
147 * accelerating vtime for other cgroups and work conservation can also be
148 * implemented by adjusting vrate dynamically. However, squaring who can
149 * donate and should take back how much requires hweight propagations
150 * anyway making it easier to implement and understand as a separate
155 * Instead of debugfs or other clumsy monitoring mechanisms, this
156 * controller uses a drgn based monitoring script -
157 * tools/cgroup/iocost_monitor.py. For details on drgn, please see
158 * https://github.com/osandov/drgn. The output looks like the following.
160 * sdb RUN per=300ms cur_per=234.218:v203.695 busy= +1 vrate= 62.12%
161 * active weight hweight% inflt% dbt delay usages%
162 * test/a * 50/ 50 33.33/ 33.33 27.65 2 0*041 033:033:033
163 * test/b * 100/ 100 66.67/ 66.67 17.56 0 0*000 066:079:077
165 * - per : Timer period
166 * - cur_per : Internal wall and device vtime clock
167 * - vrate : Device virtual time rate against wall clock
168 * - weight : Surplus-adjusted and configured weights
169 * - hweight : Surplus-adjusted and configured hierarchical weights
170 * - inflt : The percentage of in-flight IO cost at the end of last period
171 * - del_ms : Deferred issuer delay induction level and duration
172 * - usages : Usage history
175 #include <linux/kernel.h>
176 #include <linux/module.h>
177 #include <linux/timer.h>
178 #include <linux/time64.h>
179 #include <linux/parser.h>
180 #include <linux/sched/signal.h>
181 #include <linux/blk-cgroup.h>
182 #include <asm/local.h>
183 #include <asm/local64.h>
184 #include "blk-rq-qos.h"
185 #include "blk-stat.h"
188 #ifdef CONFIG_TRACEPOINTS
190 /* copied from TRACE_CGROUP_PATH, see cgroup-internal.h */
191 #define TRACE_IOCG_PATH_LEN 1024
192 static DEFINE_SPINLOCK(trace_iocg_path_lock);
193 static char trace_iocg_path[TRACE_IOCG_PATH_LEN];
195 #define TRACE_IOCG_PATH(type, iocg, ...) \
197 unsigned long flags; \
198 if (trace_iocost_##type##_enabled()) { \
199 spin_lock_irqsave(&trace_iocg_path_lock, flags); \
200 cgroup_path(iocg_to_blkg(iocg)->blkcg->css.cgroup, \
201 trace_iocg_path, TRACE_IOCG_PATH_LEN); \
202 trace_iocost_##type(iocg, trace_iocg_path, \
204 spin_unlock_irqrestore(&trace_iocg_path_lock, flags); \
208 #else /* CONFIG_TRACE_POINTS */
209 #define TRACE_IOCG_PATH(type, iocg, ...) do { } while (0)
210 #endif /* CONFIG_TRACE_POINTS */
215 /* timer period is calculated from latency requirements, bound it */
216 MIN_PERIOD = USEC_PER_MSEC,
217 MAX_PERIOD = USEC_PER_SEC,
220 * iocg->vtime is targeted at 50% behind the device vtime, which
221 * serves as its IO credit buffer. Surplus weight adjustment is
222 * immediately canceled if the vtime margin runs below 10%.
226 MARGIN_TARGET_PCT = 50,
228 INUSE_ADJ_STEP_PCT = 25,
230 /* Have some play in timer operations */
233 /* 1/64k is granular enough and can easily be handled w/ u32 */
234 WEIGHT_ONE = 1 << 16,
237 * As vtime is used to calculate the cost of each IO, it needs to
238 * be fairly high precision. For example, it should be able to
239 * represent the cost of a single page worth of discard with
240 * suffificient accuracy. At the same time, it should be able to
241 * represent reasonably long enough durations to be useful and
242 * convenient during operation.
244 * 1s worth of vtime is 2^37. This gives us both sub-nanosecond
245 * granularity and days of wrap-around time even at extreme vrates.
247 VTIME_PER_SEC_SHIFT = 37,
248 VTIME_PER_SEC = 1LLU << VTIME_PER_SEC_SHIFT,
249 VTIME_PER_USEC = VTIME_PER_SEC / USEC_PER_SEC,
250 VTIME_PER_NSEC = VTIME_PER_SEC / NSEC_PER_SEC,
252 /* bound vrate adjustments within two orders of magnitude */
253 VRATE_MIN_PPM = 10000, /* 1% */
254 VRATE_MAX_PPM = 100000000, /* 10000% */
256 VRATE_MIN = VTIME_PER_USEC * VRATE_MIN_PPM / MILLION,
257 VRATE_CLAMP_ADJ_PCT = 4,
259 /* if IOs end up waiting for requests, issue less */
260 RQ_WAIT_BUSY_PCT = 5,
262 /* unbusy hysterisis */
266 * The effect of delay is indirect and non-linear and a huge amount of
267 * future debt can accumulate abruptly while unthrottled. Linearly scale
268 * up delay as debt is going up and then let it decay exponentially.
269 * This gives us quick ramp ups while delay is accumulating and long
270 * tails which can help reducing the frequency of debt explosions on
271 * unthrottle. The parameters are experimentally determined.
273 * The delay mechanism provides adequate protection and behavior in many
274 * cases. However, this is far from ideal and falls shorts on both
275 * fronts. The debtors are often throttled too harshly costing a
276 * significant level of fairness and possibly total work while the
277 * protection against their impacts on the system can be choppy and
280 * The shortcoming primarily stems from the fact that, unlike for page
281 * cache, the kernel doesn't have well-defined back-pressure propagation
282 * mechanism and policies for anonymous memory. Fully addressing this
283 * issue will likely require substantial improvements in the area.
285 MIN_DELAY_THR_PCT = 500,
286 MAX_DELAY_THR_PCT = 25000,
288 MAX_DELAY = 250 * USEC_PER_MSEC,
290 /* halve debts if avg usage over 100ms is under 50% */
292 DFGV_PERIOD = 100 * USEC_PER_MSEC,
294 /* don't let cmds which take a very long time pin lagging for too long */
295 MAX_LAGGING_PERIODS = 10,
297 /* switch iff the conditions are met for longer than this */
298 AUTOP_CYCLE_NSEC = 10LLU * NSEC_PER_SEC,
301 * Count IO size in 4k pages. The 12bit shift helps keeping
302 * size-proportional components of cost calculation in closer
303 * numbers of digits to per-IO cost components.
306 IOC_PAGE_SIZE = 1 << IOC_PAGE_SHIFT,
307 IOC_SECT_TO_PAGE_SHIFT = IOC_PAGE_SHIFT - SECTOR_SHIFT,
309 /* if apart further than 16M, consider randio for linear model */
310 LCOEF_RANDIO_PAGES = 4096,
319 /* io.cost.qos controls including per-dev enable of the whole controller */
326 /* io.cost.qos params */
337 /* io.cost.model controls */
344 /* builtin linear cost model coefficients */
374 u32 qos[NR_QOS_PARAMS];
375 u64 i_lcoefs[NR_I_LCOEFS];
376 u64 lcoefs[NR_LCOEFS];
377 u32 too_fast_vrate_pct;
378 u32 too_slow_vrate_pct;
394 struct ioc_pcpu_stat {
395 struct ioc_missed missed[2];
397 local64_t rq_wait_ns;
407 struct ioc_params params;
408 struct ioc_margins margins;
415 struct timer_list timer;
416 struct list_head active_iocgs; /* active cgroups */
417 struct ioc_pcpu_stat __percpu *pcpu_stat;
419 enum ioc_running running;
420 atomic64_t vtime_rate;
424 seqcount_spinlock_t period_seqcount;
425 u64 period_at; /* wallclock starttime */
426 u64 period_at_vtime; /* vtime starttime */
428 atomic64_t cur_period; /* inc'd each period */
429 int busy_level; /* saturation history */
431 bool weights_updated;
432 atomic_t hweight_gen; /* for lazy hweights */
434 /* debt forgivness */
437 u64 dfgv_usage_us_sum;
439 u64 autop_too_fast_at;
440 u64 autop_too_slow_at;
442 bool user_qos_params:1;
443 bool user_cost_model:1;
446 struct iocg_pcpu_stat {
447 local64_t abs_vusage;
457 /* per device-cgroup pair */
459 struct blkg_policy_data pd;
463 * A iocg can get its weight from two sources - an explicit
464 * per-device-cgroup configuration or the default weight of the
465 * cgroup. `cfg_weight` is the explicit per-device-cgroup
466 * configuration. `weight` is the effective considering both
469 * When an idle cgroup becomes active its `active` goes from 0 to
470 * `weight`. `inuse` is the surplus adjusted active weight.
471 * `active` and `inuse` are used to calculate `hweight_active` and
474 * `last_inuse` remembers `inuse` while an iocg is idle to persist
475 * surplus adjustments.
477 * `inuse` may be adjusted dynamically during period. `saved_*` are used
478 * to determine and track adjustments.
488 sector_t cursor; /* to detect randio */
491 * `vtime` is this iocg's vtime cursor which progresses as IOs are
492 * issued. If lagging behind device vtime, the delta represents
493 * the currently available IO budget. If running ahead, the
496 * `vtime_done` is the same but progressed on completion rather
497 * than issue. The delta behind `vtime` represents the cost of
498 * currently in-flight IOs.
501 atomic64_t done_vtime;
504 /* current delay in effect and when it started */
509 * The period this iocg was last active in. Used for deactivation
510 * and invalidating `vtime`.
512 atomic64_t active_period;
513 struct list_head active_list;
515 /* see __propagate_weights() and current_hweight() for details */
516 u64 child_active_sum;
518 u64 child_adjusted_sum;
522 u32 hweight_donating;
523 u32 hweight_after_donation;
525 struct list_head walk_list;
526 struct list_head surplus_list;
528 struct wait_queue_head waitq;
529 struct hrtimer waitq_timer;
531 /* timestamp at the latest activation */
535 struct iocg_pcpu_stat __percpu *pcpu_stat;
536 struct iocg_stat local_stat;
537 struct iocg_stat desc_stat;
538 struct iocg_stat last_stat;
539 u64 last_stat_abs_vusage;
545 /* this iocg's depth in the hierarchy and ancestors including self */
547 struct ioc_gq *ancestors[];
552 struct blkcg_policy_data cpd;
553 unsigned int dfl_weight;
564 struct wait_queue_entry wait;
570 struct iocg_wake_ctx {
576 static const struct ioc_params autop[] = {
579 [QOS_RLAT] = 250000, /* 250ms */
581 [QOS_MIN] = VRATE_MIN_PPM,
582 [QOS_MAX] = VRATE_MAX_PPM,
585 [I_LCOEF_RBPS] = 174019176,
586 [I_LCOEF_RSEQIOPS] = 41708,
587 [I_LCOEF_RRANDIOPS] = 370,
588 [I_LCOEF_WBPS] = 178075866,
589 [I_LCOEF_WSEQIOPS] = 42705,
590 [I_LCOEF_WRANDIOPS] = 378,
595 [QOS_RLAT] = 25000, /* 25ms */
597 [QOS_MIN] = VRATE_MIN_PPM,
598 [QOS_MAX] = VRATE_MAX_PPM,
601 [I_LCOEF_RBPS] = 245855193,
602 [I_LCOEF_RSEQIOPS] = 61575,
603 [I_LCOEF_RRANDIOPS] = 6946,
604 [I_LCOEF_WBPS] = 141365009,
605 [I_LCOEF_WSEQIOPS] = 33716,
606 [I_LCOEF_WRANDIOPS] = 26796,
611 [QOS_RLAT] = 25000, /* 25ms */
613 [QOS_MIN] = VRATE_MIN_PPM,
614 [QOS_MAX] = VRATE_MAX_PPM,
617 [I_LCOEF_RBPS] = 488636629,
618 [I_LCOEF_RSEQIOPS] = 8932,
619 [I_LCOEF_RRANDIOPS] = 8518,
620 [I_LCOEF_WBPS] = 427891549,
621 [I_LCOEF_WSEQIOPS] = 28755,
622 [I_LCOEF_WRANDIOPS] = 21940,
624 .too_fast_vrate_pct = 500,
628 [QOS_RLAT] = 5000, /* 5ms */
630 [QOS_MIN] = VRATE_MIN_PPM,
631 [QOS_MAX] = VRATE_MAX_PPM,
634 [I_LCOEF_RBPS] = 3102524156LLU,
635 [I_LCOEF_RSEQIOPS] = 724816,
636 [I_LCOEF_RRANDIOPS] = 778122,
637 [I_LCOEF_WBPS] = 1742780862LLU,
638 [I_LCOEF_WSEQIOPS] = 425702,
639 [I_LCOEF_WRANDIOPS] = 443193,
641 .too_slow_vrate_pct = 10,
646 * vrate adjust percentages indexed by ioc->busy_level. We adjust up on
647 * vtime credit shortage and down on device saturation.
649 static u32 vrate_adj_pct[] =
651 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
652 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
653 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8, 8, 8, 16 };
655 static struct blkcg_policy blkcg_policy_iocost;
657 /* accessors and helpers */
658 static struct ioc *rqos_to_ioc(struct rq_qos *rqos)
660 return container_of(rqos, struct ioc, rqos);
663 static struct ioc *q_to_ioc(struct request_queue *q)
665 return rqos_to_ioc(rq_qos_id(q, RQ_QOS_COST));
668 static const char *q_name(struct request_queue *q)
670 if (blk_queue_registered(q))
671 return kobject_name(q->kobj.parent);
676 static const char __maybe_unused *ioc_name(struct ioc *ioc)
678 return q_name(ioc->rqos.q);
681 static struct ioc_gq *pd_to_iocg(struct blkg_policy_data *pd)
683 return pd ? container_of(pd, struct ioc_gq, pd) : NULL;
686 static struct ioc_gq *blkg_to_iocg(struct blkcg_gq *blkg)
688 return pd_to_iocg(blkg_to_pd(blkg, &blkcg_policy_iocost));
691 static struct blkcg_gq *iocg_to_blkg(struct ioc_gq *iocg)
693 return pd_to_blkg(&iocg->pd);
696 static struct ioc_cgrp *blkcg_to_iocc(struct blkcg *blkcg)
698 return container_of(blkcg_to_cpd(blkcg, &blkcg_policy_iocost),
699 struct ioc_cgrp, cpd);
703 * Scale @abs_cost to the inverse of @hw_inuse. The lower the hierarchical
704 * weight, the more expensive each IO. Must round up.
706 static u64 abs_cost_to_cost(u64 abs_cost, u32 hw_inuse)
708 return DIV64_U64_ROUND_UP(abs_cost * WEIGHT_ONE, hw_inuse);
712 * The inverse of abs_cost_to_cost(). Must round up.
714 static u64 cost_to_abs_cost(u64 cost, u32 hw_inuse)
716 return DIV64_U64_ROUND_UP(cost * hw_inuse, WEIGHT_ONE);
719 static void iocg_commit_bio(struct ioc_gq *iocg, struct bio *bio,
720 u64 abs_cost, u64 cost)
722 struct iocg_pcpu_stat *gcs;
724 bio->bi_iocost_cost = cost;
725 atomic64_add(cost, &iocg->vtime);
727 gcs = get_cpu_ptr(iocg->pcpu_stat);
728 local64_add(abs_cost, &gcs->abs_vusage);
732 static void iocg_lock(struct ioc_gq *iocg, bool lock_ioc, unsigned long *flags)
735 spin_lock_irqsave(&iocg->ioc->lock, *flags);
736 spin_lock(&iocg->waitq.lock);
738 spin_lock_irqsave(&iocg->waitq.lock, *flags);
742 static void iocg_unlock(struct ioc_gq *iocg, bool unlock_ioc, unsigned long *flags)
745 spin_unlock(&iocg->waitq.lock);
746 spin_unlock_irqrestore(&iocg->ioc->lock, *flags);
748 spin_unlock_irqrestore(&iocg->waitq.lock, *flags);
752 #define CREATE_TRACE_POINTS
753 #include <trace/events/iocost.h>
755 static void ioc_refresh_margins(struct ioc *ioc)
757 struct ioc_margins *margins = &ioc->margins;
758 u32 period_us = ioc->period_us;
759 u64 vrate = ioc->vtime_base_rate;
761 margins->min = (period_us * MARGIN_MIN_PCT / 100) * vrate;
762 margins->low = (period_us * MARGIN_LOW_PCT / 100) * vrate;
763 margins->target = (period_us * MARGIN_TARGET_PCT / 100) * vrate;
766 /* latency Qos params changed, update period_us and all the dependent params */
767 static void ioc_refresh_period_us(struct ioc *ioc)
769 u32 ppm, lat, multi, period_us;
771 lockdep_assert_held(&ioc->lock);
773 /* pick the higher latency target */
774 if (ioc->params.qos[QOS_RLAT] >= ioc->params.qos[QOS_WLAT]) {
775 ppm = ioc->params.qos[QOS_RPPM];
776 lat = ioc->params.qos[QOS_RLAT];
778 ppm = ioc->params.qos[QOS_WPPM];
779 lat = ioc->params.qos[QOS_WLAT];
783 * We want the period to be long enough to contain a healthy number
784 * of IOs while short enough for granular control. Define it as a
785 * multiple of the latency target. Ideally, the multiplier should
786 * be scaled according to the percentile so that it would nominally
787 * contain a certain number of requests. Let's be simpler and
788 * scale it linearly so that it's 2x >= pct(90) and 10x at pct(50).
791 multi = max_t(u32, (MILLION - ppm) / 50000, 2);
794 period_us = multi * lat;
795 period_us = clamp_t(u32, period_us, MIN_PERIOD, MAX_PERIOD);
797 /* calculate dependent params */
798 ioc->period_us = period_us;
799 ioc->timer_slack_ns = div64_u64(
800 (u64)period_us * NSEC_PER_USEC * TIMER_SLACK_PCT,
802 ioc_refresh_margins(ioc);
805 static int ioc_autop_idx(struct ioc *ioc)
807 int idx = ioc->autop_idx;
808 const struct ioc_params *p = &autop[idx];
813 if (!blk_queue_nonrot(ioc->rqos.q))
816 /* handle SATA SSDs w/ broken NCQ */
817 if (blk_queue_depth(ioc->rqos.q) == 1)
818 return AUTOP_SSD_QD1;
820 /* use one of the normal ssd sets */
821 if (idx < AUTOP_SSD_DFL)
822 return AUTOP_SSD_DFL;
824 /* if user is overriding anything, maintain what was there */
825 if (ioc->user_qos_params || ioc->user_cost_model)
828 /* step up/down based on the vrate */
829 vrate_pct = div64_u64(ioc->vtime_base_rate * 100, VTIME_PER_USEC);
830 now_ns = ktime_get_ns();
832 if (p->too_fast_vrate_pct && p->too_fast_vrate_pct <= vrate_pct) {
833 if (!ioc->autop_too_fast_at)
834 ioc->autop_too_fast_at = now_ns;
835 if (now_ns - ioc->autop_too_fast_at >= AUTOP_CYCLE_NSEC)
838 ioc->autop_too_fast_at = 0;
841 if (p->too_slow_vrate_pct && p->too_slow_vrate_pct >= vrate_pct) {
842 if (!ioc->autop_too_slow_at)
843 ioc->autop_too_slow_at = now_ns;
844 if (now_ns - ioc->autop_too_slow_at >= AUTOP_CYCLE_NSEC)
847 ioc->autop_too_slow_at = 0;
854 * Take the followings as input
856 * @bps maximum sequential throughput
857 * @seqiops maximum sequential 4k iops
858 * @randiops maximum random 4k iops
860 * and calculate the linear model cost coefficients.
862 * *@page per-page cost 1s / (@bps / 4096)
863 * *@seqio base cost of a seq IO max((1s / @seqiops) - *@page, 0)
864 * @randiops base cost of a rand IO max((1s / @randiops) - *@page, 0)
866 static void calc_lcoefs(u64 bps, u64 seqiops, u64 randiops,
867 u64 *page, u64 *seqio, u64 *randio)
871 *page = *seqio = *randio = 0;
874 *page = DIV64_U64_ROUND_UP(VTIME_PER_SEC,
875 DIV_ROUND_UP_ULL(bps, IOC_PAGE_SIZE));
878 v = DIV64_U64_ROUND_UP(VTIME_PER_SEC, seqiops);
884 v = DIV64_U64_ROUND_UP(VTIME_PER_SEC, randiops);
890 static void ioc_refresh_lcoefs(struct ioc *ioc)
892 u64 *u = ioc->params.i_lcoefs;
893 u64 *c = ioc->params.lcoefs;
895 calc_lcoefs(u[I_LCOEF_RBPS], u[I_LCOEF_RSEQIOPS], u[I_LCOEF_RRANDIOPS],
896 &c[LCOEF_RPAGE], &c[LCOEF_RSEQIO], &c[LCOEF_RRANDIO]);
897 calc_lcoefs(u[I_LCOEF_WBPS], u[I_LCOEF_WSEQIOPS], u[I_LCOEF_WRANDIOPS],
898 &c[LCOEF_WPAGE], &c[LCOEF_WSEQIO], &c[LCOEF_WRANDIO]);
901 static bool ioc_refresh_params(struct ioc *ioc, bool force)
903 const struct ioc_params *p;
906 lockdep_assert_held(&ioc->lock);
908 idx = ioc_autop_idx(ioc);
911 if (idx == ioc->autop_idx && !force)
914 if (idx != ioc->autop_idx)
915 atomic64_set(&ioc->vtime_rate, VTIME_PER_USEC);
917 ioc->autop_idx = idx;
918 ioc->autop_too_fast_at = 0;
919 ioc->autop_too_slow_at = 0;
921 if (!ioc->user_qos_params)
922 memcpy(ioc->params.qos, p->qos, sizeof(p->qos));
923 if (!ioc->user_cost_model)
924 memcpy(ioc->params.i_lcoefs, p->i_lcoefs, sizeof(p->i_lcoefs));
926 ioc_refresh_period_us(ioc);
927 ioc_refresh_lcoefs(ioc);
929 ioc->vrate_min = DIV64_U64_ROUND_UP((u64)ioc->params.qos[QOS_MIN] *
930 VTIME_PER_USEC, MILLION);
931 ioc->vrate_max = div64_u64((u64)ioc->params.qos[QOS_MAX] *
932 VTIME_PER_USEC, MILLION);
938 * When an iocg accumulates too much vtime or gets deactivated, we throw away
939 * some vtime, which lowers the overall device utilization. As the exact amount
940 * which is being thrown away is known, we can compensate by accelerating the
941 * vrate accordingly so that the extra vtime generated in the current period
942 * matches what got lost.
944 static void ioc_refresh_vrate(struct ioc *ioc, struct ioc_now *now)
946 s64 pleft = ioc->period_at + ioc->period_us - now->now;
947 s64 vperiod = ioc->period_us * ioc->vtime_base_rate;
948 s64 vcomp, vcomp_min, vcomp_max;
950 lockdep_assert_held(&ioc->lock);
952 /* we need some time left in this period */
957 * Calculate how much vrate should be adjusted to offset the error.
958 * Limit the amount of adjustment and deduct the adjusted amount from
961 vcomp = -div64_s64(ioc->vtime_err, pleft);
962 vcomp_min = -(ioc->vtime_base_rate >> 1);
963 vcomp_max = ioc->vtime_base_rate;
964 vcomp = clamp(vcomp, vcomp_min, vcomp_max);
966 ioc->vtime_err += vcomp * pleft;
968 atomic64_set(&ioc->vtime_rate, ioc->vtime_base_rate + vcomp);
970 /* bound how much error can accumulate */
971 ioc->vtime_err = clamp(ioc->vtime_err, -vperiod, vperiod);
974 static void ioc_adjust_base_vrate(struct ioc *ioc, u32 rq_wait_pct,
975 int nr_lagging, int nr_shortages,
976 int prev_busy_level, u32 *missed_ppm)
978 u64 vrate = ioc->vtime_base_rate;
979 u64 vrate_min = ioc->vrate_min, vrate_max = ioc->vrate_max;
981 if (!ioc->busy_level || (ioc->busy_level < 0 && nr_lagging)) {
982 if (ioc->busy_level != prev_busy_level || nr_lagging)
983 trace_iocost_ioc_vrate_adj(ioc, atomic64_read(&ioc->vtime_rate),
984 missed_ppm, rq_wait_pct,
985 nr_lagging, nr_shortages);
991 * If vrate is out of bounds, apply clamp gradually as the
992 * bounds can change abruptly. Otherwise, apply busy_level
995 if (vrate < vrate_min) {
996 vrate = div64_u64(vrate * (100 + VRATE_CLAMP_ADJ_PCT), 100);
997 vrate = min(vrate, vrate_min);
998 } else if (vrate > vrate_max) {
999 vrate = div64_u64(vrate * (100 - VRATE_CLAMP_ADJ_PCT), 100);
1000 vrate = max(vrate, vrate_max);
1002 int idx = min_t(int, abs(ioc->busy_level),
1003 ARRAY_SIZE(vrate_adj_pct) - 1);
1004 u32 adj_pct = vrate_adj_pct[idx];
1006 if (ioc->busy_level > 0)
1007 adj_pct = 100 - adj_pct;
1009 adj_pct = 100 + adj_pct;
1011 vrate = clamp(DIV64_U64_ROUND_UP(vrate * adj_pct, 100),
1012 vrate_min, vrate_max);
1015 trace_iocost_ioc_vrate_adj(ioc, vrate, missed_ppm, rq_wait_pct,
1016 nr_lagging, nr_shortages);
1018 ioc->vtime_base_rate = vrate;
1019 ioc_refresh_margins(ioc);
1022 /* take a snapshot of the current [v]time and vrate */
1023 static void ioc_now(struct ioc *ioc, struct ioc_now *now)
1027 now->now_ns = ktime_get();
1028 now->now = ktime_to_us(now->now_ns);
1029 now->vrate = atomic64_read(&ioc->vtime_rate);
1032 * The current vtime is
1034 * vtime at period start + (wallclock time since the start) * vrate
1036 * As a consistent snapshot of `period_at_vtime` and `period_at` is
1037 * needed, they're seqcount protected.
1040 seq = read_seqcount_begin(&ioc->period_seqcount);
1041 now->vnow = ioc->period_at_vtime +
1042 (now->now - ioc->period_at) * now->vrate;
1043 } while (read_seqcount_retry(&ioc->period_seqcount, seq));
1046 static void ioc_start_period(struct ioc *ioc, struct ioc_now *now)
1048 WARN_ON_ONCE(ioc->running != IOC_RUNNING);
1050 write_seqcount_begin(&ioc->period_seqcount);
1051 ioc->period_at = now->now;
1052 ioc->period_at_vtime = now->vnow;
1053 write_seqcount_end(&ioc->period_seqcount);
1055 ioc->timer.expires = jiffies + usecs_to_jiffies(ioc->period_us);
1056 add_timer(&ioc->timer);
1060 * Update @iocg's `active` and `inuse` to @active and @inuse, update level
1061 * weight sums and propagate upwards accordingly. If @save, the current margin
1062 * is saved to be used as reference for later inuse in-period adjustments.
1064 static void __propagate_weights(struct ioc_gq *iocg, u32 active, u32 inuse,
1065 bool save, struct ioc_now *now)
1067 struct ioc *ioc = iocg->ioc;
1070 lockdep_assert_held(&ioc->lock);
1072 inuse = clamp_t(u32, inuse, 1, active);
1074 iocg->last_inuse = iocg->inuse;
1076 iocg->saved_margin = now->vnow - atomic64_read(&iocg->vtime);
1078 if (active == iocg->active && inuse == iocg->inuse)
1081 for (lvl = iocg->level - 1; lvl >= 0; lvl--) {
1082 struct ioc_gq *parent = iocg->ancestors[lvl];
1083 struct ioc_gq *child = iocg->ancestors[lvl + 1];
1084 u32 parent_active = 0, parent_inuse = 0;
1086 /* update the level sums */
1087 parent->child_active_sum += (s32)(active - child->active);
1088 parent->child_inuse_sum += (s32)(inuse - child->inuse);
1089 /* apply the udpates */
1090 child->active = active;
1091 child->inuse = inuse;
1094 * The delta between inuse and active sums indicates that
1095 * much of weight is being given away. Parent's inuse
1096 * and active should reflect the ratio.
1098 if (parent->child_active_sum) {
1099 parent_active = parent->weight;
1100 parent_inuse = DIV64_U64_ROUND_UP(
1101 parent_active * parent->child_inuse_sum,
1102 parent->child_active_sum);
1105 /* do we need to keep walking up? */
1106 if (parent_active == parent->active &&
1107 parent_inuse == parent->inuse)
1110 active = parent_active;
1111 inuse = parent_inuse;
1114 ioc->weights_updated = true;
1117 static void commit_weights(struct ioc *ioc)
1119 lockdep_assert_held(&ioc->lock);
1121 if (ioc->weights_updated) {
1122 /* paired with rmb in current_hweight(), see there */
1124 atomic_inc(&ioc->hweight_gen);
1125 ioc->weights_updated = false;
1129 static void propagate_weights(struct ioc_gq *iocg, u32 active, u32 inuse,
1130 bool save, struct ioc_now *now)
1132 __propagate_weights(iocg, active, inuse, save, now);
1133 commit_weights(iocg->ioc);
1136 static void current_hweight(struct ioc_gq *iocg, u32 *hw_activep, u32 *hw_inusep)
1138 struct ioc *ioc = iocg->ioc;
1143 /* hot path - if uptodate, use cached */
1144 ioc_gen = atomic_read(&ioc->hweight_gen);
1145 if (ioc_gen == iocg->hweight_gen)
1149 * Paired with wmb in commit_weights(). If we saw the updated
1150 * hweight_gen, all the weight updates from __propagate_weights() are
1153 * We can race with weight updates during calculation and get it
1154 * wrong. However, hweight_gen would have changed and a future
1155 * reader will recalculate and we're guaranteed to discard the
1156 * wrong result soon.
1160 hwa = hwi = WEIGHT_ONE;
1161 for (lvl = 0; lvl <= iocg->level - 1; lvl++) {
1162 struct ioc_gq *parent = iocg->ancestors[lvl];
1163 struct ioc_gq *child = iocg->ancestors[lvl + 1];
1164 u64 active_sum = READ_ONCE(parent->child_active_sum);
1165 u64 inuse_sum = READ_ONCE(parent->child_inuse_sum);
1166 u32 active = READ_ONCE(child->active);
1167 u32 inuse = READ_ONCE(child->inuse);
1169 /* we can race with deactivations and either may read as zero */
1170 if (!active_sum || !inuse_sum)
1173 active_sum = max_t(u64, active, active_sum);
1174 hwa = div64_u64((u64)hwa * active, active_sum);
1176 inuse_sum = max_t(u64, inuse, inuse_sum);
1177 hwi = div64_u64((u64)hwi * inuse, inuse_sum);
1180 iocg->hweight_active = max_t(u32, hwa, 1);
1181 iocg->hweight_inuse = max_t(u32, hwi, 1);
1182 iocg->hweight_gen = ioc_gen;
1185 *hw_activep = iocg->hweight_active;
1187 *hw_inusep = iocg->hweight_inuse;
1191 * Calculate the hweight_inuse @iocg would get with max @inuse assuming all the
1192 * other weights stay unchanged.
1194 static u32 current_hweight_max(struct ioc_gq *iocg)
1196 u32 hwm = WEIGHT_ONE;
1197 u32 inuse = iocg->active;
1198 u64 child_inuse_sum;
1201 lockdep_assert_held(&iocg->ioc->lock);
1203 for (lvl = iocg->level - 1; lvl >= 0; lvl--) {
1204 struct ioc_gq *parent = iocg->ancestors[lvl];
1205 struct ioc_gq *child = iocg->ancestors[lvl + 1];
1207 child_inuse_sum = parent->child_inuse_sum + inuse - child->inuse;
1208 hwm = div64_u64((u64)hwm * inuse, child_inuse_sum);
1209 inuse = DIV64_U64_ROUND_UP(parent->active * child_inuse_sum,
1210 parent->child_active_sum);
1213 return max_t(u32, hwm, 1);
1216 static void weight_updated(struct ioc_gq *iocg, struct ioc_now *now)
1218 struct ioc *ioc = iocg->ioc;
1219 struct blkcg_gq *blkg = iocg_to_blkg(iocg);
1220 struct ioc_cgrp *iocc = blkcg_to_iocc(blkg->blkcg);
1223 lockdep_assert_held(&ioc->lock);
1225 weight = iocg->cfg_weight ?: iocc->dfl_weight;
1226 if (weight != iocg->weight && iocg->active)
1227 propagate_weights(iocg, weight, iocg->inuse, true, now);
1228 iocg->weight = weight;
1231 static bool iocg_activate(struct ioc_gq *iocg, struct ioc_now *now)
1233 struct ioc *ioc = iocg->ioc;
1234 u64 last_period, cur_period;
1239 * If seem to be already active, just update the stamp to tell the
1240 * timer that we're still active. We don't mind occassional races.
1242 if (!list_empty(&iocg->active_list)) {
1244 cur_period = atomic64_read(&ioc->cur_period);
1245 if (atomic64_read(&iocg->active_period) != cur_period)
1246 atomic64_set(&iocg->active_period, cur_period);
1250 /* racy check on internal node IOs, treat as root level IOs */
1251 if (iocg->child_active_sum)
1254 spin_lock_irq(&ioc->lock);
1259 cur_period = atomic64_read(&ioc->cur_period);
1260 last_period = atomic64_read(&iocg->active_period);
1261 atomic64_set(&iocg->active_period, cur_period);
1263 /* already activated or breaking leaf-only constraint? */
1264 if (!list_empty(&iocg->active_list))
1265 goto succeed_unlock;
1266 for (i = iocg->level - 1; i > 0; i--)
1267 if (!list_empty(&iocg->ancestors[i]->active_list))
1270 if (iocg->child_active_sum)
1274 * Always start with the target budget. On deactivation, we throw away
1275 * anything above it.
1277 vtarget = now->vnow - ioc->margins.target;
1278 vtime = atomic64_read(&iocg->vtime);
1280 atomic64_add(vtarget - vtime, &iocg->vtime);
1281 atomic64_add(vtarget - vtime, &iocg->done_vtime);
1285 * Activate, propagate weight and start period timer if not
1286 * running. Reset hweight_gen to avoid accidental match from
1289 iocg->hweight_gen = atomic_read(&ioc->hweight_gen) - 1;
1290 list_add(&iocg->active_list, &ioc->active_iocgs);
1292 propagate_weights(iocg, iocg->weight,
1293 iocg->last_inuse ?: iocg->weight, true, now);
1295 TRACE_IOCG_PATH(iocg_activate, iocg, now,
1296 last_period, cur_period, vtime);
1298 iocg->activated_at = now->now;
1300 if (ioc->running == IOC_IDLE) {
1301 ioc->running = IOC_RUNNING;
1302 ioc->dfgv_period_at = now->now;
1303 ioc->dfgv_period_rem = 0;
1304 ioc_start_period(ioc, now);
1308 spin_unlock_irq(&ioc->lock);
1312 spin_unlock_irq(&ioc->lock);
1316 static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now)
1318 struct ioc *ioc = iocg->ioc;
1319 struct blkcg_gq *blkg = iocg_to_blkg(iocg);
1320 u64 tdelta, delay, new_delay;
1321 s64 vover, vover_pct;
1324 lockdep_assert_held(&iocg->waitq.lock);
1326 /* calculate the current delay in effect - 1/2 every second */
1327 tdelta = now->now - iocg->delay_at;
1329 delay = iocg->delay >> div64_u64(tdelta, USEC_PER_SEC);
1333 /* calculate the new delay from the debt amount */
1334 current_hweight(iocg, &hwa, NULL);
1335 vover = atomic64_read(&iocg->vtime) +
1336 abs_cost_to_cost(iocg->abs_vdebt, hwa) - now->vnow;
1337 vover_pct = div64_s64(100 * vover,
1338 ioc->period_us * ioc->vtime_base_rate);
1340 if (vover_pct <= MIN_DELAY_THR_PCT)
1342 else if (vover_pct >= MAX_DELAY_THR_PCT)
1343 new_delay = MAX_DELAY;
1345 new_delay = MIN_DELAY +
1346 div_u64((MAX_DELAY - MIN_DELAY) *
1347 (vover_pct - MIN_DELAY_THR_PCT),
1348 MAX_DELAY_THR_PCT - MIN_DELAY_THR_PCT);
1350 /* pick the higher one and apply */
1351 if (new_delay > delay) {
1352 iocg->delay = new_delay;
1353 iocg->delay_at = now->now;
1357 if (delay >= MIN_DELAY) {
1358 if (!iocg->indelay_since)
1359 iocg->indelay_since = now->now;
1360 blkcg_set_delay(blkg, delay * NSEC_PER_USEC);
1363 if (iocg->indelay_since) {
1364 iocg->local_stat.indelay_us += now->now - iocg->indelay_since;
1365 iocg->indelay_since = 0;
1368 blkcg_clear_delay(blkg);
1373 static void iocg_incur_debt(struct ioc_gq *iocg, u64 abs_cost,
1374 struct ioc_now *now)
1376 struct iocg_pcpu_stat *gcs;
1378 lockdep_assert_held(&iocg->ioc->lock);
1379 lockdep_assert_held(&iocg->waitq.lock);
1380 WARN_ON_ONCE(list_empty(&iocg->active_list));
1383 * Once in debt, debt handling owns inuse. @iocg stays at the minimum
1384 * inuse donating all of it share to others until its debt is paid off.
1386 if (!iocg->abs_vdebt && abs_cost) {
1387 iocg->indebt_since = now->now;
1388 propagate_weights(iocg, iocg->active, 0, false, now);
1391 iocg->abs_vdebt += abs_cost;
1393 gcs = get_cpu_ptr(iocg->pcpu_stat);
1394 local64_add(abs_cost, &gcs->abs_vusage);
1398 static void iocg_pay_debt(struct ioc_gq *iocg, u64 abs_vpay,
1399 struct ioc_now *now)
1401 lockdep_assert_held(&iocg->ioc->lock);
1402 lockdep_assert_held(&iocg->waitq.lock);
1404 /* make sure that nobody messed with @iocg */
1405 WARN_ON_ONCE(list_empty(&iocg->active_list));
1406 WARN_ON_ONCE(iocg->inuse > 1);
1408 iocg->abs_vdebt -= min(abs_vpay, iocg->abs_vdebt);
1410 /* if debt is paid in full, restore inuse */
1411 if (!iocg->abs_vdebt) {
1412 iocg->local_stat.indebt_us += now->now - iocg->indebt_since;
1413 iocg->indebt_since = 0;
1415 propagate_weights(iocg, iocg->active, iocg->last_inuse,
1420 static int iocg_wake_fn(struct wait_queue_entry *wq_entry, unsigned mode,
1421 int flags, void *key)
1423 struct iocg_wait *wait = container_of(wq_entry, struct iocg_wait, wait);
1424 struct iocg_wake_ctx *ctx = (struct iocg_wake_ctx *)key;
1425 u64 cost = abs_cost_to_cost(wait->abs_cost, ctx->hw_inuse);
1427 ctx->vbudget -= cost;
1429 if (ctx->vbudget < 0)
1432 iocg_commit_bio(ctx->iocg, wait->bio, wait->abs_cost, cost);
1435 * autoremove_wake_function() removes the wait entry only when it
1436 * actually changed the task state. We want the wait always
1437 * removed. Remove explicitly and use default_wake_function().
1439 list_del_init(&wq_entry->entry);
1440 wait->committed = true;
1442 default_wake_function(wq_entry, mode, flags, key);
1447 * Calculate the accumulated budget, pay debt if @pay_debt and wake up waiters
1448 * accordingly. When @pay_debt is %true, the caller must be holding ioc->lock in
1449 * addition to iocg->waitq.lock.
1451 static void iocg_kick_waitq(struct ioc_gq *iocg, bool pay_debt,
1452 struct ioc_now *now)
1454 struct ioc *ioc = iocg->ioc;
1455 struct iocg_wake_ctx ctx = { .iocg = iocg };
1456 u64 vshortage, expires, oexpires;
1460 lockdep_assert_held(&iocg->waitq.lock);
1462 current_hweight(iocg, &hwa, NULL);
1463 vbudget = now->vnow - atomic64_read(&iocg->vtime);
1466 if (pay_debt && iocg->abs_vdebt && vbudget > 0) {
1467 u64 abs_vbudget = cost_to_abs_cost(vbudget, hwa);
1468 u64 abs_vpay = min_t(u64, abs_vbudget, iocg->abs_vdebt);
1469 u64 vpay = abs_cost_to_cost(abs_vpay, hwa);
1471 lockdep_assert_held(&ioc->lock);
1473 atomic64_add(vpay, &iocg->vtime);
1474 atomic64_add(vpay, &iocg->done_vtime);
1475 iocg_pay_debt(iocg, abs_vpay, now);
1479 if (iocg->abs_vdebt || iocg->delay)
1480 iocg_kick_delay(iocg, now);
1483 * Debt can still be outstanding if we haven't paid all yet or the
1484 * caller raced and called without @pay_debt. Shouldn't wake up waiters
1485 * under debt. Make sure @vbudget reflects the outstanding amount and is
1488 if (iocg->abs_vdebt) {
1489 s64 vdebt = abs_cost_to_cost(iocg->abs_vdebt, hwa);
1490 vbudget = min_t(s64, 0, vbudget - vdebt);
1494 * Wake up the ones which are due and see how much vtime we'll need for
1495 * the next one. As paying off debt restores hw_inuse, it must be read
1496 * after the above debt payment.
1498 ctx.vbudget = vbudget;
1499 current_hweight(iocg, NULL, &ctx.hw_inuse);
1501 __wake_up_locked_key(&iocg->waitq, TASK_NORMAL, &ctx);
1503 if (!waitqueue_active(&iocg->waitq)) {
1504 if (iocg->wait_since) {
1505 iocg->local_stat.wait_us += now->now - iocg->wait_since;
1506 iocg->wait_since = 0;
1511 if (!iocg->wait_since)
1512 iocg->wait_since = now->now;
1514 if (WARN_ON_ONCE(ctx.vbudget >= 0))
1517 /* determine next wakeup, add a timer margin to guarantee chunking */
1518 vshortage = -ctx.vbudget;
1519 expires = now->now_ns +
1520 DIV64_U64_ROUND_UP(vshortage, ioc->vtime_base_rate) *
1522 expires += ioc->timer_slack_ns;
1524 /* if already active and close enough, don't bother */
1525 oexpires = ktime_to_ns(hrtimer_get_softexpires(&iocg->waitq_timer));
1526 if (hrtimer_is_queued(&iocg->waitq_timer) &&
1527 abs(oexpires - expires) <= ioc->timer_slack_ns)
1530 hrtimer_start_range_ns(&iocg->waitq_timer, ns_to_ktime(expires),
1531 ioc->timer_slack_ns, HRTIMER_MODE_ABS);
1534 static enum hrtimer_restart iocg_waitq_timer_fn(struct hrtimer *timer)
1536 struct ioc_gq *iocg = container_of(timer, struct ioc_gq, waitq_timer);
1537 bool pay_debt = READ_ONCE(iocg->abs_vdebt);
1539 unsigned long flags;
1541 ioc_now(iocg->ioc, &now);
1543 iocg_lock(iocg, pay_debt, &flags);
1544 iocg_kick_waitq(iocg, pay_debt, &now);
1545 iocg_unlock(iocg, pay_debt, &flags);
1547 return HRTIMER_NORESTART;
1550 static void ioc_lat_stat(struct ioc *ioc, u32 *missed_ppm_ar, u32 *rq_wait_pct_p)
1552 u32 nr_met[2] = { };
1553 u32 nr_missed[2] = { };
1557 for_each_online_cpu(cpu) {
1558 struct ioc_pcpu_stat *stat = per_cpu_ptr(ioc->pcpu_stat, cpu);
1559 u64 this_rq_wait_ns;
1561 for (rw = READ; rw <= WRITE; rw++) {
1562 u32 this_met = local_read(&stat->missed[rw].nr_met);
1563 u32 this_missed = local_read(&stat->missed[rw].nr_missed);
1565 nr_met[rw] += this_met - stat->missed[rw].last_met;
1566 nr_missed[rw] += this_missed - stat->missed[rw].last_missed;
1567 stat->missed[rw].last_met = this_met;
1568 stat->missed[rw].last_missed = this_missed;
1571 this_rq_wait_ns = local64_read(&stat->rq_wait_ns);
1572 rq_wait_ns += this_rq_wait_ns - stat->last_rq_wait_ns;
1573 stat->last_rq_wait_ns = this_rq_wait_ns;
1576 for (rw = READ; rw <= WRITE; rw++) {
1577 if (nr_met[rw] + nr_missed[rw])
1579 DIV64_U64_ROUND_UP((u64)nr_missed[rw] * MILLION,
1580 nr_met[rw] + nr_missed[rw]);
1582 missed_ppm_ar[rw] = 0;
1585 *rq_wait_pct_p = div64_u64(rq_wait_ns * 100,
1586 ioc->period_us * NSEC_PER_USEC);
1589 /* was iocg idle this period? */
1590 static bool iocg_is_idle(struct ioc_gq *iocg)
1592 struct ioc *ioc = iocg->ioc;
1594 /* did something get issued this period? */
1595 if (atomic64_read(&iocg->active_period) ==
1596 atomic64_read(&ioc->cur_period))
1599 /* is something in flight? */
1600 if (atomic64_read(&iocg->done_vtime) != atomic64_read(&iocg->vtime))
1607 * Call this function on the target leaf @iocg's to build pre-order traversal
1608 * list of all the ancestors in @inner_walk. The inner nodes are linked through
1609 * ->walk_list and the caller is responsible for dissolving the list after use.
1611 static void iocg_build_inner_walk(struct ioc_gq *iocg,
1612 struct list_head *inner_walk)
1616 WARN_ON_ONCE(!list_empty(&iocg->walk_list));
1618 /* find the first ancestor which hasn't been visited yet */
1619 for (lvl = iocg->level - 1; lvl >= 0; lvl--) {
1620 if (!list_empty(&iocg->ancestors[lvl]->walk_list))
1624 /* walk down and visit the inner nodes to get pre-order traversal */
1625 while (++lvl <= iocg->level - 1) {
1626 struct ioc_gq *inner = iocg->ancestors[lvl];
1628 /* record traversal order */
1629 list_add_tail(&inner->walk_list, inner_walk);
1633 /* collect per-cpu counters and propagate the deltas to the parent */
1634 static void iocg_flush_stat_one(struct ioc_gq *iocg, struct ioc_now *now)
1636 struct ioc *ioc = iocg->ioc;
1637 struct iocg_stat new_stat;
1642 lockdep_assert_held(&iocg->ioc->lock);
1644 /* collect per-cpu counters */
1645 for_each_possible_cpu(cpu) {
1646 abs_vusage += local64_read(
1647 per_cpu_ptr(&iocg->pcpu_stat->abs_vusage, cpu));
1649 vusage_delta = abs_vusage - iocg->last_stat_abs_vusage;
1650 iocg->last_stat_abs_vusage = abs_vusage;
1652 iocg->usage_delta_us = div64_u64(vusage_delta, ioc->vtime_base_rate);
1653 iocg->local_stat.usage_us += iocg->usage_delta_us;
1655 /* propagate upwards */
1657 iocg->local_stat.usage_us + iocg->desc_stat.usage_us;
1659 iocg->local_stat.wait_us + iocg->desc_stat.wait_us;
1660 new_stat.indebt_us =
1661 iocg->local_stat.indebt_us + iocg->desc_stat.indebt_us;
1662 new_stat.indelay_us =
1663 iocg->local_stat.indelay_us + iocg->desc_stat.indelay_us;
1665 /* propagate the deltas to the parent */
1666 if (iocg->level > 0) {
1667 struct iocg_stat *parent_stat =
1668 &iocg->ancestors[iocg->level - 1]->desc_stat;
1670 parent_stat->usage_us +=
1671 new_stat.usage_us - iocg->last_stat.usage_us;
1672 parent_stat->wait_us +=
1673 new_stat.wait_us - iocg->last_stat.wait_us;
1674 parent_stat->indebt_us +=
1675 new_stat.indebt_us - iocg->last_stat.indebt_us;
1676 parent_stat->indelay_us +=
1677 new_stat.indelay_us - iocg->last_stat.indelay_us;
1680 iocg->last_stat = new_stat;
1683 /* get stat counters ready for reading on all active iocgs */
1684 static void iocg_flush_stat(struct list_head *target_iocgs, struct ioc_now *now)
1686 LIST_HEAD(inner_walk);
1687 struct ioc_gq *iocg, *tiocg;
1689 /* flush leaves and build inner node walk list */
1690 list_for_each_entry(iocg, target_iocgs, active_list) {
1691 iocg_flush_stat_one(iocg, now);
1692 iocg_build_inner_walk(iocg, &inner_walk);
1695 /* keep flushing upwards by walking the inner list backwards */
1696 list_for_each_entry_safe_reverse(iocg, tiocg, &inner_walk, walk_list) {
1697 iocg_flush_stat_one(iocg, now);
1698 list_del_init(&iocg->walk_list);
1703 * Determine what @iocg's hweight_inuse should be after donating unused
1704 * capacity. @hwm is the upper bound and used to signal no donation. This
1705 * function also throws away @iocg's excess budget.
1707 static u32 hweight_after_donation(struct ioc_gq *iocg, u32 old_hwi, u32 hwm,
1708 u32 usage, struct ioc_now *now)
1710 struct ioc *ioc = iocg->ioc;
1711 u64 vtime = atomic64_read(&iocg->vtime);
1712 s64 excess, delta, target, new_hwi;
1714 /* debt handling owns inuse for debtors */
1715 if (iocg->abs_vdebt)
1718 /* see whether minimum margin requirement is met */
1719 if (waitqueue_active(&iocg->waitq) ||
1720 time_after64(vtime, now->vnow - ioc->margins.min))
1723 /* throw away excess above target */
1724 excess = now->vnow - vtime - ioc->margins.target;
1726 atomic64_add(excess, &iocg->vtime);
1727 atomic64_add(excess, &iocg->done_vtime);
1729 ioc->vtime_err -= div64_u64(excess * old_hwi, WEIGHT_ONE);
1733 * Let's say the distance between iocg's and device's vtimes as a
1734 * fraction of period duration is delta. Assuming that the iocg will
1735 * consume the usage determined above, we want to determine new_hwi so
1736 * that delta equals MARGIN_TARGET at the end of the next period.
1738 * We need to execute usage worth of IOs while spending the sum of the
1739 * new budget (1 - MARGIN_TARGET) and the leftover from the last period
1742 * usage = (1 - MARGIN_TARGET + delta) * new_hwi
1744 * Therefore, the new_hwi is:
1746 * new_hwi = usage / (1 - MARGIN_TARGET + delta)
1748 delta = div64_s64(WEIGHT_ONE * (now->vnow - vtime),
1749 now->vnow - ioc->period_at_vtime);
1750 target = WEIGHT_ONE * MARGIN_TARGET_PCT / 100;
1751 new_hwi = div64_s64(WEIGHT_ONE * usage, WEIGHT_ONE - target + delta);
1753 return clamp_t(s64, new_hwi, 1, hwm);
1757 * For work-conservation, an iocg which isn't using all of its share should
1758 * donate the leftover to other iocgs. There are two ways to achieve this - 1.
1759 * bumping up vrate accordingly 2. lowering the donating iocg's inuse weight.
1761 * #1 is mathematically simpler but has the drawback of requiring synchronous
1762 * global hweight_inuse updates when idle iocg's get activated or inuse weights
1763 * change due to donation snapbacks as it has the possibility of grossly
1764 * overshooting what's allowed by the model and vrate.
1766 * #2 is inherently safe with local operations. The donating iocg can easily
1767 * snap back to higher weights when needed without worrying about impacts on
1768 * other nodes as the impacts will be inherently correct. This also makes idle
1769 * iocg activations safe. The only effect activations have is decreasing
1770 * hweight_inuse of others, the right solution to which is for those iocgs to
1771 * snap back to higher weights.
1773 * So, we go with #2. The challenge is calculating how each donating iocg's
1774 * inuse should be adjusted to achieve the target donation amounts. This is done
1775 * using Andy's method described in the following pdf.
1777 * https://drive.google.com/file/d/1PsJwxPFtjUnwOY1QJ5AeICCcsL7BM3bo
1779 * Given the weights and target after-donation hweight_inuse values, Andy's
1780 * method determines how the proportional distribution should look like at each
1781 * sibling level to maintain the relative relationship between all non-donating
1782 * pairs. To roughly summarize, it divides the tree into donating and
1783 * non-donating parts, calculates global donation rate which is used to
1784 * determine the target hweight_inuse for each node, and then derives per-level
1787 * The following pdf shows that global distribution calculated this way can be
1788 * achieved by scaling inuse weights of donating leaves and propagating the
1789 * adjustments upwards proportionally.
1791 * https://drive.google.com/file/d/1vONz1-fzVO7oY5DXXsLjSxEtYYQbOvsE
1793 * Combining the above two, we can determine how each leaf iocg's inuse should
1794 * be adjusted to achieve the target donation.
1796 * https://drive.google.com/file/d/1WcrltBOSPN0qXVdBgnKm4mdp9FhuEFQN
1798 * The inline comments use symbols from the last pdf.
1800 * b is the sum of the absolute budgets in the subtree. 1 for the root node.
1801 * f is the sum of the absolute budgets of non-donating nodes in the subtree.
1802 * t is the sum of the absolute budgets of donating nodes in the subtree.
1803 * w is the weight of the node. w = w_f + w_t
1804 * w_f is the non-donating portion of w. w_f = w * f / b
1805 * w_b is the donating portion of w. w_t = w * t / b
1806 * s is the sum of all sibling weights. s = Sum(w) for siblings
1807 * s_f and s_t are the non-donating and donating portions of s.
1809 * Subscript p denotes the parent's counterpart and ' the adjusted value - e.g.
1810 * w_pt is the donating portion of the parent's weight and w'_pt the same value
1811 * after adjustments. Subscript r denotes the root node's values.
1813 static void transfer_surpluses(struct list_head *surpluses, struct ioc_now *now)
1815 LIST_HEAD(over_hwa);
1816 LIST_HEAD(inner_walk);
1817 struct ioc_gq *iocg, *tiocg, *root_iocg;
1818 u32 after_sum, over_sum, over_target, gamma;
1821 * It's pretty unlikely but possible for the total sum of
1822 * hweight_after_donation's to be higher than WEIGHT_ONE, which will
1823 * confuse the following calculations. If such condition is detected,
1824 * scale down everyone over its full share equally to keep the sum below
1829 list_for_each_entry(iocg, surpluses, surplus_list) {
1832 current_hweight(iocg, &hwa, NULL);
1833 after_sum += iocg->hweight_after_donation;
1835 if (iocg->hweight_after_donation > hwa) {
1836 over_sum += iocg->hweight_after_donation;
1837 list_add(&iocg->walk_list, &over_hwa);
1841 if (after_sum >= WEIGHT_ONE) {
1843 * The delta should be deducted from the over_sum, calculate
1844 * target over_sum value.
1846 u32 over_delta = after_sum - (WEIGHT_ONE - 1);
1847 WARN_ON_ONCE(over_sum <= over_delta);
1848 over_target = over_sum - over_delta;
1853 list_for_each_entry_safe(iocg, tiocg, &over_hwa, walk_list) {
1855 iocg->hweight_after_donation =
1856 div_u64((u64)iocg->hweight_after_donation *
1857 over_target, over_sum);
1858 list_del_init(&iocg->walk_list);
1862 * Build pre-order inner node walk list and prepare for donation
1863 * adjustment calculations.
1865 list_for_each_entry(iocg, surpluses, surplus_list) {
1866 iocg_build_inner_walk(iocg, &inner_walk);
1869 root_iocg = list_first_entry(&inner_walk, struct ioc_gq, walk_list);
1870 WARN_ON_ONCE(root_iocg->level > 0);
1872 list_for_each_entry(iocg, &inner_walk, walk_list) {
1873 iocg->child_adjusted_sum = 0;
1874 iocg->hweight_donating = 0;
1875 iocg->hweight_after_donation = 0;
1879 * Propagate the donating budget (b_t) and after donation budget (b'_t)
1882 list_for_each_entry(iocg, surpluses, surplus_list) {
1883 struct ioc_gq *parent = iocg->ancestors[iocg->level - 1];
1885 parent->hweight_donating += iocg->hweight_donating;
1886 parent->hweight_after_donation += iocg->hweight_after_donation;
1889 list_for_each_entry_reverse(iocg, &inner_walk, walk_list) {
1890 if (iocg->level > 0) {
1891 struct ioc_gq *parent = iocg->ancestors[iocg->level - 1];
1893 parent->hweight_donating += iocg->hweight_donating;
1894 parent->hweight_after_donation += iocg->hweight_after_donation;
1899 * Calculate inner hwa's (b) and make sure the donation values are
1900 * within the accepted ranges as we're doing low res calculations with
1903 list_for_each_entry(iocg, &inner_walk, walk_list) {
1905 struct ioc_gq *parent = iocg->ancestors[iocg->level - 1];
1907 iocg->hweight_active = DIV64_U64_ROUND_UP(
1908 (u64)parent->hweight_active * iocg->active,
1909 parent->child_active_sum);
1913 iocg->hweight_donating = min(iocg->hweight_donating,
1914 iocg->hweight_active);
1915 iocg->hweight_after_donation = min(iocg->hweight_after_donation,
1916 iocg->hweight_donating - 1);
1917 if (WARN_ON_ONCE(iocg->hweight_active <= 1 ||
1918 iocg->hweight_donating <= 1 ||
1919 iocg->hweight_after_donation == 0)) {
1920 pr_warn("iocg: invalid donation weights in ");
1921 pr_cont_cgroup_path(iocg_to_blkg(iocg)->blkcg->css.cgroup);
1922 pr_cont(": active=%u donating=%u after=%u\n",
1923 iocg->hweight_active, iocg->hweight_donating,
1924 iocg->hweight_after_donation);
1929 * Calculate the global donation rate (gamma) - the rate to adjust
1930 * non-donating budgets by.
1932 * No need to use 64bit multiplication here as the first operand is
1933 * guaranteed to be smaller than WEIGHT_ONE (1<<16).
1935 * We know that there are beneficiary nodes and the sum of the donating
1936 * hweights can't be whole; however, due to the round-ups during hweight
1937 * calculations, root_iocg->hweight_donating might still end up equal to
1938 * or greater than whole. Limit the range when calculating the divider.
1940 * gamma = (1 - t_r') / (1 - t_r)
1942 gamma = DIV_ROUND_UP(
1943 (WEIGHT_ONE - root_iocg->hweight_after_donation) * WEIGHT_ONE,
1944 WEIGHT_ONE - min_t(u32, root_iocg->hweight_donating, WEIGHT_ONE - 1));
1947 * Calculate adjusted hwi, child_adjusted_sum and inuse for the inner
1950 list_for_each_entry(iocg, &inner_walk, walk_list) {
1951 struct ioc_gq *parent;
1952 u32 inuse, wpt, wptp;
1955 if (iocg->level == 0) {
1956 /* adjusted weight sum for 1st level: s' = s * b_pf / b'_pf */
1957 iocg->child_adjusted_sum = DIV64_U64_ROUND_UP(
1958 iocg->child_active_sum * (WEIGHT_ONE - iocg->hweight_donating),
1959 WEIGHT_ONE - iocg->hweight_after_donation);
1963 parent = iocg->ancestors[iocg->level - 1];
1965 /* b' = gamma * b_f + b_t' */
1966 iocg->hweight_inuse = DIV64_U64_ROUND_UP(
1967 (u64)gamma * (iocg->hweight_active - iocg->hweight_donating),
1968 WEIGHT_ONE) + iocg->hweight_after_donation;
1970 /* w' = s' * b' / b'_p */
1971 inuse = DIV64_U64_ROUND_UP(
1972 (u64)parent->child_adjusted_sum * iocg->hweight_inuse,
1973 parent->hweight_inuse);
1975 /* adjusted weight sum for children: s' = s_f + s_t * w'_pt / w_pt */
1976 st = DIV64_U64_ROUND_UP(
1977 iocg->child_active_sum * iocg->hweight_donating,
1978 iocg->hweight_active);
1979 sf = iocg->child_active_sum - st;
1980 wpt = DIV64_U64_ROUND_UP(
1981 (u64)iocg->active * iocg->hweight_donating,
1982 iocg->hweight_active);
1983 wptp = DIV64_U64_ROUND_UP(
1984 (u64)inuse * iocg->hweight_after_donation,
1985 iocg->hweight_inuse);
1987 iocg->child_adjusted_sum = sf + DIV64_U64_ROUND_UP(st * wptp, wpt);
1991 * All inner nodes now have ->hweight_inuse and ->child_adjusted_sum and
1992 * we can finally determine leaf adjustments.
1994 list_for_each_entry(iocg, surpluses, surplus_list) {
1995 struct ioc_gq *parent = iocg->ancestors[iocg->level - 1];
1999 * In-debt iocgs participated in the donation calculation with
2000 * the minimum target hweight_inuse. Configuring inuse
2001 * accordingly would work fine but debt handling expects
2002 * @iocg->inuse stay at the minimum and we don't wanna
2005 if (iocg->abs_vdebt) {
2006 WARN_ON_ONCE(iocg->inuse > 1);
2010 /* w' = s' * b' / b'_p, note that b' == b'_t for donating leaves */
2011 inuse = DIV64_U64_ROUND_UP(
2012 parent->child_adjusted_sum * iocg->hweight_after_donation,
2013 parent->hweight_inuse);
2015 TRACE_IOCG_PATH(inuse_transfer, iocg, now,
2017 iocg->hweight_inuse,
2018 iocg->hweight_after_donation);
2020 __propagate_weights(iocg, iocg->active, inuse, true, now);
2023 /* walk list should be dissolved after use */
2024 list_for_each_entry_safe(iocg, tiocg, &inner_walk, walk_list)
2025 list_del_init(&iocg->walk_list);
2029 * A low weight iocg can amass a large amount of debt, for example, when
2030 * anonymous memory gets reclaimed aggressively. If the system has a lot of
2031 * memory paired with a slow IO device, the debt can span multiple seconds or
2032 * more. If there are no other subsequent IO issuers, the in-debt iocg may end
2033 * up blocked paying its debt while the IO device is idle.
2035 * The following protects against such cases. If the device has been
2036 * sufficiently idle for a while, the debts are halved and delays are
2039 static void ioc_forgive_debts(struct ioc *ioc, u64 usage_us_sum, int nr_debtors,
2040 struct ioc_now *now)
2042 struct ioc_gq *iocg;
2043 u64 dur, usage_pct, nr_cycles;
2045 /* if no debtor, reset the cycle */
2047 ioc->dfgv_period_at = now->now;
2048 ioc->dfgv_period_rem = 0;
2049 ioc->dfgv_usage_us_sum = 0;
2054 * Debtors can pass through a lot of writes choking the device and we
2055 * don't want to be forgiving debts while the device is struggling from
2056 * write bursts. If we're missing latency targets, consider the device
2059 if (ioc->busy_level > 0)
2060 usage_us_sum = max_t(u64, usage_us_sum, ioc->period_us);
2062 ioc->dfgv_usage_us_sum += usage_us_sum;
2063 if (time_before64(now->now, ioc->dfgv_period_at + DFGV_PERIOD))
2067 * At least DFGV_PERIOD has passed since the last period. Calculate the
2068 * average usage and reset the period counters.
2070 dur = now->now - ioc->dfgv_period_at;
2071 usage_pct = div64_u64(100 * ioc->dfgv_usage_us_sum, dur);
2073 ioc->dfgv_period_at = now->now;
2074 ioc->dfgv_usage_us_sum = 0;
2076 /* if was too busy, reset everything */
2077 if (usage_pct > DFGV_USAGE_PCT) {
2078 ioc->dfgv_period_rem = 0;
2083 * Usage is lower than threshold. Let's forgive some debts. Debt
2084 * forgiveness runs off of the usual ioc timer but its period usually
2085 * doesn't match ioc's. Compensate the difference by performing the
2086 * reduction as many times as would fit in the duration since the last
2087 * run and carrying over the left-over duration in @ioc->dfgv_period_rem
2088 * - if ioc period is 75% of DFGV_PERIOD, one out of three consecutive
2089 * reductions is doubled.
2091 nr_cycles = dur + ioc->dfgv_period_rem;
2092 ioc->dfgv_period_rem = do_div(nr_cycles, DFGV_PERIOD);
2094 list_for_each_entry(iocg, &ioc->active_iocgs, active_list) {
2095 u64 __maybe_unused old_debt, __maybe_unused old_delay;
2097 if (!iocg->abs_vdebt && !iocg->delay)
2100 spin_lock(&iocg->waitq.lock);
2102 old_debt = iocg->abs_vdebt;
2103 old_delay = iocg->delay;
2105 if (iocg->abs_vdebt)
2106 iocg->abs_vdebt = iocg->abs_vdebt >> nr_cycles ?: 1;
2108 iocg->delay = iocg->delay >> nr_cycles ?: 1;
2110 iocg_kick_waitq(iocg, true, now);
2112 TRACE_IOCG_PATH(iocg_forgive_debt, iocg, now, usage_pct,
2113 old_debt, iocg->abs_vdebt,
2114 old_delay, iocg->delay);
2116 spin_unlock(&iocg->waitq.lock);
2121 * Check the active iocgs' state to avoid oversleeping and deactive
2124 * Since waiters determine the sleep durations based on the vrate
2125 * they saw at the time of sleep, if vrate has increased, some
2126 * waiters could be sleeping for too long. Wake up tardy waiters
2127 * which should have woken up in the last period and expire idle
2130 static int ioc_check_iocgs(struct ioc *ioc, struct ioc_now *now)
2133 struct ioc_gq *iocg, *tiocg;
2135 list_for_each_entry_safe(iocg, tiocg, &ioc->active_iocgs, active_list) {
2136 if (!waitqueue_active(&iocg->waitq) && !iocg->abs_vdebt &&
2137 !iocg->delay && !iocg_is_idle(iocg))
2140 spin_lock(&iocg->waitq.lock);
2142 /* flush wait and indebt stat deltas */
2143 if (iocg->wait_since) {
2144 iocg->local_stat.wait_us += now->now - iocg->wait_since;
2145 iocg->wait_since = now->now;
2147 if (iocg->indebt_since) {
2148 iocg->local_stat.indebt_us +=
2149 now->now - iocg->indebt_since;
2150 iocg->indebt_since = now->now;
2152 if (iocg->indelay_since) {
2153 iocg->local_stat.indelay_us +=
2154 now->now - iocg->indelay_since;
2155 iocg->indelay_since = now->now;
2158 if (waitqueue_active(&iocg->waitq) || iocg->abs_vdebt ||
2160 /* might be oversleeping vtime / hweight changes, kick */
2161 iocg_kick_waitq(iocg, true, now);
2162 if (iocg->abs_vdebt || iocg->delay)
2164 } else if (iocg_is_idle(iocg)) {
2165 /* no waiter and idle, deactivate */
2166 u64 vtime = atomic64_read(&iocg->vtime);
2170 * @iocg has been inactive for a full duration and will
2171 * have a high budget. Account anything above target as
2172 * error and throw away. On reactivation, it'll start
2173 * with the target budget.
2175 excess = now->vnow - vtime - ioc->margins.target;
2179 current_hweight(iocg, NULL, &old_hwi);
2180 ioc->vtime_err -= div64_u64(excess * old_hwi,
2184 TRACE_IOCG_PATH(iocg_idle, iocg, now,
2185 atomic64_read(&iocg->active_period),
2186 atomic64_read(&ioc->cur_period), vtime);
2187 __propagate_weights(iocg, 0, 0, false, now);
2188 list_del_init(&iocg->active_list);
2191 spin_unlock(&iocg->waitq.lock);
2194 commit_weights(ioc);
2198 static void ioc_timer_fn(struct timer_list *timer)
2200 struct ioc *ioc = container_of(timer, struct ioc, timer);
2201 struct ioc_gq *iocg, *tiocg;
2203 LIST_HEAD(surpluses);
2204 int nr_debtors, nr_shortages = 0, nr_lagging = 0;
2205 u64 usage_us_sum = 0;
2206 u32 ppm_rthr = MILLION - ioc->params.qos[QOS_RPPM];
2207 u32 ppm_wthr = MILLION - ioc->params.qos[QOS_WPPM];
2208 u32 missed_ppm[2], rq_wait_pct;
2210 int prev_busy_level;
2212 /* how were the latencies during the period? */
2213 ioc_lat_stat(ioc, missed_ppm, &rq_wait_pct);
2215 /* take care of active iocgs */
2216 spin_lock_irq(&ioc->lock);
2220 period_vtime = now.vnow - ioc->period_at_vtime;
2221 if (WARN_ON_ONCE(!period_vtime)) {
2222 spin_unlock_irq(&ioc->lock);
2226 nr_debtors = ioc_check_iocgs(ioc, &now);
2229 * Wait and indebt stat are flushed above and the donation calculation
2230 * below needs updated usage stat. Let's bring stat up-to-date.
2232 iocg_flush_stat(&ioc->active_iocgs, &now);
2234 /* calc usage and see whether some weights need to be moved around */
2235 list_for_each_entry(iocg, &ioc->active_iocgs, active_list) {
2236 u64 vdone, vtime, usage_us;
2237 u32 hw_active, hw_inuse;
2240 * Collect unused and wind vtime closer to vnow to prevent
2241 * iocgs from accumulating a large amount of budget.
2243 vdone = atomic64_read(&iocg->done_vtime);
2244 vtime = atomic64_read(&iocg->vtime);
2245 current_hweight(iocg, &hw_active, &hw_inuse);
2248 * Latency QoS detection doesn't account for IOs which are
2249 * in-flight for longer than a period. Detect them by
2250 * comparing vdone against period start. If lagging behind
2251 * IOs from past periods, don't increase vrate.
2253 if ((ppm_rthr != MILLION || ppm_wthr != MILLION) &&
2254 !atomic_read(&iocg_to_blkg(iocg)->use_delay) &&
2255 time_after64(vtime, vdone) &&
2256 time_after64(vtime, now.vnow -
2257 MAX_LAGGING_PERIODS * period_vtime) &&
2258 time_before64(vdone, now.vnow - period_vtime))
2262 * Determine absolute usage factoring in in-flight IOs to avoid
2263 * high-latency completions appearing as idle.
2265 usage_us = iocg->usage_delta_us;
2266 usage_us_sum += usage_us;
2268 /* see whether there's surplus vtime */
2269 WARN_ON_ONCE(!list_empty(&iocg->surplus_list));
2270 if (hw_inuse < hw_active ||
2271 (!waitqueue_active(&iocg->waitq) &&
2272 time_before64(vtime, now.vnow - ioc->margins.low))) {
2273 u32 hwa, old_hwi, hwm, new_hwi, usage;
2276 if (vdone != vtime) {
2277 u64 inflight_us = DIV64_U64_ROUND_UP(
2278 cost_to_abs_cost(vtime - vdone, hw_inuse),
2279 ioc->vtime_base_rate);
2281 usage_us = max(usage_us, inflight_us);
2284 /* convert to hweight based usage ratio */
2285 if (time_after64(iocg->activated_at, ioc->period_at))
2286 usage_dur = max_t(u64, now.now - iocg->activated_at, 1);
2288 usage_dur = max_t(u64, now.now - ioc->period_at, 1);
2290 usage = clamp_t(u32,
2291 DIV64_U64_ROUND_UP(usage_us * WEIGHT_ONE,
2296 * Already donating or accumulated enough to start.
2297 * Determine the donation amount.
2299 current_hweight(iocg, &hwa, &old_hwi);
2300 hwm = current_hweight_max(iocg);
2301 new_hwi = hweight_after_donation(iocg, old_hwi, hwm,
2303 if (new_hwi < hwm) {
2304 iocg->hweight_donating = hwa;
2305 iocg->hweight_after_donation = new_hwi;
2306 list_add(&iocg->surplus_list, &surpluses);
2308 TRACE_IOCG_PATH(inuse_shortage, iocg, &now,
2309 iocg->inuse, iocg->active,
2310 iocg->hweight_inuse, new_hwi);
2312 __propagate_weights(iocg, iocg->active,
2313 iocg->active, true, &now);
2317 /* genuinely short on vtime */
2322 if (!list_empty(&surpluses) && nr_shortages)
2323 transfer_surpluses(&surpluses, &now);
2325 commit_weights(ioc);
2327 /* surplus list should be dissolved after use */
2328 list_for_each_entry_safe(iocg, tiocg, &surpluses, surplus_list)
2329 list_del_init(&iocg->surplus_list);
2332 * If q is getting clogged or we're missing too much, we're issuing
2333 * too much IO and should lower vtime rate. If we're not missing
2334 * and experiencing shortages but not surpluses, we're too stingy
2335 * and should increase vtime rate.
2337 prev_busy_level = ioc->busy_level;
2338 if (rq_wait_pct > RQ_WAIT_BUSY_PCT ||
2339 missed_ppm[READ] > ppm_rthr ||
2340 missed_ppm[WRITE] > ppm_wthr) {
2341 /* clearly missing QoS targets, slow down vrate */
2342 ioc->busy_level = max(ioc->busy_level, 0);
2344 } else if (rq_wait_pct <= RQ_WAIT_BUSY_PCT * UNBUSY_THR_PCT / 100 &&
2345 missed_ppm[READ] <= ppm_rthr * UNBUSY_THR_PCT / 100 &&
2346 missed_ppm[WRITE] <= ppm_wthr * UNBUSY_THR_PCT / 100) {
2347 /* QoS targets are being met with >25% margin */
2350 * We're throttling while the device has spare
2351 * capacity. If vrate was being slowed down, stop.
2353 ioc->busy_level = min(ioc->busy_level, 0);
2356 * If there are IOs spanning multiple periods, wait
2357 * them out before pushing the device harder.
2363 * Nobody is being throttled and the users aren't
2364 * issuing enough IOs to saturate the device. We
2365 * simply don't know how close the device is to
2366 * saturation. Coast.
2368 ioc->busy_level = 0;
2371 /* inside the hysterisis margin, we're good */
2372 ioc->busy_level = 0;
2375 ioc->busy_level = clamp(ioc->busy_level, -1000, 1000);
2377 ioc_adjust_base_vrate(ioc, rq_wait_pct, nr_lagging, nr_shortages,
2378 prev_busy_level, missed_ppm);
2380 ioc_refresh_params(ioc, false);
2382 ioc_forgive_debts(ioc, usage_us_sum, nr_debtors, &now);
2385 * This period is done. Move onto the next one. If nothing's
2386 * going on with the device, stop the timer.
2388 atomic64_inc(&ioc->cur_period);
2390 if (ioc->running != IOC_STOP) {
2391 if (!list_empty(&ioc->active_iocgs)) {
2392 ioc_start_period(ioc, &now);
2394 ioc->busy_level = 0;
2396 ioc->running = IOC_IDLE;
2399 ioc_refresh_vrate(ioc, &now);
2402 spin_unlock_irq(&ioc->lock);
2405 static u64 adjust_inuse_and_calc_cost(struct ioc_gq *iocg, u64 vtime,
2406 u64 abs_cost, struct ioc_now *now)
2408 struct ioc *ioc = iocg->ioc;
2409 struct ioc_margins *margins = &ioc->margins;
2410 u32 __maybe_unused old_inuse = iocg->inuse, __maybe_unused old_hwi;
2413 u64 cost, new_inuse;
2415 current_hweight(iocg, NULL, &hwi);
2417 cost = abs_cost_to_cost(abs_cost, hwi);
2418 margin = now->vnow - vtime - cost;
2420 /* debt handling owns inuse for debtors */
2421 if (iocg->abs_vdebt)
2425 * We only increase inuse during period and do so if the margin has
2426 * deteriorated since the previous adjustment.
2428 if (margin >= iocg->saved_margin || margin >= margins->low ||
2429 iocg->inuse == iocg->active)
2432 spin_lock_irq(&ioc->lock);
2434 /* we own inuse only when @iocg is in the normal active state */
2435 if (iocg->abs_vdebt || list_empty(&iocg->active_list)) {
2436 spin_unlock_irq(&ioc->lock);
2441 * Bump up inuse till @abs_cost fits in the existing budget.
2442 * adj_step must be determined after acquiring ioc->lock - we might
2443 * have raced and lost to another thread for activation and could
2444 * be reading 0 iocg->active before ioc->lock which will lead to
2447 new_inuse = iocg->inuse;
2448 adj_step = DIV_ROUND_UP(iocg->active * INUSE_ADJ_STEP_PCT, 100);
2450 new_inuse = new_inuse + adj_step;
2451 propagate_weights(iocg, iocg->active, new_inuse, true, now);
2452 current_hweight(iocg, NULL, &hwi);
2453 cost = abs_cost_to_cost(abs_cost, hwi);
2454 } while (time_after64(vtime + cost, now->vnow) &&
2455 iocg->inuse != iocg->active);
2457 spin_unlock_irq(&ioc->lock);
2459 TRACE_IOCG_PATH(inuse_adjust, iocg, now,
2460 old_inuse, iocg->inuse, old_hwi, hwi);
2465 static void calc_vtime_cost_builtin(struct bio *bio, struct ioc_gq *iocg,
2466 bool is_merge, u64 *costp)
2468 struct ioc *ioc = iocg->ioc;
2469 u64 coef_seqio, coef_randio, coef_page;
2470 u64 pages = max_t(u64, bio_sectors(bio) >> IOC_SECT_TO_PAGE_SHIFT, 1);
2474 switch (bio_op(bio)) {
2476 coef_seqio = ioc->params.lcoefs[LCOEF_RSEQIO];
2477 coef_randio = ioc->params.lcoefs[LCOEF_RRANDIO];
2478 coef_page = ioc->params.lcoefs[LCOEF_RPAGE];
2481 coef_seqio = ioc->params.lcoefs[LCOEF_WSEQIO];
2482 coef_randio = ioc->params.lcoefs[LCOEF_WRANDIO];
2483 coef_page = ioc->params.lcoefs[LCOEF_WPAGE];
2490 seek_pages = abs(bio->bi_iter.bi_sector - iocg->cursor);
2491 seek_pages >>= IOC_SECT_TO_PAGE_SHIFT;
2495 if (seek_pages > LCOEF_RANDIO_PAGES) {
2496 cost += coef_randio;
2501 cost += pages * coef_page;
2506 static u64 calc_vtime_cost(struct bio *bio, struct ioc_gq *iocg, bool is_merge)
2510 calc_vtime_cost_builtin(bio, iocg, is_merge, &cost);
2514 static void calc_size_vtime_cost_builtin(struct request *rq, struct ioc *ioc,
2517 unsigned int pages = blk_rq_stats_sectors(rq) >> IOC_SECT_TO_PAGE_SHIFT;
2519 switch (req_op(rq)) {
2521 *costp = pages * ioc->params.lcoefs[LCOEF_RPAGE];
2524 *costp = pages * ioc->params.lcoefs[LCOEF_WPAGE];
2531 static u64 calc_size_vtime_cost(struct request *rq, struct ioc *ioc)
2535 calc_size_vtime_cost_builtin(rq, ioc, &cost);
2539 static void ioc_rqos_throttle(struct rq_qos *rqos, struct bio *bio)
2541 struct blkcg_gq *blkg = bio->bi_blkg;
2542 struct ioc *ioc = rqos_to_ioc(rqos);
2543 struct ioc_gq *iocg = blkg_to_iocg(blkg);
2545 struct iocg_wait wait;
2546 u64 abs_cost, cost, vtime;
2547 bool use_debt, ioc_locked;
2548 unsigned long flags;
2550 /* bypass IOs if disabled, still initializing, or for root cgroup */
2551 if (!ioc->enabled || !iocg || !iocg->level)
2554 /* calculate the absolute vtime cost */
2555 abs_cost = calc_vtime_cost(bio, iocg, false);
2559 if (!iocg_activate(iocg, &now))
2562 iocg->cursor = bio_end_sector(bio);
2563 vtime = atomic64_read(&iocg->vtime);
2564 cost = adjust_inuse_and_calc_cost(iocg, vtime, abs_cost, &now);
2567 * If no one's waiting and within budget, issue right away. The
2568 * tests are racy but the races aren't systemic - we only miss once
2569 * in a while which is fine.
2571 if (!waitqueue_active(&iocg->waitq) && !iocg->abs_vdebt &&
2572 time_before_eq64(vtime + cost, now.vnow)) {
2573 iocg_commit_bio(iocg, bio, abs_cost, cost);
2578 * We're over budget. This can be handled in two ways. IOs which may
2579 * cause priority inversions are punted to @ioc->aux_iocg and charged as
2580 * debt. Otherwise, the issuer is blocked on @iocg->waitq. Debt handling
2581 * requires @ioc->lock, waitq handling @iocg->waitq.lock. Determine
2582 * whether debt handling is needed and acquire locks accordingly.
2584 use_debt = bio_issue_as_root_blkg(bio) || fatal_signal_pending(current);
2585 ioc_locked = use_debt || READ_ONCE(iocg->abs_vdebt);
2587 iocg_lock(iocg, ioc_locked, &flags);
2590 * @iocg must stay activated for debt and waitq handling. Deactivation
2591 * is synchronized against both ioc->lock and waitq.lock and we won't
2592 * get deactivated as long as we're waiting or has debt, so we're good
2593 * if we're activated here. In the unlikely cases that we aren't, just
2596 if (unlikely(list_empty(&iocg->active_list))) {
2597 iocg_unlock(iocg, ioc_locked, &flags);
2598 iocg_commit_bio(iocg, bio, abs_cost, cost);
2603 * We're over budget. If @bio has to be issued regardless, remember
2604 * the abs_cost instead of advancing vtime. iocg_kick_waitq() will pay
2605 * off the debt before waking more IOs.
2607 * This way, the debt is continuously paid off each period with the
2608 * actual budget available to the cgroup. If we just wound vtime, we
2609 * would incorrectly use the current hw_inuse for the entire amount
2610 * which, for example, can lead to the cgroup staying blocked for a
2611 * long time even with substantially raised hw_inuse.
2613 * An iocg with vdebt should stay online so that the timer can keep
2614 * deducting its vdebt and [de]activate use_delay mechanism
2615 * accordingly. We don't want to race against the timer trying to
2616 * clear them and leave @iocg inactive w/ dangling use_delay heavily
2617 * penalizing the cgroup and its descendants.
2620 iocg_incur_debt(iocg, abs_cost, &now);
2621 if (iocg_kick_delay(iocg, &now))
2622 blkcg_schedule_throttle(rqos->q,
2623 (bio->bi_opf & REQ_SWAP) == REQ_SWAP);
2624 iocg_unlock(iocg, ioc_locked, &flags);
2628 /* guarantee that iocgs w/ waiters have maximum inuse */
2629 if (!iocg->abs_vdebt && iocg->inuse != iocg->active) {
2631 iocg_unlock(iocg, false, &flags);
2635 propagate_weights(iocg, iocg->active, iocg->active, true,
2640 * Append self to the waitq and schedule the wakeup timer if we're
2641 * the first waiter. The timer duration is calculated based on the
2642 * current vrate. vtime and hweight changes can make it too short
2643 * or too long. Each wait entry records the absolute cost it's
2644 * waiting for to allow re-evaluation using a custom wait entry.
2646 * If too short, the timer simply reschedules itself. If too long,
2647 * the period timer will notice and trigger wakeups.
2649 * All waiters are on iocg->waitq and the wait states are
2650 * synchronized using waitq.lock.
2652 init_waitqueue_func_entry(&wait.wait, iocg_wake_fn);
2653 wait.wait.private = current;
2655 wait.abs_cost = abs_cost;
2656 wait.committed = false; /* will be set true by waker */
2658 __add_wait_queue_entry_tail(&iocg->waitq, &wait.wait);
2659 iocg_kick_waitq(iocg, ioc_locked, &now);
2661 iocg_unlock(iocg, ioc_locked, &flags);
2664 set_current_state(TASK_UNINTERRUPTIBLE);
2670 /* waker already committed us, proceed */
2671 finish_wait(&iocg->waitq, &wait.wait);
2674 static void ioc_rqos_merge(struct rq_qos *rqos, struct request *rq,
2677 struct ioc_gq *iocg = blkg_to_iocg(bio->bi_blkg);
2678 struct ioc *ioc = rqos_to_ioc(rqos);
2679 sector_t bio_end = bio_end_sector(bio);
2681 u64 vtime, abs_cost, cost;
2682 unsigned long flags;
2684 /* bypass if disabled, still initializing, or for root cgroup */
2685 if (!ioc->enabled || !iocg || !iocg->level)
2688 abs_cost = calc_vtime_cost(bio, iocg, true);
2694 vtime = atomic64_read(&iocg->vtime);
2695 cost = adjust_inuse_and_calc_cost(iocg, vtime, abs_cost, &now);
2697 /* update cursor if backmerging into the request at the cursor */
2698 if (blk_rq_pos(rq) < bio_end &&
2699 blk_rq_pos(rq) + blk_rq_sectors(rq) == iocg->cursor)
2700 iocg->cursor = bio_end;
2703 * Charge if there's enough vtime budget and the existing request has
2706 if (rq->bio && rq->bio->bi_iocost_cost &&
2707 time_before_eq64(atomic64_read(&iocg->vtime) + cost, now.vnow)) {
2708 iocg_commit_bio(iocg, bio, abs_cost, cost);
2713 * Otherwise, account it as debt if @iocg is online, which it should
2714 * be for the vast majority of cases. See debt handling in
2715 * ioc_rqos_throttle() for details.
2717 spin_lock_irqsave(&ioc->lock, flags);
2718 spin_lock(&iocg->waitq.lock);
2720 if (likely(!list_empty(&iocg->active_list))) {
2721 iocg_incur_debt(iocg, abs_cost, &now);
2722 if (iocg_kick_delay(iocg, &now))
2723 blkcg_schedule_throttle(rqos->q,
2724 (bio->bi_opf & REQ_SWAP) == REQ_SWAP);
2726 iocg_commit_bio(iocg, bio, abs_cost, cost);
2729 spin_unlock(&iocg->waitq.lock);
2730 spin_unlock_irqrestore(&ioc->lock, flags);
2733 static void ioc_rqos_done_bio(struct rq_qos *rqos, struct bio *bio)
2735 struct ioc_gq *iocg = blkg_to_iocg(bio->bi_blkg);
2737 if (iocg && bio->bi_iocost_cost)
2738 atomic64_add(bio->bi_iocost_cost, &iocg->done_vtime);
2741 static void ioc_rqos_done(struct rq_qos *rqos, struct request *rq)
2743 struct ioc *ioc = rqos_to_ioc(rqos);
2744 struct ioc_pcpu_stat *ccs;
2745 u64 on_q_ns, rq_wait_ns, size_nsec;
2748 if (!ioc->enabled || !rq->alloc_time_ns || !rq->start_time_ns)
2751 switch (req_op(rq) & REQ_OP_MASK) {
2764 on_q_ns = ktime_get_ns() - rq->alloc_time_ns;
2765 rq_wait_ns = rq->start_time_ns - rq->alloc_time_ns;
2766 size_nsec = div64_u64(calc_size_vtime_cost(rq, ioc), VTIME_PER_NSEC);
2768 ccs = get_cpu_ptr(ioc->pcpu_stat);
2770 if (on_q_ns <= size_nsec ||
2771 on_q_ns - size_nsec <= ioc->params.qos[pidx] * NSEC_PER_USEC)
2772 local_inc(&ccs->missed[rw].nr_met);
2774 local_inc(&ccs->missed[rw].nr_missed);
2776 local64_add(rq_wait_ns, &ccs->rq_wait_ns);
2781 static void ioc_rqos_queue_depth_changed(struct rq_qos *rqos)
2783 struct ioc *ioc = rqos_to_ioc(rqos);
2785 spin_lock_irq(&ioc->lock);
2786 ioc_refresh_params(ioc, false);
2787 spin_unlock_irq(&ioc->lock);
2790 static void ioc_rqos_exit(struct rq_qos *rqos)
2792 struct ioc *ioc = rqos_to_ioc(rqos);
2794 blkcg_deactivate_policy(rqos->q, &blkcg_policy_iocost);
2796 spin_lock_irq(&ioc->lock);
2797 ioc->running = IOC_STOP;
2798 spin_unlock_irq(&ioc->lock);
2800 del_timer_sync(&ioc->timer);
2801 free_percpu(ioc->pcpu_stat);
2805 static struct rq_qos_ops ioc_rqos_ops = {
2806 .throttle = ioc_rqos_throttle,
2807 .merge = ioc_rqos_merge,
2808 .done_bio = ioc_rqos_done_bio,
2809 .done = ioc_rqos_done,
2810 .queue_depth_changed = ioc_rqos_queue_depth_changed,
2811 .exit = ioc_rqos_exit,
2814 static int blk_iocost_init(struct request_queue *q)
2817 struct rq_qos *rqos;
2820 ioc = kzalloc(sizeof(*ioc), GFP_KERNEL);
2824 ioc->pcpu_stat = alloc_percpu(struct ioc_pcpu_stat);
2825 if (!ioc->pcpu_stat) {
2830 for_each_possible_cpu(cpu) {
2831 struct ioc_pcpu_stat *ccs = per_cpu_ptr(ioc->pcpu_stat, cpu);
2833 for (i = 0; i < ARRAY_SIZE(ccs->missed); i++) {
2834 local_set(&ccs->missed[i].nr_met, 0);
2835 local_set(&ccs->missed[i].nr_missed, 0);
2837 local64_set(&ccs->rq_wait_ns, 0);
2841 rqos->id = RQ_QOS_COST;
2842 rqos->ops = &ioc_rqos_ops;
2845 spin_lock_init(&ioc->lock);
2846 timer_setup(&ioc->timer, ioc_timer_fn, 0);
2847 INIT_LIST_HEAD(&ioc->active_iocgs);
2849 ioc->running = IOC_IDLE;
2850 ioc->vtime_base_rate = VTIME_PER_USEC;
2851 atomic64_set(&ioc->vtime_rate, VTIME_PER_USEC);
2852 seqcount_spinlock_init(&ioc->period_seqcount, &ioc->lock);
2853 ioc->period_at = ktime_to_us(ktime_get());
2854 atomic64_set(&ioc->cur_period, 0);
2855 atomic_set(&ioc->hweight_gen, 0);
2857 spin_lock_irq(&ioc->lock);
2858 ioc->autop_idx = AUTOP_INVALID;
2859 ioc_refresh_params(ioc, true);
2860 spin_unlock_irq(&ioc->lock);
2863 * rqos must be added before activation to allow iocg_pd_init() to
2864 * lookup the ioc from q. This means that the rqos methods may get
2865 * called before policy activation completion, can't assume that the
2866 * target bio has an iocg associated and need to test for NULL iocg.
2868 rq_qos_add(q, rqos);
2869 ret = blkcg_activate_policy(q, &blkcg_policy_iocost);
2871 rq_qos_del(q, rqos);
2872 free_percpu(ioc->pcpu_stat);
2879 static struct blkcg_policy_data *ioc_cpd_alloc(gfp_t gfp)
2881 struct ioc_cgrp *iocc;
2883 iocc = kzalloc(sizeof(struct ioc_cgrp), gfp);
2887 iocc->dfl_weight = CGROUP_WEIGHT_DFL * WEIGHT_ONE;
2891 static void ioc_cpd_free(struct blkcg_policy_data *cpd)
2893 kfree(container_of(cpd, struct ioc_cgrp, cpd));
2896 static struct blkg_policy_data *ioc_pd_alloc(gfp_t gfp, struct request_queue *q,
2897 struct blkcg *blkcg)
2899 int levels = blkcg->css.cgroup->level + 1;
2900 struct ioc_gq *iocg;
2902 iocg = kzalloc_node(struct_size(iocg, ancestors, levels), gfp, q->node);
2906 iocg->pcpu_stat = alloc_percpu_gfp(struct iocg_pcpu_stat, gfp);
2907 if (!iocg->pcpu_stat) {
2915 static void ioc_pd_init(struct blkg_policy_data *pd)
2917 struct ioc_gq *iocg = pd_to_iocg(pd);
2918 struct blkcg_gq *blkg = pd_to_blkg(&iocg->pd);
2919 struct ioc *ioc = q_to_ioc(blkg->q);
2921 struct blkcg_gq *tblkg;
2922 unsigned long flags;
2927 atomic64_set(&iocg->vtime, now.vnow);
2928 atomic64_set(&iocg->done_vtime, now.vnow);
2929 atomic64_set(&iocg->active_period, atomic64_read(&ioc->cur_period));
2930 INIT_LIST_HEAD(&iocg->active_list);
2931 INIT_LIST_HEAD(&iocg->walk_list);
2932 INIT_LIST_HEAD(&iocg->surplus_list);
2933 iocg->hweight_active = WEIGHT_ONE;
2934 iocg->hweight_inuse = WEIGHT_ONE;
2936 init_waitqueue_head(&iocg->waitq);
2937 hrtimer_init(&iocg->waitq_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
2938 iocg->waitq_timer.function = iocg_waitq_timer_fn;
2940 iocg->level = blkg->blkcg->css.cgroup->level;
2942 for (tblkg = blkg; tblkg; tblkg = tblkg->parent) {
2943 struct ioc_gq *tiocg = blkg_to_iocg(tblkg);
2944 iocg->ancestors[tiocg->level] = tiocg;
2947 spin_lock_irqsave(&ioc->lock, flags);
2948 weight_updated(iocg, &now);
2949 spin_unlock_irqrestore(&ioc->lock, flags);
2952 static void ioc_pd_free(struct blkg_policy_data *pd)
2954 struct ioc_gq *iocg = pd_to_iocg(pd);
2955 struct ioc *ioc = iocg->ioc;
2956 unsigned long flags;
2959 spin_lock_irqsave(&ioc->lock, flags);
2961 if (!list_empty(&iocg->active_list)) {
2965 propagate_weights(iocg, 0, 0, false, &now);
2966 list_del_init(&iocg->active_list);
2969 WARN_ON_ONCE(!list_empty(&iocg->walk_list));
2970 WARN_ON_ONCE(!list_empty(&iocg->surplus_list));
2972 spin_unlock_irqrestore(&ioc->lock, flags);
2974 hrtimer_cancel(&iocg->waitq_timer);
2976 free_percpu(iocg->pcpu_stat);
2980 static size_t ioc_pd_stat(struct blkg_policy_data *pd, char *buf, size_t size)
2982 struct ioc_gq *iocg = pd_to_iocg(pd);
2983 struct ioc *ioc = iocg->ioc;
2989 if (iocg->level == 0) {
2990 unsigned vp10k = DIV64_U64_ROUND_CLOSEST(
2991 ioc->vtime_base_rate * 10000,
2993 pos += scnprintf(buf + pos, size - pos, " cost.vrate=%u.%02u",
2994 vp10k / 100, vp10k % 100);
2997 pos += scnprintf(buf + pos, size - pos, " cost.usage=%llu",
2998 iocg->last_stat.usage_us);
3000 if (blkcg_debug_stats)
3001 pos += scnprintf(buf + pos, size - pos,
3002 " cost.wait=%llu cost.indebt=%llu cost.indelay=%llu",
3003 iocg->last_stat.wait_us,
3004 iocg->last_stat.indebt_us,
3005 iocg->last_stat.indelay_us);
3010 static u64 ioc_weight_prfill(struct seq_file *sf, struct blkg_policy_data *pd,
3013 const char *dname = blkg_dev_name(pd->blkg);
3014 struct ioc_gq *iocg = pd_to_iocg(pd);
3016 if (dname && iocg->cfg_weight)
3017 seq_printf(sf, "%s %u\n", dname, iocg->cfg_weight / WEIGHT_ONE);
3022 static int ioc_weight_show(struct seq_file *sf, void *v)
3024 struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
3025 struct ioc_cgrp *iocc = blkcg_to_iocc(blkcg);
3027 seq_printf(sf, "default %u\n", iocc->dfl_weight / WEIGHT_ONE);
3028 blkcg_print_blkgs(sf, blkcg, ioc_weight_prfill,
3029 &blkcg_policy_iocost, seq_cft(sf)->private, false);
3033 static ssize_t ioc_weight_write(struct kernfs_open_file *of, char *buf,
3034 size_t nbytes, loff_t off)
3036 struct blkcg *blkcg = css_to_blkcg(of_css(of));
3037 struct ioc_cgrp *iocc = blkcg_to_iocc(blkcg);
3038 struct blkg_conf_ctx ctx;
3040 struct ioc_gq *iocg;
3044 if (!strchr(buf, ':')) {
3045 struct blkcg_gq *blkg;
3047 if (!sscanf(buf, "default %u", &v) && !sscanf(buf, "%u", &v))
3050 if (v < CGROUP_WEIGHT_MIN || v > CGROUP_WEIGHT_MAX)
3053 spin_lock(&blkcg->lock);
3054 iocc->dfl_weight = v * WEIGHT_ONE;
3055 hlist_for_each_entry(blkg, &blkcg->blkg_list, blkcg_node) {
3056 struct ioc_gq *iocg = blkg_to_iocg(blkg);
3059 spin_lock_irq(&iocg->ioc->lock);
3060 ioc_now(iocg->ioc, &now);
3061 weight_updated(iocg, &now);
3062 spin_unlock_irq(&iocg->ioc->lock);
3065 spin_unlock(&blkcg->lock);
3070 ret = blkg_conf_prep(blkcg, &blkcg_policy_iocost, buf, &ctx);
3074 iocg = blkg_to_iocg(ctx.blkg);
3076 if (!strncmp(ctx.body, "default", 7)) {
3079 if (!sscanf(ctx.body, "%u", &v))
3081 if (v < CGROUP_WEIGHT_MIN || v > CGROUP_WEIGHT_MAX)
3085 spin_lock(&iocg->ioc->lock);
3086 iocg->cfg_weight = v * WEIGHT_ONE;
3087 ioc_now(iocg->ioc, &now);
3088 weight_updated(iocg, &now);
3089 spin_unlock(&iocg->ioc->lock);
3091 blkg_conf_finish(&ctx);
3095 blkg_conf_finish(&ctx);
3099 static u64 ioc_qos_prfill(struct seq_file *sf, struct blkg_policy_data *pd,
3102 const char *dname = blkg_dev_name(pd->blkg);
3103 struct ioc *ioc = pd_to_iocg(pd)->ioc;
3108 seq_printf(sf, "%s enable=%d ctrl=%s rpct=%u.%02u rlat=%u wpct=%u.%02u wlat=%u min=%u.%02u max=%u.%02u\n",
3109 dname, ioc->enabled, ioc->user_qos_params ? "user" : "auto",
3110 ioc->params.qos[QOS_RPPM] / 10000,
3111 ioc->params.qos[QOS_RPPM] % 10000 / 100,
3112 ioc->params.qos[QOS_RLAT],
3113 ioc->params.qos[QOS_WPPM] / 10000,
3114 ioc->params.qos[QOS_WPPM] % 10000 / 100,
3115 ioc->params.qos[QOS_WLAT],
3116 ioc->params.qos[QOS_MIN] / 10000,
3117 ioc->params.qos[QOS_MIN] % 10000 / 100,
3118 ioc->params.qos[QOS_MAX] / 10000,
3119 ioc->params.qos[QOS_MAX] % 10000 / 100);
3123 static int ioc_qos_show(struct seq_file *sf, void *v)
3125 struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
3127 blkcg_print_blkgs(sf, blkcg, ioc_qos_prfill,
3128 &blkcg_policy_iocost, seq_cft(sf)->private, false);
3132 static const match_table_t qos_ctrl_tokens = {
3133 { QOS_ENABLE, "enable=%u" },
3134 { QOS_CTRL, "ctrl=%s" },
3135 { NR_QOS_CTRL_PARAMS, NULL },
3138 static const match_table_t qos_tokens = {
3139 { QOS_RPPM, "rpct=%s" },
3140 { QOS_RLAT, "rlat=%u" },
3141 { QOS_WPPM, "wpct=%s" },
3142 { QOS_WLAT, "wlat=%u" },
3143 { QOS_MIN, "min=%s" },
3144 { QOS_MAX, "max=%s" },
3145 { NR_QOS_PARAMS, NULL },
3148 static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
3149 size_t nbytes, loff_t off)
3151 struct block_device *bdev;
3153 u32 qos[NR_QOS_PARAMS];
3158 bdev = blkcg_conf_open_bdev(&input);
3160 return PTR_ERR(bdev);
3162 ioc = q_to_ioc(bdev->bd_disk->queue);
3164 ret = blk_iocost_init(bdev->bd_disk->queue);
3167 ioc = q_to_ioc(bdev->bd_disk->queue);
3170 spin_lock_irq(&ioc->lock);
3171 memcpy(qos, ioc->params.qos, sizeof(qos));
3172 enable = ioc->enabled;
3173 user = ioc->user_qos_params;
3174 spin_unlock_irq(&ioc->lock);
3176 while ((p = strsep(&input, " \t\n"))) {
3177 substring_t args[MAX_OPT_ARGS];
3185 switch (match_token(p, qos_ctrl_tokens, args)) {
3187 match_u64(&args[0], &v);
3191 match_strlcpy(buf, &args[0], sizeof(buf));
3192 if (!strcmp(buf, "auto"))
3194 else if (!strcmp(buf, "user"))
3201 tok = match_token(p, qos_tokens, args);
3205 if (match_strlcpy(buf, &args[0], sizeof(buf)) >=
3208 if (cgroup_parse_float(buf, 2, &v))
3210 if (v < 0 || v > 10000)
3216 if (match_u64(&args[0], &v))
3222 if (match_strlcpy(buf, &args[0], sizeof(buf)) >=
3225 if (cgroup_parse_float(buf, 2, &v))
3229 qos[tok] = clamp_t(s64, v * 100,
3230 VRATE_MIN_PPM, VRATE_MAX_PPM);
3238 if (qos[QOS_MIN] > qos[QOS_MAX])
3241 spin_lock_irq(&ioc->lock);
3244 blk_stat_enable_accounting(ioc->rqos.q);
3245 blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
3246 ioc->enabled = true;
3248 blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
3249 ioc->enabled = false;
3253 memcpy(ioc->params.qos, qos, sizeof(qos));
3254 ioc->user_qos_params = true;
3256 ioc->user_qos_params = false;
3259 ioc_refresh_params(ioc, true);
3260 spin_unlock_irq(&ioc->lock);
3262 blkdev_put_no_open(bdev);
3267 blkdev_put_no_open(bdev);
3271 static u64 ioc_cost_model_prfill(struct seq_file *sf,
3272 struct blkg_policy_data *pd, int off)
3274 const char *dname = blkg_dev_name(pd->blkg);
3275 struct ioc *ioc = pd_to_iocg(pd)->ioc;
3276 u64 *u = ioc->params.i_lcoefs;
3281 seq_printf(sf, "%s ctrl=%s model=linear "
3282 "rbps=%llu rseqiops=%llu rrandiops=%llu "
3283 "wbps=%llu wseqiops=%llu wrandiops=%llu\n",
3284 dname, ioc->user_cost_model ? "user" : "auto",
3285 u[I_LCOEF_RBPS], u[I_LCOEF_RSEQIOPS], u[I_LCOEF_RRANDIOPS],
3286 u[I_LCOEF_WBPS], u[I_LCOEF_WSEQIOPS], u[I_LCOEF_WRANDIOPS]);
3290 static int ioc_cost_model_show(struct seq_file *sf, void *v)
3292 struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
3294 blkcg_print_blkgs(sf, blkcg, ioc_cost_model_prfill,
3295 &blkcg_policy_iocost, seq_cft(sf)->private, false);
3299 static const match_table_t cost_ctrl_tokens = {
3300 { COST_CTRL, "ctrl=%s" },
3301 { COST_MODEL, "model=%s" },
3302 { NR_COST_CTRL_PARAMS, NULL },
3305 static const match_table_t i_lcoef_tokens = {
3306 { I_LCOEF_RBPS, "rbps=%u" },
3307 { I_LCOEF_RSEQIOPS, "rseqiops=%u" },
3308 { I_LCOEF_RRANDIOPS, "rrandiops=%u" },
3309 { I_LCOEF_WBPS, "wbps=%u" },
3310 { I_LCOEF_WSEQIOPS, "wseqiops=%u" },
3311 { I_LCOEF_WRANDIOPS, "wrandiops=%u" },
3312 { NR_I_LCOEFS, NULL },
3315 static ssize_t ioc_cost_model_write(struct kernfs_open_file *of, char *input,
3316 size_t nbytes, loff_t off)
3318 struct block_device *bdev;
3325 bdev = blkcg_conf_open_bdev(&input);
3327 return PTR_ERR(bdev);
3329 ioc = q_to_ioc(bdev->bd_disk->queue);
3331 ret = blk_iocost_init(bdev->bd_disk->queue);
3334 ioc = q_to_ioc(bdev->bd_disk->queue);
3337 spin_lock_irq(&ioc->lock);
3338 memcpy(u, ioc->params.i_lcoefs, sizeof(u));
3339 user = ioc->user_cost_model;
3340 spin_unlock_irq(&ioc->lock);
3342 while ((p = strsep(&input, " \t\n"))) {
3343 substring_t args[MAX_OPT_ARGS];
3351 switch (match_token(p, cost_ctrl_tokens, args)) {
3353 match_strlcpy(buf, &args[0], sizeof(buf));
3354 if (!strcmp(buf, "auto"))
3356 else if (!strcmp(buf, "user"))
3362 match_strlcpy(buf, &args[0], sizeof(buf));
3363 if (strcmp(buf, "linear"))
3368 tok = match_token(p, i_lcoef_tokens, args);
3369 if (tok == NR_I_LCOEFS)
3371 if (match_u64(&args[0], &v))
3377 spin_lock_irq(&ioc->lock);
3379 memcpy(ioc->params.i_lcoefs, u, sizeof(u));
3380 ioc->user_cost_model = true;
3382 ioc->user_cost_model = false;
3384 ioc_refresh_params(ioc, true);
3385 spin_unlock_irq(&ioc->lock);
3387 blkdev_put_no_open(bdev);
3393 blkdev_put_no_open(bdev);
3397 static struct cftype ioc_files[] = {
3400 .flags = CFTYPE_NOT_ON_ROOT,
3401 .seq_show = ioc_weight_show,
3402 .write = ioc_weight_write,
3406 .flags = CFTYPE_ONLY_ON_ROOT,
3407 .seq_show = ioc_qos_show,
3408 .write = ioc_qos_write,
3411 .name = "cost.model",
3412 .flags = CFTYPE_ONLY_ON_ROOT,
3413 .seq_show = ioc_cost_model_show,
3414 .write = ioc_cost_model_write,
3419 static struct blkcg_policy blkcg_policy_iocost = {
3420 .dfl_cftypes = ioc_files,
3421 .cpd_alloc_fn = ioc_cpd_alloc,
3422 .cpd_free_fn = ioc_cpd_free,
3423 .pd_alloc_fn = ioc_pd_alloc,
3424 .pd_init_fn = ioc_pd_init,
3425 .pd_free_fn = ioc_pd_free,
3426 .pd_stat_fn = ioc_pd_stat,
3429 static int __init ioc_init(void)
3431 return blkcg_policy_register(&blkcg_policy_iocost);
3434 static void __exit ioc_exit(void)
3436 blkcg_policy_unregister(&blkcg_policy_iocost);
3439 module_init(ioc_init);
3440 module_exit(ioc_exit);