]>
Commit | Line | Data |
---|---|---|
e43473b7 VG |
1 | /* |
2 | * Interface for controlling IO bandwidth on a request queue | |
3 | * | |
4 | * Copyright (C) 2010 Vivek Goyal <vgoyal@redhat.com> | |
5 | */ | |
6 | ||
7 | #include <linux/module.h> | |
8 | #include <linux/slab.h> | |
9 | #include <linux/blkdev.h> | |
10 | #include <linux/bio.h> | |
11 | #include <linux/blktrace_api.h> | |
12 | #include "blk-cgroup.h" | |
bc9fcbf9 | 13 | #include "blk.h" |
e43473b7 VG |
14 | |
15 | /* Max dispatch from a group in 1 round */ | |
16 | static int throtl_grp_quantum = 8; | |
17 | ||
18 | /* Total max dispatch from all groups in one round */ | |
19 | static int throtl_quantum = 32; | |
20 | ||
21 | /* Throttling is performed over 100ms slice and after that slice is renewed */ | |
22 | static unsigned long throtl_slice = HZ/10; /* 100 ms */ | |
23 | ||
3c798398 | 24 | static struct blkcg_policy blkcg_policy_throtl; |
0381411e | 25 | |
450adcbe VG |
26 | /* A workqueue to queue throttle related work */ |
27 | static struct workqueue_struct *kthrotld_workqueue; | |
450adcbe | 28 | |
c9e0332e TH |
29 | struct throtl_service_queue { |
30 | struct rb_root pending_tree; /* RB tree of active tgs */ | |
31 | struct rb_node *first_pending; /* first node in the tree */ | |
32 | unsigned int nr_pending; /* # queued in the tree */ | |
33 | unsigned long first_pending_disptime; /* disptime of the first tg */ | |
e43473b7 VG |
34 | }; |
35 | ||
c9e0332e TH |
36 | #define THROTL_SERVICE_QUEUE_INITIALIZER \ |
37 | (struct throtl_service_queue){ .pending_tree = RB_ROOT } | |
e43473b7 VG |
38 | |
39 | #define rb_entry_tg(node) rb_entry((node), struct throtl_grp, rb_node) | |
40 | ||
8a3d2615 TH |
41 | /* Per-cpu group stats */ |
42 | struct tg_stats_cpu { | |
43 | /* total bytes transferred */ | |
44 | struct blkg_rwstat service_bytes; | |
45 | /* total IOs serviced, post merge */ | |
46 | struct blkg_rwstat serviced; | |
47 | }; | |
48 | ||
e43473b7 | 49 | struct throtl_grp { |
f95a04af TH |
50 | /* must be the first member */ |
51 | struct blkg_policy_data pd; | |
52 | ||
c9e0332e | 53 | /* active throtl group service_queue member */ |
e43473b7 VG |
54 | struct rb_node rb_node; |
55 | ||
56 | /* | |
57 | * Dispatch time in jiffies. This is the estimated time when group | |
58 | * will unthrottle and is ready to dispatch more bio. It is used as | |
59 | * key to sort active groups in service tree. | |
60 | */ | |
61 | unsigned long disptime; | |
62 | ||
e43473b7 VG |
63 | unsigned int flags; |
64 | ||
65 | /* Two lists for READ and WRITE */ | |
66 | struct bio_list bio_lists[2]; | |
67 | ||
68 | /* Number of queued bios on READ and WRITE lists */ | |
69 | unsigned int nr_queued[2]; | |
70 | ||
71 | /* bytes per second rate limits */ | |
72 | uint64_t bps[2]; | |
73 | ||
8e89d13f VG |
74 | /* IOPS limits */ |
75 | unsigned int iops[2]; | |
76 | ||
e43473b7 VG |
77 | /* Number of bytes disptached in current slice */ |
78 | uint64_t bytes_disp[2]; | |
8e89d13f VG |
79 | /* Number of bio's dispatched in current slice */ |
80 | unsigned int io_disp[2]; | |
e43473b7 VG |
81 | |
82 | /* When did we start a new slice */ | |
83 | unsigned long slice_start[2]; | |
84 | unsigned long slice_end[2]; | |
fe071437 | 85 | |
8a3d2615 TH |
86 | /* Per cpu stats pointer */ |
87 | struct tg_stats_cpu __percpu *stats_cpu; | |
88 | ||
89 | /* List of tgs waiting for per cpu stats memory to be allocated */ | |
90 | struct list_head stats_alloc_node; | |
e43473b7 VG |
91 | }; |
92 | ||
93 | struct throtl_data | |
94 | { | |
e43473b7 | 95 | /* service tree for active throtl groups */ |
c9e0332e | 96 | struct throtl_service_queue service_queue; |
e43473b7 | 97 | |
e43473b7 VG |
98 | struct request_queue *queue; |
99 | ||
100 | /* Total Number of queued bios on READ and WRITE lists */ | |
101 | unsigned int nr_queued[2]; | |
102 | ||
103 | /* | |
02977e4a | 104 | * number of total undestroyed groups |
e43473b7 VG |
105 | */ |
106 | unsigned int nr_undestroyed_grps; | |
107 | ||
108 | /* Work for dispatching throttled bios */ | |
cb76199c | 109 | struct delayed_work dispatch_work; |
e43473b7 VG |
110 | }; |
111 | ||
8a3d2615 TH |
112 | /* list and work item to allocate percpu group stats */ |
113 | static DEFINE_SPINLOCK(tg_stats_alloc_lock); | |
114 | static LIST_HEAD(tg_stats_alloc_list); | |
115 | ||
116 | static void tg_stats_alloc_fn(struct work_struct *); | |
117 | static DECLARE_DELAYED_WORK(tg_stats_alloc_work, tg_stats_alloc_fn); | |
118 | ||
f95a04af TH |
119 | static inline struct throtl_grp *pd_to_tg(struct blkg_policy_data *pd) |
120 | { | |
121 | return pd ? container_of(pd, struct throtl_grp, pd) : NULL; | |
122 | } | |
123 | ||
3c798398 | 124 | static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg) |
0381411e | 125 | { |
f95a04af | 126 | return pd_to_tg(blkg_to_pd(blkg, &blkcg_policy_throtl)); |
0381411e TH |
127 | } |
128 | ||
3c798398 | 129 | static inline struct blkcg_gq *tg_to_blkg(struct throtl_grp *tg) |
0381411e | 130 | { |
f95a04af | 131 | return pd_to_blkg(&tg->pd); |
0381411e TH |
132 | } |
133 | ||
03d8e111 TH |
134 | static inline struct throtl_grp *td_root_tg(struct throtl_data *td) |
135 | { | |
136 | return blkg_to_tg(td->queue->root_blkg); | |
137 | } | |
138 | ||
e43473b7 VG |
139 | enum tg_state_flags { |
140 | THROTL_TG_FLAG_on_rr = 0, /* on round-robin busy list */ | |
141 | }; | |
142 | ||
143 | #define THROTL_TG_FNS(name) \ | |
144 | static inline void throtl_mark_tg_##name(struct throtl_grp *tg) \ | |
145 | { \ | |
146 | (tg)->flags |= (1 << THROTL_TG_FLAG_##name); \ | |
147 | } \ | |
148 | static inline void throtl_clear_tg_##name(struct throtl_grp *tg) \ | |
149 | { \ | |
150 | (tg)->flags &= ~(1 << THROTL_TG_FLAG_##name); \ | |
151 | } \ | |
152 | static inline int throtl_tg_##name(const struct throtl_grp *tg) \ | |
153 | { \ | |
154 | return ((tg)->flags & (1 << THROTL_TG_FLAG_##name)) != 0; \ | |
155 | } | |
156 | ||
157 | THROTL_TG_FNS(on_rr); | |
158 | ||
54e7ed12 TH |
159 | #define throtl_log_tg(td, tg, fmt, args...) do { \ |
160 | char __pbuf[128]; \ | |
161 | \ | |
162 | blkg_path(tg_to_blkg(tg), __pbuf, sizeof(__pbuf)); \ | |
163 | blk_add_trace_msg((td)->queue, "throtl %s " fmt, __pbuf, ##args); \ | |
164 | } while (0) | |
e43473b7 VG |
165 | |
166 | #define throtl_log(td, fmt, args...) \ | |
167 | blk_add_trace_msg((td)->queue, "throtl " fmt, ##args) | |
168 | ||
8a3d2615 TH |
169 | /* |
170 | * Worker for allocating per cpu stat for tgs. This is scheduled on the | |
3b07e9ca | 171 | * system_wq once there are some groups on the alloc_list waiting for |
8a3d2615 TH |
172 | * allocation. |
173 | */ | |
174 | static void tg_stats_alloc_fn(struct work_struct *work) | |
175 | { | |
176 | static struct tg_stats_cpu *stats_cpu; /* this fn is non-reentrant */ | |
177 | struct delayed_work *dwork = to_delayed_work(work); | |
178 | bool empty = false; | |
179 | ||
180 | alloc_stats: | |
181 | if (!stats_cpu) { | |
182 | stats_cpu = alloc_percpu(struct tg_stats_cpu); | |
183 | if (!stats_cpu) { | |
184 | /* allocation failed, try again after some time */ | |
3b07e9ca | 185 | schedule_delayed_work(dwork, msecs_to_jiffies(10)); |
8a3d2615 TH |
186 | return; |
187 | } | |
188 | } | |
189 | ||
190 | spin_lock_irq(&tg_stats_alloc_lock); | |
191 | ||
192 | if (!list_empty(&tg_stats_alloc_list)) { | |
193 | struct throtl_grp *tg = list_first_entry(&tg_stats_alloc_list, | |
194 | struct throtl_grp, | |
195 | stats_alloc_node); | |
196 | swap(tg->stats_cpu, stats_cpu); | |
197 | list_del_init(&tg->stats_alloc_node); | |
198 | } | |
199 | ||
200 | empty = list_empty(&tg_stats_alloc_list); | |
201 | spin_unlock_irq(&tg_stats_alloc_lock); | |
202 | if (!empty) | |
203 | goto alloc_stats; | |
204 | } | |
205 | ||
3c798398 | 206 | static void throtl_pd_init(struct blkcg_gq *blkg) |
a29a171e | 207 | { |
0381411e | 208 | struct throtl_grp *tg = blkg_to_tg(blkg); |
ff26eaad | 209 | unsigned long flags; |
cd1604fa | 210 | |
a29a171e VG |
211 | RB_CLEAR_NODE(&tg->rb_node); |
212 | bio_list_init(&tg->bio_lists[0]); | |
213 | bio_list_init(&tg->bio_lists[1]); | |
a29a171e | 214 | |
e56da7e2 TH |
215 | tg->bps[READ] = -1; |
216 | tg->bps[WRITE] = -1; | |
217 | tg->iops[READ] = -1; | |
218 | tg->iops[WRITE] = -1; | |
8a3d2615 TH |
219 | |
220 | /* | |
221 | * Ugh... We need to perform per-cpu allocation for tg->stats_cpu | |
222 | * but percpu allocator can't be called from IO path. Queue tg on | |
223 | * tg_stats_alloc_list and allocate from work item. | |
224 | */ | |
ff26eaad | 225 | spin_lock_irqsave(&tg_stats_alloc_lock, flags); |
8a3d2615 | 226 | list_add(&tg->stats_alloc_node, &tg_stats_alloc_list); |
3b07e9ca | 227 | schedule_delayed_work(&tg_stats_alloc_work, 0); |
ff26eaad | 228 | spin_unlock_irqrestore(&tg_stats_alloc_lock, flags); |
8a3d2615 TH |
229 | } |
230 | ||
3c798398 | 231 | static void throtl_pd_exit(struct blkcg_gq *blkg) |
8a3d2615 TH |
232 | { |
233 | struct throtl_grp *tg = blkg_to_tg(blkg); | |
ff26eaad | 234 | unsigned long flags; |
8a3d2615 | 235 | |
ff26eaad | 236 | spin_lock_irqsave(&tg_stats_alloc_lock, flags); |
8a3d2615 | 237 | list_del_init(&tg->stats_alloc_node); |
ff26eaad | 238 | spin_unlock_irqrestore(&tg_stats_alloc_lock, flags); |
8a3d2615 TH |
239 | |
240 | free_percpu(tg->stats_cpu); | |
241 | } | |
242 | ||
3c798398 | 243 | static void throtl_pd_reset_stats(struct blkcg_gq *blkg) |
8a3d2615 TH |
244 | { |
245 | struct throtl_grp *tg = blkg_to_tg(blkg); | |
246 | int cpu; | |
247 | ||
248 | if (tg->stats_cpu == NULL) | |
249 | return; | |
250 | ||
251 | for_each_possible_cpu(cpu) { | |
252 | struct tg_stats_cpu *sc = per_cpu_ptr(tg->stats_cpu, cpu); | |
253 | ||
254 | blkg_rwstat_reset(&sc->service_bytes); | |
255 | blkg_rwstat_reset(&sc->serviced); | |
256 | } | |
a29a171e VG |
257 | } |
258 | ||
3c798398 TH |
259 | static struct throtl_grp *throtl_lookup_tg(struct throtl_data *td, |
260 | struct blkcg *blkcg) | |
e43473b7 | 261 | { |
be2c6b19 | 262 | /* |
3c798398 TH |
263 | * This is the common case when there are no blkcgs. Avoid lookup |
264 | * in this case | |
cd1604fa | 265 | */ |
3c798398 | 266 | if (blkcg == &blkcg_root) |
03d8e111 | 267 | return td_root_tg(td); |
e43473b7 | 268 | |
e8989fae | 269 | return blkg_to_tg(blkg_lookup(blkcg, td->queue)); |
e43473b7 VG |
270 | } |
271 | ||
cd1604fa | 272 | static struct throtl_grp *throtl_lookup_create_tg(struct throtl_data *td, |
3c798398 | 273 | struct blkcg *blkcg) |
e43473b7 | 274 | { |
f469a7b4 | 275 | struct request_queue *q = td->queue; |
cd1604fa | 276 | struct throtl_grp *tg = NULL; |
bc16a4f9 | 277 | |
f469a7b4 | 278 | /* |
3c798398 TH |
279 | * This is the common case when there are no blkcgs. Avoid lookup |
280 | * in this case | |
f469a7b4 | 281 | */ |
3c798398 | 282 | if (blkcg == &blkcg_root) { |
03d8e111 | 283 | tg = td_root_tg(td); |
cd1604fa | 284 | } else { |
3c798398 | 285 | struct blkcg_gq *blkg; |
f469a7b4 | 286 | |
3c96cb32 | 287 | blkg = blkg_lookup_create(blkcg, q); |
f469a7b4 | 288 | |
cd1604fa TH |
289 | /* if %NULL and @q is alive, fall back to root_tg */ |
290 | if (!IS_ERR(blkg)) | |
0381411e | 291 | tg = blkg_to_tg(blkg); |
3f3299d5 | 292 | else if (!blk_queue_dying(q)) |
03d8e111 | 293 | tg = td_root_tg(td); |
f469a7b4 VG |
294 | } |
295 | ||
e43473b7 VG |
296 | return tg; |
297 | } | |
298 | ||
c9e0332e | 299 | static struct throtl_grp *throtl_rb_first(struct throtl_service_queue *sq) |
e43473b7 VG |
300 | { |
301 | /* Service tree is empty */ | |
c9e0332e | 302 | if (!sq->nr_pending) |
e43473b7 VG |
303 | return NULL; |
304 | ||
c9e0332e TH |
305 | if (!sq->first_pending) |
306 | sq->first_pending = rb_first(&sq->pending_tree); | |
e43473b7 | 307 | |
c9e0332e TH |
308 | if (sq->first_pending) |
309 | return rb_entry_tg(sq->first_pending); | |
e43473b7 VG |
310 | |
311 | return NULL; | |
312 | } | |
313 | ||
314 | static void rb_erase_init(struct rb_node *n, struct rb_root *root) | |
315 | { | |
316 | rb_erase(n, root); | |
317 | RB_CLEAR_NODE(n); | |
318 | } | |
319 | ||
c9e0332e | 320 | static void throtl_rb_erase(struct rb_node *n, struct throtl_service_queue *sq) |
e43473b7 | 321 | { |
c9e0332e TH |
322 | if (sq->first_pending == n) |
323 | sq->first_pending = NULL; | |
324 | rb_erase_init(n, &sq->pending_tree); | |
325 | --sq->nr_pending; | |
e43473b7 VG |
326 | } |
327 | ||
c9e0332e | 328 | static void update_min_dispatch_time(struct throtl_service_queue *sq) |
e43473b7 VG |
329 | { |
330 | struct throtl_grp *tg; | |
331 | ||
c9e0332e | 332 | tg = throtl_rb_first(sq); |
e43473b7 VG |
333 | if (!tg) |
334 | return; | |
335 | ||
c9e0332e | 336 | sq->first_pending_disptime = tg->disptime; |
e43473b7 VG |
337 | } |
338 | ||
c9e0332e TH |
339 | static void tg_service_queue_add(struct throtl_service_queue *sq, |
340 | struct throtl_grp *tg) | |
e43473b7 | 341 | { |
c9e0332e | 342 | struct rb_node **node = &sq->pending_tree.rb_node; |
e43473b7 VG |
343 | struct rb_node *parent = NULL; |
344 | struct throtl_grp *__tg; | |
345 | unsigned long key = tg->disptime; | |
346 | int left = 1; | |
347 | ||
348 | while (*node != NULL) { | |
349 | parent = *node; | |
350 | __tg = rb_entry_tg(parent); | |
351 | ||
352 | if (time_before(key, __tg->disptime)) | |
353 | node = &parent->rb_left; | |
354 | else { | |
355 | node = &parent->rb_right; | |
356 | left = 0; | |
357 | } | |
358 | } | |
359 | ||
360 | if (left) | |
c9e0332e | 361 | sq->first_pending = &tg->rb_node; |
e43473b7 VG |
362 | |
363 | rb_link_node(&tg->rb_node, parent, node); | |
c9e0332e | 364 | rb_insert_color(&tg->rb_node, &sq->pending_tree); |
e43473b7 VG |
365 | } |
366 | ||
367 | static void __throtl_enqueue_tg(struct throtl_data *td, struct throtl_grp *tg) | |
368 | { | |
c9e0332e | 369 | struct throtl_service_queue *sq = &td->service_queue; |
e43473b7 | 370 | |
c9e0332e | 371 | tg_service_queue_add(sq, tg); |
e43473b7 | 372 | throtl_mark_tg_on_rr(tg); |
c9e0332e | 373 | sq->nr_pending++; |
e43473b7 VG |
374 | } |
375 | ||
376 | static void throtl_enqueue_tg(struct throtl_data *td, struct throtl_grp *tg) | |
377 | { | |
378 | if (!throtl_tg_on_rr(tg)) | |
379 | __throtl_enqueue_tg(td, tg); | |
380 | } | |
381 | ||
382 | static void __throtl_dequeue_tg(struct throtl_data *td, struct throtl_grp *tg) | |
383 | { | |
c9e0332e | 384 | throtl_rb_erase(&tg->rb_node, &td->service_queue); |
e43473b7 VG |
385 | throtl_clear_tg_on_rr(tg); |
386 | } | |
387 | ||
388 | static void throtl_dequeue_tg(struct throtl_data *td, struct throtl_grp *tg) | |
389 | { | |
390 | if (throtl_tg_on_rr(tg)) | |
391 | __throtl_dequeue_tg(td, tg); | |
392 | } | |
393 | ||
a9131a27 TH |
394 | /* Call with queue lock held */ |
395 | static void throtl_schedule_delayed_work(struct throtl_data *td, | |
396 | unsigned long delay) | |
397 | { | |
398 | struct delayed_work *dwork = &td->dispatch_work; | |
399 | ||
6a525600 TH |
400 | mod_delayed_work(kthrotld_workqueue, dwork, delay); |
401 | throtl_log(td, "schedule work. delay=%lu jiffies=%lu", delay, jiffies); | |
a9131a27 TH |
402 | } |
403 | ||
e43473b7 VG |
404 | static void throtl_schedule_next_dispatch(struct throtl_data *td) |
405 | { | |
c9e0332e | 406 | struct throtl_service_queue *sq = &td->service_queue; |
e43473b7 | 407 | |
6a525600 | 408 | /* any pending children left? */ |
c9e0332e | 409 | if (!sq->nr_pending) |
e43473b7 VG |
410 | return; |
411 | ||
c9e0332e | 412 | update_min_dispatch_time(sq); |
e43473b7 | 413 | |
c9e0332e | 414 | if (time_before_eq(sq->first_pending_disptime, jiffies)) |
450adcbe | 415 | throtl_schedule_delayed_work(td, 0); |
e43473b7 | 416 | else |
c9e0332e | 417 | throtl_schedule_delayed_work(td, sq->first_pending_disptime - jiffies); |
e43473b7 VG |
418 | } |
419 | ||
420 | static inline void | |
421 | throtl_start_new_slice(struct throtl_data *td, struct throtl_grp *tg, bool rw) | |
422 | { | |
423 | tg->bytes_disp[rw] = 0; | |
8e89d13f | 424 | tg->io_disp[rw] = 0; |
e43473b7 VG |
425 | tg->slice_start[rw] = jiffies; |
426 | tg->slice_end[rw] = jiffies + throtl_slice; | |
427 | throtl_log_tg(td, tg, "[%c] new slice start=%lu end=%lu jiffies=%lu", | |
428 | rw == READ ? 'R' : 'W', tg->slice_start[rw], | |
429 | tg->slice_end[rw], jiffies); | |
430 | } | |
431 | ||
d1ae8ffd VG |
432 | static inline void throtl_set_slice_end(struct throtl_data *td, |
433 | struct throtl_grp *tg, bool rw, unsigned long jiffy_end) | |
434 | { | |
435 | tg->slice_end[rw] = roundup(jiffy_end, throtl_slice); | |
436 | } | |
437 | ||
e43473b7 VG |
438 | static inline void throtl_extend_slice(struct throtl_data *td, |
439 | struct throtl_grp *tg, bool rw, unsigned long jiffy_end) | |
440 | { | |
441 | tg->slice_end[rw] = roundup(jiffy_end, throtl_slice); | |
442 | throtl_log_tg(td, tg, "[%c] extend slice start=%lu end=%lu jiffies=%lu", | |
443 | rw == READ ? 'R' : 'W', tg->slice_start[rw], | |
444 | tg->slice_end[rw], jiffies); | |
445 | } | |
446 | ||
447 | /* Determine if previously allocated or extended slice is complete or not */ | |
448 | static bool | |
449 | throtl_slice_used(struct throtl_data *td, struct throtl_grp *tg, bool rw) | |
450 | { | |
451 | if (time_in_range(jiffies, tg->slice_start[rw], tg->slice_end[rw])) | |
452 | return 0; | |
453 | ||
454 | return 1; | |
455 | } | |
456 | ||
457 | /* Trim the used slices and adjust slice start accordingly */ | |
458 | static inline void | |
459 | throtl_trim_slice(struct throtl_data *td, struct throtl_grp *tg, bool rw) | |
460 | { | |
3aad5d3e VG |
461 | unsigned long nr_slices, time_elapsed, io_trim; |
462 | u64 bytes_trim, tmp; | |
e43473b7 VG |
463 | |
464 | BUG_ON(time_before(tg->slice_end[rw], tg->slice_start[rw])); | |
465 | ||
466 | /* | |
467 | * If bps are unlimited (-1), then time slice don't get | |
468 | * renewed. Don't try to trim the slice if slice is used. A new | |
469 | * slice will start when appropriate. | |
470 | */ | |
471 | if (throtl_slice_used(td, tg, rw)) | |
472 | return; | |
473 | ||
d1ae8ffd VG |
474 | /* |
475 | * A bio has been dispatched. Also adjust slice_end. It might happen | |
476 | * that initially cgroup limit was very low resulting in high | |
477 | * slice_end, but later limit was bumped up and bio was dispached | |
478 | * sooner, then we need to reduce slice_end. A high bogus slice_end | |
479 | * is bad because it does not allow new slice to start. | |
480 | */ | |
481 | ||
482 | throtl_set_slice_end(td, tg, rw, jiffies + throtl_slice); | |
483 | ||
e43473b7 VG |
484 | time_elapsed = jiffies - tg->slice_start[rw]; |
485 | ||
486 | nr_slices = time_elapsed / throtl_slice; | |
487 | ||
488 | if (!nr_slices) | |
489 | return; | |
3aad5d3e VG |
490 | tmp = tg->bps[rw] * throtl_slice * nr_slices; |
491 | do_div(tmp, HZ); | |
492 | bytes_trim = tmp; | |
e43473b7 | 493 | |
8e89d13f | 494 | io_trim = (tg->iops[rw] * throtl_slice * nr_slices)/HZ; |
e43473b7 | 495 | |
8e89d13f | 496 | if (!bytes_trim && !io_trim) |
e43473b7 VG |
497 | return; |
498 | ||
499 | if (tg->bytes_disp[rw] >= bytes_trim) | |
500 | tg->bytes_disp[rw] -= bytes_trim; | |
501 | else | |
502 | tg->bytes_disp[rw] = 0; | |
503 | ||
8e89d13f VG |
504 | if (tg->io_disp[rw] >= io_trim) |
505 | tg->io_disp[rw] -= io_trim; | |
506 | else | |
507 | tg->io_disp[rw] = 0; | |
508 | ||
e43473b7 VG |
509 | tg->slice_start[rw] += nr_slices * throtl_slice; |
510 | ||
3aad5d3e | 511 | throtl_log_tg(td, tg, "[%c] trim slice nr=%lu bytes=%llu io=%lu" |
e43473b7 | 512 | " start=%lu end=%lu jiffies=%lu", |
8e89d13f | 513 | rw == READ ? 'R' : 'W', nr_slices, bytes_trim, io_trim, |
e43473b7 VG |
514 | tg->slice_start[rw], tg->slice_end[rw], jiffies); |
515 | } | |
516 | ||
8e89d13f VG |
517 | static bool tg_with_in_iops_limit(struct throtl_data *td, struct throtl_grp *tg, |
518 | struct bio *bio, unsigned long *wait) | |
e43473b7 VG |
519 | { |
520 | bool rw = bio_data_dir(bio); | |
8e89d13f | 521 | unsigned int io_allowed; |
e43473b7 | 522 | unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd; |
c49c06e4 | 523 | u64 tmp; |
e43473b7 | 524 | |
8e89d13f | 525 | jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw]; |
e43473b7 | 526 | |
8e89d13f VG |
527 | /* Slice has just started. Consider one slice interval */ |
528 | if (!jiffy_elapsed) | |
529 | jiffy_elapsed_rnd = throtl_slice; | |
530 | ||
531 | jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice); | |
532 | ||
c49c06e4 VG |
533 | /* |
534 | * jiffy_elapsed_rnd should not be a big value as minimum iops can be | |
535 | * 1 then at max jiffy elapsed should be equivalent of 1 second as we | |
536 | * will allow dispatch after 1 second and after that slice should | |
537 | * have been trimmed. | |
538 | */ | |
539 | ||
540 | tmp = (u64)tg->iops[rw] * jiffy_elapsed_rnd; | |
541 | do_div(tmp, HZ); | |
542 | ||
543 | if (tmp > UINT_MAX) | |
544 | io_allowed = UINT_MAX; | |
545 | else | |
546 | io_allowed = tmp; | |
8e89d13f VG |
547 | |
548 | if (tg->io_disp[rw] + 1 <= io_allowed) { | |
e43473b7 VG |
549 | if (wait) |
550 | *wait = 0; | |
551 | return 1; | |
552 | } | |
553 | ||
8e89d13f VG |
554 | /* Calc approx time to dispatch */ |
555 | jiffy_wait = ((tg->io_disp[rw] + 1) * HZ)/tg->iops[rw] + 1; | |
556 | ||
557 | if (jiffy_wait > jiffy_elapsed) | |
558 | jiffy_wait = jiffy_wait - jiffy_elapsed; | |
559 | else | |
560 | jiffy_wait = 1; | |
561 | ||
562 | if (wait) | |
563 | *wait = jiffy_wait; | |
564 | return 0; | |
565 | } | |
566 | ||
567 | static bool tg_with_in_bps_limit(struct throtl_data *td, struct throtl_grp *tg, | |
568 | struct bio *bio, unsigned long *wait) | |
569 | { | |
570 | bool rw = bio_data_dir(bio); | |
3aad5d3e | 571 | u64 bytes_allowed, extra_bytes, tmp; |
8e89d13f | 572 | unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd; |
e43473b7 VG |
573 | |
574 | jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw]; | |
575 | ||
576 | /* Slice has just started. Consider one slice interval */ | |
577 | if (!jiffy_elapsed) | |
578 | jiffy_elapsed_rnd = throtl_slice; | |
579 | ||
580 | jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice); | |
581 | ||
5e901a2b VG |
582 | tmp = tg->bps[rw] * jiffy_elapsed_rnd; |
583 | do_div(tmp, HZ); | |
3aad5d3e | 584 | bytes_allowed = tmp; |
e43473b7 VG |
585 | |
586 | if (tg->bytes_disp[rw] + bio->bi_size <= bytes_allowed) { | |
587 | if (wait) | |
588 | *wait = 0; | |
589 | return 1; | |
590 | } | |
591 | ||
592 | /* Calc approx time to dispatch */ | |
593 | extra_bytes = tg->bytes_disp[rw] + bio->bi_size - bytes_allowed; | |
594 | jiffy_wait = div64_u64(extra_bytes * HZ, tg->bps[rw]); | |
595 | ||
596 | if (!jiffy_wait) | |
597 | jiffy_wait = 1; | |
598 | ||
599 | /* | |
600 | * This wait time is without taking into consideration the rounding | |
601 | * up we did. Add that time also. | |
602 | */ | |
603 | jiffy_wait = jiffy_wait + (jiffy_elapsed_rnd - jiffy_elapsed); | |
e43473b7 VG |
604 | if (wait) |
605 | *wait = jiffy_wait; | |
8e89d13f VG |
606 | return 0; |
607 | } | |
608 | ||
af75cd3c VG |
609 | static bool tg_no_rule_group(struct throtl_grp *tg, bool rw) { |
610 | if (tg->bps[rw] == -1 && tg->iops[rw] == -1) | |
611 | return 1; | |
612 | return 0; | |
613 | } | |
614 | ||
8e89d13f VG |
615 | /* |
616 | * Returns whether one can dispatch a bio or not. Also returns approx number | |
617 | * of jiffies to wait before this bio is with-in IO rate and can be dispatched | |
618 | */ | |
619 | static bool tg_may_dispatch(struct throtl_data *td, struct throtl_grp *tg, | |
620 | struct bio *bio, unsigned long *wait) | |
621 | { | |
622 | bool rw = bio_data_dir(bio); | |
623 | unsigned long bps_wait = 0, iops_wait = 0, max_wait = 0; | |
624 | ||
625 | /* | |
626 | * Currently whole state machine of group depends on first bio | |
627 | * queued in the group bio list. So one should not be calling | |
628 | * this function with a different bio if there are other bios | |
629 | * queued. | |
630 | */ | |
631 | BUG_ON(tg->nr_queued[rw] && bio != bio_list_peek(&tg->bio_lists[rw])); | |
e43473b7 | 632 | |
8e89d13f VG |
633 | /* If tg->bps = -1, then BW is unlimited */ |
634 | if (tg->bps[rw] == -1 && tg->iops[rw] == -1) { | |
635 | if (wait) | |
636 | *wait = 0; | |
637 | return 1; | |
638 | } | |
639 | ||
640 | /* | |
641 | * If previous slice expired, start a new one otherwise renew/extend | |
642 | * existing slice to make sure it is at least throtl_slice interval | |
643 | * long since now. | |
644 | */ | |
645 | if (throtl_slice_used(td, tg, rw)) | |
646 | throtl_start_new_slice(td, tg, rw); | |
647 | else { | |
648 | if (time_before(tg->slice_end[rw], jiffies + throtl_slice)) | |
649 | throtl_extend_slice(td, tg, rw, jiffies + throtl_slice); | |
650 | } | |
651 | ||
652 | if (tg_with_in_bps_limit(td, tg, bio, &bps_wait) | |
653 | && tg_with_in_iops_limit(td, tg, bio, &iops_wait)) { | |
654 | if (wait) | |
655 | *wait = 0; | |
656 | return 1; | |
657 | } | |
658 | ||
659 | max_wait = max(bps_wait, iops_wait); | |
660 | ||
661 | if (wait) | |
662 | *wait = max_wait; | |
663 | ||
664 | if (time_before(tg->slice_end[rw], jiffies + max_wait)) | |
665 | throtl_extend_slice(td, tg, rw, jiffies + max_wait); | |
e43473b7 VG |
666 | |
667 | return 0; | |
668 | } | |
669 | ||
3c798398 | 670 | static void throtl_update_dispatch_stats(struct blkcg_gq *blkg, u64 bytes, |
629ed0b1 TH |
671 | int rw) |
672 | { | |
8a3d2615 TH |
673 | struct throtl_grp *tg = blkg_to_tg(blkg); |
674 | struct tg_stats_cpu *stats_cpu; | |
629ed0b1 TH |
675 | unsigned long flags; |
676 | ||
677 | /* If per cpu stats are not allocated yet, don't do any accounting. */ | |
8a3d2615 | 678 | if (tg->stats_cpu == NULL) |
629ed0b1 TH |
679 | return; |
680 | ||
681 | /* | |
682 | * Disabling interrupts to provide mutual exclusion between two | |
683 | * writes on same cpu. It probably is not needed for 64bit. Not | |
684 | * optimizing that case yet. | |
685 | */ | |
686 | local_irq_save(flags); | |
687 | ||
8a3d2615 | 688 | stats_cpu = this_cpu_ptr(tg->stats_cpu); |
629ed0b1 | 689 | |
629ed0b1 TH |
690 | blkg_rwstat_add(&stats_cpu->serviced, rw, 1); |
691 | blkg_rwstat_add(&stats_cpu->service_bytes, rw, bytes); | |
692 | ||
693 | local_irq_restore(flags); | |
694 | } | |
695 | ||
e43473b7 VG |
696 | static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio) |
697 | { | |
698 | bool rw = bio_data_dir(bio); | |
e43473b7 VG |
699 | |
700 | /* Charge the bio to the group */ | |
701 | tg->bytes_disp[rw] += bio->bi_size; | |
8e89d13f | 702 | tg->io_disp[rw]++; |
e43473b7 | 703 | |
629ed0b1 | 704 | throtl_update_dispatch_stats(tg_to_blkg(tg), bio->bi_size, bio->bi_rw); |
e43473b7 VG |
705 | } |
706 | ||
707 | static void throtl_add_bio_tg(struct throtl_data *td, struct throtl_grp *tg, | |
708 | struct bio *bio) | |
709 | { | |
710 | bool rw = bio_data_dir(bio); | |
711 | ||
712 | bio_list_add(&tg->bio_lists[rw], bio); | |
713 | /* Take a bio reference on tg */ | |
1adaf3dd | 714 | blkg_get(tg_to_blkg(tg)); |
e43473b7 VG |
715 | tg->nr_queued[rw]++; |
716 | td->nr_queued[rw]++; | |
717 | throtl_enqueue_tg(td, tg); | |
718 | } | |
719 | ||
720 | static void tg_update_disptime(struct throtl_data *td, struct throtl_grp *tg) | |
721 | { | |
722 | unsigned long read_wait = -1, write_wait = -1, min_wait = -1, disptime; | |
723 | struct bio *bio; | |
724 | ||
725 | if ((bio = bio_list_peek(&tg->bio_lists[READ]))) | |
726 | tg_may_dispatch(td, tg, bio, &read_wait); | |
727 | ||
728 | if ((bio = bio_list_peek(&tg->bio_lists[WRITE]))) | |
729 | tg_may_dispatch(td, tg, bio, &write_wait); | |
730 | ||
731 | min_wait = min(read_wait, write_wait); | |
732 | disptime = jiffies + min_wait; | |
733 | ||
e43473b7 VG |
734 | /* Update dispatch time */ |
735 | throtl_dequeue_tg(td, tg); | |
736 | tg->disptime = disptime; | |
737 | throtl_enqueue_tg(td, tg); | |
738 | } | |
739 | ||
740 | static void tg_dispatch_one_bio(struct throtl_data *td, struct throtl_grp *tg, | |
741 | bool rw, struct bio_list *bl) | |
742 | { | |
743 | struct bio *bio; | |
744 | ||
745 | bio = bio_list_pop(&tg->bio_lists[rw]); | |
746 | tg->nr_queued[rw]--; | |
1adaf3dd TH |
747 | /* Drop bio reference on blkg */ |
748 | blkg_put(tg_to_blkg(tg)); | |
e43473b7 VG |
749 | |
750 | BUG_ON(td->nr_queued[rw] <= 0); | |
751 | td->nr_queued[rw]--; | |
752 | ||
753 | throtl_charge_bio(tg, bio); | |
754 | bio_list_add(bl, bio); | |
755 | bio->bi_rw |= REQ_THROTTLED; | |
756 | ||
757 | throtl_trim_slice(td, tg, rw); | |
758 | } | |
759 | ||
760 | static int throtl_dispatch_tg(struct throtl_data *td, struct throtl_grp *tg, | |
761 | struct bio_list *bl) | |
762 | { | |
763 | unsigned int nr_reads = 0, nr_writes = 0; | |
764 | unsigned int max_nr_reads = throtl_grp_quantum*3/4; | |
c2f6805d | 765 | unsigned int max_nr_writes = throtl_grp_quantum - max_nr_reads; |
e43473b7 VG |
766 | struct bio *bio; |
767 | ||
768 | /* Try to dispatch 75% READS and 25% WRITES */ | |
769 | ||
770 | while ((bio = bio_list_peek(&tg->bio_lists[READ])) | |
771 | && tg_may_dispatch(td, tg, bio, NULL)) { | |
772 | ||
773 | tg_dispatch_one_bio(td, tg, bio_data_dir(bio), bl); | |
774 | nr_reads++; | |
775 | ||
776 | if (nr_reads >= max_nr_reads) | |
777 | break; | |
778 | } | |
779 | ||
780 | while ((bio = bio_list_peek(&tg->bio_lists[WRITE])) | |
781 | && tg_may_dispatch(td, tg, bio, NULL)) { | |
782 | ||
783 | tg_dispatch_one_bio(td, tg, bio_data_dir(bio), bl); | |
784 | nr_writes++; | |
785 | ||
786 | if (nr_writes >= max_nr_writes) | |
787 | break; | |
788 | } | |
789 | ||
790 | return nr_reads + nr_writes; | |
791 | } | |
792 | ||
793 | static int throtl_select_dispatch(struct throtl_data *td, struct bio_list *bl) | |
794 | { | |
795 | unsigned int nr_disp = 0; | |
796 | struct throtl_grp *tg; | |
c9e0332e | 797 | struct throtl_service_queue *sq = &td->service_queue; |
e43473b7 VG |
798 | |
799 | while (1) { | |
c9e0332e | 800 | tg = throtl_rb_first(sq); |
e43473b7 VG |
801 | |
802 | if (!tg) | |
803 | break; | |
804 | ||
805 | if (time_before(jiffies, tg->disptime)) | |
806 | break; | |
807 | ||
808 | throtl_dequeue_tg(td, tg); | |
809 | ||
810 | nr_disp += throtl_dispatch_tg(td, tg, bl); | |
811 | ||
2db6314c | 812 | if (tg->nr_queued[0] || tg->nr_queued[1]) |
e43473b7 | 813 | tg_update_disptime(td, tg); |
e43473b7 VG |
814 | |
815 | if (nr_disp >= throtl_quantum) | |
816 | break; | |
817 | } | |
818 | ||
819 | return nr_disp; | |
820 | } | |
821 | ||
cb76199c TH |
822 | /* work function to dispatch throttled bios */ |
823 | void blk_throtl_dispatch_work_fn(struct work_struct *work) | |
e43473b7 | 824 | { |
cb76199c TH |
825 | struct throtl_data *td = container_of(to_delayed_work(work), |
826 | struct throtl_data, dispatch_work); | |
827 | struct request_queue *q = td->queue; | |
e43473b7 VG |
828 | unsigned int nr_disp = 0; |
829 | struct bio_list bio_list_on_stack; | |
830 | struct bio *bio; | |
69d60eb9 | 831 | struct blk_plug plug; |
e43473b7 VG |
832 | |
833 | spin_lock_irq(q->queue_lock); | |
834 | ||
e43473b7 VG |
835 | bio_list_init(&bio_list_on_stack); |
836 | ||
d2f31a5f | 837 | throtl_log(td, "dispatch nr_queued=%u read=%u write=%u", |
6a525600 TH |
838 | td->nr_queued[READ] + td->nr_queued[WRITE], |
839 | td->nr_queued[READ], td->nr_queued[WRITE]); | |
e43473b7 VG |
840 | |
841 | nr_disp = throtl_select_dispatch(td, &bio_list_on_stack); | |
842 | ||
843 | if (nr_disp) | |
844 | throtl_log(td, "bios disp=%u", nr_disp); | |
845 | ||
846 | throtl_schedule_next_dispatch(td); | |
6a525600 | 847 | |
e43473b7 VG |
848 | spin_unlock_irq(q->queue_lock); |
849 | ||
850 | /* | |
851 | * If we dispatched some requests, unplug the queue to make sure | |
852 | * immediate dispatch | |
853 | */ | |
854 | if (nr_disp) { | |
69d60eb9 | 855 | blk_start_plug(&plug); |
e43473b7 VG |
856 | while((bio = bio_list_pop(&bio_list_on_stack))) |
857 | generic_make_request(bio); | |
69d60eb9 | 858 | blk_finish_plug(&plug); |
e43473b7 | 859 | } |
e43473b7 VG |
860 | } |
861 | ||
f95a04af TH |
862 | static u64 tg_prfill_cpu_rwstat(struct seq_file *sf, |
863 | struct blkg_policy_data *pd, int off) | |
41b38b6d | 864 | { |
f95a04af | 865 | struct throtl_grp *tg = pd_to_tg(pd); |
41b38b6d TH |
866 | struct blkg_rwstat rwstat = { }, tmp; |
867 | int i, cpu; | |
868 | ||
869 | for_each_possible_cpu(cpu) { | |
8a3d2615 | 870 | struct tg_stats_cpu *sc = per_cpu_ptr(tg->stats_cpu, cpu); |
41b38b6d TH |
871 | |
872 | tmp = blkg_rwstat_read((void *)sc + off); | |
873 | for (i = 0; i < BLKG_RWSTAT_NR; i++) | |
874 | rwstat.cnt[i] += tmp.cnt[i]; | |
875 | } | |
876 | ||
f95a04af | 877 | return __blkg_prfill_rwstat(sf, pd, &rwstat); |
41b38b6d TH |
878 | } |
879 | ||
8a3d2615 TH |
880 | static int tg_print_cpu_rwstat(struct cgroup *cgrp, struct cftype *cft, |
881 | struct seq_file *sf) | |
41b38b6d | 882 | { |
3c798398 | 883 | struct blkcg *blkcg = cgroup_to_blkcg(cgrp); |
41b38b6d | 884 | |
3c798398 | 885 | blkcg_print_blkgs(sf, blkcg, tg_prfill_cpu_rwstat, &blkcg_policy_throtl, |
5bc4afb1 | 886 | cft->private, true); |
41b38b6d TH |
887 | return 0; |
888 | } | |
889 | ||
f95a04af TH |
890 | static u64 tg_prfill_conf_u64(struct seq_file *sf, struct blkg_policy_data *pd, |
891 | int off) | |
60c2bc2d | 892 | { |
f95a04af TH |
893 | struct throtl_grp *tg = pd_to_tg(pd); |
894 | u64 v = *(u64 *)((void *)tg + off); | |
60c2bc2d | 895 | |
af133ceb | 896 | if (v == -1) |
60c2bc2d | 897 | return 0; |
f95a04af | 898 | return __blkg_prfill_u64(sf, pd, v); |
60c2bc2d TH |
899 | } |
900 | ||
f95a04af TH |
901 | static u64 tg_prfill_conf_uint(struct seq_file *sf, struct blkg_policy_data *pd, |
902 | int off) | |
e43473b7 | 903 | { |
f95a04af TH |
904 | struct throtl_grp *tg = pd_to_tg(pd); |
905 | unsigned int v = *(unsigned int *)((void *)tg + off); | |
fe071437 | 906 | |
af133ceb TH |
907 | if (v == -1) |
908 | return 0; | |
f95a04af | 909 | return __blkg_prfill_u64(sf, pd, v); |
e43473b7 VG |
910 | } |
911 | ||
af133ceb TH |
912 | static int tg_print_conf_u64(struct cgroup *cgrp, struct cftype *cft, |
913 | struct seq_file *sf) | |
8e89d13f | 914 | { |
3c798398 TH |
915 | blkcg_print_blkgs(sf, cgroup_to_blkcg(cgrp), tg_prfill_conf_u64, |
916 | &blkcg_policy_throtl, cft->private, false); | |
af133ceb | 917 | return 0; |
8e89d13f VG |
918 | } |
919 | ||
af133ceb TH |
920 | static int tg_print_conf_uint(struct cgroup *cgrp, struct cftype *cft, |
921 | struct seq_file *sf) | |
8e89d13f | 922 | { |
3c798398 TH |
923 | blkcg_print_blkgs(sf, cgroup_to_blkcg(cgrp), tg_prfill_conf_uint, |
924 | &blkcg_policy_throtl, cft->private, false); | |
af133ceb | 925 | return 0; |
60c2bc2d TH |
926 | } |
927 | ||
af133ceb TH |
928 | static int tg_set_conf(struct cgroup *cgrp, struct cftype *cft, const char *buf, |
929 | bool is_u64) | |
60c2bc2d | 930 | { |
3c798398 | 931 | struct blkcg *blkcg = cgroup_to_blkcg(cgrp); |
60c2bc2d | 932 | struct blkg_conf_ctx ctx; |
af133ceb | 933 | struct throtl_grp *tg; |
a2b1693b | 934 | struct throtl_data *td; |
60c2bc2d TH |
935 | int ret; |
936 | ||
3c798398 | 937 | ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx); |
60c2bc2d TH |
938 | if (ret) |
939 | return ret; | |
940 | ||
af133ceb | 941 | tg = blkg_to_tg(ctx.blkg); |
a2b1693b | 942 | td = ctx.blkg->q->td; |
af133ceb | 943 | |
a2b1693b TH |
944 | if (!ctx.v) |
945 | ctx.v = -1; | |
af133ceb | 946 | |
a2b1693b TH |
947 | if (is_u64) |
948 | *(u64 *)((void *)tg + cft->private) = ctx.v; | |
949 | else | |
950 | *(unsigned int *)((void *)tg + cft->private) = ctx.v; | |
af133ceb | 951 | |
632b4493 TH |
952 | throtl_log_tg(td, tg, "limit change rbps=%llu wbps=%llu riops=%u wiops=%u", |
953 | tg->bps[READ], tg->bps[WRITE], | |
954 | tg->iops[READ], tg->iops[WRITE]); | |
955 | ||
956 | /* | |
957 | * We're already holding queue_lock and know @tg is valid. Let's | |
958 | * apply the new config directly. | |
959 | * | |
960 | * Restart the slices for both READ and WRITES. It might happen | |
961 | * that a group's limit are dropped suddenly and we don't want to | |
962 | * account recently dispatched IO with new low rate. | |
963 | */ | |
964 | throtl_start_new_slice(td, tg, 0); | |
965 | throtl_start_new_slice(td, tg, 1); | |
966 | ||
967 | if (throtl_tg_on_rr(tg)) { | |
968 | tg_update_disptime(td, tg); | |
969 | throtl_schedule_next_dispatch(td); | |
970 | } | |
60c2bc2d TH |
971 | |
972 | blkg_conf_finish(&ctx); | |
a2b1693b | 973 | return 0; |
8e89d13f VG |
974 | } |
975 | ||
af133ceb TH |
976 | static int tg_set_conf_u64(struct cgroup *cgrp, struct cftype *cft, |
977 | const char *buf) | |
60c2bc2d | 978 | { |
af133ceb | 979 | return tg_set_conf(cgrp, cft, buf, true); |
60c2bc2d TH |
980 | } |
981 | ||
af133ceb TH |
982 | static int tg_set_conf_uint(struct cgroup *cgrp, struct cftype *cft, |
983 | const char *buf) | |
60c2bc2d | 984 | { |
af133ceb | 985 | return tg_set_conf(cgrp, cft, buf, false); |
60c2bc2d TH |
986 | } |
987 | ||
988 | static struct cftype throtl_files[] = { | |
989 | { | |
990 | .name = "throttle.read_bps_device", | |
af133ceb TH |
991 | .private = offsetof(struct throtl_grp, bps[READ]), |
992 | .read_seq_string = tg_print_conf_u64, | |
993 | .write_string = tg_set_conf_u64, | |
60c2bc2d TH |
994 | .max_write_len = 256, |
995 | }, | |
996 | { | |
997 | .name = "throttle.write_bps_device", | |
af133ceb TH |
998 | .private = offsetof(struct throtl_grp, bps[WRITE]), |
999 | .read_seq_string = tg_print_conf_u64, | |
1000 | .write_string = tg_set_conf_u64, | |
60c2bc2d TH |
1001 | .max_write_len = 256, |
1002 | }, | |
1003 | { | |
1004 | .name = "throttle.read_iops_device", | |
af133ceb TH |
1005 | .private = offsetof(struct throtl_grp, iops[READ]), |
1006 | .read_seq_string = tg_print_conf_uint, | |
1007 | .write_string = tg_set_conf_uint, | |
60c2bc2d TH |
1008 | .max_write_len = 256, |
1009 | }, | |
1010 | { | |
1011 | .name = "throttle.write_iops_device", | |
af133ceb TH |
1012 | .private = offsetof(struct throtl_grp, iops[WRITE]), |
1013 | .read_seq_string = tg_print_conf_uint, | |
1014 | .write_string = tg_set_conf_uint, | |
60c2bc2d TH |
1015 | .max_write_len = 256, |
1016 | }, | |
1017 | { | |
1018 | .name = "throttle.io_service_bytes", | |
5bc4afb1 | 1019 | .private = offsetof(struct tg_stats_cpu, service_bytes), |
8a3d2615 | 1020 | .read_seq_string = tg_print_cpu_rwstat, |
60c2bc2d TH |
1021 | }, |
1022 | { | |
1023 | .name = "throttle.io_serviced", | |
5bc4afb1 | 1024 | .private = offsetof(struct tg_stats_cpu, serviced), |
8a3d2615 | 1025 | .read_seq_string = tg_print_cpu_rwstat, |
60c2bc2d TH |
1026 | }, |
1027 | { } /* terminate */ | |
1028 | }; | |
1029 | ||
da527770 | 1030 | static void throtl_shutdown_wq(struct request_queue *q) |
e43473b7 VG |
1031 | { |
1032 | struct throtl_data *td = q->td; | |
1033 | ||
cb76199c | 1034 | cancel_delayed_work_sync(&td->dispatch_work); |
e43473b7 VG |
1035 | } |
1036 | ||
3c798398 | 1037 | static struct blkcg_policy blkcg_policy_throtl = { |
f9fcc2d3 TH |
1038 | .pd_size = sizeof(struct throtl_grp), |
1039 | .cftypes = throtl_files, | |
1040 | ||
1041 | .pd_init_fn = throtl_pd_init, | |
1042 | .pd_exit_fn = throtl_pd_exit, | |
1043 | .pd_reset_stats_fn = throtl_pd_reset_stats, | |
e43473b7 VG |
1044 | }; |
1045 | ||
bc16a4f9 | 1046 | bool blk_throtl_bio(struct request_queue *q, struct bio *bio) |
e43473b7 VG |
1047 | { |
1048 | struct throtl_data *td = q->td; | |
1049 | struct throtl_grp *tg; | |
e43473b7 | 1050 | bool rw = bio_data_dir(bio), update_disptime = true; |
3c798398 | 1051 | struct blkcg *blkcg; |
bc16a4f9 | 1052 | bool throttled = false; |
e43473b7 VG |
1053 | |
1054 | if (bio->bi_rw & REQ_THROTTLED) { | |
1055 | bio->bi_rw &= ~REQ_THROTTLED; | |
bc16a4f9 | 1056 | goto out; |
e43473b7 VG |
1057 | } |
1058 | ||
af75cd3c VG |
1059 | /* |
1060 | * A throtl_grp pointer retrieved under rcu can be used to access | |
1061 | * basic fields like stats and io rates. If a group has no rules, | |
1062 | * just update the dispatch stats in lockless manner and return. | |
1063 | */ | |
af75cd3c | 1064 | rcu_read_lock(); |
3c798398 | 1065 | blkcg = bio_blkcg(bio); |
cd1604fa | 1066 | tg = throtl_lookup_tg(td, blkcg); |
af75cd3c | 1067 | if (tg) { |
af75cd3c | 1068 | if (tg_no_rule_group(tg, rw)) { |
629ed0b1 TH |
1069 | throtl_update_dispatch_stats(tg_to_blkg(tg), |
1070 | bio->bi_size, bio->bi_rw); | |
2a7f1244 | 1071 | goto out_unlock_rcu; |
af75cd3c VG |
1072 | } |
1073 | } | |
af75cd3c VG |
1074 | |
1075 | /* | |
1076 | * Either group has not been allocated yet or it is not an unlimited | |
1077 | * IO group | |
1078 | */ | |
e43473b7 | 1079 | spin_lock_irq(q->queue_lock); |
cd1604fa | 1080 | tg = throtl_lookup_create_tg(td, blkcg); |
bc16a4f9 TH |
1081 | if (unlikely(!tg)) |
1082 | goto out_unlock; | |
f469a7b4 | 1083 | |
e43473b7 VG |
1084 | if (tg->nr_queued[rw]) { |
1085 | /* | |
1086 | * There is already another bio queued in same dir. No | |
1087 | * need to update dispatch time. | |
1088 | */ | |
231d704b | 1089 | update_disptime = false; |
e43473b7 | 1090 | goto queue_bio; |
de701c74 | 1091 | |
e43473b7 VG |
1092 | } |
1093 | ||
1094 | /* Bio is with-in rate limit of group */ | |
1095 | if (tg_may_dispatch(td, tg, bio, NULL)) { | |
1096 | throtl_charge_bio(tg, bio); | |
04521db0 VG |
1097 | |
1098 | /* | |
1099 | * We need to trim slice even when bios are not being queued | |
1100 | * otherwise it might happen that a bio is not queued for | |
1101 | * a long time and slice keeps on extending and trim is not | |
1102 | * called for a long time. Now if limits are reduced suddenly | |
1103 | * we take into account all the IO dispatched so far at new | |
1104 | * low rate and * newly queued IO gets a really long dispatch | |
1105 | * time. | |
1106 | * | |
1107 | * So keep on trimming slice even if bio is not queued. | |
1108 | */ | |
1109 | throtl_trim_slice(td, tg, rw); | |
bc16a4f9 | 1110 | goto out_unlock; |
e43473b7 VG |
1111 | } |
1112 | ||
1113 | queue_bio: | |
fd16d263 | 1114 | throtl_log_tg(td, tg, "[%c] bio. bdisp=%llu sz=%u bps=%llu" |
8e89d13f VG |
1115 | " iodisp=%u iops=%u queued=%d/%d", |
1116 | rw == READ ? 'R' : 'W', | |
e43473b7 | 1117 | tg->bytes_disp[rw], bio->bi_size, tg->bps[rw], |
8e89d13f | 1118 | tg->io_disp[rw], tg->iops[rw], |
e43473b7 VG |
1119 | tg->nr_queued[READ], tg->nr_queued[WRITE]); |
1120 | ||
671058fb | 1121 | bio_associate_current(bio); |
e43473b7 | 1122 | throtl_add_bio_tg(q->td, tg, bio); |
bc16a4f9 | 1123 | throttled = true; |
e43473b7 VG |
1124 | |
1125 | if (update_disptime) { | |
1126 | tg_update_disptime(td, tg); | |
1127 | throtl_schedule_next_dispatch(td); | |
1128 | } | |
1129 | ||
bc16a4f9 | 1130 | out_unlock: |
e43473b7 | 1131 | spin_unlock_irq(q->queue_lock); |
2a7f1244 TH |
1132 | out_unlock_rcu: |
1133 | rcu_read_unlock(); | |
bc16a4f9 TH |
1134 | out: |
1135 | return throttled; | |
e43473b7 VG |
1136 | } |
1137 | ||
c9a929dd TH |
1138 | /** |
1139 | * blk_throtl_drain - drain throttled bios | |
1140 | * @q: request_queue to drain throttled bios for | |
1141 | * | |
1142 | * Dispatch all currently throttled bios on @q through ->make_request_fn(). | |
1143 | */ | |
1144 | void blk_throtl_drain(struct request_queue *q) | |
1145 | __releases(q->queue_lock) __acquires(q->queue_lock) | |
1146 | { | |
1147 | struct throtl_data *td = q->td; | |
c9e0332e | 1148 | struct throtl_service_queue *sq = &td->service_queue; |
c9a929dd TH |
1149 | struct throtl_grp *tg; |
1150 | struct bio_list bl; | |
1151 | struct bio *bio; | |
1152 | ||
8bcb6c7d | 1153 | queue_lockdep_assert_held(q); |
c9a929dd TH |
1154 | |
1155 | bio_list_init(&bl); | |
1156 | ||
c9e0332e | 1157 | while ((tg = throtl_rb_first(sq))) { |
c9a929dd TH |
1158 | throtl_dequeue_tg(td, tg); |
1159 | ||
1160 | while ((bio = bio_list_peek(&tg->bio_lists[READ]))) | |
1161 | tg_dispatch_one_bio(td, tg, bio_data_dir(bio), &bl); | |
1162 | while ((bio = bio_list_peek(&tg->bio_lists[WRITE]))) | |
1163 | tg_dispatch_one_bio(td, tg, bio_data_dir(bio), &bl); | |
1164 | } | |
1165 | spin_unlock_irq(q->queue_lock); | |
1166 | ||
1167 | while ((bio = bio_list_pop(&bl))) | |
1168 | generic_make_request(bio); | |
1169 | ||
1170 | spin_lock_irq(q->queue_lock); | |
1171 | } | |
1172 | ||
e43473b7 VG |
1173 | int blk_throtl_init(struct request_queue *q) |
1174 | { | |
1175 | struct throtl_data *td; | |
a2b1693b | 1176 | int ret; |
e43473b7 VG |
1177 | |
1178 | td = kzalloc_node(sizeof(*td), GFP_KERNEL, q->node); | |
1179 | if (!td) | |
1180 | return -ENOMEM; | |
1181 | ||
c9e0332e | 1182 | td->service_queue = THROTL_SERVICE_QUEUE_INITIALIZER; |
cb76199c | 1183 | INIT_DELAYED_WORK(&td->dispatch_work, blk_throtl_dispatch_work_fn); |
e43473b7 | 1184 | |
cd1604fa | 1185 | q->td = td; |
29b12589 | 1186 | td->queue = q; |
02977e4a | 1187 | |
a2b1693b | 1188 | /* activate policy */ |
3c798398 | 1189 | ret = blkcg_activate_policy(q, &blkcg_policy_throtl); |
a2b1693b | 1190 | if (ret) |
f51b802c | 1191 | kfree(td); |
a2b1693b | 1192 | return ret; |
e43473b7 VG |
1193 | } |
1194 | ||
1195 | void blk_throtl_exit(struct request_queue *q) | |
1196 | { | |
c875f4d0 | 1197 | BUG_ON(!q->td); |
da527770 | 1198 | throtl_shutdown_wq(q); |
3c798398 | 1199 | blkcg_deactivate_policy(q, &blkcg_policy_throtl); |
c9a929dd | 1200 | kfree(q->td); |
e43473b7 VG |
1201 | } |
1202 | ||
1203 | static int __init throtl_init(void) | |
1204 | { | |
450adcbe VG |
1205 | kthrotld_workqueue = alloc_workqueue("kthrotld", WQ_MEM_RECLAIM, 0); |
1206 | if (!kthrotld_workqueue) | |
1207 | panic("Failed to create kthrotld\n"); | |
1208 | ||
3c798398 | 1209 | return blkcg_policy_register(&blkcg_policy_throtl); |
e43473b7 VG |
1210 | } |
1211 | ||
1212 | module_init(throtl_init); |