Blame view
block/blk-throttle.c
67.8 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
e43473b7f blkio: Core imple... |
2 3 4 5 6 7 8 9 10 11 12 |
/* * Interface for controlling IO bandwidth on a request queue * * Copyright (C) 2010 Vivek Goyal <vgoyal@redhat.com> */ #include <linux/module.h> #include <linux/slab.h> #include <linux/blkdev.h> #include <linux/bio.h> #include <linux/blktrace_api.h> |
eea8f41cc blkcg: move block... |
13 |
#include <linux/blk-cgroup.h> |
bc9fcbf9c block: move blk_t... |
14 |
#include "blk.h" |
e43473b7f blkio: Core imple... |
15 16 17 18 19 20 |
/* Max dispatch from a group in 1 round */ static int throtl_grp_quantum = 8; /* Total max dispatch from all groups in one round */ static int throtl_quantum = 32; |
d61fcfa4b blk-throttle: cho... |
21 22 23 |
/* Throttling is performed over a slice and after that slice is renewed */ #define DFL_THROTL_SLICE_HD (HZ / 10) #define DFL_THROTL_SLICE_SSD (HZ / 50) |
297e3d854 blk-throttle: mak... |
24 |
#define MAX_THROTL_SLICE (HZ) |
9e234eeaf blk-throttle: add... |
25 |
#define MAX_IDLE_TIME (5L * 1000 * 1000) /* 5 s */ |
9bb67aeb9 blk-throttle: res... |
26 27 |
#define MIN_THROTL_BPS (320 * 1024) #define MIN_THROTL_IOPS (10) |
b4f428ef2 blk-throttle: for... |
28 29 |
#define DFL_LATENCY_TARGET (-1L) #define DFL_IDLE_THRESHOLD (0) |
6679a90c4 blk-throttle: set... |
30 31 32 33 34 35 36 |
#define DFL_HD_BASELINE_LATENCY (4000L) /* 4ms */ #define LATENCY_FILTERED_SSD (0) /* * For HD, very small latency comes from sequential IO. Such IO is helpless to * help determine if its IO is impacted by others, hence we ignore the IO */ #define LATENCY_FILTERED_HD (1000L) /* 1ms */ |
e43473b7f blkio: Core imple... |
37 |
|
b9147dd1b blk-throttle: add... |
38 |
#define SKIP_LATENCY (((u64)1) << BLK_STAT_RES_SHIFT) |
3c798398e blkcg: mass renam... |
39 |
static struct blkcg_policy blkcg_policy_throtl; |
0381411e4 blkcg: let blkcg ... |
40 |
|
450adcbe5 blk-throttle: Do ... |
41 42 |
/* A workqueue to queue throttle related work */ static struct workqueue_struct *kthrotld_workqueue; |
450adcbe5 blk-throttle: Do ... |
43 |
|
c5cc2070b blk-throttle: add... |
44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 |
/* * To implement hierarchical throttling, throtl_grps form a tree and bios * are dispatched upwards level by level until they reach the top and get * issued. When dispatching bios from the children and local group at each * level, if the bios are dispatched into a single bio_list, there's a risk * of a local or child group which can queue many bios at once filling up * the list starving others. * * To avoid such starvation, dispatched bios are queued separately * according to where they came from. When they are again dispatched to * the parent, they're popped in round-robin order so that no single source * hogs the dispatch window. * * throtl_qnode is used to keep the queued bios separated by their sources. * Bios are queued to throtl_qnode which in turn is queued to * throtl_service_queue and then dispatched in round-robin order. * * It's also used to track the reference counts on blkg's. A qnode always * belongs to a throtl_grp and gets queued on itself or the parent, so * incrementing the reference of the associated throtl_grp when a qnode is * queued and decrementing when dequeued is enough to keep the whole blkg * tree pinned while bios are in flight. */ struct throtl_qnode { struct list_head node; /* service_queue->queued[] */ struct bio_list bios; /* queued bios */ struct throtl_grp *tg; /* tg this qnode belongs to */ }; |
c9e0332e8 blk-throttle: ren... |
72 |
struct throtl_service_queue { |
77216b048 blk-throttle: add... |
73 |
struct throtl_service_queue *parent_sq; /* the parent service_queue */ |
73f0d49a9 blk-throttle: mov... |
74 75 76 77 |
/* * Bios queued directly to this service_queue or dispatched from * children throtl_grp's. */ |
c5cc2070b blk-throttle: add... |
78 |
struct list_head queued[2]; /* throtl_qnode [READ/WRITE] */ |
73f0d49a9 blk-throttle: mov... |
79 80 81 82 83 84 |
unsigned int nr_queued[2]; /* number of queued bios */ /* * RB tree of active children throtl_grp's, which are sorted by * their ->disptime. */ |
c9e0332e8 blk-throttle: ren... |
85 86 87 88 |
struct rb_root pending_tree; /* RB tree of active tgs */ struct rb_node *first_pending; /* first node in the tree */ unsigned int nr_pending; /* # queued in the tree */ unsigned long first_pending_disptime; /* disptime of the first tg */ |
69df0ab03 blk-throttle: sep... |
89 |
struct timer_list pending_timer; /* fires on first_pending_disptime */ |
e43473b7f blkio: Core imple... |
90 |
}; |
5b2c16aae blk-throttle: sim... |
91 92 |
enum tg_state_flags { THROTL_TG_PENDING = 1 << 0, /* on parent's pending tree */ |
0e9f4164b blk-throttle: gen... |
93 |
THROTL_TG_WAS_EMPTY = 1 << 1, /* bio_lists[] became non-empty */ |
5b2c16aae blk-throttle: sim... |
94 |
}; |
e43473b7f blkio: Core imple... |
95 |
#define rb_entry_tg(node) rb_entry((node), struct throtl_grp, rb_node) |
9f626e372 blk-throttle: pre... |
96 |
enum { |
cd5ab1b0f blk-throttle: add... |
97 |
LIMIT_LOW, |
9f626e372 blk-throttle: pre... |
98 99 100 |
LIMIT_MAX, LIMIT_CNT, }; |
e43473b7f blkio: Core imple... |
101 |
struct throtl_grp { |
f95a04afa blkcg: embed stru... |
102 103 |
/* must be the first member */ struct blkg_policy_data pd; |
c9e0332e8 blk-throttle: ren... |
104 |
/* active throtl group service_queue member */ |
e43473b7f blkio: Core imple... |
105 |
struct rb_node rb_node; |
0f3457f60 blk-throttle: add... |
106 107 |
/* throtl_data this group belongs to */ struct throtl_data *td; |
49a2f1e3f blk-throttle: add... |
108 109 |
/* this group's service queue */ struct throtl_service_queue service_queue; |
e43473b7f blkio: Core imple... |
110 |
/* |
c5cc2070b blk-throttle: add... |
111 112 113 114 115 116 117 118 119 120 121 |
* qnode_on_self is used when bios are directly queued to this * throtl_grp so that local bios compete fairly with bios * dispatched from children. qnode_on_parent is used when bios are * dispatched from this throtl_grp into its parent and will compete * with the sibling qnode_on_parents and the parent's * qnode_on_self. */ struct throtl_qnode qnode_on_self[2]; struct throtl_qnode qnode_on_parent[2]; /* |
e43473b7f blkio: Core imple... |
122 123 124 125 126 |
* Dispatch time in jiffies. This is the estimated time when group * will unthrottle and is ready to dispatch more bio. It is used as * key to sort active groups in service tree. */ unsigned long disptime; |
e43473b7f blkio: Core imple... |
127 |
unsigned int flags; |
693e751e7 blk-throttle: imp... |
128 129 |
/* are there any throtl rules between this group and td? */ bool has_rules[2]; |
cd5ab1b0f blk-throttle: add... |
130 |
/* internally used bytes per second rate limits */ |
9f626e372 blk-throttle: pre... |
131 |
uint64_t bps[2][LIMIT_CNT]; |
cd5ab1b0f blk-throttle: add... |
132 133 |
/* user configured bps limits */ uint64_t bps_conf[2][LIMIT_CNT]; |
e43473b7f blkio: Core imple... |
134 |
|
cd5ab1b0f blk-throttle: add... |
135 |
/* internally used IOPS limits */ |
9f626e372 blk-throttle: pre... |
136 |
unsigned int iops[2][LIMIT_CNT]; |
cd5ab1b0f blk-throttle: add... |
137 138 |
/* user configured IOPS limits */ unsigned int iops_conf[2][LIMIT_CNT]; |
8e89d13f4 blkio: Implementa... |
139 |
|
e43473b7f blkio: Core imple... |
140 141 |
/* Number of bytes disptached in current slice */ uint64_t bytes_disp[2]; |
8e89d13f4 blkio: Implementa... |
142 143 |
/* Number of bio's dispatched in current slice */ unsigned int io_disp[2]; |
e43473b7f blkio: Core imple... |
144 |
|
3f0abd806 blk-throttle: add... |
145 146 147 148 149 150 |
unsigned long last_low_overflow_time[2]; uint64_t last_bytes_disp[2]; unsigned int last_io_disp[2]; unsigned long last_check_time; |
ec80991d6 blk-throttle: add... |
151 |
unsigned long latency_target; /* us */ |
5b81fc3cc blk-throttle: add... |
152 |
unsigned long latency_target_conf; /* us */ |
e43473b7f blkio: Core imple... |
153 154 155 |
/* When did we start a new slice */ unsigned long slice_start[2]; unsigned long slice_end[2]; |
9e234eeaf blk-throttle: add... |
156 157 158 159 160 |
unsigned long last_finish_time; /* ns / 1024 */ unsigned long checked_last_finish_time; /* ns / 1024 */ unsigned long avg_idletime; /* ns / 1024 */ unsigned long idletime_threshold; /* us */ |
5b81fc3cc blk-throttle: add... |
161 |
unsigned long idletime_threshold_conf; /* us */ |
53696b8d2 blk-throttle: add... |
162 163 164 165 |
unsigned int bio_cnt; /* total bios */ unsigned int bad_bio_cnt; /* bios exceeding latency threshold */ unsigned long bio_cnt_reset_time; |
e43473b7f blkio: Core imple... |
166 |
}; |
b9147dd1b blk-throttle: add... |
167 168 169 170 171 172 173 174 175 176 177 178 |
/* We measure latency for request size from <= 4k to >= 1M */ #define LATENCY_BUCKET_SIZE 9 struct latency_bucket { unsigned long total_latency; /* ns / 1024 */ int samples; }; struct avg_latency_bucket { unsigned long latency; /* ns / 1024 */ bool valid; }; |
e43473b7f blkio: Core imple... |
179 180 |
struct throtl_data { |
e43473b7f blkio: Core imple... |
181 |
/* service tree for active throtl groups */ |
c9e0332e8 blk-throttle: ren... |
182 |
struct throtl_service_queue service_queue; |
e43473b7f blkio: Core imple... |
183 |
|
e43473b7f blkio: Core imple... |
184 185 186 187 |
struct request_queue *queue; /* Total Number of queued bios on READ and WRITE lists */ unsigned int nr_queued[2]; |
297e3d854 blk-throttle: mak... |
188 |
unsigned int throtl_slice; |
e43473b7f blkio: Core imple... |
189 |
/* Work for dispatching throttled bios */ |
69df0ab03 blk-throttle: sep... |
190 |
struct work_struct dispatch_work; |
9f626e372 blk-throttle: pre... |
191 192 |
unsigned int limit_index; bool limit_valid[LIMIT_CNT]; |
3f0abd806 blk-throttle: add... |
193 194 195 |
unsigned long low_upgrade_time; unsigned long low_downgrade_time; |
7394e31fa blk-throttle: mak... |
196 197 |
unsigned int scale; |
b9147dd1b blk-throttle: add... |
198 199 200 201 202 |
struct latency_bucket tmp_buckets[LATENCY_BUCKET_SIZE]; struct avg_latency_bucket avg_buckets[LATENCY_BUCKET_SIZE]; struct latency_bucket __percpu *latency_buckets; unsigned long last_calculate_time; |
6679a90c4 blk-throttle: set... |
203 |
unsigned long filtered_latency; |
b9147dd1b blk-throttle: add... |
204 205 |
bool track_bio_latency; |
e43473b7f blkio: Core imple... |
206 |
}; |
69df0ab03 blk-throttle: sep... |
207 |
static void throtl_pending_timer_fn(unsigned long arg); |
f95a04afa blkcg: embed stru... |
208 209 210 211 |
static inline struct throtl_grp *pd_to_tg(struct blkg_policy_data *pd) { return pd ? container_of(pd, struct throtl_grp, pd) : NULL; } |
3c798398e blkcg: mass renam... |
212 |
static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg) |
0381411e4 blkcg: let blkcg ... |
213 |
{ |
f95a04afa blkcg: embed stru... |
214 |
return pd_to_tg(blkg_to_pd(blkg, &blkcg_policy_throtl)); |
0381411e4 blkcg: let blkcg ... |
215 |
} |
3c798398e blkcg: mass renam... |
216 |
static inline struct blkcg_gq *tg_to_blkg(struct throtl_grp *tg) |
0381411e4 blkcg: let blkcg ... |
217 |
{ |
f95a04afa blkcg: embed stru... |
218 |
return pd_to_blkg(&tg->pd); |
0381411e4 blkcg: let blkcg ... |
219 |
} |
fda6f272c blk-throttle: imp... |
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 |
/** * sq_to_tg - return the throl_grp the specified service queue belongs to * @sq: the throtl_service_queue of interest * * Return the throtl_grp @sq belongs to. If @sq is the top-level one * embedded in throtl_data, %NULL is returned. */ static struct throtl_grp *sq_to_tg(struct throtl_service_queue *sq) { if (sq && sq->parent_sq) return container_of(sq, struct throtl_grp, service_queue); else return NULL; } /** * sq_to_td - return throtl_data the specified service queue belongs to * @sq: the throtl_service_queue of interest * |
b43daedc0 scripts/spelling.... |
239 |
* A service_queue can be embedded in either a throtl_grp or throtl_data. |
fda6f272c blk-throttle: imp... |
240 241 242 243 244 245 246 247 248 249 250 |
* Determine the associated throtl_data accordingly and return it. */ static struct throtl_data *sq_to_td(struct throtl_service_queue *sq) { struct throtl_grp *tg = sq_to_tg(sq); if (tg) return tg->td; else return container_of(sq, struct throtl_data, service_queue); } |
7394e31fa blk-throttle: mak... |
251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 |
/* * cgroup's limit in LIMIT_MAX is scaled if low limit is set. This scale is to * make the IO dispatch more smooth. * Scale up: linearly scale up according to lapsed time since upgrade. For * every throtl_slice, the limit scales up 1/2 .low limit till the * limit hits .max limit * Scale down: exponentially scale down if a cgroup doesn't hit its .low limit */ static uint64_t throtl_adjusted_limit(uint64_t low, struct throtl_data *td) { /* arbitrary value to avoid too big scale */ if (td->scale < 4096 && time_after_eq(jiffies, td->low_upgrade_time + td->scale * td->throtl_slice)) td->scale = (jiffies - td->low_upgrade_time) / td->throtl_slice; return low + (low >> 1) * td->scale; } |
9f626e372 blk-throttle: pre... |
268 269 |
static uint64_t tg_bps_limit(struct throtl_grp *tg, int rw) { |
b22c417c8 blk-throttle: con... |
270 |
struct blkcg_gq *blkg = tg_to_blkg(tg); |
7394e31fa blk-throttle: mak... |
271 |
struct throtl_data *td; |
b22c417c8 blk-throttle: con... |
272 273 274 275 |
uint64_t ret; if (cgroup_subsys_on_dfl(io_cgrp_subsys) && !blkg->parent) return U64_MAX; |
7394e31fa blk-throttle: mak... |
276 277 278 |
td = tg->td; ret = tg->bps[rw][td->limit_index]; |
9bb67aeb9 blk-throttle: res... |
279 280 281 282 283 284 285 286 |
if (ret == 0 && td->limit_index == LIMIT_LOW) { /* intermediate node or iops isn't 0 */ if (!list_empty(&blkg->blkcg->css.children) || tg->iops[rw][td->limit_index]) return U64_MAX; else return MIN_THROTL_BPS; } |
7394e31fa blk-throttle: mak... |
287 288 289 290 291 292 293 294 |
if (td->limit_index == LIMIT_MAX && tg->bps[rw][LIMIT_LOW] && tg->bps[rw][LIMIT_LOW] != tg->bps[rw][LIMIT_MAX]) { uint64_t adjusted; adjusted = throtl_adjusted_limit(tg->bps[rw][LIMIT_LOW], td); ret = min(tg->bps[rw][LIMIT_MAX], adjusted); } |
b22c417c8 blk-throttle: con... |
295 |
return ret; |
9f626e372 blk-throttle: pre... |
296 297 298 299 |
} static unsigned int tg_iops_limit(struct throtl_grp *tg, int rw) { |
b22c417c8 blk-throttle: con... |
300 |
struct blkcg_gq *blkg = tg_to_blkg(tg); |
7394e31fa blk-throttle: mak... |
301 |
struct throtl_data *td; |
b22c417c8 blk-throttle: con... |
302 303 304 305 |
unsigned int ret; if (cgroup_subsys_on_dfl(io_cgrp_subsys) && !blkg->parent) return UINT_MAX; |
9bb67aeb9 blk-throttle: res... |
306 |
|
7394e31fa blk-throttle: mak... |
307 308 |
td = tg->td; ret = tg->iops[rw][td->limit_index]; |
9bb67aeb9 blk-throttle: res... |
309 310 311 312 313 314 315 316 |
if (ret == 0 && tg->td->limit_index == LIMIT_LOW) { /* intermediate node or bps isn't 0 */ if (!list_empty(&blkg->blkcg->css.children) || tg->bps[rw][td->limit_index]) return UINT_MAX; else return MIN_THROTL_IOPS; } |
7394e31fa blk-throttle: mak... |
317 318 319 320 321 322 323 324 325 326 |
if (td->limit_index == LIMIT_MAX && tg->iops[rw][LIMIT_LOW] && tg->iops[rw][LIMIT_LOW] != tg->iops[rw][LIMIT_MAX]) { uint64_t adjusted; adjusted = throtl_adjusted_limit(tg->iops[rw][LIMIT_LOW], td); if (adjusted > UINT_MAX) adjusted = UINT_MAX; ret = min_t(unsigned int, tg->iops[rw][LIMIT_MAX], adjusted); } |
b22c417c8 blk-throttle: con... |
327 |
return ret; |
9f626e372 blk-throttle: pre... |
328 |
} |
b9147dd1b blk-throttle: add... |
329 330 |
#define request_bucket_index(sectors) \ clamp_t(int, order_base_2(sectors) - 3, 0, LATENCY_BUCKET_SIZE - 1) |
fda6f272c blk-throttle: imp... |
331 332 333 334 335 336 337 338 |
/** * throtl_log - log debug message via blktrace * @sq: the service_queue being reported * @fmt: printf format string * @args: printf args * * The messages are prefixed with "throtl BLKG_NAME" if @sq belongs to a * throtl_grp; otherwise, just "throtl". |
fda6f272c blk-throttle: imp... |
339 340 341 342 343 344 |
*/ #define throtl_log(sq, fmt, args...) do { \ struct throtl_grp *__tg = sq_to_tg((sq)); \ struct throtl_data *__td = sq_to_td((sq)); \ \ (void)__td; \ |
59fa0224c blk-throttle: don... |
345 346 |
if (likely(!blk_trace_note_message_enabled(__td->queue))) \ break; \ |
fda6f272c blk-throttle: imp... |
347 |
if ((__tg)) { \ |
35fe6d763 block: use standa... |
348 349 |
blk_add_cgroup_trace_msg(__td->queue, \ tg_to_blkg(__tg)->blkcg, "throtl " fmt, ##args);\ |
fda6f272c blk-throttle: imp... |
350 351 352 |
} else { \ blk_add_trace_msg(__td->queue, "throtl " fmt, ##args); \ } \ |
54e7ed12b blkcg: remove blk... |
353 |
} while (0) |
e43473b7f blkio: Core imple... |
354 |
|
ea0ea2bc6 blk-throttle: cap... |
355 356 357 358 359 360 361 |
static inline unsigned int throtl_bio_data_size(struct bio *bio) { /* assume it's one sector */ if (unlikely(bio_op(bio) == REQ_OP_DISCARD)) return 512; return bio->bi_iter.bi_size; } |
c5cc2070b blk-throttle: add... |
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 |
static void throtl_qnode_init(struct throtl_qnode *qn, struct throtl_grp *tg) { INIT_LIST_HEAD(&qn->node); bio_list_init(&qn->bios); qn->tg = tg; } /** * throtl_qnode_add_bio - add a bio to a throtl_qnode and activate it * @bio: bio being added * @qn: qnode to add bio to * @queued: the service_queue->queued[] list @qn belongs to * * Add @bio to @qn and put @qn on @queued if it's not already on. * @qn->tg's reference count is bumped when @qn is activated. See the * comment on top of throtl_qnode definition for details. */ static void throtl_qnode_add_bio(struct bio *bio, struct throtl_qnode *qn, struct list_head *queued) { bio_list_add(&qn->bios, bio); if (list_empty(&qn->node)) { list_add_tail(&qn->node, queued); blkg_get(tg_to_blkg(qn->tg)); } } /** * throtl_peek_queued - peek the first bio on a qnode list * @queued: the qnode list to peek */ static struct bio *throtl_peek_queued(struct list_head *queued) { struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node); struct bio *bio; if (list_empty(queued)) return NULL; bio = bio_list_peek(&qn->bios); WARN_ON_ONCE(!bio); return bio; } /** * throtl_pop_queued - pop the first bio form a qnode list * @queued: the qnode list to pop a bio from * @tg_to_put: optional out argument for throtl_grp to put * * Pop the first bio from the qnode list @queued. After popping, the first * qnode is removed from @queued if empty or moved to the end of @queued so * that the popping order is round-robin. * * When the first qnode is removed, its associated throtl_grp should be put * too. If @tg_to_put is NULL, this function automatically puts it; * otherwise, *@tg_to_put is set to the throtl_grp to put and the caller is * responsible for putting it. */ static struct bio *throtl_pop_queued(struct list_head *queued, struct throtl_grp **tg_to_put) { struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node); struct bio *bio; if (list_empty(queued)) return NULL; bio = bio_list_pop(&qn->bios); WARN_ON_ONCE(!bio); if (bio_list_empty(&qn->bios)) { list_del_init(&qn->node); if (tg_to_put) *tg_to_put = qn->tg; else blkg_put(tg_to_blkg(qn->tg)); } else { list_move_tail(&qn->node, queued); } return bio; } |
49a2f1e3f blk-throttle: add... |
444 |
/* init a service_queue, assumes the caller zeroed it */ |
b2ce2643c blk-throttle: cle... |
445 |
static void throtl_service_queue_init(struct throtl_service_queue *sq) |
49a2f1e3f blk-throttle: add... |
446 |
{ |
c5cc2070b blk-throttle: add... |
447 448 |
INIT_LIST_HEAD(&sq->queued[0]); INIT_LIST_HEAD(&sq->queued[1]); |
49a2f1e3f blk-throttle: add... |
449 |
sq->pending_tree = RB_ROOT; |
69df0ab03 blk-throttle: sep... |
450 451 452 |
setup_timer(&sq->pending_timer, throtl_pending_timer_fn, (unsigned long)sq); } |
001bea73e blkcg: replace bl... |
453 454 |
static struct blkg_policy_data *throtl_pd_alloc(gfp_t gfp, int node) { |
4fb72036f blk-throttle: rem... |
455 |
struct throtl_grp *tg; |
24bdb8ef0 blkcg: make blkcg... |
456 |
int rw; |
4fb72036f blk-throttle: rem... |
457 458 459 |
tg = kzalloc_node(sizeof(*tg), gfp, node); if (!tg) |
77ea73388 blkcg: move io_se... |
460 |
return NULL; |
4fb72036f blk-throttle: rem... |
461 |
|
b2ce2643c blk-throttle: cle... |
462 463 464 465 466 467 468 469 |
throtl_service_queue_init(&tg->service_queue); for (rw = READ; rw <= WRITE; rw++) { throtl_qnode_init(&tg->qnode_on_self[rw], tg); throtl_qnode_init(&tg->qnode_on_parent[rw], tg); } RB_CLEAR_NODE(&tg->rb_node); |
9f626e372 blk-throttle: pre... |
470 471 472 473 |
tg->bps[READ][LIMIT_MAX] = U64_MAX; tg->bps[WRITE][LIMIT_MAX] = U64_MAX; tg->iops[READ][LIMIT_MAX] = UINT_MAX; tg->iops[WRITE][LIMIT_MAX] = UINT_MAX; |
cd5ab1b0f blk-throttle: add... |
474 475 476 477 478 |
tg->bps_conf[READ][LIMIT_MAX] = U64_MAX; tg->bps_conf[WRITE][LIMIT_MAX] = U64_MAX; tg->iops_conf[READ][LIMIT_MAX] = UINT_MAX; tg->iops_conf[WRITE][LIMIT_MAX] = UINT_MAX; /* LIMIT_LOW will have default value 0 */ |
b2ce2643c blk-throttle: cle... |
479 |
|
ec80991d6 blk-throttle: add... |
480 |
tg->latency_target = DFL_LATENCY_TARGET; |
5b81fc3cc blk-throttle: add... |
481 |
tg->latency_target_conf = DFL_LATENCY_TARGET; |
b4f428ef2 blk-throttle: for... |
482 483 |
tg->idletime_threshold = DFL_IDLE_THRESHOLD; tg->idletime_threshold_conf = DFL_IDLE_THRESHOLD; |
ec80991d6 blk-throttle: add... |
484 |
|
4fb72036f blk-throttle: rem... |
485 |
return &tg->pd; |
001bea73e blkcg: replace bl... |
486 |
} |
a9520cd6f blkcg: make blkcg... |
487 |
static void throtl_pd_init(struct blkg_policy_data *pd) |
a29a171e7 blk-throttle: Do ... |
488 |
{ |
a9520cd6f blkcg: make blkcg... |
489 490 |
struct throtl_grp *tg = pd_to_tg(pd); struct blkcg_gq *blkg = tg_to_blkg(tg); |
77216b048 blk-throttle: add... |
491 |
struct throtl_data *td = blkg->q->td; |
b2ce2643c blk-throttle: cle... |
492 |
struct throtl_service_queue *sq = &tg->service_queue; |
cd1604fab blkcg: factor out... |
493 |
|
9138125be blk-throttle: imp... |
494 |
/* |
aa6ec29be cgroup: remove sa... |
495 |
* If on the default hierarchy, we switch to properly hierarchical |
9138125be blk-throttle: imp... |
496 497 498 499 500 |
* behavior where limits on a given throtl_grp are applied to the * whole subtree rather than just the group itself. e.g. If 16M * read_bps limit is set on the root group, the whole system can't * exceed 16M for the device. * |
aa6ec29be cgroup: remove sa... |
501 |
* If not on the default hierarchy, the broken flat hierarchy |
9138125be blk-throttle: imp... |
502 503 504 505 506 |
* behavior is retained where all throtl_grps are treated as if * they're all separate root groups right below throtl_data. * Limits of a group don't interact with limits of other groups * regardless of the position of the group in the hierarchy. */ |
b2ce2643c blk-throttle: cle... |
507 |
sq->parent_sq = &td->service_queue; |
9e10a130d cgroup: replace c... |
508 |
if (cgroup_subsys_on_dfl(io_cgrp_subsys) && blkg->parent) |
b2ce2643c blk-throttle: cle... |
509 |
sq->parent_sq = &blkg_to_tg(blkg->parent)->service_queue; |
77216b048 blk-throttle: add... |
510 |
tg->td = td; |
8a3d26151 blkcg: move blkio... |
511 |
} |
693e751e7 blk-throttle: imp... |
512 513 514 515 516 517 518 519 |
/* * Set has_rules[] if @tg or any of its parents have limits configured. * This doesn't require walking up to the top of the hierarchy as the * parent's has_rules[] is guaranteed to be correct. */ static void tg_update_has_rules(struct throtl_grp *tg) { struct throtl_grp *parent_tg = sq_to_tg(tg->service_queue.parent_sq); |
9f626e372 blk-throttle: pre... |
520 |
struct throtl_data *td = tg->td; |
693e751e7 blk-throttle: imp... |
521 522 523 524 |
int rw; for (rw = READ; rw <= WRITE; rw++) tg->has_rules[rw] = (parent_tg && parent_tg->has_rules[rw]) || |
9f626e372 blk-throttle: pre... |
525 526 527 |
(td->limit_valid[td->limit_index] && (tg_bps_limit(tg, rw) != U64_MAX || tg_iops_limit(tg, rw) != UINT_MAX)); |
693e751e7 blk-throttle: imp... |
528 |
} |
a9520cd6f blkcg: make blkcg... |
529 |
static void throtl_pd_online(struct blkg_policy_data *pd) |
693e751e7 blk-throttle: imp... |
530 |
{ |
aec242468 blk-throttle: det... |
531 |
struct throtl_grp *tg = pd_to_tg(pd); |
693e751e7 blk-throttle: imp... |
532 533 534 535 |
/* * We don't want new groups to escape the limits of its ancestors. * Update has_rules[] after a new group is brought online. */ |
aec242468 blk-throttle: det... |
536 |
tg_update_has_rules(tg); |
693e751e7 blk-throttle: imp... |
537 |
} |
cd5ab1b0f blk-throttle: add... |
538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 |
static void blk_throtl_update_limit_valid(struct throtl_data *td) { struct cgroup_subsys_state *pos_css; struct blkcg_gq *blkg; bool low_valid = false; rcu_read_lock(); blkg_for_each_descendant_post(blkg, pos_css, td->queue->root_blkg) { struct throtl_grp *tg = blkg_to_tg(blkg); if (tg->bps[READ][LIMIT_LOW] || tg->bps[WRITE][LIMIT_LOW] || tg->iops[READ][LIMIT_LOW] || tg->iops[WRITE][LIMIT_LOW]) low_valid = true; } rcu_read_unlock(); td->limit_valid[LIMIT_LOW] = low_valid; } |
c79892c55 blk-throttle: add... |
556 |
static void throtl_upgrade_state(struct throtl_data *td); |
cd5ab1b0f blk-throttle: add... |
557 558 559 560 561 562 563 564 565 566 |
static void throtl_pd_offline(struct blkg_policy_data *pd) { struct throtl_grp *tg = pd_to_tg(pd); tg->bps[READ][LIMIT_LOW] = 0; tg->bps[WRITE][LIMIT_LOW] = 0; tg->iops[READ][LIMIT_LOW] = 0; tg->iops[WRITE][LIMIT_LOW] = 0; blk_throtl_update_limit_valid(tg->td); |
c79892c55 blk-throttle: add... |
567 568 |
if (!tg->td->limit_valid[tg->td->limit_index]) throtl_upgrade_state(tg->td); |
cd5ab1b0f blk-throttle: add... |
569 |
} |
001bea73e blkcg: replace bl... |
570 571 |
static void throtl_pd_free(struct blkg_policy_data *pd) { |
4fb72036f blk-throttle: rem... |
572 |
struct throtl_grp *tg = pd_to_tg(pd); |
b2ce2643c blk-throttle: cle... |
573 |
del_timer_sync(&tg->service_queue.pending_timer); |
4fb72036f blk-throttle: rem... |
574 |
kfree(tg); |
001bea73e blkcg: replace bl... |
575 |
} |
0049af73b blk-throttle: reo... |
576 577 |
static struct throtl_grp * throtl_rb_first(struct throtl_service_queue *parent_sq) |
e43473b7f blkio: Core imple... |
578 579 |
{ /* Service tree is empty */ |
0049af73b blk-throttle: reo... |
580 |
if (!parent_sq->nr_pending) |
e43473b7f blkio: Core imple... |
581 |
return NULL; |
0049af73b blk-throttle: reo... |
582 583 |
if (!parent_sq->first_pending) parent_sq->first_pending = rb_first(&parent_sq->pending_tree); |
e43473b7f blkio: Core imple... |
584 |
|
0049af73b blk-throttle: reo... |
585 586 |
if (parent_sq->first_pending) return rb_entry_tg(parent_sq->first_pending); |
e43473b7f blkio: Core imple... |
587 588 589 590 591 592 593 594 595 |
return NULL; } static void rb_erase_init(struct rb_node *n, struct rb_root *root) { rb_erase(n, root); RB_CLEAR_NODE(n); } |
0049af73b blk-throttle: reo... |
596 597 |
static void throtl_rb_erase(struct rb_node *n, struct throtl_service_queue *parent_sq) |
e43473b7f blkio: Core imple... |
598 |
{ |
0049af73b blk-throttle: reo... |
599 600 601 602 |
if (parent_sq->first_pending == n) parent_sq->first_pending = NULL; rb_erase_init(n, &parent_sq->pending_tree); --parent_sq->nr_pending; |
e43473b7f blkio: Core imple... |
603 |
} |
0049af73b blk-throttle: reo... |
604 |
static void update_min_dispatch_time(struct throtl_service_queue *parent_sq) |
e43473b7f blkio: Core imple... |
605 606 |
{ struct throtl_grp *tg; |
0049af73b blk-throttle: reo... |
607 |
tg = throtl_rb_first(parent_sq); |
e43473b7f blkio: Core imple... |
608 609 |
if (!tg) return; |
0049af73b blk-throttle: reo... |
610 |
parent_sq->first_pending_disptime = tg->disptime; |
e43473b7f blkio: Core imple... |
611 |
} |
77216b048 blk-throttle: add... |
612 |
static void tg_service_queue_add(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
613 |
{ |
77216b048 blk-throttle: add... |
614 |
struct throtl_service_queue *parent_sq = tg->service_queue.parent_sq; |
0049af73b blk-throttle: reo... |
615 |
struct rb_node **node = &parent_sq->pending_tree.rb_node; |
e43473b7f blkio: Core imple... |
616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 |
struct rb_node *parent = NULL; struct throtl_grp *__tg; unsigned long key = tg->disptime; int left = 1; while (*node != NULL) { parent = *node; __tg = rb_entry_tg(parent); if (time_before(key, __tg->disptime)) node = &parent->rb_left; else { node = &parent->rb_right; left = 0; } } if (left) |
0049af73b blk-throttle: reo... |
634 |
parent_sq->first_pending = &tg->rb_node; |
e43473b7f blkio: Core imple... |
635 636 |
rb_link_node(&tg->rb_node, parent, node); |
0049af73b blk-throttle: reo... |
637 |
rb_insert_color(&tg->rb_node, &parent_sq->pending_tree); |
e43473b7f blkio: Core imple... |
638 |
} |
77216b048 blk-throttle: add... |
639 |
static void __throtl_enqueue_tg(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
640 |
{ |
77216b048 blk-throttle: add... |
641 |
tg_service_queue_add(tg); |
5b2c16aae blk-throttle: sim... |
642 |
tg->flags |= THROTL_TG_PENDING; |
77216b048 blk-throttle: add... |
643 |
tg->service_queue.parent_sq->nr_pending++; |
e43473b7f blkio: Core imple... |
644 |
} |
77216b048 blk-throttle: add... |
645 |
static void throtl_enqueue_tg(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
646 |
{ |
5b2c16aae blk-throttle: sim... |
647 |
if (!(tg->flags & THROTL_TG_PENDING)) |
77216b048 blk-throttle: add... |
648 |
__throtl_enqueue_tg(tg); |
e43473b7f blkio: Core imple... |
649 |
} |
77216b048 blk-throttle: add... |
650 |
static void __throtl_dequeue_tg(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
651 |
{ |
77216b048 blk-throttle: add... |
652 |
throtl_rb_erase(&tg->rb_node, tg->service_queue.parent_sq); |
5b2c16aae blk-throttle: sim... |
653 |
tg->flags &= ~THROTL_TG_PENDING; |
e43473b7f blkio: Core imple... |
654 |
} |
77216b048 blk-throttle: add... |
655 |
static void throtl_dequeue_tg(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
656 |
{ |
5b2c16aae blk-throttle: sim... |
657 |
if (tg->flags & THROTL_TG_PENDING) |
77216b048 blk-throttle: add... |
658 |
__throtl_dequeue_tg(tg); |
e43473b7f blkio: Core imple... |
659 |
} |
a9131a27e blk-throttle: rel... |
660 |
/* Call with queue lock held */ |
69df0ab03 blk-throttle: sep... |
661 662 |
static void throtl_schedule_pending_timer(struct throtl_service_queue *sq, unsigned long expires) |
a9131a27e blk-throttle: rel... |
663 |
{ |
a41b816c1 blk-throttle: fix... |
664 |
unsigned long max_expire = jiffies + 8 * sq_to_td(sq)->throtl_slice; |
06cceedcc blk-throttle: mak... |
665 666 667 668 669 670 671 672 673 674 |
/* * Since we are adjusting the throttle limit dynamically, the sleep * time calculated according to previous limit might be invalid. It's * possible the cgroup sleep time is very long and no other cgroups * have IO running so notify the limit changes. Make sure the cgroup * doesn't sleep too long to avoid the missed notification. */ if (time_after(expires, max_expire)) expires = max_expire; |
69df0ab03 blk-throttle: sep... |
675 676 677 |
mod_timer(&sq->pending_timer, expires); throtl_log(sq, "schedule timer. delay=%lu jiffies=%lu", expires - jiffies, jiffies); |
a9131a27e blk-throttle: rel... |
678 |
} |
7f52f98c2 blk-throttle: imp... |
679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 |
/** * throtl_schedule_next_dispatch - schedule the next dispatch cycle * @sq: the service_queue to schedule dispatch for * @force: force scheduling * * Arm @sq->pending_timer so that the next dispatch cycle starts on the * dispatch time of the first pending child. Returns %true if either timer * is armed or there's no pending child left. %false if the current * dispatch window is still open and the caller should continue * dispatching. * * If @force is %true, the dispatch timer is always scheduled and this * function is guaranteed to return %true. This is to be used when the * caller can't dispatch itself and needs to invoke pending_timer * unconditionally. Note that forced scheduling is likely to induce short * delay before dispatch starts even if @sq->first_pending_disptime is not * in the future and thus shouldn't be used in hot paths. */ static bool throtl_schedule_next_dispatch(struct throtl_service_queue *sq, bool force) |
e43473b7f blkio: Core imple... |
699 |
{ |
6a525600f blk-throttle: rem... |
700 |
/* any pending children left? */ |
c9e0332e8 blk-throttle: ren... |
701 |
if (!sq->nr_pending) |
7f52f98c2 blk-throttle: imp... |
702 |
return true; |
e43473b7f blkio: Core imple... |
703 |
|
c9e0332e8 blk-throttle: ren... |
704 |
update_min_dispatch_time(sq); |
e43473b7f blkio: Core imple... |
705 |
|
69df0ab03 blk-throttle: sep... |
706 |
/* is the next dispatch time in the future? */ |
7f52f98c2 blk-throttle: imp... |
707 |
if (force || time_after(sq->first_pending_disptime, jiffies)) { |
69df0ab03 blk-throttle: sep... |
708 |
throtl_schedule_pending_timer(sq, sq->first_pending_disptime); |
7f52f98c2 blk-throttle: imp... |
709 |
return true; |
69df0ab03 blk-throttle: sep... |
710 |
} |
7f52f98c2 blk-throttle: imp... |
711 712 |
/* tell the caller to continue dispatching */ return false; |
e43473b7f blkio: Core imple... |
713 |
} |
32ee5bc47 blk-throttle: Acc... |
714 715 716 717 718 719 720 721 722 723 724 725 726 727 |
static inline void throtl_start_new_slice_with_credit(struct throtl_grp *tg, bool rw, unsigned long start) { tg->bytes_disp[rw] = 0; tg->io_disp[rw] = 0; /* * Previous slice has expired. We must have trimmed it after last * bio dispatch. That means since start of last slice, we never used * that bandwidth. Do try to make use of that bandwidth while giving * credit. */ if (time_after_eq(start, tg->slice_start[rw])) tg->slice_start[rw] = start; |
297e3d854 blk-throttle: mak... |
728 |
tg->slice_end[rw] = jiffies + tg->td->throtl_slice; |
32ee5bc47 blk-throttle: Acc... |
729 730 731 732 733 |
throtl_log(&tg->service_queue, "[%c] new slice with credit start=%lu end=%lu jiffies=%lu", rw == READ ? 'R' : 'W', tg->slice_start[rw], tg->slice_end[rw], jiffies); } |
0f3457f60 blk-throttle: add... |
734 |
static inline void throtl_start_new_slice(struct throtl_grp *tg, bool rw) |
e43473b7f blkio: Core imple... |
735 736 |
{ tg->bytes_disp[rw] = 0; |
8e89d13f4 blkio: Implementa... |
737 |
tg->io_disp[rw] = 0; |
e43473b7f blkio: Core imple... |
738 |
tg->slice_start[rw] = jiffies; |
297e3d854 blk-throttle: mak... |
739 |
tg->slice_end[rw] = jiffies + tg->td->throtl_slice; |
fda6f272c blk-throttle: imp... |
740 741 742 743 |
throtl_log(&tg->service_queue, "[%c] new slice start=%lu end=%lu jiffies=%lu", rw == READ ? 'R' : 'W', tg->slice_start[rw], tg->slice_end[rw], jiffies); |
e43473b7f blkio: Core imple... |
744 |
} |
0f3457f60 blk-throttle: add... |
745 746 |
static inline void throtl_set_slice_end(struct throtl_grp *tg, bool rw, unsigned long jiffy_end) |
d1ae8ffdf blk-throttle: Tri... |
747 |
{ |
297e3d854 blk-throttle: mak... |
748 |
tg->slice_end[rw] = roundup(jiffy_end, tg->td->throtl_slice); |
d1ae8ffdf blk-throttle: Tri... |
749 |
} |
0f3457f60 blk-throttle: add... |
750 751 |
static inline void throtl_extend_slice(struct throtl_grp *tg, bool rw, unsigned long jiffy_end) |
e43473b7f blkio: Core imple... |
752 |
{ |
297e3d854 blk-throttle: mak... |
753 |
tg->slice_end[rw] = roundup(jiffy_end, tg->td->throtl_slice); |
fda6f272c blk-throttle: imp... |
754 755 756 757 |
throtl_log(&tg->service_queue, "[%c] extend slice start=%lu end=%lu jiffies=%lu", rw == READ ? 'R' : 'W', tg->slice_start[rw], tg->slice_end[rw], jiffies); |
e43473b7f blkio: Core imple... |
758 759 760 |
} /* Determine if previously allocated or extended slice is complete or not */ |
0f3457f60 blk-throttle: add... |
761 |
static bool throtl_slice_used(struct throtl_grp *tg, bool rw) |
e43473b7f blkio: Core imple... |
762 763 |
{ if (time_in_range(jiffies, tg->slice_start[rw], tg->slice_end[rw])) |
5cf8c2277 block/blk-throttl... |
764 |
return false; |
e43473b7f blkio: Core imple... |
765 766 767 768 769 |
return 1; } /* Trim the used slices and adjust slice start accordingly */ |
0f3457f60 blk-throttle: add... |
770 |
static inline void throtl_trim_slice(struct throtl_grp *tg, bool rw) |
e43473b7f blkio: Core imple... |
771 |
{ |
3aad5d3ee blkio-throttle: F... |
772 773 |
unsigned long nr_slices, time_elapsed, io_trim; u64 bytes_trim, tmp; |
e43473b7f blkio: Core imple... |
774 775 776 777 778 779 780 781 |
BUG_ON(time_before(tg->slice_end[rw], tg->slice_start[rw])); /* * If bps are unlimited (-1), then time slice don't get * renewed. Don't try to trim the slice if slice is used. A new * slice will start when appropriate. */ |
0f3457f60 blk-throttle: add... |
782 |
if (throtl_slice_used(tg, rw)) |
e43473b7f blkio: Core imple... |
783 |
return; |
d1ae8ffdf blk-throttle: Tri... |
784 785 786 787 788 789 790 |
/* * A bio has been dispatched. Also adjust slice_end. It might happen * that initially cgroup limit was very low resulting in high * slice_end, but later limit was bumped up and bio was dispached * sooner, then we need to reduce slice_end. A high bogus slice_end * is bad because it does not allow new slice to start. */ |
297e3d854 blk-throttle: mak... |
791 |
throtl_set_slice_end(tg, rw, jiffies + tg->td->throtl_slice); |
d1ae8ffdf blk-throttle: Tri... |
792 |
|
e43473b7f blkio: Core imple... |
793 |
time_elapsed = jiffies - tg->slice_start[rw]; |
297e3d854 blk-throttle: mak... |
794 |
nr_slices = time_elapsed / tg->td->throtl_slice; |
e43473b7f blkio: Core imple... |
795 796 797 |
if (!nr_slices) return; |
297e3d854 blk-throttle: mak... |
798 |
tmp = tg_bps_limit(tg, rw) * tg->td->throtl_slice * nr_slices; |
3aad5d3ee blkio-throttle: F... |
799 800 |
do_div(tmp, HZ); bytes_trim = tmp; |
e43473b7f blkio: Core imple... |
801 |
|
297e3d854 blk-throttle: mak... |
802 803 |
io_trim = (tg_iops_limit(tg, rw) * tg->td->throtl_slice * nr_slices) / HZ; |
e43473b7f blkio: Core imple... |
804 |
|
8e89d13f4 blkio: Implementa... |
805 |
if (!bytes_trim && !io_trim) |
e43473b7f blkio: Core imple... |
806 807 808 809 810 811 |
return; if (tg->bytes_disp[rw] >= bytes_trim) tg->bytes_disp[rw] -= bytes_trim; else tg->bytes_disp[rw] = 0; |
8e89d13f4 blkio: Implementa... |
812 813 814 815 |
if (tg->io_disp[rw] >= io_trim) tg->io_disp[rw] -= io_trim; else tg->io_disp[rw] = 0; |
297e3d854 blk-throttle: mak... |
816 |
tg->slice_start[rw] += nr_slices * tg->td->throtl_slice; |
e43473b7f blkio: Core imple... |
817 |
|
fda6f272c blk-throttle: imp... |
818 819 820 821 |
throtl_log(&tg->service_queue, "[%c] trim slice nr=%lu bytes=%llu io=%lu start=%lu end=%lu jiffies=%lu", rw == READ ? 'R' : 'W', nr_slices, bytes_trim, io_trim, tg->slice_start[rw], tg->slice_end[rw], jiffies); |
e43473b7f blkio: Core imple... |
822 |
} |
0f3457f60 blk-throttle: add... |
823 824 |
static bool tg_with_in_iops_limit(struct throtl_grp *tg, struct bio *bio, unsigned long *wait) |
e43473b7f blkio: Core imple... |
825 826 |
{ bool rw = bio_data_dir(bio); |
8e89d13f4 blkio: Implementa... |
827 |
unsigned int io_allowed; |
e43473b7f blkio: Core imple... |
828 |
unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd; |
c49c06e49 blkio-throttle: F... |
829 |
u64 tmp; |
e43473b7f blkio: Core imple... |
830 |
|
8e89d13f4 blkio: Implementa... |
831 |
jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw]; |
e43473b7f blkio: Core imple... |
832 |
|
8e89d13f4 blkio: Implementa... |
833 834 |
/* Slice has just started. Consider one slice interval */ if (!jiffy_elapsed) |
297e3d854 blk-throttle: mak... |
835 |
jiffy_elapsed_rnd = tg->td->throtl_slice; |
8e89d13f4 blkio: Implementa... |
836 |
|
297e3d854 blk-throttle: mak... |
837 |
jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, tg->td->throtl_slice); |
8e89d13f4 blkio: Implementa... |
838 |
|
c49c06e49 blkio-throttle: F... |
839 840 841 842 843 844 |
/* * jiffy_elapsed_rnd should not be a big value as minimum iops can be * 1 then at max jiffy elapsed should be equivalent of 1 second as we * will allow dispatch after 1 second and after that slice should * have been trimmed. */ |
9f626e372 blk-throttle: pre... |
845 |
tmp = (u64)tg_iops_limit(tg, rw) * jiffy_elapsed_rnd; |
c49c06e49 blkio-throttle: F... |
846 847 848 849 850 851 |
do_div(tmp, HZ); if (tmp > UINT_MAX) io_allowed = UINT_MAX; else io_allowed = tmp; |
8e89d13f4 blkio: Implementa... |
852 853 |
if (tg->io_disp[rw] + 1 <= io_allowed) { |
e43473b7f blkio: Core imple... |
854 855 |
if (wait) *wait = 0; |
5cf8c2277 block/blk-throttl... |
856 |
return true; |
e43473b7f blkio: Core imple... |
857 |
} |
8e89d13f4 blkio: Implementa... |
858 |
/* Calc approx time to dispatch */ |
9f626e372 blk-throttle: pre... |
859 |
jiffy_wait = ((tg->io_disp[rw] + 1) * HZ) / tg_iops_limit(tg, rw) + 1; |
8e89d13f4 blkio: Implementa... |
860 861 862 863 864 865 866 867 868 869 |
if (jiffy_wait > jiffy_elapsed) jiffy_wait = jiffy_wait - jiffy_elapsed; else jiffy_wait = 1; if (wait) *wait = jiffy_wait; return 0; } |
0f3457f60 blk-throttle: add... |
870 871 |
static bool tg_with_in_bps_limit(struct throtl_grp *tg, struct bio *bio, unsigned long *wait) |
8e89d13f4 blkio: Implementa... |
872 873 |
{ bool rw = bio_data_dir(bio); |
3aad5d3ee blkio-throttle: F... |
874 |
u64 bytes_allowed, extra_bytes, tmp; |
8e89d13f4 blkio: Implementa... |
875 |
unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd; |
ea0ea2bc6 blk-throttle: cap... |
876 |
unsigned int bio_size = throtl_bio_data_size(bio); |
e43473b7f blkio: Core imple... |
877 878 879 880 881 |
jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw]; /* Slice has just started. Consider one slice interval */ if (!jiffy_elapsed) |
297e3d854 blk-throttle: mak... |
882 |
jiffy_elapsed_rnd = tg->td->throtl_slice; |
e43473b7f blkio: Core imple... |
883 |
|
297e3d854 blk-throttle: mak... |
884 |
jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, tg->td->throtl_slice); |
e43473b7f blkio: Core imple... |
885 |
|
9f626e372 blk-throttle: pre... |
886 |
tmp = tg_bps_limit(tg, rw) * jiffy_elapsed_rnd; |
5e901a2b9 blkio-throttle: T... |
887 |
do_div(tmp, HZ); |
3aad5d3ee blkio-throttle: F... |
888 |
bytes_allowed = tmp; |
e43473b7f blkio: Core imple... |
889 |
|
ea0ea2bc6 blk-throttle: cap... |
890 |
if (tg->bytes_disp[rw] + bio_size <= bytes_allowed) { |
e43473b7f blkio: Core imple... |
891 892 |
if (wait) *wait = 0; |
5cf8c2277 block/blk-throttl... |
893 |
return true; |
e43473b7f blkio: Core imple... |
894 895 896 |
} /* Calc approx time to dispatch */ |
ea0ea2bc6 blk-throttle: cap... |
897 |
extra_bytes = tg->bytes_disp[rw] + bio_size - bytes_allowed; |
9f626e372 blk-throttle: pre... |
898 |
jiffy_wait = div64_u64(extra_bytes * HZ, tg_bps_limit(tg, rw)); |
e43473b7f blkio: Core imple... |
899 900 901 902 903 904 905 906 907 |
if (!jiffy_wait) jiffy_wait = 1; /* * This wait time is without taking into consideration the rounding * up we did. Add that time also. */ jiffy_wait = jiffy_wait + (jiffy_elapsed_rnd - jiffy_elapsed); |
e43473b7f blkio: Core imple... |
908 909 |
if (wait) *wait = jiffy_wait; |
8e89d13f4 blkio: Implementa... |
910 911 912 913 914 915 916 |
return 0; } /* * Returns whether one can dispatch a bio or not. Also returns approx number * of jiffies to wait before this bio is with-in IO rate and can be dispatched */ |
0f3457f60 blk-throttle: add... |
917 918 |
static bool tg_may_dispatch(struct throtl_grp *tg, struct bio *bio, unsigned long *wait) |
8e89d13f4 blkio: Implementa... |
919 920 921 922 923 924 925 926 927 928 |
{ bool rw = bio_data_dir(bio); unsigned long bps_wait = 0, iops_wait = 0, max_wait = 0; /* * Currently whole state machine of group depends on first bio * queued in the group bio list. So one should not be calling * this function with a different bio if there are other bios * queued. */ |
73f0d49a9 blk-throttle: mov... |
929 |
BUG_ON(tg->service_queue.nr_queued[rw] && |
c5cc2070b blk-throttle: add... |
930 |
bio != throtl_peek_queued(&tg->service_queue.queued[rw])); |
e43473b7f blkio: Core imple... |
931 |
|
8e89d13f4 blkio: Implementa... |
932 |
/* If tg->bps = -1, then BW is unlimited */ |
9f626e372 blk-throttle: pre... |
933 934 |
if (tg_bps_limit(tg, rw) == U64_MAX && tg_iops_limit(tg, rw) == UINT_MAX) { |
8e89d13f4 blkio: Implementa... |
935 936 |
if (wait) *wait = 0; |
5cf8c2277 block/blk-throttl... |
937 |
return true; |
8e89d13f4 blkio: Implementa... |
938 939 940 941 942 |
} /* * If previous slice expired, start a new one otherwise renew/extend * existing slice to make sure it is at least throtl_slice interval |
164c80ed8 blk-throttle: Ext... |
943 944 945 |
* long since now. New slice is started only for empty throttle group. * If there is queued bio, that means there should be an active * slice and it should be extended instead. |
8e89d13f4 blkio: Implementa... |
946 |
*/ |
164c80ed8 blk-throttle: Ext... |
947 |
if (throtl_slice_used(tg, rw) && !(tg->service_queue.nr_queued[rw])) |
0f3457f60 blk-throttle: add... |
948 |
throtl_start_new_slice(tg, rw); |
8e89d13f4 blkio: Implementa... |
949 |
else { |
297e3d854 blk-throttle: mak... |
950 951 952 953 |
if (time_before(tg->slice_end[rw], jiffies + tg->td->throtl_slice)) throtl_extend_slice(tg, rw, jiffies + tg->td->throtl_slice); |
8e89d13f4 blkio: Implementa... |
954 |
} |
0f3457f60 blk-throttle: add... |
955 956 |
if (tg_with_in_bps_limit(tg, bio, &bps_wait) && tg_with_in_iops_limit(tg, bio, &iops_wait)) { |
8e89d13f4 blkio: Implementa... |
957 958 959 960 961 962 963 964 965 966 967 |
if (wait) *wait = 0; return 1; } max_wait = max(bps_wait, iops_wait); if (wait) *wait = max_wait; if (time_before(tg->slice_end[rw], jiffies + max_wait)) |
0f3457f60 blk-throttle: add... |
968 |
throtl_extend_slice(tg, rw, jiffies + max_wait); |
e43473b7f blkio: Core imple... |
969 970 971 972 973 974 975 |
return 0; } static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio) { bool rw = bio_data_dir(bio); |
ea0ea2bc6 blk-throttle: cap... |
976 |
unsigned int bio_size = throtl_bio_data_size(bio); |
e43473b7f blkio: Core imple... |
977 978 |
/* Charge the bio to the group */ |
ea0ea2bc6 blk-throttle: cap... |
979 |
tg->bytes_disp[rw] += bio_size; |
8e89d13f4 blkio: Implementa... |
980 |
tg->io_disp[rw]++; |
ea0ea2bc6 blk-throttle: cap... |
981 |
tg->last_bytes_disp[rw] += bio_size; |
3f0abd806 blk-throttle: add... |
982 |
tg->last_io_disp[rw]++; |
e43473b7f blkio: Core imple... |
983 |
|
2a0f61e6e blk-throttle: set... |
984 |
/* |
8d2bbd4c8 block: replace RE... |
985 |
* BIO_THROTTLED is used to prevent the same bio to be throttled |
2a0f61e6e blk-throttle: set... |
986 987 988 |
* more than once as a throttled bio will go through blk-throtl the * second time when it eventually gets issued. Set it when a bio * is being charged to a tg. |
2a0f61e6e blk-throttle: set... |
989 |
*/ |
8d2bbd4c8 block: replace RE... |
990 991 |
if (!bio_flagged(bio, BIO_THROTTLED)) bio_set_flag(bio, BIO_THROTTLED); |
e43473b7f blkio: Core imple... |
992 |
} |
c5cc2070b blk-throttle: add... |
993 994 995 996 997 998 999 1000 1001 1002 1003 |
/** * throtl_add_bio_tg - add a bio to the specified throtl_grp * @bio: bio to add * @qn: qnode to use * @tg: the target throtl_grp * * Add @bio to @tg's service_queue using @qn. If @qn is not specified, * tg->qnode_on_self[] is used. */ static void throtl_add_bio_tg(struct bio *bio, struct throtl_qnode *qn, struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
1004 |
{ |
73f0d49a9 blk-throttle: mov... |
1005 |
struct throtl_service_queue *sq = &tg->service_queue; |
e43473b7f blkio: Core imple... |
1006 |
bool rw = bio_data_dir(bio); |
c5cc2070b blk-throttle: add... |
1007 1008 |
if (!qn) qn = &tg->qnode_on_self[rw]; |
0e9f4164b blk-throttle: gen... |
1009 1010 1011 1012 1013 1014 1015 1016 |
/* * If @tg doesn't currently have any bios queued in the same * direction, queueing @bio can change when @tg should be * dispatched. Mark that @tg was empty. This is automatically * cleaered on the next tg_update_disptime(). */ if (!sq->nr_queued[rw]) tg->flags |= THROTL_TG_WAS_EMPTY; |
c5cc2070b blk-throttle: add... |
1017 |
throtl_qnode_add_bio(bio, qn, &sq->queued[rw]); |
73f0d49a9 blk-throttle: mov... |
1018 |
sq->nr_queued[rw]++; |
77216b048 blk-throttle: add... |
1019 |
throtl_enqueue_tg(tg); |
e43473b7f blkio: Core imple... |
1020 |
} |
77216b048 blk-throttle: add... |
1021 |
static void tg_update_disptime(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
1022 |
{ |
73f0d49a9 blk-throttle: mov... |
1023 |
struct throtl_service_queue *sq = &tg->service_queue; |
e43473b7f blkio: Core imple... |
1024 1025 |
unsigned long read_wait = -1, write_wait = -1, min_wait = -1, disptime; struct bio *bio; |
d609af3a1 blk-throttle: Adj... |
1026 1027 |
bio = throtl_peek_queued(&sq->queued[READ]); if (bio) |
0f3457f60 blk-throttle: add... |
1028 |
tg_may_dispatch(tg, bio, &read_wait); |
e43473b7f blkio: Core imple... |
1029 |
|
d609af3a1 blk-throttle: Adj... |
1030 1031 |
bio = throtl_peek_queued(&sq->queued[WRITE]); if (bio) |
0f3457f60 blk-throttle: add... |
1032 |
tg_may_dispatch(tg, bio, &write_wait); |
e43473b7f blkio: Core imple... |
1033 1034 1035 |
min_wait = min(read_wait, write_wait); disptime = jiffies + min_wait; |
e43473b7f blkio: Core imple... |
1036 |
/* Update dispatch time */ |
77216b048 blk-throttle: add... |
1037 |
throtl_dequeue_tg(tg); |
e43473b7f blkio: Core imple... |
1038 |
tg->disptime = disptime; |
77216b048 blk-throttle: add... |
1039 |
throtl_enqueue_tg(tg); |
0e9f4164b blk-throttle: gen... |
1040 1041 1042 |
/* see throtl_add_bio_tg() */ tg->flags &= ~THROTL_TG_WAS_EMPTY; |
e43473b7f blkio: Core imple... |
1043 |
} |
32ee5bc47 blk-throttle: Acc... |
1044 1045 1046 1047 1048 1049 1050 1051 1052 |
static void start_parent_slice_with_credit(struct throtl_grp *child_tg, struct throtl_grp *parent_tg, bool rw) { if (throtl_slice_used(parent_tg, rw)) { throtl_start_new_slice_with_credit(parent_tg, rw, child_tg->slice_start[rw]); } } |
77216b048 blk-throttle: add... |
1053 |
static void tg_dispatch_one_bio(struct throtl_grp *tg, bool rw) |
e43473b7f blkio: Core imple... |
1054 |
{ |
73f0d49a9 blk-throttle: mov... |
1055 |
struct throtl_service_queue *sq = &tg->service_queue; |
6bc9c2b46 blk-throttle: mak... |
1056 1057 |
struct throtl_service_queue *parent_sq = sq->parent_sq; struct throtl_grp *parent_tg = sq_to_tg(parent_sq); |
c5cc2070b blk-throttle: add... |
1058 |
struct throtl_grp *tg_to_put = NULL; |
e43473b7f blkio: Core imple... |
1059 |
struct bio *bio; |
c5cc2070b blk-throttle: add... |
1060 1061 1062 1063 1064 1065 1066 |
/* * @bio is being transferred from @tg to @parent_sq. Popping a bio * from @tg may put its reference and @parent_sq might end up * getting released prematurely. Remember the tg to put and put it * after @bio is transferred to @parent_sq. */ bio = throtl_pop_queued(&sq->queued[rw], &tg_to_put); |
73f0d49a9 blk-throttle: mov... |
1067 |
sq->nr_queued[rw]--; |
e43473b7f blkio: Core imple... |
1068 1069 |
throtl_charge_bio(tg, bio); |
6bc9c2b46 blk-throttle: mak... |
1070 1071 1072 1073 1074 1075 1076 1077 1078 |
/* * If our parent is another tg, we just need to transfer @bio to * the parent using throtl_add_bio_tg(). If our parent is * @td->service_queue, @bio is ready to be issued. Put it on its * bio_lists[] and decrease total number queued. The caller is * responsible for issuing these bios. */ if (parent_tg) { |
c5cc2070b blk-throttle: add... |
1079 |
throtl_add_bio_tg(bio, &tg->qnode_on_parent[rw], parent_tg); |
32ee5bc47 blk-throttle: Acc... |
1080 |
start_parent_slice_with_credit(tg, parent_tg, rw); |
6bc9c2b46 blk-throttle: mak... |
1081 |
} else { |
c5cc2070b blk-throttle: add... |
1082 1083 |
throtl_qnode_add_bio(bio, &tg->qnode_on_parent[rw], &parent_sq->queued[rw]); |
6bc9c2b46 blk-throttle: mak... |
1084 1085 1086 |
BUG_ON(tg->td->nr_queued[rw] <= 0); tg->td->nr_queued[rw]--; } |
e43473b7f blkio: Core imple... |
1087 |
|
0f3457f60 blk-throttle: add... |
1088 |
throtl_trim_slice(tg, rw); |
6bc9c2b46 blk-throttle: mak... |
1089 |
|
c5cc2070b blk-throttle: add... |
1090 1091 |
if (tg_to_put) blkg_put(tg_to_blkg(tg_to_put)); |
e43473b7f blkio: Core imple... |
1092 |
} |
77216b048 blk-throttle: add... |
1093 |
static int throtl_dispatch_tg(struct throtl_grp *tg) |
e43473b7f blkio: Core imple... |
1094 |
{ |
73f0d49a9 blk-throttle: mov... |
1095 |
struct throtl_service_queue *sq = &tg->service_queue; |
e43473b7f blkio: Core imple... |
1096 1097 |
unsigned int nr_reads = 0, nr_writes = 0; unsigned int max_nr_reads = throtl_grp_quantum*3/4; |
c2f6805d4 blk-throttle: Fix... |
1098 |
unsigned int max_nr_writes = throtl_grp_quantum - max_nr_reads; |
e43473b7f blkio: Core imple... |
1099 1100 1101 |
struct bio *bio; /* Try to dispatch 75% READS and 25% WRITES */ |
c5cc2070b blk-throttle: add... |
1102 |
while ((bio = throtl_peek_queued(&sq->queued[READ])) && |
0f3457f60 blk-throttle: add... |
1103 |
tg_may_dispatch(tg, bio, NULL)) { |
e43473b7f blkio: Core imple... |
1104 |
|
77216b048 blk-throttle: add... |
1105 |
tg_dispatch_one_bio(tg, bio_data_dir(bio)); |
e43473b7f blkio: Core imple... |
1106 1107 1108 1109 1110 |
nr_reads++; if (nr_reads >= max_nr_reads) break; } |
c5cc2070b blk-throttle: add... |
1111 |
while ((bio = throtl_peek_queued(&sq->queued[WRITE])) && |
0f3457f60 blk-throttle: add... |
1112 |
tg_may_dispatch(tg, bio, NULL)) { |
e43473b7f blkio: Core imple... |
1113 |
|
77216b048 blk-throttle: add... |
1114 |
tg_dispatch_one_bio(tg, bio_data_dir(bio)); |
e43473b7f blkio: Core imple... |
1115 1116 1117 1118 1119 1120 1121 1122 |
nr_writes++; if (nr_writes >= max_nr_writes) break; } return nr_reads + nr_writes; } |
651930bc1 blk-throttle: dis... |
1123 |
static int throtl_select_dispatch(struct throtl_service_queue *parent_sq) |
e43473b7f blkio: Core imple... |
1124 1125 |
{ unsigned int nr_disp = 0; |
e43473b7f blkio: Core imple... |
1126 1127 |
while (1) { |
73f0d49a9 blk-throttle: mov... |
1128 1129 |
struct throtl_grp *tg = throtl_rb_first(parent_sq); struct throtl_service_queue *sq = &tg->service_queue; |
e43473b7f blkio: Core imple... |
1130 1131 1132 1133 1134 1135 |
if (!tg) break; if (time_before(jiffies, tg->disptime)) break; |
77216b048 blk-throttle: add... |
1136 |
throtl_dequeue_tg(tg); |
e43473b7f blkio: Core imple... |
1137 |
|
77216b048 blk-throttle: add... |
1138 |
nr_disp += throtl_dispatch_tg(tg); |
e43473b7f blkio: Core imple... |
1139 |
|
73f0d49a9 blk-throttle: mov... |
1140 |
if (sq->nr_queued[0] || sq->nr_queued[1]) |
77216b048 blk-throttle: add... |
1141 |
tg_update_disptime(tg); |
e43473b7f blkio: Core imple... |
1142 1143 1144 1145 1146 1147 1148 |
if (nr_disp >= throtl_quantum) break; } return nr_disp; } |
c79892c55 blk-throttle: add... |
1149 1150 |
static bool throtl_can_upgrade(struct throtl_data *td, struct throtl_grp *this_tg); |
6e1a5704c blk-throttle: dis... |
1151 1152 1153 1154 1155 1156 1157 |
/** * throtl_pending_timer_fn - timer function for service_queue->pending_timer * @arg: the throtl_service_queue being serviced * * This timer is armed when a child throtl_grp with active bio's become * pending and queued on the service_queue's pending_tree and expires when * the first child throtl_grp should be dispatched. This function |
2e48a530a blk-throttle: mak... |
1158 1159 1160 1161 1162 1163 1164 |
* dispatches bio's from the children throtl_grps to the parent * service_queue. * * If the parent's parent is another throtl_grp, dispatching is propagated * by either arming its pending_timer or repeating dispatch directly. If * the top-level service_tree is reached, throtl_data->dispatch_work is * kicked so that the ready bio's are issued. |
6e1a5704c blk-throttle: dis... |
1165 |
*/ |
69df0ab03 blk-throttle: sep... |
1166 1167 1168 |
static void throtl_pending_timer_fn(unsigned long arg) { struct throtl_service_queue *sq = (void *)arg; |
2e48a530a blk-throttle: mak... |
1169 |
struct throtl_grp *tg = sq_to_tg(sq); |
69df0ab03 blk-throttle: sep... |
1170 |
struct throtl_data *td = sq_to_td(sq); |
cb76199c3 blk-throttle: col... |
1171 |
struct request_queue *q = td->queue; |
2e48a530a blk-throttle: mak... |
1172 1173 |
struct throtl_service_queue *parent_sq; bool dispatched; |
6e1a5704c blk-throttle: dis... |
1174 |
int ret; |
e43473b7f blkio: Core imple... |
1175 1176 |
spin_lock_irq(q->queue_lock); |
c79892c55 blk-throttle: add... |
1177 1178 |
if (throtl_can_upgrade(td, NULL)) throtl_upgrade_state(td); |
2e48a530a blk-throttle: mak... |
1179 1180 1181 |
again: parent_sq = sq->parent_sq; dispatched = false; |
e43473b7f blkio: Core imple... |
1182 |
|
7f52f98c2 blk-throttle: imp... |
1183 1184 |
while (true) { throtl_log(sq, "dispatch nr_queued=%u read=%u write=%u", |
2e48a530a blk-throttle: mak... |
1185 1186 |
sq->nr_queued[READ] + sq->nr_queued[WRITE], sq->nr_queued[READ], sq->nr_queued[WRITE]); |
7f52f98c2 blk-throttle: imp... |
1187 1188 1189 |
ret = throtl_select_dispatch(sq); if (ret) { |
7f52f98c2 blk-throttle: imp... |
1190 1191 1192 |
throtl_log(sq, "bios disp=%u", ret); dispatched = true; } |
e43473b7f blkio: Core imple... |
1193 |
|
7f52f98c2 blk-throttle: imp... |
1194 1195 |
if (throtl_schedule_next_dispatch(sq, false)) break; |
e43473b7f blkio: Core imple... |
1196 |
|
7f52f98c2 blk-throttle: imp... |
1197 1198 1199 1200 |
/* this dispatch windows is still open, relax and repeat */ spin_unlock_irq(q->queue_lock); cpu_relax(); spin_lock_irq(q->queue_lock); |
651930bc1 blk-throttle: dis... |
1201 |
} |
e43473b7f blkio: Core imple... |
1202 |
|
2e48a530a blk-throttle: mak... |
1203 1204 |
if (!dispatched) goto out_unlock; |
6e1a5704c blk-throttle: dis... |
1205 |
|
2e48a530a blk-throttle: mak... |
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 |
if (parent_sq) { /* @parent_sq is another throl_grp, propagate dispatch */ if (tg->flags & THROTL_TG_WAS_EMPTY) { tg_update_disptime(tg); if (!throtl_schedule_next_dispatch(parent_sq, false)) { /* window is already open, repeat dispatching */ sq = parent_sq; tg = sq_to_tg(sq); goto again; } } } else { /* reached the top-level, queue issueing */ queue_work(kthrotld_workqueue, &td->dispatch_work); } out_unlock: |
e43473b7f blkio: Core imple... |
1222 |
spin_unlock_irq(q->queue_lock); |
6e1a5704c blk-throttle: dis... |
1223 |
} |
e43473b7f blkio: Core imple... |
1224 |
|
6e1a5704c blk-throttle: dis... |
1225 1226 1227 1228 1229 1230 1231 1232 |
/** * blk_throtl_dispatch_work_fn - work function for throtl_data->dispatch_work * @work: work item being executed * * This function is queued for execution when bio's reach the bio_lists[] * of throtl_data->service_queue. Those bio's are ready and issued by this * function. */ |
8876e140e block/blk-throttl... |
1233 |
static void blk_throtl_dispatch_work_fn(struct work_struct *work) |
6e1a5704c blk-throttle: dis... |
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 |
{ struct throtl_data *td = container_of(work, struct throtl_data, dispatch_work); struct throtl_service_queue *td_sq = &td->service_queue; struct request_queue *q = td->queue; struct bio_list bio_list_on_stack; struct bio *bio; struct blk_plug plug; int rw; bio_list_init(&bio_list_on_stack); spin_lock_irq(q->queue_lock); |
c5cc2070b blk-throttle: add... |
1247 1248 1249 |
for (rw = READ; rw <= WRITE; rw++) while ((bio = throtl_pop_queued(&td_sq->queued[rw], NULL))) bio_list_add(&bio_list_on_stack, bio); |
6e1a5704c blk-throttle: dis... |
1250 1251 1252 |
spin_unlock_irq(q->queue_lock); if (!bio_list_empty(&bio_list_on_stack)) { |
69d60eb96 blk-throttle: Use... |
1253 |
blk_start_plug(&plug); |
e43473b7f blkio: Core imple... |
1254 1255 |
while((bio = bio_list_pop(&bio_list_on_stack))) generic_make_request(bio); |
69d60eb96 blk-throttle: Use... |
1256 |
blk_finish_plug(&plug); |
e43473b7f blkio: Core imple... |
1257 |
} |
e43473b7f blkio: Core imple... |
1258 |
} |
f95a04afa blkcg: embed stru... |
1259 1260 |
static u64 tg_prfill_conf_u64(struct seq_file *sf, struct blkg_policy_data *pd, int off) |
60c2bc2d5 blkcg: move conf/... |
1261 |
{ |
f95a04afa blkcg: embed stru... |
1262 1263 |
struct throtl_grp *tg = pd_to_tg(pd); u64 v = *(u64 *)((void *)tg + off); |
60c2bc2d5 blkcg: move conf/... |
1264 |
|
2ab5492de blk-throttle: use... |
1265 |
if (v == U64_MAX) |
60c2bc2d5 blkcg: move conf/... |
1266 |
return 0; |
f95a04afa blkcg: embed stru... |
1267 |
return __blkg_prfill_u64(sf, pd, v); |
60c2bc2d5 blkcg: move conf/... |
1268 |
} |
f95a04afa blkcg: embed stru... |
1269 1270 |
static u64 tg_prfill_conf_uint(struct seq_file *sf, struct blkg_policy_data *pd, int off) |
e43473b7f blkio: Core imple... |
1271 |
{ |
f95a04afa blkcg: embed stru... |
1272 1273 |
struct throtl_grp *tg = pd_to_tg(pd); unsigned int v = *(unsigned int *)((void *)tg + off); |
fe0714377 blkio: Recalculat... |
1274 |
|
2ab5492de blk-throttle: use... |
1275 |
if (v == UINT_MAX) |
af133ceb2 blkcg: move blkio... |
1276 |
return 0; |
f95a04afa blkcg: embed stru... |
1277 |
return __blkg_prfill_u64(sf, pd, v); |
e43473b7f blkio: Core imple... |
1278 |
} |
2da8ca822 cgroup: replace c... |
1279 |
static int tg_print_conf_u64(struct seq_file *sf, void *v) |
8e89d13f4 blkio: Implementa... |
1280 |
{ |
2da8ca822 cgroup: replace c... |
1281 1282 |
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_u64, &blkcg_policy_throtl, seq_cft(sf)->private, false); |
af133ceb2 blkcg: move blkio... |
1283 |
return 0; |
8e89d13f4 blkio: Implementa... |
1284 |
} |
2da8ca822 cgroup: replace c... |
1285 |
static int tg_print_conf_uint(struct seq_file *sf, void *v) |
8e89d13f4 blkio: Implementa... |
1286 |
{ |
2da8ca822 cgroup: replace c... |
1287 1288 |
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_uint, &blkcg_policy_throtl, seq_cft(sf)->private, false); |
af133ceb2 blkcg: move blkio... |
1289 |
return 0; |
60c2bc2d5 blkcg: move conf/... |
1290 |
} |
9bb67aeb9 blk-throttle: res... |
1291 |
static void tg_conf_updated(struct throtl_grp *tg, bool global) |
60c2bc2d5 blkcg: move conf/... |
1292 |
{ |
69948b070 blkcg: separate o... |
1293 |
struct throtl_service_queue *sq = &tg->service_queue; |
492eb21b9 cgroup: make hier... |
1294 |
struct cgroup_subsys_state *pos_css; |
69948b070 blkcg: separate o... |
1295 |
struct blkcg_gq *blkg; |
af133ceb2 blkcg: move blkio... |
1296 |
|
fda6f272c blk-throttle: imp... |
1297 1298 |
throtl_log(&tg->service_queue, "limit change rbps=%llu wbps=%llu riops=%u wiops=%u", |
9f626e372 blk-throttle: pre... |
1299 1300 |
tg_bps_limit(tg, READ), tg_bps_limit(tg, WRITE), tg_iops_limit(tg, READ), tg_iops_limit(tg, WRITE)); |
632b44935 blk-throttle: rem... |
1301 1302 |
/* |
693e751e7 blk-throttle: imp... |
1303 1304 1305 1306 1307 1308 |
* Update has_rules[] flags for the updated tg's subtree. A tg is * considered to have rules if either the tg itself or any of its * ancestors has rules. This identifies groups without any * restrictions in the whole hierarchy and allows them to bypass * blk-throttle. */ |
9bb67aeb9 blk-throttle: res... |
1309 1310 |
blkg_for_each_descendant_pre(blkg, pos_css, global ? tg->td->queue->root_blkg : tg_to_blkg(tg)) { |
5b81fc3cc blk-throttle: add... |
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 |
struct throtl_grp *this_tg = blkg_to_tg(blkg); struct throtl_grp *parent_tg; tg_update_has_rules(this_tg); /* ignore root/second level */ if (!cgroup_subsys_on_dfl(io_cgrp_subsys) || !blkg->parent || !blkg->parent->parent) continue; parent_tg = blkg_to_tg(blkg->parent); /* * make sure all children has lower idle time threshold and * higher latency target */ this_tg->idletime_threshold = min(this_tg->idletime_threshold, parent_tg->idletime_threshold); this_tg->latency_target = max(this_tg->latency_target, parent_tg->latency_target); } |
693e751e7 blk-throttle: imp... |
1329 1330 |
/* |
632b44935 blk-throttle: rem... |
1331 1332 1333 1334 1335 1336 1337 |
* We're already holding queue_lock and know @tg is valid. Let's * apply the new config directly. * * Restart the slices for both READ and WRITES. It might happen * that a group's limit are dropped suddenly and we don't want to * account recently dispatched IO with new low rate. */ |
0f3457f60 blk-throttle: add... |
1338 1339 |
throtl_start_new_slice(tg, 0); throtl_start_new_slice(tg, 1); |
632b44935 blk-throttle: rem... |
1340 |
|
5b2c16aae blk-throttle: sim... |
1341 |
if (tg->flags & THROTL_TG_PENDING) { |
77216b048 blk-throttle: add... |
1342 |
tg_update_disptime(tg); |
7f52f98c2 blk-throttle: imp... |
1343 |
throtl_schedule_next_dispatch(sq->parent_sq, true); |
632b44935 blk-throttle: rem... |
1344 |
} |
69948b070 blkcg: separate o... |
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 |
} static ssize_t tg_set_conf(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off, bool is_u64) { struct blkcg *blkcg = css_to_blkcg(of_css(of)); struct blkg_conf_ctx ctx; struct throtl_grp *tg; int ret; u64 v; ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx); if (ret) return ret; ret = -EINVAL; if (sscanf(ctx.body, "%llu", &v) != 1) goto out_finish; if (!v) |
2ab5492de blk-throttle: use... |
1364 |
v = U64_MAX; |
69948b070 blkcg: separate o... |
1365 1366 1367 1368 1369 1370 1371 |
tg = blkg_to_tg(ctx.blkg); if (is_u64) *(u64 *)((void *)tg + of_cft(of)->private) = v; else *(unsigned int *)((void *)tg + of_cft(of)->private) = v; |
60c2bc2d5 blkcg: move conf/... |
1372 |
|
9bb67aeb9 blk-throttle: res... |
1373 |
tg_conf_updated(tg, false); |
36aa9e5f5 blkcg: move body ... |
1374 1375 |
ret = 0; out_finish: |
60c2bc2d5 blkcg: move conf/... |
1376 |
blkg_conf_finish(&ctx); |
36aa9e5f5 blkcg: move body ... |
1377 |
return ret ?: nbytes; |
8e89d13f4 blkio: Implementa... |
1378 |
} |
451af504d cgroup: replace c... |
1379 1380 |
static ssize_t tg_set_conf_u64(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
60c2bc2d5 blkcg: move conf/... |
1381 |
{ |
451af504d cgroup: replace c... |
1382 |
return tg_set_conf(of, buf, nbytes, off, true); |
60c2bc2d5 blkcg: move conf/... |
1383 |
} |
451af504d cgroup: replace c... |
1384 1385 |
static ssize_t tg_set_conf_uint(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
60c2bc2d5 blkcg: move conf/... |
1386 |
{ |
451af504d cgroup: replace c... |
1387 |
return tg_set_conf(of, buf, nbytes, off, false); |
60c2bc2d5 blkcg: move conf/... |
1388 |
} |
880f50e22 blkcg: mark exist... |
1389 |
static struct cftype throtl_legacy_files[] = { |
60c2bc2d5 blkcg: move conf/... |
1390 1391 |
{ .name = "throttle.read_bps_device", |
9f626e372 blk-throttle: pre... |
1392 |
.private = offsetof(struct throtl_grp, bps[READ][LIMIT_MAX]), |
2da8ca822 cgroup: replace c... |
1393 |
.seq_show = tg_print_conf_u64, |
451af504d cgroup: replace c... |
1394 |
.write = tg_set_conf_u64, |
60c2bc2d5 blkcg: move conf/... |
1395 1396 1397 |
}, { .name = "throttle.write_bps_device", |
9f626e372 blk-throttle: pre... |
1398 |
.private = offsetof(struct throtl_grp, bps[WRITE][LIMIT_MAX]), |
2da8ca822 cgroup: replace c... |
1399 |
.seq_show = tg_print_conf_u64, |
451af504d cgroup: replace c... |
1400 |
.write = tg_set_conf_u64, |
60c2bc2d5 blkcg: move conf/... |
1401 1402 1403 |
}, { .name = "throttle.read_iops_device", |
9f626e372 blk-throttle: pre... |
1404 |
.private = offsetof(struct throtl_grp, iops[READ][LIMIT_MAX]), |
2da8ca822 cgroup: replace c... |
1405 |
.seq_show = tg_print_conf_uint, |
451af504d cgroup: replace c... |
1406 |
.write = tg_set_conf_uint, |
60c2bc2d5 blkcg: move conf/... |
1407 1408 1409 |
}, { .name = "throttle.write_iops_device", |
9f626e372 blk-throttle: pre... |
1410 |
.private = offsetof(struct throtl_grp, iops[WRITE][LIMIT_MAX]), |
2da8ca822 cgroup: replace c... |
1411 |
.seq_show = tg_print_conf_uint, |
451af504d cgroup: replace c... |
1412 |
.write = tg_set_conf_uint, |
60c2bc2d5 blkcg: move conf/... |
1413 1414 1415 |
}, { .name = "throttle.io_service_bytes", |
77ea73388 blkcg: move io_se... |
1416 1417 |
.private = (unsigned long)&blkcg_policy_throtl, .seq_show = blkg_print_stat_bytes, |
60c2bc2d5 blkcg: move conf/... |
1418 1419 |
}, { |
1b9137ac5 blk-throttle: exp... |
1420 1421 1422 1423 1424 |
.name = "throttle.io_service_bytes_recursive", .private = (unsigned long)&blkcg_policy_throtl, .seq_show = blkg_print_stat_bytes_recursive, }, { |
60c2bc2d5 blkcg: move conf/... |
1425 |
.name = "throttle.io_serviced", |
77ea73388 blkcg: move io_se... |
1426 1427 |
.private = (unsigned long)&blkcg_policy_throtl, .seq_show = blkg_print_stat_ios, |
60c2bc2d5 blkcg: move conf/... |
1428 |
}, |
1b9137ac5 blk-throttle: exp... |
1429 1430 1431 1432 1433 |
{ .name = "throttle.io_serviced_recursive", .private = (unsigned long)&blkcg_policy_throtl, .seq_show = blkg_print_stat_ios_recursive, }, |
60c2bc2d5 blkcg: move conf/... |
1434 1435 |
{ } /* terminate */ }; |
cd5ab1b0f blk-throttle: add... |
1436 |
static u64 tg_prfill_limit(struct seq_file *sf, struct blkg_policy_data *pd, |
2ee867dcf blkcg: implement ... |
1437 1438 1439 1440 1441 |
int off) { struct throtl_grp *tg = pd_to_tg(pd); const char *dname = blkg_dev_name(pd->blkg); char bufs[4][21] = { "max", "max", "max", "max" }; |
cd5ab1b0f blk-throttle: add... |
1442 1443 |
u64 bps_dft; unsigned int iops_dft; |
ada75b6e5 blk-throttle: add... |
1444 |
char idle_time[26] = ""; |
ec80991d6 blk-throttle: add... |
1445 |
char latency_time[26] = ""; |
2ee867dcf blkcg: implement ... |
1446 1447 1448 |
if (!dname) return 0; |
9f626e372 blk-throttle: pre... |
1449 |
|
cd5ab1b0f blk-throttle: add... |
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 |
if (off == LIMIT_LOW) { bps_dft = 0; iops_dft = 0; } else { bps_dft = U64_MAX; iops_dft = UINT_MAX; } if (tg->bps_conf[READ][off] == bps_dft && tg->bps_conf[WRITE][off] == bps_dft && tg->iops_conf[READ][off] == iops_dft && |
ada75b6e5 blk-throttle: add... |
1461 |
tg->iops_conf[WRITE][off] == iops_dft && |
ec80991d6 blk-throttle: add... |
1462 |
(off != LIMIT_LOW || |
b4f428ef2 blk-throttle: for... |
1463 |
(tg->idletime_threshold_conf == DFL_IDLE_THRESHOLD && |
5b81fc3cc blk-throttle: add... |
1464 |
tg->latency_target_conf == DFL_LATENCY_TARGET))) |
2ee867dcf blkcg: implement ... |
1465 |
return 0; |
9bb67aeb9 blk-throttle: res... |
1466 |
if (tg->bps_conf[READ][off] != U64_MAX) |
9f626e372 blk-throttle: pre... |
1467 |
snprintf(bufs[0], sizeof(bufs[0]), "%llu", |
cd5ab1b0f blk-throttle: add... |
1468 |
tg->bps_conf[READ][off]); |
9bb67aeb9 blk-throttle: res... |
1469 |
if (tg->bps_conf[WRITE][off] != U64_MAX) |
9f626e372 blk-throttle: pre... |
1470 |
snprintf(bufs[1], sizeof(bufs[1]), "%llu", |
cd5ab1b0f blk-throttle: add... |
1471 |
tg->bps_conf[WRITE][off]); |
9bb67aeb9 blk-throttle: res... |
1472 |
if (tg->iops_conf[READ][off] != UINT_MAX) |
9f626e372 blk-throttle: pre... |
1473 |
snprintf(bufs[2], sizeof(bufs[2]), "%u", |
cd5ab1b0f blk-throttle: add... |
1474 |
tg->iops_conf[READ][off]); |
9bb67aeb9 blk-throttle: res... |
1475 |
if (tg->iops_conf[WRITE][off] != UINT_MAX) |
9f626e372 blk-throttle: pre... |
1476 |
snprintf(bufs[3], sizeof(bufs[3]), "%u", |
cd5ab1b0f blk-throttle: add... |
1477 |
tg->iops_conf[WRITE][off]); |
ada75b6e5 blk-throttle: add... |
1478 |
if (off == LIMIT_LOW) { |
5b81fc3cc blk-throttle: add... |
1479 |
if (tg->idletime_threshold_conf == ULONG_MAX) |
ada75b6e5 blk-throttle: add... |
1480 1481 1482 |
strcpy(idle_time, " idle=max"); else snprintf(idle_time, sizeof(idle_time), " idle=%lu", |
5b81fc3cc blk-throttle: add... |
1483 |
tg->idletime_threshold_conf); |
ec80991d6 blk-throttle: add... |
1484 |
|
5b81fc3cc blk-throttle: add... |
1485 |
if (tg->latency_target_conf == ULONG_MAX) |
ec80991d6 blk-throttle: add... |
1486 1487 1488 |
strcpy(latency_time, " latency=max"); else snprintf(latency_time, sizeof(latency_time), |
5b81fc3cc blk-throttle: add... |
1489 |
" latency=%lu", tg->latency_target_conf); |
ada75b6e5 blk-throttle: add... |
1490 |
} |
2ee867dcf blkcg: implement ... |
1491 |
|
ec80991d6 blk-throttle: add... |
1492 1493 1494 1495 |
seq_printf(sf, "%s rbps=%s wbps=%s riops=%s wiops=%s%s%s ", dname, bufs[0], bufs[1], bufs[2], bufs[3], idle_time, latency_time); |
2ee867dcf blkcg: implement ... |
1496 1497 |
return 0; } |
cd5ab1b0f blk-throttle: add... |
1498 |
static int tg_print_limit(struct seq_file *sf, void *v) |
2ee867dcf blkcg: implement ... |
1499 |
{ |
cd5ab1b0f blk-throttle: add... |
1500 |
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_limit, |
2ee867dcf blkcg: implement ... |
1501 1502 1503 |
&blkcg_policy_throtl, seq_cft(sf)->private, false); return 0; } |
cd5ab1b0f blk-throttle: add... |
1504 |
static ssize_t tg_set_limit(struct kernfs_open_file *of, |
2ee867dcf blkcg: implement ... |
1505 1506 1507 1508 1509 1510 |
char *buf, size_t nbytes, loff_t off) { struct blkcg *blkcg = css_to_blkcg(of_css(of)); struct blkg_conf_ctx ctx; struct throtl_grp *tg; u64 v[4]; |
ada75b6e5 blk-throttle: add... |
1511 |
unsigned long idle_time; |
ec80991d6 blk-throttle: add... |
1512 |
unsigned long latency_time; |
2ee867dcf blkcg: implement ... |
1513 |
int ret; |
cd5ab1b0f blk-throttle: add... |
1514 |
int index = of_cft(of)->private; |
2ee867dcf blkcg: implement ... |
1515 1516 1517 1518 1519 1520 |
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx); if (ret) return ret; tg = blkg_to_tg(ctx.blkg); |
cd5ab1b0f blk-throttle: add... |
1521 1522 1523 1524 |
v[0] = tg->bps_conf[READ][index]; v[1] = tg->bps_conf[WRITE][index]; v[2] = tg->iops_conf[READ][index]; v[3] = tg->iops_conf[WRITE][index]; |
2ee867dcf blkcg: implement ... |
1525 |
|
5b81fc3cc blk-throttle: add... |
1526 1527 |
idle_time = tg->idletime_threshold_conf; latency_time = tg->latency_target_conf; |
2ee867dcf blkcg: implement ... |
1528 1529 1530 |
while (true) { char tok[27]; /* wiops=18446744073709551616 */ char *p; |
2ab5492de blk-throttle: use... |
1531 |
u64 val = U64_MAX; |
2ee867dcf blkcg: implement ... |
1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 |
int len; if (sscanf(ctx.body, "%26s%n", tok, &len) != 1) break; if (tok[0] == '\0') break; ctx.body += len; ret = -EINVAL; p = tok; strsep(&p, "="); if (!p || (sscanf(p, "%llu", &val) != 1 && strcmp(p, "max"))) goto out_finish; ret = -ERANGE; if (!val) goto out_finish; ret = -EINVAL; if (!strcmp(tok, "rbps")) v[0] = val; else if (!strcmp(tok, "wbps")) v[1] = val; else if (!strcmp(tok, "riops")) v[2] = min_t(u64, val, UINT_MAX); else if (!strcmp(tok, "wiops")) v[3] = min_t(u64, val, UINT_MAX); |
ada75b6e5 blk-throttle: add... |
1559 1560 |
else if (off == LIMIT_LOW && !strcmp(tok, "idle")) idle_time = val; |
ec80991d6 blk-throttle: add... |
1561 1562 |
else if (off == LIMIT_LOW && !strcmp(tok, "latency")) latency_time = val; |
2ee867dcf blkcg: implement ... |
1563 1564 1565 |
else goto out_finish; } |
cd5ab1b0f blk-throttle: add... |
1566 1567 1568 1569 |
tg->bps_conf[READ][index] = v[0]; tg->bps_conf[WRITE][index] = v[1]; tg->iops_conf[READ][index] = v[2]; tg->iops_conf[WRITE][index] = v[3]; |
2ee867dcf blkcg: implement ... |
1570 |
|
cd5ab1b0f blk-throttle: add... |
1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 |
if (index == LIMIT_MAX) { tg->bps[READ][index] = v[0]; tg->bps[WRITE][index] = v[1]; tg->iops[READ][index] = v[2]; tg->iops[WRITE][index] = v[3]; } tg->bps[READ][LIMIT_LOW] = min(tg->bps_conf[READ][LIMIT_LOW], tg->bps_conf[READ][LIMIT_MAX]); tg->bps[WRITE][LIMIT_LOW] = min(tg->bps_conf[WRITE][LIMIT_LOW], tg->bps_conf[WRITE][LIMIT_MAX]); tg->iops[READ][LIMIT_LOW] = min(tg->iops_conf[READ][LIMIT_LOW], tg->iops_conf[READ][LIMIT_MAX]); tg->iops[WRITE][LIMIT_LOW] = min(tg->iops_conf[WRITE][LIMIT_LOW], tg->iops_conf[WRITE][LIMIT_MAX]); |
b4f428ef2 blk-throttle: for... |
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 |
tg->idletime_threshold_conf = idle_time; tg->latency_target_conf = latency_time; /* force user to configure all settings for low limit */ if (!(tg->bps[READ][LIMIT_LOW] || tg->iops[READ][LIMIT_LOW] || tg->bps[WRITE][LIMIT_LOW] || tg->iops[WRITE][LIMIT_LOW]) || tg->idletime_threshold_conf == DFL_IDLE_THRESHOLD || tg->latency_target_conf == DFL_LATENCY_TARGET) { tg->bps[READ][LIMIT_LOW] = 0; tg->bps[WRITE][LIMIT_LOW] = 0; tg->iops[READ][LIMIT_LOW] = 0; tg->iops[WRITE][LIMIT_LOW] = 0; tg->idletime_threshold = DFL_IDLE_THRESHOLD; tg->latency_target = DFL_LATENCY_TARGET; } else if (index == LIMIT_LOW) { |
5b81fc3cc blk-throttle: add... |
1600 |
tg->idletime_threshold = tg->idletime_threshold_conf; |
5b81fc3cc blk-throttle: add... |
1601 |
tg->latency_target = tg->latency_target_conf; |
cd5ab1b0f blk-throttle: add... |
1602 |
} |
b4f428ef2 blk-throttle: for... |
1603 1604 1605 1606 1607 1608 1609 |
blk_throtl_update_limit_valid(tg->td); if (tg->td->limit_valid[LIMIT_LOW]) { if (index == LIMIT_LOW) tg->td->limit_index = LIMIT_LOW; } else tg->td->limit_index = LIMIT_MAX; |
9bb67aeb9 blk-throttle: res... |
1610 1611 |
tg_conf_updated(tg, index == LIMIT_LOW && tg->td->limit_valid[LIMIT_LOW]); |
2ee867dcf blkcg: implement ... |
1612 1613 1614 1615 1616 1617 1618 |
ret = 0; out_finish: blkg_conf_finish(&ctx); return ret ?: nbytes; } static struct cftype throtl_files[] = { |
cd5ab1b0f blk-throttle: add... |
1619 1620 1621 1622 1623 1624 1625 1626 1627 |
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW { .name = "low", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = tg_print_limit, .write = tg_set_limit, .private = LIMIT_LOW, }, #endif |
2ee867dcf blkcg: implement ... |
1628 1629 1630 |
{ .name = "max", .flags = CFTYPE_NOT_ON_ROOT, |
cd5ab1b0f blk-throttle: add... |
1631 1632 1633 |
.seq_show = tg_print_limit, .write = tg_set_limit, .private = LIMIT_MAX, |
2ee867dcf blkcg: implement ... |
1634 1635 1636 |
}, { } /* terminate */ }; |
da5277700 block: Move blk_t... |
1637 |
static void throtl_shutdown_wq(struct request_queue *q) |
e43473b7f blkio: Core imple... |
1638 1639 |
{ struct throtl_data *td = q->td; |
69df0ab03 blk-throttle: sep... |
1640 |
cancel_work_sync(&td->dispatch_work); |
e43473b7f blkio: Core imple... |
1641 |
} |
3c798398e blkcg: mass renam... |
1642 |
static struct blkcg_policy blkcg_policy_throtl = { |
2ee867dcf blkcg: implement ... |
1643 |
.dfl_cftypes = throtl_files, |
880f50e22 blkcg: mark exist... |
1644 |
.legacy_cftypes = throtl_legacy_files, |
f9fcc2d39 blkcg: collapse b... |
1645 |
|
001bea73e blkcg: replace bl... |
1646 |
.pd_alloc_fn = throtl_pd_alloc, |
f9fcc2d39 blkcg: collapse b... |
1647 |
.pd_init_fn = throtl_pd_init, |
693e751e7 blk-throttle: imp... |
1648 |
.pd_online_fn = throtl_pd_online, |
cd5ab1b0f blk-throttle: add... |
1649 |
.pd_offline_fn = throtl_pd_offline, |
001bea73e blkcg: replace bl... |
1650 |
.pd_free_fn = throtl_pd_free, |
e43473b7f blkio: Core imple... |
1651 |
}; |
3f0abd806 blk-throttle: add... |
1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 |
static unsigned long __tg_last_low_overflow_time(struct throtl_grp *tg) { unsigned long rtime = jiffies, wtime = jiffies; if (tg->bps[READ][LIMIT_LOW] || tg->iops[READ][LIMIT_LOW]) rtime = tg->last_low_overflow_time[READ]; if (tg->bps[WRITE][LIMIT_LOW] || tg->iops[WRITE][LIMIT_LOW]) wtime = tg->last_low_overflow_time[WRITE]; return min(rtime, wtime); } /* tg should not be an intermediate node */ static unsigned long tg_last_low_overflow_time(struct throtl_grp *tg) { struct throtl_service_queue *parent_sq; struct throtl_grp *parent = tg; unsigned long ret = __tg_last_low_overflow_time(tg); while (true) { parent_sq = parent->service_queue.parent_sq; parent = sq_to_tg(parent_sq); if (!parent) break; /* * The parent doesn't have low limit, it always reaches low * limit. Its overflow time is useless for children */ if (!parent->bps[READ][LIMIT_LOW] && !parent->iops[READ][LIMIT_LOW] && !parent->bps[WRITE][LIMIT_LOW] && !parent->iops[WRITE][LIMIT_LOW]) continue; if (time_after(__tg_last_low_overflow_time(parent), ret)) ret = __tg_last_low_overflow_time(parent); } return ret; } |
9e234eeaf blk-throttle: add... |
1690 1691 1692 1693 1694 |
static bool throtl_tg_is_idle(struct throtl_grp *tg) { /* * cgroup is idle if: * - single idle is too long, longer than a fixed value (in case user |
b4f428ef2 blk-throttle: for... |
1695 |
* configure a too big threshold) or 4 times of idletime threshold |
9e234eeaf blk-throttle: add... |
1696 |
* - average think time is more than threshold |
53696b8d2 blk-throttle: add... |
1697 |
* - IO latency is largely below threshold |
9e234eeaf blk-throttle: add... |
1698 |
*/ |
b4f428ef2 blk-throttle: for... |
1699 |
unsigned long time; |
4cff729f6 blk-throttle: out... |
1700 |
bool ret; |
9e234eeaf blk-throttle: add... |
1701 |
|
b4f428ef2 blk-throttle: for... |
1702 1703 1704 1705 1706 1707 |
time = min_t(unsigned long, MAX_IDLE_TIME, 4 * tg->idletime_threshold); ret = tg->latency_target == DFL_LATENCY_TARGET || tg->idletime_threshold == DFL_IDLE_THRESHOLD || (ktime_get_ns() >> 10) - tg->last_finish_time > time || tg->avg_idletime > tg->idletime_threshold || (tg->latency_target && tg->bio_cnt && |
53696b8d2 blk-throttle: add... |
1708 |
tg->bad_bio_cnt * 5 < tg->bio_cnt); |
4cff729f6 blk-throttle: out... |
1709 1710 1711 1712 1713 |
throtl_log(&tg->service_queue, "avg_idle=%ld, idle_threshold=%ld, bad_bio=%d, total_bio=%d, is_idle=%d, scale=%d", tg->avg_idletime, tg->idletime_threshold, tg->bad_bio_cnt, tg->bio_cnt, ret, tg->td->scale); return ret; |
9e234eeaf blk-throttle: add... |
1714 |
} |
c79892c55 blk-throttle: add... |
1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 |
static bool throtl_tg_can_upgrade(struct throtl_grp *tg) { struct throtl_service_queue *sq = &tg->service_queue; bool read_limit, write_limit; /* * if cgroup reaches low limit (if low limit is 0, the cgroup always * reaches), it's ok to upgrade to next limit */ read_limit = tg->bps[READ][LIMIT_LOW] || tg->iops[READ][LIMIT_LOW]; write_limit = tg->bps[WRITE][LIMIT_LOW] || tg->iops[WRITE][LIMIT_LOW]; if (!read_limit && !write_limit) return true; if (read_limit && sq->nr_queued[READ] && (!write_limit || sq->nr_queued[WRITE])) return true; if (write_limit && sq->nr_queued[WRITE] && (!read_limit || sq->nr_queued[READ])) return true; |
aec242468 blk-throttle: det... |
1734 1735 |
if (time_after_eq(jiffies, |
fa6fb5aab blk-throttle: ign... |
1736 1737 |
tg_last_low_overflow_time(tg) + tg->td->throtl_slice) && throtl_tg_is_idle(tg)) |
aec242468 blk-throttle: det... |
1738 |
return true; |
c79892c55 blk-throttle: add... |
1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 |
return false; } static bool throtl_hierarchy_can_upgrade(struct throtl_grp *tg) { while (true) { if (throtl_tg_can_upgrade(tg)) return true; tg = sq_to_tg(tg->service_queue.parent_sq); if (!tg || !tg_to_blkg(tg)->parent) return false; } return false; } static bool throtl_can_upgrade(struct throtl_data *td, struct throtl_grp *this_tg) { struct cgroup_subsys_state *pos_css; struct blkcg_gq *blkg; if (td->limit_index != LIMIT_LOW) return false; |
297e3d854 blk-throttle: mak... |
1762 |
if (time_before(jiffies, td->low_downgrade_time + td->throtl_slice)) |
3f0abd806 blk-throttle: add... |
1763 |
return false; |
c79892c55 blk-throttle: add... |
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 |
rcu_read_lock(); blkg_for_each_descendant_post(blkg, pos_css, td->queue->root_blkg) { struct throtl_grp *tg = blkg_to_tg(blkg); if (tg == this_tg) continue; if (!list_empty(&tg_to_blkg(tg)->blkcg->css.children)) continue; if (!throtl_hierarchy_can_upgrade(tg)) { rcu_read_unlock(); return false; } } rcu_read_unlock(); return true; } |
fa6fb5aab blk-throttle: ign... |
1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 |
static void throtl_upgrade_check(struct throtl_grp *tg) { unsigned long now = jiffies; if (tg->td->limit_index != LIMIT_LOW) return; if (time_after(tg->last_check_time + tg->td->throtl_slice, now)) return; tg->last_check_time = now; if (!time_after_eq(now, __tg_last_low_overflow_time(tg) + tg->td->throtl_slice)) return; if (throtl_can_upgrade(tg->td, NULL)) throtl_upgrade_state(tg->td); } |
c79892c55 blk-throttle: add... |
1799 1800 1801 1802 |
static void throtl_upgrade_state(struct throtl_data *td) { struct cgroup_subsys_state *pos_css; struct blkcg_gq *blkg; |
4cff729f6 blk-throttle: out... |
1803 |
throtl_log(&td->service_queue, "upgrade to max"); |
c79892c55 blk-throttle: add... |
1804 |
td->limit_index = LIMIT_MAX; |
3f0abd806 blk-throttle: add... |
1805 |
td->low_upgrade_time = jiffies; |
7394e31fa blk-throttle: mak... |
1806 |
td->scale = 0; |
c79892c55 blk-throttle: add... |
1807 1808 1809 1810 1811 1812 1813 |
rcu_read_lock(); blkg_for_each_descendant_post(blkg, pos_css, td->queue->root_blkg) { struct throtl_grp *tg = blkg_to_tg(blkg); struct throtl_service_queue *sq = &tg->service_queue; tg->disptime = jiffies - 1; throtl_select_dispatch(sq); |
4f02fb761 blk-throttle: fix... |
1814 |
throtl_schedule_next_dispatch(sq, true); |
c79892c55 blk-throttle: add... |
1815 1816 1817 |
} rcu_read_unlock(); throtl_select_dispatch(&td->service_queue); |
4f02fb761 blk-throttle: fix... |
1818 |
throtl_schedule_next_dispatch(&td->service_queue, true); |
c79892c55 blk-throttle: add... |
1819 1820 |
queue_work(kthrotld_workqueue, &td->dispatch_work); } |
3f0abd806 blk-throttle: add... |
1821 1822 |
static void throtl_downgrade_state(struct throtl_data *td, int new) { |
7394e31fa blk-throttle: mak... |
1823 |
td->scale /= 2; |
4cff729f6 blk-throttle: out... |
1824 |
throtl_log(&td->service_queue, "downgrade, scale %d", td->scale); |
7394e31fa blk-throttle: mak... |
1825 1826 1827 1828 |
if (td->scale) { td->low_upgrade_time = jiffies - td->scale * td->throtl_slice; return; } |
3f0abd806 blk-throttle: add... |
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 |
td->limit_index = new; td->low_downgrade_time = jiffies; } static bool throtl_tg_can_downgrade(struct throtl_grp *tg) { struct throtl_data *td = tg->td; unsigned long now = jiffies; /* * If cgroup is below low limit, consider downgrade and throttle other * cgroups */ |
297e3d854 blk-throttle: mak... |
1842 1843 |
if (time_after_eq(now, td->low_upgrade_time + td->throtl_slice) && time_after_eq(now, tg_last_low_overflow_time(tg) + |
fa6fb5aab blk-throttle: ign... |
1844 1845 1846 |
td->throtl_slice) && (!throtl_tg_is_idle(tg) || !list_empty(&tg_to_blkg(tg)->blkcg->css.children))) |
3f0abd806 blk-throttle: add... |
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 |
return true; return false; } static bool throtl_hierarchy_can_downgrade(struct throtl_grp *tg) { while (true) { if (!throtl_tg_can_downgrade(tg)) return false; tg = sq_to_tg(tg->service_queue.parent_sq); if (!tg || !tg_to_blkg(tg)->parent) break; } return true; } static void throtl_downgrade_check(struct throtl_grp *tg) { uint64_t bps; unsigned int iops; unsigned long elapsed_time; unsigned long now = jiffies; if (tg->td->limit_index != LIMIT_MAX || !tg->td->limit_valid[LIMIT_LOW]) return; if (!list_empty(&tg_to_blkg(tg)->blkcg->css.children)) return; |
297e3d854 blk-throttle: mak... |
1875 |
if (time_after(tg->last_check_time + tg->td->throtl_slice, now)) |
3f0abd806 blk-throttle: add... |
1876 1877 1878 1879 |
return; elapsed_time = now - tg->last_check_time; tg->last_check_time = now; |
297e3d854 blk-throttle: mak... |
1880 1881 |
if (time_before(now, tg_last_low_overflow_time(tg) + tg->td->throtl_slice)) |
3f0abd806 blk-throttle: add... |
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 |
return; if (tg->bps[READ][LIMIT_LOW]) { bps = tg->last_bytes_disp[READ] * HZ; do_div(bps, elapsed_time); if (bps >= tg->bps[READ][LIMIT_LOW]) tg->last_low_overflow_time[READ] = now; } if (tg->bps[WRITE][LIMIT_LOW]) { bps = tg->last_bytes_disp[WRITE] * HZ; do_div(bps, elapsed_time); if (bps >= tg->bps[WRITE][LIMIT_LOW]) tg->last_low_overflow_time[WRITE] = now; } if (tg->iops[READ][LIMIT_LOW]) { iops = tg->last_io_disp[READ] * HZ / elapsed_time; if (iops >= tg->iops[READ][LIMIT_LOW]) tg->last_low_overflow_time[READ] = now; } if (tg->iops[WRITE][LIMIT_LOW]) { iops = tg->last_io_disp[WRITE] * HZ / elapsed_time; if (iops >= tg->iops[WRITE][LIMIT_LOW]) tg->last_low_overflow_time[WRITE] = now; } /* * If cgroup is below low limit, consider downgrade and throttle other * cgroups */ if (throtl_hierarchy_can_downgrade(tg)) throtl_downgrade_state(tg->td, LIMIT_LOW); tg->last_bytes_disp[READ] = 0; tg->last_bytes_disp[WRITE] = 0; tg->last_io_disp[READ] = 0; tg->last_io_disp[WRITE] = 0; } |
9e234eeaf blk-throttle: add... |
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 |
static void blk_throtl_update_idletime(struct throtl_grp *tg) { unsigned long now = ktime_get_ns() >> 10; unsigned long last_finish_time = tg->last_finish_time; if (now <= last_finish_time || last_finish_time == 0 || last_finish_time == tg->checked_last_finish_time) return; tg->avg_idletime = (tg->avg_idletime * 7 + now - last_finish_time) >> 3; tg->checked_last_finish_time = last_finish_time; } |
b9147dd1b blk-throttle: add... |
1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 |
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW static void throtl_update_latency_buckets(struct throtl_data *td) { struct avg_latency_bucket avg_latency[LATENCY_BUCKET_SIZE]; int i, cpu; unsigned long last_latency = 0; unsigned long latency; if (!blk_queue_nonrot(td->queue)) return; if (time_before(jiffies, td->last_calculate_time + HZ)) return; td->last_calculate_time = jiffies; memset(avg_latency, 0, sizeof(avg_latency)); for (i = 0; i < LATENCY_BUCKET_SIZE; i++) { struct latency_bucket *tmp = &td->tmp_buckets[i]; for_each_possible_cpu(cpu) { struct latency_bucket *bucket; /* this isn't race free, but ok in practice */ bucket = per_cpu_ptr(td->latency_buckets, cpu); tmp->total_latency += bucket[i].total_latency; tmp->samples += bucket[i].samples; bucket[i].total_latency = 0; bucket[i].samples = 0; } if (tmp->samples >= 32) { int samples = tmp->samples; latency = tmp->total_latency; tmp->total_latency = 0; tmp->samples = 0; latency /= samples; if (latency == 0) continue; avg_latency[i].latency = latency; } } for (i = 0; i < LATENCY_BUCKET_SIZE; i++) { if (!avg_latency[i].latency) { if (td->avg_buckets[i].latency < last_latency) td->avg_buckets[i].latency = last_latency; continue; } if (!td->avg_buckets[i].valid) latency = avg_latency[i].latency; else latency = (td->avg_buckets[i].latency * 7 + avg_latency[i].latency) >> 3; td->avg_buckets[i].latency = max(latency, last_latency); td->avg_buckets[i].valid = true; last_latency = td->avg_buckets[i].latency; } |
4cff729f6 blk-throttle: out... |
1994 1995 1996 1997 1998 |
for (i = 0; i < LATENCY_BUCKET_SIZE; i++) throtl_log(&td->service_queue, "Latency bucket %d: latency=%ld, valid=%d", i, td->avg_buckets[i].latency, td->avg_buckets[i].valid); |
b9147dd1b blk-throttle: add... |
1999 2000 2001 2002 2003 2004 |
} #else static inline void throtl_update_latency_buckets(struct throtl_data *td) { } #endif |
2bc19cd5f blk-throttle: fix... |
2005 2006 2007 |
static void blk_throtl_assoc_bio(struct throtl_grp *tg, struct bio *bio) { #ifdef CONFIG_BLK_DEV_THROTTLING_LOW |
007cc56b7 block: always att... |
2008 |
if (bio->bi_css) |
2bc19cd5f blk-throttle: fix... |
2009 2010 |
bio->bi_cg_private = tg; blk_stat_set_issue(&bio->bi_issue_stat, bio_sectors(bio)); |
2bc19cd5f blk-throttle: fix... |
2011 2012 |
#endif } |
ae1188963 blkcg: consolidat... |
2013 2014 |
bool blk_throtl_bio(struct request_queue *q, struct blkcg_gq *blkg, struct bio *bio) |
e43473b7f blkio: Core imple... |
2015 |
{ |
c5cc2070b blk-throttle: add... |
2016 |
struct throtl_qnode *qn = NULL; |
ae1188963 blkcg: consolidat... |
2017 |
struct throtl_grp *tg = blkg_to_tg(blkg ?: q->root_blkg); |
73f0d49a9 blk-throttle: mov... |
2018 |
struct throtl_service_queue *sq; |
0e9f4164b blk-throttle: gen... |
2019 |
bool rw = bio_data_dir(bio); |
bc16a4f93 block: reorganize... |
2020 |
bool throttled = false; |
b9147dd1b blk-throttle: add... |
2021 |
struct throtl_data *td = tg->td; |
e43473b7f blkio: Core imple... |
2022 |
|
ae1188963 blkcg: consolidat... |
2023 |
WARN_ON_ONCE(!rcu_read_lock_held()); |
2a0f61e6e blk-throttle: set... |
2024 |
/* see throtl_charge_bio() */ |
8d2bbd4c8 block: replace RE... |
2025 |
if (bio_flagged(bio, BIO_THROTTLED) || !tg->has_rules[rw]) |
bc16a4f93 block: reorganize... |
2026 |
goto out; |
e43473b7f blkio: Core imple... |
2027 2028 |
spin_lock_irq(q->queue_lock); |
c9589f03e blk-throttle: imp... |
2029 |
|
b9147dd1b blk-throttle: add... |
2030 |
throtl_update_latency_buckets(td); |
c9589f03e blk-throttle: imp... |
2031 |
if (unlikely(blk_queue_bypass(q))) |
bc16a4f93 block: reorganize... |
2032 |
goto out_unlock; |
f469a7b4d blk-cgroup: Allow... |
2033 |
|
2bc19cd5f blk-throttle: fix... |
2034 |
blk_throtl_assoc_bio(tg, bio); |
9e234eeaf blk-throttle: add... |
2035 |
blk_throtl_update_idletime(tg); |
73f0d49a9 blk-throttle: mov... |
2036 |
sq = &tg->service_queue; |
c79892c55 blk-throttle: add... |
2037 |
again: |
9e660acff blk-throttle: mak... |
2038 |
while (true) { |
3f0abd806 blk-throttle: add... |
2039 2040 2041 |
if (tg->last_low_overflow_time[rw] == 0) tg->last_low_overflow_time[rw] = jiffies; throtl_downgrade_check(tg); |
fa6fb5aab blk-throttle: ign... |
2042 |
throtl_upgrade_check(tg); |
9e660acff blk-throttle: mak... |
2043 2044 2045 |
/* throtl is FIFO - if bios are already queued, should queue */ if (sq->nr_queued[rw]) break; |
de701c74a blk-throttle: Som... |
2046 |
|
9e660acff blk-throttle: mak... |
2047 |
/* if above limits, break to queue */ |
c79892c55 blk-throttle: add... |
2048 |
if (!tg_may_dispatch(tg, bio, NULL)) { |
3f0abd806 blk-throttle: add... |
2049 |
tg->last_low_overflow_time[rw] = jiffies; |
b9147dd1b blk-throttle: add... |
2050 2051 |
if (throtl_can_upgrade(td, tg)) { throtl_upgrade_state(td); |
c79892c55 blk-throttle: add... |
2052 2053 |
goto again; } |
9e660acff blk-throttle: mak... |
2054 |
break; |
c79892c55 blk-throttle: add... |
2055 |
} |
9e660acff blk-throttle: mak... |
2056 2057 |
/* within limits, let's charge and dispatch directly */ |
e43473b7f blkio: Core imple... |
2058 |
throtl_charge_bio(tg, bio); |
04521db04 blk-throttle: Res... |
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 |
/* * We need to trim slice even when bios are not being queued * otherwise it might happen that a bio is not queued for * a long time and slice keeps on extending and trim is not * called for a long time. Now if limits are reduced suddenly * we take into account all the IO dispatched so far at new * low rate and * newly queued IO gets a really long dispatch * time. * * So keep on trimming slice even if bio is not queued. */ |
0f3457f60 blk-throttle: add... |
2071 |
throtl_trim_slice(tg, rw); |
9e660acff blk-throttle: mak... |
2072 2073 2074 2075 2076 2077 |
/* * @bio passed through this layer without being throttled. * Climb up the ladder. If we''re already at the top, it * can be executed directly. */ |
c5cc2070b blk-throttle: add... |
2078 |
qn = &tg->qnode_on_parent[rw]; |
9e660acff blk-throttle: mak... |
2079 2080 2081 2082 |
sq = sq->parent_sq; tg = sq_to_tg(sq); if (!tg) goto out_unlock; |
e43473b7f blkio: Core imple... |
2083 |
} |
9e660acff blk-throttle: mak... |
2084 |
/* out-of-limit, queue to @tg */ |
fda6f272c blk-throttle: imp... |
2085 2086 |
throtl_log(sq, "[%c] bio. bdisp=%llu sz=%u bps=%llu iodisp=%u iops=%u queued=%d/%d", rw == READ ? 'R' : 'W', |
9f626e372 blk-throttle: pre... |
2087 2088 2089 |
tg->bytes_disp[rw], bio->bi_iter.bi_size, tg_bps_limit(tg, rw), tg->io_disp[rw], tg_iops_limit(tg, rw), |
fda6f272c blk-throttle: imp... |
2090 |
sq->nr_queued[READ], sq->nr_queued[WRITE]); |
e43473b7f blkio: Core imple... |
2091 |
|
3f0abd806 blk-throttle: add... |
2092 |
tg->last_low_overflow_time[rw] = jiffies; |
b9147dd1b blk-throttle: add... |
2093 |
td->nr_queued[rw]++; |
c5cc2070b blk-throttle: add... |
2094 |
throtl_add_bio_tg(bio, qn, tg); |
bc16a4f93 block: reorganize... |
2095 |
throttled = true; |
e43473b7f blkio: Core imple... |
2096 |
|
7f52f98c2 blk-throttle: imp... |
2097 2098 2099 2100 2101 2102 |
/* * Update @tg's dispatch time and force schedule dispatch if @tg * was empty before @bio. The forced scheduling isn't likely to * cause undue delay as @bio is likely to be dispatched directly if * its @tg's disptime is not in the future. */ |
0e9f4164b blk-throttle: gen... |
2103 |
if (tg->flags & THROTL_TG_WAS_EMPTY) { |
77216b048 blk-throttle: add... |
2104 |
tg_update_disptime(tg); |
7f52f98c2 blk-throttle: imp... |
2105 |
throtl_schedule_next_dispatch(tg->service_queue.parent_sq, true); |
e43473b7f blkio: Core imple... |
2106 |
} |
bc16a4f93 block: reorganize... |
2107 |
out_unlock: |
e43473b7f blkio: Core imple... |
2108 |
spin_unlock_irq(q->queue_lock); |
bc16a4f93 block: reorganize... |
2109 |
out: |
3ef1c33f9 block-throttle: a... |
2110 |
bio_set_flag(bio, BIO_THROTTLED); |
b9147dd1b blk-throttle: add... |
2111 2112 2113 2114 2115 |
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW if (throttled || !td->track_bio_latency) bio->bi_issue_stat.stat |= SKIP_LATENCY; #endif |
bc16a4f93 block: reorganize... |
2116 |
return throttled; |
e43473b7f blkio: Core imple... |
2117 |
} |
9e234eeaf blk-throttle: add... |
2118 |
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW |
b9147dd1b blk-throttle: add... |
2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 |
static void throtl_track_latency(struct throtl_data *td, sector_t size, int op, unsigned long time) { struct latency_bucket *latency; int index; if (!td || td->limit_index != LIMIT_LOW || op != REQ_OP_READ || !blk_queue_nonrot(td->queue)) return; index = request_bucket_index(size); latency = get_cpu_ptr(td->latency_buckets); latency[index].total_latency += time; latency[index].samples++; put_cpu_ptr(td->latency_buckets); } void blk_throtl_stat_add(struct request *rq, u64 time_ns) { struct request_queue *q = rq->q; struct throtl_data *td = q->td; throtl_track_latency(td, blk_stat_size(&rq->issue_stat), req_op(rq), time_ns >> 10); } |
9e234eeaf blk-throttle: add... |
2145 2146 2147 |
void blk_throtl_bio_endio(struct bio *bio) { struct throtl_grp *tg; |
b9147dd1b blk-throttle: add... |
2148 2149 2150 2151 |
u64 finish_time_ns; unsigned long finish_time; unsigned long start_time; unsigned long lat; |
9e234eeaf blk-throttle: add... |
2152 2153 2154 2155 2156 |
tg = bio->bi_cg_private; if (!tg) return; bio->bi_cg_private = NULL; |
b9147dd1b blk-throttle: add... |
2157 2158 2159 2160 2161 |
finish_time_ns = ktime_get_ns(); tg->last_finish_time = finish_time_ns >> 10; start_time = blk_stat_time(&bio->bi_issue_stat) >> 10; finish_time = __blk_stat_time(finish_time_ns) >> 10; |
53696b8d2 blk-throttle: add... |
2162 2163 2164 2165 |
if (!start_time || finish_time <= start_time) return; lat = finish_time - start_time; |
b9147dd1b blk-throttle: add... |
2166 |
/* this is only for bio based driver */ |
53696b8d2 blk-throttle: add... |
2167 |
if (!(bio->bi_issue_stat.stat & SKIP_LATENCY)) |
b9147dd1b blk-throttle: add... |
2168 2169 |
throtl_track_latency(tg->td, blk_stat_size(&bio->bi_issue_stat), bio_op(bio), lat); |
53696b8d2 blk-throttle: add... |
2170 |
|
6679a90c4 blk-throttle: set... |
2171 |
if (tg->latency_target && lat >= tg->td->filtered_latency) { |
53696b8d2 blk-throttle: add... |
2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 |
int bucket; unsigned int threshold; bucket = request_bucket_index( blk_stat_size(&bio->bi_issue_stat)); threshold = tg->td->avg_buckets[bucket].latency + tg->latency_target; if (lat > threshold) tg->bad_bio_cnt++; /* * Not race free, could get wrong count, which means cgroups * will be throttled */ tg->bio_cnt++; } if (time_after(jiffies, tg->bio_cnt_reset_time) || tg->bio_cnt > 1024) { tg->bio_cnt_reset_time = tg->td->throtl_slice + jiffies; tg->bio_cnt /= 2; tg->bad_bio_cnt /= 2; |
b9147dd1b blk-throttle: add... |
2192 |
} |
9e234eeaf blk-throttle: add... |
2193 2194 |
} #endif |
2a12f0dcd blk-throttle: mak... |
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 |
/* * Dispatch all bios from all children tg's queued on @parent_sq. On * return, @parent_sq is guaranteed to not have any active children tg's * and all bios from previously active tg's are on @parent_sq->bio_lists[]. */ static void tg_drain_bios(struct throtl_service_queue *parent_sq) { struct throtl_grp *tg; while ((tg = throtl_rb_first(parent_sq))) { struct throtl_service_queue *sq = &tg->service_queue; struct bio *bio; throtl_dequeue_tg(tg); |
c5cc2070b blk-throttle: add... |
2209 |
while ((bio = throtl_peek_queued(&sq->queued[READ]))) |
2a12f0dcd blk-throttle: mak... |
2210 |
tg_dispatch_one_bio(tg, bio_data_dir(bio)); |
c5cc2070b blk-throttle: add... |
2211 |
while ((bio = throtl_peek_queued(&sq->queued[WRITE]))) |
2a12f0dcd blk-throttle: mak... |
2212 2213 2214 |
tg_dispatch_one_bio(tg, bio_data_dir(bio)); } } |
c9a929dde block: fix reques... |
2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 |
/** * blk_throtl_drain - drain throttled bios * @q: request_queue to drain throttled bios for * * Dispatch all currently throttled bios on @q through ->make_request_fn(). */ void blk_throtl_drain(struct request_queue *q) __releases(q->queue_lock) __acquires(q->queue_lock) { struct throtl_data *td = q->td; |
2a12f0dcd blk-throttle: mak... |
2225 |
struct blkcg_gq *blkg; |
492eb21b9 cgroup: make hier... |
2226 |
struct cgroup_subsys_state *pos_css; |
c9a929dde block: fix reques... |
2227 |
struct bio *bio; |
651930bc1 blk-throttle: dis... |
2228 |
int rw; |
c9a929dde block: fix reques... |
2229 |
|
8bcb6c7d4 block: use lockde... |
2230 |
queue_lockdep_assert_held(q); |
2a12f0dcd blk-throttle: mak... |
2231 |
rcu_read_lock(); |
c9a929dde block: fix reques... |
2232 |
|
2a12f0dcd blk-throttle: mak... |
2233 2234 2235 2236 2237 2238 |
/* * Drain each tg while doing post-order walk on the blkg tree, so * that all bios are propagated to td->service_queue. It'd be * better to walk service_queue tree directly but blkg walk is * easier. */ |
492eb21b9 cgroup: make hier... |
2239 |
blkg_for_each_descendant_post(blkg, pos_css, td->queue->root_blkg) |
2a12f0dcd blk-throttle: mak... |
2240 |
tg_drain_bios(&blkg_to_tg(blkg)->service_queue); |
73f0d49a9 blk-throttle: mov... |
2241 |
|
2a12f0dcd blk-throttle: mak... |
2242 2243 2244 2245 |
/* finally, transfer bios from top-level tg's into the td */ tg_drain_bios(&td->service_queue); rcu_read_unlock(); |
c9a929dde block: fix reques... |
2246 |
spin_unlock_irq(q->queue_lock); |
2a12f0dcd blk-throttle: mak... |
2247 |
/* all bios now should be in td->service_queue, issue them */ |
651930bc1 blk-throttle: dis... |
2248 |
for (rw = READ; rw <= WRITE; rw++) |
c5cc2070b blk-throttle: add... |
2249 2250 |
while ((bio = throtl_pop_queued(&td->service_queue.queued[rw], NULL))) |
651930bc1 blk-throttle: dis... |
2251 |
generic_make_request(bio); |
c9a929dde block: fix reques... |
2252 2253 2254 |
spin_lock_irq(q->queue_lock); } |
e43473b7f blkio: Core imple... |
2255 2256 2257 |
int blk_throtl_init(struct request_queue *q) { struct throtl_data *td; |
a2b1693ba blkcg: implement ... |
2258 |
int ret; |
e43473b7f blkio: Core imple... |
2259 2260 2261 2262 |
td = kzalloc_node(sizeof(*td), GFP_KERNEL, q->node); if (!td) return -ENOMEM; |
b9147dd1b blk-throttle: add... |
2263 2264 2265 2266 2267 2268 |
td->latency_buckets = __alloc_percpu(sizeof(struct latency_bucket) * LATENCY_BUCKET_SIZE, __alignof__(u64)); if (!td->latency_buckets) { kfree(td); return -ENOMEM; } |
e43473b7f blkio: Core imple... |
2269 |
|
69df0ab03 blk-throttle: sep... |
2270 |
INIT_WORK(&td->dispatch_work, blk_throtl_dispatch_work_fn); |
b2ce2643c blk-throttle: cle... |
2271 |
throtl_service_queue_init(&td->service_queue); |
e43473b7f blkio: Core imple... |
2272 |
|
cd1604fab blkcg: factor out... |
2273 |
q->td = td; |
29b125892 blk-throttle: Dyn... |
2274 |
td->queue = q; |
02977e4af blkio: Add root g... |
2275 |
|
9f626e372 blk-throttle: pre... |
2276 |
td->limit_valid[LIMIT_MAX] = true; |
cd5ab1b0f blk-throttle: add... |
2277 |
td->limit_index = LIMIT_MAX; |
3f0abd806 blk-throttle: add... |
2278 2279 |
td->low_upgrade_time = jiffies; td->low_downgrade_time = jiffies; |
9e234eeaf blk-throttle: add... |
2280 |
|
a2b1693ba blkcg: implement ... |
2281 |
/* activate policy */ |
3c798398e blkcg: mass renam... |
2282 |
ret = blkcg_activate_policy(q, &blkcg_policy_throtl); |
b9147dd1b blk-throttle: add... |
2283 2284 |
if (ret) { free_percpu(td->latency_buckets); |
f51b802c1 blkcg: use the us... |
2285 |
kfree(td); |
b9147dd1b blk-throttle: add... |
2286 |
} |
a2b1693ba blkcg: implement ... |
2287 |
return ret; |
e43473b7f blkio: Core imple... |
2288 2289 2290 2291 |
} void blk_throtl_exit(struct request_queue *q) { |
c875f4d02 blkcg: drop unnec... |
2292 |
BUG_ON(!q->td); |
da5277700 block: Move blk_t... |
2293 |
throtl_shutdown_wq(q); |
3c798398e blkcg: mass renam... |
2294 |
blkcg_deactivate_policy(q, &blkcg_policy_throtl); |
b9147dd1b blk-throttle: add... |
2295 |
free_percpu(q->td->latency_buckets); |
c9a929dde block: fix reques... |
2296 |
kfree(q->td); |
e43473b7f blkio: Core imple... |
2297 |
} |
d61fcfa4b blk-throttle: cho... |
2298 2299 2300 |
void blk_throtl_register_queue(struct request_queue *q) { struct throtl_data *td; |
6679a90c4 blk-throttle: set... |
2301 |
int i; |
d61fcfa4b blk-throttle: cho... |
2302 2303 2304 |
td = q->td; BUG_ON(!td); |
6679a90c4 blk-throttle: set... |
2305 |
if (blk_queue_nonrot(q)) { |
d61fcfa4b blk-throttle: cho... |
2306 |
td->throtl_slice = DFL_THROTL_SLICE_SSD; |
6679a90c4 blk-throttle: set... |
2307 2308 |
td->filtered_latency = LATENCY_FILTERED_SSD; } else { |
d61fcfa4b blk-throttle: cho... |
2309 |
td->throtl_slice = DFL_THROTL_SLICE_HD; |
6679a90c4 blk-throttle: set... |
2310 2311 2312 2313 |
td->filtered_latency = LATENCY_FILTERED_HD; for (i = 0; i < LATENCY_BUCKET_SIZE; i++) td->avg_buckets[i].latency = DFL_HD_BASELINE_LATENCY; } |
d61fcfa4b blk-throttle: cho... |
2314 2315 2316 2317 |
#ifndef CONFIG_BLK_DEV_THROTTLING_LOW /* if no low limit, use previous default */ td->throtl_slice = DFL_THROTL_SLICE_HD; #endif |
9e234eeaf blk-throttle: add... |
2318 |
|
b9147dd1b blk-throttle: add... |
2319 2320 2321 |
td->track_bio_latency = !q->mq_ops && !q->request_fn; if (!td->track_bio_latency) blk_stat_enable_accounting(q); |
d61fcfa4b blk-throttle: cho... |
2322 |
} |
297e3d854 blk-throttle: mak... |
2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 |
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW ssize_t blk_throtl_sample_time_show(struct request_queue *q, char *page) { if (!q->td) return -EINVAL; return sprintf(page, "%u ", jiffies_to_msecs(q->td->throtl_slice)); } ssize_t blk_throtl_sample_time_store(struct request_queue *q, const char *page, size_t count) { unsigned long v; unsigned long t; if (!q->td) return -EINVAL; if (kstrtoul(page, 10, &v)) return -EINVAL; t = msecs_to_jiffies(v); if (t == 0 || t > MAX_THROTL_SLICE) return -EINVAL; q->td->throtl_slice = t; return count; } #endif |
e43473b7f blkio: Core imple... |
2349 2350 |
static int __init throtl_init(void) { |
450adcbe5 blk-throttle: Do ... |
2351 2352 2353 2354 |
kthrotld_workqueue = alloc_workqueue("kthrotld", WQ_MEM_RECLAIM, 0); if (!kthrotld_workqueue) panic("Failed to create kthrotld "); |
3c798398e blkcg: mass renam... |
2355 |
return blkcg_policy_register(&blkcg_policy_throtl); |
e43473b7f blkio: Core imple... |
2356 2357 2358 |
} module_init(throtl_init); |