block/cfq-iosched.c

   1 /*
   2  *  CFQ, or complete fairness queueing, disk scheduler.
   3  *
   4  *  Based on ideas from a previously unfinished io
   5  *  scheduler (round robin per-process disk scheduling) and Andrea Arcangeli.
   6  *
   7  *  Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
   8  */
   9 #include <linux/module.h>
  10 #include <linux/slab.h>
  11 #include <linux/sched/clock.h>
  12 #include <linux/blkdev.h>
  13 #include <linux/elevator.h>
  14 #include <linux/ktime.h>
  15 #include <linux/rbtree.h>
  16 #include <linux/ioprio.h>
  17 #include <linux/blktrace_api.h>
  18 #include <linux/blk-cgroup.h>
  19 #include "blk.h"
  20 #include "blk-wbt.h"
  21
  22 /*
  23  * tunables
  24  */
  25 /* max queue in one round of service */
  26 static const int cfq_quantum = 8;
  27 static const u64 cfq_fifo_expire[2] = { NSEC_PER_SEC / 4, NSEC_PER_SEC / 8 };
  28 /* maximum backwards seek, in KiB */
  29 static const int cfq_back_max = 16 * 1024;
  30 /* penalty of a backwards seek */
  31 static const int cfq_back_penalty = 2;
  32 static const u64 cfq_slice_sync = NSEC_PER_SEC / 10;
  33 static u64 cfq_slice_async = NSEC_PER_SEC / 25;
  34 static const int cfq_slice_async_rq = 2;
  35 static u64 cfq_slice_idle = NSEC_PER_SEC / 125;
  36 static u64 cfq_group_idle = NSEC_PER_SEC / 125;
  37 static const u64 cfq_target_latency = (u64)NSEC_PER_SEC * 3/10; /* 300 ms */
  38 static const int cfq_hist_divisor = 4;
  39
  40 /*
  41  * offset from end of queue service tree for idle class
  42  */
  43 #define CFQ_IDLE_DELAY          (NSEC_PER_SEC / 5)
  44 /* offset from end of group service tree under time slice mode */
  45 #define CFQ_SLICE_MODE_GROUP_DELAY (NSEC_PER_SEC / 5)
  46 /* offset from end of group service under IOPS mode */
  47 #define CFQ_IOPS_MODE_GROUP_DELAY (HZ / 5)
  48
  49 /*
  50  * below this threshold, we consider thinktime immediate
  51  */
  52 #define CFQ_MIN_TT              (2 * NSEC_PER_SEC / HZ)
  53
  54 #define CFQ_SLICE_SCALE         (5)
  55 #define CFQ_HW_QUEUE_MIN        (5)
  56 #define CFQ_SERVICE_SHIFT       12
  57
  58 #define CFQQ_SEEK_THR           (sector_t)(8 * 100)
  59 #define CFQQ_CLOSE_THR          (sector_t)(8 * 1024)
  60 #define CFQQ_SECT_THR_NONROT    (sector_t)(2 * 32)
  61 #define CFQQ_SEEKY(cfqq)        (hweight32(cfqq->seek_history) > 32/8)
  62
  63 #define RQ_CIC(rq)              icq_to_cic((rq)->elv.icq)
  64 #define RQ_CFQQ(rq)             (struct cfq_queue *) ((rq)->elv.priv[0])
  65 #define RQ_CFQG(rq)             (struct cfq_group *) ((rq)->elv.priv[1])
  66
  67 static struct kmem_cache *cfq_pool;
  68
  69 #define CFQ_PRIO_LISTS          IOPRIO_BE_NR
  70 #define cfq_class_idle(cfqq)    ((cfqq)->ioprio_class == IOPRIO_CLASS_IDLE)
  71 #define cfq_class_rt(cfqq)      ((cfqq)->ioprio_class == IOPRIO_CLASS_RT)
  72
  73 #define sample_valid(samples)   ((samples) > 80)
  74 #define rb_entry_cfqg(node)     rb_entry((node), struct cfq_group, rb_node)
  75
  76 /* blkio-related constants */
  77 #define CFQ_WEIGHT_LEGACY_MIN   10
  78 #define CFQ_WEIGHT_LEGACY_DFL   500
  79 #define CFQ_WEIGHT_LEGACY_MAX   1000
  80
  81 struct cfq_ttime {
  82         u64 last_end_request;
  83
  84         u64 ttime_total;
  85         u64 ttime_mean;
  86         unsigned long ttime_samples;
  87 };
  88
  89 /*
  90  * Most of our rbtree usage is for sorting with min extraction, so
  91  * if we cache the leftmost node we don't have to walk down the tree
  92  * to find it. Idea borrowed from Ingo Molnars CFS scheduler. We should
  93  * move this into the elevator for the rq sorting as well.
  94  */
  95 struct cfq_rb_root {
  96         struct rb_root_cached rb;
  97         struct rb_node *rb_rightmost;
  98         unsigned count;
  99         u64 min_vdisktime;
 100         struct cfq_ttime ttime;
 101 };
 102 #define CFQ_RB_ROOT     (struct cfq_rb_root) { .rb = RB_ROOT_CACHED, \
 103                         .rb_rightmost = NULL,                        \
 104                         .ttime = {.last_end_request = ktime_get_ns(),},}
 105
 106 /*
 107  * Per process-grouping structure
 108  */
 109 struct cfq_queue {
 110         /* reference count */
 111         int ref;
 112         /* various state flags, see below */
 113         unsigned int flags;
 114         /* parent cfq_data */
 115         struct cfq_data *cfqd;
 116         /* service_tree member */
 117         struct rb_node rb_node;
 118         /* service_tree key */
 119         u64 rb_key;
 120         /* prio tree member */
 121         struct rb_node p_node;
 122         /* prio tree root we belong to, if any */
 123         struct rb_root *p_root;
 124         /* sorted list of pending requests */
 125         struct rb_root sort_list;
 126         /* if fifo isn't expired, next request to serve */
 127         struct request *next_rq;
 128         /* requests queued in sort_list */
 129         int queued[2];
 130         /* currently allocated requests */
 131         int allocated[2];
 132         /* fifo list of requests in sort_list */
 133         struct list_head fifo;
 134
 135         /* time when queue got scheduled in to dispatch first request. */
 136         u64 dispatch_start;
 137         u64 allocated_slice;
 138         u64 slice_dispatch;
 139         /* time when first request from queue completed and slice started. */
 140         u64 slice_start;
 141         u64 slice_end;
 142         s64 slice_resid;
 143
 144         /* pending priority requests */
 145         int prio_pending;
 146         /* number of requests that are on the dispatch list or inside driver */
 147         int dispatched;
 148
 149         /* io prio of this group */
 150         unsigned short ioprio, org_ioprio;
 151         unsigned short ioprio_class, org_ioprio_class;
 152
 153         pid_t pid;
 154
 155         u32 seek_history;
 156         sector_t last_request_pos;
 157
 158         struct cfq_rb_root *service_tree;
 159         struct cfq_queue *new_cfqq;
 160         struct cfq_group *cfqg;
 161         /* Number of sectors dispatched from queue in single dispatch round */
 162         unsigned long nr_sectors;
 163 };
 164
 165 /*
 166  * First index in the service_trees.
 167  * IDLE is handled separately, so it has negative index
 168  */
 169 enum wl_class_t {
 170         BE_WORKLOAD = 0,
 171         RT_WORKLOAD = 1,
 172         IDLE_WORKLOAD = 2,
 173         CFQ_PRIO_NR,
 174 };
 175
 176 /*
 177  * Second index in the service_trees.
 178  */
 179 enum wl_type_t {
 180         ASYNC_WORKLOAD = 0,
 181         SYNC_NOIDLE_WORKLOAD = 1,
 182         SYNC_WORKLOAD = 2
 183 };
 184
 185 struct cfqg_stats {
 186 #ifdef CONFIG_CFQ_GROUP_IOSCHED
 187         /* number of ios merged */
 188         struct blkg_rwstat              merged;
 189         /* total time spent on device in ns, may not be accurate w/ queueing */
 190         struct blkg_rwstat              service_time;
 191         /* total time spent waiting in scheduler queue in ns */
 192         struct blkg_rwstat              wait_time;
 193         /* number of IOs queued up */
 194         struct blkg_rwstat              queued;
 195         /* total disk time and nr sectors dispatched by this group */
 196         struct blkg_stat                time;
 197 #ifdef CONFIG_DEBUG_BLK_CGROUP
 198         /* time not charged to this cgroup */
 199         struct blkg_stat                unaccounted_time;
 200         /* sum of number of ios queued across all samples */
 201         struct blkg_stat                avg_queue_size_sum;
 202         /* count of samples taken for average */
 203         struct blkg_stat                avg_queue_size_samples;
 204         /* how many times this group has been removed from service tree */
 205         struct blkg_stat                dequeue;
 206         /* total time spent waiting for it to be assigned a timeslice. */
 207         struct blkg_stat                group_wait_time;
 208         /* time spent idling for this blkcg_gq */
 209         struct blkg_stat                idle_time;
 210         /* total time with empty current active q with other requests queued */
 211         struct blkg_stat                empty_time;
 212         /* fields after this shouldn't be cleared on stat reset */
 213         u64                             start_group_wait_time;
 214         u64                             start_idle_time;
 215         u64                             start_empty_time;
 216         uint16_t                        flags;
 217 #endif  /* CONFIG_DEBUG_BLK_CGROUP */
 218 #endif  /* CONFIG_CFQ_GROUP_IOSCHED */
 219 };
 220
 221 /* Per-cgroup data */
 222 struct cfq_group_data {
 223         /* must be the first member */
 224         struct blkcg_policy_data cpd;
 225
 226         unsigned int weight;
 227         unsigned int leaf_weight;
 228 };
 229
 230 /* This is per cgroup per device grouping structure */
 231 struct cfq_group {
 232         /* must be the first member */
 233         struct blkg_policy_data pd;
 234
 235         /* group service_tree member */
 236         struct rb_node rb_node;
 237
 238         /* group service_tree key */
 239         u64 vdisktime;
 240
 241         /*
 242          * The number of active cfqgs and sum of their weights under this
 243          * cfqg.  This covers this cfqg's leaf_weight and all children's
 244          * weights, but does not cover weights of further descendants.
 245          *
 246          * If a cfqg is on the service tree, it's active.  An active cfqg
 247          * also activates its parent and contributes to the children_weight
 248          * of the parent.
 249          */
 250         int nr_active;
 251         unsigned int children_weight;
 252
 253         /*
 254          * vfraction is the fraction of vdisktime that the tasks in this
 255          * cfqg are entitled to.  This is determined by compounding the
 256          * ratios walking up from this cfqg to the root.
 257          *
 258          * It is in fixed point w/ CFQ_SERVICE_SHIFT and the sum of all
 259          * vfractions on a service tree is approximately 1.  The sum may
 260          * deviate a bit due to rounding errors and fluctuations caused by
 261          * cfqgs entering and leaving the service tree.
 262          */
 263         unsigned int vfraction;
 264
 265         /*
 266          * There are two weights - (internal) weight is the weight of this
 267          * cfqg against the sibling cfqgs.  leaf_weight is the wight of
 268          * this cfqg against the child cfqgs.  For the root cfqg, both
 269          * weights are kept in sync for backward compatibility.
 270          */
 271         unsigned int weight;
 272         unsigned int new_weight;
 273         unsigned int dev_weight;
 274
 275         unsigned int leaf_weight;
 276         unsigned int new_leaf_weight;
 277         unsigned int dev_leaf_weight;
 278
 279         /* number of cfqq currently on this group */
 280         int nr_cfqq;
 281
 282         /*
 283          * Per group busy queues average. Useful for workload slice calc. We
 284          * create the array for each prio class but at run time it is used
 285          * only for RT and BE class and slot for IDLE class remains unused.
 286          * This is primarily done to avoid confusion and a gcc warning.
 287          */
 288         unsigned int busy_queues_avg[CFQ_PRIO_NR];
 289         /*
 290          * rr lists of queues with requests. We maintain service trees for
 291          * RT and BE classes. These trees are subdivided in subclasses
 292          * of SYNC, SYNC_NOIDLE and ASYNC based on workload type. For IDLE
 293          * class there is no subclassification and all the cfq queues go on
 294          * a single tree service_tree_idle.
 295          * Counts are embedded in the cfq_rb_root
 296          */
 297         struct cfq_rb_root service_trees[2][3];
 298         struct cfq_rb_root service_tree_idle;
 299
 300         u64 saved_wl_slice;
 301         enum wl_type_t saved_wl_type;
 302         enum wl_class_t saved_wl_class;
 303
 304         /* number of requests that are on the dispatch list or inside driver */
 305         int dispatched;
 306         struct cfq_ttime ttime;
 307         struct cfqg_stats stats;        /* stats for this cfqg */
 308
 309         /* async queue for each priority case */
 310         struct cfq_queue *async_cfqq[2][IOPRIO_BE_NR];
 311         struct cfq_queue *async_idle_cfqq;
 312
 313 };
 314
 315 struct cfq_io_cq {
 316         struct io_cq            icq;            /* must be the first member */
 317         struct cfq_queue        *cfqq[2];
 318         struct cfq_ttime        ttime;
 319         int                     ioprio;         /* the current ioprio */
 320 #ifdef CONFIG_CFQ_GROUP_IOSCHED
 321         uint64_t                blkcg_serial_nr; /* the current blkcg serial */
 322 #endif
 323 };
 324
 325 /*
 326  * Per block device queue structure
 327  */
 328 struct cfq_data {
 329         struct request_queue *queue;
 330         /* Root service tree for cfq_groups */
 331         struct cfq_rb_root grp_service_tree;
 332         struct cfq_group *root_group;
 333
 334         /*
 335          * The priority currently being served
 336          */
 337         enum wl_class_t serving_wl_class;
 338         enum wl_type_t serving_wl_type;
 339         u64 workload_expires;
 340         struct cfq_group *serving_group;
 341
 342         /*
 343          * Each priority tree is sorted by next_request position.  These
 344          * trees are used when determining if two or more queues are
 345          * interleaving requests (see cfq_close_cooperator).
 346          */
 347         struct rb_root prio_trees[CFQ_PRIO_LISTS];
 348
 349         unsigned int busy_queues;
 350         unsigned int busy_sync_queues;
 351
 352         int rq_in_driver;
 353         int rq_in_flight[2];
 354
 355         /*
 356          * queue-depth detection
 357          */
 358         int rq_queued;
 359         int hw_tag;
 360         /*
 361          * hw_tag can be
 362          * -1 => indeterminate, (cfq will behave as if NCQ is present, to allow better detection)
 363          *  1 => NCQ is present (hw_tag_est_depth is the estimated max depth)
 364          *  0 => no NCQ
 365          */
 366         int hw_tag_est_depth;
 367         unsigned int hw_tag_samples;
 368
 369         /*
 370          * idle window management
 371          */
 372         struct hrtimer idle_slice_timer;
 373         struct work_struct unplug_work;
 374
 375         struct cfq_queue *active_queue;
 376         struct cfq_io_cq *active_cic;
 377
 378         sector_t last_position;
 379
 380         /*
 381          * tunables, see top of file
 382          */
 383         unsigned int cfq_quantum;
 384         unsigned int cfq_back_penalty;
 385         unsigned int cfq_back_max;
 386         unsigned int cfq_slice_async_rq;
 387         unsigned int cfq_latency;
 388         u64 cfq_fifo_expire[2];
 389         u64 cfq_slice[2];
 390         u64 cfq_slice_idle;
 391         u64 cfq_group_idle;
 392         u64 cfq_target_latency;
 393
 394         /*
 395          * Fallback dummy cfqq for extreme OOM conditions
 396          */
 397         struct cfq_queue oom_cfqq;
 398
 399         u64 last_delayed_sync;
 400 };
 401
 402 static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd);
 403 static void cfq_put_queue(struct cfq_queue *cfqq);
 404
 405 static struct cfq_rb_root *st_for(struct cfq_group *cfqg,
 406                                             enum wl_class_t class,
 407                                             enum wl_type_t type)
 408 {
 409         if (!cfqg)
 410                 return NULL;
 411
 412         if (class == IDLE_WORKLOAD)
 413                 return &cfqg->service_tree_idle;
 414
 415         return &cfqg->service_trees[class][type];
 416 }
 417
 418 enum cfqq_state_flags {
 419         CFQ_CFQQ_FLAG_on_rr = 0,        /* on round-robin busy list */
 420         CFQ_CFQQ_FLAG_wait_request,     /* waiting for a request */
 421         CFQ_CFQQ_FLAG_must_dispatch,    /* must be allowed a dispatch */
 422         CFQ_CFQQ_FLAG_must_alloc_slice, /* per-slice must_alloc flag */
 423         CFQ_CFQQ_FLAG_fifo_expire,      /* FIFO checked in this slice */
 424         CFQ_CFQQ_FLAG_idle_window,      /* slice idling enabled */
 425         CFQ_CFQQ_FLAG_prio_changed,     /* task priority has changed */
 426         CFQ_CFQQ_FLAG_slice_new,        /* no requests dispatched in slice */
 427         CFQ_CFQQ_FLAG_sync,             /* synchronous queue */
 428         CFQ_CFQQ_FLAG_coop,             /* cfqq is shared */
 429         CFQ_CFQQ_FLAG_split_coop,       /* shared cfqq will be splitted */
 430         CFQ_CFQQ_FLAG_deep,             /* sync cfqq experienced large depth */
 431         CFQ_CFQQ_FLAG_wait_busy,        /* Waiting for next request */
 432 };
 433
 434 #define CFQ_CFQQ_FNS(name)                                              \
 435 static inline void cfq_mark_cfqq_##name(struct cfq_queue *cfqq)         \
 436 {                                                                       \
 437         (cfqq)->flags |= (1 << CFQ_CFQQ_FLAG_##name);                   \
 438 }                                                                       \
 439 static inline void cfq_clear_cfqq_##name(struct cfq_queue *cfqq)        \
 440 {                                                                       \
 441         (cfqq)->flags &= ~(1 << CFQ_CFQQ_FLAG_##name);                  \
 442 }                                                                       \
 443 static inline int cfq_cfqq_##name(const struct cfq_queue *cfqq)         \
 444 {                                                                       \
 445         return ((cfqq)->flags & (1 << CFQ_CFQQ_FLAG_##name)) != 0;      \
 446 }
 447
 448 CFQ_CFQQ_FNS(on_rr);
 449 CFQ_CFQQ_FNS(wait_request);
 450 CFQ_CFQQ_FNS(must_dispatch);
 451 CFQ_CFQQ_FNS(must_alloc_slice);
 452 CFQ_CFQQ_FNS(fifo_expire);
 453 CFQ_CFQQ_FNS(idle_window);
 454 CFQ_CFQQ_FNS(prio_changed);
 455 CFQ_CFQQ_FNS(slice_new);
 456 CFQ_CFQQ_FNS(sync);
 457 CFQ_CFQQ_FNS(coop);
 458 CFQ_CFQQ_FNS(split_coop);
 459 CFQ_CFQQ_FNS(deep);
 460 CFQ_CFQQ_FNS(wait_busy);
 461 #undef CFQ_CFQQ_FNS
 462
 463 #if defined(CONFIG_CFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
 464
 465 /* cfqg stats flags */
 466 enum cfqg_stats_flags {
 467         CFQG_stats_waiting = 0,
 468         CFQG_stats_idling,
 469         CFQG_stats_empty,
 470 };
 471
 472 #define CFQG_FLAG_FNS(name)                                             \
 473 static inline void cfqg_stats_mark_##name(struct cfqg_stats *stats)     \
 474 {                                                                       \
 475         stats->flags |= (1 << CFQG_stats_##name);                       \
 476 }                                                                       \
 477 static inline void cfqg_stats_clear_##name(struct cfqg_stats *stats)    \
 478 {                                                                       \
 479         stats->flags &= ~(1 << CFQG_stats_##name);                      \
 480 }                                                                       \
 481 static inline int cfqg_stats_##name(struct cfqg_stats *stats)           \
 482 {                                                                       \
 483         return (stats->flags & (1 << CFQG_stats_##name)) != 0;          \
 484 }                                                                       \
 485
 486 CFQG_FLAG_FNS(waiting)
 487 CFQG_FLAG_FNS(idling)
 488 CFQG_FLAG_FNS(empty)
 489 #undef CFQG_FLAG_FNS
 490
 491 /* This should be called with the queue_lock held. */
 492 static void cfqg_stats_update_group_wait_time(struct cfqg_stats *stats)
 493 {
 494         u64 now;
 495
 496         if (!cfqg_stats_waiting(stats))
 497                 return;
 498
 499         now = ktime_get_ns();
 500         if (now > stats->start_group_wait_time)
 501                 blkg_stat_add(&stats->group_wait_time,
 502                               now - stats->start_group_wait_time);
 503         cfqg_stats_clear_waiting(stats);
 504 }
 505
 506 /* This should be called with the queue_lock held. */
 507 static void cfqg_stats_set_start_group_wait_time(struct cfq_group *cfqg,
 508                                                  struct cfq_group *curr_cfqg)
 509 {
 510         struct cfqg_stats *stats = &cfqg->stats;
 511
 512         if (cfqg_stats_waiting(stats))
 513                 return;
 514         if (cfqg == curr_cfqg)
 515                 return;
 516         stats->start_group_wait_time = ktime_get_ns();
 517         cfqg_stats_mark_waiting(stats);
 518 }
 519
 520 /* This should be called with the queue_lock held. */
 521 static void cfqg_stats_end_empty_time(struct cfqg_stats *stats)
 522 {
 523         u64 now;
 524
 525         if (!cfqg_stats_empty(stats))
 526                 return;
 527
 528         now = ktime_get_ns();
 529         if (now > stats->start_empty_time)
 530                 blkg_stat_add(&stats->empty_time,
 531                               now - stats->start_empty_time);
 532         cfqg_stats_clear_empty(stats);
 533 }
 534
 535 static void cfqg_stats_update_dequeue(struct cfq_group *cfqg)
 536 {
 537         blkg_stat_add(&cfqg->stats.dequeue, 1);
 538 }
 539
 540 static void cfqg_stats_set_start_empty_time(struct cfq_group *cfqg)
 541 {
 542         struct cfqg_stats *stats = &cfqg->stats;
 543
 544         if (blkg_rwstat_total(&stats->queued))
 545                 return;
 546
 547         /*
 548          * group is already marked empty. This can happen if cfqq got new
 549          * request in parent group and moved to this group while being added
 550          * to service tree. Just ignore the event and move on.
 551          */
 552         if (cfqg_stats_empty(stats))
 553                 return;
 554
 555         stats->start_empty_time = ktime_get_ns();
 556         cfqg_stats_mark_empty(stats);
 557 }
 558
 559 static void cfqg_stats_update_idle_time(struct cfq_group *cfqg)
 560 {
 561         struct cfqg_stats *stats = &cfqg->stats;
 562
 563         if (cfqg_stats_idling(stats)) {
 564                 u64 now = ktime_get_ns();
 565
 566                 if (now > stats->start_idle_time)
 567                         blkg_stat_add(&stats->idle_time,
 568                                       now - stats->start_idle_time);
 569                 cfqg_stats_clear_idling(stats);
 570         }
 571 }
 572
 573 static void cfqg_stats_set_start_idle_time(struct cfq_group *cfqg)
 574 {
 575         struct cfqg_stats *stats = &cfqg->stats;
 576
 577         BUG_ON(cfqg_stats_idling(stats));
 578
 579         stats->start_idle_time = ktime_get_ns();
 580         cfqg_stats_mark_idling(stats);
 581 }
 582
 583 static void cfqg_stats_update_avg_queue_size(struct cfq_group *cfqg)
 584 {
 585         struct cfqg_stats *stats = &cfqg->stats;
 586
 587         blkg_stat_add(&stats->avg_queue_size_sum,
 588                       blkg_rwstat_total(&stats->queued));
 589         blkg_stat_add(&stats->avg_queue_size_samples, 1);
 590         cfqg_stats_update_group_wait_time(stats);
 591 }
 592
 593 #else   /* CONFIG_CFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
 594
 595 static inline void cfqg_stats_set_start_group_wait_time(struct cfq_group *cfqg, struct cfq_group *curr_cfqg) { }
 596 static inline void cfqg_stats_end_empty_time(struct cfqg_stats *stats) { }
 597 static inline void cfqg_stats_update_dequeue(struct cfq_group *cfqg) { }
 598 static inline void cfqg_stats_set_start_empty_time(struct cfq_group *cfqg) { }
 599 static inline void cfqg_stats_update_idle_time(struct cfq_group *cfqg) { }
 600 static inline void cfqg_stats_set_start_idle_time(struct cfq_group *cfqg) { }
 601 static inline void cfqg_stats_update_avg_queue_size(struct cfq_group *cfqg) { }
 602
 603 #endif  /* CONFIG_CFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
 604
 605 #ifdef CONFIG_CFQ_GROUP_IOSCHED
 606
 607 static inline struct cfq_group *pd_to_cfqg(struct blkg_policy_data *pd)
 608 {
 609         return pd ? container_of(pd, struct cfq_group, pd) : NULL;
 610 }
 611
 612 static struct cfq_group_data
 613 *cpd_to_cfqgd(struct blkcg_policy_data *cpd)
 614 {
 615         return cpd ? container_of(cpd, struct cfq_group_data, cpd) : NULL;
 616 }
 617
 618 static inline struct blkcg_gq *cfqg_to_blkg(struct cfq_group *cfqg)
 619 {
 620         return pd_to_blkg(&cfqg->pd);
 621 }
 622
 623 static struct blkcg_policy blkcg_policy_cfq;
 624
 625 static inline struct cfq_group *blkg_to_cfqg(struct blkcg_gq *blkg)
 626 {
 627         return pd_to_cfqg(blkg_to_pd(blkg, &blkcg_policy_cfq));
 628 }
 629
 630 static struct cfq_group_data *blkcg_to_cfqgd(struct blkcg *blkcg)
 631 {
 632         return cpd_to_cfqgd(blkcg_to_cpd(blkcg, &blkcg_policy_cfq));
 633 }
 634
 635 static inline struct cfq_group *cfqg_parent(struct cfq_group *cfqg)
 636 {
 637         struct blkcg_gq *pblkg = cfqg_to_blkg(cfqg)->parent;
 638
 639         return pblkg ? blkg_to_cfqg(pblkg) : NULL;
 640 }
 641
 642 static inline bool cfqg_is_descendant(struct cfq_group *cfqg,
 643                                       struct cfq_group *ancestor)
 644 {
 645         return cgroup_is_descendant(cfqg_to_blkg(cfqg)->blkcg->css.cgroup,
 646                                     cfqg_to_blkg(ancestor)->blkcg->css.cgroup);
 647 }
 648
 649 static inline void cfqg_get(struct cfq_group *cfqg)
 650 {
 651         return blkg_get(cfqg_to_blkg(cfqg));
 652 }
 653
 654 static inline void cfqg_put(struct cfq_group *cfqg)
 655 {
 656         return blkg_put(cfqg_to_blkg(cfqg));
 657 }
 658
 659 #define cfq_log_cfqq(cfqd, cfqq, fmt, args...)  do {                    \
 660         blk_add_cgroup_trace_msg((cfqd)->queue,                         \
 661                         cfqg_to_blkg((cfqq)->cfqg)->blkcg,              \
 662                         "cfq%d%c%c " fmt, (cfqq)->pid,                  \
 663                         cfq_cfqq_sync((cfqq)) ? 'S' : 'A',              \
 664                         cfqq_type((cfqq)) == SYNC_NOIDLE_WORKLOAD ? 'N' : ' ',\
 665                           ##args);                                      \
 666 } while (0)
 667
 668 #define cfq_log_cfqg(cfqd, cfqg, fmt, args...)  do {                    \
 669         blk_add_cgroup_trace_msg((cfqd)->queue,                         \
 670                         cfqg_to_blkg(cfqg)->blkcg, fmt, ##args);        \
 671 } while (0)
 672
 673 static inline void cfqg_stats_update_io_add(struct cfq_group *cfqg,
 674                                             struct cfq_group *curr_cfqg,
 675                                             unsigned int op)
 676 {
 677         blkg_rwstat_add(&cfqg->stats.queued, op, 1);
 678         cfqg_stats_end_empty_time(&cfqg->stats);
 679         cfqg_stats_set_start_group_wait_time(cfqg, curr_cfqg);
 680 }
 681
 682 static inline void cfqg_stats_update_timeslice_used(struct cfq_group *cfqg,
 683                         uint64_t time, unsigned long unaccounted_time)
 684 {
 685         blkg_stat_add(&cfqg->stats.time, time);
 686 #ifdef CONFIG_DEBUG_BLK_CGROUP
 687         blkg_stat_add(&cfqg->stats.unaccounted_time, unaccounted_time);
 688 #endif
 689 }
 690
 691 static inline void cfqg_stats_update_io_remove(struct cfq_group *cfqg,
 692                                                unsigned int op)
 693 {
 694         blkg_rwstat_add(&cfqg->stats.queued, op, -1);
 695 }
 696
 697 static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg,
 698                                                unsigned int op)
 699 {
 700         blkg_rwstat_add(&cfqg->stats.merged, op, 1);
 701 }
 702
 703 static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
 704                                                 u64 start_time_ns,
 705                                                 u64 io_start_time_ns,
 706                                                 unsigned int op)
 707 {
 708         struct cfqg_stats *stats = &cfqg->stats;
 709         u64 now = ktime_get_ns();
 710
 711         if (now > io_start_time_ns)
 712                 blkg_rwstat_add(&stats->service_time, op,
 713                                 now - io_start_time_ns);
 714         if (io_start_time_ns > start_time_ns)
 715                 blkg_rwstat_add(&stats->wait_time, op,
 716                                 io_start_time_ns - start_time_ns);
 717 }
 718
 719 /* @stats = 0 */
 720 static void cfqg_stats_reset(struct cfqg_stats *stats)
 721 {
 722         /* queued stats shouldn't be cleared */
 723         blkg_rwstat_reset(&stats->merged);
 724         blkg_rwstat_reset(&stats->service_time);
 725         blkg_rwstat_reset(&stats->wait_time);
 726         blkg_stat_reset(&stats->time);
 727 #ifdef CONFIG_DEBUG_BLK_CGROUP
 728         blkg_stat_reset(&stats->unaccounted_time);
 729         blkg_stat_reset(&stats->avg_queue_size_sum);
 730         blkg_stat_reset(&stats->avg_queue_size_samples);
 731         blkg_stat_reset(&stats->dequeue);
 732         blkg_stat_reset(&stats->group_wait_time);
 733         blkg_stat_reset(&stats->idle_time);
 734         blkg_stat_reset(&stats->empty_time);
 735 #endif
 736 }
 737
 738 /* @to += @from */
 739 static void cfqg_stats_add_aux(struct cfqg_stats *to, struct cfqg_stats *from)
 740 {
 741         /* queued stats shouldn't be cleared */
 742         blkg_rwstat_add_aux(&to->merged, &from->merged);
 743         blkg_rwstat_add_aux(&to->service_time, &from->service_time);
 744         blkg_rwstat_add_aux(&to->wait_time, &from->wait_time);
 745         blkg_stat_add_aux(&from->time, &from->time);
 746 #ifdef CONFIG_DEBUG_BLK_CGROUP
 747         blkg_stat_add_aux(&to->unaccounted_time, &from->unaccounted_time);
 748         blkg_stat_add_aux(&to->avg_queue_size_sum, &from->avg_queue_size_sum);
 749         blkg_stat_add_aux(&to->avg_queue_size_samples, &from->avg_queue_size_samples);
 750         blkg_stat_add_aux(&to->dequeue, &from->dequeue);
 751         blkg_stat_add_aux(&to->group_wait_time, &from->group_wait_time);
 752         blkg_stat_add_aux(&to->idle_time, &from->idle_time);
 753         blkg_stat_add_aux(&to->empty_time, &from->empty_time);
 754 #endif
 755 }
 756
 757 /*
 758  * Transfer @cfqg's stats to its parent's aux counts so that the ancestors'
 759  * recursive stats can still account for the amount used by this cfqg after
 760  * it's gone.
 761  */
 762 static void cfqg_stats_xfer_dead(struct cfq_group *cfqg)
 763 {
 764         struct cfq_group *parent = cfqg_parent(cfqg);
 765
 766         lockdep_assert_held(cfqg_to_blkg(cfqg)->q->queue_lock);
 767
 768         if (unlikely(!parent))
 769                 return;
 770
 771         cfqg_stats_add_aux(&parent->stats, &cfqg->stats);
 772         cfqg_stats_reset(&cfqg->stats);
 773 }
 774
 775 #else   /* CONFIG_CFQ_GROUP_IOSCHED */
 776
 777 static inline struct cfq_group *cfqg_parent(struct cfq_group *cfqg) { return NULL; }
 778 static inline bool cfqg_is_descendant(struct cfq_group *cfqg,
 779                                       struct cfq_group *ancestor)
 780 {
 781         return true;
 782 }
 783 static inline void cfqg_get(struct cfq_group *cfqg) { }
 784 static inline void cfqg_put(struct cfq_group *cfqg) { }
 785
 786 #define cfq_log_cfqq(cfqd, cfqq, fmt, args...)  \
 787         blk_add_trace_msg((cfqd)->queue, "cfq%d%c%c " fmt, (cfqq)->pid, \
 788                         cfq_cfqq_sync((cfqq)) ? 'S' : 'A',              \
 789                         cfqq_type((cfqq)) == SYNC_NOIDLE_WORKLOAD ? 'N' : ' ',\
 790                                 ##args)
 791 #define cfq_log_cfqg(cfqd, cfqg, fmt, args...)          do {} while (0)
 792
 793 static inline void cfqg_stats_update_io_add(struct cfq_group *cfqg,
 794                         struct cfq_group *curr_cfqg, unsigned int op) { }
 795 static inline void cfqg_stats_update_timeslice_used(struct cfq_group *cfqg,
 796                         uint64_t time, unsigned long unaccounted_time) { }
 797 static inline void cfqg_stats_update_io_remove(struct cfq_group *cfqg,
 798                         unsigned int op) { }
 799 static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg,
 800                         unsigned int op) { }
 801 static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
 802                                                 u64 start_time_ns,
 803                                                 u64 io_start_time_ns,
 804                                                 unsigned int op) { }
 805
 806 #endif  /* CONFIG_CFQ_GROUP_IOSCHED */
 807
 808 #define cfq_log(cfqd, fmt, args...)     \
 809         blk_add_trace_msg((cfqd)->queue, "cfq " fmt, ##args)
 810
 811 /* Traverses through cfq group service trees */
 812 #define for_each_cfqg_st(cfqg, i, j, st) \
 813         for (i = 0; i <= IDLE_WORKLOAD; i++) \
 814                 for (j = 0, st = i < IDLE_WORKLOAD ? &cfqg->service_trees[i][j]\
 815                         : &cfqg->service_tree_idle; \
 816                         (i < IDLE_WORKLOAD && j <= SYNC_WORKLOAD) || \
 817                         (i == IDLE_WORKLOAD && j == 0); \
 818                         j++, st = i < IDLE_WORKLOAD ? \
 819                         &cfqg->service_trees[i][j]: NULL) \
 820
 821 static inline bool cfq_io_thinktime_big(struct cfq_data *cfqd,
 822         struct cfq_ttime *ttime, bool group_idle)
 823 {
 824         u64 slice;
 825         if (!sample_valid(ttime->ttime_samples))
 826                 return false;
 827         if (group_idle)
 828                 slice = cfqd->cfq_group_idle;
 829         else
 830                 slice = cfqd->cfq_slice_idle;
 831         return ttime->ttime_mean > slice;
 832 }
 833
 834 static inline bool iops_mode(struct cfq_data *cfqd)
 835 {
 836         /*
 837          * If we are not idling on queues and it is a NCQ drive, parallel
 838          * execution of requests is on and measuring time is not possible
 839          * in most of the cases until and unless we drive shallower queue
 840          * depths and that becomes a performance bottleneck. In such cases
 841          * switch to start providing fairness in terms of number of IOs.
 842          */
 843         if (!cfqd->cfq_slice_idle && cfqd->hw_tag)
 844                 return true;
 845         else
 846                 return false;
 847 }
 848
 849 static inline enum wl_class_t cfqq_class(struct cfq_queue *cfqq)
 850 {
 851         if (cfq_class_idle(cfqq))
 852                 return IDLE_WORKLOAD;
 853         if (cfq_class_rt(cfqq))
 854                 return RT_WORKLOAD;
 855         return BE_WORKLOAD;
 856 }
 857
 858
 859 static enum wl_type_t cfqq_type(struct cfq_queue *cfqq)
 860 {
 861         if (!cfq_cfqq_sync(cfqq))
 862                 return ASYNC_WORKLOAD;
 863         if (!cfq_cfqq_idle_window(cfqq))
 864                 return SYNC_NOIDLE_WORKLOAD;
 865         return SYNC_WORKLOAD;
 866 }
 867
 868 static inline int cfq_group_busy_queues_wl(enum wl_class_t wl_class,
 869                                         struct cfq_data *cfqd,
 870                                         struct cfq_group *cfqg)
 871 {
 872         if (wl_class == IDLE_WORKLOAD)
 873                 return cfqg->service_tree_idle.count;
 874
 875         return cfqg->service_trees[wl_class][ASYNC_WORKLOAD].count +
 876                 cfqg->service_trees[wl_class][SYNC_NOIDLE_WORKLOAD].count +
 877                 cfqg->service_trees[wl_class][SYNC_WORKLOAD].count;
 878 }
 879
 880 static inline int cfqg_busy_async_queues(struct cfq_data *cfqd,
 881                                         struct cfq_group *cfqg)
 882 {
 883         return cfqg->service_trees[RT_WORKLOAD][ASYNC_WORKLOAD].count +
 884                 cfqg->service_trees[BE_WORKLOAD][ASYNC_WORKLOAD].count;
 885 }
 886
 887 static void cfq_dispatch_insert(struct request_queue *, struct request *);
 888 static struct cfq_queue *cfq_get_queue(struct cfq_data *cfqd, bool is_sync,
 889                                        struct cfq_io_cq *cic, struct bio *bio);
 890
 891 static inline struct cfq_io_cq *icq_to_cic(struct io_cq *icq)
 892 {
 893         /* cic->icq is the first member, %NULL will convert to %NULL */
 894         return container_of(icq, struct cfq_io_cq, icq);
 895 }
 896
 897 static inline struct cfq_io_cq *cfq_cic_lookup(struct cfq_data *cfqd,
 898                                                struct io_context *ioc)
 899 {
 900         if (ioc)
 901                 return icq_to_cic(ioc_lookup_icq(ioc, cfqd->queue));
 902         return NULL;
 903 }
 904
 905 static inline struct cfq_queue *cic_to_cfqq(struct cfq_io_cq *cic, bool is_sync)
 906 {
 907         return cic->cfqq[is_sync];
 908 }
 909
 910 static inline void cic_set_cfqq(struct cfq_io_cq *cic, struct cfq_queue *cfqq,
 911                                 bool is_sync)
 912 {
 913         cic->cfqq[is_sync] = cfqq;
 914 }
 915
 916 static inline struct cfq_data *cic_to_cfqd(struct cfq_io_cq *cic)
 917 {
 918         return cic->icq.q->elevator->elevator_data;
 919 }
 920
 921 /*
 922  * scheduler run of queue, if there are requests pending and no one in the
 923  * driver that will restart queueing
 924  */
 925 static inline void cfq_schedule_dispatch(struct cfq_data *cfqd)
 926 {
 927         if (cfqd->busy_queues) {
 928                 cfq_log(cfqd, "schedule dispatch");
 929                 kblockd_schedule_work(&cfqd->unplug_work);
 930         }
 931 }
 932
 933 /*
 934  * Scale schedule slice based on io priority. Use the sync time slice only
 935  * if a queue is marked sync and has sync io queued. A sync queue with async
 936  * io only, should not get full sync slice length.
 937  */
 938 static inline u64 cfq_prio_slice(struct cfq_data *cfqd, bool sync,
 939                                  unsigned short prio)
 940 {
 941         u64 base_slice = cfqd->cfq_slice[sync];
 942         u64 slice = div_u64(base_slice, CFQ_SLICE_SCALE);
 943
 944         WARN_ON(prio >= IOPRIO_BE_NR);
 945
 946         return base_slice + (slice * (4 - prio));
 947 }
 948
 949 static inline u64
 950 cfq_prio_to_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 951 {
 952         return cfq_prio_slice(cfqd, cfq_cfqq_sync(cfqq), cfqq->ioprio);
 953 }
 954
 955 /**
 956  * cfqg_scale_charge - scale disk time charge according to cfqg weight
 957  * @charge: disk time being charged
 958  * @vfraction: vfraction of the cfqg, fixed point w/ CFQ_SERVICE_SHIFT
 959  *
 960  * Scale @charge according to @vfraction, which is in range (0, 1].  The
 961  * scaling is inversely proportional.
 962  *
 963  * scaled = charge / vfraction
 964  *
 965  * The result is also in fixed point w/ CFQ_SERVICE_SHIFT.
 966  */
 967 static inline u64 cfqg_scale_charge(u64 charge,
 968                                     unsigned int vfraction)
 969 {
 970         u64 c = charge << CFQ_SERVICE_SHIFT;    /* make it fixed point */
 971
 972         /* charge / vfraction */
 973         c <<= CFQ_SERVICE_SHIFT;
 974         return div_u64(c, vfraction);
 975 }
 976
 977 static inline u64 max_vdisktime(u64 min_vdisktime, u64 vdisktime)
 978 {
 979         s64 delta = (s64)(vdisktime - min_vdisktime);
 980         if (delta > 0)
 981                 min_vdisktime = vdisktime;
 982
 983         return min_vdisktime;
 984 }
 985
 986 static void update_min_vdisktime(struct cfq_rb_root *st)
 987 {
 988         if (!RB_EMPTY_ROOT(&st->rb.rb_root)) {
 989                 struct cfq_group *cfqg = rb_entry_cfqg(st->rb.rb_leftmost);
 990
 991                 st->min_vdisktime = max_vdisktime(st->min_vdisktime,
 992                                                   cfqg->vdisktime);
 993         }
 994 }
 995
 996 /*
 997  * get averaged number of queues of RT/BE priority.
 998  * average is updated, with a formula that gives more weight to higher numbers,
 999  * to quickly follows sudden increases and decrease slowly
1000  */
1001
1002 static inline unsigned cfq_group_get_avg_queues(struct cfq_data *cfqd,
1003                                         struct cfq_group *cfqg, bool rt)
1004 {
1005         unsigned min_q, max_q;
1006         unsigned mult  = cfq_hist_divisor - 1;
1007         unsigned round = cfq_hist_divisor / 2;
1008         unsigned busy = cfq_group_busy_queues_wl(rt, cfqd, cfqg);
1009
1010         min_q = min(cfqg->busy_queues_avg[rt], busy);
1011         max_q = max(cfqg->busy_queues_avg[rt], busy);
1012         cfqg->busy_queues_avg[rt] = (mult * max_q + min_q + round) /
1013                 cfq_hist_divisor;
1014         return cfqg->busy_queues_avg[rt];
1015 }
1016
1017 static inline u64
1018 cfq_group_slice(struct cfq_data *cfqd, struct cfq_group *cfqg)
1019 {
1020         return cfqd->cfq_target_latency * cfqg->vfraction >> CFQ_SERVICE_SHIFT;
1021 }
1022
1023 static inline u64
1024 cfq_scaled_cfqq_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
1025 {
1026         u64 slice = cfq_prio_to_slice(cfqd, cfqq);
1027         if (cfqd->cfq_latency) {
1028                 /*
1029                  * interested queues (we consider only the ones with the same
1030                  * priority class in the cfq group)
1031                  */
1032                 unsigned iq = cfq_group_get_avg_queues(cfqd, cfqq->cfqg,
1033                                                 cfq_class_rt(cfqq));
1034                 u64 sync_slice = cfqd->cfq_slice[1];
1035                 u64 expect_latency = sync_slice * iq;
1036                 u64 group_slice = cfq_group_slice(cfqd, cfqq->cfqg);
1037
1038                 if (expect_latency > group_slice) {
1039                         u64 base_low_slice = 2 * cfqd->cfq_slice_idle;
1040                         u64 low_slice;
1041
1042                         /* scale low_slice according to IO priority
1043                          * and sync vs async */
1044                         low_slice = div64_u64(base_low_slice*slice, sync_slice);
1045                         low_slice = min(slice, low_slice);
1046                         /* the adapted slice value is scaled to fit all iqs
1047                          * into the target latency */
1048                         slice = div64_u64(slice*group_slice, expect_latency);
1049                         slice = max(slice, low_slice);
1050                 }
1051         }
1052         return slice;
1053 }
1054
1055 static inline void
1056 cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
1057 {
1058         u64 slice = cfq_scaled_cfqq_slice(cfqd, cfqq);
1059         u64 now = ktime_get_ns();
1060
1061         cfqq->slice_start = now;
1062         cfqq->slice_end = now + slice;
1063         cfqq->allocated_slice = slice;
1064         cfq_log_cfqq(cfqd, cfqq, "set_slice=%llu", cfqq->slice_end - now);
1065 }
1066
1067 /*
1068  * We need to wrap this check in cfq_cfqq_slice_new(), since ->slice_end
1069  * isn't valid until the first request from the dispatch is activated
1070  * and the slice time set.
1071  */
1072 static inline bool cfq_slice_used(struct cfq_queue *cfqq)
1073 {
1074         if (cfq_cfqq_slice_new(cfqq))
1075                 return false;
1076         if (ktime_get_ns() < cfqq->slice_end)
1077                 return false;
1078
1079         return true;
1080 }
1081
1082 /*
1083  * Lifted from AS - choose which of rq1 and rq2 that is best served now.
1084  * We choose the request that is closest to the head right now. Distance
1085  * behind the head is penalized and only allowed to a certain extent.
1086  */
1087 static struct request *
1088 cfq_choose_req(struct cfq_data *cfqd, struct request *rq1, struct request *rq2, sector_t last)
1089 {
1090         sector_t s1, s2, d1 = 0, d2 = 0;
1091         unsigned long back_max;
1092 #define CFQ_RQ1_WRAP    0x01 /* request 1 wraps */
1093 #define CFQ_RQ2_WRAP    0x02 /* request 2 wraps */
1094         unsigned wrap = 0; /* bit mask: requests behind the disk head? */
1095
1096         if (rq1 == NULL || rq1 == rq2)
1097                 return rq2;
1098         if (rq2 == NULL)
1099                 return rq1;
1100
1101         if (rq_is_sync(rq1) != rq_is_sync(rq2))
1102                 return rq_is_sync(rq1) ? rq1 : rq2;
1103
1104         if ((rq1->cmd_flags ^ rq2->cmd_flags) & REQ_PRIO)
1105                 return rq1->cmd_flags & REQ_PRIO ? rq1 : rq2;
1106
1107         s1 = blk_rq_pos(rq1);
1108         s2 = blk_rq_pos(rq2);
1109
1110         /*
1111          * by definition, 1KiB is 2 sectors
1112          */
1113         back_max = cfqd->cfq_back_max * 2;
1114
1115         /*
1116          * Strict one way elevator _except_ in the case where we allow
1117          * short backward seeks which are biased as twice the cost of a
1118          * similar forward seek.
1119          */
1120         if (s1 >= last)
1121                 d1 = s1 - last;
1122         else if (s1 + back_max >= last)
1123                 d1 = (last - s1) * cfqd->cfq_back_penalty;
1124         else
1125                 wrap |= CFQ_RQ1_WRAP;
1126
1127         if (s2 >= last)
1128                 d2 = s2 - last;
1129         else if (s2 + back_max >= last)
1130                 d2 = (last - s2) * cfqd->cfq_back_penalty;
1131         else
1132                 wrap |= CFQ_RQ2_WRAP;
1133
1134         /* Found required data */
1135
1136         /*
1137          * By doing switch() on the bit mask "wrap" we avoid having to
1138          * check two variables for all permutations: --> faster!
1139          */
1140         switch (wrap) {
1141         case 0: /* common case for CFQ: rq1 and rq2 not wrapped */
1142                 if (d1 < d2)
1143                         return rq1;
1144                 else if (d2 < d1)
1145                         return rq2;
1146                 else {
1147                         if (s1 >= s2)
1148                                 return rq1;
1149                         else
1150                                 return rq2;
1151                 }
1152
1153         case CFQ_RQ2_WRAP:
1154                 return rq1;
1155         case CFQ_RQ1_WRAP:
1156                 return rq2;
1157         case (CFQ_RQ1_WRAP|CFQ_RQ2_WRAP): /* both rqs wrapped */
1158         default:
1159                 /*
1160                  * Since both rqs are wrapped,
1161                  * start with the one that's further behind head
1162                  * (--> only *one* back seek required),
1163                  * since back seek takes more time than forward.
1164                  */
1165                 if (s1 <= s2)
1166                         return rq1;
1167                 else
1168                         return rq2;
1169         }
1170 }
1171
1172 static struct cfq_queue *cfq_rb_first(struct cfq_rb_root *root)
1173 {
1174         /* Service tree is empty */
1175         if (!root->count)
1176                 return NULL;
1177
1178         return rb_entry(rb_first_cached(&root->rb), struct cfq_queue, rb_node);
1179 }
1180
1181 static struct cfq_group *cfq_rb_first_group(struct cfq_rb_root *root)
1182 {
1183         return rb_entry_cfqg(rb_first_cached(&root->rb));
1184 }
1185
1186 static void cfq_rb_erase(struct rb_node *n, struct cfq_rb_root *root)
1187 {
1188         if (root->rb_rightmost == n)
1189                 root->rb_rightmost = rb_prev(n);
1190
1191         rb_erase_cached(n, &root->rb);
1192         RB_CLEAR_NODE(n);
1193
1194         --root->count;
1195 }
1196
1197 /*
1198  * would be nice to take fifo expire time into account as well
1199  */
1200 static struct request *
1201 cfq_find_next_rq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
1202                   struct request *last)
1203 {
1204         struct rb_node *rbnext = rb_next(&last->rb_node);
1205         struct rb_node *rbprev = rb_prev(&last->rb_node);
1206         struct request *next = NULL, *prev = NULL;
1207
1208         BUG_ON(RB_EMPTY_NODE(&last->rb_node));
1209
1210         if (rbprev)
1211                 prev = rb_entry_rq(rbprev);
1212
1213         if (rbnext)
1214                 next = rb_entry_rq(rbnext);
1215         else {
1216                 rbnext = rb_first(&cfqq->sort_list);
1217                 if (rbnext && rbnext != &last->rb_node)
1218                         next = rb_entry_rq(rbnext);
1219         }
1220
1221         return cfq_choose_req(cfqd, next, prev, blk_rq_pos(last));
1222 }
1223
1224 static u64 cfq_slice_offset(struct cfq_data *cfqd,
1225                             struct cfq_queue *cfqq)
1226 {
1227         /*
1228          * just an approximation, should be ok.
1229          */
1230         return (cfqq->cfqg->nr_cfqq - 1) * (cfq_prio_slice(cfqd, 1, 0) -
1231                        cfq_prio_slice(cfqd, cfq_cfqq_sync(cfqq), cfqq->ioprio));
1232 }
1233
1234 static inline s64
1235 cfqg_key(struct cfq_rb_root *st, struct cfq_group *cfqg)
1236 {
1237         return cfqg->vdisktime - st->min_vdisktime;
1238 }
1239
1240 static void
1241 __cfq_group_service_tree_add(struct cfq_rb_root *st, struct cfq_group *cfqg)
1242 {
1243         struct rb_node **node = &st->rb.rb_root.rb_node;
1244         struct rb_node *parent = NULL;
1245         struct cfq_group *__cfqg;
1246         s64 key = cfqg_key(st, cfqg);
1247         bool leftmost = true, rightmost = true;
1248
1249         while (*node != NULL) {
1250                 parent = *node;
1251                 __cfqg = rb_entry_cfqg(parent);
1252
1253                 if (key < cfqg_key(st, __cfqg)) {
1254                         node = &parent->rb_left;
1255                         rightmost = false;
1256                 } else {
1257                         node = &parent->rb_right;
1258                         leftmost = false;
1259                 }
1260         }
1261
1262         if (rightmost)
1263                 st->rb_rightmost = &cfqg->rb_node;
1264
1265         rb_link_node(&cfqg->rb_node, parent, node);
1266         rb_insert_color_cached(&cfqg->rb_node, &st->rb, leftmost);
1267 }
1268
1269 /*
1270  * This has to be called only on activation of cfqg
1271  */
1272 static void
1273 cfq_update_group_weight(struct cfq_group *cfqg)
1274 {
1275         if (cfqg->new_weight) {
1276                 cfqg->weight = cfqg->new_weight;
1277                 cfqg->new_weight = 0;
1278         }
1279 }
1280
1281 static void
1282 cfq_update_group_leaf_weight(struct cfq_group *cfqg)
1283 {
1284         BUG_ON(!RB_EMPTY_NODE(&cfqg->rb_node));
1285
1286         if (cfqg->new_leaf_weight) {
1287                 cfqg->leaf_weight = cfqg->new_leaf_weight;
1288                 cfqg->new_leaf_weight = 0;
1289         }
1290 }
1291
1292 static void
1293 cfq_group_service_tree_add(struct cfq_rb_root *st, struct cfq_group *cfqg)
1294 {
1295         unsigned int vfr = 1 << CFQ_SERVICE_SHIFT;      /* start with 1 */
1296         struct cfq_group *pos = cfqg;
1297         struct cfq_group *parent;
1298         bool propagate;
1299
1300         /* add to the service tree */
1301         BUG_ON(!RB_EMPTY_NODE(&cfqg->rb_node));
1302
1303         /*
1304          * Update leaf_weight.  We cannot update weight at this point
1305          * because cfqg might already have been activated and is
1306          * contributing its current weight to the parent's child_weight.
1307          */
1308         cfq_update_group_leaf_weight(cfqg);
1309         __cfq_group_service_tree_add(st, cfqg);
1310
1311         /*
1312          * Activate @cfqg and calculate the portion of vfraction @cfqg is
1313          * entitled to.  vfraction is calculated by walking the tree
1314          * towards the root calculating the fraction it has at each level.
1315          * The compounded ratio is how much vfraction @cfqg owns.
1316          *
1317          * Start with the proportion tasks in this cfqg has against active
1318          * children cfqgs - its leaf_weight against children_weight.
1319          */
1320         propagate = !pos->nr_active++;
1321         pos->children_weight += pos->leaf_weight;
1322         vfr = vfr * pos->leaf_weight / pos->children_weight;
1323
1324         /*
1325          * Compound ->weight walking up the tree.  Both activation and
1326          * vfraction calculation are done in the same loop.  Propagation
1327          * stops once an already activated node is met.  vfraction
1328          * calculation should always continue to the root.
1329          */
1330         while ((parent = cfqg_parent(pos))) {
1331                 if (propagate) {
1332                         cfq_update_group_weight(pos);
1333                         propagate = !parent->nr_active++;
1334                         parent->children_weight += pos->weight;
1335                 }
1336                 vfr = vfr * pos->weight / parent->children_weight;
1337                 pos = parent;
1338         }
1339
1340         cfqg->vfraction = max_t(unsigned, vfr, 1);
1341 }
1342
1343 static inline u64 cfq_get_cfqg_vdisktime_delay(struct cfq_data *cfqd)
1344 {
1345         if (!iops_mode(cfqd))
1346                 return CFQ_SLICE_MODE_GROUP_DELAY;
1347         else
1348                 return CFQ_IOPS_MODE_GROUP_DELAY;
1349 }
1350
1351 static void
1352 cfq_group_notify_queue_add(struct cfq_data *cfqd, struct cfq_group *cfqg)
1353 {
1354         struct cfq_rb_root *st = &cfqd->grp_service_tree;
1355         struct cfq_group *__cfqg;
1356         struct rb_node *n;
1357
1358         cfqg->nr_cfqq++;
1359         if (!RB_EMPTY_NODE(&cfqg->rb_node))
1360                 return;
1361
1362         /*
1363          * Currently put the group at the end. Later implement something
1364          * so that groups get lesser vtime based on their weights, so that
1365          * if group does not loose all if it was not continuously backlogged.
1366          */
1367         n = st->rb_rightmost;
1368         if (n) {
1369                 __cfqg = rb_entry_cfqg(n);
1370                 cfqg->vdisktime = __cfqg->vdisktime +
1371                         cfq_get_cfqg_vdisktime_delay(cfqd);
1372         } else
1373                 cfqg->vdisktime = st->min_vdisktime;
1374         cfq_group_service_tree_add(st, cfqg);
1375 }
1376
1377 static void
1378 cfq_group_service_tree_del(struct cfq_rb_root *st, struct cfq_group *cfqg)
1379 {
1380         struct cfq_group *pos = cfqg;
1381         bool propagate;
1382
1383         /*
1384          * Undo activation from cfq_group_service_tree_add().  Deactivate
1385          * @cfqg and propagate deactivation upwards.
1386          */
1387         propagate = !--pos->nr_active;
1388         pos->children_weight -= pos->leaf_weight;
1389
1390         while (propagate) {
1391                 struct cfq_group *parent = cfqg_parent(pos);
1392
1393                 /* @pos has 0 nr_active at this point */
1394                 WARN_ON_ONCE(pos->children_weight);
1395                 pos->vfraction = 0;
1396
1397                 if (!parent)
1398                         break;
1399
1400                 propagate = !--parent->nr_active;
1401                 parent->children_weight -= pos->weight;
1402                 pos = parent;
1403         }
1404
1405         /* remove from the service tree */
1406         if (!RB_EMPTY_NODE(&cfqg->rb_node))
1407                 cfq_rb_erase(&cfqg->rb_node, st);
1408 }
1409
1410 static void
1411 cfq_group_notify_queue_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
1412 {
1413         struct cfq_rb_root *st = &cfqd->grp_service_tree;
1414
1415         BUG_ON(cfqg->nr_cfqq < 1);
1416         cfqg->nr_cfqq--;
1417
1418         /* If there are other cfq queues under this group, don't delete it */
1419         if (cfqg->nr_cfqq)
1420                 return;
1421
1422         cfq_log_cfqg(cfqd, cfqg, "del_from_rr group");
1423         cfq_group_service_tree_del(st, cfqg);
1424         cfqg->saved_wl_slice = 0;
1425         cfqg_stats_update_dequeue(cfqg);
1426 }
1427
1428 static inline u64 cfq_cfqq_slice_usage(struct cfq_queue *cfqq,
1429                                        u64 *unaccounted_time)
1430 {
1431         u64 slice_used;
1432         u64 now = ktime_get_ns();
1433
1434         /*
1435          * Queue got expired before even a single request completed or
1436          * got expired immediately after first request completion.
1437          */
1438         if (!cfqq->slice_start || cfqq->slice_start == now) {
1439                 /*
1440                  * Also charge the seek time incurred to the group, otherwise
1441                  * if there are mutiple queues in the group, each can dispatch
1442                  * a single request on seeky media and cause lots of seek time
1443                  * and group will never know it.
1444                  */
1445                 slice_used = max_t(u64, (now - cfqq->dispatch_start),
1446                                         jiffies_to_nsecs(1));
1447         } else {
1448                 slice_used = now - cfqq->slice_start;
1449                 if (slice_used > cfqq->allocated_slice) {
1450                         *unaccounted_time = slice_used - cfqq->allocated_slice;
1451                         slice_used = cfqq->allocated_slice;
1452                 }
1453                 if (cfqq->slice_start > cfqq->dispatch_start)
1454                         *unaccounted_time += cfqq->slice_start -
1455                                         cfqq->dispatch_start;
1456         }
1457
1458         return slice_used;
1459 }
1460
1461 static void cfq_group_served(struct cfq_data *cfqd, struct cfq_group *cfqg,
1462                                 struct cfq_queue *cfqq)
1463 {
1464         struct cfq_rb_root *st = &cfqd->grp_service_tree;
1465         u64 used_sl, charge, unaccounted_sl = 0;
1466         int nr_sync = cfqg->nr_cfqq - cfqg_busy_async_queues(cfqd, cfqg)
1467                         - cfqg->service_tree_idle.count;
1468         unsigned int vfr;
1469         u64 now = ktime_get_ns();
1470
1471         BUG_ON(nr_sync < 0);
1472         used_sl = charge = cfq_cfqq_slice_usage(cfqq, &unaccounted_sl);
1473
1474         if (iops_mode(cfqd))
1475                 charge = cfqq->slice_dispatch;
1476         else if (!cfq_cfqq_sync(cfqq) && !nr_sync)
1477                 charge = cfqq->allocated_slice;
1478
1479         /*
1480          * Can't update vdisktime while on service tree and cfqg->vfraction
1481          * is valid only while on it.  Cache vfr, leave the service tree,
1482          * update vdisktime and go back on.  The re-addition to the tree
1483          * will also update the weights as necessary.
1484          */
1485         vfr = cfqg->vfraction;
1486         cfq_group_service_tree_del(st, cfqg);
1487         cfqg->vdisktime += cfqg_scale_charge(charge, vfr);
1488         cfq_group_service_tree_add(st, cfqg);
1489
1490         /* This group is being expired. Save the context */
1491         if (cfqd->workload_expires > now) {
1492                 cfqg->saved_wl_slice = cfqd->workload_expires - now;
1493                 cfqg->saved_wl_type = cfqd->serving_wl_type;
1494                 cfqg->saved_wl_class = cfqd->serving_wl_class;
1495         } else
1496                 cfqg->saved_wl_slice = 0;
1497
1498         cfq_log_cfqg(cfqd, cfqg, "served: vt=%llu min_vt=%llu", cfqg->vdisktime,
1499                                         st->min_vdisktime);
1500         cfq_log_cfqq(cfqq->cfqd, cfqq,
1501                      "sl_used=%llu disp=%llu charge=%llu iops=%u sect=%lu",
1502                      used_sl, cfqq->slice_dispatch, charge,
1503                      iops_mode(cfqd), cfqq->nr_sectors);
1504         cfqg_stats_update_timeslice_used(cfqg, used_sl, unaccounted_sl);
1505         cfqg_stats_set_start_empty_time(cfqg);
1506 }
1507
1508 /**
1509  * cfq_init_cfqg_base - initialize base part of a cfq_group
1510  * @cfqg: cfq_group to initialize
1511  *
1512  * Initialize the base part which is used whether %CONFIG_CFQ_GROUP_IOSCHED
1513  * is enabled or not.
1514  */
1515 static void cfq_init_cfqg_base(struct cfq_group *cfqg)
1516 {
1517         struct cfq_rb_root *st;
1518         int i, j;
1519
1520         for_each_cfqg_st(cfqg, i, j, st)
1521                 *st = CFQ_RB_ROOT;
1522         RB_CLEAR_NODE(&cfqg->rb_node);
1523
1524         cfqg->ttime.last_end_request = ktime_get_ns();
1525 }
1526
1527 #ifdef CONFIG_CFQ_GROUP_IOSCHED
1528 static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
1529                             bool on_dfl, bool reset_dev, bool is_leaf_weight);
1530
1531 static void cfqg_stats_exit(struct cfqg_stats *stats)
1532 {
1533         blkg_rwstat_exit(&stats->merged);
1534         blkg_rwstat_exit(&stats->service_time);
1535         blkg_rwstat_exit(&stats->wait_time);
1536         blkg_rwstat_exit(&stats->queued);
1537         blkg_stat_exit(&stats->time);
1538 #ifdef CONFIG_DEBUG_BLK_CGROUP
1539         blkg_stat_exit(&stats->unaccounted_time);
1540         blkg_stat_exit(&stats->avg_queue_size_sum);
1541         blkg_stat_exit(&stats->avg_queue_size_samples);
1542         blkg_stat_exit(&stats->dequeue);
1543         blkg_stat_exit(&stats->group_wait_time);
1544         blkg_stat_exit(&stats->idle_time);
1545         blkg_stat_exit(&stats->empty_time);
1546 #endif
1547 }
1548
1549 static int cfqg_stats_init(struct cfqg_stats *stats, gfp_t gfp)
1550 {
1551         if (blkg_rwstat_init(&stats->merged, gfp) ||
1552             blkg_rwstat_init(&stats->service_time, gfp) ||
1553             blkg_rwstat_init(&stats->wait_time, gfp) ||
1554             blkg_rwstat_init(&stats->queued, gfp) ||
1555             blkg_stat_init(&stats->time, gfp))
1556                 goto err;
1557
1558 #ifdef CONFIG_DEBUG_BLK_CGROUP
1559         if (blkg_stat_init(&stats->unaccounted_time, gfp) ||
1560             blkg_stat_init(&stats->avg_queue_size_sum, gfp) ||
1561             blkg_stat_init(&stats->avg_queue_size_samples, gfp) ||
1562             blkg_stat_init(&stats->dequeue, gfp) ||
1563             blkg_stat_init(&stats->group_wait_time, gfp) ||
1564             blkg_stat_init(&stats->idle_time, gfp) ||
1565             blkg_stat_init(&stats->empty_time, gfp))
1566                 goto err;
1567 #endif
1568         return 0;
1569 err:
1570         cfqg_stats_exit(stats);
1571         return -ENOMEM;
1572 }
1573
1574 static struct blkcg_policy_data *cfq_cpd_alloc(gfp_t gfp)
1575 {
1576         struct cfq_group_data *cgd;
1577
1578         cgd = kzalloc(sizeof(*cgd), gfp);
1579         if (!cgd)
1580                 return NULL;
1581         return &cgd->cpd;
1582 }
1583
1584 static void cfq_cpd_init(struct blkcg_policy_data *cpd)
1585 {
1586         struct cfq_group_data *cgd = cpd_to_cfqgd(cpd);
1587         unsigned int weight = cgroup_subsys_on_dfl(io_cgrp_subsys) ?
1588                               CGROUP_WEIGHT_DFL : CFQ_WEIGHT_LEGACY_DFL;
1589
1590         if (cpd_to_blkcg(cpd) == &blkcg_root)
1591                 weight *= 2;
1592
1593         cgd->weight = weight;
1594         cgd->leaf_weight = weight;
1595 }
1596
1597 static void cfq_cpd_free(struct blkcg_policy_data *cpd)
1598 {
1599         kfree(cpd_to_cfqgd(cpd));
1600 }
1601
1602 static void cfq_cpd_bind(struct blkcg_policy_data *cpd)
1603 {
1604         struct blkcg *blkcg = cpd_to_blkcg(cpd);
1605         bool on_dfl = cgroup_subsys_on_dfl(io_cgrp_subsys);
1606         unsigned int weight = on_dfl ? CGROUP_WEIGHT_DFL : CFQ_WEIGHT_LEGACY_DFL;
1607
1608         if (blkcg == &blkcg_root)
1609                 weight *= 2;
1610
1611         WARN_ON_ONCE(__cfq_set_weight(&blkcg->css, weight, on_dfl, true, false));
1612         WARN_ON_ONCE(__cfq_set_weight(&blkcg->css, weight, on_dfl, true, true));
1613 }
1614
1615 static struct blkg_policy_data *cfq_pd_alloc(gfp_t gfp, int node)
1616 {
1617         struct cfq_group *cfqg;
1618
1619         cfqg = kzalloc_node(sizeof(*cfqg), gfp, node);
1620         if (!cfqg)
1621                 return NULL;
1622
1623         cfq_init_cfqg_base(cfqg);
1624         if (cfqg_stats_init(&cfqg->stats, gfp)) {
1625                 kfree(cfqg);
1626                 return NULL;
1627         }
1628
1629         return &cfqg->pd;
1630 }
1631
1632 static void cfq_pd_init(struct blkg_policy_data *pd)
1633 {
1634         struct cfq_group *cfqg = pd_to_cfqg(pd);
1635         struct cfq_group_data *cgd = blkcg_to_cfqgd(pd->blkg->blkcg);
1636
1637         cfqg->weight = cgd->weight;
1638         cfqg->leaf_weight = cgd->leaf_weight;
1639 }
1640
1641 static void cfq_pd_offline(struct blkg_policy_data *pd)
1642 {
1643         struct cfq_group *cfqg = pd_to_cfqg(pd);
1644         int i;
1645
1646         for (i = 0; i < IOPRIO_BE_NR; i++) {
1647                 if (cfqg->async_cfqq[0][i]) {
1648                         cfq_put_queue(cfqg->async_cfqq[0][i]);
1649                         cfqg->async_cfqq[0][i] = NULL;
1650                 }
1651                 if (cfqg->async_cfqq[1][i]) {
1652                         cfq_put_queue(cfqg->async_cfqq[1][i]);
1653                         cfqg->async_cfqq[1][i] = NULL;
1654                 }
1655         }
1656
1657         if (cfqg->async_idle_cfqq) {
1658                 cfq_put_queue(cfqg->async_idle_cfqq);
1659                 cfqg->async_idle_cfqq = NULL;
1660         }
1661
1662         /*
1663          * @blkg is going offline and will be ignored by
1664          * blkg_[rw]stat_recursive_sum().  Transfer stats to the parent so
1665          * that they don't get lost.  If IOs complete after this point, the
1666          * stats for them will be lost.  Oh well...
1667          */
1668         cfqg_stats_xfer_dead(cfqg);
1669 }
1670
1671 static void cfq_pd_free(struct blkg_policy_data *pd)
1672 {
1673         struct cfq_group *cfqg = pd_to_cfqg(pd);
1674
1675         cfqg_stats_exit(&cfqg->stats);
1676         return kfree(cfqg);
1677 }
1678
1679 static void cfq_pd_reset_stats(struct blkg_policy_data *pd)
1680 {
1681         struct cfq_group *cfqg = pd_to_cfqg(pd);
1682
1683         cfqg_stats_reset(&cfqg->stats);
1684 }
1685
1686 static struct cfq_group *cfq_lookup_cfqg(struct cfq_data *cfqd,
1687                                          struct blkcg *blkcg)
1688 {
1689         struct blkcg_gq *blkg;
1690
1691         blkg = blkg_lookup(blkcg, cfqd->queue);
1692         if (likely(blkg))
1693                 return blkg_to_cfqg(blkg);
1694         return NULL;
1695 }
1696
1697 static void cfq_link_cfqq_cfqg(struct cfq_queue *cfqq, struct cfq_group *cfqg)
1698 {
1699         cfqq->cfqg = cfqg;
1700         /* cfqq reference on cfqg */
1701         cfqg_get(cfqg);
1702 }
1703
1704 static u64 cfqg_prfill_weight_device(struct seq_file *sf,
1705                                      struct blkg_policy_data *pd, int off)
1706 {
1707         struct cfq_group *cfqg = pd_to_cfqg(pd);
1708
1709         if (!cfqg->dev_weight)
1710                 return 0;
1711         return __blkg_prfill_u64(sf, pd, cfqg->dev_weight);
1712 }
1713
1714 static int cfqg_print_weight_device(struct seq_file *sf, void *v)
1715 {
1716         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1717                           cfqg_prfill_weight_device, &blkcg_policy_cfq,
1718                           0, false);
1719         return 0;
1720 }
1721
1722 static u64 cfqg_prfill_leaf_weight_device(struct seq_file *sf,
1723                                           struct blkg_policy_data *pd, int off)
1724 {
1725         struct cfq_group *cfqg = pd_to_cfqg(pd);
1726
1727         if (!cfqg->dev_leaf_weight)
1728                 return 0;
1729         return __blkg_prfill_u64(sf, pd, cfqg->dev_leaf_weight);
1730 }
1731
1732 static int cfqg_print_leaf_weight_device(struct seq_file *sf, void *v)
1733 {
1734         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1735                           cfqg_prfill_leaf_weight_device, &blkcg_policy_cfq,
1736                           0, false);
1737         return 0;
1738 }
1739
1740 static int cfq_print_weight(struct seq_file *sf, void *v)
1741 {
1742         struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
1743         struct cfq_group_data *cgd = blkcg_to_cfqgd(blkcg);
1744         unsigned int val = 0;
1745
1746         if (cgd)
1747                 val = cgd->weight;
1748
1749         seq_printf(sf, "%u\n", val);
1750         return 0;
1751 }
1752
1753 static int cfq_print_leaf_weight(struct seq_file *sf, void *v)
1754 {
1755         struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
1756         struct cfq_group_data *cgd = blkcg_to_cfqgd(blkcg);
1757         unsigned int val = 0;
1758
1759         if (cgd)
1760                 val = cgd->leaf_weight;
1761
1762         seq_printf(sf, "%u\n", val);
1763         return 0;
1764 }
1765
1766 static ssize_t __cfqg_set_weight_device(struct kernfs_open_file *of,
1767                                         char *buf, size_t nbytes, loff_t off,
1768                                         bool on_dfl, bool is_leaf_weight)
1769 {
1770         unsigned int min = on_dfl ? CGROUP_WEIGHT_MIN : CFQ_WEIGHT_LEGACY_MIN;
1771         unsigned int max = on_dfl ? CGROUP_WEIGHT_MAX : CFQ_WEIGHT_LEGACY_MAX;
1772         struct blkcg *blkcg = css_to_blkcg(of_css(of));
1773         struct blkg_conf_ctx ctx;
1774         struct cfq_group *cfqg;
1775         struct cfq_group_data *cfqgd;
1776         int ret;
1777         u64 v;
1778
1779         ret = blkg_conf_prep(blkcg, &blkcg_policy_cfq, buf, &ctx);
1780         if (ret)
1781                 return ret;
1782
1783         if (sscanf(ctx.body, "%llu", &v) == 1) {
1784                 /* require "default" on dfl */
1785                 ret = -ERANGE;
1786                 if (!v && on_dfl)
1787                         goto out_finish;
1788         } else if (!strcmp(strim(ctx.body), "default")) {
1789                 v = 0;
1790         } else {
1791                 ret = -EINVAL;
1792                 goto out_finish;
1793         }
1794
1795         cfqg = blkg_to_cfqg(ctx.blkg);
1796         cfqgd = blkcg_to_cfqgd(blkcg);
1797
1798         ret = -ERANGE;
1799         if (!v || (v >= min && v <= max)) {
1800                 if (!is_leaf_weight) {
1801                         cfqg->dev_weight = v;
1802                         cfqg->new_weight = v ?: cfqgd->weight;
1803                 } else {
1804                         cfqg->dev_leaf_weight = v;
1805                         cfqg->new_leaf_weight = v ?: cfqgd->leaf_weight;
1806                 }
1807                 ret = 0;
1808         }
1809 out_finish:
1810         blkg_conf_finish(&ctx);
1811         return ret ?: nbytes;
1812 }
1813
1814 static ssize_t cfqg_set_weight_device(struct kernfs_open_file *of,
1815                                       char *buf, size_t nbytes, loff_t off)
1816 {
1817         return __cfqg_set_weight_device(of, buf, nbytes, off, false, false);
1818 }
1819
1820 static ssize_t cfqg_set_leaf_weight_device(struct kernfs_open_file *of,
1821                                            char *buf, size_t nbytes, loff_t off)
1822 {
1823         return __cfqg_set_weight_device(of, buf, nbytes, off, false, true);
1824 }
1825
1826 static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
1827                             bool on_dfl, bool reset_dev, bool is_leaf_weight)
1828 {
1829         unsigned int min = on_dfl ? CGROUP_WEIGHT_MIN : CFQ_WEIGHT_LEGACY_MIN;
1830         unsigned int max = on_dfl ? CGROUP_WEIGHT_MAX : CFQ_WEIGHT_LEGACY_MAX;
1831         struct blkcg *blkcg = css_to_blkcg(css);
1832         struct blkcg_gq *blkg;
1833         struct cfq_group_data *cfqgd;
1834         int ret = 0;
1835
1836         if (val < min || val > max)
1837                 return -ERANGE;
1838
1839         spin_lock_irq(&blkcg->lock);
1840         cfqgd = blkcg_to_cfqgd(blkcg);
1841         if (!cfqgd) {
1842                 ret = -EINVAL;
1843                 goto out;
1844         }
1845
1846         if (!is_leaf_weight)
1847                 cfqgd->weight = val;
1848         else
1849                 cfqgd->leaf_weight = val;
1850
1851         hlist_for_each_entry(blkg, &blkcg->blkg_list, blkcg_node) {
1852                 struct cfq_group *cfqg = blkg_to_cfqg(blkg);
1853
1854                 if (!cfqg)
1855                         continue;
1856
1857                 if (!is_leaf_weight) {
1858                         if (reset_dev)
1859                                 cfqg->dev_weight = 0;
1860                         if (!cfqg->dev_weight)
1861                                 cfqg->new_weight = cfqgd->weight;
1862                 } else {
1863                         if (reset_dev)
1864                                 cfqg->dev_leaf_weight = 0;
1865                         if (!cfqg->dev_leaf_weight)
1866                                 cfqg->new_leaf_weight = cfqgd->leaf_weight;
1867                 }
1868         }
1869
1870 out:
1871         spin_unlock_irq(&blkcg->lock);
1872         return ret;
1873 }
1874
1875 static int cfq_set_weight(struct cgroup_subsys_state *css, struct cftype *cft,
1876                           u64 val)
1877 {
1878         return __cfq_set_weight(css, val, false, false, false);
1879 }
1880
1881 static int cfq_set_leaf_weight(struct cgroup_subsys_state *css,
1882                                struct cftype *cft, u64 val)
1883 {
1884         return __cfq_set_weight(css, val, false, false, true);
1885 }
1886
1887 static int cfqg_print_stat(struct seq_file *sf, void *v)
1888 {
1889         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), blkg_prfill_stat,
1890                           &blkcg_policy_cfq, seq_cft(sf)->private, false);
1891         return 0;
1892 }
1893
1894 static int cfqg_print_rwstat(struct seq_file *sf, void *v)
1895 {
1896         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), blkg_prfill_rwstat,
1897                           &blkcg_policy_cfq, seq_cft(sf)->private, true);
1898         return 0;
1899 }
1900
1901 static u64 cfqg_prfill_stat_recursive(struct seq_file *sf,
1902                                       struct blkg_policy_data *pd, int off)
1903 {
1904         u64 sum = blkg_stat_recursive_sum(pd_to_blkg(pd),
1905                                           &blkcg_policy_cfq, off);
1906         return __blkg_prfill_u64(sf, pd, sum);
1907 }
1908
1909 static u64 cfqg_prfill_rwstat_recursive(struct seq_file *sf,
1910                                         struct blkg_policy_data *pd, int off)
1911 {
1912         struct blkg_rwstat sum = blkg_rwstat_recursive_sum(pd_to_blkg(pd),
1913                                                         &blkcg_policy_cfq, off);
1914         return __blkg_prfill_rwstat(sf, pd, &sum);
1915 }
1916
1917 static int cfqg_print_stat_recursive(struct seq_file *sf, void *v)
1918 {
1919         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1920                           cfqg_prfill_stat_recursive, &blkcg_policy_cfq,
1921                           seq_cft(sf)->private, false);
1922         return 0;
1923 }
1924
1925 static int cfqg_print_rwstat_recursive(struct seq_file *sf, void *v)
1926 {
1927         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1928                           cfqg_prfill_rwstat_recursive, &blkcg_policy_cfq,
1929                           seq_cft(sf)->private, true);
1930         return 0;
1931 }
1932
1933 static u64 cfqg_prfill_sectors(struct seq_file *sf, struct blkg_policy_data *pd,
1934                                int off)
1935 {
1936         u64 sum = blkg_rwstat_total(&pd->blkg->stat_bytes);
1937
1938         return __blkg_prfill_u64(sf, pd, sum >> 9);
1939 }
1940
1941 static int cfqg_print_stat_sectors(struct seq_file *sf, void *v)
1942 {
1943         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1944                           cfqg_prfill_sectors, &blkcg_policy_cfq, 0, false);
1945         return 0;
1946 }
1947
1948 static u64 cfqg_prfill_sectors_recursive(struct seq_file *sf,
1949                                          struct blkg_policy_data *pd, int off)
1950 {
1951         struct blkg_rwstat tmp = blkg_rwstat_recursive_sum(pd->blkg, NULL,
1952                                         offsetof(struct blkcg_gq, stat_bytes));
1953         u64 sum = atomic64_read(&tmp.aux_cnt[BLKG_RWSTAT_READ]) +
1954                 atomic64_read(&tmp.aux_cnt[BLKG_RWSTAT_WRITE]);
1955
1956         return __blkg_prfill_u64(sf, pd, sum >> 9);
1957 }
1958
1959 static int cfqg_print_stat_sectors_recursive(struct seq_file *sf, void *v)
1960 {
1961         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1962                           cfqg_prfill_sectors_recursive, &blkcg_policy_cfq, 0,
1963                           false);
1964         return 0;
1965 }
1966
1967 #ifdef CONFIG_DEBUG_BLK_CGROUP
1968 static u64 cfqg_prfill_avg_queue_size(struct seq_file *sf,
1969                                       struct blkg_policy_data *pd, int off)
1970 {
1971         struct cfq_group *cfqg = pd_to_cfqg(pd);
1972         u64 samples = blkg_stat_read(&cfqg->stats.avg_queue_size_samples);
1973         u64 v = 0;
1974
1975         if (samples) {
1976                 v = blkg_stat_read(&cfqg->stats.avg_queue_size_sum);
1977                 v = div64_u64(v, samples);
1978         }
1979         __blkg_prfill_u64(sf, pd, v);
1980         return 0;
1981 }
1982
1983 /* print avg_queue_size */
1984 static int cfqg_print_avg_queue_size(struct seq_file *sf, void *v)
1985 {
1986         blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
1987                           cfqg_prfill_avg_queue_size, &blkcg_policy_cfq,
1988                           0, false);
1989         return 0;
1990 }
1991 #endif  /* CONFIG_DEBUG_BLK_CGROUP */
1992
1993 static struct cftype cfq_blkcg_legacy_files[] = {
1994         /* on root, weight is mapped to leaf_weight */
1995         {
1996                 .name = "weight_device",
1997                 .flags = CFTYPE_ONLY_ON_ROOT,
1998                 .seq_show = cfqg_print_leaf_weight_device,
1999                 .write = cfqg_set_leaf_weight_device,
2000         },
2001         {
2002                 .name = "weight",
2003                 .flags = CFTYPE_ONLY_ON_ROOT,
2004                 .seq_show = cfq_print_leaf_weight,
2005                 .write_u64 = cfq_set_leaf_weight,
2006         },
2007
2008         /* no such mapping necessary for !roots */
2009         {
2010                 .name = "weight_device",
2011                 .flags = CFTYPE_NOT_ON_ROOT,
2012                 .seq_show = cfqg_print_weight_device,
2013                 .write = cfqg_set_weight_device,
2014         },
2015         {
2016                 .name = "weight",
2017                 .flags = CFTYPE_NOT_ON_ROOT,
2018                 .seq_show = cfq_print_weight,
2019                 .write_u64 = cfq_set_weight,
2020         },
2021
2022         {
2023                 .name = "leaf_weight_device",
2024                 .seq_show = cfqg_print_leaf_weight_device,
2025                 .write = cfqg_set_leaf_weight_device,
2026         },
2027         {
2028                 .name = "leaf_weight",
2029                 .seq_show = cfq_print_leaf_weight,
2030                 .write_u64 = cfq_set_leaf_weight,
2031         },
2032
2033         /* statistics, covers only the tasks in the cfqg */
2034         {
2035                 .name = "time",
2036                 .private = offsetof(struct cfq_group, stats.time),
2037                 .seq_show = cfqg_print_stat,
2038         },
2039         {
2040                 .name = "sectors",
2041                 .seq_show = cfqg_print_stat_sectors,
2042         },
2043         {
2044                 .name = "io_service_bytes",
2045                 .private = (unsigned long)&blkcg_policy_cfq,
2046                 .seq_show = blkg_print_stat_bytes,
2047         },
2048         {
2049                 .name = "io_serviced",
2050                 .private = (unsigned long)&blkcg_policy_cfq,
2051                 .seq_show = blkg_print_stat_ios,
2052         },
2053         {
2054                 .name = "io_service_time",
2055                 .private = offsetof(struct cfq_group, stats.service_time),
2056                 .seq_show = cfqg_print_rwstat,
2057         },
2058         {
2059                 .name = "io_wait_time",
2060                 .private = offsetof(struct cfq_group, stats.wait_time),
2061                 .seq_show = cfqg_print_rwstat,
2062         },
2063         {
2064                 .name = "io_merged",
2065                 .private = offsetof(struct cfq_group, stats.merged),
2066                 .seq_show = cfqg_print_rwstat,
2067         },
2068         {
2069                 .name = "io_queued",
2070                 .private = offsetof(struct cfq_group, stats.queued),
2071                 .seq_show = cfqg_print_rwstat,
2072         },
2073
2074         /* the same statictics which cover the cfqg and its descendants */
2075         {
2076                 .name = "time_recursive",
2077                 .private = offsetof(struct cfq_group, stats.time),
2078                 .seq_show = cfqg_print_stat_recursive,
2079         },
2080         {
2081                 .name = "sectors_recursive",
2082                 .seq_show = cfqg_print_stat_sectors_recursive,
2083         },
2084         {
2085                 .name = "io_service_bytes_recursive",
2086                 .private = (unsigned long)&blkcg_policy_cfq,
2087                 .seq_show = blkg_print_stat_bytes_recursive,
2088         },
2089         {
2090                 .name = "io_serviced_recursive",
2091                 .private = (unsigned long)&blkcg_policy_cfq,
2092                 .seq_show = blkg_print_stat_ios_recursive,
2093         },
2094         {
2095                 .name = "io_service_time_recursive",
2096                 .private = offsetof(struct cfq_group, stats.service_time),
2097                 .seq_show = cfqg_print_rwstat_recursive,
2098         },
2099         {
2100                 .name = "io_wait_time_recursive",
2101                 .private = offsetof(struct cfq_group, stats.wait_time),
2102                 .seq_show = cfqg_print_rwstat_recursive,
2103         },
2104         {
2105                 .name = "io_merged_recursive",
2106                 .private = offsetof(struct cfq_group, stats.merged),
2107                 .seq_show = cfqg_print_rwstat_recursive,
2108         },
2109         {
2110                 .name = "io_queued_recursive",
2111                 .private = offsetof(struct cfq_group, stats.queued),
2112                 .seq_show = cfqg_print_rwstat_recursive,
2113         },
2114 #ifdef CONFIG_DEBUG_BLK_CGROUP
2115         {
2116                 .name = "avg_queue_size",
2117                 .seq_show = cfqg_print_avg_queue_size,
2118         },
2119         {
2120                 .name = "group_wait_time",
2121                 .private = offsetof(struct cfq_group, stats.group_wait_time),
2122                 .seq_show = cfqg_print_stat,
2123         },
2124         {
2125                 .name = "idle_time",
2126                 .private = offsetof(struct cfq_group, stats.idle_time),
2127                 .seq_show = cfqg_print_stat,
2128         },
2129         {
2130                 .name = "empty_time",
2131                 .private = offsetof(struct cfq_group, stats.empty_time),
2132                 .seq_show = cfqg_print_stat,
2133         },
2134         {
2135                 .name = "dequeue",
2136                 .private = offsetof(struct cfq_group, stats.dequeue),
2137                 .seq_show = cfqg_print_stat,
2138         },
2139         {
2140                 .name = "unaccounted_time",
2141                 .private = offsetof(struct cfq_group, stats.unaccounted_time),
2142                 .seq_show = cfqg_print_stat,
2143         },
2144 #endif  /* CONFIG_DEBUG_BLK_CGROUP */
2145         { }     /* terminate */
2146 };
2147
2148 static int cfq_print_weight_on_dfl(struct seq_file *sf, void *v)
2149 {
2150         struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
2151         struct cfq_group_data *cgd = blkcg_to_cfqgd(blkcg);
2152
2153         seq_printf(sf, "default %u\n", cgd->weight);
2154         blkcg_print_blkgs(sf, blkcg, cfqg_prfill_weight_device,
2155                           &blkcg_policy_cfq, 0, false);
2156         return 0;
2157 }
2158
2159 static ssize_t cfq_set_weight_on_dfl(struct kernfs_open_file *of,
2160                                      char *buf, size_t nbytes, loff_t off)
2161 {
2162         char *endp;
2163         int ret;
2164         u64 v;
2165
2166         buf = strim(buf);
2167
2168         /* "WEIGHT" or "default WEIGHT" sets the default weight */
2169         v = simple_strtoull(buf, &endp, 0);
2170         if (*endp == '\0' || sscanf(buf, "default %llu", &v) == 1) {
2171                 ret = __cfq_set_weight(of_css(of), v, true, false, false);
2172                 return ret ?: nbytes;
2173         }
2174
2175         /* "MAJ:MIN WEIGHT" */
2176         return __cfqg_set_weight_device(of, buf, nbytes, off, true, false);
2177 }
2178
2179 static struct cftype cfq_blkcg_files[] = {
2180         {
2181                 .name = "weight",
2182                 .flags = CFTYPE_NOT_ON_ROOT,
2183                 .seq_show = cfq_print_weight_on_dfl,
2184                 .write = cfq_set_weight_on_dfl,
2185         },
2186         { }     /* terminate */
2187 };
2188
2189 #else /* GROUP_IOSCHED */
2190 static struct cfq_group *cfq_lookup_cfqg(struct cfq_data *cfqd,
2191                                          struct blkcg *blkcg)
2192 {
2193         return cfqd->root_group;
2194 }
2195
2196 static inline void
2197 cfq_link_cfqq_cfqg(struct cfq_queue *cfqq, struct cfq_group *cfqg) {
2198         cfqq->cfqg = cfqg;
2199 }
2200
2201 #endif /* GROUP_IOSCHED */
2202
2203 /*
2204  * The cfqd->service_trees holds all pending cfq_queue's that have
2205  * requests waiting to be processed. It is sorted in the order that
2206  * we will service the queues.
2207  */
2208 static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
2209                                  bool add_front)
2210 {
2211         struct rb_node **p, *parent;
2212         struct cfq_queue *__cfqq;
2213         u64 rb_key;
2214         struct cfq_rb_root *st;
2215         bool leftmost = true;
2216         int new_cfqq = 1;
2217         u64 now = ktime_get_ns();
2218
2219         st = st_for(cfqq->cfqg, cfqq_class(cfqq), cfqq_type(cfqq));
2220         if (cfq_class_idle(cfqq)) {
2221                 rb_key = CFQ_IDLE_DELAY;
2222                 parent = st->rb_rightmost;
2223                 if (parent && parent != &cfqq->rb_node) {
2224                         __cfqq = rb_entry(parent, struct cfq_queue, rb_node);
2225                         rb_key += __cfqq->rb_key;
2226                 } else
2227                         rb_key += now;
2228         } else if (!add_front) {
2229                 /*
2230                  * Get our rb key offset. Subtract any residual slice
2231                  * value carried from last service. A negative resid
2232                  * count indicates slice overrun, and this should position
2233                  * the next service time further away in the tree.
2234                  */
2235                 rb_key = cfq_slice_offset(cfqd, cfqq) + now;
2236                 rb_key -= cfqq->slice_resid;
2237                 cfqq->slice_resid = 0;
2238         } else {
2239                 rb_key = -NSEC_PER_SEC;
2240                 __cfqq = cfq_rb_first(st);
2241                 rb_key += __cfqq ? __cfqq->rb_key : now;
2242         }
2243
2244         if (!RB_EMPTY_NODE(&cfqq->rb_node)) {
2245                 new_cfqq = 0;
2246                 /*
2247                  * same position, nothing more to do
2248                  */
2249                 if (rb_key == cfqq->rb_key && cfqq->service_tree == st)
2250                         return;
2251
2252                 cfq_rb_erase(&cfqq->rb_node, cfqq->service_tree);
2253                 cfqq->service_tree = NULL;
2254         }
2255
2256         parent = NULL;
2257         cfqq->service_tree = st;
2258         p = &st->rb.rb_root.rb_node;
2259         while (*p) {
2260                 parent = *p;
2261                 __cfqq = rb_entry(parent, struct cfq_queue, rb_node);
2262
2263                 /*
2264                  * sort by key, that represents service time.
2265                  */
2266                 if (rb_key < __cfqq->rb_key)
2267                         p = &parent->rb_left;
2268                 else {
2269                         p = &parent->rb_right;
2270                         leftmost = false;
2271                 }
2272         }
2273
2274         cfqq->rb_key = rb_key;
2275         rb_link_node(&cfqq->rb_node, parent, p);
2276         rb_insert_color_cached(&cfqq->rb_node, &st->rb, leftmost);
2277         st->count++;
2278         if (add_front || !new_cfqq)
2279                 return;
2280         cfq_group_notify_queue_add(cfqd, cfqq->cfqg);
2281 }
2282
2283 static struct cfq_queue *
2284 cfq_prio_tree_lookup(struct cfq_data *cfqd, struct rb_root *root,
2285                      sector_t sector, struct rb_node **ret_parent,
2286                      struct rb_node ***rb_link)
2287 {
2288         struct rb_node **p, *parent;
2289         struct cfq_queue *cfqq = NULL;
2290
2291         parent = NULL;
2292         p = &root->rb_node;
2293         while (*p) {
2294                 struct rb_node **n;
2295
2296                 parent = *p;
2297                 cfqq = rb_entry(parent, struct cfq_queue, p_node);
2298
2299                 /*
2300                  * Sort strictly based on sector.  Smallest to the left,
2301                  * largest to the right.
2302                  */
2303                 if (sector > blk_rq_pos(cfqq->next_rq))
2304                         n = &(*p)->rb_right;
2305                 else if (sector < blk_rq_pos(cfqq->next_rq))
2306                         n = &(*p)->rb_left;
2307                 else
2308                         break;
2309                 p = n;
2310                 cfqq = NULL;
2311         }
2312
2313         *ret_parent = parent;
2314         if (rb_link)
2315                 *rb_link = p;
2316         return cfqq;
2317 }
2318
2319 static void cfq_prio_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2320 {
2321         struct rb_node **p, *parent;
2322         struct cfq_queue *__cfqq;
2323
2324         if (cfqq->p_root) {
2325                 rb_erase(&cfqq->p_node, cfqq->p_root);
2326                 cfqq->p_root = NULL;
2327         }
2328
2329         if (cfq_class_idle(cfqq))
2330                 return;
2331         if (!cfqq->next_rq)
2332                 return;
2333
2334         cfqq->p_root = &cfqd->prio_trees[cfqq->org_ioprio];
2335         __cfqq = cfq_prio_tree_lookup(cfqd, cfqq->p_root,
2336                                       blk_rq_pos(cfqq->next_rq), &parent, &p);
2337         if (!__cfqq) {
2338                 rb_link_node(&cfqq->p_node, parent, p);
2339                 rb_insert_color(&cfqq->p_node, cfqq->p_root);
2340         } else
2341                 cfqq->p_root = NULL;
2342 }
2343
2344 /*
2345  * Update cfqq's position in the service tree.
2346  */
2347 static void cfq_resort_rr_list(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2348 {
2349         /*
2350          * Resorting requires the cfqq to be on the RR list already.
2351          */
2352         if (cfq_cfqq_on_rr(cfqq)) {
2353                 cfq_service_tree_add(cfqd, cfqq, 0);
2354                 cfq_prio_tree_add(cfqd, cfqq);
2355         }
2356 }
2357
2358 /*
2359  * add to busy list of queues for service, trying to be fair in ordering
2360  * the pending list according to last request service
2361  */
2362 static void cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2363 {
2364         cfq_log_cfqq(cfqd, cfqq, "add_to_rr");
2365         BUG_ON(cfq_cfqq_on_rr(cfqq));
2366         cfq_mark_cfqq_on_rr(cfqq);
2367         cfqd->busy_queues++;
2368         if (cfq_cfqq_sync(cfqq))
2369                 cfqd->busy_sync_queues++;
2370
2371         cfq_resort_rr_list(cfqd, cfqq);
2372 }
2373
2374 /*
2375  * Called when the cfqq no longer has requests pending, remove it from
2376  * the service tree.
2377  */
2378 static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2379 {
2380         cfq_log_cfqq(cfqd, cfqq, "del_from_rr");
2381         BUG_ON(!cfq_cfqq_on_rr(cfqq));
2382         cfq_clear_cfqq_on_rr(cfqq);
2383
2384         if (!RB_EMPTY_NODE(&cfqq->rb_node)) {
2385                 cfq_rb_erase(&cfqq->rb_node, cfqq->service_tree);
2386                 cfqq->service_tree = NULL;
2387         }
2388         if (cfqq->p_root) {
2389                 rb_erase(&cfqq->p_node, cfqq->p_root);
2390                 cfqq->p_root = NULL;
2391         }
2392
2393         cfq_group_notify_queue_del(cfqd, cfqq->cfqg);
2394         BUG_ON(!cfqd->busy_queues);
2395         cfqd->busy_queues--;
2396         if (cfq_cfqq_sync(cfqq))
2397                 cfqd->busy_sync_queues--;
2398 }
2399
2400 /*
2401  * rb tree support functions
2402  */
2403 static void cfq_del_rq_rb(struct request *rq)
2404 {
2405         struct cfq_queue *cfqq = RQ_CFQQ(rq);
2406         const int sync = rq_is_sync(rq);
2407
2408         BUG_ON(!cfqq->queued[sync]);
2409         cfqq->queued[sync]--;
2410
2411         elv_rb_del(&cfqq->sort_list, rq);
2412
2413         if (cfq_cfqq_on_rr(cfqq) && RB_EMPTY_ROOT(&cfqq->sort_list)) {
2414                 /*
2415                  * Queue will be deleted from service tree when we actually
2416                  * expire it later. Right now just remove it from prio tree
2417                  * as it is empty.
2418                  */
2419                 if (cfqq->p_root) {
2420                         rb_erase(&cfqq->p_node, cfqq->p_root);
2421                         cfqq->p_root = NULL;
2422                 }
2423         }
2424 }
2425
2426 static void cfq_add_rq_rb(struct request *rq)
2427 {
2428         struct cfq_queue *cfqq = RQ_CFQQ(rq);
2429         struct cfq_data *cfqd = cfqq->cfqd;
2430         struct request *prev;
2431
2432         cfqq->queued[rq_is_sync(rq)]++;
2433
2434         elv_rb_add(&cfqq->sort_list, rq);
2435
2436         if (!cfq_cfqq_on_rr(cfqq))
2437                 cfq_add_cfqq_rr(cfqd, cfqq);
2438
2439         /*
2440          * check if this request is a better next-serve candidate
2441          */
2442         prev = cfqq->next_rq;
2443         cfqq->next_rq = cfq_choose_req(cfqd, cfqq->next_rq, rq, cfqd->last_position);
2444
2445         /*
2446          * adjust priority tree position, if ->next_rq changes
2447          */
2448         if (prev != cfqq->next_rq)
2449                 cfq_prio_tree_add(cfqd, cfqq);
2450
2451         BUG_ON(!cfqq->next_rq);
2452 }
2453
2454 static void cfq_reposition_rq_rb(struct cfq_queue *cfqq, struct request *rq)
2455 {
2456         elv_rb_del(&cfqq->sort_list, rq);
2457         cfqq->queued[rq_is_sync(rq)]--;
2458         cfqg_stats_update_io_remove(RQ_CFQG(rq), rq->cmd_flags);
2459         cfq_add_rq_rb(rq);
2460         cfqg_stats_update_io_add(RQ_CFQG(rq), cfqq->cfqd->serving_group,
2461                                  rq->cmd_flags);
2462 }
2463
2464 static struct request *
2465 cfq_find_rq_fmerge(struct cfq_data *cfqd, struct bio *bio)
2466 {
2467         struct task_struct *tsk = current;
2468         struct cfq_io_cq *cic;
2469         struct cfq_queue *cfqq;
2470
2471         cic = cfq_cic_lookup(cfqd, tsk->io_context);
2472         if (!cic)
2473                 return NULL;
2474
2475         cfqq = cic_to_cfqq(cic, op_is_sync(bio->bi_opf));
2476         if (cfqq)
2477                 return elv_rb_find(&cfqq->sort_list, bio_end_sector(bio));
2478
2479         return NULL;
2480 }
2481
2482 static void cfq_activate_request(struct request_queue *q, struct request *rq)
2483 {
2484         struct cfq_data *cfqd = q->elevator->elevator_data;
2485
2486         cfqd->rq_in_driver++;
2487         cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "activate rq, drv=%d",
2488                                                 cfqd->rq_in_driver);
2489
2490         cfqd->last_position = blk_rq_pos(rq) + blk_rq_sectors(rq);
2491 }
2492
2493 static void cfq_deactivate_request(struct request_queue *q, struct request *rq)
2494 {
2495         struct cfq_data *cfqd = q->elevator->elevator_data;
2496
2497         WARN_ON(!cfqd->rq_in_driver);
2498         cfqd->rq_in_driver--;
2499         cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "deactivate rq, drv=%d",
2500                                                 cfqd->rq_in_driver);
2501 }
2502
2503 static void cfq_remove_request(struct request *rq)
2504 {
2505         struct cfq_queue *cfqq = RQ_CFQQ(rq);
2506
2507         if (cfqq->next_rq == rq)
2508                 cfqq->next_rq = cfq_find_next_rq(cfqq->cfqd, cfqq, rq);
2509
2510         list_del_init(&rq->queuelist);
2511         cfq_del_rq_rb(rq);
2512
2513         cfqq->cfqd->rq_queued--;
2514         cfqg_stats_update_io_remove(RQ_CFQG(rq), rq->cmd_flags);
2515         if (rq->cmd_flags & REQ_PRIO) {
2516                 WARN_ON(!cfqq->prio_pending);
2517                 cfqq->prio_pending--;
2518         }
2519 }
2520
2521 static enum elv_merge cfq_merge(struct request_queue *q, struct request **req,
2522                      struct bio *bio)
2523 {
2524         struct cfq_data *cfqd = q->elevator->elevator_data;
2525         struct request *__rq;
2526
2527         __rq = cfq_find_rq_fmerge(cfqd, bio);
2528         if (__rq && elv_bio_merge_ok(__rq, bio)) {
2529                 *req = __rq;
2530                 return ELEVATOR_FRONT_MERGE;
2531         }
2532
2533         return ELEVATOR_NO_MERGE;
2534 }
2535
2536 static void cfq_merged_request(struct request_queue *q, struct request *req,
2537                                enum elv_merge type)
2538 {
2539         if (type == ELEVATOR_FRONT_MERGE) {
2540                 struct cfq_queue *cfqq = RQ_CFQQ(req);
2541
2542                 cfq_reposition_rq_rb(cfqq, req);
2543         }
2544 }
2545
2546 static void cfq_bio_merged(struct request_queue *q, struct request *req,
2547                                 struct bio *bio)
2548 {
2549         cfqg_stats_update_io_merged(RQ_CFQG(req), bio->bi_opf);
2550 }
2551
2552 static void
2553 cfq_merged_requests(struct request_queue *q, struct request *rq,
2554                     struct request *next)
2555 {
2556         struct cfq_queue *cfqq = RQ_CFQQ(rq);
2557         struct cfq_data *cfqd = q->elevator->elevator_data;
2558
2559         /*
2560          * reposition in fifo if next is older than rq
2561          */
2562         if (!list_empty(&rq->queuelist) && !list_empty(&next->queuelist) &&
2563             next->fifo_time < rq->fifo_time &&
2564             cfqq == RQ_CFQQ(next)) {
2565                 list_move(&rq->queuelist, &next->queuelist);
2566                 rq->fifo_time = next->fifo_time;
2567         }
2568
2569         if (cfqq->next_rq == next)
2570                 cfqq->next_rq = rq;
2571         cfq_remove_request(next);
2572         cfqg_stats_update_io_merged(RQ_CFQG(rq), next->cmd_flags);
2573
2574         cfqq = RQ_CFQQ(next);
2575         /*
2576          * all requests of this queue are merged to other queues, delete it
2577          * from the service tree. If it's the active_queue,
2578          * cfq_dispatch_requests() will choose to expire it or do idle
2579          */
2580         if (cfq_cfqq_on_rr(cfqq) && RB_EMPTY_ROOT(&cfqq->sort_list) &&
2581             cfqq != cfqd->active_queue)
2582                 cfq_del_cfqq_rr(cfqd, cfqq);
2583 }
2584
2585 static int cfq_allow_bio_merge(struct request_queue *q, struct request *rq,
2586                                struct bio *bio)
2587 {
2588         struct cfq_data *cfqd = q->elevator->elevator_data;
2589         bool is_sync = op_is_sync(bio->bi_opf);
2590         struct cfq_io_cq *cic;
2591         struct cfq_queue *cfqq;
2592
2593         /*
2594          * Disallow merge of a sync bio into an async request.
2595          */
2596         if (is_sync && !rq_is_sync(rq))
2597                 return false;
2598
2599         /*
2600          * Lookup the cfqq that this bio will be queued with and allow
2601          * merge only if rq is queued there.
2602          */
2603         cic = cfq_cic_lookup(cfqd, current->io_context);
2604         if (!cic)
2605                 return false;
2606
2607         cfqq = cic_to_cfqq(cic, is_sync);
2608         return cfqq == RQ_CFQQ(rq);
2609 }
2610
2611 static int cfq_allow_rq_merge(struct request_queue *q, struct request *rq,
2612                               struct request *next)
2613 {
2614         return RQ_CFQQ(rq) == RQ_CFQQ(next);
2615 }
2616
2617 static inline void cfq_del_timer(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2618 {
2619         hrtimer_try_to_cancel(&cfqd->idle_slice_timer);
2620         cfqg_stats_update_idle_time(cfqq->cfqg);
2621 }
2622
2623 static void __cfq_set_active_queue(struct cfq_data *cfqd,
2624                                    struct cfq_queue *cfqq)
2625 {
2626         if (cfqq) {
2627                 cfq_log_cfqq(cfqd, cfqq, "set_active wl_class:%d wl_type:%d",
2628                                 cfqd->serving_wl_class, cfqd->serving_wl_type);
2629                 cfqg_stats_update_avg_queue_size(cfqq->cfqg);
2630                 cfqq->slice_start = 0;
2631                 cfqq->dispatch_start = ktime_get_ns();
2632                 cfqq->allocated_slice = 0;
2633                 cfqq->slice_end = 0;
2634                 cfqq->slice_dispatch = 0;
2635                 cfqq->nr_sectors = 0;
2636
2637                 cfq_clear_cfqq_wait_request(cfqq);
2638                 cfq_clear_cfqq_must_dispatch(cfqq);
2639                 cfq_clear_cfqq_must_alloc_slice(cfqq);
2640                 cfq_clear_cfqq_fifo_expire(cfqq);
2641                 cfq_mark_cfqq_slice_new(cfqq);
2642
2643                 cfq_del_timer(cfqd, cfqq);
2644         }
2645
2646         cfqd->active_queue = cfqq;
2647 }
2648
2649 /*
2650  * current cfqq expired its slice (or was too idle), select new one
2651  */
2652 static void
2653 __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
2654                     bool timed_out)
2655 {
2656         cfq_log_cfqq(cfqd, cfqq, "slice expired t=%d", timed_out);
2657
2658         if (cfq_cfqq_wait_request(cfqq))
2659                 cfq_del_timer(cfqd, cfqq);
2660
2661         cfq_clear_cfqq_wait_request(cfqq);
2662         cfq_clear_cfqq_wait_busy(cfqq);
2663
2664         /*
2665          * If this cfqq is shared between multiple processes, check to
2666          * make sure that those processes are still issuing I/Os within
2667          * the mean seek distance.  If not, it may be time to break the
2668          * queues apart again.
2669          */
2670         if (cfq_cfqq_coop(cfqq) && CFQQ_SEEKY(cfqq))
2671                 cfq_mark_cfqq_split_coop(cfqq);
2672
2673         /*
2674          * store what was left of this slice, if the queue idled/timed out
2675          */
2676         if (timed_out) {
2677                 if (cfq_cfqq_slice_new(cfqq))
2678                         cfqq->slice_resid = cfq_scaled_cfqq_slice(cfqd, cfqq);
2679                 else
2680                         cfqq->slice_resid = cfqq->slice_end - ktime_get_ns();
2681                 cfq_log_cfqq(cfqd, cfqq, "resid=%lld", cfqq->slice_resid);
2682         }
2683
2684         cfq_group_served(cfqd, cfqq->cfqg, cfqq);
2685
2686         if (cfq_cfqq_on_rr(cfqq) && RB_EMPTY_ROOT(&cfqq->sort_list))
2687                 cfq_del_cfqq_rr(cfqd, cfqq);
2688
2689         cfq_resort_rr_list(cfqd, cfqq);
2690
2691         if (cfqq == cfqd->active_queue)
2692                 cfqd->active_queue = NULL;
2693
2694         if (cfqd->active_cic) {
2695                 put_io_context(cfqd->active_cic->icq.ioc);
2696                 cfqd->active_cic = NULL;
2697         }
2698 }
2699
2700 static inline void cfq_slice_expired(struct cfq_data *cfqd, bool timed_out)
2701 {
2702         struct cfq_queue *cfqq = cfqd->active_queue;
2703
2704         if (cfqq)
2705                 __cfq_slice_expired(cfqd, cfqq, timed_out);
2706 }
2707
2708 /*
2709  * Get next queue for service. Unless we have a queue preemption,
2710  * we'll simply select the first cfqq in the service tree.
2711  */
2712 static struct cfq_queue *cfq_get_next_queue(struct cfq_data *cfqd)
2713 {
2714         struct cfq_rb_root *st = st_for(cfqd->serving_group,
2715                         cfqd->serving_wl_class, cfqd->serving_wl_type);
2716
2717         if (!cfqd->rq_queued)
2718                 return NULL;
2719
2720         /* There is nothing to dispatch */
2721         if (!st)
2722                 return NULL;
2723         if (RB_EMPTY_ROOT(&st->rb.rb_root))
2724                 return NULL;
2725         return cfq_rb_first(st);
2726 }
2727
2728 static struct cfq_queue *cfq_get_next_queue_forced(struct cfq_data *cfqd)
2729 {
2730         struct cfq_group *cfqg;
2731         struct cfq_queue *cfqq;
2732         int i, j;
2733         struct cfq_rb_root *st;
2734
2735         if (!cfqd->rq_queued)
2736                 return NULL;
2737
2738         cfqg = cfq_get_next_cfqg(cfqd);
2739         if (!cfqg)
2740                 return NULL;
2741
2742         for_each_cfqg_st(cfqg, i, j, st) {
2743                 cfqq = cfq_rb_first(st);
2744                 if (cfqq)
2745                         return cfqq;
2746         }
2747         return NULL;
2748 }
2749
2750 /*
2751  * Get and set a new active queue for service.
2752  */
2753 static struct cfq_queue *cfq_set_active_queue(struct cfq_data *cfqd,
2754                                               struct cfq_queue *cfqq)
2755 {
2756         if (!cfqq)
2757                 cfqq = cfq_get_next_queue(cfqd);
2758
2759         __cfq_set_active_queue(cfqd, cfqq);
2760         return cfqq;
2761 }
2762
2763 static inline sector_t cfq_dist_from_last(struct cfq_data *cfqd,
2764                                           struct request *rq)
2765 {
2766         if (blk_rq_pos(rq) >= cfqd->last_position)
2767                 return blk_rq_pos(rq) - cfqd->last_position;
2768         else
2769                 return cfqd->last_position - blk_rq_pos(rq);
2770 }
2771
2772 static inline int cfq_rq_close(struct cfq_data *cfqd, struct cfq_queue *cfqq,
2773                                struct request *rq)
2774 {
2775         return cfq_dist_from_last(cfqd, rq) <= CFQQ_CLOSE_THR;
2776 }
2777
2778 static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
2779                                     struct cfq_queue *cur_cfqq)
2780 {
2781         struct rb_root *root = &cfqd->prio_trees[cur_cfqq->org_ioprio];
2782         struct rb_node *parent, *node;
2783         struct cfq_queue *__cfqq;
2784         sector_t sector = cfqd->last_position;
2785
2786         if (RB_EMPTY_ROOT(root))
2787                 return NULL;
2788
2789         /*
2790          * First, if we find a request starting at the end of the last
2791          * request, choose it.
2792          */
2793         __cfqq = cfq_prio_tree_lookup(cfqd, root, sector, &parent, NULL);
2794         if (__cfqq)
2795                 return __cfqq;
2796
2797         /*
2798          * If the exact sector wasn't found, the parent of the NULL leaf
2799          * will contain the closest sector.
2800          */
2801         __cfqq = rb_entry(parent, struct cfq_queue, p_node);
2802         if (cfq_rq_close(cfqd, cur_cfqq, __cfqq->next_rq))
2803                 return __cfqq;
2804
2805         if (blk_rq_pos(__cfqq->next_rq) < sector)
2806                 node = rb_next(&__cfqq->p_node);
2807         else
2808                 node = rb_prev(&__cfqq->p_node);
2809         if (!node)
2810                 return NULL;
2811
2812         __cfqq = rb_entry(node, struct cfq_queue, p_node);
2813         if (cfq_rq_close(cfqd, cur_cfqq, __cfqq->next_rq))
2814                 return __cfqq;
2815
2816         return NULL;
2817 }
2818
2819 /*
2820  * cfqd - obvious
2821  * cur_cfqq - passed in so that we don't decide that the current queue is
2822  *            closely cooperating with itself.
2823  *
2824  * So, basically we're assuming that that cur_cfqq has dispatched at least
2825  * one request, and that cfqd->last_position reflects a position on the disk
2826  * associated with the I/O issued by cur_cfqq.  I'm not sure this is a valid
2827  * assumption.
2828  */
2829 static struct cfq_queue *cfq_close_cooperator(struct cfq_data *cfqd,
2830                                               struct cfq_queue *cur_cfqq)
2831 {
2832         struct cfq_queue *cfqq;
2833
2834         if (cfq_class_idle(cur_cfqq))
2835                 return NULL;
2836         if (!cfq_cfqq_sync(cur_cfqq))
2837                 return NULL;
2838         if (CFQQ_SEEKY(cur_cfqq))
2839                 return NULL;
2840
2841         /*
2842          * Don't search priority tree if it's the only queue in the group.
2843          */
2844         if (cur_cfqq->cfqg->nr_cfqq == 1)
2845                 return NULL;
2846
2847         /*
2848          * We should notice if some of the queues are cooperating, eg
2849          * working closely on the same area of the disk. In that case,
2850          * we can group them together and don't waste time idling.
2851          */
2852         cfqq = cfqq_close(cfqd, cur_cfqq);
2853         if (!cfqq)
2854                 return NULL;
2855
2856         /* If new queue belongs to different cfq_group, don't choose it */
2857         if (cur_cfqq->cfqg != cfqq->cfqg)
2858                 return NULL;
2859
2860         /*
2861          * It only makes sense to merge sync queues.
2862          */
2863         if (!cfq_cfqq_sync(cfqq))
2864                 return NULL;
2865         if (CFQQ_SEEKY(cfqq))
2866                 return NULL;
2867
2868         /*
2869          * Do not merge queues of different priority classes
2870          */
2871         if (cfq_class_rt(cfqq) != cfq_class_rt(cur_cfqq))
2872                 return NULL;
2873
2874         return cfqq;
2875 }
2876
2877 /*
2878  * Determine whether we should enforce idle window for this queue.
2879  */
2880
2881 static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq)
2882 {
2883         enum wl_class_t wl_class = cfqq_class(cfqq);
2884         struct cfq_rb_root *st = cfqq->service_tree;
2885
2886         BUG_ON(!st);
2887         BUG_ON(!st->count);
2888
2889         if (!cfqd->cfq_slice_idle)
2890                 return false;
2891
2892         /* We never do for idle class queues. */
2893         if (wl_class == IDLE_WORKLOAD)
2894                 return false;
2895
2896         /* We do for queues that were marked with idle window flag. */
2897         if (cfq_cfqq_idle_window(cfqq) &&
2898            !(blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag))
2899                 return true;
2900
2901         /*
2902          * Otherwise, we do only if they are the last ones
2903          * in their service tree.
2904          */
2905         if (st->count == 1 && cfq_cfqq_sync(cfqq) &&
2906            !cfq_io_thinktime_big(cfqd, &st->ttime, false))
2907                 return true;
2908         cfq_log_cfqq(cfqd, cfqq, "Not idling. st->count:%d", st->count);
2909         return false;
2910 }
2911
2912 static void cfq_arm_slice_timer(struct cfq_data *cfqd)
2913 {
2914         struct cfq_queue *cfqq = cfqd->active_queue;
2915         struct cfq_rb_root *st = cfqq->service_tree;
2916         struct cfq_io_cq *cic;
2917         u64 sl, group_idle = 0;
2918         u64 now = ktime_get_ns();
2919
2920         /*
2921          * SSD device without seek penalty, disable idling. But only do so
2922          * for devices that support queuing, otherwise we still have a problem
2923          * with sync vs async workloads.
2924          */
2925         if (blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag &&
2926                 !cfqd->cfq_group_idle)
2927                 return;
2928
2929         WARN_ON(!RB_EMPTY_ROOT(&cfqq->sort_list));
2930         WARN_ON(cfq_cfqq_slice_new(cfqq));
2931
2932         /*
2933          * idle is disabled, either manually or by past process history
2934          */
2935         if (!cfq_should_idle(cfqd, cfqq)) {
2936                 /* no queue idling. Check for group idling */
2937                 if (cfqd->cfq_group_idle)
2938                         group_idle = cfqd->cfq_group_idle;
2939                 else
2940                         return;
2941         }
2942
2943         /*
2944          * still active requests from this queue, don't idle
2945          */
2946         if (cfqq->dispatched)
2947                 return;
2948
2949         /*
2950          * task has exited, don't wait
2951          */
2952         cic = cfqd->active_cic;
2953         if (!cic || !atomic_read(&cic->icq.ioc->active_ref))
2954                 return;
2955
2956         /*
2957          * If our average think time is larger than the remaining time
2958          * slice, then don't idle. This avoids overrunning the allotted
2959          * time slice.
2960          */
2961         if (sample_valid(cic->ttime.ttime_samples) &&
2962             (cfqq->slice_end - now < cic->ttime.ttime_mean)) {
2963                 cfq_log_cfqq(cfqd, cfqq, "Not idling. think_time:%llu",
2964                              cic->ttime.ttime_mean);
2965                 return;
2966         }
2967
2968         /*
2969          * There are other queues in the group or this is the only group and
2970          * it has too big thinktime, don't do group idle.
2971          */
2972         if (group_idle &&
2973             (cfqq->cfqg->nr_cfqq > 1 ||
2974              cfq_io_thinktime_big(cfqd, &st->ttime, true)))
2975                 return;
2976
2977         cfq_mark_cfqq_wait_request(cfqq);
2978
2979         if (group_idle)
2980                 sl = cfqd->cfq_group_idle;
2981         else
2982                 sl = cfqd->cfq_slice_idle;
2983
2984         hrtimer_start(&cfqd->idle_slice_timer, ns_to_ktime(sl),
2985                       HRTIMER_MODE_REL);
2986         cfqg_stats_set_start_idle_time(cfqq->cfqg);
2987         cfq_log_cfqq(cfqd, cfqq, "arm_idle: %llu group_idle: %d", sl,
2988                         group_idle ? 1 : 0);
2989 }
2990
2991 /*
2992  * Move request from internal lists to the request queue dispatch list.
2993  */
2994 static void cfq_dispatch_insert(struct request_queue *q, struct request *rq)
2995 {
2996         struct cfq_data *cfqd = q->elevator->elevator_data;
2997         struct cfq_queue *cfqq = RQ_CFQQ(rq);
2998
2999         cfq_log_cfqq(cfqd, cfqq, "dispatch_insert");
3000
3001         cfqq->next_rq = cfq_find_next_rq(cfqd, cfqq, rq);
3002         cfq_remove_request(rq);
3003         cfqq->dispatched++;
3004         (RQ_CFQG(rq))->dispatched++;
3005         elv_dispatch_sort(q, rq);
3006
3007         cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]++;
3008         cfqq->nr_sectors += blk_rq_sectors(rq);
3009 }
3010
3011 /*
3012  * return expired entry, or NULL to just start from scratch in rbtree
3013  */
3014 static struct request *cfq_check_fifo(struct cfq_queue *cfqq)
3015 {
3016         struct request *rq = NULL;
3017
3018         if (cfq_cfqq_fifo_expire(cfqq))
3019                 return NULL;
3020
3021         cfq_mark_cfqq_fifo_expire(cfqq);
3022
3023         if (list_empty(&cfqq->fifo))
3024                 return NULL;
3025
3026         rq = rq_entry_fifo(cfqq->fifo.next);
3027         if (ktime_get_ns() < rq->fifo_time)
3028                 rq = NULL;
3029
3030         return rq;
3031 }
3032
3033 static inline int
3034 cfq_prio_to_maxrq(struct cfq_data *cfqd, struct cfq_queue *cfqq)
3035 {
3036         const int base_rq = cfqd->cfq_slice_async_rq;
3037
3038         WARN_ON(cfqq->ioprio >= IOPRIO_BE_NR);
3039
3040         return 2 * base_rq * (IOPRIO_BE_NR - cfqq->ioprio);
3041 }
3042
3043 /*
3044  * Must be called with the queue_lock held.
3045  */
3046 static int cfqq_process_refs(struct cfq_queue *cfqq)
3047 {
3048         int process_refs, io_refs;
3049
3050         io_refs = cfqq->allocated[READ] + cfqq->allocated[WRITE];
3051         process_refs = cfqq->ref - io_refs;
3052         BUG_ON(process_refs < 0);
3053         return process_refs;
3054 }
3055
3056 static void cfq_setup_merge(struct cfq_queue *cfqq, struct cfq_queue *new_cfqq)
3057 {
3058         int process_refs, new_process_refs;
3059         struct cfq_queue *__cfqq;
3060
3061         /*
3062          * If there are no process references on the new_cfqq, then it is
3063          * unsafe to follow the ->new_cfqq chain as other cfqq's in the
3064          * chain may have dropped their last reference (not just their
3065          * last process reference).
3066          */
3067         if (!cfqq_process_refs(new_cfqq))
3068                 return;
3069
3070         /* Avoid a circular list and skip interim queue merges */
3071         while ((__cfqq = new_cfqq->new_cfqq)) {
3072                 if (__cfqq == cfqq)
3073                         return;
3074                 new_cfqq = __cfqq;
3075         }
3076
3077         process_refs = cfqq_process_refs(cfqq);
3078         new_process_refs = cfqq_process_refs(new_cfqq);
3079         /*
3080          * If the process for the cfqq has gone away, there is no
3081          * sense in merging the queues.
3082          */
3083         if (process_refs == 0 || new_process_refs == 0)
3084                 return;
3085
3086         /*
3087          * Merge in the direction of the lesser amount of work.
3088          */
3089         if (new_process_refs >= process_refs) {
3090                 cfqq->new_cfqq = new_cfqq;
3091                 new_cfqq->ref += process_refs;
3092         } else {
3093                 new_cfqq->new_cfqq = cfqq;
3094                 cfqq->ref += new_process_refs;
3095         }
3096 }
3097
3098 static enum wl_type_t cfq_choose_wl_type(struct cfq_data *cfqd,
3099                         struct cfq_group *cfqg, enum wl_class_t wl_class)
3100 {
3101         struct cfq_queue *queue;
3102         int i;
3103         bool key_valid = false;
3104         u64 lowest_key = 0;
3105         enum wl_type_t cur_best = SYNC_NOIDLE_WORKLOAD;
3106
3107         for (i = 0; i <= SYNC_WORKLOAD; ++i) {
3108                 /* select the one with lowest rb_key */
3109                 queue = cfq_rb_first(st_for(cfqg, wl_class, i));
3110                 if (queue &&
3111                     (!key_valid || queue->rb_key < lowest_key)) {
3112                         lowest_key = queue->rb_key;
3113                         cur_best = i;
3114                         key_valid = true;
3115                 }
3116         }
3117
3118         return cur_best;
3119 }
3120
3121 static void
3122 choose_wl_class_and_type(struct cfq_data *cfqd, struct cfq_group *cfqg)
3123 {
3124         u64 slice;
3125         unsigned count;
3126         struct cfq_rb_root *st;
3127         u64 group_slice;
3128         enum wl_class_t original_class = cfqd->serving_wl_class;
3129         u64 now = ktime_get_ns();
3130
3131         /* Choose next priority. RT > BE > IDLE */
3132         if (cfq_group_busy_queues_wl(RT_WORKLOAD, cfqd, cfqg))
3133                 cfqd->serving_wl_class = RT_WORKLOAD;
3134         else if (cfq_group_busy_queues_wl(BE_WORKLOAD, cfqd, cfqg))
3135                 cfqd->serving_wl_class = BE_WORKLOAD;
3136         else {
3137                 cfqd->serving_wl_class = IDLE_WORKLOAD;
3138                 cfqd->workload_expires = now + jiffies_to_nsecs(1);
3139                 return;
3140         }
3141
3142         if (original_class != cfqd->serving_wl_class)
3143                 goto new_workload;
3144
3145         /*
3146          * For RT and BE, we have to choose also the type
3147          * (SYNC, SYNC_NOIDLE, ASYNC), and to compute a workload
3148          * expiration time
3149          */
3150         st = st_for(cfqg, cfqd->serving_wl_class, cfqd->serving_wl_type);
3151         count = st->count;
3152
3153         /*
3154          * check workload expiration, and that we still have other queues ready
3155          */
3156         if (count && !(now > cfqd->workload_expires))
3157                 return;
3158
3159 new_workload:
3160         /* otherwise select new workload type */
3161         cfqd->serving_wl_type = cfq_choose_wl_type(cfqd, cfqg,
3162                                         cfqd->serving_wl_class);
3163         st = st_for(cfqg, cfqd->serving_wl_class, cfqd->serving_wl_type);
3164         count = st->count;
3165
3166         /*
3167          * the workload slice is computed as a fraction of target latency
3168          * proportional to the number of queues in that workload, over
3169          * all the queues in the same priority class
3170          */
3171         group_slice = cfq_group_slice(cfqd, cfqg);
3172
3173         slice = div_u64(group_slice * count,
3174                 max_t(unsigned, cfqg->busy_queues_avg[cfqd->serving_wl_class],
3175                       cfq_group_busy_queues_wl(cfqd->serving_wl_class, cfqd,
3176                                         cfqg)));
3177
3178         if (cfqd->serving_wl_type == ASYNC_WORKLOAD) {
3179                 u64 tmp;
3180
3181                 /*
3182                  * Async queues are currently system wide. Just taking
3183                  * proportion of queues with-in same group will lead to higher
3184                  * async ratio system wide as generally root group is going
3185                  * to have higher weight. A more accurate thing would be to
3186                  * calculate system wide asnc/sync ratio.
3187                  */
3188                 tmp = cfqd->cfq_target_latency *
3189                         cfqg_busy_async_queues(cfqd, cfqg);
3190                 tmp = div_u64(tmp, cfqd->busy_queues);
3191                 slice = min_t(u64, slice, tmp);
3192
3193                 /* async workload slice is scaled down according to
3194                  * the sync/async slice ratio. */
3195                 slice = div64_u64(slice*cfqd->cfq_slice[0], cfqd->cfq_slice[1]);
3196         } else
3197                 /* sync workload slice is at least 2 * cfq_slice_idle */
3198                 slice = max(slice, 2 * cfqd->cfq_slice_idle);
3199
3200         slice = max_t(u64, slice, CFQ_MIN_TT);
3201         cfq_log(cfqd, "workload slice:%llu", slice);
3202         cfqd->workload_expires = now + slice;
3203 }
3204
3205 static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd)
3206 {
3207         struct cfq_rb_root *st = &cfqd->grp_service_tree;
3208         struct cfq_group *cfqg;
3209
3210         if (RB_EMPTY_ROOT(&st->rb.rb_root))
3211                 return NULL;
3212         cfqg = cfq_rb_first_group(st);
3213         update_min_vdisktime(st);
3214         return cfqg;
3215 }
3216
3217 static void cfq_choose_cfqg(struct cfq_data *cfqd)
3218 {
3219         struct cfq_group *cfqg = cfq_get_next_cfqg(cfqd);
3220         u64 now = ktime_get_ns();
3221
3222         cfqd->serving_group = cfqg;
3223
3224         /* Restore the workload type data */
3225         if (cfqg->saved_wl_slice) {
3226                 cfqd->workload_expires = now + cfqg->saved_wl_slice;
3227                 cfqd->serving_wl_type = cfqg->saved_wl_type;
3228                 cfqd->serving_wl_class = cfqg->saved_wl_class;
3229         } else
3230                 cfqd->workload_expires = now - 1;
3231
3232         choose_wl_class_and_type(cfqd, cfqg);
3233 }
3234
3235 /*
3236  * Select a queue for service. If we have a current active queue,
3237  * check whether to continue servicing it, or retrieve and set a new one.
3238  */
3239 static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
3240 {
3241         struct cfq_queue *cfqq, *new_cfqq = NULL;
3242         u64 now = ktime_get_ns();
3243
3244         cfqq = cfqd->active_queue;
3245         if (!cfqq)
3246                 goto new_queue;
3247
3248         if (!cfqd->rq_queued)
3249                 return NULL;
3250
3251         /*
3252          * We were waiting for group to get backlogged. Expire the queue
3253          */
3254         if (cfq_cfqq_wait_busy(cfqq) && !RB_EMPTY_ROOT(&cfqq->sort_list))
3255                 goto expire;
3256
3257         /*
3258          * The active queue has run out of time, expire it and select new.
3259          */
3260         if (cfq_slice_used(cfqq) && !cfq_cfqq_must_dispatch(cfqq)) {
3261                 /*
3262                  * If slice had not expired at the completion of last request
3263                  * we might not have turned on wait_busy flag. Don't expire
3264                  * the queue yet. Allow the group to get backlogged.
3265                  *
3266                  * The very fact that we have used the slice, that means we
3267                  * have been idling all along on this queue and it should be
3268                  * ok to wait for this request to complete.
3269                  */
3270                 if (cfqq->cfqg->nr_cfqq == 1 && RB_EMPTY_ROOT(&cfqq->sort_list)
3271                     && cfqq->dispatched && cfq_should_idle(cfqd, cfqq)) {
3272                         cfqq = NULL;
3273                         goto keep_queue;
3274                 } else
3275                         goto check_group_idle;
3276         }
3277
3278         /*
3279          * The active queue has requests and isn't expired, allow it to
3280          * dispatch.
3281          */
3282         if (!RB_EMPTY_ROOT(&cfqq->sort_list))
3283                 goto keep_queue;
3284
3285         /*
3286          * If another queue has a request waiting within our mean seek
3287          * distance, let it run.  The expire code will check for close
3288          * cooperators and put the close queue at the front of the service
3289          * tree.  If possible, merge the expiring queue with the new cfqq.
3290          */
3291         new_cfqq = cfq_close_cooperator(cfqd, cfqq);
3292         if (new_cfqq) {
3293                 if (!cfqq->new_cfqq)
3294                         cfq_setup_merge(cfqq, new_cfqq);
3295                 goto expire;
3296         }
3297
3298         /*
3299          * No requests pending. If the active queue still has requests in
3300          * flight or is idling for a new request, allow either of these
3301          * conditions to happen (or time out) before selecting a new queue.
3302          */
3303         if (hrtimer_active(&cfqd->idle_slice_timer)) {
3304                 cfqq = NULL;
3305                 goto keep_queue;
3306         }
3307
3308         /*
3309          * This is a deep seek queue, but the device is much faster than
3310          * the queue can deliver, don't idle
3311          **/
3312         if (CFQQ_SEEKY(cfqq) && cfq_cfqq_idle_window(cfqq) &&
3313             (cfq_cfqq_slice_new(cfqq) ||
3314             (cfqq->slice_end - now > now - cfqq->slice_start))) {
3315                 cfq_clear_cfqq_deep(cfqq);
3316                 cfq_clear_cfqq_idle_window(cfqq);
3317         }
3318
3319         if (cfqq->dispatched && cfq_should_idle(cfqd, cfqq)) {
3320                 cfqq = NULL;
3321                 goto keep_queue;
3322         }
3323
3324         /*
3325          * If group idle is enabled and there are requests dispatched from
3326          * this group, wait for requests to complete.
3327          */
3328 check_group_idle:
3329         if (cfqd->cfq_group_idle && cfqq->cfqg->nr_cfqq == 1 &&
3330             cfqq->cfqg->dispatched &&
3331             !cfq_io_thinktime_big(cfqd, &cfqq->cfqg->ttime, true)) {
3332                 cfqq = NULL;
3333                 goto keep_queue;
3334         }
3335
3336 expire:
3337         cfq_slice_expired(cfqd, 0);
3338 new_queue:
3339         /*
3340          * Current queue expired. Check if we have to switch to a new
3341          * service tree
3342          */
3343         if (!new_cfqq)
3344                 cfq_choose_cfqg(cfqd);
3345
3346         cfqq = cfq_set_active_queue(cfqd, new_cfqq);
3347 keep_queue:
3348         return cfqq;
3349 }
3350
3351 static int __cfq_forced_dispatch_cfqq(struct cfq_queue *cfqq)
3352 {
3353         int dispatched = 0;
3354
3355         while (cfqq->next_rq) {
3356                 cfq_dispatch_insert(cfqq->cfqd->queue, cfqq->next_rq);
3357                 dispatched++;
3358         }
3359
3360         BUG_ON(!list_empty(&cfqq->fifo));
3361
3362         /* By default cfqq is not expired if it is empty. Do it explicitly */
3363         __cfq_slice_expired(cfqq->cfqd, cfqq, 0);
3364         return dispatched;
3365 }
3366
3367 /*
3368  * Drain our current requests. Used for barriers and when switching
3369  * io schedulers on-the-fly.
3370  */
3371 static int cfq_forced_dispatch(struct cfq_data *cfqd)
3372 {
3373         struct cfq_queue *cfqq;
3374         int dispatched = 0;
3375
3376         /* Expire the timeslice of the current active queue first */
3377         cfq_slice_expired(cfqd, 0);
3378         while ((cfqq = cfq_get_next_queue_forced(cfqd)) != NULL) {
3379                 __cfq_set_active_queue(cfqd, cfqq);
3380                 dispatched += __cfq_forced_dispatch_cfqq(cfqq);
3381         }
3382
3383         BUG_ON(cfqd->busy_queues);
3384
3385         cfq_log(cfqd, "forced_dispatch=%d", dispatched);
3386         return dispatched;
3387 }
3388
3389 static inline bool cfq_slice_used_soon(struct cfq_data *cfqd,
3390         struct cfq_queue *cfqq)
3391 {
3392         u64 now = ktime_get_ns();
3393
3394         /* the queue hasn't finished any request, can't estimate */
3395         if (cfq_cfqq_slice_new(cfqq))
3396                 return true;
3397         if (now + cfqd->cfq_slice_idle * cfqq->dispatched > cfqq->slice_end)
3398                 return true;
3399
3400         return false;
3401 }
3402
3403 static bool cfq_may_dispatch(struct cfq_data *cfqd, struct cfq_queue *cfqq)
3404 {
3405         unsigned int max_dispatch;
3406
3407         if (cfq_cfqq_must_dispatch(cfqq))
3408                 return true;
3409
3410         /*
3411          * Drain async requests before we start sync IO
3412          */
3413         if (cfq_should_idle(cfqd, cfqq) && cfqd->rq_in_flight[BLK_RW_ASYNC])
3414                 return false;
3415
3416         /*
3417          * If this is an async queue and we have sync IO in flight, let it wait
3418          */
3419         if (cfqd->rq_in_flight[BLK_RW_SYNC] && !cfq_cfqq_sync(cfqq))
3420                 return false;
3421
3422         max_dispatch = max_t(unsigned int, cfqd->cfq_quantum / 2, 1);
3423         if (cfq_class_idle(cfqq))
3424                 max_dispatch = 1;
3425
3426         /*
3427          * Does this cfqq already have too much IO in flight?
3428          */
3429         if (cfqq->dispatched >= max_dispatch) {
3430                 bool promote_sync = false;
3431                 /*
3432                  * idle queue must always only have a single IO in flight
3433                  */
3434                 if (cfq_class_idle(cfqq))
3435                         return false;
3436
3437                 /*
3438                  * If there is only one sync queue
3439                  * we can ignore async queue here and give the sync
3440                  * queue no dispatch limit. The reason is a sync queue can
3441                  * preempt async queue, limiting the sync queue doesn't make
3442                  * sense. This is useful for aiostress test.
3443                  */
3444                 if (cfq_cfqq_sync(cfqq) && cfqd->busy_sync_queues == 1)
3445                         promote_sync = true;
3446
3447                 /*
3448                  * We have other queues, don't allow more IO from this one
3449                  */
3450                 if (cfqd->busy_queues > 1 && cfq_slice_used_soon(cfqd, cfqq) &&
3451                                 !promote_sync)
3452                         return false;
3453
3454                 /*
3455                  * Sole queue user, no limit
3456                  */
3457                 if (cfqd->busy_queues == 1 || promote_sync)
3458                         max_dispatch = -1;
3459                 else
3460                         /*
3461                          * Normally we start throttling cfqq when cfq_quantum/2
3462                          * requests have been dispatched. But we can drive
3463                          * deeper queue depths at the beginning of slice
3464                          * subjected to upper limit of cfq_quantum.
3465                          * */
3466                         max_dispatch = cfqd->cfq_quantum;
3467         }
3468
3469         /*
3470          * Async queues must wait a bit before being allowed dispatch.
3471          * We also ramp up the dispatch depth gradually for async IO,
3472          * based on the last sync IO we serviced
3473          */
3474         if (!cfq_cfqq_sync(cfqq) && cfqd->cfq_latency) {
3475                 u64 last_sync = ktime_get_ns() - cfqd->last_delayed_sync;
3476                 unsigned int depth;
3477
3478                 depth = div64_u64(last_sync, cfqd->cfq_slice[1]);
3479                 if (!depth && !cfqq->dispatched)
3480                         depth = 1;
3481                 if (depth < max_dispatch)
3482                         max_dispatch = depth;
3483         }
3484
3485         /*
3486          * If we're below the current max, allow a dispatch
3487          */
3488         return cfqq->dispatched < max_dispatch;
3489 }
3490
3491 /*
3492  * Dispatch a request from cfqq, moving them to the request queue
3493  * dispatch list.
3494  */
3495 static bool cfq_dispatch_request(struct cfq_data *cfqd, struct cfq_queue *cfqq)
3496 {
3497         struct request *rq;
3498
3499         BUG_ON(RB_EMPTY_ROOT(&cfqq->sort_list));
3500
3501         rq = cfq_check_fifo(cfqq);
3502         if (rq)
3503                 cfq_mark_cfqq_must_dispatch(cfqq);
3504
3505         if (!cfq_may_dispatch(cfqd, cfqq))
3506                 return false;
3507
3508         /*
3509          * follow expired path, else get first next available
3510          */
3511         if (!rq)
3512                 rq = cfqq->next_rq;
3513         else
3514                 cfq_log_cfqq(cfqq->cfqd, cfqq, "fifo=%p", rq);
3515
3516         /*
3517          * insert request into driver dispatch list
3518          */
3519         cfq_dispatch_insert(cfqd->queue, rq);
3520
3521         if (!cfqd->active_cic) {
3522                 struct cfq_io_cq *cic = RQ_CIC(rq);
3523
3524                 atomic_long_inc(&cic->icq.ioc->refcount);
3525                 cfqd->active_cic = cic;
3526         }
3527
3528         return true;
3529 }
3530
3531 /*
3532  * Find the cfqq that we need to service and move a request from that to the
3533  * dispatch list
3534  */
3535 static int cfq_dispatch_requests(struct request_queue *q, int force)
3536 {
3537         struct cfq_data *cfqd = q->elevator->elevator_data;
3538         struct cfq_queue *cfqq;
3539
3540         if (!cfqd->busy_queues)
3541                 return 0;
3542
3543         if (unlikely(force))
3544                 return cfq_forced_dispatch(cfqd);
3545
3546         cfqq = cfq_select_queue(cfqd);
3547         if (!cfqq)
3548                 return 0;
3549
3550         /*
3551          * Dispatch a request from this cfqq, if it is allowed
3552          */
3553         if (!cfq_dispatch_request(cfqd, cfqq))
3554                 return 0;
3555
3556         cfqq->slice_dispatch++;
3557         cfq_clear_cfqq_must_dispatch(cfqq);
3558
3559         /*
3560          * expire an async queue immediately if it has used up its slice. idle
3561          * queue always expire after 1 dispatch round.
3562          */
3563         if (cfqd->busy_queues > 1 && ((!cfq_cfqq_sync(cfqq) &&
3564             cfqq->slice_dispatch >= cfq_prio_to_maxrq(cfqd, cfqq)) ||
3565             cfq_class_idle(cfqq))) {
3566                 cfqq->slice_end = ktime_get_ns() + 1;
3567                 cfq_slice_expired(cfqd, 0);
3568         }
3569
3570         cfq_log_cfqq(cfqd, cfqq, "dispatched a request");
3571         return 1;
3572 }
3573
3574 /*
3575  * task holds one reference to the queue, dropped when task exits. each rq
3576  * in-flight on this queue also holds a reference, dropped when rq is freed.
3577  *
3578  * Each cfq queue took a reference on the parent group. Drop it now.
3579  * queue lock must be held here.
3580  */
3581 static void cfq_put_queue(struct cfq_queue *cfqq)
3582 {
3583         struct cfq_data *cfqd = cfqq->cfqd;
3584         struct cfq_group *cfqg;
3585
3586         BUG_ON(cfqq->ref <= 0);
3587
3588         cfqq->ref--;
3589         if (cfqq->ref)
3590                 return;
3591
3592         cfq_log_cfqq(cfqd, cfqq, "put_queue");
3593         BUG_ON(rb_first(&cfqq->sort_list));
3594         BUG_ON(cfqq->allocated[READ] + cfqq->allocated[WRITE]);
3595         cfqg = cfqq->cfqg;
3596
3597         if (unlikely(cfqd->active_queue == cfqq)) {
3598                 __cfq_slice_expired(cfqd, cfqq, 0);
3599                 cfq_schedule_dispatch(cfqd);
3600         }
3601
3602         BUG_ON(cfq_cfqq_on_rr(cfqq));
3603         kmem_cache_free(cfq_pool, cfqq);
3604         cfqg_put(cfqg);
3605 }
3606
3607 static void cfq_put_cooperator(struct cfq_queue *cfqq)
3608 {
3609         struct cfq_queue *__cfqq, *next;
3610
3611         /*
3612          * If this queue was scheduled to merge with another queue, be
3613          * sure to drop the reference taken on that queue (and others in
3614          * the merge chain).  See cfq_setup_merge and cfq_merge_cfqqs.
3615          */
3616         __cfqq = cfqq->new_cfqq;
3617         while (__cfqq) {
3618                 if (__cfqq == cfqq) {
3619                         WARN(1, "cfqq->new_cfqq loop detected\n");
3620                         break;
3621                 }
3622                 next = __cfqq->new_cfqq;
3623                 cfq_put_queue(__cfqq);
3624                 __cfqq = next;
3625         }
3626 }
3627
3628 static void cfq_exit_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq)
3629 {
3630         if (unlikely(cfqq == cfqd->active_queue)) {
3631                 __cfq_slice_expired(cfqd, cfqq, 0);
3632                 cfq_schedule_dispatch(cfqd);
3633         }
3634
3635         cfq_put_cooperator(cfqq);
3636
3637         cfq_put_queue(cfqq);
3638 }
3639
3640 static void cfq_init_icq(struct io_cq *icq)
3641 {
3642         struct cfq_io_cq *cic = icq_to_cic(icq);
3643
3644         cic->ttime.last_end_request = ktime_get_ns();
3645 }
3646
3647 static void cfq_exit_icq(struct io_cq *icq)
3648 {
3649         struct cfq_io_cq *cic = icq_to_cic(icq);
3650         struct cfq_data *cfqd = cic_to_cfqd(cic);
3651
3652         if (cic_to_cfqq(cic, false)) {
3653                 cfq_exit_cfqq(cfqd, cic_to_cfqq(cic, false));
3654                 cic_set_cfqq(cic, NULL, false);
3655         }
3656
3657         if (cic_to_cfqq(cic, true)) {
3658                 cfq_exit_cfqq(cfqd, cic_to_cfqq(cic, true));
3659                 cic_set_cfqq(cic, NULL, true);
3660         }
3661 }
3662
3663 static void cfq_init_prio_data(struct cfq_queue *cfqq, struct cfq_io_cq *cic)
3664 {
3665         struct task_struct *tsk = current;
3666         int ioprio_class;
3667
3668         if (!cfq_cfqq_prio_changed(cfqq))
3669                 return;
3670
3671         ioprio_class = IOPRIO_PRIO_CLASS(cic->ioprio);
3672         switch (ioprio_class) {
3673         default:
3674                 printk(KERN_ERR "cfq: bad prio %x\n", ioprio_class);
3675                 /* fall through */
3676         case IOPRIO_CLASS_NONE:
3677                 /*
3678                  * no prio set, inherit CPU scheduling settings
3679                  */
3680                 cfqq->ioprio = task_nice_ioprio(tsk);
3681                 cfqq->ioprio_class = task_nice_ioclass(tsk);
3682                 break;
3683         case IOPRIO_CLASS_RT:
3684                 cfqq->ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
3685                 cfqq->ioprio_class = IOPRIO_CLASS_RT;
3686                 break;
3687         case IOPRIO_CLASS_BE:
3688                 cfqq->ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
3689                 cfqq->ioprio_class = IOPRIO_CLASS_BE;
3690                 break;
3691         case IOPRIO_CLASS_IDLE:
3692                 cfqq->ioprio_class = IOPRIO_CLASS_IDLE;
3693                 cfqq->ioprio = 7;
3694                 cfq_clear_cfqq_idle_window(cfqq);
3695                 break;
3696         }
3697
3698         /*
3699          * keep track of original prio settings in case we have to temporarily
3700          * elevate the priority of this queue
3701          */
3702         cfqq->org_ioprio = cfqq->ioprio;
3703         cfqq->org_ioprio_class = cfqq->ioprio_class;
3704         cfq_clear_cfqq_prio_changed(cfqq);
3705 }
3706
3707 static void check_ioprio_changed(struct cfq_io_cq *cic, struct bio *bio)
3708 {
3709         int ioprio = cic->icq.ioc->ioprio;
3710         struct cfq_data *cfqd = cic_to_cfqd(cic);
3711         struct cfq_queue *cfqq;
3712
3713         /*
3714          * Check whether ioprio has changed.  The condition may trigger
3715          * spuriously on a newly created cic but there's no harm.
3716          */
3717         if (unlikely(!cfqd) || likely(cic->ioprio == ioprio))
3718                 return;
3719
3720         cfqq = cic_to_cfqq(cic, false);
3721         if (cfqq) {
3722                 cfq_put_queue(cfqq);
3723                 cfqq = cfq_get_queue(cfqd, BLK_RW_ASYNC, cic, bio);
3724                 cic_set_cfqq(cic, cfqq, false);
3725         }
3726
3727         cfqq = cic_to_cfqq(cic, true);
3728         if (cfqq)
3729                 cfq_mark_cfqq_prio_changed(cfqq);
3730
3731         cic->ioprio = ioprio;
3732 }
3733
3734 static void cfq_init_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
3735                           pid_t pid, bool is_sync)
3736 {
3737         RB_CLEAR_NODE(&cfqq->rb_node);
3738         RB_CLEAR_NODE(&cfqq->p_node);
3739         INIT_LIST_HEAD(&cfqq->fifo);
3740
3741         cfqq->ref = 0;
3742         cfqq->cfqd = cfqd;
3743
3744         cfq_mark_cfqq_prio_changed(cfqq);
3745
3746         if (is_sync) {
3747                 if (!cfq_class_idle(cfqq))
3748                         cfq_mark_cfqq_idle_window(cfqq);
3749                 cfq_mark_cfqq_sync(cfqq);
3750         }
3751         cfqq->pid = pid;
3752 }
3753
3754 #ifdef CONFIG_CFQ_GROUP_IOSCHED
3755 static void check_blkcg_changed(struct cfq_io_cq *cic, struct bio *bio)
3756 {
3757         struct cfq_data *cfqd = cic_to_cfqd(cic);
3758         struct cfq_queue *cfqq;
3759         uint64_t serial_nr;
3760
3761         rcu_read_lock();
3762         serial_nr = __bio_blkcg(bio)->css.serial_nr;
3763         rcu_read_unlock();
3764
3765         /*
3766          * Check whether blkcg has changed.  The condition may trigger
3767          * spuriously on a newly created cic but there's no harm.
3768          */
3769         if (unlikely(!cfqd) || likely(cic->blkcg_serial_nr == serial_nr))
3770                 return;
3771
3772         /*
3773          * Drop reference to queues.  New queues will be assigned in new
3774          * group upon arrival of fresh requests.
3775          */
3776         cfqq = cic_to_cfqq(cic, false);
3777         if (cfqq) {
3778                 cfq_log_cfqq(cfqd, cfqq, "changed cgroup");
3779                 cic_set_cfqq(cic, NULL, false);
3780                 cfq_put_queue(cfqq);
3781         }
3782
3783         cfqq = cic_to_cfqq(cic, true);
3784         if (cfqq) {
3785                 cfq_log_cfqq(cfqd, cfqq, "changed cgroup");
3786                 cic_set_cfqq(cic, NULL, true);
3787                 cfq_put_queue(cfqq);
3788         }
3789
3790         cic->blkcg_serial_nr = serial_nr;
3791 }
3792 #else
3793 static inline void check_blkcg_changed(struct cfq_io_cq *cic, struct bio *bio)
3794 {
3795 }
3796 #endif  /* CONFIG_CFQ_GROUP_IOSCHED */
3797
3798 static struct cfq_queue **
3799 cfq_async_queue_prio(struct cfq_group *cfqg, int ioprio_class, int ioprio)
3800 {
3801         switch (ioprio_class) {
3802         case IOPRIO_CLASS_RT:
3803                 return &cfqg->async_cfqq[0][ioprio];
3804         case IOPRIO_CLASS_NONE:
3805                 ioprio = IOPRIO_NORM;
3806                 /* fall through */
3807         case IOPRIO_CLASS_BE:
3808                 return &cfqg->async_cfqq[1][ioprio];
3809         case IOPRIO_CLASS_IDLE:
3810                 return &cfqg->async_idle_cfqq;
3811         default:
3812                 BUG();
3813         }
3814 }
3815
3816 static struct cfq_queue *
3817 cfq_get_queue(struct cfq_data *cfqd, bool is_sync, struct cfq_io_cq *cic,
3818               struct bio *bio)
3819 {
3820         int ioprio_class = IOPRIO_PRIO_CLASS(cic->ioprio);
3821         int ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
3822         struct cfq_queue **async_cfqq = NULL;
3823         struct cfq_queue *cfqq;
3824         struct cfq_group *cfqg;
3825
3826         rcu_read_lock();
3827         cfqg = cfq_lookup_cfqg(cfqd, __bio_blkcg(bio));
3828         if (!cfqg) {
3829                 cfqq = &cfqd->oom_cfqq;
3830                 goto out;
3831         }
3832
3833         if (!is_sync) {
3834                 if (!ioprio_valid(cic->ioprio)) {
3835                         struct task_struct *tsk = current;
3836                         ioprio = task_nice_ioprio(tsk);
3837                         ioprio_class = task_nice_ioclass(tsk);
3838                 }
3839                 async_cfqq = cfq_async_queue_prio(cfqg, ioprio_class, ioprio);
3840                 cfqq = *async_cfqq;
3841                 if (cfqq)
3842                         goto out;
3843         }
3844
3845         cfqq = kmem_cache_alloc_node(cfq_pool,
3846                                      GFP_NOWAIT | __GFP_ZERO | __GFP_NOWARN,
3847                                      cfqd->queue->node);
3848         if (!cfqq) {
3849                 cfqq = &cfqd->oom_cfqq;
3850                 goto out;
3851         }
3852
3853         /* cfq_init_cfqq() assumes cfqq->ioprio_class is initialized. */
3854         cfqq->ioprio_class = IOPRIO_CLASS_NONE;
3855         cfq_init_cfqq(cfqd, cfqq, current->pid, is_sync);
3856         cfq_init_prio_data(cfqq, cic);
3857         cfq_link_cfqq_cfqg(cfqq, cfqg);
3858         cfq_log_cfqq(cfqd, cfqq, "alloced");
3859
3860         if (async_cfqq) {
3861                 /* a new async queue is created, pin and remember */
3862                 cfqq->ref++;
3863                 *async_cfqq = cfqq;
3864         }
3865 out:
3866         cfqq->ref++;
3867         rcu_read_unlock();
3868         return cfqq;
3869 }
3870
3871 static void
3872 __cfq_update_io_thinktime(struct cfq_ttime *ttime, u64 slice_idle)
3873 {
3874         u64 elapsed = ktime_get_ns() - ttime->last_end_request;
3875         elapsed = min(elapsed, 2UL * slice_idle);
3876
3877         ttime->ttime_samples = (7*ttime->ttime_samples + 256) / 8;
3878         ttime->ttime_total = div_u64(7*ttime->ttime_total + 256*elapsed,  8);
3879         ttime->ttime_mean = div64_ul(ttime->ttime_total + 128,
3880                                      ttime->ttime_samples);
3881 }
3882
3883 static void
3884 cfq_update_io_thinktime(struct cfq_data *cfqd, struct cfq_queue *cfqq,
3885                         struct cfq_io_cq *cic)
3886 {
3887         if (cfq_cfqq_sync(cfqq)) {
3888                 __cfq_update_io_thinktime(&cic->ttime, cfqd->cfq_slice_idle);
3889                 __cfq_update_io_thinktime(&cfqq->service_tree->ttime,
3890                         cfqd->cfq_slice_idle);
3891         }
3892 #ifdef CONFIG_CFQ_GROUP_IOSCHED
3893         __cfq_update_io_thinktime(&cfqq->cfqg->ttime, cfqd->cfq_group_idle);
3894 #endif
3895 }
3896
3897 static void
3898 cfq_update_io_seektime(struct cfq_data *cfqd, struct cfq_queue *cfqq,
3899                        struct request *rq)
3900 {
3901         sector_t sdist = 0;
3902         sector_t n_sec = blk_rq_sectors(rq);
3903         if (cfqq->last_request_pos) {
3904                 if (cfqq->last_request_pos < blk_rq_pos(rq))
3905                         sdist = blk_rq_pos(rq) - cfqq->last_request_pos;
3906                 else
3907                         sdist = cfqq->last_request_pos - blk_rq_pos(rq);
3908         }
3909
3910         cfqq->seek_history <<= 1;
3911         if (blk_queue_nonrot(cfqd->queue))
3912                 cfqq->seek_history |= (n_sec < CFQQ_SECT_THR_NONROT);
3913         else
3914                 cfqq->seek_history |= (sdist > CFQQ_SEEK_THR);
3915 }
3916
3917 static inline bool req_noidle(struct request *req)
3918 {
3919         return req_op(req) == REQ_OP_WRITE &&
3920                 (req->cmd_flags & (REQ_SYNC | REQ_IDLE)) == REQ_SYNC;
3921 }
3922
3923 /*
3924  * Disable idle window if the process thinks too long or seeks so much that
3925  * it doesn't matter
3926  */
3927 static void
3928 cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
3929                        struct cfq_io_cq *cic)
3930 {
3931         int old_idle, enable_idle;
3932
3933         /*
3934          * Don't idle for async or idle io prio class
3935          */
3936         if (!cfq_cfqq_sync(cfqq) || cfq_class_idle(cfqq))
3937                 return;
3938
3939         enable_idle = old_idle = cfq_cfqq_idle_window(cfqq);
3940
3941         if (cfqq->queued[0] + cfqq->queued[1] >= 4)
3942                 cfq_mark_cfqq_deep(cfqq);
3943
3944         if (cfqq->next_rq && req_noidle(cfqq->next_rq))
3945                 enable_idle = 0;
3946         else if (!atomic_read(&cic->icq.ioc->active_ref) ||
3947                  !cfqd->cfq_slice_idle ||
3948                  (!cfq_cfqq_deep(cfqq) && CFQQ_SEEKY(cfqq)))
3949                 enable_idle = 0;
3950         else if (sample_valid(cic->ttime.ttime_samples)) {
3951                 if (cic->ttime.ttime_mean > cfqd->cfq_slice_idle)
3952                         enable_idle = 0;
3953                 else
3954                         enable_idle = 1;
3955         }
3956
3957         if (old_idle != enable_idle) {
3958                 cfq_log_cfqq(cfqd, cfqq, "idle=%d", enable_idle);
3959                 if (enable_idle)
3960                         cfq_mark_cfqq_idle_window(cfqq);
3961                 else
3962                         cfq_clear_cfqq_idle_window(cfqq);
3963         }
3964 }
3965
3966 /*
3967  * Check if new_cfqq should preempt the currently active queue. Return 0 for
3968  * no or if we aren't sure, a 1 will cause a preempt.
3969  */
3970 static bool
3971 cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
3972                    struct request *rq)
3973 {
3974         struct cfq_queue *cfqq;
3975
3976         cfqq = cfqd->active_queue;
3977         if (!cfqq)
3978                 return false;
3979
3980         if (cfq_class_idle(new_cfqq))
3981                 return false;
3982
3983         if (cfq_class_idle(cfqq))
3984                 return true;
3985
3986         /*
3987          * Don't allow a non-RT request to preempt an ongoing RT cfqq timeslice.
3988          */
3989         if (cfq_class_rt(cfqq) && !cfq_class_rt(new_cfqq))
3990                 return false;
3991
3992         /*
3993          * if the new request is sync, but the currently running queue is
3994          * not, let the sync request have priority.
3995          */
3996         if (rq_is_sync(rq) && !cfq_cfqq_sync(cfqq) && !cfq_cfqq_must_dispatch(cfqq))
3997                 return true;
3998
3999         /*
4000          * Treat ancestors of current cgroup the same way as current cgroup.
4001          * For anybody else we disallow preemption to guarantee service
4002          * fairness among cgroups.
4003          */
4004         if (!cfqg_is_descendant(cfqq->cfqg, new_cfqq->cfqg))
4005                 return false;
4006
4007         if (cfq_slice_used(cfqq))
4008                 return true;
4009
4010         /*
4011          * Allow an RT request to pre-empt an ongoing non-RT cfqq timeslice.
4012          */
4013         if (cfq_class_rt(new_cfqq) && !cfq_class_rt(cfqq))
4014                 return true;
4015
4016         WARN_ON_ONCE(cfqq->ioprio_class != new_cfqq->ioprio_class);
4017         /* Allow preemption only if we are idling on sync-noidle tree */
4018         if (cfqd->serving_wl_type == SYNC_NOIDLE_WORKLOAD &&
4019             cfqq_type(new_cfqq) == SYNC_NOIDLE_WORKLOAD &&
4020             RB_EMPTY_ROOT(&cfqq->sort_list))
4021                 return true;
4022
4023         /*
4024          * So both queues are sync. Let the new request get disk time if
4025          * it's a metadata request and the current queue is doing regular IO.
4026          */
4027         if ((rq->cmd_flags & REQ_PRIO) && !cfqq->prio_pending)
4028                 return true;
4029
4030         /* An idle queue should not be idle now for some reason */
4031         if (RB_EMPTY_ROOT(&cfqq->sort_list) && !cfq_should_idle(cfqd, cfqq))
4032                 return true;
4033
4034         if (!cfqd->active_cic || !cfq_cfqq_wait_request(cfqq))
4035                 return false;
4036
4037         /*
4038          * if this request is as-good as one we would expect from the
4039          * current cfqq, let it preempt
4040          */
4041         if (cfq_rq_close(cfqd, cfqq, rq))
4042                 return true;
4043
4044         return false;
4045 }
4046
4047 /*
4048  * cfqq preempts the active queue. if we allowed preempt with no slice left,
4049  * let it have half of its nominal slice.
4050  */
4051 static void cfq_preempt_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
4052 {
4053         enum wl_type_t old_type = cfqq_type(cfqd->active_queue);
4054
4055         cfq_log_cfqq(cfqd, cfqq, "preempt");
4056         cfq_slice_expired(cfqd, 1);
4057
4058         /*
4059          * workload type is changed, don't save slice, otherwise preempt
4060          * doesn't happen
4061          */
4062         if (old_type != cfqq_type(cfqq))
4063                 cfqq->cfqg->saved_wl_slice = 0;
4064
4065         /*
4066          * Put the new queue at the front of the of the current list,
4067          * so we know that it will be selected next.
4068          */
4069         BUG_ON(!cfq_cfqq_on_rr(cfqq));
4070
4071         cfq_service_tree_add(cfqd, cfqq, 1);
4072
4073         cfqq->slice_end = 0;
4074         cfq_mark_cfqq_slice_new(cfqq);
4075 }
4076
4077 /*
4078  * Called when a new fs request (rq) is added (to cfqq). Check if there's
4079  * something we should do about it
4080  */
4081 static void
4082 cfq_rq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
4083                 struct request *rq)
4084 {
4085         struct cfq_io_cq *cic = RQ_CIC(rq);
4086
4087         cfqd->rq_queued++;
4088         if (rq->cmd_flags & REQ_PRIO)
4089                 cfqq->prio_pending++;
4090
4091         cfq_update_io_thinktime(cfqd, cfqq, cic);
4092         cfq_update_io_seektime(cfqd, cfqq, rq);
4093         cfq_update_idle_window(cfqd, cfqq, cic);
4094
4095         cfqq->last_request_pos = blk_rq_pos(rq) + blk_rq_sectors(rq);
4096
4097         if (cfqq == cfqd->active_queue) {
4098                 /*
4099                  * Remember that we saw a request from this process, but
4100                  * don't start queuing just yet. Otherwise we risk seeing lots
4101                  * of tiny requests, because we disrupt the normal plugging
4102                  * and merging. If the request is already larger than a single
4103                  * page, let it rip immediately. For that case we assume that
4104                  * merging is already done. Ditto for a busy system that
4105                  * has other work pending, don't risk delaying until the
4106                  * idle timer unplug to continue working.
4107                  */
4108                 if (cfq_cfqq_wait_request(cfqq)) {
4109                         if (blk_rq_bytes(rq) > PAGE_SIZE ||
4110                             cfqd->busy_queues > 1) {
4111                                 cfq_del_timer(cfqd, cfqq);
4112                                 cfq_clear_cfqq_wait_request(cfqq);
4113                                 __blk_run_queue(cfqd->queue);
4114                         } else {
4115                                 cfqg_stats_update_idle_time(cfqq->cfqg);
4116                                 cfq_mark_cfqq_must_dispatch(cfqq);
4117                         }
4118                 }
4119         } else if (cfq_should_preempt(cfqd, cfqq, rq)) {
4120                 /*
4121                  * not the active queue - expire current slice if it is
4122                  * idle and has expired it's mean thinktime or this new queue
4123                  * has some old slice time left and is of higher priority or
4124                  * this new queue is RT and the current one is BE
4125                  */
4126                 cfq_preempt_queue(cfqd, cfqq);
4127                 __blk_run_queue(cfqd->queue);
4128         }
4129 }
4130
4131 static void cfq_insert_request(struct request_queue *q, struct request *rq)
4132 {
4133         struct cfq_data *cfqd = q->elevator->elevator_data;
4134         struct cfq_queue *cfqq = RQ_CFQQ(rq);
4135
4136         cfq_log_cfqq(cfqd, cfqq, "insert_request");
4137         cfq_init_prio_data(cfqq, RQ_CIC(rq));
4138
4139         rq->fifo_time = ktime_get_ns() + cfqd->cfq_fifo_expire[rq_is_sync(rq)];
4140         list_add_tail(&rq->queuelist, &cfqq->fifo);
4141         cfq_add_rq_rb(rq);
4142         cfqg_stats_update_io_add(RQ_CFQG(rq), cfqd->serving_group,
4143                                  rq->cmd_flags);
4144         cfq_rq_enqueued(cfqd, cfqq, rq);
4145 }
4146
4147 /*
4148  * Update hw_tag based on peak queue depth over 50 samples under
4149  * sufficient load.
4150  */
4151 static void cfq_update_hw_tag(struct cfq_data *cfqd)
4152 {
4153         struct cfq_queue *cfqq = cfqd->active_queue;
4154
4155         if (cfqd->rq_in_driver > cfqd->hw_tag_est_depth)
4156                 cfqd->hw_tag_est_depth = cfqd->rq_in_driver;
4157
4158         if (cfqd->hw_tag == 1)
4159                 return;
4160
4161         if (cfqd->rq_queued <= CFQ_HW_QUEUE_MIN &&
4162             cfqd->rq_in_driver <= CFQ_HW_QUEUE_MIN)
4163                 return;
4164
4165         /*
4166          * If active queue hasn't enough requests and can idle, cfq might not
4167          * dispatch sufficient requests to hardware. Don't zero hw_tag in this
4168          * case
4169          */
4170         if (cfqq && cfq_cfqq_idle_window(cfqq) &&
4171             cfqq->dispatched + cfqq->queued[0] + cfqq->queued[1] <
4172             CFQ_HW_QUEUE_MIN && cfqd->rq_in_driver < CFQ_HW_QUEUE_MIN)
4173                 return;
4174
4175         if (cfqd->hw_tag_samples++ < 50)
4176                 return;
4177
4178         if (cfqd->hw_tag_est_depth >= CFQ_HW_QUEUE_MIN)
4179                 cfqd->hw_tag = 1;
4180         else
4181                 cfqd->hw_tag = 0;
4182 }
4183
4184 static bool cfq_should_wait_busy(struct cfq_data *cfqd, struct cfq_queue *cfqq)
4185 {
4186         struct cfq_io_cq *cic = cfqd->active_cic;
4187         u64 now = ktime_get_ns();
4188
4189         /* If the queue already has requests, don't wait */
4190         if (!RB_EMPTY_ROOT(&cfqq->sort_list))
4191                 return false;
4192
4193         /* If there are other queues in the group, don't wait */
4194         if (cfqq->cfqg->nr_cfqq > 1)
4195                 return false;
4196
4197         /* the only queue in the group, but think time is big */
4198         if (cfq_io_thinktime_big(cfqd, &cfqq->cfqg->ttime, true))
4199                 return false;
4200
4201         if (cfq_slice_used(cfqq))
4202                 return true;
4203
4204         /* if slice left is less than think time, wait busy */
4205         if (cic && sample_valid(cic->ttime.ttime_samples)
4206             && (cfqq->slice_end - now < cic->ttime.ttime_mean))
4207                 return true;
4208
4209         /*
4210          * If think times is less than a jiffy than ttime_mean=0 and above
4211          * will not be true. It might happen that slice has not expired yet
4212          * but will expire soon (4-5 ns) during select_queue(). To cover the
4213          * case where think time is less than a jiffy, mark the queue wait
4214          * busy if only 1 jiffy is left in the slice.
4215          */
4216         if (cfqq->slice_end - now <= jiffies_to_nsecs(1))
4217                 return true;
4218
4219         return false;
4220 }
4221
4222 static void cfq_completed_request(struct request_queue *q, struct request *rq)
4223 {
4224         struct cfq_queue *cfqq = RQ_CFQQ(rq);
4225         struct cfq_data *cfqd = cfqq->cfqd;
4226         const int sync = rq_is_sync(rq);
4227         u64 now = ktime_get_ns();
4228
4229         cfq_log_cfqq(cfqd, cfqq, "complete rqnoidle %d", req_noidle(rq));
4230
4231         cfq_update_hw_tag(cfqd);
4232
4233         WARN_ON(!cfqd->rq_in_driver);
4234         WARN_ON(!cfqq->dispatched);
4235         cfqd->rq_in_driver--;
4236         cfqq->dispatched--;
4237         (RQ_CFQG(rq))->dispatched--;
4238         cfqg_stats_update_completion(cfqq->cfqg, rq->start_time_ns,
4239                                      rq->io_start_time_ns, rq->cmd_flags);
4240
4241         cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]--;
4242
4243         if (sync) {
4244                 struct cfq_rb_root *st;
4245
4246                 RQ_CIC(rq)->ttime.last_end_request = now;
4247
4248                 if (cfq_cfqq_on_rr(cfqq))
4249                         st = cfqq->service_tree;
4250                 else
4251                         st = st_for(cfqq->cfqg, cfqq_class(cfqq),
4252                                         cfqq_type(cfqq));
4253
4254                 st->ttime.last_end_request = now;
4255                 if (rq->start_time_ns + cfqd->cfq_fifo_expire[1] <= now)
4256                         cfqd->last_delayed_sync = now;
4257         }
4258
4259 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4260         cfqq->cfqg->ttime.last_end_request = now;
4261 #endif
4262
4263         /*
4264          * If this is the active queue, check if it needs to be expired,
4265          * or if we want to idle in case it has no pending requests.
4266          */
4267         if (cfqd->active_queue == cfqq) {
4268                 const bool cfqq_empty = RB_EMPTY_ROOT(&cfqq->sort_list);
4269
4270                 if (cfq_cfqq_slice_new(cfqq)) {
4271                         cfq_set_prio_slice(cfqd, cfqq);
4272                         cfq_clear_cfqq_slice_new(cfqq);
4273                 }
4274
4275                 /*
4276                  * Should we wait for next request to come in before we expire
4277                  * the queue.
4278                  */
4279                 if (cfq_should_wait_busy(cfqd, cfqq)) {
4280                         u64 extend_sl = cfqd->cfq_slice_idle;
4281                         if (!cfqd->cfq_slice_idle)
4282                                 extend_sl = cfqd->cfq_group_idle;
4283                         cfqq->slice_end = now + extend_sl;
4284                         cfq_mark_cfqq_wait_busy(cfqq);
4285                         cfq_log_cfqq(cfqd, cfqq, "will busy wait");
4286                 }
4287
4288                 /*
4289                  * Idling is not enabled on:
4290                  * - expired queues
4291                  * - idle-priority queues
4292                  * - async queues
4293                  * - queues with still some requests queued
4294                  * - when there is a close cooperator
4295                  */
4296                 if (cfq_slice_used(cfqq) || cfq_class_idle(cfqq))
4297                         cfq_slice_expired(cfqd, 1);
4298                 else if (sync && cfqq_empty &&
4299                          !cfq_close_cooperator(cfqd, cfqq)) {
4300                         cfq_arm_slice_timer(cfqd);
4301                 }
4302         }
4303
4304         if (!cfqd->rq_in_driver)
4305                 cfq_schedule_dispatch(cfqd);
4306 }
4307
4308 static void cfqq_boost_on_prio(struct cfq_queue *cfqq, unsigned int op)
4309 {
4310         /*
4311          * If REQ_PRIO is set, boost class and prio level, if it's below
4312          * BE/NORM. If prio is not set, restore the potentially boosted
4313          * class/prio level.
4314          */
4315         if (!(op & REQ_PRIO)) {
4316                 cfqq->ioprio_class = cfqq->org_ioprio_class;
4317                 cfqq->ioprio = cfqq->org_ioprio;
4318         } else {
4319                 if (cfq_class_idle(cfqq))
4320                         cfqq->ioprio_class = IOPRIO_CLASS_BE;
4321                 if (cfqq->ioprio > IOPRIO_NORM)
4322                         cfqq->ioprio = IOPRIO_NORM;
4323         }
4324 }
4325
4326 static inline int __cfq_may_queue(struct cfq_queue *cfqq)
4327 {
4328         if (cfq_cfqq_wait_request(cfqq) && !cfq_cfqq_must_alloc_slice(cfqq)) {
4329                 cfq_mark_cfqq_must_alloc_slice(cfqq);
4330                 return ELV_MQUEUE_MUST;
4331         }
4332
4333         return ELV_MQUEUE_MAY;
4334 }
4335
4336 static int cfq_may_queue(struct request_queue *q, unsigned int op)
4337 {
4338         struct cfq_data *cfqd = q->elevator->elevator_data;
4339         struct task_struct *tsk = current;
4340         struct cfq_io_cq *cic;
4341         struct cfq_queue *cfqq;
4342
4343         /*
4344          * don't force setup of a queue from here, as a call to may_queue
4345          * does not necessarily imply that a request actually will be queued.
4346          * so just lookup a possibly existing queue, or return 'may queue'
4347          * if that fails
4348          */
4349         cic = cfq_cic_lookup(cfqd, tsk->io_context);
4350         if (!cic)
4351                 return ELV_MQUEUE_MAY;
4352
4353         cfqq = cic_to_cfqq(cic, op_is_sync(op));
4354         if (cfqq) {
4355                 cfq_init_prio_data(cfqq, cic);
4356                 cfqq_boost_on_prio(cfqq, op);
4357
4358                 return __cfq_may_queue(cfqq);
4359         }
4360
4361         return ELV_MQUEUE_MAY;
4362 }
4363
4364 /*
4365  * queue lock held here
4366  */
4367 static void cfq_put_request(struct request *rq)
4368 {
4369         struct cfq_queue *cfqq = RQ_CFQQ(rq);
4370
4371         if (cfqq) {
4372                 const int rw = rq_data_dir(rq);
4373
4374                 BUG_ON(!cfqq->allocated[rw]);
4375                 cfqq->allocated[rw]--;
4376
4377                 /* Put down rq reference on cfqg */
4378                 cfqg_put(RQ_CFQG(rq));
4379                 rq->elv.priv[0] = NULL;
4380                 rq->elv.priv[1] = NULL;
4381
4382                 cfq_put_queue(cfqq);
4383         }
4384 }
4385
4386 static struct cfq_queue *
4387 cfq_merge_cfqqs(struct cfq_data *cfqd, struct cfq_io_cq *cic,
4388                 struct cfq_queue *cfqq)
4389 {
4390         cfq_log_cfqq(cfqd, cfqq, "merging with queue %p", cfqq->new_cfqq);
4391         cic_set_cfqq(cic, cfqq->new_cfqq, 1);
4392         cfq_mark_cfqq_coop(cfqq->new_cfqq);
4393         cfq_put_queue(cfqq);
4394         return cic_to_cfqq(cic, 1);
4395 }
4396
4397 /*
4398  * Returns NULL if a new cfqq should be allocated, or the old cfqq if this
4399  * was the last process referring to said cfqq.
4400  */
4401 static struct cfq_queue *
4402 split_cfqq(struct cfq_io_cq *cic, struct cfq_queue *cfqq)
4403 {
4404         if (cfqq_process_refs(cfqq) == 1) {
4405                 cfqq->pid = current->pid;
4406                 cfq_clear_cfqq_coop(cfqq);
4407                 cfq_clear_cfqq_split_coop(cfqq);
4408                 return cfqq;
4409         }
4410
4411         cic_set_cfqq(cic, NULL, 1);
4412
4413         cfq_put_cooperator(cfqq);
4414
4415         cfq_put_queue(cfqq);
4416         return NULL;
4417 }
4418 /*
4419  * Allocate cfq data structures associated with this request.
4420  */
4421 static int
4422 cfq_set_request(struct request_queue *q, struct request *rq, struct bio *bio,
4423                 gfp_t gfp_mask)
4424 {
4425         struct cfq_data *cfqd = q->elevator->elevator_data;
4426         struct cfq_io_cq *cic = icq_to_cic(rq->elv.icq);
4427         const int rw = rq_data_dir(rq);
4428         const bool is_sync = rq_is_sync(rq);
4429         struct cfq_queue *cfqq;
4430
4431         spin_lock_irq(q->queue_lock);
4432
4433         check_ioprio_changed(cic, bio);
4434         check_blkcg_changed(cic, bio);
4435 new_queue:
4436         cfqq = cic_to_cfqq(cic, is_sync);
4437         if (!cfqq || cfqq == &cfqd->oom_cfqq) {
4438                 if (cfqq)
4439                         cfq_put_queue(cfqq);
4440                 cfqq = cfq_get_queue(cfqd, is_sync, cic, bio);
4441                 cic_set_cfqq(cic, cfqq, is_sync);
4442         } else {
4443                 /*
4444                  * If the queue was seeky for too long, break it apart.
4445                  */
4446                 if (cfq_cfqq_coop(cfqq) && cfq_cfqq_split_coop(cfqq)) {
4447                         cfq_log_cfqq(cfqd, cfqq, "breaking apart cfqq");
4448                         cfqq = split_cfqq(cic, cfqq);
4449                         if (!cfqq)
4450                                 goto new_queue;
4451                 }
4452
4453                 /*
4454                  * Check to see if this queue is scheduled to merge with
4455                  * another, closely cooperating queue.  The merging of
4456                  * queues happens here as it must be done in process context.
4457                  * The reference on new_cfqq was taken in merge_cfqqs.
4458                  */
4459                 if (cfqq->new_cfqq)
4460                         cfqq = cfq_merge_cfqqs(cfqd, cic, cfqq);
4461         }
4462
4463         cfqq->allocated[rw]++;
4464
4465         cfqq->ref++;
4466         cfqg_get(cfqq->cfqg);
4467         rq->elv.priv[0] = cfqq;
4468         rq->elv.priv[1] = cfqq->cfqg;
4469         spin_unlock_irq(q->queue_lock);
4470
4471         return 0;
4472 }
4473
4474 static void cfq_kick_queue(struct work_struct *work)
4475 {
4476         struct cfq_data *cfqd =
4477                 container_of(work, struct cfq_data, unplug_work);
4478         struct request_queue *q = cfqd->queue;
4479
4480         spin_lock_irq(q->queue_lock);
4481         __blk_run_queue(cfqd->queue);
4482         spin_unlock_irq(q->queue_lock);
4483 }
4484
4485 /*
4486  * Timer running if the active_queue is currently idling inside its time slice
4487  */
4488 static enum hrtimer_restart cfq_idle_slice_timer(struct hrtimer *timer)
4489 {
4490         struct cfq_data *cfqd = container_of(timer, struct cfq_data,
4491                                              idle_slice_timer);
4492         struct cfq_queue *cfqq;
4493         unsigned long flags;
4494         int timed_out = 1;
4495
4496         cfq_log(cfqd, "idle timer fired");
4497
4498         spin_lock_irqsave(cfqd->queue->queue_lock, flags);
4499
4500         cfqq = cfqd->active_queue;
4501         if (cfqq) {
4502                 timed_out = 0;
4503
4504                 /*
4505                  * We saw a request before the queue expired, let it through
4506                  */
4507                 if (cfq_cfqq_must_dispatch(cfqq))
4508                         goto out_kick;
4509
4510                 /*
4511                  * expired
4512                  */
4513                 if (cfq_slice_used(cfqq))
4514                         goto expire;
4515
4516                 /*
4517                  * only expire and reinvoke request handler, if there are
4518                  * other queues with pending requests
4519                  */
4520                 if (!cfqd->busy_queues)
4521                         goto out_cont;
4522
4523                 /*
4524                  * not expired and it has a request pending, let it dispatch
4525                  */
4526                 if (!RB_EMPTY_ROOT(&cfqq->sort_list))
4527                         goto out_kick;
4528
4529                 /*
4530                  * Queue depth flag is reset only when the idle didn't succeed
4531                  */
4532                 cfq_clear_cfqq_deep(cfqq);
4533         }
4534 expire:
4535         cfq_slice_expired(cfqd, timed_out);
4536 out_kick:
4537         cfq_schedule_dispatch(cfqd);
4538 out_cont:
4539         spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
4540         return HRTIMER_NORESTART;
4541 }
4542
4543 static void cfq_shutdown_timer_wq(struct cfq_data *cfqd)
4544 {
4545         hrtimer_cancel(&cfqd->idle_slice_timer);
4546         cancel_work_sync(&cfqd->unplug_work);
4547 }
4548
4549 static void cfq_exit_queue(struct elevator_queue *e)
4550 {
4551         struct cfq_data *cfqd = e->elevator_data;
4552         struct request_queue *q = cfqd->queue;
4553
4554         cfq_shutdown_timer_wq(cfqd);
4555
4556         spin_lock_irq(q->queue_lock);
4557
4558         if (cfqd->active_queue)
4559                 __cfq_slice_expired(cfqd, cfqd->active_queue, 0);
4560
4561         spin_unlock_irq(q->queue_lock);
4562
4563         cfq_shutdown_timer_wq(cfqd);
4564
4565 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4566         blkcg_deactivate_policy(q, &blkcg_policy_cfq);
4567 #else
4568         kfree(cfqd->root_group);
4569 #endif
4570         kfree(cfqd);
4571 }
4572
4573 static int cfq_init_queue(struct request_queue *q, struct elevator_type *e)
4574 {
4575         struct cfq_data *cfqd;
4576         struct blkcg_gq *blkg __maybe_unused;
4577         int i, ret;
4578         struct elevator_queue *eq;
4579
4580         eq = elevator_alloc(q, e);
4581         if (!eq)
4582                 return -ENOMEM;
4583
4584         cfqd = kzalloc_node(sizeof(*cfqd), GFP_KERNEL, q->node);
4585         if (!cfqd) {
4586                 kobject_put(&eq->kobj);
4587                 return -ENOMEM;
4588         }
4589         eq->elevator_data = cfqd;
4590
4591         cfqd->queue = q;
4592         spin_lock_irq(q->queue_lock);
4593         q->elevator = eq;
4594         spin_unlock_irq(q->queue_lock);
4595
4596         /* Init root service tree */
4597         cfqd->grp_service_tree = CFQ_RB_ROOT;
4598
4599         /* Init root group and prefer root group over other groups by default */
4600 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4601         ret = blkcg_activate_policy(q, &blkcg_policy_cfq);
4602         if (ret)
4603                 goto out_free;
4604
4605         cfqd->root_group = blkg_to_cfqg(q->root_blkg);
4606 #else
4607         ret = -ENOMEM;
4608         cfqd->root_group = kzalloc_node(sizeof(*cfqd->root_group),
4609                                         GFP_KERNEL, cfqd->queue->node);
4610         if (!cfqd->root_group)
4611                 goto out_free;
4612
4613         cfq_init_cfqg_base(cfqd->root_group);
4614         cfqd->root_group->weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
4615         cfqd->root_group->leaf_weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
4616 #endif
4617
4618         /*
4619          * Not strictly needed (since RB_ROOT just clears the node and we
4620          * zeroed cfqd on alloc), but better be safe in case someone decides
4621          * to add magic to the rb code
4622          */
4623         for (i = 0; i < CFQ_PRIO_LISTS; i++)
4624                 cfqd->prio_trees[i] = RB_ROOT;
4625
4626         /*
4627          * Our fallback cfqq if cfq_get_queue() runs into OOM issues.
4628          * Grab a permanent reference to it, so that the normal code flow
4629          * will not attempt to free it.  oom_cfqq is linked to root_group
4630          * but shouldn't hold a reference as it'll never be unlinked.  Lose
4631          * the reference from linking right away.
4632          */
4633         cfq_init_cfqq(cfqd, &cfqd->oom_cfqq, 1, 0);
4634         cfqd->oom_cfqq.ref++;
4635
4636         spin_lock_irq(q->queue_lock);
4637         cfq_link_cfqq_cfqg(&cfqd->oom_cfqq, cfqd->root_group);
4638         cfqg_put(cfqd->root_group);
4639         spin_unlock_irq(q->queue_lock);
4640
4641         hrtimer_init(&cfqd->idle_slice_timer, CLOCK_MONOTONIC,
4642                      HRTIMER_MODE_REL);
4643         cfqd->idle_slice_timer.function = cfq_idle_slice_timer;
4644
4645         INIT_WORK(&cfqd->unplug_work, cfq_kick_queue);
4646
4647         cfqd->cfq_quantum = cfq_quantum;
4648         cfqd->cfq_fifo_expire[0] = cfq_fifo_expire[0];
4649         cfqd->cfq_fifo_expire[1] = cfq_fifo_expire[1];
4650         cfqd->cfq_back_max = cfq_back_max;
4651         cfqd->cfq_back_penalty = cfq_back_penalty;
4652         cfqd->cfq_slice[0] = cfq_slice_async;
4653         cfqd->cfq_slice[1] = cfq_slice_sync;
4654         cfqd->cfq_target_latency = cfq_target_latency;
4655         cfqd->cfq_slice_async_rq = cfq_slice_async_rq;
4656         cfqd->cfq_slice_idle = cfq_slice_idle;
4657         cfqd->cfq_group_idle = cfq_group_idle;
4658         cfqd->cfq_latency = 1;
4659         cfqd->hw_tag = -1;
4660         /*
4661          * we optimistically start assuming sync ops weren't delayed in last
4662          * second, in order to have larger depth for async operations.
4663          */
4664         cfqd->last_delayed_sync = ktime_get_ns() - NSEC_PER_SEC;
4665         return 0;
4666
4667 out_free:
4668         kfree(cfqd);
4669         kobject_put(&eq->kobj);
4670         return ret;
4671 }
4672
4673 static void cfq_registered_queue(struct request_queue *q)
4674 {
4675         struct elevator_queue *e = q->elevator;
4676         struct cfq_data *cfqd = e->elevator_data;
4677
4678         /*
4679          * Default to IOPS mode with no idling for SSDs
4680          */
4681         if (blk_queue_nonrot(q))
4682                 cfqd->cfq_slice_idle = 0;
4683         wbt_disable_default(q);
4684 }
4685
4686 /*
4687  * sysfs parts below -->
4688  */
4689 static ssize_t
4690 cfq_var_show(unsigned int var, char *page)
4691 {
4692         return sprintf(page, "%u\n", var);
4693 }
4694
4695 static void
4696 cfq_var_store(unsigned int *var, const char *page)
4697 {
4698         char *p = (char *) page;
4699
4700         *var = simple_strtoul(p, &p, 10);
4701 }
4702
4703 #define SHOW_FUNCTION(__FUNC, __VAR, __CONV)                            \
4704 static ssize_t __FUNC(struct elevator_queue *e, char *page)             \
4705 {                                                                       \
4706         struct cfq_data *cfqd = e->elevator_data;                       \
4707         u64 __data = __VAR;                                             \
4708         if (__CONV)                                                     \
4709                 __data = div_u64(__data, NSEC_PER_MSEC);                        \
4710         return cfq_var_show(__data, (page));                            \
4711 }
4712 SHOW_FUNCTION(cfq_quantum_show, cfqd->cfq_quantum, 0);
4713 SHOW_FUNCTION(cfq_fifo_expire_sync_show, cfqd->cfq_fifo_expire[1], 1);
4714 SHOW_FUNCTION(cfq_fifo_expire_async_show, cfqd->cfq_fifo_expire[0], 1);
4715 SHOW_FUNCTION(cfq_back_seek_max_show, cfqd->cfq_back_max, 0);
4716 SHOW_FUNCTION(cfq_back_seek_penalty_show, cfqd->cfq_back_penalty, 0);
4717 SHOW_FUNCTION(cfq_slice_idle_show, cfqd->cfq_slice_idle, 1);
4718 SHOW_FUNCTION(cfq_group_idle_show, cfqd->cfq_group_idle, 1);
4719 SHOW_FUNCTION(cfq_slice_sync_show, cfqd->cfq_slice[1], 1);
4720 SHOW_FUNCTION(cfq_slice_async_show, cfqd->cfq_slice[0], 1);
4721 SHOW_FUNCTION(cfq_slice_async_rq_show, cfqd->cfq_slice_async_rq, 0);
4722 SHOW_FUNCTION(cfq_low_latency_show, cfqd->cfq_latency, 0);
4723 SHOW_FUNCTION(cfq_target_latency_show, cfqd->cfq_target_latency, 1);
4724 #undef SHOW_FUNCTION
4725
4726 #define USEC_SHOW_FUNCTION(__FUNC, __VAR)                               \
4727 static ssize_t __FUNC(struct elevator_queue *e, char *page)             \
4728 {                                                                       \
4729         struct cfq_data *cfqd = e->elevator_data;                       \
4730         u64 __data = __VAR;                                             \
4731         __data = div_u64(__data, NSEC_PER_USEC);                        \
4732         return cfq_var_show(__data, (page));                            \
4733 }
4734 USEC_SHOW_FUNCTION(cfq_slice_idle_us_show, cfqd->cfq_slice_idle);
4735 USEC_SHOW_FUNCTION(cfq_group_idle_us_show, cfqd->cfq_group_idle);
4736 USEC_SHOW_FUNCTION(cfq_slice_sync_us_show, cfqd->cfq_slice[1]);
4737 USEC_SHOW_FUNCTION(cfq_slice_async_us_show, cfqd->cfq_slice[0]);
4738 USEC_SHOW_FUNCTION(cfq_target_latency_us_show, cfqd->cfq_target_latency);
4739 #undef USEC_SHOW_FUNCTION
4740
4741 #define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX, __CONV)                 \
4742 static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count) \
4743 {                                                                       \
4744         struct cfq_data *cfqd = e->elevator_data;                       \
4745         unsigned int __data, __min = (MIN), __max = (MAX);              \
4746                                                                         \
4747         cfq_var_store(&__data, (page));                                 \
4748         if (__data < __min)                                             \
4749                 __data = __min;                                         \
4750         else if (__data > __max)                                        \
4751                 __data = __max;                                         \
4752         if (__CONV)                                                     \
4753                 *(__PTR) = (u64)__data * NSEC_PER_MSEC;                 \
4754         else                                                            \
4755                 *(__PTR) = __data;                                      \
4756         return count;                                                   \
4757 }
4758 STORE_FUNCTION(cfq_quantum_store, &cfqd->cfq_quantum, 1, UINT_MAX, 0);
4759 STORE_FUNCTION(cfq_fifo_expire_sync_store, &cfqd->cfq_fifo_expire[1], 1,
4760                 UINT_MAX, 1);
4761 STORE_FUNCTION(cfq_fifo_expire_async_store, &cfqd->cfq_fifo_expire[0], 1,
4762                 UINT_MAX, 1);
4763 STORE_FUNCTION(cfq_back_seek_max_store, &cfqd->cfq_back_max, 0, UINT_MAX, 0);
4764 STORE_FUNCTION(cfq_back_seek_penalty_store, &cfqd->cfq_back_penalty, 1,
4765                 UINT_MAX, 0);
4766 STORE_FUNCTION(cfq_slice_idle_store, &cfqd->cfq_slice_idle, 0, UINT_MAX, 1);
4767 STORE_FUNCTION(cfq_group_idle_store, &cfqd->cfq_group_idle, 0, UINT_MAX, 1);
4768 STORE_FUNCTION(cfq_slice_sync_store, &cfqd->cfq_slice[1], 1, UINT_MAX, 1);
4769 STORE_FUNCTION(cfq_slice_async_store, &cfqd->cfq_slice[0], 1, UINT_MAX, 1);
4770 STORE_FUNCTION(cfq_slice_async_rq_store, &cfqd->cfq_slice_async_rq, 1,
4771                 UINT_MAX, 0);
4772 STORE_FUNCTION(cfq_low_latency_store, &cfqd->cfq_latency, 0, 1, 0);
4773 STORE_FUNCTION(cfq_target_latency_store, &cfqd->cfq_target_latency, 1, UINT_MAX, 1);
4774 #undef STORE_FUNCTION
4775
4776 #define USEC_STORE_FUNCTION(__FUNC, __PTR, MIN, MAX)                    \
4777 static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count) \
4778 {                                                                       \
4779         struct cfq_data *cfqd = e->elevator_data;                       \
4780         unsigned int __data, __min = (MIN), __max = (MAX);              \
4781                                                                         \
4782         cfq_var_store(&__data, (page));                                 \
4783         if (__data < __min)                                             \
4784                 __data = __min;                                         \
4785         else if (__data > __max)                                        \
4786                 __data = __max;                                         \
4787         *(__PTR) = (u64)__data * NSEC_PER_USEC;                         \
4788         return count;                                                   \
4789 }
4790 USEC_STORE_FUNCTION(cfq_slice_idle_us_store, &cfqd->cfq_slice_idle, 0, UINT_MAX);
4791 USEC_STORE_FUNCTION(cfq_group_idle_us_store, &cfqd->cfq_group_idle, 0, UINT_MAX);
4792 USEC_STORE_FUNCTION(cfq_slice_sync_us_store, &cfqd->cfq_slice[1], 1, UINT_MAX);
4793 USEC_STORE_FUNCTION(cfq_slice_async_us_store, &cfqd->cfq_slice[0], 1, UINT_MAX);
4794 USEC_STORE_FUNCTION(cfq_target_latency_us_store, &cfqd->cfq_target_latency, 1, UINT_MAX);
4795 #undef USEC_STORE_FUNCTION
4796
4797 #define CFQ_ATTR(name) \
4798         __ATTR(name, 0644, cfq_##name##_show, cfq_##name##_store)
4799
4800 static struct elv_fs_entry cfq_attrs[] = {
4801         CFQ_ATTR(quantum),
4802         CFQ_ATTR(fifo_expire_sync),
4803         CFQ_ATTR(fifo_expire_async),
4804         CFQ_ATTR(back_seek_max),
4805         CFQ_ATTR(back_seek_penalty),
4806         CFQ_ATTR(slice_sync),
4807         CFQ_ATTR(slice_sync_us),
4808         CFQ_ATTR(slice_async),
4809         CFQ_ATTR(slice_async_us),
4810         CFQ_ATTR(slice_async_rq),
4811         CFQ_ATTR(slice_idle),
4812         CFQ_ATTR(slice_idle_us),
4813         CFQ_ATTR(group_idle),
4814         CFQ_ATTR(group_idle_us),
4815         CFQ_ATTR(low_latency),
4816         CFQ_ATTR(target_latency),
4817         CFQ_ATTR(target_latency_us),
4818         __ATTR_NULL
4819 };
4820
4821 static struct elevator_type iosched_cfq = {
4822         .ops.sq = {
4823                 .elevator_merge_fn =            cfq_merge,
4824                 .elevator_merged_fn =           cfq_merged_request,
4825                 .elevator_merge_req_fn =        cfq_merged_requests,
4826                 .elevator_allow_bio_merge_fn =  cfq_allow_bio_merge,
4827                 .elevator_allow_rq_merge_fn =   cfq_allow_rq_merge,
4828                 .elevator_bio_merged_fn =       cfq_bio_merged,
4829                 .elevator_dispatch_fn =         cfq_dispatch_requests,
4830                 .elevator_add_req_fn =          cfq_insert_request,
4831                 .elevator_activate_req_fn =     cfq_activate_request,
4832                 .elevator_deactivate_req_fn =   cfq_deactivate_request,
4833                 .elevator_completed_req_fn =    cfq_completed_request,
4834                 .elevator_former_req_fn =       elv_rb_former_request,
4835                 .elevator_latter_req_fn =       elv_rb_latter_request,
4836                 .elevator_init_icq_fn =         cfq_init_icq,
4837                 .elevator_exit_icq_fn =         cfq_exit_icq,
4838                 .elevator_set_req_fn =          cfq_set_request,
4839                 .elevator_put_req_fn =          cfq_put_request,
4840                 .elevator_may_queue_fn =        cfq_may_queue,
4841                 .elevator_init_fn =             cfq_init_queue,
4842                 .elevator_exit_fn =             cfq_exit_queue,
4843                 .elevator_registered_fn =       cfq_registered_queue,
4844         },
4845         .icq_size       =       sizeof(struct cfq_io_cq),
4846         .icq_align      =       __alignof__(struct cfq_io_cq),
4847         .elevator_attrs =       cfq_attrs,
4848         .elevator_name  =       "cfq",
4849         .elevator_owner =       THIS_MODULE,
4850 };
4851
4852 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4853 static struct blkcg_policy blkcg_policy_cfq = {
4854         .dfl_cftypes            = cfq_blkcg_files,
4855         .legacy_cftypes         = cfq_blkcg_legacy_files,
4856
4857         .cpd_alloc_fn           = cfq_cpd_alloc,
4858         .cpd_init_fn            = cfq_cpd_init,
4859         .cpd_free_fn            = cfq_cpd_free,
4860         .cpd_bind_fn            = cfq_cpd_bind,
4861
4862         .pd_alloc_fn            = cfq_pd_alloc,
4863         .pd_init_fn             = cfq_pd_init,
4864         .pd_offline_fn          = cfq_pd_offline,
4865         .pd_free_fn             = cfq_pd_free,
4866         .pd_reset_stats_fn      = cfq_pd_reset_stats,
4867 };
4868 #endif
4869
4870 static int __init cfq_init(void)
4871 {
4872         int ret;
4873
4874 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4875         ret = blkcg_policy_register(&blkcg_policy_cfq);
4876         if (ret)
4877                 return ret;
4878 #else
4879         cfq_group_idle = 0;
4880 #endif
4881
4882         ret = -ENOMEM;
4883         cfq_pool = KMEM_CACHE(cfq_queue, 0);
4884         if (!cfq_pool)
4885                 goto err_pol_unreg;
4886
4887         ret = elv_register(&iosched_cfq);
4888         if (ret)
4889                 goto err_free_pool;
4890
4891         return 0;
4892
4893 err_free_pool:
4894         kmem_cache_destroy(cfq_pool);
4895 err_pol_unreg:
4896 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4897         blkcg_policy_unregister(&blkcg_policy_cfq);
4898 #endif
4899         return ret;
4900 }
4901
4902 static void __exit cfq_exit(void)
4903 {
4904 #ifdef CONFIG_CFQ_GROUP_IOSCHED
4905         blkcg_policy_unregister(&blkcg_policy_cfq);
4906 #endif
4907         elv_unregister(&iosched_cfq);
4908         kmem_cache_destroy(cfq_pool);
4909 }
4910
4911 module_init(cfq_init);
4912 module_exit(cfq_exit);
4913
4914 MODULE_AUTHOR("Jens Axboe");
4915 MODULE_LICENSE("GPL");
4916 MODULE_DESCRIPTION("Completely Fair Queueing IO scheduler");