net/sched/sch_fq.c

   1 /*
   2  * net/sched/sch_fq.c Fair Queue Packet Scheduler (per flow pacing)
   3  *
   4  *  Copyright (C) 2013-2015 Eric Dumazet <edumazet@google.com>
   5  *
   6  *      This program is free software; you can redistribute it and/or
   7  *      modify it under the terms of the GNU General Public License
   8  *      as published by the Free Software Foundation; either version
   9  *      2 of the License, or (at your option) any later version.
  10  *
  11  *  Meant to be mostly used for locally generated traffic :
  12  *  Fast classification depends on skb->sk being set before reaching us.
  13  *  If not, (router workload), we use rxhash as fallback, with 32 bits wide hash.
  14  *  All packets belonging to a socket are considered as a 'flow'.
  15  *
  16  *  Flows are dynamically allocated and stored in a hash table of RB trees
  17  *  They are also part of one Round Robin 'queues' (new or old flows)
  18  *
  19  *  Burst avoidance (aka pacing) capability :
  20  *
  21  *  Transport (eg TCP) can set in sk->sk_pacing_rate a rate, enqueue a
  22  *  bunch of packets, and this packet scheduler adds delay between
  23  *  packets to respect rate limitation.
  24  *
  25  *  enqueue() :
  26  *   - lookup one RB tree (out of 1024 or more) to find the flow.
  27  *     If non existent flow, create it, add it to the tree.
  28  *     Add skb to the per flow list of skb (fifo).
  29  *   - Use a special fifo for high prio packets
  30  *
  31  *  dequeue() : serves flows in Round Robin
  32  *  Note : When a flow becomes empty, we do not immediately remove it from
  33  *  rb trees, for performance reasons (its expected to send additional packets,
  34  *  or SLAB cache will reuse socket for another flow)
  35  */
  36
  37 #include <linux/module.h>
  38 #include <linux/types.h>
  39 #include <linux/kernel.h>
  40 #include <linux/jiffies.h>
  41 #include <linux/string.h>
  42 #include <linux/in.h>
  43 #include <linux/errno.h>
  44 #include <linux/init.h>
  45 #include <linux/skbuff.h>
  46 #include <linux/slab.h>
  47 #include <linux/rbtree.h>
  48 #include <linux/hash.h>
  49 #include <linux/prefetch.h>
  50 #include <linux/vmalloc.h>
  51 #include <net/netlink.h>
  52 #include <net/pkt_sched.h>
  53 #include <net/sock.h>
  54 #include <net/tcp_states.h>
  55 #include <net/tcp.h>
  56
  57 /*
  58  * Per flow structure, dynamically allocated
  59  */
  60 struct fq_flow {
  61         struct sk_buff  *head;          /* list of skbs for this flow : first skb */
  62         union {
  63                 struct sk_buff *tail;   /* last skb in the list */
  64                 unsigned long  age;     /* jiffies when flow was emptied, for gc */
  65         };
  66         struct rb_node  fq_node;        /* anchor in fq_root[] trees */
  67         struct sock     *sk;
  68         int             qlen;           /* number of packets in flow queue */
  69         int             credit;
  70         u32             socket_hash;    /* sk_hash */
  71         struct fq_flow *next;           /* next pointer in RR lists, or &detached */
  72
  73         struct rb_node  rate_node;      /* anchor in q->delayed tree */
  74         u64             time_next_packet;
  75 };
  76
  77 struct fq_flow_head {
  78         struct fq_flow *first;
  79         struct fq_flow *last;
  80 };
  81
  82 struct fq_sched_data {
  83         struct fq_flow_head new_flows;
  84
  85         struct fq_flow_head old_flows;
  86
  87         struct rb_root  delayed;        /* for rate limited flows */
  88         u64             time_next_delayed_flow;
  89
  90         struct fq_flow  internal;       /* for non classified or high prio packets */
  91         u32             quantum;
  92         u32             initial_quantum;
  93         u32             flow_refill_delay;
  94         u32             flow_max_rate;  /* optional max rate per flow */
  95         u32             flow_plimit;    /* max packets per flow */
  96         u32             orphan_mask;    /* mask for orphaned skb */
  97         struct rb_root  *fq_root;
  98         u8              rate_enable;
  99         u8              fq_trees_log;
 100
 101         u32             flows;
 102         u32             inactive_flows;
 103         u32             throttled_flows;
 104
 105         u64             stat_gc_flows;
 106         u64             stat_internal_packets;
 107         u64             stat_tcp_retrans;
 108         u64             stat_throttled;
 109         u64             stat_flows_plimit;
 110         u64             stat_pkts_too_long;
 111         u64             stat_allocation_errors;
 112         struct qdisc_watchdog watchdog;
 113 };
 114
 115 /* special value to mark a detached flow (not on old/new list) */
 116 static struct fq_flow detached, throttled;
 117
 118 static void fq_flow_set_detached(struct fq_flow *f)
 119 {
 120         f->next = &detached;
 121         f->age = jiffies;
 122 }
 123
 124 static bool fq_flow_is_detached(const struct fq_flow *f)
 125 {
 126         return f->next == &detached;
 127 }
 128
 129 static void fq_flow_set_throttled(struct fq_sched_data *q, struct fq_flow *f)
 130 {
 131         struct rb_node **p = &q->delayed.rb_node, *parent = NULL;
 132
 133         while (*p) {
 134                 struct fq_flow *aux;
 135
 136                 parent = *p;
 137                 aux = container_of(parent, struct fq_flow, rate_node);
 138                 if (f->time_next_packet >= aux->time_next_packet)
 139                         p = &parent->rb_right;
 140                 else
 141                         p = &parent->rb_left;
 142         }
 143         rb_link_node(&f->rate_node, parent, p);
 144         rb_insert_color(&f->rate_node, &q->delayed);
 145         q->throttled_flows++;
 146         q->stat_throttled++;
 147
 148         f->next = &throttled;
 149         if (q->time_next_delayed_flow > f->time_next_packet)
 150                 q->time_next_delayed_flow = f->time_next_packet;
 151 }
 152
 153
 154 static struct kmem_cache *fq_flow_cachep __read_mostly;
 155
 156 static void fq_flow_add_tail(struct fq_flow_head *head, struct fq_flow *flow)
 157 {
 158         if (head->first)
 159                 head->last->next = flow;
 160         else
 161                 head->first = flow;
 162         head->last = flow;
 163         flow->next = NULL;
 164 }
 165
 166 /* limit number of collected flows per round */
 167 #define FQ_GC_MAX 8
 168 #define FQ_GC_AGE (3*HZ)
 169
 170 static bool fq_gc_candidate(const struct fq_flow *f)
 171 {
 172         return fq_flow_is_detached(f) &&
 173                time_after(jiffies, f->age + FQ_GC_AGE);
 174 }
 175
 176 static void fq_gc(struct fq_sched_data *q,
 177                   struct rb_root *root,
 178                   struct sock *sk)
 179 {
 180         struct fq_flow *f, *tofree[FQ_GC_MAX];
 181         struct rb_node **p, *parent;
 182         int fcnt = 0;
 183
 184         p = &root->rb_node;
 185         parent = NULL;
 186         while (*p) {
 187                 parent = *p;
 188
 189                 f = container_of(parent, struct fq_flow, fq_node);
 190                 if (f->sk == sk)
 191                         break;
 192
 193                 if (fq_gc_candidate(f)) {
 194                         tofree[fcnt++] = f;
 195                         if (fcnt == FQ_GC_MAX)
 196                                 break;
 197                 }
 198
 199                 if (f->sk > sk)
 200                         p = &parent->rb_right;
 201                 else
 202                         p = &parent->rb_left;
 203         }
 204
 205         q->flows -= fcnt;
 206         q->inactive_flows -= fcnt;
 207         q->stat_gc_flows += fcnt;
 208         while (fcnt) {
 209                 struct fq_flow *f = tofree[--fcnt];
 210
 211                 rb_erase(&f->fq_node, root);
 212                 kmem_cache_free(fq_flow_cachep, f);
 213         }
 214 }
 215
 216 static struct fq_flow *fq_classify(struct sk_buff *skb, struct fq_sched_data *q)
 217 {
 218         struct rb_node **p, *parent;
 219         struct sock *sk = skb->sk;
 220         struct rb_root *root;
 221         struct fq_flow *f;
 222
 223         /* warning: no starvation prevention... */
 224         if (unlikely((skb->priority & TC_PRIO_MAX) == TC_PRIO_CONTROL))
 225                 return &q->internal;
 226
 227         /* SYNACK messages are attached to a TCP_NEW_SYN_RECV request socket
 228          * or a listener (SYNCOOKIE mode)
 229          * 1) request sockets are not full blown,
 230          *    they do not contain sk_pacing_rate
 231          * 2) They are not part of a 'flow' yet
 232          * 3) We do not want to rate limit them (eg SYNFLOOD attack),
 233          *    especially if the listener set SO_MAX_PACING_RATE
 234          * 4) We pretend they are orphaned
 235          */
 236         if (!sk || sk_listener(sk)) {
 237                 unsigned long hash = skb_get_hash(skb) & q->orphan_mask;
 238
 239                 /* By forcing low order bit to 1, we make sure to not
 240                  * collide with a local flow (socket pointers are word aligned)
 241                  */
 242                 sk = (struct sock *)((hash << 1) | 1UL);
 243                 skb_orphan(skb);
 244         }
 245
 246         root = &q->fq_root[hash_32((u32)(long)sk, q->fq_trees_log)];
 247
 248         if (q->flows >= (2U << q->fq_trees_log) &&
 249             q->inactive_flows > q->flows/2)
 250                 fq_gc(q, root, sk);
 251
 252         p = &root->rb_node;
 253         parent = NULL;
 254         while (*p) {
 255                 parent = *p;
 256
 257                 f = container_of(parent, struct fq_flow, fq_node);
 258                 if (f->sk == sk) {
 259                         /* socket might have been reallocated, so check
 260                          * if its sk_hash is the same.
 261                          * It not, we need to refill credit with
 262                          * initial quantum
 263                          */
 264                         if (unlikely(skb->sk &&
 265                                      f->socket_hash != sk->sk_hash)) {
 266                                 f->credit = q->initial_quantum;
 267                                 f->socket_hash = sk->sk_hash;
 268                                 f->time_next_packet = 0ULL;
 269                         }
 270                         return f;
 271                 }
 272                 if (f->sk > sk)
 273                         p = &parent->rb_right;
 274                 else
 275                         p = &parent->rb_left;
 276         }
 277
 278         f = kmem_cache_zalloc(fq_flow_cachep, GFP_ATOMIC | __GFP_NOWARN);
 279         if (unlikely(!f)) {
 280                 q->stat_allocation_errors++;
 281                 return &q->internal;
 282         }
 283         fq_flow_set_detached(f);
 284         f->sk = sk;
 285         if (skb->sk)
 286                 f->socket_hash = sk->sk_hash;
 287         f->credit = q->initial_quantum;
 288
 289         rb_link_node(&f->fq_node, parent, p);
 290         rb_insert_color(&f->fq_node, root);
 291
 292         q->flows++;
 293         q->inactive_flows++;
 294         return f;
 295 }
 296
 297
 298 /* remove one skb from head of flow queue */
 299 static struct sk_buff *fq_dequeue_head(struct Qdisc *sch, struct fq_flow *flow)
 300 {
 301         struct sk_buff *skb = flow->head;
 302
 303         if (skb) {
 304                 flow->head = skb->next;
 305                 skb->next = NULL;
 306                 flow->qlen--;
 307                 qdisc_qstats_backlog_dec(sch, skb);
 308                 sch->q.qlen--;
 309         }
 310         return skb;
 311 }
 312
 313 /* We might add in the future detection of retransmits
 314  * For the time being, just return false
 315  */
 316 static bool skb_is_retransmit(struct sk_buff *skb)
 317 {
 318         return false;
 319 }
 320
 321 /* add skb to flow queue
 322  * flow queue is a linked list, kind of FIFO, except for TCP retransmits
 323  * We special case tcp retransmits to be transmitted before other packets.
 324  * We rely on fact that TCP retransmits are unlikely, so we do not waste
 325  * a separate queue or a pointer.
 326  * head->  [retrans pkt 1]
 327  *         [retrans pkt 2]
 328  *         [ normal pkt 1]
 329  *         [ normal pkt 2]
 330  *         [ normal pkt 3]
 331  * tail->  [ normal pkt 4]
 332  */
 333 static void flow_queue_add(struct fq_flow *flow, struct sk_buff *skb)
 334 {
 335         struct sk_buff *prev, *head = flow->head;
 336
 337         skb->next = NULL;
 338         if (!head) {
 339                 flow->head = skb;
 340                 flow->tail = skb;
 341                 return;
 342         }
 343         if (likely(!skb_is_retransmit(skb))) {
 344                 flow->tail->next = skb;
 345                 flow->tail = skb;
 346                 return;
 347         }
 348
 349         /* This skb is a tcp retransmit,
 350          * find the last retrans packet in the queue
 351          */
 352         prev = NULL;
 353         while (skb_is_retransmit(head)) {
 354                 prev = head;
 355                 head = head->next;
 356                 if (!head)
 357                         break;
 358         }
 359         if (!prev) { /* no rtx packet in queue, become the new head */
 360                 skb->next = flow->head;
 361                 flow->head = skb;
 362         } else {
 363                 if (prev == flow->tail)
 364                         flow->tail = skb;
 365                 else
 366                         skb->next = prev->next;
 367                 prev->next = skb;
 368         }
 369 }
 370
 371 static int fq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 372                       struct sk_buff **to_free)
 373 {
 374         struct fq_sched_data *q = qdisc_priv(sch);
 375         struct fq_flow *f;
 376
 377         if (unlikely(sch->q.qlen >= sch->limit))
 378                 return qdisc_drop(skb, sch, to_free);
 379
 380         f = fq_classify(skb, q);
 381         if (unlikely(f->qlen >= q->flow_plimit && f != &q->internal)) {
 382                 q->stat_flows_plimit++;
 383                 return qdisc_drop(skb, sch, to_free);
 384         }
 385
 386         f->qlen++;
 387         if (skb_is_retransmit(skb))
 388                 q->stat_tcp_retrans++;
 389         qdisc_qstats_backlog_inc(sch, skb);
 390         if (fq_flow_is_detached(f)) {
 391                 fq_flow_add_tail(&q->new_flows, f);
 392                 if (time_after(jiffies, f->age + q->flow_refill_delay))
 393                         f->credit = max_t(u32, f->credit, q->quantum);
 394                 q->inactive_flows--;
 395         }
 396
 397         /* Note: this overwrites f->age */
 398         flow_queue_add(f, skb);
 399
 400         if (unlikely(f == &q->internal)) {
 401                 q->stat_internal_packets++;
 402         }
 403         sch->q.qlen++;
 404
 405         return NET_XMIT_SUCCESS;
 406 }
 407
 408 static void fq_check_throttled(struct fq_sched_data *q, u64 now)
 409 {
 410         struct rb_node *p;
 411
 412         if (q->time_next_delayed_flow > now)
 413                 return;
 414
 415         q->time_next_delayed_flow = ~0ULL;
 416         while ((p = rb_first(&q->delayed)) != NULL) {
 417                 struct fq_flow *f = container_of(p, struct fq_flow, rate_node);
 418
 419                 if (f->time_next_packet > now) {
 420                         q->time_next_delayed_flow = f->time_next_packet;
 421                         break;
 422                 }
 423                 rb_erase(p, &q->delayed);
 424                 q->throttled_flows--;
 425                 fq_flow_add_tail(&q->old_flows, f);
 426         }
 427 }
 428
 429 static struct sk_buff *fq_dequeue(struct Qdisc *sch)
 430 {
 431         struct fq_sched_data *q = qdisc_priv(sch);
 432         u64 now = ktime_get_ns();
 433         struct fq_flow_head *head;
 434         struct sk_buff *skb;
 435         struct fq_flow *f;
 436         u32 rate;
 437
 438         skb = fq_dequeue_head(sch, &q->internal);
 439         if (skb)
 440                 goto out;
 441         fq_check_throttled(q, now);
 442 begin:
 443         head = &q->new_flows;
 444         if (!head->first) {
 445                 head = &q->old_flows;
 446                 if (!head->first) {
 447                         if (q->time_next_delayed_flow != ~0ULL)
 448                                 qdisc_watchdog_schedule_ns(&q->watchdog,
 449                                                            q->time_next_delayed_flow);
 450                         return NULL;
 451                 }
 452         }
 453         f = head->first;
 454
 455         if (f->credit <= 0) {
 456                 f->credit += q->quantum;
 457                 head->first = f->next;
 458                 fq_flow_add_tail(&q->old_flows, f);
 459                 goto begin;
 460         }
 461
 462         skb = f->head;
 463         if (unlikely(skb && now < f->time_next_packet &&
 464                      !skb_is_tcp_pure_ack(skb))) {
 465                 head->first = f->next;
 466                 fq_flow_set_throttled(q, f);
 467                 goto begin;
 468         }
 469
 470         skb = fq_dequeue_head(sch, f);
 471         if (!skb) {
 472                 head->first = f->next;
 473                 /* force a pass through old_flows to prevent starvation */
 474                 if ((head == &q->new_flows) && q->old_flows.first) {
 475                         fq_flow_add_tail(&q->old_flows, f);
 476                 } else {
 477                         fq_flow_set_detached(f);
 478                         q->inactive_flows++;
 479                 }
 480                 goto begin;
 481         }
 482         prefetch(&skb->end);
 483         f->credit -= qdisc_pkt_len(skb);
 484
 485         if (f->credit > 0 || !q->rate_enable)
 486                 goto out;
 487
 488         /* Do not pace locally generated ack packets */
 489         if (skb_is_tcp_pure_ack(skb))
 490                 goto out;
 491
 492         rate = q->flow_max_rate;
 493         if (skb->sk)
 494                 rate = min(skb->sk->sk_pacing_rate, rate);
 495
 496         if (rate != ~0U) {
 497                 u32 plen = max(qdisc_pkt_len(skb), q->quantum);
 498                 u64 len = (u64)plen * NSEC_PER_SEC;
 499
 500                 if (likely(rate))
 501                         do_div(len, rate);
 502                 /* Since socket rate can change later,
 503                  * clamp the delay to 1 second.
 504                  * Really, providers of too big packets should be fixed !
 505                  */
 506                 if (unlikely(len > NSEC_PER_SEC)) {
 507                         len = NSEC_PER_SEC;
 508                         q->stat_pkts_too_long++;
 509                 }
 510
 511                 f->time_next_packet = now + len;
 512         }
 513 out:
 514         qdisc_bstats_update(sch, skb);
 515         return skb;
 516 }
 517
 518 static void fq_flow_purge(struct fq_flow *flow)
 519 {
 520         rtnl_kfree_skbs(flow->head, flow->tail);
 521         flow->head = NULL;
 522         flow->qlen = 0;
 523 }
 524
 525 static void fq_reset(struct Qdisc *sch)
 526 {
 527         struct fq_sched_data *q = qdisc_priv(sch);
 528         struct rb_root *root;
 529         struct rb_node *p;
 530         struct fq_flow *f;
 531         unsigned int idx;
 532
 533         sch->q.qlen = 0;
 534         sch->qstats.backlog = 0;
 535
 536         fq_flow_purge(&q->internal);
 537
 538         if (!q->fq_root)
 539                 return;
 540
 541         for (idx = 0; idx < (1U << q->fq_trees_log); idx++) {
 542                 root = &q->fq_root[idx];
 543                 while ((p = rb_first(root)) != NULL) {
 544                         f = container_of(p, struct fq_flow, fq_node);
 545                         rb_erase(p, root);
 546
 547                         fq_flow_purge(f);
 548
 549                         kmem_cache_free(fq_flow_cachep, f);
 550                 }
 551         }
 552         q->new_flows.first      = NULL;
 553         q->old_flows.first      = NULL;
 554         q->delayed              = RB_ROOT;
 555         q->flows                = 0;
 556         q->inactive_flows       = 0;
 557         q->throttled_flows      = 0;
 558 }
 559
 560 static void fq_rehash(struct fq_sched_data *q,
 561                       struct rb_root *old_array, u32 old_log,
 562                       struct rb_root *new_array, u32 new_log)
 563 {
 564         struct rb_node *op, **np, *parent;
 565         struct rb_root *oroot, *nroot;
 566         struct fq_flow *of, *nf;
 567         int fcnt = 0;
 568         u32 idx;
 569
 570         for (idx = 0; idx < (1U << old_log); idx++) {
 571                 oroot = &old_array[idx];
 572                 while ((op = rb_first(oroot)) != NULL) {
 573                         rb_erase(op, oroot);
 574                         of = container_of(op, struct fq_flow, fq_node);
 575                         if (fq_gc_candidate(of)) {
 576                                 fcnt++;
 577                                 kmem_cache_free(fq_flow_cachep, of);
 578                                 continue;
 579                         }
 580                         nroot = &new_array[hash_32((u32)(long)of->sk, new_log)];
 581
 582                         np = &nroot->rb_node;
 583                         parent = NULL;
 584                         while (*np) {
 585                                 parent = *np;
 586
 587                                 nf = container_of(parent, struct fq_flow, fq_node);
 588                                 BUG_ON(nf->sk == of->sk);
 589
 590                                 if (nf->sk > of->sk)
 591                                         np = &parent->rb_right;
 592                                 else
 593                                         np = &parent->rb_left;
 594                         }
 595
 596                         rb_link_node(&of->fq_node, parent, np);
 597                         rb_insert_color(&of->fq_node, nroot);
 598                 }
 599         }
 600         q->flows -= fcnt;
 601         q->inactive_flows -= fcnt;
 602         q->stat_gc_flows += fcnt;
 603 }
 604
 605 static void *fq_alloc_node(size_t sz, int node)
 606 {
 607         void *ptr;
 608
 609         ptr = kmalloc_node(sz, GFP_KERNEL | __GFP_REPEAT | __GFP_NOWARN, node);
 610         if (!ptr)
 611                 ptr = vmalloc_node(sz, node);
 612         return ptr;
 613 }
 614
 615 static void fq_free(void *addr)
 616 {
 617         kvfree(addr);
 618 }
 619
 620 static int fq_resize(struct Qdisc *sch, u32 log)
 621 {
 622         struct fq_sched_data *q = qdisc_priv(sch);
 623         struct rb_root *array;
 624         void *old_fq_root;
 625         u32 idx;
 626
 627         if (q->fq_root && log == q->fq_trees_log)
 628                 return 0;
 629
 630         /* If XPS was setup, we can allocate memory on right NUMA node */
 631         array = fq_alloc_node(sizeof(struct rb_root) << log,
 632                               netdev_queue_numa_node_read(sch->dev_queue));
 633         if (!array)
 634                 return -ENOMEM;
 635
 636         for (idx = 0; idx < (1U << log); idx++)
 637                 array[idx] = RB_ROOT;
 638
 639         sch_tree_lock(sch);
 640
 641         old_fq_root = q->fq_root;
 642         if (old_fq_root)
 643                 fq_rehash(q, old_fq_root, q->fq_trees_log, array, log);
 644
 645         q->fq_root = array;
 646         q->fq_trees_log = log;
 647
 648         sch_tree_unlock(sch);
 649
 650         fq_free(old_fq_root);
 651
 652         return 0;
 653 }
 654
 655 static const struct nla_policy fq_policy[TCA_FQ_MAX + 1] = {
 656         [TCA_FQ_PLIMIT]                 = { .type = NLA_U32 },
 657         [TCA_FQ_FLOW_PLIMIT]            = { .type = NLA_U32 },
 658         [TCA_FQ_QUANTUM]                = { .type = NLA_U32 },
 659         [TCA_FQ_INITIAL_QUANTUM]        = { .type = NLA_U32 },
 660         [TCA_FQ_RATE_ENABLE]            = { .type = NLA_U32 },
 661         [TCA_FQ_FLOW_DEFAULT_RATE]      = { .type = NLA_U32 },
 662         [TCA_FQ_FLOW_MAX_RATE]          = { .type = NLA_U32 },
 663         [TCA_FQ_BUCKETS_LOG]            = { .type = NLA_U32 },
 664         [TCA_FQ_FLOW_REFILL_DELAY]      = { .type = NLA_U32 },
 665 };
 666
 667 static int fq_change(struct Qdisc *sch, struct nlattr *opt)
 668 {
 669         struct fq_sched_data *q = qdisc_priv(sch);
 670         struct nlattr *tb[TCA_FQ_MAX + 1];
 671         int err, drop_count = 0;
 672         unsigned drop_len = 0;
 673         u32 fq_log;
 674
 675         if (!opt)
 676                 return -EINVAL;
 677
 678         err = nla_parse_nested(tb, TCA_FQ_MAX, opt, fq_policy);
 679         if (err < 0)
 680                 return err;
 681
 682         sch_tree_lock(sch);
 683
 684         fq_log = q->fq_trees_log;
 685
 686         if (tb[TCA_FQ_BUCKETS_LOG]) {
 687                 u32 nval = nla_get_u32(tb[TCA_FQ_BUCKETS_LOG]);
 688
 689                 if (nval >= 1 && nval <= ilog2(256*1024))
 690                         fq_log = nval;
 691                 else
 692                         err = -EINVAL;
 693         }
 694         if (tb[TCA_FQ_PLIMIT])
 695                 sch->limit = nla_get_u32(tb[TCA_FQ_PLIMIT]);
 696
 697         if (tb[TCA_FQ_FLOW_PLIMIT])
 698                 q->flow_plimit = nla_get_u32(tb[TCA_FQ_FLOW_PLIMIT]);
 699
 700         if (tb[TCA_FQ_QUANTUM]) {
 701                 u32 quantum = nla_get_u32(tb[TCA_FQ_QUANTUM]);
 702
 703                 if (quantum > 0)
 704                         q->quantum = quantum;
 705                 else
 706                         err = -EINVAL;
 707         }
 708
 709         if (tb[TCA_FQ_INITIAL_QUANTUM])
 710                 q->initial_quantum = nla_get_u32(tb[TCA_FQ_INITIAL_QUANTUM]);
 711
 712         if (tb[TCA_FQ_FLOW_DEFAULT_RATE])
 713                 pr_warn_ratelimited("sch_fq: defrate %u ignored.\n",
 714                                     nla_get_u32(tb[TCA_FQ_FLOW_DEFAULT_RATE]));
 715
 716         if (tb[TCA_FQ_FLOW_MAX_RATE])
 717                 q->flow_max_rate = nla_get_u32(tb[TCA_FQ_FLOW_MAX_RATE]);
 718
 719         if (tb[TCA_FQ_RATE_ENABLE]) {
 720                 u32 enable = nla_get_u32(tb[TCA_FQ_RATE_ENABLE]);
 721
 722                 if (enable <= 1)
 723                         q->rate_enable = enable;
 724                 else
 725                         err = -EINVAL;
 726         }
 727
 728         if (tb[TCA_FQ_FLOW_REFILL_DELAY]) {
 729                 u32 usecs_delay = nla_get_u32(tb[TCA_FQ_FLOW_REFILL_DELAY]) ;
 730
 731                 q->flow_refill_delay = usecs_to_jiffies(usecs_delay);
 732         }
 733
 734         if (tb[TCA_FQ_ORPHAN_MASK])
 735                 q->orphan_mask = nla_get_u32(tb[TCA_FQ_ORPHAN_MASK]);
 736
 737         if (!err) {
 738                 sch_tree_unlock(sch);
 739                 err = fq_resize(sch, fq_log);
 740                 sch_tree_lock(sch);
 741         }
 742         while (sch->q.qlen > sch->limit) {
 743                 struct sk_buff *skb = fq_dequeue(sch);
 744
 745                 if (!skb)
 746                         break;
 747                 drop_len += qdisc_pkt_len(skb);
 748                 rtnl_kfree_skbs(skb, skb);
 749                 drop_count++;
 750         }
 751         qdisc_tree_reduce_backlog(sch, drop_count, drop_len);
 752
 753         sch_tree_unlock(sch);
 754         return err;
 755 }
 756
 757 static void fq_destroy(struct Qdisc *sch)
 758 {
 759         struct fq_sched_data *q = qdisc_priv(sch);
 760
 761         fq_reset(sch);
 762         fq_free(q->fq_root);
 763         qdisc_watchdog_cancel(&q->watchdog);
 764 }
 765
 766 static int fq_init(struct Qdisc *sch, struct nlattr *opt)
 767 {
 768         struct fq_sched_data *q = qdisc_priv(sch);
 769         int err;
 770
 771         sch->limit              = 10000;
 772         q->flow_plimit          = 100;
 773         q->quantum              = 2 * psched_mtu(qdisc_dev(sch));
 774         q->initial_quantum      = 10 * psched_mtu(qdisc_dev(sch));
 775         q->flow_refill_delay    = msecs_to_jiffies(40);
 776         q->flow_max_rate        = ~0U;
 777         q->rate_enable          = 1;
 778         q->new_flows.first      = NULL;
 779         q->old_flows.first      = NULL;
 780         q->delayed              = RB_ROOT;
 781         q->fq_root              = NULL;
 782         q->fq_trees_log         = ilog2(1024);
 783         q->orphan_mask          = 1024 - 1;
 784         qdisc_watchdog_init(&q->watchdog, sch);
 785
 786         if (opt)
 787                 err = fq_change(sch, opt);
 788         else
 789                 err = fq_resize(sch, q->fq_trees_log);
 790
 791         return err;
 792 }
 793
 794 static int fq_dump(struct Qdisc *sch, struct sk_buff *skb)
 795 {
 796         struct fq_sched_data *q = qdisc_priv(sch);
 797         struct nlattr *opts;
 798
 799         opts = nla_nest_start(skb, TCA_OPTIONS);
 800         if (opts == NULL)
 801                 goto nla_put_failure;
 802
 803         /* TCA_FQ_FLOW_DEFAULT_RATE is not used anymore */
 804
 805         if (nla_put_u32(skb, TCA_FQ_PLIMIT, sch->limit) ||
 806             nla_put_u32(skb, TCA_FQ_FLOW_PLIMIT, q->flow_plimit) ||
 807             nla_put_u32(skb, TCA_FQ_QUANTUM, q->quantum) ||
 808             nla_put_u32(skb, TCA_FQ_INITIAL_QUANTUM, q->initial_quantum) ||
 809             nla_put_u32(skb, TCA_FQ_RATE_ENABLE, q->rate_enable) ||
 810             nla_put_u32(skb, TCA_FQ_FLOW_MAX_RATE, q->flow_max_rate) ||
 811             nla_put_u32(skb, TCA_FQ_FLOW_REFILL_DELAY,
 812                         jiffies_to_usecs(q->flow_refill_delay)) ||
 813             nla_put_u32(skb, TCA_FQ_ORPHAN_MASK, q->orphan_mask) ||
 814             nla_put_u32(skb, TCA_FQ_BUCKETS_LOG, q->fq_trees_log))
 815                 goto nla_put_failure;
 816
 817         return nla_nest_end(skb, opts);
 818
 819 nla_put_failure:
 820         return -1;
 821 }
 822
 823 static int fq_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
 824 {
 825         struct fq_sched_data *q = qdisc_priv(sch);
 826         u64 now = ktime_get_ns();
 827         struct tc_fq_qd_stats st = {
 828                 .gc_flows               = q->stat_gc_flows,
 829                 .highprio_packets       = q->stat_internal_packets,
 830                 .tcp_retrans            = q->stat_tcp_retrans,
 831                 .throttled              = q->stat_throttled,
 832                 .flows_plimit           = q->stat_flows_plimit,
 833                 .pkts_too_long          = q->stat_pkts_too_long,
 834                 .allocation_errors      = q->stat_allocation_errors,
 835                 .flows                  = q->flows,
 836                 .inactive_flows         = q->inactive_flows,
 837                 .throttled_flows        = q->throttled_flows,
 838                 .time_next_delayed_flow = q->time_next_delayed_flow - now,
 839         };
 840
 841         return gnet_stats_copy_app(d, &st, sizeof(st));
 842 }
 843
 844 static struct Qdisc_ops fq_qdisc_ops __read_mostly = {
 845         .id             =       "fq",
 846         .priv_size      =       sizeof(struct fq_sched_data),
 847
 848         .enqueue        =       fq_enqueue,
 849         .dequeue        =       fq_dequeue,
 850         .peek           =       qdisc_peek_dequeued,
 851         .init           =       fq_init,
 852         .reset          =       fq_reset,
 853         .destroy        =       fq_destroy,
 854         .change         =       fq_change,
 855         .dump           =       fq_dump,
 856         .dump_stats     =       fq_dump_stats,
 857         .owner          =       THIS_MODULE,
 858 };
 859
 860 static int __init fq_module_init(void)
 861 {
 862         int ret;
 863
 864         fq_flow_cachep = kmem_cache_create("fq_flow_cache",
 865                                            sizeof(struct fq_flow),
 866                                            0, 0, NULL);
 867         if (!fq_flow_cachep)
 868                 return -ENOMEM;
 869
 870         ret = register_qdisc(&fq_qdisc_ops);
 871         if (ret)
 872                 kmem_cache_destroy(fq_flow_cachep);
 873         return ret;
 874 }
 875
 876 static void __exit fq_module_exit(void)
 877 {
 878         unregister_qdisc(&fq_qdisc_ops);
 879         kmem_cache_destroy(fq_flow_cachep);
 880 }
 881
 882 module_init(fq_module_init)
 883 module_exit(fq_module_exit)
 884 MODULE_AUTHOR("Eric Dumazet");
 885 MODULE_LICENSE("GPL");