Commit afe4fd06 authored by Eric Dumazet's avatar Eric Dumazet Committed by David S. Miller
Browse files

pkt_sched: fq: Fair Queue packet scheduler



- Uses perfect flow match (not stochastic hash like SFQ/FQ_codel)
- Uses the new_flow/old_flow separation from FQ_codel
- New flows get an initial credit allowing IW10 without added delay.
- Special FIFO queue for high prio packets (no need for PRIO + FQ)
- Uses a hash table of RB trees to locate the flows at enqueue() time
- Smart on demand gc (at enqueue() time, RB tree lookup evicts old
  unused flows)
- Dynamic memory allocations.
- Designed to allow millions of concurrent flows per Qdisc.
- Small memory footprint : ~8K per Qdisc, and 104 bytes per flow.
- Single high resolution timer for throttled flows (if any).
- One RB tree to link throttled flows.
- Ability to have a max rate per flow. We might add a socket option
  to add per socket limitation.

Attempts have been made to add TCP pacing in TCP stack, but this
seems to add complex code to an already complex stack.

TCP pacing is welcomed for flows having idle times, as the cwnd
permits TCP stack to queue a possibly large number of packets.

This removes the 'slow start after idle' choice, hitting badly
large BDP flows, and applications delivering chunks of data
as video streams.

Nicely spaced packets :
Here interface is 10Gbit, but flow bottleneck is ~20Mbit

cwin is big, yet FQ avoids the typical bursts generated by TCP
(as in netperf TCP_RR -- -r 100000,100000)

15:01:23.545279 IP A > B: . 78193:81089(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.545394 IP B > A: . ack 81089 win 3668 <nop,nop,timestamp 11597985 1115>
15:01:23.546488 IP A > B: . 81089:83985(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.546565 IP B > A: . ack 83985 win 3668 <nop,nop,timestamp 11597986 1115>
15:01:23.547713 IP A > B: . 83985:86881(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.547778 IP B > A: . ack 86881 win 3668 <nop,nop,timestamp 11597987 1115>
15:01:23.548911 IP A > B: . 86881:89777(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.548949 IP B > A: . ack 89777 win 3668 <nop,nop,timestamp 11597988 1115>
15:01:23.550116 IP A > B: . 89777:92673(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.550182 IP B > A: . ack 92673 win 3668 <nop,nop,timestamp 11597989 1115>
15:01:23.551333 IP A > B: . 92673:95569(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.551406 IP B > A: . ack 95569 win 3668 <nop,nop,timestamp 11597991 1115>
15:01:23.552539 IP A > B: . 95569:98465(2896) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.552576 IP B > A: . ack 98465 win 3668 <nop,nop,timestamp 11597992 1115>
15:01:23.553756 IP A > B: . 98465:99913(1448) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.554138 IP A > B: P 99913:100001(88) ack 65248 win 3125 <nop,nop,timestamp 1115 11597805>
15:01:23.554204 IP B > A: . ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.554234 IP B > A: . 65248:68144(2896) ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.555620 IP B > A: . 68144:71040(2896) ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.557005 IP B > A: . 71040:73936(2896) ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.558390 IP B > A: . 73936:76832(2896) ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.559773 IP B > A: . 76832:79728(2896) ack 100001 win 3668 <nop,nop,timestamp 11597993 1115>
15:01:23.561158 IP B > A: . 79728:82624(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.562543 IP B > A: . 82624:85520(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.563928 IP B > A: . 85520:88416(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.565313 IP B > A: . 88416:91312(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.566698 IP B > A: . 91312:94208(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.568083 IP B > A: . 94208:97104(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.569467 IP B > A: . 97104:100000(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.570852 IP B > A: . 100000:102896(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.572237 IP B > A: . 102896:105792(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.573639 IP B > A: . 105792:108688(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.575024 IP B > A: . 108688:111584(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.576408 IP B > A: . 111584:114480(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>
15:01:23.577793 IP B > A: . 114480:117376(2896) ack 100001 win 3668 <nop,nop,timestamp 11597994 1115>

TCP timestamps show that most packets from B were queued in the same ms
timeframe (TSval 1159799{3,4}), but FQ managed to send them right
in time to avoid a big burst.

In slow start or steady state, very few packets are throttled [1]

FQ gets a bunch of tunables as :

  limit : max number of packets on whole Qdisc (default 10000)

  flow_limit : max number of packets per flow (default 100)

  quantum : the credit per RR round (default is 2 MTU)

  initial_quantum : initial credit for new flows (default is 10 MTU)

  maxrate : max per flow rate (default : unlimited)

  buckets : number of RB trees (default : 1024) in hash table.
               (consumes 8 bytes per bucket)

  [no]pacing : disable/enable pacing (default is enable)

All of them can be changed on a live qdisc.

$ tc qd add dev eth0 root fq help
Usage: ... fq [ limit PACKETS ] [ flow_limit PACKETS ]
              [ quantum BYTES ] [ initial_quantum BYTES ]
              [ maxrate RATE  ] [ buckets NUMBER ]
              [ [no]pacing ]

$ tc -s -d qd
qdisc fq 8002: dev eth0 root refcnt 32 limit 10000p flow_limit 100p buckets 256 quantum 3028 initial_quantum 15140
 Sent 216532416 bytes 148395 pkt (dropped 0, overlimits 0 requeues 14)
 backlog 0b 0p requeues 14
  511 flows, 511 inactive, 0 throttled
  110 gc, 0 highprio, 0 retrans, 1143 throttled, 0 flows_plimit

[1] Except if initial srtt is overestimated, as if using
cached srtt in tcp metrics. We'll provide a fix for this issue.
Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 7ec06da8
......@@ -744,4 +744,45 @@ struct tc_fq_codel_xstats {
};
};
/* FQ */
enum {
TCA_FQ_UNSPEC,
TCA_FQ_PLIMIT, /* limit of total number of packets in queue */
TCA_FQ_FLOW_PLIMIT, /* limit of packets per flow */
TCA_FQ_QUANTUM, /* RR quantum */
TCA_FQ_INITIAL_QUANTUM, /* RR quantum for new flow */
TCA_FQ_RATE_ENABLE, /* enable/disable rate limiting */
TCA_FQ_FLOW_DEFAULT_RATE,/* for sockets with unspecified sk_rate,
* use the following rate
*/
TCA_FQ_FLOW_MAX_RATE, /* per flow max rate */
TCA_FQ_BUCKETS_LOG, /* log2(number of buckets) */
__TCA_FQ_MAX
};
#define TCA_FQ_MAX (__TCA_FQ_MAX - 1)
struct tc_fq_qd_stats {
__u64 gc_flows;
__u64 highprio_packets;
__u64 tcp_retrans;
__u64 throttled;
__u64 flows_plimit;
__u64 pkts_too_long;
__u64 allocation_errors;
__s64 time_next_delayed_flow;
__u32 flows;
__u32 inactive_flows;
__u32 throttled_flows;
__u32 pad;
};
#endif
......@@ -272,6 +272,20 @@ config NET_SCH_FQ_CODEL
If unsure, say N.
config NET_SCH_FQ
tristate "Fair Queue"
help
Say Y here if you want to use the FQ packet scheduling algorithm.
FQ does flow separation, and is able to respect pacing requirements
set by TCP stack into sk->sk_pacing_rate (for localy generated
traffic)
To compile this driver as a module, choose M here: the module
will be called sch_fq.
If unsure, say N.
config NET_SCH_INGRESS
tristate "Ingress Qdisc"
depends on NET_CLS_ACT
......
......@@ -39,6 +39,7 @@ obj-$(CONFIG_NET_SCH_CHOKE) += sch_choke.o
obj-$(CONFIG_NET_SCH_QFQ) += sch_qfq.o
obj-$(CONFIG_NET_SCH_CODEL) += sch_codel.o
obj-$(CONFIG_NET_SCH_FQ_CODEL) += sch_fq_codel.o
obj-$(CONFIG_NET_SCH_FQ) += sch_fq.o
obj-$(CONFIG_NET_CLS_U32) += cls_u32.o
obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o
......
/*
* net/sched/sch_fq.c Fair Queue Packet Scheduler (per flow pacing)
*
* Copyright (C) 2013 Eric Dumazet <edumazet@google.com>
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Meant to be mostly used for localy generated traffic :
* Fast classification depends on skb->sk being set before reaching us.
* If not, (router workload), we use rxhash as fallback, with 32 bits wide hash.
* All packets belonging to a socket are considered as a 'flow'.
*
* Flows are dynamically allocated and stored in a hash table of RB trees
* They are also part of one Round Robin 'queues' (new or old flows)
*
* Burst avoidance (aka pacing) capability :
*
* Transport (eg TCP) can set in sk->sk_pacing_rate a rate, enqueue a
* bunch of packets, and this packet scheduler adds delay between
* packets to respect rate limitation.
*
* enqueue() :
* - lookup one RB tree (out of 1024 or more) to find the flow.
* If non existent flow, create it, add it to the tree.
* Add skb to the per flow list of skb (fifo).
* - Use a special fifo for high prio packets
*
* dequeue() : serves flows in Round Robin
* Note : When a flow becomes empty, we do not immediately remove it from
* rb trees, for performance reasons (its expected to send additional packets,
* or SLAB cache will reuse socket for another flow)
*/
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/jiffies.h>
#include <linux/string.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/skbuff.h>
#include <linux/slab.h>
#include <linux/rbtree.h>
#include <linux/hash.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
#include <net/sock.h>
#include <net/tcp_states.h>
/*
* Per flow structure, dynamically allocated
*/
struct fq_flow {
struct sk_buff *head; /* list of skbs for this flow : first skb */
union {
struct sk_buff *tail; /* last skb in the list */
unsigned long age; /* jiffies when flow was emptied, for gc */
};
struct rb_node fq_node; /* anchor in fq_root[] trees */
struct sock *sk;
int qlen; /* number of packets in flow queue */
int credit;
u32 socket_hash; /* sk_hash */
struct fq_flow *next; /* next pointer in RR lists, or &detached */
struct rb_node rate_node; /* anchor in q->delayed tree */
u64 time_next_packet;
};
struct fq_flow_head {
struct fq_flow *first;
struct fq_flow *last;
};
struct fq_sched_data {
struct fq_flow_head new_flows;
struct fq_flow_head old_flows;
struct rb_root delayed; /* for rate limited flows */
u64 time_next_delayed_flow;
struct fq_flow internal; /* for non classified or high prio packets */
u32 quantum;
u32 initial_quantum;
u32 flow_default_rate;/* rate per flow : bytes per second */
u32 flow_max_rate; /* optional max rate per flow */
u32 flow_plimit; /* max packets per flow */
struct rb_root *fq_root;
u8 rate_enable;
u8 fq_trees_log;
u32 flows;
u32 inactive_flows;
u32 throttled_flows;
u64 stat_gc_flows;
u64 stat_internal_packets;
u64 stat_tcp_retrans;
u64 stat_throttled;
u64 stat_flows_plimit;
u64 stat_pkts_too_long;
u64 stat_allocation_errors;
struct qdisc_watchdog watchdog;
};
/* special value to mark a detached flow (not on old/new list) */
static struct fq_flow detached, throttled;
static void fq_flow_set_detached(struct fq_flow *f)
{
f->next = &detached;
}
static bool fq_flow_is_detached(const struct fq_flow *f)
{
return f->next == &detached;
}
static void fq_flow_set_throttled(struct fq_sched_data *q, struct fq_flow *f)
{
struct rb_node **p = &q->delayed.rb_node, *parent = NULL;
while (*p) {
struct fq_flow *aux;
parent = *p;
aux = container_of(parent, struct fq_flow, rate_node);
if (f->time_next_packet >= aux->time_next_packet)
p = &parent->rb_right;
else
p = &parent->rb_left;
}
rb_link_node(&f->rate_node, parent, p);
rb_insert_color(&f->rate_node, &q->delayed);
q->throttled_flows++;
q->stat_throttled++;
f->next = &throttled;
if (q->time_next_delayed_flow > f->time_next_packet)
q->time_next_delayed_flow = f->time_next_packet;
}
static struct kmem_cache *fq_flow_cachep __read_mostly;
static void fq_flow_add_tail(struct fq_flow_head *head, struct fq_flow *flow)
{
if (head->first)
head->last->next = flow;
else
head->first = flow;
head->last = flow;
flow->next = NULL;
}
/* limit number of collected flows per round */
#define FQ_GC_MAX 8
#define FQ_GC_AGE (3*HZ)
static bool fq_gc_candidate(const struct fq_flow *f)
{
return fq_flow_is_detached(f) &&
time_after(jiffies, f->age + FQ_GC_AGE);
}
static void fq_gc(struct fq_sched_data *q,
struct rb_root *root,
struct sock *sk)
{
struct fq_flow *f, *tofree[FQ_GC_MAX];
struct rb_node **p, *parent;
int fcnt = 0;
p = &root->rb_node;
parent = NULL;
while (*p) {
parent = *p;
f = container_of(parent, struct fq_flow, fq_node);
if (f->sk == sk)
break;
if (fq_gc_candidate(f)) {
tofree[fcnt++] = f;
if (fcnt == FQ_GC_MAX)
break;
}
if (f->sk > sk)
p = &parent->rb_right;
else
p = &parent->rb_left;
}
q->flows -= fcnt;
q->inactive_flows -= fcnt;
q->stat_gc_flows += fcnt;
while (fcnt) {
struct fq_flow *f = tofree[--fcnt];
rb_erase(&f->fq_node, root);
kmem_cache_free(fq_flow_cachep, f);
}
}
static const u8 prio2band[TC_PRIO_MAX + 1] = {
1, 2, 2, 2, 1, 2, 0, 0 , 1, 1, 1, 1, 1, 1, 1, 1
};
static struct fq_flow *fq_classify(struct sk_buff *skb, struct fq_sched_data *q)
{
struct rb_node **p, *parent;
struct sock *sk = skb->sk;
struct rb_root *root;
struct fq_flow *f;
int band;
/* warning: no starvation prevention... */
band = prio2band[skb->priority & TC_PRIO_MAX];
if (unlikely(band == 0))
return &q->internal;
if (unlikely(!sk)) {
/* By forcing low order bit to 1, we make sure to not
* collide with a local flow (socket pointers are word aligned)
*/
sk = (struct sock *)(skb_get_rxhash(skb) | 1L);
}
root = &q->fq_root[hash_32((u32)(long)sk, q->fq_trees_log)];
if (q->flows >= (2U << q->fq_trees_log) &&
q->inactive_flows > q->flows/2)
fq_gc(q, root, sk);
p = &root->rb_node;
parent = NULL;
while (*p) {
parent = *p;
f = container_of(parent, struct fq_flow, fq_node);
if (f->sk == sk) {
/* socket might have been reallocated, so check
* if its sk_hash is the same.
* It not, we need to refill credit with
* initial quantum
*/
if (unlikely(skb->sk &&
f->socket_hash != sk->sk_hash)) {
f->credit = q->initial_quantum;
f->socket_hash = sk->sk_hash;
}
return f;
}
if (f->sk > sk)
p = &parent->rb_right;
else
p = &parent->rb_left;
}
f = kmem_cache_zalloc(fq_flow_cachep, GFP_ATOMIC | __GFP_NOWARN);
if (unlikely(!f)) {
q->stat_allocation_errors++;
return &q->internal;
}
fq_flow_set_detached(f);
f->sk = sk;
if (skb->sk)
f->socket_hash = sk->sk_hash;
f->credit = q->initial_quantum;
rb_link_node(&f->fq_node, parent, p);
rb_insert_color(&f->fq_node, root);
q->flows++;
q->inactive_flows++;
return f;
}
/* remove one skb from head of flow queue */
static struct sk_buff *fq_dequeue_head(struct fq_flow *flow)
{
struct sk_buff *skb = flow->head;
if (skb) {
flow->head = skb->next;
skb->next = NULL;
flow->qlen--;
}
return skb;
}
/* We might add in the future detection of retransmits
* For the time being, just return false
*/
static bool skb_is_retransmit(struct sk_buff *skb)
{
return false;
}
/* add skb to flow queue
* flow queue is a linked list, kind of FIFO, except for TCP retransmits
* We special case tcp retransmits to be transmitted before other packets.
* We rely on fact that TCP retransmits are unlikely, so we do not waste
* a separate queue or a pointer.
* head-> [retrans pkt 1]
* [retrans pkt 2]
* [ normal pkt 1]
* [ normal pkt 2]
* [ normal pkt 3]
* tail-> [ normal pkt 4]
*/
static void flow_queue_add(struct fq_flow *flow, struct sk_buff *skb)
{
struct sk_buff *prev, *head = flow->head;
skb->next = NULL;
if (!head) {
flow->head = skb;
flow->tail = skb;
return;
}
if (likely(!skb_is_retransmit(skb))) {
flow->tail->next = skb;
flow->tail = skb;
return;
}
/* This skb is a tcp retransmit,
* find the last retrans packet in the queue
*/
prev = NULL;
while (skb_is_retransmit(head)) {
prev = head;
head = head->next;
if (!head)
break;
}
if (!prev) { /* no rtx packet in queue, become the new head */
skb->next = flow->head;
flow->head = skb;
} else {
if (prev == flow->tail)
flow->tail = skb;
else
skb->next = prev->next;
prev->next = skb;
}
}
static int fq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct fq_sched_data *q = qdisc_priv(sch);
struct fq_flow *f;
if (unlikely(sch->q.qlen >= sch->limit))
return qdisc_drop(skb, sch);
f = fq_classify(skb, q);
if (unlikely(f->qlen >= q->flow_plimit && f != &q->internal)) {
q->stat_flows_plimit++;
return qdisc_drop(skb, sch);
}
f->qlen++;
flow_queue_add(f, skb);
if (skb_is_retransmit(skb))
q->stat_tcp_retrans++;
sch->qstats.backlog += qdisc_pkt_len(skb);
if (fq_flow_is_detached(f)) {
fq_flow_add_tail(&q->new_flows, f);
if (q->quantum > f->credit)
f->credit = q->quantum;
q->inactive_flows--;
qdisc_unthrottled(sch);
}
if (unlikely(f == &q->internal)) {
q->stat_internal_packets++;
qdisc_unthrottled(sch);
}
sch->q.qlen++;
return NET_XMIT_SUCCESS;
}
static void fq_check_throttled(struct fq_sched_data *q, u64 now)
{
struct rb_node *p;
if (q->time_next_delayed_flow > now)
return;
q->time_next_delayed_flow = ~0ULL;
while ((p = rb_first(&q->delayed)) != NULL) {
struct fq_flow *f = container_of(p, struct fq_flow, rate_node);
if (f->time_next_packet > now) {
q->time_next_delayed_flow = f->time_next_packet;
break;
}
rb_erase(p, &q->delayed);
q->throttled_flows--;
fq_flow_add_tail(&q->old_flows, f);
}
}
static struct sk_buff *fq_dequeue(struct Qdisc *sch)
{
struct fq_sched_data *q = qdisc_priv(sch);
u64 now = ktime_to_ns(ktime_get());
struct fq_flow_head *head;
struct sk_buff *skb;
struct fq_flow *f;
skb = fq_dequeue_head(&q->internal);
if (skb)
goto out;
fq_check_throttled(q, now);
begin:
head = &q->new_flows;
if (!head->first) {
head = &q->old_flows;
if (!head->first) {
if (q->time_next_delayed_flow != ~0ULL)
qdisc_watchdog_schedule_ns(&q->watchdog,
q->time_next_delayed_flow);
return NULL;
}
}
f = head->first;
if (f->credit <= 0) {
f->credit += q->quantum;
head->first = f->next;
fq_flow_add_tail(&q->old_flows, f);
goto begin;
}
if (unlikely(f->head && now < f->time_next_packet)) {
head->first = f->next;
fq_flow_set_throttled(q, f);
goto begin;
}
skb = fq_dequeue_head(f);
if (!skb) {
head->first = f->next;
/* force a pass through old_flows to prevent starvation */
if ((head == &q->new_flows) && q->old_flows.first) {
fq_flow_add_tail(&q->old_flows, f);
} else {
fq_flow_set_detached(f);
f->age = jiffies;
q->inactive_flows++;
}
goto begin;
}
f->time_next_packet = now;
f->credit -= qdisc_pkt_len(skb);
if (f->credit <= 0 &&
q->rate_enable &&
skb->sk && skb->sk->sk_state != TCP_TIME_WAIT) {
u32 rate = skb->sk->sk_pacing_rate ?: q->flow_default_rate;
rate = min(rate, q->flow_max_rate);
if (rate) {
u64 len = (u64)qdisc_pkt_len(skb) * NSEC_PER_SEC;
do_div(len, rate);
/* Since socket rate can change later,
* clamp the delay to 125 ms.
* TODO: maybe segment the too big skb, as in commit
* e43ac79a4bc ("sch_tbf: segment too big GSO packets")
*/
if (unlikely(len > 125 * NSEC_PER_MSEC)) {
len = 125 * NSEC_PER_MSEC;
q->stat_pkts_too_long++;
}
f->time_next_packet = now + len;
}
}
out:
prefetch(&skb->end);
sch->qstats.backlog -= qdisc_pkt_len(skb);
qdisc_bstats_update(sch, skb);
sch->q.qlen--;
qdisc_unthrottled(sch);
return skb;
}
static void fq_reset(struct Qdisc *sch)
{
struct sk_buff *skb;
while ((skb = fq_dequeue(sch)) != NULL)
kfree_skb(skb);
}
static void fq_rehash(struct fq_sched_data *q,
struct rb_root *old_array, u32 old_log,
struct rb_root *new_array, u32 new_log)
{
struct rb_node *op, **np, *parent;
struct rb_root *oroot, *nroot;
struct fq_flow *of, *nf;
int fcnt = 0;
u32 idx;
for (idx = 0; idx < (1U << old_log); idx++) {
oroot = &old_array[idx];
while ((op = rb_first(oroot)) != NULL) {
rb_erase(op, oroot);
of = container_of(op, struct fq_flow, fq_node);
if (fq_gc_candidate(of)) {
fcnt++;
kmem_cache_free(fq_flow_cachep, of);
continue;
}
nroot = &new_array[hash_32((u32)(long)of->sk, new_log)];
np = &nroot->rb_node;
parent = NULL;
while (*np) {
parent = *np;
nf = container_of(parent, struct fq_flow, fq_node);
BUG_ON(nf->sk == of->sk);
if (nf->sk > of->sk)
np = &parent->rb_right;
else
np = &parent->rb_left;
}
rb_link_node(&of->fq_node, parent, np);
rb_insert_color(&of->fq_node, nroot);
}
}
q->flows -= fcnt;
q->inactive_flows -= fcnt;
q->stat_gc_flows += fcnt;
}