linux_dsm_epyc7002/net/sched/sch_fq_codel.c
Eric Dumazet a5d2809040 codel: fix maxpacket/mtu confusion
Under presence of TSO/GSO/GRO packets, codel at low rates can be quite
useless. In following example, not a single packet was ever dropped,
while average delay in codel queue is ~100 ms !

qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms
 Sent 134376498 bytes 88797 pkt (dropped 0, overlimits 0 requeues 0)
 backlog 13626b 3p requeues 0
  count 0 lastcount 0 ldelay 96.9ms drop_next 0us
  maxpacket 9084 ecn_mark 0 drop_overlimit 0

This comes from a confusion of what should be the minimal backlog. It is
pretty clear it is not 64KB or whatever max GSO packet ever reached the
qdisc.

codel intent was to use MTU of the device.

After the fix, we finally drop some packets, and rtt/cwnd of my single
TCP flow are meeting our expectations.

qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms
 Sent 102798497 bytes 67912 pkt (dropped 1365, overlimits 0 requeues 0)
 backlog 6056b 3p requeues 0
  count 1 lastcount 1 ldelay 36.3ms drop_next 0us
  maxpacket 10598 ecn_mark 0 drop_overlimit 0

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Kathleen Nichols <nichols@pollere.com>
Cc: Dave Taht <dave.taht@gmail.com>
Cc: Van Jacobson <vanj@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-03 22:17:40 -04:00

625 lines
16 KiB
C

/*
* Fair Queue CoDel discipline
*
* 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.
*
* Copyright (C) 2012 Eric Dumazet <edumazet@google.com>
*/
#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/jhash.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
#include <net/flow_keys.h>
#include <net/codel.h>
/* Fair Queue CoDel.
*
* Principles :
* Packets are classified (internal classifier or external) on flows.
* This is a Stochastic model (as we use a hash, several flows
* might be hashed on same slot)
* Each flow has a CoDel managed queue.
* Flows are linked onto two (Round Robin) lists,
* so that new flows have priority on old ones.
*
* For a given flow, packets are not reordered (CoDel uses a FIFO)
* head drops only.
* ECN capability is on by default.
* Low memory footprint (64 bytes per flow)
*/
struct fq_codel_flow {
struct sk_buff *head;
struct sk_buff *tail;
struct list_head flowchain;
int deficit;
u32 dropped; /* number of drops (or ECN marks) on this flow */
struct codel_vars cvars;
}; /* please try to keep this structure <= 64 bytes */
struct fq_codel_sched_data {
struct tcf_proto __rcu *filter_list; /* optional external classifier */
struct fq_codel_flow *flows; /* Flows table [flows_cnt] */
u32 *backlogs; /* backlog table [flows_cnt] */
u32 flows_cnt; /* number of flows */
u32 perturbation; /* hash perturbation */
u32 quantum; /* psched_mtu(qdisc_dev(sch)); */
struct codel_params cparams;
struct codel_stats cstats;
u32 drop_overlimit;
u32 new_flow_count;
struct list_head new_flows; /* list of new flows */
struct list_head old_flows; /* list of old flows */
};
static unsigned int fq_codel_hash(const struct fq_codel_sched_data *q,
const struct sk_buff *skb)
{
struct flow_keys keys;
unsigned int hash;
skb_flow_dissect(skb, &keys);
hash = jhash_3words((__force u32)keys.dst,
(__force u32)keys.src ^ keys.ip_proto,
(__force u32)keys.ports, q->perturbation);
return reciprocal_scale(hash, q->flows_cnt);
}
static unsigned int fq_codel_classify(struct sk_buff *skb, struct Qdisc *sch,
int *qerr)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct tcf_proto *filter;
struct tcf_result res;
int result;
if (TC_H_MAJ(skb->priority) == sch->handle &&
TC_H_MIN(skb->priority) > 0 &&
TC_H_MIN(skb->priority) <= q->flows_cnt)
return TC_H_MIN(skb->priority);
filter = rcu_dereference_bh(q->filter_list);
if (!filter)
return fq_codel_hash(q, skb) + 1;
*qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
result = tc_classify(skb, filter, &res);
if (result >= 0) {
#ifdef CONFIG_NET_CLS_ACT
switch (result) {
case TC_ACT_STOLEN:
case TC_ACT_QUEUED:
*qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN;
case TC_ACT_SHOT:
return 0;
}
#endif
if (TC_H_MIN(res.classid) <= q->flows_cnt)
return TC_H_MIN(res.classid);
}
return 0;
}
/* helper functions : might be changed when/if skb use a standard list_head */
/* remove one skb from head of slot queue */
static inline struct sk_buff *dequeue_head(struct fq_codel_flow *flow)
{
struct sk_buff *skb = flow->head;
flow->head = skb->next;
skb->next = NULL;
return skb;
}
/* add skb to flow queue (tail add) */
static inline void flow_queue_add(struct fq_codel_flow *flow,
struct sk_buff *skb)
{
if (flow->head == NULL)
flow->head = skb;
else
flow->tail->next = skb;
flow->tail = skb;
skb->next = NULL;
}
static unsigned int fq_codel_drop(struct Qdisc *sch)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb;
unsigned int maxbacklog = 0, idx = 0, i, len;
struct fq_codel_flow *flow;
/* Queue is full! Find the fat flow and drop packet from it.
* This might sound expensive, but with 1024 flows, we scan
* 4KB of memory, and we dont need to handle a complex tree
* in fast path (packet queue/enqueue) with many cache misses.
*/
for (i = 0; i < q->flows_cnt; i++) {
if (q->backlogs[i] > maxbacklog) {
maxbacklog = q->backlogs[i];
idx = i;
}
}
flow = &q->flows[idx];
skb = dequeue_head(flow);
len = qdisc_pkt_len(skb);
q->backlogs[idx] -= len;
kfree_skb(skb);
sch->q.qlen--;
qdisc_qstats_drop(sch);
qdisc_qstats_backlog_dec(sch, skb);
flow->dropped++;
return idx;
}
static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
unsigned int idx;
struct fq_codel_flow *flow;
int uninitialized_var(ret);
idx = fq_codel_classify(skb, sch, &ret);
if (idx == 0) {
if (ret & __NET_XMIT_BYPASS)
qdisc_qstats_drop(sch);
kfree_skb(skb);
return ret;
}
idx--;
codel_set_enqueue_time(skb);
flow = &q->flows[idx];
flow_queue_add(flow, skb);
q->backlogs[idx] += qdisc_pkt_len(skb);
qdisc_qstats_backlog_inc(sch, skb);
if (list_empty(&flow->flowchain)) {
list_add_tail(&flow->flowchain, &q->new_flows);
q->new_flow_count++;
flow->deficit = q->quantum;
flow->dropped = 0;
}
if (++sch->q.qlen <= sch->limit)
return NET_XMIT_SUCCESS;
q->drop_overlimit++;
/* Return Congestion Notification only if we dropped a packet
* from this flow.
*/
if (fq_codel_drop(sch) == idx)
return NET_XMIT_CN;
/* As we dropped a packet, better let upper stack know this */
qdisc_tree_decrease_qlen(sch, 1);
return NET_XMIT_SUCCESS;
}
/* This is the specific function called from codel_dequeue()
* to dequeue a packet from queue. Note: backlog is handled in
* codel, we dont need to reduce it here.
*/
static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct fq_codel_flow *flow;
struct sk_buff *skb = NULL;
flow = container_of(vars, struct fq_codel_flow, cvars);
if (flow->head) {
skb = dequeue_head(flow);
q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb);
sch->q.qlen--;
}
return skb;
}
static struct sk_buff *fq_codel_dequeue(struct Qdisc *sch)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb;
struct fq_codel_flow *flow;
struct list_head *head;
u32 prev_drop_count, prev_ecn_mark;
begin:
head = &q->new_flows;
if (list_empty(head)) {
head = &q->old_flows;
if (list_empty(head))
return NULL;
}
flow = list_first_entry(head, struct fq_codel_flow, flowchain);
if (flow->deficit <= 0) {
flow->deficit += q->quantum;
list_move_tail(&flow->flowchain, &q->old_flows);
goto begin;
}
prev_drop_count = q->cstats.drop_count;
prev_ecn_mark = q->cstats.ecn_mark;
skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats,
dequeue);
flow->dropped += q->cstats.drop_count - prev_drop_count;
flow->dropped += q->cstats.ecn_mark - prev_ecn_mark;
if (!skb) {
/* force a pass through old_flows to prevent starvation */
if ((head == &q->new_flows) && !list_empty(&q->old_flows))
list_move_tail(&flow->flowchain, &q->old_flows);
else
list_del_init(&flow->flowchain);
goto begin;
}
qdisc_bstats_update(sch, skb);
flow->deficit -= qdisc_pkt_len(skb);
/* We cant call qdisc_tree_decrease_qlen() if our qlen is 0,
* or HTB crashes. Defer it for next round.
*/
if (q->cstats.drop_count && sch->q.qlen) {
qdisc_tree_decrease_qlen(sch, q->cstats.drop_count);
q->cstats.drop_count = 0;
}
return skb;
}
static void fq_codel_reset(struct Qdisc *sch)
{
struct sk_buff *skb;
while ((skb = fq_codel_dequeue(sch)) != NULL)
kfree_skb(skb);
}
static const struct nla_policy fq_codel_policy[TCA_FQ_CODEL_MAX + 1] = {
[TCA_FQ_CODEL_TARGET] = { .type = NLA_U32 },
[TCA_FQ_CODEL_LIMIT] = { .type = NLA_U32 },
[TCA_FQ_CODEL_INTERVAL] = { .type = NLA_U32 },
[TCA_FQ_CODEL_ECN] = { .type = NLA_U32 },
[TCA_FQ_CODEL_FLOWS] = { .type = NLA_U32 },
[TCA_FQ_CODEL_QUANTUM] = { .type = NLA_U32 },
};
static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct nlattr *tb[TCA_FQ_CODEL_MAX + 1];
int err;
if (!opt)
return -EINVAL;
err = nla_parse_nested(tb, TCA_FQ_CODEL_MAX, opt, fq_codel_policy);
if (err < 0)
return err;
if (tb[TCA_FQ_CODEL_FLOWS]) {
if (q->flows)
return -EINVAL;
q->flows_cnt = nla_get_u32(tb[TCA_FQ_CODEL_FLOWS]);
if (!q->flows_cnt ||
q->flows_cnt > 65536)
return -EINVAL;
}
sch_tree_lock(sch);
if (tb[TCA_FQ_CODEL_TARGET]) {
u64 target = nla_get_u32(tb[TCA_FQ_CODEL_TARGET]);
q->cparams.target = (target * NSEC_PER_USEC) >> CODEL_SHIFT;
}
if (tb[TCA_FQ_CODEL_INTERVAL]) {
u64 interval = nla_get_u32(tb[TCA_FQ_CODEL_INTERVAL]);
q->cparams.interval = (interval * NSEC_PER_USEC) >> CODEL_SHIFT;
}
if (tb[TCA_FQ_CODEL_LIMIT])
sch->limit = nla_get_u32(tb[TCA_FQ_CODEL_LIMIT]);
if (tb[TCA_FQ_CODEL_ECN])
q->cparams.ecn = !!nla_get_u32(tb[TCA_FQ_CODEL_ECN]);
if (tb[TCA_FQ_CODEL_QUANTUM])
q->quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM]));
while (sch->q.qlen > sch->limit) {
struct sk_buff *skb = fq_codel_dequeue(sch);
kfree_skb(skb);
q->cstats.drop_count++;
}
qdisc_tree_decrease_qlen(sch, q->cstats.drop_count);
q->cstats.drop_count = 0;
sch_tree_unlock(sch);
return 0;
}
static void *fq_codel_zalloc(size_t sz)
{
void *ptr = kzalloc(sz, GFP_KERNEL | __GFP_NOWARN);
if (!ptr)
ptr = vzalloc(sz);
return ptr;
}
static void fq_codel_free(void *addr)
{
kvfree(addr);
}
static void fq_codel_destroy(struct Qdisc *sch)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
tcf_destroy_chain(&q->filter_list);
fq_codel_free(q->backlogs);
fq_codel_free(q->flows);
}
static int fq_codel_init(struct Qdisc *sch, struct nlattr *opt)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
int i;
sch->limit = 10*1024;
q->flows_cnt = 1024;
q->quantum = psched_mtu(qdisc_dev(sch));
q->perturbation = prandom_u32();
INIT_LIST_HEAD(&q->new_flows);
INIT_LIST_HEAD(&q->old_flows);
codel_params_init(&q->cparams, sch);
codel_stats_init(&q->cstats);
q->cparams.ecn = true;
if (opt) {
int err = fq_codel_change(sch, opt);
if (err)
return err;
}
if (!q->flows) {
q->flows = fq_codel_zalloc(q->flows_cnt *
sizeof(struct fq_codel_flow));
if (!q->flows)
return -ENOMEM;
q->backlogs = fq_codel_zalloc(q->flows_cnt * sizeof(u32));
if (!q->backlogs) {
fq_codel_free(q->flows);
return -ENOMEM;
}
for (i = 0; i < q->flows_cnt; i++) {
struct fq_codel_flow *flow = q->flows + i;
INIT_LIST_HEAD(&flow->flowchain);
codel_vars_init(&flow->cvars);
}
}
if (sch->limit >= 1)
sch->flags |= TCQ_F_CAN_BYPASS;
else
sch->flags &= ~TCQ_F_CAN_BYPASS;
return 0;
}
static int fq_codel_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct nlattr *opts;
opts = nla_nest_start(skb, TCA_OPTIONS);
if (opts == NULL)
goto nla_put_failure;
if (nla_put_u32(skb, TCA_FQ_CODEL_TARGET,
codel_time_to_us(q->cparams.target)) ||
nla_put_u32(skb, TCA_FQ_CODEL_LIMIT,
sch->limit) ||
nla_put_u32(skb, TCA_FQ_CODEL_INTERVAL,
codel_time_to_us(q->cparams.interval)) ||
nla_put_u32(skb, TCA_FQ_CODEL_ECN,
q->cparams.ecn) ||
nla_put_u32(skb, TCA_FQ_CODEL_QUANTUM,
q->quantum) ||
nla_put_u32(skb, TCA_FQ_CODEL_FLOWS,
q->flows_cnt))
goto nla_put_failure;
return nla_nest_end(skb, opts);
nla_put_failure:
return -1;
}
static int fq_codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
struct tc_fq_codel_xstats st = {
.type = TCA_FQ_CODEL_XSTATS_QDISC,
};
struct list_head *pos;
st.qdisc_stats.maxpacket = q->cstats.maxpacket;
st.qdisc_stats.drop_overlimit = q->drop_overlimit;
st.qdisc_stats.ecn_mark = q->cstats.ecn_mark;
st.qdisc_stats.new_flow_count = q->new_flow_count;
list_for_each(pos, &q->new_flows)
st.qdisc_stats.new_flows_len++;
list_for_each(pos, &q->old_flows)
st.qdisc_stats.old_flows_len++;
return gnet_stats_copy_app(d, &st, sizeof(st));
}
static struct Qdisc *fq_codel_leaf(struct Qdisc *sch, unsigned long arg)
{
return NULL;
}
static unsigned long fq_codel_get(struct Qdisc *sch, u32 classid)
{
return 0;
}
static unsigned long fq_codel_bind(struct Qdisc *sch, unsigned long parent,
u32 classid)
{
/* we cannot bypass queue discipline anymore */
sch->flags &= ~TCQ_F_CAN_BYPASS;
return 0;
}
static void fq_codel_put(struct Qdisc *q, unsigned long cl)
{
}
static struct tcf_proto __rcu **fq_codel_find_tcf(struct Qdisc *sch,
unsigned long cl)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
if (cl)
return NULL;
return &q->filter_list;
}
static int fq_codel_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
tcm->tcm_handle |= TC_H_MIN(cl);
return 0;
}
static int fq_codel_dump_class_stats(struct Qdisc *sch, unsigned long cl,
struct gnet_dump *d)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
u32 idx = cl - 1;
struct gnet_stats_queue qs = { 0 };
struct tc_fq_codel_xstats xstats;
if (idx < q->flows_cnt) {
const struct fq_codel_flow *flow = &q->flows[idx];
const struct sk_buff *skb = flow->head;
memset(&xstats, 0, sizeof(xstats));
xstats.type = TCA_FQ_CODEL_XSTATS_CLASS;
xstats.class_stats.deficit = flow->deficit;
xstats.class_stats.ldelay =
codel_time_to_us(flow->cvars.ldelay);
xstats.class_stats.count = flow->cvars.count;
xstats.class_stats.lastcount = flow->cvars.lastcount;
xstats.class_stats.dropping = flow->cvars.dropping;
if (flow->cvars.dropping) {
codel_tdiff_t delta = flow->cvars.drop_next -
codel_get_time();
xstats.class_stats.drop_next = (delta >= 0) ?
codel_time_to_us(delta) :
-codel_time_to_us(-delta);
}
while (skb) {
qs.qlen++;
skb = skb->next;
}
qs.backlog = q->backlogs[idx];
qs.drops = flow->dropped;
}
if (gnet_stats_copy_queue(d, NULL, &qs, 0) < 0)
return -1;
if (idx < q->flows_cnt)
return gnet_stats_copy_app(d, &xstats, sizeof(xstats));
return 0;
}
static void fq_codel_walk(struct Qdisc *sch, struct qdisc_walker *arg)
{
struct fq_codel_sched_data *q = qdisc_priv(sch);
unsigned int i;
if (arg->stop)
return;
for (i = 0; i < q->flows_cnt; i++) {
if (list_empty(&q->flows[i].flowchain) ||
arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(sch, i + 1, arg) < 0) {
arg->stop = 1;
break;
}
arg->count++;
}
}
static const struct Qdisc_class_ops fq_codel_class_ops = {
.leaf = fq_codel_leaf,
.get = fq_codel_get,
.put = fq_codel_put,
.tcf_chain = fq_codel_find_tcf,
.bind_tcf = fq_codel_bind,
.unbind_tcf = fq_codel_put,
.dump = fq_codel_dump_class,
.dump_stats = fq_codel_dump_class_stats,
.walk = fq_codel_walk,
};
static struct Qdisc_ops fq_codel_qdisc_ops __read_mostly = {
.cl_ops = &fq_codel_class_ops,
.id = "fq_codel",
.priv_size = sizeof(struct fq_codel_sched_data),
.enqueue = fq_codel_enqueue,
.dequeue = fq_codel_dequeue,
.peek = qdisc_peek_dequeued,
.drop = fq_codel_drop,
.init = fq_codel_init,
.reset = fq_codel_reset,
.destroy = fq_codel_destroy,
.change = fq_codel_change,
.dump = fq_codel_dump,
.dump_stats = fq_codel_dump_stats,
.owner = THIS_MODULE,
};
static int __init fq_codel_module_init(void)
{
return register_qdisc(&fq_codel_qdisc_ops);
}
static void __exit fq_codel_module_exit(void)
{
unregister_qdisc(&fq_codel_qdisc_ops);
}
module_init(fq_codel_module_init)
module_exit(fq_codel_module_exit)
MODULE_AUTHOR("Eric Dumazet");
MODULE_LICENSE("GPL");