linux/net/sched/sch_red.c
Jarek Poplawski 378a2f090f net_sched: Add qdisc __NET_XMIT_STOLEN flag
Patrick McHardy <kaber@trash.net> noticed:
"The other problem that affects all qdiscs supporting actions is
TC_ACT_QUEUED/TC_ACT_STOLEN getting mapped to NET_XMIT_SUCCESS
even though the packet is not queued, corrupting upper qdiscs'
qlen counters."

and later explained:
"The reason why it translates it at all seems to be to not increase
the drops counter. Within a single qdisc this could be avoided by
other means easily, upper qdiscs would still increase the counter
when we return anything besides NET_XMIT_SUCCESS though.

This means we need a new NET_XMIT return value to indicate this to
the upper qdiscs. So I'd suggest to introduce NET_XMIT_STOLEN,
return that to upper qdiscs and translate it to NET_XMIT_SUCCESS
in dev_queue_xmit, similar to NET_XMIT_BYPASS."

David Miller <davem@davemloft.net> noticed:
"Maybe these NET_XMIT_* values being passed around should be a set of
bits. They could be composed of base meanings, combined with specific
attributes.

So you could say "NET_XMIT_DROP | __NET_XMIT_NO_DROP_COUNT"

The attributes get masked out by the top-level ->enqueue() caller,
such that the base meanings are the only thing that make their
way up into the stack. If it's only about communication within the
qdisc tree, let's simply code it that way."

This patch is trying to realize these ideas.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-04 22:31:03 -07:00

389 lines
8.5 KiB
C

/*
* net/sched/sch_red.c Random Early Detection queue.
*
* 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.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Changes:
* J Hadi Salim 980914: computation fixes
* Alexey Makarenko <makar@phoenix.kharkov.ua> 990814: qave on idle link was calculated incorrectly.
* J Hadi Salim 980816: ECN support
*/
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/skbuff.h>
#include <net/pkt_sched.h>
#include <net/inet_ecn.h>
#include <net/red.h>
/* Parameters, settable by user:
-----------------------------
limit - bytes (must be > qth_max + burst)
Hard limit on queue length, should be chosen >qth_max
to allow packet bursts. This parameter does not
affect the algorithms behaviour and can be chosen
arbitrarily high (well, less than ram size)
Really, this limit will never be reached
if RED works correctly.
*/
struct red_sched_data
{
u32 limit; /* HARD maximal queue length */
unsigned char flags;
struct red_parms parms;
struct red_stats stats;
struct Qdisc *qdisc;
};
static inline int red_use_ecn(struct red_sched_data *q)
{
return q->flags & TC_RED_ECN;
}
static inline int red_use_harddrop(struct red_sched_data *q)
{
return q->flags & TC_RED_HARDDROP;
}
static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
struct Qdisc *child = q->qdisc;
int ret;
q->parms.qavg = red_calc_qavg(&q->parms, child->qstats.backlog);
if (red_is_idling(&q->parms))
red_end_of_idle_period(&q->parms);
switch (red_action(&q->parms, q->parms.qavg)) {
case RED_DONT_MARK:
break;
case RED_PROB_MARK:
sch->qstats.overlimits++;
if (!red_use_ecn(q) || !INET_ECN_set_ce(skb)) {
q->stats.prob_drop++;
goto congestion_drop;
}
q->stats.prob_mark++;
break;
case RED_HARD_MARK:
sch->qstats.overlimits++;
if (red_use_harddrop(q) || !red_use_ecn(q) ||
!INET_ECN_set_ce(skb)) {
q->stats.forced_drop++;
goto congestion_drop;
}
q->stats.forced_mark++;
break;
}
ret = qdisc_enqueue(skb, child);
if (likely(ret == NET_XMIT_SUCCESS)) {
sch->bstats.bytes += qdisc_pkt_len(skb);
sch->bstats.packets++;
sch->q.qlen++;
} else if (net_xmit_drop_count(ret)) {
q->stats.pdrop++;
sch->qstats.drops++;
}
return ret;
congestion_drop:
qdisc_drop(skb, sch);
return NET_XMIT_CN;
}
static int red_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
struct Qdisc *child = q->qdisc;
int ret;
if (red_is_idling(&q->parms))
red_end_of_idle_period(&q->parms);
ret = child->ops->requeue(skb, child);
if (likely(ret == NET_XMIT_SUCCESS)) {
sch->qstats.requeues++;
sch->q.qlen++;
}
return ret;
}
static struct sk_buff * red_dequeue(struct Qdisc* sch)
{
struct sk_buff *skb;
struct red_sched_data *q = qdisc_priv(sch);
struct Qdisc *child = q->qdisc;
skb = child->dequeue(child);
if (skb)
sch->q.qlen--;
else if (!red_is_idling(&q->parms))
red_start_of_idle_period(&q->parms);
return skb;
}
static unsigned int red_drop(struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
struct Qdisc *child = q->qdisc;
unsigned int len;
if (child->ops->drop && (len = child->ops->drop(child)) > 0) {
q->stats.other++;
sch->qstats.drops++;
sch->q.qlen--;
return len;
}
if (!red_is_idling(&q->parms))
red_start_of_idle_period(&q->parms);
return 0;
}
static void red_reset(struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
qdisc_reset(q->qdisc);
sch->q.qlen = 0;
red_restart(&q->parms);
}
static void red_destroy(struct Qdisc *sch)
{
struct red_sched_data *q = qdisc_priv(sch);
qdisc_destroy(q->qdisc);
}
static const struct nla_policy red_policy[TCA_RED_MAX + 1] = {
[TCA_RED_PARMS] = { .len = sizeof(struct tc_red_qopt) },
[TCA_RED_STAB] = { .len = RED_STAB_SIZE },
};
static int red_change(struct Qdisc *sch, struct nlattr *opt)
{
struct red_sched_data *q = qdisc_priv(sch);
struct nlattr *tb[TCA_RED_MAX + 1];
struct tc_red_qopt *ctl;
struct Qdisc *child = NULL;
int err;
if (opt == NULL)
return -EINVAL;
err = nla_parse_nested(tb, TCA_RED_MAX, opt, red_policy);
if (err < 0)
return err;
if (tb[TCA_RED_PARMS] == NULL ||
tb[TCA_RED_STAB] == NULL)
return -EINVAL;
ctl = nla_data(tb[TCA_RED_PARMS]);
if (ctl->limit > 0) {
child = fifo_create_dflt(sch, &bfifo_qdisc_ops, ctl->limit);
if (IS_ERR(child))
return PTR_ERR(child);
}
sch_tree_lock(sch);
q->flags = ctl->flags;
q->limit = ctl->limit;
if (child) {
qdisc_tree_decrease_qlen(q->qdisc, q->qdisc->q.qlen);
qdisc_destroy(xchg(&q->qdisc, child));
}
red_set_parms(&q->parms, ctl->qth_min, ctl->qth_max, ctl->Wlog,
ctl->Plog, ctl->Scell_log,
nla_data(tb[TCA_RED_STAB]));
if (skb_queue_empty(&sch->q))
red_end_of_idle_period(&q->parms);
sch_tree_unlock(sch);
return 0;
}
static int red_init(struct Qdisc* sch, struct nlattr *opt)
{
struct red_sched_data *q = qdisc_priv(sch);
q->qdisc = &noop_qdisc;
return red_change(sch, opt);
}
static int red_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct red_sched_data *q = qdisc_priv(sch);
struct nlattr *opts = NULL;
struct tc_red_qopt opt = {
.limit = q->limit,
.flags = q->flags,
.qth_min = q->parms.qth_min >> q->parms.Wlog,
.qth_max = q->parms.qth_max >> q->parms.Wlog,
.Wlog = q->parms.Wlog,
.Plog = q->parms.Plog,
.Scell_log = q->parms.Scell_log,
};
opts = nla_nest_start(skb, TCA_OPTIONS);
if (opts == NULL)
goto nla_put_failure;
NLA_PUT(skb, TCA_RED_PARMS, sizeof(opt), &opt);
return nla_nest_end(skb, opts);
nla_put_failure:
nla_nest_cancel(skb, opts);
return -EMSGSIZE;
}
static int red_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
{
struct red_sched_data *q = qdisc_priv(sch);
struct tc_red_xstats st = {
.early = q->stats.prob_drop + q->stats.forced_drop,
.pdrop = q->stats.pdrop,
.other = q->stats.other,
.marked = q->stats.prob_mark + q->stats.forced_mark,
};
return gnet_stats_copy_app(d, &st, sizeof(st));
}
static int red_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
struct red_sched_data *q = qdisc_priv(sch);
if (cl != 1)
return -ENOENT;
tcm->tcm_handle |= TC_H_MIN(1);
tcm->tcm_info = q->qdisc->handle;
return 0;
}
static int red_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
struct Qdisc **old)
{
struct red_sched_data *q = qdisc_priv(sch);
if (new == NULL)
new = &noop_qdisc;
sch_tree_lock(sch);
*old = xchg(&q->qdisc, new);
qdisc_tree_decrease_qlen(*old, (*old)->q.qlen);
qdisc_reset(*old);
sch_tree_unlock(sch);
return 0;
}
static struct Qdisc *red_leaf(struct Qdisc *sch, unsigned long arg)
{
struct red_sched_data *q = qdisc_priv(sch);
return q->qdisc;
}
static unsigned long red_get(struct Qdisc *sch, u32 classid)
{
return 1;
}
static void red_put(struct Qdisc *sch, unsigned long arg)
{
return;
}
static int red_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
struct nlattr **tca, unsigned long *arg)
{
return -ENOSYS;
}
static int red_delete(struct Qdisc *sch, unsigned long cl)
{
return -ENOSYS;
}
static void red_walk(struct Qdisc *sch, struct qdisc_walker *walker)
{
if (!walker->stop) {
if (walker->count >= walker->skip)
if (walker->fn(sch, 1, walker) < 0) {
walker->stop = 1;
return;
}
walker->count++;
}
}
static struct tcf_proto **red_find_tcf(struct Qdisc *sch, unsigned long cl)
{
return NULL;
}
static const struct Qdisc_class_ops red_class_ops = {
.graft = red_graft,
.leaf = red_leaf,
.get = red_get,
.put = red_put,
.change = red_change_class,
.delete = red_delete,
.walk = red_walk,
.tcf_chain = red_find_tcf,
.dump = red_dump_class,
};
static struct Qdisc_ops red_qdisc_ops __read_mostly = {
.id = "red",
.priv_size = sizeof(struct red_sched_data),
.cl_ops = &red_class_ops,
.enqueue = red_enqueue,
.dequeue = red_dequeue,
.requeue = red_requeue,
.drop = red_drop,
.init = red_init,
.reset = red_reset,
.destroy = red_destroy,
.change = red_change,
.dump = red_dump,
.dump_stats = red_dump_stats,
.owner = THIS_MODULE,
};
static int __init red_module_init(void)
{
return register_qdisc(&red_qdisc_ops);
}
static void __exit red_module_exit(void)
{
unregister_qdisc(&red_qdisc_ops);
}
module_init(red_module_init)
module_exit(red_module_exit)
MODULE_LICENSE("GPL");