static unsigned int gred_drop(struct Qdisc *sch)
{
struct sk_buff *skb;
struct gred_sched *t = qdisc_priv(sch);
skb = qdisc_dequeue_tail(sch);
if (skb) {
unsigned int len = qdisc_pkt_len(skb);
struct gred_sched_data *q;
u16 dp = tc_index_to_dp(skb);
if (dp >= t->DPs || (q = t->tab[dp]) == NULL) {
if (net_ratelimit())
pr_warning("GRED: Unable to relocate VQ 0x%x "
"while dropping, screwing up "
"backlog.\n", tc_index_to_dp(skb));
} else {
q->backlog -= len;
q->stats.other++;
if (!q->backlog && !gred_wred_mode(t))
red_start_of_idle_period(&q->vars);
}
qdisc_drop(skb, sch);
return len;
}
if (gred_wred_mode(t) && !red_is_idling(&t->wred_set))
red_start_of_idle_period(&t->wred_set);
return 0;
}
static struct sk_buff *gred_dequeue(struct Qdisc *sch)
{
struct sk_buff *skb;
struct gred_sched *t = qdisc_priv(sch);
skb = qdisc_dequeue_head(sch);
if (skb) {
struct gred_sched_data *q;
u16 dp = tc_index_to_dp(skb);
if (dp >= t->DPs || (q = t->tab[dp]) == NULL) {
if (net_ratelimit())
pr_warning("GRED: Unable to relocate VQ 0x%x "
"after dequeue, screwing up "
"backlog.\n", tc_index_to_dp(skb));
} else {
q->backlog -= qdisc_pkt_len(skb);
if (!q->backlog && !gred_wred_mode(t))
red_start_of_idle_period(&q->vars);
}
return skb;
}
if (gred_wred_mode(t) && !red_is_idling(&t->wred_set))
red_start_of_idle_period(&t->wred_set);
return NULL;
}
static struct sk_buff *gred_dequeue(struct Qdisc *sch)
{
struct sk_buff *skb;
struct gred_sched *t = qdisc_priv(sch);
skb = qdisc_dequeue_head(sch);
if (skb) {
struct gred_sched_data *q;
u16 dp = tc_index_to_dp(skb);
if (dp >= t->DPs || (q = t->tab[dp]) == NULL) {
net_warn_ratelimited("GRED: Unable to relocate VQ 0x%x after dequeue, screwing up backlog\n",
tc_index_to_dp(skb));
} else {
q->backlog -= qdisc_pkt_len(skb);
if (gred_wred_mode(t)) {
if (!sch->qstats.backlog)
red_start_of_idle_period(&t->wred_set);
} else {
if (!q->backlog)
red_start_of_idle_period(&q->vars);
}
}
return skb;
}
return NULL;
}
static inline unsigned int gred_backlog(struct gred_sched *table,
struct gred_sched_data *q,
struct Qdisc *sch)
{
if (gred_wred_mode(table))
return sch->qstats.backlog;
else
return q->backlog;
}
static void gred_offload(struct Qdisc *sch, enum tc_gred_command command)
{
struct gred_sched *table = qdisc_priv(sch);
struct net_device *dev = qdisc_dev(sch);
struct tc_gred_qopt_offload opt = {
.command = command,
.handle = sch->handle,
.parent = sch->parent,
};
if (!tc_can_offload(dev) || !dev->netdev_ops->ndo_setup_tc)
return;
if (command == TC_GRED_REPLACE) {
unsigned int i;
opt.set.grio_on = gred_rio_mode(table);
opt.set.wred_on = gred_wred_mode(table);
opt.set.dp_cnt = table->DPs;
opt.set.dp_def = table->def;
for (i = 0; i < table->DPs; i++) {
struct gred_sched_data *q = table->tab[i];
if (!q)
continue;
opt.set.tab[i].present = true;
opt.set.tab[i].limit = q->limit;
opt.set.tab[i].prio = q->prio;
opt.set.tab[i].min = q->parms.qth_min >> q->parms.Wlog;
opt.set.tab[i].max = q->parms.qth_max >> q->parms.Wlog;
opt.set.tab[i].is_ecn = gred_use_ecn(q);
opt.set.tab[i].is_harddrop = gred_use_harddrop(q);
opt.set.tab[i].probability = q->parms.max_P;
opt.set.tab[i].backlog = &q->backlog;
}
opt.set.qstats = &sch->qstats;
}
dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_QDISC_GRED, &opt);
}
static int gred_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct gred_sched *table = qdisc_priv(sch);
struct nlattr *parms, *vqs, *opts = NULL;
int i;
u32 max_p[MAX_DPs];
struct tc_gred_sopt sopt = {
.DPs = table->DPs,
.def_DP = table->def,
.grio = gred_rio_mode(table),
.flags = table->red_flags,
};
if (gred_offload_dump_stats(sch))
goto nla_put_failure;
opts = nla_nest_start_noflag(skb, TCA_OPTIONS);
if (opts == NULL)
goto nla_put_failure;
if (nla_put(skb, TCA_GRED_DPS, sizeof(sopt), &sopt))
goto nla_put_failure;
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
max_p[i] = q ? q->parms.max_P : 0;
}
if (nla_put(skb, TCA_GRED_MAX_P, sizeof(max_p), max_p))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_LIMIT, sch->limit))
goto nla_put_failure;
/* Old style all-in-one dump of VQs */
parms = nla_nest_start_noflag(skb, TCA_GRED_PARMS);
if (parms == NULL)
goto nla_put_failure;
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
struct tc_gred_qopt opt;
unsigned long qavg;
memset(&opt, 0, sizeof(opt));
if (!q) {
/* hack -- fix at some point with proper message
This is how we indicate to tc that there is no VQ
at this DP */
opt.DP = MAX_DPs + i;
goto append_opt;
}
opt.limit = q->limit;
opt.DP = q->DP;
opt.backlog = gred_backlog(table, q, sch);
opt.prio = q->prio;
opt.qth_min = q->parms.qth_min >> q->parms.Wlog;
opt.qth_max = q->parms.qth_max >> q->parms.Wlog;
opt.Wlog = q->parms.Wlog;
opt.Plog = q->parms.Plog;
opt.Scell_log = q->parms.Scell_log;
opt.other = q->stats.other;
opt.early = q->stats.prob_drop;
opt.forced = q->stats.forced_drop;
opt.pdrop = q->stats.pdrop;
opt.packets = q->packetsin;
opt.bytesin = q->bytesin;
if (gred_wred_mode(table))
gred_load_wred_set(table, q);
qavg = red_calc_qavg(&q->parms, &q->vars,
q->vars.qavg >> q->parms.Wlog);
opt.qave = qavg >> q->parms.Wlog;
append_opt:
if (nla_append(skb, sizeof(opt), &opt) < 0)
goto nla_put_failure;
}
nla_nest_end(skb, parms);
/* Dump the VQs again, in more structured way */
vqs = nla_nest_start_noflag(skb, TCA_GRED_VQ_LIST);
if (!vqs)
goto nla_put_failure;
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
struct nlattr *vq;
if (!q)
continue;
vq = nla_nest_start_noflag(skb, TCA_GRED_VQ_ENTRY);
if (!vq)
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_DP, q->DP))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_FLAGS, q->red_flags))
goto nla_put_failure;
/* Stats */
if (nla_put_u64_64bit(skb, TCA_GRED_VQ_STAT_BYTES, q->bytesin,
TCA_GRED_VQ_PAD))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_PACKETS, q->packetsin))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_BACKLOG,
gred_backlog(table, q, sch)))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_PROB_DROP,
q->stats.prob_drop))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_PROB_MARK,
q->stats.prob_mark))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_FORCED_DROP,
q->stats.forced_drop))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_FORCED_MARK,
q->stats.forced_mark))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_PDROP, q->stats.pdrop))
goto nla_put_failure;
if (nla_put_u32(skb, TCA_GRED_VQ_STAT_OTHER, q->stats.other))
goto nla_put_failure;
nla_nest_end(skb, vq);
}
nla_nest_end(skb, vqs);
return nla_nest_end(skb, opts);
nla_put_failure:
nla_nest_cancel(skb, opts);
return -EMSGSIZE;
}
static void gred_destroy(struct Qdisc *sch)
{
struct gred_sched *table = qdisc_priv(sch);
int i;
for (i = 0; i < table->DPs; i++) {
if (table->tab[i])
gred_destroy_vq(table->tab[i]);
}
gred_offload(sch, TC_GRED_DESTROY);
}
static struct Qdisc_ops gred_qdisc_ops __read_mostly = {
.id = "gred",
.priv_size = sizeof(struct gred_sched),
.enqueue = gred_enqueue,
.dequeue = gred_dequeue,
.peek = qdisc_peek_head,
.init = gred_init,
.reset = gred_reset,
.destroy = gred_destroy,
.change = gred_change,
.dump = gred_dump,
.owner = THIS_MODULE,
};
static int __init gred_module_init(void)
{
return register_qdisc(&gred_qdisc_ops);
}
static void __exit gred_module_exit(void)
{
unregister_qdisc(&gred_qdisc_ops);
}
module_init(gred_module_init)
module_exit(gred_module_exit)
MODULE_LICENSE("GPL");
static int gred_enqueue(struct sk_buff *skb, struct Qdisc *sch,
struct sk_buff **to_free)
{
struct gred_sched_data *q = NULL;
struct gred_sched *t = qdisc_priv(sch);
unsigned long qavg = 0;
u16 dp = tc_index_to_dp(skb);
if (dp >= t->DPs || (q = t->tab[dp]) == NULL) {
dp = t->def;
q = t->tab[dp];
if (!q) {
/* Pass through packets not assigned to a DP
* if no default DP has been configured. This
* allows for DP flows to be left untouched.
*/
if (likely(sch->qstats.backlog + qdisc_pkt_len(skb) <=
sch->limit))
return qdisc_enqueue_tail(skb, sch);
else
goto drop;
}
/* fix tc_index? --could be controversial but needed for
requeueing */
skb->tc_index = (skb->tc_index & ~GRED_VQ_MASK) | dp;
}
/* sum up all the qaves of prios < ours to get the new qave */
if (!gred_wred_mode(t) && gred_rio_mode(t)) {
int i;
for (i = 0; i < t->DPs; i++) {
if (t->tab[i] && t->tab[i]->prio < q->prio &&
!red_is_idling(&t->tab[i]->vars))
qavg += t->tab[i]->vars.qavg;
}
}
q->packetsin++;
q->bytesin += qdisc_pkt_len(skb);
if (gred_wred_mode(t))
gred_load_wred_set(t, q);
q->vars.qavg = red_calc_qavg(&q->parms,
&q->vars,
gred_backlog(t, q, sch));
if (red_is_idling(&q->vars))
red_end_of_idle_period(&q->vars);
if (gred_wred_mode(t))
gred_store_wred_set(t, q);
switch (red_action(&q->parms, &q->vars, q->vars.qavg + qavg)) {
case RED_DONT_MARK:
break;
case RED_PROB_MARK:
qdisc_qstats_overlimit(sch);
if (!gred_use_ecn(q) || !INET_ECN_set_ce(skb)) {
q->stats.prob_drop++;
goto congestion_drop;
}
q->stats.prob_mark++;
break;
case RED_HARD_MARK:
qdisc_qstats_overlimit(sch);
if (gred_use_harddrop(q) || !gred_use_ecn(q) ||
!INET_ECN_set_ce(skb)) {
q->stats.forced_drop++;
goto congestion_drop;
}
q->stats.forced_mark++;
break;
}
if (gred_backlog(t, q, sch) + qdisc_pkt_len(skb) <= q->limit) {
q->backlog += qdisc_pkt_len(skb);
return qdisc_enqueue_tail(skb, sch);
}
q->stats.pdrop++;
drop:
return qdisc_drop(skb, sch, to_free);
congestion_drop:
qdisc_drop(skb, sch, to_free);
return NET_XMIT_CN;
}
static int gred_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct gred_sched *table = qdisc_priv(sch);
struct nlattr *parms, *opts = NULL;
int i;
u32 max_p[MAX_DPs];
struct tc_gred_sopt sopt = {
.DPs = table->DPs,
.def_DP = table->def,
.grio = gred_rio_mode(table),
.flags = table->red_flags,
};
opts = nla_nest_start(skb, TCA_OPTIONS);
if (opts == NULL)
goto nla_put_failure;
nla_put(skb, TCA_GRED_DPS, sizeof(sopt), &sopt);
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
max_p[i] = q ? q->parms.max_P : 0;
}
nla_put(skb, TCA_GRED_MAX_P, sizeof(max_p), max_p);
parms = nla_nest_start(skb, TCA_GRED_PARMS);
if (parms == NULL)
goto nla_put_failure;
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
struct tc_gred_qopt opt;
memset(&opt, 0, sizeof(opt));
if (!q) {
/* hack -- fix at some point with proper message
This is how we indicate to tc that there is no VQ
at this DP */
opt.DP = MAX_DPs + i;
goto append_opt;
}
opt.limit = q->limit;
opt.DP = q->DP;
opt.backlog = q->backlog;
opt.prio = q->prio;
opt.qth_min = q->parms.qth_min >> q->parms.Wlog;
opt.qth_max = q->parms.qth_max >> q->parms.Wlog;
opt.Wlog = q->parms.Wlog;
opt.Plog = q->parms.Plog;
opt.Scell_log = q->parms.Scell_log;
opt.other = q->stats.other;
opt.early = q->stats.prob_drop;
opt.forced = q->stats.forced_drop;
opt.pdrop = q->stats.pdrop;
opt.packets = q->packetsin;
opt.bytesin = q->bytesin;
if (gred_wred_mode(table))
gred_load_wred_set(table, q);
opt.qave = red_calc_qavg(&q->parms, &q->vars, q->vars.qavg);
append_opt:
if (nla_append(skb, sizeof(opt), &opt) < 0)
goto nla_put_failure;
}
nla_nest_end(skb, parms);
return nla_nest_end(skb, opts);
nla_put_failure:
nla_nest_cancel(skb, opts);
return -EMSGSIZE;
}
static void gred_destroy(struct Qdisc *sch)
{
struct gred_sched *table = qdisc_priv(sch);
int i;
for (i = 0; i < table->DPs; i++) {
if (table->tab[i])
gred_destroy_vq(table->tab[i]);
}
}
static struct Qdisc_ops gred_qdisc_ops __read_mostly = {
.id = "gred",
.priv_size = sizeof(struct gred_sched),
.enqueue = gred_enqueue,
.dequeue = gred_dequeue,
.peek = qdisc_peek_head,
.drop = gred_drop,
.init = gred_init,
.reset = gred_reset,
.destroy = gred_destroy,
.change = gred_change,
.dump = gred_dump,
.owner = THIS_MODULE,
};
static int __init gred_module_init(void)
{
return register_qdisc(&gred_qdisc_ops);
}
static void __exit gred_module_exit(void)
{
unregister_qdisc(&gred_qdisc_ops);
}
module_init(gred_module_init)
module_exit(gred_module_exit)
MODULE_LICENSE("GPL");
static int gred_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct gred_sched *table = qdisc_priv(sch);
struct rtattr *parms, *opts = NULL;
int i;
struct tc_gred_sopt sopt = {
.DPs = table->DPs,
.def_DP = table->def,
.grio = gred_rio_mode(table),
.flags = table->red_flags,
};
opts = RTA_NEST(skb, TCA_OPTIONS);
RTA_PUT(skb, TCA_GRED_DPS, sizeof(sopt), &sopt);
parms = RTA_NEST(skb, TCA_GRED_PARMS);
for (i = 0; i < MAX_DPs; i++) {
struct gred_sched_data *q = table->tab[i];
struct tc_gred_qopt opt;
memset(&opt, 0, sizeof(opt));
if (!q) {
/* hack -- fix at some point with proper message
This is how we indicate to tc that there is no VQ
at this DP */
opt.DP = MAX_DPs + i;
goto append_opt;
}
opt.limit = q->limit;
opt.DP = q->DP;
opt.backlog = q->backlog;
opt.prio = q->prio;
opt.qth_min = q->parms.qth_min >> q->parms.Wlog;
opt.qth_max = q->parms.qth_max >> q->parms.Wlog;
opt.Wlog = q->parms.Wlog;
opt.Plog = q->parms.Plog;
opt.Scell_log = q->parms.Scell_log;
opt.other = q->stats.other;
opt.early = q->stats.prob_drop;
opt.forced = q->stats.forced_drop;
opt.pdrop = q->stats.pdrop;
opt.packets = q->packetsin;
opt.bytesin = q->bytesin;
if (gred_wred_mode(table)) {
q->parms.qidlestart =
table->tab[table->def]->parms.qidlestart;
q->parms.qavg = table->tab[table->def]->parms.qavg;
}
opt.qave = red_calc_qavg(&q->parms, q->parms.qavg);
append_opt:
RTA_APPEND(skb, sizeof(opt), &opt);
}
RTA_NEST_END(skb, parms);
return RTA_NEST_END(skb, opts);
rtattr_failure:
return RTA_NEST_CANCEL(skb, opts);
}
static void gred_destroy(struct Qdisc *sch)
{
struct gred_sched *table = qdisc_priv(sch);
int i;
for (i = 0; i < table->DPs; i++) {
if (table->tab[i])
gred_destroy_vq(table->tab[i]);
}
}
static struct Qdisc_ops gred_qdisc_ops = {
.id = "gred",
.priv_size = sizeof(struct gred_sched),
.enqueue = gred_enqueue,
.dequeue = gred_dequeue,
.requeue = gred_requeue,
.drop = gred_drop,
.init = gred_init,
.reset = gred_reset,
.destroy = gred_destroy,
.change = gred_change,
.dump = gred_dump,
.owner = THIS_MODULE,
};
static int __init gred_module_init(void)
{
return register_qdisc(&gred_qdisc_ops);
}
static void __exit gred_module_exit(void)
{
unregister_qdisc(&gred_qdisc_ops);
}
module_init(gred_module_init)
module_exit(gred_module_exit)
MODULE_LICENSE("GPL");