/* Queue length based ECN marking: per-queue abd per-port */
void wfq_qlen_marking(struct sk_buff *skb,
struct wfq_sched_data *q,
struct wfq_class *cl)
{
switch (wfq_ecn_scheme)
{
/* Per-queue ECN marking */
case wfq_queue_ecn:
{
if (cl->len_bytes > wfq_queue_thresh_bytes[cl->id])
INET_ECN_set_ce(skb);
break;
}
/* Per-port ECN marking */
case wfq_port_ecn:
{
if (q->sum_len_bytes > wfq_port_thresh_bytes)
INET_ECN_set_ce(skb);
break;
}
default:
{
break;
}
}
}
/* CoDel ECN marking. Borrow from codel_dequeue in Linux kernel */
static void codel_marking(struct sk_buff *skb, struct wfq_class *cl)
{
s64 now_ns = ktime_get_ns();
codel_time_t now = ns_to_codel_time(now_ns);
bool mark = codel_should_mark(skb, cl, now_ns);
if (cl->marking)
{
if (!mark)
{
/* sojourn time below target - leave marking state */
cl->marking = false;
}
else if (codel_time_after_eq(now, cl->mark_next))
{
/* It's time for the next mark */
cl->count++;
codel_Newton_step(cl);
cl->mark_next = codel_control_law(cl->mark_next,
wfq_codel_interval,
cl->rec_inv_sqrt);
INET_ECN_set_ce(skb);
}
}
else if (mark)
{
u32 delta;
INET_ECN_set_ce(skb);
cl->marking = true;
/* if min went above target close to when we last went below it
* assume that the drop rate that controlled the queue on the
* last cycle is a good starting point to control it now.
*/
delta = cl->count - cl->lastcount;
if (delta > 1 &&
codel_time_before(now - cl->mark_next,
(codel_time_t)wfq_codel_interval * 16))
{
cl->count = delta;
/* we dont care if rec_inv_sqrt approximation
* is not very precise :
* Next Newton steps will correct it quadratically.
*/
codel_Newton_step(cl);
}
else
{
cl->count = 1;
cl->rec_inv_sqrt = ~0U >> REC_INV_SQRT_SHIFT;
}
cl->lastcount = cl->count;
cl->mark_next = codel_control_law(now,
wfq_codel_interval,
cl->rec_inv_sqrt);
}
}
static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct pie_sched_data *q = qdisc_priv(sch);
bool enqueue = false;
if (unlikely(qdisc_qlen(sch) >= sch->limit)) {
q->stats.overlimit++;
goto out;
}
if (!drop_early(sch, &q->params, &q->vars, skb->len)) {
enqueue = true;
} else if (q->params.ecn && (q->vars.prob <= MAX_PROB / 10) &&
INET_ECN_set_ce(skb)) {
/* If packet is ecn capable, mark it if drop probability
* is lower than 10%, else drop it.
*/
q->stats.ecn_mark++;
enqueue = true;
}
/* we can enqueue the packet */
if (enqueue) {
q->stats.packets_in++;
if (qdisc_qlen(sch) > q->stats.maxq)
q->stats.maxq = qdisc_qlen(sch);
return qdisc_enqueue_tail(skb, sch);
}
out:
q->stats.dropped++;
return qdisc_drop(skb, sch);
}
/* TCN marking scheme */
static inline void tcn_marking(struct sk_buff *skb)
{
codel_time_t delay;
delay = ns_to_codel_time(ktime_get_ns() - skb->tstamp.tv64);
if (codel_time_after(delay, (codel_time_t)wfq_tcn_thresh))
INET_ECN_set_ce(skb);
}
static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
q->parms.qavg = red_calc_qavg(&q->parms, sch->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;
}
if (sch->qstats.backlog + skb->len <= q->limit)
return qdisc_enqueue_tail(skb, sch);
q->stats.pdrop++;
return qdisc_drop(skb, sch);
congestion_drop:
qdisc_drop(skb, sch);
return NET_XMIT_CN;
}
static int
sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
unsigned int hash;
sfq_index x, qlen;
struct sfq_slot *slot;
int uninitialized_var(ret);
struct sk_buff *head;
int delta;
hash = sfq_classify(skb, sch, &ret);
if (hash == 0) {
if (ret & __NET_XMIT_BYPASS)
sch->qstats.drops++;
kfree_skb(skb);
return ret;
}
hash--;
x = q->ht[hash];
slot = &q->slots[x];
if (x == SFQ_EMPTY_SLOT) {
x = q->dep[0].next; /* get a free slot */
if (x >= SFQ_MAX_FLOWS)
return qdisc_drop(skb, sch);
q->ht[hash] = x;
slot = &q->slots[x];
slot->hash = hash;
slot->backlog = 0; /* should already be 0 anyway... */
red_set_vars(&slot->vars);
goto enqueue;
}
if (q->red_parms) {
slot->vars.qavg = red_calc_qavg_no_idle_time(q->red_parms,
&slot->vars,
slot->backlog);
switch (red_action(q->red_parms,
&slot->vars,
slot->vars.qavg)) {
case RED_DONT_MARK:
break;
case RED_PROB_MARK:
sch->qstats.overlimits++;
if (sfq_prob_mark(q)) {
/* We know we have at least one packet in queue */
if (sfq_headdrop(q) &&
INET_ECN_set_ce(slot->skblist_next)) {
q->stats.prob_mark_head++;
break;
}
if (INET_ECN_set_ce(skb)) {
q->stats.prob_mark++;
break;
}
}
q->stats.prob_drop++;
goto congestion_drop;
case RED_HARD_MARK:
sch->qstats.overlimits++;
if (sfq_hard_mark(q)) {
/* We know we have at least one packet in queue */
if (sfq_headdrop(q) &&
INET_ECN_set_ce(slot->skblist_next)) {
q->stats.forced_mark_head++;
break;
}
if (INET_ECN_set_ce(skb)) {
q->stats.forced_mark++;
break;
}
}
q->stats.forced_drop++;
goto congestion_drop;
}
}
if (slot->qlen >= q->maxdepth) {
congestion_drop:
if (!sfq_headdrop(q))
return qdisc_drop(skb, sch);
/* We know we have at least one packet in queue */
head = slot_dequeue_head(slot);
delta = qdisc_pkt_len(head) - qdisc_pkt_len(skb);
sch->qstats.backlog -= delta;
slot->backlog -= delta;
qdisc_drop(head, sch);
slot_queue_add(slot, skb);
return NET_XMIT_CN;
}
enqueue:
sch->qstats.backlog += qdisc_pkt_len(skb);
slot->backlog += qdisc_pkt_len(skb);
slot_queue_add(slot, skb);
sfq_inc(q, x);
if (slot->qlen == 1) { /* The flow is new */
if (q->tail == NULL) { /* It is the first flow */
slot->next = x;
} else {
slot->next = q->tail->next;
q->tail->next = x;
}
/* We put this flow at the end of our flow list.
* This might sound unfair for a new flow to wait after old ones,
* but we could endup servicing new flows only, and freeze old ones.
*/
q->tail = slot;
/* We could use a bigger initial quantum for new flows */
slot->allot = q->scaled_quantum;
}
if (++sch->q.qlen <= q->limit)
return NET_XMIT_SUCCESS;
qlen = slot->qlen;
sfq_drop(sch);
/* Return Congestion Notification only if we dropped a packet
* from this flow.
*/
if (qlen != slot->qlen)
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;
}
static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct choke_sched_data *q = qdisc_priv(sch);
const struct red_parms *p = &q->parms;
int ret = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
if (q->filter_list) {
/* If using external classifiers, get result and record it. */
if (!choke_classify(skb, sch, &ret))
goto other_drop; /* Packet was eaten by filter */
}
choke_skb_cb(skb)->keys_valid = 0;
/* Compute average queue usage (see RED) */
q->vars.qavg = red_calc_qavg(p, &q->vars, sch->q.qlen);
if (red_is_idling(&q->vars))
red_end_of_idle_period(&q->vars);
/* Is queue small? */
if (q->vars.qavg <= p->qth_min)
q->vars.qcount = -1;
else {
unsigned int idx;
/* Draw a packet at random from queue and compare flow */
if (choke_match_random(q, skb, &idx)) {
q->stats.matched++;
choke_drop_by_idx(sch, idx);
goto congestion_drop;
}
/* Queue is large, always mark/drop */
if (q->vars.qavg > p->qth_max) {
q->vars.qcount = -1;
sch->qstats.overlimits++;
if (use_harddrop(q) || !use_ecn(q) ||
!INET_ECN_set_ce(skb)) {
q->stats.forced_drop++;
goto congestion_drop;
}
q->stats.forced_mark++;
} else if (++q->vars.qcount) {
if (red_mark_probability(p, &q->vars, q->vars.qavg)) {
q->vars.qcount = 0;
q->vars.qR = red_random(p);
sch->qstats.overlimits++;
if (!use_ecn(q) || !INET_ECN_set_ce(skb)) {
q->stats.prob_drop++;
goto congestion_drop;
}
q->stats.prob_mark++;
}
} else
q->vars.qR = red_random(p);
}
/* Admit new packet */
if (sch->q.qlen < q->limit) {
q->tab[q->tail] = skb;
q->tail = (q->tail + 1) & q->tab_mask;
++sch->q.qlen;
sch->qstats.backlog += qdisc_pkt_len(skb);
return NET_XMIT_SUCCESS;
}
q->stats.pdrop++;
sch->qstats.drops++;
kfree_skb(skb);
return NET_XMIT_DROP;
congestion_drop:
qdisc_drop(skb, sch);
return NET_XMIT_CN;
other_drop:
if (ret & __NET_XMIT_BYPASS)
sch->qstats.drops++;
kfree_skb(skb);
return ret;
}
static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch,
struct sk_buff **to_free)
{
struct choke_sched_data *q = qdisc_priv(sch);
const struct red_parms *p = &q->parms;
choke_skb_cb(skb)->keys_valid = 0;
/* Compute average queue usage (see RED) */
q->vars.qavg = red_calc_qavg(p, &q->vars, sch->q.qlen);
if (red_is_idling(&q->vars))
red_end_of_idle_period(&q->vars);
/* Is queue small? */
if (q->vars.qavg <= p->qth_min)
q->vars.qcount = -1;
else {
unsigned int idx;
/* Draw a packet at random from queue and compare flow */
if (choke_match_random(q, skb, &idx)) {
q->stats.matched++;
choke_drop_by_idx(sch, idx, to_free);
goto congestion_drop;
}
/* Queue is large, always mark/drop */
if (q->vars.qavg > p->qth_max) {
q->vars.qcount = -1;
qdisc_qstats_overlimit(sch);
if (use_harddrop(q) || !use_ecn(q) ||
!INET_ECN_set_ce(skb)) {
q->stats.forced_drop++;
goto congestion_drop;
}
q->stats.forced_mark++;
} else if (++q->vars.qcount) {
if (red_mark_probability(p, &q->vars, q->vars.qavg)) {
q->vars.qcount = 0;
q->vars.qR = red_random(p);
qdisc_qstats_overlimit(sch);
if (!use_ecn(q) || !INET_ECN_set_ce(skb)) {
q->stats.prob_drop++;
goto congestion_drop;
}
q->stats.prob_mark++;
}
} else
q->vars.qR = red_random(p);
}
/* Admit new packet */
if (sch->q.qlen < q->limit) {
q->tab[q->tail] = skb;
q->tail = (q->tail + 1) & q->tab_mask;
++sch->q.qlen;
qdisc_qstats_backlog_inc(sch, skb);
return NET_XMIT_SUCCESS;
}
q->stats.pdrop++;
return qdisc_drop(skb, sch, to_free);
congestion_drop:
qdisc_drop(skb, sch, to_free);
return NET_XMIT_CN;
}
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
sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
unsigned int hash;
sfq_index x, qlen;
struct sfq_slot *slot;
int uninitialized_var(ret);
struct sk_buff *head;
int delta;
hash = sfq_classify(skb, sch, &ret);
if (hash == 0) {
if (ret & __NET_XMIT_BYPASS)
sch->qstats.drops++;
kfree_skb(skb);
return ret;
}
hash--;
x = q->ht[hash];
slot = &q->slots[x];
if (x == SFQ_EMPTY_SLOT) {
x = q->dep[0].next; /* */
if (x >= SFQ_MAX_FLOWS)
return qdisc_drop(skb, sch);
q->ht[hash] = x;
slot = &q->slots[x];
slot->hash = hash;
slot->backlog = 0; /* */
red_set_vars(&slot->vars);
goto enqueue;
}
if (q->red_parms) {
slot->vars.qavg = red_calc_qavg_no_idle_time(q->red_parms,
&slot->vars,
slot->backlog);
switch (red_action(q->red_parms,
&slot->vars,
slot->vars.qavg)) {
case RED_DONT_MARK:
break;
case RED_PROB_MARK:
sch->qstats.overlimits++;
if (sfq_prob_mark(q)) {
/* */
if (sfq_headdrop(q) &&
INET_ECN_set_ce(slot->skblist_next)) {
q->stats.prob_mark_head++;
break;
}
if (INET_ECN_set_ce(skb)) {
q->stats.prob_mark++;
break;
}
}
q->stats.prob_drop++;
goto congestion_drop;
case RED_HARD_MARK:
sch->qstats.overlimits++;
if (sfq_hard_mark(q)) {
/* */
if (sfq_headdrop(q) &&
INET_ECN_set_ce(slot->skblist_next)) {
q->stats.forced_mark_head++;
break;
}
if (INET_ECN_set_ce(skb)) {
q->stats.forced_mark++;
break;
}
}
q->stats.forced_drop++;
goto congestion_drop;
}
}
if (slot->qlen >= q->maxdepth) {
congestion_drop:
if (!sfq_headdrop(q))
return qdisc_drop(skb, sch);
/* */
head = slot_dequeue_head(slot);
delta = qdisc_pkt_len(head) - qdisc_pkt_len(skb);
sch->qstats.backlog -= delta;
slot->backlog -= delta;
qdisc_drop(head, sch);
slot_queue_add(slot, skb);
return NET_XMIT_CN;
}
enqueue:
sch->qstats.backlog += qdisc_pkt_len(skb);
slot->backlog += qdisc_pkt_len(skb);
slot_queue_add(slot, skb);
sfq_inc(q, x);
if (slot->qlen == 1) { /* */
if (q->tail == NULL) { /* */
slot->next = x;
} else {
slot->next = q->tail->next;
q->tail->next = x;
}
/*
*/
q->tail = slot;
/* */
slot->allot = q->scaled_quantum;
}
if (++sch->q.qlen <= q->limit)
return NET_XMIT_SUCCESS;
qlen = slot->qlen;
sfq_drop(sch);
/*
*/
if (qlen != slot->qlen)
return NET_XMIT_CN;
/* */
qdisc_tree_decrease_qlen(sch, 1);
return NET_XMIT_SUCCESS;
}
