^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /* Copyright (C) 2013 Cisco Systems, Inc, 2013.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * Author: Vijay Subramanian <vijaynsu@cisco.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Author: Mythili Prabhu <mysuryan@cisco.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * ECN support is added by Naeem Khademi <naeemk@ifi.uio.no>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * University of Oslo, Norway.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * References:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) * RFC 8033: https://tools.ietf.org/html/rfc8033
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/types.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <linux/kernel.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/errno.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/skbuff.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <net/pkt_sched.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <net/inet_ecn.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include <net/pie.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) /* private data for the Qdisc */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) struct pie_sched_data {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) struct pie_vars vars;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) struct pie_params params;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) struct pie_stats stats;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) struct timer_list adapt_timer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) struct Qdisc *sch;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) bool pie_drop_early(struct Qdisc *sch, struct pie_params *params,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) struct pie_vars *vars, u32 backlog, u32 packet_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) u64 rnd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) u64 local_prob = vars->prob;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) u32 mtu = psched_mtu(qdisc_dev(sch));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) /* If there is still burst allowance left skip random early drop */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) if (vars->burst_time > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) /* If current delay is less than half of target, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) * if drop prob is low already, disable early_drop
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) if ((vars->qdelay < params->target / 2) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) (vars->prob < MAX_PROB / 5))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) /* If we have fewer than 2 mtu-sized packets, disable pie_drop_early,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) * similar to min_th in RED
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) if (backlog < 2 * mtu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) /* If bytemode is turned on, use packet size to compute new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) * probablity. Smaller packets will have lower drop prob in this case
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) if (params->bytemode && packet_size <= mtu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) local_prob = (u64)packet_size * div_u64(local_prob, mtu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) local_prob = vars->prob;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) if (local_prob == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) vars->accu_prob = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) vars->accu_prob += local_prob;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) if (vars->accu_prob < (MAX_PROB / 100) * 85)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) if (vars->accu_prob >= (MAX_PROB / 2) * 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) prandom_bytes(&rnd, 8);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) if ((rnd >> BITS_PER_BYTE) < local_prob) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) vars->accu_prob = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) EXPORT_SYMBOL_GPL(pie_drop_early);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) struct sk_buff **to_free)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) bool enqueue = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) if (unlikely(qdisc_qlen(sch) >= sch->limit)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) q->stats.overlimit++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) if (!pie_drop_early(sch, &q->params, &q->vars, sch->qstats.backlog,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) skb->len)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) enqueue = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) } else if (q->params.ecn && (q->vars.prob <= MAX_PROB / 10) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) INET_ECN_set_ce(skb)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) /* If packet is ecn capable, mark it if drop probability
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) * is lower than 10%, else drop it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) q->stats.ecn_mark++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) enqueue = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) /* we can enqueue the packet */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) if (enqueue) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) /* Set enqueue time only when dq_rate_estimator is disabled. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) if (!q->params.dq_rate_estimator)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) pie_set_enqueue_time(skb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) q->stats.packets_in++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) if (qdisc_qlen(sch) > q->stats.maxq)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) q->stats.maxq = qdisc_qlen(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) return qdisc_enqueue_tail(skb, sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) q->stats.dropped++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) q->vars.accu_prob = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) return qdisc_drop(skb, sch, to_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) static const struct nla_policy pie_policy[TCA_PIE_MAX + 1] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) [TCA_PIE_TARGET] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) [TCA_PIE_LIMIT] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) [TCA_PIE_TUPDATE] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) [TCA_PIE_ALPHA] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) [TCA_PIE_BETA] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) [TCA_PIE_ECN] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) [TCA_PIE_BYTEMODE] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) [TCA_PIE_DQ_RATE_ESTIMATOR] = {.type = NLA_U32},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) static int pie_change(struct Qdisc *sch, struct nlattr *opt,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) struct netlink_ext_ack *extack)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) struct nlattr *tb[TCA_PIE_MAX + 1];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) unsigned int qlen, dropped = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) if (!opt)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) err = nla_parse_nested_deprecated(tb, TCA_PIE_MAX, opt, pie_policy,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) if (err < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) sch_tree_lock(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) /* convert from microseconds to pschedtime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) if (tb[TCA_PIE_TARGET]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) /* target is in us */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) u32 target = nla_get_u32(tb[TCA_PIE_TARGET]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) /* convert to pschedtime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) q->params.target = PSCHED_NS2TICKS((u64)target * NSEC_PER_USEC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) /* tupdate is in jiffies */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) if (tb[TCA_PIE_TUPDATE])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) q->params.tupdate =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) usecs_to_jiffies(nla_get_u32(tb[TCA_PIE_TUPDATE]));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) if (tb[TCA_PIE_LIMIT]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) u32 limit = nla_get_u32(tb[TCA_PIE_LIMIT]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) q->params.limit = limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) sch->limit = limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) if (tb[TCA_PIE_ALPHA])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) q->params.alpha = nla_get_u32(tb[TCA_PIE_ALPHA]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) if (tb[TCA_PIE_BETA])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) q->params.beta = nla_get_u32(tb[TCA_PIE_BETA]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) if (tb[TCA_PIE_ECN])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) q->params.ecn = nla_get_u32(tb[TCA_PIE_ECN]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) if (tb[TCA_PIE_BYTEMODE])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) q->params.bytemode = nla_get_u32(tb[TCA_PIE_BYTEMODE]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) if (tb[TCA_PIE_DQ_RATE_ESTIMATOR])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) q->params.dq_rate_estimator =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) nla_get_u32(tb[TCA_PIE_DQ_RATE_ESTIMATOR]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) /* Drop excess packets if new limit is lower */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) qlen = sch->q.qlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) while (sch->q.qlen > sch->limit) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) struct sk_buff *skb = __qdisc_dequeue_head(&sch->q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) dropped += qdisc_pkt_len(skb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) qdisc_qstats_backlog_dec(sch, skb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) rtnl_qdisc_drop(skb, sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) qdisc_tree_reduce_backlog(sch, qlen - sch->q.qlen, dropped);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) sch_tree_unlock(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) void pie_process_dequeue(struct sk_buff *skb, struct pie_params *params,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) struct pie_vars *vars, u32 backlog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) psched_time_t now = psched_get_time();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) u32 dtime = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) /* If dq_rate_estimator is disabled, calculate qdelay using the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) * packet timestamp.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) if (!params->dq_rate_estimator) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) vars->qdelay = now - pie_get_enqueue_time(skb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) if (vars->dq_tstamp != DTIME_INVALID)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) dtime = now - vars->dq_tstamp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) vars->dq_tstamp = now;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) if (backlog == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) vars->qdelay = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) if (dtime == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) goto burst_allowance_reduction;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) /* If current queue is about 10 packets or more and dq_count is unset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) * we have enough packets to calculate the drain rate. Save
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) * current time as dq_tstamp and start measurement cycle.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) if (backlog >= QUEUE_THRESHOLD && vars->dq_count == DQCOUNT_INVALID) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) vars->dq_tstamp = psched_get_time();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) vars->dq_count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) /* Calculate the average drain rate from this value. If queue length
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) * has receded to a small value viz., <= QUEUE_THRESHOLD bytes, reset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) * the dq_count to -1 as we don't have enough packets to calculate the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) * drain rate anymore. The following if block is entered only when we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) * have a substantial queue built up (QUEUE_THRESHOLD bytes or more)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) * and we calculate the drain rate for the threshold here. dq_count is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) * in bytes, time difference in psched_time, hence rate is in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) * bytes/psched_time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) if (vars->dq_count != DQCOUNT_INVALID) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) vars->dq_count += skb->len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) if (vars->dq_count >= QUEUE_THRESHOLD) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) u32 count = vars->dq_count << PIE_SCALE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) dtime = now - vars->dq_tstamp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) if (dtime == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) count = count / dtime;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) if (vars->avg_dq_rate == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) vars->avg_dq_rate = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) vars->avg_dq_rate =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) (vars->avg_dq_rate -
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) (vars->avg_dq_rate >> 3)) + (count >> 3);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) /* If the queue has receded below the threshold, we hold
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) * on to the last drain rate calculated, else we reset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) * dq_count to 0 to re-enter the if block when the next
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) * packet is dequeued
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) if (backlog < QUEUE_THRESHOLD) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) vars->dq_count = DQCOUNT_INVALID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) vars->dq_count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) vars->dq_tstamp = psched_get_time();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) goto burst_allowance_reduction;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) burst_allowance_reduction:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) if (vars->burst_time > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) if (vars->burst_time > dtime)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) vars->burst_time -= dtime;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) vars->burst_time = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) EXPORT_SYMBOL_GPL(pie_process_dequeue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) void pie_calculate_probability(struct pie_params *params, struct pie_vars *vars,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) u32 backlog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) psched_time_t qdelay = 0; /* in pschedtime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) psched_time_t qdelay_old = 0; /* in pschedtime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) s64 delta = 0; /* determines the change in probability */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) u64 oldprob;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) u64 alpha, beta;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) u32 power;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) bool update_prob = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) if (params->dq_rate_estimator) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) qdelay_old = vars->qdelay;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) vars->qdelay_old = vars->qdelay;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) if (vars->avg_dq_rate > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) qdelay = (backlog << PIE_SCALE) / vars->avg_dq_rate;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) qdelay = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) qdelay = vars->qdelay;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) qdelay_old = vars->qdelay_old;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) /* If qdelay is zero and backlog is not, it means backlog is very small,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) * so we do not update probabilty in this round.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) if (qdelay == 0 && backlog != 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) update_prob = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) /* In the algorithm, alpha and beta are between 0 and 2 with typical
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) * value for alpha as 0.125. In this implementation, we use values 0-32
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) * passed from user space to represent this. Also, alpha and beta have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) * unit of HZ and need to be scaled before they can used to update
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) * probability. alpha/beta are updated locally below by scaling down
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) * by 16 to come to 0-2 range.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) alpha = ((u64)params->alpha * (MAX_PROB / PSCHED_TICKS_PER_SEC)) >> 4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) beta = ((u64)params->beta * (MAX_PROB / PSCHED_TICKS_PER_SEC)) >> 4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) /* We scale alpha and beta differently depending on how heavy the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) * congestion is. Please see RFC 8033 for details.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) if (vars->prob < MAX_PROB / 10) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) alpha >>= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) beta >>= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) power = 100;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) while (vars->prob < div_u64(MAX_PROB, power) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) power <= 1000000) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) alpha >>= 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) beta >>= 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) power *= 10;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) /* alpha and beta should be between 0 and 32, in multiples of 1/16 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) delta += alpha * (qdelay - params->target);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) delta += beta * (qdelay - qdelay_old);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) oldprob = vars->prob;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) /* to ensure we increase probability in steps of no more than 2% */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) if (delta > (s64)(MAX_PROB / (100 / 2)) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) vars->prob >= MAX_PROB / 10)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) delta = (MAX_PROB / 100) * 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) /* Non-linear drop:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) * Tune drop probability to increase quickly for high delays(>= 250ms)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) * 250ms is derived through experiments and provides error protection
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) if (qdelay > (PSCHED_NS2TICKS(250 * NSEC_PER_MSEC)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) delta += MAX_PROB / (100 / 2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) vars->prob += delta;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) if (delta > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) /* prevent overflow */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) if (vars->prob < oldprob) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) vars->prob = MAX_PROB;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) /* Prevent normalization error. If probability is at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) * maximum value already, we normalize it here, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) * skip the check to do a non-linear drop in the next
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) * section.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) update_prob = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) /* prevent underflow */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) if (vars->prob > oldprob)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) vars->prob = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) /* Non-linear drop in probability: Reduce drop probability quickly if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) * delay is 0 for 2 consecutive Tupdate periods.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) if (qdelay == 0 && qdelay_old == 0 && update_prob)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) /* Reduce drop probability to 98.4% */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) vars->prob -= vars->prob / 64;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) vars->qdelay = qdelay;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) vars->backlog_old = backlog;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) /* We restart the measurement cycle if the following conditions are met
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) * 1. If the delay has been low for 2 consecutive Tupdate periods
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) * 2. Calculated drop probability is zero
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) * 3. If average dq_rate_estimator is enabled, we have atleast one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) * estimate for the avg_dq_rate ie., is a non-zero value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) if ((vars->qdelay < params->target / 2) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) (vars->qdelay_old < params->target / 2) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) vars->prob == 0 &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) (!params->dq_rate_estimator || vars->avg_dq_rate > 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) pie_vars_init(vars);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) if (!params->dq_rate_estimator)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) vars->qdelay_old = qdelay;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) EXPORT_SYMBOL_GPL(pie_calculate_probability);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) static void pie_timer(struct timer_list *t)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) struct pie_sched_data *q = from_timer(q, t, adapt_timer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) struct Qdisc *sch = q->sch;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) spinlock_t *root_lock = qdisc_lock(qdisc_root_sleeping(sch));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) spin_lock(root_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) pie_calculate_probability(&q->params, &q->vars, sch->qstats.backlog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) /* reset the timer to fire after 'tupdate'. tupdate is in jiffies. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) if (q->params.tupdate)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) mod_timer(&q->adapt_timer, jiffies + q->params.tupdate);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) spin_unlock(root_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) static int pie_init(struct Qdisc *sch, struct nlattr *opt,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) struct netlink_ext_ack *extack)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) pie_params_init(&q->params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) pie_vars_init(&q->vars);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) sch->limit = q->params.limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) q->sch = sch;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) timer_setup(&q->adapt_timer, pie_timer, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) if (opt) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) int err = pie_change(sch, opt, extack);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) mod_timer(&q->adapt_timer, jiffies + HZ / 2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) static int pie_dump(struct Qdisc *sch, struct sk_buff *skb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) struct nlattr *opts;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) opts = nla_nest_start_noflag(skb, TCA_OPTIONS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) if (!opts)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) /* convert target from pschedtime to us */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) if (nla_put_u32(skb, TCA_PIE_TARGET,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) ((u32)PSCHED_TICKS2NS(q->params.target)) /
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) NSEC_PER_USEC) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) nla_put_u32(skb, TCA_PIE_LIMIT, sch->limit) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) nla_put_u32(skb, TCA_PIE_TUPDATE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) jiffies_to_usecs(q->params.tupdate)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) nla_put_u32(skb, TCA_PIE_ALPHA, q->params.alpha) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) nla_put_u32(skb, TCA_PIE_BETA, q->params.beta) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) nla_put_u32(skb, TCA_PIE_ECN, q->params.ecn) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) nla_put_u32(skb, TCA_PIE_BYTEMODE, q->params.bytemode) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) nla_put_u32(skb, TCA_PIE_DQ_RATE_ESTIMATOR,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) q->params.dq_rate_estimator))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) return nla_nest_end(skb, opts);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) nla_put_failure:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) nla_nest_cancel(skb, opts);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) static int pie_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) struct tc_pie_xstats st = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) .prob = q->vars.prob << BITS_PER_BYTE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) .delay = ((u32)PSCHED_TICKS2NS(q->vars.qdelay)) /
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) NSEC_PER_USEC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) .packets_in = q->stats.packets_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) .overlimit = q->stats.overlimit,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) .maxq = q->stats.maxq,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) .dropped = q->stats.dropped,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) .ecn_mark = q->stats.ecn_mark,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) /* avg_dq_rate is only valid if dq_rate_estimator is enabled */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) st.dq_rate_estimating = q->params.dq_rate_estimator;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) /* unscale and return dq_rate in bytes per sec */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) if (q->params.dq_rate_estimator)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) st.avg_dq_rate = q->vars.avg_dq_rate *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) (PSCHED_TICKS_PER_SEC) >> PIE_SCALE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) return gnet_stats_copy_app(d, &st, sizeof(st));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) static struct sk_buff *pie_qdisc_dequeue(struct Qdisc *sch)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) struct sk_buff *skb = qdisc_dequeue_head(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) if (!skb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) pie_process_dequeue(skb, &q->params, &q->vars, sch->qstats.backlog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) return skb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) static void pie_reset(struct Qdisc *sch)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) qdisc_reset_queue(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) pie_vars_init(&q->vars);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) static void pie_destroy(struct Qdisc *sch)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) struct pie_sched_data *q = qdisc_priv(sch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) q->params.tupdate = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) del_timer_sync(&q->adapt_timer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) static struct Qdisc_ops pie_qdisc_ops __read_mostly = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) .id = "pie",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) .priv_size = sizeof(struct pie_sched_data),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) .enqueue = pie_qdisc_enqueue,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) .dequeue = pie_qdisc_dequeue,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) .peek = qdisc_peek_dequeued,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) .init = pie_init,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) .destroy = pie_destroy,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) .reset = pie_reset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) .change = pie_change,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) .dump = pie_dump,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) .dump_stats = pie_dump_stats,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) .owner = THIS_MODULE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) static int __init pie_module_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) return register_qdisc(&pie_qdisc_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) static void __exit pie_module_exit(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) unregister_qdisc(&pie_qdisc_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) module_init(pie_module_init);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) module_exit(pie_module_exit);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) MODULE_DESCRIPTION("Proportional Integral controller Enhanced (PIE) scheduler");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) MODULE_AUTHOR("Vijay Subramanian");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) MODULE_AUTHOR("Mythili Prabhu");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) MODULE_LICENSE("GPL");