Commit c6d14ff1 authored by Jiri Pirko's avatar Jiri Pirko Committed by David S. Miller

act_police: improved accuracy at high rates

Current act_police uses rate table computed by the "tc" userspace
program, which has the following issue:

The rate table has 256 entries to map packet lengths to token (time
units).  With TSO sized packets, the 256 entry granularity leads to
loss/gain of rate, making the token bucket inaccurate.

Thus, instead of relying on rate table, this patch explicitly computes
the time and accounts for packet transmission times with nanosecond
granularity.

This is a followup to 56b765b7
("htb: improved accuracy at high rates").
Signed-off-by: default avatarJiri Pirko <jiri@resnulli.us>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 0e243218
...@@ -26,20 +26,20 @@ struct tcf_police { ...@@ -26,20 +26,20 @@ struct tcf_police {
struct tcf_common common; struct tcf_common common;
int tcfp_result; int tcfp_result;
u32 tcfp_ewma_rate; u32 tcfp_ewma_rate;
u32 tcfp_burst; s64 tcfp_burst;
u32 tcfp_mtu; u32 tcfp_mtu;
u32 tcfp_toks; s64 tcfp_toks;
u32 tcfp_ptoks; s64 tcfp_ptoks;
psched_time_t tcfp_t_c; s64 tcfp_mtu_ptoks;
struct qdisc_rate_table *tcfp_R_tab; s64 tcfp_t_c;
struct qdisc_rate_table *tcfp_P_tab; struct psched_ratecfg rate;
bool rate_present;
struct psched_ratecfg peak;
bool peak_present;
}; };
#define to_police(pc) \ #define to_police(pc) \
container_of(pc, struct tcf_police, common) container_of(pc, struct tcf_police, common)
#define L2T(p, L) qdisc_l2t((p)->tcfp_R_tab, L)
#define L2T_P(p, L) qdisc_l2t((p)->tcfp_P_tab, L)
#define POL_TAB_MASK 15 #define POL_TAB_MASK 15
static struct tcf_common *tcf_police_ht[POL_TAB_MASK + 1]; static struct tcf_common *tcf_police_ht[POL_TAB_MASK + 1];
static u32 police_idx_gen; static u32 police_idx_gen;
...@@ -123,10 +123,6 @@ static void tcf_police_destroy(struct tcf_police *p) ...@@ -123,10 +123,6 @@ static void tcf_police_destroy(struct tcf_police *p)
write_unlock_bh(&police_lock); write_unlock_bh(&police_lock);
gen_kill_estimator(&p->tcf_bstats, gen_kill_estimator(&p->tcf_bstats,
&p->tcf_rate_est); &p->tcf_rate_est);
if (p->tcfp_R_tab)
qdisc_put_rtab(p->tcfp_R_tab);
if (p->tcfp_P_tab)
qdisc_put_rtab(p->tcfp_P_tab);
/* /*
* gen_estimator est_timer() might access p->tcf_lock * gen_estimator est_timer() might access p->tcf_lock
* or bstats, wait a RCU grace period before freeing p * or bstats, wait a RCU grace period before freeing p
...@@ -227,26 +223,36 @@ static int tcf_act_police_locate(struct net *net, struct nlattr *nla, ...@@ -227,26 +223,36 @@ static int tcf_act_police_locate(struct net *net, struct nlattr *nla,
} }
/* No failure allowed after this point */ /* No failure allowed after this point */
if (R_tab != NULL) { police->tcfp_mtu = parm->mtu;
qdisc_put_rtab(police->tcfp_R_tab); if (police->tcfp_mtu == 0) {
police->tcfp_R_tab = R_tab; police->tcfp_mtu = ~0;
if (R_tab)
police->tcfp_mtu = 255 << R_tab->rate.cell_log;
}
if (R_tab) {
police->rate_present = true;
psched_ratecfg_precompute(&police->rate, R_tab->rate.rate);
qdisc_put_rtab(R_tab);
} else {
police->rate_present = false;
} }
if (P_tab != NULL) { if (P_tab) {
qdisc_put_rtab(police->tcfp_P_tab); police->peak_present = true;
police->tcfp_P_tab = P_tab; psched_ratecfg_precompute(&police->peak, P_tab->rate.rate);
qdisc_put_rtab(P_tab);
} else {
police->peak_present = false;
} }
if (tb[TCA_POLICE_RESULT]) if (tb[TCA_POLICE_RESULT])
police->tcfp_result = nla_get_u32(tb[TCA_POLICE_RESULT]); police->tcfp_result = nla_get_u32(tb[TCA_POLICE_RESULT]);
police->tcfp_toks = police->tcfp_burst = parm->burst; police->tcfp_burst = PSCHED_TICKS2NS(parm->burst);
police->tcfp_mtu = parm->mtu; police->tcfp_toks = police->tcfp_burst;
if (police->tcfp_mtu == 0) { if (police->peak_present) {
police->tcfp_mtu = ~0; police->tcfp_mtu_ptoks = (s64) psched_l2t_ns(&police->peak,
if (police->tcfp_R_tab) police->tcfp_mtu);
police->tcfp_mtu = 255<<police->tcfp_R_tab->rate.cell_log; police->tcfp_ptoks = police->tcfp_mtu_ptoks;
} }
if (police->tcfp_P_tab)
police->tcfp_ptoks = L2T_P(police, police->tcfp_mtu);
police->tcf_action = parm->action; police->tcf_action = parm->action;
if (tb[TCA_POLICE_AVRATE]) if (tb[TCA_POLICE_AVRATE])
...@@ -256,7 +262,7 @@ static int tcf_act_police_locate(struct net *net, struct nlattr *nla, ...@@ -256,7 +262,7 @@ static int tcf_act_police_locate(struct net *net, struct nlattr *nla,
if (ret != ACT_P_CREATED) if (ret != ACT_P_CREATED)
return ret; return ret;
police->tcfp_t_c = psched_get_time(); police->tcfp_t_c = ktime_to_ns(ktime_get());
police->tcf_index = parm->index ? parm->index : police->tcf_index = parm->index ? parm->index :
tcf_hash_new_index(&police_idx_gen, &police_hash_info); tcf_hash_new_index(&police_idx_gen, &police_hash_info);
h = tcf_hash(police->tcf_index, POL_TAB_MASK); h = tcf_hash(police->tcf_index, POL_TAB_MASK);
...@@ -302,9 +308,9 @@ static int tcf_act_police(struct sk_buff *skb, const struct tc_action *a, ...@@ -302,9 +308,9 @@ static int tcf_act_police(struct sk_buff *skb, const struct tc_action *a,
struct tcf_result *res) struct tcf_result *res)
{ {
struct tcf_police *police = a->priv; struct tcf_police *police = a->priv;
psched_time_t now; s64 now;
long toks; s64 toks;
long ptoks = 0; s64 ptoks = 0;
spin_lock(&police->tcf_lock); spin_lock(&police->tcf_lock);
...@@ -320,24 +326,25 @@ static int tcf_act_police(struct sk_buff *skb, const struct tc_action *a, ...@@ -320,24 +326,25 @@ static int tcf_act_police(struct sk_buff *skb, const struct tc_action *a,
} }
if (qdisc_pkt_len(skb) <= police->tcfp_mtu) { if (qdisc_pkt_len(skb) <= police->tcfp_mtu) {
if (police->tcfp_R_tab == NULL) { if (!police->rate_present) {
spin_unlock(&police->tcf_lock); spin_unlock(&police->tcf_lock);
return police->tcfp_result; return police->tcfp_result;
} }
now = psched_get_time(); now = ktime_to_ns(ktime_get());
toks = psched_tdiff_bounded(now, police->tcfp_t_c, toks = min_t(s64, now - police->tcfp_t_c,
police->tcfp_burst); police->tcfp_burst);
if (police->tcfp_P_tab) { if (police->peak_present) {
ptoks = toks + police->tcfp_ptoks; ptoks = toks + police->tcfp_ptoks;
if (ptoks > (long)L2T_P(police, police->tcfp_mtu)) if (ptoks > police->tcfp_mtu_ptoks)
ptoks = (long)L2T_P(police, police->tcfp_mtu); ptoks = police->tcfp_mtu_ptoks;
ptoks -= L2T_P(police, qdisc_pkt_len(skb)); ptoks -= (s64) psched_l2t_ns(&police->peak,
qdisc_pkt_len(skb));
} }
toks += police->tcfp_toks; toks += police->tcfp_toks;
if (toks > (long)police->tcfp_burst) if (toks > police->tcfp_burst)
toks = police->tcfp_burst; toks = police->tcfp_burst;
toks -= L2T(police, qdisc_pkt_len(skb)); toks -= (s64) psched_l2t_ns(&police->rate, qdisc_pkt_len(skb));
if ((toks|ptoks) >= 0) { if ((toks|ptoks) >= 0) {
police->tcfp_t_c = now; police->tcfp_t_c = now;
police->tcfp_toks = toks; police->tcfp_toks = toks;
...@@ -363,15 +370,15 @@ tcf_act_police_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref) ...@@ -363,15 +370,15 @@ tcf_act_police_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
.index = police->tcf_index, .index = police->tcf_index,
.action = police->tcf_action, .action = police->tcf_action,
.mtu = police->tcfp_mtu, .mtu = police->tcfp_mtu,
.burst = police->tcfp_burst, .burst = PSCHED_NS2TICKS(police->tcfp_burst),
.refcnt = police->tcf_refcnt - ref, .refcnt = police->tcf_refcnt - ref,
.bindcnt = police->tcf_bindcnt - bind, .bindcnt = police->tcf_bindcnt - bind,
}; };
if (police->tcfp_R_tab) if (police->rate_present)
opt.rate = police->tcfp_R_tab->rate; opt.rate.rate = psched_ratecfg_getrate(&police->rate);
if (police->tcfp_P_tab) if (police->peak_present)
opt.peakrate = police->tcfp_P_tab->rate; opt.peakrate.rate = psched_ratecfg_getrate(&police->peak);
if (nla_put(skb, TCA_POLICE_TBF, sizeof(opt), &opt)) if (nla_put(skb, TCA_POLICE_TBF, sizeof(opt), &opt))
goto nla_put_failure; goto nla_put_failure;
if (police->tcfp_result && if (police->tcfp_result &&
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment