/*
* Update the fraction of marked bytes represented as 'alpha'.
* Also initialize several internal parameters at the end of this function.
*/
static void
dctcp_update_alpha(struct cc_var *ccv)
{
struct dctcp *dctcp_data;
int alpha_prev;
dctcp_data = ccv->cc_data;
alpha_prev = dctcp_data->alpha;
dctcp_data->bytes_total = max(dctcp_data->bytes_total, 1);
/*
* Update alpha: alpha = (1 - g) * alpha + g * F.
* Here:
* g is weight factor
* recommaded to be set to 1/16
* small g = slow convergence between competitive DCTCP flows
* large g = impacts low utilization of bandwidth at switches
* F is fraction of marked segments in last RTT
* updated every RTT
* Alpha must be round to 0 - MAX_ALPHA_VALUE.
*/
dctcp_data->alpha = min(alpha_prev - (alpha_prev >> V_dctcp_shift_g) +
(dctcp_data->bytes_ecn << (10 - V_dctcp_shift_g)) /
dctcp_data->bytes_total, MAX_ALPHA_VALUE);
/* Initialize internal parameters for next alpha calculation */
dctcp_data->bytes_ecn = 0;
dctcp_data->bytes_total = 0;
dctcp_data->save_sndnxt = CCV(ccv, snd_nxt);
}
/*
* Perform any necessary tasks before we exit congestion recovery.
*/
static void
dctcp_post_recovery(struct cc_var *ccv)
{
dctcp_cc_algo.post_recovery = newreno_cc_algo.post_recovery;
if (CCV(ccv, t_flags) & TF_ECN_PERMIT)
dctcp_update_alpha(ccv);
}
/*
* Perform any necessary tasks before we exit congestion recovery.
*/
static void
newreno_post_recovery(struct cc_var *ccv)
{
int pipe;
pipe = 0;
if (IN_FASTRECOVERY(CCV(ccv, t_flags))) {
/*
* Fast recovery will conclude after returning from this
* function. Window inflation should have left us with
* approximately snd_ssthresh outstanding data. But in case we
* would be inclined to send a burst, better to do it via the
* slow start mechanism.
*
* XXXLAS: Find a way to do this without needing curack
*/
if (V_tcp_do_rfc6675_pipe)
pipe = tcp_compute_pipe(ccv->ccvc.tcp);
else
pipe = CCV(ccv, snd_max) - ccv->curack;
if (pipe < CCV(ccv, snd_ssthresh))
CCV(ccv, snd_cwnd) = pipe + CCV(ccv, t_maxseg);
else
CCV(ccv, snd_cwnd) = CCV(ccv, snd_ssthresh);
}
}
static void
dctcp_ecnpkt_handler(struct cc_var *ccv)
{
struct dctcp *dctcp_data;
uint32_t ccflag;
int delay_ack;
dctcp_data = ccv->cc_data;
ccflag = ccv->flags;
delay_ack = 1;
/*
* DCTCP responses an ACK immediately when the CE state
* in between this segment and the last segment is not same.
*/
if (ccflag & CCF_IPHDR_CE) {
if (!dctcp_data->ce_prev && (ccflag & CCF_DELACK))
delay_ack = 0;
dctcp_data->ce_prev = 1;
CCV(ccv, t_flags) |= TF_ECN_SND_ECE;
} else {
if (dctcp_data->ce_prev && (ccflag & CCF_DELACK))
delay_ack = 0;
dctcp_data->ce_prev = 0;
CCV(ccv, t_flags) &= ~TF_ECN_SND_ECE;
}
/* DCTCP sets delayed ack when this segment sets the CWR flag. */
if ((ccflag & CCF_DELACK) && (ccflag & CCF_TCPHDR_CWR))
delay_ack = 1;
if (delay_ack == 0)
ccv->flags |= CCF_ACKNOW;
else
ccv->flags &= ~CCF_ACKNOW;
}
static void
dctcp_after_idle(struct cc_var *ccv)
{
struct dctcp *dctcp_data;
dctcp_data = ccv->cc_data;
/* Initialize internal parameters after idle time */
dctcp_data->bytes_ecn = 0;
dctcp_data->bytes_total = 0;
dctcp_data->save_sndnxt = CCV(ccv, snd_nxt);
dctcp_data->alpha = V_dctcp_alpha;
dctcp_data->ece_curr = 0;
dctcp_data->ece_prev = 0;
dctcp_data->num_cong_events = 0;
dctcp_cc_algo.after_idle = newreno_cc_algo.after_idle;
}
static void
newreno_ack_received(struct cc_var *ccv, uint16_t type)
{
if (type == CC_ACK && !IN_RECOVERY(CCV(ccv, t_flags)) &&
(ccv->flags & CCF_CWND_LIMITED)) {
u_int cw = CCV(ccv, snd_cwnd);
u_int incr = CCV(ccv, t_maxseg);
/*
* Regular in-order ACK, open the congestion window.
* Method depends on which congestion control state we're
* in (slow start or cong avoid) and if ABC (RFC 3465) is
* enabled.
*
* slow start: cwnd <= ssthresh
* cong avoid: cwnd > ssthresh
*
* slow start and ABC (RFC 3465):
* Grow cwnd exponentially by the amount of data
* ACKed capping the max increment per ACK to
* (abc_l_var * maxseg) bytes.
*
* slow start without ABC (RFC 5681):
* Grow cwnd exponentially by maxseg per ACK.
*
* cong avoid and ABC (RFC 3465):
* Grow cwnd linearly by maxseg per RTT for each
* cwnd worth of ACKed data.
*
* cong avoid without ABC (RFC 5681):
* Grow cwnd linearly by approximately maxseg per RTT using
* maxseg^2 / cwnd per ACK as the increment.
* If cwnd > maxseg^2, fix the cwnd increment at 1 byte to
* avoid capping cwnd.
*/
if (cw > CCV(ccv, snd_ssthresh)) {
if (V_tcp_do_rfc3465) {
if (ccv->flags & CCF_ABC_SENTAWND)
ccv->flags &= ~CCF_ABC_SENTAWND;
else
incr = 0;
} else
incr = max((incr * incr / cw), 1);
} else if (V_tcp_do_rfc3465) {
/*
* In slow-start with ABC enabled and no RTO in sight?
* (Must not use abc_l_var > 1 if slow starting after
* an RTO. On RTO, snd_nxt = snd_una, so the
* snd_nxt == snd_max check is sufficient to
* handle this).
*
* XXXLAS: Find a way to signal SS after RTO that
* doesn't rely on tcpcb vars.
*/
if (CCV(ccv, snd_nxt) == CCV(ccv, snd_max))
incr = min(ccv->bytes_this_ack,
V_tcp_abc_l_var * CCV(ccv, t_maxseg));
else
incr = min(ccv->bytes_this_ack, CCV(ccv, t_maxseg));
}
/* ABC is on by default, so incr equals 0 frequently. */
if (incr > 0)
CCV(ccv, snd_cwnd) = min(cw + incr,
TCP_MAXWIN << CCV(ccv, snd_scale));
}
}
/*
* Perform any necessary tasks before we enter congestion recovery.
*/
static void
newreno_cong_signal(struct cc_var *ccv, uint32_t type)
{
u_int win;
/* Catch algos which mistakenly leak private signal types. */
KASSERT((type & CC_SIGPRIVMASK) == 0,
("%s: congestion signal type 0x%08x is private\n", __func__, type));
win = max(CCV(ccv, snd_cwnd) / 2 / CCV(ccv, t_maxseg), 2) *
CCV(ccv, t_maxseg);
switch (type) {
case CC_NDUPACK:
if (!IN_FASTRECOVERY(CCV(ccv, t_flags))) {
if (!IN_CONGRECOVERY(CCV(ccv, t_flags)))
CCV(ccv, snd_ssthresh) = win;
ENTER_RECOVERY(CCV(ccv, t_flags));
}
break;
case CC_ECN:
if (!IN_CONGRECOVERY(CCV(ccv, t_flags))) {
CCV(ccv, snd_ssthresh) = win;
CCV(ccv, snd_cwnd) = win;
ENTER_CONGRECOVERY(CCV(ccv, t_flags));
}
break;
}
}
/*
* Perform any necessary tasks before we enter congestion recovery.
*/
static void
newreno_cong_signal(struct cc_var *ccv, uint32_t type)
{
uint32_t cwin, ssthresh_on_loss;
u_int mss;
cwin = CCV(ccv, snd_cwnd);
mss = CCV(ccv, t_maxseg);
ssthresh_on_loss =
max((CCV(ccv, snd_max) - CCV(ccv, snd_una)) / 2 / mss, 2)
* mss;
/* Catch algos which mistakenly leak private signal types. */
KASSERT((type & CC_SIGPRIVMASK) == 0,
("%s: congestion signal type 0x%08x is private\n", __func__, type));
cwin = max(cwin / 2 / mss, 2) * mss;
switch (type) {
case CC_NDUPACK:
if (!IN_FASTRECOVERY(CCV(ccv, t_flags))) {
if (!IN_CONGRECOVERY(CCV(ccv, t_flags))) {
CCV(ccv, snd_ssthresh) = ssthresh_on_loss;
CCV(ccv, snd_cwnd) = cwin;
}
ENTER_RECOVERY(CCV(ccv, t_flags));
}
break;
case CC_ECN:
if (!IN_CONGRECOVERY(CCV(ccv, t_flags))) {
CCV(ccv, snd_ssthresh) = ssthresh_on_loss;
CCV(ccv, snd_cwnd) = cwin;
ENTER_CONGRECOVERY(CCV(ccv, t_flags));
}
break;
case CC_RTO:
CCV(ccv, snd_ssthresh) = ssthresh_on_loss;
CCV(ccv, snd_cwnd) = mss;
break;
}
}
/*
* Perform any necessary tasks before we enter congestion recovery.
*/
static void
dctcp_cong_signal(struct cc_var *ccv, uint32_t type)
{
struct dctcp *dctcp_data;
u_int win, mss;
dctcp_data = ccv->cc_data;
win = CCV(ccv, snd_cwnd);
mss = CCV(ccv, t_maxseg);
switch (type) {
case CC_NDUPACK:
if (!IN_FASTRECOVERY(CCV(ccv, t_flags))) {
if (!IN_CONGRECOVERY(CCV(ccv, t_flags))) {
CCV(ccv, snd_ssthresh) = mss *
max(win / 2 / mss, 2);
dctcp_data->num_cong_events++;
} else {
/* cwnd has already updated as congestion
* recovery. Reverse cwnd value using
* snd_cwnd_prev and recalculate snd_ssthresh
*/
win = CCV(ccv, snd_cwnd_prev);
CCV(ccv, snd_ssthresh) =
max(win / 2 / mss, 2) * mss;
}
ENTER_RECOVERY(CCV(ccv, t_flags));
}
break;
case CC_ECN:
/*
* Save current snd_cwnd when the host encounters both
* congestion recovery and fast recovery.
*/
CCV(ccv, snd_cwnd_prev) = win;
if (!IN_CONGRECOVERY(CCV(ccv, t_flags))) {
if (V_dctcp_slowstart &&
dctcp_data->num_cong_events++ == 0) {
CCV(ccv, snd_ssthresh) =
mss * max(win / 2 / mss, 2);
dctcp_data->alpha = MAX_ALPHA_VALUE;
dctcp_data->bytes_ecn = 0;
dctcp_data->bytes_total = 0;
dctcp_data->save_sndnxt = CCV(ccv, snd_nxt);
} else
CCV(ccv, snd_ssthresh) = max((win - ((win *
dctcp_data->alpha) >> 11)) / mss, 2) * mss;
CCV(ccv, snd_cwnd) = CCV(ccv, snd_ssthresh);
ENTER_CONGRECOVERY(CCV(ccv, t_flags));
}
dctcp_data->ece_curr = 1;
break;
case CC_RTO:
if (CCV(ccv, t_flags) & TF_ECN_PERMIT) {
CCV(ccv, t_flags) |= TF_ECN_SND_CWR;
dctcp_update_alpha(ccv);
dctcp_data->save_sndnxt += CCV(ccv, t_maxseg);
dctcp_data->num_cong_events++;
}
break;
}
}
static void
dctcp_ack_received(struct cc_var *ccv, uint16_t type)
{
struct dctcp *dctcp_data;
int bytes_acked = 0;
dctcp_data = ccv->cc_data;
if (CCV(ccv, t_flags) & TF_ECN_PERMIT) {
/*
* DCTCP doesn't treat receipt of ECN marked packet as a
* congestion event. Thus, DCTCP always executes the ACK
* processing out of congestion recovery.
*/
if (IN_CONGRECOVERY(CCV(ccv, t_flags))) {
EXIT_CONGRECOVERY(CCV(ccv, t_flags));
newreno_cc_algo.ack_received(ccv, type);
ENTER_CONGRECOVERY(CCV(ccv, t_flags));
} else
newreno_cc_algo.ack_received(ccv, type);
if (type == CC_DUPACK)
bytes_acked = CCV(ccv, t_maxseg);
if (type == CC_ACK)
bytes_acked = ccv->bytes_this_ack;
/* Update total bytes. */
dctcp_data->bytes_total += bytes_acked;
/* Update total marked bytes. */
if (dctcp_data->ece_curr) {
if (!dctcp_data->ece_prev
&& bytes_acked > CCV(ccv, t_maxseg)) {
dctcp_data->bytes_ecn +=
(bytes_acked - CCV(ccv, t_maxseg));
} else
dctcp_data->bytes_ecn += bytes_acked;
dctcp_data->ece_prev = 1;
} else {
if (dctcp_data->ece_prev
&& bytes_acked > CCV(ccv, t_maxseg))
dctcp_data->bytes_ecn += CCV(ccv, t_maxseg);
dctcp_data->ece_prev = 0;
}
dctcp_data->ece_curr = 0;
/*
* Update the fraction of marked bytes at the end of
* current window size.
*/
if ((IN_FASTRECOVERY(CCV(ccv, t_flags)) &&
SEQ_GEQ(ccv->curack, CCV(ccv, snd_recover))) ||
(!IN_FASTRECOVERY(CCV(ccv, t_flags)) &&
SEQ_GT(ccv->curack, dctcp_data->save_sndnxt)))
dctcp_update_alpha(ccv);
} else
newreno_cc_algo.ack_received(ccv, type);
}
