Searched refs:losses (Results 1 - 5 of 5) sorted by relevance
/linux-master/net/ipv4/ |
H A D | tcp_rate.c | 135 rs->losses = lost; /* freshly marked lost */ 171 * for connections suffer heavy or prolonged losses.
|
H A D | tcp_bbr.c | 12 * The core algorithm does not react directly to packet losses or delays, 320 /* Save "last known good" cwnd so we can restore it after losses or PROBE_RTT */ 472 /* An optimization in BBR to reduce losses: On the first round of recovery, we 492 if (rs->losses > 0) 493 cwnd = max_t(s32, cwnd - rs->losses, 1); 580 (rs->losses || /* perhaps pacing_gain*BDP won't fit */ 683 * explicitly models their policed rate, to reduce unnecessary losses. We 710 if (!rs->losses) 735 if (!rs->losses) 1088 * always reduce cwnd on losses (se [all...] |
H A D | tcp_input.c | 457 * throughput and the higher sensitivity of the connection to losses. 8) 748 * To cope with packet losses, we need a 2x factor 760 /* minimal window to cope with packet losses, assuming 2384 /* Use SACK to deduce losses of new sequences sent during recovery */ 2547 tp->rack.advanced = 1; /* Force RACK to re-exam losses */ 2646 * But when SND_UNA is acked without further losses, 3548 tcp_cwnd_reduction(sk, acked_sacked, rs->losses, flag); 3853 int rexmit = REXMIT_NONE; /* Flag to (re)transmit to recover losses */
|
/linux-master/tools/testing/selftests/bpf/ |
H A D | bpf_tcp_helpers.h | 148 int losses; /* number of packets marked lost upon ACK */ member in struct:rate_sample
|
/linux-master/include/net/ |
H A D | tcp.h | 261 #define TCP_RACK_LOSS_DETECTION 0x1 /* Use RACK to detect losses */ 1134 int losses; /* number of packets marked lost upon ACK */ member in struct:rate_sample
|
Completed in 172 milliseconds