Author: rrs Date: Tue Jun 16 12:26:23 2020 New Revision: 362225 URL: https://svnweb.freebsd.org/changeset/base/362225
Log: So it turns out rack has a shortcoming in dup-ack counting. It counts the dupacks but then does not properly respond to them. This is because a few missing bits are not present. BBR actually does properly respond (though it also sends a TLP which is interesting and maybe something to fix).. Sponsored by: Netflix Inc. Differential Revision: https://reviews.freebsd.org/D25294 Modified: head/sys/netinet/tcp_stacks/rack.c Modified: head/sys/netinet/tcp_stacks/rack.c ============================================================================== --- head/sys/netinet/tcp_stacks/rack.c Tue Jun 16 12:21:55 2020 (r362224) +++ head/sys/netinet/tcp_stacks/rack.c Tue Jun 16 12:26:23 2020 (r362225) @@ -4588,7 +4588,7 @@ activate_rxt: goto activate_rxt; } /* Convert from ms to usecs */ - if (rsm->r_flags & RACK_SACK_PASSED) { + if ((rsm->r_flags & RACK_SACK_PASSED) || (rsm->r_dupack >= DUP_ACK_THRESHOLD)) { if ((tp->t_flags & TF_SENTFIN) && ((tp->snd_max - tp->snd_una) == 1) && (rsm->r_flags & RACK_HAS_FIN)) { @@ -6237,7 +6237,7 @@ rack_log_output(struct tcpcb *tp, struct tcpopt *to, i * or FIN if seq_out is adding more on and a FIN is present * (and we are not resending). */ - if ((th_flags & TH_SYN) && (seq_out == tp->iss)) + if ((th_flags & TH_SYN) && (seq_out == tp->iss)) len++; if (th_flags & TH_FIN) len++; @@ -8190,6 +8190,7 @@ rack_strike_dupack(struct tcp_rack *rack) rsm->r_dupack++; if (rsm->r_dupack >= DUP_ACK_THRESHOLD) { rack->r_wanted_output = 1; + rack->r_timer_override = 1; rack_log_retran_reason(rack, rsm, __LINE__, 1, 3); } else { rack_log_retran_reason(rack, rsm, __LINE__, 0, 3); @@ -11359,7 +11360,8 @@ check_it: if (rsm->r_flags & RACK_ACKED) { return (NULL); } - if ((rsm->r_flags & RACK_SACK_PASSED) == 0) { + if (((rsm->r_flags & RACK_SACK_PASSED) == 0) && + (rsm->r_dupack < DUP_ACK_THRESHOLD)) { /* Its not yet ready */ return (NULL); } _______________________________________________ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
