Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Merge branch 'accecn-protocol-case-handling-series'
Chia-Yu Chang says:
====================
AccECN protocol case handling series
Plesae find the v13 AccECN case handling patch series, which covers
several excpetional case handling of Accurate ECN spec (RFC9768),
adds new identifiers to be used by CC modules, adds ecn_delta into
rate_sample, and keeps the ACE counter for computation, etc.
This patch series is part of the full AccECN patch series, which is at
https://github.com/L4STeam/linux-net-next/commits/upstream_l4steam/
---
Chia-Yu Chang (13):
selftests/net: gro: add self-test for TCP CWR flag
tcp: ECT_1_NEGOTIATION and NEEDS_ACCECN identifiers
tcp: disable RFC3168 fallback identifier for CC modules
tcp: accecn: handle unexpected AccECN negotiation feedback
tcp: accecn: retransmit downgraded SYN in AccECN negotiation
tcp: add TCP_SYNACK_RETRANS synack_type
tcp: accecn: retransmit SYN/ACK without AccECN option or non-AccECN
SYN/ACK
tcp: accecn: unset ECT if receive or send ACE=0 in AccECN negotiaion
tcp: accecn: fallback outgoing half link to non-AccECN
tcp: accecn: detect loss ACK w/ AccECN option and add
TCP_ACCECN_OPTION_PERSIST
tcp: accecn: add tcpi_ecn_mode and tcpi_option2 in tcp_info
tcp: accecn: enable AccECN
selftests/net: packetdrill: add TCP Accurate ECN cases
Ilpo Järvinen (2):
tcp: try to avoid safer when ACKs are thinned
gro: flushing when CWR is set negatively affects AccECN
Documentation/networking/ip-sysctl.rst | 4 +-
.../networking/net_cachelines/tcp_sock.rst | 1 +
include/linux/tcp.h | 4 +-
include/net/inet_ecn.h | 20 +++-
include/net/tcp.h | 32 +++++-
include/net/tcp_ecn.h | 103 ++++++++++++------
include/uapi/linux/tcp.h | 26 ++++-
net/ipv4/inet_connection_sock.c | 3 +
net/ipv4/sysctl_net_ipv4.c | 4 +-
net/ipv4/tcp.c | 10 ++
net/ipv4/tcp_cong.c | 5 +-
net/ipv4/tcp_input.c | 40 ++++++-
net/ipv4/tcp_minisocks.c | 43 +++++---
net/ipv4/tcp_offload.c | 3 +-
net/ipv4/tcp_output.c | 34 ++++--
net/ipv4/tcp_timer.c | 3 +
tools/testing/selftests/drivers/net/gro.c | 81 ++++++++++----
tools/testing/selftests/drivers/net/gro.py | 3 +-
.../tcp_accecn_2nd_data_as_first.pkt | 24 ++++
.../tcp_accecn_2nd_data_as_first_connect.pkt | 30 +++++
.../tcp_accecn_3rd_ack_after_synack_rxmt.pkt | 19 ++++
..._accecn_3rd_ack_ce_updates_received_ce.pkt | 18 +++
.../tcp_accecn_3rd_ack_lost_data_ce.pkt | 22 ++++
.../net/packetdrill/tcp_accecn_3rd_dups.pkt | 26 +++++
.../tcp_accecn_acc_ecn_disabled.pkt | 13 +++
.../tcp_accecn_accecn_then_notecn_syn.pkt | 28 +++++
.../tcp_accecn_accecn_to_rfc3168.pkt | 18 +++
.../tcp_accecn_client_accecn_options_drop.pkt | 34 ++++++
.../tcp_accecn_client_accecn_options_lost.pkt | 38 +++++++
.../tcp_accecn_clientside_disabled.pkt | 12 ++
...cecn_close_local_close_then_remote_fin.pkt | 25 +++++
.../tcp_accecn_delivered_2ndlargeack.pkt | 25 +++++
..._accecn_delivered_falseoverflow_detect.pkt | 31 ++++++
.../tcp_accecn_delivered_largeack.pkt | 24 ++++
.../tcp_accecn_delivered_largeack2.pkt | 25 +++++
.../tcp_accecn_delivered_maxack.pkt | 25 +++++
.../tcp_accecn_delivered_updates.pkt | 70 ++++++++++++
.../net/packetdrill/tcp_accecn_ecn3.pkt | 12 ++
.../tcp_accecn_ecn_field_updates_opt.pkt | 35 ++++++
.../packetdrill/tcp_accecn_ipflags_drop.pkt | 14 +++
.../tcp_accecn_listen_opt_drop.pkt | 16 +++
.../tcp_accecn_multiple_syn_ack_drop.pkt | 28 +++++
.../tcp_accecn_multiple_syn_drop.pkt | 18 +++
.../tcp_accecn_negotiation_bleach.pkt | 23 ++++
.../tcp_accecn_negotiation_connect.pkt | 23 ++++
.../tcp_accecn_negotiation_listen.pkt | 26 +++++
.../tcp_accecn_negotiation_noopt_connect.pkt | 23 ++++
.../tcp_accecn_negotiation_optenable.pkt | 23 ++++
.../tcp_accecn_no_ecn_after_accecn.pkt | 20 ++++
.../net/packetdrill/tcp_accecn_noopt.pkt | 27 +++++
.../net/packetdrill/tcp_accecn_noprogress.pkt | 27 +++++
.../tcp_accecn_notecn_then_accecn_syn.pkt | 28 +++++
.../tcp_accecn_rfc3168_to_fallback.pkt | 18 +++
.../tcp_accecn_rfc3168_to_rfc3168.pkt | 18 +++
.../tcp_accecn_sack_space_grab.pkt | 28 +++++
.../tcp_accecn_sack_space_grab_with_ts.pkt | 39 +++++++
...tcp_accecn_serverside_accecn_disabled1.pkt | 20 ++++
...tcp_accecn_serverside_accecn_disabled2.pkt | 20 ++++
.../tcp_accecn_serverside_broken.pkt | 19 ++++
.../tcp_accecn_serverside_ecn_disabled.pkt | 19 ++++
.../tcp_accecn_serverside_only.pkt | 18 +++
...n_syn_ace_flags_acked_after_retransmit.pkt | 18 +++
.../tcp_accecn_syn_ace_flags_drop.pkt | 16 +++
...n_ack_ace_flags_acked_after_retransmit.pkt | 27 +++++
.../tcp_accecn_syn_ack_ace_flags_drop.pkt | 26 +++++
.../net/packetdrill/tcp_accecn_syn_ce.pkt | 13 +++
.../net/packetdrill/tcp_accecn_syn_ect0.pkt | 13 +++
.../net/packetdrill/tcp_accecn_syn_ect1.pkt | 13 +++
.../net/packetdrill/tcp_accecn_synack_ce.pkt | 27 +++++
..._accecn_synack_ce_updates_delivered_ce.pkt | 22 ++++
.../packetdrill/tcp_accecn_synack_ect0.pkt | 24 ++++
.../packetdrill/tcp_accecn_synack_ect1.pkt | 24 ++++
.../packetdrill/tcp_accecn_synack_rexmit.pkt | 15 +++
.../packetdrill/tcp_accecn_synack_rxmt.pkt | 25 +++++
.../packetdrill/tcp_accecn_tsnoprogress.pkt | 26 +++++
.../net/packetdrill/tcp_accecn_tsprogress.pkt | 25 +++++
76 files changed, 1680 insertions(+), 102 deletions(-)
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first_connect.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_after_synack_rxmt.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_ce_updates_received_ce.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_lost_data_ce.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_dups.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_acc_ecn_disabled.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_then_notecn_syn.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_to_rfc3168.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_lost.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_clientside_disabled.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_close_local_close_then_remote_fin.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_2ndlargeack.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_falseoverflow_detect.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack2.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_maxack.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_updates.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_ecn3.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_ecn_field_updates_opt.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_ipflags_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_listen_opt_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_ack_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_bleach.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_connect.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_listen.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_noopt_connect.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_optenable.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_no_ecn_after_accecn.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_noopt.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_noprogress.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_notecn_then_accecn_syn.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_fallback.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_rfc3168.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab_with_ts.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled1.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled2.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_broken.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_ecn_disabled.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_only.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_acked_after_retransmit.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_acked_after_retransmit.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_drop.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ce.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect0.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect1.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce_updates_delivered_ce.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect0.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect1.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rexmit.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rxmt.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_tsnoprogress.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_accecn_tsprogress.pkt
====================
Link: https://patch.msgid.link/20260131222515.8485-1-chia-yu.chang@nokia-bell-labs.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
+1678
-100
+3
-1
Documentation/networking/ip-sysctl.rst
+3
-1
Documentation/networking/ip-sysctl.rst
···
482
482
1 Send AccECN option sparingly according to the minimum option
483
483
rules outlined in draft-ietf-tcpm-accurate-ecn.
484
484
2 Send AccECN option on every packet whenever it fits into TCP
485
-
option space.
485
+
option space except when AccECN fallback is triggered.
486
+
3 Send AccECN option on every packet whenever it fits into TCP
487
+
option space even when AccECN fallback is triggered.
486
488
= ============================================================
487
489
488
490
Default: 2
+1
Documentation/networking/net_cachelines/tcp_sock.rst
+1
Documentation/networking/net_cachelines/tcp_sock.rst
···
105
105
u32[3] received_ecn_bytes read_mostly read_write
106
106
u8:4 received_ce_pending read_mostly read_write
107
107
u32[3] delivered_ecn_bytes read_write
108
+
u16 pkts_acked_ewma read_write
108
109
u8:2 syn_ect_snt write_mostly read_write
109
110
u8:2 syn_ect_rcv read_mostly read_write
110
111
u8:2 accecn_minlen write_mostly read_write
+3
-1
include/linux/tcp.h
+3
-1
include/linux/tcp.h
···
291
291
u8 nonagle : 4,/* Disable Nagle algorithm? */
292
292
rate_app_limited:1; /* rate_{delivered,interval_us} limited? */
293
293
u8 received_ce_pending:4, /* Not yet transmit cnt of received_ce */
294
-
unused2:4;
294
+
accecn_opt_sent_w_dsack:1,/* Sent ACCECN opt in previous ACK w/ D-SACK */
295
+
unused2:3;
295
296
u8 accecn_minlen:2,/* Minimum length of AccECN option sent */
296
297
est_ecnfield:2,/* ECN field for AccECN delivered estimates */
297
298
accecn_opt_demand:2,/* Demand AccECN option for n next ACKs */
···
343
342
u32 rate_interval_us; /* saved rate sample: time elapsed */
344
343
u32 rcv_rtt_last_tsecr;
345
344
u32 delivered_ecn_bytes[3];
345
+
u16 pkts_acked_ewma;/* Pkts acked EWMA for AccECN cep heuristic */
346
346
u64 first_tx_mstamp; /* start of window send phase */
347
347
u64 delivered_mstamp; /* time we reached "delivered" */
348
348
u64 bytes_acked; /* RFC4898 tcpEStatsAppHCThruOctetsAcked
+17
-3
include/net/inet_ecn.h
+17
-3
include/net/inet_ecn.h
···
51
51
return outer;
52
52
}
53
53
54
+
/* Apply either ECT(0) or ECT(1) */
55
+
static inline void __INET_ECN_xmit(struct sock *sk, bool use_ect_1)
56
+
{
57
+
__u8 ect = use_ect_1 ? INET_ECN_ECT_1 : INET_ECN_ECT_0;
58
+
59
+
/* Mask the complete byte in case the connection alternates between
60
+
* ECT(0) and ECT(1).
61
+
*/
62
+
inet_sk(sk)->tos &= ~INET_ECN_MASK;
63
+
inet_sk(sk)->tos |= ect;
64
+
if (inet6_sk(sk)) {
65
+
inet6_sk(sk)->tclass &= ~INET_ECN_MASK;
66
+
inet6_sk(sk)->tclass |= ect;
67
+
}
68
+
}
69
+
54
70
static inline void INET_ECN_xmit(struct sock *sk)
55
71
{
56
-
inet_sk(sk)->tos |= INET_ECN_ECT_0;
57
-
if (inet6_sk(sk) != NULL)
58
-
inet6_sk(sk)->tclass |= INET_ECN_ECT_0;
72
+
__INET_ECN_xmit(sk, false);
59
73
}
60
74
61
75
static inline void INET_ECN_dontxmit(struct sock *sk)
+31
-1
include/net/tcp.h
+31
-1
include/net/tcp.h
···
552
552
TCP_SYNACK_NORMAL,
553
553
TCP_SYNACK_FASTOPEN,
554
554
TCP_SYNACK_COOKIE,
555
+
TCP_SYNACK_RETRANS,
555
556
};
556
557
struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
557
558
struct request_sock *req,
···
1216
1215
#define TCP_CONG_NON_RESTRICTED BIT(0)
1217
1216
/* Requires ECN/ECT set on all packets */
1218
1217
#define TCP_CONG_NEEDS_ECN BIT(1)
1219
-
#define TCP_CONG_MASK (TCP_CONG_NON_RESTRICTED | TCP_CONG_NEEDS_ECN)
1218
+
/* Require successfully negotiated AccECN capability */
1219
+
#define TCP_CONG_NEEDS_ACCECN BIT(2)
1220
+
/* Use ECT(1) instead of ECT(0) while the CA is uninitialized */
1221
+
#define TCP_CONG_ECT_1_NEGOTIATION BIT(3)
1222
+
/* Cannot fallback to RFC3168 during AccECN negotiation */
1223
+
#define TCP_CONG_NO_FALLBACK_RFC3168 BIT(4)
1224
+
#define TCP_CONG_MASK (TCP_CONG_NON_RESTRICTED | TCP_CONG_NEEDS_ECN | \
1225
+
TCP_CONG_NEEDS_ACCECN | TCP_CONG_ECT_1_NEGOTIATION | \
1226
+
TCP_CONG_NO_FALLBACK_RFC3168)
1220
1227
1221
1228
union tcp_cc_info;
1222
1229
···
1363
1354
const struct inet_connection_sock *icsk = inet_csk(sk);
1364
1355
1365
1356
return icsk->icsk_ca_ops->flags & TCP_CONG_NEEDS_ECN;
1357
+
}
1358
+
1359
+
static inline bool tcp_ca_needs_accecn(const struct sock *sk)
1360
+
{
1361
+
const struct inet_connection_sock *icsk = inet_csk(sk);
1362
+
1363
+
return icsk->icsk_ca_ops->flags & TCP_CONG_NEEDS_ACCECN;
1364
+
}
1365
+
1366
+
static inline bool tcp_ca_ect_1_negotiation(const struct sock *sk)
1367
+
{
1368
+
const struct inet_connection_sock *icsk = inet_csk(sk);
1369
+
1370
+
return icsk->icsk_ca_ops->flags & TCP_CONG_ECT_1_NEGOTIATION;
1371
+
}
1372
+
1373
+
static inline bool tcp_ca_no_fallback_rfc3168(const struct sock *sk)
1374
+
{
1375
+
const struct inet_connection_sock *icsk = inet_csk(sk);
1376
+
1377
+
return icsk->icsk_ca_ops->flags & TCP_CONG_NO_FALLBACK_RFC3168;
1366
1378
}
1367
1379
1368
1380
static inline void tcp_ca_event(struct sock *sk, const enum tcp_ca_event event)
+68
-35
include/net/tcp_ecn.h
+68
-35
include/net/tcp_ecn.h
···
29
29
TCP_ACCECN_OPTION_DISABLED = 0,
30
30
TCP_ACCECN_OPTION_MINIMUM = 1,
31
31
TCP_ACCECN_OPTION_FULL = 2,
32
+
TCP_ACCECN_OPTION_PERSIST = 3,
32
33
};
34
+
35
+
/* Apply either ECT(0) or ECT(1) based on TCP_CONG_ECT_1_NEGOTIATION flag */
36
+
static inline void INET_ECN_xmit_ect_1_negotiation(struct sock *sk)
37
+
{
38
+
__INET_ECN_xmit(sk, tcp_ca_ect_1_negotiation(sk));
39
+
}
33
40
34
41
static inline void tcp_ecn_queue_cwr(struct tcp_sock *tp)
35
42
{
···
67
60
tp->ecn_flags &= ~TCP_ECN_QUEUE_CWR;
68
61
}
69
62
70
-
/* tp->accecn_fail_mode */
71
-
#define TCP_ACCECN_ACE_FAIL_SEND BIT(0)
72
-
#define TCP_ACCECN_ACE_FAIL_RECV BIT(1)
73
-
#define TCP_ACCECN_OPT_FAIL_SEND BIT(2)
74
-
#define TCP_ACCECN_OPT_FAIL_RECV BIT(3)
75
-
76
63
static inline bool tcp_accecn_ace_fail_send(const struct tcp_sock *tp)
77
64
{
78
65
return tp->accecn_fail_mode & TCP_ACCECN_ACE_FAIL_SEND;
···
91
90
{
92
91
tp->accecn_fail_mode |= mode;
93
92
}
94
-
95
-
#define TCP_ACCECN_OPT_NOT_SEEN 0x0
96
-
#define TCP_ACCECN_OPT_EMPTY_SEEN 0x1
97
-
#define TCP_ACCECN_OPT_COUNTER_SEEN 0x2
98
-
#define TCP_ACCECN_OPT_FAIL_SEEN 0x3
99
93
100
94
static inline u8 tcp_accecn_ace(const struct tcphdr *th)
101
95
{
···
165
169
switch (ace) {
166
170
case 0x0:
167
171
/* Invalid value */
168
-
tcp_accecn_fail_mode_set(tp, TCP_ACCECN_ACE_FAIL_RECV);
172
+
if (!TCP_SKB_CB(skb)->sacked)
173
+
tcp_accecn_fail_mode_set(tp, TCP_ACCECN_ACE_FAIL_RECV |
174
+
TCP_ACCECN_OPT_FAIL_RECV);
169
175
break;
170
176
case 0x7:
171
177
case 0x5:
···
396
398
tp->received_ce_pending = 0;
397
399
__tcp_accecn_init_bytes_counters(tp->received_ecn_bytes);
398
400
__tcp_accecn_init_bytes_counters(tp->delivered_ecn_bytes);
401
+
tp->accecn_opt_sent_w_dsack = 0;
399
402
tp->accecn_minlen = 0;
400
403
tp->accecn_opt_demand = 0;
401
404
tp->est_ecnfield = 0;
···
466
467
return TCP_ACCECN_OPT_COUNTER_SEEN;
467
468
}
468
469
470
+
static inline void tcp_ecn_rcv_synack_accecn(struct sock *sk,
471
+
const struct sk_buff *skb, u8 dsf)
472
+
{
473
+
struct tcp_sock *tp = tcp_sk(sk);
474
+
475
+
tcp_ecn_mode_set(tp, TCP_ECN_MODE_ACCECN);
476
+
tp->syn_ect_rcv = dsf & INET_ECN_MASK;
477
+
/* Demand Accurate ECN option in response to the SYN on the SYN/ACK
478
+
* and the TCP server will try to send one more packet with an AccECN
479
+
* Option at a later point during the connection.
480
+
*/
481
+
if (tp->rx_opt.accecn &&
482
+
tp->saw_accecn_opt < TCP_ACCECN_OPT_COUNTER_SEEN) {
483
+
u8 saw_opt = tcp_accecn_option_init(skb, tp->rx_opt.accecn);
484
+
485
+
tcp_accecn_saw_opt_fail_recv(tp, saw_opt);
486
+
tp->accecn_opt_demand = 2;
487
+
}
488
+
}
489
+
469
490
/* See Table 2 of the AccECN draft */
470
491
static inline void tcp_ecn_rcv_synack(struct sock *sk, const struct sk_buff *skb,
471
492
const struct tcphdr *th, u8 ip_dsfield)
···
508
489
tcp_ecn_mode_set(tp, TCP_ECN_DISABLED);
509
490
break;
510
491
case 0x1:
511
-
case 0x5:
512
492
/* +========+========+============+=============+
513
493
* | A | B | SYN/ACK | Feedback |
514
494
* | | | B->A | Mode of A |
515
495
* | | | AE CWR ECE | |
516
496
* +========+========+============+=============+
517
-
* | AccECN | Nonce | 1 0 1 | (Reserved) |
518
497
* | AccECN | ECN | 0 0 1 | Classic ECN |
519
498
* | Nonce | AccECN | 0 0 1 | Classic ECN |
520
499
* | ECN | AccECN | 0 0 1 | Classic ECN |
521
500
* +========+========+============+=============+
522
501
*/
523
-
if (tcp_ecn_mode_pending(tp))
524
-
/* Downgrade from AccECN, or requested initially */
502
+
if (tcp_ca_no_fallback_rfc3168(sk))
503
+
tcp_ecn_mode_set(tp, TCP_ECN_DISABLED);
504
+
else
525
505
tcp_ecn_mode_set(tp, TCP_ECN_MODE_RFC3168);
526
506
break;
527
-
default:
528
-
tcp_ecn_mode_set(tp, TCP_ECN_MODE_ACCECN);
529
-
tp->syn_ect_rcv = ip_dsfield & INET_ECN_MASK;
530
-
if (tp->rx_opt.accecn &&
531
-
tp->saw_accecn_opt < TCP_ACCECN_OPT_COUNTER_SEEN) {
532
-
u8 saw_opt = tcp_accecn_option_init(skb, tp->rx_opt.accecn);
533
-
534
-
tcp_accecn_saw_opt_fail_recv(tp, saw_opt);
535
-
tp->accecn_opt_demand = 2;
507
+
case 0x5:
508
+
if (tcp_ecn_mode_pending(tp)) {
509
+
tcp_ecn_rcv_synack_accecn(sk, skb, ip_dsfield);
510
+
if (INET_ECN_is_ce(ip_dsfield)) {
511
+
tp->received_ce++;
512
+
tp->received_ce_pending++;
513
+
}
536
514
}
515
+
break;
516
+
default:
517
+
tcp_ecn_rcv_synack_accecn(sk, skb, ip_dsfield);
537
518
if (INET_ECN_is_ce(ip_dsfield) &&
538
519
tcp_accecn_validate_syn_feedback(sk, ace,
539
520
tp->syn_ect_snt)) {
···
544
525
}
545
526
}
546
527
547
-
static inline void tcp_ecn_rcv_syn(struct tcp_sock *tp, const struct tcphdr *th,
528
+
static inline void tcp_ecn_rcv_syn(struct sock *sk, const struct tcphdr *th,
548
529
const struct sk_buff *skb)
549
530
{
531
+
struct tcp_sock *tp = tcp_sk(sk);
532
+
550
533
if (tcp_ecn_mode_pending(tp)) {
551
534
if (!tcp_accecn_syn_requested(th)) {
552
535
/* Downgrade to classic ECN feedback */
···
560
539
tcp_ecn_mode_set(tp, TCP_ECN_MODE_ACCECN);
561
540
}
562
541
}
563
-
if (tcp_ecn_mode_rfc3168(tp) && (!th->ece || !th->cwr))
542
+
if (tcp_ecn_mode_rfc3168(tp) &&
543
+
(!th->ece || !th->cwr || tcp_ca_no_fallback_rfc3168(sk)))
564
544
tcp_ecn_mode_set(tp, TCP_ECN_DISABLED);
565
545
}
566
546
···
583
561
TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_ECE;
584
562
else if (tcp_ca_needs_ecn(sk) ||
585
563
tcp_bpf_ca_needs_ecn(sk))
586
-
INET_ECN_xmit(sk);
564
+
INET_ECN_xmit_ect_1_negotiation(sk);
587
565
588
566
if (tp->ecn_flags & TCP_ECN_MODE_ACCECN) {
589
567
TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_ACE;
···
601
579
bool use_ecn, use_accecn;
602
580
u8 tcp_ecn = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_ecn);
603
581
604
-
use_accecn = tcp_ecn == TCP_ECN_IN_ACCECN_OUT_ACCECN;
582
+
use_accecn = tcp_ecn == TCP_ECN_IN_ACCECN_OUT_ACCECN ||
583
+
tcp_ca_needs_accecn(sk);
605
584
use_ecn = tcp_ecn == TCP_ECN_IN_ECN_OUT_ECN ||
606
585
tcp_ecn == TCP_ECN_IN_ACCECN_OUT_ECN ||
607
586
tcp_ca_needs_ecn(sk) || bpf_needs_ecn || use_accecn;
···
618
595
619
596
if (use_ecn) {
620
597
if (tcp_ca_needs_ecn(sk) || bpf_needs_ecn)
621
-
INET_ECN_xmit(sk);
598
+
INET_ECN_xmit_ect_1_negotiation(sk);
622
599
623
600
TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_ECE | TCPHDR_CWR;
624
601
if (use_accecn) {
···
642
619
}
643
620
644
621
static inline void
645
-
tcp_ecn_make_synack(const struct request_sock *req, struct tcphdr *th)
622
+
tcp_ecn_make_synack(const struct request_sock *req, struct tcphdr *th,
623
+
enum tcp_synack_type synack_type)
646
624
{
647
-
if (tcp_rsk(req)->accecn_ok)
648
-
tcp_accecn_echo_syn_ect(th, tcp_rsk(req)->syn_ect_rcv);
649
-
else if (inet_rsk(req)->ecn_ok)
650
-
th->ece = 1;
625
+
/* Accurate ECN shall retransmit SYN/ACK with ACE=0 if the
626
+
* previously retransmitted SYN/ACK also times out.
627
+
*/
628
+
if (!req->num_timeout || synack_type != TCP_SYNACK_RETRANS) {
629
+
if (tcp_rsk(req)->accecn_ok)
630
+
tcp_accecn_echo_syn_ect(th, tcp_rsk(req)->syn_ect_rcv);
631
+
else if (inet_rsk(req)->ecn_ok)
632
+
th->ece = 1;
633
+
} else if (tcp_rsk(req)->accecn_ok) {
634
+
th->ae = 0;
635
+
th->cwr = 0;
636
+
th->ece = 0;
637
+
}
651
638
}
652
639
653
640
static inline bool tcp_accecn_option_beacon_check(const struct sock *sk)
+23
-3
include/uapi/linux/tcp.h
+23
-3
include/uapi/linux/tcp.h
···
226
226
#define TCPF_CA_Loss (1<<TCP_CA_Loss)
227
227
};
228
228
229
+
/* Values for tcpi_ecn_mode after negotiation */
230
+
#define TCPI_ECN_MODE_DISABLED 0x0
231
+
#define TCPI_ECN_MODE_RFC3168 0x1
232
+
#define TCPI_ECN_MODE_ACCECN 0x2
233
+
#define TCPI_ECN_MODE_PENDING 0x3
234
+
235
+
/* Values for accecn_opt_seen */
236
+
#define TCP_ACCECN_OPT_NOT_SEEN 0x0
237
+
#define TCP_ACCECN_OPT_EMPTY_SEEN 0x1
238
+
#define TCP_ACCECN_OPT_COUNTER_SEEN 0x2
239
+
#define TCP_ACCECN_OPT_FAIL_SEEN 0x3
240
+
241
+
/* Values for accecn_fail_mode */
242
+
#define TCP_ACCECN_ACE_FAIL_SEND BIT(0)
243
+
#define TCP_ACCECN_ACE_FAIL_RECV BIT(1)
244
+
#define TCP_ACCECN_OPT_FAIL_SEND BIT(2)
245
+
#define TCP_ACCECN_OPT_FAIL_RECV BIT(3)
246
+
229
247
struct tcp_info {
230
248
__u8 tcpi_state;
231
249
__u8 tcpi_ca_state;
···
334
316
* in milliseconds, including any
335
317
* unfinished recovery.
336
318
*/
337
-
__u32 tcpi_received_ce; /* # of CE marks received */
319
+
__u32 tcpi_received_ce; /* # of CE marked segments received */
338
320
__u32 tcpi_delivered_e1_bytes; /* Accurate ECN byte counters */
339
321
__u32 tcpi_delivered_e0_bytes;
340
322
__u32 tcpi_delivered_ce_bytes;
341
323
__u32 tcpi_received_e1_bytes;
342
324
__u32 tcpi_received_e0_bytes;
343
325
__u32 tcpi_received_ce_bytes;
344
-
__u16 tcpi_accecn_fail_mode;
345
-
__u16 tcpi_accecn_opt_seen;
326
+
__u32 tcpi_ecn_mode:2,
327
+
tcpi_accecn_opt_seen:2,
328
+
tcpi_accecn_fail_mode:4,
329
+
tcpi_options2:24;
346
330
};
347
331
348
332
/* netlink attributes types for SCM_TIMESTAMPING_OPT_STATS */
+3
net/ipv4/inet_connection_sock.c
+3
net/ipv4/inet_connection_sock.c
···
20
20
#include <net/tcp_states.h>
21
21
#include <net/xfrm.h>
22
22
#include <net/tcp.h>
23
+
#include <net/tcp_ecn.h>
23
24
#include <net/sock_reuseport.h>
24
25
#include <net/addrconf.h>
25
26
···
1104
1103
(!resend ||
1105
1104
!tcp_rtx_synack(sk_listener, req) ||
1106
1105
inet_rsk(req)->acked)) {
1106
+
if (req->num_retrans > 1 && tcp_rsk(req)->accecn_ok)
1107
+
tcp_rsk(req)->accecn_fail_mode |= TCP_ACCECN_ACE_FAIL_SEND;
1107
1108
if (req->num_timeout++ == 0)
1108
1109
atomic_dec(&queue->young);
1109
1110
mod_timer(&req->rsk_timer, jiffies + tcp_reqsk_timeout(req));
+2
-2
net/ipv4/sysctl_net_ipv4.c
+2
-2
net/ipv4/sysctl_net_ipv4.c
···
47
47
static int tcp_plb_max_rounds = 31;
48
48
static int tcp_plb_max_cong_thresh = 256;
49
49
static unsigned int tcp_tw_reuse_delay_max = TCP_PAWS_MSL * MSEC_PER_SEC;
50
-
static int tcp_ecn_mode_max = 2;
50
+
static int tcp_ecn_mode_max = 5;
51
51
static u32 icmp_errors_extension_mask_all =
52
52
GENMASK_U8(ICMP_ERR_EXT_COUNT - 1, 0);
53
53
···
749
749
.mode = 0644,
750
750
.proc_handler = proc_dou8vec_minmax,
751
751
.extra1 = SYSCTL_ZERO,
752
-
.extra2 = SYSCTL_TWO,
752
+
.extra2 = SYSCTL_THREE,
753
753
},
754
754
{
755
755
.procname = "tcp_ecn_option_beacon",
+10
net/ipv4/tcp.c
+10
net/ipv4/tcp.c
···
3470
3470
tcp_accecn_init_counters(tp);
3471
3471
tp->prev_ecnfield = 0;
3472
3472
tp->accecn_opt_tstamp = 0;
3473
+
tp->pkts_acked_ewma = 0;
3473
3474
if (icsk->icsk_ca_initialized && icsk->icsk_ca_ops->release)
3474
3475
icsk->icsk_ca_ops->release(sk);
3475
3476
memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
···
4373
4372
if (tp->rto_stamp)
4374
4373
info->tcpi_total_rto_time += tcp_clock_ms() - tp->rto_stamp;
4375
4374
4375
+
if (tcp_ecn_disabled(tp))
4376
+
info->tcpi_ecn_mode = TCPI_ECN_MODE_DISABLED;
4377
+
else if (tcp_ecn_mode_rfc3168(tp))
4378
+
info->tcpi_ecn_mode = TCPI_ECN_MODE_RFC3168;
4379
+
else if (tcp_ecn_mode_accecn(tp))
4380
+
info->tcpi_ecn_mode = TCPI_ECN_MODE_ACCECN;
4381
+
else if (tcp_ecn_mode_pending(tp))
4382
+
info->tcpi_ecn_mode = TCPI_ECN_MODE_PENDING;
4376
4383
info->tcpi_accecn_fail_mode = tp->accecn_fail_mode;
4377
4384
info->tcpi_accecn_opt_seen = tp->saw_accecn_opt;
4378
4385
info->tcpi_received_ce = tp->received_ce;
···
5252
5243
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, rate_interval_us);
5253
5244
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, rcv_rtt_last_tsecr);
5254
5245
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, delivered_ecn_bytes);
5246
+
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, pkts_acked_ewma);
5255
5247
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, first_tx_mstamp);
5256
5248
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, delivered_mstamp);
5257
5249
CACHELINE_ASSERT_GROUP_MEMBER(struct tcp_sock, tcp_sock_write_rx, bytes_acked);
+3
-2
net/ipv4/tcp_cong.c
+3
-2
net/ipv4/tcp_cong.c
···
16
16
#include <linux/gfp.h>
17
17
#include <linux/jhash.h>
18
18
#include <net/tcp.h>
19
+
#include <net/tcp_ecn.h>
19
20
#include <trace/events/tcp.h>
20
21
21
22
static DEFINE_SPINLOCK(tcp_cong_list_lock);
···
228
227
229
228
memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
230
229
if (ca->flags & TCP_CONG_NEEDS_ECN)
231
-
INET_ECN_xmit(sk);
230
+
INET_ECN_xmit_ect_1_negotiation(sk);
232
231
else
233
232
INET_ECN_dontxmit(sk);
234
233
}
···
258
257
memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
259
258
260
259
if (ca->flags & TCP_CONG_NEEDS_ECN)
261
-
INET_ECN_xmit(sk);
260
+
INET_ECN_xmit_ect_1_negotiation(sk);
262
261
else
263
262
INET_ECN_dontxmit(sk);
264
263
+36
-4
net/ipv4/tcp_input.c
+36
-4
net/ipv4/tcp_input.c
···
488
488
tcp_count_delivered_ce(tp, delivered);
489
489
}
490
490
491
+
#define PKTS_ACKED_WEIGHT 6
492
+
#define PKTS_ACKED_PREC 6
493
+
#define ACK_COMP_THRESH 4
494
+
491
495
/* Returns the ECN CE delta */
492
496
static u32 __tcp_accecn_process(struct sock *sk, const struct sk_buff *skb,
493
497
u32 delivered_pkts, u32 delivered_bytes,
···
503
499
u32 delta, safe_delta, d_ceb;
504
500
bool opt_deltas_valid;
505
501
u32 corrected_ace;
502
+
u32 ewma;
506
503
507
504
/* Reordered ACK or uncertain due to lack of data to send and ts */
508
505
if (!(flag & (FLAG_FORWARD_PROGRESS | FLAG_TS_PROGRESS)))
···
511
506
512
507
opt_deltas_valid = tcp_accecn_process_option(tp, skb,
513
508
delivered_bytes, flag);
509
+
510
+
if (delivered_pkts) {
511
+
if (!tp->pkts_acked_ewma) {
512
+
ewma = delivered_pkts << PKTS_ACKED_PREC;
513
+
} else {
514
+
ewma = tp->pkts_acked_ewma;
515
+
ewma = (((ewma << PKTS_ACKED_WEIGHT) - ewma) +
516
+
(delivered_pkts << PKTS_ACKED_PREC)) >>
517
+
PKTS_ACKED_WEIGHT;
518
+
}
519
+
tp->pkts_acked_ewma = min_t(u32, ewma, 0xFFFFU);
520
+
}
514
521
515
522
if (!(flag & FLAG_SLOWPATH)) {
516
523
/* AccECN counter might overflow on large ACKs */
···
572
555
if (d_ceb <
573
556
safe_delta * tp->mss_cache >> TCP_ACCECN_SAFETY_SHIFT)
574
557
return delta;
575
-
}
558
+
} else if (tp->pkts_acked_ewma > (ACK_COMP_THRESH << PKTS_ACKED_PREC))
559
+
return delta;
576
560
577
561
return safe_delta;
578
562
}
···
5046
5028
tcp_sack_extend(tp->duplicate_sack, seq, end_seq);
5047
5029
}
5048
5030
5049
-
static void tcp_rcv_spurious_retrans(struct sock *sk, const struct sk_buff *skb)
5031
+
static void tcp_rcv_spurious_retrans(struct sock *sk,
5032
+
const struct sk_buff *skb)
5050
5033
{
5034
+
struct tcp_sock *tp = tcp_sk(sk);
5035
+
5051
5036
/* When the ACK path fails or drops most ACKs, the sender would
5052
5037
* timeout and spuriously retransmit the same segment repeatedly.
5053
5038
* If it seems our ACKs are not reaching the other side,
···
5070
5049
/* Save last flowlabel after a spurious retrans. */
5071
5050
tcp_save_lrcv_flowlabel(sk, skb);
5072
5051
#endif
5052
+
/* Check DSACK info to detect that the previous ACK carrying the
5053
+
* AccECN option was lost after the second retransmision, and then
5054
+
* stop sending AccECN option in all subsequent ACKs.
5055
+
*/
5056
+
if (tcp_ecn_mode_accecn(tp) &&
5057
+
tp->accecn_opt_sent_w_dsack &&
5058
+
TCP_SKB_CB(skb)->seq == tp->duplicate_sack[0].start_seq)
5059
+
tcp_accecn_fail_mode_set(tp, TCP_ACCECN_OPT_FAIL_SEND);
5073
5060
}
5074
5061
5075
5062
static void tcp_send_dupack(struct sock *sk, const struct sk_buff *skb)
···
6480
6451
if (th->syn) {
6481
6452
if (tcp_ecn_mode_accecn(tp)) {
6482
6453
accecn_reflector = true;
6454
+
tp->syn_ect_rcv = TCP_SKB_CB(skb)->ip_dsfield &
6455
+
INET_ECN_MASK;
6483
6456
if (tp->rx_opt.accecn &&
6484
6457
tp->saw_accecn_opt < TCP_ACCECN_OPT_COUNTER_SEEN) {
6485
6458
u8 saw_opt = tcp_accecn_option_init(skb, tp->rx_opt.accecn);
···
7103
7072
tp->snd_wl1 = TCP_SKB_CB(skb)->seq;
7104
7073
tp->max_window = tp->snd_wnd;
7105
7074
7106
-
tcp_ecn_rcv_syn(tp, th, skb);
7075
+
tcp_ecn_rcv_syn(sk, th, skb);
7107
7076
7108
7077
tcp_mtup_init(sk);
7109
7078
tcp_sync_mss(sk, icsk->icsk_pmtu_cookie);
···
7508
7477
u32 ecn_ok_dst;
7509
7478
7510
7479
if (tcp_accecn_syn_requested(th) &&
7511
-
READ_ONCE(net->ipv4.sysctl_tcp_ecn) >= 3) {
7480
+
(READ_ONCE(net->ipv4.sysctl_tcp_ecn) >= 3 ||
7481
+
tcp_ca_needs_accecn(listen_sk))) {
7512
7482
inet_rsk(req)->ecn_ok = 1;
7513
7483
tcp_rsk(req)->accecn_ok = 1;
7514
7484
tcp_rsk(req)->syn_ect_rcv = TCP_SKB_CB(skb)->ip_dsfield &
+28
-11
net/ipv4/tcp_minisocks.c
+28
-11
net/ipv4/tcp_minisocks.c
···
481
481
tp->syn_ect_snt = treq->syn_ect_snt;
482
482
tcp_accecn_third_ack(sk, skb, treq->syn_ect_snt);
483
483
tp->saw_accecn_opt = treq->saw_accecn_opt;
484
+
if (treq->accecn_fail_mode & TCP_ACCECN_ACE_FAIL_SEND)
485
+
tcp_accecn_fail_mode_set(tp, TCP_ACCECN_ACE_FAIL_SEND);
486
+
if (treq->accecn_fail_mode & TCP_ACCECN_ACE_FAIL_RECV)
487
+
tcp_accecn_fail_mode_set(tp, TCP_ACCECN_ACE_FAIL_RECV);
484
488
tp->prev_ecnfield = treq->syn_ect_rcv;
485
489
tp->accecn_opt_demand = 1;
486
490
tcp_ecn_received_counters_payload(sk, skb);
487
491
} else {
488
-
tcp_ecn_mode_set(tp, inet_rsk(req)->ecn_ok ?
489
-
TCP_ECN_MODE_RFC3168 :
490
-
TCP_ECN_DISABLED);
492
+
if (inet_rsk(req)->ecn_ok && !tcp_ca_no_fallback_rfc3168(sk))
493
+
tcp_ecn_mode_set(tp, TCP_ECN_MODE_RFC3168);
494
+
else
495
+
tcp_ecn_mode_set(tp, TCP_ECN_DISABLED);
491
496
}
492
497
}
493
498
···
753
748
*/
754
749
if (!tcp_oow_rate_limited(sock_net(sk), skb,
755
750
LINUX_MIB_TCPACKSKIPPEDSYNRECV,
756
-
&tcp_rsk(req)->last_oow_ack_time) &&
751
+
&tcp_rsk(req)->last_oow_ack_time)) {
752
+
if (tcp_rsk(req)->accecn_ok) {
753
+
u8 ect_rcv = TCP_SKB_CB(skb)->ip_dsfield &
754
+
INET_ECN_MASK;
757
755
758
-
!tcp_rtx_synack(sk, req)) {
759
-
unsigned long expires = jiffies;
756
+
tcp_rsk(req)->syn_ect_rcv = ect_rcv;
757
+
if (tcp_accecn_ace(tcp_hdr(skb)) == 0x0)
758
+
tcp_rsk(req)->accecn_fail_mode |= TCP_ACCECN_ACE_FAIL_RECV;
759
+
}
760
+
if (!tcp_rtx_synack(sk, req)) {
761
+
unsigned long expires = jiffies;
760
762
761
-
expires += tcp_reqsk_timeout(req);
762
-
if (!fastopen)
763
-
mod_timer_pending(&req->rsk_timer, expires);
764
-
else
765
-
req->rsk_timer.expires = expires;
763
+
if (req->num_retrans > 1 && tcp_rsk(req)->accecn_ok)
764
+
tcp_rsk(req)->accecn_fail_mode |= TCP_ACCECN_ACE_FAIL_SEND;
765
+
766
+
expires += tcp_reqsk_timeout(req);
767
+
if (!fastopen)
768
+
mod_timer_pending(&req->rsk_timer,
769
+
expires);
770
+
else
771
+
req->rsk_timer.expires = expires;
772
+
}
766
773
}
767
774
return NULL;
768
775
}
+1
-2
net/ipv4/tcp_offload.c
+1
-2
net/ipv4/tcp_offload.c
···
304
304
goto out_check_final;
305
305
306
306
th2 = tcp_hdr(p);
307
-
flush = (__force int)(flags & TCP_FLAG_CWR);
308
-
flush |= (__force int)((flags ^ tcp_flag_word(th2)) &
307
+
flush = (__force int)((flags ^ tcp_flag_word(th2)) &
309
308
~(TCP_FLAG_FIN | TCP_FLAG_PSH));
310
309
flush |= (__force int)(th->ack_seq ^ th2->ack_seq);
311
310
for (i = sizeof(*th); i < thlen; i += 4)
+23
-11
net/ipv4/tcp_output.c
+23
-11
net/ipv4/tcp_output.c
···
334
334
return;
335
335
336
336
if (tcp_ecn_mode_accecn(tp)) {
337
-
if (!tcp_accecn_ace_fail_recv(tp))
337
+
if (!tcp_accecn_ace_fail_recv(tp) &&
338
+
!tcp_accecn_ace_fail_send(tp))
338
339
INET_ECN_xmit(sk);
340
+
else
341
+
INET_ECN_dontxmit(sk);
339
342
tcp_accecn_set_ace(tp, skb, th);
340
343
skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ACCECN;
341
344
} else {
···
715
712
if (tp) {
716
713
tp->accecn_minlen = 0;
717
714
tp->accecn_opt_tstamp = tp->tcp_mstamp;
715
+
tp->accecn_opt_sent_w_dsack = tp->rx_opt.dsack;
718
716
if (tp->accecn_opt_demand)
719
717
tp->accecn_opt_demand--;
720
718
}
719
+
} else if (tp) {
720
+
tp->accecn_opt_sent_w_dsack = 0;
721
721
}
722
722
723
723
if (unlikely(OPTION_SACK_ADVERTISE & options)) {
···
1112
1106
1113
1107
if (treq->accecn_ok &&
1114
1108
READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_ecn_option) &&
1115
-
req->num_timeout < 1 && remaining >= TCPOLEN_ACCECN_BASE) {
1109
+
synack_type != TCP_SYNACK_RETRANS && remaining >= TCPOLEN_ACCECN_BASE) {
1116
1110
opts->use_synack_ecn_bytes = 1;
1117
1111
remaining -= tcp_options_fit_accecn(opts, 0, remaining);
1118
1112
}
···
1192
1186
if (tcp_ecn_mode_accecn(tp)) {
1193
1187
int ecn_opt = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_ecn_option);
1194
1188
1195
-
if (ecn_opt && tp->saw_accecn_opt && !tcp_accecn_opt_fail_send(tp) &&
1189
+
if (ecn_opt && tp->saw_accecn_opt &&
1190
+
(ecn_opt >= TCP_ACCECN_OPTION_PERSIST ||
1191
+
!tcp_accecn_opt_fail_send(tp)) &&
1196
1192
(ecn_opt >= TCP_ACCECN_OPTION_FULL || tp->accecn_opt_demand ||
1197
1193
tcp_accecn_option_beacon_check(sk))) {
1198
1194
opts->use_synack_ecn_bytes = 0;
···
3614
3606
tcp_retrans_try_collapse(sk, skb, avail_wnd);
3615
3607
}
3616
3608
3617
-
/* RFC3168, section 6.1.1.1. ECN fallback
3618
-
* As AccECN uses the same SYN flags (+ AE), this check covers both
3619
-
* cases.
3620
-
*/
3621
-
if ((TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN_ECN) == TCPHDR_SYN_ECN)
3622
-
tcp_ecn_clear_syn(sk, skb);
3609
+
if (!tcp_ecn_mode_pending(tp) || icsk->icsk_retransmits > 1) {
3610
+
/* RFC3168, section 6.1.1.1. ECN fallback
3611
+
* As AccECN uses the same SYN flags (+ AE), this check
3612
+
* covers both cases.
3613
+
*/
3614
+
if ((TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN_ECN) ==
3615
+
TCPHDR_SYN_ECN)
3616
+
tcp_ecn_clear_syn(sk, skb);
3617
+
}
3623
3618
3624
3619
/* Update global and local TCP statistics. */
3625
3620
segs = tcp_skb_pcount(skb);
···
3937
3926
3938
3927
switch (synack_type) {
3939
3928
case TCP_SYNACK_NORMAL:
3929
+
case TCP_SYNACK_RETRANS:
3940
3930
skb_set_owner_edemux(skb, req_to_sk(req));
3941
3931
break;
3942
3932
case TCP_SYNACK_COOKIE:
···
4020
4008
memset(th, 0, sizeof(struct tcphdr));
4021
4009
th->syn = 1;
4022
4010
th->ack = 1;
4023
-
tcp_ecn_make_synack(req, th);
4011
+
tcp_ecn_make_synack(req, th, synack_type);
4024
4012
th->source = htons(ireq->ir_num);
4025
4013
th->dest = ireq->ir_rmt_port;
4026
4014
skb->mark = ireq->ir_mark;
···
4623
4611
/* Paired with WRITE_ONCE() in sock_setsockopt() */
4624
4612
if (READ_ONCE(sk->sk_txrehash) == SOCK_TXREHASH_ENABLED)
4625
4613
WRITE_ONCE(tcp_rsk(req)->txhash, net_tx_rndhash());
4626
-
res = af_ops->send_synack(sk, NULL, &fl, req, NULL, TCP_SYNACK_NORMAL,
4614
+
res = af_ops->send_synack(sk, NULL, &fl, req, NULL, TCP_SYNACK_RETRANS,
4627
4615
NULL);
4628
4616
if (!res) {
4629
4617
TCP_INC_STATS(sock_net(sk), TCP_MIB_RETRANSSEGS);
+3
net/ipv4/tcp_timer.c
+3
net/ipv4/tcp_timer.c
···
22
22
#include <linux/module.h>
23
23
#include <linux/gfp.h>
24
24
#include <net/tcp.h>
25
+
#include <net/tcp_ecn.h>
25
26
#include <net/rstreason.h>
26
27
27
28
static u32 tcp_clamp_rto_to_user_timeout(const struct sock *sk)
···
480
479
* it's not good to give up too easily.
481
480
*/
482
481
tcp_rtx_synack(sk, req);
482
+
if (req->num_retrans > 1 && tcp_rsk(req)->accecn_ok)
483
+
tcp_rsk(req)->accecn_fail_mode |= TCP_ACCECN_ACE_FAIL_SEND;
483
484
req->num_timeout++;
484
485
tcp_update_rto_stats(sk);
485
486
if (!tp->retrans_stamp)
+58
-23
tools/testing/selftests/drivers/net/gro.c
+58
-23
tools/testing/selftests/drivers/net/gro.c
···
17
17
* Pure ACK does not coalesce.
18
18
*
19
19
* flags_*:
20
-
* No packets with PSH, SYN, URG, RST set will be coalesced.
21
-
* - flags_psh, flags_syn, flags_rst, flags_urg
20
+
* No packets with PSH, SYN, URG, RST, CWR set will be coalesced.
21
+
* - flags_psh, flags_syn, flags_rst, flags_urg, flags_cwr
22
22
*
23
23
* tcp_*:
24
24
* Packets with incorrect checksum, non-consecutive seqno and
···
360
360
fill_datalinklayer(buf);
361
361
}
362
362
363
-
/* send one extra flag, not first and not last pkt */
364
-
static void send_flags(int fd, struct sockaddr_ll *daddr, int psh, int syn,
365
-
int rst, int urg)
363
+
#ifndef TH_CWR
364
+
#define TH_CWR 0x80
365
+
#endif
366
+
static void set_flags(struct tcphdr *tcph, int payload_len, int psh, int syn,
367
+
int rst, int urg, int cwr)
366
368
{
367
-
static char flag_buf[MAX_HDR_LEN + PAYLOAD_LEN];
368
-
static char buf[MAX_HDR_LEN + PAYLOAD_LEN];
369
-
int payload_len, pkt_size, flag, i;
370
-
struct tcphdr *tcph;
371
-
372
-
payload_len = PAYLOAD_LEN * psh;
373
-
pkt_size = total_hdr_len + payload_len;
374
-
flag = NUM_PACKETS / 2;
375
-
376
-
create_packet(flag_buf, flag * payload_len, 0, payload_len, 0);
377
-
378
-
tcph = (struct tcphdr *)(flag_buf + tcp_offset);
379
369
tcph->psh = psh;
380
370
tcph->syn = syn;
381
371
tcph->rst = rst;
382
372
tcph->urg = urg;
373
+
if (cwr)
374
+
tcph->th_flags |= TH_CWR;
375
+
else
376
+
tcph->th_flags &= ~TH_CWR;
383
377
tcph->check = 0;
384
378
tcph->check = tcp_checksum(tcph, payload_len);
379
+
}
380
+
381
+
/* send extra flags of the (NUM_PACKETS / 2) and (NUM_PACKETS / 2 - 1)
382
+
* pkts, not first and not last pkt
383
+
*/
384
+
static void send_flags(int fd, struct sockaddr_ll *daddr, int psh, int syn,
385
+
int rst, int urg, int cwr)
386
+
{
387
+
static char flag_buf[2][MAX_HDR_LEN + PAYLOAD_LEN];
388
+
static char buf[MAX_HDR_LEN + PAYLOAD_LEN];
389
+
int payload_len, pkt_size, i;
390
+
struct tcphdr *tcph;
391
+
int flag[2];
392
+
393
+
payload_len = PAYLOAD_LEN * (psh || cwr);
394
+
pkt_size = total_hdr_len + payload_len;
395
+
flag[0] = NUM_PACKETS / 2;
396
+
flag[1] = NUM_PACKETS / 2 - 1;
397
+
398
+
/* Create and configure packets with flags
399
+
*/
400
+
for (i = 0; i < 2; i++) {
401
+
if (flag[i] > 0) {
402
+
create_packet(flag_buf[i], flag[i] * payload_len, 0,
403
+
payload_len, 0);
404
+
tcph = (struct tcphdr *)(flag_buf[i] + tcp_offset);
405
+
set_flags(tcph, payload_len, psh, syn, rst, urg, cwr);
406
+
}
407
+
}
385
408
386
409
for (i = 0; i < NUM_PACKETS + 1; i++) {
387
-
if (i == flag) {
388
-
write_packet(fd, flag_buf, pkt_size, daddr);
410
+
if (i == flag[0]) {
411
+
write_packet(fd, flag_buf[0], pkt_size, daddr);
412
+
continue;
413
+
} else if (i == flag[1] && cwr) {
414
+
write_packet(fd, flag_buf[1], pkt_size, daddr);
389
415
continue;
390
416
}
391
417
create_packet(buf, i * PAYLOAD_LEN, 0, PAYLOAD_LEN, 0);
···
1094
1068
1095
1069
/* flags sub-tests */
1096
1070
} else if (strcmp(testname, "flags_psh") == 0) {
1097
-
send_flags(txfd, &daddr, 1, 0, 0, 0);
1071
+
send_flags(txfd, &daddr, 1, 0, 0, 0, 0);
1098
1072
write_packet(txfd, fin_pkt, total_hdr_len, &daddr);
1099
1073
} else if (strcmp(testname, "flags_syn") == 0) {
1100
-
send_flags(txfd, &daddr, 0, 1, 0, 0);
1074
+
send_flags(txfd, &daddr, 0, 1, 0, 0, 0);
1101
1075
write_packet(txfd, fin_pkt, total_hdr_len, &daddr);
1102
1076
} else if (strcmp(testname, "flags_rst") == 0) {
1103
-
send_flags(txfd, &daddr, 0, 0, 1, 0);
1077
+
send_flags(txfd, &daddr, 0, 0, 1, 0, 0);
1104
1078
write_packet(txfd, fin_pkt, total_hdr_len, &daddr);
1105
1079
} else if (strcmp(testname, "flags_urg") == 0) {
1106
-
send_flags(txfd, &daddr, 0, 0, 0, 1);
1080
+
send_flags(txfd, &daddr, 0, 0, 0, 1, 0);
1081
+
write_packet(txfd, fin_pkt, total_hdr_len, &daddr);
1082
+
} else if (strcmp(testname, "flags_cwr") == 0) {
1083
+
send_flags(txfd, &daddr, 0, 0, 0, 0, 1);
1107
1084
write_packet(txfd, fin_pkt, total_hdr_len, &daddr);
1108
1085
1109
1086
/* tcp sub-tests */
···
1267
1238
correct_payload[1] = 0;
1268
1239
correct_payload[2] = PAYLOAD_LEN * 2;
1269
1240
printf("urg flag ends coalescing: ");
1241
+
check_recv_pkts(rxfd, correct_payload, 3);
1242
+
} else if (strcmp(testname, "flags_cwr") == 0) {
1243
+
correct_payload[0] = PAYLOAD_LEN;
1244
+
correct_payload[1] = PAYLOAD_LEN * 2;
1245
+
correct_payload[2] = PAYLOAD_LEN * 2;
1246
+
printf("cwr flag ends coalescing: ");
1270
1247
check_recv_pkts(rxfd, correct_payload, 3);
1271
1248
1272
1249
/* tcp sub-tests */
+2
-1
tools/testing/selftests/drivers/net/gro.py
+2
-1
tools/testing/selftests/drivers/net/gro.py
···
17
17
- flags_syn: Packets with SYN flag don't coalesce
18
18
- flags_rst: Packets with RST flag don't coalesce
19
19
- flags_urg: Packets with URG flag don't coalesce
20
+
- flags_cwr: Packets with CWR flag don't coalesce
20
21
- tcp_csum: Packets with incorrect checksum don't coalesce
21
22
- tcp_seq: Packets with non-consecutive seqno don't coalesce
22
23
- tcp_ts: Packets with different timestamp options don't coalesce
···
192
191
common_tests = [
193
192
"data_same", "data_lrg_sml", "data_sml_lrg",
194
193
"ack",
195
-
"flags_psh", "flags_syn", "flags_rst", "flags_urg",
194
+
"flags_psh", "flags_syn", "flags_rst", "flags_urg", "flags_cwr",
196
195
"tcp_csum", "tcp_seq", "tcp_ts", "tcp_opt",
197
196
"ip_ecn", "ip_tos",
198
197
"large_max", "large_rem",
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first.pkt
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first.pkt
···
1
+
// 3rd ACK + 1st data segment lost, data segments with ce
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0.05 < SEWA 0:0(0) win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
// 3rd ACK lost
15
+
// 1st data segment lost
16
+
+0.05 < [ce] EAP. 1001:2001(1000) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
17
+
+.002 > [ect0] WA. 1:1(0) ack 1 <ECN e1b 1 ceb 1000 e0b 1,nop,nop,nop,sack 1001:2001>
18
+
+.002 accept(3, ..., ...) = 4
19
+
20
+
+0.2 < [ce] EAP. 1:1001(1000) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
21
+
+.001 > [ect0] EWA. 1:1(0) ack 2001 <ECN e1b 1 ceb 2000 e0b 1,nop>
22
+
23
+
+0.05 < [ce] EAP. 2001:3001(1000) ack 1 win 264
24
+
+.001 > [ect0] . 1:1(0) ack 3001 <ECN e1b 1 ceb 3000 e0b 1,nop>
+30
tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first_connect.pkt
+30
tools/testing/selftests/net/packetdrill/tcp_accecn_2nd_data_as_first_connect.pkt
···
1
+
// 3rd ACK + 1st data segment lost, 2nd data segments with ce
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [noecn] SW. 0:0(0) ack 1 win 32767 <mss 1016,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
// 3rd ACK lost
13
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
14
+
15
+
+0.01 write(4, ..., 2000) = 2000
16
+
// 1st data segment lost + 2nd gets CE
17
+
+.002 > [ect0] .5 1:1005(1004) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
18
+
+.000 > [ect0] P.5 1005:2001(996) ack 1 <ECN e1b 1 ceb 0 e0b 1, nop>
19
+
+0.05 < [ect0] .6 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 996 e1b 1,nop,nop,nop,sack 1005:2001>
20
+
21
+
+0.01 %{ assert tcpi_delivered_ce == 1, tcpi_delivered_ce }%
22
+
23
+
+0.002~+0.1 > [ect0] .5 1:1005(1004) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
24
+
+.05 < [ect0] .6 1:1(0) ack 2001 win 264 <ECN e0b 1005 ceb 996 e1b 1,nop>
25
+
26
+
+0.01 write(4, ..., 1000) = 1000
27
+
+0~+0.002 > [ect0] P.5 2001:3001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
28
+
29
+
+0.1 < [ect0] .5 1:1001(1000) ack 3001 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
30
+
+0~+0.01 > [ect0] .5 3001:3001(0) ack 1001 <ECN e1b 1 ceb 0 e0b 1001,nop>
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_after_synack_rxmt.pkt
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_after_synack_rxmt.pkt
···
1
+
// Test 3rd ACK flags when SYN-ACK is rexmitted
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [noecn] SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
13
+
14
+
+0.1 < [ect0] S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
16
+
// Our code currently sends a challenge ACK
17
+
// when it receives a SYN in ESTABLISHED state
18
+
// based on the latest SYN
19
+
+.002 > [ect0] A. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_ce_updates_received_ce.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_ce_updates_received_ce.pkt
···
1
+
// Third ACK CE increases r.cep
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0.05 < SEWA 0:0(0) win 32767 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ce] W. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 write(4, ..., 1000) = 1000
18
+
+.002 > [ect0] WAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+22
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_lost_data_ce.pkt
+22
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_ack_lost_data_ce.pkt
···
1
+
// 3rd ACK lost, CE for the first data segment
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0.05 < SEWA 0:0(0) win 32767 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
// 3rd ACK lost
15
+
+0.05 < [ce] EAP. 1:1001(1000) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
16
+
+.002 > [ect0] WA. 1:1(0) ack 1001 <ECN e1b 1 ceb 1000 e0b 1,nop>
17
+
+.002 accept(3, ..., ...) = 4
18
+
19
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
20
+
21
+
+0.05 < [ce] EAP. 1001:2001(1000) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
22
+
+.001 > [ect0] EWA. 1:1(0) ack 2001 <ECN e1b 1 ceb 2000 e0b 1 ,nop>
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_dups.pkt
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_3rd_dups.pkt
···
1
+
// Test SYN/ACK rexmit triggered 3rd ACK duplicate + CE on first data seg
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
15
+
// SYN/ACK rexmitted => two 3rd ACKs in-flight
16
+
+1.0~+1.1 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
17
+
// Delivered 1st 3rd ACK
18
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
19
+
+.002 accept(3, ..., ...) = 4
20
+
21
+
// Duplicate 3rd ACK delivered
22
+
+1.05 < [ect0] W. 1:1(0) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
23
+
24
+
+0.05 < [ce] EAP. 1:1001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
25
+
+.002 > [ect0] WA. 1:1(0) ack 1001 <ECN e1b 1 ceb 1000 e0b 1,nop>
26
+
+0 read(4, ..., 1000) = 1000
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_acc_ecn_disabled.pkt
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_acc_ecn_disabled.pkt
···
1
+
// Test that when accurate ECN is disabled,
2
+
// client uses RFC3168 ECN for SYN
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=1
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
9
+
+.002 ... 0.052 connect(4, ..., ...) = 0
10
+
11
+
+.002 > [noecn] SEW 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
12
+
+0.05 < [noecn] S. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,nop,nop,nop,wscale 8>
13
+
+.002 > [noecn] . 1:1(0) ack 1
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_then_notecn_syn.pkt
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_then_notecn_syn.pkt
···
1
+
// Test that SYN-ACK with ACE flags and without
2
+
// ACE flags got dropped. Although we disable ECN,
3
+
// we shouldn't consider this as blackholed as
4
+
// these are dropped due to congestion
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
sysctl -q net.ipv4.tcp_ecn_option=2
9
+
`
10
+
11
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
12
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
13
+
+0 bind(3, ..., ...) = 0
14
+
+0 listen(3, 1) = 0
15
+
16
+
+0 < [ect0] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
17
+
+.002 > [noecn] SA. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
18
+
19
+
// Retransmit SYN
20
+
+0.1 < [noecn] S 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
21
+
+.002 > [noecn] SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
22
+
23
+
+0.1 < [noecn] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
24
+
+.002 accept(3, ..., ...) = 4
25
+
26
+
// Write with AccECN option but with ip-noecn since we received one SYN with ACE=0
27
+
+0.01 write(4, ..., 100) = 100
28
+
+.002 > [noecn] P5. 1:101(100) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_to_rfc3168.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_accecn_to_rfc3168.pkt
···
1
+
// Test AccECN -> RFC3168 fallback when sysctl asks for RFC3168 ECN
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=1
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SE. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < . 1:1(0) ack 1 win 320
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 write(4, ..., 1000) = 1000
18
+
+.002 > [ect0] P. 1:1001(1000) ack 1
+34
tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_drop.pkt
+34
tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_drop.pkt
···
1
+
// Client negotiates AccECN and starts sending
2
+
// AccECN option in last ACK and data segments
3
+
// Middlebox drops AccECN option and client
4
+
// reverts to ACE flags only
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
sysctl -q net.ipv4.tcp_ecn_option=2
9
+
sysctl -q net.ipv4.tcp_ecn_option_beacon=1
10
+
`
11
+
12
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
13
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
14
+
+0 bind(3, ..., ...) = 0
15
+
+0 listen(3, 1) = 0
16
+
17
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
18
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
19
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
20
+
+.002 accept(3, ..., ...) = 4
21
+
22
+
+0.05 < [ect0] EAP. 1:1001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
23
+
+.002 > [ect0] EA. 1:1(0) ack 1001 <ECN e1b 1 ceb 0 e0b 1001,nop>
24
+
+0 read(4, ..., 1000) = 1000
25
+
26
+
+0.05 < [ect0] EAP. 1:1001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
27
+
+.002 > [ect0] EA. 1:1(0) ack 1001 <ECN e1b 1 ceb 0 e0b 2001,nop,nop,nop,sack 1:1001>
28
+
29
+
+0.05 < [ect0] EAP. 1:1001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
30
+
+.002 > [ect0] EA. 1:1(0) ack 1001 <nop,nop,sack 1:1001>
31
+
32
+
+0.05 < [ect0] EAP. 1001:2001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
33
+
+.002 > [ect0] EA. 1:1(0) ack 2001
34
+
+0 read(4, ..., 1000) = 1000
+38
tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_lost.pkt
+38
tools/testing/selftests/net/packetdrill/tcp_accecn_client_accecn_options_lost.pkt
···
1
+
// Client negotiates AccECN and starts sending
2
+
// AccECN option in last ACK and data segments
3
+
// Middlebox accepts AccECN option but some packets
4
+
// are lost due to congestion. Client should
5
+
// continue to send AccECN option
6
+
7
+
`./defaults.sh
8
+
sysctl -q net.ipv4.tcp_ecn=3
9
+
sysctl -q net.ipv4.tcp_ecn_option=2
10
+
`
11
+
12
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
13
+
+.002 ... 0.102 connect(4, ..., ...) = 0
14
+
15
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
16
+
+0.1 < [ect0] SW. 0:0(0) ack 1 win 32767 <mss 1024,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
17
+
+.002 > [ect0] A. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
18
+
19
+
// Send
20
+
+0.01 write(4, ..., 3000) = 3000
21
+
+.002 > [ect0] .5 1:1013(1012) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
22
+
+.002 > [ect0] P.5 1013:2025(1012) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
23
+
+.002 > [ect0] P.5 2025:3001(976) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
24
+
25
+
// First two segments were lost due to congestion as SACK was
26
+
// received acknowledging 3rd segment
27
+
+0.1 < [ect0] .5 1:1(0) ack 1 win 264 <ECN e1b 1 ceb 0 e0b 977,nop,nop,nop,sack 2025:3001>
28
+
29
+
// Since data with option was SACKed, we can
30
+
// continue to use AccECN option for the rest of
31
+
// the connection. This one is a rexmt
32
+
+.02~+0.5 > [ect0] .5 1:1013(1012) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
33
+
+0.1 < [ect0] .5 1:1(0) ack 3001 win 264 <ECN e1b 1 ceb 0 e0b 3000,nop>
34
+
35
+
// Send new data, it should contain AccECN option
36
+
+0.01 write(4, ..., 2000) = 2000
37
+
+.002 > [ect0] .5 3001:4013(1012) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
38
+
+.002 > [ect0] P.5 4013:5001(988) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+12
tools/testing/selftests/net/packetdrill/tcp_accecn_clientside_disabled.pkt
+12
tools/testing/selftests/net/packetdrill/tcp_accecn_clientside_disabled.pkt
···
1
+
// AccECN sysctl server-side only, no ECN/AccECN
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=5
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < S. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,nop,nop,nop,wscale 8>
12
+
+.002 > . 1:1(0) ack 1
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_close_local_close_then_remote_fin.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_close_local_close_then_remote_fin.pkt
···
1
+
// Test basic connection teardown where local process closes first:
2
+
// the local process calls close() first, so we send a FIN, and receive an ACK.
3
+
// Then we receive a FIN and ACK it.
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
sysctl -q net.ipv4.tcp_ecn_option=0
8
+
`
9
+
10
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
11
+
+.01...0.011 connect(3, ..., ...) = 0
12
+
+0 > [noecn] SEWA 0:0(0) <...>
13
+
+0 < [ect1] SW. 0:0(0) ack 1 win 32768 <mss 1000,nop,wscale 6,nop,nop,sackOK>
14
+
+0 > [ect0] EW. 1:1(0) ack 1
15
+
16
+
+0 write(3, ..., 1000) = 1000
17
+
+0 > [ect0] P5. 1:1001(1000) ack 1
18
+
+0 < [ect0] .5 1:1(0) ack 1001 win 257
19
+
20
+
+0 close(3) = 0
21
+
+0 > [ect0] F5. 1001:1001(0) ack 1
22
+
+0 < [ect0] .5 1:1(0) ack 1002 win 257
23
+
24
+
+0 < [ect0] F5. 1:1(0) ack 1002 win 257
25
+
+0 > [ect0] . 1002:1002(0) ack 2
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_2ndlargeack.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_2ndlargeack.pkt
···
1
+
// Test a large ACK (> ACE field max)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 14600) = 14600
21
+
+.002 > [ect0] P.5 1:14601(14600) ack 1
22
+
+0.05 < [ect0] .5 1:1(0) ack 1461 win 264
23
+
+0.05 < [ect0] .5 1:1(0) ack 14601 win 264
24
+
25
+
+0.01 %{ assert tcpi_delivered_ce == 8, tcpi_delivered_ce }%
+31
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_falseoverflow_detect.pkt
+31
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_falseoverflow_detect.pkt
···
1
+
// Test false overflow detection with option used to rule out overflow
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
18
+
19
+
// Stop sending option to allow easier testing
20
+
+0 `sysctl -q net.ipv4.tcp_ecn_option=0`
21
+
22
+
+0.002 write(4, ..., 14600) = 14600
23
+
+.002 > [ect0] P.5 1:14601(14600) ack 1
24
+
25
+
+0.05 < [ect0] .5 1:1(0) ack 1460 win 264 <ECN e0b 1461 ceb 0 e1b 1,nop>
26
+
+0.05 < [ect0] .5 1:1(0) ack 14601 win 264 <ECN e0b 14601 ceb 0 e1b 1,nop>
27
+
28
+
+0.01 %{
29
+
assert tcpi_delivered_ce == 0, tcpi_delivered_ce
30
+
assert tcpi_delivered_e0_bytes == 14600, tcpi_delivered_e0_bytes
31
+
}%
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack.pkt
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack.pkt
···
1
+
// Test a large ACK (> ACE field max)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 14600) = 14600
21
+
+.002 > [ect0] P.5 1:14601(14600) ack 1
22
+
+0.05 < [ect0] .5 1:1(0) ack 14601 win 264
23
+
24
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack2.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_largeack2.pkt
···
1
+
// Test a large ACK (> ACE field max)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 14600) = 14600
21
+
+.002 > [ect0] P.5 1:14601(14600) ack 1
22
+
// Fake CE
23
+
+0.05 < [ect0] .6 1:1(0) ack 14601 win 264
24
+
25
+
+0.01 %{ assert tcpi_delivered_ce == 1, tcpi_delivered_ce }%
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_maxack.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_maxack.pkt
···
1
+
// Test a large ACK (at ACE field max delta)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 14600) = 14600
21
+
+.002 > [ect0] P.5 1:14601(14600) ack 1
22
+
// Fake CE
23
+
+0.05 < [ect0] .4 1:1(0) ack 14601 win 264
24
+
25
+
+0.01 %{ assert tcpi_delivered_ce == 7, tcpi_delivered_ce }%
+70
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_updates.pkt
+70
tools/testing/selftests/net/packetdrill/tcp_accecn_delivered_updates.pkt
···
1
+
// Test basic AccECN CEP/CEB/E0B/E1B functionality & CEP wrapping
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 %{
18
+
assert tcpi_delivered_ce == 0, tcpi_delivered_ce
19
+
assert tcpi_delivered_ce_bytes == 0, tcpi_delivered_ce_bytes
20
+
}%
21
+
22
+
+0.01 write(4, ..., 1000) = 1000
23
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
24
+
// Fake CE
25
+
+0.05 < [ect0] WA. 1:1(0) ack 1001 win 264 <ECN e0b 1 ceb 1000 e1b 1,nop>
26
+
27
+
+0.01 %{
28
+
assert tcpi_delivered_ce == 1, tcpi_delivered_ce
29
+
assert tcpi_delivered_ce_bytes == 1000, tcpi_delivered_ce_bytes
30
+
}%
31
+
32
+
+0.01 write(4, ..., 1000) = 1000
33
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
34
+
// Fake ect0
35
+
+0.05 < [ect0] WA. 1:1(0) ack 2001 win 264 <ECN e0b 1001 ceb 1000 e1b 1,nop>
36
+
37
+
+0.01 %{
38
+
assert tcpi_delivered_ce == 1, tcpi_delivered_ce
39
+
assert tcpi_delivered_e0_bytes == 1000, tcpi_delivered_e0_bytes
40
+
}%
41
+
42
+
+0.01 write(4, ..., 1000) = 1000
43
+
+.002 > [ect0] EAP. 2001:3001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
44
+
// Fake ce
45
+
+0.05 < [ect0] EWA. 1:1(0) ack 3001 win 264 <ECN e0b 1001 ceb 2000 e1b 1,nop>
46
+
47
+
+0.01 %{
48
+
assert tcpi_delivered_ce == 2, tcpi_delivered_ce
49
+
assert tcpi_delivered_ce_bytes == 2000, tcpi_delivered_ce_bytes
50
+
}%
51
+
52
+
+0.01 write(4, ..., 1000) = 1000
53
+
+.002 > [ect0] EAP. 3001:4001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
54
+
// Fake ect1
55
+
+0.05 < [ect0] EWA. 1:1(0) ack 4001 win 264 <ECN e0b 1001 ceb 2000 e1b 1001,nop>
56
+
57
+
+0.01 %{
58
+
assert tcpi_delivered_ce == 2, tcpi_delivered_ce
59
+
assert tcpi_delivered_e1_bytes == 1000, tcpi_delivered_e1_bytes
60
+
}%
61
+
62
+
+0.01 write(4, ..., 1000) = 1000
63
+
+.002 > [ect0] EAP. 4001:5001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
64
+
// Fake ce
65
+
+0.05 < [ect0] . 1:1(0) ack 5001 win 264 <ECN e0b 1001 ceb 3000 e1b 1001,nop>
66
+
67
+
+0.01 %{
68
+
assert tcpi_delivered_ce == 3, tcpi_delivered_ce
69
+
assert tcpi_delivered_ce_bytes == 3000, tcpi_delivered_ce_bytes
70
+
}%
+12
tools/testing/selftests/net/packetdrill/tcp_accecn_ecn3.pkt
+12
tools/testing/selftests/net/packetdrill/tcp_accecn_ecn3.pkt
···
1
+
// Test that tcp_ecn=4 uses RFC3168 ECN for SYN
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=4
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.05 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEW 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > . 1:1(0) ack 1
+35
tools/testing/selftests/net/packetdrill/tcp_accecn_ecn_field_updates_opt.pkt
+35
tools/testing/selftests/net/packetdrill/tcp_accecn_ecn_field_updates_opt.pkt
···
1
+
// Test basic AccECN CEP/CEB/E0B/E1B functionality & CEP wrapping
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.05 < [ce] EAP. 1:1001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
18
+
+.002 > [ect0] WA. 1:1(0) ack 1001 <ECN e1b 1 ceb 1000 e0b 1,nop>
19
+
+0 read(4, ..., 1000) = 1000
20
+
21
+
+0.05 < [ect0] EAP. 1001:2001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
22
+
+.002 > [ect0] WA. 1:1(0) ack 2001 <ECN e1b 1 ceb 1000 e0b 1001,nop>
23
+
+0 read(4, ..., 1000) = 1000
24
+
25
+
+0.05 < [ce] EAP. 2001:3001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
26
+
+.002 > [ect0] EWA. 1:1(0) ack 3001 <ECN e1b 1 ceb 2000 e0b 1001,nop>
27
+
+0 read(4, ..., 1000) = 1000
28
+
29
+
+0.05 < [ect1] EAP. 3001:4001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
30
+
+.002 > [ect0] EWA. 1:1(0) ack 4001 <ECN e1b 1001 ceb 2000 e0b 1001,nop>
31
+
+0 read(4, ..., 1000) = 1000
32
+
33
+
+0.05 < [ce] EAP. 4001:5001(1000) ack 1 win 257 <ECN e0b 1 ceb 0 e1b 1,nop>
34
+
+.002 > [ect0] . 1:1(0) ack 5001 <ECN e1b 1001 ceb 3000 e0b 1001,nop>
35
+
+0 read(4, ..., 1000) = 1000
+14
tools/testing/selftests/net/packetdrill/tcp_accecn_ipflags_drop.pkt
+14
tools/testing/selftests/net/packetdrill/tcp_accecn_ipflags_drop.pkt
···
1
+
// Test IP flags drop
2
+
--tolerance_usecs=50000
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
9
+
+.002 ... 1.1 connect(4, ..., ...) = 0
10
+
11
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
12
+
+.02 ~ +1.1 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
13
+
+0.05 < S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > [noecn] . 1:1(0) ack 1
+16
tools/testing/selftests/net/packetdrill/tcp_accecn_listen_opt_drop.pkt
+16
tools/testing/selftests/net/packetdrill/tcp_accecn_listen_opt_drop.pkt
···
1
+
// SYN/ACK option drop test
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+.02 ~+2 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+.02 ~+5 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
16
+
+.02 ~+8 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_ack_drop.pkt
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_ack_drop.pkt
···
1
+
// Test that SYN-ACK with ACE flags and without
2
+
// ACE flags got dropped. Although we disable ECN,
3
+
// we shouldn't consider this as blackholed as
4
+
// these are dropped due to congestion
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
sysctl -q net.ipv4.tcp_ecn_option=2
9
+
`
10
+
11
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
12
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
13
+
+0 bind(3, ..., ...) = 0
14
+
+0 listen(3, 1) = 0
15
+
16
+
+0 < [noecn] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
17
+
+.002 > [noecn] SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
18
+
19
+
// Retransmit SYN-ACK without option
20
+
+1~+1.1 > [noecn] SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
21
+
22
+
// SYN-ACK maybe getting blackholed, disable ECN
23
+
+2~+2.2 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
24
+
+4~+4.4 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
25
+
26
+
// Received an ACK after sending 3rd retransmission, not a blackhole
27
+
+0.1 < [noecn] . 1:1(0) ack 1 win 320
28
+
+.002 accept(3, ..., ...) = 4
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_drop.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_multiple_syn_drop.pkt
···
1
+
// Test that SYN with ACE flags and without
2
+
// ACE flags got dropped. Although we disable
3
+
// ECN, we shouldn't consider this as blackholed
4
+
// as these are dropped due to congestion
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
`
9
+
10
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
11
+
+.002 ... 3.1 connect(4, ..., ...) = 0
12
+
13
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
14
+
+.02~+1.1 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
15
+
+.02~+1.1 > [noecn] S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
16
+
+.02~+1.1 > [noecn] S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
17
+
+0.1 < [noecn] S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
18
+
+0~+0.01 > [noecn] . 1:1(0) ack 1
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_bleach.pkt
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_bleach.pkt
···
1
+
// Test AccECN flags bleach
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] . 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
18
+
19
+
+0.01 write(4, ..., 1000) = 1000
20
+
+.002 > [noecn] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
21
+
+0.05 < [ect0] EAP. 1:1(0) ack 1001 win 320
22
+
23
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_connect.pkt
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_connect.pkt
···
1
+
// Test basic AccECN negotiation
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
13
+
14
+
+0.01 write(4, ..., 1000) = 1000
15
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
16
+
+.05 < [ect0] EAP. 1:1(0) ack 1001 win 256 <ECN e0b 1001 ceb 0 e1b 0,nop>
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 1000) = 1000
21
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
22
+
23
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_listen.pkt
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_listen.pkt
···
1
+
// Test basic AccECN negotiation
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
18
+
19
+
+0.01 write(4, ..., 1000) = 1000
20
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
21
+
+0.05 < [ect0] EAP. 1:1(0) ack 1001 win 320
22
+
23
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
24
+
25
+
+0.01 write(4, ..., 1000) = 1000
26
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_noopt_connect.pkt
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_noopt_connect.pkt
···
1
+
// Test basic AccECN negotiation without option
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < SW. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] W. 1:1(0) ack 1
13
+
14
+
+0.01 write(4, ..., 1000) = 1000
15
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1
16
+
+.05 < [ect0] EAP. 1:1(0) ack 1001 win 256
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 1000) = 1000
21
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1
22
+
23
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_optenable.pkt
+23
tools/testing/selftests/net/packetdrill/tcp_accecn_negotiation_optenable.pkt
···
1
+
// Test basic AccECN negotiation, late option enable
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < SW. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] W. 1:1(0) ack 1
13
+
14
+
+0.01 write(4, ..., 1000) = 1000
15
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1
16
+
+.05 < [ect0] EAP. 1:1(0) ack 1001 win 256 <ECN e0b 1001 ceb 0 e1b 1,nop>
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 1000) = 1000
21
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
22
+
23
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_no_ecn_after_accecn.pkt
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_no_ecn_after_accecn.pkt
···
1
+
// Test client behavior on receiving a non ECN SYN-ACK
2
+
// after receiving an AccECN SYN-ACK and moving to
3
+
// ESTABLISHED state
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
sysctl -q net.ipv4.tcp_ecn_option=2
8
+
`
9
+
10
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
11
+
+.002 ... 0.052 connect(4, ..., ...) = 0
12
+
13
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
14
+
// Receive an AccECN SYN-ACK and move to ESTABLISHED
15
+
+0.05 < [noecn] SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
16
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
17
+
18
+
// Receive a non ECN SYN-ACK and send a challenge ACK with ACE feedback
19
+
+0.1 < [noecn] S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
20
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_noopt.pkt
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_noopt.pkt
···
1
+
// Test basic AccECN negotiation with option off using sysctl
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 1000) = 1000
21
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1
22
+
+0.05 < [ect0] EAP. 1:1(0) ack 1001 win 320
23
+
24
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
25
+
26
+
+0.01 write(4, ..., 1000) = 1000
27
+
+.002 > [ect0] EAP. 1001:2001(1000) ack 1
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_noprogress.pkt
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_noprogress.pkt
···
1
+
// Test no progress filtering
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
18
+
19
+
+0.01 write(4, ..., 1000) = 1000
20
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
21
+
// Fake CE and claim no progress
22
+
+0.05 < [ect0] WA. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 1000 e1b 1,nop>
23
+
24
+
+0.01 %{
25
+
assert tcpi_delivered_ce == 0, tcpi_delivered_ce
26
+
assert tcpi_delivered_ce_bytes == 0, tcpi_delivered_ce_bytes
27
+
}%
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_notecn_then_accecn_syn.pkt
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_notecn_then_accecn_syn.pkt
···
1
+
// Test that SYN-ACK with ACE flags and without
2
+
// ACE flags got dropped. Although we disable ECN,
3
+
// we shouldn't consider this as blackholed as
4
+
// these are dropped due to congestion
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
sysctl -q net.ipv4.tcp_ecn_option=2
9
+
`
10
+
11
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
12
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
13
+
+0 bind(3, ..., ...) = 0
14
+
+0 listen(3, 1) = 0
15
+
16
+
+0 < [noecn] S 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
17
+
+.002 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
18
+
19
+
// Retransmit SYN
20
+
+0.1 < [ect0] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
21
+
+.002 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
22
+
23
+
+0.1 < [noecn] . 1:1(0) ack 1 win 320
24
+
+.002 accept(3, ..., ...) = 4
25
+
26
+
// Write with AccECN option but with ip-noecn since we received one SYN with ACE=0
27
+
+0.01 write(4, ..., 100) = 100
28
+
+.002 > [noecn] P. 1:101(100) ack 1
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_fallback.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_fallback.pkt
···
1
+
// Test RFC3168 fallback when sysctl asks for AccECN
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEW 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SE. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < . 1:1(0) ack 1 win 320
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 write(4, ..., 1000) = 1000
18
+
+.002 > [ect0] P. 1:1001(1000) ack 1
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_rfc3168.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_rfc3168_to_rfc3168.pkt
···
1
+
// Test RFC3168 ECN when sysctl asks for RFC3168 ECN
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=1
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEW 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SE. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < . 1:1(0) ack 1 win 320
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 write(4, ..., 1000) = 1000
18
+
+.002 > [ect0] P. 1:1001(1000) ack 1
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab.pkt
+28
tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab.pkt
···
1
+
// Test SACK space grab to fit AccECN option
2
+
--tcp_ts_tick_usecs=1000
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+.01 < [ect1] EAP. 1001:2001(1000) ack 1 win 264
19
+
+0.002 > [ect0] EA. 1:1(0) ack 1 <ECN e1b 1001 ceb 0 e0b 1,nop,nop,nop,sack 1001:2001>
20
+
+.01 < [ect0] EAP. 3001:4001(1000) ack 1 win 264
21
+
+0.002 > [ect0] EA. 1:1(0) ack 1 <ECN e1b 1001 ceb 0 e0b 1001,nop,nop,nop,sack 3001:4001 1001:2001>
22
+
+.01 < [ce] EAP. 5001:6001(1000) ack 1 win 264
23
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <ECN e1b 1001 ceb 1000 e0b 1001,nop,nop,nop,sack 5001:6001 3001:4001 1001:2001>
24
+
// DSACK works?
25
+
+.01 < [ect0] EAP. 5001:6001(1000) ack 1 win 264
26
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <ECN e1b 1001 ceb 1000 e0b 2001,nop,nop,nop,sack 5001:6001 5001:6001 3001:4001>
27
+
+.01 < [ect1] EAP. 6001:7001(1000) ack 1 win 264
28
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <ECN e1b 2001 ceb 1000 e0b 2001,nop,nop,nop,sack 5001:7001 3001:4001 1001:2001>
+39
tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab_with_ts.pkt
+39
tools/testing/selftests/net/packetdrill/tcp_accecn_sack_space_grab_with_ts.pkt
···
1
+
// Test SACK space grab to fit AccECN option
2
+
--tcp_ts_tick_usecs=1000
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,sackOK,TS val 1 ecr 0,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,sackOK,TS val 100 ecr 1,ECN e1b 1 ceb 0 e0b 1,nop,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <nop,nop,TS val 2 ecr 100,ECN e0b 1 ceb 0 e1b 1,nop>
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
// One SACK block should allow all 3 AccECN fields:
19
+
+.01 < [ect1] EAP. 1001:2001(1000) ack 1 win 264 <nop,nop,TS val 3 ecr 100>
20
+
+0.002 > [ect0] EA. 1:1(0) ack 1 <nop,nop,TS val 160 ecr 2,ECN e1b 1001 ceb 0 e0b 1,nop,nop,nop,sack 1001:2001>
21
+
22
+
// Two SACK blocks should fit w/ AccECN if we only need to use 2 AccECN fields: check ect1 arriving.
23
+
+.01 < [ect1] EAP. 3001:4001(1000) ack 1 win 264 <nop,nop,TS val 4 ecr 100>
24
+
+0.002 > [ect0] EA. 1:1(0) ack 1 <nop,nop,TS val 172 ecr 2,ECN e1b 2001 ceb 0,nop,nop,sack 3001:4001 1001:2001>
25
+
26
+
// Two SACK blocks should fit w/ AccECN if we only need to use 2 AccECN fields: check CE arriving.
27
+
+.01 < [ce] EAP. 5001:6001(1000) ack 1 win 264 <nop,nop,TS val 5 ecr 100>
28
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <nop,nop,TS val 184 ecr 2,ECN e1b 2001 ceb 1000,nop,nop,sack 5001:6001 3001:4001>
29
+
30
+
// Check that DSACK works, using 2 SACK blocks in total, if we only need to use 2 AccECN fields: check ect1 arriving.
31
+
+.01 < [ect1] EAP. 5001:6001(1000) ack 1 win 264 <nop,nop,TS val 5 ecr 100>
32
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <nop,nop,TS val 196 ecr 2,ECN e1b 3001 ceb 1000,nop,nop,sack 5001:6001 5001:6001>
33
+
34
+
// Check the case where the AccECN option doesn't fit, because sending ect0
35
+
// with order 1 would rquire 3 AccECN fields,
36
+
// and TS (12 bytes) + 2 SACK blocks (20 bytes) + 3 AccECN fields (2 + 3*3 bytes) > 40 bytes.
37
+
// That's OK; Linux TCP AccECN is optimized for the ECT1 case, not ECT0.
38
+
+.01 < [ect0] EAP. 6001:7001(1000) ack 1 win 264 <nop,nop,TS val 5 ecr 100>
39
+
+0.002 > [ect0] WA. 1:1(0) ack 1 <nop,nop,TS val 204 ecr 2,nop,nop,sack 5001:7001 3001:4001 1001:2001>
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled1.pkt
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled1.pkt
···
1
+
// Test against classic ECN server
2
+
// Not-ECT on SYN and server sets 1|0|1 (AE is unused for classic ECN)
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
9
+
+.002 ... 0.052 connect(4, ..., ...) = 0
10
+
11
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
12
+
+0.05 < [noecn] SEA. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,TS val 700 ecr 100,nop,wscale 8>
13
+
+.002 > [ect0] W. 1:1(0) ack 1 <nop, nop, TS val 200 ecr 700>
14
+
15
+
+0 write(4, ..., 100) = 100
16
+
+.002 > [ect0] P.5 1:101(100) ack 1 <nop,nop,TS val 300 ecr 700>
17
+
+0 close(4) = 0
18
+
19
+
+.002 > [ect0] F.5 101:101(0) ack 1 <nop,nop,TS val 400 ecr 700>
20
+
+0.1 < [noecn] R. 1:1(0) ack 102 win 4242
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled2.pkt
+20
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_accecn_disabled2.pkt
···
1
+
// Test against classic ECN server
2
+
// Not-ECT on SYN and server sets 0|0|1
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
9
+
+.002 ... 0.052 connect(4, ..., ...) = 0
10
+
11
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
12
+
+0.05 < [noecn] SE. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,TS val 700 ecr 100,nop,wscale 8>
13
+
+.002 > [noecn] . 1:1(0) ack 1 <nop, nop, TS val 200 ecr 700>
14
+
15
+
+0 write(4, ..., 100) = 100
16
+
+.002 > [ect0] P. 1:101(100) ack 1 <nop,nop,TS val 300 ecr 700>
17
+
+0 close(4) = 0
18
+
19
+
+0 > [noecn] F. 101:101(0) ack 1 <...>
20
+
+0.1 < R. 1:1(0) ack 102 win 4242
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_broken.pkt
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_broken.pkt
···
1
+
// Test against broken server (1|1|1)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [noecn] SEWA. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,TS val 700 ecr 100,nop,wscale 8>
12
+
+.002 > [noecn] . 1:1(0) ack 1 <nop, nop, TS val 200 ecr 700>
13
+
14
+
+0 write(4, ..., 100) = 100
15
+
+.002 > [noecn] P. 1:101(100) ack 1 <nop,nop,TS val 300 ecr 700>
16
+
+0 close(4) = 0
17
+
18
+
+.002 > [noecn] F. 101:101(0) ack 1 <...>
19
+
+0.1 < [noecn] R. 1:1(0) ack 102 win 4242
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_ecn_disabled.pkt
+19
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_ecn_disabled.pkt
···
1
+
// Test against Non ECN server (0|0|0)
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [noecn] S. 0:0(0) ack 1 win 32767 <mss 1460,sackOK,TS val 700 ecr 100,nop,wscale 8>
12
+
+.002 > [noecn] . 1:1(0) ack 1 <nop, nop, TS val 200 ecr 700>
13
+
14
+
+0 write(4, ..., 100) = 100
15
+
+.002 > [noecn] P. 1:101(100) ack 1 <nop,nop,TS val 300 ecr 700>
16
+
+0 close(4) = 0
17
+
18
+
+.002 > [noecn] F. 101:101(0) ack 1 <nop,nop,TS val 400 ecr 700>
19
+
+0.1 < [noecn] R. 1:1(0) ack 102 win 4242
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_only.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_serverside_only.pkt
···
1
+
// Test AccECN with sysctl set to server-side only
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=5
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
14
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
15
+
+.002 accept(3, ..., ...) = 4
16
+
17
+
+0.01 write(4, ..., 1000) = 1000
18
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_acked_after_retransmit.pkt
+18
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_acked_after_retransmit.pkt
···
1
+
// Test that SYN with ACE flags was Acked
2
+
// after 2nd retransmission. In this case,
3
+
// since we got SYN-ACK that supports Accurate
4
+
// ECN, we consider this as successful negotiation
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
`
9
+
10
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
11
+
+.002 ... 2.1 connect(4, ..., ...) = 0
12
+
13
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
14
+
+1~+1.1 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
15
+
+1~+1.1 > [noecn] S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
16
+
17
+
+0.1 < [noecn] SW. 0:0(0) ack 1 win 32767 <mss 1016,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
18
+
+0~+0.01 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+16
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_drop.pkt
+16
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ace_flags_drop.pkt
···
1
+
// Test that SYN with ACE flags got dropped
2
+
// We retry one more time with ACE and then
3
+
// fallback to disabled ECN
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
`
8
+
9
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
10
+
+.002 ... 2.1 connect(4, ..., ...) = 0
11
+
12
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
13
+
+1~+1.1 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
14
+
+1~+1.1 > [noecn] S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
15
+
+0.1 < [noecn] S. 0:0(0) ack 1 win 32767 <mss 1460,nop,nop,sackOK,nop,wscale 8>
16
+
+0~+0.01 > [noecn] . 1:1(0) ack 1
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_acked_after_retransmit.pkt
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_acked_after_retransmit.pkt
···
1
+
// Test that SYN-ACK with ACE flags was Acked
2
+
// after 2nd retransmission. In this case,
3
+
// since we got the last ACK that supports Accurate
4
+
// ECN, we consider this as successful negotiation
5
+
6
+
`./defaults.sh
7
+
sysctl -q net.ipv4.tcp_ecn=3
8
+
sysctl -q net.ipv4.tcp_ecn_option=2
9
+
`
10
+
11
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
12
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
13
+
+0 bind(3, ..., ...) = 0
14
+
+0 listen(3, 1) = 0
15
+
16
+
+0 < [noecn] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
17
+
+.002 > [noecn] SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
18
+
19
+
// Retransmit SYN-ACK without option
20
+
+1~+1.1 > [noecn] SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
21
+
22
+
// SYN-ACK maybe getting blackholed, disable ECN
23
+
+2~+2.2 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
24
+
25
+
// Received an ACK with ACE flags, state should be set to negotiation succeeded
26
+
+0.1 < [noecn] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
27
+
+.002 accept(3, ..., ...) = 4
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_drop.pkt
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ack_ace_flags_drop.pkt
···
1
+
// Test that SYN-ACK with ACE flags got dropped
2
+
// We retry one more time with ACE and then
3
+
// fallback to disabled ECN
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
sysctl -q net.ipv4.tcp_ecn_option=2
8
+
`
9
+
10
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
11
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
12
+
+0 bind(3, ..., ...) = 0
13
+
+0 listen(3, 1) = 0
14
+
15
+
+0 < [noecn] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
16
+
+.002 > [noecn] SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
17
+
18
+
// Retransmit SYN-ACK without option
19
+
+1~+1.1 > [noecn] SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
20
+
21
+
// SYN-ACK maybe getting blackholed, disable ECN
22
+
+2~+2.2 > [noecn] S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
23
+
24
+
// Received an ACK with no ACE flags, state should be set to blackholed
25
+
+0.1 < [noecn] . 1:1(0) ack 1 win 320
26
+
+0 accept(3, ..., ...) = 4
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ce.pkt
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ce.pkt
···
1
+
// Test AccECN ECN field reflector in SYNACK
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < [ce] SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SWA. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect0.pkt
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect0.pkt
···
1
+
// Test AccECN ECN field reflector in SYNACK
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < [ect0] SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SA. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect1.pkt
+13
tools/testing/selftests/net/packetdrill/tcp_accecn_syn_ect1.pkt
···
1
+
// Test AccECN ECN field reflector in SYNACK
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
8
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
9
+
+0 bind(3, ..., ...) = 0
10
+
+0 listen(3, 1) = 0
11
+
12
+
+0 < [ect1] SEWA 0:0(0) win 32792 <mss 1050,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > SEW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce.pkt
+27
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce.pkt
···
1
+
// Test SYNACK CE & received_ce update
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_option=2
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
9
+
+.002 ... 0.052 connect(4, ..., ...) = 0
10
+
11
+
+.002 > [noecn] SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
12
+
+0.05 < [ce] SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
13
+
+.002 > [ect0] WA. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
14
+
15
+
+0.01 write(4, ..., 100) = 100
16
+
+.002 > [ect0] P.6 1:101(100) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
17
+
+0.05 < [ect0] P.5 1:101(100) ack 101 win 256 <ECN e0b 101 ceb 0 e1b 1,nop>
18
+
+.002 > [ect0] .6 101:101(0) ack 101 <ECN e1b 1 ceb 0 e0b 101,nop>
19
+
20
+
+0.01 write(4, ..., 100) = 100
21
+
+.002 > [ect0] P.6 101:201(100) ack 101 <ECN e1b 1 ceb 0 e0b 101,nop>
22
+
23
+
+0.1 < [ect1] P.5 201:301(100) ack 201 win 256 <ECN e0b 101 ceb 0 e1b 1,nop>
24
+
+.002 > [ect0] .6 201:201(0) ack 101 <ECN e1b 101 ceb 0 e0b 101,nop,nop,nop,sack 201:301>
25
+
26
+
+0.01 < [ce] .6 401:501(100) ack 201 win 256 <ECN e0b 101 ceb 0 e1b 1,nop>
27
+
+.002 > [ect0] .7 201:201(0) ack 101 <ECN e1b 101 ceb 100 e0b 101,nop,nop,nop,sack 401:501 201:301>
+22
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce_updates_delivered_ce.pkt
+22
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ce_updates_delivered_ce.pkt
···
1
+
// Reflected SYNACK CE mark increases delivered_ce
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
sysctl -q net.ipv4.tcp_ecn_fallback=0
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0.05 < SEWA 0:0(0) win 32767 <mss 1050,nop,nop,sackOK,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
15
+
// Fake ce for prev, ECT validator must be disabled for this to work
16
+
+0.05 < [ect0] WA. 1:1(0) ack 1 win 264 <ECN e0b 1 ceb 0 e1b 1,nop>
17
+
+.002 accept(3, ..., ...) = 4
18
+
19
+
+0.01 %{ assert tcpi_delivered_ce == 1, tcpi_delivered_ce }%
20
+
21
+
+0.01 write(4, ..., 1000) = 1000
22
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect0.pkt
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect0.pkt
···
1
+
// Test SYN=0 reflector
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [ect0] SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] A. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
13
+
14
+
+0.01 write(4, ..., 100) = 100
15
+
+.002 > [ect0] P.5 1:101(100) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
16
+
+0.05 < [ect0] P.5 1:1(0) ack 101 win 256 <ECN e0b 101 ceb 0 e1b 1,nop>
17
+
18
+
+0.01 < [ect0] P.5 1:101(100) ack 101 win 256 <ECN e0b 1 ceb 0 e1b 1,nop>
19
+
+.002 > [ect0] .5 101:101(0) ack 101 <ECN e1b 1 ceb 0 e0b 101,nop>
20
+
+0 read(4, ..., 100) = 100
21
+
22
+
+0 close(4) = 0
23
+
+0 > F.5 101:101(0) ack 101 <...>
24
+
+0.1 < R. 101:101(0) ack 102 win 4242
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect1.pkt
+24
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_ect1.pkt
···
1
+
// Test SYN=0 reflector
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < [ect1] SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] EW. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
13
+
14
+
+0.01 write(4, ..., 100) = 100
15
+
+.002 > [ect0] P.5 1:101(100) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
16
+
+0.05 < [ect1] P.5 1:1(0) ack 101 win 256 <ECN e0b 101 ceb 0 e1b 1,nop>
17
+
18
+
+0.01 < [ect1] P.5 1:101(100) ack 101 win 256 <ECN e0b 1 ceb 0 e1b 1,nop>
19
+
+.002 > [ect0] .5 101:101(0) ack 101 <ECN e1b 101 ceb 0 e0b 1,nop>
20
+
+0 read(4, ..., 100) = 100
21
+
22
+
+0 close(4) = 0
23
+
+0 > F5. 101:101(0) ack 101 <...>
24
+
+0.1 < R. 101:101(0) ack 102 win 4242
+15
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rexmit.pkt
+15
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rexmit.pkt
···
1
+
// Test 3rd ACK flags when SYN-ACK is rexmitted
2
+
3
+
`./defaults.sh
4
+
sysctl -q net.ipv4.tcp_ecn=3
5
+
`
6
+
7
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
8
+
+.002 ... 0.052 connect(4, ..., ...) = 0
9
+
10
+
+.002 > SEWA 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8>
11
+
+0.05 < SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
12
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
13
+
14
+
+0.05 < SW. 0:0(0) ack 1 win 32767 <mss 1460,ECN e0b 1 ceb 0 e1b 1,nop,nop,nop,sackOK,nop,wscale 8>
15
+
+.002 > [ect0] W. 1:1(0) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rxmt.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_synack_rxmt.pkt
···
1
+
// Test that we retransmit SYN-ACK with ACE and without
2
+
// AccECN options after
3
+
// SYN-ACK was lost and TCP moved to TCPS_SYN_RECEIVED
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
sysctl -q net.ipv4.tcp_ecn_option=2
8
+
`
9
+
10
+
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
11
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
12
+
+0 bind(3, ..., ...) = 0
13
+
+0 listen(3, 1) = 0
14
+
15
+
+0 < [noecn] SEWA 0:0(0) win 32792 <mss 1460,nop,nop,sackOK,nop,wscale 8>
16
+
+.002 > [noecn] SW. 0:0(0) ack 1 <mss 1460,ECN e1b 1 ceb 0 e0b 1,nop,nop,nop,sackOK,nop,wscale 8>
17
+
18
+
// Retransmit SYN-ACK without option
19
+
+1~+1.1 > [noecn] SW. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
20
+
+0.1 < [noecn] W. 1:1(0) ack 1 win 320 <ECN e0b 1 ceb 0 e1b 1,nop>
21
+
+.002 accept(3, ..., ...) = 4
22
+
23
+
// We try to write with AccECN option
24
+
+0.01 write(4, ..., 100) = 100
25
+
+.002 > [ect0] P5. 1:101(100) ack 1 <ECN e1b 1 ceb 0 e0b 1,nop>
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_tsnoprogress.pkt
+26
tools/testing/selftests/net/packetdrill/tcp_accecn_tsnoprogress.pkt
···
1
+
// Test TS progress filtering
2
+
--tcp_ts_tick_usecs=1000
3
+
--tolerance_usecs=7000
4
+
5
+
`./defaults.sh
6
+
sysctl -q net.ipv4.tcp_ecn=3
7
+
`
8
+
9
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
10
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
11
+
+0 bind(3, ..., ...) = 0
12
+
+0 listen(3, 1) = 0
13
+
14
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,sackOK,TS val 1 ecr 0,nop,wscale 8>
15
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,sackOK,TS val 10 ecr 1,ECN e1b 1 ceb 0 e0b 1,nop,nop,wscale 8>
16
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <nop,nop,TS val 2 ecr 10>
17
+
+.002 accept(3, ..., ...) = 4
18
+
19
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
20
+
21
+
+0.01 write(4, ..., 1000) = 1000
22
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <nop,nop,TS val 83 ecr 2>
23
+
// Fake CE and claim no progress
24
+
+0.05 < [ect0] WA. 1:1(0) ack 1 win 264 <nop,nop,TS val 2 ecr 83>
25
+
26
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_tsprogress.pkt
+25
tools/testing/selftests/net/packetdrill/tcp_accecn_tsprogress.pkt
···
1
+
// Test TS progress filtering
2
+
--tcp_ts_tick_usecs=1000
3
+
4
+
`./defaults.sh
5
+
sysctl -q net.ipv4.tcp_ecn=3
6
+
`
7
+
8
+
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
9
+
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
10
+
+0 bind(3, ..., ...) = 0
11
+
+0 listen(3, 1) = 0
12
+
13
+
+0 < SEWA 0:0(0) win 32792 <mss 1050,sackOK,TS val 1 ecr 0,nop,wscale 8>
14
+
+.002 > SW. 0:0(0) ack 1 <mss 1460,sackOK,TS val 10 ecr 1,ECN e1b 1 ceb 0 e0b 1,nop,nop,wscale 8>
15
+
+0.05 < [ect0] W. 1:1(0) ack 1 win 264 <nop,nop,TS val 2 ecr 10>
16
+
+.002 accept(3, ..., ...) = 4
17
+
18
+
+0.01 %{ assert tcpi_delivered_ce == 0, tcpi_delivered_ce }%
19
+
20
+
+0.01 write(4, ..., 1000) = 1000
21
+
+.002 > [ect0] EAP. 1:1001(1000) ack 1 <nop,nop,TS val 83 ecr 2>
22
+
// Fake CE and claim no progress
23
+
+0.05 < [ect0] WA. 1:1(0) ack 1 win 264 <nop,nop,TS val 3 ecr 83>
24
+
25
+
+0.01 %{ assert tcpi_delivered_ce == 1, tcpi_delivered_ce }%