selftests/net: packetdrill: import multiple tests
Subsystems (3)
cover selftests/net: packetdrill: import multiple tests
From: Soham Chakradeo <[email protected]>
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO (tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018
Recent Timestamp (ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)
Procedure to import and test the packetdrill tests into upstream linux is explained in the first patch of this series
These tests have many authors. We only import them here from github.com/google/packetdrill. Thanks to the following authors fo their contributions over the years to these tests: Neal Cardwell, Shuo Chen, Yuchung Cheng, Jerry Chu, Eric Dumazet, Luke Hsiao, Priyaranjan Jha, Chonggang Li, Tanner Love, John Sperbeck, Wei Wang and Maciej Żenczykowski. For more info see the original github commits, such as
https://github.com/google/packetdrill/commit/8229c94928ac.
Soham Chakradeo (4):
selftests/net: packetdrill: import tcp/ecn , tcp/close , tcp/sack ,
tcp/tcp_info
selftests/net: packetdrill: import tcp/fast_recovery , tcp/nagle ,
tcp/timestamping
selftests/net: packetdrill: import tcp/eor , tcp/splice ,
tcp/ts_recent , tcp/blocking
selftests/net: packetdrill: import tcp/user_timeout , tcp/validate ,
tcp/sendfile , tcp/limited-transmit , tcp/syscall_bad_arg
.../tcp_blocking_blocking-accept.pkt | 18 +++ .../tcp_blocking_blocking-connect.pkt | 13 ++ .../tcp_blocking_blocking-read.pkt | 29 ++++ .../tcp_blocking_blocking-write.pkt | 35 +++++ ...lose_close-local-close-then-remote-fin.pkt | 23 +++ .../tcp_close_close-on-syn-sent.pkt | 21 +++ .../tcp_close_close-remote-fin-then-close.pkt | 36 +++++ .../net/packetdrill/tcp_ecn_ecn-uses-ect0.pkt | 21 +++ .../packetdrill/tcp_eor_no-coalesce-large.pkt | 38 +++++ .../tcp_eor_no-coalesce-retrans.pkt | 72 +++++++++ .../packetdrill/tcp_eor_no-coalesce-small.pkt | 36 +++++ .../tcp_eor_no-coalesce-subsequent.pkt | 66 ++++++++ .../tcp_fast_recovery_prr-ss-10pkt-lost-1.pkt | 72 +++++++++ ...t_recovery_prr-ss-30pkt-lost-1_4-11_16.pkt | 50 ++++++ ...tcp_fast_recovery_prr-ss-30pkt-lost1_4.pkt | 43 ++++++ ...ecovery_prr-ss-ack-below-snd_una-cubic.pkt | 41 +++++ ...ited_transmit_limited-transmit-no-sack.pkt | 53 +++++++ ...limited_transmit_limited-transmit-sack.pkt | 50 ++++++ .../packetdrill/tcp_nagle_https_client.pkt | 40 +++++ .../tcp_nagle_sendmsg_msg_more.pkt | 66 ++++++++ .../tcp_nagle_sockopt_cork_nodelay.pkt | 43 ++++++ .../tcp_sack_sack-route-refresh-ip-tos.pkt | 37 +++++ ...ack_sack-shift-sacked-2-6-8-3-9-nofack.pkt | 64 ++++++++ ..._sack_sack-shift-sacked-7-3-4-8-9-fack.pkt | 66 ++++++++ ..._sack_sack-shift-sacked-7-5-6-8-9-fack.pkt | 62 ++++++++ .../tcp_sendfile_sendfile-simple.pkt | 26 ++++ .../tcp_splice_tcp_splice_loop_test.pkt | 20 +++ ...scall_bad_arg_fastopen-invalid-buf-ptr.pkt | 42 +++++ .../tcp_syscall_bad_arg_sendmsg-empty-iov.pkt | 31 ++++ ...yscall_bad_arg_syscall-invalid-buf-ptr.pkt | 25 +++ .../tcp_tcp_info_tcp-info-last_data_recv.pkt | 21 +++ .../tcp_tcp_info_tcp-info-rwnd-limited.pkt | 54 +++++++ .../tcp_tcp_info_tcp-info-sndbuf-limited.pkt | 39 +++++ ...tcp_timestamping_client-only-last-byte.pkt | 92 +++++++++++ .../packetdrill/tcp_timestamping_partial.pkt | 91 +++++++++++ .../packetdrill/tcp_timestamping_server.pkt | 145 ++++++++++++++++++ .../packetdrill/tcp_ts_recent_fin_tsval.pkt | 23 +++ .../packetdrill/tcp_ts_recent_invalid_ack.pkt | 25 +++ .../packetdrill/tcp_ts_recent_reset_tsval.pkt | 25 +++ .../tcp_user_timeout_user-timeout-probe.pkt | 37 +++++ .../tcp_user_timeout_user_timeout.pkt | 33 ++++ ...validate_validate-established-no-flags.pkt | 24 +++ 42 files changed, 1848 insertions(+) create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-accept.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-connect.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-read.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-write.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-local-close-then-remote-fin.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-on-syn-sent.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-remote-fin-then-close.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ecn_ecn-uses-ect0.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-large.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-retrans.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-small.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-subsequent.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-10pkt-lost-1.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost-1_4-11_16.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost1_4.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-ack-below-snd_una-cubic.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-no-sack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-sack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_https_client.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_sendmsg_msg_more.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_sockopt_cork_nodelay.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-route-refresh-ip-tos.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-2-6-8-3-9-nofack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-3-4-8-9-fack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-5-6-8-9-fack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sendfile_sendfile-simple.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_splice_tcp_splice_loop_test.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_fastopen-invalid-buf-ptr.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_sendmsg-empty-iov.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_syscall-invalid-buf-ptr.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-last_data_recv.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-rwnd-limited.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-sndbuf-limited.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_client-only-last-byte.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_partial.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_server.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_fin_tsval.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_invalid_ack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_reset_tsval.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_user_timeout_user-timeout-probe.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_user_timeout_user_timeout.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_validate_validate-established-no-flags.pkt
Signature · 1 line
2.47.0.338.g60cca15819-goog
PATCH 1/4 selftests/net: packetdrill: import tcp/ecn, tcp/close, tcp/sack, tcp/tcp_info
From: Soham Chakradeo <[email protected]>
Same as initial tests, import verbatim from
github.com/google/packetdrill, aside from:
- update source ./defaults.sh path to adjust for flat dir
- add SPDX headers
- remove author statements if any
- drop blank lines at EOF
Same test process as previous tests. Both with and without debug mode. Recording the steps once:
make mrproper
vng --build \
--config tools/testing/selftests/net/packetdrill/config \
--config kernel/configs/debug.config
vng -v --run . --user root --cpus 4 -- \
make -C tools/testing/selftests TARGETS=net/packetdrill run_tests
--- ...lose_close-local-close-then-remote-fin.pkt | 23 +++++++ .../tcp_close_close-on-syn-sent.pkt | 21 ++++++ .../tcp_close_close-remote-fin-then-close.pkt | 36 ++++++++++ .../net/packetdrill/tcp_ecn_ecn-uses-ect0.pkt | 21 ++++++ .../tcp_sack_sack-route-refresh-ip-tos.pkt | 37 +++++++++++ ...ack_sack-shift-sacked-2-6-8-3-9-nofack.pkt | 64 ++++++++++++++++++ ..._sack_sack-shift-sacked-7-3-4-8-9-fack.pkt | 66 +++++++++++++++++++ ..._sack_sack-shift-sacked-7-5-6-8-9-fack.pkt | 62 +++++++++++++++++ .../tcp_tcp_info_tcp-info-last_data_recv.pkt | 21 ++++++ .../tcp_tcp_info_tcp-info-rwnd-limited.pkt | 54 +++++++++++++++ .../tcp_tcp_info_tcp-info-sndbuf-limited.pkt | 39 +++++++++++ 11 files changed, 444 insertions(+) create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-local-close-then-remote-fin.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-on-syn-sent.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_close_close-remote-fin-then-close.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ecn_ecn-uses-ect0.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-route-refresh-ip-tos.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-2-6-8-3-9-nofack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-3-4-8-9-fack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-5-6-8-9-fack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-last_data_recv.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-rwnd-limited.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-sndbuf-limited.pkt
tools/testing/selftests/net/packetdrill/tcp_close_close-local-close-then-remote-fin.pkt new file +23 -0
@@ -0,0 +1,23 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test basic connection teardown where local process closes first: +// the local process calls close() first, so we send a FIN, and receive an ACK. +// Then we receive a FIN and ACK it. + +`./defaults.sh` + + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +.01...0.011 connect(3, ..., ...) = 0 + +0 > S 0:0(0) <...> + +0 < S. 0:0(0) ack 1 win 32768 <mss 1000,nop,wscale 6,nop,nop,sackOK> + +0 > . 1:1(0) ack 1 + + +0 write(3, ..., 1000) = 1000 + +0 > P. 1:1001(1000) ack 1 + +0 < . 1:1(0) ack 1001 win 257 + + +0 close(3) = 0 + +0 > F. 1001:1001(0) ack 1 + +0 < . 1:1(0) ack 1002 win 257 + + +0 < F. 1:1(0) ack 1002 win 257 + +0 > . 1002:1002(0) ack 2
tools/testing/selftests/net/packetdrill/tcp_close_close-on-syn-sent.pkt new file +21 -0
@@ -0,0 +1,21 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test to make sure no RST is being sent when close() +// is called on a socket with SYN_SENT state. + +`./defaults.sh` + + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0 + + +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress) + +0 > S 0:0(0) <...> + +// Application decideds to close the socket in SYN_SENT state +// Make sure no RST is sent after close(). + +0 close(3) = 0 + +// Receive syn-ack to trigger the send side packet examination: +// If a RESET were sent right after close(), it would have failed with +// a mismatched timestamp. + +.1 < S. 0:0(0) ack 1 win 32000 <mss 1460,nop,wscale 7> + +0 > R 1:1(0)
tools/testing/selftests/net/packetdrill/tcp_close_close-remote-fin-then-close.pkt new file +36 -0
@@ -0,0 +1,36 @@ +// SPDX-License-Identifier: GPL-2.0 +// Verify behavior for the sequence: remote side sends FIN, then we close(). +// Since the remote side (client) closes first, we test our LAST_ACK code path. + +`./defaults.sh` + +// Initialize a server socket. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> + +0 < . 1:1(0) ack 1 win 257 + + +0 accept(3, ..., ...) = 4 + +// Client closes first. + +.01 < F. 1:1(0) ack 1 win 257 + +0 > . 1:1(0) ack 2 + +// App notices that client closed. + +0 read(4, ..., 1000) = 0 + +// Then we close. + +.01 close(4) = 0 + +0 > F. 1:1(0) ack 2 + +// Client ACKs our FIN. + +.01 < . 2:2(0) ack 2 win 257 + +// Verify that we send RST in response to any incoming segments +// (because the kernel no longer has any record of this socket). + +.01 < . 2:2(0) ack 2 win 257 + +0 > R 2:2(0)
tools/testing/selftests/net/packetdrill/tcp_ecn_ecn-uses-ect0.pkt new file +21 -0
@@ -0,0 +1,21 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test ECN: verify that Linux TCP ECN sending code uses ECT0 (not ECT1). +// +`./defaults.sh +sysctl -q net.ipv4.tcp_ecn=1 # fully enabled +` + +// Initialize connection + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4 + +// ECN handshake: send EW flags in SYN packet, E flag in SYN-ACK response ++.002 ... 0.004 connect(4, ..., ...) = 0 + + +0 > SEW 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8> ++.002 < SE. 0:0(0) ack 1 win 32767 <mss 1000,nop,wscale 6,nop,nop,sackOK> + +0 > . 1:1(0) ack 1 + +// Write 1 MSS. ++.002 write(4, ..., 1000) = 1000 +// Send 1 MSS with ect0. + +0 > [ect0] P. 1:1001(1000) ack 1
tools/testing/selftests/net/packetdrill/tcp_sack_sack-route-refresh-ip-tos.pkt new file +37 -0
@@ -0,0 +1,37 @@ +// SPDX-License-Identifier: GPL-2.0 +// Verify that setsockopt calls that force a route refresh do not +// cause problems matching SACKs with packets in the write queue. +// This variant tests IP_TOS. + +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_IP, IP_MTU_DISCOVER, [IP_PMTUDISC_DONT], 1) = 0 + +0...0.010 connect(3, ..., ...) = 0 + + +0 > S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8> + +.01 < S. 0:0(0) ack 1 win 65535 <mss 1460,nop,wscale 2,nop,nop,sackOK> + +0 > . 1:1(0) ack 1 + + +.01 write(3, ..., 5840) = 5840 + +0 > P. 1:5841(5840) ack 1 + +.01 < . 1:1(0) ack 5841 win 65535 + + +.01 write(3, ..., 5840) = 5840 + +0 > P. 5841:11681(5840) ack 1 + +.01 < . 1:1(0) ack 11681 win 65535 + + +.01 write(3, ..., 14600) = 14600 + +0 > P. 11681:26281(14600) ack 1 + +// Try the socket option that we know can force a route refresh. + +0 setsockopt(3, SOL_IP, IP_TOS, [4], 1) = 0 +// Then revert to avoid routing/mangling/etc implications of that setting. + +0 setsockopt(3, SOL_IP, IP_TOS, [0], 1) = 0 + +// Verify that we do not retransmit the SACKed segments. + +.01 < . 1:1(0) ack 13141 win 65535 <sack 16061:17521 20441:26281,nop,nop> + +0 > . 13141:16061(2920) ack 1 + +0 > P. 17521:20441(2920) ack 1 + +.01 < . 1:1(0) ack 26281 win 65535
tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-2-6-8-3-9-nofack.pkt new file +64 -0
64 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-3-4-8-9-fack.pkt new file +66 -0
66 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_sack_sack-shift-sacked-7-5-6-8-9-fack.pkt new file +62 -0
62 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-last_data_recv.pkt new file +21 -0
@@ -0,0 +1,21 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test tcpi_last_data_recv for active session +`./defaults.sh` + +// Create a socket and set it to non-blocking. ++0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 ++0 fcntl(3, F_GETFL) = 0x2 (flags O_RDWR) ++0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0 + ++0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress) ++0 > S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8> ++.030 < S. 0:0(0) ack 1 win 10000 <mss 1000,sackOK,nop,nop,nop,wscale 8> ++0 > . 1:1(0) ack 1 + ++1 %{ assert 990 <= tcpi_last_data_recv <= 1010, tcpi_last_data_recv }% + ++0 < . 1:1001(1000) ack 1 win 300 ++0 > . 1:1(0) ack 1001 + ++0 %{ assert tcpi_last_data_recv <= 10, tcpi_last_data_recv }% +
tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-rwnd-limited.pkt new file +54 -0
54 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_tcp_info_tcp-info-sndbuf-limited.pkt new file +39 -0
@@ -0,0 +1,39 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test send-buffer-limited time in tcp_info for client side. +`./defaults.sh` + +// Create a socket and set it to non-blocking. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 fcntl(3, F_GETFL) = 0x2 (flags O_RDWR) + +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0 + + +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress) + +0 > S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8> + +.01 < S. 0:0(0) ack 1 win 10000 <mss 1000,sackOK,nop,nop,nop,wscale 8> + +0 > . 1:1(0) ack 1 + +0 getsockopt(3, SOL_SOCKET, SO_ERROR, [0], [4]) = 0 + +0 fcntl(3, F_SETFL, O_RDWR) = 0 // set back to blocking + +0 setsockopt(3, SOL_SOCKET, SO_SNDBUF, [10000], 4) = 0 + +0 getsockopt(3, SOL_SOCKET, SO_SNDBUF, [20000], [4]) = 0 + + +.09...0.14 write(3, ..., 150000) = 150000 + + +.01 < . 1:1(0) ack 10001 win 10000 + + +.01 < . 1:1(0) ack 30001 win 10000 + +// cwnd goes from 40(60KB) to 80(120KB), and that we hit the tiny sndbuf limit 10KB + +.01 < . 1:1(0) ack 70001 win 10000 + + +.02 < . 1:1(0) ack 95001 win 10000 + +0 %{ assert 19000 <= tcpi_sndbuf_limited <= 21000, tcpi_sndbuf_limited; \ + assert 49000 <= tcpi_busy_time <= 52000, tcpi_busy_time; \ + assert 0 == tcpi_rwnd_limited, tcpi_rwnd_limited }% + +// This ack frees up enough buffer so we are no longer +// buffer limited (socket flag SOCK_NOSPACE is cleared) + +.02 < . 1:1(0) ack 150001 win 10000 + +0 %{ assert 19000 <= tcpi_sndbuf_limited <= 21000, tcpi_sndbuf_limited;\ + assert 69000 <= tcpi_busy_time <= 73000, tcpi_busy_time;\ + assert 0 == tcpi_rwnd_limited, tcpi_rwnd_limited }% +
Signature · 1 line
2.47.1.613.gc27f4b7a9f-goog
PATCH 2/4 selftests/net: packetdrill: import tcp/fast_recovery, tcp/nagle, tcp/timestamping
In reply to Soham ChakradeoFrom: Soham Chakradeo <[email protected]>
Use the standard import and testing method, as described in the import of tcp/ecn , tcp/close , tcp/sack , tcp/tcp_info.
--- .../tcp_fast_recovery_prr-ss-10pkt-lost-1.pkt | 72 +++++++++ ...t_recovery_prr-ss-30pkt-lost-1_4-11_16.pkt | 50 ++++++ ...tcp_fast_recovery_prr-ss-30pkt-lost1_4.pkt | 43 ++++++ ...ecovery_prr-ss-ack-below-snd_una-cubic.pkt | 41 +++++ .../packetdrill/tcp_nagle_https_client.pkt | 40 +++++ .../tcp_nagle_sendmsg_msg_more.pkt | 66 ++++++++ .../tcp_nagle_sockopt_cork_nodelay.pkt | 43 ++++++ ...tcp_timestamping_client-only-last-byte.pkt | 92 +++++++++++ .../packetdrill/tcp_timestamping_partial.pkt | 91 +++++++++++ .../packetdrill/tcp_timestamping_server.pkt | 145 ++++++++++++++++++ 10 files changed, 683 insertions(+) create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-10pkt-lost-1.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost-1_4-11_16.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost1_4.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-ack-below-snd_una-cubic.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_https_client.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_sendmsg_msg_more.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_nagle_sockopt_cork_nodelay.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_client-only-last-byte.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_partial.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_timestamping_server.pkt
tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-10pkt-lost-1.pkt new file +72 -0
@@ -0,0 +1,72 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test PRR-slowstart implementation. +// In this variant we test a simple case where in-flight == ssthresh +// all the way through recovery, so during fast recovery we send one segment +// for each segment SACKed/ACKed. + +// Set up config. +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +.1 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> +// RTT 100ms + +.1 < . 1:1(0) ack 1 win 320 + +0 accept(3, ..., ...) = 4 + +// Send 10 data segments. + +0 write(4, ..., 10000) = 10000 + +0 > P. 1:10001(10000) ack 1 + +// Lost packet 1:1001. + +.11 < . 1:1(0) ack 1 win 320 <sack 1001:2001,nop,nop> + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:3001,nop,nop> + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:4001,nop,nop> +// Enter fast recovery. + +0 > . 1:1001(1000) ack 1 + +.01 %{ +assert tcpi_ca_state == TCP_CA_Recovery, tcpi_ca_state +assert tcpi_snd_cwnd == 7, tcpi_snd_cwnd +assert tcpi_snd_ssthresh == 7, tcpi_snd_ssthresh +}% + +// Write some more, which we will send 1 MSS at a time, +// as in-flight segments are SACKed or ACKed. + +.01 write(4, ..., 7000) = 7000 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:5001,nop,nop> + +0 > . 10001:11001(1000) ack 1 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:6001,nop,nop> + +0 > . 11001:12001(1000) ack 1 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:7001,nop,nop> + +0 > . 12001:13001(1000) ack 1 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:8001,nop,nop> + +0 > . 13001:14001(1000) ack 1 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:9001,nop,nop> + +0 > . 14001:15001(1000) ack 1 + + +.01 < . 1:1(0) ack 1 win 320 <sack 1001:10001,nop,nop> + +0 > . 15001:16001(1000) ack 1 + + +.02 < . 1:1(0) ack 10001 win 320 + +0 > P. 16001:17001(1000) ack 1 +// Leave fast recovery. + +.01 %{ +assert tcpi_ca_state == TCP_CA_Open, tcpi_ca_state +assert tcpi_snd_cwnd == 7, tcpi_snd_cwnd +assert tcpi_snd_ssthresh == 7, tcpi_snd_ssthresh +}% + + +.03 < . 1:1(0) ack 12001 win 320 + +.02 < . 1:1(0) ack 14001 win 320 + +.02 < . 1:1(0) ack 16001 win 320 + +.02 < . 1:1(0) ack 17001 win 320
tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost-1_4-11_16.pkt new file +50 -0
50 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-30pkt-lost1_4.pkt new file +43 -0
43 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_fast_recovery_prr-ss-ack-below-snd_una-cubic.pkt new file +41 -0
41 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_nagle_https_client.pkt new file +40 -0
@@ -0,0 +1,40 @@ +// SPDX-License-Identifier: GPL-2.0 +// This is a test inspired by an Android client app using SSL. This +// test verifies using TCP_NODELAY would save application latency +// (Perhaps even better with TCP_NAGLE). +// +`./defaults.sh +ethtool -K tun0 tso off gso off +./set_sysctls.py /proc/sys/net/ipv4/tcp_timestamps=0` + + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4 + +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0 + +0 setsockopt(4, SOL_TCP, TCP_NODELAY, [1], 4) = 0 + + +0 connect(4, ..., ...) = -1 EINPROGRESS (Operation now in progress) + +0 > S 0:0(0) <mss 1460,nop,nop,sackOK,nop,wscale 8> + +.1 < S. 0:0(0) ack 1 win 5792 <mss 974,nop,nop,sackOK,nop,wscale 7> + +0 > . 1:1(0) ack 1 + +// SSL handshake (resumed session) + +0 write(4, ..., 517) = 517 + +0 > P. 1:518(517) ack 1 + +.1 < . 1:1(0) ack 518 win 229 + + +0 < P. 1:144(143) ack 1 win 229 + +0 > . 518:518(0) ack 144 + +0 read(4, ..., 1000) = 143 + +// Application POST header (51B) and body (2002B) + +0 write(4, ..., 51) = 51 + +0 > P. 518:569(51) ack 144 + +.03 write(4, ..., 2002) = 2002 + +0 > . 569:1543(974) ack 144 + +0 > P. 1543:2517(974) ack 144 +// Without disabling Nagle, this packet will not happen until the remote ACK. + +0 > P. 2517:2571(54) ack 144 + + +.1 < . 1:1(0) ack 2571 win 229 + +// Reset sysctls +`/tmp/sysctl_restore_${PPID}.sh`
tools/testing/selftests/net/packetdrill/tcp_nagle_sendmsg_msg_more.pkt new file +66 -0
66 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_nagle_sockopt_cork_nodelay.pkt new file +43 -0
43 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_timestamping_client-only-last-byte.pkt new file +92 -0
92 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_timestamping_partial.pkt new file +91 -0
91 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_timestamping_server.pkt new file +145 -0
145 lines not on this page · raw message
Signature · 1 line
2.47.1.613.gc27f4b7a9f-goog
PATCH 3/4 selftests/net: packetdrill: import tcp/eor, tcp/splice, tcp/ts_recent, tcp/blocking
In reply to Soham ChakradeoFrom: Soham Chakradeo <[email protected]>
Use the standard import and testing method, as described in the import of tcp/ecn and tcp/close , tcp/sack , tcp/tcp_info.
--- .../tcp_blocking_blocking-accept.pkt | 18 +++++ .../tcp_blocking_blocking-connect.pkt | 13 ++++ .../tcp_blocking_blocking-read.pkt | 29 ++++++++ .../tcp_blocking_blocking-write.pkt | 35 +++++++++ .../packetdrill/tcp_eor_no-coalesce-large.pkt | 38 ++++++++++ .../tcp_eor_no-coalesce-retrans.pkt | 72 +++++++++++++++++++ .../packetdrill/tcp_eor_no-coalesce-small.pkt | 36 ++++++++++ .../tcp_eor_no-coalesce-subsequent.pkt | 66 +++++++++++++++++ .../tcp_splice_tcp_splice_loop_test.pkt | 20 ++++++ .../packetdrill/tcp_ts_recent_fin_tsval.pkt | 23 ++++++ .../packetdrill/tcp_ts_recent_invalid_ack.pkt | 25 +++++++ .../packetdrill/tcp_ts_recent_reset_tsval.pkt | 25 +++++++ 12 files changed, 400 insertions(+) create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-accept.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-connect.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-read.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-write.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-large.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-retrans.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-small.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-subsequent.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_splice_tcp_splice_loop_test.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_fin_tsval.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_invalid_ack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_ts_recent_reset_tsval.pkt
tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-accept.pkt new file +18 -0
@@ -0,0 +1,18 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test for blocking accept. + +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +0...0.200 accept(3, ..., ...) = 4 + + +.1 < S 0:0(0) win 32792 <mss 1000,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8> + +.1 < . 1:1(0) ack 1 win 257 + + +.1 write(4, ..., 2000) = 2000 + +0 > P. 1:2001(2000) ack 1
tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-connect.pkt new file +13 -0
@@ -0,0 +1,13 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test for blocking connect. + +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + + +.1...0.200 connect(3, ..., ...) = 0 + + +0 > S 0:0(0) <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 8> + +.1 < S. 0:0(0) ack 1 win 5792 <mss 1460,nop,wscale 2,nop,nop,sackOK> + +0 > . 1:1(0) ack 1
tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-read.pkt new file +29 -0
@@ -0,0 +1,29 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test for blocking read. +--tolerance_usecs=10000 + +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +.1 < S 0:0(0) win 32792 <mss 1000,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8> + +.1 < . 1:1(0) ack 1 win 257 + +0 accept(3, ..., ...) = 4 + + +0...0.100 read(4, ..., 2000) = 2000 + +.1 < P. 1:2001(2000) ack 1 win 257 + +0 > . 1:1(0) ack 2001 + + +.1...0.200 read(4, ..., 2000) = 2000 + +.1 < P. 2001:4001(2000) ack 1 win 257 + +0 > . 1:1(0) ack 4001 + + +.1 < P. 4001:6001(2000) ack 1 win 257 + +0 > . 1:1(0) ack 6001 + +0...0.000 read(4, ..., 1000) = 1000 + +0...0.000 read(4, ..., 1000) = 1000
tools/testing/selftests/net/packetdrill/tcp_blocking_blocking-write.pkt new file +35 -0
@@ -0,0 +1,35 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test for blocking write. +--tolerance_usecs=10000 + +`./defaults.sh +./set_sysctls.py /proc/sys/net/ipv4/tcp_min_tso_segs=10 +` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +.1 < S 0:0(0) win 50000 <mss 1000,nop,wscale 0> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8> + +.1 < . 1:1(0) ack 1 win 50000 + +0 accept(3, ..., ...) = 4 + +// Kernel doubles our value -> sk->sk_sndbuf is set to 42000 + +0 setsockopt(4, SOL_SOCKET, SO_SNDBUF, [21000], 4) = 0 + +0 getsockopt(4, SOL_SOCKET, SO_SNDBUF, [42000], [4]) = 0 + +// A write of 60000 does not block. + +0...0.300 write(4, ..., 61000) = 61000 // this write() blocks + + +.1 < . 1:1(0) ack 10001 win 50000 + + +.1 < . 1:1(0) ack 30001 win 50000 + +// This ACK should wakeup the write(). An ACK of 35001 does not. + +.1 < . 1:1(0) ack 36001 win 50000 + +// Reset to sysctls defaults. +`/tmp/sysctl_restore_${PPID}.sh`
tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-large.pkt new file +38 -0
@@ -0,0 +1,38 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test TCP does not append any data from consequent writes to the tail +// skb created for the chunk. The large chunk itself should be packetized as +// usual. +`./defaults.sh +` + +// Initialize connection + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 10> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> + +.01 < . 1:1(0) ack 1 win 514 + + +0 accept(3, ..., ...) = 4 + +// Write a 10400B chunk to fill the ICW, and have a 400 byte skb sitting on +// the tail. + +0 write(4, ..., 10400) = 10400 + +// Write another 10040B chunk with no coalescing options. + +0 send(4, ..., 10400, MSG_EOR) = 10400 + +// Write a 2KB chunk. This chunk should not be appended to the packets created +// the previous chunk. + +0 write(4, ..., 2000) = 2000 + + +0 > P. 1:10001(10000) ack 1 ++.001 < . 1:1(0) ack 10001 win 514 +// Now we have enough room to send out the 2 x 400B packets out. + +0 > P. 10001:20801(10800) ack 1 ++.001 < . 1:1(0) ack 20801 win 514 +// This 2KB packet should be sent alone. + +0 > P. 20801:22801(2000) ack 1 ++.001 < . 1:1(0) ack 22801 win 514
tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-retrans.pkt new file +72 -0
72 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-small.pkt new file +36 -0
@@ -0,0 +1,36 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test TCP does not append any data from consequent writes to the tail +// skb created for the chunk. +`./defaults.sh +` + +// Initialize connection + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 10> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> + +.01 < . 1:1(0) ack 1 win 514 + + +0 accept(3, ..., ...) = 4 + +// Write a 10400B chunk to fill the ICW, and have a 400 byte skb sitting on +// the tail. + +0 write(4, ..., 10400) = 10400 + +// Write a 400B chunk with no coalescing options. + +0 send(4, ..., 400, MSG_EOR) = 400 + +// This chunk should not be appended to the skbs created for the previous chunk. + +0 write(4, ..., 10000) = 10000 + + +0 > P. 1:10001(10000) ack 1 ++.001 < . 1:1(0) ack 10001 win 514 +// Now we have enough room to send out the 2 x 400B packets out. + +0 > P. 10001:10801(800) ack 1 + +0 > P. 10801:20801(10000) ack 1 ++.001 < . 1:1(0) ack 10401 win 514 ++.001 < . 1:1(0) ack 10801 win 514 ++.001 < . 1:1(0) ack 20801 win 514
tools/testing/selftests/net/packetdrill/tcp_eor_no-coalesce-subsequent.pkt new file +66 -0
66 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_splice_tcp_splice_loop_test.pkt new file +20 -0
@@ -0,0 +1,20 @@ +// SPDX-License-Identifier: GPL-2.0 +`./defaults.sh` + +// Initialize a server socket + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 setsockopt(3, SOL_IP, IP_FREEBIND, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +// Connection should get accepted + +0 < S 0:0(0) win 32972 <mss 1460,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <...> + +0 < . 1:1(0) ack 1 win 257 + +0 accept(3, ..., ...) = 4 + + +0 pipe([5, 6]) = 0 + +0 < U. 1:101(100) ack 1 win 257 urg 100 + +0 splice(4, NULL, 6, NULL, 99, 0) = 99 + +0 splice(4, NULL, 6, NULL, 1, 0) = 0
tools/testing/selftests/net/packetdrill/tcp_ts_recent_fin_tsval.pkt new file +23 -0
@@ -0,0 +1,23 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test that we send FIN packet with correct TSval +--tcp_ts_tick_usecs=1000 +--tolerance_usecs=7000 + +`./defaults.sh` + +// Create a socket. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +// Establish a connection. + +0 < S 0:0(0) win 20000 <mss 1000,sackOK,TS val 100 ecr 0> + +0 > S. 0:0(0) ack 1 <mss 1460,sackOK,TS val 100 ecr 100> + +.1 < . 1:1(0) ack 1 win 20000 <nop,nop,TS val 200 ecr 100> + +0 accept(3, ..., ...) = 4 + + +1 close(4) = 0 +// Check that FIN TSval is updated properly, one second has passed since last sent packet. + +0 > F. 1:1(0) ack 1 <nop,nop,TS val 1200 ecr 200>
tools/testing/selftests/net/packetdrill/tcp_ts_recent_invalid_ack.pkt new file +25 -0
@@ -0,0 +1,25 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test that we reject TS val updates on a packet with invalid ACK sequence + +`./defaults.sh +` + +// Create a socket. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +// Establish a connection. + +.1 < S 0:0(0) win 20000 <mss 1000,sackOK,TS val 100 ecr 0> + +0 > S. 0:0(0) ack 1 <mss 1460,sackOK,TS val 100 ecr 100> + +.1 < . 1:1(0) ack 1 win 20000 <nop,nop,TS val 200 ecr 100> + +0 accept(3, ..., ...) = 4 + +// bad packet with high tsval (its ACK sequence is above our sndnxt) + +0 < F. 1:1(0) ack 9999 win 20000 <nop,nop,TS val 200000 ecr 100> + + + +0 < . 1:1001(1000) ack 1 win 20000 <nop,nop,TS val 201 ecr 100> + +0 > . 1:1(0) ack 1001 <nop,nop,TS val 200 ecr 201>
tools/testing/selftests/net/packetdrill/tcp_ts_recent_reset_tsval.pkt new file +25 -0
@@ -0,0 +1,25 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test that we send RST packet with correct TSval +--tcp_ts_tick_usecs=1000 + +`./defaults.sh` + +// Create a socket. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +// Establish a connection. + +0 < S 0:0(0) win 20000 <mss 1000,sackOK,TS val 100 ecr 0> + +0 > S. 0:0(0) ack 1 <mss 1460,sackOK,TS val 100 ecr 100> + +.1 < . 1:1(0) ack 1 win 20000 <nop,nop,TS val 200 ecr 100> + +0 accept(3, ..., ...) = 4 + + +0 < . 1:1001(1000) ack 1 win 20000 <nop,nop,TS val 201 ecr 100> + +0 > . 1:1(0) ack 1001 <nop,nop,TS val 200 ecr 201> + + +1 close(4) = 0 +// Check that RST TSval is updated properly, one second has passed since last sent packet. + +0 > R. 1:1(0) ack 1001 <nop,nop,TS val 1200 ecr 201>
Signature · 1 line
2.47.1.613.gc27f4b7a9f-goog
PATCH 4/4 selftests/net: packetdrill: import tcp/user_timeout, tcp/validate, tcp/sendfile, tcp/limited-transmit, tcp/syscall_bad_arg
In reply to Soham ChakradeoFrom: Soham Chakradeo <[email protected]>
Use the standard import and testing method, as described in the import of tcp/ecn and tcp/close , tcp/sack , tcp/tcp_info.
--- ...ited_transmit_limited-transmit-no-sack.pkt | 53 +++++++++++++++++++ ...limited_transmit_limited-transmit-sack.pkt | 50 +++++++++++++++++ .../tcp_sendfile_sendfile-simple.pkt | 26 +++++++++ ...scall_bad_arg_fastopen-invalid-buf-ptr.pkt | 42 +++++++++++++++ .../tcp_syscall_bad_arg_sendmsg-empty-iov.pkt | 31 +++++++++++ ...yscall_bad_arg_syscall-invalid-buf-ptr.pkt | 25 +++++++++ .../tcp_user_timeout_user-timeout-probe.pkt | 37 +++++++++++++ .../tcp_user_timeout_user_timeout.pkt | 33 ++++++++++++ ...validate_validate-established-no-flags.pkt | 24 +++++++++ 9 files changed, 321 insertions(+) create mode 100644 tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-no-sack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-sack.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_sendfile_sendfile-simple.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_fastopen-invalid-buf-ptr.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_sendmsg-empty-iov.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_syscall-invalid-buf-ptr.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_user_timeout_user-timeout-probe.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_user_timeout_user_timeout.pkt create mode 100644 tools/testing/selftests/net/packetdrill/tcp_validate_validate-established-no-flags.pkt
tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-no-sack.pkt new file +53 -0
@@ -0,0 +1,53 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test RFC 3042 "Limited Transmit": "sending a new data segment in +// response to each of the first two duplicate acknowledgments that +// arrive at the sender". +// This variation tests a receiver that doesn't support SACK. + +`./defaults.sh` + +// Establish a connection. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +.1 < S 0:0(0) win 32792 <mss 1000,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8> + +.1 < . 1:1(0) ack 1 win 320 + +0 accept(3, ..., ...) = 4 + +// Write some data, and send the initial congestion window. + +0 write(4, ..., 15000) = 15000 + +0 > P. 1:10001(10000) ack 1 + +// Limited transmit: on first dupack, send a new data segment. + +.11 < . 1:1(0) ack 1 win 320 + +0 > . 10001:11001(1000) ack 1 + +// Limited transmit: on second dupack, send a new data segment. + +.01 < . 1:1(0) ack 1 win 320 + +0 > . 11001:12001(1000) ack 1 + +// It turned out to be reordering, not loss. +// We have one packet newly acked (1001:3001 were DUP-ACK'd) +// So we revert state back to Open. Slow start cwnd from 10 to 11 +// and send 11 - 9 = 2 packets + +.01 < . 1:1(0) ack 3001 win 320 + +0 > P. 12001:14001(2000) ack 1 + + +.02 < . 1:1(0) ack 5001 win 320 + +0 > P. 14001:15001(1000) ack 1 + +// Client gradually ACKs all data. + +.02 < . 1:1(0) ack 7001 win 320 + +.02 < . 1:1(0) ack 9001 win 320 + +.02 < . 1:1(0) ack 11001 win 320 + +.02 < . 1:1(0) ack 13001 win 320 + +.02 < . 1:1(0) ack 15001 win 320 + +// Clean up. + +.17 close(4) = 0 + +0 > F. 15001:15001(0) ack 1 + +.1 < F. 1:1(0) ack 15002 win 257 + +0 > . 15002:15002(0) ack 2
tools/testing/selftests/net/packetdrill/tcp_limited_transmit_limited-transmit-sack.pkt new file +50 -0
50 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_sendfile_sendfile-simple.pkt new file +26 -0
@@ -0,0 +1,26 @@ +// SPDX-License-Identifier: GPL-2.0 +// Simplest possible test of open() and then sendfile(). +// We write some zeroes into a file (since packetdrill expects payloads +// to be all zeroes) and then open() the file, then use sendfile() +// and verify that the correct number of zeroes goes out. + +`./defaults.sh +/bin/rm -f /tmp/testfile +/bin/dd bs=1 count=5 if=/dev/zero of=/tmp/testfile status=none +` + +// Initialize connection + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 10> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> + +0 < . 1:1(0) ack 1 win 514 + + +0 accept(3, ..., ...) = 4 + + +0 open("/tmp/testfile", O_RDONLY) = 5 + +0 sendfile(4, 5, [0], 5) = 5 + +0 > P. 1:6(5) ack 1
tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_fastopen-invalid-buf-ptr.pkt new file +42 -0
42 lines not on this page · raw message
tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_sendmsg-empty-iov.pkt new file +31 -0
@@ -0,0 +1,31 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test that we correctly skip zero-length IOVs. +`./defaults.sh` + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,nop,wscale 7> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8> + +.01 < . 1:1(0) ack 1 win 257 + +0 accept(3, ..., ...) = 4 + +0 setsockopt(4, SOL_TCP, TCP_NODELAY, [1], 4) = 0 + + +0 sendmsg(4, {msg_name(...)=..., + msg_iov(4)=[{..., 0}, {..., 40}, {..., 0}, {..., 20}], + msg_flags=0}, 0) = 60 + +0 > P. 1:61(60) ack 1 + +.01 < . 1:1(0) ack 61 win 257 + + +0 sendmsg(4, {msg_name(...)=..., + msg_iov(4)=[{..., 0}, {..., 0}, {..., 0}, {..., 0}], + msg_flags=0}, MSG_ZEROCOPY) = 0 + + +0 sendmsg(4, {msg_name(...)=..., + msg_iov(4)=[{..., 0}, {..., 10}, {..., 0}, {..., 50}], + msg_flags=0}, MSG_ZEROCOPY) = 60 + +0 > P. 61:121(60) ack 1 + +.01 < . 1:1(0) ack 121 win 257 +
tools/testing/selftests/net/packetdrill/tcp_syscall_bad_arg_syscall-invalid-buf-ptr.pkt new file +25 -0
@@ -0,0 +1,25 @@ +// SPDX-License-Identifier: GPL-2.0 +// Test kernel behavior with NULL as buffer pointer + +`./defaults.sh` + + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 10> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8> + +.2 < . 1:1(0) ack 1 win 514 + + +0 accept(3, ..., ...) = 4 + + +0 write(4, NULL, 1000) = -1 EFAULT (Bad address) + +0 send(4, NULL, 1000, 0) = -1 EFAULT (Bad address) + +0 sendto(4, NULL, 1000, 0, ..., ...) = -1 EFAULT (Bad address) + + +0 < . 1:1001(1000) ack 1 win 200 + +0 read(4, NULL, 1000) = -1 EFAULT (Bad address) + +0 recv(4, NULL, 1000, 0) = -1 EFAULT (Bad address) + +0 recvfrom(4, NULL, 1000, 0, ..., ...) = -1 EFAULT (Bad address)
tools/testing/selftests/net/packetdrill/tcp_user_timeout_user-timeout-probe.pkt new file +37 -0
@@ -0,0 +1,37 @@ +// SPDX-License-Identifier: GPL-2.0 + +`./defaults.sh` + + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + + +0 < S 0:0(0) win 0 <mss 1460> + +0 > S. 0:0(0) ack 1 <mss 1460> + + +.1 < . 1:1(0) ack 1 win 65530 + +0 accept(3, ..., ...) = 4 + + +0 setsockopt(4, SOL_TCP, TCP_USER_TIMEOUT, [3000], 4) = 0 + +0 write(4, ..., 24) = 24 + +0 > P. 1:25(24) ack 1 + +.1 < . 1:1(0) ack 25 win 65530 + +0 %{ assert tcpi_probes == 0, tcpi_probes; \ + assert tcpi_backoff == 0, tcpi_backoff }% + +// install a qdisc dropping all packets + +0 `tc qdisc delete dev tun0 root 2>/dev/null ; tc qdisc add dev tun0 root pfifo limit 0` + +0 write(4, ..., 24) = 24 + // When qdisc is congested we retry every 500ms + // (TCP_RESOURCE_PROBE_INTERVAL) and therefore + // we retry 6 times before hitting 3s timeout. + // First verify that the connection is alive: ++3.250 write(4, ..., 24) = 24 + // Now verify that shortly after that the socket is dead: + +.100 write(4, ..., 24) = -1 ETIMEDOUT (Connection timed out) + + +0 %{ assert tcpi_probes == 6, tcpi_probes; \ + assert tcpi_backoff == 0, tcpi_backoff }% + +0 close(4) = 0
tools/testing/selftests/net/packetdrill/tcp_user_timeout_user_timeout.pkt new file +33 -0
@@ -0,0 +1,33 @@ +// SPDX-License-Identifier: GPL-2.0 +`./defaults.sh` + +// Initialize connection + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + + +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK> + +.1 < . 1:1(0) ack 1 win 32792 + + + +0 accept(3, ..., ...) = 4 + +// Okay, we received nothing, and decide to close this idle socket. +// We set TCP_USER_TIMEOUT to 3 seconds because really it is not worth +// trying hard to cleanly close this flow, at the price of keeping +// a TCP structure in kernel for about 1 minute ! + +2 setsockopt(4, SOL_TCP, TCP_USER_TIMEOUT, [3000], 4) = 0 + +0 close(4) = 0 + + +0 > F. 1:1(0) ack 1 + +.3~+.400 > F. 1:1(0) ack 1 + +.3~+.400 > F. 1:1(0) ack 1 + +.6~+.800 > F. 1:1(0) ack 1 + +// We finally receive something from the peer, but it is way too late +// Our socket vanished because TCP_USER_TIMEOUT was really small + +0 < . 1:2(1) ack 1 win 32792 + +0 > R 1:1(0) +
tools/testing/selftests/net/packetdrill/tcp_validate_validate-established-no-flags.pkt new file +24 -0
@@ -0,0 +1,24 @@ +// SPDX-License-Identifier: GPL-2.0 +// Verify that established connections drop a segment without the ACK flag set. + +`./defaults.sh` + +// Create a socket. + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 + +0 bind(3, ..., ...) = 0 + +0 listen(3, 1) = 0 + +// Establish a connection. + +0 < S 0:0(0) win 20000 <mss 1000,sackOK,nop,nop> + +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK> + +.01 < . 1:1(0) ack 1 win 20000 + +0 accept(3, ..., ...) = 4 + +// Receive a segment with no flags set, verify that it's not enqueued. + +.01 < - 1:1001(1000) win 20000 + +0 ioctl(4, SIOCINQ, [0]) = 0 + +// Receive a segment with ACK flag set, verify that it is enqueued. + +.01 < . 1:1001(1000) ack 1 win 20000 + +0 ioctl(4, SIOCINQ, [1000]) = 0
Signature · 1 line
2.47.1.613.gc27f4b7a9f-goog
Jakub Kicinski
maintainerQuoted 8 lines · On Tue, 17 Dec 2024 18:51:57 +0000 Soham Chakradeo wrote:
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO(tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018
Quoted 8 lines
Recent Timestamp (
ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)
Excellent, thanks for adding all these! I will merge the patches momentarily but I do see a number of flakes on our VMs with debug configs enabled:
https://netdev.bots.linux.dev/flakes.html?min-flip=0&tn-needle=packetdrill-dbg
In the 7 runs so far we got 2 flakes on:
tcp-timestamping-client-only-last-byte-pkt tcp-fast-recovery-prr-ss-ack-below-snd-una-cubic-pkt tcp-timestamping-server-pkt
1 flake on:
tcp-timestamping-partial-pkt
tcp-eor-no-coalesce-retrans-pkt
LMK if you can't find the outputs, they should be there within a couple of clicks.
I'll set these cases to be ignored for now, but would be great if we could find the way for them to be less time sensitive, perhaps?
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:
Quoted 19 lines · On Tue, 17 Dec 2024 18:51:57 +0000 you wrote:
From: Soham Chakradeo <[email protected]>
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO(tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018
Recent Timestamp (ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)
Quoted 1 line
[...]
Here is the summary with links:
- [net-next,1/4] selftests/net: packetdrill: import tcp/ecn, tcp/close, tcp/sack, tcp/tcp_info (no matching commit)
- [net-next,2/4] selftests/net: packetdrill: import tcp/fast_recovery, tcp/nagle, tcp/timestamping
https://git.kernel.org/netdev/net-next/c/eab35989cc37 - [net-next,3/4] selftests/net: packetdrill: import tcp/eor, tcp/splice, tcp/ts_recent, tcp/blocking https://git.kernel.org/netdev/net-next/c/6f6692053939
- [net-next,4/4] selftests/net: packetdrill: import tcp/user_timeout, tcp/validate, tcp/sendfile, tcp/limited-transmit, tcp/syscall_bad_arg (no matching commit)
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
Quoted 31 lines · Jakub Kicinski
Quoted 8 lines · On Tue, 17 Dec 2024 18:51:57 +0000 Soham Chakradeo wrote:
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO(tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018Quoted 8 lines
Recent Timestamp (
ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)Excellent, thanks for adding all these! I will merge the patches momentarily but I do see a number of flakes on our VMs with debug configs enabled:
https://netdev.bots.linux.dev/flakes.html?min-flip=0&tn-needle=packetdrill-dbgIn the 7 runs so far we got 2 flakes on:
tcp-timestamping-client-only-last-byte-pkt tcp-fast-recovery-prr-ss-ack-below-snd-una-cubic-pkt tcp-timestamping-server-pkt
1 flake on:
Quoted 8 lines
tcp-timestamping-partial-pkt
tcp-eor-no-coalesce-retrans-pktLMK if you can't find the outputs, they should be there within a couple of clicks.
I'll set these cases to be ignored for now, but would be great if we could find the way for them to be less time sensitive, perhaps?
Yes, let's get a bit more data how flaky they are and investigate.
Hopefully it's just tuning. Worst case we could back out these few tests (temporarily) to avoid polluting the dash.
We did not observe this on our end during repeated debug runs.
Paolo Abeni
maintainerQuoted 18 lines · Jakub Kicinski
Quoted 8 lines · On Tue, 17 Dec 2024 18:51:57 +0000 Soham Chakradeo wrote:
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO(tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018Quoted 8 lines
Recent Timestamp (
ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)
Quoted 8 lines
Excellent, thanks for adding all these! I will merge the patches momentarily but I do see a number of flakes on our VMs with debug configs enabled:
https://netdev.bots.linux.dev/flakes.html?min-flip=0&tn-needle=packetdrill-dbgIn the 7 runs so far we got 2 flakes on:
tcp-timestamping-client-only-last-byte-pkt
Quickly skimming over this one, it looks like it does not account for the increased default 'tolerance_us'. Kernel packetdrill set it by default to 14K (instead of 10K IIRC).
I guess this statement:
// SCM_TSTAMP_SCHED for the last byte should be received almost immediately // once 10001 is acked at t=20ms.
the the follow-up check should be updated accordingly. In the failures observed so far the max timestamp is > 35ms.
Cheers,
Paolo
Quoted 32 lines · Paolo Abeni
Quoted 18 lines · Jakub Kicinski
Quoted 8 lines · On Tue, 17 Dec 2024 18:51:57 +0000 Soham Chakradeo wrote:
Import tests for the following features (folder names in brackets):
ECN (ecn) : RFC 3168
Close (close) : RFC 9293TCP_INFO(tcp_info) : RFC 9293
Fast recovery (fast_recovery) : RFC 5681
Timestamping (timestamping) : RFC 1323
Nagle (nagle) : RFC 896
Selective Acknowledgments (sack) : RFC 2018Quoted 8 lines
Recent Timestamp (
ts_recent) : RFC 1323
Send file (sendfile)
Syscall bad arg (syscall_bad_arg)
Validate (validate)
Blocking (blocking)
Splice (splice)
End of record (eor)
Limited transmit (limited_transmit)Quoted 8 lines
Excellent, thanks for adding all these! I will merge the patches momentarily but I do see a number of flakes on our VMs with debug configs enabled:
https://netdev.bots.linux.dev/flakes.html?min-flip=0&tn-needle=packetdrill-dbgIn the 7 runs so far we got 2 flakes on:
tcp-timestamping-client-only-last-byte-pkt
Quickly skimming over this one, it looks like it does not account for the increased default '
tolerance_us'. Kernel packetdrill set it by default to 14K (instead of 10K IIRC).
Quoted 7 lines
I guess this statement:
//
SCM_TSTAMP_SCHEDfor the last byte should be received almost immediately // once 10001 is acked at t=20ms.the the follow-up check should be updated accordingly. In the failures observed so far the max timestamp is > 35ms.
Thanks Paolo.
All three timestamping flakes are instances where the script expects the timestamp to be taken essentially instantaneously after the send call.
This is not the case, and the delay is outside even the 14K tolerance. I see occurrences of 20K. At some point we cannot keep increasing the tolerance, perhaps.
Jakub Kicinski
maintainerQuoted 7 lines · On Thu, 19 Dec 2024 14:31:44 -0500 Willem de Bruijn wrote:
All three timestamping flakes are instances where the script expects the timestamp to be taken essentially instantaneously after the send call.
This is not the case, and the delay is outside even the 14K tolerance. I see occurrences of 20K. At some point we cannot keep increasing the tolerance, perhaps.
I pinned the other services away and gave the packetdrill tester its own cores. Let's see how much of a difference this makes. The net-next-2024-12-20--03-00 branch will be the first to have this.
Quoted 5 lines · Jakub Kicinski
Quoted 4 lines · On Thu, 19 Dec 2024 14:31:44 -0500 Willem de Bruijn wrote:
All three timestamping flakes are instances where the script expects the timestamp to be taken essentially instantaneously after the send call.
Quoted 7 lines
Quoted 3 lines
This is not the case, and the delay is outside even the 14K tolerance. I see occurrences of 20K. At some point we cannot keep increasing the tolerance, perhaps.
I pinned the other services away and gave the packetdrill tester its own cores. Let's see how much of a difference this makes. The net-next-2024-12-20--03-00 branch will be the first to have this.
Thanks. It does not seem to resolve the flakes.
At this point I think the best path is to run them in debug mode to get coverage, but ignore errors. With the below draft patch, error output is still logged. For instance:
# tcp_timestamping_partial.pkt:58: runtime error in recvmsg call: Bad timestamp 0 in scm_timestamping 0: expected=1734924748967958 (20000) actual=1734924748982069 (34111) start=1734924748947958
# ok 2 ipv6 # SKIP
Such timestamping test failures are fairly straightforward. We could just increase the KSFT_MACHINE_SLOW timeout. But other tests see an actual difference in TCP stack behavior, e.g., size of packet. That is not addressed by a further relaxation of the tolerance.
+++ b/tools/testing/selftests/net/packetdrill/ksft_runner.sh @@ -23,7 +23,7 @@ if [ $# -ne 1 ]; then
ktap_exit_fail_msg "usage: $0 <script>"
exit "$KSFT_FAIL" fi
-script="$1"
+script="$(basename $1)"
if [ -z "$(which packetdrill)" ]; then
ktap_skip_all "packetdrill not found in PATH"
@@ -31,16 +31,27 @@ if [ -z "$(which packetdrill)" ]; then
fi
declare -a optargs
+failfunc=ktap_test_fail
+
if [[ -n "${KSFT_MACHINE_SLOW}" ]]; then
optargs+=('--tolerance_usecs=14000')
+
+ declare -ar skip_list=(
+ "tcp_fast_recovery_prr-ss.*.pkt"
+ "tcp_timestamping.*.pkt"
+ "tcp_user_timeout_user-timeout-probe.pkt"
+ "tcp_zerocopy_epoll_.*.pkt"
+ )
+ readonly skip_pattern="^($(printf '%s|' "${skip_list[@]}"))$"
+ [[ "$script" =~ ${skip_pattern} ]] && failfunc=ktap_test_skipfi
ktap_print_header
ktap_set_plan 2
-unshare -n packetdrill ${ipv4_args[@]} ${optargs[@]} $(basename $script) > /dev/null \
- && ktap_test_pass "ipv4" || ktap_test_fail "ipv4"
-unshare -n packetdrill ${ipv6_args[@]} ${optargs[@]} $(basename $script) > /dev/null \
- && ktap_test_pass "ipv6" || ktap_test_fail "ipv6"
+unshare -n packetdrill ${ipv4_args[@]} ${optargs[@]} $script > /dev/null \
+ && ktap_test_pass "ipv4" || $failfunc "ipv4"
+unshare -n packetdrill ${ipv6_args[@]} ${optargs[@]} $script > /dev/null \
+ && ktap_test_pass "ipv6" || $failfunc "ipv6"
Jakub Kicinski
maintainerQuoted 14 lines · On Sun, 22 Dec 2024 22:46:26 -0500 Willem de Bruijn wrote:
Quoted 5 lines · Jakub Kicinski
Quoted 4 lines · On Thu, 19 Dec 2024 14:31:44 -0500 Willem de Bruijn wrote:
All three timestamping flakes are instances where the script expects the timestamp to be taken essentially instantaneously after the send call.
Quoted 7 lines
Quoted 3 lines
This is not the case, and the delay is outside even the 14K tolerance. I see occurrences of 20K. At some point we cannot keep increasing the tolerance, perhaps.
I pinned the other services away and gave the packetdrill tester its own cores. Let's see how much of a difference this makes. The net-next-2024-12-20--03-00 branch will be the first to have this.
Quoted 8 lines
Thanks. It does not seem to resolve the flakes.
At this point I think the best path is to run them in debug mode to get coverage, but ignore errors. With the below draft patch, error output is still logged. For instance:
# tcp_timestamping_partial.pkt:58: runtime error in recvmsg call: Bad timestamp 0 in
scm_timestamping0: expected=1734924748967958 (20000) actual=1734924748982069 (34111) start=1734924748947958# ok 2 ipv6 # SKIP
Makes sense. Can we make this XFAIL instead of SKIP, tho? Not exactly accurate but we try to use SKIP for reporting env / setup problems like missing commands. We have FAIL_TO_XFAIL and xfail_on_slow() in the lib for netdev bash tests, already.
Quoted 21 lines · Jakub Kicinski
Quoted 20 lines · On Sun, 22 Dec 2024 22:46:26 -0500 Willem de Bruijn wrote:
Quoted 5 lines · Jakub Kicinski
Quoted 4 lines · On Thu, 19 Dec 2024 14:31:44 -0500 Willem de Bruijn wrote:
All three timestamping flakes are instances where the script expects the timestamp to be taken essentially instantaneously after the send call.
Quoted 7 lines
Quoted 3 lines
This is not the case, and the delay is outside even the 14K tolerance. I see occurrences of 20K. At some point we cannot keep increasing the tolerance, perhaps.
I pinned the other services away and gave the packetdrill tester its own cores. Let's see how much of a difference this makes. The net-next-2024-12-20--03-00 branch will be the first to have this.
Thanks. It does not seem to resolve the flakes.
At this point I think the best path is to run them in debug mode to get coverage, but ignore errors. With the below draft patch, error output is still logged. For instance:
Quoted 7 lines
Quoted 2 lines
# tcp_timestamping_partial.pkt:58: runtime error in recvmsg call: Bad timestamp 0 in
scm_timestamping0: expected=1734924748967958 (20000) actual=1734924748982069 (34111) start=1734924748947958# ok 2 ipv6 # SKIPMakes sense. Can we make this XFAIL instead of SKIP, tho? Not exactly accurate but we try to use SKIP for reporting env / setup problems like missing commands. We have
FAIL_TO_XFAILandxfail_on_slow()in the lib for netdev bash tests, already.
Sounds good. I'll add a ktap_test_xfail() to stay with that API. I see no clean way to make use of xfail_on_slow directly.
When net-next reopens, unless the noisy dash is annoying.
Jakub Kicinski
maintainerQuoted 9 lines · On Tue, 24 Dec 2024 10:59:38 -0500 Willem de Bruijn wrote:
Quoted 9 lines
Quoted 8 lines
Thanks. It does not seem to resolve the flakes.
At this point I think the best path is to run them in debug mode to get coverage, but ignore errors. With the below draft patch, error output is still logged. For instance:
# tcp_timestamping_partial.pkt:58: runtime error in recvmsg call: Bad timestamp 0 in
scm_timestamping0: expected=1734924748967958 (20000) actual=1734924748982069 (34111) start=1734924748947958# ok 2 ipv6 # SKIP
Quoted 7 lines
Quoted 4 lines
Makes sense. Can we make this XFAIL instead of SKIP, tho? Not exactly accurate but we try to use SKIP for reporting env / setup problems like missing commands. We have
FAIL_TO_XFAILandxfail_on_slow()in the lib for netdev bash tests, already.Sounds good. I'll add a
ktap_test_xfail()to stay with that API. I see no clean way to make use ofxfail_on_slowdirectly.
Ack.
Quoted 1 line
When net-next reopens, unless the noisy dash is annoying.
No huge rush, once we mark the test as ignored it's not very annoying. As long as we have a fix before the next merge window we'll be good.