Date:   Mon, 18 Dec 2017 17:10:45 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     Ingo Molnar <mingo@...nel.org>,
        Ian McDonald <ian.mcdonald@...di.co.nz>,
        Vlad Yasevich <vyasevich@...il.com>,
        Stephen Hemminger <stephen@...workplumber.org>,
        Steven Rostedt <rostedt@...dmis.org>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        LKML <linux-kernel@...r.kernel.org>,
        "H . Peter Anvin" <hpa@...or.com>,
        Gerrit Renker <gerrit@....abdn.ac.uk>,
        "David S . Miller" <davem@...emloft.net>,
        Neil Horman <nhorman@...driver.com>, dccp@...r.kernel.org,
        netdev@...r.kernel.org, linux-sctp@...r.kernel.org,
        Stephen Rothwell <sfr@...b.auug.org.au>, mhiramat@...nel.org
Subject: [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events

Hi,

This series is v2 of the replacement of jprobe usage with trace
events, and is just rebased on current tip/master branch.

Please review it and test it.
Anyway, this removes all {tcp,sctp,dccp}_probe.ko because those
depend on jprobe.

This introduce new trace events which allows user to
trace network congestion window etc. via ftrace or perftools.
And remove jprobe usages (tcp_probe/dccp_probe/sctp_probe).
So this series removes all register_jprobe users from the kernel
tree.

So following example in

https://wiki.linuxfoundation.org/networking/tcpprobe

 # modprobe tcp_probe port=5001
 # cat /proc/net/tcpprobe >/tmp/data.out &
 # pid=$!
 # iperf -c otherhost
 # kill $pid

will be changed as below;

 # cd <debugfs or tracefs>/tracing
 # echo 1 > events/tcp/tcp_probe/enable
 # echo "sport == 5001 || dport == 5001"  > events/tcp/tcp_probe/filter
 # tail -f trace_pipe > /tmp/data.out &
 # pid=$!
 # iperf -c otherhost
 # kill $pid

And it outouts logs lile below;

# tracer: nop
#
#                              _-----=> irqs-off
#                             / _----=> need-resched
#                            | / _---=> hardirq/softirq
#                            || / _--=> preempt-depth
#                            ||| /     delay
#           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
#              | |       |   ||||       |         |
          <idle>-0     [000] ..s2  1089.238049: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=37 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28960
          <idle>-0     [000] ..s2  1090.156938: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=37 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28992
          <idle>-0     [000] ..s2  1091.333729: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=38 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28992
          <idle>-0     [000] ..s2  1092.300330: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=37 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28992
          <idle>-0     [000] ..s2  1095.044739: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=36 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28992
          <idle>-0     [000] ..s2  1096.573825: tcp_probe: src=[::ffff:192.168.139.2]:5001 dest=[::ffff:192.168.139.1]:56256 mark=0x0 length=32 snd_nxt=0xee4abe9c snd_una=0xee4abe9c snd_cwnd=10 ssthresh=2147483647 snd_wnd=29312 srtt=478 rcv_wnd=28992

I need your feedback for this change, like formatting etc.
Also, I need more test for this events by who can setup
DCCP and SCTP, since those are special protocols, I have
no environment to test it.

Steve, I also wrote a hack in sctp_probe event. In that event
it calls trace_sctp_probe_path() directly and record it.
This means an event invokes another event inside.
As far as I can see, that seems OK.
But I need your review too.

Thank you,

---

Masami Hiramatsu (6):
      net: tcp: Add trace events for TCP congestion window tracing
      net: tcp: Remove TCP probe module
      net: sctp: Add SCTP ACK tracking trace event
      net: sctp: Remove debug SCTP probe module
      net: dccp: Add DCCP sendmsg trace event
      net: dccp: Remove dccpprobe module


 include/trace/events/sctp.h |   96 ++++++++++++++
 include/trace/events/tcp.h  |   86 ++++++++++++
 net/Kconfig                 |   17 --
 net/core/net-traces.c       |    1 
 net/dccp/Kconfig            |   17 --
 net/dccp/Makefile           |    2 
 net/dccp/probe.c            |  203 -----------------------------
 net/dccp/proto.c            |    5 +
 net/dccp/trace.h            |  105 +++++++++++++++
 net/ipv4/Makefile           |    1 
 net/ipv4/tcp_input.c        |    3 
 net/ipv4/tcp_probe.c        |  301 -------------------------------------------
 net/sctp/Kconfig            |   12 --
 net/sctp/Makefile           |    3 
 net/sctp/probe.c            |  244 -----------------------------------
 net/sctp/sm_statefuns.c     |    5 +
 16 files changed, 301 insertions(+), 800 deletions(-)
 create mode 100644 include/trace/events/sctp.h
 delete mode 100644 net/dccp/probe.c
 create mode 100644 net/dccp/trace.h
 delete mode 100644 net/ipv4/tcp_probe.c
 delete mode 100644 net/sctp/probe.c

--
Masami Hiramatsu (Linaro) <mhiramat@...nel.org>

Powered by blists - more mailing lists