* [v2 PATCH -tip 1/6] net: tcp: Add trace events for TCP congestion window tracing
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
@ 2017-12-18 8:11 ` Masami Hiramatsu
2017-12-20 1:44 ` kbuild test robot
2017-12-18 8:11 ` [v2 PATCH -tip 2/6] net: tcp: Remove TCP probe module Masami Hiramatsu
` (5 subsequent siblings)
6 siblings, 1 reply; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:11 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
This adds an event to trace TCP stat variables with
slightly intrusive trace-event. This uses ftrace/perf
event log buffer to trace those state, no needs to
prepare own ring-buffer, nor custom user apps.
User can use ftrace to trace this event as below;
# cd /sys/kernel/debug/tracing
# echo 1 > events/tcp/tcp_probe/enable
(run workloads)
# cat trace
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
include/trace/events/tcp.h | 86 ++++++++++++++++++++++++++++++++++++++++++++
net/core/net-traces.c | 1 +
net/ipv4/tcp_input.c | 3 ++
3 files changed, 90 insertions(+)
diff --git a/include/trace/events/tcp.h b/include/trace/events/tcp.h
index 07cccca6cbf1..77c26a37d5ce 100644
--- a/include/trace/events/tcp.h
+++ b/include/trace/events/tcp.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: GPL-2.0 */
#undef TRACE_SYSTEM
#define TRACE_SYSTEM tcp
@@ -293,6 +294,91 @@ TRACE_EVENT(tcp_retransmit_synack,
__entry->saddr_v6, __entry->daddr_v6)
);
+#include <net/tcp.h>
+#include <linux/tcp.h>
+#include <linux/ipv6.h>
+#include <uapi/linux/in.h>
+#include <uapi/linux/in6.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(tcp_probe,
+
+ TP_PROTO(struct sock *sk, struct sk_buff *skb),
+
+ TP_ARGS(sk, skb),
+
+ TP_STRUCT__entry(
+ /* sockaddr_in6 is always bigger than sockaddr_in */
+ __array(__u8, saddr, sizeof(struct sockaddr_in6))
+ __array(__u8, daddr, sizeof(struct sockaddr_in6))
+ __field(__u16, sport)
+ __field(__u16, dport)
+ __field(__u32, mark)
+ __field(__u16, length)
+ __field(__u32, snd_nxt)
+ __field(__u32, snd_una)
+ __field(__u32, snd_cwnd)
+ __field(__u32, ssthresh)
+ __field(__u32, snd_wnd)
+ __field(__u32, srtt)
+ __field(__u32, rcv_wnd)
+ ),
+
+ TP_fast_assign(
+ const struct tcp_sock *tp = tcp_sk(sk);
+ const struct inet_sock *inet = inet_sk(sk);
+
+ memset(__entry->saddr, 0, sizeof(struct sockaddr_in6));
+ memset(__entry->daddr, 0, sizeof(struct sockaddr_in6));
+
+ if (sk->sk_family == AF_INET) {
+ struct sockaddr_in *v4 = (void *)__entry->saddr;
+
+ v4->sin_family = AF_INET;
+ v4->sin_port = inet->inet_sport;
+ v4->sin_addr.s_addr = inet->inet_saddr;
+ v4 = (void *)__entry->daddr;
+ v4->sin_family = AF_INET;
+ v4->sin_port = inet->inet_dport;
+ v4->sin_addr.s_addr = inet->inet_daddr;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (sk->sk_family == AF_INET6) {
+ struct sockaddr_in6 *v6 = (void *)__entry->saddr;
+
+ v6->sin6_family = AF_INET6;
+ v6->sin6_port = inet->inet_sport;
+ v6->sin6_addr = inet6_sk(sk)->saddr;
+ v6 = (void *)__entry->daddr;
+ v6->sin6_family = AF_INET6;
+ v6->sin6_port = inet->inet_dport;
+ v6->sin6_addr = sk->sk_v6_daddr;
+#endif
+ }
+
+ /* For filtering use */
+ __entry->sport = ntohs(inet->inet_sport);
+ __entry->dport = ntohs(inet->inet_dport);
+ __entry->mark = skb->mark;
+
+ __entry->length = skb->len;
+ __entry->snd_nxt = tp->snd_nxt;
+ __entry->snd_una = tp->snd_una;
+ __entry->snd_cwnd = tp->snd_cwnd;
+ __entry->snd_wnd = tp->snd_wnd;
+ __entry->rcv_wnd = tp->rcv_wnd;
+ __entry->ssthresh = tcp_current_ssthresh(sk);
+ __entry->srtt = tp->srtt_us >> 3;
+ ),
+
+ TP_printk("src=%pISpc dest=%pISpc mark=%#x length=%d snd_nxt=%#x "
+ "snd_una=%#x snd_cwnd=%u ssthresh=%u snd_wnd=%u srtt=%u "
+ "rcv_wnd=%u",
+ __entry->saddr, __entry->daddr, __entry->mark,
+ __entry->length, __entry->snd_nxt, __entry->snd_una,
+ __entry->snd_cwnd, __entry->ssthresh, __entry->snd_wnd,
+ __entry->srtt, __entry->rcv_wnd)
+);
+
#endif /* _TRACE_TCP_H */
/* This part must be outside protection */
diff --git a/net/core/net-traces.c b/net/core/net-traces.c
index 380934580fa1..25b9334ef065 100644
--- a/net/core/net-traces.c
+++ b/net/core/net-traces.c
@@ -31,6 +31,7 @@
#include <trace/events/net.h>
#include <trace/events/napi.h>
#include <trace/events/sock.h>
+#include <trace/events/tcp.h>
#include <trace/events/udp.h>
#include <trace/events/tcp.h>
#include <trace/events/fib.h>
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 45f750e85714..fa19cc86b2ad 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5306,6 +5306,9 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
unsigned int len = skb->len;
struct tcp_sock *tp = tcp_sk(sk);
+ /* TCP congestion window tracking */
+ trace_tcp_probe(sk, skb);
+
tcp_mstamp_refresh(tp);
if (unlikely(!sk->sk_rx_dst))
inet_csk(sk)->icsk_af_ops->sk_rx_dst_set(sk, skb);
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [v2 PATCH -tip 1/6] net: tcp: Add trace events for TCP congestion window tracing
2017-12-18 8:11 ` [v2 PATCH -tip 1/6] net: tcp: Add trace events for TCP congestion window tracing Masami Hiramatsu
@ 2017-12-20 1:44 ` kbuild test robot
0 siblings, 0 replies; 12+ messages in thread
From: kbuild test robot @ 2017-12-20 1:44 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: kbuild-all, Ingo Molnar, Ian McDonald, Vlad Yasevich,
Stephen Hemminger, Steven Rostedt, Peter Zijlstra,
Thomas Gleixner, LKML, H . Peter Anvin, Gerrit Renker,
David S . Miller, Neil Horman, dccp, netdev, linux-sctp,
Stephen Rothwell, mhiramat
[-- Attachment #1: Type: text/plain, Size: 11009 bytes --]
Hi Masami,
I love your patch! Yet something to improve:
[auto build test ERROR on net/master]
[also build test ERROR on v4.15-rc4 next-20171219]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Masami-Hiramatsu/net-tcp-sctp-dccp-Replace-jprobe-usage-with-trace-events/20171220-081035
config: sh-allmodconfig (attached as .config)
compiler: sh4-linux-gnu-gcc (Debian 7.2.0-11) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=sh
All error/warnings (new ones prefixed by >>):
In file included from include/trace/events/udp.h:9:0,
from net//core/net-traces.c:35:
>> include/trace/events/tcp.h:37:11: error: expected ')' before 'const'
TP_PROTO(const struct sock *sk, const struct sk_buff *skb),
^
include/linux/tracepoint.h:105:27: note: in definition of macro 'TP_PROTO'
#define TP_PROTO(args...) args
^~~~
>> include/linux/tracepoint.h:237:20: error: redefinition of '__tpstrtab_tcp_retransmit_skb'
static const char __tpstrtab_##name[] \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
>> include/trace/events/tcp.h:90:1: note: in expansion of macro 'DEFINE_EVENT'
DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
^~~~~~~~~~~~
In file included from include/trace/events/tcp.h:10:0,
from net//core/net-traces.c:34:
include/linux/tracepoint.h:237:20: note: previous definition of '__tpstrtab_tcp_retransmit_skb' was here
static const char __tpstrtab_##name[] \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
>> include/trace/events/tcp.h:90:1: note: in expansion of macro 'DEFINE_EVENT'
DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
^~~~~~~~~~~~
In file included from include/trace/events/udp.h:9:0,
from net//core/net-traces.c:35:
include/linux/tracepoint.h:239:20: error: redefinition of '__tracepoint_tcp_retransmit_skb'
struct tracepoint __tracepoint_##name \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
>> include/trace/events/tcp.h:90:1: note: in expansion of macro 'DEFINE_EVENT'
DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
^~~~~~~~~~~~
In file included from include/trace/events/tcp.h:10:0,
from net//core/net-traces.c:34:
include/linux/tracepoint.h:239:20: note: previous definition of '__tracepoint_tcp_retransmit_skb' was here
struct tracepoint __tracepoint_##name \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
>> include/trace/events/tcp.h:90:1: note: in expansion of macro 'DEFINE_EVENT'
DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
^~~~~~~~~~~~
In file included from include/trace/events/udp.h:9:0,
from net//core/net-traces.c:35:
>> include/linux/tracepoint.h:242:35: error: redefinition of '__tracepoint_ptr_tcp_retransmit_skb'
static struct tracepoint * const __tracepoint_ptr_##name __used \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
>> include/trace/events/tcp.h:90:1: note: in expansion of macro 'DEFINE_EVENT'
DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
^~~~~~~~~~~~
In file included from include/trace/events/tcp.h:10:0,
from net//core/net-traces.c:34:
include/linux/tracepoint.h:242:35: note: previous definition of '__tracepoint_ptr_tcp_retransmit_skb' was here
static struct tracepoint * const __tracepoint_ptr_##name __used \
^
>> include/linux/tracepoint.h:247:2: note: in expansion of macro 'DEFINE_TRACE_FN'
DEFINE_TRACE_FN(name, NULL, NULL);
^~~~~~~~~~~~~~~
>> include/trace/define_trace.h:51:2: note: in expansion of macro 'DEFINE_TRACE'
DEFINE_TRACE(name)
^~~~~~~~~~~~
vim +37 include/trace/events/tcp.h
e086101b Cong Wang 2017-10-13 12
e8fce239 Song Liu 2017-10-23 13 #define tcp_state_name(state) { state, #state }
e8fce239 Song Liu 2017-10-23 14 #define show_tcp_state_name(val) \
e8fce239 Song Liu 2017-10-23 15 __print_symbolic(val, \
e8fce239 Song Liu 2017-10-23 16 tcp_state_name(TCP_ESTABLISHED), \
e8fce239 Song Liu 2017-10-23 17 tcp_state_name(TCP_SYN_SENT), \
e8fce239 Song Liu 2017-10-23 18 tcp_state_name(TCP_SYN_RECV), \
e8fce239 Song Liu 2017-10-23 19 tcp_state_name(TCP_FIN_WAIT1), \
e8fce239 Song Liu 2017-10-23 20 tcp_state_name(TCP_FIN_WAIT2), \
e8fce239 Song Liu 2017-10-23 21 tcp_state_name(TCP_TIME_WAIT), \
e8fce239 Song Liu 2017-10-23 22 tcp_state_name(TCP_CLOSE), \
e8fce239 Song Liu 2017-10-23 23 tcp_state_name(TCP_CLOSE_WAIT), \
e8fce239 Song Liu 2017-10-23 24 tcp_state_name(TCP_LAST_ACK), \
e8fce239 Song Liu 2017-10-23 25 tcp_state_name(TCP_LISTEN), \
e8fce239 Song Liu 2017-10-23 26 tcp_state_name(TCP_CLOSING), \
e8fce239 Song Liu 2017-10-23 27 tcp_state_name(TCP_NEW_SYN_RECV))
e8fce239 Song Liu 2017-10-23 28
f6e37b25 Song Liu 2017-10-23 29 /*
f6e37b25 Song Liu 2017-10-23 30 * tcp event with arguments sk and skb
f6e37b25 Song Liu 2017-10-23 31 *
f6e37b25 Song Liu 2017-10-23 32 * Note: this class requires a valid sk pointer; while skb pointer could
f6e37b25 Song Liu 2017-10-23 33 * be NULL.
f6e37b25 Song Liu 2017-10-23 34 */
f6e37b25 Song Liu 2017-10-23 @35 DECLARE_EVENT_CLASS(tcp_event_sk_skb,
e086101b Cong Wang 2017-10-13 36
7344e29f Song Liu 2017-10-23 @37 TP_PROTO(const struct sock *sk, const struct sk_buff *skb),
e086101b Cong Wang 2017-10-13 38
e086101b Cong Wang 2017-10-13 39 TP_ARGS(sk, skb),
e086101b Cong Wang 2017-10-13 40
e086101b Cong Wang 2017-10-13 41 TP_STRUCT__entry(
7344e29f Song Liu 2017-10-23 42 __field(const void *, skbaddr)
7344e29f Song Liu 2017-10-23 43 __field(const void *, skaddr)
e086101b Cong Wang 2017-10-13 44 __field(__u16, sport)
e086101b Cong Wang 2017-10-13 45 __field(__u16, dport)
e086101b Cong Wang 2017-10-13 46 __array(__u8, saddr, 4)
e086101b Cong Wang 2017-10-13 47 __array(__u8, daddr, 4)
e086101b Cong Wang 2017-10-13 48 __array(__u8, saddr_v6, 16)
e086101b Cong Wang 2017-10-13 49 __array(__u8, daddr_v6, 16)
e086101b Cong Wang 2017-10-13 50 ),
e086101b Cong Wang 2017-10-13 51
e086101b Cong Wang 2017-10-13 52 TP_fast_assign(
e086101b Cong Wang 2017-10-13 53 struct inet_sock *inet = inet_sk(sk);
e086101b Cong Wang 2017-10-13 54 struct in6_addr *pin6;
e086101b Cong Wang 2017-10-13 55 __be32 *p32;
e086101b Cong Wang 2017-10-13 56
e086101b Cong Wang 2017-10-13 57 __entry->skbaddr = skb;
e086101b Cong Wang 2017-10-13 58 __entry->skaddr = sk;
e086101b Cong Wang 2017-10-13 59
e086101b Cong Wang 2017-10-13 60 __entry->sport = ntohs(inet->inet_sport);
e086101b Cong Wang 2017-10-13 61 __entry->dport = ntohs(inet->inet_dport);
e086101b Cong Wang 2017-10-13 62
e086101b Cong Wang 2017-10-13 63 p32 = (__be32 *) __entry->saddr;
e086101b Cong Wang 2017-10-13 64 *p32 = inet->inet_saddr;
e086101b Cong Wang 2017-10-13 65
e086101b Cong Wang 2017-10-13 66 p32 = (__be32 *) __entry->daddr;
e086101b Cong Wang 2017-10-13 67 *p32 = inet->inet_daddr;
e086101b Cong Wang 2017-10-13 68
89005678 David Ahern 2017-10-18 69 #if IS_ENABLED(CONFIG_IPV6)
89005678 David Ahern 2017-10-18 70 if (sk->sk_family == AF_INET6) {
e086101b Cong Wang 2017-10-13 71 pin6 = (struct in6_addr *)__entry->saddr_v6;
386fd5da David Ahern 2017-10-16 72 *pin6 = sk->sk_v6_rcv_saddr;
e086101b Cong Wang 2017-10-13 73 pin6 = (struct in6_addr *)__entry->daddr_v6;
386fd5da David Ahern 2017-10-16 74 *pin6 = sk->sk_v6_daddr;
89005678 David Ahern 2017-10-18 75 } else
89005678 David Ahern 2017-10-18 76 #endif
89005678 David Ahern 2017-10-18 77 {
e086101b Cong Wang 2017-10-13 78 pin6 = (struct in6_addr *)__entry->saddr_v6;
e086101b Cong Wang 2017-10-13 79 ipv6_addr_set_v4mapped(inet->inet_saddr, pin6);
e086101b Cong Wang 2017-10-13 80 pin6 = (struct in6_addr *)__entry->daddr_v6;
e086101b Cong Wang 2017-10-13 81 ipv6_addr_set_v4mapped(inet->inet_daddr, pin6);
e086101b Cong Wang 2017-10-13 82 }
e086101b Cong Wang 2017-10-13 83 ),
e086101b Cong Wang 2017-10-13 84
fb6ff75e David Ahern 2017-10-16 85 TP_printk("sport=%hu dport=%hu saddr=%pI4 daddr=%pI4 saddrv6=%pI6c daddrv6=%pI6c",
e086101b Cong Wang 2017-10-13 86 __entry->sport, __entry->dport, __entry->saddr, __entry->daddr,
e086101b Cong Wang 2017-10-13 87 __entry->saddr_v6, __entry->daddr_v6)
e086101b Cong Wang 2017-10-13 88 );
e086101b Cong Wang 2017-10-13 89
f6e37b25 Song Liu 2017-10-23 @90 DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb,
f6e37b25 Song Liu 2017-10-23 91
7344e29f Song Liu 2017-10-23 92 TP_PROTO(const struct sock *sk, const struct sk_buff *skb),
f6e37b25 Song Liu 2017-10-23 93
f6e37b25 Song Liu 2017-10-23 94 TP_ARGS(sk, skb)
f6e37b25 Song Liu 2017-10-23 95 );
f6e37b25 Song Liu 2017-10-23 96
:::::: The code at line 37 was first introduced by commit
:::::: 7344e29f285a94b965075599731811c352f3ab40 tcp: mark trace event arguments sk and skb as const
:::::: TO: Song Liu <songliubraving@fb.com>
:::::: CC: David S. Miller <davem@davemloft.net>
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 47602 bytes --]
^ permalink raw reply [flat|nested] 12+ messages in thread
* [v2 PATCH -tip 2/6] net: tcp: Remove TCP probe module
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
2017-12-18 8:11 ` [v2 PATCH -tip 1/6] net: tcp: Add trace events for TCP congestion window tracing Masami Hiramatsu
@ 2017-12-18 8:11 ` Masami Hiramatsu
2017-12-18 8:12 ` [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event Masami Hiramatsu
` (4 subsequent siblings)
6 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:11 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
Remove TCP probe module since jprobe has been deprecated.
That function is now replaced by tcp/tcp_probe trace-event.
You can use it via ftrace or perftools.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
net/Kconfig | 17 ---
net/ipv4/Makefile | 1
net/ipv4/tcp_probe.c | 301 --------------------------------------------------
3 files changed, 319 deletions(-)
delete mode 100644 net/ipv4/tcp_probe.c
diff --git a/net/Kconfig b/net/Kconfig
index 9dba2715919d..efe930db3c08 100644
--- a/net/Kconfig
+++ b/net/Kconfig
@@ -336,23 +336,6 @@ config NET_PKTGEN
To compile this code as a module, choose M here: the
module will be called pktgen.
-config NET_TCPPROBE
- tristate "TCP connection probing"
- depends on INET && PROC_FS && KPROBES
- ---help---
- This module allows for capturing the changes to TCP connection
- state in response to incoming packets. It is used for debugging
- TCP congestion avoidance modules. If you don't understand
- what was just said, you don't need it: say N.
-
- Documentation on how to use TCP connection probing can be found
- at:
-
- http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe
-
- To compile this code as a module, choose M here: the
- module will be called tcp_probe.
-
config NET_DROP_MONITOR
tristate "Network packet drop alerting service"
depends on INET && TRACEPOINTS
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index c6c8ad1d4b6d..47a0a6649a9d 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -43,7 +43,6 @@ obj-$(CONFIG_INET_DIAG) += inet_diag.o
obj-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o
obj-$(CONFIG_INET_UDP_DIAG) += udp_diag.o
obj-$(CONFIG_INET_RAW_DIAG) += raw_diag.o
-obj-$(CONFIG_NET_TCPPROBE) += tcp_probe.o
obj-$(CONFIG_TCP_CONG_BBR) += tcp_bbr.o
obj-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o
obj-$(CONFIG_TCP_CONG_CDG) += tcp_cdg.o
diff --git a/net/ipv4/tcp_probe.c b/net/ipv4/tcp_probe.c
deleted file mode 100644
index 697f4c67b2e3..000000000000
--- a/net/ipv4/tcp_probe.c
+++ /dev/null
@@ -1,301 +0,0 @@
-/*
- * tcpprobe - Observe the TCP flow with kprobes.
- *
- * The idea for this came from Werner Almesberger's umlsim
- * Copyright (C) 2004, Stephen Hemminger <shemminger@osdl.org>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- */
-
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
-#include <linux/kernel.h>
-#include <linux/kprobes.h>
-#include <linux/socket.h>
-#include <linux/tcp.h>
-#include <linux/slab.h>
-#include <linux/proc_fs.h>
-#include <linux/module.h>
-#include <linux/ktime.h>
-#include <linux/time.h>
-#include <net/net_namespace.h>
-
-#include <net/tcp.h>
-
-MODULE_AUTHOR("Stephen Hemminger <shemminger@linux-foundation.org>");
-MODULE_DESCRIPTION("TCP cwnd snooper");
-MODULE_LICENSE("GPL");
-MODULE_VERSION("1.1");
-
-static int port __read_mostly;
-MODULE_PARM_DESC(port, "Port to match (0=all)");
-module_param(port, int, 0);
-
-static unsigned int bufsize __read_mostly = 4096;
-MODULE_PARM_DESC(bufsize, "Log buffer size in packets (4096)");
-module_param(bufsize, uint, 0);
-
-static unsigned int fwmark __read_mostly;
-MODULE_PARM_DESC(fwmark, "skb mark to match (0=no mark)");
-module_param(fwmark, uint, 0);
-
-static int full __read_mostly;
-MODULE_PARM_DESC(full, "Full log (1=every ack packet received, 0=only cwnd changes)");
-module_param(full, int, 0);
-
-static const char procname[] = "tcpprobe";
-
-struct tcp_log {
- ktime_t tstamp;
- union {
- struct sockaddr raw;
- struct sockaddr_in v4;
- struct sockaddr_in6 v6;
- } src, dst;
- u16 length;
- u32 snd_nxt;
- u32 snd_una;
- u32 snd_wnd;
- u32 rcv_wnd;
- u32 snd_cwnd;
- u32 ssthresh;
- u32 srtt;
-};
-
-static struct {
- spinlock_t lock;
- wait_queue_head_t wait;
- ktime_t start;
- u32 lastcwnd;
-
- unsigned long head, tail;
- struct tcp_log *log;
-} tcp_probe;
-
-static inline int tcp_probe_used(void)
-{
- return (tcp_probe.head - tcp_probe.tail) & (bufsize - 1);
-}
-
-static inline int tcp_probe_avail(void)
-{
- return bufsize - tcp_probe_used() - 1;
-}
-
-#define tcp_probe_copy_fl_to_si4(inet, si4, mem) \
- do { \
- si4.sin_family = AF_INET; \
- si4.sin_port = inet->inet_##mem##port; \
- si4.sin_addr.s_addr = inet->inet_##mem##addr; \
- } while (0) \
-
-/*
- * Hook inserted to be called before each receive packet.
- * Note: arguments must match tcp_rcv_established()!
- */
-static void jtcp_rcv_established(struct sock *sk, struct sk_buff *skb,
- const struct tcphdr *th)
-{
- unsigned int len = skb->len;
- const struct tcp_sock *tp = tcp_sk(sk);
- const struct inet_sock *inet = inet_sk(sk);
-
- /* Only update if port or skb mark matches */
- if (((port == 0 && fwmark == 0) ||
- ntohs(inet->inet_dport) == port ||
- ntohs(inet->inet_sport) == port ||
- (fwmark > 0 && skb->mark == fwmark)) &&
- (full || tp->snd_cwnd != tcp_probe.lastcwnd)) {
-
- spin_lock(&tcp_probe.lock);
- /* If log fills, just silently drop */
- if (tcp_probe_avail() > 1) {
- struct tcp_log *p = tcp_probe.log + tcp_probe.head;
-
- p->tstamp = ktime_get();
- switch (sk->sk_family) {
- case AF_INET:
- tcp_probe_copy_fl_to_si4(inet, p->src.v4, s);
- tcp_probe_copy_fl_to_si4(inet, p->dst.v4, d);
- break;
- case AF_INET6:
- memset(&p->src.v6, 0, sizeof(p->src.v6));
- memset(&p->dst.v6, 0, sizeof(p->dst.v6));
-#if IS_ENABLED(CONFIG_IPV6)
- p->src.v6.sin6_family = AF_INET6;
- p->src.v6.sin6_port = inet->inet_sport;
- p->src.v6.sin6_addr = inet6_sk(sk)->saddr;
-
- p->dst.v6.sin6_family = AF_INET6;
- p->dst.v6.sin6_port = inet->inet_dport;
- p->dst.v6.sin6_addr = sk->sk_v6_daddr;
-#endif
- break;
- default:
- BUG();
- }
-
- p->length = len;
- p->snd_nxt = tp->snd_nxt;
- p->snd_una = tp->snd_una;
- p->snd_cwnd = tp->snd_cwnd;
- p->snd_wnd = tp->snd_wnd;
- p->rcv_wnd = tp->rcv_wnd;
- p->ssthresh = tcp_current_ssthresh(sk);
- p->srtt = tp->srtt_us >> 3;
-
- tcp_probe.head = (tcp_probe.head + 1) & (bufsize - 1);
- }
- tcp_probe.lastcwnd = tp->snd_cwnd;
- spin_unlock(&tcp_probe.lock);
-
- wake_up(&tcp_probe.wait);
- }
-
- jprobe_return();
-}
-
-static struct jprobe tcp_jprobe = {
- .kp = {
- .symbol_name = "tcp_rcv_established",
- },
- .entry = jtcp_rcv_established,
-};
-
-static int tcpprobe_open(struct inode *inode, struct file *file)
-{
- /* Reset (empty) log */
- spin_lock_bh(&tcp_probe.lock);
- tcp_probe.head = tcp_probe.tail = 0;
- tcp_probe.start = ktime_get();
- spin_unlock_bh(&tcp_probe.lock);
-
- return 0;
-}
-
-static int tcpprobe_sprint(char *tbuf, int n)
-{
- const struct tcp_log *p
- = tcp_probe.log + tcp_probe.tail;
- struct timespec64 ts
- = ktime_to_timespec64(ktime_sub(p->tstamp, tcp_probe.start));
-
- return scnprintf(tbuf, n,
- "%lu.%09lu %pISpc %pISpc %d %#x %#x %u %u %u %u %u\n",
- (unsigned long)ts.tv_sec,
- (unsigned long)ts.tv_nsec,
- &p->src, &p->dst, p->length, p->snd_nxt, p->snd_una,
- p->snd_cwnd, p->ssthresh, p->snd_wnd, p->srtt, p->rcv_wnd);
-}
-
-static ssize_t tcpprobe_read(struct file *file, char __user *buf,
- size_t len, loff_t *ppos)
-{
- int error = 0;
- size_t cnt = 0;
-
- if (!buf)
- return -EINVAL;
-
- while (cnt < len) {
- char tbuf[256];
- int width;
-
- /* Wait for data in buffer */
- error = wait_event_interruptible(tcp_probe.wait,
- tcp_probe_used() > 0);
- if (error)
- break;
-
- spin_lock_bh(&tcp_probe.lock);
- if (tcp_probe.head == tcp_probe.tail) {
- /* multiple readers race? */
- spin_unlock_bh(&tcp_probe.lock);
- continue;
- }
-
- width = tcpprobe_sprint(tbuf, sizeof(tbuf));
-
- if (cnt + width < len)
- tcp_probe.tail = (tcp_probe.tail + 1) & (bufsize - 1);
-
- spin_unlock_bh(&tcp_probe.lock);
-
- /* if record greater than space available
- return partial buffer (so far) */
- if (cnt + width >= len)
- break;
-
- if (copy_to_user(buf + cnt, tbuf, width))
- return -EFAULT;
- cnt += width;
- }
-
- return cnt == 0 ? error : cnt;
-}
-
-static const struct file_operations tcpprobe_fops = {
- .owner = THIS_MODULE,
- .open = tcpprobe_open,
- .read = tcpprobe_read,
- .llseek = noop_llseek,
-};
-
-static __init int tcpprobe_init(void)
-{
- int ret = -ENOMEM;
-
- /* Warning: if the function signature of tcp_rcv_established,
- * has been changed, you also have to change the signature of
- * jtcp_rcv_established, otherwise you end up right here!
- */
- BUILD_BUG_ON(__same_type(tcp_rcv_established,
- jtcp_rcv_established) == 0);
-
- init_waitqueue_head(&tcp_probe.wait);
- spin_lock_init(&tcp_probe.lock);
-
- if (bufsize == 0)
- return -EINVAL;
-
- bufsize = roundup_pow_of_two(bufsize);
- tcp_probe.log = kcalloc(bufsize, sizeof(struct tcp_log), GFP_KERNEL);
- if (!tcp_probe.log)
- goto err0;
-
- if (!proc_create(procname, S_IRUSR, init_net.proc_net, &tcpprobe_fops))
- goto err0;
-
- ret = register_jprobe(&tcp_jprobe);
- if (ret)
- goto err1;
-
- pr_info("probe registered (port=%d/fwmark=%u) bufsize=%u\n",
- port, fwmark, bufsize);
- return 0;
- err1:
- remove_proc_entry(procname, init_net.proc_net);
- err0:
- kfree(tcp_probe.log);
- return ret;
-}
-module_init(tcpprobe_init);
-
-static __exit void tcpprobe_exit(void)
-{
- remove_proc_entry(procname, init_net.proc_net);
- unregister_jprobe(&tcp_jprobe);
- kfree(tcp_probe.log);
-}
-module_exit(tcpprobe_exit);
^ permalink raw reply related [flat|nested] 12+ messages in thread* [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
2017-12-18 8:11 ` [v2 PATCH -tip 1/6] net: tcp: Add trace events for TCP congestion window tracing Masami Hiramatsu
2017-12-18 8:11 ` [v2 PATCH -tip 2/6] net: tcp: Remove TCP probe module Masami Hiramatsu
@ 2017-12-18 8:12 ` Masami Hiramatsu
2017-12-18 17:05 ` Steven Rostedt
2017-12-20 0:48 ` kbuild test robot
2017-12-18 8:12 ` [v2 PATCH -tip 4/6] net: sctp: Remove debug SCTP probe module Masami Hiramatsu
` (3 subsequent siblings)
6 siblings, 2 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:12 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
Add SCTP ACK tracking trace event to trace the changes of SCTP
association state in response to incoming packets.
It is used for debugging SCTP congestion control algorithms,
and will replace sctp_probe module.
Note that this event a bit tricky. Since this consists of 2
events (sctp_probe and sctp_probe_path) so you have to enable
both events as below.
# cd /sys/kernel/debug/tracing
# echo 1 > events/sctp/sctp_probe/enable
# echo 1 > events/sctp/sctp_probe_path/enable
Or, you can enable all the events under sctp.
# echo 1 > events/sctp/enable
Since sctp_probe_path event is always invoked from sctp_probe
event, you can not see any output if you only enable
sctp_probe_path.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
include/trace/events/sctp.h | 96 +++++++++++++++++++++++++++++++++++++++++++
net/sctp/sm_statefuns.c | 5 ++
2 files changed, 101 insertions(+)
create mode 100644 include/trace/events/sctp.h
diff --git a/include/trace/events/sctp.h b/include/trace/events/sctp.h
new file mode 100644
index 000000000000..32c2dc72311e
--- /dev/null
+++ b/include/trace/events/sctp.h
@@ -0,0 +1,96 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM sctp
+
+#if !defined(_TRACE_SCTP_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_SCTP_H
+
+#include <net/sctp/structs.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(sctp_probe_path,
+
+ TP_PROTO(struct sctp_transport *sp,
+ const struct sctp_association *asoc),
+
+ TP_ARGS(sp, asoc),
+
+ TP_STRUCT__entry(
+ __field(__u64, asoc)
+ __field(__u32, primary)
+ __array(__u8, ipaddr, sizeof(union sctp_addr))
+ __field(__u32, state)
+ __field(__u32, cwnd)
+ __field(__u32, ssthresh)
+ __field(__u32, flight_size)
+ __field(__u32, partial_bytes_acked)
+ __field(__u32, pathmtu)
+ ),
+
+ TP_fast_assign(
+ __entry->asoc = (__u64)asoc;
+ __entry->primary = (sp == asoc->peer.primary_path);
+ memcpy(__entry->ipaddr, &sp->ipaddr, sizeof(union sctp_addr));
+ __entry->state = sp->state;
+ __entry->cwnd = sp->cwnd;
+ __entry->ssthresh = sp->ssthresh;
+ __entry->flight_size = sp->flight_size;
+ __entry->partial_bytes_acked = sp->partial_bytes_acked;
+ __entry->pathmtu = sp->pathmtu;
+ ),
+
+ TP_printk("asoc=%#llx%s ipaddr=%pISpc state=%u cwnd=%u ssthresh=%u "
+ "flight_size=%u partial_bytes_acked=%u pathmtu=%u",
+ __entry->asoc, __entry->primary ? "(*)" : "",
+ __entry->ipaddr, __entry->state, __entry->cwnd,
+ __entry->ssthresh, __entry->flight_size,
+ __entry->partial_bytes_acked, __entry->pathmtu)
+);
+
+TRACE_EVENT(sctp_probe,
+
+ TP_PROTO(const struct sctp_endpoint *ep,
+ const struct sctp_association *asoc,
+ struct sctp_chunk *chunk),
+
+ TP_ARGS(ep, asoc, chunk),
+
+ TP_STRUCT__entry(
+ __field(__u64, asoc)
+ __field(__u32, mark)
+ __field(__u16, bind_port)
+ __field(__u16, peer_port)
+ __field(__u32, pathmtu)
+ __field(__u32, rwnd)
+ __field(__u16, unack_data)
+ ),
+
+ TP_fast_assign(
+ struct sctp_transport *sp;
+ struct sk_buff *skb = chunk->skb;
+
+ __entry->asoc = (__u64)asoc;
+ __entry->mark = skb->mark;
+ __entry->bind_port = ep->base.bind_addr.port;
+ __entry->peer_port = asoc->peer.port;
+ __entry->pathmtu = asoc->pathmtu;
+ __entry->rwnd = asoc->peer.rwnd;
+ __entry->unack_data = asoc->unack_data;
+
+ list_for_each_entry(sp, &asoc->peer.transport_addr_list,
+ transports) {
+ trace_sctp_probe_path(sp, asoc);
+ }
+ ),
+
+ TP_printk("asoc=%#llx mark=%#x bind_port=%d peer_port=%d pathmtu=%d "
+ "rwnd=%u unack_data=%d",
+ __entry->asoc, __entry->mark, __entry->bind_port,
+ __entry->peer_port, __entry->pathmtu, __entry->rwnd,
+ __entry->unack_data)
+);
+
+#endif /* _TRACE_SCTP_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 8f8ccded13e4..c5f92b2cc5c3 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -59,6 +59,9 @@
#include <net/sctp/sm.h>
#include <net/sctp/structs.h>
+#define CREATE_TRACE_POINTS
+#include <trace/events/sctp.h>
+
static struct sctp_packet *sctp_abort_pkt_new(
struct net *net,
const struct sctp_endpoint *ep,
@@ -3219,6 +3222,8 @@ enum sctp_disposition sctp_sf_eat_sack_6_2(struct net *net,
struct sctp_sackhdr *sackh;
__u32 ctsn;
+ trace_sctp_probe(ep, asoc, chunk);
+
if (!sctp_vtag_verify(chunk, asoc))
return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event
2017-12-18 8:12 ` [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event Masami Hiramatsu
@ 2017-12-18 17:05 ` Steven Rostedt
2017-12-19 1:31 ` Masami Hiramatsu
2017-12-20 0:48 ` kbuild test robot
1 sibling, 1 reply; 12+ messages in thread
From: Steven Rostedt @ 2017-12-18 17:05 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell
On Mon, 18 Dec 2017 17:12:15 +0900
Masami Hiramatsu <mhiramat@kernel.org> wrote:
> Add SCTP ACK tracking trace event to trace the changes of SCTP
> association state in response to incoming packets.
> It is used for debugging SCTP congestion control algorithms,
> and will replace sctp_probe module.
>
> Note that this event a bit tricky. Since this consists of 2
> events (sctp_probe and sctp_probe_path) so you have to enable
> both events as below.
>
> # cd /sys/kernel/debug/tracing
> # echo 1 > events/sctp/sctp_probe/enable
> # echo 1 > events/sctp/sctp_probe_path/enable
>
> Or, you can enable all the events under sctp.
>
> # echo 1 > events/sctp/enable
>
> Since sctp_probe_path event is always invoked from sctp_probe
> event, you can not see any output if you only enable
> sctp_probe_path.
I have to ask, why did you do it this way?
> +#include <trace/define_trace.h>
> diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
> index 8f8ccded13e4..c5f92b2cc5c3 100644
> --- a/net/sctp/sm_statefuns.c
> +++ b/net/sctp/sm_statefuns.c
> @@ -59,6 +59,9 @@
> #include <net/sctp/sm.h>
> #include <net/sctp/structs.h>
>
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/sctp.h>
> +
> static struct sctp_packet *sctp_abort_pkt_new(
> struct net *net,
> const struct sctp_endpoint *ep,
> @@ -3219,6 +3222,8 @@ enum sctp_disposition sctp_sf_eat_sack_6_2(struct net *net,
> struct sctp_sackhdr *sackh;
> __u32 ctsn;
>
> + trace_sctp_probe(ep, asoc, chunk);
What about doing this right after this probe:
if (trace_sctp_probe_path_enabled()) {
struct sctp_transport *sp;
list_for_each_entry(sp, &asoc->peer.transpor_addr_list,
transports) {
trace_sctp_probe_path(sp, asoc);
}
}
The "trace_sctp_probe_path_enabled()" is a static branch, which means
it's a nop just like a tracepoint is, and will not add any overhead if
the trace_sctp_probe_path is not enabled.
-- Steve
> +
> if (!sctp_vtag_verify(chunk, asoc))
> return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
>
^ permalink raw reply [flat|nested] 12+ messages in thread* Re: [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event
2017-12-18 17:05 ` Steven Rostedt
@ 2017-12-19 1:31 ` Masami Hiramatsu
0 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-19 1:31 UTC (permalink / raw)
To: Steven Rostedt
Cc: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell
On Mon, 18 Dec 2017 12:05:16 -0500
Steven Rostedt <rostedt@goodmis.org> wrote:
> On Mon, 18 Dec 2017 17:12:15 +0900
> Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> > Add SCTP ACK tracking trace event to trace the changes of SCTP
> > association state in response to incoming packets.
> > It is used for debugging SCTP congestion control algorithms,
> > and will replace sctp_probe module.
> >
> > Note that this event a bit tricky. Since this consists of 2
> > events (sctp_probe and sctp_probe_path) so you have to enable
> > both events as below.
> >
> > # cd /sys/kernel/debug/tracing
> > # echo 1 > events/sctp/sctp_probe/enable
> > # echo 1 > events/sctp/sctp_probe_path/enable
> >
> > Or, you can enable all the events under sctp.
> >
> > # echo 1 > events/sctp/enable
> >
> > Since sctp_probe_path event is always invoked from sctp_probe
> > event, you can not see any output if you only enable
> > sctp_probe_path.
>
> I have to ask, why did you do it this way?
>
>
> > +#include <trace/define_trace.h>
> > diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
> > index 8f8ccded13e4..c5f92b2cc5c3 100644
> > --- a/net/sctp/sm_statefuns.c
> > +++ b/net/sctp/sm_statefuns.c
> > @@ -59,6 +59,9 @@
> > #include <net/sctp/sm.h>
> > #include <net/sctp/structs.h>
> >
> > +#define CREATE_TRACE_POINTS
> > +#include <trace/events/sctp.h>
> > +
> > static struct sctp_packet *sctp_abort_pkt_new(
> > struct net *net,
> > const struct sctp_endpoint *ep,
> > @@ -3219,6 +3222,8 @@ enum sctp_disposition sctp_sf_eat_sack_6_2(struct net *net,
> > struct sctp_sackhdr *sackh;
> > __u32 ctsn;
> >
> > + trace_sctp_probe(ep, asoc, chunk);
>
> What about doing this right after this probe:
>
> if (trace_sctp_probe_path_enabled()) {
> struct sctp_transport *sp;
>
> list_for_each_entry(sp, &asoc->peer.transpor_addr_list,
> transports) {
> trace_sctp_probe_path(sp, asoc);
> }
> }
>
> The "trace_sctp_probe_path_enabled()" is a static branch, which means
> it's a nop just like a tracepoint is, and will not add any overhead if
> the trace_sctp_probe_path is not enabled.
That's a good idea! I'll update to use it :)
Thank you,
>
> -- Steve
>
> > +
> > if (!sctp_vtag_verify(chunk, asoc))
> > return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
> >
>
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event
2017-12-18 8:12 ` [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event Masami Hiramatsu
2017-12-18 17:05 ` Steven Rostedt
@ 2017-12-20 0:48 ` kbuild test robot
1 sibling, 0 replies; 12+ messages in thread
From: kbuild test robot @ 2017-12-20 0:48 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: kbuild-all, Ingo Molnar, Ian McDonald, Vlad Yasevich,
Stephen Hemminger, Steven Rostedt, Peter Zijlstra,
Thomas Gleixner, LKML, H . Peter Anvin, Gerrit Renker,
David S . Miller, Neil Horman, dccp, netdev, linux-sctp,
Stephen Rothwell, mhiramat
[-- Attachment #1: Type: text/plain, Size: 7154 bytes --]
Hi Masami,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net/master]
[also build test WARNING on v4.15-rc4 next-20171219]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Masami-Hiramatsu/net-tcp-sctp-dccp-Replace-jprobe-usage-with-trace-events/20171220-081035
config: i386-randconfig-x011-201751 (attached as .config)
compiler: gcc-7 (Debian 7.2.0-12) 7.2.1 20171025
reproduce:
# save the attached .config to linux build tree
make ARCH=i386
All warnings (new ones prefixed by >>):
In file included from include/trace/define_trace.h:96:0,
from include/trace/events/sctp.h:96,
from net//sctp/sm_statefuns.c:63:
include/trace/events/sctp.h: In function 'trace_event_raw_event_sctp_probe_path':
>> include/trace/events/sctp.h:31:19: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
__entry->asoc = (__u64)asoc;
^
include/trace/trace_events.h:719:4: note: in definition of macro 'DECLARE_EVENT_CLASS'
{ assign; } \
^~~~~~
include/trace/trace_events.h:78:9: note: in expansion of macro 'PARAMS'
PARAMS(assign), \
^~~~~~
>> include/trace/events/sctp.h:11:1: note: in expansion of macro 'TRACE_EVENT'
TRACE_EVENT(sctp_probe_path,
^~~~~~~~~~~
>> include/trace/events/sctp.h:30:2: note: in expansion of macro 'TP_fast_assign'
TP_fast_assign(
^~~~~~~~~~~~~~
include/trace/events/sctp.h: In function 'trace_event_raw_event_sctp_probe':
include/trace/events/sctp.h:72:19: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
__entry->asoc = (__u64)asoc;
^
include/trace/trace_events.h:719:4: note: in definition of macro 'DECLARE_EVENT_CLASS'
{ assign; } \
^~~~~~
include/trace/trace_events.h:78:9: note: in expansion of macro 'PARAMS'
PARAMS(assign), \
^~~~~~
include/trace/events/sctp.h:50:1: note: in expansion of macro 'TRACE_EVENT'
TRACE_EVENT(sctp_probe,
^~~~~~~~~~~
include/trace/events/sctp.h:68:2: note: in expansion of macro 'TP_fast_assign'
TP_fast_assign(
^~~~~~~~~~~~~~
In file included from include/trace/define_trace.h:97:0,
from include/trace/events/sctp.h:96,
from net//sctp/sm_statefuns.c:63:
include/trace/events/sctp.h: In function 'perf_trace_sctp_probe_path':
>> include/trace/events/sctp.h:31:19: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
__entry->asoc = (__u64)asoc;
^
include/trace/perf.h:66:4: note: in definition of macro 'DECLARE_EVENT_CLASS'
{ assign; } \
^~~~~~
include/trace/trace_events.h:78:9: note: in expansion of macro 'PARAMS'
PARAMS(assign), \
^~~~~~
>> include/trace/events/sctp.h:11:1: note: in expansion of macro 'TRACE_EVENT'
TRACE_EVENT(sctp_probe_path,
^~~~~~~~~~~
>> include/trace/events/sctp.h:30:2: note: in expansion of macro 'TP_fast_assign'
TP_fast_assign(
^~~~~~~~~~~~~~
include/trace/events/sctp.h: In function 'perf_trace_sctp_probe':
include/trace/events/sctp.h:72:19: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
__entry->asoc = (__u64)asoc;
^
include/trace/perf.h:66:4: note: in definition of macro 'DECLARE_EVENT_CLASS'
{ assign; } \
^~~~~~
include/trace/trace_events.h:78:9: note: in expansion of macro 'PARAMS'
PARAMS(assign), \
^~~~~~
include/trace/events/sctp.h:50:1: note: in expansion of macro 'TRACE_EVENT'
TRACE_EVENT(sctp_probe,
^~~~~~~~~~~
include/trace/events/sctp.h:68:2: note: in expansion of macro 'TP_fast_assign'
TP_fast_assign(
^~~~~~~~~~~~~~
vim +31 include/trace/events/sctp.h
10
> 11 TRACE_EVENT(sctp_probe_path,
12
13 TP_PROTO(struct sctp_transport *sp,
14 const struct sctp_association *asoc),
15
16 TP_ARGS(sp, asoc),
17
18 TP_STRUCT__entry(
19 __field(__u64, asoc)
20 __field(__u32, primary)
21 __array(__u8, ipaddr, sizeof(union sctp_addr))
22 __field(__u32, state)
23 __field(__u32, cwnd)
24 __field(__u32, ssthresh)
25 __field(__u32, flight_size)
26 __field(__u32, partial_bytes_acked)
27 __field(__u32, pathmtu)
28 ),
29
> 30 TP_fast_assign(
> 31 __entry->asoc = (__u64)asoc;
32 __entry->primary = (sp == asoc->peer.primary_path);
33 memcpy(__entry->ipaddr, &sp->ipaddr, sizeof(union sctp_addr));
34 __entry->state = sp->state;
35 __entry->cwnd = sp->cwnd;
36 __entry->ssthresh = sp->ssthresh;
37 __entry->flight_size = sp->flight_size;
38 __entry->partial_bytes_acked = sp->partial_bytes_acked;
39 __entry->pathmtu = sp->pathmtu;
40 ),
41
42 TP_printk("asoc=%#llx%s ipaddr=%pISpc state=%u cwnd=%u ssthresh=%u "
43 "flight_size=%u partial_bytes_acked=%u pathmtu=%u",
44 __entry->asoc, __entry->primary ? "(*)" : "",
45 __entry->ipaddr, __entry->state, __entry->cwnd,
46 __entry->ssthresh, __entry->flight_size,
47 __entry->partial_bytes_acked, __entry->pathmtu)
48 );
49
50 TRACE_EVENT(sctp_probe,
51
52 TP_PROTO(const struct sctp_endpoint *ep,
53 const struct sctp_association *asoc,
54 struct sctp_chunk *chunk),
55
56 TP_ARGS(ep, asoc, chunk),
57
58 TP_STRUCT__entry(
59 __field(__u64, asoc)
60 __field(__u32, mark)
61 __field(__u16, bind_port)
62 __field(__u16, peer_port)
63 __field(__u32, pathmtu)
64 __field(__u32, rwnd)
65 __field(__u16, unack_data)
66 ),
67
68 TP_fast_assign(
69 struct sctp_transport *sp;
70 struct sk_buff *skb = chunk->skb;
71
72 __entry->asoc = (__u64)asoc;
73 __entry->mark = skb->mark;
74 __entry->bind_port = ep->base.bind_addr.port;
75 __entry->peer_port = asoc->peer.port;
76 __entry->pathmtu = asoc->pathmtu;
77 __entry->rwnd = asoc->peer.rwnd;
78 __entry->unack_data = asoc->unack_data;
79
80 list_for_each_entry(sp, &asoc->peer.transport_addr_list,
81 transports) {
82 trace_sctp_probe_path(sp, asoc);
83 }
84 ),
85
86 TP_printk("asoc=%#llx mark=%#x bind_port=%d peer_port=%d pathmtu=%d "
87 "rwnd=%u unack_data=%d",
88 __entry->asoc, __entry->mark, __entry->bind_port,
89 __entry->peer_port, __entry->pathmtu, __entry->rwnd,
90 __entry->unack_data)
91 );
92
93 #endif /* _TRACE_SCTP_H */
94
95 /* This part must be outside protection */
> 96 #include <trace/define_trace.h>
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 26730 bytes --]
^ permalink raw reply [flat|nested] 12+ messages in thread
* [v2 PATCH -tip 4/6] net: sctp: Remove debug SCTP probe module
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
` (2 preceding siblings ...)
2017-12-18 8:12 ` [v2 PATCH -tip 3/6] net: sctp: Add SCTP ACK tracking trace event Masami Hiramatsu
@ 2017-12-18 8:12 ` Masami Hiramatsu
2017-12-18 8:13 ` [v2 PATCH -tip 5/6] net: dccp: Add DCCP sendmsg trace event Masami Hiramatsu
` (2 subsequent siblings)
6 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:12 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
Remove SCTP probe module since jprobe has been deprecated.
That function is now replaced by sctp/sctp_probe and
sctp/sctp_probe_path trace-events.
You can use it via ftrace or perftools.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
net/sctp/Kconfig | 12 ---
net/sctp/Makefile | 3 -
net/sctp/probe.c | 244 -----------------------------------------------------
3 files changed, 259 deletions(-)
delete mode 100644 net/sctp/probe.c
diff --git a/net/sctp/Kconfig b/net/sctp/Kconfig
index d9c04dc1b3f3..c740b189d4ba 100644
--- a/net/sctp/Kconfig
+++ b/net/sctp/Kconfig
@@ -37,18 +37,6 @@ menuconfig IP_SCTP
if IP_SCTP
-config NET_SCTPPROBE
- tristate "SCTP: Association probing"
- depends on PROC_FS && KPROBES
- ---help---
- This module allows for capturing the changes to SCTP association
- state in response to incoming packets. It is used for debugging
- SCTP congestion control algorithms. If you don't understand
- what was just said, you don't need it: say N.
-
- To compile this code as a module, choose M here: the
- module will be called sctp_probe.
-
config SCTP_DBG_OBJCNT
bool "SCTP: Debug object counts"
depends on PROC_FS
diff --git a/net/sctp/Makefile b/net/sctp/Makefile
index 1ca84a288443..b52dfc6f8b4c 100644
--- a/net/sctp/Makefile
+++ b/net/sctp/Makefile
@@ -4,7 +4,6 @@
#
obj-$(CONFIG_IP_SCTP) += sctp.o
-obj-$(CONFIG_NET_SCTPPROBE) += sctp_probe.o
obj-$(CONFIG_INET_SCTP_DIAG) += sctp_diag.o
sctp-y := sm_statetable.o sm_statefuns.o sm_sideeffect.o \
@@ -16,8 +15,6 @@ sctp-y := sm_statetable.o sm_statefuns.o sm_sideeffect.o \
offload.o stream_sched.o stream_sched_prio.o \
stream_sched_rr.o
-sctp_probe-y := probe.o
-
sctp-$(CONFIG_SCTP_DBG_OBJCNT) += objcnt.o
sctp-$(CONFIG_PROC_FS) += proc.o
sctp-$(CONFIG_SYSCTL) += sysctl.o
diff --git a/net/sctp/probe.c b/net/sctp/probe.c
deleted file mode 100644
index 1280f85a598d..000000000000
--- a/net/sctp/probe.c
+++ /dev/null
@@ -1,244 +0,0 @@
-/*
- * sctp_probe - Observe the SCTP flow with kprobes.
- *
- * The idea for this came from Werner Almesberger's umlsim
- * Copyright (C) 2004, Stephen Hemminger <shemminger@osdl.org>
- *
- * Modified for SCTP from Stephen Hemminger's code
- * Copyright (C) 2010, Wei Yongjun <yjwei@cn.fujitsu.com>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- */
-
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
-#include <linux/kernel.h>
-#include <linux/kprobes.h>
-#include <linux/socket.h>
-#include <linux/sctp.h>
-#include <linux/proc_fs.h>
-#include <linux/vmalloc.h>
-#include <linux/module.h>
-#include <linux/kfifo.h>
-#include <linux/time.h>
-#include <net/net_namespace.h>
-
-#include <net/sctp/sctp.h>
-#include <net/sctp/sm.h>
-
-MODULE_SOFTDEP("pre: sctp");
-MODULE_AUTHOR("Wei Yongjun <yjwei@cn.fujitsu.com>");
-MODULE_DESCRIPTION("SCTP snooper");
-MODULE_LICENSE("GPL");
-
-static int port __read_mostly = 0;
-MODULE_PARM_DESC(port, "Port to match (0=all)");
-module_param(port, int, 0);
-
-static unsigned int fwmark __read_mostly = 0;
-MODULE_PARM_DESC(fwmark, "skb mark to match (0=no mark)");
-module_param(fwmark, uint, 0);
-
-static int bufsize __read_mostly = 64 * 1024;
-MODULE_PARM_DESC(bufsize, "Log buffer size (default 64k)");
-module_param(bufsize, int, 0);
-
-static int full __read_mostly = 1;
-MODULE_PARM_DESC(full, "Full log (1=every ack packet received, 0=only cwnd changes)");
-module_param(full, int, 0);
-
-static const char procname[] = "sctpprobe";
-
-static struct {
- struct kfifo fifo;
- spinlock_t lock;
- wait_queue_head_t wait;
- struct timespec64 tstart;
-} sctpw;
-
-static __printf(1, 2) void printl(const char *fmt, ...)
-{
- va_list args;
- int len;
- char tbuf[256];
-
- va_start(args, fmt);
- len = vscnprintf(tbuf, sizeof(tbuf), fmt, args);
- va_end(args);
-
- kfifo_in_locked(&sctpw.fifo, tbuf, len, &sctpw.lock);
- wake_up(&sctpw.wait);
-}
-
-static int sctpprobe_open(struct inode *inode, struct file *file)
-{
- kfifo_reset(&sctpw.fifo);
- ktime_get_ts64(&sctpw.tstart);
-
- return 0;
-}
-
-static ssize_t sctpprobe_read(struct file *file, char __user *buf,
- size_t len, loff_t *ppos)
-{
- int error = 0, cnt = 0;
- unsigned char *tbuf;
-
- if (!buf)
- return -EINVAL;
-
- if (len == 0)
- return 0;
-
- tbuf = vmalloc(len);
- if (!tbuf)
- return -ENOMEM;
-
- error = wait_event_interruptible(sctpw.wait,
- kfifo_len(&sctpw.fifo) != 0);
- if (error)
- goto out_free;
-
- cnt = kfifo_out_locked(&sctpw.fifo, tbuf, len, &sctpw.lock);
- error = copy_to_user(buf, tbuf, cnt) ? -EFAULT : 0;
-
-out_free:
- vfree(tbuf);
-
- return error ? error : cnt;
-}
-
-static const struct file_operations sctpprobe_fops = {
- .owner = THIS_MODULE,
- .open = sctpprobe_open,
- .read = sctpprobe_read,
- .llseek = noop_llseek,
-};
-
-static enum sctp_disposition jsctp_sf_eat_sack(
- struct net *net,
- const struct sctp_endpoint *ep,
- const struct sctp_association *asoc,
- const union sctp_subtype type,
- void *arg,
- struct sctp_cmd_seq *commands)
-{
- struct sctp_chunk *chunk = arg;
- struct sk_buff *skb = chunk->skb;
- struct sctp_transport *sp;
- static __u32 lcwnd = 0;
- struct timespec64 now;
-
- sp = asoc->peer.primary_path;
-
- if (((port == 0 && fwmark == 0) ||
- asoc->peer.port == port ||
- ep->base.bind_addr.port == port ||
- (fwmark > 0 && skb->mark == fwmark)) &&
- (full || sp->cwnd != lcwnd)) {
- lcwnd = sp->cwnd;
-
- ktime_get_ts64(&now);
- now = timespec64_sub(now, sctpw.tstart);
-
- printl("%lu.%06lu ", (unsigned long) now.tv_sec,
- (unsigned long) now.tv_nsec / NSEC_PER_USEC);
-
- printl("%p %5d %5d %5d %8d %5d ", asoc,
- ep->base.bind_addr.port, asoc->peer.port,
- asoc->pathmtu, asoc->peer.rwnd, asoc->unack_data);
-
- list_for_each_entry(sp, &asoc->peer.transport_addr_list,
- transports) {
- if (sp == asoc->peer.primary_path)
- printl("*");
-
- printl("%pISc %2u %8u %8u %8u %8u %8u ",
- &sp->ipaddr, sp->state, sp->cwnd, sp->ssthresh,
- sp->flight_size, sp->partial_bytes_acked,
- sp->pathmtu);
- }
- printl("\n");
- }
-
- jprobe_return();
- return 0;
-}
-
-static struct jprobe sctp_recv_probe = {
- .kp = {
- .symbol_name = "sctp_sf_eat_sack_6_2",
- },
- .entry = jsctp_sf_eat_sack,
-};
-
-static __init int sctp_setup_jprobe(void)
-{
- int ret = register_jprobe(&sctp_recv_probe);
-
- if (ret) {
- if (request_module("sctp"))
- goto out;
- ret = register_jprobe(&sctp_recv_probe);
- }
-
-out:
- return ret;
-}
-
-static __init int sctpprobe_init(void)
-{
- int ret = -ENOMEM;
-
- /* Warning: if the function signature of sctp_sf_eat_sack_6_2,
- * has been changed, you also have to change the signature of
- * jsctp_sf_eat_sack, otherwise you end up right here!
- */
- BUILD_BUG_ON(__same_type(sctp_sf_eat_sack_6_2,
- jsctp_sf_eat_sack) == 0);
-
- init_waitqueue_head(&sctpw.wait);
- spin_lock_init(&sctpw.lock);
- if (kfifo_alloc(&sctpw.fifo, bufsize, GFP_KERNEL))
- return ret;
-
- if (!proc_create(procname, S_IRUSR, init_net.proc_net,
- &sctpprobe_fops))
- goto free_kfifo;
-
- ret = sctp_setup_jprobe();
- if (ret)
- goto remove_proc;
-
- pr_info("probe registered (port=%d/fwmark=%u) bufsize=%u\n",
- port, fwmark, bufsize);
- return 0;
-
-remove_proc:
- remove_proc_entry(procname, init_net.proc_net);
-free_kfifo:
- kfifo_free(&sctpw.fifo);
- return ret;
-}
-
-static __exit void sctpprobe_exit(void)
-{
- kfifo_free(&sctpw.fifo);
- remove_proc_entry(procname, init_net.proc_net);
- unregister_jprobe(&sctp_recv_probe);
-}
-
-module_init(sctpprobe_init);
-module_exit(sctpprobe_exit);
^ permalink raw reply related [flat|nested] 12+ messages in thread* [v2 PATCH -tip 5/6] net: dccp: Add DCCP sendmsg trace event
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
` (3 preceding siblings ...)
2017-12-18 8:12 ` [v2 PATCH -tip 4/6] net: sctp: Remove debug SCTP probe module Masami Hiramatsu
@ 2017-12-18 8:13 ` Masami Hiramatsu
2017-12-18 8:13 ` [v2 PATCH -tip 6/6] net: dccp: Remove dccpprobe module Masami Hiramatsu
2017-12-19 8:31 ` [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
6 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:13 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
Add DCCP sendmsg trace event (dccp/dccp_probe) for
replacing dccpprobe. User can trace this event via
ftrace or perftools.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
net/dccp/proto.c | 5 +++
net/dccp/trace.h | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 110 insertions(+)
create mode 100644 net/dccp/trace.h
diff --git a/net/dccp/proto.c b/net/dccp/proto.c
index 9d43c1f40274..e57b5db495cd 100644
--- a/net/dccp/proto.c
+++ b/net/dccp/proto.c
@@ -38,6 +38,9 @@
#include "dccp.h"
#include "feat.h"
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+
DEFINE_SNMP_STAT(struct dccp_mib, dccp_statistics) __read_mostly;
EXPORT_SYMBOL_GPL(dccp_statistics);
@@ -761,6 +764,8 @@ int dccp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
int rc, size;
long timeo;
+ trace_dccp_probe(sk, len);
+
if (len > dp->dccps_mss_cache)
return -EMSGSIZE;
diff --git a/net/dccp/trace.h b/net/dccp/trace.h
new file mode 100644
index 000000000000..aa01321a6c37
--- /dev/null
+++ b/net/dccp/trace.h
@@ -0,0 +1,105 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM dccp
+
+#if !defined(_TRACE_DCCP_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_DCCP_H
+
+#include <net/sock.h>
+#include "dccp.h"
+#include "ccids/ccid3.h"
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(dccp_probe,
+
+ TP_PROTO(struct sock *sk, size_t size),
+
+ TP_ARGS(sk, size),
+
+ TP_STRUCT__entry(
+ /* sockaddr_in6 is always bigger than sockaddr_in */
+ __array(__u8, saddr, sizeof(struct sockaddr_in6))
+ __array(__u8, daddr, sizeof(struct sockaddr_in6))
+ __field(__u16, sport)
+ __field(__u16, dport)
+ __field(__u16, size)
+ __field(__u16, tx_s)
+ __field(__u32, tx_rtt)
+ __field(__u32, tx_p)
+ __field(__u32, tx_x_calc)
+ __field(__u64, tx_x_recv)
+ __field(__u64, tx_x)
+ __field(__u32, tx_t_ipi)
+ ),
+
+ TP_fast_assign(
+ const struct inet_sock *inet = inet_sk(sk);
+ struct ccid3_hc_tx_sock *hc = NULL;
+
+ if (ccid_get_current_tx_ccid(dccp_sk(sk)) == DCCPC_CCID3)
+ hc = ccid3_hc_tx_sk(sk);
+
+ memset(__entry->saddr, 0, sizeof(struct sockaddr_in6));
+ memset(__entry->daddr, 0, sizeof(struct sockaddr_in6));
+
+ if (sk->sk_family == AF_INET) {
+ struct sockaddr_in *v4 = (void *)__entry->saddr;
+
+ v4->sin_family = AF_INET;
+ v4->sin_port = inet->inet_sport;
+ v4->sin_addr.s_addr = inet->inet_saddr;
+ v4 = (void *)__entry->daddr;
+ v4->sin_family = AF_INET;
+ v4->sin_port = inet->inet_dport;
+ v4->sin_addr.s_addr = inet->inet_daddr;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (sk->sk_family == AF_INET6) {
+ struct sockaddr_in6 *v6 = (void *)__entry->saddr;
+
+ v6->sin6_family = AF_INET6;
+ v6->sin6_port = inet->inet_sport;
+ v6->sin6_addr = inet6_sk(sk)->saddr;
+ v6 = (void *)__entry->daddr;
+ v6->sin6_family = AF_INET6;
+ v6->sin6_port = inet->inet_dport;
+ v6->sin6_addr = sk->sk_v6_daddr;
+#endif
+ }
+
+ /* For filtering use */
+ __entry->sport = ntohs(inet->inet_sport);
+ __entry->dport = ntohs(inet->inet_dport);
+
+ __entry->size = size;
+ if (hc) {
+ __entry->tx_s = hc->tx_s;
+ __entry->tx_rtt = hc->tx_rtt;
+ __entry->tx_p = hc->tx_p;
+ __entry->tx_x_calc = hc->tx_x_calc;
+ __entry->tx_x_recv = hc->tx_x_recv >> 6;
+ __entry->tx_x = hc->tx_x >> 6;
+ __entry->tx_t_ipi = hc->tx_t_ipi;
+ } else {
+ __entry->tx_s = 0;
+ memset(&__entry->tx_rtt, 0, (void *)&__entry->tx_t_ipi -
+ (void *)&__entry->tx_rtt +
+ sizeof(__entry->tx_t_ipi));
+ }
+ ),
+
+ TP_printk("src=%pISpc dest=%pISpc size=%d tx_s=%d tx_rtt=%d "
+ "tx_p=%d tx_x_calc=%u tx_x_recv=%llu tx_x=%llu tx_t_ipi=%d",
+ __entry->saddr, __entry->daddr, __entry->size,
+ __entry->tx_s, __entry->tx_rtt, __entry->tx_p,
+ __entry->tx_x_calc, __entry->tx_x_recv, __entry->tx_x,
+ __entry->tx_t_ipi)
+);
+
+#endif /* _TRACE_TCP_H */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+#include <trace/define_trace.h>
^ permalink raw reply related [flat|nested] 12+ messages in thread* [v2 PATCH -tip 6/6] net: dccp: Remove dccpprobe module
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
` (4 preceding siblings ...)
2017-12-18 8:13 ` [v2 PATCH -tip 5/6] net: dccp: Add DCCP sendmsg trace event Masami Hiramatsu
@ 2017-12-18 8:13 ` Masami Hiramatsu
2017-12-19 8:31 ` [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
6 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-18 8:13 UTC (permalink / raw)
To: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt
Cc: Peter Zijlstra, Thomas Gleixner, LKML, H . Peter Anvin,
Gerrit Renker, David S . Miller, Neil Horman, dccp, netdev,
linux-sctp, Stephen Rothwell, mhiramat
Remove DCCP probe module since jprobe has been deprecated.
That function is now replaced by dccp/dccp_probe trace-event.
You can use it via ftrace or perftools.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
net/dccp/Kconfig | 17 ----
net/dccp/Makefile | 2 -
net/dccp/probe.c | 203 -----------------------------------------------------
3 files changed, 222 deletions(-)
delete mode 100644 net/dccp/probe.c
diff --git a/net/dccp/Kconfig b/net/dccp/Kconfig
index 8c0ef71bed2f..b270e84d9c13 100644
--- a/net/dccp/Kconfig
+++ b/net/dccp/Kconfig
@@ -39,23 +39,6 @@ config IP_DCCP_DEBUG
Just say N.
-config NET_DCCPPROBE
- tristate "DCCP connection probing"
- depends on PROC_FS && KPROBES
- ---help---
- This module allows for capturing the changes to DCCP connection
- state in response to incoming packets. It is used for debugging
- DCCP congestion avoidance modules. If you don't understand
- what was just said, you don't need it: say N.
-
- Documentation on how to use DCCP connection probing can be found
- at:
-
- http://www.linuxfoundation.org/collaborate/workgroups/networking/dccpprobe
-
- To compile this code as a module, choose M here: the
- module will be called dccp_probe.
-
endmenu
diff --git a/net/dccp/Makefile b/net/dccp/Makefile
index 2e7b56097bc4..9d0383d2f277 100644
--- a/net/dccp/Makefile
+++ b/net/dccp/Makefile
@@ -21,9 +21,7 @@ obj-$(subst y,$(CONFIG_IP_DCCP),$(CONFIG_IPV6)) += dccp_ipv6.o
dccp_ipv6-y := ipv6.o
obj-$(CONFIG_INET_DCCP_DIAG) += dccp_diag.o
-obj-$(CONFIG_NET_DCCPPROBE) += dccp_probe.o
dccp-$(CONFIG_SYSCTL) += sysctl.o
dccp_diag-y := diag.o
-dccp_probe-y := probe.o
diff --git a/net/dccp/probe.c b/net/dccp/probe.c
deleted file mode 100644
index 3d3fda05b32d..000000000000
--- a/net/dccp/probe.c
+++ /dev/null
@@ -1,203 +0,0 @@
-/*
- * dccp_probe - Observe the DCCP flow with kprobes.
- *
- * The idea for this came from Werner Almesberger's umlsim
- * Copyright (C) 2004, Stephen Hemminger <shemminger@osdl.org>
- *
- * Modified for DCCP from Stephen Hemminger's code
- * Copyright (C) 2006, Ian McDonald <ian.mcdonald@jandi.co.nz>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- */
-
-#include <linux/kernel.h>
-#include <linux/kprobes.h>
-#include <linux/socket.h>
-#include <linux/dccp.h>
-#include <linux/proc_fs.h>
-#include <linux/module.h>
-#include <linux/kfifo.h>
-#include <linux/vmalloc.h>
-#include <linux/time64.h>
-#include <linux/gfp.h>
-#include <net/net_namespace.h>
-
-#include "dccp.h"
-#include "ccid.h"
-#include "ccids/ccid3.h"
-
-static int port;
-
-static int bufsize = 64 * 1024;
-
-static const char procname[] = "dccpprobe";
-
-static struct {
- struct kfifo fifo;
- spinlock_t lock;
- wait_queue_head_t wait;
- struct timespec64 tstart;
-} dccpw;
-
-static void printl(const char *fmt, ...)
-{
- va_list args;
- int len;
- struct timespec64 now;
- char tbuf[256];
-
- va_start(args, fmt);
- getnstimeofday64(&now);
-
- now = timespec64_sub(now, dccpw.tstart);
-
- len = sprintf(tbuf, "%lu.%06lu ",
- (unsigned long) now.tv_sec,
- (unsigned long) now.tv_nsec / NSEC_PER_USEC);
- len += vscnprintf(tbuf+len, sizeof(tbuf)-len, fmt, args);
- va_end(args);
-
- kfifo_in_locked(&dccpw.fifo, tbuf, len, &dccpw.lock);
- wake_up(&dccpw.wait);
-}
-
-static int jdccp_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
-{
- const struct inet_sock *inet = inet_sk(sk);
- struct ccid3_hc_tx_sock *hc = NULL;
-
- if (ccid_get_current_tx_ccid(dccp_sk(sk)) == DCCPC_CCID3)
- hc = ccid3_hc_tx_sk(sk);
-
- if (port == 0 || ntohs(inet->inet_dport) == port ||
- ntohs(inet->inet_sport) == port) {
- if (hc)
- printl("%pI4:%u %pI4:%u %d %d %d %d %u %llu %llu %d\n",
- &inet->inet_saddr, ntohs(inet->inet_sport),
- &inet->inet_daddr, ntohs(inet->inet_dport), size,
- hc->tx_s, hc->tx_rtt, hc->tx_p,
- hc->tx_x_calc, hc->tx_x_recv >> 6,
- hc->tx_x >> 6, hc->tx_t_ipi);
- else
- printl("%pI4:%u %pI4:%u %d\n",
- &inet->inet_saddr, ntohs(inet->inet_sport),
- &inet->inet_daddr, ntohs(inet->inet_dport),
- size);
- }
-
- jprobe_return();
- return 0;
-}
-
-static struct jprobe dccp_send_probe = {
- .kp = {
- .symbol_name = "dccp_sendmsg",
- },
- .entry = jdccp_sendmsg,
-};
-
-static int dccpprobe_open(struct inode *inode, struct file *file)
-{
- kfifo_reset(&dccpw.fifo);
- getnstimeofday64(&dccpw.tstart);
- return 0;
-}
-
-static ssize_t dccpprobe_read(struct file *file, char __user *buf,
- size_t len, loff_t *ppos)
-{
- int error = 0, cnt = 0;
- unsigned char *tbuf;
-
- if (!buf)
- return -EINVAL;
-
- if (len == 0)
- return 0;
-
- tbuf = vmalloc(len);
- if (!tbuf)
- return -ENOMEM;
-
- error = wait_event_interruptible(dccpw.wait,
- kfifo_len(&dccpw.fifo) != 0);
- if (error)
- goto out_free;
-
- cnt = kfifo_out_locked(&dccpw.fifo, tbuf, len, &dccpw.lock);
- error = copy_to_user(buf, tbuf, cnt) ? -EFAULT : 0;
-
-out_free:
- vfree(tbuf);
-
- return error ? error : cnt;
-}
-
-static const struct file_operations dccpprobe_fops = {
- .owner = THIS_MODULE,
- .open = dccpprobe_open,
- .read = dccpprobe_read,
- .llseek = noop_llseek,
-};
-
-static __init int dccpprobe_init(void)
-{
- int ret = -ENOMEM;
-
- init_waitqueue_head(&dccpw.wait);
- spin_lock_init(&dccpw.lock);
- if (kfifo_alloc(&dccpw.fifo, bufsize, GFP_KERNEL))
- return ret;
- if (!proc_create(procname, S_IRUSR, init_net.proc_net, &dccpprobe_fops))
- goto err0;
-
- ret = register_jprobe(&dccp_send_probe);
- if (ret) {
- ret = request_module("dccp");
- if (!ret)
- ret = register_jprobe(&dccp_send_probe);
- }
-
- if (ret)
- goto err1;
-
- pr_info("DCCP watch registered (port=%d)\n", port);
- return 0;
-err1:
- remove_proc_entry(procname, init_net.proc_net);
-err0:
- kfifo_free(&dccpw.fifo);
- return ret;
-}
-module_init(dccpprobe_init);
-
-static __exit void dccpprobe_exit(void)
-{
- kfifo_free(&dccpw.fifo);
- remove_proc_entry(procname, init_net.proc_net);
- unregister_jprobe(&dccp_send_probe);
-
-}
-module_exit(dccpprobe_exit);
-
-MODULE_PARM_DESC(port, "Port to match (0=all)");
-module_param(port, int, 0);
-
-MODULE_PARM_DESC(bufsize, "Log buffer size (default 64k)");
-module_param(bufsize, int, 0);
-
-MODULE_AUTHOR("Ian McDonald <ian.mcdonald@jandi.co.nz>");
-MODULE_DESCRIPTION("DCCP snooper");
-MODULE_LICENSE("GPL");
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events
2017-12-18 8:10 [v2 PATCH -tip 0/6] net: tcp: sctp: dccp: Replace jprobe usage with trace events Masami Hiramatsu
` (5 preceding siblings ...)
2017-12-18 8:13 ` [v2 PATCH -tip 6/6] net: dccp: Remove dccpprobe module Masami Hiramatsu
@ 2017-12-19 8:31 ` Masami Hiramatsu
6 siblings, 0 replies; 12+ messages in thread
From: Masami Hiramatsu @ 2017-12-19 8:31 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Ingo Molnar, Ian McDonald, Vlad Yasevich, Stephen Hemminger,
Steven Rostedt, Peter Zijlstra, Thomas Gleixner, LKML,
H . Peter Anvin, Gerrit Renker, David S . Miller, Neil Horman,
dccp, netdev, linux-sctp, Stephen Rothwell
Oops, I found a build error for tcp.h
I'll update it.
On Mon, 18 Dec 2017 17:10:45 +0900
Masami Hiramatsu <mhiramat@kernel.org> wrote:
> 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@kernel.org>
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 12+ messages in thread