forked from iovisor/bcc
-
Notifications
You must be signed in to change notification settings - Fork 2
/
tcpdrop.py
executable file
·218 lines (197 loc) · 6.49 KB
/
tcpdrop.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
#!/usr/bin/python
# @lint-avoid-python-3-compatibility-imports
#
# tcpdrop Trace TCP kernel-dropped packets/segments.
# For Linux, uses BCC, eBPF. Embedded C.
#
# This provides information such as packet details, socket state, and kernel
# stack trace for packets/segments that were dropped via tcp_drop().
#
# USAGE: tcpdrop [-4 | -6] [-h]
#
# This uses dynamic tracing of kernel functions, and will need to be updated
# to match kernel changes.
#
# Copyright 2018 Netflix, Inc.
# Licensed under the Apache License, Version 2.0 (the "License")
#
# 30-May-2018 Brendan Gregg Created this.
from __future__ import print_function
from bcc import BPF
import argparse
from time import strftime
from socket import inet_ntop, AF_INET, AF_INET6
from struct import pack
from time import sleep
from bcc import tcp
# arguments
examples = """examples:
./tcpdrop # trace kernel TCP drops
./tcpdrop -4 # trace IPv4 family only
./tcpdrop -6 # trace IPv6 family only
"""
parser = argparse.ArgumentParser(
description="Trace TCP drops by the kernel",
formatter_class=argparse.RawDescriptionHelpFormatter,
epilog=examples)
group = parser.add_mutually_exclusive_group()
group.add_argument("-4", "--ipv4", action="store_true",
help="trace IPv4 family only")
group.add_argument("-6", "--ipv6", action="store_true",
help="trace IPv6 family only")
parser.add_argument("--ebpf", action="store_true",
help=argparse.SUPPRESS)
args = parser.parse_args()
debug = 0
# define BPF program
bpf_text = """
#include <uapi/linux/ptrace.h>
#include <uapi/linux/tcp.h>
#include <uapi/linux/ip.h>
#include <net/sock.h>
#include <bcc/proto.h>
BPF_STACK_TRACE(stack_traces, 1024);
// separate data structs for ipv4 and ipv6
struct ipv4_data_t {
u32 pid;
u64 ip;
u32 saddr;
u32 daddr;
u16 sport;
u16 dport;
u8 state;
u8 tcpflags;
u32 stack_id;
};
BPF_PERF_OUTPUT(ipv4_events);
struct ipv6_data_t {
u32 pid;
u64 ip;
unsigned __int128 saddr;
unsigned __int128 daddr;
u16 sport;
u16 dport;
u8 state;
u8 tcpflags;
u32 stack_id;
};
BPF_PERF_OUTPUT(ipv6_events);
static struct tcphdr *skb_to_tcphdr(const struct sk_buff *skb)
{
// unstable API. verify logic in tcp_hdr() -> skb_transport_header().
return (struct tcphdr *)(skb->head + skb->transport_header);
}
static inline struct iphdr *skb_to_iphdr(const struct sk_buff *skb)
{
// unstable API. verify logic in ip_hdr() -> skb_network_header().
return (struct iphdr *)(skb->head + skb->network_header);
}
// from include/net/tcp.h:
#ifndef tcp_flag_byte
#define tcp_flag_byte(th) (((u_int8_t *)th)[13])
#endif
int trace_tcp_drop(struct pt_regs *ctx, struct sock *sk, struct sk_buff *skb)
{
if (sk == NULL)
return 0;
u32 pid = bpf_get_current_pid_tgid() >> 32;
// pull in details from the packet headers and the sock struct
u16 family = sk->__sk_common.skc_family;
char state = sk->__sk_common.skc_state;
u16 sport = 0, dport = 0;
struct tcphdr *tcp = skb_to_tcphdr(skb);
struct iphdr *ip = skb_to_iphdr(skb);
u8 tcpflags = ((u_int8_t *)tcp)[13];
sport = tcp->source;
dport = tcp->dest;
sport = ntohs(sport);
dport = ntohs(dport);
FILTER_FAMILY
if (family == AF_INET) {
struct ipv4_data_t data4 = {};
data4.pid = pid;
data4.ip = 4;
data4.saddr = ip->saddr;
data4.daddr = ip->daddr;
data4.dport = dport;
data4.sport = sport;
data4.state = state;
data4.tcpflags = tcpflags;
data4.stack_id = stack_traces.get_stackid(ctx, 0);
ipv4_events.perf_submit(ctx, &data4, sizeof(data4));
} else if (family == AF_INET6) {
struct ipv6_data_t data6 = {};
data6.pid = pid;
data6.ip = 6;
// The remote address (skc_v6_daddr) was the source
bpf_probe_read_kernel(&data6.saddr, sizeof(data6.saddr),
sk->__sk_common.skc_v6_daddr.in6_u.u6_addr32);
// The local address (skc_v6_rcv_saddr) was the destination
bpf_probe_read_kernel(&data6.daddr, sizeof(data6.daddr),
sk->__sk_common.skc_v6_rcv_saddr.in6_u.u6_addr32);
data6.dport = dport;
data6.sport = sport;
data6.state = state;
data6.tcpflags = tcpflags;
data6.stack_id = stack_traces.get_stackid(ctx, 0);
ipv6_events.perf_submit(ctx, &data6, sizeof(data6));
}
// else drop
return 0;
}
"""
if debug or args.ebpf:
print(bpf_text)
if args.ebpf:
exit()
if args.ipv4:
bpf_text = bpf_text.replace('FILTER_FAMILY',
'if (family != AF_INET) { return 0; }')
elif args.ipv6:
bpf_text = bpf_text.replace('FILTER_FAMILY',
'if (family != AF_INET6) { return 0; }')
else:
bpf_text = bpf_text.replace('FILTER_FAMILY', '')
# process event
def print_ipv4_event(cpu, data, size):
event = b["ipv4_events"].event(data)
print("%-8s %-7d %-2d %-20s > %-20s %s (%s)" % (
strftime("%H:%M:%S"), event.pid, event.ip,
"%s:%d" % (inet_ntop(AF_INET, pack('I', event.saddr)), event.sport),
"%s:%s" % (inet_ntop(AF_INET, pack('I', event.daddr)), event.dport),
tcp.tcpstate[event.state], tcp.flags2str(event.tcpflags)))
for addr in stack_traces.walk(event.stack_id):
sym = b.ksym(addr, show_offset=True)
print("\t%s" % sym)
print("")
def print_ipv6_event(cpu, data, size):
event = b["ipv6_events"].event(data)
print("%-8s %-7d %-2d %-20s > %-20s %s (%s)" % (
strftime("%H:%M:%S"), event.pid, event.ip,
"%s:%d" % (inet_ntop(AF_INET6, event.saddr), event.sport),
"%s:%d" % (inet_ntop(AF_INET6, event.daddr), event.dport),
tcp.tcpstate[event.state], tcp.flags2str(event.tcpflags)))
for addr in stack_traces.walk(event.stack_id):
sym = b.ksym(addr, show_offset=True)
print("\t%s" % sym)
print("")
# initialize BPF
b = BPF(text=bpf_text)
if b.get_kprobe_functions(b"tcp_drop"):
b.attach_kprobe(event="tcp_drop", fn_name="trace_tcp_drop")
else:
print("ERROR: tcp_drop() kernel function not found or traceable. "
"The kernel might be too old or the the function has been inlined.")
exit()
stack_traces = b.get_table("stack_traces")
# header
print("%-8s %-7s %-2s %-20s > %-20s %s (%s)" % ("TIME", "PID", "IP",
"SADDR:SPORT", "DADDR:DPORT", "STATE", "FLAGS"))
# read events
b["ipv4_events"].open_perf_buffer(print_ipv4_event)
b["ipv6_events"].open_perf_buffer(print_ipv6_event)
while 1:
try:
b.perf_buffer_poll()
except KeyboardInterrupt:
exit()