Skip to content

Commit

Permalink
tracing: Tracepoints for in- and outbound P2P msgs
Browse files Browse the repository at this point in the history
Can be used to monitor in- and outbound node traffic.

Based on ealier work by jb55.

Co-authored-by: William Casarin <[email protected]>
  • Loading branch information
0xB10C and jb55 committed Jul 27, 2021
1 parent 469b71a commit 4224dec
Show file tree
Hide file tree
Showing 7 changed files with 639 additions and 1 deletion.
120 changes: 120 additions & 0 deletions contrib/tracing/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -43,3 +43,123 @@ example, to point to release builds if needed. See the
kernel VM. This means the bpftrace and BCC examples must be executed with root
privileges. Make sure to carefully review any scripts that you run with root
privileges first!**

### log_p2p_traffic.bt

A bpftrace script logging information about inbound and outbound P2P network
messages. Based on the `net:inbound_message` and `net:outbound_message`
tracepoints.

By default, `bpftrace` limits strings to 64 bytes due to the limited stack size
in the eBPF VM. For example, Tor v3 addresses exceed the string size limit which
results in the port being cut off during logging. The string size limit can be
increased with the `BPFTRACE_STRLEN` environment variable (`BPFTRACE_STRLEN=70`
works fine).

```
$ bpftrace contrib/tracing/log_p2p_traffic.bt
```

Output
```
outbound 'ping' msg to peer 11 (outbound-full-relay, [2a02:b10c:f747:1:ef:fake:ipv6:addr]:8333) with 8 bytes
inbound 'pong' msg from peer 11 (outbound-full-relay, [2a02:b10c:f747:1:ef:fake:ipv6:addr]:8333) with 8 bytes
inbound 'inv' msg from peer 16 (outbound-full-relay, XX.XX.XXX.121:8333) with 37 bytes
outbound 'getdata' msg to peer 16 (outbound-full-relay, XX.XX.XXX.121:8333) with 37 bytes
inbound 'tx' msg from peer 16 (outbound-full-relay, XX.XX.XXX.121:8333) with 222 bytes
outbound 'inv' msg to peer 9 (outbound-full-relay, faketorv3addressa2ufa6odvoi3s77j4uegey0xb10csyfyve2t33curbyd.onion:8333) with 37 bytes
outbound 'inv' msg to peer 7 (outbound-full-relay, XX.XX.XXX.242:8333) with 37 bytes
```

### p2p_monitor.py

A BCC Python script using curses for an interactive P2P message monitor. Based
on the `net:inbound_message` and `net:outbound_message` tracepoints.

Inbound and outbound traffic is listed for each peer together with information
about the connection. Peers can be selected individually to view recent P2P
messages.

```
$ python3 contrib/tracing/p2p_monitor.py ./src/bitcoind
```

Lists selectable peers and traffic and connection information.
```
P2P Message Monitor
Navigate with UP/DOWN or J/K and select a peer with ENTER or SPACE to see individual P2P messages
PEER OUTBOUND INBOUND TYPE ADDR
0 46 398 byte 61 1407590 byte block-relay-only XX.XX.XXX.196:8333
11 1156 253570 byte 3431 2394924 byte outbound-full-relay XXX.X.XX.179:8333
13 3425 1809620 byte 1236 305458 byte inbound XXX.X.X.X:60380
16 1046 241633 byte 1589 1199220 byte outbound-full-relay 4faketorv2pbfu7x.onion:8333
19 577 181679 byte 390 148951 byte outbound-full-relay kfake4vctorjv2o2.onion:8333
20 11 1248 byte 13 1283 byte block-relay-only [2600:fake:64d9:b10c:4436:aaaa:fe:bb]:8333
21 11 1248 byte 13 1299 byte block-relay-only XX.XXX.X.155:8333
22 5 103 byte 1 102 byte feeler XX.XX.XXX.173:8333
23 11 1248 byte 12 1255 byte block-relay-only XX.XXX.XXX.220:8333
24 3 103 byte 1 102 byte feeler XXX.XXX.XXX.64:8333
```

Showing recent P2P messages between our node and a selected peer.

```
----------------------------------------------------------------------
| PEER 16 (4faketorv2pbfu7x.onion:8333) |
| OUR NODE outbound-full-relay PEER |
| <--- sendcmpct (9 bytes) |
| inv (37 byte) ---> |
| <--- ping (8 bytes) |
| pong (8 byte) ---> |
| inv (37 byte) ---> |
| <--- addr (31 bytes) |
| inv (37 byte) ---> |
| <--- getheaders (1029 bytes) |
| headers (1 byte) ---> |
| <--- feefilter (8 bytes) |
| <--- pong (8 bytes) |
| <--- headers (82 bytes) |
| <--- addr (30003 bytes) |
| inv (1261 byte) ---> |
| … |
```

### log_raw_p2p_msgs.py

A BCC Python script showcasing eBPF and USDT limitations when passing data
larger than about 32kb. Based on the `net:inbound_message` and
`net:outbound_message` tracepoints.

Bitcoin P2P messages can be larger than 32kb (e.g. `tx`, `block`, ...). The
eBPF VM's stack is limited to 512 bytes, and we can't allocate more than about
32kb for a P2P message in the eBPF VM. The **message data is cut off** when the
message is larger than MAX_MSG_DATA_LENGTH (see script). This can be detected
in user-space by comparing the data length to the message length variable. The
message is cut off when the data length is smaller than the message length.
A warning is included with the printed message data.

Data is submitted to user-space (i.e. to this script) via a ring buffer. The
throughput of the ring buffer is limited. Each p2p_message is about 32kb in
size. In- or outbound messages submitted to the ring buffer in rapid
succession fill the ring buffer faster than it can be read. Some messages are
lost. BCC prints: `Possibly lost 2 samples` on lost messages.


```
$ python3 contrib/tracing/log_raw_p2p_msgs.py ./src/bitcoind
```

```
Logging raw P2P messages.
Messages larger that about 32kb will be cut off!
Some messages might be lost!
outbound msg 'inv' from peer 4 (outbound-full-relay, XX.XXX.XX.4:8333) with 253 bytes: 0705000000be2245c8f844c9f763748e1a7…
Warning: incomplete message (only 32568 out of 53552 bytes)! inbound msg 'tx' from peer 32 (outbound-full-relay, XX.XXX.XXX.43:8333) with 53552 bytes: 020000000001fd3c01939c85ad6756ed9fc…
Possibly lost 2 samples
```
28 changes: 28 additions & 0 deletions contrib/tracing/log_p2p_traffic.bt
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
#!/usr/bin/env bpftrace

BEGIN
{
printf("Logging P2P traffic\n")
}

usdt:./src/bitcoind:net:inbound_message
{
$peer_id = (int64) arg0;
$peer_addr = str(arg1);
$peer_type = str(arg2);
$msg_type = str(arg3);
$msg_len = arg4;
printf("inbound '%s' msg from peer %d (%s, %s) with %d bytes\n", $msg_type, $peer_id, $peer_type, $peer_addr, $msg_len);
}

usdt:./src/bitcoind:net:outbound_message
{
$peer_id = (int64) arg0;
$peer_addr = str(arg1);
$peer_type = str(arg2);
$msg_type = str(arg3);
$msg_len = arg4;

printf("outbound '%s' msg to peer %d (%s, %s) with %d bytes\n", $msg_type, $peer_id, $peer_type, $peer_addr, $msg_len);
}

180 changes: 180 additions & 0 deletions contrib/tracing/log_raw_p2p_msgs.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,180 @@
#!/usr/bin/env python3

""" Demonstration of eBPF limitations and the effect on USDT with the
net:inbound_message and net:outbound_message tracepoints. """

# This script shows a limitation of eBPF when data larger than 32kb is passed to
# user-space. It uses BCC (https://github.com/iovisor/bcc) to load a sandboxed
# eBPF program into the Linux kernel (root privileges are required). The eBPF
# program attaches to two statically defined tracepoints. The tracepoint
# 'net:inbound_message' is called when a new P2P message is received, and
# 'net:outbound_message' is called on outbound P2P messages. The eBPF program
# submits the P2P messages to this script via a BPF ring buffer. The submitted
# messages are printed.

# eBPF Limitations:
#
# Bitcoin P2P messages can be larger than 32kb (e.g. tx, block, ...). The eBPF
# VM's stack is limited to 512 bytes, and we can't allocate more than about 32kb
# for a P2P message in the eBPF VM. The message data is cut off when the message
# is larger than MAX_MSG_DATA_LENGTH (see definition below). This can be detected
# in user-space by comparing the data length to the message length variable. The
# message is cut off when the data length is smaller than the message length.
# A warning is included with the printed message data.
#
# Data is submitted to user-space (i.e. to this script) via a ring buffer. The
# throughput of the ring buffer is limited. Each p2p_message is about 32kb in
# size. In- or outbound messages submitted to the ring buffer in rapid
# succession fill the ring buffer faster than it can be read. Some messages are
# lost.
#
# BCC prints: "Possibly lost 2 samples" on lost messages.

import sys
from bcc import BPF, USDT

# BCC: The C program to be compiled to an eBPF program (by BCC) and loaded into
# a sandboxed Linux kernel VM.
program = """
#include <uapi/linux/ptrace.h>
#define MIN(a,b) ({ __typeof__ (a) _a = (a); __typeof__ (b) _b = (b); _a < _b ? _a : _b; })
// Maximum possible allocation size
// from include/linux/percpu.h in the Linux kernel
#define PCPU_MIN_UNIT_SIZE (32 << 10)
// Tor v3 addresses are 62 chars + 6 chars for the port (':12345').
#define MAX_PEER_ADDR_LENGTH 62 + 6
#define MAX_PEER_CONN_TYPE_LENGTH 20
#define MAX_MSG_TYPE_LENGTH 20
#define MAX_MSG_DATA_LENGTH PCPU_MIN_UNIT_SIZE - 200
struct p2p_message
{
u64 peer_id;
char peer_addr[MAX_PEER_ADDR_LENGTH];
char peer_conn_type[MAX_PEER_CONN_TYPE_LENGTH];
char msg_type[MAX_MSG_TYPE_LENGTH];
u64 msg_size;
u8 msg[MAX_MSG_DATA_LENGTH];
};
// We can't store the p2p_message struct on the eBPF stack as it is limited to
// 512 bytes and P2P message can be bigger than 512 bytes. However, we can use
// an BPF-array with a length of 1 to allocate up to 32768 bytes (this is
// defined by PCPU_MIN_UNIT_SIZE in include/linux/percpu.h in the Linux kernel).
// Also see https://github.com/iovisor/bcc/issues/2306
BPF_ARRAY(msg_arr, struct p2p_message, 1);
// Two BPF perf buffers for pushing data (here P2P messages) to user-space.
BPF_PERF_OUTPUT(inbound_messages);
BPF_PERF_OUTPUT(outbound_messages);
int trace_inbound_message(struct pt_regs *ctx) {
int idx = 0;
struct p2p_message *msg = msg_arr.lookup(&idx);
// lookup() does not return a NULL pointer. However, the BPF verifier
// requires an explicit check that that the `msg` pointer isn't a NULL
// pointer. See https://github.com/iovisor/bcc/issues/2595
if (msg == NULL) return 1;
bpf_usdt_readarg(1, ctx, &msg->peer_id);
bpf_usdt_readarg_p(2, ctx, &msg->peer_addr, MAX_PEER_ADDR_LENGTH);
bpf_usdt_readarg_p(3, ctx, &msg->peer_conn_type, MAX_PEER_CONN_TYPE_LENGTH);
bpf_usdt_readarg_p(4, ctx, &msg->msg_type, MAX_MSG_TYPE_LENGTH);
bpf_usdt_readarg(5, ctx, &msg->msg_size);
bpf_usdt_readarg_p(6, ctx, &msg->msg, MIN(msg->msg_size, MAX_MSG_DATA_LENGTH));
inbound_messages.perf_submit(ctx, msg, sizeof(*msg));
return 0;
};
int trace_outbound_message(struct pt_regs *ctx) {
int idx = 0;
struct p2p_message *msg = msg_arr.lookup(&idx);
// lookup() does not return a NULL pointer. However, the BPF verifier
// requires an explicit check that that the `msg` pointer isn't a NULL
// pointer. See https://github.com/iovisor/bcc/issues/2595
if (msg == NULL) return 1;
bpf_usdt_readarg(1, ctx, &msg->peer_id);
bpf_usdt_readarg_p(2, ctx, &msg->peer_addr, MAX_PEER_ADDR_LENGTH);
bpf_usdt_readarg_p(3, ctx, &msg->peer_conn_type, MAX_PEER_CONN_TYPE_LENGTH);
bpf_usdt_readarg_p(4, ctx, &msg->msg_type, MAX_MSG_TYPE_LENGTH);
bpf_usdt_readarg(5, ctx, &msg->msg_size);
bpf_usdt_readarg_p(6, ctx, &msg->msg, MIN(msg->msg_size, MAX_MSG_DATA_LENGTH));
outbound_messages.perf_submit(ctx, msg, sizeof(*msg));
return 0;
};
"""


def print_message(event, inbound):
print(f"%s %s msg '%s' from peer %d (%s, %s) with %d bytes: %s" %
(
f"Warning: incomplete message (only %d out of %d bytes)!" % (
len(event.msg), event.msg_size) if len(event.msg) < event.msg_size else "",
"inbound" if inbound else "outbound",
event.msg_type.decode("utf-8"),
event.peer_id,
event.peer_conn_type.decode("utf-8"),
event.peer_addr.decode("utf-8"),
event.msg_size,
bytes(event.msg[:event.msg_size]).hex(),
)
)


def main(bitcoind_path):
bitcoind_with_usdts = USDT(path=str(bitcoind_path))

# attaching the trace functions defined in the BPF program to the tracepoints
bitcoind_with_usdts.enable_probe(
probe="inbound_message", fn_name="trace_inbound_message")
bitcoind_with_usdts.enable_probe(
probe="outbound_message", fn_name="trace_outbound_message")
bpf = BPF(text=program, usdt_contexts=[bitcoind_with_usdts])

# BCC: perf buffer handle function for inbound_messages
def handle_inbound(_, data, size):
""" Inbound message handler.
Called each time a message is submitted to the inbound_messages BPF table."""

event = bpf["inbound_messages"].event(data)
print_message(event, True)

# BCC: perf buffer handle function for outbound_messages

def handle_outbound(_, data, size):
""" Outbound message handler.
Called each time a message is submitted to the outbound_messages BPF table."""

event = bpf["outbound_messages"].event(data)
print_message(event, False)

# BCC: add handlers to the inbound and outbound perf buffers
bpf["inbound_messages"].open_perf_buffer(handle_inbound)
bpf["outbound_messages"].open_perf_buffer(handle_outbound)

print("Logging raw P2P messages.")
print("Messages larger that about 32kb will be cut off!")
print("Some messages might be lost!")
while True:
try:
bpf.perf_buffer_poll()
except KeyboardInterrupt:
exit()


if __name__ == "__main__":
if len(sys.argv) < 2:
print("USAGE:", sys.argv[0], "path/to/bitcoind")
exit()
path = sys.argv[1]
main(path)
Loading

0 comments on commit 4224dec

Please sign in to comment.