-
Notifications
You must be signed in to change notification settings - Fork 3.8k
/
offcputime.py
executable file
·196 lines (171 loc) · 5.61 KB
/
offcputime.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
#!/usr/bin/python
#
# offcputime Summarize off-CPU time by kernel stack trace
# For Linux, uses BCC, eBPF.
#
# USAGE: offcputime [-h] [-u] [-p PID] [-v] [-f] [duration]
#
# The current implementation uses an unrolled loop for x86_64, and was written
# as a proof of concept. This implementation should be replaced in the future
# with an appropriate bpf_ call, when available.
#
# Copyright 2016 Netflix, Inc.
# Licensed under the Apache License, Version 2.0 (the "License")
#
# 13-Jan-2016 Brendan Gregg Created this.
from __future__ import print_function
from bcc import BPF
from sys import stderr
from time import sleep, strftime
import argparse
import signal
# arg validation
def positive_int(val):
try:
ival = int(val)
except ValueError:
raise argparse.ArgumentTypeError("must be an integer")
if ival < 0:
raise argparse.ArgumentTypeError("must be positive")
return ival
def positive_nonzero_int(val):
ival = positive_int(val)
if ival == 0:
raise argparse.ArgumentTypeError("must be nonzero")
return ival
# arguments
examples = """examples:
./offcputime # trace off-CPU stack time until Ctrl-C
./offcputime 5 # trace for 5 seconds only
./offcputime -f 5 # 5 seconds, and output in folded format
./offcputime -u # don't include kernel threads (user only)
./offcputime -p 185 # trace for PID 185 only
"""
parser = argparse.ArgumentParser(
description="Summarize off-CPU time by kernel stack trace",
formatter_class=argparse.RawDescriptionHelpFormatter,
epilog=examples)
thread_group = parser.add_mutually_exclusive_group()
thread_group.add_argument("-u", "--useronly", action="store_true",
help="user threads only (no kernel threads)")
thread_group.add_argument("-p", "--pid", type=positive_int,
help="trace this PID only")
parser.add_argument("-v", "--verbose", action="store_true",
help="show raw addresses")
parser.add_argument("-f", "--folded", action="store_true",
help="output folded format")
parser.add_argument("--stack-storage-size", default=1024,
type=positive_nonzero_int,
help="the number of unique stack traces that can be stored and " \
"displayed (default 1024)")
parser.add_argument("duration", nargs="?", default=99999999,
type=positive_nonzero_int,
help="duration of trace, in seconds")
args = parser.parse_args()
folded = args.folded
duration = int(args.duration)
# signal handler
def signal_ignore(signal, frame):
print()
# define BPF program
bpf_text = """
#include <uapi/linux/ptrace.h>
#include <linux/sched.h>
#define MINBLOCK_US 1
struct key_t {
char name[TASK_COMM_LEN];
int stack_id;
};
BPF_HASH(counts, struct key_t);
BPF_HASH(start, u32);
BPF_STACK_TRACE(stack_traces, STACK_STORAGE_SIZE)
int oncpu(struct pt_regs *ctx, struct task_struct *prev) {
u32 pid;
u64 ts, *tsp;
// record previous thread sleep time
if (FILTER) {
pid = prev->pid;
ts = bpf_ktime_get_ns();
start.update(&pid, &ts);
}
// calculate current thread's delta time
pid = bpf_get_current_pid_tgid();
tsp = start.lookup(&pid);
if (tsp == 0)
return 0; // missed start or filtered
u64 delta = bpf_ktime_get_ns() - *tsp;
start.delete(&pid);
delta = delta / 1000;
if (delta < MINBLOCK_US)
return 0;
// create map key
u64 zero = 0, *val;
struct key_t key = {};
bpf_get_current_comm(&key.name, sizeof(key.name));
key.stack_id = stack_traces.get_stackid(ctx, BPF_F_REUSE_STACKID);
val = counts.lookup_or_init(&key, &zero);
(*val) += delta;
return 0;
}
"""
# set thread filter
if args.pid is not None:
filter = 'pid == %s' % args.pid
elif args.useronly:
filter = '!(prev->flags & PF_KTHREAD)'
else:
filter = '1'
bpf_text = bpf_text.replace('FILTER', filter)
# set stack storage size
bpf_text = bpf_text.replace('STACK_STORAGE_SIZE', str(args.stack_storage_size))
# initialize BPF
b = BPF(text=bpf_text)
b.attach_kprobe(event="finish_task_switch", fn_name="oncpu")
matched = b.num_open_kprobes()
if matched == 0:
print("error: 0 functions traced. Exiting.", file=stderr)
exit(1)
# header
if not folded:
print("Tracing off-CPU time (us) by kernel stack", end="")
if duration < 99999999:
print(" for %d secs." % duration)
else:
print("... Hit Ctrl-C to end.")
try:
sleep(duration)
except KeyboardInterrupt:
# as cleanup can take many seconds, trap Ctrl-C:
signal.signal(signal.SIGINT, signal_ignore)
if not folded:
print()
missing_stacks = 0
has_enomem = False
counts = b.get_table("counts")
stack_traces = b.get_table("stack_traces")
for k, v in sorted(counts.items(), key=lambda counts: counts[1].value):
# handle get_stackid erorrs
if k.stack_id < 0:
missing_stacks += 1
# check for an ENOMEM error
if k.stack_id == -12:
has_enomem = True
continue
stack = stack_traces.walk(k.stack_id)
if folded:
# print folded stack output
stack = list(stack)[1:]
line = [k.name.decode()] + [b.ksym(addr) for addr in reversed(stack)]
print("%s %d" % (";".join(line), v.value))
else:
# print default multi-line stack output
for addr in stack:
print(" %-16x %s" % (addr, b.ksym(addr)))
print(" %-16s %s" % ("-", k.name))
print(" %d\n" % v.value)
if missing_stacks > 0:
enomem_str = "" if not has_enomem else \
" Consider increasing --stack-storage-size."
print("WARNING: %d stack traces could not be displayed.%s" %
(missing_stacks, enomem_str),
file=stderr)