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
|
// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
// Copyright (c) 2019 Facebook
#include <argp.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/resource.h>
#include <time.h>
#include <bpf/libbpf.h>
#include <bpf/bpf.h>
#include "runqslower.h"
#include "runqslower.skel.h"
struct env {
pid_t pid;
__u64 min_us;
bool verbose;
} env = {
.min_us = 10000,
};
const char *argp_program_version = "runqslower 0.1";
const char *argp_program_bug_address = "<bpf@vger.kernel.org>";
const char argp_program_doc[] =
"runqslower Trace long process scheduling delays.\n"
" For Linux, uses eBPF, BPF CO-RE, libbpf, BTF.\n"
"\n"
"This script traces high scheduling delays between tasks being\n"
"ready to run and them running on CPU after that.\n"
"\n"
"USAGE: runqslower [-p PID] [min_us]\n"
"\n"
"EXAMPLES:\n"
" runqslower # trace run queue latency higher than 10000 us (default)\n"
" runqslower 1000 # trace run queue latency higher than 1000 us\n"
" runqslower -p 123 # trace pid 123 only\n";
static const struct argp_option opts[] = {
{ "pid", 'p', "PID", 0, "Process PID to trace"},
{ "verbose", 'v', NULL, 0, "Verbose debug output" },
{},
};
static error_t parse_arg(int key, char *arg, struct argp_state *state)
{
static int pos_args;
int pid;
long long min_us;
switch (key) {
case 'v':
env.verbose = true;
break;
case 'p':
errno = 0;
pid = strtol(arg, NULL, 10);
if (errno || pid <= 0) {
fprintf(stderr, "Invalid PID: %s\n", arg);
argp_usage(state);
}
env.pid = pid;
break;
case ARGP_KEY_ARG:
if (pos_args++) {
fprintf(stderr,
"Unrecognized positional argument: %s\n", arg);
argp_usage(state);
}
errno = 0;
min_us = strtoll(arg, NULL, 10);
if (errno || min_us <= 0) {
fprintf(stderr, "Invalid delay (in us): %s\n", arg);
argp_usage(state);
}
env.min_us = min_us;
break;
default:
return ARGP_ERR_UNKNOWN;
}
return 0;
}
int libbpf_print_fn(enum libbpf_print_level level,
const char *format, va_list args)
{
if (level == LIBBPF_DEBUG && !env.verbose)
return 0;
return vfprintf(stderr, format, args);
}
static int bump_memlock_rlimit(void)
{
struct rlimit rlim_new = {
.rlim_cur = RLIM_INFINITY,
.rlim_max = RLIM_INFINITY,
};
return setrlimit(RLIMIT_MEMLOCK, &rlim_new);
}
void handle_event(void *ctx, int cpu, void *data, __u32 data_sz)
{
const struct event *e = data;
struct tm *tm;
char ts[32];
time_t t;
time(&t);
tm = localtime(&t);
strftime(ts, sizeof(ts), "%H:%M:%S", tm);
printf("%-8s %-16s %-6d %14llu\n", ts, e->task, e->pid, e->delta_us);
}
void handle_lost_events(void *ctx, int cpu, __u64 lost_cnt)
{
printf("Lost %llu events on CPU #%d!\n", lost_cnt, cpu);
}
int main(int argc, char **argv)
{
static const struct argp argp = {
.options = opts,
.parser = parse_arg,
.doc = argp_program_doc,
};
struct perf_buffer_opts pb_opts;
struct perf_buffer *pb = NULL;
struct runqslower_bpf *obj;
int err;
err = argp_parse(&argp, argc, argv, 0, NULL, NULL);
if (err)
return err;
libbpf_set_print(libbpf_print_fn);
err = bump_memlock_rlimit();
if (err) {
fprintf(stderr, "failed to increase rlimit: %d", err);
return 1;
}
obj = runqslower_bpf__open();
if (!obj) {
fprintf(stderr, "failed to open and/or load BPF object\n");
return 1;
}
/* initialize global data (filtering options) */
obj->rodata->targ_pid = env.pid;
obj->rodata->min_us = env.min_us;
err = runqslower_bpf__load(obj);
if (err) {
fprintf(stderr, "failed to load BPF object: %d\n", err);
goto cleanup;
}
err = runqslower_bpf__attach(obj);
if (err) {
fprintf(stderr, "failed to attach BPF programs\n");
goto cleanup;
}
printf("Tracing run queue latency higher than %llu us\n", env.min_us);
printf("%-8s %-16s %-6s %14s\n", "TIME", "COMM", "PID", "LAT(us)");
pb_opts.sample_cb = handle_event;
pb_opts.lost_cb = handle_lost_events;
pb = perf_buffer__new(bpf_map__fd(obj->maps.events), 64, &pb_opts);
err = libbpf_get_error(pb);
if (err) {
pb = NULL;
fprintf(stderr, "failed to open perf buffer: %d\n", err);
goto cleanup;
}
while ((err = perf_buffer__poll(pb, 100)) >= 0)
;
printf("Error polling perf buffer: %d\n", err);
cleanup:
perf_buffer__free(pb);
runqslower_bpf__destroy(obj);
return err != 0;
}
|
