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
|
// Copyright 2017 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include <dirent.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/time.h>
#include <sys/timerfd.h>
#include <sys/types.h>
#include <limits>
#include <memory>
#include "file_utils.h"
#include "logging.h"
#include "process_info.h"
namespace {
using ProcessMap = std::map<int, std::unique_ptr<ProcessInfo>>;
int g_timer;
std::unique_ptr<ProcessMap> CollectStatsForAllProcs(bool full_mem_stats) {
std::unique_ptr<ProcessMap> procs(new ProcessMap());
file_utils::ForEachPidInProcPath("/proc", [&procs, full_mem_stats](int pid) {
if (!ProcessInfo::IsProcess(pid))
return;
CHECK(procs->count(pid) == 0);
std::unique_ptr<ProcessInfo> pinfo(new ProcessInfo(pid));
if (!(pinfo->ReadProcessName() && pinfo->ReadThreadNames() &&
pinfo->ReadOOMStats() && pinfo->ReadPageFaultsAndCPUTimeStats()))
return;
if (full_mem_stats) {
if (!pinfo->memory()->ReadFullStats())
return;
} else {
if (!pinfo->memory()->ReadLightStats())
return;
}
(*procs)[pid] = std::move(pinfo);
});
return procs;
}
void SerializeSnapshot(const ProcessMap& procs,
FILE* stream,
bool full_mem_stats) {
struct timespec ts = {};
CHECK(clock_gettime(CLOCK_MONOTONIC_COARSE, &ts) == 0);
fprintf(stream, "{\n");
fprintf(stream, " \"ts\": %lu,\n",
(ts.tv_sec * 1000 + ts.tv_nsec / 1000000ul));
fprintf(stream, " \"processes\": [\n");
for (auto it = procs.begin(); it != procs.end();) {
int pid = it->first;
const ProcessInfo& pinfo = *it->second;
fprintf(stream, " {\"pid\": %d, \"name\": \"%s\", \"exe\": \"%s\"", pid,
pinfo.name(), pinfo.exe());
fprintf(stream, ", \"threads\": [");
for (auto t = pinfo.threads()->begin(); t != pinfo.threads()->end();) {
fprintf(stream, "{\"tid\": %d, \"name\":\"%s\"", t->first,
t->second->name);
t++;
fprintf(stream, t != pinfo.threads()->end() ? "}, " : "}");
}
fprintf(stream, "]");
const ProcessMemoryStats* mem_info = pinfo.memory();
fprintf(stream, ", \"mem\": {\"vm\": %llu, \"rss\": %llu",
mem_info->virt_kb(), mem_info->rss_kb());
if (full_mem_stats) {
fprintf(stream,
", \"pss\": %llu, \"swp\": %llu, \"pc\": %llu, \"pd\": %llu, "
"\"sc\": %llu, \"sd\": %llu",
mem_info->rss_kb(), mem_info->swapped_kb(),
mem_info->private_clean_kb(), mem_info->private_dirty_kb(),
mem_info->shared_clean_kb(), mem_info->shared_dirty_kb());
}
fprintf(stream, "}");
fprintf(stream,
", \"oom\": {\"adj\": %d, \"score_adj\": %d, \"score\": %d}",
pinfo.oom_adj(), pinfo.oom_score_adj(), pinfo.oom_score());
fprintf(stream,
", \"stat\": {\"minflt\": %lu, \"majflt\": %lu, "
"\"utime\": %lu, \"stime\": %lu }",
pinfo.minflt(), pinfo.majflt(), pinfo.utime(), pinfo.stime());
fprintf(stream, "}");
it++;
fprintf(stream, it != procs.end() ? ",\n" : "\n");
}
fprintf(stream, " ]\n");
fprintf(stream, "}\n");
}
} // namespace
int main(int argc, char** argv) {
bool background = false;
int dump_interval_ms = 5000;
char out_file[PATH_MAX] = {};
bool dump_to_file = false;
bool full_mem_stats = false;
int count = std::numeric_limits<int>::max();
int opt;
while ((opt = getopt(argc, argv, "bmt:o:c:")) != -1) {
switch (opt) {
case 'b':
background = true;
break;
case 'm':
full_mem_stats = true;
break;
case 't':
dump_interval_ms = atoi(optarg);
CHECK(dump_interval_ms > 0);
break;
case 'c':
count = atoi(optarg);
CHECK(count > 0);
break;
case 'o':
strncpy(out_file, optarg, sizeof(out_file));
dump_to_file = true;
break;
default:
fprintf(stderr,
"Usage: %s [-b] [-t dump_interval_ms] [-c dumps_count] "
"[-o out.json]\n",
argv[0]);
exit(EXIT_FAILURE);
}
}
if (geteuid()) {
fprintf(stderr, "Must run as root\n");
exit(EXIT_FAILURE);
}
FILE* out_stream = stdout;
char tmp_file[PATH_MAX];
if (dump_to_file) {
unlink(out_file);
sprintf(tmp_file, "%s.tmp", out_file);
out_stream = fopen(tmp_file, "w");
CHECK(out_stream);
}
if (background) {
if (!dump_to_file) {
fprintf(stderr, "-b requires -o for output dump path\n");
exit(EXIT_FAILURE);
}
printf("Continuing in background. kill -TERM to terminate the daemon.\n");
CHECK(daemon(0 /* nochdir */, 0 /* noclose */) == 0);
}
g_timer = timerfd_create(CLOCK_MONOTONIC, 0);
CHECK(g_timer >= 0);
struct itimerspec ts = {};
ts.it_value.tv_nsec = 1; // Get the first snapshot immediately.
ts.it_interval.tv_nsec = (dump_interval_ms % 1000) * 1000000ul;
ts.it_interval.tv_sec = dump_interval_ms / 1000;
CHECK(timerfd_settime(g_timer, 0, &ts, nullptr) == 0);
// Closing the g_timer fd on SIGINT/SIGTERM will cause the read() below to
// unblock and fail with EBADF, hence allowing the loop below to finalize
// the file and exit.
auto on_exit = [](int) { close(g_timer); };
signal(SIGINT, on_exit);
signal(SIGTERM, on_exit);
fprintf(out_stream, "{\"snapshots\": [\n");
bool is_first_snapshot = true;
for (; count > 0; count--) {
uint64_t missed = 0;
int res = read(g_timer, &missed, sizeof(missed));
if (res < 0 && errno == EBADF)
break; // Received SIGINT/SIGTERM signal.
CHECK(res > 0);
if (!is_first_snapshot)
fprintf(out_stream, ",");
is_first_snapshot = false;
std::unique_ptr<ProcessMap> procs = CollectStatsForAllProcs(full_mem_stats);
SerializeSnapshot(*procs, out_stream, full_mem_stats);
fflush(out_stream);
}
fprintf(out_stream, "]}\n");
fclose(out_stream);
if (dump_to_file)
rename(tmp_file, out_file);
}
|
