c - 如何从 PROC 获取有关子进程的信息
问题描述
我正在尝试编写一个将几个进程作为参数的程序。然后父进程执行每个子进程并打印出一些关于它们的统计信息。
示例:/generate ls -l //将导致程序打印出一些关于 ls -l 的统计信息(特别是它的系统时间、用户时间和上下文切换次数)。
我不想使用 getrusage() 函数,而是想从 Proc 文件系统中获取必要的信息。现在我的理解是,如果我要使用 wait() 函数,它最终会从我的 proc 文件系统中删除信息。我在下面包含了我的代码
#include <time.h>
#include <stdbool.h>
#include <assert.h>
#include <errno.h>
#include <string.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/resource.h>
void inputted_command(int a, char **b){
for(int i=1;i<a;i++)
printf("%s ",b[i]);
}
int main(int argc, char **argv){
int status;
pid_t childpid;
pid_t get_information;
if (argc < 2)
{
return 1;
}
bool handle_signals = (signal(SIGINT, SIG_IGN) != SIG_IGN);
clock_t t;
t= clock();
pid_t pid = fork();
if(pid<0)
{
printf("fork: error no = %s\n",strerror(errno));
return 1;
}
else if(pid>0){
signal(SIGINT,SIG_IGN);
sleep(60);
/*
get_information=fork();
if(get_information==0){
execlp(___);
}else
waitpid(pid, &status, 0);
*/
waitpid(childpid, &status, 0);
t= clock()-t;
double real_time_taken = ((double)t)/CLOCKS_PER_SEC;
printf("The command ");
inputted_command(argc,argv);
if(WIFSIGNALED(status)){
printf("is interrupted by the signal number = %d (Insert Name Here) real: %.2f, user: , system: , context switch: \n",WTERMSIG(status),real_time_taken);
}
else{
printf("terminated with return status code = %d real: %.2f, user: , system: , context switch: \n",WEXITSTATUS(status), real_time_taken);
}
}
else if(pid==0){
childpid=getpid();
printf("Process with id: %d created for the command: ",(int)getpid());
inputted_command(argc,argv);
printf("\n");
assert(pid == 0);
if (handle_signals)
signal(SIGINT, SIG_DFL);
execvp(argv[1], &argv[1]);
printf(" experienced an error in starting the command: ");
inputted_command(argc,argv);
printf("\n");
exit(-1);
}
}
- 我的部分代码已被评论,我不确定如何去做。
- 我的想法是首先让父进程进入睡眠状态,以便子进程完成终止。
- 然后父进程创建一个新的子进程来访问PROC/Fie系统并获取必要的数据(已注释)。
- 最后,我再次调用等待函数并终止初始子进程
所以我的主要问题是,这是否是获取子进程信息的合适方式以及如何获取信息(主要是系统时间、用户时间以及自愿和非自愿上下文切换?
解决方案
在Craig Estey的回答的推动下,在我的评论之后,并使用来自 POSIX for 的信息sigaction()
,它指向Signal Actions
和<signal.h>
,我想出了以下代码,它使用信号SA_SIGINFO
处理SIGCHLD
,允许程序收集/proc
在子进程终止之后但在等待之前,来自文件系统的信息。
siginfo47.c
#define _XOPEN_SOURCE 700
#include "stderr.h"
#include <assert.h>
#include <errno.h>
#include <fcntl.h>
#include <inttypes.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/wait.h>
#include <unistd.h>
static int got_signal = 0;
static siginfo_t child_info = { 0 };
static void sigchld(int signum, siginfo_t *info, void *ctxt)
{
assert(info != 0);
assert(ctxt != 0);
assert(signum == info->si_signo);
got_signal = signum;
child_info = *info;
}
struct si_code_names
{
int si_code;
char si_code_name[16];
char si_code_meaning[64];
};
static struct si_code_names si_codes[] =
{
[CLD_EXITED] = { CLD_EXITED, "CLD_EXITED", "Child has exited." },
[CLD_KILLED] = { CLD_KILLED, "CLD_KILLED", "Child has terminated abnormally and did not create a core file." },
[CLD_DUMPED] = { CLD_DUMPED, "CLD_DUMPED", "Child has terminated abnormally and created a core file." },
[CLD_TRAPPED] = { CLD_TRAPPED, "CLD_TRAPPED", "Traced child has trapped." },
[CLD_STOPPED] = { CLD_STOPPED, "CLD_STOPPED", "Child has stopped." },
[CLD_CONTINUED] = { CLD_CONTINUED, "CLD_CONTINUED", "Stopped child has continued." },
};
static void cat_proc_file(int pid, const char *base)
{
char buffer[1024];
int rc;
rc = snprintf(buffer, sizeof(buffer), "/proc/%d/%s", pid, base);
if (rc < 0 || rc >= (int)sizeof(buffer))
err_error("snprintf() failed - can't happen!?!\n");
int fd = open(buffer, O_RDONLY);
if (fd < 0)
err_syserr("failed to open file '%s' for reading: ", buffer);
printf("Contents of %s:\n", buffer);
int nbytes;
while ((nbytes = read(fd, buffer, sizeof(buffer))) > 0)
printf("%*.*s", nbytes, nbytes, buffer);
putchar('\n');
fflush(stdout);
close(fd);
}
int main(int argc, char **argv)
{
char *cmdv[] = { "ls", "-l", 0 };
err_setarg0(argv[0]);
if (argc <= 1)
{
argc = 2;
argv = cmdv;
}
else
{
argc--;
argv++;
}
pid_t pid = fork();
if (pid < 0)
err_syserr("failed to fork: ");
else if (pid > 0)
{
struct sigaction sa = { 0 };
sa.sa_sigaction = sigchld;
sa.sa_flags = SA_SIGINFO;
if (sigaction(SIGCHLD, &sa, 0) != 0)
err_syserr("failed to set signal handling: ");
printf("Parent PID %d: pausing while PID %d runs\n", (int)getpid(), (int)pid);
fflush(stdout);
pause();
printf("Parent PID %d: unpaused\n", (int)getpid());
printf("Stashed information:\n");
printf(" Signal: %d\n", got_signal);
printf(" si_signo: %d\n", child_info.si_signo);
printf(" si_code: %d\n", child_info.si_code);
if (child_info.si_signo == SIGCHLD)
{
struct si_code_names *code = &si_codes[child_info.si_code];
printf(" [%s] %s\n", code->si_code_name,
code->si_code_meaning);
}
printf(" si_pid: %d\n", (int)child_info.si_pid);
printf(" si_uid: %d\n", (int)child_info.si_uid);
printf(" si_addr: 0x%12" PRIXPTR "\n", (uintptr_t)child_info.si_addr);
printf(" si_status: %d\n", child_info.si_code);
printf(" si_value.int: %d\n", child_info.si_value.sival_int);
cat_proc_file(pid, "stat");
cat_proc_file(pid, "status");
int status;
int corpse;
if ((corpse = waitpid(pid, &status, 0)) != pid)
err_syserr("failed to wait for child %d", pid);
if (WIFSIGNALED(status))
printf("PID %d died from signal number = %d (0x%.4X)\n",
corpse, WTERMSIG(status), status);
else if (WIFEXITED(status))
printf("PID %d exited normally with status = %d (0x%.4X)\n",
corpse, WEXITSTATUS(status), status);
else
printf("PID %d was neither signalled nor exited normally (0x%.4X)\n",
corpse, status);
}
else if (pid == 0)
{
printf("PID: %d:", (int)getpid());
for (int i = 0; argv[i] != 0; i++)
printf(" %s", argv[i]);
putchar('\n');
fflush(stdout);
execvp(argv[0], &argv[0]);
err_syserr("failed to execute %s: ", argv[0]);
/*NOTREACHED*/
}
}
其中一些代码可在我的 GitHub 上的SOQ (堆栈溢出问题)存储库中找到。具体来说,这些文件stderr.c
可以stderr.h
在src/libsoq子目录中找到。它们极大地简化了错误报告。
示例运行包括:
$ siginfo47
Parent PID 15016: pausing while PID 15017 runs
PID: 15017: ls -l
total 400
drwxr-xr-x 2 jleffler pd 4096 Oct 21 15:16 bin
drwxr-xr-x 5 jleffler pd 256 Oct 21 15:15 doc
drwxr-xr-x 2 jleffler pd 4096 Oct 21 15:15 etc
drwxr-xr-x 2 jleffler pd 4096 Oct 21 15:16 inc
drwxr-xr-x 2 jleffler pd 256 Oct 21 15:16 lib
-rw-r--r-- 1 jleffler pd 22072 Oct 21 15:15 LICENSE.md
-rw-r--r-- 1 jleffler pd 390 Oct 21 15:15 makefile
drwxr-xr-x 2 jleffler pd 256 Oct 21 15:15 packages
-rw-r--r-- 1 jleffler pd 2694 Oct 21 15:15 README.md
-rwxr-xr-x 1 jleffler pd 64968 Oct 21 15:17 siginfo41
-rw-r--r-- 1 jleffler pd 5990 Oct 21 15:17 siginfo41.c
-rwxr-xr-x 1 jleffler pd 66104 Oct 21 15:34 siginfo47
-rw-r--r-- 1 jleffler pd 7417 Oct 21 15:33 siginfo47.c
drwxr-xr-x 230 jleffler pd 8192 Oct 21 15:15 src
Parent PID 15016: unpaused
Stashed information:
Signal: 17
si_signo: 17
si_code: 1
[CLD_EXITED] Child has exited.
si_pid: 15017
si_uid: 9508
si_addr: 0x252400003AA9
si_status: 1
si_value.int: 0
Contents of /proc/15017/stat:
15017 (ls) Z 15016 15016 13211 34827 15016 4227084 452 0 0 0 0 0 0 0 20 0 1 0 511347844 0 0 18446744073709551615 0 0 0 0 0 0 0 0 0 18446744073709551615 0 0 17 6 0 0 0 0 0 0 0 0 0 0 0 0 0
Contents of /proc/15017/status:
Name: ls
State: Z (zombie)
Tgid: 15017
Ngid: 0
Pid: 15017
PPid: 15016
TracerPid: 0
Uid: 9508 9508 9508 9508
Gid: 1240 1240 1240 1240
FDSize: 0
Groups: 297 1240 1360 8714
Threads: 1
SigQ: 0/71487
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: 0000000000000000
SigCgt: 0000000180000000
CapInh: 0000000000000000
CapPrm: 0000000000000000
CapEff: 0000000000000000
CapBnd: 0000001fffffffff
CapAmb: 0000000000000000
Seccomp: 0
Cpus_allowed: ff
Cpus_allowed_list: 0-7
Mems_allowed: 00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 3
nonvoluntary_ctxt_switches: 1
PID 15017 exited normally with status = 0 (0x0000)
$ siginfo47 exitcode 23
Parent PID 15032: pausing while PID 15033 runs
PID: 15033: exitcode 23
Parent PID 15032: unpaused
Stashed information:
Signal: 17
si_signo: 17
si_code: 1
[CLD_EXITED] Child has exited.
si_pid: 15033
si_uid: 9508
si_addr: 0x252400003AB9
si_status: 1
si_value.int: 23
Contents of /proc/15033/stat:
15033 (exitcode) Z 15032 15032 13211 34827 15032 4227084 179 0 0 0 0 0 0 0 20 0 1 0 511349111 0 0 18446744073709551615 0 0 0 0 0 0 0 0 0 18446744073709551615 0 0 17 5 0 0 0 0 0 0 0 0 0 0 0 0 0
Contents of /proc/15033/status:
Name: exitcode
State: Z (zombie)
Tgid: 15033
Ngid: 0
Pid: 15033
PPid: 15032
TracerPid: 0
Uid: 9508 9508 9508 9508
Gid: 1240 1240 1240 1240
FDSize: 0
Groups: 297 1240 1360 8714
Threads: 1
SigQ: 0/71487
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: 0000000000000000
SigCgt: 0000000000000000
CapInh: 0000000000000000
CapPrm: 0000000000000000
CapEff: 0000000000000000
CapBnd: 0000001fffffffff
CapAmb: 0000000000000000
Seccomp: 0
Cpus_allowed: ff
Cpus_allowed_list: 0-7
Mems_allowed: 00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 3
nonvoluntary_ctxt_switches: 1
PID 15033 exited normally with status = 23 (0x1700)
$ siginfo47 exitcode -s 13
PID: 15057: exitcode -s 13
Parent PID 15056: pausing while PID 15057 runs
Parent PID 15056: unpaused
Stashed information:
Signal: 17
si_signo: 17
si_code: 2
[CLD_KILLED] Child has terminated abnormally and did not create a core file.
si_pid: 15057
si_uid: 9508
si_addr: 0x252400003AD1
si_status: 2
si_value.int: 13
Contents of /proc/15057/stat:
15057 (exitcode) Z 15056 15056 13211 34827 15056 4228108 177 0 0 0 0 0 0 0 20 0 1 0 511350462 0 0 18446744073709551615 0 0 0 0 0 4096 0 0 0 18446744073709551615 0 0 17 5 0 0 0 0 0 0 0 0 0 0 0 0 0
Contents of /proc/15057/status:
Name: exitcode
State: Z (zombie)
Tgid: 15057
Ngid: 0
Pid: 15057
PPid: 15056
TracerPid: 0
Uid: 9508 9508 9508 9508
Gid: 1240 1240 1240 1240
FDSize: 0
Groups: 297 1240 1360 8714
Threads: 1
SigQ: 1/71487
SigPnd: 0000000000001000
ShdPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: 0000000000000000
SigCgt: 0000000000000000
CapInh: 0000000000000000
CapPrm: 0000000000000000
CapEff: 0000000000000000
CapBnd: 0000001fffffffff
CapAmb: 0000000000000000
Seccomp: 0
Cpus_allowed: ff
Cpus_allowed_list: 0-7
Mems_allowed: 00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 2
nonvoluntary_ctxt_switches: 1
PID 15057 died from signal number = 13 (0x000D)
$ exitcode -h
Usage: exitcode [-hV] [-s signal] [exit-status]
-h Print this help message and exit
-s signal Kill self with signal number
-V Print version information and exit
$
正如帮助消息所指出的,exitcode
程序以退出状态终止,通常是 ( exitcode 23
) 或作为信号 ( exitcode -s 13
) 的结果。
推荐阅读
- linux - Puppet - 如何基于 Role/Profile 方法编写 yaml 文件
- android - 在 Kotlin 中,如何在创建时在一个表达式中设置对象成员?
- javascript - 单击添加/删除类并一次添加/删除一组 DIV 的值
- windows - Windows 证书存储问题仍然使用 wincrypt 列出已删除的证书
- windows - 在 PowerShell 命令提示符中显示正确的用户名
- sql - 获取行格式的数据
- javascript - AJAX 错误地获取和发布多个数据
- c++ - 通过引用返回并复制值
- amazon-web-services - 从 EC2 安全组获取被阻止的 IP 而不打开 0.0.0.0
- mongodb - Quarkus Panache MongoDB:将 ObjectId 序列化为字符串