如何在Linux上用clone()创build一个真正的线程?

我正在尝试使用clone()创build一个新的线程。 用下面的代码(…):

#include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _SCHED_H 1 #define __USE_GNU 1 #include <bits/sched.h> #define STACK_SIZE 4096 int func(void *arg) { printf("Inside func.\n"); sleep(1); printf("Terminating func...\n"); return 0; } int main() { printf("This process pid: %u\n", getpid()); char status_file[] = "/proc/self/status"; void *child_stack = malloc(STACK_SIZE); int thread_pid; printf("Creating new thread...\n"); thread_pid = clone(&func, child_stack+STACK_SIZE, CLONE_SIGHAND|CLONE_FS|CLONE_VM|CLONE_FILES, NULL); printf("Done! Thread pid: %d\n", thread_pid); FILE *fp = fopen(status_file, "rb"); printf("Looking into %s...\n", status_file); while(1) { char ch = fgetc(fp); if(feof(fp)) break; printf("%c", ch); } fclose(fp); getchar(); return 0; } 

我得到以下内容:

 This process pid: 10839 Creating new thread... Done! Thread pid: 10840 Inside func. Looking into /proc/self/status... Name: threadTest02 State: R (running) Tgid: 10839 Pid: 10839 PPid: 4777 TracerPid: 0 Uid: 1000 1000 1000 1000 Gid: 1000 1000 1000 1000 FDSize: 256 Groups: 4 20 24 27 30 46 107 123 124 1000 VmPeak: 4300 kB VmSize: 4300 kB VmLck: 0 kB VmPin: 0 kB VmHWM: 356 kB VmRSS: 356 kB VmData: 188 kB VmStk: 136 kB VmExe: 4 kB VmLib: 1884 kB VmPTE: 32 kB VmSwap: 0 kB Threads: 1 SigQ: 0/22869 SigPnd: 0000000000000000 ShdPnd: 0000000000000000 SigBlk: 0000000000000000 SigIgn: 0000000000000000 SigCgt: 0000000000000000 CapInh: 0000000000000000 CapPrm: 0000000000000000 CapEff: 0000000000000000 CapBnd: ffffffffffffffff Cpus_allowed: 3 Cpus_allowed_list: 0-1 Mems_allowed: 00000000,00000001 Mems_allowed_list: 0 voluntary_ctxt_switches: 1 nonvoluntary_ctxt_switches: 1 Terminating func... 

总之,我的程序是做什么的? 它创build一个clone的新线程,并打印它的/proc/self/status ,以便我可以看到它的状态。 由于我的线程hibernate1秒,当/proc/self/status被打印时它仍然活着。

然而,至less有两件事情,使我的线程不像一个共同的线程行为。 首先,如上所见,进程的PID是10839,我的线程的PID是10840.所以,进程和我的线程没有相同的PID,就像普通线程中发生的那样。 其次,即使在创buildThreads: ,我的进程“ /proc/self/status文件中的Threads:字段仍为1.所以,我的线程似乎不被识别为线程。

我的问题是,我的代码中缺less什么? 我需要做些什么来使我的线程像一个共同的线程行为? clone的第三个参数中是否有任何选项丢失?

您可能希望看到CLONE_THREAD标志,它会将新线程放置在与调用进程相同的线程组中。

一旦你给了CLONE_THREAD,它将使新线程与调用进程具有相同的pid和ppid。 它用在posix线程中。 以下是我的系统输出。 LWP专栏说,现在是一个轻量级的过程,并有不同的TID

 UID PID PPID LWP C NLWP SZ RSS PSR STIME TTY TIME CMD anukalp 18398 9638 18398 0 2 464 456 0 10:56 pts/3 00:00:00 ./a.out anukalp 18398 9638 18399 0 2 464 456 1 10:56 pts/3 00:00:00 ./a.out 

/ proc / self / status的输出也改变了,我添加了几个printfs:

[anukalp @ localhost〜] $ ./a.out

 This process pid: 18398 Creating new thread... Done! Thread pid: 18399 /* This is now thread id, available to caller of clone */ getpid(): ad pid: 18399 Inside func. getpid(): 18398 getppid(): 9638 Looking into /proc/self/status... Name: a.out State: R (running) Tgid: 18398 Pid: 18398 PPid: 9638 TracerPid: 0 Uid: 500 500 500 500 Gid: 500 500 500 500 FDSize: 256 Groups: 7 19 22 80 81 82 83 100 490 500 VmPeak: 1856 kB VmSize: 1856 kB VmLck: 0 kB VmPin: 0 kB VmHWM: 248 kB VmRSS: 248 kB VmData: 168 kB VmStk: 140 kB VmExe: 4 kB VmLib: 1516 kB VmPTE: 16 kB VmSwap: 0 kB Threads: 2 SigQ: 1/14050 SigPnd: 0000000000000000 ShdPnd: 0000000000000000 SigBlk: 0000000000000000 SigIgn: 0000000000000000 SigCgt: 0000000000000000 CapInh: 0000000000000000 CapPrm: 0000000000000000 CapEff: 0000000000000000 CapBnd: ffffffffffffffff Cpus_allowed: 00000000,000000ff Cpus_allowed_list: 0-7 voluntary_ctxt_switches: 1 nonvoluntary_ctxt_switches: 0 Inside thread: thread pid = 18398 Inside thread: thread ppid = 9638 

请让我知道这可不可以帮你!