c语言thread函数,内核线程函数kernel_thread解析

设备驱动程序中，如果需要几个并发执行的任务，可以启动内核线程，启动内核线程的函数为:

int kernel_thread (int ( * fn )( void * ), void * arg, unsigned long flags);

kernel_thread函数的作用是产生一个新的线程

内核线程实际上就是一个共享父进程地址空间的进程,它有自己的系统堆栈.

内核线程和进程都是通过do_fork()函数来产生的，系统中规定的最大进程数与线程数由fork_init来决定:

[/arch/kernel/process.c/fork_init()]

void __init fork_init(unsigned long mempages)

{

#ifndef __HAVE_ARCH_TASK_STRUCT_ALLOCATOR

#ifndef ARCH_MIN_TASKALIGN

#define ARCH_MIN_TASKALIGN   L1_CACHE_BYTES

#endif

/*在slab高速缓存中建立task_struct结构专用的缓冲区队列*/

task_struct_cachep =

kmem_cache_create("task_struct", sizeof(struct task_struct),

ARCH_MIN_TASKALIGN, SLAB_PANIC, NULL, NULL);

#endif

/*

把默认线程数设置到一个安全值,因为内核中总的线程占用的空间

可能要内存一半还要多.

参数mempages系统中总的物理内存结构大小，它等于mempages/PAGESIZE.

比如我机器的内存是512m，那么在我的系统最多能同时产生线程数为

(512*2^20/2^12) / 2^3 = 512*2^5 = 16384

*/

max_threads = mempages / (8 * THREAD_SIZE / PAGE_SIZE);

/*

*启动系统的时候至少需要20个线程

*/

if(max_threads < 20)

max_threads = 20;

/*

*每个进程最多产生max_threads/2,也就是线程总数的一半,在我的机器上为8192.

*/

init_task.signal->rlim[RLIMIT_NPROC].rlim_cur = max_threads/2;

init_task.signal->rlim[RLIMIT_NPROC].rlim_max = max_threads/2;

}

kernel_thread原形在/arch/kernel/process.c中.

(*fn)(void *)为要执行的函数的指针,arg为函数参数，flags为do_fork产生线程时的标志.

int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)

{

struct pt_regs regs;

memset(&regs, 0, sizeof(regs));

regs.ebx = (unsigned long) fn;   /* ebx指向函数地址*/

regs.edx = (unsigned long) arg;   /* edx指向参数*/

regs.xds = __USER_DS;

regs.xes = __USER_DS;

regs.orig_eax = -1;

regs.eip = (unsigned long) kernel_thread_helper;

regs.xcs = __KERNEL_CS;

regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2;

/*利用do_fork来产生一个新的线程,共享父进程地址空间,并且不允许调试子进程*/

return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);

}

[/arch/i386/kernel/process.c/kernel_thread_helper]

extern void kernel_thread_helper(void); /*定义成全局变量*/

__asm__(".section .text\n"

".align 4\n"

"kernel_thread_helper:\n\t"

"movl %edx,%eax\n\t"

"pushl %edx\n\t"   /* edx指向参数,压入堆栈*/

"call *%ebx\n\t"   /* ebx指向函数地址,执行函数*/

"pushl %eax\n\t"

"call do_exit\n"   /*结束线程*/

".previous");

在kernel_thread中调用了do_fork，那么do_fork是怎样转入kernel_thread_helper去执行的呢，继续跟踪下do_fork函数.

[kernel/fork.c/do_fork()]

long do_fork(unsigned long clone_flags,

unsigned long stack_start,

struct pt_regs *regs,

unsigned long stack_size,

int __user *parent_tidptr,

int __user *child_tidptr)

{

....

....

p = copy_process(clone_flags, stack_start, regs, stack_size, parent_tidptr, child_tidptr, pid);

....

....

}

它调用copy_process函数来向子进程拷贝父进程的进程环境和全部寄存器副本.

[kernel/fork.c/do_fork()->copy_process]

static task_t *copy_process(unsigned long clone_flags,

unsigned long stack_start,

struct pt_regs *regs,

unsigned long stack_size,

int __user *parent_tidptr,

int __user *child_tidptr,

int pid)

{

...

...

retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs);

...

...

}

它又调用copy_thread来拷贝父进程的系统堆栈并做相应的调整.

[/arch/i386/kernel/process.c/copy_thread]:

int copy_thread(int nr, unsigned long clone_flags, unsigned long esp,

unsigned long unused,

struct task_struct * p, struct pt_regs * regs)

{

...

...

p->thread.eip = (unsigned long) ret_from_fork;

}

在这里把ret_from_fork的地址赋值给p->thread.eip,p->thread.eip表示当进程下一次调度时的指令开始地址，

所以当线程创建后被调度时，是从ret_from_fork地址处开始的.

[/arch/i386/kernel/entry.s]

到这里说明，新的线程已经产生了.

ENTRY(ret_from_fork)

pushl %eax

call schedule_tail

GET_THREAD_INFO(%ebp)

popl %eax

jmp syscall_exit

syscall_exit:

...

work_resched:

call schedule

...

当它从ret_from_fork退出时，会从堆栈中弹出原来保存的ip，而ip指向kernel_thread_helper,

至此kernel_thread_helper被调用，它就可以运行我们的指定的函数了.

请注意在kernel_thread是如何调用系统调用的,我们知道kernel_thread是在内核中

调用,所以他是可以直接调用系统调用的,像sys_open()等

关于kernel的flags的标志，做如下记述：

/*

* cloning flags:

*/

#define CSIGNAL        0x000000ff    /* signal mask to be sent at exit */

#define CLONE_VM    0x00000100    /* set if VM shared between processes */

#define CLONE_FS    0x00000200    /* set if fs info shared between processes */

#define CLONE_FILES    0x00000400    /* set if open files shared between processes */

#define CLONE_SIGHAND    0x00000800    /* set if signal handlers and blocked signals shared */

#define CLONE_IDLETASK    0x00001000    /* set if new pid should be 0 (kernel only)*/

#define CLONE_PTRACE    0x00002000    /* set if we want to let tracing continue on the child too */

#define CLONE_VFORK    0x00004000    /* set if the parent wants the child to wake it up on mm_release */

#define CLONE_PARENT    0x00008000    /* set if we want to have the same parent as the cloner */

#define CLONE_THREAD    0x00010000    /* Same thread group? */

#define CLONE_NEWNS    0x00020000    /* New namespace group? */

#define CLONE_SYSVSEM    0x00040000    /* share system V SEM_UNDO semantics */

#define CLONE_SETTLS    0x00080000    /* create a new TLS for the child */

#define CLONE_PARENT_SETTID    0x00100000    /* set the TID in the parent */

#define CLONE_CHILD_CLEARTID    0x00200000    /* clear the TID in the child */

#define CLONE_DETACHED        0x00400000    /* Unused, ignored */

#define CLONE_UNTRACED        0x00800000    /* set if the tracing process can’t force CLONE_PTRACE on this clone */

#define CLONE_CHILD_SETTID    0x01000000    /* set the TID in the child */

#define CLONE_STOPPED        0x02000000    /* Start in stopped state */

/*

* List of flags we want to share for kernel threads,

* if only because they are not used by them anyway.

*/

#define CLONE_KERNEL    (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)

内核线程常用的flags就是CLONE_KERNEL。