kproc_exit(9) kernel processes

Other Alias

kproc_start, kproc_shutdown, kproc_create, kproc_resume, kproc_suspend, kproc_suspend_check

SYNOPSIS

In sys/kthread.h Ft void Fn kproc_start const void *udata Ft void Fn kproc_shutdown void *arg int howto Ft int Fo kproc_create Fa void (*func)(void *) void *arg struct proc **newpp Fa int flags int pages Fa const char *fmt ... Fc Ft void Fn kproc_exit int ecode Ft int Fn kproc_resume struct proc *p Ft int Fn kproc_suspend struct proc *p int timo Ft void Fn kproc_suspend_check struct proc *p Ft int Fo kproc_kthread_add Fa void (*func)(void *) void *arg Fa struct proc **procptr struct thread **tdptr Fa int flags int pages char * procname const char *fmt ... Fc

DESCRIPTION

In Fx 8.0 , the Fn kthread* 9 family of functions was renamed to be the Fn kproc* 9 family of functions, as they were misnamed and actually produced kernel processes. A new family of different Fn kthread_* 9 functions was added to produce real kernel threads See the kthread(9) man page for more information on those calls. Also note that the Fn kproc_kthread_add 9 function appears in both pages as its functionality is split.

The function Fn kproc_start is used to start ``internal'' daemons such as bufdaemon , pagedaemon , vmdaemon and the syncer and is intended to be called from SYSINIT(9). The Fa udata argument is actually a pointer to a Vt struct kproc_desc which describes the kernel process that should be created:

struct kproc_desc {
        char            *arg0;
        void            (*func)(void);
        struct proc     **global_procpp;
};

The structure members are used by Fn kproc_start as follows:

arg0
String to be used for the name of the process. This string will be copied into the p_comm member of the new process' Vt struct proc .
func
The main function for this kernel process to run.
global_procpp
A pointer to a Vt struct proc pointer that should be updated to point to the newly created process' process structure. If this variable is NULL then it is ignored.

The Fn kproc_create function is used to create a kernel process. The new process shares its address space with process 0, the swapper process, and runs in kernel mode only. The Fa func argument specifies the function that the process should execute. The Fa arg argument is an arbitrary pointer that is passed in as the only argument to Fa func when it is called by the new process. The Fa newpp pointer points to a Vt struct proc pointer that is to be updated to point to the newly created process. If this argument is NULL then it is ignored. The Fa flags argument specifies a set of flags as described in rfork(2). The Fa pages argument specifies the size of the new kernel process's stack in pages. If 0 is used, the default kernel stack size is allocated. The rest of the arguments form a printf(9) argument list that is used to build the name of the new process and is stored in the p_comm member of the new process's Vt struct proc .

The Fn kproc_exit function is used to terminate kernel processes. It should be called by the main function of the kernel process rather than letting the main function return to its caller. The Fa ecode argument specifies the exit status of the process. While exiting, the function exit1(9) will initiate a call to wakeup(9) on the process handle.

The Fn kproc_resume , Fn kproc_suspend , and Fn kproc_suspend_check functions are used to suspend and resume a kernel process. During the main loop of its execution, a kernel process that wishes to allow itself to be suspended should call Fn kproc_suspend_check passing in curproc as the only argument. This function checks to see if the kernel process has been asked to suspend. If it has, it will tsleep(9) until it is told to resume. Once it has been told to resume it will return allowing execution of the kernel process to continue. The other two functions are used to notify a kernel process of a suspend or resume request. The Fa p argument points to the Vt struct proc of the kernel process to suspend or resume. For Fn kproc_suspend , the Fa timo argument specifies a timeout to wait for the kernel process to acknowledge the suspend request and suspend itself.

The Fn kproc_shutdown function is meant to be registered as a shutdown event for kernel processes that need to be suspended voluntarily during system shutdown so as not to interfere with system shutdown activities. The actual suspension of the kernel process is done with Fn kproc_suspend .

The Fn kproc_kthread_add function is much like the Fn kproc_create function above except that if the kproc already exists, then only a new thread (see kthread(9)) is created on the existing process. The Fa func argument specifies the function that the process should execute. The Fa arg argument is an arbitrary pointer that is passed in as the only argument to Fa func when it is called by the new process. The Fa procptr pointer points to a Vt struct proc pointer that is the location to be updated with the new proc pointer if a new process is created, or if not NULL must contain the process pointer for the already existing process. If this argument points to NULL then a new process is created and the field updated. If not NULL, the Fa tdptr pointer points to a Vt struct thread pointer that is the location to be updated with the new thread pointer. The Fa flags argument specifies a set of flags as described in rfork(2). The Fa pages argument specifies the size of the new kernel thread's stack in pages. If 0 is used, the default kernel stack size is allocated. The procname argument is the name the new process should be given if it needs to be created. It is NOT a printf style format specifier but a simple string. The rest of the arguments form a printf(9) argument list that is used to build the name of the new thread and is stored in the td_name member of the new thread's Vt struct thread .

RETURN VALUES

The Fn kproc_create , Fn kproc_resume , and Fn kproc_suspend functions return zero on success and non-zero on failure.

EXAMPLES

This example demonstrates the use of a Vt struct kproc_desc and the functions Fn kproc_start , Fn kproc_shutdown , and Fn kproc_suspend_check to run the bufdaemon process.
static struct proc *bufdaemonproc;
static struct kproc_desc buf_kp = {
        "bufdaemon",
        buf_daemon,
        &bufdaemonproc
};
SYSINIT(bufdaemon, SI_SUB_KTHREAD_BUF, SI_ORDER_FIRST, kproc_start,
    &buf_kp)
static void
buf_daemon()
{
        ...
        /*
         * This process needs to be suspended prior to shutdown sync.
         */
        EVENTHANDLER_REGISTER(shutdown_pre_sync, kproc_shutdown,
            bufdaemonproc, SHUTDOWN_PRI_LAST);
        ...
        for (;;) {
                kproc_suspend_check(bufdaemonproc);
                ...
        }
}

ERRORS

The Fn kproc_resume and Fn kproc_suspend functions will fail if:

Bq Er EINVAL
The Fa p argument does not reference a kernel process.

The Fn kproc_create function will fail if:

Bq Er EAGAIN
The system-imposed limit on the total number of processes under execution would be exceeded. The limit is given by the sysctl(3) MIB variable KERN_MAXPROC
Bq Er EINVAL
The RFCFDG flag was specified in the Fa flags parameter.

HISTORY

The Fn kproc_start function first appeared in Fx 2.2 . The Fn kproc_shutdown , Fn kproc_create , Fn kproc_exit , Fn kproc_resume , Fn kproc_suspend , and Fn kproc_suspend_check functions were introduced in Fx 4.0 . Prior to Fx 5.0 , the Fn kproc_shutdown , Fn kproc_resume , Fn kproc_suspend , and Fn kproc_suspend_check functions were named Fn shutdown_kproc , Fn resume_kproc , Fn shutdown_kproc , and Fn kproc_suspend_loop , respectively. Originally they had the names Fn kthread_* but were changed to Fn kproc_* when real kthreads became available.