[unix-history] / sys / kern / kern_fork.c

/*
 * Copyright (c) 1982, 1986, 1989, 1991 Regents of the University of California.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	from: @(#)kern_fork.c	7.29 (Berkeley) 5/15/91
 *	$Id: kern_fork.c,v 1.6 1993/12/19 00:51:25 wollman Exp $
 */

#include "param.h"
#include "systm.h"
#include "filedesc.h"
#include "kernel.h"
#include "malloc.h"
#include "proc.h"
#include "resourcevar.h"
#include "vnode.h"
#include "file.h"
#include "acct.h"
#include "ktrace.h"
#include "vm/vm.h"

static int fork1(struct proc *, int, int *);

/* ARGSUSED */
int
fork(p, uap, retval)
	struct proc *p;
	void *uap;
	int retval[];
{

	return (fork1(p, 0, retval));
}

/* ARGSUSED */
int
vfork(p, uap, retval)
	struct proc *p;
	void *uap;
	int retval[];
{

	return (fork1(p, 1, retval));
}

int	nprocs = 1;		/* process 0 */

static int
fork1(p1, isvfork, retval)
	register struct proc *p1;
	int isvfork, retval[];
{
	register struct proc *p2;
	register int count, uid;
	static int nextpid, pidchecked = 0;

	count = 0;
	if ((uid = p1->p_ucred->cr_uid) != 0) {
		for (p2 = allproc; p2; p2 = p2->p_nxt)
			if (p2->p_ucred->cr_uid == uid)
				count++;
		for (p2 = zombproc; p2; p2 = p2->p_nxt)
			if (p2->p_ucred->cr_uid == uid)
				count++;
	}
	/*
	 * Although process entries are dynamically entries,
	 * we still keep a global limit on the maximum number
	 * we will create.  Don't allow a nonprivileged user
	 * to exceed its current limit or to bring us within one
	 * of the global limit; don't let root exceed the limit.
	 * nprocs is the current number of processes,
	 * maxproc is the limit.
	 */
	if (nprocs >= maxproc || uid == 0 && nprocs >= maxproc + 1) {
		tablefull("proc");
		return (EAGAIN);
	}
	if (count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur)
		return (EAGAIN);

	/*
	 * Find an unused process ID.
	 * We remember a range of unused IDs ready to use
	 * (from nextpid+1 through pidchecked-1).
	 */
	nextpid++;
retry:
	/*
	 * If the process ID prototype has wrapped around,
	 * restart somewhat above 0, as the low-numbered procs
	 * tend to include daemons that don't exit.
	 */
	if (nextpid >= PID_MAX) {
		nextpid = 100;
		pidchecked = 0;
	}
	if (nextpid >= pidchecked) {
		int doingzomb = 0;

		pidchecked = PID_MAX;
		/*
		 * Scan the active and zombie procs to check whether this pid
		 * is in use.  Remember the lowest pid that's greater
		 * than nextpid, so we can avoid checking for a while.
		 */
		p2 = allproc;
again:
		for (; p2 != NULL; p2 = p2->p_nxt) {
			if (p2->p_pid == nextpid ||
			    p2->p_pgrp->pg_id == nextpid) {
				nextpid++;
				if (nextpid >= pidchecked)
					goto retry;
			}
			if (p2->p_pid > nextpid && pidchecked > p2->p_pid)
				pidchecked = p2->p_pid;
			if (p2->p_pgrp->pg_id > nextpid && 
			    pidchecked > p2->p_pgrp->pg_id)
				pidchecked = p2->p_pgrp->pg_id;
		}
		if (!doingzomb) {
			doingzomb = 1;
			p2 = zombproc;
			goto again;
		}
	}


	/*
	 * Allocate new proc.
	 * Link onto allproc (this should probably be delayed).
	 */
	MALLOC(p2, struct proc *, sizeof(struct proc), M_PROC, M_WAITOK);
	nprocs++;

	/* Initialize all fields to zero */
	bzero((struct proc *)p2, sizeof(struct proc));

	p2->p_nxt = allproc;
	p2->p_nxt->p_prev = &p2->p_nxt;		/* allproc is never NULL */
	p2->p_prev = &allproc;
	allproc = p2;

	/*
	 * Make a proc table entry for the new process.
	 * Copy the section that is copied directly from the parent.
	 */
	bcopy(&p1->p_startcopy, &p2->p_startcopy,
	    (unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy));

	/*
	 * Duplicate sub-structures as needed.
	 * Increase reference counts on shared objects.
	 * The p_stats and p_sigacts substructs are set in vm_fork.
	 */
	MALLOC(p2->p_cred, struct pcred *, sizeof(struct pcred),
	    M_SUBPROC, M_WAITOK);
	bcopy(p1->p_cred, p2->p_cred, sizeof(*p2->p_cred));
	p2->p_cred->p_refcnt = 1;
	crhold(p1->p_ucred);

	p2->p_fd = fdcopy(p1);
	/*
	 * If p_limit is still copy-on-write, bump refcnt,
	 * otherwise get a copy that won't be modified.
	 * (If PL_SHAREMOD is clear, the structure is shared
	 * copy-on-write.)
	 */
	if (p1->p_limit->p_lflags & PL_SHAREMOD)
		p2->p_limit = limcopy(p1->p_limit);
	else {
		p2->p_limit = p1->p_limit;
		p2->p_limit->p_refcnt++;
	}

	p2->p_flag = SLOAD | (p1->p_flag & SHPUX);
	if (p1->p_session->s_ttyvp != NULL && p1->p_flag & SCTTY)
		p2->p_flag |= SCTTY;
	if (isvfork)
		p2->p_flag |= SPPWAIT;
	p2->p_stat = SIDL;
	p2->p_pid = nextpid;
	{
	struct proc **hash = &pidhash[PIDHASH(p2->p_pid)];

	p2->p_hash = *hash;
	*hash = p2;
	}
	p2->p_pgrpnxt = p1->p_pgrpnxt;
	p1->p_pgrpnxt = p2;
	p2->p_pptr = p1;
	p2->p_osptr = p1->p_cptr;
	if (p1->p_cptr)
		p1->p_cptr->p_ysptr = p2;
	p1->p_cptr = p2;
#ifdef KTRACE
	/*
	 * Copy traceflag and tracefile if enabled.
	 * If not inherited, these were zeroed above.
	 */
	if (p1->p_traceflag&KTRFAC_INHERIT) {
		p2->p_traceflag = p1->p_traceflag;
		if ((p2->p_tracep = p1->p_tracep) != NULL)
			VREF(p2->p_tracep);
	}
#endif

#if defined(tahoe)
	p2->p_vmspace->p_ckey = p1->p_vmspace->p_ckey; /* XXX move this */
#endif

	/*
	 * set priority of child to be that of parent
	 */
	p2->p_cpu = p1->p_cpu;

	/*
	 * This begins the section where we must prevent the parent
	 * from being swapped.
	 */
	p1->p_flag |= SKEEP;
	/*
	 * Set return values for child before vm_fork,
	 * so they can be copied to child stack.
	 * We return parent pid, and mark as child in retval[1].
	 * NOTE: the kernel stack may be at a different location in the child
	 * process, and thus addresses of automatic variables (including retval)
	 * may be invalid after vm_fork returns in the child process.
	 */
	retval[0] = p1->p_pid;
	retval[1] = 1;
	if (vm_fork(p1, p2, isvfork)) {
		/*
		 * Child process.  Set start time and get to work.
		 */
		(void) splclock();
		p2->p_stats->p_start = time;
		(void) spl0();
		p2->p_acflag = AFORK;
/*
		vm_map_init_pmap(&p2->p_vmspace->vm_map);
*/
		return (0);
	}

	/*
	 * Make child runnable and add to run queue.
	 */
	(void) splhigh();
	p2->p_stat = SRUN;
	setrq(p2);
	(void) spl0();

	/*
	 * Now can be swapped.
	 */
	p1->p_flag &= ~SKEEP;

	/*
	 * Preserve synchronization semantics of vfork.
	 * If waiting for child to exec or exit, set SPPWAIT
	 * on child, and sleep on our proc (in case of exit).
	 */
	if (isvfork)
		while (p2->p_flag & SPPWAIT)
			tsleep((caddr_t)p1, PWAIT, "ppwait", 0);

	/*
	 * Return child pid to parent process,
	 * marking us as parent via retval[1].
	 */
	retval[0] = p2->p_pid;
	retval[1] = 0;
	return (0);
}
Commit	Line	Data
15637ed4 RG	1	/*
	2	* Copyright (c) 1982, 1986, 1989, 1991 Regents of the University of California.
	3	* All rights reserved.
	4	*
	5	* Redistribution and use in source and binary forms, with or without
	6	* modification, are permitted provided that the following conditions
	7	* are met:
	8	* 1. Redistributions of source code must retain the above copyright
	9	* notice, this list of conditions and the following disclaimer.
	10	* 2. Redistributions in binary form must reproduce the above copyright
	11	* notice, this list of conditions and the following disclaimer in the
	12	* documentation and/or other materials provided with the distribution.
	13	* 3. All advertising materials mentioning features or use of this software
	14	* must display the following acknowledgement:
	15	* This product includes software developed by the University of
	16	* California, Berkeley and its contributors.
	17	* 4. Neither the name of the University nor the names of its contributors
	18	* may be used to endorse or promote products derived from this software
	19	* without specific prior written permission.
	20	*
	21	* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
	22	* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
	23	* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
	24	* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
	25	* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
	26	* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
	27	* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
	28	* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
	29	* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
	30	* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
	31	* SUCH DAMAGE.
	32	*
600f7f07	33	* from: @(#)kern_fork.c 7.29 (Berkeley) 5/15/91
41aefbec	34	* $Id: kern_fork.c,v 1.6 1993/12/19 00:51:25 wollman Exp $
15637ed4 RG	35	*/
	36
	37	#include "param.h"
	38	#include "systm.h"
	39	#include "filedesc.h"
	40	#include "kernel.h"
	41	#include "malloc.h"
	42	#include "proc.h"
	43	#include "resourcevar.h"
	44	#include "vnode.h"
	45	#include "file.h"
	46	#include "acct.h"
	47	#include "ktrace.h"
	48	#include "vm/vm.h"
	49
fde1aeb2 GW	50	static int fork1(struct proc , int, int );
fde1aeb2 GW	51
15637ed4	52	/* ARGSUSED */
4c45483e	53	int
15637ed4 RG	54	fork(p, uap, retval)
	55	struct proc *p;
	56	void *uap;
	57	int retval[];
	58	{
	59
	60	return (fork1(p, 0, retval));
	61	}
	62
	63	/* ARGSUSED */
4c45483e	64	int
15637ed4 RG	65	vfork(p, uap, retval)
	66	struct proc *p;
	67	void *uap;
	68	int retval[];
	69	{
	70
	71	return (fork1(p, 1, retval));
	72	}
	73
	74	int nprocs = 1; /* process 0 */
	75
fde1aeb2	76	static int
15637ed4 RG	77	fork1(p1, isvfork, retval)
	78	register struct proc *p1;
	79	int isvfork, retval[];
	80	{
	81	register struct proc *p2;
	82	register int count, uid;
	83	static int nextpid, pidchecked = 0;
	84
	85	count = 0;
	86	if ((uid = p1->p_ucred->cr_uid) != 0) {
	87	for (p2 = allproc; p2; p2 = p2->p_nxt)
	88	if (p2->p_ucred->cr_uid == uid)
	89	count++;
	90	for (p2 = zombproc; p2; p2 = p2->p_nxt)
	91	if (p2->p_ucred->cr_uid == uid)
	92	count++;
	93	}
	94	/*
	95	* Although process entries are dynamically entries,
	96	* we still keep a global limit on the maximum number
	97	* we will create. Don't allow a nonprivileged user
	98	* to exceed its current limit or to bring us within one
	99	* of the global limit; don't let root exceed the limit.
	100	* nprocs is the current number of processes,
	101	* maxproc is the limit.
	102	*/
	103	if (nprocs >= maxproc \|\| uid == 0 && nprocs >= maxproc + 1) {
	104	tablefull("proc");
	105	return (EAGAIN);
	106	}
	107	if (count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur)
	108	return (EAGAIN);
	109
	110	/*
	111	* Find an unused process ID.
	112	* We remember a range of unused IDs ready to use
	113	* (from nextpid+1 through pidchecked-1).
	114	*/
	115	nextpid++;
	116	retry:
	117	/*
	118	* If the process ID prototype has wrapped around,
	119	* restart somewhat above 0, as the low-numbered procs
	120	* tend to include daemons that don't exit.
	121	*/
	122	if (nextpid >= PID_MAX) {
	123	nextpid = 100;
	124	pidchecked = 0;
	125	}
	126	if (nextpid >= pidchecked) {
	127	int doingzomb = 0;
	128
	129	pidchecked = PID_MAX;
	130	/*
	131	* Scan the active and zombie procs to check whether this pid
	132	* is in use. Remember the lowest pid that's greater
	133	* than nextpid, so we can avoid checking for a while.
	134	*/
	135	p2 = allproc;
	136	again:
	137	for (; p2 != NULL; p2 = p2->p_nxt) {
	138	if (p2->p_pid == nextpid \|\|
	139	p2->p_pgrp->pg_id == nextpid) {
	140	nextpid++;
141	if (nextpid >= pidchecked)
142	goto retry;
143	}
144	if (p2->p_pid > nextpid && pidchecked > p2->p_pid)
145	pidchecked = p2->p_pid;
146	if (p2->p_pgrp->pg_id > nextpid &&
147	pidchecked > p2->p_pgrp->pg_id)
148	pidchecked = p2->p_pgrp->pg_id;
149	}
150	if (!doingzomb) {
151	doingzomb = 1;
152	p2 = zombproc;
153	goto again;
154	}
155	}
156
157
158	/*
159	* Allocate new proc.
160	* Link onto allproc (this should probably be delayed).
161	*/
162	MALLOC(p2, struct proc *, sizeof(struct proc), M_PROC, M_WAITOK);
163	nprocs++;
a9d36f61 DG	164
	165	/* Initialize all fields to zero */
	166	bzero((struct proc *)p2, sizeof(struct proc));
	167
15637ed4 RG	168	p2->p_nxt = allproc;
	169	p2->p_nxt->p_prev = &p2->p_nxt; /* allproc is never NULL */
	170	p2->p_prev = &allproc;
	171	allproc = p2;
15637ed4 RG	172
	173	/*
	174	* Make a proc table entry for the new process.
a9d36f61	175	* Copy the section that is copied directly from the parent.
15637ed4	176	*/
15637ed4 RG	177	bcopy(&p1->p_startcopy, &p2->p_startcopy,
15637ed4 RG	178	(unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy));
15637ed4 RG	179
	180	/*
	181	* Duplicate sub-structures as needed.
	182	* Increase reference counts on shared objects.
	183	* The p_stats and p_sigacts substructs are set in vm_fork.
	184	*/
	185	MALLOC(p2->p_cred, struct pcred *, sizeof(struct pcred),
	186	M_SUBPROC, M_WAITOK);
	187	bcopy(p1->p_cred, p2->p_cred, sizeof(*p2->p_cred));
	188	p2->p_cred->p_refcnt = 1;
	189	crhold(p1->p_ucred);
	190
	191	p2->p_fd = fdcopy(p1);
	192	/*
	193	* If p_limit is still copy-on-write, bump refcnt,
	194	* otherwise get a copy that won't be modified.
	195	* (If PL_SHAREMOD is clear, the structure is shared
	196	* copy-on-write.)
	197	*/
	198	if (p1->p_limit->p_lflags & PL_SHAREMOD)
	199	p2->p_limit = limcopy(p1->p_limit);
	200	else {
	201	p2->p_limit = p1->p_limit;
	202	p2->p_limit->p_refcnt++;
	203	}
	204
	205	p2->p_flag = SLOAD \| (p1->p_flag & SHPUX);
	206	if (p1->p_session->s_ttyvp != NULL && p1->p_flag & SCTTY)
	207	p2->p_flag \|= SCTTY;
	208	if (isvfork)
	209	p2->p_flag \|= SPPWAIT;
	210	p2->p_stat = SIDL;
	211	p2->p_pid = nextpid;
	212	{
	213	struct proc **hash = &pidhash[PIDHASH(p2->p_pid)];
	214
	215	p2->p_hash = *hash;
	216	*hash = p2;
	217	}
	218	p2->p_pgrpnxt = p1->p_pgrpnxt;
	219	p1->p_pgrpnxt = p2;
	220	p2->p_pptr = p1;
	221	p2->p_osptr = p1->p_cptr;
	222	if (p1->p_cptr)
	223	p1->p_cptr->p_ysptr = p2;
	224	p1->p_cptr = p2;
	225	#ifdef KTRACE
	226	/*
	227	* Copy traceflag and tracefile if enabled.
	228	* If not inherited, these were zeroed above.
	229	*/
	230	if (p1->p_traceflag&KTRFAC_INHERIT) {
	231	p2->p_traceflag = p1->p_traceflag;
	232	if ((p2->p_tracep = p1->p_tracep) != NULL)
	233	VREF(p2->p_tracep);
	234	}
	235	#endif
	236
	237	#if defined(tahoe)
	238	p2->p_vmspace->p_ckey = p1->p_vmspace->p_ckey; /* XXX move this */
	239	#endif
	240
41aefbec DG	241	/*
	242	* set priority of child to be that of parent
	243	*/
	244	p2->p_cpu = p1->p_cpu;
	245
15637ed4 RG	246	/*
	247	* This begins the section where we must prevent the parent
	248	* from being swapped.
	249	*/
	250	p1->p_flag \|= SKEEP;
	251	/*
	252	* Set return values for child before vm_fork,
	253	* so they can be copied to child stack.
	254	* We return parent pid, and mark as child in retval[1].
	255	* NOTE: the kernel stack may be at a different location in the child
	256	* process, and thus addresses of automatic variables (including retval)
	257	* may be invalid after vm_fork returns in the child process.
	258	*/
	259	retval[0] = p1->p_pid;
	260	retval[1] = 1;
	261	if (vm_fork(p1, p2, isvfork)) {
	262	/*
	263	* Child process. Set start time and get to work.
	264	*/
	265	(void) splclock();
	266	p2->p_stats->p_start = time;
	267	(void) spl0();
	268	p2->p_acflag = AFORK;
41aefbec DG	269	/*
	270	vm_map_init_pmap(&p2->p_vmspace->vm_map);
	271	*/
15637ed4 RG	272	return (0);
	273	}
	274
	275	/*
	276	* Make child runnable and add to run queue.
	277	*/
	278	(void) splhigh();
	279	p2->p_stat = SRUN;
	280	setrq(p2);
	281	(void) spl0();
	282
	283	/*
	284	* Now can be swapped.
	285	*/
	286	p1->p_flag &= ~SKEEP;
	287
	288	/*
	289	* Preserve synchronization semantics of vfork.
	290	* If waiting for child to exec or exit, set SPPWAIT
	291	* on child, and sleep on our proc (in case of exit).
	292	*/
	293	if (isvfork)
	294	while (p2->p_flag & SPPWAIT)
	295	tsleep((caddr_t)p1, PWAIT, "ppwait", 0);
	296
	297	/*
	298	* Return child pid to parent process,
	299	* marking us as parent via retval[1].
	300	*/
	301	retval[0] = p2->p_pid;
	302	retval[1] = 0;
	303	return (0);
	304	}