NetBSD/sys/kern/kern_fork.c

411 lines
11 KiB
C
Raw Normal View History

/* $NetBSD: kern_fork.c,v 1.64 2000/05/08 19:59:21 thorpej Exp $ */
1994-05-17 08:21:49 +04:00
/*
1994-05-19 09:57:44 +04:00
* Copyright (c) 1982, 1986, 1989, 1991, 1993
* The Regents of the University of California. All rights reserved.
1994-05-17 08:21:49 +04:00
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* 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.
*
1998-03-01 05:20:01 +03:00
* @(#)kern_fork.c 8.8 (Berkeley) 2/14/95
1994-05-17 08:21:49 +04:00
*/
1998-06-26 01:17:15 +04:00
#include "opt_ktrace.h"
1994-05-17 08:21:49 +04:00
#include <sys/param.h>
#include <sys/systm.h>
1994-05-19 09:57:44 +04:00
#include <sys/map.h>
1994-05-17 08:21:49 +04:00
#include <sys/filedesc.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
1998-08-02 08:41:32 +04:00
#include <sys/pool.h>
1996-02-09 21:59:18 +03:00
#include <sys/mount.h>
1994-05-17 08:21:49 +04:00
#include <sys/proc.h>
#include <sys/resourcevar.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/acct.h>
#include <sys/ktrace.h>
#include <sys/vmmeter.h>
#include <sys/sched.h>
#include <sys/signalvar.h>
1994-05-17 08:21:49 +04:00
1996-02-09 21:59:18 +03:00
#include <sys/syscallargs.h>
#include <vm/vm.h>
#include <vm/vm_kern.h>
1996-02-09 21:59:18 +03:00
#include <uvm/uvm_extern.h>
int nprocs = 1; /* process 0 */
1996-02-04 05:15:01 +03:00
/*ARGSUSED*/
int
sys_fork(p, v, retval)
1994-05-17 08:21:49 +04:00
struct proc *p;
void *v;
register_t *retval;
1994-05-17 08:21:49 +04:00
{
return (fork1(p, 0, SIGCHLD, NULL, 0, retval, NULL));
1994-05-17 08:21:49 +04:00
}
/*
* vfork(2) system call compatible with 4.4BSD (i.e. BSD with Mach VM).
* Address space is not shared, but parent is blocked until child exit.
*/
1996-02-04 05:15:01 +03:00
/*ARGSUSED*/
int
sys_vfork(p, v, retval)
1994-05-17 08:21:49 +04:00
struct proc *p;
void *v;
register_t *retval;
1994-05-17 08:21:49 +04:00
{
return (fork1(p, FORK_PPWAIT, SIGCHLD, NULL, 0, retval, NULL));
1994-05-17 08:21:49 +04:00
}
/*
* New vfork(2) system call for NetBSD, which implements original 3BSD vfork(2)
* semantics. Address space is shared, and parent is blocked until child exit.
*/
/*ARGSUSED*/
int
sys___vfork14(p, v, retval)
struct proc *p;
void *v;
register_t *retval;
{
return (fork1(p, FORK_PPWAIT|FORK_SHAREVM, SIGCHLD, NULL, 0,
retval, NULL));
}
int
fork1(p1, flags, exitsig, stack, stacksize, retval, rnewprocp)
2000-03-30 13:27:11 +04:00
struct proc *p1;
int flags;
int exitsig;
void *stack;
size_t stacksize;
register_t *retval;
struct proc **rnewprocp;
1994-05-17 08:21:49 +04:00
{
2000-03-30 13:27:11 +04:00
struct proc *p2;
uid_t uid;
1994-05-19 09:57:44 +04:00
struct proc *newproc;
int count, s;
vaddr_t uaddr;
1994-05-17 08:21:49 +04:00
static int nextpid, pidchecked = 0;
/*
1994-05-19 09:57:44 +04:00
* Although process entries are dynamically created, we still keep
1994-05-17 08:21:49 +04:00
* a global limit on the maximum number we will create. Don't allow
* a nonprivileged user to use the last process; don't let root
1994-05-19 09:57:44 +04:00
* exceed the limit. The variable nprocs is the current number of
1994-05-17 08:21:49 +04:00
* processes, maxproc is the limit.
*/
1994-05-19 09:57:44 +04:00
uid = p1->p_cred->p_ruid;
if (__predict_false((nprocs >= maxproc - 1 && uid != 0) ||
nprocs >= maxproc)) {
1994-05-17 08:21:49 +04:00
tablefull("proc");
return (EAGAIN);
}
1994-08-30 10:16:25 +04:00
1994-05-19 09:57:44 +04:00
/*
* Increment the count of procs running with this uid. Don't allow
* a nonprivileged user to exceed their current limit.
*/
count = chgproccnt(uid, 1);
if (__predict_false(uid != 0 && count >
p1->p_rlimit[RLIMIT_NPROC].rlim_cur)) {
1994-05-19 09:57:44 +04:00
(void)chgproccnt(uid, -1);
1994-05-17 08:21:49 +04:00
return (EAGAIN);
1994-05-19 09:57:44 +04:00
}
/*
* Allocate virtual address space for the U-area now, while it
* is still easy to abort the fork operation if we're out of
* kernel virtual address space. The actual U-area pages will
* be allocated and wired in vm_fork().
*/
uaddr = uvm_km_valloc(kernel_map, USPACE);
if (__predict_false(uaddr == 0)) {
(void)chgproccnt(uid, -1);
return (ENOMEM);
}
/*
* We are now committed to the fork. From here on, we may
* block on resources, but resource allocation may NOT fail.
*/
1994-05-19 09:57:44 +04:00
/* Allocate new proc. */
1998-08-02 08:41:32 +04:00
newproc = pool_get(&proc_pool, PR_WAITOK);
1994-05-17 08:21:49 +04:00
/*
* BEGIN PID ALLOCATION.
*/
s = proclist_lock_write();
1994-05-17 08:21:49 +04:00
/*
1994-05-19 09:57:44 +04:00
* Find an unused process ID. We remember a range of unused IDs
* ready to use (from nextpid+1 through pidchecked-1).
1994-05-17 08:21:49 +04:00
*/
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) {
const struct proclist_desc *pd;
1994-05-17 08:21:49 +04:00
pidchecked = PID_MAX;
/*
* Scan the process lists to check whether this pid
1994-05-17 08:21:49 +04:00
* is in use. Remember the lowest pid that's greater
* than nextpid, so we can avoid checking for a while.
*/
pd = proclists;
1994-05-17 08:21:49 +04:00
again:
for (p2 = LIST_FIRST(pd->pd_list); p2 != 0;
p2 = LIST_NEXT(p2, p_list)) {
1994-05-17 08:21:49 +04:00
while (p2->p_pid == nextpid ||
p2->p_pgrp->pg_id == nextpid ||
p2->p_session->s_sid == nextpid) {
1994-05-17 08:21:49 +04:00
nextpid++;
if (nextpid >= pidchecked)
goto retry;
}
if (p2->p_pid > nextpid && pidchecked > p2->p_pid)
pidchecked = p2->p_pid;
1994-05-17 08:21:49 +04:00
if (p2->p_pgrp->pg_id > nextpid &&
pidchecked > p2->p_pgrp->pg_id)
pidchecked = p2->p_pgrp->pg_id;
if (p2->p_session->s_sid > nextpid &&
pidchecked > p2->p_session->s_sid)
pidchecked = p2->p_session->s_sid;
1994-05-17 08:21:49 +04:00
}
/*
* If there's another list, scan it. If we have checked
* them all, we've found one!
*/
pd++;
if (pd->pd_list != NULL)
1994-05-17 08:21:49 +04:00
goto again;
}
nprocs++;
1994-05-19 09:57:44 +04:00
p2 = newproc;
/* Record the pid we've allocated. */
p2->p_pid = nextpid;
/* Record the signal to be delivered to the parent on exit. */
p2->p_exitsig = exitsig;
/*
* Put the proc on allproc before unlocking PID allocation
* so that waiters won't grab it as soon as we unlock.
*/
1999-07-22 22:28:30 +04:00
p2->p_stat = SIDL; /* protect against others */
p2->p_forw = p2->p_back = NULL; /* shouldn't be necessary */
LIST_INSERT_HEAD(&allproc, p2, p_list);
1999-07-22 22:28:30 +04:00
LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
/*
* END PID ALLOCATION.
*/
proclist_unlock_write(s);
1994-05-17 08:21:49 +04:00
/*
* Make a proc table entry for the new process.
* Start by zeroing the section of proc that is zero-initialized,
* then copy the section that is copied directly from the parent.
*/
memset(&p2->p_startzero, 0,
1994-05-17 08:21:49 +04:00
(unsigned) ((caddr_t)&p2->p_endzero - (caddr_t)&p2->p_startzero));
memcpy(&p2->p_startcopy, &p1->p_startcopy,
1994-05-17 08:21:49 +04:00
(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.
*/
1997-02-18 04:30:37 +03:00
p2->p_flag = P_INMEM | (p1->p_flag & P_SUGID);
1994-08-30 10:16:25 +04:00
p2->p_emul = p1->p_emul;
1994-05-19 09:57:44 +04:00
if (p1->p_flag & P_PROFIL)
startprofclock(p2);
p2->p_cred = pool_get(&pcred_pool, PR_WAITOK);
memcpy(p2->p_cred, p1->p_cred, sizeof(*p2->p_cred));
1994-05-17 08:21:49 +04:00
p2->p_cred->p_refcnt = 1;
crhold(p1->p_ucred);
1994-05-19 09:57:44 +04:00
/* bump references to the text vnode (for procfs) */
p2->p_textvp = p1->p_textvp;
if (p2->p_textvp)
1994-05-17 08:21:49 +04:00
VREF(p2->p_textvp);
if (flags & FORK_SHAREFILES)
fdshare(p1, p2);
else
p2->p_fd = fdcopy(p1);
if (flags & FORK_SHARECWD)
cwdshare(p1, p2);
else
p2->p_cwdi = cwdinit(p1);
1994-05-17 08:21:49 +04:00
/*
* 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++;
}
if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
p2->p_flag |= P_CONTROLT;
if (flags & FORK_PPWAIT)
1994-05-17 08:21:49 +04:00
p2->p_flag |= P_PPWAIT;
LIST_INSERT_AFTER(p1, p2, p_pglist);
1994-05-17 08:21:49 +04:00
p2->p_pptr = p1;
LIST_INSERT_HEAD(&p1->p_children, p2, p_sibling);
LIST_INIT(&p2->p_children);
callout_init(&p2->p_realit_ch);
callout_init(&p2->p_tsleep_ch);
1994-05-17 08:21:49 +04:00
#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)
ktradref(p2);
1994-05-17 08:21:49 +04:00
}
#endif
scheduler_fork_hook(p1, p2);
/*
* Create signal actions for the child process.
*/
if (flags & FORK_SHARESIGS)
sigactsshare(p1, p2);
else
p2->p_sigacts = sigactsinit(p1);
1994-05-17 08:21:49 +04:00
/*
* This begins the section where we must prevent the parent
* from being swapped.
*/
PHOLD(p1);
/*
* Finish creating the child process. It will return through a
* different path later.
*/
p2->p_addr = (struct user *)uaddr;
uvm_fork(p1, p2, (flags & FORK_SHAREVM) ? TRUE : FALSE,
stack, stacksize);
1994-05-17 08:21:49 +04:00
/*
* Make child runnable, set start time, and add to run queue.
1994-05-17 08:21:49 +04:00
*/
s = splstatclock();
p2->p_stats->p_start = time;
p2->p_acflag = AFORK;
1994-05-17 08:21:49 +04:00
p2->p_stat = SRUN;
setrunqueue(p2);
splx(s);
1994-05-17 08:21:49 +04:00
/*
* Now can be swapped.
*/
PRELE(p1);
1994-05-17 08:21:49 +04:00
/*
* Update stats now that we know the fork was successful.
*/
uvmexp.forks++;
if (flags & FORK_PPWAIT)
uvmexp.forks_ppwait++;
if (flags & FORK_SHAREVM)
uvmexp.forks_sharevm++;
/*
* Pass a pointer to the new process to the caller.
*/
if (rnewprocp != NULL)
*rnewprocp = p2;
1994-05-17 08:21:49 +04:00
/*
1994-05-19 09:57:44 +04:00
* Preserve synchronization semantics of vfork. If waiting for
* child to exec or exit, set P_PPWAIT on child, and sleep on our
* proc (in case of exit).
1994-05-17 08:21:49 +04:00
*/
if (flags & FORK_PPWAIT)
1994-05-17 08:21:49 +04:00
while (p2->p_flag & P_PPWAIT)
1994-05-19 09:57:44 +04:00
tsleep(p1, PWAIT, "ppwait", 0);
1994-05-17 08:21:49 +04:00
/*
* Return child pid to parent process,
* marking us as parent via retval[1].
*/
if (retval != NULL) {
retval[0] = p2->p_pid;
retval[1] = 0;
}
1994-05-17 08:21:49 +04:00
return (0);
}