kern_fork.c revision 1.20 1 /* $NetBSD: kern_fork.c,v 1.20 1994/08/30 03:05:35 mycroft Exp $ */
2
3 /*
4 * Copyright (c) 1982, 1986, 1989, 1991, 1993
5 * The Regents of the University of California. All rights reserved.
6 * (c) UNIX System Laboratories, Inc.
7 * All or some portions of this file are derived from material licensed
8 * to the University of California by American Telephone and Telegraph
9 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
10 * the permission of UNIX System Laboratories, Inc.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
20 * 3. All advertising materials mentioning features or use of this software
21 * must display the following acknowledgement:
22 * This product includes software developed by the University of
23 * California, Berkeley and its contributors.
24 * 4. Neither the name of the University nor the names of its contributors
25 * may be used to endorse or promote products derived from this software
26 * without specific prior written permission.
27 *
28 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
29 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
30 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
31 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
32 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
33 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
34 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
35 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
37 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38 * SUCH DAMAGE.
39 *
40 * @(#)kern_fork.c 8.6 (Berkeley) 4/8/94
41 */
42
43 #include <sys/param.h>
44 #include <sys/systm.h>
45 #include <sys/map.h>
46 #include <sys/filedesc.h>
47 #include <sys/kernel.h>
48 #include <sys/malloc.h>
49 #include <sys/proc.h>
50 #include <sys/resourcevar.h>
51 #include <sys/vnode.h>
52 #include <sys/file.h>
53 #include <sys/acct.h>
54 #include <sys/ktrace.h>
55
56 /* ARGSUSED */
57 fork(p, uap, retval)
58 struct proc *p;
59 void *uap;
60 int retval[];
61 {
62
63 return (fork1(p, 0, retval));
64 }
65
66 /* ARGSUSED */
67 vfork(p, uap, retval)
68 struct proc *p;
69 void *uap;
70 int retval[];
71 {
72
73 return (fork1(p, 1, retval));
74 }
75
76 int nprocs = 1; /* process 0 */
77
78 fork1(p1, isvfork, retval)
79 register struct proc *p1;
80 int isvfork, retval[];
81 {
82 register struct proc *p2;
83 register uid_t uid;
84 struct proc *newproc;
85 struct proc **hash;
86 int count;
87 static int nextpid, pidchecked = 0;
88
89 /*
90 * Although process entries are dynamically created, we still keep
91 * a global limit on the maximum number we will create. Don't allow
92 * a nonprivileged user to use the last process; don't let root
93 * exceed the limit. The variable nprocs is the current number of
94 * processes, maxproc is the limit.
95 */
96 uid = p1->p_cred->p_ruid;
97 if ((nprocs >= maxproc - 1 && uid != 0) || nprocs >= maxproc) {
98 tablefull("proc");
99 return (EAGAIN);
100 }
101 /*
102 * Increment the count of procs running with this uid. Don't allow
103 * a nonprivileged user to exceed their current limit.
104 */
105 count = chgproccnt(uid, 1);
106 if (uid != 0 && count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur) {
107 (void)chgproccnt(uid, -1);
108 return (EAGAIN);
109 }
110
111 /* Allocate new proc. */
112 MALLOC(newproc, struct proc *, sizeof(struct proc), M_PROC, M_WAITOK);
113
114 /*
115 * Find an unused process ID. We remember a range of unused IDs
116 * ready to use (from nextpid+1 through pidchecked-1).
117 */
118 nextpid++;
119 retry:
120 /*
121 * If the process ID prototype has wrapped around,
122 * restart somewhat above 0, as the low-numbered procs
123 * tend to include daemons that don't exit.
124 */
125 if (nextpid >= PID_MAX) {
126 nextpid = 100;
127 pidchecked = 0;
128 }
129 if (nextpid >= pidchecked) {
130 int doingzomb = 0;
131
132 pidchecked = PID_MAX;
133 /*
134 * Scan the active and zombie procs to check whether this pid
135 * is in use. Remember the lowest pid that's greater
136 * than nextpid, so we can avoid checking for a while.
137 */
138 p2 = allproc.lh_first;
139 again:
140 for (; p2 != 0; p2 = p2->p_list.le_next) {
141 while (p2->p_pid == nextpid ||
142 p2->p_pgrp->pg_id == nextpid) {
143 nextpid++;
144 if (nextpid >= pidchecked)
145 goto retry;
146 }
147 if (p2->p_pid > nextpid && pidchecked > p2->p_pid)
148 pidchecked = p2->p_pid;
149 if (p2->p_pgrp->pg_id > nextpid &&
150 pidchecked > p2->p_pgrp->pg_id)
151 pidchecked = p2->p_pgrp->pg_id;
152 }
153 if (!doingzomb) {
154 doingzomb = 1;
155 p2 = zombproc.lh_first;
156 goto again;
157 }
158 }
159
160
161 nprocs++;
162 p2 = newproc;
163 p2->p_stat = SIDL; /* protect against others */
164 p2->p_pid = nextpid;
165 LIST_INSERT_HEAD(&allproc, p2, p_list);
166 p2->p_forw = p2->p_back = NULL; /* shouldn't be necessary */
167 LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
168
169 /*
170 * Make a proc table entry for the new process.
171 * Start by zeroing the section of proc that is zero-initialized,
172 * then copy the section that is copied directly from the parent.
173 */
174 bzero(&p2->p_startzero,
175 (unsigned) ((caddr_t)&p2->p_endzero - (caddr_t)&p2->p_startzero));
176 bcopy(&p1->p_startcopy, &p2->p_startcopy,
177 (unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy));
178
179 /*
180 * Duplicate sub-structures as needed.
181 * Increase reference counts on shared objects.
182 * The p_stats and p_sigacts substructs are set in vm_fork.
183 */
184 p2->p_flag = P_INMEM;
185 if (p1->p_flag & P_PROFIL)
186 startprofclock(p2);
187 MALLOC(p2->p_cred, struct pcred *, sizeof(struct pcred),
188 M_SUBPROC, M_WAITOK);
189 bcopy(p1->p_cred, p2->p_cred, sizeof(*p2->p_cred));
190 p2->p_cred->p_refcnt = 1;
191 crhold(p1->p_ucred);
192
193 /* bump references to the text vnode (for procfs) */
194 p2->p_textvp = p1->p_textvp;
195 if (p2->p_textvp)
196 VREF(p2->p_textvp);
197
198 p2->p_fd = fdcopy(p1);
199 /*
200 * If p_limit is still copy-on-write, bump refcnt,
201 * otherwise get a copy that won't be modified.
202 * (If PL_SHAREMOD is clear, the structure is shared
203 * copy-on-write.)
204 */
205 if (p1->p_limit->p_lflags & PL_SHAREMOD)
206 p2->p_limit = limcopy(p1->p_limit);
207 else {
208 p2->p_limit = p1->p_limit;
209 p2->p_limit->p_refcnt++;
210 }
211
212 if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
213 p2->p_flag |= P_CONTROLT;
214 if (isvfork)
215 p2->p_flag |= P_PPWAIT;
216 LIST_INSERT_AFTER(p1, p2, p_pglist);
217 p2->p_pptr = p1;
218 LIST_INSERT_HEAD(&p1->p_children, p2, p_sibling);
219 LIST_INIT(&p2->p_children);
220
221 #ifdef KTRACE
222 /*
223 * Copy traceflag and tracefile if enabled.
224 * If not inherited, these were zeroed above.
225 */
226 if (p1->p_traceflag&KTRFAC_INHERIT) {
227 p2->p_traceflag = p1->p_traceflag;
228 if ((p2->p_tracep = p1->p_tracep) != NULL)
229 VREF(p2->p_tracep);
230 }
231 #endif
232
233 /*
234 * This begins the section where we must prevent the parent
235 * from being swapped.
236 */
237 p1->p_holdcnt++;
238 /*
239 * Set return values for child before vm_fork,
240 * so they can be copied to child stack.
241 * We return parent pid, and mark as child in retval[1].
242 * NOTE: the kernel stack may be at a different location in the child
243 * process, and thus addresses of automatic variables (including retval)
244 * may be invalid after vm_fork returns in the child process.
245 */
246 retval[0] = p1->p_pid;
247 retval[1] = 1;
248 if (vm_fork(p1, p2, isvfork)) {
249 /*
250 * Child process. Set start time and get to work.
251 */
252 (void) splclock();
253 p2->p_stats->p_start = time;
254 (void) spl0();
255 p2->p_acflag = AFORK;
256 return (0);
257 }
258
259 /*
260 * Make child runnable and add to run queue.
261 */
262 (void) splhigh();
263 p2->p_stat = SRUN;
264 setrunqueue(p2);
265 (void) spl0();
266
267 /*
268 * Now can be swapped.
269 */
270 p1->p_holdcnt--;
271
272 /*
273 * Preserve synchronization semantics of vfork. If waiting for
274 * child to exec or exit, set P_PPWAIT on child, and sleep on our
275 * proc (in case of exit).
276 */
277 if (isvfork)
278 while (p2->p_flag & P_PPWAIT)
279 tsleep(p1, PWAIT, "ppwait", 0);
280
281 /*
282 * Return child pid to parent process,
283 * marking us as parent via retval[1].
284 */
285 retval[0] = p2->p_pid;
286 retval[1] = 0;
287 return (0);
288 }
289