| 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 | * |
| 33 | * @(#)kern_fork.c 7.29 (Berkeley) 5/15/91 |
| 34 | */ |
| 35 | |
| 36 | #include "param.h" |
| 37 | #include "systm.h" |
| 38 | #include "filedesc.h" |
| 39 | #include "kernel.h" |
| 40 | #include "malloc.h" |
| 41 | #include "proc.h" |
| 42 | #include "resourcevar.h" |
| 43 | #include "vnode.h" |
| 44 | #include "file.h" |
| 45 | #include "acct.h" |
| 46 | #include "ktrace.h" |
| 47 | #include "vm/vm.h" |
| 48 | |
| 49 | /* ARGSUSED */ |
| 50 | fork(p, uap, retval) |
| 51 | struct proc *p; |
| 52 | void *uap; |
| 53 | int retval[]; |
| 54 | { |
| 55 | |
| 56 | return (fork1(p, 0, retval)); |
| 57 | } |
| 58 | |
| 59 | /* ARGSUSED */ |
| 60 | vfork(p, uap, retval) |
| 61 | struct proc *p; |
| 62 | void *uap; |
| 63 | int retval[]; |
| 64 | { |
| 65 | |
| 66 | return (fork1(p, 1, retval)); |
| 67 | } |
| 68 | |
| 69 | int nprocs = 1; /* process 0 */ |
| 70 | |
| 71 | fork1(p1, isvfork, retval) |
| 72 | register struct proc *p1; |
| 73 | int isvfork, retval[]; |
| 74 | { |
| 75 | register struct proc *p2; |
| 76 | register int count, uid; |
| 77 | static int nextpid, pidchecked = 0; |
| 78 | |
| 79 | count = 0; |
| 80 | if ((uid = p1->p_ucred->cr_uid) != 0) { |
| 81 | for (p2 = allproc; p2; p2 = p2->p_nxt) |
| 82 | if (p2->p_ucred->cr_uid == uid) |
| 83 | count++; |
| 84 | for (p2 = zombproc; p2; p2 = p2->p_nxt) |
| 85 | if (p2->p_ucred->cr_uid == uid) |
| 86 | count++; |
| 87 | } |
| 88 | /* |
| 89 | * Although process entries are dynamically entries, |
| 90 | * we still keep a global limit on the maximum number |
| 91 | * we will create. Don't allow a nonprivileged user |
| 92 | * to exceed its current limit or to bring us within one |
| 93 | * of the global limit; don't let root exceed the limit. |
| 94 | * nprocs is the current number of processes, |
| 95 | * maxproc is the limit. |
| 96 | */ |
| 97 | if (nprocs >= maxproc || uid == 0 && nprocs >= maxproc + 1) { |
| 98 | tablefull("proc"); |
| 99 | return (EAGAIN); |
| 100 | } |
| 101 | if (count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur) |
| 102 | return (EAGAIN); |
| 103 | |
| 104 | /* |
| 105 | * Find an unused process ID. |
| 106 | * We remember a range of unused IDs ready to use |
| 107 | * (from nextpid+1 through pidchecked-1). |
| 108 | */ |
| 109 | nextpid++; |
| 110 | retry: |
| 111 | /* |
| 112 | * If the process ID prototype has wrapped around, |
| 113 | * restart somewhat above 0, as the low-numbered procs |
| 114 | * tend to include daemons that don't exit. |
| 115 | */ |
| 116 | if (nextpid >= PID_MAX) { |
| 117 | nextpid = 100; |
| 118 | pidchecked = 0; |
| 119 | } |
| 120 | if (nextpid >= pidchecked) { |
| 121 | int doingzomb = 0; |
| 122 | |
| 123 | pidchecked = PID_MAX; |
| 124 | /* |
| 125 | * Scan the active and zombie procs to check whether this pid |
| 126 | * is in use. Remember the lowest pid that's greater |
| 127 | * than nextpid, so we can avoid checking for a while. |
| 128 | */ |
| 129 | p2 = allproc; |
| 130 | again: |
| 131 | for (; p2 != NULL; p2 = p2->p_nxt) { |
| 132 | if (p2->p_pid == nextpid || |
| 133 | p2->p_pgrp->pg_id == nextpid) { |
| 134 | nextpid++; |
| 135 | if (nextpid >= pidchecked) |
| 136 | goto retry; |
| 137 | } |
| 138 | if (p2->p_pid > nextpid && pidchecked > p2->p_pid) |
| 139 | pidchecked = p2->p_pid; |
| 140 | if (p2->p_pgrp->pg_id > nextpid && |
| 141 | pidchecked > p2->p_pgrp->pg_id) |
| 142 | pidchecked = p2->p_pgrp->pg_id; |
| 143 | } |
| 144 | if (!doingzomb) { |
| 145 | doingzomb = 1; |
| 146 | p2 = zombproc; |
| 147 | goto again; |
| 148 | } |
| 149 | } |
| 150 | |
| 151 | |
| 152 | /* |
| 153 | * Allocate new proc. |
| 154 | * Link onto allproc (this should probably be delayed). |
| 155 | */ |
| 156 | MALLOC(p2, struct proc *, sizeof(struct proc), M_PROC, M_WAITOK); |
| 157 | nprocs++; |
| 158 | p2->p_nxt = allproc; |
| 159 | p2->p_nxt->p_prev = &p2->p_nxt; /* allproc is never NULL */ |
| 160 | p2->p_prev = &allproc; |
| 161 | allproc = p2; |
| 162 | p2->p_link = NULL; /* shouldn't be necessary */ |
| 163 | p2->p_rlink = NULL; /* shouldn't be necessary */ |
| 164 | |
| 165 | /* |
| 166 | * Make a proc table entry for the new process. |
| 167 | * Start by zeroing the section of proc that is zero-initialized, |
| 168 | * then copy the section that is copied directly from the parent. |
| 169 | */ |
| 170 | bzero(&p2->p_startzero, |
| 171 | (unsigned) ((caddr_t)&p2->p_endzero - (caddr_t)&p2->p_startzero)); |
| 172 | bcopy(&p1->p_startcopy, &p2->p_startcopy, |
| 173 | (unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy)); |
| 174 | p2->p_spare[0] = 0; /* XXX - should be in zero range */ |
| 175 | p2->p_spare[1] = 0; /* XXX - should be in zero range */ |
| 176 | p2->p_spare[2] = 0; /* XXX - should be in zero range */ |
| 177 | p2->p_spare[3] = 0; /* XXX - should be in zero range */ |
| 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 | MALLOC(p2->p_cred, struct pcred *, sizeof(struct pcred), |
| 185 | M_SUBPROC, M_WAITOK); |
| 186 | bcopy(p1->p_cred, p2->p_cred, sizeof(*p2->p_cred)); |
| 187 | p2->p_cred->p_refcnt = 1; |
| 188 | crhold(p1->p_ucred); |
| 189 | |
| 190 | p2->p_fd = fdcopy(p1); |
| 191 | /* |
| 192 | * If p_limit is still copy-on-write, bump refcnt, |
| 193 | * otherwise get a copy that won't be modified. |
| 194 | * (If PL_SHAREMOD is clear, the structure is shared |
| 195 | * copy-on-write.) |
| 196 | */ |
| 197 | if (p1->p_limit->p_lflags & PL_SHAREMOD) |
| 198 | p2->p_limit = limcopy(p1->p_limit); |
| 199 | else { |
| 200 | p2->p_limit = p1->p_limit; |
| 201 | p2->p_limit->p_refcnt++; |
| 202 | } |
| 203 | |
| 204 | p2->p_flag = SLOAD | (p1->p_flag & SHPUX); |
| 205 | if (p1->p_session->s_ttyvp != NULL && p1->p_flag & SCTTY) |
| 206 | p2->p_flag |= SCTTY; |
| 207 | if (isvfork) |
| 208 | p2->p_flag |= SPPWAIT; |
| 209 | p2->p_stat = SIDL; |
| 210 | p2->p_pid = nextpid; |
| 211 | { |
| 212 | struct proc **hash = &pidhash[PIDHASH(p2->p_pid)]; |
| 213 | |
| 214 | p2->p_hash = *hash; |
| 215 | *hash = p2; |
| 216 | } |
| 217 | p2->p_pgrpnxt = p1->p_pgrpnxt; |
| 218 | p1->p_pgrpnxt = p2; |
| 219 | p2->p_pptr = p1; |
| 220 | p2->p_osptr = p1->p_cptr; |
| 221 | if (p1->p_cptr) |
| 222 | p1->p_cptr->p_ysptr = p2; |
| 223 | p1->p_cptr = p2; |
| 224 | #ifdef KTRACE |
| 225 | /* |
| 226 | * Copy traceflag and tracefile if enabled. |
| 227 | * If not inherited, these were zeroed above. |
| 228 | */ |
| 229 | if (p1->p_traceflag&KTRFAC_INHERIT) { |
| 230 | p2->p_traceflag = p1->p_traceflag; |
| 231 | if ((p2->p_tracep = p1->p_tracep) != NULL) |
| 232 | VREF(p2->p_tracep); |
| 233 | } |
| 234 | #endif |
| 235 | |
| 236 | #if defined(tahoe) |
| 237 | p2->p_vmspace->p_ckey = p1->p_vmspace->p_ckey; /* XXX move this */ |
| 238 | #endif |
| 239 | |
| 240 | /* |
| 241 | * This begins the section where we must prevent the parent |
| 242 | * from being swapped. |
| 243 | */ |
| 244 | p1->p_flag |= SKEEP; |
| 245 | /* |
| 246 | * Set return values for child before vm_fork, |
| 247 | * so they can be copied to child stack. |
| 248 | * We return parent pid, and mark as child in retval[1]. |
| 249 | * NOTE: the kernel stack may be at a different location in the child |
| 250 | * process, and thus addresses of automatic variables (including retval) |
| 251 | * may be invalid after vm_fork returns in the child process. |
| 252 | */ |
| 253 | retval[0] = p1->p_pid; |
| 254 | retval[1] = 1; |
| 255 | if (vm_fork(p1, p2, isvfork)) { |
| 256 | /* |
| 257 | * Child process. Set start time and get to work. |
| 258 | */ |
| 259 | (void) splclock(); |
| 260 | p2->p_stats->p_start = time; |
| 261 | (void) spl0(); |
| 262 | p2->p_acflag = AFORK; |
| 263 | return (0); |
| 264 | } |
| 265 | |
| 266 | /* |
| 267 | * Make child runnable and add to run queue. |
| 268 | */ |
| 269 | (void) splhigh(); |
| 270 | p2->p_stat = SRUN; |
| 271 | setrq(p2); |
| 272 | (void) spl0(); |
| 273 | |
| 274 | /* |
| 275 | * Now can be swapped. |
| 276 | */ |
| 277 | p1->p_flag &= ~SKEEP; |
| 278 | |
| 279 | /* |
| 280 | * Preserve synchronization semantics of vfork. |
| 281 | * If waiting for child to exec or exit, set SPPWAIT |
| 282 | * on child, and sleep on our proc (in case of exit). |
| 283 | */ |
| 284 | if (isvfork) |
| 285 | while (p2->p_flag & SPPWAIT) |
| 286 | tsleep((caddr_t)p1, PWAIT, "ppwait", 0); |
| 287 | |
| 288 | /* |
| 289 | * Return child pid to parent process, |
| 290 | * marking us as parent via retval[1]. |
| 291 | */ |
| 292 | retval[0] = p2->p_pid; |
| 293 | retval[1] = 0; |
| 294 | return (0); |
| 295 | } |