The Design and Implementation of the FreeBSD Operating System, Second Edition
Now available: The Design and Implementation of the FreeBSD Operating System (Second Edition)


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]

FreeBSD/Linux Kernel Cross Reference
sys/kern/kern_fork.c

Version: -  FREEBSD  -  FREEBSD-12-STABLE  -  FREEBSD-12-0  -  FREEBSD-11-STABLE  -  FREEBSD-11-2  -  FREEBSD-11-1  -  FREEBSD-11-0  -  FREEBSD-10-STABLE  -  FREEBSD-10-4  -  FREEBSD-10-3  -  FREEBSD-10-2  -  FREEBSD-10-1  -  FREEBSD-10-0  -  FREEBSD-9-STABLE  -  FREEBSD-9-3  -  FREEBSD-9-2  -  FREEBSD-9-1  -  FREEBSD-9-0  -  FREEBSD-8-STABLE  -  FREEBSD-8-4  -  FREEBSD-8-3  -  FREEBSD-8-2  -  FREEBSD-8-1  -  FREEBSD-8-0  -  FREEBSD-7-STABLE  -  FREEBSD-7-4  -  FREEBSD-7-3  -  FREEBSD-7-2  -  FREEBSD-7-1  -  FREEBSD-7-0  -  FREEBSD-6-STABLE  -  FREEBSD-6-4  -  FREEBSD-6-3  -  FREEBSD-6-2  -  FREEBSD-6-1  -  FREEBSD-6-0  -  FREEBSD-5-STABLE  -  FREEBSD-5-5  -  FREEBSD-5-4  -  FREEBSD-5-3  -  FREEBSD-5-2  -  FREEBSD-5-1  -  FREEBSD-5-0  -  FREEBSD-4-STABLE  -  FREEBSD-3-STABLE  -  FREEBSD22  -  linux-2.6  -  linux-2.4.22  -  MK83  -  MK84  -  PLAN9  -  DFBSD  -  NETBSD  -  NETBSD5  -  NETBSD4  -  NETBSD3  -  NETBSD20  -  OPENBSD  -  xnu-517  -  xnu-792  -  xnu-792.6.70  -  xnu-1228  -  xnu-1456.1.26  -  xnu-1699.24.8  -  xnu-2050.18.24  -  OPENSOLARIS  -  minix-3-1-1 
SearchContext: -  none  -  3  -  10 

    1 /*      $OpenBSD: kern_fork.c,v 1.101 2008/11/11 02:13:14 tedu Exp $    */
    2 /*      $NetBSD: kern_fork.c,v 1.29 1996/02/09 18:59:34 christos Exp $  */
    3 
    4 /*
    5  * Copyright (c) 1982, 1986, 1989, 1991, 1993
    6  *      The Regents of the University of California.  All rights reserved.
    7  * (c) UNIX System Laboratories, Inc.
    8  * All or some portions of this file are derived from material licensed
    9  * to the University of California by American Telephone and Telegraph
   10  * Co. or Unix System Laboratories, Inc. and are reproduced herein with
   11  * the permission of UNIX System Laboratories, Inc.
   12  *
   13  * Redistribution and use in source and binary forms, with or without
   14  * modification, are permitted provided that the following conditions
   15  * are met:
   16  * 1. Redistributions of source code must retain the above copyright
   17  *    notice, this list of conditions and the following disclaimer.
   18  * 2. Redistributions in binary form must reproduce the above copyright
   19  *    notice, this list of conditions and the following disclaimer in the
   20  *    documentation and/or other materials provided with the distribution.
   21  * 3. Neither the name of the University nor the names of its contributors
   22  *    may be used to endorse or promote products derived from this software
   23  *    without specific prior written permission.
   24  *
   25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
   26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
   27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
   28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
   29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
   30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
   31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
   32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
   33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
   34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
   35  * SUCH DAMAGE.
   36  *
   37  *      @(#)kern_fork.c 8.6 (Berkeley) 4/8/94
   38  */
   39 
   40 #include <sys/param.h>
   41 #include <sys/systm.h>
   42 #include <sys/filedesc.h>
   43 #include <sys/kernel.h>
   44 #include <sys/malloc.h>
   45 #include <sys/mount.h>
   46 #include <sys/proc.h>
   47 #include <sys/exec.h>
   48 #include <sys/resourcevar.h>
   49 #include <sys/signalvar.h>
   50 #include <sys/vnode.h>
   51 #include <sys/file.h>
   52 #include <sys/acct.h>
   53 #include <sys/ktrace.h>
   54 #include <sys/sched.h>
   55 #include <dev/rndvar.h>
   56 #include <sys/pool.h>
   57 #include <sys/mman.h>
   58 #include <sys/ptrace.h>
   59 
   60 #include <sys/syscallargs.h>
   61 
   62 #include "systrace.h"
   63 #include <dev/systrace.h>
   64 
   65 #include <uvm/uvm_extern.h>
   66 #include <uvm/uvm_map.h>
   67 
   68 int     nprocs = 1;             /* process 0 */
   69 int     randompid;              /* when set to 1, pid's go random */
   70 pid_t   lastpid;
   71 struct  forkstat forkstat;
   72 
   73 void fork_return(void *);
   74 int pidtaken(pid_t);
   75 
   76 void process_new(struct proc *, struct proc *);
   77 
   78 void
   79 fork_return(void *arg)
   80 {
   81         struct proc *p = (struct proc *)arg;
   82 
   83         if (p->p_flag & P_TRACED)
   84                 psignal(p, SIGTRAP);
   85 
   86         child_return(p);
   87 }
   88 
   89 /*ARGSUSED*/
   90 int
   91 sys_fork(struct proc *p, void *v, register_t *retval)
   92 {
   93         int flags;
   94 
   95         flags = FORK_FORK;
   96         if (p->p_ptmask & PTRACE_FORK)
   97                 flags |= FORK_PTRACE;
   98         return (fork1(p, SIGCHLD, flags, NULL, 0,
   99             fork_return, NULL, retval, NULL));
  100 }
  101 
  102 /*ARGSUSED*/
  103 int
  104 sys_vfork(struct proc *p, void *v, register_t *retval)
  105 {
  106         return (fork1(p, SIGCHLD, FORK_VFORK|FORK_PPWAIT, NULL, 0, NULL,
  107             NULL, retval, NULL));
  108 }
  109 
  110 int
  111 sys_rfork(struct proc *p, void *v, register_t *retval)
  112 {
  113         struct sys_rfork_args /* {
  114                 syscallarg(int) flags;
  115         } */ *uap = v;
  116 
  117         int rforkflags;
  118         int flags;
  119 
  120         flags = FORK_RFORK;
  121         rforkflags = SCARG(uap, flags);
  122 
  123         if ((rforkflags & RFPROC) == 0)
  124                 return (EINVAL);
  125 
  126         switch(rforkflags & (RFFDG|RFCFDG)) {
  127         case (RFFDG|RFCFDG):
  128                 return EINVAL;
  129         case RFCFDG:
  130                 flags |= FORK_CLEANFILES;
  131                 break;
  132         case RFFDG:
  133                 break;
  134         default:
  135                 flags |= FORK_SHAREFILES;
  136                 break;
  137         }
  138 
  139         if (rforkflags & RFNOWAIT)
  140                 flags |= FORK_NOZOMBIE;
  141 
  142         if (rforkflags & RFMEM)
  143                 flags |= FORK_SHAREVM;
  144 #ifdef RTHREADS
  145         if (rforkflags & RFTHREAD)
  146                 flags |= FORK_THREAD | FORK_SIGHAND;
  147 #endif
  148 
  149         return (fork1(p, SIGCHLD, flags, NULL, 0, NULL, NULL, retval, NULL));
  150 }
  151 
  152 /*
  153  * Allocate and initialize a new process.
  154  */
  155 void
  156 process_new(struct proc *newproc, struct proc *parent)
  157 {
  158         struct process *pr;
  159 
  160         pr = pool_get(&process_pool, PR_WAITOK);
  161         pr->ps_mainproc = newproc;
  162         TAILQ_INIT(&pr->ps_threads);
  163         TAILQ_INSERT_TAIL(&pr->ps_threads, newproc, p_thr_link);
  164         pr->ps_refcnt = 1;
  165         newproc->p_p = pr;
  166 }
  167 
  168 /* print the 'table full' message once per 10 seconds */
  169 struct timeval fork_tfmrate = { 10, 0 };
  170 
  171 int
  172 fork1(struct proc *p1, int exitsig, int flags, void *stack, size_t stacksize,
  173     void (*func)(void *), void *arg, register_t *retval,
  174     struct proc **rnewprocp)
  175 {
  176         struct proc *p2;
  177         uid_t uid;
  178         struct vmspace *vm;
  179         int count;
  180         vaddr_t uaddr;
  181         int s;
  182         extern void endtsleep(void *);
  183         extern void realitexpire(void *);
  184         struct  ptrace_state *newptstat = NULL;
  185 #if NSYSTRACE > 0
  186         void *newstrp = NULL;
  187 #endif
  188 
  189         /*
  190          * Although process entries are dynamically created, we still keep
  191          * a global limit on the maximum number we will create. We reserve
  192          * the last 5 processes to root. The variable nprocs is the current
  193          * number of processes, maxproc is the limit.
  194          */
  195         uid = p1->p_cred->p_ruid;
  196         if ((nprocs >= maxproc - 5 && uid != 0) || nprocs >= maxproc) {
  197                 static struct timeval lasttfm;
  198 
  199                 if (ratecheck(&lasttfm, &fork_tfmrate))
  200                         tablefull("proc");
  201                 return (EAGAIN);
  202         }
  203         nprocs++;
  204 
  205         /*
  206          * Increment the count of procs running with this uid. Don't allow
  207          * a nonprivileged user to exceed their current limit.
  208          */
  209         count = chgproccnt(uid, 1);
  210         if (uid != 0 && count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur) {
  211                 (void)chgproccnt(uid, -1);
  212                 nprocs--;
  213                 return (EAGAIN);
  214         }
  215 
  216         uaddr = uvm_km_alloc1(kernel_map, USPACE, USPACE_ALIGN, 1);
  217         if (uaddr == 0) {
  218                 chgproccnt(uid, -1);
  219                 nprocs--;
  220                 return (ENOMEM);
  221         }
  222 
  223         /*
  224          * From now on, we're committed to the fork and cannot fail.
  225          */
  226 
  227         /* Allocate new proc. */
  228         p2 = pool_get(&proc_pool, PR_WAITOK);
  229 
  230         p2->p_stat = SIDL;                      /* protect against others */
  231         p2->p_exitsig = exitsig;
  232         p2->p_flag = 0;
  233 
  234 #ifdef RTHREADS
  235         if (flags & FORK_THREAD) {
  236                 atomic_setbits_int(&p2->p_flag, P_THREAD);
  237                 p2->p_p = p1->p_p;
  238                 TAILQ_INSERT_TAIL(&p2->p_p->ps_threads, p2, p_thr_link);
  239                 p2->p_p->ps_refcnt++;
  240         } else {
  241                 process_new(p2, p1);
  242         }
  243 #else
  244         process_new(p2, p1);
  245 #endif
  246 
  247         /*
  248          * Make a proc table entry for the new process.
  249          * Start by zeroing the section of proc that is zero-initialized,
  250          * then copy the section that is copied directly from the parent.
  251          */
  252         bzero(&p2->p_startzero,
  253             (unsigned) ((caddr_t)&p2->p_endzero - (caddr_t)&p2->p_startzero));
  254         bcopy(&p1->p_startcopy, &p2->p_startcopy,
  255             (unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy));
  256 
  257         /*
  258          * Initialize the timeouts.
  259          */
  260         timeout_set(&p2->p_sleep_to, endtsleep, p2);
  261         timeout_set(&p2->p_realit_to, realitexpire, p2);
  262 
  263         p2->p_cpu = p1->p_cpu;
  264 
  265         /*
  266          * Duplicate sub-structures as needed.
  267          * Increase reference counts on shared objects.
  268          * The p_stats and p_sigacts substructs are set in vm_fork.
  269          */
  270         p2->p_emul = p1->p_emul;
  271         if (p1->p_flag & P_PROFIL)
  272                 startprofclock(p2);
  273         atomic_setbits_int(&p2->p_flag, p1->p_flag & (P_SUGID | P_SUGIDEXEC));
  274         if (flags & FORK_PTRACE)
  275                 atomic_setbits_int(&p2->p_flag, p1->p_flag & P_TRACED);
  276 #ifdef RTHREADS
  277         if (flags & FORK_THREAD) {
  278                 /* nothing */
  279         } else
  280 #endif
  281         {
  282                 p2->p_p->ps_cred = pool_get(&pcred_pool, PR_WAITOK);
  283                 bcopy(p1->p_p->ps_cred, p2->p_p->ps_cred, sizeof(*p2->p_p->ps_cred));
  284                 p2->p_p->ps_cred->p_refcnt = 1;
  285                 crhold(p1->p_ucred);
  286         }
  287 
  288         /* bump references to the text vnode (for procfs) */
  289         p2->p_textvp = p1->p_textvp;
  290         if (p2->p_textvp)
  291                 VREF(p2->p_textvp);
  292 
  293         if (flags & FORK_CLEANFILES)
  294                 p2->p_fd = fdinit(p1);
  295         else if (flags & FORK_SHAREFILES)
  296                 p2->p_fd = fdshare(p1);
  297         else
  298                 p2->p_fd = fdcopy(p1);
  299 
  300         /*
  301          * If ps_limit is still copy-on-write, bump refcnt,
  302          * otherwise get a copy that won't be modified.
  303          * (If PL_SHAREMOD is clear, the structure is shared
  304          * copy-on-write.)
  305          */
  306 #ifdef RTHREADS
  307         if (flags & FORK_THREAD) {
  308                 /* nothing */
  309         } else
  310 #endif
  311         {
  312                 if (p1->p_p->ps_limit->p_lflags & PL_SHAREMOD)
  313                         p2->p_p->ps_limit = limcopy(p1->p_p->ps_limit);
  314                 else {
  315                         p2->p_p->ps_limit = p1->p_p->ps_limit;
  316                         p2->p_p->ps_limit->p_refcnt++;
  317                 }
  318         }
  319 
  320         if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
  321                 atomic_setbits_int(&p2->p_flag, P_CONTROLT);
  322         if (flags & FORK_PPWAIT)
  323                 atomic_setbits_int(&p2->p_flag, P_PPWAIT);
  324         p2->p_pptr = p1;
  325         if (flags & FORK_NOZOMBIE)
  326                 atomic_setbits_int(&p2->p_flag, P_NOZOMBIE);
  327         LIST_INIT(&p2->p_children);
  328 
  329 #ifdef KTRACE
  330         /*
  331          * Copy traceflag and tracefile if enabled.
  332          * If not inherited, these were zeroed above.
  333          */
  334         if (p1->p_traceflag & KTRFAC_INHERIT) {
  335                 p2->p_traceflag = p1->p_traceflag;
  336                 if ((p2->p_tracep = p1->p_tracep) != NULL)
  337                         VREF(p2->p_tracep);
  338         }
  339 #endif
  340 
  341         /*
  342          * set priority of child to be that of parent
  343          * XXX should move p_estcpu into the region of struct proc which gets
  344          * copied.
  345          */
  346         scheduler_fork_hook(p1, p2);
  347 
  348         /*
  349          * Create signal actions for the child process.
  350          */
  351         if (flags & FORK_SIGHAND)
  352                 sigactsshare(p1, p2);
  353         else
  354                 p2->p_sigacts = sigactsinit(p1);
  355 
  356         /*
  357          * If emulation has process fork hook, call it now.
  358          */
  359         if (p2->p_emul->e_proc_fork)
  360                 (*p2->p_emul->e_proc_fork)(p2, p1);
  361 
  362         p2->p_addr = (struct user *)uaddr;
  363 
  364         /*
  365          * Finish creating the child process.  It will return through a
  366          * different path later.
  367          */
  368         uvm_fork(p1, p2, ((flags & FORK_SHAREVM) ? TRUE : FALSE), stack,
  369             stacksize, func ? func : child_return, arg ? arg : p2);
  370 
  371         timeout_set(&p2->p_stats->p_virt_to, virttimer_trampoline, p2);
  372         timeout_set(&p2->p_stats->p_prof_to, proftimer_trampoline, p2);
  373 
  374         vm = p2->p_vmspace;
  375 
  376         if (flags & FORK_FORK) {
  377                 forkstat.cntfork++;
  378                 forkstat.sizfork += vm->vm_dsize + vm->vm_ssize;
  379         } else if (flags & FORK_VFORK) {
  380                 forkstat.cntvfork++;
  381                 forkstat.sizvfork += vm->vm_dsize + vm->vm_ssize;
  382         } else if (flags & FORK_RFORK) {
  383                 forkstat.cntrfork++;
  384                 forkstat.sizrfork += vm->vm_dsize + vm->vm_ssize;
  385         } else {
  386                 forkstat.cntkthread++;
  387                 forkstat.sizkthread += vm->vm_dsize + vm->vm_ssize;
  388         }
  389 
  390         if (p2->p_flag & P_TRACED && flags & FORK_FORK)
  391                 newptstat = malloc(sizeof(*newptstat), M_SUBPROC, M_WAITOK);
  392 #if NSYSTRACE > 0
  393         if (ISSET(p1->p_flag, P_SYSTRACE))
  394                 newstrp = systrace_getproc();
  395 #endif
  396 
  397         /* Find an unused pid satisfying 1 <= lastpid <= PID_MAX */
  398         do {
  399                 lastpid = 1 + (randompid ? arc4random() : lastpid) % PID_MAX;
  400         } while (pidtaken(lastpid));
  401         p2->p_pid = lastpid;
  402 
  403         LIST_INSERT_HEAD(&allproc, p2, p_list);
  404         LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
  405         LIST_INSERT_HEAD(&p1->p_children, p2, p_sibling);
  406         LIST_INSERT_AFTER(p1, p2, p_pglist);
  407         if (p2->p_flag & P_TRACED) {
  408                 p2->p_oppid = p1->p_pid;
  409                 if (p2->p_pptr != p1->p_pptr)
  410                         proc_reparent(p2, p1->p_pptr);
  411 
  412                 /*
  413                  * Set ptrace status.
  414                  */
  415                 if (flags & FORK_FORK) {
  416                         p2->p_ptstat = newptstat;
  417                         newptstat = NULL;
  418                         p1->p_ptstat->pe_report_event = PTRACE_FORK;
  419                         p2->p_ptstat->pe_report_event = PTRACE_FORK;
  420                         p1->p_ptstat->pe_other_pid = p2->p_pid;
  421                         p2->p_ptstat->pe_other_pid = p1->p_pid;
  422                 }
  423         }
  424 
  425 #if NSYSTRACE > 0
  426         if (newstrp)
  427                 systrace_fork(p1, p2, newstrp);
  428 #endif
  429 
  430         /*
  431          * Make child runnable, set start time, and add to run queue.
  432          */
  433         SCHED_LOCK(s);
  434         getmicrotime(&p2->p_stats->p_start);
  435         p2->p_acflag = AFORK;
  436         p2->p_stat = SRUN;
  437         setrunqueue(p2);
  438         SCHED_UNLOCK(s);
  439 
  440         if (newptstat)
  441                 free(newptstat, M_SUBPROC);
  442 
  443         /*
  444          * Notify any interested parties about the new process.
  445          */
  446         KNOTE(&p1->p_klist, NOTE_FORK | p2->p_pid);
  447 
  448         /*
  449          * Update stats now that we know the fork was successfull.
  450          */
  451         uvmexp.forks++;
  452         if (flags & FORK_PPWAIT)
  453                 uvmexp.forks_ppwait++;
  454         if (flags & FORK_SHAREVM)
  455                 uvmexp.forks_sharevm++;
  456 
  457         /*
  458          * Pass a pointer to the new process to the caller.
  459          */
  460         if (rnewprocp != NULL)
  461                 *rnewprocp = p2;
  462 
  463         /*
  464          * Preserve synchronization semantics of vfork.  If waiting for
  465          * child to exec or exit, set P_PPWAIT on child, and sleep on our
  466          * proc (in case of exit).
  467          */
  468         if (flags & FORK_PPWAIT)
  469                 while (p2->p_flag & P_PPWAIT)
  470                         tsleep(p1, PWAIT, "ppwait", 0);
  471 
  472         /*
  473          * If we're tracing the child, alert the parent too.
  474          */
  475         if ((flags & FORK_PTRACE) && (p1->p_flag & P_TRACED))
  476                 psignal(p1, SIGTRAP);
  477 
  478         /*
  479          * Return child pid to parent process,
  480          * marking us as parent via retval[1].
  481          */
  482         if (retval != NULL) {
  483                 retval[0] = p2->p_pid;
  484                 retval[1] = 0;
  485         }
  486         return (0);
  487 }
  488 
  489 /*
  490  * Checks for current use of a pid, either as a pid or pgid.
  491  */
  492 int
  493 pidtaken(pid_t pid)
  494 {
  495         struct proc *p;
  496 
  497         if (pfind(pid) != NULL)
  498                 return (1);
  499         if (pgfind(pid) != NULL)
  500                 return (1);
  501         LIST_FOREACH(p, &zombproc, p_list)
  502                 if (p->p_pid == pid || p->p_pgid == pid)
  503                         return (1);
  504         return (0);
  505 }
  506 
  507 #if defined(MULTIPROCESSOR)
  508 /*
  509  * XXX This is a slight hack to get newly-formed processes to
  510  * XXX acquire the kernel lock as soon as they run.
  511  */
  512 void
  513 proc_trampoline_mp(void)
  514 {
  515         struct proc *p;
  516 
  517         p = curproc;
  518 
  519         SCHED_ASSERT_LOCKED();
  520         __mp_unlock(&sched_lock);
  521         spl0();
  522         SCHED_ASSERT_UNLOCKED();
  523         KASSERT(__mp_lock_held(&kernel_lock) == 0);
  524 
  525         KERNEL_PROC_LOCK(p);
  526 }
  527 #endif

Cache object: 952f9c643dbb46285c1d753ceeef14a7


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]


This page is part of the FreeBSD/Linux Linux Kernel Cross-Reference, and was automatically generated using a modified version of the LXR engine.