BSD 4_4 release

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

diff --git a/usr/src/sys/kern/kern_fork.c b/usr/src/sys/kern/kern_fork.c
index 27a22ab..eec00ae 100644 (file)
--- a/usr/src/sys/kern/kern_fork.c
+++ b/usr/src/sys/kern/kern_fork.c
@@ -1,336 +1,306 @@
 /*
 /*

- * Copyright (c) 1982, 1986, 1989 Regents of the University of California.
- * All rights reserved.
+ * Copyright (c) 1982, 1986, 1989, 1991, 1993
+ *     The Regents of the University of California.  All rights reserved.

  *
  *

- * Redistribution is only permitted until one year after the first shipment
- * of 4.4BSD by the Regents.  Otherwise, redistribution and use in source and
- * binary forms are permitted provided that: (1) source distributions retain
- * this entire copyright notice and comment, and (2) distributions including
- * binaries display the following acknowledgement:  This product includes
- * software developed by the University of California, Berkeley and its
- * contributors'' in the documentation or other materials provided with the
- * distribution and in all advertising materials mentioning features or use
- * of this software.  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 AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
- * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+ * 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.

  *
  *

- *     @(#)kern_fork.c 7.20 (Berkeley) 7/27/90
+ * 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.
+ *
+ *     @(#)kern_fork.c 8.1 (Berkeley) 6/10/93

  */
 
  */
 

-#include "param.h"
-#include "systm.h"
-#include "map.h"
-#include "user.h"
-#include "kernel.h"
-#include "proc.h"
-#include "vnode.h"
-#include "seg.h"
-#include "vm.h"
-#include "text.h"
-#include "file.h"
-#include "acct.h"
-#include "ktrace.h"
-#include "../ufs/quota.h"
-
-#include "machine/reg.h"
-#include "machine/pte.h"
-#include "machine/psl.h"
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/map.h>
+#include <sys/filedesc.h>
+#include <sys/kernel.h>
+#include <sys/malloc.h>
+#include <sys/proc.h>
+#include <sys/resourcevar.h>
+#include <sys/vnode.h>
+#include <sys/file.h>
+#include <sys/acct.h>
+#include <sys/ktrace.h>

 
 

-/*
- * fork system call.
- */
+struct fork_args {
+       int     dummy;
+};

 /* ARGSUSED */
 fork(p, uap, retval)
        struct proc *p;
 /* ARGSUSED */
 fork(p, uap, retval)
        struct proc *p;

-       struct args *uap;
+       struct fork_args *uap;

        int retval[];
 {
        int retval[];
 {

-       int error;

 
 

-       u.u_cdmap = zdmap;
-       u.u_csmap = zdmap;
-       if (error = swpexpand(u.u_dsize, u.u_ssize, &u.u_cdmap, &u.u_csmap)) {
-               retval[1] = 0;
-               return (error);
-       }

        return (fork1(p, 0, retval));
 }
 
 /* ARGSUSED */
 vfork(p, uap, retval)
        struct proc *p;
        return (fork1(p, 0, retval));
 }
 
 /* ARGSUSED */
 vfork(p, uap, retval)
        struct proc *p;

-       struct args *uap;
+       struct fork_args *uap;

        int retval[];
 {
 
        return (fork1(p, 1, retval));
 }
 
        int retval[];
 {
 
        return (fork1(p, 1, retval));
 }
 

+int    nprocs = 1;             /* process 0 */
+

 fork1(p1, isvfork, retval)
        register struct proc *p1;
        int isvfork, retval[];
 {
        register struct proc *p2;
 fork1(p1, isvfork, retval)
        register struct proc *p1;
        int isvfork, retval[];
 {
        register struct proc *p2;

-       register int a;
+       register uid_t uid;
+       struct proc *newproc;
+       struct proc **hash;
+       int count;
+       static int nextpid, pidchecked = 0;

 
 

-       a = 0;
-       if (p1->p_uid != 0) {
-               for (p2 = allproc; p2; p2 = p2->p_nxt)
-                       if (p2->p_uid == p1->p_uid)
-                               a++;
-               for (p2 = zombproc; p2; p2 = p2->p_nxt)
-                       if (p2->p_uid == p1->p_uid)
-                               a++;
-       }

        /*
        /*

-        * Disallow if
-        *  No processes at all;
-        *  not su and too many procs owned; or
-        *  not su and would take last slot.
+        * Although process entries are dynamically created, we still keep
+        * a global limit on the maximum number we will create.  Don't allow
+        * a nonprivileged user to bring the system within one of the global
+        * limit; don't let root exceed the limit. The variable nprocs is
+        * the current number of processes, maxproc is the limit.

         */
         */

-       p2 = freeproc;
-       if (p2==NULL)
+       uid = p1->p_cred->p_ruid;
+       if (nprocs >= maxproc || uid == 0 && nprocs >= maxproc + 1) {

                tablefull("proc");
                tablefull("proc");

-       if (p2 == NULL ||
-           (p1->p_uid != 0 && (p2->p_nxt == NULL || a > MAXUPRC))) {
-               if (!isvfork) {
-                       (void) vsexpand((segsz_t)0, &u.u_cdmap, 1);
-                       (void) vsexpand((segsz_t)0, &u.u_csmap, 1);
-               }
-               retval[1] = 0;

                return (EAGAIN);
        }
                return (EAGAIN);
        }

-       if (newproc(isvfork)) {
-               retval[0] = p1->p_pid;
-               retval[1] = 1;  /* child */
-               u.u_acflag = AFORK;
-               return (0);
+       /*
+        * 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 (uid != 0 && count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur) {
+               (void)chgproccnt(uid, -1);
+               return (EAGAIN);

        }
        }

-       retval[0] = p2->p_pid;
-       retval[1] = 0;
-       return (0);
-}

 
 

-/*
- * Create a new process-- the internal version of
- * sys fork.
- * It returns 1 in the new process, 0 in the old.
- */
-newproc(isvfork)
-       int isvfork;
-{
-       register struct proc *rpp, *rip;
-       register int n;
-       register struct file *fp;
-       static int pidchecked = 0;
+       /* Allocate new proc. */
+       MALLOC(newproc, struct proc *, sizeof(struct proc), M_PROC, M_WAITOK);

 
        /*
 
        /*

-        * First, just locate a slot for a process
-        * and copy the useful info from this process into it.
-        * The panic "cannot happen" because fork has already
-        * checked for the existence of a slot.
+        * Find an unused process ID.  We remember a range of unused IDs
+        * ready to use (from nextpid+1 through pidchecked-1).

         */
         */

-       mpid++;
+       nextpid++;

 retry:
 retry:

-       if (mpid >= PID_MAX) {
-               mpid = 100;
+       /*
+        * 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;
        }
                pidchecked = 0;
        }

-       if (mpid >= pidchecked) {
+       if (nextpid >= pidchecked) {

                int doingzomb = 0;
 
                pidchecked = PID_MAX;
                /*
                int doingzomb = 0;
 
                pidchecked = PID_MAX;
                /*

-                * Scan the proc table to check whether this pid
+                * Scan the active and zombie procs to check whether this pid

                 * is in use.  Remember the lowest pid that's greater
                 * is in use.  Remember the lowest pid that's greater

-                * than mpid, so we can avoid checking for a while.
+                * than nextpid, so we can avoid checking for a while.

                 */
                 */

-               rpp = allproc;
+               p2 = (struct proc *)allproc;

 again:
 again:

-               for (; rpp != NULL; rpp = rpp->p_nxt) {
-                       if (rpp->p_pid == mpid || rpp->p_pgrp->pg_id == mpid) {
-                               mpid++;
-                               if (mpid >= pidchecked)
+               for (; p2 != NULL; p2 = p2->p_nxt) {
+                       while (p2->p_pid == nextpid ||
+                           p2->p_pgrp->pg_id == nextpid) {
+                               nextpid++;
+                               if (nextpid >= pidchecked)

                                        goto retry;
                        }
                                        goto retry;
                        }

-                       if (rpp->p_pid > mpid && pidchecked > rpp->p_pid)
-                               pidchecked = rpp->p_pid;
-                       if (rpp->p_pgrp->pg_id > mpid && 
-                           pidchecked > rpp->p_pgrp->pg_id)
-                               pidchecked = rpp->p_pgrp->pg_id;
+                       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;
                }
                if (!doingzomb) {
                        doingzomb = 1;

-                       rpp = zombproc;
+                       p2 = zombproc;

                        goto again;
                }
        }
                        goto again;
                }
        }

-       if ((rpp = freeproc) == NULL)
-               panic("no procs");

 
 

-       freeproc = rpp->p_nxt;                  /* off freeproc */
-       rpp->p_nxt = allproc;                   /* onto allproc */
-       rpp->p_nxt->p_prev = &rpp->p_nxt;       /*   (allproc is never NULL) */
-       rpp->p_prev = &allproc;
-       allproc = rpp;
+
+       /*
+        * Link onto allproc (this should probably be delayed).
+        * Heavy use of volatile here to prevent the compiler from
+        * rearranging code.  Yes, it *is* terribly ugly, but at least
+        * it works.
+        */
+       nprocs++;
+       p2 = newproc;
+#define        Vp2 ((volatile struct proc *)p2)
+       Vp2->p_stat = SIDL;                     /* protect against others */
+       Vp2->p_pid = nextpid;
+       /*
+        * This is really:
+        *      p2->p_nxt = allproc;
+        *      allproc->p_prev = &p2->p_nxt;
+        *      p2->p_prev = &allproc;
+        *      allproc = p2;
+        * The assignment via allproc is legal since it is never NULL.
+        */
+       *(volatile struct proc **)&Vp2->p_nxt = allproc;
+       *(volatile struct proc ***)&allproc->p_prev =
+           (volatile struct proc **)&Vp2->p_nxt;
+       *(volatile struct proc ***)&Vp2->p_prev = &allproc;
+       allproc = Vp2;
+#undef Vp2
+       p2->p_link = NULL;                      /* shouldn't be necessary */
+       p2->p_rlink = NULL;                     /* shouldn't be necessary */
+
+       /* Insert on the hash chain. */
+       hash = &pidhash[PIDHASH(p2->p_pid)];
+       p2->p_hash = *hash;
+       *hash = p2;

 
        /*
         * Make a proc table entry for the new process.
 
        /*
         * 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.

         */
         */

-       rip = u.u_procp;
-#if defined(tahoe)
-       rpp->p_ckey = rip->p_ckey;
-       rpp->p_dkey = 0;
-#endif
-       rpp->p_stat = SIDL;
-       timerclear(&rpp->p_realtimer.it_value);
-       rpp->p_flag = SLOAD | (rip->p_flag & (SPAGV|SHPUX));
-       if (rip->p_session->s_ttyvp != NULL && rip->p_flag & SCTTY)
-               rpp->p_flag |= SCTTY;
-       if (isvfork) {
-               rpp->p_flag |= SVFORK;
-               rpp->p_ndx = rip->p_ndx;
-       } else
-               rpp->p_ndx = rpp - proc;
-       bcopy(rip->p_comm, rpp->p_comm, MAXCOMLEN+1);
-       bcopy(rip->p_logname, rpp->p_logname, MAXLOGNAME);
-       rpp->p_uid = rip->p_uid;
-       rpp->p_ruid = rip->p_ruid;
-       rpp->p_rgid = rip->p_rgid;
-       rpp->p_pgrp = rip->p_pgrp;
-       rpp->p_pgrpnxt = rip->p_pgrpnxt;
-       rip->p_pgrpnxt = rpp;
-       rpp->p_nice = rip->p_nice;
-       rpp->p_textp = isvfork ? 0 : rip->p_textp;
-       rpp->p_pid = mpid;
-       rpp->p_ppid = rip->p_pid;
-       rpp->p_pptr = rip;
-       rpp->p_osptr = rip->p_cptr;
-       if (rip->p_cptr)
-               rip->p_cptr->p_ysptr = rpp;
-       rpp->p_ysptr = NULL;
-       rpp->p_cptr = NULL;
-       rip->p_cptr = rpp;
-       rpp->p_time = 0;
-       bzero((caddr_t)&rpp->p_utime, sizeof (struct timeval));
-       bzero((caddr_t)&rpp->p_stime, sizeof (struct timeval));
-       rpp->p_cpu = 0;
-       rpp->p_sigmask = rip->p_sigmask;
-       rpp->p_sigcatch = rip->p_sigcatch;
-       rpp->p_sigignore = rip->p_sigignore;
-       /* take along any pending signals like stops? */
-       if (isvfork) {
-               rpp->p_tsize = rpp->p_dsize = rpp->p_mmsize = rpp->p_ssize = 0;
-               rpp->p_szpt = clrnd(ctopt(HIGHPAGES));
-               forkstat.cntvfork++;
-               forkstat.sizvfork += rip->p_dsize + rip->p_ssize;
-       } else {
-               rpp->p_tsize = rip->p_tsize;
-               rpp->p_dsize = rip->p_dsize;
-               rpp->p_mmsize = rip->p_mmsize;
-               rpp->p_ssize = rip->p_ssize;
-               rpp->p_szpt = rip->p_szpt;
-               forkstat.cntfork++;
-               forkstat.sizfork += rip->p_dsize + rip->p_ssize;
-       }
-#ifdef KTRACE
-       if (rip->p_traceflag&KTRFAC_INHERIT) {
-               rpp->p_traceflag = rip->p_traceflag;
-               if ((rpp->p_tracep = rip->p_tracep) != NULL)
-                       VREF(rpp->p_tracep);
-       } else {
-               rpp->p_tracep = NULL;
-               rpp->p_traceflag = 0;
-       }
-#endif
-       rpp->p_rssize = 0;
-       rpp->p_maxrss = rip->p_maxrss;
-       rpp->p_wchan = 0;
-       rpp->p_slptime = 0;
-       rpp->p_pctcpu = 0;
-       rpp->p_cpticks = 0;
-       {
-       struct proc **hash = &pidhash[PIDHASH(rpp->p_pid)];
+       bzero(&p2->p_startzero,
+           (unsigned) ((caddr_t)&p2->p_endzero - (caddr_t)&p2->p_startzero));
+       bcopy(&p1->p_startcopy, &p2->p_startcopy,
+           (unsigned) ((caddr_t)&p2->p_endcopy - (caddr_t)&p2->p_startcopy));

 
 

-       rpp->p_hash = *hash;
-       *hash = rpp;
+       /*
+        * Duplicate sub-structures as needed.
+        * Increase reference counts on shared objects.
+        * The p_stats and p_sigacts substructs are set in vm_fork.
+        */
+       p2->p_flag = SLOAD;
+       if (p1->p_flag & SPROFIL)
+               startprofclock(p2);
+       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++;

        }
        }

-       multprog++;

 
 

+       if (p1->p_session->s_ttyvp != NULL && p1->p_flag & SCTTY)
+               p2->p_flag |= SCTTY;
+       if (isvfork)
+               p2->p_flag |= SPPWAIT;
+       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

        /*
        /*

-        * Increase reference counts on shared objects.
+        * Copy traceflag and tracefile if enabled.
+        * If not inherited, these were zeroed above.

         */
         */

-       for (n = 0; n <= u.u_lastfile; n++) {
-               fp = u.u_ofile[n];
-               if (fp == NULL)
-                       continue;
-               fp->f_count++;
+       if (p1->p_traceflag&KTRFAC_INHERIT) {
+               p2->p_traceflag = p1->p_traceflag;
+               if ((p2->p_tracep = p1->p_tracep) != NULL)
+                       VREF(p2->p_tracep);

        }
        }

-       VREF(u.u_cdir);
-       if (u.u_rdir)
-               VREF(u.u_rdir);
-       crhold(u.u_cred);
+#endif

 
        /*
         * This begins the section where we must prevent the parent
         * from being swapped.
         */
 
        /*
         * This begins the section where we must prevent the parent
         * from being swapped.
         */

-       rip->p_flag |= SKEEP;
-       if (procdup(rpp, isvfork)) {
+       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();
                (void) splclock();

-               u.u_start = time;
+               p2->p_stats->p_start = time;

                (void) spl0();
                (void) spl0();

-               return (1);
+               p2->p_acflag = AFORK;
+               return (0);

        }
 
        /*
         * Make child runnable and add to run queue.
         */
        }
 
        /*
         * Make child runnable and add to run queue.
         */

-       (void) splclock();
-       rpp->p_stat = SRUN;
-       setrq(rpp);
+       (void) splhigh();
+       p2->p_stat = SRUN;
+       setrq(p2);

        (void) spl0();
 
        (void) spl0();
 

-       /*
-        * Cause child to take a non-local goto as soon as it runs.
-        * On older systems this was done with SSWAP bit in proc
-        * table; on VAX we use u.u_pcb.pcb_sswap so don't need
-        * to do rpp->p_flag |= SSWAP.  Actually do nothing here.
-        */
-       /* rpp->p_flag |= SSWAP; */
-

        /*
         * Now can be swapped.
         */
        /*
         * Now can be swapped.
         */

-       rip->p_flag &= ~SKEEP;
+       p1->p_flag &= ~SKEEP;

 
        /*
 
        /*

-        * If vfork make chain from parent process to child
-        * (where virtal memory is temporarily).  Wait for
-        * child to finish, steal virtual memory back,
-        * and wakeup child to let it die.
+        * 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) {
-               u.u_procp->p_xlink = rpp;
-               u.u_procp->p_flag |= SNOVM;
-               while (rpp->p_flag & SVFORK)
-                       sleep((caddr_t)rpp, PZERO - 1);
-               if ((rpp->p_flag & SLOAD) == 0)
-                       panic("newproc vfork");
-               uaccess(rpp, Vfmap, &vfutl);
-               u.u_procp->p_xlink = 0;
-               vpassvm(rpp, u.u_procp, &vfutl, &u, Vfmap);
-               u.u_procp->p_flag &= ~SNOVM;
-               rpp->p_ndx = rpp - proc;
-               rpp->p_flag |= SVFDONE;
-               wakeup((caddr_t)rpp);
-       }
+       if (isvfork)
+               while (p2->p_flag & SPPWAIT)
+                       tsleep((caddr_t)p1, PWAIT, "ppwait", 0);

 
        /*
 
        /*

-        * 0 return means parent.
+        * Return child pid to parent process,
+        * marking us as parent via retval[1].

         */
         */

+       retval[0] = p2->p_pid;
+       retval[1] = 0;

        return (0);
 }
        return (0);
 }