freebsd-dev/sys/kern/kern_fork.c

551 lines
13 KiB
C
Raw Normal View History

1994-05-24 10:09:53 +00:00
/*
* Copyright (c) 1982, 1986, 1989, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (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.
*
* @(#)kern_fork.c 8.6 (Berkeley) 4/8/94
* $Id: kern_fork.c,v 1.63 1999/07/03 20:58:44 peter Exp $
1994-05-24 10:09:53 +00:00
*/
#include "opt_ktrace.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysproto.h>
1994-05-24 10:09:53 +00:00
#include <sys/filedesc.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/proc.h>
#include <sys/resourcevar.h>
#include <sys/vnode.h>
#include <sys/acct.h>
#include <sys/ktrace.h>
#include <sys/unistd.h>
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
#include <sys/jail.h>
1994-05-24 10:09:53 +00:00
#include <vm/vm.h>
#include <sys/lock.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_extern.h>
VM level code cleanups. 1) Start using TSM. Struct procs continue to point to upages structure, after being freed. Struct vmspace continues to point to pte object and kva space for kstack. u_map is now superfluous. 2) vm_map's don't need to be reference counted. They always exist either in the kernel or in a vmspace. The vmspaces are managed by reference counts. 3) Remove the "wired" vm_map nonsense. 4) No need to keep a cache of kernel stack kva's. 5) Get rid of strange looking ++var, and change to var++. 6) Change more data structures to use our "zone" allocator. Added struct proc, struct vmspace and struct vnode. This saves a significant amount of kva space and physical memory. Additionally, this enables TSM for the zone managed memory. 7) Keep ioopt disabled for now. 8) Remove the now bogus "single use" map concept. 9) Use generation counts or id's for data structures residing in TSM, where it allows us to avoid unneeded restart overhead during traversals, where blocking might occur. 10) Account better for memory deficits, so the pageout daemon will be able to make enough memory available (experimental.) 11) Fix some vnode locking problems. (From Tor, I think.) 12) Add a check in ufs_lookup, to avoid lots of unneeded calls to bcmp. (experimental.) 13) Significantly shrink, cleanup, and make slightly faster the vm_fault.c code. Use generation counts, get rid of unneded collpase operations, and clean up the cluster code. 14) Make vm_zone more suitable for TSM. This commit is partially as a result of discussions and contributions from other people, including DG, Tor Egge, PHK, and probably others that I have forgotten to attribute (so let me know, if I forgot.) This is not the infamous, final cleanup of the vnode stuff, but a necessary step. Vnode mgmt should be correct, but things might still change, and there is still some missing stuff (like ioopt, and physical backing of non-merged cache files, debugging of layering concepts.)
1998-01-22 17:30:44 +00:00
#include <vm/vm_zone.h>
#include <machine/frame.h>
#include <sys/user.h>
static int fast_vfork = 1;
SYSCTL_INT(_kern, OID_AUTO, fast_vfork, CTLFLAG_RW, &fast_vfork, 0, "");
/*
* These are the stuctures used to create a callout list for things to do
* when forking a process
*/
typedef struct fork_list_element {
struct fork_list_element *next;
forklist_fn function;
} *fle_p;
static fle_p fork_list;
#ifndef _SYS_SYSPROTO_H_
struct fork_args {
int dummy;
};
#endif
1994-05-24 10:09:53 +00:00
/* ARGSUSED */
int
fork(p, uap)
1994-05-24 10:09:53 +00:00
struct proc *p;
struct fork_args *uap;
{
int error;
struct proc *p2;
error = fork1(p, RFFDG | RFPROC, &p2);
if (error == 0) {
p->p_retval[0] = p2->p_pid;
p->p_retval[1] = 0;
}
return error;
1994-05-24 10:09:53 +00:00
}
/* ARGSUSED */
int
vfork(p, uap)
1994-05-24 10:09:53 +00:00
struct proc *p;
struct vfork_args *uap;
1994-05-24 10:09:53 +00:00
{
int error;
struct proc *p2;
error = fork1(p, RFFDG | RFPROC | RFPPWAIT | RFMEM, &p2);
if (error == 0) {
p->p_retval[0] = p2->p_pid;
p->p_retval[1] = 0;
}
return error;
}
1994-05-24 10:09:53 +00:00
int
rfork(p, uap)
struct proc *p;
struct rfork_args *uap;
{
int error;
struct proc *p2;
error = fork1(p, uap->flags, &p2);
if (error == 0) {
p->p_retval[0] = p2 ? p2->p_pid : 0;
p->p_retval[1] = 0;
}
return error;
1994-05-24 10:09:53 +00:00
}
1994-05-24 10:09:53 +00:00
int nprocs = 1; /* process 0 */
static int nextpid = 0;
1994-05-24 10:09:53 +00:00
int
fork1(p1, flags, procp)
struct proc *p1;
int flags;
struct proc **procp;
1994-05-24 10:09:53 +00:00
{
struct proc *p2, *pptr;
uid_t uid;
1994-05-24 10:09:53 +00:00
struct proc *newproc;
int count;
static int pidchecked = 0;
fle_p ep ;
ep = fork_list;
if ((flags & (RFFDG|RFCFDG)) == (RFFDG|RFCFDG))
return (EINVAL);
1994-05-24 10:09:53 +00:00
/*
* Here we don't create a new process, but we divorce
* certain parts of a process from itself.
*/
if ((flags & RFPROC) == 0) {
/*
* Divorce the memory, if it is shared, essentially
* this changes shared memory amongst threads, into
* COW locally.
*/
if ((flags & RFMEM) == 0) {
if (p1->p_vmspace->vm_refcnt > 1) {
vmspace_unshare(p1);
}
}
/*
* Close all file descriptors.
*/
if (flags & RFCFDG) {
struct filedesc *fdtmp;
fdtmp = fdinit(p1);
fdfree(p1);
p1->p_fd = fdtmp;
}
/*
* Unshare file descriptors (from parent.)
*/
if (flags & RFFDG) {
if (p1->p_fd->fd_refcnt > 1) {
struct filedesc *newfd;
newfd = fdcopy(p1);
fdfree(p1);
p1->p_fd = newfd;
}
}
*procp = NULL;
return (0);
}
1994-05-24 10:09:53 +00:00
/*
* 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 use the last process; don't let root
* exceed the limit. The variable nprocs is the current number of
* processes, maxproc is the limit.
*/
uid = p1->p_cred->p_ruid;
if ((nprocs >= maxproc - 1 && uid != 0) || nprocs >= maxproc) {
tablefull("proc");
return (EAGAIN);
}
/*
* Increment the nprocs resource before blocking can occur. There
* are hard-limits as to the number of processes that can run.
*/
nprocs++;
1994-05-24 10:09:53 +00:00
/*
* 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 (uid != 0 && count > p1->p_rlimit[RLIMIT_NPROC].rlim_cur) {
(void)chgproccnt(uid, -1);
/*
* Back out the process count
*/
nprocs--;
1994-05-24 10:09:53 +00:00
return (EAGAIN);
}
/* Allocate new proc. */
VM level code cleanups. 1) Start using TSM. Struct procs continue to point to upages structure, after being freed. Struct vmspace continues to point to pte object and kva space for kstack. u_map is now superfluous. 2) vm_map's don't need to be reference counted. They always exist either in the kernel or in a vmspace. The vmspaces are managed by reference counts. 3) Remove the "wired" vm_map nonsense. 4) No need to keep a cache of kernel stack kva's. 5) Get rid of strange looking ++var, and change to var++. 6) Change more data structures to use our "zone" allocator. Added struct proc, struct vmspace and struct vnode. This saves a significant amount of kva space and physical memory. Additionally, this enables TSM for the zone managed memory. 7) Keep ioopt disabled for now. 8) Remove the now bogus "single use" map concept. 9) Use generation counts or id's for data structures residing in TSM, where it allows us to avoid unneeded restart overhead during traversals, where blocking might occur. 10) Account better for memory deficits, so the pageout daemon will be able to make enough memory available (experimental.) 11) Fix some vnode locking problems. (From Tor, I think.) 12) Add a check in ufs_lookup, to avoid lots of unneeded calls to bcmp. (experimental.) 13) Significantly shrink, cleanup, and make slightly faster the vm_fault.c code. Use generation counts, get rid of unneded collpase operations, and clean up the cluster code. 14) Make vm_zone more suitable for TSM. This commit is partially as a result of discussions and contributions from other people, including DG, Tor Egge, PHK, and probably others that I have forgotten to attribute (so let me know, if I forgot.) This is not the infamous, final cleanup of the vnode stuff, but a necessary step. Vnode mgmt should be correct, but things might still change, and there is still some missing stuff (like ioopt, and physical backing of non-merged cache files, debugging of layering concepts.)
1998-01-22 17:30:44 +00:00
newproc = zalloc(proc_zone);
1994-05-24 10:09:53 +00:00
/*
* Setup linkage for kernel based threading
*/
if((flags & RFTHREAD) != 0) {
newproc->p_peers = p1->p_peers;
p1->p_peers = newproc;
newproc->p_leader = p1->p_leader;
} else {
newproc->p_peers = 0;
newproc->p_leader = newproc;
}
newproc->p_wakeup = 0;
newproc->p_vmspace = NULL;
1994-05-24 10:09:53 +00:00
/*
* Find an unused process ID. We remember a range of unused IDs
* ready to use (from nextpid+1 through pidchecked-1).
*/
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) {
int doingzomb = 0;
pidchecked = PID_MAX;
/*
* Scan the active and zombie procs to check whether this pid
* is in use. Remember the lowest pid that's greater
* than nextpid, so we can avoid checking for a while.
*/
p2 = allproc.lh_first;
1994-05-24 10:09:53 +00:00
again:
for (; p2 != 0; p2 = p2->p_list.le_next) {
1994-05-24 10:09:53 +00:00
while (p2->p_pid == nextpid ||
p2->p_pgrp->pg_id == nextpid ||
p2->p_session->s_sid == nextpid) {
1994-05-24 10:09:53 +00:00
nextpid++;
if (nextpid >= pidchecked)
goto retry;
}
if (p2->p_pid > nextpid && pidchecked > p2->p_pid)
pidchecked = p2->p_pid;
1995-05-30 08:16:23 +00:00
if (p2->p_pgrp->pg_id > nextpid &&
1994-05-24 10:09:53 +00:00
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;
1994-05-24 10:09:53 +00:00
}
if (!doingzomb) {
doingzomb = 1;
p2 = zombproc.lh_first;
1994-05-24 10:09:53 +00:00
goto again;
}
}
p2 = newproc;
p2->p_stat = SIDL; /* protect against others */
p2->p_pid = nextpid;
LIST_INSERT_HEAD(&allproc, p2, p_list);
LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
1994-05-24 10:09:53 +00:00
/*
* 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.
*/
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));
p2->p_aioinfo = NULL;
1994-05-24 10:09:53 +00:00
/*
* 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 = P_INMEM;
if (p1->p_flag & P_PROFIL)
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);
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
if (p2->p_prison) {
p2->p_prison->pr_ref++;
p2->p_flag |= P_JAILED;
}
if (flags & RFSIGSHARE) {
p2->p_procsig = p1->p_procsig;
p2->p_procsig->ps_refcnt++;
if (p1->p_sigacts == &p1->p_addr->u_sigacts) {
struct sigacts *newsigacts;
int s;
/* Create the shared sigacts structure */
MALLOC(newsigacts, struct sigacts *,
sizeof(struct sigacts), M_SUBPROC, M_WAITOK);
s = splhigh();
/*
* Set p_sigacts to the new shared structure.
* Note that this is updating p1->p_sigacts at the
* same time, since p_sigacts is just a pointer to
* the shared p_procsig->ps_sigacts.
*/
p2->p_sigacts = newsigacts;
bcopy(&p1->p_addr->u_sigacts, p2->p_sigacts,
sizeof(*p2->p_sigacts));
*p2->p_sigacts = p1->p_addr->u_sigacts;
splx(s);
}
} else {
MALLOC(p2->p_procsig, struct procsig *, sizeof(struct procsig),
M_SUBPROC, M_WAITOK);
bcopy(p1->p_procsig, p2->p_procsig, sizeof(*p2->p_procsig));
p2->p_procsig->ps_refcnt = 1;
p2->p_sigacts = NULL; /* finished in vm_fork() */
}
if (flags & RFLINUXTHPN)
p2->p_sigparent = SIGUSR1;
else
p2->p_sigparent = SIGCHLD;
1994-05-24 10:09:53 +00:00
/* bump references to the text vnode (for procfs) */
p2->p_textvp = p1->p_textvp;
if (p2->p_textvp)
VREF(p2->p_textvp);
if (flags & RFCFDG)
p2->p_fd = fdinit(p1);
else if (flags & RFFDG)
p2->p_fd = fdcopy(p1);
else
p2->p_fd = fdshare(p1);
1994-05-24 10:09:53 +00:00
/*
* 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++;
}
/*
* Preserve some more flags in subprocess. P_PROFIL has already
* been preserved.
*/
p2->p_flag |= p1->p_flag & P_SUGID;
1994-05-24 10:09:53 +00:00
if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
p2->p_flag |= P_CONTROLT;
if (flags & RFPPWAIT)
1994-05-24 10:09:53 +00:00
p2->p_flag |= P_PPWAIT;
LIST_INSERT_AFTER(p1, p2, p_pglist);
/*
* Attach the new process to its parent.
*
* If RFNOWAIT is set, the newly created process becomes a child
* of init. This effectively disassociates the child from the
* parent.
*/
if (flags & RFNOWAIT)
pptr = initproc;
else
pptr = p1;
p2->p_pptr = pptr;
LIST_INSERT_HEAD(&pptr->p_children, p2, p_sibling);
LIST_INIT(&p2->p_children);
1994-05-24 10:09:53 +00:00
#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)
VREF(p2->p_tracep);
}
#endif
/*
* set priority of child to be that of parent
*/
p2->p_estcpu = p1->p_estcpu;
1994-05-24 10:09:53 +00:00
/*
* This begins the section where we must prevent the parent
* from being swapped.
*/
PHOLD(p1);
/*
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
* Finish creating the child process. It will return via a different
* execution path later. (ie: directly into user mode)
*/
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
vm_fork(p1, p2, flags);
1994-05-24 10:09:53 +00:00
/*
* Both processes are set up, now check if any loadable modules want
* to adjust anything.
* What if they have an error? XXX
*/
while (ep) {
(*ep->function)(p1, p2, flags);
ep = ep->next;
}
1994-05-24 10:09:53 +00:00
/*
* Make child runnable and add to run queue.
*/
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
microtime(&(p2->p_stats->p_start));
p2->p_acflag = AFORK;
1994-05-24 10:09:53 +00:00
(void) splhigh();
p2->p_stat = SRUN;
setrunqueue(p2);
(void) spl0();
/*
* Now can be swapped.
*/
PRELE(p1);
1994-05-24 10:09:53 +00:00
/*
* 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).
*/
while (p2->p_flag & P_PPWAIT)
tsleep(p1, PWAIT, "ppwait", 0);
1994-05-24 10:09:53 +00:00
/*
* Return child proc pointer to parent.
1994-05-24 10:09:53 +00:00
*/
*procp = p2;
1994-05-24 10:09:53 +00:00
return (0);
}
/*
* The next two functionms are general routines to handle adding/deleting
* items on the fork callout list.
*
* at_fork():
* Take the arguments given and put them onto the fork callout list,
* However first make sure that it's not already there.
* Returns 0 on success or a standard error number.
*/
int
1997-08-26 00:15:04 +00:00
at_fork(function)
forklist_fn function;
{
fle_p ep;
/* let the programmer know if he's been stupid */
if (rm_at_fork(function))
printf("fork callout entry already present\n");
ep = malloc(sizeof(*ep), M_TEMP, M_NOWAIT);
if (ep == NULL)
return (ENOMEM);
ep->next = fork_list;
ep->function = function;
fork_list = ep;
return (0);
}
/*
* Scan the exit callout list for the given items and remove them.
* Returns the number of items removed.
* Theoretically this value can only be 0 or 1.
*/
int
1997-08-26 00:15:04 +00:00
rm_at_fork(function)
forklist_fn function;
{
fle_p *epp, ep;
int count;
count= 0;
epp = &fork_list;
ep = *epp;
while (ep) {
if (ep->function == function) {
*epp = ep->next;
free(ep, M_TEMP);
count++;
} else {
epp = &ep->next;
}
ep = *epp;
}
return (count);
}