NetBSD/dist/am-utils/amd/autil.c

784 lines
22 KiB
C

/* $NetBSD: autil.c,v 1.9 2006/03/18 21:10:12 christos Exp $ */
/*
* Copyright (c) 1997-2005 Erez Zadok
* Copyright (c) 1990 Jan-Simon Pendry
* Copyright (c) 1990 Imperial College of Science, Technology & Medicine
* Copyright (c) 1990 The Regents of the University of California.
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Jan-Simon Pendry at Imperial College, London.
*
* 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 acknowledgment:
* 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.
*
*
* File: am-utils/amd/autil.c
*
*/
/*
* utilities specified to amd, taken out of the older amd/util.c.
*/
#ifdef HAVE_CONFIG_H
# include <config.h>
#endif /* HAVE_CONFIG_H */
#include <am_defs.h>
#include <amd.h>
int NumChildren = 0; /* number of children of primary amd */
static char invalid_keys[] = "\"'!;@ \t\n";
/****************************************************************************
*** MACROS ***
****************************************************************************/
#ifdef HAVE_TRANSPORT_TYPE_TLI
# define PARENT_USLEEP_TIME 100000 /* 0.1 seconds */
#endif /* HAVE_TRANSPORT_TYPE_TLI */
/****************************************************************************
*** FORWARD DEFINITIONS ***
****************************************************************************/
static void domain_strip(char *otherdom, char *localdom);
static int dofork(void);
/****************************************************************************
*** FUNCTIONS ***
****************************************************************************/
/*
* Copy s into p, reallocating p if necessary
*/
char *
strealloc(char *p, char *s)
{
size_t len = strlen(s) + 1;
p = (char *) xrealloc((voidp) p, len);
xstrlcpy(p, s, len);
#ifdef DEBUG_MEM
# if defined(HAVE_MALLINFO) && defined(HAVE_MALLOC_VERIFY)
malloc_verify();
# endif /* not defined(HAVE_MALLINFO) && defined(HAVE_MALLOC_VERIFY) */
#endif /* DEBUG_MEM */
return p;
}
/*
* Strip off the trailing part of a domain
* to produce a short-form domain relative
* to the local host domain.
* Note that this has no effect if the domain
* names do not have the same number of
* components. If that restriction proves
* to be a problem then the loop needs recoding
* to skip from right to left and do partial
* matches along the way -- ie more expensive.
*/
static void
domain_strip(char *otherdom, char *localdom)
{
char *p1, *p2;
if ((p1 = strchr(otherdom, '.')) &&
(p2 = strchr(localdom, '.')) &&
STREQ(p1 + 1, p2 + 1))
*p1 = '\0';
}
/*
* Normalize a host name: replace cnames with real names, and decide if to
* strip domain name or not.
*/
void
host_normalize(char **chp)
{
/*
* Normalize hosts is used to resolve host name aliases
* and replace them with the standard-form name.
* Invoked with "-n" command line option.
*/
if (gopt.flags & CFM_NORMALIZE_HOSTNAMES) {
struct hostent *hp;
hp = gethostbyname(*chp);
if (hp && hp->h_addrtype == AF_INET) {
dlog("Hostname %s normalized to %s", *chp, hp->h_name);
*chp = strealloc(*chp, (char *) hp->h_name);
}
}
if (gopt.flags & CFM_DOMAIN_STRIP) {
domain_strip(*chp, hostd);
}
}
/*
* Keys are not allowed to contain " ' ! or ; to avoid
* problems with macro expansions.
*/
int
valid_key(char *key)
{
while (*key)
if (strchr(invalid_keys, *key++))
return FALSE;
return TRUE;
}
void
forcibly_timeout_mp(am_node *mp)
{
mntfs *mf = mp->am_mnt;
/*
* Arrange to timeout this node
*/
if (mf && ((mp->am_flags & AMF_ROOT) ||
(mf->mf_flags & (MFF_MOUNTING | MFF_UNMOUNTING)))) {
if (mf->mf_flags & MFF_UNMOUNTING)
plog(XLOG_WARNING, "node %s is currently being unmounted, ignoring timeout request", mp->am_path);
else
plog(XLOG_WARNING, "ignoring timeout request for active node %s", mp->am_path);
} else {
plog(XLOG_INFO, "\"%s\" forcibly timed out", mp->am_path);
mp->am_flags &= ~AMF_NOTIMEOUT;
mp->am_ttl = clocktime(NULL);
/*
* Force mtime update of parent dir, to prevent DNLC/dcache from caching
* the old entry, which could result in ESTALE errors, bad symlinks, and
* more.
*/
clocktime(&mp->am_parent->am_fattr.na_mtime);
reschedule_timeout_mp();
}
}
void
mf_mounted(mntfs *mf, bool_t call_free_opts)
{
int quoted;
int wasmounted = mf->mf_flags & MFF_MOUNTED;
if (!wasmounted) {
/*
* If this is a freshly mounted
* filesystem then update the
* mntfs structure...
*/
mf->mf_flags |= MFF_MOUNTED;
mf->mf_error = 0;
/*
* Do mounted callback
*/
if (mf->mf_ops->mounted)
mf->mf_ops->mounted(mf);
/*
* Be careful when calling free_ops and XFREE here. Some pseudo file
* systems like nfsx call this function (mf_mounted), even though it
* would be called by the lower-level amd file system functions. nfsx
* needs to call this function because of the other actions it takes.
* So we pass a boolean from the caller (yes, not so clean workaround)
* to determine if we should free or not. If we're not freeing (often
* because we're called from a callback function), then just to be sure,
* we'll zero out the am_opts structure and set the pointer to NULL.
* The parent mntfs node owns this memory and is going to free it with a
* call to mf_mounted(mntfs,TRUE) (see comment in the am_mounted code).
*/
if (call_free_opts) {
free_opts(mf->mf_fo); /* this free is needed to prevent leaks */
XFREE(mf->mf_fo); /* (also this one) */
} else {
memset(mf->mf_fo, 0, sizeof(am_opts));
mf->mf_fo = NULL;
}
}
if (mf->mf_flags & MFF_RESTART) {
mf->mf_flags &= ~MFF_RESTART;
dlog("Restarted filesystem %s, flags 0x%x", mf->mf_mount, mf->mf_flags);
}
/*
* Log message
*/
quoted = strchr(mf->mf_info, ' ') != 0;
plog(XLOG_INFO, "%s%s%s %s fstype %s on %s",
quoted ? "\"" : "",
mf->mf_info,
quoted ? "\"" : "",
wasmounted ? "referenced" : "mounted",
mf->mf_ops->fs_type, mf->mf_mount);
}
void
am_mounted(am_node *mp)
{
int notimeout = 0; /* assume normal timeouts initially */
mntfs *mf = mp->am_mnt;
/*
* This is the parent mntfs which does the mf->mf_fo (am_opts type), and
* we're passing TRUE here to tell mf_mounted to actually free the
* am_opts. See a related comment in mf_mounted().
*/
mf_mounted(mf, TRUE);
#ifdef HAVE_FS_AUTOFS
if (mf->mf_flags & MFF_IS_AUTOFS)
autofs_mounted(mp);
#endif /* HAVE_FS_AUTOFS */
/*
* Patch up path for direct mounts
*/
if (mp->am_parent && mp->am_parent->am_mnt->mf_fsflags & FS_DIRECT)
mp->am_path = str3cat(mp->am_path, mp->am_parent->am_path, "/", ".");
/*
* Check whether this mount should be cached permanently or not,
* and handle user-requested timeouts.
*/
/* first check if file system was set to never timeout */
if (mf->mf_fsflags & FS_NOTIMEOUT)
notimeout = 1;
/* next, alter that decision by map flags */
if (mf->mf_mopts) {
mntent_t mnt;
mnt.mnt_opts = mf->mf_mopts;
/* umount option: user wants to unmount this entry */
if (amu_hasmntopt(&mnt, "unmount") || amu_hasmntopt(&mnt, "umount"))
notimeout = 0;
/* noumount option: user does NOT want to unmount this entry */
if (amu_hasmntopt(&mnt, "nounmount") || amu_hasmntopt(&mnt, "noumount"))
notimeout = 1;
/* utimeout=N option: user wants to unmount this option AND set timeout */
if ((mp->am_timeo = hasmntval(&mnt, "utimeout")) == 0)
mp->am_timeo = gopt.am_timeo; /* otherwise use default timeout */
else
notimeout = 0;
/* special case: don't try to unmount "/" (it can never succeed) */
if (mf->mf_mount[0] == '/' && mf->mf_mount[1] == '\0')
notimeout = 1;
}
/* finally set actual flags */
if (notimeout) {
mp->am_flags |= AMF_NOTIMEOUT;
plog(XLOG_INFO, "%s set to never timeout", mp->am_path);
} else {
mp->am_flags &= ~AMF_NOTIMEOUT;
plog(XLOG_INFO, "%s set to timeout in %d seconds", mp->am_path, mp->am_timeo);
}
/*
* If this node is a symlink then
* compute the length of the returned string.
*/
if (mp->am_fattr.na_type == NFLNK)
mp->am_fattr.na_size = strlen(mp->am_link ? mp->am_link : mf->mf_mount);
/*
* Record mount time, and update am_stats at the same time.
*/
mp->am_stats.s_mtime = clocktime(&mp->am_fattr.na_mtime);
new_ttl(mp);
/*
* Update mtime of parent node (copying "struct nfstime" in '=' below)
*/
if (mp->am_parent && mp->am_parent->am_mnt)
mp->am_parent->am_fattr.na_mtime = mp->am_fattr.na_mtime;
/*
* This is ugly, but essentially unavoidable
* Sublinks must be treated separately as type==link
* when the base type is different.
*/
if (mp->am_link && mf->mf_ops != &amfs_link_ops)
amfs_link_ops.mount_fs(mp, mf);
/*
* Now, if we can, do a reply to our client here
* to speed things up.
*/
#ifdef HAVE_FS_AUTOFS
if (mp->am_flags & AMF_AUTOFS)
autofs_mount_succeeded(mp);
else
#endif /* HAVE_FS_AUTOFS */
nfs_quick_reply(mp, 0);
/*
* Update stats
*/
amd_stats.d_mok++;
}
/*
* Replace mount point with a reference to an error filesystem.
* The mount point (struct mntfs) is NOT discarded,
* the caller must do it if it wants to _before_ calling this function.
*/
void
assign_error_mntfs(am_node *mp)
{
int error;
dlog("assign_error_mntfs");
/*
* Save the old error code
*/
error = mp->am_error;
if (error <= 0)
error = mp->am_mnt->mf_error;
/*
* Allocate a new error reference
*/
mp->am_mnt = new_mntfs();
/*
* Put back the error code
*/
mp->am_mnt->mf_error = error;
mp->am_mnt->mf_flags |= MFF_ERROR;
/*
* Zero the error in the mount point
*/
mp->am_error = 0;
}
/*
* Build a new map cache for this node, or re-use
* an existing cache for the same map.
*/
void
amfs_mkcacheref(mntfs *mf)
{
char *cache;
if (mf->mf_fo && mf->mf_fo->opt_cache)
cache = mf->mf_fo->opt_cache;
else
cache = "none";
mf->mf_private = (opaque_t) mapc_find(mf->mf_info, cache,
mf->mf_fo ? mf->mf_fo->opt_maptype : NULL);
mf->mf_prfree = mapc_free;
}
/*
* Locate next node in sibling list which is mounted
* and is not an error node.
*/
am_node *
next_nonerror_node(am_node *xp)
{
mntfs *mf;
/*
* Bug report (7/12/89) from Rein Tollevik <rein@ifi.uio.no>
* Fixes a race condition when mounting direct automounts.
* Also fixes a problem when doing a readdir on a directory
* containing hung automounts.
*/
while (xp &&
(!(mf = xp->am_mnt) || /* No mounted filesystem */
mf->mf_error != 0 || /* There was a mntfs error */
xp->am_error != 0 || /* There was a mount error */
!(mf->mf_flags & MFF_MOUNTED) || /* The fs is not mounted */
(mf->mf_server->fs_flags & FSF_DOWN)) /* The fs may be down */
)
xp = xp->am_osib;
return xp;
}
/*
* Mount an automounter directory.
* The automounter is connected into the system
* as a user-level NFS server. amfs_mount constructs
* the necessary NFS parameters to be given to the
* kernel so that it will talk back to us.
*
* NOTE: automounter mounts in themselves are using NFS Version 2 (UDP).
*
* NEW: on certain systems, mounting can be done using the
* kernel-level automount (autofs) support. In that case,
* we don't need NFS at all here.
*/
int
amfs_mount(am_node *mp, mntfs *mf, char *opts)
{
char fs_hostname[MAXHOSTNAMELEN + MAXPATHLEN + 1];
int retry, error = 0, genflags;
int on_autofs = mf->mf_flags & MFF_ON_AUTOFS;
char *dir = mf->mf_mount;
mntent_t mnt;
MTYPE_TYPE type;
int forced_unmount = 0; /* are we using forced unmounts? */
memset((voidp) &mnt, 0, sizeof(mnt));
mnt.mnt_dir = dir;
mnt.mnt_fsname = pid_fsname;
mnt.mnt_opts = opts;
#ifdef HAVE_FS_AUTOFS
if (mf->mf_flags & MFF_IS_AUTOFS) {
type = MOUNT_TYPE_AUTOFS;
/*
* Make sure that amd's top-level autofs mounts are hidden by default
* from df.
* XXX: It works ok on Linux, might not work on other systems.
*/
mnt.mnt_type = "autofs";
} else
#endif /* HAVE_FS_AUTOFS */
{
type = MOUNT_TYPE_NFS;
/*
* Make sure that amd's top-level NFS mounts are hidden by default
* from df.
* If they don't appear to support the either the "ignore" mnttab
* option entry, or the "auto" one, set the mount type to "nfs".
*/
mnt.mnt_type = HIDE_MOUNT_TYPE;
}
retry = hasmntval(&mnt, MNTTAB_OPT_RETRY);
if (retry <= 0)
retry = 2; /* XXX: default to 2 retries */
/*
* SET MOUNT ARGS
*/
/*
* Make a ``hostname'' string for the kernel
*/
xsnprintf(fs_hostname, sizeof(fs_hostname), "pid%ld@%s:%s",
get_server_pid(), am_get_hostname(), dir);
/*
* Most kernels have a name length restriction (64 bytes)...
*/
if (strlen(fs_hostname) >= MAXHOSTNAMELEN)
xstrlcpy(fs_hostname + MAXHOSTNAMELEN - 3, "..",
sizeof(fs_hostname) - MAXHOSTNAMELEN + 3);
#ifdef HOSTNAMESZ
/*
* ... and some of these restrictions are 32 bytes (HOSTNAMESZ)
* If you need to get the definition for HOSTNAMESZ found, you may
* add the proper header file to the conf/nfs_prot/nfs_prot_*.h file.
*/
if (strlen(fs_hostname) >= HOSTNAMESZ)
xstrlcpy(fs_hostname + HOSTNAMESZ - 3, "..",
sizeof(fs_hostname) - HOSTNAMESZ + 3);
#endif /* HOSTNAMESZ */
/*
* Finally we can compute the mount genflags set above,
* and add any automounter specific flags.
*/
genflags = compute_mount_flags(&mnt);
#ifdef HAVE_FS_AUTOFS
if (on_autofs)
genflags |= autofs_compute_mount_flags(&mnt);
#endif /* HAVE_FS_AUTOFS */
genflags |= compute_automounter_mount_flags(&mnt);
again:
if (!(mf->mf_flags & MFF_IS_AUTOFS)) {
nfs_args_t nfs_args;
am_nfs_fh *fhp;
am_nfs_handle_t anh;
#ifndef HAVE_TRANSPORT_TYPE_TLI
u_short port;
struct sockaddr_in sin;
#endif /* not HAVE_TRANSPORT_TYPE_TLI */
/*
* get fhandle of remote path for automount point
*/
fhp = get_root_nfs_fh(dir);
if (!fhp) {
plog(XLOG_FATAL, "Can't find root file handle for %s", dir);
return EINVAL;
}
#ifndef HAVE_TRANSPORT_TYPE_TLI
/*
* Create sockaddr to point to the local machine.
*/
memset((voidp) &sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_addr = myipaddr;
port = hasmntval(&mnt, MNTTAB_OPT_PORT);
if (port) {
sin.sin_port = htons(port);
} else {
plog(XLOG_ERROR, "no port number specified for %s", dir);
return EINVAL;
}
#endif /* not HAVE_TRANSPORT_TYPE_TLI */
/* setup the many fields and flags within nfs_args */
memmove(&anh.v2, fhp, sizeof(*fhp));
#ifdef HAVE_TRANSPORT_TYPE_TLI
compute_nfs_args(&nfs_args,
&mnt,
genflags,
nfsncp,
NULL, /* remote host IP addr is set below */
NFS_VERSION, /* version 2 */
"udp",
&anh,
fs_hostname,
pid_fsname);
/*
* IMPORTANT: set the correct IP address AFTERWARDS. It cannot
* be done using the normal mechanism of compute_nfs_args(), because
* that one will allocate a new address and use NFS_SA_DREF() to copy
* parts to it, while assuming that the ip_addr passed is always
* a "struct sockaddr_in". That assumption is incorrect on TLI systems,
* because they define a special macro HOST_SELF which is DIFFERENT
* than localhost (127.0.0.1)!
*/
nfs_args.addr = &nfsxprt->xp_ltaddr;
#else /* not HAVE_TRANSPORT_TYPE_TLI */
compute_nfs_args(&nfs_args,
&mnt,
genflags,
NULL,
&sin,
NFS_VERSION, /* version 2 */
"udp",
&anh,
fs_hostname,
pid_fsname);
#endif /* not HAVE_TRANSPORT_TYPE_TLI */
/*************************************************************************
* NOTE: while compute_nfs_args() works ok for regular NFS mounts *
* the toplvl one is not quite regular, and so some options must be *
* corrected by hand more carefully, *after* compute_nfs_args() runs. *
*************************************************************************/
compute_automounter_nfs_args(&nfs_args, &mnt);
if (amuDebug(D_TRACE)) {
print_nfs_args(&nfs_args, 0);
plog(XLOG_DEBUG, "Generic mount flags 0x%x", genflags);
}
/* This is it! Here we try to mount amd on its mount points */
error = mount_fs(&mnt, genflags, (caddr_t) &nfs_args,
retry, type, 0, NULL, mnttab_file_name, on_autofs);
#ifdef HAVE_TRANSPORT_TYPE_TLI
free_knetconfig(nfs_args.knconf);
/*
* local automounter mounts do not allocate a special address, so
* no need to XFREE(nfs_args.addr) under TLI.
*/
#endif /* HAVE_TRANSPORT_TYPE_TLI */
#ifdef HAVE_FS_AUTOFS
} else {
/* This is it! Here we try to mount amd on its mount points */
error = mount_fs(&mnt, genflags, (caddr_t) mp->am_autofs_fh,
retry, type, 0, NULL, mnttab_file_name, on_autofs);
#endif /* HAVE_FS_AUTOFS */
}
if (error == 0 || forced_unmount)
return error;
/*
* If user wants forced/lazy unmount semantics, then try it iff the
* current mount failed with EIO or ESTALE.
*/
if (gopt.flags & CFM_FORCED_UNMOUNTS) {
switch (errno) {
case ESTALE:
case EIO:
forced_unmount = errno;
plog(XLOG_WARNING, "Mount %s failed (%m); force unmount.", mp->am_path);
if ((error = UMOUNT_FS(mp->am_path, mnttab_file_name,
AMU_UMOUNT_FORCE | AMU_UMOUNT_DETACH)) < 0) {
plog(XLOG_WARNING, "Forced umount %s failed: %m.", mp->am_path);
errno = forced_unmount;
} else
goto again;
default:
break;
}
}
return error;
}
void
am_unmounted(am_node *mp)
{
mntfs *mf = mp->am_mnt;
if (!foreground) /* firewall - should never happen */
return;
/*
* Do unmounted callback
*/
if (mf->mf_ops->umounted)
mf->mf_ops->umounted(mf);
/*
* This is ugly, but essentially unavoidable.
* Sublinks must be treated separately as type==link
* when the base type is different.
*/
if (mp->am_link && mf->mf_ops != &amfs_link_ops)
amfs_link_ops.umount_fs(mp, mf);
#ifdef HAVE_FS_AUTOFS
if (mf->mf_flags & MFF_IS_AUTOFS)
autofs_release_fh(mp);
if (mp->am_flags & AMF_AUTOFS)
autofs_umount_succeeded(mp);
#endif /* HAVE_FS_AUTOFS */
/*
* Clean up any directories that were made
*
* If we remove the mount point of a pending mount, any queued access
* to it will fail. So don't do it in that case.
* Also don't do it if the refcount is > 1.
*/
if (mf->mf_flags & MFF_MKMNT &&
mf->mf_refc == 1 &&
!(mp->am_flags & AMF_REMOUNT)) {
plog(XLOG_INFO, "removing mountpoint directory '%s'", mf->mf_mount);
rmdirs(mf->mf_mount);
mf->mf_flags &= ~MFF_MKMNT;
}
/*
* If this is a pseudo-directory then adjust the link count
* in the parent
*/
if (mp->am_parent && mp->am_fattr.na_type == NFDIR)
--mp->am_parent->am_fattr.na_nlink;
/*
* Update mtime of parent node
*/
if (mp->am_parent && mp->am_parent->am_mnt)
clocktime(&mp->am_parent->am_fattr.na_mtime);
if (mp->am_parent && (mp->am_flags & AMF_REMOUNT)) {
char *fname = strdup(mp->am_name);
am_node *mp_parent = mp->am_parent;
mntfs *mf_parent = mp_parent->am_mnt;
int error = 0;
free_map(mp);
plog(XLOG_INFO, "am_unmounted: remounting %s", fname);
mp = mf_parent->mf_ops->lookup_child(mp_parent, fname, &error, VLOOK_CREATE);
if (mp && error < 0)
mp = mf_parent->mf_ops->mount_child(mp, &error);
if (error > 0) {
errno = error;
plog(XLOG_ERROR, "am_unmounted: could not remount %s: %m", fname);
}
XFREE(fname);
} else
/*
* We have a race here.
* If this node has a pending mount and amd is going down (unmounting
* everything in the process), then we could potentially free it here
* while a struct continuation still has a reference to it. So when
* amfs_cont is called, it blows up.
* We avoid the race by refusing to free any nodes that have
* pending mounts (defined as having a non-NULL am_mfarray).
*/
if (!mp->am_mfarray)
free_map(mp);
}
/*
* Fork the automounter
*
* TODO: Need a better strategy for handling errors
*/
static int
dofork(void)
{
int pid;
top:
pid = fork();
if (pid < 0) { /* fork error, retry in 1 second */
sleep(1);
goto top;
}
if (pid == 0) { /* child process (foreground==false) */
am_set_mypid();
foreground = 0;
} else { /* parent process, has one more child */
NumChildren++;
}
return pid;
}
int
background(void)
{
int pid = dofork();
if (pid == 0) {
dlog("backgrounded");
foreground = 0;
} else
dlog("forked process %d", pid);
return pid;
}