1683 lines
37 KiB
C
1683 lines
37 KiB
C
/* $NetBSD: svr4_misc.c,v 1.61 1998/08/02 04:41:33 thorpej Exp $ */
|
|
|
|
/*
|
|
* Copyright (c) 1994 Christos Zoulas
|
|
* All rights reserved.
|
|
*
|
|
* 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. The name of the author may not be used to endorse or promote products
|
|
* derived from this software without specific prior written permission
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
|
|
*/
|
|
|
|
/*
|
|
* SVR4 compatibility module.
|
|
*
|
|
* SVR4 system calls that are implemented differently in BSD are
|
|
* handled here.
|
|
*/
|
|
|
|
#include "opt_uvm.h"
|
|
|
|
#include <sys/param.h>
|
|
#include <sys/systm.h>
|
|
#include <sys/namei.h>
|
|
#include <sys/dirent.h>
|
|
#include <sys/proc.h>
|
|
#include <sys/file.h>
|
|
#include <sys/stat.h>
|
|
#include <sys/time.h>
|
|
#include <sys/filedesc.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/kernel.h>
|
|
#include <sys/malloc.h>
|
|
#include <sys/pool.h>
|
|
#include <sys/mbuf.h>
|
|
#include <sys/mman.h>
|
|
#include <sys/mount.h>
|
|
#include <sys/resource.h>
|
|
#include <sys/resourcevar.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/vnode.h>
|
|
#include <sys/uio.h>
|
|
#include <sys/wait.h>
|
|
#include <sys/utsname.h>
|
|
#include <sys/unistd.h>
|
|
#include <sys/times.h>
|
|
#include <sys/sem.h>
|
|
#include <sys/msg.h>
|
|
#include <sys/ptrace.h>
|
|
#include <sys/signalvar.h>
|
|
|
|
#include <netinet/in.h>
|
|
#include <sys/syscallargs.h>
|
|
|
|
#include <miscfs/specfs/specdev.h>
|
|
|
|
#include <compat/svr4/svr4_types.h>
|
|
#include <compat/svr4/svr4_signal.h>
|
|
#include <compat/svr4/svr4_syscallargs.h>
|
|
#include <compat/svr4/svr4_util.h>
|
|
#include <compat/svr4/svr4_time.h>
|
|
#include <compat/svr4/svr4_dirent.h>
|
|
#include <compat/svr4/svr4_ulimit.h>
|
|
#include <compat/svr4/svr4_hrt.h>
|
|
#include <compat/svr4/svr4_wait.h>
|
|
#include <compat/svr4/svr4_statvfs.h>
|
|
#include <compat/svr4/svr4_sysconfig.h>
|
|
#include <compat/svr4/svr4_acl.h>
|
|
#include <compat/svr4/svr4_mman.h>
|
|
|
|
#include <vm/vm.h>
|
|
#if defined(UVM)
|
|
#include <uvm/uvm_extern.h>
|
|
#endif
|
|
|
|
static __inline clock_t timeval_to_clock_t __P((struct timeval *));
|
|
static int svr4_setinfo __P((struct proc *, int, svr4_siginfo_t *));
|
|
|
|
struct svr4_hrtcntl_args;
|
|
static int svr4_hrtcntl __P((struct proc *, struct svr4_hrtcntl_args *,
|
|
register_t *));
|
|
static void bsd_statfs_to_svr4_statvfs __P((const struct statfs *,
|
|
struct svr4_statvfs *));
|
|
static void bsd_statfs_to_svr4_statvfs64 __P((const struct statfs *,
|
|
struct svr4_statvfs64 *));
|
|
static struct proc *svr4_pfind __P((pid_t pid));
|
|
|
|
static int svr4_mknod __P((struct proc *, register_t *, char *,
|
|
svr4_mode_t, svr4_dev_t));
|
|
|
|
int
|
|
svr4_sys_wait(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_wait_args *uap = v;
|
|
struct sys_wait4_args w4;
|
|
int error;
|
|
size_t sz = sizeof(*SCARG(&w4, status));
|
|
|
|
SCARG(&w4, rusage) = NULL;
|
|
SCARG(&w4, options) = 0;
|
|
|
|
if (SCARG(uap, status) == NULL) {
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
SCARG(&w4, status) = stackgap_alloc(&sg, sz);
|
|
}
|
|
else
|
|
SCARG(&w4, status) = SCARG(uap, status);
|
|
|
|
SCARG(&w4, pid) = WAIT_ANY;
|
|
|
|
if ((error = sys_wait4(p, &w4, retval)) != 0)
|
|
return error;
|
|
|
|
/*
|
|
* It looks like wait(2) on svr4/solaris/2.4 returns
|
|
* the status in retval[1], and the pid on retval[0].
|
|
* NB: this can break if register_t stops being an int.
|
|
*/
|
|
return copyin(SCARG(&w4, status), &retval[1], sz);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_execv(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_execv_args /* {
|
|
syscallarg(char *) path;
|
|
syscallarg(char **) argv;
|
|
} */ *uap = v;
|
|
struct sys_execve_args ap;
|
|
caddr_t sg;
|
|
|
|
sg = stackgap_init(p->p_emul);
|
|
SVR4_CHECK_ALT_EXIST(p, &sg, SCARG(uap, path));
|
|
|
|
SCARG(&ap, path) = SCARG(uap, path);
|
|
SCARG(&ap, argp) = SCARG(uap, argp);
|
|
SCARG(&ap, envp) = NULL;
|
|
|
|
return sys_execve(p, &ap, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_execve(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_execve_args /* {
|
|
syscallarg(char *) path;
|
|
syscallarg(char **) argv;
|
|
syscallarg(char **) envp;
|
|
} */ *uap = v;
|
|
struct sys_execve_args ap;
|
|
caddr_t sg;
|
|
|
|
sg = stackgap_init(p->p_emul);
|
|
SVR4_CHECK_ALT_EXIST(p, &sg, SCARG(uap, path));
|
|
|
|
SCARG(&ap, path) = SCARG(uap, path);
|
|
SCARG(&ap, argp) = SCARG(uap, argp);
|
|
SCARG(&ap, envp) = SCARG(uap, envp);
|
|
|
|
return sys_execve(p, &ap, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_time(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_time_args *uap = v;
|
|
int error = 0;
|
|
struct timeval tv;
|
|
|
|
microtime(&tv);
|
|
if (SCARG(uap, t))
|
|
error = copyout(&tv.tv_sec, SCARG(uap, t),
|
|
sizeof(*(SCARG(uap, t))));
|
|
*retval = (int) tv.tv_sec;
|
|
|
|
return error;
|
|
}
|
|
|
|
|
|
/*
|
|
* Read SVR4-style directory entries. We suck them into kernel space so
|
|
* that they can be massaged before being copied out to user code. Like
|
|
* SunOS, we squish out `empty' entries.
|
|
*
|
|
* This is quite ugly, but what do you expect from compatibility code?
|
|
*/
|
|
int
|
|
svr4_sys_getdents64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_getdents64_args *uap = v;
|
|
register struct dirent *bdp;
|
|
struct vnode *vp;
|
|
caddr_t inp, buf; /* BSD-format */
|
|
int len, reclen; /* BSD-format */
|
|
caddr_t outp; /* SVR4-format */
|
|
int resid, svr4_reclen; /* SVR4-format */
|
|
struct file *fp;
|
|
struct uio auio;
|
|
struct iovec aiov;
|
|
struct svr4_dirent64 idb;
|
|
off_t off; /* true file offset */
|
|
int buflen, error, eofflag;
|
|
off_t *cookiebuf = NULL, *cookie;
|
|
int ncookies;
|
|
|
|
if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
|
|
return (error);
|
|
|
|
if ((fp->f_flag & FREAD) == 0)
|
|
return (EBADF);
|
|
|
|
vp = (struct vnode *)fp->f_data;
|
|
|
|
buflen = min(MAXBSIZE, SCARG(uap, nbytes));
|
|
buf = malloc(buflen, M_TEMP, M_WAITOK);
|
|
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
|
|
off = fp->f_offset;
|
|
again:
|
|
aiov.iov_base = buf;
|
|
aiov.iov_len = buflen;
|
|
auio.uio_iov = &aiov;
|
|
auio.uio_iovcnt = 1;
|
|
auio.uio_rw = UIO_READ;
|
|
auio.uio_segflg = UIO_SYSSPACE;
|
|
auio.uio_procp = p;
|
|
auio.uio_resid = buflen;
|
|
auio.uio_offset = off;
|
|
/*
|
|
* First we read into the malloc'ed buffer, then
|
|
* we massage it into user space, one record at a time.
|
|
*/
|
|
error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, &cookiebuf,
|
|
&ncookies);
|
|
if (error)
|
|
goto out;
|
|
|
|
inp = buf;
|
|
outp = (char *) SCARG(uap, dp);
|
|
resid = SCARG(uap, nbytes);
|
|
if ((len = buflen - auio.uio_resid) == 0)
|
|
goto eof;
|
|
|
|
for (cookie = cookiebuf; len > 0; len -= reclen) {
|
|
bdp = (struct dirent *)inp;
|
|
reclen = bdp->d_reclen;
|
|
if (reclen & 3)
|
|
panic("svr4_getdents64: bad reclen");
|
|
if (bdp->d_fileno == 0) {
|
|
inp += reclen; /* it is a hole; squish it out */
|
|
off = *cookie++;
|
|
continue;
|
|
}
|
|
svr4_reclen = SVR4_RECLEN(&idb, bdp->d_namlen);
|
|
if (reclen > len || resid < svr4_reclen) {
|
|
/* entry too big for buffer, so just stop */
|
|
outp++;
|
|
break;
|
|
}
|
|
off = *cookie++; /* each entry points to the next */
|
|
/*
|
|
* Massage in place to make a SVR4-shaped dirent (otherwise
|
|
* we have to worry about touching user memory outside of
|
|
* the copyout() call).
|
|
*/
|
|
idb.d_ino = (svr4_ino64_t)bdp->d_fileno;
|
|
idb.d_off = (svr4_off64_t)off;
|
|
idb.d_reclen = (u_short)svr4_reclen;
|
|
strcpy(idb.d_name, bdp->d_name);
|
|
if ((error = copyout((caddr_t)&idb, outp, svr4_reclen)))
|
|
goto out;
|
|
/* advance past this real entry */
|
|
inp += reclen;
|
|
/* advance output past SVR4-shaped entry */
|
|
outp += svr4_reclen;
|
|
resid -= svr4_reclen;
|
|
}
|
|
|
|
/* if we squished out the whole block, try again */
|
|
if (outp == (char *) SCARG(uap, dp))
|
|
goto again;
|
|
fp->f_offset = off; /* update the vnode offset */
|
|
|
|
eof:
|
|
*retval = SCARG(uap, nbytes) - resid;
|
|
out:
|
|
VOP_UNLOCK(vp, 0);
|
|
if (cookiebuf)
|
|
free(cookiebuf, M_TEMP);
|
|
free(buf, M_TEMP);
|
|
return error;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_getdents(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_getdents_args *uap = v;
|
|
register struct dirent *bdp;
|
|
struct vnode *vp;
|
|
caddr_t inp, buf; /* BSD-format */
|
|
int len, reclen; /* BSD-format */
|
|
caddr_t outp; /* SVR4-format */
|
|
int resid, svr4_reclen; /* SVR4-format */
|
|
struct file *fp;
|
|
struct uio auio;
|
|
struct iovec aiov;
|
|
struct svr4_dirent idb;
|
|
off_t off; /* true file offset */
|
|
int buflen, error, eofflag;
|
|
off_t *cookiebuf = NULL, *cookie;
|
|
int ncookies;
|
|
|
|
if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
|
|
return (error);
|
|
|
|
if ((fp->f_flag & FREAD) == 0)
|
|
return (EBADF);
|
|
|
|
vp = (struct vnode *)fp->f_data;
|
|
|
|
buflen = min(MAXBSIZE, SCARG(uap, nbytes));
|
|
buf = malloc(buflen, M_TEMP, M_WAITOK);
|
|
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
|
|
off = fp->f_offset;
|
|
again:
|
|
aiov.iov_base = buf;
|
|
aiov.iov_len = buflen;
|
|
auio.uio_iov = &aiov;
|
|
auio.uio_iovcnt = 1;
|
|
auio.uio_rw = UIO_READ;
|
|
auio.uio_segflg = UIO_SYSSPACE;
|
|
auio.uio_procp = p;
|
|
auio.uio_resid = buflen;
|
|
auio.uio_offset = off;
|
|
/*
|
|
* First we read into the malloc'ed buffer, then
|
|
* we massage it into user space, one record at a time.
|
|
*/
|
|
error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, &cookiebuf,
|
|
&ncookies);
|
|
if (error)
|
|
goto out;
|
|
|
|
inp = buf;
|
|
outp = SCARG(uap, buf);
|
|
resid = SCARG(uap, nbytes);
|
|
if ((len = buflen - auio.uio_resid) == 0)
|
|
goto eof;
|
|
|
|
for (cookie = cookiebuf; len > 0; len -= reclen) {
|
|
bdp = (struct dirent *)inp;
|
|
reclen = bdp->d_reclen;
|
|
if (reclen & 3)
|
|
panic("svr4_getdents: bad reclen");
|
|
off = *cookie++; /* each entry points to the next */
|
|
if ((off >> 32) != 0) {
|
|
compat_offseterr(vp, "svr4_getdents");
|
|
error = EINVAL;
|
|
goto out;
|
|
}
|
|
if (bdp->d_fileno == 0) {
|
|
inp += reclen; /* it is a hole; squish it out */
|
|
continue;
|
|
}
|
|
svr4_reclen = SVR4_RECLEN(&idb, bdp->d_namlen);
|
|
if (reclen > len || resid < svr4_reclen) {
|
|
/* entry too big for buffer, so just stop */
|
|
outp++;
|
|
break;
|
|
}
|
|
/*
|
|
* Massage in place to make a SVR4-shaped dirent (otherwise
|
|
* we have to worry about touching user memory outside of
|
|
* the copyout() call).
|
|
*/
|
|
idb.d_ino = (svr4_ino_t)bdp->d_fileno;
|
|
idb.d_off = (svr4_off_t)off;
|
|
idb.d_reclen = (u_short)svr4_reclen;
|
|
strcpy(idb.d_name, bdp->d_name);
|
|
if ((error = copyout((caddr_t)&idb, outp, svr4_reclen)))
|
|
goto out;
|
|
/* advance past this real entry */
|
|
inp += reclen;
|
|
/* advance output past SVR4-shaped entry */
|
|
outp += svr4_reclen;
|
|
resid -= svr4_reclen;
|
|
}
|
|
|
|
/* if we squished out the whole block, try again */
|
|
if (outp == SCARG(uap, buf))
|
|
goto again;
|
|
fp->f_offset = off; /* update the vnode offset */
|
|
|
|
eof:
|
|
*retval = SCARG(uap, nbytes) - resid;
|
|
out:
|
|
VOP_UNLOCK(vp, 0);
|
|
if (cookiebuf)
|
|
free(cookiebuf, M_TEMP);
|
|
free(buf, M_TEMP);
|
|
return error;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_mmap(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_mmap_args *uap = v;
|
|
struct sys_mmap_args mm;
|
|
void *rp;
|
|
#define _MAP_NEW 0x80000000
|
|
/*
|
|
* Verify the arguments.
|
|
*/
|
|
if (SCARG(uap, prot) & ~(PROT_READ | PROT_WRITE | PROT_EXEC))
|
|
return EINVAL; /* XXX still needed? */
|
|
|
|
if (SCARG(uap, len) == 0)
|
|
return EINVAL;
|
|
|
|
SCARG(&mm, prot) = SCARG(uap, prot);
|
|
SCARG(&mm, len) = SCARG(uap, len);
|
|
SCARG(&mm, flags) = SCARG(uap, flags) & ~_MAP_NEW;
|
|
SCARG(&mm, fd) = SCARG(uap, fd);
|
|
SCARG(&mm, addr) = SCARG(uap, addr);
|
|
SCARG(&mm, pos) = SCARG(uap, pos);
|
|
|
|
rp = (void *) round_page(p->p_vmspace->vm_daddr + MAXDSIZ);
|
|
if ((SCARG(&mm, flags) & MAP_FIXED) == 0 &&
|
|
SCARG(&mm, addr) != 0 && SCARG(&mm, addr) < rp)
|
|
SCARG(&mm, addr) = rp;
|
|
|
|
return sys_mmap(p, &mm, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_mmap64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_mmap64_args *uap = v;
|
|
struct sys_mmap_args mm;
|
|
void *rp;
|
|
#define _MAP_NEW 0x80000000
|
|
/*
|
|
* Verify the arguments.
|
|
*/
|
|
if (SCARG(uap, prot) & ~(PROT_READ | PROT_WRITE | PROT_EXEC))
|
|
return EINVAL; /* XXX still needed? */
|
|
|
|
if (SCARG(uap, len) == 0)
|
|
return EINVAL;
|
|
|
|
SCARG(&mm, prot) = SCARG(uap, prot);
|
|
SCARG(&mm, len) = SCARG(uap, len);
|
|
SCARG(&mm, flags) = SCARG(uap, flags) & ~_MAP_NEW;
|
|
SCARG(&mm, fd) = SCARG(uap, fd);
|
|
SCARG(&mm, addr) = SCARG(uap, addr);
|
|
SCARG(&mm, pos) = SCARG(uap, pos);
|
|
|
|
rp = (void *) round_page(p->p_vmspace->vm_daddr + MAXDSIZ);
|
|
if ((SCARG(&mm, flags) & MAP_FIXED) == 0 &&
|
|
SCARG(&mm, addr) != 0 && SCARG(&mm, addr) < rp)
|
|
SCARG(&mm, addr) = rp;
|
|
|
|
return sys_mmap(p, &mm, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_fchroot(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_fchroot_args *uap = v;
|
|
struct filedesc *fdp = p->p_fd;
|
|
struct vnode *vp;
|
|
struct file *fp;
|
|
int error;
|
|
|
|
if ((error = suser(p->p_ucred, &p->p_acflag)) != 0)
|
|
return error;
|
|
if ((error = getvnode(fdp, SCARG(uap, fd), &fp)) != 0)
|
|
return error;
|
|
vp = (struct vnode *) fp->f_data;
|
|
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
|
|
if (vp->v_type != VDIR)
|
|
error = ENOTDIR;
|
|
else
|
|
error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p);
|
|
VOP_UNLOCK(vp, 0);
|
|
if (error)
|
|
return error;
|
|
VREF(vp);
|
|
if (fdp->fd_rdir != NULL)
|
|
vrele(fdp->fd_rdir);
|
|
fdp->fd_rdir = vp;
|
|
return 0;
|
|
}
|
|
|
|
|
|
static int
|
|
svr4_mknod(p, retval, path, mode, dev)
|
|
struct proc *p;
|
|
register_t *retval;
|
|
char *path;
|
|
svr4_mode_t mode;
|
|
svr4_dev_t dev;
|
|
{
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
|
|
SVR4_CHECK_ALT_EXIST(p, &sg, path);
|
|
|
|
if (S_ISFIFO(mode)) {
|
|
struct sys_mkfifo_args ap;
|
|
SCARG(&ap, path) = path;
|
|
SCARG(&ap, mode) = mode;
|
|
return sys_mkfifo(p, &ap, retval);
|
|
} else {
|
|
struct sys_mknod_args ap;
|
|
SCARG(&ap, path) = path;
|
|
SCARG(&ap, mode) = mode;
|
|
SCARG(&ap, dev) = dev;
|
|
return sys_mknod(p, &ap, retval);
|
|
}
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_mknod(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_mknod_args *uap = v;
|
|
return svr4_mknod(p, retval,
|
|
SCARG(uap, path), SCARG(uap, mode),
|
|
svr4_to_bsd_odev_t(SCARG(uap, dev)));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_xmknod(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_xmknod_args *uap = v;
|
|
return svr4_mknod(p, retval,
|
|
SCARG(uap, path), SCARG(uap, mode),
|
|
svr4_to_bsd_dev_t(SCARG(uap, dev)));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_vhangup(p, v, retval)
|
|
struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_sysconfig(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_sysconfig_args *uap = v;
|
|
extern int maxfiles;
|
|
|
|
switch (SCARG(uap, name)) {
|
|
case SVR4_CONFIG_UNUSED:
|
|
*retval = 0;
|
|
break;
|
|
case SVR4_CONFIG_NGROUPS:
|
|
*retval = NGROUPS_MAX;
|
|
break;
|
|
case SVR4_CONFIG_CHILD_MAX:
|
|
*retval = maxproc;
|
|
break;
|
|
case SVR4_CONFIG_OPEN_FILES:
|
|
*retval = maxfiles;
|
|
break;
|
|
case SVR4_CONFIG_POSIX_VER:
|
|
*retval = 198808;
|
|
break;
|
|
case SVR4_CONFIG_PAGESIZE:
|
|
*retval = NBPG;
|
|
break;
|
|
case SVR4_CONFIG_CLK_TCK:
|
|
*retval = 60; /* should this be `hz', ie. 100? */
|
|
break;
|
|
case SVR4_CONFIG_XOPEN_VER:
|
|
*retval = 2; /* XXX: What should that be? */
|
|
break;
|
|
case SVR4_CONFIG_PROF_TCK:
|
|
*retval = 60; /* XXX: What should that be? */
|
|
break;
|
|
case SVR4_CONFIG_NPROC_CONF:
|
|
*retval = 1; /* Only one processor for now */
|
|
break;
|
|
case SVR4_CONFIG_NPROC_ONLN:
|
|
*retval = 1; /* And it better be online */
|
|
break;
|
|
case SVR4_CONFIG_AIO_LISTIO_MAX:
|
|
case SVR4_CONFIG_AIO_MAX:
|
|
case SVR4_CONFIG_AIO_PRIO_DELTA_MAX:
|
|
*retval = 0; /* No aio support */
|
|
break;
|
|
case SVR4_CONFIG_DELAYTIMER_MAX:
|
|
*retval = 0; /* No delaytimer support */
|
|
break;
|
|
case SVR4_CONFIG_MQ_OPEN_MAX:
|
|
*retval = msginfo.msgmni;
|
|
break;
|
|
case SVR4_CONFIG_MQ_PRIO_MAX:
|
|
*retval = 0; /* XXX: Don't know */
|
|
break;
|
|
case SVR4_CONFIG_RTSIG_MAX:
|
|
*retval = 0;
|
|
break;
|
|
case SVR4_CONFIG_SEM_NSEMS_MAX:
|
|
*retval = seminfo.semmni;
|
|
break;
|
|
case SVR4_CONFIG_SEM_VALUE_MAX:
|
|
*retval = seminfo.semvmx;
|
|
break;
|
|
case SVR4_CONFIG_SIGQUEUE_MAX:
|
|
*retval = 0; /* XXX: Don't know */
|
|
break;
|
|
case SVR4_CONFIG_SIGRT_MIN:
|
|
case SVR4_CONFIG_SIGRT_MAX:
|
|
*retval = 0; /* No real time signals */
|
|
break;
|
|
case SVR4_CONFIG_TIMER_MAX:
|
|
*retval = 3; /* XXX: real, virtual, profiling */
|
|
break;
|
|
case SVR4_CONFIG_PHYS_PAGES:
|
|
#if defined(UVM)
|
|
*retval = uvmexp.free; /* XXX: free instead of total */
|
|
#else
|
|
*retval = cnt.v_free_count; /* XXX: free instead of total */
|
|
#endif
|
|
break;
|
|
case SVR4_CONFIG_AVPHYS_PAGES:
|
|
#if defined(UVM)
|
|
*retval = uvmexp.active; /* XXX: active instead of avg */
|
|
#else
|
|
*retval = cnt.v_active_count; /* XXX: active instead of avg */
|
|
#endif
|
|
break;
|
|
default:
|
|
return EINVAL;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
#define SVR4_RLIMIT_NOFILE 5 /* Other RLIMIT_* are the same */
|
|
#define SVR4_RLIMIT_VMEM 6 /* Other RLIMIT_* are the same */
|
|
#define SVR4_RLIM_NLIMITS 7
|
|
|
|
int
|
|
svr4_sys_getrlimit(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_getrlimit_args *uap = v;
|
|
struct compat_43_sys_getrlimit_args ap;
|
|
|
|
if (SCARG(uap, which) >= SVR4_RLIM_NLIMITS)
|
|
return EINVAL;
|
|
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_NOFILE)
|
|
SCARG(uap, which) = RLIMIT_NOFILE;
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_VMEM)
|
|
SCARG(uap, which) = RLIMIT_RSS;
|
|
|
|
SCARG(&ap, which) = SCARG(uap, which);
|
|
SCARG(&ap, rlp) = SCARG(uap, rlp);
|
|
|
|
return compat_43_sys_getrlimit(p, &ap, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_setrlimit(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_setrlimit_args *uap = v;
|
|
struct compat_43_sys_setrlimit_args ap;
|
|
|
|
if (SCARG(uap, which) >= SVR4_RLIM_NLIMITS)
|
|
return EINVAL;
|
|
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_NOFILE)
|
|
SCARG(uap, which) = RLIMIT_NOFILE;
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_VMEM)
|
|
SCARG(uap, which) = RLIMIT_RSS;
|
|
|
|
SCARG(&ap, which) = SCARG(uap, which);
|
|
SCARG(&ap, rlp) = SCARG(uap, rlp);
|
|
|
|
return compat_43_sys_setrlimit(p, uap, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_getrlimit64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_getrlimit64_args *uap = v;
|
|
|
|
if (SCARG(uap, which) >= SVR4_RLIM_NLIMITS)
|
|
return EINVAL;
|
|
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_NOFILE)
|
|
SCARG(uap, which) = RLIMIT_NOFILE;
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_VMEM)
|
|
SCARG(uap, which) = RLIMIT_RSS;
|
|
|
|
return sys_getrlimit(p, uap, retval);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_setrlimit64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_setrlimit64_args *uap = v;
|
|
|
|
if (SCARG(uap, which) >= SVR4_RLIM_NLIMITS)
|
|
return EINVAL;
|
|
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_NOFILE)
|
|
SCARG(uap, which) = RLIMIT_NOFILE;
|
|
if (SCARG(uap, which) == SVR4_RLIMIT_VMEM)
|
|
SCARG(uap, which) = RLIMIT_RSS;
|
|
|
|
return sys_setrlimit(p, uap, retval);
|
|
}
|
|
|
|
|
|
/* ARGSUSED */
|
|
int
|
|
svr4_sys_break(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_break_args *uap = v;
|
|
register struct vmspace *vm = p->p_vmspace;
|
|
vm_offset_t new, old;
|
|
int rv;
|
|
register int diff;
|
|
|
|
old = (vm_offset_t) vm->vm_daddr;
|
|
new = round_page(SCARG(uap, nsize));
|
|
diff = new - old;
|
|
|
|
DPRINTF(("break(1): old %lx new %lx diff %x\n", old, new, diff));
|
|
|
|
if (diff > p->p_rlimit[RLIMIT_DATA].rlim_cur)
|
|
return ENOMEM;
|
|
|
|
old = round_page(old + ctob(vm->vm_dsize));
|
|
DPRINTF(("break(2): dsize = %x ctob %x\n",
|
|
vm->vm_dsize, ctob(vm->vm_dsize)));
|
|
|
|
diff = new - old;
|
|
DPRINTF(("break(3): old %lx new %lx diff %x\n", old, new, diff));
|
|
|
|
if (diff > 0) {
|
|
#if defined(UVM)
|
|
rv = uvm_map(&vm->vm_map, &old, diff, NULL, UVM_UNKNOWN_OFFSET,
|
|
UVM_MAPFLAG(UVM_PROT_ALL, UVM_PROT_ALL, UVM_INH_COPY,
|
|
UVM_ADV_NORMAL,
|
|
UVM_FLAG_AMAPPAD|UVM_FLAG_FIXED|
|
|
UVM_FLAG_OVERLAY|UVM_FLAG_COPYONW));
|
|
#else
|
|
rv = vm_allocate(&vm->vm_map, &old, diff, FALSE);
|
|
#endif
|
|
if (rv != KERN_SUCCESS) {
|
|
uprintf("sbrk: grow failed, return = %d\n", rv);
|
|
return ENOMEM;
|
|
}
|
|
vm->vm_dsize += btoc(diff);
|
|
} else if (diff < 0) {
|
|
diff = -diff;
|
|
#if defined(UVM)
|
|
rv = uvm_deallocate(&vm->vm_map, new, diff);
|
|
#else
|
|
rv = vm_deallocate(&vm->vm_map, new, diff);
|
|
#endif
|
|
if (rv != KERN_SUCCESS) {
|
|
uprintf("sbrk: shrink failed, return = %d\n", rv);
|
|
return ENOMEM;
|
|
}
|
|
vm->vm_dsize -= btoc(diff);
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
|
|
static __inline clock_t
|
|
timeval_to_clock_t(tv)
|
|
struct timeval *tv;
|
|
{
|
|
return tv->tv_sec * hz + tv->tv_usec / (1000000 / hz);
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_times(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_times_args *uap = v;
|
|
int error;
|
|
struct tms tms;
|
|
struct timeval t;
|
|
struct rusage *ru;
|
|
struct rusage r;
|
|
struct sys_getrusage_args ga;
|
|
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
ru = stackgap_alloc(&sg, sizeof(struct rusage));
|
|
|
|
SCARG(&ga, who) = RUSAGE_SELF;
|
|
SCARG(&ga, rusage) = ru;
|
|
|
|
error = sys_getrusage(p, &ga, retval);
|
|
if (error)
|
|
return error;
|
|
|
|
if ((error = copyin(ru, &r, sizeof r)) != 0)
|
|
return error;
|
|
|
|
tms.tms_utime = timeval_to_clock_t(&r.ru_utime);
|
|
tms.tms_stime = timeval_to_clock_t(&r.ru_stime);
|
|
|
|
SCARG(&ga, who) = RUSAGE_CHILDREN;
|
|
error = sys_getrusage(p, &ga, retval);
|
|
if (error)
|
|
return error;
|
|
|
|
if ((error = copyin(ru, &r, sizeof r)) != 0)
|
|
return error;
|
|
|
|
tms.tms_cutime = timeval_to_clock_t(&r.ru_utime);
|
|
tms.tms_cstime = timeval_to_clock_t(&r.ru_stime);
|
|
|
|
microtime(&t);
|
|
*retval = timeval_to_clock_t(&t);
|
|
|
|
return copyout(&tms, SCARG(uap, tp), sizeof(tms));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_ulimit(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_ulimit_args *uap = v;
|
|
|
|
switch (SCARG(uap, cmd)) {
|
|
case SVR4_GFILLIM:
|
|
*retval = p->p_rlimit[RLIMIT_FSIZE].rlim_cur / 512;
|
|
if (*retval == -1)
|
|
*retval = 0x7fffffff;
|
|
return 0;
|
|
|
|
case SVR4_SFILLIM:
|
|
{
|
|
int error;
|
|
struct sys_setrlimit_args srl;
|
|
struct rlimit krl;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
struct rlimit *url = (struct rlimit *)
|
|
stackgap_alloc(&sg, sizeof *url);
|
|
|
|
krl.rlim_cur = SCARG(uap, newlimit) * 512;
|
|
krl.rlim_max = p->p_rlimit[RLIMIT_FSIZE].rlim_max;
|
|
|
|
error = copyout(&krl, url, sizeof(*url));
|
|
if (error)
|
|
return error;
|
|
|
|
SCARG(&srl, which) = RLIMIT_FSIZE;
|
|
SCARG(&srl, rlp) = url;
|
|
|
|
error = sys_setrlimit(p, &srl, retval);
|
|
if (error)
|
|
return error;
|
|
|
|
*retval = p->p_rlimit[RLIMIT_FSIZE].rlim_cur;
|
|
if (*retval == -1)
|
|
*retval = 0x7fffffff;
|
|
return 0;
|
|
}
|
|
|
|
case SVR4_GMEMLIM:
|
|
{
|
|
struct vmspace *vm = p->p_vmspace;
|
|
register_t r = p->p_rlimit[RLIMIT_DATA].rlim_cur;
|
|
|
|
if (r == -1)
|
|
r = 0x7fffffff;
|
|
r += (long) vm->vm_daddr;
|
|
if (r < 0)
|
|
r = 0x7fffffff;
|
|
*retval = r;
|
|
return 0;
|
|
}
|
|
|
|
case SVR4_GDESLIM:
|
|
*retval = p->p_rlimit[RLIMIT_NOFILE].rlim_cur;
|
|
if (*retval == -1)
|
|
*retval = 0x7fffffff;
|
|
return 0;
|
|
|
|
default:
|
|
return EINVAL;
|
|
}
|
|
}
|
|
|
|
|
|
static struct proc *
|
|
svr4_pfind(pid)
|
|
pid_t pid;
|
|
{
|
|
struct proc *p;
|
|
|
|
/* look in the live processes */
|
|
if ((p = pfind(pid)) != NULL)
|
|
return p;
|
|
|
|
/* look in the zombies */
|
|
for (p = zombproc.lh_first; p != 0; p = p->p_list.le_next)
|
|
if (p->p_pid == pid)
|
|
return p;
|
|
|
|
return NULL;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_pgrpsys(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_pgrpsys_args *uap = v;
|
|
|
|
switch (SCARG(uap, cmd)) {
|
|
case 1: /* setpgrp() */
|
|
/*
|
|
* SVR4 setpgrp() (which takes no arguments) has the
|
|
* semantics that the session ID is also created anew, so
|
|
* in almost every sense, setpgrp() is identical to
|
|
* setsid() for SVR4. (Under BSD, the difference is that
|
|
* a setpgid(0,0) will not create a new session.)
|
|
*/
|
|
sys_setsid(p, NULL, retval);
|
|
/*FALLTHROUGH*/
|
|
|
|
case 0: /* getpgrp() */
|
|
*retval = p->p_pgrp->pg_id;
|
|
return 0;
|
|
|
|
case 2: /* getsid(pid) */
|
|
if (SCARG(uap, pid) != 0 &&
|
|
(p = svr4_pfind(SCARG(uap, pid))) == NULL)
|
|
return ESRCH;
|
|
/*
|
|
* This has already been initialized to the pid of
|
|
* the session leader.
|
|
*/
|
|
*retval = (register_t) p->p_session->s_sid;
|
|
return 0;
|
|
|
|
case 3: /* setsid() */
|
|
return sys_setsid(p, NULL, retval);
|
|
|
|
case 4: /* getpgid(pid) */
|
|
|
|
if (SCARG(uap, pid) != 0 &&
|
|
(p = svr4_pfind(SCARG(uap, pid))) == NULL)
|
|
return ESRCH;
|
|
|
|
*retval = (int) p->p_pgrp->pg_id;
|
|
return 0;
|
|
|
|
case 5: /* setpgid(pid, pgid); */
|
|
{
|
|
struct sys_setpgid_args sa;
|
|
|
|
SCARG(&sa, pid) = SCARG(uap, pid);
|
|
SCARG(&sa, pgid) = SCARG(uap, pgid);
|
|
return sys_setpgid(p, &sa, retval);
|
|
}
|
|
|
|
default:
|
|
return EINVAL;
|
|
}
|
|
}
|
|
|
|
#define syscallarg(x) union { x datum; register_t pad; }
|
|
|
|
struct svr4_hrtcntl_args {
|
|
syscallarg(int) cmd;
|
|
syscallarg(int) fun;
|
|
syscallarg(int) clk;
|
|
syscallarg(svr4_hrt_interval_t *) iv;
|
|
syscallarg(svr4_hrt_time_t *) ti;
|
|
};
|
|
|
|
|
|
static int
|
|
svr4_hrtcntl(p, uap, retval)
|
|
register struct proc *p;
|
|
register struct svr4_hrtcntl_args *uap;
|
|
register_t *retval;
|
|
{
|
|
switch (SCARG(uap, fun)) {
|
|
case SVR4_HRT_CNTL_RES:
|
|
DPRINTF(("htrcntl(RES)\n"));
|
|
*retval = SVR4_HRT_USEC;
|
|
return 0;
|
|
|
|
case SVR4_HRT_CNTL_TOFD:
|
|
DPRINTF(("htrcntl(TOFD)\n"));
|
|
{
|
|
struct timeval tv;
|
|
svr4_hrt_time_t t;
|
|
if (SCARG(uap, clk) != SVR4_HRT_CLK_STD) {
|
|
DPRINTF(("clk == %d\n", SCARG(uap, clk)));
|
|
return EINVAL;
|
|
}
|
|
if (SCARG(uap, ti) == NULL) {
|
|
DPRINTF(("ti NULL\n"));
|
|
return EINVAL;
|
|
}
|
|
microtime(&tv);
|
|
t.h_sec = tv.tv_sec;
|
|
t.h_rem = tv.tv_usec;
|
|
t.h_res = SVR4_HRT_USEC;
|
|
return copyout(&t, SCARG(uap, ti), sizeof(t));
|
|
}
|
|
|
|
case SVR4_HRT_CNTL_START:
|
|
DPRINTF(("htrcntl(START)\n"));
|
|
return ENOSYS;
|
|
|
|
case SVR4_HRT_CNTL_GET:
|
|
DPRINTF(("htrcntl(GET)\n"));
|
|
return ENOSYS;
|
|
default:
|
|
DPRINTF(("Bad htrcntl command %d\n", SCARG(uap, fun)));
|
|
return ENOSYS;
|
|
}
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_hrtsys(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_hrtsys_args *uap = v;
|
|
|
|
switch (SCARG(uap, cmd)) {
|
|
case SVR4_HRT_CNTL:
|
|
return svr4_hrtcntl(p, (struct svr4_hrtcntl_args *) uap,
|
|
retval);
|
|
|
|
case SVR4_HRT_ALRM:
|
|
DPRINTF(("hrtalarm\n"));
|
|
return ENOSYS;
|
|
|
|
case SVR4_HRT_SLP:
|
|
DPRINTF(("hrtsleep\n"));
|
|
return ENOSYS;
|
|
|
|
case SVR4_HRT_CAN:
|
|
DPRINTF(("hrtcancel\n"));
|
|
return ENOSYS;
|
|
|
|
default:
|
|
DPRINTF(("Bad hrtsys command %d\n", SCARG(uap, cmd)));
|
|
return EINVAL;
|
|
}
|
|
}
|
|
|
|
|
|
static int
|
|
svr4_setinfo(p, st, s)
|
|
struct proc *p;
|
|
int st;
|
|
svr4_siginfo_t *s;
|
|
{
|
|
svr4_siginfo_t i;
|
|
|
|
bzero(&i, sizeof(i));
|
|
|
|
i.si_signo = SVR4_SIGCHLD;
|
|
i.si_errno = 0; /* XXX? */
|
|
|
|
if (p) {
|
|
i.si_pid = p->p_pid;
|
|
if (p->p_stat == SZOMB) {
|
|
i.si_stime = p->p_ru->ru_stime.tv_sec;
|
|
i.si_utime = p->p_ru->ru_utime.tv_sec;
|
|
}
|
|
else {
|
|
i.si_stime = p->p_stats->p_ru.ru_stime.tv_sec;
|
|
i.si_utime = p->p_stats->p_ru.ru_utime.tv_sec;
|
|
}
|
|
}
|
|
|
|
if (WIFEXITED(st)) {
|
|
i.si_status = WEXITSTATUS(st);
|
|
i.si_code = SVR4_CLD_EXITED;
|
|
}
|
|
else if (WIFSTOPPED(st)) {
|
|
i.si_status = bsd_to_svr4_sig[WSTOPSIG(st)];
|
|
|
|
if (i.si_status == SVR4_SIGCONT)
|
|
i.si_code = SVR4_CLD_CONTINUED;
|
|
else
|
|
i.si_code = SVR4_CLD_STOPPED;
|
|
} else {
|
|
i.si_status = bsd_to_svr4_sig[WTERMSIG(st)];
|
|
|
|
if (WCOREDUMP(st))
|
|
i.si_code = SVR4_CLD_DUMPED;
|
|
else
|
|
i.si_code = SVR4_CLD_KILLED;
|
|
}
|
|
|
|
DPRINTF(("siginfo [pid %ld signo %d code %d errno %d status %d]\n",
|
|
i.si_pid, i.si_signo, i.si_code, i.si_errno, i.si_status));
|
|
|
|
return copyout(&i, s, sizeof(i));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_waitsys(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_waitsys_args *uap = v;
|
|
int nfound;
|
|
int error;
|
|
struct proc *q, *t;
|
|
|
|
|
|
switch (SCARG(uap, grp)) {
|
|
case SVR4_P_PID:
|
|
break;
|
|
|
|
case SVR4_P_PGID:
|
|
SCARG(uap, id) = -p->p_pgid;
|
|
break;
|
|
|
|
case SVR4_P_ALL:
|
|
SCARG(uap, id) = WAIT_ANY;
|
|
break;
|
|
|
|
default:
|
|
return EINVAL;
|
|
}
|
|
|
|
DPRINTF(("waitsys(%d, %d, %p, %x)\n",
|
|
SCARG(uap, grp), SCARG(uap, id),
|
|
SCARG(uap, info), SCARG(uap, options)));
|
|
|
|
loop:
|
|
nfound = 0;
|
|
for (q = p->p_children.lh_first; q != 0; q = q->p_sibling.le_next) {
|
|
if (SCARG(uap, id) != WAIT_ANY &&
|
|
q->p_pid != SCARG(uap, id) &&
|
|
q->p_pgid != -SCARG(uap, id)) {
|
|
DPRINTF(("pid %d pgid %d != %d\n", q->p_pid,
|
|
q->p_pgid, SCARG(uap, id)));
|
|
continue;
|
|
}
|
|
nfound++;
|
|
if (q->p_stat == SZOMB &&
|
|
((SCARG(uap, options) & (SVR4_WEXITED|SVR4_WTRAPPED)))) {
|
|
*retval = 0;
|
|
DPRINTF(("found %d\n", q->p_pid));
|
|
if ((error = svr4_setinfo(q, q->p_xstat,
|
|
SCARG(uap, info))) != 0)
|
|
return error;
|
|
|
|
|
|
if ((SCARG(uap, options) & SVR4_WNOWAIT)) {
|
|
DPRINTF(("Don't wait\n"));
|
|
return 0;
|
|
}
|
|
|
|
/*
|
|
* If we got the child via a ptrace 'attach',
|
|
* we need to give it back to the old parent.
|
|
*/
|
|
if (q->p_oppid && (t = pfind(q->p_oppid))) {
|
|
q->p_oppid = 0;
|
|
proc_reparent(q, t);
|
|
psignal(t, SIGCHLD);
|
|
wakeup((caddr_t)t);
|
|
return 0;
|
|
}
|
|
q->p_xstat = 0;
|
|
ruadd(&p->p_stats->p_cru, q->p_ru);
|
|
FREE(q->p_ru, M_ZOMBIE);
|
|
|
|
/*
|
|
* Decrement the count of procs running with this uid.
|
|
*/
|
|
(void)chgproccnt(q->p_cred->p_ruid, -1);
|
|
|
|
/*
|
|
* Free up credentials.
|
|
*/
|
|
if (--q->p_cred->p_refcnt == 0) {
|
|
crfree(q->p_cred->pc_ucred);
|
|
FREE(q->p_cred, M_SUBPROC);
|
|
}
|
|
|
|
/*
|
|
* Release reference to text vnode
|
|
*/
|
|
if (q->p_textvp)
|
|
vrele(q->p_textvp);
|
|
|
|
/*
|
|
* Finally finished with old proc entry.
|
|
* Unlink it from its process group and free it.
|
|
*/
|
|
leavepgrp(q);
|
|
LIST_REMOVE(q, p_list); /* off zombproc */
|
|
LIST_REMOVE(q, p_sibling);
|
|
|
|
/*
|
|
* Give machine-dependent layer a chance
|
|
* to free anything that cpu_exit couldn't
|
|
* release while still running in process context.
|
|
*/
|
|
cpu_wait(q);
|
|
pool_put(&proc_pool, q);
|
|
nprocs--;
|
|
return 0;
|
|
}
|
|
if (q->p_stat == SSTOP && (q->p_flag & P_WAITED) == 0 &&
|
|
(q->p_flag & P_TRACED ||
|
|
(SCARG(uap, options) & (SVR4_WSTOPPED|SVR4_WCONTINUED)))) {
|
|
DPRINTF(("jobcontrol %d\n", q->p_pid));
|
|
if (((SCARG(uap, options) & SVR4_WNOWAIT)) == 0)
|
|
q->p_flag |= P_WAITED;
|
|
*retval = 0;
|
|
return svr4_setinfo(q, W_STOPCODE(q->p_xstat),
|
|
SCARG(uap, info));
|
|
}
|
|
}
|
|
|
|
if (nfound == 0)
|
|
return ECHILD;
|
|
|
|
if (SCARG(uap, options) & SVR4_WNOHANG) {
|
|
*retval = 0;
|
|
if ((error = svr4_setinfo(NULL, 0, SCARG(uap, info))) != 0)
|
|
return error;
|
|
return 0;
|
|
}
|
|
|
|
if ((error = tsleep((caddr_t)p, PWAIT | PCATCH, "svr4_wait", 0)) != 0)
|
|
return error;
|
|
goto loop;
|
|
}
|
|
|
|
|
|
static void
|
|
bsd_statfs_to_svr4_statvfs(bfs, sfs)
|
|
const struct statfs *bfs;
|
|
struct svr4_statvfs *sfs;
|
|
{
|
|
sfs->f_bsize = bfs->f_iosize; /* XXX */
|
|
sfs->f_frsize = bfs->f_bsize;
|
|
sfs->f_blocks = bfs->f_blocks;
|
|
sfs->f_bfree = bfs->f_bfree;
|
|
sfs->f_bavail = bfs->f_bavail;
|
|
sfs->f_files = bfs->f_files;
|
|
sfs->f_ffree = bfs->f_ffree;
|
|
sfs->f_favail = bfs->f_ffree;
|
|
sfs->f_fsid = bfs->f_fsid.val[0];
|
|
bcopy(bfs->f_fstypename, sfs->f_basetype, sizeof(sfs->f_basetype));
|
|
sfs->f_flag = 0;
|
|
if (bfs->f_flags & MNT_RDONLY)
|
|
sfs->f_flag |= SVR4_ST_RDONLY;
|
|
if (bfs->f_flags & MNT_NOSUID)
|
|
sfs->f_flag |= SVR4_ST_NOSUID;
|
|
sfs->f_namemax = MAXNAMLEN;
|
|
bcopy(bfs->f_fstypename, sfs->f_fstr, sizeof(sfs->f_fstr)); /* XXX */
|
|
bzero(sfs->f_filler, sizeof(sfs->f_filler));
|
|
}
|
|
|
|
|
|
static void
|
|
bsd_statfs_to_svr4_statvfs64(bfs, sfs)
|
|
const struct statfs *bfs;
|
|
struct svr4_statvfs64 *sfs;
|
|
{
|
|
sfs->f_bsize = bfs->f_iosize; /* XXX */
|
|
sfs->f_frsize = bfs->f_bsize;
|
|
sfs->f_blocks = bfs->f_blocks;
|
|
sfs->f_bfree = bfs->f_bfree;
|
|
sfs->f_bavail = bfs->f_bavail;
|
|
sfs->f_files = bfs->f_files;
|
|
sfs->f_ffree = bfs->f_ffree;
|
|
sfs->f_favail = bfs->f_ffree;
|
|
sfs->f_fsid = bfs->f_fsid.val[0];
|
|
bcopy(bfs->f_fstypename, sfs->f_basetype, sizeof(sfs->f_basetype));
|
|
sfs->f_flag = 0;
|
|
if (bfs->f_flags & MNT_RDONLY)
|
|
sfs->f_flag |= SVR4_ST_RDONLY;
|
|
if (bfs->f_flags & MNT_NOSUID)
|
|
sfs->f_flag |= SVR4_ST_NOSUID;
|
|
sfs->f_namemax = MAXNAMLEN;
|
|
bcopy(bfs->f_fstypename, sfs->f_fstr, sizeof(sfs->f_fstr)); /* XXX */
|
|
bzero(sfs->f_filler, sizeof(sfs->f_filler));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_statvfs(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_statvfs_args *uap = v;
|
|
struct sys_statfs_args fs_args;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
struct statfs *fs = stackgap_alloc(&sg, sizeof(struct statfs));
|
|
struct statfs bfs;
|
|
struct svr4_statvfs sfs;
|
|
int error;
|
|
|
|
SVR4_CHECK_ALT_EXIST(p, &sg, SCARG(uap, path));
|
|
SCARG(&fs_args, path) = SCARG(uap, path);
|
|
SCARG(&fs_args, buf) = fs;
|
|
|
|
if ((error = sys_statfs(p, &fs_args, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = copyin(fs, &bfs, sizeof(bfs))) != 0)
|
|
return error;
|
|
|
|
bsd_statfs_to_svr4_statvfs(&bfs, &sfs);
|
|
|
|
return copyout(&sfs, SCARG(uap, fs), sizeof(sfs));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_fstatvfs(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_fstatvfs_args *uap = v;
|
|
struct sys_fstatfs_args fs_args;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
struct statfs *fs = stackgap_alloc(&sg, sizeof(struct statfs));
|
|
struct statfs bfs;
|
|
struct svr4_statvfs sfs;
|
|
int error;
|
|
|
|
SCARG(&fs_args, fd) = SCARG(uap, fd);
|
|
SCARG(&fs_args, buf) = fs;
|
|
|
|
if ((error = sys_fstatfs(p, &fs_args, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = copyin(fs, &bfs, sizeof(bfs))) != 0)
|
|
return error;
|
|
|
|
bsd_statfs_to_svr4_statvfs(&bfs, &sfs);
|
|
|
|
return copyout(&sfs, SCARG(uap, fs), sizeof(sfs));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_statvfs64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_statvfs64_args *uap = v;
|
|
struct sys_statfs_args fs_args;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
struct statfs *fs = stackgap_alloc(&sg, sizeof(struct statfs));
|
|
struct statfs bfs;
|
|
struct svr4_statvfs64 sfs;
|
|
int error;
|
|
|
|
SVR4_CHECK_ALT_EXIST(p, &sg, SCARG(uap, path));
|
|
SCARG(&fs_args, path) = SCARG(uap, path);
|
|
SCARG(&fs_args, buf) = fs;
|
|
|
|
if ((error = sys_statfs(p, &fs_args, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = copyin(fs, &bfs, sizeof(bfs))) != 0)
|
|
return error;
|
|
|
|
bsd_statfs_to_svr4_statvfs64(&bfs, &sfs);
|
|
|
|
return copyout(&sfs, SCARG(uap, fs), sizeof(sfs));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_fstatvfs64(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_fstatvfs64_args *uap = v;
|
|
struct sys_fstatfs_args fs_args;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
struct statfs *fs = stackgap_alloc(&sg, sizeof(struct statfs));
|
|
struct statfs bfs;
|
|
struct svr4_statvfs64 sfs;
|
|
int error;
|
|
|
|
SCARG(&fs_args, fd) = SCARG(uap, fd);
|
|
SCARG(&fs_args, buf) = fs;
|
|
|
|
if ((error = sys_fstatfs(p, &fs_args, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = copyin(fs, &bfs, sizeof(bfs))) != 0)
|
|
return error;
|
|
|
|
bsd_statfs_to_svr4_statvfs64(&bfs, &sfs);
|
|
|
|
return copyout(&sfs, SCARG(uap, fs), sizeof(sfs));
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_alarm(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_alarm_args *uap = v;
|
|
int error;
|
|
struct itimerval *ntp, *otp, tp;
|
|
struct sys_setitimer_args sa;
|
|
caddr_t sg = stackgap_init(p->p_emul);
|
|
|
|
ntp = stackgap_alloc(&sg, sizeof(struct itimerval));
|
|
otp = stackgap_alloc(&sg, sizeof(struct itimerval));
|
|
|
|
timerclear(&tp.it_interval);
|
|
tp.it_value.tv_sec = SCARG(uap, sec);
|
|
tp.it_value.tv_usec = 0;
|
|
|
|
if ((error = copyout(&tp, ntp, sizeof(tp))) != 0)
|
|
return error;
|
|
|
|
SCARG(&sa, which) = ITIMER_REAL;
|
|
SCARG(&sa, itv) = ntp;
|
|
SCARG(&sa, oitv) = otp;
|
|
|
|
if ((error = sys_setitimer(p, &sa, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = copyin(otp, &tp, sizeof(tp))) != 0)
|
|
return error;
|
|
|
|
if (tp.it_value.tv_usec)
|
|
tp.it_value.tv_sec++;
|
|
|
|
*retval = (register_t) tp.it_value.tv_sec;
|
|
|
|
return 0;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_gettimeofday(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_gettimeofday_args *uap = v;
|
|
|
|
if (SCARG(uap, tp)) {
|
|
struct timeval atv;
|
|
|
|
microtime(&atv);
|
|
return copyout(&atv, SCARG(uap, tp), sizeof (atv));
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_facl(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_facl_args *uap = v;
|
|
|
|
*retval = 0;
|
|
|
|
switch (SCARG(uap, cmd)) {
|
|
case SVR4_SYS_SETACL:
|
|
/* We don't support acls on any filesystem */
|
|
return ENOSYS;
|
|
|
|
case SVR4_SYS_GETACL:
|
|
return copyout(retval, &SCARG(uap, num),
|
|
sizeof(SCARG(uap, num)));
|
|
|
|
case SVR4_SYS_GETACLCNT:
|
|
return 0;
|
|
|
|
default:
|
|
return EINVAL;
|
|
}
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_acl(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
return svr4_sys_facl(p, v, retval); /* XXX: for now the same */
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_auditsys(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
/*
|
|
* XXX: Big brother is *not* watching.
|
|
*/
|
|
return 0;
|
|
}
|
|
|
|
int
|
|
svr4_sys_memcntl(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_memcntl_args *uap = v;
|
|
switch (SCARG(uap, cmd)) {
|
|
case SVR4_MC_SYNC:
|
|
{
|
|
struct sys___msync13_args msa;
|
|
|
|
SCARG(&msa, addr) = SCARG(uap, addr);
|
|
SCARG(&msa, len) = SCARG(uap, len);
|
|
SCARG(&msa, flags) = (int)SCARG(uap, arg);
|
|
|
|
return sys___msync13(p, &msa, retval);
|
|
}
|
|
case SVR4_MC_ADVISE:
|
|
{
|
|
struct sys_madvise_args maa;
|
|
|
|
SCARG(&maa, addr) = SCARG(uap, addr);
|
|
SCARG(&maa, len) = SCARG(uap, len);
|
|
SCARG(&maa, behav) = (int)SCARG(uap, arg);
|
|
|
|
return sys_madvise(p, &maa, retval);
|
|
}
|
|
case SVR4_MC_LOCK:
|
|
case SVR4_MC_UNLOCK:
|
|
case SVR4_MC_LOCKAS:
|
|
case SVR4_MC_UNLOCKAS:
|
|
return EOPNOTSUPP;
|
|
default:
|
|
return ENOSYS;
|
|
}
|
|
}
|
|
|
|
|
|
int
|
|
svr4_sys_nice(p, v, retval)
|
|
register struct proc *p;
|
|
void *v;
|
|
register_t *retval;
|
|
{
|
|
struct svr4_sys_nice_args *uap = v;
|
|
struct sys_setpriority_args ap;
|
|
int error;
|
|
|
|
SCARG(&ap, which) = PRIO_PROCESS;
|
|
SCARG(&ap, who) = 0;
|
|
SCARG(&ap, prio) = SCARG(uap, prio);
|
|
|
|
if ((error = sys_setpriority(p, &ap, retval)) != 0)
|
|
return error;
|
|
|
|
if ((error = sys_getpriority(p, &ap, retval)) != 0)
|
|
return error;
|
|
|
|
return 0;
|
|
}
|