1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Copyright (c) 1982, 1986, 1989, 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.
|
|
|
|
* 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.
|
|
|
|
*
|
|
|
|
* @(#)sys_generic.c 8.5 (Berkeley) 1/21/94
|
|
|
|
*/
|
|
|
|
|
2003-06-11 00:56:59 +00:00
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
1996-01-03 21:42:35 +00:00
|
|
|
#include "opt_ktrace.h"
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
1995-11-12 06:43:28 +00:00
|
|
|
#include <sys/sysproto.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/filedesc.h>
|
1997-03-24 11:52:29 +00:00
|
|
|
#include <sys/filio.h>
|
1997-03-23 03:37:54 +00:00
|
|
|
#include <sys/fcntl.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/file.h>
|
|
|
|
#include <sys/proc.h>
|
1994-10-02 17:35:40 +00:00
|
|
|
#include <sys/signalvar.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/socketvar.h>
|
|
|
|
#include <sys/uio.h>
|
|
|
|
#include <sys/kernel.h>
|
2003-04-29 13:36:06 +00:00
|
|
|
#include <sys/limits.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/malloc.h>
|
1997-09-14 02:30:32 +00:00
|
|
|
#include <sys/poll.h>
|
2001-02-07 23:28:01 +00:00
|
|
|
#include <sys/resourcevar.h>
|
2001-01-09 04:33:49 +00:00
|
|
|
#include <sys/selinfo.h>
|
Switch the sleep/wakeup and condition variable implementations to use the
sleep queue interface:
- Sleep queues attempt to merge some of the benefits of both sleep queues
and condition variables. Having sleep qeueus in a hash table avoids
having to allocate a queue head for each wait channel. Thus, struct cv
has shrunk down to just a single char * pointer now. However, the
hash table does not hold threads directly, but queue heads. This means
that once you have located a queue in the hash bucket, you no longer have
to walk the rest of the hash chain looking for threads. Instead, you have
a list of all the threads sleeping on that wait channel.
- Outside of the sleepq code and the sleep/cv code the kernel no longer
differentiates between cv's and sleep/wakeup. For example, calls to
abortsleep() and cv_abort() are replaced with a call to sleepq_abort().
Thus, the TDF_CVWAITQ flag is removed. Also, calls to unsleep() and
cv_waitq_remove() have been replaced with calls to sleepq_remove().
- The sched_sleep() function no longer accepts a priority argument as
sleep's no longer inherently bump the priority. Instead, this is soley
a propery of msleep() which explicitly calls sched_prio() before
blocking.
- The TDF_ONSLEEPQ flag has been dropped as it was never used. The
associated TDF_SET_ONSLEEPQ and TDF_CLR_ON_SLEEPQ macros have also been
dropped and replaced with a single explicit clearing of td_wchan.
TD_SET_ONSLEEPQ() would really have only made sense if it had taken
the wait channel and message as arguments anyway. Now that that only
happens in one place, a macro would be overkill.
2004-02-27 18:52:44 +00:00
|
|
|
#include <sys/sleepqueue.h>
|
2002-09-01 20:37:28 +00:00
|
|
|
#include <sys/syscallsubr.h>
|
2000-01-05 19:40:17 +00:00
|
|
|
#include <sys/sysctl.h>
|
1997-09-14 02:30:32 +00:00
|
|
|
#include <sys/sysent.h>
|
2004-01-20 01:27:42 +00:00
|
|
|
#include <sys/vnode.h>
|
2000-11-18 21:01:04 +00:00
|
|
|
#include <sys/bio.h>
|
|
|
|
#include <sys/buf.h>
|
2001-05-14 05:26:48 +00:00
|
|
|
#include <sys/condvar.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
|
|
|
#include <sys/ktrace.h>
|
|
|
|
#endif
|
2000-11-18 21:01:04 +00:00
|
|
|
#include <vm/vm.h>
|
|
|
|
#include <vm/vm_page.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1997-10-12 20:26:33 +00:00
|
|
|
static MALLOC_DEFINE(M_IOCTLOPS, "ioctlops", "ioctl data buffer");
|
|
|
|
static MALLOC_DEFINE(M_SELECT, "select", "select() buffer");
|
|
|
|
MALLOC_DEFINE(M_IOV, "iov", "large iov's");
|
1997-10-11 18:31:40 +00:00
|
|
|
|
2002-03-09 22:44:37 +00:00
|
|
|
static int pollscan(struct thread *, struct pollfd *, u_int);
|
|
|
|
static int selscan(struct thread *, fd_mask **, fd_mask **, int);
|
|
|
|
static int dofileread(struct thread *, struct file *, int, void *,
|
|
|
|
size_t, off_t, int);
|
|
|
|
static int dofilewrite(struct thread *, struct file *, int,
|
|
|
|
const void *, size_t, off_t, int);
|
2003-11-09 09:17:26 +00:00
|
|
|
static void doselwakeup(struct selinfo *, int);
|
1999-04-04 21:41:28 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Read system call.
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct read_args {
|
|
|
|
int fd;
|
1998-09-05 14:30:11 +00:00
|
|
|
void *buf;
|
|
|
|
size_t nbyte;
|
1994-05-24 10:09:53 +00:00
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
read(td, uap)
|
|
|
|
struct thread *td;
|
2001-11-14 06:30:36 +00:00
|
|
|
struct read_args *uap;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2000-11-18 21:01:04 +00:00
|
|
|
int error;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2001-11-14 06:30:36 +00:00
|
|
|
if ((error = fget_read(td, uap->fd, &fp)) == 0) {
|
2001-09-12 08:38:13 +00:00
|
|
|
error = dofileread(td, fp, uap->fd, uap->buf,
|
2001-09-01 19:34:23 +00:00
|
|
|
uap->nbyte, (off_t)-1, 0);
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
2001-09-01 19:34:23 +00:00
|
|
|
}
|
2000-11-18 21:01:04 +00:00
|
|
|
return(error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
1999-03-27 21:16:58 +00:00
|
|
|
/*
|
1999-04-04 21:41:28 +00:00
|
|
|
* Pread system call
|
1999-03-27 21:16:58 +00:00
|
|
|
*/
|
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
|
|
|
struct pread_args {
|
|
|
|
int fd;
|
|
|
|
void *buf;
|
|
|
|
size_t nbyte;
|
1999-04-04 21:41:28 +00:00
|
|
|
int pad;
|
|
|
|
off_t offset;
|
1999-03-27 21:16:58 +00:00
|
|
|
};
|
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1999-03-27 21:16:58 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
pread(td, uap)
|
|
|
|
struct thread *td;
|
2001-11-14 06:30:36 +00:00
|
|
|
struct pread_args *uap;
|
1999-03-27 21:16:58 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2000-11-18 21:01:04 +00:00
|
|
|
int error;
|
1999-04-04 21:41:28 +00:00
|
|
|
|
2002-01-23 08:22:59 +00:00
|
|
|
if ((error = fget_read(td, uap->fd, &fp)) != 0)
|
|
|
|
return (error);
|
2004-01-20 01:27:42 +00:00
|
|
|
if (!(fp->f_ops->fo_flags & DFLAG_SEEKABLE))
|
2002-01-13 11:58:06 +00:00
|
|
|
error = ESPIPE;
|
2004-01-20 01:27:42 +00:00
|
|
|
else if (uap->offset < 0 && fp->f_vnode->v_type != VCHR)
|
|
|
|
error = EINVAL;
|
|
|
|
else {
|
2002-01-13 11:58:06 +00:00
|
|
|
error = dofileread(td, fp, uap->fd, uap->buf, uap->nbyte,
|
|
|
|
uap->offset, FOF_OFFSET);
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
2002-01-13 11:58:06 +00:00
|
|
|
fdrop(fp, td);
|
2000-11-18 21:01:04 +00:00
|
|
|
return(error);
|
1999-04-04 21:41:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Code common for read and pread
|
|
|
|
*/
|
2002-09-28 17:15:38 +00:00
|
|
|
static int
|
2001-09-12 08:38:13 +00:00
|
|
|
dofileread(td, fp, fd, buf, nbyte, offset, flags)
|
|
|
|
struct thread *td;
|
1999-04-04 21:41:28 +00:00
|
|
|
struct file *fp;
|
|
|
|
int fd, flags;
|
|
|
|
void *buf;
|
|
|
|
size_t nbyte;
|
|
|
|
off_t offset;
|
|
|
|
{
|
1999-03-27 21:16:58 +00:00
|
|
|
struct uio auio;
|
|
|
|
struct iovec aiov;
|
|
|
|
long cnt, error = 0;
|
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *ktruio = NULL;
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
|
|
|
|
2002-07-22 15:44:27 +00:00
|
|
|
aiov.iov_base = buf;
|
1999-04-04 21:41:28 +00:00
|
|
|
aiov.iov_len = nbyte;
|
1999-03-27 21:16:58 +00:00
|
|
|
auio.uio_iov = &aiov;
|
|
|
|
auio.uio_iovcnt = 1;
|
1999-04-04 21:41:28 +00:00
|
|
|
auio.uio_offset = offset;
|
|
|
|
if (nbyte > INT_MAX)
|
1999-03-27 21:16:58 +00:00
|
|
|
return (EINVAL);
|
1999-04-04 21:41:28 +00:00
|
|
|
auio.uio_resid = nbyte;
|
1999-03-27 21:16:58 +00:00
|
|
|
auio.uio_rw = UIO_READ;
|
|
|
|
auio.uio_segflg = UIO_USERSPACE;
|
2001-09-12 08:38:13 +00:00
|
|
|
auio.uio_td = td;
|
1999-03-27 21:16:58 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (KTRPOINT(td, KTR_GENIO))
|
|
|
|
ktruio = cloneuio(&auio);
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
1999-04-04 21:41:28 +00:00
|
|
|
cnt = nbyte;
|
2000-11-18 21:01:04 +00:00
|
|
|
|
In order to better support flexible and extensible access control,
make a series of modifications to the credential arguments relating
to file read and write operations to cliarfy which credential is
used for what:
- Change fo_read() and fo_write() to accept "active_cred" instead of
"cred", and change the semantics of consumers of fo_read() and
fo_write() to pass the active credential of the thread requesting
an operation rather than the cached file cred. The cached file
cred is still available in fo_read() and fo_write() consumers
via fp->f_cred. These changes largely in sys_generic.c.
For each implementation of fo_read() and fo_write(), update cred
usage to reflect this change and maintain current semantics:
- badfo_readwrite() unchanged
- kqueue_read/write() unchanged
pipe_read/write() now authorize MAC using active_cred rather
than td->td_ucred
- soo_read/write() unchanged
- vn_read/write() now authorize MAC using active_cred but
VOP_READ/WRITE() with fp->f_cred
Modify vn_rdwr() to accept two credential arguments instead of a
single credential: active_cred and file_cred. Use active_cred
for MAC authorization, and select a credential for use in
VOP_READ/WRITE() based on whether file_cred is NULL or not. If
file_cred is provided, authorize the VOP using that cred,
otherwise the active credential, matching current semantics.
Modify current vn_rdwr() consumers to pass a file_cred if used
in the context of a struct file, and to always pass active_cred.
When vn_rdwr() is used without a file_cred, pass NOCRED.
These changes should maintain current semantics for read/write,
but avoid a redundant passing of fp->f_cred, as well as making
it more clear what the origin of each credential is in file
descriptor read/write operations.
Follow-up commits will make similar changes to other file descriptor
operations, and modify the MAC framework to pass both credentials
to MAC policy modules so they can implement either semantic for
revocation.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs
2002-08-15 20:55:08 +00:00
|
|
|
if ((error = fo_read(fp, &auio, td->td_ucred, flags, td))) {
|
1999-03-27 21:16:58 +00:00
|
|
|
if (auio.uio_resid != cnt && (error == ERESTART ||
|
|
|
|
error == EINTR || error == EWOULDBLOCK))
|
|
|
|
error = 0;
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
1999-03-27 21:16:58 +00:00
|
|
|
cnt -= auio.uio_resid;
|
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (ktruio != NULL) {
|
|
|
|
ktruio->uio_resid = cnt;
|
|
|
|
ktrgenio(fd, UIO_READ, ktruio, error);
|
2000-07-02 08:08:09 +00:00
|
|
|
}
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = cnt;
|
1999-03-27 21:16:58 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Scatter read system call.
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct readv_args {
|
1995-11-11 06:57:34 +00:00
|
|
|
int fd;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct iovec *iovp;
|
|
|
|
u_int iovcnt;
|
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2004-07-10 15:42:16 +00:00
|
|
|
readv(struct thread *td, struct readv_args *uap)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *auio = NULL;
|
|
|
|
long cnt;
|
2002-05-09 02:30:41 +00:00
|
|
|
int error;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *ktruio = NULL;
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
|
|
|
|
2004-07-10 15:42:16 +00:00
|
|
|
error = fget_read(td, uap->fd, &fp);
|
|
|
|
if (error)
|
|
|
|
return (error);
|
|
|
|
error = copyinuio(uap->iovp, uap->iovcnt, &auio);
|
|
|
|
if (error) {
|
|
|
|
fdrop(fp, td);
|
2002-05-09 02:30:41 +00:00
|
|
|
return (error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2004-07-10 15:42:16 +00:00
|
|
|
auio->uio_rw = UIO_READ;
|
|
|
|
auio->uio_td = td;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (KTRPOINT(td, KTR_GENIO))
|
|
|
|
ktruio = cloneuio(auio);
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
2004-07-10 15:42:16 +00:00
|
|
|
cnt = auio->uio_resid;
|
|
|
|
if ((error = fo_read(fp, auio, td->td_ucred, 0, td))) {
|
|
|
|
if (auio->uio_resid != cnt && (error == ERESTART ||
|
1994-05-24 10:09:53 +00:00
|
|
|
error == EINTR || error == EWOULDBLOCK))
|
|
|
|
error = 0;
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
2004-07-10 15:42:16 +00:00
|
|
|
cnt -= auio->uio_resid;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (ktruio != NULL) {
|
|
|
|
ktruio->uio_resid = cnt;
|
|
|
|
ktrgenio(uap->fd, UIO_READ, ktruio, error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
#endif
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = cnt;
|
2004-07-10 15:42:16 +00:00
|
|
|
free(auio, M_IOV);
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
1994-05-24 10:09:53 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Write system call
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct write_args {
|
|
|
|
int fd;
|
1998-09-05 14:30:11 +00:00
|
|
|
const void *buf;
|
|
|
|
size_t nbyte;
|
1994-05-24 10:09:53 +00:00
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
write(td, uap)
|
|
|
|
struct thread *td;
|
2001-11-14 06:30:36 +00:00
|
|
|
struct write_args *uap;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2000-11-18 21:01:04 +00:00
|
|
|
int error;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2001-11-14 06:30:36 +00:00
|
|
|
if ((error = fget_write(td, uap->fd, &fp)) == 0) {
|
2001-09-12 08:38:13 +00:00
|
|
|
error = dofilewrite(td, fp, uap->fd, uap->buf, uap->nbyte,
|
2001-09-01 19:34:23 +00:00
|
|
|
(off_t)-1, 0);
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
2001-09-01 19:34:23 +00:00
|
|
|
} else {
|
2001-11-14 06:30:36 +00:00
|
|
|
error = EBADF; /* XXX this can't be right */
|
2001-09-01 19:34:23 +00:00
|
|
|
}
|
2000-11-18 21:01:04 +00:00
|
|
|
return(error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
1999-03-27 21:16:58 +00:00
|
|
|
/*
|
1999-04-04 21:41:28 +00:00
|
|
|
* Pwrite system call
|
1999-03-27 21:16:58 +00:00
|
|
|
*/
|
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
|
|
|
struct pwrite_args {
|
|
|
|
int fd;
|
|
|
|
const void *buf;
|
|
|
|
size_t nbyte;
|
1999-04-04 21:41:28 +00:00
|
|
|
int pad;
|
|
|
|
off_t offset;
|
1999-03-27 21:16:58 +00:00
|
|
|
};
|
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1999-03-27 21:16:58 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
pwrite(td, uap)
|
|
|
|
struct thread *td;
|
2001-11-14 06:30:36 +00:00
|
|
|
struct pwrite_args *uap;
|
1999-03-27 21:16:58 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2000-11-18 21:01:04 +00:00
|
|
|
int error;
|
1999-04-04 21:41:28 +00:00
|
|
|
|
2001-11-14 06:30:36 +00:00
|
|
|
if ((error = fget_write(td, uap->fd, &fp)) == 0) {
|
2004-01-20 01:27:42 +00:00
|
|
|
if (!(fp->f_ops->fo_flags & DFLAG_SEEKABLE))
|
2003-06-18 19:53:59 +00:00
|
|
|
error = ESPIPE;
|
2004-01-20 01:27:42 +00:00
|
|
|
else if (uap->offset < 0 && fp->f_vnode->v_type != VCHR)
|
|
|
|
error = EINVAL;
|
|
|
|
else {
|
2001-11-14 06:30:36 +00:00
|
|
|
error = dofilewrite(td, fp, uap->fd, uap->buf,
|
|
|
|
uap->nbyte, uap->offset, FOF_OFFSET);
|
|
|
|
}
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
2000-11-18 21:01:04 +00:00
|
|
|
} else {
|
2001-11-14 06:30:36 +00:00
|
|
|
error = EBADF; /* this can't be right */
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
|
|
|
return(error);
|
1999-04-04 21:41:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2001-09-12 08:38:13 +00:00
|
|
|
dofilewrite(td, fp, fd, buf, nbyte, offset, flags)
|
|
|
|
struct thread *td;
|
1999-04-04 21:41:28 +00:00
|
|
|
struct file *fp;
|
|
|
|
int fd, flags;
|
|
|
|
const void *buf;
|
|
|
|
size_t nbyte;
|
|
|
|
off_t offset;
|
|
|
|
{
|
1999-03-27 21:16:58 +00:00
|
|
|
struct uio auio;
|
|
|
|
struct iovec aiov;
|
|
|
|
long cnt, error = 0;
|
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *ktruio = NULL;
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
|
|
|
|
2000-07-28 22:17:42 +00:00
|
|
|
aiov.iov_base = (void *)(uintptr_t)buf;
|
1999-04-04 21:41:28 +00:00
|
|
|
aiov.iov_len = nbyte;
|
1999-03-27 21:16:58 +00:00
|
|
|
auio.uio_iov = &aiov;
|
|
|
|
auio.uio_iovcnt = 1;
|
1999-04-04 21:41:28 +00:00
|
|
|
auio.uio_offset = offset;
|
|
|
|
if (nbyte > INT_MAX)
|
1999-03-27 21:16:58 +00:00
|
|
|
return (EINVAL);
|
1999-04-04 21:41:28 +00:00
|
|
|
auio.uio_resid = nbyte;
|
1999-03-27 21:16:58 +00:00
|
|
|
auio.uio_rw = UIO_WRITE;
|
|
|
|
auio.uio_segflg = UIO_USERSPACE;
|
2001-09-12 08:38:13 +00:00
|
|
|
auio.uio_td = td;
|
1999-03-27 21:16:58 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (KTRPOINT(td, KTR_GENIO))
|
|
|
|
ktruio = cloneuio(&auio);
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
1999-04-04 21:41:28 +00:00
|
|
|
cnt = nbyte;
|
2000-11-30 20:23:14 +00:00
|
|
|
if (fp->f_type == DTYPE_VNODE)
|
|
|
|
bwillwrite();
|
In order to better support flexible and extensible access control,
make a series of modifications to the credential arguments relating
to file read and write operations to cliarfy which credential is
used for what:
- Change fo_read() and fo_write() to accept "active_cred" instead of
"cred", and change the semantics of consumers of fo_read() and
fo_write() to pass the active credential of the thread requesting
an operation rather than the cached file cred. The cached file
cred is still available in fo_read() and fo_write() consumers
via fp->f_cred. These changes largely in sys_generic.c.
For each implementation of fo_read() and fo_write(), update cred
usage to reflect this change and maintain current semantics:
- badfo_readwrite() unchanged
- kqueue_read/write() unchanged
pipe_read/write() now authorize MAC using active_cred rather
than td->td_ucred
- soo_read/write() unchanged
- vn_read/write() now authorize MAC using active_cred but
VOP_READ/WRITE() with fp->f_cred
Modify vn_rdwr() to accept two credential arguments instead of a
single credential: active_cred and file_cred. Use active_cred
for MAC authorization, and select a credential for use in
VOP_READ/WRITE() based on whether file_cred is NULL or not. If
file_cred is provided, authorize the VOP using that cred,
otherwise the active credential, matching current semantics.
Modify current vn_rdwr() consumers to pass a file_cred if used
in the context of a struct file, and to always pass active_cred.
When vn_rdwr() is used without a file_cred, pass NOCRED.
These changes should maintain current semantics for read/write,
but avoid a redundant passing of fp->f_cred, as well as making
it more clear what the origin of each credential is in file
descriptor read/write operations.
Follow-up commits will make similar changes to other file descriptor
operations, and modify the MAC framework to pass both credentials
to MAC policy modules so they can implement either semantic for
revocation.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs
2002-08-15 20:55:08 +00:00
|
|
|
if ((error = fo_write(fp, &auio, td->td_ucred, flags, td))) {
|
1999-03-27 21:16:58 +00:00
|
|
|
if (auio.uio_resid != cnt && (error == ERESTART ||
|
|
|
|
error == EINTR || error == EWOULDBLOCK))
|
|
|
|
error = 0;
|
2002-06-20 18:52:54 +00:00
|
|
|
/* Socket layer is responsible for issuing SIGPIPE. */
|
|
|
|
if (error == EPIPE && fp->f_type != DTYPE_SOCKET) {
|
2001-09-12 08:38:13 +00:00
|
|
|
PROC_LOCK(td->td_proc);
|
|
|
|
psignal(td->td_proc, SIGPIPE);
|
|
|
|
PROC_UNLOCK(td->td_proc);
|
2001-03-07 03:37:06 +00:00
|
|
|
}
|
1999-03-27 21:16:58 +00:00
|
|
|
}
|
|
|
|
cnt -= auio.uio_resid;
|
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (ktruio != NULL) {
|
|
|
|
ktruio->uio_resid = cnt;
|
|
|
|
ktrgenio(fd, UIO_WRITE, ktruio, error);
|
2000-07-02 08:08:09 +00:00
|
|
|
}
|
1999-03-27 21:16:58 +00:00
|
|
|
#endif
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = cnt;
|
1999-03-27 21:16:58 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Gather write system call
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct writev_args {
|
|
|
|
int fd;
|
|
|
|
struct iovec *iovp;
|
|
|
|
u_int iovcnt;
|
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2004-07-10 15:42:16 +00:00
|
|
|
writev(struct thread *td, struct writev_args *uap)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2001-11-14 06:30:36 +00:00
|
|
|
struct file *fp;
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *auio = NULL;
|
|
|
|
long cnt;
|
|
|
|
int error;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
struct uio *ktruio = NULL;
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
|
|
|
|
2004-07-10 15:42:16 +00:00
|
|
|
error = fget_write(td, uap->fd, &fp);
|
|
|
|
if (error)
|
2003-08-01 02:21:54 +00:00
|
|
|
return (EBADF);
|
2004-07-10 15:42:16 +00:00
|
|
|
error = copyinuio(uap->iovp, uap->iovcnt, &auio);
|
|
|
|
if (error) {
|
|
|
|
fdrop(fp, td);
|
|
|
|
return (error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2004-07-10 15:42:16 +00:00
|
|
|
auio->uio_rw = UIO_WRITE;
|
|
|
|
auio->uio_td = td;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (KTRPOINT(td, KTR_GENIO))
|
|
|
|
ktruio = cloneuio(auio);
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
2004-07-10 15:42:16 +00:00
|
|
|
cnt = auio->uio_resid;
|
2000-12-07 23:45:57 +00:00
|
|
|
if (fp->f_type == DTYPE_VNODE)
|
|
|
|
bwillwrite();
|
2004-07-10 15:42:16 +00:00
|
|
|
if ((error = fo_write(fp, auio, td->td_ucred, 0, td))) {
|
|
|
|
if (auio->uio_resid != cnt && (error == ERESTART ||
|
1994-05-24 10:09:53 +00:00
|
|
|
error == EINTR || error == EWOULDBLOCK))
|
|
|
|
error = 0;
|
2001-03-07 03:37:06 +00:00
|
|
|
if (error == EPIPE) {
|
2001-09-12 08:38:13 +00:00
|
|
|
PROC_LOCK(td->td_proc);
|
|
|
|
psignal(td->td_proc, SIGPIPE);
|
|
|
|
PROC_UNLOCK(td->td_proc);
|
2001-03-07 03:37:06 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2004-07-10 15:42:16 +00:00
|
|
|
cnt -= auio->uio_resid;
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef KTRACE
|
2004-07-10 15:42:16 +00:00
|
|
|
if (ktruio != NULL) {
|
|
|
|
ktruio->uio_resid = cnt;
|
|
|
|
ktrgenio(uap->fd, UIO_WRITE, ktruio, error);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
#endif
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = cnt;
|
|
|
|
fdrop(fp, td);
|
2004-07-10 15:42:16 +00:00
|
|
|
free(auio, M_IOV);
|
1994-05-24 10:09:53 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Ioctl system call
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct ioctl_args {
|
|
|
|
int fd;
|
1998-08-24 08:39:39 +00:00
|
|
|
u_long com;
|
1994-05-24 10:09:53 +00:00
|
|
|
caddr_t data;
|
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-24 10:09:53 +00:00
|
|
|
/* ARGSUSED */
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
ioctl(td, uap)
|
|
|
|
struct thread *td;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct ioctl_args *uap;
|
|
|
|
{
|
2002-01-14 00:13:45 +00:00
|
|
|
struct file *fp;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct filedesc *fdp;
|
1998-06-10 10:29:31 +00:00
|
|
|
register u_long com;
|
2001-09-01 19:34:23 +00:00
|
|
|
int error = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
register u_int size;
|
|
|
|
caddr_t data, memp;
|
|
|
|
int tmp;
|
|
|
|
#define STK_PARAMS 128
|
2000-05-09 17:43:21 +00:00
|
|
|
union {
|
|
|
|
char stkbuf[STK_PARAMS];
|
|
|
|
long align;
|
|
|
|
} ubuf;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-01-14 00:13:45 +00:00
|
|
|
if ((error = fget(td, uap->fd, &fp)) != 0)
|
|
|
|
return (error);
|
2002-01-19 01:03:54 +00:00
|
|
|
mtx_lock(&Giant);
|
2001-09-01 19:34:23 +00:00
|
|
|
if ((fp->f_flag & (FREAD | FWRITE)) == 0) {
|
2002-01-13 11:58:06 +00:00
|
|
|
fdrop(fp, td);
|
2002-01-19 01:03:54 +00:00
|
|
|
mtx_unlock(&Giant);
|
2002-01-13 11:58:06 +00:00
|
|
|
return (EBADF);
|
2001-09-01 19:34:23 +00:00
|
|
|
}
|
2002-01-13 11:58:06 +00:00
|
|
|
fdp = td->td_proc->p_fd;
|
1994-05-24 10:09:53 +00:00
|
|
|
switch (com = uap->com) {
|
|
|
|
case FIONCLEX:
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_LOCK(fdp);
|
1994-05-24 10:09:53 +00:00
|
|
|
fdp->fd_ofileflags[uap->fd] &= ~UF_EXCLOSE;
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_UNLOCK(fdp);
|
|
|
|
fdrop(fp, td);
|
2002-01-19 01:03:54 +00:00
|
|
|
mtx_unlock(&Giant);
|
2002-01-13 11:58:06 +00:00
|
|
|
return (0);
|
1994-05-24 10:09:53 +00:00
|
|
|
case FIOCLEX:
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_LOCK(fdp);
|
1994-05-24 10:09:53 +00:00
|
|
|
fdp->fd_ofileflags[uap->fd] |= UF_EXCLOSE;
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_UNLOCK(fdp);
|
|
|
|
fdrop(fp, td);
|
2002-01-19 01:03:54 +00:00
|
|
|
mtx_unlock(&Giant);
|
2002-01-13 11:58:06 +00:00
|
|
|
return (0);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Interpret high order word to find amount of data to be
|
|
|
|
* copied to/from the user's address space.
|
|
|
|
*/
|
|
|
|
size = IOCPARM_LEN(com);
|
2001-09-01 19:34:23 +00:00
|
|
|
if (size > IOCPARM_MAX) {
|
2002-01-13 11:58:06 +00:00
|
|
|
fdrop(fp, td);
|
2002-01-19 01:03:54 +00:00
|
|
|
mtx_unlock(&Giant);
|
2002-01-13 11:58:06 +00:00
|
|
|
return (ENOTTY);
|
2001-09-01 19:34:23 +00:00
|
|
|
}
|
2000-11-18 21:01:04 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
memp = NULL;
|
2000-05-09 17:43:21 +00:00
|
|
|
if (size > sizeof (ubuf.stkbuf)) {
|
2003-02-19 05:47:46 +00:00
|
|
|
memp = malloc((u_long)size, M_IOCTLOPS, M_WAITOK);
|
1994-05-24 10:09:53 +00:00
|
|
|
data = memp;
|
2000-11-18 21:01:04 +00:00
|
|
|
} else {
|
2000-05-09 17:43:21 +00:00
|
|
|
data = ubuf.stkbuf;
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
if (com&IOC_IN) {
|
|
|
|
if (size) {
|
|
|
|
error = copyin(uap->data, data, (u_int)size);
|
|
|
|
if (error) {
|
|
|
|
if (memp)
|
|
|
|
free(memp, M_IOCTLOPS);
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
2002-01-13 11:58:06 +00:00
|
|
|
goto done;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2000-11-18 21:01:04 +00:00
|
|
|
} else {
|
1994-05-24 10:09:53 +00:00
|
|
|
*(caddr_t *)data = uap->data;
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
|
|
|
} else if ((com&IOC_OUT) && size) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Zero the buffer so the user always
|
|
|
|
* gets back something deterministic.
|
|
|
|
*/
|
|
|
|
bzero(data, size);
|
2000-11-18 21:01:04 +00:00
|
|
|
} else if (com&IOC_VOID) {
|
1994-05-24 10:09:53 +00:00
|
|
|
*(caddr_t *)data = uap->data;
|
2000-11-18 21:01:04 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
switch (com) {
|
|
|
|
|
|
|
|
case FIONBIO:
|
2002-01-13 11:58:06 +00:00
|
|
|
FILE_LOCK(fp);
|
1994-09-25 19:34:02 +00:00
|
|
|
if ((tmp = *(int *)data))
|
1994-05-24 10:09:53 +00:00
|
|
|
fp->f_flag |= FNONBLOCK;
|
|
|
|
else
|
|
|
|
fp->f_flag &= ~FNONBLOCK;
|
2002-01-13 11:58:06 +00:00
|
|
|
FILE_UNLOCK(fp);
|
2002-08-17 02:36:16 +00:00
|
|
|
error = fo_ioctl(fp, FIONBIO, &tmp, td->td_ucred, td);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
case FIOASYNC:
|
2002-01-13 11:58:06 +00:00
|
|
|
FILE_LOCK(fp);
|
1994-09-25 19:34:02 +00:00
|
|
|
if ((tmp = *(int *)data))
|
1994-05-24 10:09:53 +00:00
|
|
|
fp->f_flag |= FASYNC;
|
|
|
|
else
|
|
|
|
fp->f_flag &= ~FASYNC;
|
2002-01-13 11:58:06 +00:00
|
|
|
FILE_UNLOCK(fp);
|
2002-08-17 02:36:16 +00:00
|
|
|
error = fo_ioctl(fp, FIOASYNC, &tmp, td->td_ucred, td);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
2002-08-17 02:36:16 +00:00
|
|
|
error = fo_ioctl(fp, com, data, td->td_ucred, td);
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Copy any data to user, size was
|
|
|
|
* already set and checked above.
|
|
|
|
*/
|
|
|
|
if (error == 0 && (com&IOC_OUT) && size)
|
|
|
|
error = copyout(data, uap->data, (u_int)size);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (memp)
|
|
|
|
free(memp, M_IOCTLOPS);
|
2001-09-12 08:38:13 +00:00
|
|
|
fdrop(fp, td);
|
2002-01-13 11:58:06 +00:00
|
|
|
done:
|
2001-09-01 19:34:23 +00:00
|
|
|
mtx_unlock(&Giant);
|
1994-05-24 10:09:53 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
2002-03-14 01:32:30 +00:00
|
|
|
/*
|
|
|
|
* sellock and selwait are initialized in selectinit() via SYSINIT.
|
|
|
|
*/
|
|
|
|
struct mtx sellock;
|
2001-05-14 05:26:48 +00:00
|
|
|
struct cv selwait;
|
2002-06-12 02:08:18 +00:00
|
|
|
u_int nselcoll; /* Select collisions since boot */
|
|
|
|
SYSCTL_UINT(_kern, OID_AUTO, nselcoll, CTLFLAG_RD, &nselcoll, 0, "");
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Select system call.
|
|
|
|
*/
|
1995-11-12 06:43:28 +00:00
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
1994-05-24 10:09:53 +00:00
|
|
|
struct select_args {
|
1996-08-20 07:18:10 +00:00
|
|
|
int nd;
|
1994-05-24 10:09:53 +00:00
|
|
|
fd_set *in, *ou, *ex;
|
|
|
|
struct timeval *tv;
|
|
|
|
};
|
1995-11-12 06:43:28 +00:00
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1994-05-25 09:21:21 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
select(td, uap)
|
|
|
|
register struct thread *td;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct select_args *uap;
|
2002-09-01 20:37:28 +00:00
|
|
|
{
|
|
|
|
struct timeval tv, *tvp;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
if (uap->tv != NULL) {
|
|
|
|
error = copyin(uap->tv, &tv, sizeof(tv));
|
|
|
|
if (error)
|
|
|
|
return (error);
|
|
|
|
tvp = &tv;
|
|
|
|
} else
|
|
|
|
tvp = NULL;
|
|
|
|
|
|
|
|
return (kern_select(td, uap->nd, uap->in, uap->ou, uap->ex, tvp));
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
kern_select(struct thread *td, int nd, fd_set *fd_in, fd_set *fd_ou,
|
|
|
|
fd_set *fd_ex, struct timeval *tvp)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2002-01-13 11:58:06 +00:00
|
|
|
struct filedesc *fdp;
|
1997-02-20 11:51:52 +00:00
|
|
|
/*
|
|
|
|
* The magic 2048 here is chosen to be just enough for FD_SETSIZE
|
|
|
|
* infds with the new FD_SETSIZE of 1024, and more than enough for
|
|
|
|
* FD_SETSIZE infds, outfds and exceptfds with the old FD_SETSIZE
|
|
|
|
* of 256.
|
|
|
|
*/
|
|
|
|
fd_mask s_selbits[howmany(2048, NFDBITS)];
|
2002-01-29 22:54:19 +00:00
|
|
|
fd_mask *ibits[3], *obits[3], *selbits, *sbp;
|
1998-04-04 13:26:20 +00:00
|
|
|
struct timeval atv, rtv, ttv;
|
2002-06-12 02:08:18 +00:00
|
|
|
int error, timo;
|
|
|
|
u_int ncoll, nbufbytes, ncpbytes, nfdbits;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-09-01 20:37:28 +00:00
|
|
|
if (nd < 0)
|
1996-08-20 15:03:41 +00:00
|
|
|
return (EINVAL);
|
2002-01-13 11:58:06 +00:00
|
|
|
fdp = td->td_proc->p_fd;
|
2004-03-13 05:58:57 +00:00
|
|
|
/*
|
|
|
|
* XXX: kern_select() currently requires that we acquire Giant
|
|
|
|
* even if none of the file descriptors we poll requires Giant.
|
|
|
|
*/
|
2001-09-01 19:34:23 +00:00
|
|
|
mtx_lock(&Giant);
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_LOCK(fdp);
|
2001-09-01 19:34:23 +00:00
|
|
|
|
2002-09-01 20:37:28 +00:00
|
|
|
if (nd > td->td_proc->p_fd->fd_nfiles)
|
|
|
|
nd = td->td_proc->p_fd->fd_nfiles; /* forgiving; slightly wrong */
|
2002-01-13 11:58:06 +00:00
|
|
|
FILEDESC_UNLOCK(fdp);
|
1996-08-20 07:18:10 +00:00
|
|
|
|
|
|
|
/*
|
1997-02-20 11:51:52 +00:00
|
|
|
* Allocate just enough bits for the non-null fd_sets. Use the
|
|
|
|
* preallocated auto buffer if possible.
|
1996-08-20 07:18:10 +00:00
|
|
|
*/
|
2002-09-01 20:37:28 +00:00
|
|
|
nfdbits = roundup(nd, NFDBITS);
|
1997-02-20 11:51:52 +00:00
|
|
|
ncpbytes = nfdbits / NBBY;
|
|
|
|
nbufbytes = 0;
|
2002-09-01 20:37:28 +00:00
|
|
|
if (fd_in != NULL)
|
1997-02-20 11:51:52 +00:00
|
|
|
nbufbytes += 2 * ncpbytes;
|
2002-09-01 20:37:28 +00:00
|
|
|
if (fd_ou != NULL)
|
1997-02-20 11:51:52 +00:00
|
|
|
nbufbytes += 2 * ncpbytes;
|
2002-09-01 20:37:28 +00:00
|
|
|
if (fd_ex != NULL)
|
1997-02-20 11:51:52 +00:00
|
|
|
nbufbytes += 2 * ncpbytes;
|
|
|
|
if (nbufbytes <= sizeof s_selbits)
|
|
|
|
selbits = &s_selbits[0];
|
|
|
|
else
|
2003-02-19 05:47:46 +00:00
|
|
|
selbits = malloc(nbufbytes, M_SELECT, M_WAITOK);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1997-02-20 11:51:52 +00:00
|
|
|
/*
|
|
|
|
* Assign pointers into the bit buffers and fetch the input bits.
|
|
|
|
* Put the output buffers together so that they can be bzeroed
|
|
|
|
* together.
|
|
|
|
*/
|
|
|
|
sbp = selbits;
|
1994-05-24 10:09:53 +00:00
|
|
|
#define getbits(name, x) \
|
1997-02-20 11:51:52 +00:00
|
|
|
do { \
|
2002-09-01 20:37:28 +00:00
|
|
|
if (name == NULL) \
|
1997-02-20 11:51:52 +00:00
|
|
|
ibits[x] = NULL; \
|
|
|
|
else { \
|
|
|
|
ibits[x] = sbp + nbufbytes / 2 / sizeof *sbp; \
|
|
|
|
obits[x] = sbp; \
|
|
|
|
sbp += ncpbytes / sizeof *sbp; \
|
2002-09-01 20:37:28 +00:00
|
|
|
error = copyin(name, ibits[x], ncpbytes); \
|
2001-05-14 05:26:48 +00:00
|
|
|
if (error != 0) \
|
2002-03-14 01:32:30 +00:00
|
|
|
goto done_nosellock; \
|
1997-02-20 11:51:52 +00:00
|
|
|
} \
|
|
|
|
} while (0)
|
2002-09-01 20:37:28 +00:00
|
|
|
getbits(fd_in, 0);
|
|
|
|
getbits(fd_ou, 1);
|
|
|
|
getbits(fd_ex, 2);
|
1994-05-24 10:09:53 +00:00
|
|
|
#undef getbits
|
1997-02-20 11:51:52 +00:00
|
|
|
if (nbufbytes != 0)
|
|
|
|
bzero(selbits, nbufbytes / 2);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-09-01 20:37:28 +00:00
|
|
|
if (tvp != NULL) {
|
|
|
|
atv = *tvp;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (itimerfix(&atv)) {
|
|
|
|
error = EINVAL;
|
2002-03-14 01:32:30 +00:00
|
|
|
goto done_nosellock;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1998-05-17 11:53:46 +00:00
|
|
|
getmicrouptime(&rtv);
|
1998-04-04 13:26:20 +00:00
|
|
|
timevaladd(&atv, &rtv);
|
2000-07-13 02:12:25 +00:00
|
|
|
} else {
|
1998-04-04 13:26:20 +00:00
|
|
|
atv.tv_sec = 0;
|
2000-07-13 02:12:25 +00:00
|
|
|
atv.tv_usec = 0;
|
|
|
|
}
|
1998-04-04 13:26:20 +00:00
|
|
|
timo = 0;
|
2002-08-23 22:43:28 +00:00
|
|
|
TAILQ_INIT(&td->td_selq);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
1994-05-24 10:09:53 +00:00
|
|
|
retry:
|
|
|
|
ncoll = nselcoll;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_flags |= TDF_SELECT;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock(&sellock);
|
|
|
|
|
2002-09-01 20:37:28 +00:00
|
|
|
error = selscan(td, ibits, obits, nd);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
2001-09-12 08:38:13 +00:00
|
|
|
if (error || td->td_retval[0])
|
1994-05-24 10:09:53 +00:00
|
|
|
goto done;
|
2000-07-12 22:46:40 +00:00
|
|
|
if (atv.tv_sec || atv.tv_usec) {
|
1998-05-17 11:53:46 +00:00
|
|
|
getmicrouptime(&rtv);
|
2002-03-14 01:32:30 +00:00
|
|
|
if (timevalcmp(&rtv, &atv, >=))
|
1998-04-04 13:26:20 +00:00
|
|
|
goto done;
|
|
|
|
ttv = atv;
|
|
|
|
timevalsub(&ttv, &rtv);
|
|
|
|
timo = ttv.tv_sec > 24 * 60 * 60 ?
|
|
|
|
24 * 60 * 60 * hz : tvtohz(&ttv);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2002-03-14 01:32:30 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* An event of interest may occur while we do not hold
|
|
|
|
* sellock, so check TDF_SELECT and the number of
|
|
|
|
* collisions and rescan the file descriptors if
|
|
|
|
* necessary.
|
|
|
|
*/
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
if ((td->td_flags & TDF_SELECT) == 0 || nselcoll != ncoll) {
|
|
|
|
mtx_unlock_spin(&sched_lock);
|
|
|
|
goto retry;
|
|
|
|
}
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2000-01-14 02:53:29 +00:00
|
|
|
|
2001-05-14 05:26:48 +00:00
|
|
|
if (timo > 0)
|
2002-03-14 01:32:30 +00:00
|
|
|
error = cv_timedwait_sig(&selwait, &sellock, timo);
|
2001-05-14 05:26:48 +00:00
|
|
|
else
|
2002-03-14 01:32:30 +00:00
|
|
|
error = cv_wait_sig(&selwait, &sellock);
|
2000-01-14 02:53:29 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
if (error == 0)
|
|
|
|
goto retry;
|
2001-05-14 05:26:48 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
done:
|
2002-03-14 01:32:30 +00:00
|
|
|
clear_selinfo_list(td);
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_flags &= ~TDF_SELECT;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock(&sellock);
|
|
|
|
|
|
|
|
done_nosellock:
|
1994-05-24 10:09:53 +00:00
|
|
|
/* select is not restarted after signals... */
|
|
|
|
if (error == ERESTART)
|
|
|
|
error = EINTR;
|
|
|
|
if (error == EWOULDBLOCK)
|
|
|
|
error = 0;
|
|
|
|
#define putbits(name, x) \
|
2002-09-01 20:37:28 +00:00
|
|
|
if (name && (error2 = copyout(obits[x], name, ncpbytes))) \
|
1994-05-24 10:09:53 +00:00
|
|
|
error = error2;
|
|
|
|
if (error == 0) {
|
|
|
|
int error2;
|
|
|
|
|
2002-09-01 20:37:28 +00:00
|
|
|
putbits(fd_in, 0);
|
|
|
|
putbits(fd_ou, 1);
|
|
|
|
putbits(fd_ex, 2);
|
1994-05-24 10:09:53 +00:00
|
|
|
#undef putbits
|
|
|
|
}
|
1997-02-20 11:51:52 +00:00
|
|
|
if (selbits != &s_selbits[0])
|
|
|
|
free(selbits, M_SELECT);
|
2001-09-01 19:34:23 +00:00
|
|
|
|
|
|
|
mtx_unlock(&Giant);
|
1994-05-24 10:09:53 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
1995-12-14 08:32:45 +00:00
|
|
|
static int
|
2001-09-12 08:38:13 +00:00
|
|
|
selscan(td, ibits, obits, nfd)
|
|
|
|
struct thread *td;
|
1996-08-20 07:18:10 +00:00
|
|
|
fd_mask **ibits, **obits;
|
1997-11-06 19:29:57 +00:00
|
|
|
int nfd;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2000-02-20 13:36:26 +00:00
|
|
|
int msk, i, fd;
|
|
|
|
fd_mask bits;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct file *fp;
|
|
|
|
int n = 0;
|
1997-11-23 10:30:50 +00:00
|
|
|
/* Note: backend also returns POLLHUP/POLLERR if appropriate. */
|
|
|
|
static int flag[3] = { POLLRDNORM, POLLWRNORM, POLLRDBAND };
|
2002-01-29 22:54:19 +00:00
|
|
|
struct filedesc *fdp = td->td_proc->p_fd;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-01-29 22:54:19 +00:00
|
|
|
FILEDESC_LOCK(fdp);
|
1994-05-24 10:09:53 +00:00
|
|
|
for (msk = 0; msk < 3; msk++) {
|
1997-02-20 11:51:52 +00:00
|
|
|
if (ibits[msk] == NULL)
|
|
|
|
continue;
|
1994-05-24 10:09:53 +00:00
|
|
|
for (i = 0; i < nfd; i += NFDBITS) {
|
1996-08-20 07:18:10 +00:00
|
|
|
bits = ibits[msk][i/NFDBITS];
|
2000-02-20 13:36:26 +00:00
|
|
|
/* ffs(int mask) not portable, fd_mask is long */
|
|
|
|
for (fd = i; bits && fd < nfd; fd++, bits >>= 1) {
|
|
|
|
if (!(bits & 1))
|
|
|
|
continue;
|
2002-01-29 22:54:19 +00:00
|
|
|
if ((fp = fget_locked(fdp, fd)) == NULL) {
|
|
|
|
FILEDESC_UNLOCK(fdp);
|
1994-05-24 10:09:53 +00:00
|
|
|
return (EBADF);
|
2002-01-29 22:54:19 +00:00
|
|
|
}
|
Make similar changes to fo_stat() and fo_poll() as made earlier to
fo_read() and fo_write(): explicitly use the cred argument to fo_poll()
as "active_cred" using the passed file descriptor's f_cred reference
to provide access to the file credential. Add an active_cred
argument to fo_stat() so that implementers have access to the active
credential as well as the file credential. Generally modify callers
of fo_stat() to pass in td->td_ucred rather than fp->f_cred, which
was redundantly provided via the fp argument. This set of modifications
also permits threads to perform these operations on behalf of another
thread without modifying their credential.
Trickle this change down into fo_stat/poll() implementations:
- badfo_poll(), badfo_stat(): modify/add arguments.
- kqueue_poll(), kqueue_stat(): modify arguments.
- pipe_poll(), pipe_stat(): modify/add arguments, pass active_cred to
MAC checks rather than td->td_ucred.
- soo_poll(), soo_stat(): modify/add arguments, pass fp->f_cred rather
than cred to pru_sopoll() to maintain current semantics.
- sopoll(): moidfy arguments.
- vn_poll(), vn_statfile(): modify/add arguments, pass new arguments
to vn_stat(). Pass active_cred to MAC and fp->f_cred to VOP_POLL()
to maintian current semantics.
- vn_close(): rename cred to file_cred to reflect reality while I'm here.
- vn_stat(): Add active_cred and file_cred arguments to vn_stat()
and consumers so that this distinction is maintained at the VFS
as well as 'struct file' layer. Pass active_cred instead of
td->td_ucred to MAC and to VOP_GETATTR() to maintain current semantics.
- fifofs: modify the creation of a "filetemp" so that the file
credential is properly initialized and can be used in the socket
code if desired. Pass ap->a_td->td_ucred as the active
credential to soo_poll(). If we teach the vnop interface about
the distinction between file and active credentials, we would use
the active credential here.
Note that current inconsistent passing of active_cred vs. file_cred to
VOP's is maintained. It's not clear why GETATTR would be authorized
using active_cred while POLL would be authorized using file_cred at
the file system level.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs
2002-08-16 12:52:03 +00:00
|
|
|
if (fo_poll(fp, flag[msk], td->td_ucred,
|
|
|
|
td)) {
|
1997-02-20 11:51:52 +00:00
|
|
|
obits[msk][(fd)/NFDBITS] |=
|
2000-02-20 13:36:26 +00:00
|
|
|
((fd_mask)1 << ((fd) % NFDBITS));
|
1994-05-24 10:09:53 +00:00
|
|
|
n++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2002-01-29 22:54:19 +00:00
|
|
|
FILEDESC_UNLOCK(fdp);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = n;
|
1994-05-24 10:09:53 +00:00
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
1997-09-14 02:30:32 +00:00
|
|
|
/*
|
|
|
|
* Poll system call.
|
|
|
|
*/
|
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
|
|
|
struct poll_args {
|
|
|
|
struct pollfd *fds;
|
|
|
|
u_int nfds;
|
|
|
|
int timeout;
|
|
|
|
};
|
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1997-09-14 02:30:32 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
poll(td, uap)
|
|
|
|
struct thread *td;
|
2001-02-27 15:11:31 +00:00
|
|
|
struct poll_args *uap;
|
1997-09-14 02:30:32 +00:00
|
|
|
{
|
2004-08-27 21:23:50 +00:00
|
|
|
struct pollfd *bits;
|
|
|
|
struct pollfd smallbits[32];
|
1998-04-04 13:26:20 +00:00
|
|
|
struct timeval atv, rtv, ttv;
|
2002-06-12 02:08:18 +00:00
|
|
|
int error = 0, timo;
|
|
|
|
u_int ncoll, nfds;
|
1997-09-14 02:30:32 +00:00
|
|
|
size_t ni;
|
|
|
|
|
2002-12-14 01:56:26 +00:00
|
|
|
nfds = uap->nfds;
|
2001-09-01 19:34:23 +00:00
|
|
|
|
2004-03-13 05:58:57 +00:00
|
|
|
/*
|
|
|
|
* XXX: poll() currently requires that we acquire Giant even if
|
|
|
|
* none of the file descriptors we poll requires Giant.
|
|
|
|
*/
|
2001-09-01 19:34:23 +00:00
|
|
|
mtx_lock(&Giant);
|
2001-02-07 23:28:01 +00:00
|
|
|
/*
|
2001-02-09 08:10:22 +00:00
|
|
|
* This is kinda bogus. We have fd limits, but that is not
|
|
|
|
* really related to the size of the pollfd array. Make sure
|
|
|
|
* we let the process use at least FD_SETSIZE entries and at
|
|
|
|
* least enough for the current limits. We want to be reasonably
|
|
|
|
* safe, but not overly restrictive.
|
2001-02-07 23:28:01 +00:00
|
|
|
*/
|
2004-02-04 21:52:57 +00:00
|
|
|
PROC_LOCK(td->td_proc);
|
|
|
|
if ((nfds > lim_cur(td->td_proc, RLIMIT_NOFILE)) &&
|
2001-09-12 08:38:13 +00:00
|
|
|
(nfds > FD_SETSIZE)) {
|
2004-02-04 21:52:57 +00:00
|
|
|
PROC_UNLOCK(td->td_proc);
|
2001-09-01 19:34:23 +00:00
|
|
|
error = EINVAL;
|
|
|
|
goto done2;
|
|
|
|
}
|
2004-02-04 21:52:57 +00:00
|
|
|
PROC_UNLOCK(td->td_proc);
|
2001-02-07 23:28:01 +00:00
|
|
|
ni = nfds * sizeof(struct pollfd);
|
1997-09-14 02:30:32 +00:00
|
|
|
if (ni > sizeof(smallbits))
|
2003-02-19 05:47:46 +00:00
|
|
|
bits = malloc(ni, M_TEMP, M_WAITOK);
|
1997-09-14 02:30:32 +00:00
|
|
|
else
|
|
|
|
bits = smallbits;
|
2002-12-14 01:56:26 +00:00
|
|
|
error = copyin(uap->fds, bits, ni);
|
1997-09-14 02:30:32 +00:00
|
|
|
if (error)
|
2002-03-14 01:32:30 +00:00
|
|
|
goto done_nosellock;
|
2002-12-14 01:56:26 +00:00
|
|
|
if (uap->timeout != INFTIM) {
|
|
|
|
atv.tv_sec = uap->timeout / 1000;
|
|
|
|
atv.tv_usec = (uap->timeout % 1000) * 1000;
|
1997-09-14 02:30:32 +00:00
|
|
|
if (itimerfix(&atv)) {
|
|
|
|
error = EINVAL;
|
2002-03-14 01:32:30 +00:00
|
|
|
goto done_nosellock;
|
1997-09-14 02:30:32 +00:00
|
|
|
}
|
1998-05-17 11:53:46 +00:00
|
|
|
getmicrouptime(&rtv);
|
1998-04-04 13:26:20 +00:00
|
|
|
timevaladd(&atv, &rtv);
|
2000-07-13 02:12:25 +00:00
|
|
|
} else {
|
1998-04-04 13:26:20 +00:00
|
|
|
atv.tv_sec = 0;
|
2000-07-13 02:12:25 +00:00
|
|
|
atv.tv_usec = 0;
|
|
|
|
}
|
1998-04-04 13:26:20 +00:00
|
|
|
timo = 0;
|
2002-08-23 22:43:28 +00:00
|
|
|
TAILQ_INIT(&td->td_selq);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
1997-09-14 02:30:32 +00:00
|
|
|
retry:
|
|
|
|
ncoll = nselcoll;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_flags |= TDF_SELECT;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock(&sellock);
|
|
|
|
|
2004-08-27 21:23:50 +00:00
|
|
|
error = pollscan(td, bits, nfds);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
2001-09-12 08:38:13 +00:00
|
|
|
if (error || td->td_retval[0])
|
1997-09-14 02:30:32 +00:00
|
|
|
goto done;
|
2000-07-12 22:46:40 +00:00
|
|
|
if (atv.tv_sec || atv.tv_usec) {
|
1998-05-17 11:53:46 +00:00
|
|
|
getmicrouptime(&rtv);
|
2002-03-14 01:32:30 +00:00
|
|
|
if (timevalcmp(&rtv, &atv, >=))
|
1998-04-04 13:26:20 +00:00
|
|
|
goto done;
|
|
|
|
ttv = atv;
|
|
|
|
timevalsub(&ttv, &rtv);
|
|
|
|
timo = ttv.tv_sec > 24 * 60 * 60 ?
|
|
|
|
24 * 60 * 60 * hz : tvtohz(&ttv);
|
1997-09-14 02:30:32 +00:00
|
|
|
}
|
2002-03-14 01:32:30 +00:00
|
|
|
/*
|
|
|
|
* An event of interest may occur while we do not hold
|
|
|
|
* sellock, so check TDF_SELECT and the number of collisions
|
|
|
|
* and rescan the file descriptors if necessary.
|
|
|
|
*/
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
if ((td->td_flags & TDF_SELECT) == 0 || nselcoll != ncoll) {
|
|
|
|
mtx_unlock_spin(&sched_lock);
|
|
|
|
goto retry;
|
|
|
|
}
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
|
2001-05-14 05:26:48 +00:00
|
|
|
if (timo > 0)
|
2002-03-14 01:32:30 +00:00
|
|
|
error = cv_timedwait_sig(&selwait, &sellock, timo);
|
2001-05-14 05:26:48 +00:00
|
|
|
else
|
2002-03-14 01:32:30 +00:00
|
|
|
error = cv_wait_sig(&selwait, &sellock);
|
|
|
|
|
1997-09-14 02:30:32 +00:00
|
|
|
if (error == 0)
|
|
|
|
goto retry;
|
2001-05-14 05:26:48 +00:00
|
|
|
|
1997-09-14 02:30:32 +00:00
|
|
|
done:
|
2002-03-14 01:32:30 +00:00
|
|
|
clear_selinfo_list(td);
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_flags &= ~TDF_SELECT;
|
2001-09-21 22:06:22 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock(&sellock);
|
|
|
|
|
|
|
|
done_nosellock:
|
1997-09-14 02:30:32 +00:00
|
|
|
/* poll is not restarted after signals... */
|
|
|
|
if (error == ERESTART)
|
|
|
|
error = EINTR;
|
|
|
|
if (error == EWOULDBLOCK)
|
|
|
|
error = 0;
|
|
|
|
if (error == 0) {
|
2002-12-14 01:56:26 +00:00
|
|
|
error = copyout(bits, uap->fds, ni);
|
1997-09-14 02:30:32 +00:00
|
|
|
if (error)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
if (ni > sizeof(smallbits))
|
|
|
|
free(bits, M_TEMP);
|
2001-09-01 19:34:23 +00:00
|
|
|
done2:
|
|
|
|
mtx_unlock(&Giant);
|
1997-09-14 02:30:32 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2001-09-12 08:38:13 +00:00
|
|
|
pollscan(td, fds, nfd)
|
|
|
|
struct thread *td;
|
1997-09-14 02:30:32 +00:00
|
|
|
struct pollfd *fds;
|
2001-02-27 15:11:31 +00:00
|
|
|
u_int nfd;
|
1997-09-14 02:30:32 +00:00
|
|
|
{
|
2001-09-12 08:38:13 +00:00
|
|
|
register struct filedesc *fdp = td->td_proc->p_fd;
|
1997-09-14 02:30:32 +00:00
|
|
|
int i;
|
|
|
|
struct file *fp;
|
|
|
|
int n = 0;
|
|
|
|
|
2002-01-29 22:54:19 +00:00
|
|
|
FILEDESC_LOCK(fdp);
|
1997-09-14 02:30:32 +00:00
|
|
|
for (i = 0; i < nfd; i++, fds++) {
|
1998-12-10 01:53:26 +00:00
|
|
|
if (fds->fd >= fdp->fd_nfiles) {
|
1997-09-14 02:30:32 +00:00
|
|
|
fds->revents = POLLNVAL;
|
|
|
|
n++;
|
1998-12-10 01:53:26 +00:00
|
|
|
} else if (fds->fd < 0) {
|
|
|
|
fds->revents = 0;
|
1997-09-14 02:30:32 +00:00
|
|
|
} else {
|
|
|
|
fp = fdp->fd_ofiles[fds->fd];
|
2000-11-18 21:01:04 +00:00
|
|
|
if (fp == NULL) {
|
1997-09-14 02:30:32 +00:00
|
|
|
fds->revents = POLLNVAL;
|
|
|
|
n++;
|
|
|
|
} else {
|
1997-11-23 10:30:50 +00:00
|
|
|
/*
|
|
|
|
* Note: backend also returns POLLHUP and
|
|
|
|
* POLLERR if appropriate.
|
|
|
|
*/
|
This is what was "fdfix2.patch," a fix for fd sharing. It's pretty
far-reaching in fd-land, so you'll want to consult the code for
changes. The biggest change is that now, you don't use
fp->f_ops->fo_foo(fp, bar)
but instead
fo_foo(fp, bar),
which increments and decrements the fp refcount upon entry and exit.
Two new calls, fhold() and fdrop(), are provided. Each does what it
seems like it should, and if fdrop() brings the refcount to zero, the
fd is freed as well.
Thanks to peter ("to hell with it, it looks ok to me.") for his review.
Thanks to msmith for keeping me from putting locks everywhere :)
Reviewed by: peter
1999-09-19 17:00:25 +00:00
|
|
|
fds->revents = fo_poll(fp, fds->events,
|
Make similar changes to fo_stat() and fo_poll() as made earlier to
fo_read() and fo_write(): explicitly use the cred argument to fo_poll()
as "active_cred" using the passed file descriptor's f_cred reference
to provide access to the file credential. Add an active_cred
argument to fo_stat() so that implementers have access to the active
credential as well as the file credential. Generally modify callers
of fo_stat() to pass in td->td_ucred rather than fp->f_cred, which
was redundantly provided via the fp argument. This set of modifications
also permits threads to perform these operations on behalf of another
thread without modifying their credential.
Trickle this change down into fo_stat/poll() implementations:
- badfo_poll(), badfo_stat(): modify/add arguments.
- kqueue_poll(), kqueue_stat(): modify arguments.
- pipe_poll(), pipe_stat(): modify/add arguments, pass active_cred to
MAC checks rather than td->td_ucred.
- soo_poll(), soo_stat(): modify/add arguments, pass fp->f_cred rather
than cred to pru_sopoll() to maintain current semantics.
- sopoll(): moidfy arguments.
- vn_poll(), vn_statfile(): modify/add arguments, pass new arguments
to vn_stat(). Pass active_cred to MAC and fp->f_cred to VOP_POLL()
to maintian current semantics.
- vn_close(): rename cred to file_cred to reflect reality while I'm here.
- vn_stat(): Add active_cred and file_cred arguments to vn_stat()
and consumers so that this distinction is maintained at the VFS
as well as 'struct file' layer. Pass active_cred instead of
td->td_ucred to MAC and to VOP_GETATTR() to maintain current semantics.
- fifofs: modify the creation of a "filetemp" so that the file
credential is properly initialized and can be used in the socket
code if desired. Pass ap->a_td->td_ucred as the active
credential to soo_poll(). If we teach the vnop interface about
the distinction between file and active credentials, we would use
the active credential here.
Note that current inconsistent passing of active_cred vs. file_cred to
VOP's is maintained. It's not clear why GETATTR would be authorized
using active_cred while POLL would be authorized using file_cred at
the file system level.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs
2002-08-16 12:52:03 +00:00
|
|
|
td->td_ucred, td);
|
1997-09-14 02:30:32 +00:00
|
|
|
if (fds->revents != 0)
|
|
|
|
n++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2002-01-29 22:54:19 +00:00
|
|
|
FILEDESC_UNLOCK(fdp);
|
2001-09-12 08:38:13 +00:00
|
|
|
td->td_retval[0] = n;
|
1997-09-14 02:30:32 +00:00
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* OpenBSD poll system call.
|
|
|
|
* XXX this isn't quite a true representation.. OpenBSD uses select ops.
|
|
|
|
*/
|
|
|
|
#ifndef _SYS_SYSPROTO_H_
|
|
|
|
struct openbsd_poll_args {
|
|
|
|
struct pollfd *fds;
|
|
|
|
u_int nfds;
|
|
|
|
int timeout;
|
|
|
|
};
|
|
|
|
#endif
|
2001-09-01 19:34:23 +00:00
|
|
|
/*
|
|
|
|
* MPSAFE
|
|
|
|
*/
|
1997-09-14 02:30:32 +00:00
|
|
|
int
|
2001-09-12 08:38:13 +00:00
|
|
|
openbsd_poll(td, uap)
|
|
|
|
register struct thread *td;
|
1997-09-14 02:30:32 +00:00
|
|
|
register struct openbsd_poll_args *uap;
|
|
|
|
{
|
2001-09-12 08:38:13 +00:00
|
|
|
return (poll(td, (struct poll_args *)uap));
|
1997-09-14 02:30:32 +00:00
|
|
|
}
|
|
|
|
|
2002-03-14 01:32:30 +00:00
|
|
|
/*
|
|
|
|
* Remove the references to the thread from all of the objects
|
|
|
|
* we were polling.
|
|
|
|
*
|
|
|
|
* This code assumes that the underlying owner of the selinfo
|
|
|
|
* structure will hold sellock before it changes it, and that
|
|
|
|
* it will unlink itself from our list if it goes away.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
clear_selinfo_list(td)
|
|
|
|
struct thread *td;
|
|
|
|
{
|
|
|
|
struct selinfo *si;
|
|
|
|
|
|
|
|
mtx_assert(&sellock, MA_OWNED);
|
|
|
|
TAILQ_FOREACH(si, &td->td_selq, si_thrlist)
|
|
|
|
si->si_thread = NULL;
|
|
|
|
TAILQ_INIT(&td->td_selq);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Record a select request.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
selrecord(selector, sip)
|
2001-09-12 08:38:13 +00:00
|
|
|
struct thread *selector;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct selinfo *sip;
|
|
|
|
{
|
|
|
|
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
|
|
|
/*
|
2002-07-24 00:29:22 +00:00
|
|
|
* If the selinfo's thread pointer is NULL then take ownership of it.
|
|
|
|
*
|
|
|
|
* If the thread pointer is not NULL and it points to another
|
|
|
|
* thread, then we have a collision.
|
|
|
|
*
|
|
|
|
* If the thread pointer is not NULL and points back to us then leave
|
|
|
|
* it alone as we've already added pointed it at us and added it to
|
|
|
|
* our list.
|
2002-03-14 01:32:30 +00:00
|
|
|
*/
|
|
|
|
if (sip->si_thread == NULL) {
|
|
|
|
sip->si_thread = selector;
|
|
|
|
TAILQ_INSERT_TAIL(&selector->td_selq, sip, si_thrlist);
|
|
|
|
} else if (sip->si_thread != selector) {
|
|
|
|
sip->si_flags |= SI_COLL;
|
2001-01-24 11:12:37 +00:00
|
|
|
}
|
2002-03-14 01:32:30 +00:00
|
|
|
|
|
|
|
mtx_unlock(&sellock);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2003-11-09 09:17:26 +00:00
|
|
|
/* Wake up a selecting thread. */
|
|
|
|
void
|
|
|
|
selwakeup(sip)
|
|
|
|
struct selinfo *sip;
|
|
|
|
{
|
|
|
|
doselwakeup(sip, -1);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Wake up a selecting thread, and set its priority. */
|
|
|
|
void
|
|
|
|
selwakeuppri(sip, pri)
|
|
|
|
struct selinfo *sip;
|
|
|
|
int pri;
|
|
|
|
{
|
|
|
|
doselwakeup(sip, pri);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Do a wakeup when a selectable event occurs.
|
|
|
|
*/
|
2003-11-09 09:17:26 +00:00
|
|
|
static void
|
|
|
|
doselwakeup(sip, pri)
|
2002-03-14 01:32:30 +00:00
|
|
|
struct selinfo *sip;
|
2003-11-09 09:17:26 +00:00
|
|
|
int pri;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2001-09-12 08:38:13 +00:00
|
|
|
struct thread *td;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_lock(&sellock);
|
|
|
|
td = sip->si_thread;
|
|
|
|
if ((sip->si_flags & SI_COLL) != 0) {
|
1994-05-24 10:09:53 +00:00
|
|
|
nselcoll++;
|
|
|
|
sip->si_flags &= ~SI_COLL;
|
2003-11-09 09:17:26 +00:00
|
|
|
cv_broadcastpri(&selwait, pri);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2002-03-14 01:32:30 +00:00
|
|
|
if (td == NULL) {
|
|
|
|
mtx_unlock(&sellock);
|
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2002-03-14 01:32:30 +00:00
|
|
|
TAILQ_REMOVE(&td->td_selq, sip, si_thrlist);
|
|
|
|
sip->si_thread = NULL;
|
|
|
|
mtx_lock_spin(&sched_lock);
|
Switch the sleep/wakeup and condition variable implementations to use the
sleep queue interface:
- Sleep queues attempt to merge some of the benefits of both sleep queues
and condition variables. Having sleep qeueus in a hash table avoids
having to allocate a queue head for each wait channel. Thus, struct cv
has shrunk down to just a single char * pointer now. However, the
hash table does not hold threads directly, but queue heads. This means
that once you have located a queue in the hash bucket, you no longer have
to walk the rest of the hash chain looking for threads. Instead, you have
a list of all the threads sleeping on that wait channel.
- Outside of the sleepq code and the sleep/cv code the kernel no longer
differentiates between cv's and sleep/wakeup. For example, calls to
abortsleep() and cv_abort() are replaced with a call to sleepq_abort().
Thus, the TDF_CVWAITQ flag is removed. Also, calls to unsleep() and
cv_waitq_remove() have been replaced with calls to sleepq_remove().
- The sched_sleep() function no longer accepts a priority argument as
sleep's no longer inherently bump the priority. Instead, this is soley
a propery of msleep() which explicitly calls sched_prio() before
blocking.
- The TDF_ONSLEEPQ flag has been dropped as it was never used. The
associated TDF_SET_ONSLEEPQ and TDF_CLR_ON_SLEEPQ macros have also been
dropped and replaced with a single explicit clearing of td_wchan.
TD_SET_ONSLEEPQ() would really have only made sense if it had taken
the wait channel and message as arguments anyway. Now that that only
happens in one place, a macro would be overkill.
2004-02-27 18:52:44 +00:00
|
|
|
td->td_flags &= ~TDF_SELECT;
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
Switch the sleep/wakeup and condition variable implementations to use the
sleep queue interface:
- Sleep queues attempt to merge some of the benefits of both sleep queues
and condition variables. Having sleep qeueus in a hash table avoids
having to allocate a queue head for each wait channel. Thus, struct cv
has shrunk down to just a single char * pointer now. However, the
hash table does not hold threads directly, but queue heads. This means
that once you have located a queue in the hash bucket, you no longer have
to walk the rest of the hash chain looking for threads. Instead, you have
a list of all the threads sleeping on that wait channel.
- Outside of the sleepq code and the sleep/cv code the kernel no longer
differentiates between cv's and sleep/wakeup. For example, calls to
abortsleep() and cv_abort() are replaced with a call to sleepq_abort().
Thus, the TDF_CVWAITQ flag is removed. Also, calls to unsleep() and
cv_waitq_remove() have been replaced with calls to sleepq_remove().
- The sched_sleep() function no longer accepts a priority argument as
sleep's no longer inherently bump the priority. Instead, this is soley
a propery of msleep() which explicitly calls sched_prio() before
blocking.
- The TDF_ONSLEEPQ flag has been dropped as it was never used. The
associated TDF_SET_ONSLEEPQ and TDF_CLR_ON_SLEEPQ macros have also been
dropped and replaced with a single explicit clearing of td_wchan.
TD_SET_ONSLEEPQ() would really have only made sense if it had taken
the wait channel and message as arguments anyway. Now that that only
happens in one place, a macro would be overkill.
2004-02-27 18:52:44 +00:00
|
|
|
sleepq_remove(td, &selwait);
|
2002-03-14 01:32:30 +00:00
|
|
|
mtx_unlock(&sellock);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2001-05-14 05:26:48 +00:00
|
|
|
|
2002-03-19 21:25:46 +00:00
|
|
|
static void selectinit(void *);
|
2001-05-14 05:26:48 +00:00
|
|
|
SYSINIT(select, SI_SUB_LOCK, SI_ORDER_FIRST, selectinit, NULL)
|
|
|
|
|
|
|
|
/* ARGSUSED*/
|
|
|
|
static void
|
|
|
|
selectinit(dummy)
|
|
|
|
void *dummy;
|
|
|
|
{
|
|
|
|
cv_init(&selwait, "select");
|
2002-04-04 21:03:38 +00:00
|
|
|
mtx_init(&sellock, "sellck", NULL, MTX_DEF);
|
2001-05-14 05:26:48 +00:00
|
|
|
}
|