freebsd-nq/sys/kern/vfs_vnops.c

1331 lines
30 KiB
C
Raw Normal View History

/*-
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.
*
* @(#)vfs_vnops.c 8.2 (Berkeley) 1/21/94
*/
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_mac.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/fcntl.h>
1994-05-24 10:09:53 +00:00
#include <sys/file.h>
#include <sys/kdb.h>
1994-05-24 10:09:53 +00:00
#include <sys/stat.h>
#include <sys/priv.h>
1994-05-24 10:09:53 +00:00
#include <sys/proc.h>
#include <sys/limits.h>
2001-10-11 17:52:20 +00:00
#include <sys/lock.h>
1994-05-24 10:09:53 +00:00
#include <sys/mount.h>
#include <sys/mutex.h>
1994-05-24 10:09:53 +00:00
#include <sys/namei.h>
#include <sys/vnode.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/filio.h>
#include <sys/sx.h>
#include <sys/ttycom.h>
1999-08-13 11:22:48 +00:00
#include <sys/conf.h>
#include <sys/syslog.h>
#include <sys/unistd.h>
1994-05-24 10:09:53 +00:00
#include <security/mac/mac_framework.h>
2002-12-24 09:44:51 +00:00
static fo_rdwr_t vn_read;
static fo_rdwr_t vn_write;
static fo_truncate_t vn_truncate;
2002-12-24 09:44:51 +00:00
static fo_ioctl_t vn_ioctl;
static fo_poll_t vn_poll;
static fo_kqfilter_t vn_kqfilter;
static fo_stat_t vn_statfile;
static fo_close_t vn_closefile;
struct fileops vnops = {
.fo_read = vn_read,
.fo_write = vn_write,
.fo_truncate = vn_truncate,
.fo_ioctl = vn_ioctl,
.fo_poll = vn_poll,
.fo_kqfilter = vn_kqfilter,
.fo_stat = vn_statfile,
.fo_close = vn_closefile,
.fo_flags = DFLAG_PASSABLE | DFLAG_SEEKABLE
};
int
vn_open(ndp, flagp, cmode, fp)
struct nameidata *ndp;
int *flagp, cmode;
struct file *fp;
{
struct thread *td = ndp->ni_cnd.cn_thread;
return (vn_open_cred(ndp, flagp, cmode, td->td_ucred, fp));
}
1994-05-24 10:09:53 +00:00
/*
* Common code for vnode open operations.
* Check permissions, and call the VOP_OPEN or VOP_CREATE routine.
*
* Note that this does NOT free nameidata for the successful case,
* due to the NDINIT being done elsewhere.
1994-05-24 10:09:53 +00:00
*/
int
vn_open_cred(ndp, flagp, cmode, cred, fp)
struct nameidata *ndp;
int *flagp, cmode;
struct ucred *cred;
struct file *fp;
1994-05-24 10:09:53 +00:00
{
struct vnode *vp;
struct mount *mp;
struct thread *td = ndp->ni_cnd.cn_thread;
1994-05-24 10:09:53 +00:00
struct vattr vat;
struct vattr *vap = &vat;
int fmode, error;
accmode_t accmode;
int vfslocked, mpsafe;
1994-05-24 10:09:53 +00:00
mpsafe = ndp->ni_cnd.cn_flags & MPSAFE;
restart:
vfslocked = 0;
fmode = *flagp;
1994-05-24 10:09:53 +00:00
if (fmode & O_CREAT) {
ndp->ni_cnd.cn_nameiop = CREATE;
ndp->ni_cnd.cn_flags = ISOPEN | LOCKPARENT | LOCKLEAF |
MPSAFE | AUDITVNODE1;
if ((fmode & O_EXCL) == 0 && (fmode & O_NOFOLLOW) == 0)
1994-05-24 10:09:53 +00:00
ndp->ni_cnd.cn_flags |= FOLLOW;
bwillwrite();
if ((error = namei(ndp)) != 0)
1994-05-24 10:09:53 +00:00
return (error);
vfslocked = NDHASGIANT(ndp);
if (!mpsafe)
ndp->ni_cnd.cn_flags &= ~MPSAFE;
1994-05-24 10:09:53 +00:00
if (ndp->ni_vp == NULL) {
VATTR_NULL(vap);
vap->va_type = VREG;
vap->va_mode = cmode;
if (fmode & O_EXCL)
vap->va_vaflags |= VA_EXCLUSIVE;
if (vn_start_write(ndp->ni_dvp, &mp, V_NOWAIT) != 0) {
NDFREE(ndp, NDF_ONLY_PNBUF);
vput(ndp->ni_dvp);
VFS_UNLOCK_GIANT(vfslocked);
if ((error = vn_start_write(NULL, &mp,
V_XSLEEP | PCATCH)) != 0)
return (error);
goto restart;
}
#ifdef MAC
error = mac_vnode_check_create(cred, ndp->ni_dvp,
&ndp->ni_cnd, vap);
if (error == 0)
#endif
error = VOP_CREATE(ndp->ni_dvp, &ndp->ni_vp,
&ndp->ni_cnd, vap);
vput(ndp->ni_dvp);
vn_finished_write(mp);
if (error) {
VFS_UNLOCK_GIANT(vfslocked);
NDFREE(ndp, NDF_ONLY_PNBUF);
1994-05-24 10:09:53 +00:00
return (error);
}
1994-05-24 10:09:53 +00:00
fmode &= ~O_TRUNC;
vp = ndp->ni_vp;
} else {
if (ndp->ni_dvp == ndp->ni_vp)
vrele(ndp->ni_dvp);
else
vput(ndp->ni_dvp);
ndp->ni_dvp = NULL;
vp = ndp->ni_vp;
if (fmode & O_EXCL) {
error = EEXIST;
goto bad;
}
fmode &= ~O_CREAT;
}
} else {
ndp->ni_cnd.cn_nameiop = LOOKUP;
ndp->ni_cnd.cn_flags = ISOPEN |
((fmode & O_NOFOLLOW) ? NOFOLLOW : FOLLOW) |
LOCKLEAF | MPSAFE | AUDITVNODE1;
if (!(fmode & FWRITE))
ndp->ni_cnd.cn_flags |= LOCKSHARED;
if ((error = namei(ndp)) != 0)
1994-05-24 10:09:53 +00:00
return (error);
if (!mpsafe)
ndp->ni_cnd.cn_flags &= ~MPSAFE;
vfslocked = NDHASGIANT(ndp);
1994-05-24 10:09:53 +00:00
vp = ndp->ni_vp;
}
if (vp->v_type == VLNK) {
error = EMLINK;
goto bad;
}
if (vp->v_type == VSOCK) {
1994-05-24 10:09:53 +00:00
error = EOPNOTSUPP;
goto bad;
}
accmode = 0;
if (fmode & (FWRITE | O_TRUNC)) {
if (vp->v_type == VDIR) {
error = EISDIR;
goto bad;
}
accmode |= VWRITE;
}
if (fmode & FREAD)
accmode |= VREAD;
if (fmode & FEXEC)
accmode |= VEXEC;
if (fmode & O_APPEND)
accmode |= VAPPEND;
#ifdef MAC
error = mac_vnode_check_open(cred, vp, accmode);
if (error)
goto bad;
#endif
1994-05-24 10:09:53 +00:00
if ((fmode & O_CREAT) == 0) {
if (accmode & VWRITE) {
error = vn_writechk(vp);
if (error)
goto bad;
}
if (accmode) {
error = VOP_ACCESS(vp, accmode, cred, td);
if (error)
1994-05-24 10:09:53 +00:00
goto bad;
}
}
if ((error = VOP_OPEN(vp, fmode, cred, td, fp)) != 0)
1994-05-24 10:09:53 +00:00
goto bad;
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct proc or any VM system structure will have to be rebuilt!!! Much needed overhaul of the VM system. Included in this first round of changes: 1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages, haspage, and sync operations are supported. The haspage interface now provides information about clusterability. All pager routines now take struct vm_object's instead of "pagers". 2) Improved data structures. In the previous paradigm, there is constant confusion caused by pagers being both a data structure ("allocate a pager") and a collection of routines. The idea of a pager structure has escentially been eliminated. Objects now have types, and this type is used to index the appropriate pager. In most cases, items in the pager structure were duplicated in the object data structure and thus were unnecessary. In the few cases that remained, a un_pager structure union was created in the object to contain these items. 3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now be removed. For instance, vm_object_enter(), vm_object_lookup(), vm_object_remove(), and the associated object hash list were some of the things that were removed. 4) simple_lock's removed. Discussion with several people reveals that the SMP locking primitives used in the VM system aren't likely the mechanism that we'll be adopting. Even if it were, the locking that was in the code was very inadequate and would have to be mostly re-done anyway. The locking in a uni-processor kernel was a no-op but went a long way toward making the code difficult to read and debug. 5) Places that attempted to kludge-up the fact that we don't have kernel thread support have been fixed to reflect the reality that we are really dealing with processes, not threads. The VM system didn't have complete thread support, so the comments and mis-named routines were just wrong. We now use tsleep and wakeup directly in the lock routines, for instance. 6) Where appropriate, the pagers have been improved, especially in the pager_alloc routines. Most of the pager_allocs have been rewritten and are now faster and easier to maintain. 7) The pagedaemon pageout clustering algorithm has been rewritten and now tries harder to output an even number of pages before and after the requested page. This is sort of the reverse of the ideal pagein algorithm and should provide better overall performance. 8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup have been removed. Some other unnecessary casts have also been removed. 9) Some almost useless debugging code removed. 10) Terminology of shadow objects vs. backing objects straightened out. The fact that the vm_object data structure escentially had this backwards really confused things. The use of "shadow" and "backing object" throughout the code is now internally consistent and correct in the Mach terminology. 11) Several minor bug fixes, including one in the vm daemon that caused 0 RSS objects to not get purged as intended. 12) A "default pager" has now been created which cleans up the transition of objects to the "swap" type. The previous checks throughout the code for swp->pg_data != NULL were really ugly. This change also provides the rudiments for future backing of "anonymous" memory by something other than the swap pager (via the vnode pager, for example), and it allows the decision about which of these pagers to use to be made dynamically (although will need some additional decision code to do this, of course). 13) (dyson) MAP_COPY has been deprecated and the corresponding "copy object" code has been removed. MAP_COPY was undocumented and non- standard. It was furthermore broken in several ways which caused its behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will continue to work correctly, but via the slightly different semantics of MAP_PRIVATE. 14) (dyson) Sharing maps have been removed. It's marginal usefulness in a threads design can be worked around in other ways. Both #12 and #13 were done to simplify the code and improve readability and maintain- ability. (As were most all of these changes) TODO: 1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing this will reduce the vnode pager to a mere fraction of its current size. 2) Rewrite vm_fault and the swap/vnode pagers to use the clustering information provided by the new haspage pager interface. This will substantially reduce the overhead by eliminating a large number of VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be improved to provide both a "behind" and "ahead" indication of contiguousness. 3) Implement the extended features of pager_haspage in swap_pager_haspage(). It currently just says 0 pages ahead/behind. 4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps via a much more general mechanism that could also be used for disk striping of regular filesystems. 5) Do something to improve the architecture of vm_object_collapse(). The fact that it makes calls into the swap pager and knows too much about how the swap pager operates really bothers me. It also doesn't allow for collapsing of non-swap pager objects ("unnamed" objects backed by other pagers).
1995-07-13 08:48:48 +00:00
if (fmode & FWRITE)
vp->v_writecount++;
*flagp = fmode;
ASSERT_VOP_LOCKED(vp, "vn_open_cred");
if (!mpsafe)
VFS_UNLOCK_GIANT(vfslocked);
1994-05-24 10:09:53 +00:00
return (0);
bad:
NDFREE(ndp, NDF_ONLY_PNBUF);
1994-05-24 10:09:53 +00:00
vput(vp);
VFS_UNLOCK_GIANT(vfslocked);
*flagp = fmode;
ndp->ni_vp = NULL;
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* Check for write permissions on the specified vnode.
* Prototype text segments cannot be written.
1994-05-24 10:09:53 +00:00
*/
int
1994-05-24 10:09:53 +00:00
vn_writechk(vp)
register struct vnode *vp;
{
ASSERT_VOP_LOCKED(vp, "vn_writechk");
1994-05-24 10:09:53 +00:00
/*
* If there's shared text associated with
* the vnode, try to free it up once. If
* we fail, we can't allow writing.
*/
if (vp->v_vflag & VV_TEXT)
1994-05-24 10:09:53 +00:00
return (ETXTBSY);
1994-05-24 10:09:53 +00:00
return (0);
}
/*
* Vnode close call
*/
int
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
vn_close(vp, flags, file_cred, td)
1994-05-24 10:09:53 +00:00
register struct vnode *vp;
int flags;
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
struct ucred *file_cred;
struct thread *td;
1994-05-24 10:09:53 +00:00
{
struct mount *mp;
int error, lock_flags;
if (!(flags & FWRITE) && vp->v_mount != NULL &&
vp->v_mount->mnt_kern_flag & MNTK_EXTENDED_SHARED)
lock_flags = LK_SHARED;
else
lock_flags = LK_EXCLUSIVE;
1994-05-24 10:09:53 +00:00
VFS_ASSERT_GIANT(vp->v_mount);
vn_start_write(vp, &mp, V_WAIT);
vn_lock(vp, lock_flags | LK_RETRY);
if (flags & FWRITE) {
VNASSERT(vp->v_writecount > 0, vp,
("vn_close: negative writecount"));
1994-05-24 10:09:53 +00:00
vp->v_writecount--;
}
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
error = VOP_CLOSE(vp, flags, file_cred, td);
vput(vp);
vn_finished_write(mp);
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* Heuristic to detect sequential operation.
*/
static int
sequential_heuristic(struct uio *uio, struct file *fp)
{
/*
* Offset 0 is handled specially. open() sets f_seqcount to 1 so
* that the first I/O is normally considered to be slightly
* sequential. Seeking to offset 0 doesn't change sequentiality
* unless previous seeks have reduced f_seqcount to 0, in which
* case offset 0 is not special.
*/
if ((uio->uio_offset == 0 && fp->f_seqcount > 0) ||
uio->uio_offset == fp->f_nextoff) {
/*
* f_seqcount is in units of fixed-size blocks so that it
* depends mainly on the amount of sequential I/O and not
* much on the number of sequential I/O's. The fixed size
* of 16384 is hard-coded here since it is (not quite) just
* a magic size that works well here. This size is more
* closely related to the best I/O size for real disks than
* to any block size used by software.
*/
fp->f_seqcount += howmany(uio->uio_resid, 16384);
if (fp->f_seqcount > IO_SEQMAX)
fp->f_seqcount = IO_SEQMAX;
return (fp->f_seqcount << IO_SEQSHIFT);
}
/* Not sequential. Quickly draw-down sequentiality. */
if (fp->f_seqcount > 1)
fp->f_seqcount = 1;
else
fp->f_seqcount = 0;
return (0);
}
1994-05-24 10:09:53 +00:00
/*
* Package up an I/O request on a vnode into a uio and do it.
*/
int
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
vn_rdwr(rw, vp, base, len, offset, segflg, ioflg, active_cred, file_cred,
aresid, td)
1994-05-24 10:09:53 +00:00
enum uio_rw rw;
struct vnode *vp;
2005-12-14 00:49:52 +00:00
void *base;
1994-05-24 10:09:53 +00:00
int len;
off_t offset;
enum uio_seg segflg;
int ioflg;
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
struct ucred *active_cred;
struct ucred *file_cred;
1994-05-24 10:09:53 +00:00
int *aresid;
struct thread *td;
1994-05-24 10:09:53 +00:00
{
struct uio auio;
struct iovec aiov;
struct mount *mp;
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
struct ucred *cred;
int error, lock_flags;
1994-05-24 10:09:53 +00:00
VFS_ASSERT_GIANT(vp->v_mount);
if ((ioflg & IO_NODELOCKED) == 0) {
mp = NULL;
if (rw == UIO_WRITE) {
if (vp->v_type != VCHR &&
(error = vn_start_write(vp, &mp, V_WAIT | PCATCH))
!= 0)
return (error);
if (mp != NULL &&
(mp->mnt_kern_flag & MNTK_SHARED_WRITES)) {
lock_flags = LK_SHARED;
} else {
lock_flags = LK_EXCLUSIVE;
}
vn_lock(vp, lock_flags | LK_RETRY);
} else
vn_lock(vp, LK_SHARED | LK_RETRY);
}
ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");
1994-05-24 10:09:53 +00:00
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
aiov.iov_base = base;
aiov.iov_len = len;
auio.uio_resid = len;
auio.uio_offset = offset;
auio.uio_segflg = segflg;
auio.uio_rw = rw;
auio.uio_td = td;
error = 0;
#ifdef MAC
if ((ioflg & IO_NOMACCHECK) == 0) {
if (rw == UIO_READ)
error = mac_vnode_check_read(active_cred, file_cred,
vp);
else
error = mac_vnode_check_write(active_cred, file_cred,
vp);
}
#endif
if (error == 0) {
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 (file_cred)
cred = file_cred;
else
cred = active_cred;
if (rw == UIO_READ)
error = VOP_READ(vp, &auio, ioflg, cred);
else
error = VOP_WRITE(vp, &auio, ioflg, cred);
}
1994-05-24 10:09:53 +00:00
if (aresid)
*aresid = auio.uio_resid;
else
if (auio.uio_resid && error == 0)
error = EIO;
if ((ioflg & IO_NODELOCKED) == 0) {
if (rw == UIO_WRITE && vp->v_type != VCHR)
vn_finished_write(mp);
VOP_UNLOCK(vp, 0);
}
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* Package up an I/O request on a vnode into a uio and do it. The I/O
* request is split up into smaller chunks and we try to avoid saturating
* the buffer cache while potentially holding a vnode locked, so we
* check bwillwrite() before calling vn_rdwr(). We also call uio_yield()
* to give other processes a chance to lock the vnode (either other processes
* core'ing the same binary, or unrelated processes scanning the directory).
*/
int
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
vn_rdwr_inchunks(rw, vp, base, len, offset, segflg, ioflg, active_cred,
file_cred, aresid, td)
enum uio_rw rw;
struct vnode *vp;
2005-12-14 00:49:52 +00:00
void *base;
size_t len;
off_t offset;
enum uio_seg segflg;
int ioflg;
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
struct ucred *active_cred;
struct ucred *file_cred;
size_t *aresid;
struct thread *td;
{
int error = 0;
int iaresid;
VFS_ASSERT_GIANT(vp->v_mount);
do {
int chunk;
/*
* Force `offset' to a multiple of MAXBSIZE except possibly
* for the first chunk, so that filesystems only need to
* write full blocks except possibly for the first and last
* chunks.
*/
chunk = MAXBSIZE - (uoff_t)offset % MAXBSIZE;
if (chunk > len)
chunk = len;
if (rw != UIO_READ && vp->v_type == VREG)
bwillwrite();
iaresid = 0;
error = vn_rdwr(rw, vp, base, chunk, offset, segflg,
ioflg, active_cred, file_cred, &iaresid, td);
len -= chunk; /* aresid calc already includes length */
if (error)
break;
offset += chunk;
2005-12-14 00:49:52 +00:00
base = (char *)base + chunk;
uio_yield();
} while (len);
if (aresid)
*aresid = len + iaresid;
return (error);
}
1994-05-24 10:09:53 +00:00
/*
* File table vnode read routine.
*/
static int
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
vn_read(fp, uio, active_cred, flags, td)
1994-05-24 10:09:53 +00:00
struct file *fp;
struct uio *uio;
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
struct ucred *active_cred;
struct thread *td;
int flags;
1994-05-24 10:09:53 +00:00
{
struct vnode *vp;
int error, ioflag;
struct mtx *mtxp;
int vfslocked;
1994-05-24 10:09:53 +00:00
KASSERT(uio->uio_td == td, ("uio_td %p is not td %p",
uio->uio_td, td));
mtxp = NULL;
vp = fp->f_vnode;
ioflag = 0;
if (fp->f_flag & FNONBLOCK)
ioflag |= IO_NDELAY;
if (fp->f_flag & O_DIRECT)
ioflag |= IO_DIRECT;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
/*
* According to McKusick the vn lock was protecting f_offset here.
* It is now protected by the FOFFSET_LOCKED flag.
*/
if ((flags & FOF_OFFSET) == 0) {
mtxp = mtx_pool_find(mtxpool_sleep, fp);
mtx_lock(mtxp);
while(fp->f_vnread_flags & FOFFSET_LOCKED) {
fp->f_vnread_flags |= FOFFSET_LOCK_WAITING;
msleep(&fp->f_vnread_flags, mtxp, PUSER -1,
"vnread offlock", 0);
}
fp->f_vnread_flags |= FOFFSET_LOCKED;
mtx_unlock(mtxp);
vn_lock(vp, LK_SHARED | LK_RETRY);
uio->uio_offset = fp->f_offset;
} else
vn_lock(vp, LK_SHARED | LK_RETRY);
ioflag |= sequential_heuristic(uio, fp);
#ifdef MAC
error = mac_vnode_check_read(active_cred, fp->f_cred, vp);
if (error == 0)
#endif
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
error = VOP_READ(vp, uio, ioflag, fp->f_cred);
if ((flags & FOF_OFFSET) == 0) {
fp->f_offset = uio->uio_offset;
mtx_lock(mtxp);
if (fp->f_vnread_flags & FOFFSET_LOCK_WAITING)
wakeup(&fp->f_vnread_flags);
fp->f_vnread_flags = 0;
mtx_unlock(mtxp);
}
fp->f_nextoff = uio->uio_offset;
VOP_UNLOCK(vp, 0);
VFS_UNLOCK_GIANT(vfslocked);
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* File table vnode write routine.
*/
static int
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
vn_write(fp, uio, active_cred, flags, td)
1994-05-24 10:09:53 +00:00
struct file *fp;
struct uio *uio;
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
struct ucred *active_cred;
struct thread *td;
int flags;
1994-05-24 10:09:53 +00:00
{
These changes appear to give us benefits with both small (32MB) and large (1G) memory machine configurations. I was able to run 'dbench 32' on a 32MB system without bring the machine to a grinding halt. * buffer cache hash table now dynamically allocated. This will have no effect on memory consumption for smaller systems and will help scale the buffer cache for larger systems. * minor enhancement to pmap_clearbit(). I noticed that all the calls to it used constant arguments. Making it an inline allows the constants to propogate to deeper inlines and should produce better code. * removal of inherent vfs_ioopt support through the emplacement of appropriate #ifdef's, with John's permission. If we do not find a use for it by the end of the year we will remove it entirely. * removal of getnewbufloops* counters & sysctl's - no longer necessary for debugging, getnewbuf() is now optimal. * buffer hash table functions removed from sys/buf.h and localized to vfs_bio.c * VFS_BIO_NEED_DIRTYFLUSH flag and support code added ( bwillwrite() ), allowing processes to block when too many dirty buffers are present in the system. * removal of a softdep test in bdwrite() that is no longer necessary now that bdwrite() no longer attempts to flush dirty buffers. * slight optimization added to bqrelse() - there is no reason to test for available buffer space on B_DELWRI buffers. * addition of reverse-scanning code to vfs_bio_awrite(). vfs_bio_awrite() will attempt to locate clusterable areas in both the forward and reverse direction relative to the offset of the buffer passed to it. This will probably not make much of a difference now, but I believe we will start to rely on it heavily in the future if we decide to shift some of the burden of the clustering closer to the actual I/O initiation. * Removal of the newbufcnt and lastnewbuf counters that Kirk added. They do not fix any race conditions that haven't already been fixed by the gbincore() test done after the only call to getnewbuf(). getnewbuf() is a static, so there is no chance of it being misused by other modules. ( Unless Kirk can think of a specific thing that this code fixes. I went through it very carefully and didn't see anything ). * removal of VOP_ISLOCKED() check in flushbufqueues(). I do not think this check is necessary, the buffer should flush properly whether the vnode is locked or not. ( yes? ). * removal of extra arguments passed to getnewbuf() that are not necessary. * missed cluster_wbuild() that had to be a cluster_wbuild_wb() in vfs_cluster.c * vn_write() now calls bwillwrite() *PRIOR* to locking the vnode, which should greatly aid flushing operations in heavy load situations - both the pageout and update daemons will be able to operate more efficiently. * removal of b_usecount. We may add it back in later but for now it is useless. Prior implementations of the buffer cache never had enough buffers for it to be useful, and current implementations which make more buffers available might not benefit relative to the amount of sophistication required to implement a b_usecount. Straight LRU should work just as well, especially when most things are VMIO backed. I expect that (even though John will not like this assumption) directories will become VMIO backed some point soon. Submitted by: Matthew Dillon <dillon@backplane.com> Reviewed by: Kirk McKusick <mckusick@mckusick.com>
1999-07-08 06:06:00 +00:00
struct vnode *vp;
struct mount *mp;
int error, ioflag, lock_flags;
int vfslocked;
1994-05-24 10:09:53 +00:00
KASSERT(uio->uio_td == td, ("uio_td %p is not td %p",
uio->uio_td, td));
vp = fp->f_vnode;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
These changes appear to give us benefits with both small (32MB) and large (1G) memory machine configurations. I was able to run 'dbench 32' on a 32MB system without bring the machine to a grinding halt. * buffer cache hash table now dynamically allocated. This will have no effect on memory consumption for smaller systems and will help scale the buffer cache for larger systems. * minor enhancement to pmap_clearbit(). I noticed that all the calls to it used constant arguments. Making it an inline allows the constants to propogate to deeper inlines and should produce better code. * removal of inherent vfs_ioopt support through the emplacement of appropriate #ifdef's, with John's permission. If we do not find a use for it by the end of the year we will remove it entirely. * removal of getnewbufloops* counters & sysctl's - no longer necessary for debugging, getnewbuf() is now optimal. * buffer hash table functions removed from sys/buf.h and localized to vfs_bio.c * VFS_BIO_NEED_DIRTYFLUSH flag and support code added ( bwillwrite() ), allowing processes to block when too many dirty buffers are present in the system. * removal of a softdep test in bdwrite() that is no longer necessary now that bdwrite() no longer attempts to flush dirty buffers. * slight optimization added to bqrelse() - there is no reason to test for available buffer space on B_DELWRI buffers. * addition of reverse-scanning code to vfs_bio_awrite(). vfs_bio_awrite() will attempt to locate clusterable areas in both the forward and reverse direction relative to the offset of the buffer passed to it. This will probably not make much of a difference now, but I believe we will start to rely on it heavily in the future if we decide to shift some of the burden of the clustering closer to the actual I/O initiation. * Removal of the newbufcnt and lastnewbuf counters that Kirk added. They do not fix any race conditions that haven't already been fixed by the gbincore() test done after the only call to getnewbuf(). getnewbuf() is a static, so there is no chance of it being misused by other modules. ( Unless Kirk can think of a specific thing that this code fixes. I went through it very carefully and didn't see anything ). * removal of VOP_ISLOCKED() check in flushbufqueues(). I do not think this check is necessary, the buffer should flush properly whether the vnode is locked or not. ( yes? ). * removal of extra arguments passed to getnewbuf() that are not necessary. * missed cluster_wbuild() that had to be a cluster_wbuild_wb() in vfs_cluster.c * vn_write() now calls bwillwrite() *PRIOR* to locking the vnode, which should greatly aid flushing operations in heavy load situations - both the pageout and update daemons will be able to operate more efficiently. * removal of b_usecount. We may add it back in later but for now it is useless. Prior implementations of the buffer cache never had enough buffers for it to be useful, and current implementations which make more buffers available might not benefit relative to the amount of sophistication required to implement a b_usecount. Straight LRU should work just as well, especially when most things are VMIO backed. I expect that (even though John will not like this assumption) directories will become VMIO backed some point soon. Submitted by: Matthew Dillon <dillon@backplane.com> Reviewed by: Kirk McKusick <mckusick@mckusick.com>
1999-07-08 06:06:00 +00:00
if (vp->v_type == VREG)
bwillwrite();
ioflag = IO_UNIT;
if (vp->v_type == VREG && (fp->f_flag & O_APPEND))
1994-05-24 10:09:53 +00:00
ioflag |= IO_APPEND;
if (fp->f_flag & FNONBLOCK)
ioflag |= IO_NDELAY;
if (fp->f_flag & O_DIRECT)
ioflag |= IO_DIRECT;
if ((fp->f_flag & O_FSYNC) ||
(vp->v_mount && (vp->v_mount->mnt_flag & MNT_SYNCHRONOUS)))
ioflag |= IO_SYNC;
mp = NULL;
if (vp->v_type != VCHR &&
(error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
goto unlock;
if (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) &&
(flags & FOF_OFFSET) != 0) {
lock_flags = LK_SHARED;
} else {
lock_flags = LK_EXCLUSIVE;
}
vn_lock(vp, lock_flags | LK_RETRY);
if ((flags & FOF_OFFSET) == 0)
uio->uio_offset = fp->f_offset;
ioflag |= sequential_heuristic(uio, fp);
#ifdef MAC
error = mac_vnode_check_write(active_cred, fp->f_cred, vp);
if (error == 0)
#endif
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
error = VOP_WRITE(vp, uio, ioflag, fp->f_cred);
if ((flags & FOF_OFFSET) == 0)
1994-05-24 10:09:53 +00:00
fp->f_offset = uio->uio_offset;
fp->f_nextoff = uio->uio_offset;
VOP_UNLOCK(vp, 0);
if (vp->v_type != VCHR)
vn_finished_write(mp);
unlock:
VFS_UNLOCK_GIANT(vfslocked);
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* File table truncate routine.
*/
static int
vn_truncate(fp, length, active_cred, td)
struct file *fp;
off_t length;
struct ucred *active_cred;
struct thread *td;
{
struct vattr vattr;
struct mount *mp;
struct vnode *vp;
int vfslocked;
int error;
vp = fp->f_vnode;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
error = vn_start_write(vp, &mp, V_WAIT | PCATCH);
if (error) {
VFS_UNLOCK_GIANT(vfslocked);
return (error);
}
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
if (vp->v_type == VDIR) {
error = EISDIR;
goto out;
}
#ifdef MAC
error = mac_vnode_check_write(active_cred, fp->f_cred, vp);
if (error)
goto out;
#endif
error = vn_writechk(vp);
if (error == 0) {
VATTR_NULL(&vattr);
vattr.va_size = length;
error = VOP_SETATTR(vp, &vattr, fp->f_cred);
}
out:
VOP_UNLOCK(vp, 0);
vn_finished_write(mp);
VFS_UNLOCK_GIANT(vfslocked);
return (error);
}
1994-05-24 10:09:53 +00:00
/*
* File table vnode stat routine.
*/
1999-11-08 03:32:15 +00:00
static int
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
vn_statfile(fp, sb, active_cred, td)
1999-11-08 03:32:15 +00:00
struct file *fp;
struct stat *sb;
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
struct ucred *active_cred;
struct thread *td;
1999-11-08 03:32:15 +00:00
{
struct vnode *vp = fp->f_vnode;
int vfslocked;
int error;
1999-11-08 03:32:15 +00:00
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
vn_lock(vp, LK_SHARED | LK_RETRY);
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
error = vn_stat(vp, sb, active_cred, fp->f_cred, td);
VOP_UNLOCK(vp, 0);
VFS_UNLOCK_GIANT(vfslocked);
return (error);
1999-11-08 03:32:15 +00:00
}
/*
* Stat a vnode; implementation for the stat syscall
*/
int
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
vn_stat(vp, sb, active_cred, file_cred, td)
1994-05-24 10:09:53 +00:00
struct vnode *vp;
register struct stat *sb;
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
struct ucred *active_cred;
struct ucred *file_cred;
struct thread *td;
1994-05-24 10:09:53 +00:00
{
struct vattr vattr;
register struct vattr *vap;
int error;
u_short mode;
#ifdef MAC
error = mac_vnode_check_stat(active_cred, file_cred, vp);
if (error)
return (error);
#endif
1994-05-24 10:09:53 +00:00
vap = &vattr;
/*
* Initialize defaults for new and unusual fields, so that file
* systems which don't support these fields don't need to know
* about them.
*/
vap->va_birthtime.tv_sec = -1;
vap->va_birthtime.tv_nsec = 0;
vap->va_fsid = VNOVAL;
vap->va_rdev = NODEV;
error = VOP_GETATTR(vp, vap, active_cred);
1994-05-24 10:09:53 +00:00
if (error)
return (error);
/*
* Zero the spare stat fields
*/
bzero(sb, sizeof *sb);
1994-05-24 10:09:53 +00:00
/*
* Copy from vattr table
*/
if (vap->va_fsid != VNOVAL)
sb->st_dev = vap->va_fsid;
else
sb->st_dev = vp->v_mount->mnt_stat.f_fsid.val[0];
1994-05-24 10:09:53 +00:00
sb->st_ino = vap->va_fileid;
mode = vap->va_mode;
switch (vap->va_type) {
1994-05-24 10:09:53 +00:00
case VREG:
mode |= S_IFREG;
break;
case VDIR:
mode |= S_IFDIR;
break;
case VBLK:
mode |= S_IFBLK;
break;
case VCHR:
mode |= S_IFCHR;
break;
case VLNK:
mode |= S_IFLNK;
break;
case VSOCK:
mode |= S_IFSOCK;
break;
case VFIFO:
mode |= S_IFIFO;
break;
default:
return (EBADF);
};
sb->st_mode = mode;
sb->st_nlink = vap->va_nlink;
sb->st_uid = vap->va_uid;
sb->st_gid = vap->va_gid;
sb->st_rdev = vap->va_rdev;
if (vap->va_size > OFF_MAX)
return (EOVERFLOW);
1994-05-24 10:09:53 +00:00
sb->st_size = vap->va_size;
sb->st_atimespec = vap->va_atime;
sb->st_mtimespec = vap->va_mtime;
1994-05-24 10:09:53 +00:00
sb->st_ctimespec = vap->va_ctime;
sb->st_birthtimespec = vap->va_birthtime;
/*
* According to www.opengroup.org, the meaning of st_blksize is
* "a filesystem-specific preferred I/O block size for this
* object. In some filesystem types, this may vary from file
* to file"
* Default to PAGE_SIZE after much discussion.
* XXX: min(PAGE_SIZE, vp->v_bufobj.bo_bsize) may be more correct.
*/
sb->st_blksize = PAGE_SIZE;
1994-05-24 10:09:53 +00:00
sb->st_flags = vap->va_flags;
if (priv_check(td, PRIV_VFS_GENERATION))
sb->st_gen = 0;
else
sb->st_gen = vap->va_gen;
1994-05-24 10:09:53 +00:00
sb->st_blocks = vap->va_bytes / S_BLKSIZE;
return (0);
}
/*
* File table vnode ioctl routine.
*/
static int
vn_ioctl(fp, com, data, active_cred, td)
1994-05-24 10:09:53 +00:00
struct file *fp;
u_long com;
void *data;
struct ucred *active_cred;
struct thread *td;
1994-05-24 10:09:53 +00:00
{
struct vnode *vp = fp->f_vnode;
1994-05-24 10:09:53 +00:00
struct vattr vattr;
int vfslocked;
1994-05-24 10:09:53 +00:00
int error;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
error = ENOTTY;
1994-05-24 10:09:53 +00:00
switch (vp->v_type) {
case VREG:
case VDIR:
if (com == FIONREAD) {
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
error = VOP_GETATTR(vp, &vattr, active_cred);
VOP_UNLOCK(vp, 0);
if (!error)
*(int *)data = vattr.va_size - fp->f_offset;
1994-05-24 10:09:53 +00:00
}
if (com == FIONBIO || com == FIOASYNC) /* XXX */
error = 0;
else
error = VOP_IOCTL(vp, com, data, fp->f_flag,
active_cred, td);
break;
1994-05-24 10:09:53 +00:00
default:
break;
1994-05-24 10:09:53 +00:00
}
VFS_UNLOCK_GIANT(vfslocked);
return (error);
1994-05-24 10:09:53 +00:00
}
/*
1997-09-14 02:51:16 +00:00
* File table vnode poll routine.
1994-05-24 10:09:53 +00:00
*/
static int
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
vn_poll(fp, events, active_cred, td)
1994-05-24 10:09:53 +00:00
struct file *fp;
1997-09-14 02:51:16 +00:00
int 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
struct ucred *active_cred;
struct thread *td;
1994-05-24 10:09:53 +00:00
{
struct vnode *vp;
int vfslocked;
int error;
vp = fp->f_vnode;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
#ifdef MAC
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
error = mac_vnode_check_poll(active_cred, fp->f_cred, vp);
VOP_UNLOCK(vp, 0);
if (!error)
#endif
1994-05-24 10:09:53 +00:00
error = VOP_POLL(vp, events, fp->f_cred, td);
VFS_UNLOCK_GIANT(vfslocked);
return (error);
1994-05-24 10:09:53 +00:00
}
/*
* Acquire the requested lock and then check for validity. LK_RETRY
* permits vn_lock to return doomed vnodes.
*/
int
_vn_lock(struct vnode *vp, int flags, char *file, int line)
{
int error;
VNASSERT((flags & LK_TYPE_MASK) != 0, vp,
("vn_lock called with no locktype."));
do {
#ifdef DEBUG_VFS_LOCKS
KASSERT(vp->v_holdcnt != 0,
("vn_lock %p: zero hold count", vp));
#endif
error = VOP_LOCK1(vp, flags, file, line);
flags &= ~LK_INTERLOCK; /* Interlock is always dropped. */
KASSERT((flags & LK_RETRY) == 0 || error == 0,
("LK_RETRY set with incompatible flags (0x%x) or an error occured (%d)",
flags, error));
/*
* Callers specify LK_RETRY if they wish to get dead vnodes.
* If RETRY is not set, we return ENOENT instead.
*/
if (error == 0 && vp->v_iflag & VI_DOOMED &&
(flags & LK_RETRY) == 0) {
VOP_UNLOCK(vp, 0);
error = ENOENT;
break;
}
} while (flags & LK_RETRY && error != 0);
return (error);
}
/*
* File table vnode close routine.
*/
static int
vn_closefile(fp, td)
struct file *fp;
struct thread *td;
{
struct vnode *vp;
struct flock lf;
int vfslocked;
int error;
vp = fp->f_vnode;
vfslocked = VFS_LOCK_GIANT(vp->v_mount);
if (fp->f_type == DTYPE_VNODE && fp->f_flag & FHASLOCK) {
lf.l_whence = SEEK_SET;
lf.l_start = 0;
lf.l_len = 0;
lf.l_type = F_UNLCK;
2005-12-14 00:49:52 +00:00
(void) VOP_ADVLOCK(vp, fp, F_UNLCK, &lf, F_FLOCK);
}
fp->f_ops = &badfileops;
error = vn_close(vp, fp->f_flag, fp->f_cred, td);
VFS_UNLOCK_GIANT(vfslocked);
return (error);
}
/*
* Preparing to start a filesystem write operation. If the operation is
* permitted, then we bump the count of operations in progress and
* proceed. If a suspend request is in progress, we wait until the
* suspension is over, and then proceed.
*/
int
vn_start_write(vp, mpp, flags)
struct vnode *vp;
struct mount **mpp;
int flags;
{
struct mount *mp;
int error;
error = 0;
/*
* If a vnode is provided, get and return the mount point that
* to which it will write.
*/
if (vp != NULL) {
if ((error = VOP_GETWRITEMOUNT(vp, mpp)) != 0) {
*mpp = NULL;
if (error != EOPNOTSUPP)
return (error);
return (0);
}
}
if ((mp = *mpp) == NULL)
return (0);
2008-11-02 10:15:42 +00:00
/*
* VOP_GETWRITEMOUNT() returns with the mp refcount held through
* a vfs_ref().
* As long as a vnode is not provided we need to acquire a
* refcount for the provided mountpoint too, in order to
* emulate a vfs_ref().
*/
MNT_ILOCK(mp);
if (vp == NULL)
MNT_REF(mp);
2008-11-02 10:15:42 +00:00
/*
* Check on status of suspension.
*/
if ((curthread->td_pflags & TDP_IGNSUSP) == 0 ||
mp->mnt_susp_owner != curthread) {
while ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0) {
if (flags & V_NOWAIT) {
error = EWOULDBLOCK;
goto unlock;
}
error = msleep(&mp->mnt_flag, MNT_MTX(mp),
(PUSER - 1) | (flags & PCATCH), "suspfs", 0);
if (error)
goto unlock;
}
}
if (flags & V_XSLEEP)
goto unlock;
mp->mnt_writeopcount++;
unlock:
MNT_REL(mp);
MNT_IUNLOCK(mp);
return (error);
}
/*
* Secondary suspension. Used by operations such as vop_inactive
* routines that are needed by the higher level functions. These
* are allowed to proceed until all the higher level functions have
* completed (indicated by mnt_writeopcount dropping to zero). At that
* time, these operations are halted until the suspension is over.
*/
int
vn_start_secondary_write(vp, mpp, flags)
struct vnode *vp;
struct mount **mpp;
int flags;
{
struct mount *mp;
int error;
retry:
if (vp != NULL) {
if ((error = VOP_GETWRITEMOUNT(vp, mpp)) != 0) {
*mpp = NULL;
if (error != EOPNOTSUPP)
return (error);
return (0);
}
}
/*
* If we are not suspended or have not yet reached suspended
* mode, then let the operation proceed.
*/
if ((mp = *mpp) == NULL)
return (0);
2008-11-02 10:15:42 +00:00
/*
* VOP_GETWRITEMOUNT() returns with the mp refcount held through
* a vfs_ref().
* As long as a vnode is not provided we need to acquire a
* refcount for the provided mountpoint too, in order to
* emulate a vfs_ref().
*/
MNT_ILOCK(mp);
if (vp == NULL)
MNT_REF(mp);
if ((mp->mnt_kern_flag & (MNTK_SUSPENDED | MNTK_SUSPEND2)) == 0) {
mp->mnt_secondary_writes++;
mp->mnt_secondary_accwrites++;
MNT_REL(mp);
MNT_IUNLOCK(mp);
return (0);
}
if (flags & V_NOWAIT) {
MNT_REL(mp);
MNT_IUNLOCK(mp);
return (EWOULDBLOCK);
}
/*
* Wait for the suspension to finish.
*/
error = msleep(&mp->mnt_flag, MNT_MTX(mp),
(PUSER - 1) | (flags & PCATCH) | PDROP, "suspfs", 0);
vfs_rel(mp);
if (error == 0)
goto retry;
return (error);
}
/*
* Filesystem write operation has completed. If we are suspending and this
* operation is the last one, notify the suspender that the suspension is
* now in effect.
*/
void
vn_finished_write(mp)
struct mount *mp;
{
if (mp == NULL)
return;
MNT_ILOCK(mp);
mp->mnt_writeopcount--;
if (mp->mnt_writeopcount < 0)
panic("vn_finished_write: neg cnt");
if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0 &&
mp->mnt_writeopcount <= 0)
wakeup(&mp->mnt_writeopcount);
MNT_IUNLOCK(mp);
}
/*
* Filesystem secondary write operation has completed. If we are
* suspending and this operation is the last one, notify the suspender
* that the suspension is now in effect.
*/
void
vn_finished_secondary_write(mp)
struct mount *mp;
{
if (mp == NULL)
return;
MNT_ILOCK(mp);
mp->mnt_secondary_writes--;
if (mp->mnt_secondary_writes < 0)
panic("vn_finished_secondary_write: neg cnt");
if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0 &&
mp->mnt_secondary_writes <= 0)
wakeup(&mp->mnt_secondary_writes);
MNT_IUNLOCK(mp);
}
/*
* Request a filesystem to suspend write operations.
*/
int
vfs_write_suspend(mp)
struct mount *mp;
{
int error;
MNT_ILOCK(mp);
if (mp->mnt_susp_owner == curthread) {
MNT_IUNLOCK(mp);
return (EALREADY);
}
while (mp->mnt_kern_flag & MNTK_SUSPEND)
msleep(&mp->mnt_flag, MNT_MTX(mp), PUSER - 1, "wsuspfs", 0);
mp->mnt_kern_flag |= MNTK_SUSPEND;
mp->mnt_susp_owner = curthread;
if (mp->mnt_writeopcount > 0)
(void) msleep(&mp->mnt_writeopcount,
MNT_MTX(mp), (PUSER - 1)|PDROP, "suspwt", 0);
else
MNT_IUNLOCK(mp);
if ((error = VFS_SYNC(mp, MNT_SUSPEND)) != 0)
vfs_write_resume(mp);
return (error);
}
/*
* Request a filesystem to resume write operations.
*/
void
vfs_write_resume(mp)
struct mount *mp;
{
MNT_ILOCK(mp);
if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0) {
KASSERT(mp->mnt_susp_owner == curthread, ("mnt_susp_owner"));
mp->mnt_kern_flag &= ~(MNTK_SUSPEND | MNTK_SUSPEND2 |
MNTK_SUSPENDED);
mp->mnt_susp_owner = NULL;
wakeup(&mp->mnt_writeopcount);
wakeup(&mp->mnt_flag);
curthread->td_pflags &= ~TDP_IGNSUSP;
MNT_IUNLOCK(mp);
VFS_SUSP_CLEAN(mp);
} else
MNT_IUNLOCK(mp);
}
/*
* Implement kqueues for files by translating it to vnode operation.
*/
static int
vn_kqfilter(struct file *fp, struct knote *kn)
{
int vfslocked;
int error;
vfslocked = VFS_LOCK_GIANT(fp->f_vnode->v_mount);
error = VOP_KQFILTER(fp->f_vnode, kn);
VFS_UNLOCK_GIANT(vfslocked);
return error;
}
/*
* Simplified in-kernel wrapper calls for extended attribute access.
* Both calls pass in a NULL credential, authorizing as "kernel" access.
* Set IO_NODELOCKED in ioflg if the vnode is already locked.
*/
int
vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace,
const char *attrname, int *buflen, char *buf, struct thread *td)
{
struct uio auio;
struct iovec iov;
int error;
iov.iov_len = *buflen;
iov.iov_base = buf;
auio.uio_iov = &iov;
auio.uio_iovcnt = 1;
auio.uio_rw = UIO_READ;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_td = td;
auio.uio_offset = 0;
auio.uio_resid = *buflen;
if ((ioflg & IO_NODELOCKED) == 0)
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");
/* authorize attribute retrieval as kernel */
error = VOP_GETEXTATTR(vp, attrnamespace, attrname, &auio, NULL, NULL,
td);
if ((ioflg & IO_NODELOCKED) == 0)
VOP_UNLOCK(vp, 0);
if (error == 0) {
*buflen = *buflen - auio.uio_resid;
}
return (error);
}
/*
* XXX failure mode if partially written?
*/
int
vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace,
const char *attrname, int buflen, char *buf, struct thread *td)
{
struct uio auio;
struct iovec iov;
struct mount *mp;
int error;
iov.iov_len = buflen;
iov.iov_base = buf;
auio.uio_iov = &iov;
auio.uio_iovcnt = 1;
auio.uio_rw = UIO_WRITE;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_td = td;
auio.uio_offset = 0;
auio.uio_resid = buflen;
if ((ioflg & IO_NODELOCKED) == 0) {
if ((error = vn_start_write(vp, &mp, V_WAIT)) != 0)
return (error);
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
}
ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");
/* authorize attribute setting as kernel */
error = VOP_SETEXTATTR(vp, attrnamespace, attrname, &auio, NULL, td);
if ((ioflg & IO_NODELOCKED) == 0) {
vn_finished_write(mp);
VOP_UNLOCK(vp, 0);
}
return (error);
}
int
vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
const char *attrname, struct thread *td)
{
struct mount *mp;
int error;
if ((ioflg & IO_NODELOCKED) == 0) {
if ((error = vn_start_write(vp, &mp, V_WAIT)) != 0)
return (error);
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
}
ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");
/* authorize attribute removal as kernel */
error = VOP_DELETEEXTATTR(vp, attrnamespace, attrname, NULL, td);
if (error == EOPNOTSUPP)
error = VOP_SETEXTATTR(vp, attrnamespace, attrname, NULL,
NULL, td);
if ((ioflg & IO_NODELOCKED) == 0) {
vn_finished_write(mp);
VOP_UNLOCK(vp, 0);
}
return (error);
}
int
vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags, struct vnode **rvp)
{
struct mount *mp;
int ltype, error;
mp = vp->v_mount;
ltype = VOP_ISLOCKED(vp);
KASSERT(ltype == LK_EXCLUSIVE || ltype == LK_SHARED,
("vn_vget_ino: vp not locked"));
error = vfs_busy(mp, MBF_NOWAIT);
if (error != 0) {
VOP_UNLOCK(vp, 0);
error = vfs_busy(mp, 0);
vn_lock(vp, ltype | LK_RETRY);
if (error != 0)
return (ENOENT);
if (vp->v_iflag & VI_DOOMED) {
vfs_unbusy(mp);
return (ENOENT);
}
}
VOP_UNLOCK(vp, 0);
error = VFS_VGET(mp, ino, lkflags, rvp);
vfs_unbusy(mp);
vn_lock(vp, ltype | LK_RETRY);
if (vp->v_iflag & VI_DOOMED) {
if (error == 0)
vput(*rvp);
error = ENOENT;
}
return (error);
}