of a process pointer.
- Move the p_candebug() at the start of procfs_control() a bit to make
locking feasible. We still perform the access check before doing
anything, we just now perform it after acquiring locks.
- Don't lock the sched_lock for TRACE_WAIT_P() and when checking to see if
p_stat is SSTOP. We lock the process while setting p_stat to SSTOP
so locking the process is sufficient to do a read to see if p_stat is
SSTOP or not.
Setting of timestamps on devices had no effect visible to userland
because timestamps for devices were set in places that are never used.
This broke:
- update of file change time after a change of an attribute
- setting of file access and modification times.
The VA_UTIMES_NULL case did not work. Revs 1.31-1.32 were supposed to
fix this by copying correct bits from ufs, but had little or no effect
because the old checks were not removed.
files. We didn't clear the update marks when we set the times, so
some of the settings were sometimes clobbered with the current time a
little later. This caused cp -p even by root to almost always fail
to preserve any times despite not reporting any errors in attempting
to preserve them.
Don't forget to set the archive attribute when we set the read-only
attribute. We should only set the archive attribute if we actually
change something, but we mostly don't bother avoiding setting it
elsewhere, so don't bother here yet.
MFC after: 1 week
most cases NULL is passed, but in some cases such as network driver locks
(which use the MTX_NETWORK_LOCK macro) and UMA zone locks, a name is used.
Tested on: i386, alpha, sparc64
they aren't in the usual path of execution for syscalls and traps.
The main complication for this is that we have to set flags to control
ast() everywhere that changes the signal mask.
Avoid locking in userret() in most of the remaining cases.
Submitted by: luoqi (first part only, long ago, reorganized by me)
Reminded by: dillon
general cleanup of the API. The entire API now consists of two functions
similar to the pre-KSE API. The suser() function takes a thread pointer
as its only argument. The td_ucred member of this thread must be valid
so the only valid thread pointers are curthread and a few kernel threads
such as thread0. The suser_cred() function takes a pointer to a struct
ucred as its first argument and an integer flag as its second argument.
The flag is currently only used for the PRISON_ROOT flag.
Discussed on: smp@
provided the latter is nonzero. At this point, the former is a fairly
arbitrary default value (DFTPHYS), so changing it to any reasonable
value specified by the device driver is safe. Using the maximum of
these limits broke ffs clustered i/o for devices whose si_iosize_max
is < DFLTPHYS. Using the minimum would break device drivers' ability
to increase the active limit from DFTLPHYS up to MAXPHYS.
Copied the code for this and the associated (unnecessary?) fixup of
mp_iosize_max to all other filesystems that use clustering (ext2fs and
msdosfs). It was completely missing.
PR: 36309
MFC-after: 1 week
as it leaves the nullfs vnode allocated, but with no identity. The
effect is that a null mount can slowly accumulate all the vnodes
in the system, reclaiming them only when it is unmounted. Thus
the null_inactive state instead accelerates the release of the
null vnode by calling vrecycle which will in turn call the
null_reclaim operator. The null_reclaim routine then does the
freeing actions previosuly (incorrectly) done in null_inactive.
locking flags when acquiring a vnode. The immediate purpose is
to allow polling lock requests (LK_NOWAIT) needed by soft updates
to avoid deadlock when enlisting other processes to help with
the background cleanup. For the future it will allow the use of
shared locks for read access to vnodes. This change touches a
lot of files as it affects most filesystems within the system.
It has been well tested on FFS, loopback, and CD-ROM filesystems.
only lightly on the others, so if you find a problem there, please
let me (mckusick@mckusick.com) know.
the bio and buffer structures to have daddr64_t bio_pblkno,
b_blkno, and b_lblkno fields which allows access to disks
larger than a Terabyte in size. This change also requires
that the VOP_BMAP vnode operation accept and return daddr64_t
blocks. This delta should not affect system operation in
any way. It merely sets up the necessary interfaces to allow
the development of disk drivers that work with these larger
disk block addresses. It also allows for the development of
UFS2 which will use 64-bit block addresses.
New locks are:
- pgrpsess_lock which locks the whole pgrps and sessions,
- pg_mtx which protects the pgrp members, and
- s_mtx which protects the session members.
Please refer to sys/proc.h for the coverage of these locks.
Changes on the pgrp/session interface:
- pgfind() needs the pgrpsess_lock held.
- The caller of enterpgrp() is responsible to allocate a new pgrp and
session.
- Call enterthispgrp() in order to enter an existing pgrp.
- pgsignal() requires a pgrp lock held.
Reviewed by: jhb, alfred
Tested on: cvsup.jp.FreeBSD.org
(which is a quad-CPU machine running -current)
- clobbering of jsp's $Id$ by FreeBSD's old $Id$.
- long lines in recent KSE changes (procfs_ctl.c).
- other style bugs in KSE changes (most related to an shadowed variable
in procfs_status.c -- the td in the outer scope is obfuscated by
PFS_FILL_ARGS).
Approved by: des
- clobbering of jsp's $Id$ by FreeBSD's old $Id$.
- lost Berkeley id in procfs_dbregs.c
- long lines in recent KSE changes.
- various gratuitous differences between procfs_*regs.c.
o Modify the system call syntax for extattr_{get,set}_{fd,file}() so
as not to use the scatter gather API (which appeared not to be used
by any consumers, and be less portable), rather, accepts 'data'
and 'nbytes' in the style of other simple read/write interfaces.
This changes the API and ABI.
o Modify system call semantics so that extattr_get_{fd,file}() return
a size_t. When performing a read, the number of bytes read will
be returned, unless the data pointer is NULL, in which case the
number of bytes of data are returned. This changes the API only.
o Modify the VOP_GETEXTATTR() vnode operation to accept a *size_t
argument so as to return the size, if desirable. If set to NULL,
the size will not be returned.
o Update various filesystems (pseodofs, ufs) to DTRT.
These changes should make extended attributes more useful and more
portable. More commits to rebuild the system call files, as well
as update userland utilities to follow.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs
this is a low-functionality change that changes the kernel to access the main
thread of a process via the linked list of threads rather than
assuming that it is embedded in the process. It IS still embeded there
but remove all teh code that assumes that in preparation for the next commit
which will actually move it out.
Reviewed by: peter@freebsd.org, gallatin@cs.duke.edu, benno rice,
Backout revision 1.56 and 1.57 of fifo_vnops.c.
Introduce a new poll op "POLLINIGNEOF" that can be used to ignore
EOF on a fifo, POLLIN/POLLRDNORM is converted to POLLINIGNEOF within
the FIFO implementation to effect the correct behavior.
This should allow one to view a fifo pretty much as a data source
rather than worry about connections coming and going.
Reviewed by: bde
functions just grab f_data and don't muck with anything else so this
should be ok.
this fixes a panic with invariants where it thinks we've doubly initialized
the filetmp mutex even though all we've done is neglect to bzero it.
Seigo Tanimura (tanimura) posted the initial delta.
I've polished it quite a bit reducing the need for locking and
adapting it for KSE.
Locks:
1 mutex in each filedesc
protects all the fields.
protects "struct file" initialization, while a struct file
is being changed from &badfileops -> &pipeops or something
the filedesc should be locked.
1 mutex in each struct file
protects the refcount fields.
doesn't protect anything else.
the flags used for garbage collection have been moved to
f_gcflag which was the FILLER short, this doesn't need
locking because the garbage collection is a single threaded
container.
could likely be made to use a pool mutex.
1 sx lock for the global filelist.
struct file * fhold(struct file *fp);
/* increments reference count on a file */
struct file * fhold_locked(struct file *fp);
/* like fhold but expects file to locked */
struct file * ffind_hold(struct thread *, int fd);
/* finds the struct file in thread, adds one reference and
returns it unlocked */
struct file * ffind_lock(struct thread *, int fd);
/* ffind_hold, but returns file locked */
I still have to smp-safe the fget cruft, I'll get to that asap.
socreate(), rather than getting it implicitly from the thread
argument.
o Make NFS cache the credential provided at mount-time, and use
the cached credential (nfsmount->nm_cred) when making calls to
socreate() on initially connecting, or reconnecting the socket.
This fixes bugs involving NFS over TCP and ipfw uid/gid rules, as well
as bugs involving NFS and mandatory access control implementations.
Reviewed by: freebsd-arch
against VM_WAIT in the pageout code. Both fixes involve adjusting
the lockmgr's timeout capability so locks obtained with timeouts do not
interfere with locks obtained without a timeout.
Hopefully MFC: before the 4.5 release
the wrong VOP descriptor. This misuse caused VFS-cached vnodes to be
re-cached, resulting in the leak. This commit is an interim fix until DES
has a chance to rework the code involved.
With this change, mounting an smb share (using mount_smb, which is not
yet included in the tree) without any of smbfs, libiconv or libmchain
compiled into the kernel or loaded works.
not the calling process. While we're here, also unstaticize procfs_doprocfile() and
procfs_docurproc() so linprocfs can call them directly instead of duplicating them.
Submitted by: Dominic Mitchell <dom@semantico.com>
mutable contents of struct prison (hostname, securelevel, refcount,
pr_linux, ...)
o Generally introduce mtx_lock()/mtx_unlock() calls throughout kern/
so as to enforce these protections, in particular, in kern_mib.c
protection sysctl access to the hostname and securelevel, as well as
kern_prot.c access to the securelevel for access control purposes.
o Rewrite linux emulator abstractions for accessing per-jail linux
mib entries (osname, osrelease, osversion) so that they don't return
a pointer to the text in the struct linux_prison, rather, a copy
to an array passed into the calls. Likewise, update linprocfs to
use these primitives.
o Update in_pcb.c to always use prison_getip() rather than directly
accessing struct prison.
Reviewed by: jhb
The problem was that the ISO9660 code wasn't opening the device prior to
issuing ioctl calls. In particular, the device must be open before
iso_get_ssector() is called in iso_mountroot().
If the device isn't opened first, the disk layer blows up due to an
uninitialized variable.
The solution was to open the device, call iso_get_ssector() and then close
it again.
The ATAPI CDROM driver doesn't have this problem because it doesn't use the
disk layer, and evidently doesn't mind if someone issues an ioctl without
first issuing an open call.
Thanks to phk for pointing me at the source of this problem.
Tested by: dirk
MFC after: 1 week
pathconf() variables for directories, and set st_size and st_blocks
(of struct stat) for directories as appropriate. Note that st_size is
always set to DEV_BSIZE, since the size of the directories is not
currently kept.
Reviewed by: phk, bde
Basically FIFOs become a real pain to abuse as a rendevous point without
this change because you can't really select(2) on them because they always
return ready even though there is no writer (to signal EOF).
Obtained from: BSD/os
quad_t cannot be printed with %lld on 64 bit systems.
Dont waste cpu to round user and system times up to long long, it is
highly improbable that a process will have accumulated 68 years of
user or system cpu time (not wall clock time) before a reboot or
process restart.
structure changes now rather then piecemeal later on. mnt_nvnodelist
currently holds all the vnodes under the mount point. This will eventually
be split into a 'dirty' and 'clean' list. This way we only break kld's once
rather then twice. nvnodelist will eventually turn into the dirty list
and should remain compatible with the klds.
that a buffer's b_blkno would be valid. This is true when vmiodirenable
is turned off because the B_MALLOC'd buffer's data is invalidated when
the buffer is destroyed. But when vmiodirenable is turned on a buffer
can be reconstituted from its VMIO backing store. The reconstituted buffer
will have no knowledge of the physical block translation and the result is
serious directory corruption of the CDROM.
The solution is to fix cd9660_blkatoff() to always BMAP the buffer if
b_lblkno == b_blkno.
MFC after: 0 days
- crhold() returns a reference to the ucred whose refcount it bumps.
- crcopy() now simply copies the credentials from one credential to
another and has no return value.
- a new crshared() primitive is added which returns true if a ucred's
refcount is > 1 and false (0) otherwise.
Until now, the ptrace syscall was implemented as a wrapper that called
various functions in procfs depending on which ptrace operation was
requested. Most of these functions were themselves wrappers around
procfs_{read,write}_{,db,fp}regs(), with only some extra error checks,
which weren't necessary in the ptrace case anyway.
This commit moves procfs_rwmem() from procfs_mem.c into sys_process.c
(renaming it to proc_rwmem() in the process), and implements ptrace()
directly in terms of procfs_{read,write}_{,db,fp}regs() instead of
having it fake up a struct uio and then call procfs_do{,db,fp}regs().
It also moves the prototypes for procfs_{read,write}_{,db,fp}regs()
and proc_rwmem() from proc.h to ptrace.h, and marks all procfs files
except procfs_machdep.c as "optional procfs" instead of "standard".
the target process was being held locked during the uiomove() call. If the
process calling readdir() was the same as the target process (for instance
'ls /proc/curproc/'), and uiomove() caused a page fault, the result would
be a proc lock recursion. I have no idea how long this has been broken -
possibly ever since pfind() was changed to lock the process it returns.
Also replace the one and only call to procfs_findtextvp() with a direct
test of td->td_proc->p_textvp.
YA pseudofs megacommit, part 2:
- Merge the pfs_vnode and pfs_vdata structures, and make the vnode cache
a doubly-linked list. This eliminates the need to walk the list in
pfs_vncache_free().
- Add an exit callout which revokes vnodes associated with the process
that just exited. Since it needs to lock the cache when it does this,
pfs_vncache_mutex needs MTX_RECURSE.
- Add a third callback to the pfs_node structure. This one simply returns
non-zero if the specified requesting process is allowed to access the
specified node for the specified target process. This is used in
addition to the usual permission checks, e.g. when certain files don't
make sense for certain (system) processes.
- Make sure that pfs_lookup() and pfs_readdir() don't yap about files
which aren't pfs_visible(). Also check pfs_visible() before performing
reads and writes, to prevent the kind of races reported in SA-00:77 and
SA-01:55 (fork a child, open /proc/child/ctl, have that child fork a
setuid binary, and assume control of it).
- Add some more trace points.
- Rearrange the flag constants a little to simplify specifying and testing
for readability and writeability.
pseudofs_vnops.c:
- Track the aforementioned change.
- Add checks to pfs_open() to prevent opening read-only files for writing
or vice versa (pfs_{read,write} would block the actual reads and writes,
but it's still a bug to allow the open() to succeed). Also, return
EOPNOTSUPP if the caller attempts to lock the file.
- Add more trace points.
- Remove hardcoded uid, gid, mode from struct pfs_node; make pfs_getattr()
smart enough to get it right most of the time, and allow for callbacks
to handle the remaining cases. Rework the definition macros to match.
- Add lots of (conditional) debugging output.
- Fix a long-standing bug inherited from procfs: don't pretend to be a
read-only file system. Instead, return EOPNOTSUPP for operations we
truly can't support and allow others to fail silently. In particular,
pfs_lookup() now treats CREATE as LOOKUP. This may need more work.
- In pfs_lookup(), if the parent node is process-dependent, check that
the process in question still exists.
- Implement pfs_open() - its only current function is to check that the
process opening the file can see the process it belongs to.
- Finish adding support for writeable nodes.
- Bump module version number.
- Introduce lots of new bugs.
- Use some simple #define's at the top of the files for proc -> thread
changes instead of having lots of needless #ifdef's in the code.
- Don't try to use struct thread in !FreeBSD code.
- Don't use a few struct lwp's in some of the NetBSD code since it isn't
in their HEAD.
The new diff relative to before KSE is now signficantly smaller and easier
to maintain.
Note ALL MODULES MUST BE RECOMPILED
make the kernel aware that there are smaller units of scheduling than the
process. (but only allow one thread per process at this time).
This is functionally equivalent to teh previousl -current except
that there is a thread associated with each process.
Sorry john! (your next MFC will be a doosie!)
Reviewed by: peter@freebsd.org, dillon@freebsd.org
X-MFC after: ha ha ha ha
linux_getdents uses VOP_READDIR( ..., &ncookies, &cookies ) instead of
VOP_READDIR( ..., NULL, NULL ) because it seems to need the offsets for
linux_dirent and sizeof(dirent) != sizeof(linux_dirent)...
PR: 29467
Submitted by: Michael Reifenberger <root@nihil.plaut.de>
Reviewed by: phk
as there are now "unusual" protection properties to Pmem that differ
from the other files. While I'm at it, introduce proc locking for
the other files, which was previously present only in the Pmem case.
Obtained from: TrustedBSD Project
and so special-casing was introduced to provide extra procfs privilege
to the kmem group. With the advent of non-setgid kmem ps, this code
is no longer required, and in fact, can is potentially harmful as it
allocates privilege to a gid that is increasingly less meaningful.
Knowledge of specific gid's in kernel is also generally bad precedent,
as the kernel security policy doesn't distinguish gid's specifically,
only uid 0.
This commit removes reference to kmem in procfs, both in terms of
access control decisions, and the applying of gid kmem to the
/proc/*/mem file, simplifying the associated code considerably.
Processes are still permitted to access the mem file based on
the debugging policy, so ps -e still works fine for normal
processes and use.
Reviewed by: tmm
Obtained from: TrustedBSD Project
The p_can(...) construct was a premature (and, it turns out,
awkward) abstraction. The individual calls to p_canxxx() better
reflect differences between the inter-process authorization checks,
such as differing checks based on the type of signal. This has
a side effect of improving code readability.
o Replace direct credential authorization checks in ktrace() with
invocation of p_candebug(), while maintaining the special case
check of KTR_ROOT. This allows ktrace() to "play more nicely"
with new mandatory access control schemes, as well as making its
authorization checks consistent with other "debugging class"
checks.
o Eliminate "privused" construct for p_can*() calls which allowed the
caller to determine if privilege was required for successful
evaluation of the access control check. This primitive is currently
unused, and as such, serves only to complicate the API.
Approved by: ({procfs,linprocfs} changes) des
Obtained from: TrustedBSD Project
(this commit is just the first stage). Also add various GIANT_ macros to
formalize the removal of Giant, making it easy to test in a more piecemeal
fashion. These macros will allow us to test fine-grained locks to a degree
before removing Giant, and also after, and to remove Giant in a piecemeal
fashion via sysctl's on those subsystems which the authors believe can
operate without Giant.
Replace the a.out emulation of 'struct linker_set' with something
a little more flexible. <sys/linker_set.h> now provides macros for
accessing elements and completely hides the implementation.
The linker_set.h macros have been on the back burner in various
forms since 1998 and has ideas and code from Mike Smith (SET_FOREACH()),
John Polstra (ELF clue) and myself (cleaned up API and the conversion
of the rest of the kernel to use it).
The macros declare a strongly typed set. They return elements with the
type that you declare the set with, rather than a generic void *.
For ELF, we use the magic ld symbols (__start_<setname> and
__stop_<setname>). Thanks to Richard Henderson <rth@redhat.com> for the
trick about how to force ld to provide them for kld's.
For a.out, we use the old linker_set struct.
NOTE: the item lists are no longer null terminated. This is why
the code impact is high in certain areas.
The runtime linker has a new method to find the linker set
boundaries depending on which backend format is in use.
linker sets are still module/kld unfriendly and should never be used
for anything that may be modular one day.
Reviewed by: eivind
not behaving correctly. Fix by attaching to the correct socket.
Also call so{rw}wakeup in addition to the fifo wakeup, so that any
kqfilters attached to the socket buffer get poked.
Only tun0 -> tun32767 may now be opened as struct ifnet's if_unit
is a short.
It's now possible to open /dev/tun and get a handle back for an available
tun device (use devname to find out what you got).
The implementation uses rman by popular demand (and against my judgement)
to track opened devices and uses the new dev_depends() to ensure that
all make_dev()d devices go away before the module is unloaded.
Reviewed by: phk
dev_t. The dev_depends(dev_t, dev_t) function is for tying them
to each other.
When destroy_dev() is called on a dev_t, all dev_t's depending
on it will also be destroyed (depth first order).
Rewrite the make_dev_alias() to use this dependency facility.
kern/subr_disk.c:
Make the disk mini-layer use dependencies to make sure all
relevant dev_t's are removed when the disk disappears.
Make the disk mini-layer precreate some magic sub devices
which the disk/slice/label code expects to be there.
kern/subr_disklabel.c:
Remove some now unneeded variables.
kern/subr_diskmbr.c:
Remove some ancient, commented out code.
kern/subr_diskslice.c:
Minor cleanup. Use name from dev_t instead of dsname()
real uid, saved uid, real gid, and saved gid to ucred, as well as the
pcred->pc_uidinfo, which was associated with the real uid, only rename
it to cr_ruidinfo so as not to conflict with cr_uidinfo, which
corresponds to the effective uid.
o Remove p_cred from struct proc; add p_ucred to struct proc, replacing
original macro that pointed.
p->p_ucred to p->p_cred->pc_ucred.
o Universally update code so that it makes use of ucred instead of pcred,
p->p_ucred instead of p->p_pcred, cr_ruidinfo instead of p_uidinfo,
cr_{r,sv}{u,g}id instead of p_*, etc.
o Remove pcred0 and its initialization from init_main.c; initialize
cr_ruidinfo there.
o Restruction many credential modification chunks to always crdup while
we figure out locking and optimizations; generally speaking, this
means moving to a structure like this:
newcred = crdup(oldcred);
...
p->p_ucred = newcred;
crfree(oldcred);
It's not race-free, but better than nothing. There are also races
in sys_process.c, all inter-process authorization, fork, exec, and
exit.
o Remove sigio->sio_ruid since sigio->sio_ucred now contains the ruid;
remove comments indicating that the old arrangement was a problem.
o Restructure exec1() a little to use newcred/oldcred arrangement, and
use improved uid management primitives.
o Clean up exit1() so as to do less work in credential cleanup due to
pcred removal.
o Clean up fork1() so as to do less work in credential cleanup and
allocation.
o Clean up ktrcanset() to take into account changes, and move to using
suser_xxx() instead of performing a direct uid==0 comparision.
o Improve commenting in various kern_prot.c credential modification
calls to better document current behavior. In a couple of places,
current behavior is a little questionable and we need to check
POSIX.1 to make sure it's "right". More commenting work still
remains to be done.
o Update credential management calls, such as crfree(), to take into
account new ruidinfo reference.
o Modify or add the following uid and gid helper routines:
change_euid()
change_egid()
change_ruid()
change_rgid()
change_svuid()
change_svgid()
In each case, the call now acts on a credential not a process, and as
such no longer requires more complicated process locking/etc. They
now assume the caller will do any necessary allocation of an
exclusive credential reference. Each is commented to document its
reference requirements.
o CANSIGIO() is simplified to require only credentials, not processes
and pcreds.
o Remove lots of (p_pcred==NULL) checks.
o Add an XXX to authorization code in nfs_lock.c, since it's
questionable, and needs to be considered carefully.
o Simplify posix4 authorization code to require only credentials, not
processes and pcreds. Note that this authorization, as well as
CANSIGIO(), needs to be updated to use the p_cansignal() and
p_cansched() centralized authorization routines, as they currently
do not take into account some desirable restrictions that are handled
by the centralized routines, as well as being inconsistent with other
similar authorization instances.
o Update libkvm to take these changes into account.
Obtained from: TrustedBSD Project
Reviewed by: green, bde, jhb, freebsd-arch, freebsd-audit
This fixes a number of warnings relating to removed cloned devices.
It also makes it possible to recreate deleted devices with
mknod(2). The major/minor arguments are ignored.
systems were repo-copied from sys/miscfs to sys/fs.
- Renamed the following file systems and their modules:
fdesc -> fdescfs, portal -> portalfs, union -> unionfs.
- Renamed corresponding kernel options:
FDESC -> FDESCFS, PORTAL -> PORTALFS, UNION -> UNIONFS.
- Install header files for the above file systems.
- Removed bogus -I${.CURDIR}/../../sys CFLAGS from userland
Makefiles.
vm_mtx does not recurse and is required for most low level
vm operations.
faults can not be taken without holding Giant.
Memory subsystems can now call the base page allocators safely.
Almost all atomic ops were removed as they are covered under the
vm mutex.
Alpha and ia64 now need to catch up to i386's trap handlers.
FFS and NFS have been tested, other filesystems will need minor
changes (grabbing the vm lock when twiddling page properties).
Reviewed (partially) by: jake, jhb
vn_start_write() on the given vnode will be successful. VOP_LEASE() may
help to solve this problem, but its return value ignored nearly everywhere.
For now just assume that the missing upper layer on write means insufficient
access rights (which is correct for most cases).
the number of references on the filesystem root vnode to be both
expected and released. Many filesystems hold an extra reference on
the filesystem root vnode, which must be accounted for when
determining if the filesystem is busy and then released if it isn't
busy. The old `skipvp' approach required individual filesystem
xxx_unmount functions to re-implement much of vflush()'s logic to
deal with the root vnode.
All 9 filesystems that hold an extra reference on the root vnode
got the logic wrong in the case of forced unmounts, so `umount -f'
would always fail if there were any extra root vnode references.
Fix this issue centrally in vflush(), now that we can.
This commit also fixes a vnode reference leak in devfs, which could
result in idle devfs filesystems that refuse to unmount.
Reviewed by: phk, bp
If for some reason DEVFS is undesired, the "NODEVFS" option is
needed now.
Pending any significant issues, DEVFS will be made mandatory in
-current on july 1st so that we can start reaping the full
benefits of having it.
