Commit Graph

2287 Commits

Author SHA1 Message Date
Andriy Gapon
84206c741f fs/udf: fix incorrect error return (-1) when reading a large dir
Not enough space in user-land buffer is not an error, userland
will read further until eof is reached. So instead of propagating
-1 to caller we convert it to zero/success.

cd9660 code works exactly the same way.

PR:		kern/78987
Reviewed by:	jhb (mentor)
Approved by:	jhb (mentor)
2009-02-19 15:05:30 +00:00
Dag-Erling Smørgrav
655fcdaa00 Fix a logic bug that caused the pfs_attr method to be called only for
PFS_PROCDEP nodes.

Submitted by:	Andrew Brampton <brampton@gmail.com>
MFC after:	2 weeks
2009-02-16 15:17:26 +00:00
John Baldwin
ea77ff0a15 Use shared vnode locks when invoking VOP_READDIR().
MFC after:	1 month
2009-02-13 18:18:14 +00:00
John Baldwin
0b2ab5ec8e - Consolidate error handling in the cd9660 and udf mount routines.
- Always read the character device pointer while the associated devfs vnode
  is locked.  Also, use dev_ref() to obtain a new reference on the vnode for
  the mountpoint.  This reference is released on unmount.  This mirrors the
  earlier fix to FFS.

Reviewed by:	kib
2009-02-11 22:22:26 +00:00
John Baldwin
4ad0d60bb8 Mark udf(4) MPSAFE and add support for shared vnode locks during pathname
lookups:
- Honor the caller's locking flags in udf_root() and udf_vget().
- Set VV_ROOT for the root vnode in udf_vget() instead of only doing it in
  udf_root().
- Honor the requested locking flags during pathname lookups in udf_lookup().
- Release the buffer holding the directory data before looking up the vnode
  for a given file to avoid a LOR between the "udf" vnode locks and
  "bufwait".
- Use vn_vget_ino() to handle ".." lookups.
- Special case "." lookups instead of calling udf_vget().  We have to do
  extra checking for the vnode lock for "." lookups.
2009-02-09 20:50:23 +00:00
John Baldwin
a40ad858de Use the same style as the rest of the file for the optional data string
after each path component rather than a GCC-ism.
2009-02-09 20:42:51 +00:00
Konstantin Belousov
e3c7e75305 Lookup up the directory entry for the tmpfs node that are deleted by
both node pointer and name component. This does the right thing for
hardlinks to the same node in the same directory.

Submitted by:	Yoshihiro Ota <ota j email ne jp>
PR:	kern/131356
MFC after:	2 weeks
2009-02-08 19:18:33 +00:00
John Baldwin
e3024df2e0 Add rudimentary support for symbolic links on UDF. Links are stored as a
sequence of pathname components.  We walk the list building a string in
the caller's passed in buffer.  Currently this only handles path names
in CS8 (character set 8) as that is what mkisofs generates for UDF images.

MFC after:	1 month
2009-02-06 22:24:03 +00:00
John Baldwin
61e69c80e4 Add support for fifos to UDF:
- Add a separate set of vnode operations that inherits from the fifo ops
  and use it for fifo nodes.
- Add a VOP_SETATTR() method that allows setting the size (by silently
  ignoring the requests) of fifos.  This is to allow O_TRUNC opens of
  fifo devices (e.g. I/O redirection in shells using ">").
- Add a VOP_PRINT() handler while I'm here.
2009-02-06 20:09:14 +00:00
John Baldwin
8941aad19b Tweak the output of VOP_PRINT/vn_printf() some.
- Align the fifo output in fifo_print() with other vn_printf() output.
- Remove the leading space from lockmgr_printinfo() so its output lines up
  in vn_printf().
- lockmgr_printinfo() now ends with a newline, so remove an extra newline
  from vn_printf().
2009-02-06 20:06:48 +00:00
Bjoern A. Zeeb
13fd4d2163 After r186194 the *fs_strategy() functions always return 0.
So we are no longer interested in the error returned from
the *fs_doio() functions. With that we can remove the
error variable as its value is unused now.

Submitted by:	Christoph Mallon christoph.mallon@gmx.de
2009-01-31 18:06:34 +00:00
Bjoern A. Zeeb
7956d34b95 Remove unused local variables.
Submitted by:	Christoph Mallon christoph.mallon@gmx.de
Reviewed by:	kib
MFC after:	2 weeks
2009-01-31 17:36:22 +00:00
Ed Schouten
f3b86a5fd7 Mark most often used sysctl's as MPSAFE.
After running a `make buildkernel', I noticed most of the Giant locks in
sysctl are only caused by a very small amount of sysctl's:

- sysctl.name2oid. This one is locked by SYSCTL_LOCK, just like
  sysctl.oidfmt.

- kern.ident, kern.osrelease, kern.version, etc. These are just constant
  strings.

- kern.arandom, used by the stack protector. It is already protected by
  arc4_mtx.

I also saw the following sysctl's show up. Not as often as the ones
above, but still quite often:

- security.jail.jailed. Also mark security.jail.list as MPSAFE. They
  don't need locking or already use allprison_lock.

- kern.devname, used by devname(3), ttyname(3), etc.

This seems to reduce Giant locking inside sysctl by ~75% in my primitive
test setup.
2009-01-28 19:58:05 +00:00
Warner Losh
bb5d3b71d3 Use the correct field name for the size of the sierra_id. While this
is the same size as id, and is unlikely to change, it seems better to
use the correct field here.  There's no difference in the generated
code.
2009-01-28 19:09:49 +00:00
John Baldwin
c222ece0cc Mark cd9660 MPSAFE and add support for using shared vnode locks during
pathname lookups.
- Remove 'i_offset' and 'i_ino' from the ISO node structure and replace
  them with local variables in the lookup routine instead.
- Cache a copy of 'i_diroff' for use during a lookup in a local variable.
- Save a copy of the found directory entry in a malloc'd buffer after a
  successfull lookup before getting the vnode.  This allows us to release
  the buffer holding the directory block before calling vget() which
  otherwise resulted in a LOR between "bufwait" and the vnode lock.
- Use an inlined version of vn_vget_ino() to handle races with ..
  lookups.  I had to inline the code here since cd9660 uses an internal
  vget routine to save a disk I/O that would otherwise re-read the
  directory block.
- Honor the requested locking flags during lookups to allow for shared
  locking.
- Honor the requested locking flags passed to VFS_ROOT() and VFS_VGET()
  similar to UFS.
- Don't make every ISO 9660 vnode hold a reference on the vnode of the
  underlying device vnode of the mountpoint.  The mountpoint already
  holds a suitable reference.
2009-01-28 18:54:56 +00:00
John Baldwin
04c98d464f Sync with ufs_vnops.c:1.245 and remove support for accessing device nodes
in ISO 9660 filesystems.
2009-01-28 18:46:29 +00:00
John Baldwin
be09858abd Assert an exclusive vnode lock for fifo_cleanup() and fifo_close() since
they change v_fifoinfo.

Discussed with:	ups (a while ago)
2009-01-28 18:10:57 +00:00
Ed Schouten
a4611ab612 Last step of splitting up minor and unit numbers: remove minor().
Inside the kernel, the minor() function was responsible for obtaining
the device minor number of a character device. Because we made device
numbers dynamically allocated and independent of the unit number passed
to make_dev() a long time ago, it was actually a misnomer. If you really
want to obtain the device number, you should use dev2udev().

We already converted all the drivers to use dev2unit() to obtain the
device unit number, which is still used by a lot of drivers. I've
noticed not a single driver passes NULL to dev2unit(). Even if they
would, its behaviour would make little sense. This is why I've removed
the NULL check.

Ths commit removes minor(), minor2unit() and unit2minor() from the
kernel. Because there was a naming collision with uminor(), we can
rename umajor() and uminor() back to major() and minor(). This means
that the makedev(3) manual page also applies to kernel space code now.

I suspect umajor() and uminor() isn't used that often in external code,
but to make it easier for other parties to port their code, I've
increased __FreeBSD_version to 800062.
2009-01-28 17:57:16 +00:00
Konstantin Belousov
e442a285a6 The kernel may do unbalanced calls to fifo_close() for fifo vnode,
without corresponding number of fifo_open(). This causes assertion
failure in fifo_close() due to vp->v_fifoinfo being NULL for kernel
with INVARIANTS, or NULL pointer dereference otherwise. In fact, we may
ignore excess calls to fifo_close() without bad consequences.

Turn KASSERT() into the return, and print warning for now.

Tested by:	pho
Reviewed by:	rwatson
MFC after:	2 weeks
2009-01-26 14:21:00 +00:00
Edward Tomasz Napierala
abb0cbf9c9 Turn a "panic: non-decreasing id" into an error printf. This seems
to be caused by a metadata corruption that occurs quite often after
unplugging a pendrive during write activity.

Reviewed by:	scottl
Approved by:	rwatson (mentor)
Sponsored by:	FreeBSD Foundation
2009-01-13 22:35:26 +00:00
Edward Tomasz Napierala
f99f675d5a Fix msdosfs_print(), which in turn fixes "show lockedvnods" for msdosfs
vnodes.

Reviewed by:	kib
Approved by:	rwatson (mentor)
Sponsored by:	FreeBSD Foundation
2009-01-11 17:11:01 +00:00
Joe Marcus Clarke
4424c9d053 Fix a deadlock which can occur due to a pseudofs vnode not getting unlocked.
Reported by:	Richard Todd <rmtodd@ichotolot.servalan.com>
Reviewed by:	kib
Approved by:	kib
2009-01-09 22:06:48 +00:00
Edward Tomasz Napierala
71624181c8 Don't panic with "vinvalbuf: dirty bufs" when the mounted device that was
being written to goes away.

Reviewed by:	kib, scottl
Approved by:	rwatson (mentor)
Sponsored by:	FreeBSD Foundation
2009-01-08 19:13:34 +00:00
Joe Marcus Clarke
e7f54c1b71 Add a VOP_VPTOCNP implementation for pseudofs which covers file systems
such as procfs and linprocfs.

This implementation's locking was enhanced by kib.

Reviewed by:	kib
		des
Approved by:	des
		kib
Tested by:	pho
2008-12-30 21:49:39 +00:00
Konstantin Belousov
78e4cea909 When the insmntque() in the pfs_vncache_alloc() fails, vop_reclaim calls
pfs_vncache_free() that removes pvd from the list, while it is not yet
put on the list.

Prevent the invalid removal from the list by clearing pvd_next and
pvd_prev for the newly allocated pvd, and only move pfs_vncache list
head when the pvd was at the head.

Suggested and approved by:	des
MFC after:	2 weeks
2008-12-29 13:25:58 +00:00
Konstantin Belousov
22a448c4d9 vm_map_lock_read() does not increment map->timestamp, so we should
compare map->timestamp with saved timestamp after map read lock is
reacquired, not with saved timestamp + 1. The only consequence of the +1
was unconditional lookup of the next map entry, though.

Tested by:	pho
Approved by:	des
MFC after:	2 weeks
2008-12-29 12:45:11 +00:00
Konstantin Belousov
c990bf0896 Use curproc->p_sysent->sv_flags bit SV_ILP32 for detection of the 32 bit
caller, instead of direct comparision with ia32_freebsd_sysvec.

Tested by:	pho
Approved by:	des
MFC after:	2 weeks
2008-12-29 12:41:32 +00:00
Konstantin Belousov
505d02eebe Drop the pseudofs vnode lock around call to pfs_read handler. The handler
may need to lock arbitrary vnodes, causing either lock order reversal or
recursive vnode lock acquisition.

Tested by:	pho
Approved by:	des
MFC after:	2 weeks
2008-12-29 12:12:23 +00:00
Konstantin Belousov
99ec92c962 After the pfs_vncache_mutex is dropped, another thread may attempt to
do pfs_vncache_alloc() for the same pfs_node and pid. In this case, we
could end up with two vnodes for the pair. Recheck the cache under the
locked pfs_vncache_mutex after all sleeping operations are done [1].

This case mostly cannot happen now because pseudofs uses exclusive vnode
locking for lookup. But it does drop the vnode lock for dotdot lookups,
and Marcus' pseudofs_vptocnp implementation is vulnerable too.

Do not call free() on the struct pfs_vdata after insmntque() failure,
because vp->v_data points to the structure, and pseudofs_reclaim()
frees it by the call to pfs_vncache_free().

Tested by:	pho [1]
Approved by:	des
MFC after:	2 weeks
2008-12-29 12:07:18 +00:00
Edward Tomasz Napierala
0da50f6ef8 According to phk@, VOP_STRATEGY should never, _ever_, return
anything other than 0.  Make it so.  This fixes
"panic: VOP_STRATEGY failed bp=0xc320dd90 vp=0xc3b9f648",
encountered when writing to an orphaned filesystem.  Reason
for the panic was the following assert:
KASSERT(i == 0, ("VOP_STRATEGY failed bp=%p vp=%p", bp, bp->b_vp));
at vfs_bio:bufstrategy().

Reviewed by:	scottl, phk
Approved by:	rwatson (mentor)
Sponsored by:	FreeBSD Foundation
2008-12-16 21:13:11 +00:00
Konstantin Belousov
c7462f4387 Reference the vmspace of the process being inspected by procfs, linprocfs
and sysctl kern_proc_vmmap handlers.

Reported and tested by:	pho
Reviewed by:	rwatson, des
MFC after:	1 week
2008-12-12 12:12:36 +00:00
Konstantin Belousov
c7c7520a95 Do not leak defs_de_interlock on error.
Another pointy hat for my collection.
2008-12-12 11:10:10 +00:00
Joe Marcus Clarke
4c44fd376a Implement VOP_VPTOCNP for devfs. Directory and character device vnodes are
properly translated to their component names.

Reviewed by:	arch
Approved by:	kib
2008-12-12 01:00:38 +00:00
Joe Marcus Clarke
933efde2e5 Add a simple VOP_VPTOCNP implementation for deadfs which returns EBADF.
Reviewed by:	arch
Approved by:	kib
2008-12-12 00:59:36 +00:00
Konstantin Belousov
c96f374195 Relock user map earlier, to have the lock held when break leaves the
loop earlier due to sbuf error.

Pointy hat to:	me
Submitted by:	dchagin
2008-12-10 16:11:09 +00:00
Konstantin Belousov
9499cb83bf Make two style changes to create new commit and document proper commit
message for r185765.

Noted by:	rdivacky
Requested by:	des

Commit message for r185765 should be:
In procfs map handler, and in linprocfs maps handler, do not call
vn_fullpath() while having vm map locked. This is done in anticipation
of the vop_vptocnp commit, that would make vn_fullpath sometime
acquire vnode lock.

Also, in linprocfs, maps handler already acquires vnode lock.

No objections from:	des
MFC after:	2 week
2008-12-08 13:15:31 +00:00
Konstantin Belousov
5a66e0259b Change the linprocfs <pid>/maps and procfs <pid>/map handlers to use
sbuf instead of doing uiomove. This allows for reads from non-zero
offsets to work.

Patch is forward-ported des@' one, and was adopted to current code
by dchagin@ and me.

Reviewed by:	des (linprocfs part)
PR:	kern/101453
MFC after:	1 week
2008-12-08 12:34:52 +00:00
Tim Kientzle
5c423e0640 The timezone byte is a signed value, treat it as such.
Otherwise, time zone information for time zones west of
GMT gets discarded.

PR:		kern/128934
Submitted by:	J.R. Oldroyd
MFC after:	4 days
2008-11-27 06:21:04 +00:00
Konstantin Belousov
5147a76a0e In null_lookup(), do the needed cleanup instead of panicing saying
the cleanup is needed.

Reported by:	kris, pho
Tested by:	pho
MFC after:	2 weeks
2008-11-26 13:41:15 +00:00
Ulf Lilleengen
f7b8cfa890 - Support IEEE_P1282 and IEEE_1282 tags in the rock ridge extensions record.
PR:		kern/128942
Submitted by:	"J.R. Oldroyd" <fbsd - at - opal.com>
2008-11-26 13:09:45 +00:00
Daichi GOTO
16385727ce Simplify mode_t check treatment (suggested by trasz).
By semantical view, trasz's code is better than prior one.

Submitted by:	trasz
Reviewed by:	Masanori OZAWA <ozawa@ongs.co.jp>
2008-11-25 03:49:41 +00:00
Daichi GOTO
1e5da15a63 Fixes Unionfs socket issue reported as kern/118346.
PR:		118346
Submitted by:	Masanori OZAWA <ozawa@ongs.co.jp>
Discussed at:	devsummit Strassburg, EuroBSDCon2008
Discussed with:	rwatson, gnn, hrs
MFC after:	2 week
2008-11-25 03:18:35 +00:00
John Baldwin
66a6ea1de2 - Fix a typo in a comment.
- Whitespace fix.
- Remove #if 0'd BSD 4.x code for flushing busy buffers from a mountpoint
  during an unmount.  FreeBSD uses vflush() for this.
2008-11-18 23:19:43 +00:00
John Baldwin
77ddca67d5 When looking up the vnode for the device to mount the filesystem on,
ask NDINIT to return a locked vnode instead of letting it drop the
lock and return a referenced vnode and then relock the vnode a few
lines down.  This matches the behavior of other filesystem mount routines.
2008-11-18 23:18:37 +00:00
John Baldwin
1ea456e7a6 Remove copy/paste code from UFS to handle sparse blocks. While Rock
Ridge does support sparse files, the cd9660 code does not currently
support them.
2008-11-18 23:15:17 +00:00
John Baldwin
05b1d36516 Remove unused i_flags field and IN_ACCESS flag from cd9660 in-memory
i-nodes.  cd9660 doesn't support access times.
2008-11-18 23:13:40 +00:00
John Baldwin
2ff47c5f18 Remove unnecessary locking around vn_fullpath(). The vnode lock for the
vnode in question does not need to be held.  All the data structures used
during the name lookup are protected by the global name cache lock.
Instead, the caller merely needs to ensure a reference is held on the
vnode (such as vhold()) to keep it from being freed.

In the case of procfs' <pid>/file entry, grab the process lock while we
gain a new reference (via vhold()) on p_textvp to fully close races with
execve(2).

For the kern.proc.vmmap sysctl handler, use a shared vnode lock around
the call to VOP_GETATTR() rather than an exclusive lock.

MFC after:	1 month
2008-11-04 19:04:01 +00:00
John Baldwin
7265164f53 Don't pass WANTPARENT to the pathname lookup of the mount point for a
unionfs mount just so we can immediately drop the reference on the parent
directory vnode without using it.
2008-11-04 18:54:44 +00:00
Edward Tomasz Napierala
ea49f15447 Fix few missed accmode changes in coda.
Approved by:	rwatson (mentor)
2008-11-03 16:36:23 +00:00
Doug Rabson
a9148abd9d Implement support for RPCSEC_GSS authentication to both the NFS client
and server. This replaces the RPC implementation of the NFS client and
server with the newer RPC implementation originally developed
(actually ported from the userland sunrpc code) to support the NFS
Lock Manager.  I have tested this code extensively and I believe it is
stable and that performance is at least equal to the legacy RPC
implementation.

The NFS code currently contains support for both the new RPC
implementation and the older legacy implementation inherited from the
original NFS codebase. The default is to use the new implementation -
add the NFS_LEGACYRPC option to fall back to the old code. When I
merge this support back to RELENG_7, I will probably change this so
that users have to 'opt in' to get the new code.

To use RPCSEC_GSS on either client or server, you must build a kernel
which includes the KGSSAPI option and the crypto device. On the
userland side, you must build at least a new libc, mountd, mount_nfs
and gssd. You must install new versions of /etc/rc.d/gssd and
/etc/rc.d/nfsd and add 'gssd_enable=YES' to /etc/rc.conf.

As long as gssd is running, you should be able to mount an NFS
filesystem from a server that requires RPCSEC_GSS authentication. The
mount itself can happen without any kerberos credentials but all
access to the filesystem will be denied unless the accessing user has
a valid ticket file in the standard place (/tmp/krb5cc_<uid>). There
is currently no support for situations where the ticket file is in a
different place, such as when the user logged in via SSH and has
delegated credentials from that login. This restriction is also
present in Solaris and Linux. In theory, we could improve this in
future, possibly using Brooks Davis' implementation of variant
symlinks.

Supporting RPCSEC_GSS on a server is nearly as simple. You must create
service creds for the server in the form 'nfs/<fqdn>@<REALM>' and
install them in /etc/krb5.keytab. The standard heimdal utility ktutil
makes this fairly easy. After the service creds have been created, you
can add a '-sec=krb5' option to /etc/exports and restart both mountd
and nfsd.

The only other difference an administrator should notice is that nfsd
doesn't fork to create service threads any more. In normal operation,
there will be two nfsd processes, one in userland waiting for TCP
connections and one in the kernel handling requests. The latter
process will create as many kthreads as required - these should be
visible via 'top -H'. The code has some support for varying the number
of service threads according to load but initially at least, nfsd uses
a fixed number of threads according to the value supplied to its '-n'
option.

Sponsored by:	Isilon Systems
MFC after:	1 month
2008-11-03 10:38:00 +00:00
Robert Watson
2b7da2dbf1 Catch up with netsmb locking: explicit thread arguments no longer required. 2008-11-02 23:20:27 +00:00
Edward Tomasz Napierala
2a9e5e2e7c Remove the call to getinoquota() from ntfs_access. How did it get there?!
Approved by:	rwatson (mentor)
2008-11-02 11:49:19 +00:00
Edward Tomasz Napierala
15bc6b2bd8 Introduce accmode_t. This is required for NFSv4 ACLs - it will be neccessary
to add more V* constants, and the variables changed by this patch were often
being assigned to mode_t variables, which is 16 bit.

Approved by:	rwatson (mentor)
2008-10-28 13:44:11 +00:00
Dag-Erling Smørgrav
e11e3f187d Fix a number of style issues in the MALLOC / FREE commit. I've tried to
be careful not to fix anything that was already broken; the NFSv4 code is
particularly bad in this respect.
2008-10-23 20:26:15 +00:00
Dag-Erling Smørgrav
1ede983cc9 Retire the MALLOC and FREE macros. They are an abomination unto style(9).
MFC after:	3 months
2008-10-23 15:53:51 +00:00
Robert Watson
f17c6f031e The locking in portalfs's socket connect code is no less correct than
identical code in connect(2), so remove XXX that it might be incorrect.

MFC after:	3 days
2008-10-12 19:23:02 +00:00
Attilio Rao
0d7935fd01 Remove the struct thread unuseful argument from bufobj interface.
In particular following functions KPI results modified:
- bufobj_invalbuf()
- bufsync()

and BO_SYNC() "virtual method" of the buffer objects set.
Main consumers of bufobj functions are affected by this change too and,
in particular, functions which changed their KPI are:
- vinvalbuf()
- g_vfs_close()

Due to the KPI breakage, __FreeBSD_version will be bumped in a later
commit.

As a side note, please consider just temporary the 'curthread' argument
passing to VOP_SYNC() (in bufsync()) as it will be axed out ASAP

Reviewed by:	kib
Tested by:	Giovanni Trematerra <giovanni dot trematerra at gmail dot com>
2008-10-10 21:23:50 +00:00
Robert Watson
4759ebf015 Use soconnect2() rather than directly invoking uipc_connect2() to
interconnect two UNIX domain sockets.

MFC after:	3 days
2008-10-06 18:38:50 +00:00
Konstantin Belousov
9a1e630dfd Change the linprocfs <pid>/maps and procfs <pid>/map handlers to use
sbuf instead of doing uiomove. This allows for reads from non-zero
offsets to work.

Patch is forward-ported des@' one, and was adopted to current code
by dchagin@ and me.

Reviewed by:	des (linprocfs part)
PR:	kern/101453
MFC after:	1 week
2008-10-04 14:08:16 +00:00
Edward Tomasz Napierala
a37d6ec935 Fix Vflags abuse in fdescfs. There should be no functional changes.
Approved by:	rwatson (mentor)
2008-10-03 23:21:14 +00:00
Edward Tomasz Napierala
464119c422 Fix Vflags abuse in cd9660. There should be no functional changes.
Approved by:	rwatson (mentor)
2008-10-03 23:17:22 +00:00
Marko Zec
8b615593fc Step 1.5 of importing the network stack virtualization infrastructure
from the vimage project, as per plan established at devsummit 08/08:
http://wiki.freebsd.org/Image/Notes200808DevSummit

Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator
macros, and CURVNET_SET() context setting macros, all currently
resolving to NOPs.

Prepare for virtualization of selected SYSCTL objects by introducing a
family of SYSCTL_V_*() macros, currently resolving to their global
counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT().

Move selected #defines from sys/sys/vimage.h to newly introduced header
files specific to virtualized subsystems (sys/net/vnet.h,
sys/netinet/vinet.h etc.).

All the changes are verified to have zero functional impact at this
point in time by doing MD5 comparision between pre- and post-change
object files(*).

(*) netipsec/keysock.c did not validate depending on compile time options.

Implemented by:	julian, bz, brooks, zec
Reviewed by:	julian, bz, brooks, kris, rwatson, ...
Approved by:	julian (mentor)
Obtained from:	//depot/projects/vimage-commit2/...
X-MFC after:	never
Sponsored by:	NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
Konstantin Belousov
7818e0a545 Save previous content of the td_fpop before storing the current
filedescriptor into it. Make sure that td_fpop is NULL when calling
d_mmap from dev_pager_getpages().

Change guards against td_fpop field being non-NULL with private state
for another device, and against sudden clearing the td_fpop. This
could occur when either a driver method calls another driver through
the filedescriptor operation, or a page fault happen while driver is
writing to a memory backed by another driver.

Noted by:	rwatson
Tested by:	rnoland
MFC after:	3 days
2008-09-26 14:50:49 +00:00
Ed Schouten
d3ce832719 Remove unit2minor() use from kernel code.
When I changed kern_conf.c three months ago I made device unit numbers
equal to (unneeded) device minor numbers. We used to require
bitshifting, because there were eight bits in the middle that were
reserved for a device major number. Not very long after I turned
dev2unit(), minor(), unit2minor() and minor2unit() into macro's.
The unit2minor() and minor2unit() macro's were no-ops.

We'd better not remove these four macro's from the kernel, because there
is a lot of (external) code that may still depend on them. For now it's
harmless to remove all invocations of unit2minor() and minor2unit().

Reviewed by:	kib
2008-09-26 14:19:52 +00:00
David E. O'Brien
ae72afe0f2 The kernel implemented 'memcmp' is an alias for 'bcmp'. However, memcmp
and bcmp are not the same thing.  'man bcmp' states that the return is
"non-zero" if the two byte strings are not identical.  Where as,
'man memcmp' states that the return is the "difference between the
first two differing bytes (treated as unsigned char values" if the
two byte strings are not identical.

So provide a proper memcmp(9), but it is a C implementation not a tuned
assembly implementation.  Therefore bcmp(9) should be preferred over memcmp(9).
2008-09-23 14:45:10 +00:00
Ed Schouten
219cc94999 Already initialize the vfs timestamps inside the cdev upon allocation.
In the MPSAFE TTY branch I noticed the vfs timestamps inside devfs were
allocated with 0, where the getattr() routine bumps the timestamps to
boottime if the value is below 3600. The reason why it has been designed
like this, is because timestamps during boot are likely to be invalid.

This means that device nodes that are created on demand (posix_openpt())
have timestamps with a value of boottime, which is not what we want.
Solve this by calling vfs_timestamp() inside devfs_alloc().

Discussed with:	kib
2008-09-21 14:02:43 +00:00
Konstantin Belousov
caf8aec886 fdescfs, devfs, mqueuefs, nfs, portalfs, pseudofs, tmpfs and xfs
initialize the vattr structure in VOP_GETATTR() with VATTR_NULL(),
vattr_null() or by zeroing it. Remove these to allow preinitialization
of fields work in vn_stat(). This is needed to get birthtime initialized
correctly.

Submitted by:   Jaakko Heinonen <jh saunalahti fi>
Discussed on:   freebsd-fs
MFC after:	1 month
2008-09-20 19:50:52 +00:00
Konstantin Belousov
4c5a20e3da Initialize va_rdev to NODEV instead of 0 or VNOVAL in VOP_GETATTR().
NODEV is more appropriate when va_rdev doesn't have a meaningful value.

Submitted by:   Jaakko Heinonen <jh saunalahti fi>
Suggested by:   bde
Discussed on:   freebsd-fs
MFC after:	1 month
2008-09-20 19:49:15 +00:00
Konstantin Belousov
86dacdfe2b Initialize va_flags and va_filerev properly in VOP_GETATTR(). Don't
initialize va_vaflags and va_spare because they are not part of the
VOP_GETATTR() API. Also don't initialize birthtime to ctime or zero.

Submitted by:   Jaakko Heinonen <jh saunalahti fi>
Reviewed by:	bde
Discussed on:   freebsd-fs
MFC after:	1 month
2008-09-20 19:46:45 +00:00
Ed Schouten
19c5cd6288 Fix two small typo's in comments in the nullfs vnops code.
Submitted by:	Jille Timmermans <jille quis cx>
2008-09-11 20:15:34 +00:00
Xin LI
e08d55674d Reflect license change of NetBSD code.
Obtained from:	NetBSD
MFC after:	3 days
2008-09-03 18:53:48 +00:00
Konstantin Belousov
67c7bbf39c In rev. 1.17 (r33548) of msdosfs_fat.c, relative cluster numbers were
replaced by file relative sector numbers as the buffer block number when
zero-padding a file during extension. Revert the change, it causes wrong
blocks filled with zeroes on seeking beyond end of file.

PR:	kern/47628
Submitted by:	tegge
MFC after:	3 days
2008-09-01 13:18:16 +00:00
Attilio Rao
0359a12ead Decontextualize the couplet VOP_GETATTR / VOP_SETATTR as the passed thread
was always curthread and totally unuseful.

Tested by: Giovanni Trematerra <giovanni dot trematerra at gmail dot com>
2008-08-28 15:23:18 +00:00
Ed Schouten
bc093719ca Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:

- Improved driver model:

  The old TTY layer has a driver model that is not abstract enough to
  make it friendly to use. A good example is the output path, where the
  device drivers directly access the output buffers. This means that an
  in-kernel PPP implementation must always convert network buffers into
  TTY buffers.

  If a PPP implementation would be built on top of the new TTY layer
  (still needs a hooks layer, though), it would allow the PPP
  implementation to directly hand the data to the TTY driver.

- Improved hotplugging:

  With the old TTY layer, it isn't entirely safe to destroy TTY's from
  the system. This implementation has a two-step destructing design,
  where the driver first abandons the TTY. After all threads have left
  the TTY, the TTY layer calls a routine in the driver, which can be
  used to free resources (unit numbers, etc).

  The pts(4) driver also implements this feature, which means
  posix_openpt() will now return PTY's that are created on the fly.

- Improved performance:

  One of the major improvements is the per-TTY mutex, which is expected
  to improve scalability when compared to the old Giant locking.
  Another change is the unbuffered copying to userspace, which is both
  used on TTY device nodes and PTY masters.

Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.

Obtained from:		//depot/projects/mpsafetty/...
Approved by:		philip (ex-mentor)
Discussed:		on the lists, at BSDCan, at the DevSummit
Sponsored by:		Snow B.V., the Netherlands
dcons(4) fixed by:	kan
2008-08-20 08:31:58 +00:00
Bjoern A. Zeeb
603724d3ab Commit step 1 of the vimage project, (network stack)
virtualization work done by Marko Zec (zec@).

This is the first in a series of commits over the course
of the next few weeks.

Mark all uses of global variables to be virtualized
with a V_ prefix.
Use macros to map them back to their global names for
now, so this is a NOP change only.

We hope to have caught at least 85-90% of what is needed
so we do not invalidate a lot of outstanding patches again.

Obtained from:	//depot/projects/vimage-commit2/...
Reviewed by:	brooks, des, ed, mav, julian,
		jamie, kris, rwatson, zec, ...
		(various people I forgot, different versions)
		md5 (with a bit of help)
Sponsored by:	NLnet Foundation, The FreeBSD Foundation
X-MFC after:	never
V_Commit_Message_Reviewed_By:	more people than the patch
2008-08-17 23:27:27 +00:00
Konstantin Belousov
f35db5f7ca Remove unnecessary locking around pointer fetch.
Requested by:   jhb
2008-08-12 19:34:45 +00:00
Robert Watson
4f7d1876d5 Introduce a new lock, hostname_mtx, and use it to synchronize access
to global hostname and domainname variables.  Where necessary, copy
to or from a stack-local buffer before performing copyin() or
copyout().  A few uses, such as in cd9660 and daemon_saver, remain
under-synchronized and will require further updates.

Correct a bug in which a failed copyin() of domainname would leave
domainname potentially corrupted.

MFC after:	3 weeks
2008-07-05 13:10:10 +00:00
Konstantin Belousov
813d71de08 The uniqdosname() function takes char[12] as it third argument.
Found by:	-fstack-protector
Reported by:	dougb
Tested by:	dougb, Rainer Hurling <rhurlin gwdg de>
MFC after:	3 days
2008-07-04 09:40:52 +00:00
Robert Watson
e54fdca237 Remove unused 'td' arguments from smbfs_hash_lock() and
smbfs_hash_unlock().

MFC after:	3 days
2008-07-01 07:51:16 +00:00
Oleksandr Tymoshenko
2da528a74f Get pointer to devfs_ruleset struct after garbage collection has been
performed. Otherwise if ruleset is used by given mountpoint and is empty
it's freed by devfs_ruleset_reap and pointer becomes bogus.

Submitted by:   Mateusz Guzik <mjguzik@gmail.com>
PR:             kern/124853
2008-06-22 14:34:38 +00:00
Konstantin Belousov
05427aafc6 Struct cdev is always the member of the struct cdev_priv. When devfs
needed to promote cdev to cdev_priv, the si_priv pointer was followed.

Use member2struct() to calculate address of the wrapping cdev_priv.
Rename si_priv to __si_reserved.

Tested by:	pho
Reviewed by:	ed
MFC after:	2 weeks
2008-06-16 17:34:59 +00:00
Konstantin Belousov
a0b454dc4b Do not redo the vnode tear-down work already done by insmntque() when
vnode cannot be put on the vnode list for mount.

Reported and tested by:	marck
Guilty party:	me
MFC after:	3 days
2008-06-15 18:40:58 +00:00
Ed Schouten
29d4cb241b Don't enforce unique device minor number policy anymore.
Except for the case where we use the cloner library (clone_create() and
friends), there is no reason to enforce a unique device minor number
policy. There are various drivers in the source tree that allocate unr
pools and such to provide minor numbers, without using them themselves.

Because we still need to support unique device minor numbers for the
cloner library, introduce a new flag called D_NEEDMINOR. All cdevsw's
that are used in combination with the cloner library should be marked
with this flag to make the cloning work.

This means drivers can now freely use si_drv0 to store their own flags
and state, making it effectively the same as si_drv1 and si_drv2. We
still keep the minor() and dev2unit() routines around to make drivers
happy.

The NTFS code also used the minor number in its hash table. We should
not do this anymore. If the si_drv0 field would be changed, it would no
longer end up in the same list.

Approved by:	philip (mentor)
2008-06-11 18:55:19 +00:00
Konstantin Belousov
ac8b6edd89 In cd9660_readdir vop, always initialize the idp->uio_off member.
The while loop that is assumed to initialize the uio_off later, may
be not entered at all, causing uninitialized value to be returned in
uio->uio_offset.

PR:  122925
Submitted by:	Jaakko Heinonen <jh saunalahti fi>
MFC after:	1 weeks
2008-06-11 12:46:09 +00:00
Konstantin Belousov
9e40a5f827 When devfs_allocv() committed to create new vnode, since de_vnode is NULL,
the dm_lock is held while the newly allocated vnode is locked. Since no
other threads may try to lock the new vnode yet, the LOR there cannot
result in the deadlock.

Shut down the witness warning to note this fact.

Tested by:	pho
Prodded by:	attilio
2008-06-05 09:15:47 +00:00
Ed Schouten
16151645c2 Revert the changes I made to devfs_setattr() in r179457.
As discussed with Robert Watson and John Baldwin, it would be better if
PTY's are created with proper permissions, turning grantpt() into a
no-op.

Bypassing security frameworks like MAC by passing NOCRED to
VOP_SETATTR() will only make things more complex.

Approved by:	philip (mentor)
2008-06-01 14:02:46 +00:00
Ed Schouten
34d1dcf0cc Merge back devfs changes from the mpsafetty branch.
In the mpsafetty branch, PTY's are allocated through the posix_openpt()
system call. The controller side of a PTY now uses its own file
descriptor type (just like sockets, vnodes, pipes, etc).

To remain compatible with existing FreeBSD and Linux C libraries, we can
still create PTY's by opening /dev/ptmx or /dev/ptyXX. These nodes
implement d_fdopen(). Devfs has been slightly changed here, to allow
finit() to be called from d_fdopen().

The routine grantpt() has also been moved into the kernel. This routine
is a little odd, because it needs to bypass standard UNIX permissions.
It needs to change the owner/group/mode of the slave device node, which
may often not be possible. The old implementation solved this by
spawning a setuid utility.

When VOP_SETATTR() is called with NOCRED, devfs_setattr() dereferences
ap->a_cred, causing a kernel panic. Change the de_{uid,gid,mode} code to
allow changes when a->a_cred is set to NOCRED.

Approved by:	philip (mentor)
2008-05-31 14:06:37 +00:00
Ulf Lilleengen
60af8a6a7a - Add locking to all filesystem operations in fdescfs and flag it as MPSAFE.
- Use proper synhronization primitives to protect the internal fdesc node cache
  used in fdescfs.
- Properly initialize and uninitalize hash.
- Remove unused functions.

Since fdescfs might recurse on itself, adding proper locking to it needed some
tricky workarounds in some parts to make it work. For instance, a descriptor in
fdescfs could refer to an open descriptor to itself, thus forcing the thread to
recurse on vnode locks. Because of this, other race conditions also had to be
fixed.

Tested by:	pho
Reviewed by:	kib (mentor)
Approved by:	kib (mentor)
2008-05-24 14:51:30 +00:00
Konstantin Belousov
772e245341 When vget() fails (because the vnode has been reclaimed), there is no
sense to loop trying to vget() the vnode again.

PR:	122977
Submitted by:	Arthur Hartwig <arthur.hartwig nokia com>
Tested by:	pho
Reviewed by:	jhb
MFC after:	1 week
2008-05-23 16:36:39 +00:00
Konstantin Belousov
82f4d64035 Implement the per-open file data for the cdev.
The patch does not change the cdevsw KBI. Management of the data is
provided by the functions
int	devfs_set_cdevpriv(void *priv, cdevpriv_dtr_t dtr);
int	devfs_get_cdevpriv(void **datap);
void	devfs_clear_cdevpriv(void);
All of the functions are supposed to be called from the cdevsw method
contexts.

- devfs_set_cdevpriv assigns the priv as private data for the file
  descriptor which is used to initiate currently performed driver
  operation. dtr is the function that will be called when either the
  last refernce to the file goes away, the device is destroyed  or
  devfs_clear_cdevpriv is called.
- devfs_get_cdevpriv is the obvious accessor.
- devfs_clear_cdevpriv allows to clear the private data for the still
  open file.

Implementation keeps the driver-supplied pointers in the struct
cdev_privdata, that is referenced both from the struct file and struct
cdev, and cannot outlive any of the referee.

Man pages will be provided after the KPI stabilizes.

Reviewed by:	jhb
Useful suggestions from:	jeff, antoine
Debugging help and tested by:	pho
MFC after:	1 month
2008-05-21 09:31:44 +00:00
Markus Brueffer
9c2bf69d32 Fix and speedup timestamp calculations which is roughly based on the patch in
the mentioned PR:

- bounds check time->month as it is used as an array index
- fix usage of time->month as array index (month is 1-12)
- fix calculation based on time->day (day is 1-31)
- fix the speedup code as it doesn't calculate correct timestamps before
  the year 2000 and reduce the number of calculation in the year-by-year code
- speedup month calculations by replacing the array content with cumulative
  values
- add microseconds calculation
- fix an endian problem

PR:		kern/97786
Submitted by:	Andriy Gapon <avg@topspin.kiev.ua>
Reviewed by:	scottl (earlier version)
Approved by:	emax (mentor)
MFC after:	1 week
2008-05-16 22:31:17 +00:00
Attilio Rao
58c5a5eb70 lockinit() can't accept LK_EXCLUSIVE as an initializaiton flag, so just
drop it.

Reported by:	Josh Carroll <josh dot carroll at gmail dot com>
Submitted by:	jhb
2008-05-15 21:39:25 +00:00
John Baldwin
06d0d0e274 Don't explicitly drop Giant around d_open/d_fdopen/d_close for MPSAFE
drivers.  Since devfs is already marked MPSAFE it shouldn't be held
anyway.

MFC after:	2 weeks
Discussed with:	phk
2008-05-07 19:03:57 +00:00
Daichi GOTO
3af387c9d2 - change function name from *_vdir to *_vnode because
VSOCK has been added as cache target. Now they process
  not only VDIR but also VSOCK.
- fixed panic issue caused by cache incorrect free process
  by "umount -f"

Submitted by:	Masanori OZAWA <ozawa@ongs.co.jp>
MFC after:	1 week
2008-05-07 05:32:55 +00:00
Daichi GOTO
fe5f08cda3 o Fixed multi thread access issue reported by Alexander V. Chernikov
(admin@su29.net)
  fixed: kern/109950

PR:		kern/109950
Submitted by:	Alexander V. Chernikov (admin@su29.net)
Reviewed by:	Masanori OZAWA (ozawa@ongs.co.jp)
MFC after:	1 week
2008-04-25 11:37:20 +00:00
Daichi GOTO
938161d61a o Improved unix socket connection issue
fixed: kern/118346

PR:		kern/118346
Submitted by:	Masanori OZAWA (ozawa@ongs.co.jp)
MFC after:	1 week
2008-04-25 09:53:52 +00:00
Daichi GOTO
5307411cbe o Fixed rename panic issue
Submitted by:	Masanori OZAWA (ozawa@ongs.co.jp)
MFC after:	1 week
2008-04-25 09:44:47 +00:00
Daichi GOTO
a9b794ff5e o Fixed inaccessible issue especially including devfs on unionfs case.
fixed also: kern/117829

PR:		kern/117829
Submitted by:	Masanori OZAWA (ozawa@ongs.co.jp)
MFC after:	1 week
2008-04-25 09:38:48 +00:00
Daichi GOTO
a68ae31c71 o Added system hang-up process when VOP_READDIR of unionfs_nodeget()
returns not end of the file status on debug mode (DIAGNOSTIC defined)
  kernel.

Submitted by:	Masanori OZAWA (ozawa@ongs.co.jp)
MFC after:	1 week
2008-04-25 07:58:19 +00:00
Konstantin Belousov
eab626f110 Move the head of byte-level advisory lock list from the
filesystem-specific vnode data to the struct vnode. Provide the
default implementation for the vop_advlock and vop_advlockasync.
Purge the locks on the vnode reclaim by using the lf_purgelocks().
The default implementation is augmented for the nfs and smbfs.
In the nfs_advlock, push the Giant inside the nfs_dolock.

Before the change, the vop_advlock and vop_advlockasync have taken the
unlocked vnode and dereferenced the fs-private inode data, racing with
with the vnode reclamation due to forced unmount. Now, the vop_getattr
under the shared vnode lock is used to obtain the inode size, and
later, in the lf_advlockasync, after locking the vnode interlock, the
VI_DOOMED flag is checked to prevent an operation on the doomed vnode.

The implementation of the lf_purgelocks() is submitted by dfr.

Reported by:	kris
Tested by:	kris, pho
Discussed with:	jeff, dfr
MFC after:	2 weeks
2008-04-16 11:33:32 +00:00
Doug Rabson
18121c17f5 When calling lf_advlock to unlock a record, make sure that ap->a_fl->l_type
is F_UNLCK otherwise we trigger a LOCKF_DEBUG panic.

MFC after: 3 days
2008-04-14 09:22:48 +00:00
Attilio Rao
047dd67e96 Optimize lockmgr in order to get rid of the pool mutex interlock, of the
state transitioning flags and of msleep(9) callings.
Use, instead, an algorithm very similar to what sx(9) and rwlock(9)
alredy do and direct accesses to the sleepqueue(9) primitive.

In order to avoid writer starvation a mechanism very similar to what
rwlock(9) uses now is implemented, with the correspective per-thread
shared lockmgrs counter.

This patch also adds 2 new functions to lockmgr KPI: lockmgr_rw() and
lockmgr_args_rw().  These two are like the 2 "normal" versions, but they
both accept a rwlock as interlock.  In order to realize this, the general
lockmgr manager function "__lockmgr_args()" has been implemented through
the generic lock layer. It supports all the blocking primitives, but
currently only these 2 mappers live.

The patch drops the support for WITNESS atm, but it will be probabilly
added soon. Also, there is a little race in the draining code which is
also present in the current CVS stock implementation: if some sharers,
once they wakeup, are in the runqueue they can contend the lock with
the exclusive drainer.  This is hard to be fixed but the now committed
code mitigate this issue a lot better than the (past) CVS version.
In addition assertive KA_HELD and KA_UNHELD have been made mute
assertions because they are dangerous and they will be nomore supported
soon.

In order to avoid namespace pollution, stack.h is splitted into two
parts: one which includes only the "struct stack" definition (_stack.h)
and one defining the KPI.  In this way, newly added _lockmgr.h can
just include _stack.h.

Kernel ABI results heavilly changed by this commit (the now committed
version of "struct lock" is a lot smaller than the previous one) and
KPI results broken by lockmgr_rw() / lockmgr_args_rw() introduction,
so manpages and __FreeBSD_version will be updated accordingly.

Tested by:      kris, pho, jeff, danger
Reviewed by:    jeff
Sponsored by:   Google, Summer of Code program 2007
2008-04-06 20:08:51 +00:00
Konstantin Belousov
8eb6b6ecb6 The temporary workaround for the call to the vget() without lock type in
the fdesc_allocvp(). The caller of the fdesc_allocvp() expects that the
returned vnode is not reclaimed. Do lock the vnode exclusive and drop
the lock after.

Reported by:	pho
Reviewed by:	jeff
2008-04-04 09:37:57 +00:00
Konstantin Belousov
57b4252e45 Add the support for the AT_FDCWD and fd-relative name lookups to the
namei(9).

Based on the submission by rdivacky,
	sponsored by Google Summer of Code 2007
Reviewed by:	rwatson, rdivacky
Tested by:	pho
2008-03-31 12:01:21 +00:00
Jeff Roberson
4c65d593e2 - Simplify null_hashget() and null_hashins() by using vref() rather
than a complex series of steps involving vget() without a lock type
   to emulate the same thing.
2008-03-29 23:24:54 +00:00
Doug Rabson
dfdcada31e Add the new kernel-mode NFS Lock Manager. To use it instead of the
user-mode lock manager, build a kernel with the NFSLOCKD option and
add '-k' to 'rpc_lockd_flags' in rc.conf.

Highlights include:

* Thread-safe kernel RPC client - many threads can use the same RPC
  client handle safely with replies being de-multiplexed at the socket
  upcall (typically driven directly by the NIC interrupt) and handed
  off to whichever thread matches the reply. For UDP sockets, many RPC
  clients can share the same socket. This allows the use of a single
  privileged UDP port number to talk to an arbitrary number of remote
  hosts.

* Single-threaded kernel RPC server. Adding support for multi-threaded
  server would be relatively straightforward and would follow
  approximately the Solaris KPI. A single thread should be sufficient
  for the NLM since it should rarely block in normal operation.

* Kernel mode NLM server supporting cancel requests and granted
  callbacks. I've tested the NLM server reasonably extensively - it
  passes both my own tests and the NFS Connectathon locking tests
  running on Solaris, Mac OS X and Ubuntu Linux.

* Userland NLM client supported. While the NLM server doesn't have
  support for the local NFS client's locking needs, it does have to
  field async replies and granted callbacks from remote NLMs that the
  local client has contacted. We relay these replies to the userland
  rpc.lockd over a local domain RPC socket.

* Robust deadlock detection for the local lock manager. In particular
  it will detect deadlocks caused by a lock request that covers more
  than one blocking request. As required by the NLM protocol, all
  deadlock detection happens synchronously - a user is guaranteed that
  if a lock request isn't rejected immediately, the lock will
  eventually be granted. The old system allowed for a 'deferred
  deadlock' condition where a blocked lock request could wake up and
  find that some other deadlock-causing lock owner had beaten them to
  the lock.

* Since both local and remote locks are managed by the same kernel
  locking code, local and remote processes can safely use file locks
  for mutual exclusion. Local processes have no fairness advantage
  compared to remote processes when contending to lock a region that
  has just been unlocked - the local lock manager enforces a strict
  first-come first-served model for both local and remote lockers.

Sponsored by:	Isilon Systems
PR:		95247 107555 115524 116679
MFC after:	2 weeks
2008-03-26 15:23:12 +00:00
Jeff Roberson
698b1a6643 - Complete part of the unfinished bufobj work by consistently using
BO_LOCK/UNLOCK/MTX when manipulating the bufobj.
 - Create a new lock in the bufobj to lock bufobj fields independently.
   This leaves the vnode interlock as an 'identity' lock while the bufobj
   is an io lock.  The bufobj lock is ordered before the vnode interlock
   and also before the mnt ilock.
 - Exploit this new lock order to simplify softdep_check_suspend().
 - A few sync related functions are marked with a new XXX to note that
   we may not properly interlock against a non-zero bv_cnt when
   attempting to sync all vnodes on a mountlist.  I do not believe this
   race is important.  If I'm wrong this will make these locations easier
   to find.

Reviewed by:	kib (earlier diff)
Tested by:	kris, pho (earlier diff)
2008-03-22 09:15:16 +00:00
Konstantin Belousov
91a35e7870 Do not dereference cdev->si_cdevsw, use the dev_refthread() to properly
obtain the reference. In particular, this fixes the panic reported in
the PR. Remove the comments stating that this needs to be done.

PR:	kern/119422
MFC after:	1 week
2008-03-20 16:08:42 +00:00
Jeff Roberson
6617724c5f Remove kernel support for M:N threading.
While the KSE project was quite successful in bringing threading to
FreeBSD, the M:N approach taken by the kse library was never developed
to its full potential.  Backwards compatibility will be provided via
libmap.conf for dynamically linked binaries and static binaries will
be broken.
2008-03-12 10:12:01 +00:00
Robert Watson
970a2d8770 Replace lockmgr lock protecting nwfs vnode hash table with an sx lock.
MFC after:	1 month
2008-03-02 19:02:30 +00:00
Robert Watson
7947229ff6 Replace lockmgr lock protecting smbfs node hash table with sx lock.
MFC after:	1 month
2008-03-02 18:56:13 +00:00
Attilio Rao
7fbfba7bf8 - Handle buffer lock waiters count directly in the buffer cache instead
than rely on the lockmgr support [1]:
  * bump the waiters only if the interlock is held
  * let brelvp() return the waiters count
  * rely on brelvp() instead than BUF_LOCKWAITERS() in order to check
    for the waiters number
- Remove a namespace pollution introduced recently with lockmgr.h
  including lock.h by including lock.h directly in the consumers and
  making it mandatory for using lockmgr.
- Modify flags accepted by lockinit():
  * introduce LK_NOPROFILE which disables lock profiling for the
    specified lockmgr
  * introduce LK_QUIET which disables ktr tracing for the specified
    lockmgr [2]
  * disallow LK_SLEEPFAIL and LK_NOWAIT to be passed there so that it
    can only be used on a per-instance basis
- Remove BUF_LOCKWAITERS() and lockwaiters() as they are no longer
  used

This patch breaks KPI so __FreBSD_version will be bumped and manpages
updated by further commits. Additively, 'struct buf' changes results in
a disturbed ABI also.

[2] Really, currently there is no ktr tracing in the lockmgr, but it
will be added soon.

[1] Submitted by:	kib
Tested by:	pho, Andrea Barberio <insomniac at slackware dot it>
2008-03-01 19:47:50 +00:00
Konstantin Belousov
e6591b84ff Rename fdescfs vnode from "fdesc" to "fdescfs" to avoid name collision
of the vnode lock with the fdesc_mtx mutex. Having different kinds of
locks with the same name confuses witness.
2008-02-26 10:10:55 +00:00
Robert Watson
18ff731caa Add "Make MPSAFE" to the Coda todo list.
MFC after:	3 days
2008-02-26 09:27:47 +00:00
Attilio Rao
81c794f998 Axe the 'thread' argument from VOP_ISLOCKED() and lockstatus() as it is
always curthread.

As KPI gets broken by this patch, manpages and __FreeBSD_version will be
updated by further commits.

Tested by:	Andrea Barberio <insomniac at slackware dot it>
2008-02-25 18:45:57 +00:00
Attilio Rao
628f51d275 Introduce some functions in the vnode locks namespace and in the ffs
namespace in order to handle lockmgr fields in a controlled way instead
than spreading all around bogus stubs:
- VN_LOCK_AREC() allows lock recursion for a specified vnode
- VN_LOCK_ASHARE() allows lock sharing for a specified vnode

In FFS land:
- BUF_AREC() allows lock recursion for a specified buffer lock
- BUF_NOREC() disallows recursion for a specified buffer lock

Side note: union_subr.c::unionfs_node_update() is the only other function
directly handling lockmgr fields. As this is not simple to fix, it has
been left behind as "sole" exception.
2008-02-24 16:38:58 +00:00
Marcel Moolenaar
043ec583dc Don't check the bpbSecPerTrack and bpbHeads fields of the BPB.
They are typically 0 on new ia64 systems. Since we don't use
either field, there's no harm in not checking.
2008-02-21 03:19:46 +00:00
Robert Watson
fa8003c6b9 Remove custom queue macros in Coda, replacing them with queue(9) tailq
macros.  The only semantic change was the need to add a vc_opened field
to struct vcomm since we can no longer use the request queue returning
to an uninitialized state to hold whether or not the device is open.

MFC after:	1 month
2008-02-17 14:33:28 +00:00
Robert Watson
b15ce9be2e Remove namecache performance-tuning todo for Coda: we now use the FreeBSD
name cache.

MFC after:	1 month
2008-02-17 12:40:27 +00:00
Robert Watson
a8c34e8ee0 The possibly interruptible msleep in coda_call() means well, but is
fundamentally fairly confused about how signals work and when it is
appropriate for upcalls to be interrupted.  In particular, we should
be exempting certain upcalls from interruption, we should not always
eventually time out sleeping on a upcall, and we should not be
interrupting the sleep for certain signals that we currently are
(including SIGINFO).  This code needs to be reworked in the style of
NFS interruptible mounts.

MFC after:	1 month
2008-02-15 13:31:35 +00:00
Robert Watson
c30ddc8d99 Spell replys as replies.
MFC after:	1 month
2008-02-15 12:11:45 +00:00
Robert Watson
93b510870f Reorder and clean up make_coda_node(), annotate weaknesses in the
implementation.

MFC after:	1 month
2008-02-15 11:58:11 +00:00
Robert Watson
c0964f549b Remove debugging code under OLD_DIAGNOSTIC; this is all >10 years old and
hasn't been used in that time.

MFC after:	1 month
2008-02-14 00:55:03 +00:00
Robert Watson
57a77b811f In Coda, flush the attribute cache for a cnode when its fid is
changed, as its synthesized inode number may have changed and we
want stat(2) to pick up the new inode number.

MFC after:	1 month
2008-02-14 00:30:06 +00:00
Robert Watson
89d1d7886a Update cache flushing behavior in light of recent namecache and
access cache improvements:

- Flush just access control state on CODA_PURGEUSER, not the full
  namecache for /coda.

- When replacing a fid on a cnode as a result of, e.g.,
  reintegration after offline operation, we no longer need to
  purge the namecache entries associated with its vnode.

MFC after:	1 month
2008-02-13 19:50:17 +00:00
Robert Watson
38ab9a906a Implement a rudimentary access cache for the Coda kernel module,
modeled on the access cache found in NFS, smbfs, and the Linux coda
module.  This is a positive access cache of a single entry per file,
tracking recently granted rights, but unlike NFS and smbfs,
supporting explicit invalidation by the distributed file system.

For each cnode, maintain a C_ACCCACHE flag indicating the validity
of the cache, and a cached uid and mode tracking recently granted
positive access control decisions.

Prefer the cache to venus_access() in VOP_ACCESS() if it is valid,
and when we must fall back to venus_access(), update the cache.

Allow Venus to clear the access cache, either the whole cache on
CODA_FLUSH, or just entries for a specific uid on CODA_PURGEUSER.
Unlike the Coda module on Linux, we don't flush all entries on a
user purge using a generation number, we instead walk present
cnodes and clear only entries for the specific user, meaning it is
somewhat more expensive but won't hit all users.

Since the Coda module is agressive about not keeping around
unopened cnodes, the utility of the cache is somewhat limited for
files, but works will for directories.  We should make Coda less
agressive about GCing cnodes in VOP_INACTIVE() in order to improve
the effectiveness of in-kernel caching of attributes and access
rights.

MFC after:	1 month
2008-02-13 15:45:12 +00:00
Robert Watson
d25a3c4c44 Remove now-unused Coda namecache.
MFC after:	1 month
2008-02-13 13:26:01 +00:00
Robert Watson
44abffb44b Rather than having the Coda module use its own namecache, use the global
VFS namecache, as is done by the Coda module on Linux.  Unlike the Coda
namecache, the global VFS namecache isn't tagged by credential, so use
ore conservative flushing behavior (for now) when CODA_PURGEUSER is
issued by Venus.

This improves overall integration with the FreeBSD VFS, including
allowing __getcwd() to work better, procfs/procstat monitoring, and so
on.  This improves shell behavior in many cases, and improves ".."
handling.  It may lead to some slowdown until we've implemented a
specific access cache, which should net improve performance, but in the
mean time, lookup access control now always goes to Venus, whereas
previously it didn't.

MFC after:	1 month
2008-02-13 13:06:22 +00:00
Attilio Rao
d1215e10d2 Fix a lock leak in the ntfs locking scheme:
When ntfs_ntput() reaches 0 in the refcount the inode lockmgr is not
released and directly destroyed. Fix this by unlocking the lockmgr() even
in the case of zero-refcount.

Reported by: dougb, yar, Scot Hetzel <swhetzel at gmail dot com>
Submitted by: yar
2008-02-13 13:02:12 +00:00
Robert Watson
4f52b754df Clean up coda_pathconf() slightly while debugging a problem there.
MFC after:	1 month
2008-02-11 00:01:45 +00:00
Robert Watson
21bb029533 Since we're now actively maintaining the Coda module in the FreeBSD source
tree, restyle everything but coda.h (which is more explicitly shared
across systems) into a closer approximation to style(9).

Remove a few more unused function prototypes.

Add or clarify some comments.

MFC after:	1 month
2008-02-10 11:18:12 +00:00
Robert Watson
d57786ec68 Various further non-functional cleanups to coda:
- Rename print_vattr to coda_print_vattr and make static, rename
  print_cred to coda_print_cred.
- Remove unused coda_vop_nop.
- Add XXX comment because coda_readdir forwards to the cache vnode's
  readdir rather than venus_readdir, and annotate venus_readdir as
  unused.
- Rename vc_nb_* to vc_*.
- Use d_open_t, d_close_t, d_read_t, d_write_t, d_ioctl_t and d_poll_t
  for prototyping vc_* as that is the intent, don't use our own
  definitions.
- Rename coda_nb_statfs to coda_statfs, rename NB_SFS_SIZ to
  CODA_SFS_SIZ.
- Replace one more OBE reference to NetBSD with a reference to FreeBSD.
- Tidy up a little vertical whitespace here and there.
- Annotate coda_nc_zapvnode as unused.
- Remove unused vcodattach.
- Annotate VM_INTR as unused.
- Annotate that coda_fhtovp is unused and doesn't match the FreeBSD
  prototype, so isn't hooked up to vfs_fhtovp.  If we want NFS export of
  Coda to work someday, this needs to be fixed.
- Remove unused getNewVnode.
- Remove unused coda_vget, coda_init, coda_quotactl prototypes.

MFC after:	1 month
2008-02-09 12:49:18 +00:00
Robert Watson
fc9d8f0057 No reason not to maintain stats on statfs in Coda, as it's done for
other VFS operations, so uncomment the existing statistics gathering.

MFC after:	1 month
2008-02-09 11:40:49 +00:00
Robert Watson
8571e9a189 Remove unused devtomp(), which exploited UFS-specific knowledge to find
the mountpoint for a specific device.  This was implemented incorrectly,
a bad idea in a fundamental sense, and also never used, so presumably
a long-idle debugging function.

MFC after:	1 month
2008-02-09 11:12:18 +00:00
Robert Watson
82e4904ffb Since Coda is effectively a stacked file system, use VOP_EOPNOTSUPP
for vop_bmap; delete the existing stub that returned either EINVAL
or EOPNOTSUPP, and had unreachable calls to VOP_BMAP on the cache
vnode.

MFC after:	1 month
2008-02-09 09:33:19 +00:00
Robert Watson
37245e3742 Lock cache vnode when VOP_FSYNC() is called on a Coda vnode.
MFC after:	1 month
2008-02-09 00:12:22 +00:00
Robert Watson
6dc70a9dec Make all calls to vn_lock() in Coda, including recently added ones,
use LK_RETRY, since failure is undesirable (and not handled).

MFC after:	1 month
Pointed out by:	kib
2008-02-09 00:03:22 +00:00
Robert Watson
7a246a6314 The Coda module was originally ported to NetBSD from Mach by rvb, and
then later to FreeBSD.  Update various NetBSD-related comments: in some
cases delete them because they don't appply, in others update to say
FreeBSD as they still apply but in FreeBSD (and might for that matter
no longer apply on NetBSD), and flag one case where I'm not sure
whether it applies.

MFC after:	1 month
2008-02-08 23:15:36 +00:00
Robert Watson
efeac2fb25 Before invoking vnode operations on cache vnodes, acquire the vnode
locks of those vnodes.  Probably, Coda should do the same lock sharing/
pass-through that is done for nullfs, but in the mean time this ensures
that locks are adequately held to prevent corruption of data structures
in the cache file system.

Assuming most operations came from the top layer of Coda and weren't
performed directly on the cache vnodes, in practice this corruption was
relatively unlikely as the Coda vnode locks were ensuring exclusive
access for most consumers.

This causes WITNESS to squeal like a pig immediately when Coda is used,
rather than waiting until file close; I noticed these problems because
of the lack of said squealing.

MFC after:	1 month
2008-02-08 23:01:40 +00:00
Robert Watson
99a2317ed3 Remove undefined coda excluded by #if 1 #else, which previously protected
vget() calls using inode numbers to query the root of /coda, which is not
needed since we now cache the root vnode with the mountpoint.

MFC after:	1 month
2008-02-08 22:37:15 +00:00
Attilio Rao
2433c4883e Conver all explicit instances to VOP_ISLOCKED(arg, NULL) into
VOP_ISLOCKED(arg, curthread). Now, VOP_ISLOCKED() and lockstatus() should
only acquire curthread as argument; this will lead in axing the additional
argument from both functions, making the code cleaner.

Reviewed by: jeff, kib
2008-02-08 21:45:47 +00:00
Robert Watson
c55376e791 Remove Giant acquisition around soreceive() and sosend() in fifofs. The
bug that caused us to reintroduce it is believed to be fixed, and Kris
says he no longer sees problems with fifofs in highly parallel builds.
If this works out, we'll MFC it for 7.1.

MFC after:	3 months
Pointed out by:	kris
2008-01-26 12:34:23 +00:00
Attilio Rao
0e9eb108f0 Cleanup lockmgr interface and exported KPI:
- Remove the "thread" argument from the lockmgr() function as it is
  always curthread now
- Axe lockcount() function as it is no longer used
- Axe LOCKMGR_ASSERT() as it is bogus really and no currently used.
  Hopefully this will be soonly replaced by something suitable for it.
- Remove the prototype for dumplockinfo() as the function is no longer
  present

Addictionally:
- Introduce a KASSERT() in lockstatus() in order to let it accept only
  curthread or NULL as they should only be passed
- Do a little bit of style(9) cleanup on lockmgr.h

KPI results heavilly broken by this change, so manpages and
FreeBSD_version will be modified accordingly by further commits.

Tested by: matteo
2008-01-24 12:34:30 +00:00
Robert Watson
9d3e5c0e2b Put "coda_rdwr: Internally Opening" printf generated by in-kernel writes
to files, such as ktrace output, under CODA_VERBOSE.  Otherwise, each
such call to VOP_WRITE() results in a kernel printf.

MFC after:	3 days
Obtained from:	NetBSD
2008-01-21 21:39:08 +00:00
Robert Watson
e866951b59 Replace references to VOP_LOCK() w/o LK_RETRY to vn_lock() with LK_RETRY,
avoiding extra error handling, or in some cases, missing error handling.

MFC after:	3 days
Discussed with:	kib
2008-01-21 21:19:07 +00:00
Robert Watson
9440b9f7ea Remove unused oldhash definition from Coda namecache.
MFC after:	3 days
2008-01-19 19:21:07 +00:00
Robert Watson
de5910460a Improve default vnode operation handling for Coda:
- Don't specify vnode operations for mknod, lease, and advlock--let them
  fall through to vop_default.

- Implement vop_default with &default_vnodeops, rather than with VOP_PANIC,
  so that unimplemented vnode operations are handled in more sensible ways
  than panicking, such as EOPNOTSUPP on ACL queries generated by bsdtar,
  or mknod.

MFC after:	3 days
2008-01-19 17:12:44 +00:00
Robert Watson
aeab4f72a0 Rework coda_statfs(): no longer need to zero the statfs structure or
fill out all fields, just fill out the ones the file system knows
about.  Among other things, this causes the outpuf of "mount" and
"df" to make quite a bit more sense as /dev/cfs0 is specified as the
mountfrom name.

MFC after:	3 days
2008-01-19 16:39:14 +00:00
Robert Watson
82bf4517ef Zero mi_rotovp and coda_ctlvp immediately after calling vrele() on the
vnodes during coda_unmount() in order to detect errant use of them
after the vnode references may no longer be valid.

No need to clear the VV_ROOT flag on mi_rootvp flag (especially after
the vnode reference is no longer valid) as this isn't done on other
file systems.

MFC after:	3 days
2008-01-19 15:40:46 +00:00
Robert Watson
96b1e9b015 Don't acquire an additional vnode reference to a vnode when it is opened
and then release it when it is closed: we rely on the caller to keep the
vnode around with a valid reference.  This avoids vrele() destroying the
vnode vop_close() is being called from during a call to vop_close(), and
a crash due to lockmgr recursing the vnode lock when a Coda unmount
occurs.

MFC after:	3 days
2008-01-19 15:39:10 +00:00