Commit Graph

865 Commits

Author SHA1 Message Date
Robert Watson
15b27e726e o Minor style fix on #endif, missing '_' in comment. 2002-02-20 15:44:43 +00:00
Poul-Henning Kamp
68edc1b939 Make v_addpollinfo() visible and non-inline.
Have callers only call it as needed.
Add necessary call in ufs_kqfilter().

Test-case found by:	Andrew Gallatin <gallatin@cs.duke.edu>
2002-02-18 16:18:02 +00:00
Poul-Henning Kamp
4b55dbe36b Move the stuff related to select and poll out of struct vnode.
The use of the zone allocator may or may not be overkill.
There is an XXX: over in ufs/ufs/ufs_vnops.c that jlemon may need
to revisit.

This shaves about 60 bytes of struct vnode which on my laptop means
600k less RAM used for vnodes.
2002-02-17 21:15:36 +00:00
Poul-Henning Kamp
e8b26e995e Collect the VN_KNOTE() macro definitions on vnode.h 2002-02-17 21:07:57 +00:00
Julian Elischer
2c1007663f In a threaded world, differnt priorirites become properties of
different entities.  Make it so.

Reviewed by:	jhb@freebsd.org (john baldwin)
2002-02-11 20:37:54 +00:00
Robert Watson
cfcd3c783e Minor style tweaks.
Remove an unneeded comment and commented out code that won't be
needed.
2002-02-10 04:57:08 +00:00
Robert Watson
41d5a43fa1 Copyright + license update. 2002-02-10 04:50:24 +00:00
Robert Watson
74237f55b0 Part I: Update extended attribute API and ABI:
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
2002-02-10 04:43:22 +00:00
Poul-Henning Kamp
b6e1c37356 Remove di_inumber since LFS is long gone. 2002-02-10 00:55:49 +00:00
Kirk McKusick
b06051cf7c Occationally background fsck would cause a spurious ``freeing free
inode'' panic. This change corrects that problem by setting the
fs_active flag when the inode map changes to notify the snapshot
code that the cylinder group must be rescanned.

Submitted by:	Robert Watson <rwatson@FreeBSD.org>
2002-02-07 22:13:56 +00:00
Kirk McKusick
cfdaa88697 Occationally deleted files would hang around for hours or days
without being reclaimed. This bug was introduced in revision 1.95
dealing with filenames placed in newly allocated directory blocks,
thus is not present in 4.X systems. The bug is triggered when a
new entry is made in a directory after the data block containing
the original new entry has been written, but before the inode
that references the data block has been written.

Submitted by:	Bill Fenner <fenner@research.att.com>
2002-02-07 00:54:32 +00:00
Kirk McKusick
c9f96392c7 When taking a snapshot, we must check for active files that have
been unlinked (e.g., with a zero link count). We have to expunge
all trace of these files from the snapshot so that they are neither
reclaimed prematurely by fsck nor saved unnecessarily by dump.
2002-02-02 01:42:44 +00:00
Kirk McKusick
7b60855308 Add a stub for softdep_request_cleanup() so that compilation without
SOFTUPDATES option works properly.

Submitted by:	Benno Rice <benno@jeamland.net>
2002-01-23 02:18:56 +00:00
Kirk McKusick
03a2057a5b This patch fixes a long standing complaint with soft updates in
which small and/or nearly full filesystems would fail with `file
system full' messages when trying to replace a number of existing
files (for example during a system installation). When the allocation
routines are about to fail with a file system full condition, they
make a call to softdep_request_cleanup() which attempts to accelerate
the flushing of pending deletion requests in an effort to free up
space. In the face of filesystem I/O requests that exceed the
available disk transfer capacity, the cleanup request could take
an unbounded amount of time. Thus, the softdep_request_cleanup()
routine will only try for tickdelay seconds (default 2 seconds)
before giving up and returning a filesystem full error. Under typical
conditions, the softdep_request_cleanup() routine is able to free
up space in under fifty milliseconds.
2002-01-22 06:17:22 +00:00
Kirk McKusick
99bef8782b Fix a bug introduced in ffs_snapshot.c -r1.25 and fs.h -r1.26
which caused incomplete snapshots to be taken. When background
fsck would run on these snapshots, the result would be files
being incorrectly released which would subsequently panic the
kernel with ``handle_workitem_freefile: inodedep survived'',
``handle_written_inodeblock: live inodedep'', and
``handle_workitem_remove: lost inodedep'' errors.
2002-01-17 08:33:32 +00:00
Kirk McKusick
8af31e7b46 Put write on read-only filesystem panic after we have weeded out
block and character devices, fifo's, etc.

Submitted by:	Bruce Evans <bde@zeta.org.au>
2002-01-16 04:59:09 +00:00
Kirk McKusick
cd6005961f When downgrading a filesystem from read-write to read-only, operations
involving file removal or file update were not always being fully
committed to disk. The result was lost files or corrupted file data.
This change ensures that the filesystem is properly synced to disk
before the filesystem is down-graded.

This delta also fixes a long standing bug in which a file open for
reading has been unlinked. When the last open reference to the file
is closed, the inode is reclaimed by the filesystem. Previously,
if the filesystem had been down-graded to read-only, the inode could
not be reclaimed, and thus was lost and had to be later recovered
by fsck.  With this change, such files are found at the time of the
down-grade.  Normally they will result in the filesystem down-grade
failing with `device busy'. If a forcible down-grade is done, then
the affected files will be revoked causing the inode to be released
and the open file descriptors to begin failing on attempts to read.

Submitted by:	"Sam Leffler" <sam@errno.com>
2002-01-15 07:17:12 +00:00
Alfred Perlstein
426da3bcfb SMP Lock struct file, filedesc and the global file list.
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.
2002-01-13 11:58:06 +00:00
Kirk McKusick
0bc7a833ec When going to sleep, we must save our SPL so that it does not get
lost if some other process uses the lock while we are sleeping. We
restore it after we have slept. This functionality is provided by
a new routine interlocked_sleep() that wraps the interlocking with
functions that sleep. This function is then used in place of the
old ACQUIRE_LOCK_INTERLOCKED() and FREE_LOCK_INTERLOCKED() macros.

Submitted by:	Debbie Chu <dchu@juniper.net>
2002-01-12 20:57:36 +00:00
Kirk McKusick
794ef3471f Must call drain_output() before checking the dirty block list
in softdep_sync_metadata(). Otherwise we may miss dependencies
that need to be flushed which will result in a later panic
with the message ``vinvalbuf: dirty bufs''.

Submitted by:	Matthew Dillon <dillon@apollo.backplane.com>
MFC after:	1 week
2002-01-11 19:59:27 +00:00
Poul-Henning Kamp
9c643340bb Do not pull quota entries of the cache-list if they have already
been removed from the cache-list as part of a previous unmount.

This would result in panics (page fault in dqflush()) during subsequent
umounts provided that enough distinct UID's to actually make the
hash do something are active.

This can probably explain a number of weird quota related behaviours.

PR:		32331 maybe more.
Reproduced by:	Søren Schrørder <sch@cybercity.dk>
2002-01-10 15:02:57 +00:00
Mike Smith
b9a4338d29 Initialise the bioops vector hack at runtime rather than at link time. This
avoids the use of common variables.

Reviewed by:	mckusick
2002-01-08 19:32:18 +00:00
Matthew Dillon
23b590188f Fix a BUF_TIMELOCK race against BUF_LOCK and fix a deadlock in vget()
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
2001-12-20 22:42:27 +00:00
Kirk McKusick
f305c5d199 Change the atomic_set_char to atomic_set_int and atomic_clear_char
to atomic_clear_int to ease the implementation for the sparc64.

Requested by:	Jake Burkholder <jake@locore.ca>
2001-12-18 18:05:17 +00:00
Ian Dowse
143a5346c9 Make sure we ignore the value of `fs_active' when reloading the
superblock, and move the initialisation of it to beside where other
pointer fields are initialised.
2001-12-16 18:54:09 +00:00
Ian Dowse
3fa4044e34 Move the new superblock field `fs_active' into the region of the
superblock that is already set up to handle pointer types. This
fixes an accidental change in the superblock size on 64-bit platforms
caused by revision 1.24.
2001-12-16 18:51:11 +00:00
Kirk McKusick
cc5a92334f Minimize the time necessary to suspend operations on a filesystem
when taking a snapshot. The two time consuming operations are
scanning all the filesystem bitmaps to determine which blocks
are in use and scanning all the other snapshots so as to be able
to expunge their blocks from the view of the current snapshot.
The bitmap scanning is broken into two passes. Before suspending
the filesystem all bitmaps are scanned. After the suspension,
those bitmaps that changed after being scanned the first time
are rescanned. Typically there are few bitmaps that need to be
rescanned. The expunging of other snapshots is now done after
the suspension is released by observing that we can easily
identify any blocks that were allocated to them after the
suspension (they will be maked as `not needing to be copied'
in the just created snapshot). For all the gory details, see
the ``Running fsck in the Background'' paper in the Usenix
BSDCon 2002 Conference Proceedings, pages 55-64.
2001-12-14 00:15:06 +00:00
Kirk McKusick
9db12e5108 When a file is partially truncated, we first check to see if the
new file end will land in the middle of a file hole. Since the last
block of a file must always be allocated, the hole is filled by
allocating a block at that location. If the hole being filled is
a direct block, then the truncation may eventually reduce the
full sized block down to a fragment. When running with soft
updates, it is necessary to FSYNC the file after allocating the
block and before creating the fragment to avoid triggering a
soft updates inconsistency when the block unexpectedly shrinks.

Found by:	Matthew Dillon <dillon@apollo.backplane.com>
MFC after:	1 week
2001-12-13 05:07:48 +00:00
Robert Watson
24373ce6ed Use 'mkdir -p /.attribute/system' instead of breaking it into
two seperate mkdir targets.

Submitted by:	jedgar
2001-11-30 15:32:07 +00:00
Robert Watson
cff9580525 Use 'mkdir -p /.attribute/system' instead of breaking it into
two seperate mkdir targets.
2001-11-30 15:21:20 +00:00
Robert Watson
15f1c8d3d2 README.extattr incorrectly specified sample command lines for
UFS_EXTATTR_AUTOSTART.  Insert the missing 'initattr' arguments
to extattrctl.

Noticed by:	green
2001-11-30 15:15:27 +00:00
Guido van Rooij
40e294f796 When mkdir()-ing, the parent dir gets is linkcount increased.
Fix VN_KNOTE to reflect that.

Found by: tobez@freebsd.org
MFC after:	2 days
2001-11-22 15:33:12 +00:00
Ian Dowse
4202b366fc Oops, when trying the dirhash sequential-access optimisation,
compare the slot offset against the predicted offset, not a boolean
flag. This typo effectively disabled the sequential optimisation,
but was otherwise harmless.

Not surprisingly, fixing this improves performance in the sequential
access case. I am seeing a 7% speedup on one machine here; using
dirhash when sequentially looking up directory entries is now about
5% faster instead of 2% slower than the non-dirhash case.

Submitted by:	KOIE Hidetaka <koie@suri.co.jp>
MFC after:	1 week
2001-11-14 15:08:07 +00:00
Matthew Dillon
7e76bb562e Implement IO_NOWDRAIN and B_NOWDRAIN - prevents the buffer cache from blocking
in wdrain during a write.  This flag needs to be used in devices whos
strategy routines turn-around and issue another high level I/O, such as
when MD turns around and issues a VOP_WRITE to vnode backing store, in order
to avoid deadlocking the dirty buffer draining code.

Remove a vprintf() warning from MD when the backing vnode is found to be
in-use.  The syncer of buf_daemon could be flushing the backing vnode at
the time of an MD operation so the warning is not correct.

MFC after:	1 week
2001-11-05 18:48:54 +00:00
Robert Watson
6d8785434f o Update copyright dates.
o Add reference to TrustedBSD Project in license header.
o Update dated comments, including comment in extattr.h claiming that
  no file systems support extended attributes.
o Improve comment consistency.
2001-11-01 21:37:07 +00:00
Robert Watson
b6e0472987 o Althought this is not specified in POSIX.1e, the UFS ACL implementation
coerces the deletion of a default ACL on a directory when no default
  ACL EA is present to success.  Because the UFS EA implementation doesn't
  disinguish the EA failure modes "that EA name has not been
  administratively enabled" from "that EA name has no defined data",
  there's a potential conflict in error return values.  Normally, the
  lack of administratively configured EA support is coerced to
  EOPNOTSUPP to indicate that ACLs are not available; in this case,
  it is possible to get a successful return, even if ACLs are not
  available because EA support for them has not been enabled.

  Expand the comment in ufs_setacl() to identify this case.

Obtained from:	TrustedBSD Project
2001-10-27 05:39:17 +00:00
Robert Watson
ac8b3dd7dc o Clarify a comment about the locking condition of the vnode upon exit
from ufs_extattr_enable_with_open().
o Print auto-start notifications if (bootverbose).  This was previously
  commented out since it didn't know how to check for bootverbose.
o Drop in comments throughout indicating where ENOENT should be replaced
  with ENOATTR once that is available.

Obtained from:	TrustedBSD Project
2001-10-27 05:19:14 +00:00
Robert Watson
29543004bd o The comment about ordering the destruction of the lock and the removal of
the flag indicating that the structure was initialized didn't need
  an XXX, since it didn't need fixing.

Obtained from:	TrustedBSD Project
2001-10-27 05:05:39 +00:00
Robert Watson
9444746795 o Wrap a number of long lines of code, many of which were introduced
due to KSE-related (p) expansions.

Obtained from:	TrustedBSD Project
2001-10-27 05:03:05 +00:00
Robert Watson
ce5ddec25f Since namespace support was added to the UFS extended attribute
implementation to replace single-character namespace prefixes, '$' is no
longer an invalid attribute name, and the namespace is relevant to
validity determination.

o Remove '$' case from ufs_extattr_valid_attrname()
o Add attrnamespace argument to ufs_extattr_valid_attrname(), and
  fill out appropriately.

Currently no decisions are made based on the namespace argument, but
may be in the future.

Obtained from:	TrustedBSD Project
2001-10-27 04:58:28 +00:00
Matthew Dillon
245df27cee Implement kern.maxvnodes. adjusting kern.maxvnodes now actually has a
real effect.

Optimize vfs_msync().  Avoid having to continually drop and re-obtain
mutexes when scanning the vnode list.  Improves looping case by 500%.

Optimize ffs_sync().  Avoid having to continually drop and re-obtain
mutexes when scanning the vnode list.  This makes a couple of assumptions,
which I believe are ok, in regards to vnode stability when the mount list
mutex is held.  Improves looping case by 500%.

(more optimization work is needed on top of these fixes)

MFC after:	1 week
2001-10-26 00:08:05 +00:00
Ian Dowse
71fc5e11c7 Default to not performing ufs_dirhash's extensive directory-block
sanity check after every directory modification. This check can be
re-enabled at any time by setting the sysctl "vfs.ufs.dirhash_docheck"
to 1.

This group of sanity tests was there to ensure that any UFS_DIRHASH
bugs could be caught by a panic before a potentially corrupted
directory block would be written to disk. It has served its main
purpose now, so disable it in the interest of performance.

MFC after:	1 week
2001-10-25 22:55:59 +00:00
Matthew Dillon
c72ccd014d Change the vnode list under the mount point from a LIST to a TAILQ
in preparation for an implementation of limiting code for kern.maxvnodes.

MFC after:	3 days
2001-10-23 01:21:29 +00:00
John Baldwin
bd78cece5d Change the kernel's ucred API as follows:
- 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.
2001-10-11 23:38:17 +00:00
John Baldwin
7106ca0d1a Add missing includes of sys/lock.h. 2001-10-11 17:52:20 +00:00
Matthew Dillon
962922dcd2 Remove panics for rename() race conditions. The panics are inappropriate
because the IN_RENAME flag only fixes a few of the huge number of race
conditions that can result in the source path becoming invalid even
prior to the VOP_RENAME() call.  The panics created a serious security
issue whereby an attacker could fairly easily cause the panic to
occur, crashing the machine.

The correct solution requires a great deal of work in the namei
path cache code.

MFC after:	0 days
2001-10-08 00:37:54 +00:00
Robert Watson
ab66aa1468 o Replace two direct uid!=0 comparisons with suser_xxx() calls.
Obtained from:	TrustedBSD Project
2001-10-02 14:41:43 +00:00
Robert Watson
b73d2870cd o Replace two direct uid!=0 comparisons with suser_td() calls.
Obtained from:	TrustedBSD Project
2001-10-02 14:34:22 +00:00
Matthew Dillon
4c94c7bfb9 Backout the last commit. The problem is actually much worse then I
first thought and may require serious work to the VOP_RENAME() api itself.
Basically, by the time the VOP_RENAME() function is called, it's already
too late.
2001-10-02 04:26:58 +00:00
Matthew Dillon
be2a975a9f IN_RENAME should only be cleared by the routine that set it. This fixes
a rename/rmdir race that has been shown to cause a panic.

Bug reported by: Yevgeniy Aleynikov <eugenea@infospace.com>
MFC after:	3 days
2001-10-02 02:58:48 +00:00