Commit Graph

165 Commits

Author SHA1 Message Date
Poul-Henning Kamp
4be2eb8c49 I got tired of seeing all the cdevsw[major(foo)] all over the place.
Made a new (inline) function devsw(dev_t dev) and substituted it.

Changed to the BDEV variant to this format as well: bdevsw(dev_t dev)

DEVFS will eventually benefit from this change too.
1999-05-08 06:40:31 +00:00
Poul-Henning Kamp
46eede0058 Continue where Julian left off in July 1998:
Virtualize bdevsw[] from cdevsw.  bdevsw() is now an (inline)
        function.

        Join CDEV_MODULE and BDEV_MODULE to DEV_MODULE (please pay attention
        to the order of the cmaj/bmaj arguments!)

        Join CDEV_DRIVER_MODULE and BDEV_DRIVER_MODULE to DEV_DRIVER_MODULE
        (ditto!)

(Next step will be to convert all bdev dev_t's to cdev dev_t's
before they get to do any damage^H^H^H^H^H^Hwork in the kernel.)
1999-05-07 10:11:40 +00:00
Alan Cox
4221e284a3 The VFS/BIO subsystem contained a number of hacks in order to optimize
piecemeal, middle-of-file writes for NFS.  These hacks have caused no
end of trouble, especially when combined with mmap().  I've removed
them.  Instead, NFS will issue a read-before-write to fully
instantiate the struct buf containing the write.  NFS does, however,
optimize piecemeal appends to files.  For most common file operations,
you will not notice the difference.  The sole remaining fragment in
the VFS/BIO system is b_dirtyoff/end, which NFS uses to avoid cache
coherency issues with read-merge-write style operations.  NFS also
optimizes the write-covers-entire-buffer case by avoiding the
read-before-write.  There is quite a bit of room for further
optimization in these areas.

The VM system marks pages fully-valid (AKA vm_page_t->valid =
VM_PAGE_BITS_ALL) in several places, most noteably in vm_fault.  This
is not correct operation.  The vm_pager_get_pages() code is now
responsible for marking VM pages all-valid.  A number of VM helper
routines have been added to aid in zeroing-out the invalid portions of
a VM page prior to the page being marked all-valid.  This operation is
necessary to properly support mmap().  The zeroing occurs most often
when dealing with file-EOF situations.  Several bugs have been fixed
in the NFS subsystem, including bits handling file and directory EOF
situations and buf->b_flags consistancy issues relating to clearing
B_ERROR & B_INVAL, and handling B_DONE.

getblk() and allocbuf() have been rewritten.  B_CACHE operation is now
formally defined in comments and more straightforward in
implementation.  B_CACHE for VMIO buffers is based on the validity of
the backing store.  B_CACHE for non-VMIO buffers is based simply on
whether the buffer is B_INVAL or not (B_CACHE set if B_INVAL clear,
and vise-versa).  biodone() is now responsible for setting B_CACHE
when a successful read completes.  B_CACHE is also set when a bdwrite()
is initiated and when a bwrite() is initiated.  VFS VOP_BWRITE
routines (there are only two - nfs_bwrite() and bwrite()) are now
expected to set B_CACHE.  This means that bowrite() and bawrite() also
set B_CACHE indirectly.

There are a number of places in the code which were previously using
buf->b_bufsize (which is DEV_BSIZE aligned) when they should have
been using buf->b_bcount.  These have been fixed.  getblk() now clears
B_DONE on return because the rest of the system is so bad about
dealing with B_DONE.

Major fixes to NFS/TCP have been made.  A server-side bug could cause
requests to be lost by the server due to nfs_realign() overwriting
other rpc's in the same TCP mbuf chain.  The server's kernel must be
recompiled to get the benefit of the fixes.

Submitted by:	Matthew Dillon <dillon@apollo.backplane.com>
1999-05-02 23:57:16 +00:00
Poul-Henning Kamp
75c1354190 This Implements the mumbled about "Jail" feature.
This is a seriously beefed up chroot kind of thing.  The process
is jailed along the same lines as a chroot does it, but with
additional tough restrictions imposed on what the superuser can do.

For all I know, it is safe to hand over the root bit inside a
prison to the customer living in that prison, this is what
it was developed for in fact:  "real virtual servers".

Each prison has an ip number associated with it, which all IP
communications will be coerced to use and each prison has its own
hostname.

Needless to say, you need more RAM this way, but the advantage is
that each customer can run their own particular version of apache
and not stomp on the toes of their neighbors.

It generally does what one would expect, but setting up a jail
still takes a little knowledge.

A few notes:

   I have no scripts for setting up a jail, don't ask me for them.

   The IP number should be an alias on one of the interfaces.

   mount a /proc in each jail, it will make ps more useable.

   /proc/<pid>/status tells the hostname of the prison for
   jailed processes.

   Quotas are only sensible if you have a mountpoint per prison.

   There are no privisions for stopping resource-hogging.

   Some "#ifdef INET" and similar may be missing (send patches!)

If somebody wants to take it from here and develop it into
more of a "virtual machine" they should be most welcome!

Tools, comments, patches & documentation most welcome.

Have fun...

Sponsored by:   http://www.rndassociates.com/
Run for almost a year by:       http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
Poul-Henning Kamp
f711d546d2 Suser() simplification:
1:
  s/suser/suser_xxx/

2:
  Add new function: suser(struct proc *), prototyped in <sys/proc.h>.

3:
  s/suser_xxx(\([a-zA-Z0-9_]*\)->p_ucred, \&\1->p_acflag)/suser(\1)/

The remaining suser_xxx() calls will be scrutinized and dealt with
later.

There may be some unneeded #include <sys/cred.h>, but they are left
as an exercise for Bruce.

More changes to the suser() API will come along with the "jail" code.
1999-04-27 11:18:52 +00:00
Bruce Evans
44f332052d Don't depend on <ufs/ufs/quota.h> or another (old) prerequisite including
<sys/queue.h>.  This fixes my recent breakage of biosboot by unpolluting
<ufs/ufs/quota.h> in the !KERNEL case.
1999-03-06 05:21:09 +00:00
Warner Losh
5369eb85ca Merge patch to ufs_vnops.c's ufs_rename to the copy of ufs_rename that
lives in ext2_vnops.c for ext2fs.  Also remove cast from comparision.
Bruce pointed out that it was bogus since we'd force a signed
comparision when we really wanted an unsigned comparison.
1999-03-02 05:31:47 +00:00
Bruce Evans
a5c9bce777 Added a used #include (don't depend on "vnode_if.h" including <sys/buf.h>). 1999-02-25 15:54:06 +00:00
Bruce Evans
ae4d334421 Fixed parenthesization botch in previous commit. Async update of inodes
was broken.
1999-01-29 15:36:05 +00:00
Matthew Dillon
8aef171243 Fix warnings in preparation for adding -Wall -Wcast-qual to the
kernel compile
1999-01-28 00:57:57 +00:00
Matthew Dillon
fe08c21a53 Fix warnings in preparation for adding -Wall -Wcast-qual to the
kernel compile.

    This commit includes significant work to proper handle const arguments
    for the DDB symbol routines.
1999-01-27 23:45:44 +00:00
Matthew Dillon
d254af07a1 Fix warnings in preparation for adding -Wall -Wcast-qual to the
kernel compile
1999-01-27 21:50:00 +00:00
Eivind Eklund
65c0c7b08e Avoid warning for unused variable. 1999-01-11 23:32:35 +00:00
Bruce Evans
de5d1ba57c Don't pass unused unused timestamp args to UFS_UPDATE() or waste
time initializing them.  This almost finishes centralizing (in-core)
timestamp updates in ufs_itimes().
1999-01-07 16:14:19 +00:00
Bruce Evans
4591d9bb7e UFS_UPDATE() takes a boolean `waitfor' arg, so don't pass it the value
MNT_WAIT when we mean boolean `true' or check for that value not being
passed.  There was no problem in practice because MNT_WAIT had the
magic value of 1.
1999-01-06 18:18:06 +00:00
Archie Cobbs
f1d19042b0 The "easy" fixes for compiling the kernel -Wunused: remove unreferenced static
and local variables, goto labels, and functions declared but not defined.
1998-12-07 21:58:50 +00:00
Bruce Evans
b54e74eb87 Fixed a misspelling of boolean true as MNT_WAIT. 1998-11-15 15:46:33 +00:00
Peter Wemm
40c8cfe552 Use TAILQ macros for clean/dirty block list processing. Set b_xflags
rather than abusing the list next pointer with a magic number.
1998-10-31 15:31:29 +00:00
Peter Wemm
91ecc00e71 error return assignment was less than ideal. Fix the part that caused
warnings to be the same as the ffs code.  Previously, any error from
the UFS_UPDATE() call was lost (I think).
1998-10-29 09:44:12 +00:00
Peter Wemm
f6020599aa Use vtruncbuf() to clean out cached blocks on a file shorten rather than
the more expensive vinvalbuf(), based on the FFS version of the same
routine.  I don't have any ext2fs filesystems to test this on.
1998-10-29 09:30:52 +00:00
Bruce Evans
b5ee16407f Oops, the redundant tests for major numbers weren't redundant here.
They checked for the magic major number for the "device" behind mfs
mount points.  Use a more obvious check for this device.

Debugged by:		Andrew Gallatin <gallatin@cs.duke.edu>
1998-10-27 11:47:08 +00:00
Bruce Evans
569555b969 Removed redundant bitrotted checks for major numbers instead of updating
them.
1998-10-26 08:53:13 +00:00
Bruce Evans
65baf8f06b Don't follow null bdevsw pointers. The `major(dev) < nblkdev' test rotted
when bdevsw[] became sparse.  We still depend on magic to avoid having to
check that (v_rdev) device numbers in vnodes are not NODEV.
1998-10-25 19:26:18 +00:00
Bruce Evans
d2165c2f7d Fixed bloatage of `struct inode'. We used 5 "spare" fields for ext2fs,
but when i_effnlink was added to support soft updates, there was only
room for 4 spares.  The number of spares was not reduced, so the inode
size became 260 (on i386's), or 512 after rounding up by malloc().
Use one spare field in `struct dinode' instead of the 5th spare field
in the inode and reduced to 4 spares in the inode so that the size is
256 again.

Changed the types of the spares in the inode from int to u_int32_t
so that the inode size has more chance of being <= 256 under other
arches, and downdated ext2fs to match (it was broken to use ints
before rev.1.1).
1998-10-13 15:45:43 +00:00
Bruce Evans
8f359bc68c Quick fix for not being able to sync all the buffers in boot() if
an ext2fs file system is mounted.  The soft update changes added
a check for B_DELWRI buffers.  This exposed the complete brokenness
of the previous quick fix for failing syncs (PR 3571, committed on
1997/08/04).  Use a new buffer flag B_DIRTY and don't abuse B_DELWRI.
B_DIRTY buffers are still written too late, as broken in the previous
fix.  This is fairly harmless, because B_DIRTY is only used for
bitmap buffers and fsck.ext2 can fix up the bitmaps perfectly.

Fixed a race in ULCK_BUF() (bremfree() was outside of the splbio()
section).
1998-10-03 16:19:28 +00:00
Bruce Evans
9702cd0422 Fixed initialization of new inodes. ext2fs doesn't clear inodes when
they are deleted, so inodes must be cleared when they are reused, but
we didn't clear the indirect blocks.  This caused serious filesystem
corruption.
1998-09-29 08:07:32 +00:00
Bruce Evans
6674be30f6 Updated ext2_reload() and ext2_sync(). Locking was broken, and MNT_LAZY
syncs weren't optimized properly (they probably still aren't, but are bug
for bug compatible with ffs).  These fixes are mostly academic, since
ext2fs is too broken to mount read-write (it apparently doesn't clear
indirect blocks).

Obtained from:	mostly from Lite2
1998-09-26 12:42:17 +00:00
Bruce Evans
7cff8977ca Fixed missing newlines in messages in ext2_check_descriptors().
Fixed vnode and memory leaks after an unlikely (?) error in
ext2_mountfs().

Fixed an unconditional memory leak in ext2_unmount().
1998-09-26 07:16:41 +00:00
Bruce Evans
a094db128f Fixed clean flag handling:
Fixes for bugs not shared with ffs:
- don't mount unclean filesystems rw unless forced to.
- accept EXT2_ERROR_FS (treat it like !EXT2_VALID_FS).  We still don't set
  this or honour the maximal mount count.
- don't attempt to print the name of the mount point when mounting an
  unclean file system, since the name of the previous mount point is
  unknown and the name of the current mount point is still "".

Fixes for bugs shared with ffs until recently:
- don't set the clean flag on unmount of an initially-unclean filesystem
  that was (forcibly) mounted rw.
- set the clean flag on rw -> ro update of a mounted initially-clean
  filesystem.
- fixed some style bugs (mostly long lines).

The fixes are slightly simpler than for ffs, because the relevant on-disk
state is not a simple boolean variable, and the superblock has a core-only
extension.

Obtained from:	parts from ffs_vfsops.c, parts from NetBSD
1998-09-26 06:18:59 +00:00
Bruce Evans
5d207357db Fixed the usual missing permissions checks in mount(). As for cd9660,
the damage was limited by the default of 0 for vfs.usermount.

Obtained from:	Lite2 via the -current ffs_vfsops.c
1998-09-09 20:21:18 +00:00
Bruce Evans
05d46b3cd6 Don't forget to initialize the inode lock. This bug caused
surprisingly few problems.  Most fields were initialized to the
correct values by bzero(), but lk_prio was 0 instead of PINOD (=8),
the lk_wmsg was NULL instead of "ext2in", and lk_lockholder was 0
instead of -1.

Obtained from:	Lite2 via the -current ffs_vfsops.c
1998-09-09 13:09:24 +00:00
Bruce Evans
d6c54caabe Support compiling with `gcc -pedantic' (don't use hard newlines in
(asm) string constants).
1998-09-09 12:22:17 +00:00
Bruce Evans
8994ca3ce9 Removed statically configured mount type numbers (MOUNT_*) and all
references to them.

The change a couple of days ago to ignore these numbers in statically
configured vfsconf structs was slightly premature because the cd9660,
cfs, devfs, ext2fs, nfs vfs's still used MOUNT_* instead of the number
in their vfsconf struct.
1998-09-07 13:17:06 +00:00
Bruce Evans
1874ef935c Quick fix for breakage of read clustering on non-IDE drives. Read
clustering is obsolescent technology so hardly anyone noticed.  On
a DORS 32160 SCSI drive with 4 tags, read clustering makes very
little difference even for huge sequential reads.  However, on a
ZIP SCSI drive with 0 tags, the minimum overhead per block is about
40 msec, so very large clusters must be used to get anywhere near
the maximum transfer rate.  Using clusters consisting of 1 8K block
reduces the transfer rate to about 250K/sec.  Under msdosfs, missing
read clustering is normal and a cluster size of 1 512 byte block
reduces the transfer rate to about 25K/sec.

Broken in:	rev.1.18
1998-08-18 03:54:39 +00:00
Mike Smith
f01beb610a "The releaseing of the reference and lock is not temporary and belongs
where it is.  The reference and lock(s) are acquired just above the
 code in VREF() and relookup()."

Submitted by:	Michael Hancock <michaelh@cet.co.jp>
1998-08-12 21:42:54 +00:00
Bruce Evans
85badd7eba Fixed printf format errors. 1998-07-30 17:12:39 +00:00
Julian Elischer
49cc016a39 add anti-panic workaround from chris radek (cradek@in221.inetnebr.com)
Not sure why this is needed but but does stop crashes.
1998-07-30 03:22:52 +00:00
Bruce Evans
ac1e407b32 Fixed printf format errors. 1998-07-11 07:46:16 +00:00
Julian Elischer
6deaf84b1f Catch a few corner cases where FreeBSD differs enough from BSD 4.4 to
confuse Soft updates..
Should solve several "dangling deps" panics.
1998-07-08 01:04:33 +00:00
Julian Elischer
fd5d1124e2 VOP_STRATEGY grows an (struct vnode *) argument
as the value in b_vp is often not really what you want.
(and needs to be frobbed). more cleanups will follow this.
Reviewed by: Bruce Evans <bde@freebsd.org>
1998-07-04 20:45:42 +00:00
Bruce Evans
3055187290 Sync timestamp changes for inodes of special files to disk as late
as possible (when the inode is reclaimed).  Temporarily only do
this if option UFS_LAZYMOD configured and softupdates aren't enabled.
UFS_LAZYMOD is intentionally left out of /sys/conf/options.

This is mainly to avoid almost useless disk i/o on battery powered
machines.  It's silly to write to disk (on the next sync or when the
inode becomes inactive) just because someone hit a key or something
wrote to the screen or /dev/null.

PR:		5577
Previous version reviewed by:	phk
1998-07-03 22:17:03 +00:00
Bruce Evans
33cc029eab Centralized in-core inode update. Update the in-core inode directly
in ufs_setattr() so that there is no need to pass timestamps to
UFS_UPDATE() (everything else just needs the current time).  Ignore
the passed-in timestamps in UFS_UPDATE() and always call ufs_itimes()
(was: itimes()) to do the update.  The timestamps are still passed
so that all the callers don't need to be changed yet.
1998-07-03 18:46:52 +00:00
Bruce Evans
add4ae9324 Fixed (?) races in mark_buffer_dirty(). We abuse the buffer cache
by hacking on locked buffers without getblk()ing them, and we didn't
even use splbio() to prevent biodone() changing the buffer underneath
use when a write completes.  I think there was no problem in practice
on i386's because the operations on b_flags and numdirtybufs happen to
be atomic.  We still depend on biodone()'s operations on b_flags not
interfering with ours.  I think there is only interference for B_ERROR,
and this is harmless because errors for async writes are ignored anyway.

Don't use mark_buffer_dirty() except for superblock-related metadata.
It was used in just one case where ordinary BSD buffering is more
natural.
1998-06-21 21:06:04 +00:00
Bruce Evans
9b7a8fb7d8 Removed unused function ll_w_block(). It has always had races due
to not using splbio(), and has rotted a little.  The races were
probably harmless in practice because this function was only used
for superblock updates, and separate superblock updates are probably
prevented from running into each other by doing part of the update
synchronously.
1998-06-21 19:56:31 +00:00
Bruce Evans
be160d60ab Removed unused includes. 1998-06-21 18:02:50 +00:00
Bruce Evans
e5b19842ef Removed unused includes. 1998-06-21 14:53:44 +00:00
Bruce Evans
4344f492c4 Added a missing options include. 1998-06-21 12:36:12 +00:00
Bruce Evans
dae50f6c50 Don't use "ffs" in an ext2fs sleep message string.
Don't forget to clear the inode hash lock before returning from ext2_vget()
after getnewvnode() fails.  Obtained from: rev.1.24 of ffs_vfsops.c (the
original patch for the getnewvnode() race).  Forgotten in: rev.1.4 here.

Removed a duplicate comment.  Duplicated in: rev.1.4 here.

Fixed the MALLOC() vs getnewvnode() race in ext2_vget().  Obtained from:
rev.1.39 of ffs_vfsops.c.
1998-05-16 17:47:44 +00:00
Bruce Evans
2c8838fec4 Abbreviate "ext2fs_fsync" as "e2fsyn" instead of as "extfsn" in a sleep
message string.
1998-05-16 16:52:20 +00:00
Mike Smith
7be2d30077 In the words of the submitter:
---------
Make callers of namei() responsible for releasing references or locks
instead of having the underlying filesystems do it.  This eliminates
redundancy in all terminal filesystems and makes it possible for stacked
transport layers such as umapfs or nullfs to operate correctly.

Quality testing was done with testvn, and lat_fs from the lmbench suite.

Some NFS client testing courtesy of Patrik Kudo.

vop_mknod and vop_symlink still release the returned vpp.  vop_rename
still releases 4 vnode arguments before it returns.  These remaining cases
will be corrected in the next set of patches.
---------

Submitted by:	Michael Hancock <michaelh@cet.co.jp>
1998-05-07 04:58:58 +00:00