Commit Graph

476 Commits

Author SHA1 Message Date
Matthew Dillon
dfbdd7d2b3 A tentative agreement has been reached in regards to a procedure
to remove 'b'lock devices.  The agreement is, essentially, that
    block devices will be collapsed into character devices as a first
    step (though I don't particularly agree), and raw device names 'rxxx'
    will become simply 'xxx' in devfs in the second step (i.e. no 'rxxx'
    names will exist).  The renaming will not effect the original /dev
    and the expectation is that devfs will eventually (but not immediately)
    become the standard way to access devices in the system.

    If it is determined that a reimplementation of block device access
    characteristics is beneficial, a number of alternatives will
    be possible that do not involve resurrecting the 'b'lock device class.
    For example, an ioctl() that might be made on an open character device
    descriptor or a generic buffered overlay device.

    This commit removes the blockdev disablement sysctl which does not
    apply to the solution that was reached.
1999-10-20 06:31:49 +00:00
Poul-Henning Kamp
8342096f29 Change the default for the vfs.bdev_buffered sysctl to zero.
This means that access to block devices nodes will act the
same as char device nodes for disk-like devices.

If you encounter problems after this, where programs accessing
disks directly fail to operate, please use the following command
to revert to previous behaviour:

        sysctl -w vfs.bdev_buffered=1

And verify that this was indeed the cause of your trouble.

See the mail-archives of the arch@FreeBSD.org list for background.
1999-10-18 16:59:50 +00:00
Poul-Henning Kamp
1201869007 Add a couple of strategic KASSERTs 1999-10-08 19:07:23 +00:00
Poul-Henning Kamp
856de19089 Add back sysctl vfs.enable_userblk_io 1999-10-08 18:25:19 +00:00
Poul-Henning Kamp
adab70d67a Warn once per driver about dev_t's not registered with make_dev(). 1999-10-04 12:33:05 +00:00
Poul-Henning Kamp
aa4f4b695e Move the buffered read/write code out of spec_{read|write} and into
two new functions spec_buf{read|write}.

Add sysctl vfs.bdev_buffered which defaults to 1 == true.  This
sysctl can be used to experimentally turn buffered behaviour for
bdevs off.  I should not be changed while any blockdevices are
open.  Remove the misplaced sysctl vfs.enable_userblk_io.

No other changes in behaviour.
1999-10-04 11:23:10 +00:00
Poul-Henning Kamp
3b6fb88590 Before we start to mess with the VFS name-cache clean things up a little bit:
Isolate the namecache in its own file, and give it a dedicated malloc type.
1999-10-03 12:18:29 +00:00
Poul-Henning Kamp
1b5464ef9d Remove v_maxio from struct vnode.
Replace it with mnt_iosize_max in struct mount.

Nits from:	bde
1999-09-29 20:05:33 +00:00
Marcel Moolenaar
2c42a14602 sigset_t change (part 2 of 5)
-----------------------------

The core of the signalling code has been rewritten to operate
on the new sigset_t. No methodological changes have been made.
Most references to a sigset_t object are through macros (see
signalvar.h) to create a level of abstraction and to provide
a basis for further improvements.

The NSIG constant has not been changed to reflect the maximum
number of signals possible. The reason is that it breaks
programs (especially shells) which assume that all signals
have a non-null name in sys_signame. See src/bin/sh/trap.c
for an example. Instead _SIG_MAXSIG has been introduced to
hold the maximum signal possible with the new sigset_t.

struct sigprop has been moved from signalvar.h to kern_sig.c
because a) it is only used there, and b) access must be done
though function sigprop(). The latter because the table doesn't
holds properties for all signals, but only for the first NSIG
signals.

signal.h has been reorganized to make reading easier and to
add the new and/or modified structures. The "old" structures
are moved to signalvar.h to prevent namespace polution.

Especially the coda filesystem suffers from the change, because
it contained lines like (p->p_sigmask == SIGIO), which is easy
to do for integral types, but not for compound types.

NOTE: kdump (and port linux_kdump) must be recompiled.

Thanks to Garrett Wollman and Daniel Eischen for pressing the
importance of changing sigreturn as well.
1999-09-29 15:03:48 +00:00
Matthew Dillon
e3a285c715 Make sure file after VOP_OPEN is VMIO'd when transfering control from
a lower layer to an upper layer.  I'm not sure how necessary this is
    for reading.

    Fix bug in union_lookup() (note: there are probably still several bugs
    in union_lookup()).  This one set lerror as a side effect without
    setting lowervp, causing copyup code further on down to crash on a null
    lowervp pointer.  Changed the side effect to use a temporary variable
    instead.
1999-09-28 05:48:39 +00:00
Matthew Dillon
2a31267e43 This is a major fixup of unionfs. At least 30 serious bugs have been
fixed (many due to changing semantics in other parts of the kernel and not
    the original author's fault), including one critical one: unionfs could
    cause UFS corruption in the fronting store due to calling VOP_OPEN for
    writing without turning on vmio for the UFS vnode.

    Most of the bugs were related to semantics changes in VOP calls, lock
    ordering problems (causing deadlocks), improper handling of a read-only
    backing store (such as an NFS mount), improper referencing and locking
    of vnodes, not using real struct locks for vnode locking, not using
    recursive locks when accessing the fronting store, and things like that.

    New functionality has been added:  unionfs now has mmap() support, but
    only partially tested, and rename has been enhanced considerably.

    There are still some things that unionfs cannot do.   You cannot
    rename a directory without confusing unionfs, and there are issues
    with softlinks, hardlinks, and special files.  unionfs mostly doesn't
    understand them (and never did).

    There are probably still panic situations, but hopefully no where near
    as many as before this commit.

    The unionfs in this commit has been tested overlayed on /usr/src
    (backing /usr/src being a read-only NFS mount, fronting /usr/src being
    a local filesystem).  kernel builds have been tested, buildworld is
    undergoing testing.  More testing is necessary.
1999-09-26 20:52:41 +00:00
Poul-Henning Kamp
9ab8e01991 Remove a warning check which was too general. 1999-09-25 18:52:03 +00:00
Poul-Henning Kamp
d6a0e38a1b Remove five now unused fields from struct cdevsw. They should never
have been there in the first place.  A GENERIC kernel shrinks almost 1k.

Add a slightly different safetybelt under nostop for tty drivers.

Add some missing FreeBSD tags
1999-09-25 18:24:47 +00:00
Poul-Henning Kamp
ae8e1d08d7 This patch clears the way for removing a number of tty related
fields in struct cdevsw:

        d_stop          moved to struct tty.
        d_reset         already unused.
        d_devtotty      linkage now provided by dev_t->si_tty.

These fields will be removed from struct cdevsw together with
d_params and d_maxio Real Soon Now.

The changes in this patch consist of:

        initialize dev->si_tty in *_open()
        initialize tty->t_stop
        remove devtotty functions
        rename ttpoll to ttypoll
        a few adjustments to these changes in the generic code
        a bump of __FreeBSD_version
        add a couple of FreeBSD tags
1999-09-25 16:21:39 +00:00
Poul-Henning Kamp
c428d4c048 Kill the cdevsw->d_maxio field.
d_maxio is replaced by the dev->si_iosize_max field which the driver
should be set in all calls to cdevsw->d_open if it has a better
idea than the system wide default.

The field is a generic dev_t field (ie: not disk specific) so that
tapes and other devices can use physio as well.
1999-09-22 19:56:14 +00:00
Matthew Dillon
8411bf657d Fix handling of a device EOF that occurs in the middle of a block. The
transfer size calculation was incorrect resulting in the last read being
    potentially larger then the actual extent of the device.

    EOF and write handling has not yet been fixed.

Reviewed by:	Tor.Egge@fast.no
1999-09-20 23:17:47 +00:00
Poul-Henning Kamp
fae03f66d1 Step one of replacing devsw->d_maxio with si_bsize_max.
Rename dev->si_bsize_max to si_iosize_max and set it in spec_open
if the device didn't.

Set vp->v_maxio from dev->si_bsize_max in spec_open rather than
in ufs_bmap.c
1999-09-20 19:57:28 +00:00
Matthew Dillon
bb01f28e97 Add vfs.enable_userblk_io sysctl to control whether user reads and writes
to buffered block devices are allowed.  The default is to be backwards
    compatible, i.e. reads and writes are allowed.

    The idea is for a larger crowd to start running with this disabled and
    see what problems, if any, crop up, and then to change the default to
    off and see if any problems crop up in the next 6 months prior to
    potentially removing support entirely.  There are still a few people,
    Julian and myself included, who believe the buffered block device
    access from usermode to be useful.

    Remove use of vnode->v_lastr from buffered block device I/O in
    preparation for removal of vnode->v_lastr field, replacing it with
    the already existing seqcount metric to detect sequential operation.

Reviewed by:	Alan Cox <alc@cs.rice.edu>, David Greenman <dg@root.com>
1999-09-17 06:10:27 +00:00
Alfred Perlstein
c24fda81c9 Seperate the export check in VFS_FHTOVP, exports are now checked via
VFS_CHECKEXP.

Add fh(open|stat|stafs) syscalls to allow userland to query filesystems
based on (network) filehandle.

Obtained from:	NetBSD
1999-09-11 00:46:08 +00:00
Julian Elischer
85a219d201 Changes to centralise the default blocksize behaviour.
More likely to follow.

Submitted by: phk@freebsd.org
1999-09-09 19:08:44 +00:00
Alfred Perlstein
5a5fccc8e7 All unimplemented VFS ops now have entries in kern/vfs_default.c that return
reasonable defaults.

This avoids confusing and ugly casting to eopnotsupp or making dummy functions.
Bogus casting of filesystem sysctls to eopnotsupp() have been removed.

This should make *_vfsops.c more readable and reduce bloat.

Reviewed by:	msmith, eivind
Approved by:	phk
Tested by:	Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>
1999-09-07 22:42:38 +00:00
Bruce Evans
a0f40f54aa Get rid of the NULLFS_DIAGNOSTIC option. This option was as useful as
the other XXXFS_DIAGNOSTIC options (not very) and mostly controlled
tracing of normal operation.  Use `#ifdef DEBUG' for non-diagnostics
and `#ifdef DIAGNOSTIC' for diagnostics.
1999-09-04 12:35:09 +00:00
Bruce Evans
622163afbe Fixed the previous change. Some more code controlled by UMAPFS_DIAGNOSTIC
is actually for diagnostics; control it with DIAGNOSTIC and not DDB.
1999-09-04 11:51:41 +00:00
Julian Elischer
1f7e07329d Print out the device name when there is an uninitialised IO size or IO error
in spec_getpages().

Submitted by:	phk suggested the idea.
1999-09-03 09:14:36 +00:00
Julian Elischer
c5d6480694 Add a catchall to set default blocksize values for disk like devices.
Submitted by:	phk@freebsd.org
1999-09-03 08:26:46 +00:00
Julian Elischer
7012bab988 Revert a bunch of contraversial changes by PHK. After
a quick think and discussion among various people some form of some of
these changes will probably be recommitted.

The reversion requested was requested by dg while discussions proceed.
PHK has indicated that he can live with this, and it has been agreed
that some form of some of these changes may return shortly after further
discussion.
1999-09-03 05:16:59 +00:00
Poul-Henning Kamp
3b7df19ba4 Set the buffersize for non BSDFFS labeled partitions to
max(dev->si_bsize_phys, BLKDEV_IOSIZE).

Requested by:   davidg
1999-08-31 21:46:42 +00:00
Poul-Henning Kamp
586e1b7b46 Make buffered acces to bdevs from userland controllable with
a sysctl vfs.bdev_access.
1999-08-31 21:01:57 +00:00
Poul-Henning Kamp
02e1576966 Make bdev userland access work like cdev userland access unless
the highly non-recommended option ALLOW_BDEV_ACCESS is used.

(bdev access is evil because you don't get write errors reported.)

Kill si_bsize_best before it kills Matt :-)

Use the specfs routines rather having cloned copies in devfs.
1999-08-30 07:56:23 +00:00
Bruce Evans
4047cd0bb2 Converted the silly SAFTEY option into a new-style option by renaming it to
DIAGNOSTIC.

Fixed an English style bug in the panic messages controlled by SAFETY.
1999-08-30 07:08:04 +00:00
Bruce Evans
8021ba879a Changed old-style option UNION_DIAGNOSTIC to DEBUG and fixed printf
format errors exposed by this.  It has nothing to do with diagnostics
since it does little more than control tracing of normal operation.
Actual diagnostics for the union file system are still controlled by
the DIAGNOSTIC option.
1999-08-29 10:03:35 +00:00
Bruce Evans
b00b6d73dd Changed old-style options UMAPFS_DIAGNOSTIC and UMAP_DIAGNOSTIC to DEBUG
or DDB and fixed printf format errors exposed by this.  The options had
little to do with diagnostics; they mostly controlled tracing of normal
operation.
1999-08-29 09:54:17 +00:00
Bruce Evans
47966f4c8a Changed old-style option KERNFS_DIAGNOSTIC to DEBUG and fixed printf
format errors exposed by this.  It has nothing to do with diagnostics
since it does little more than control tracing of normal operation.
1999-08-29 09:38:25 +00:00
Poul-Henning Kamp
94db13fe98 Fix various trivial warnings from LINT 1999-08-28 19:44:07 +00:00
Peter Wemm
c3aac50f28 $Id$ -> $FreeBSD$ 1999-08-28 01:08:13 +00:00
Poul-Henning Kamp
dbafb3660f Simplify the handling of VCHR and VBLK vnodes using the new dev_t:
Make the alias list a SLIST.

        Drop the "fast recycling" optimization of vnodes (including
        the returning of a prexisting but stale vnode from checkalias).
        It doesn't buy us anything now that we don't hardlimit
        vnodes anymore.

        Rename checkalias2() and checkalias() to addalias() and
        addaliasu() - which takes dev_t and udev_t arg respectively.

        Make the revoke syscalls use vcount() instead of VALIASED.

        Remove VALIASED flag, we don't need it now and it is faster
        to traverse the much shorter lists than to maintain the
        flag.

        vfs_mountedon() can check the dev_t directly, all the vnodes
        point to the same one.

Print the devicename in specfs/vprint().

Remove a couple of stale LFS vnode flags.

Remove unimplemented/unused LK_DRAINED;
1999-08-26 14:53:31 +00:00
Julian Elischer
2a1be8333d Make a place to store the devfs hook for the block device, as the same
specinfo is used to identify both raw and block version sof a device.
1999-08-25 22:50:12 +00:00
Julian Elischer
0ff7b13acd Make DEVFS use PHK's specinfo struct as the source of dev_t and devsw.
In lookup() however it's the other way around as we need to supply the
dev_t for the vnode, so devfs still has a copy of it stashed away.

Sourcing it from the vnode in the vnops however is useful as it makes
a lot of the code almost the same as that in specfs.
1999-08-25 04:55:20 +00:00
Julian Elischer
2053751093 Make DEVFS ops reflect some of the changes that have recently occured in the
specfs version of the same ops. Merging with phk's dev_t changes is not yet
complete.

i.e. devfs still uses it's own devsw pointer rather than following
devsw(a_vp->v_rdev), and it's own copy of the dev_t.

This fixes some broken actions re: syscons and DEVFS due to bitrot in
devfs vops.
1999-08-25 02:04:40 +00:00
Julian Elischer
bb70475a92 Fix comment to match reality..
vop_strategy gets a vnode argument these days.
1999-08-25 00:26:34 +00:00
Julian Elischer
792b73b97b Devfs now adds itself to the hook that PHK supplies.
block devices are still not handled correctly, though a kludge
may make them "ok".
1999-08-24 20:30:15 +00:00
Marcel Moolenaar
63a9927353 Let processes retrieve their argv through procfs. Revert to the original
behaviour in all other cases.

Submitted by: Andrew Gordon <arg@arg1.demon.co.uk>
1999-08-19 19:41:08 +00:00
Bill Fumerola
98f3a0a19e Remove an unused variable.
Reviewed by:	bde
1999-08-17 22:08:55 +00:00
Alan Cox
2c28a10540 Add the (inline) function vm_page_undirty for clearing the dirty bitmask
of a vm_page.

Use it.

Submitted by:	dillon
1999-08-17 04:02:34 +00:00
Poul-Henning Kamp
49ff4debd3 Spring cleaning around strategy and disklabels/slices:
Introduce BUF_STRATEGY(struct buf *, int flag) macro, and use it throughout.
please see comment in sys/conf.h about the flag argument.

Remove strategy argument from all the diskslice/label/bad144
implementations, it should be found from the dev_t.

Remove bogus and unused strategy1 routines.

Remove open/close arguments from dssize().  Pick them up from dev_t.

Remove unused and unfinished setgeom support from diskslice/label/bad144 code.
1999-08-14 11:40:51 +00:00
Poul-Henning Kamp
2820b2e762 Add support for device drivers which want to track all open/close
operations.  This allows a device driver better insight into
what is going on that the current:

        proc1:  open /dev/foo R/O
                        devsw->open( R/O, proc1 )
        proc2:  open /dev/foo R/W
                        devsw->open( R/W, proc2 )
        proc2:  close
                        /* nothing, but device is
                           really only R/O open */
        proc1:  close
                        devsw->close( R/O, proc1 )
1999-08-13 16:29:27 +00:00
Poul-Henning Kamp
608bb3ffdf Remove spec_getattr(), which as far as I can tell can never be called from the current code-paths, and if it were, would panic on any unmounted bdev. 1999-08-13 10:53:58 +00:00
Poul-Henning Kamp
7dc5cd047f The bdevsw() and cdevsw() are now identical, so kill the former. 1999-08-13 10:29:38 +00:00
Poul-Henning Kamp
4d4f932326 s/v_specinfo/v_rdev/ 1999-08-13 10:10:12 +00:00
Poul-Henning Kamp
0ef1c82630 Decommision miscfs/specfs/specdev.h. Most of it goes into <sys/conf.h>,
a few lines into <sys/vnode.h>.

Add a few fields to struct specinfo, paving the way for the fun part.
1999-08-08 18:43:05 +00:00