]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoint/long printf format problems on the Alpha.
peter [Thu, 1 Jul 1999 19:56:50 +0000 (19:56 +0000)]
int/long printf format problems on the Alpha.

25 years agoFix some int/long printf problems for the Alpha
peter [Thu, 1 Jul 1999 19:53:43 +0000 (19:53 +0000)]
Fix some int/long printf problems for the Alpha

25 years agoQuiet warnings on an Alpha. CBSIZE has long type and causes the other
peter [Thu, 1 Jul 1999 19:46:36 +0000 (19:46 +0000)]
Quiet warnings on an Alpha. CBSIZE has long type and causes the other
ints to promote to long.

25 years agoQuiet warnings on Alpha. (db_expr_t is a long on alpha, int on x86)
peter [Thu, 1 Jul 1999 19:42:56 +0000 (19:42 +0000)]
Quiet warnings on Alpha.  (db_expr_t is a long on alpha, int on x86)

25 years agoRemove the now-bogus comment about using iosiz with npx0 for memory
jkh [Thu, 1 Jul 1999 18:39:23 +0000 (18:39 +0000)]
Remove the now-bogus comment about using iosiz with npx0 for memory
sizing - environment does this properly  now.  Thanks, Peter!

25 years agoLook up the kernel environment for MAXMEM as a final override for the
peter [Thu, 1 Jul 1999 18:33:22 +0000 (18:33 +0000)]
Look up the kernel environment for MAXMEM as a final override for the
memory size.  If somebody wants to change the name, fine - I used this
since it's consistant with the config variable it replaces.
This is intended to replace the npx0 msize hack (which no longer works).

25 years agoMove kern_envp and preload initialization a little earlier so that we
peter [Thu, 1 Jul 1999 18:27:15 +0000 (18:27 +0000)]
Move kern_envp and preload initialization a little earlier so that we
can do a getenv_int() inside the memory sizing routines to override the
memory limit.

25 years agoFix ambigious else.
billf [Thu, 1 Jul 1999 17:59:17 +0000 (17:59 +0000)]
Fix ambigious else.

Reviewed By: eivind

25 years agoDon't include trailing whitespace in ID tokens before comments. ie:
peter [Thu, 1 Jul 1999 16:20:16 +0000 (16:20 +0000)]
Don't include trailing whitespace in ID tokens before comments. ie:
options FOO=10 # comment
would give FOO the value of "10 " and that caused unwanted
touches on the opt_*.h files.

I hope I've got this right..

25 years agoDetect and remove defunct or unknown options from opt_*.h files. This
peter [Thu, 1 Jul 1999 16:17:13 +0000 (16:17 +0000)]
Detect and remove defunct or unknown options from opt_*.h files.  This
can happen when options are removed from the options files.

25 years agoFix two easy warnings when using BRIDGE without IPFIREWALL.
peter [Thu, 1 Jul 1999 15:07:08 +0000 (15:07 +0000)]
Fix two easy warnings when using BRIDGE without IPFIREWALL.

25 years agoFix some warnings, unused functions etc.
peter [Thu, 1 Jul 1999 15:05:11 +0000 (15:05 +0000)]
Fix some warnings, unused functions etc.

25 years agoDocument that FTP_PASSIVE_MODE should be explicitly
ru [Thu, 1 Jul 1999 14:32:07 +0000 (14:32 +0000)]
Document that FTP_PASSIVE_MODE should be explicitly
set to `YES' for passive mode to be used by default.
Just setting FTP_PASSIVE_MODE is not sufficient, as
it was before.

Noted by: eivind
Reviewed by: des

25 years agoTurn System V command line syntax ``on'' by default.
ru [Thu, 1 Jul 1999 13:33:56 +0000 (13:33 +0000)]
Turn System V command line syntax ``on'' by default.

Requested by: peter
Reviewed by: des, billf

25 years agoFix warning. va_fsid is udev_t, which is int32_t. No need to use %lx.
peter [Thu, 1 Jul 1999 13:32:54 +0000 (13:32 +0000)]
Fix warning. va_fsid is udev_t, which is int32_t.  No need to use %lx.

25 years agoSpelll 'asynchronous' coriectly
kris [Thu, 1 Jul 1999 13:31:30 +0000 (13:31 +0000)]
Spelll 'asynchronous' coriectly

25 years agoSlight reorganization of kernel thread/process creation. Instead of using
peter [Thu, 1 Jul 1999 13:21:46 +0000 (13:21 +0000)]
Slight reorganization of kernel thread/process creation.  Instead of using
SYSINIT_KT() etc (which is a static, compile-time procedure), use a
NetBSD-style kthread_create() interface.  kproc_start is still available
as a SYSINIT() hook.  This allowed simplification of chunks of the
sysinit code in the process.  This kthread_create() is our old kproc_start
internals, with the SYSINIT_KT fork hooks grafted in and tweaked to work
the same as the NetBSD one.

One thing I'd like to do shortly is get rid of nfsiod as a user initiated
process.  It makes sense for the nfs client code to create them on the
fly as needed up to a user settable limit.  This means that nfsiod
doesn't need to be in /sbin and is always "available".  This is a fair bit
easier to do outside of the SYSINIT_KT() framework.

25 years agoFix the NAME section of the kldload.2 manpage, which made reference to
sheldonh [Thu, 1 Jul 1999 12:16:06 +0000 (12:16 +0000)]
Fix the NAME section of the kldload.2 manpage, which made reference to
kldunload. Add proper cross-references to the whole family.

PR: 12472
Submitted by: Chris Costello <chris@calldei.com>

25 years agoSigh. It's always the smallest changes that generate the largest public
jkh [Thu, 1 Jul 1999 11:36:38 +0000 (11:36 +0000)]
Sigh. It's always the smallest changes that generate the largest public
commentary. :)  3rd and final frob of this.  Leave enough comments
behind that anyone running into trouble at least has some cited clues
on dealing with it and jump into the brave new world with uncommented
IANA port assignments.

25 years ago(1) Don't perror() when at.deny doesn't exist, that's an accepted behavior.
billf [Thu, 1 Jul 1999 01:07:30 +0000 (01:07 +0000)]
(1) Don't perror() when at.deny doesn't exist, that's an accepted behavior.

(2) Die when there is a problem opening at.allow other then it not existing.
    An error other then it not existing might be a trick to somehow
    circumvent system security.

Mostly Reviewed By: msmith

25 years agoCompromise proposal: Comment out both old and new radius ports, explain
jkh [Thu, 1 Jul 1999 00:08:08 +0000 (00:08 +0000)]
Compromise proposal:  Comment out both old and new radius ports, explain
the differences in the comments, let the USER select which assignment to
use now.

Suggested by: rgrimes

25 years agoOmmitted in previous commit message:
sheldonh [Wed, 30 Jun 1999 23:47:46 +0000 (23:47 +0000)]
Ommitted in previous commit message:

Submitted by: David Malone <dwmalone@maths.tcd.ie>

25 years agoEnable wrapping for dgram services and fix logging so that -l really
sheldonh [Wed, 30 Jun 1999 23:36:39 +0000 (23:36 +0000)]
Enable wrapping for dgram services and fix logging so that -l really
does log all connections.

25 years agoIncrease the size of the route reference count from 15 bits to 31 bits.
msmith [Wed, 30 Jun 1999 23:11:15 +0000 (23:11 +0000)]
Increase the size of the route reference count from 15 bits to 31 bits.

This doesn't change the size or alignment of the structure on either i386
or Alpha, and thus should be binary-compatible (modulo problems with old
applications and routes with more than 2^15 references).

Reviewed by: peter

25 years agoAllow the use of `-' as an argument to the -w option so that standard
sheldonh [Wed, 30 Jun 1999 21:46:03 +0000 (21:46 +0000)]
Allow the use of `-' as an argument to the -w option so that standard
input may be used (e.g. gunzip -c /var/log/wtmp.Jan.gz | ac -w - ).

PR: 12467
Submitted by: wollman

25 years agoSlight tweak to fork1() calling conventions. Add a third argument so
peter [Wed, 30 Jun 1999 15:33:41 +0000 (15:33 +0000)]
Slight tweak to fork1() calling conventions.  Add a third argument so
the caller can easily find the child proc struct.  fork(), rfork() etc
syscalls set p->p_retval[] themselves.  Simplify the SYSINIT_KT() code
and other kernel thread creators to not need to use pfind() to find the
child based on the pid.  While here, partly tidy up some of the fork1()
code for RF_SIGSHARE etc.

25 years agoSync with sys/i386/i386/userconfig.c revision 1.147.
kato [Wed, 30 Jun 1999 13:30:58 +0000 (13:30 +0000)]
Sync with sys/i386/i386/userconfig.c revision 1.147.

25 years agoSync with sys/i386/i386/machdep.c revision 1.345.
kato [Wed, 30 Jun 1999 13:30:10 +0000 (13:30 +0000)]
Sync with sys/i386/i386/machdep.c revision 1.345.

25 years agoretreived -> retrieved
kris [Wed, 30 Jun 1999 12:53:06 +0000 (12:53 +0000)]
retreived -> retrieved

Obtained from: OpenBSD

25 years agoGrammar and spelling fixes
kris [Wed, 30 Jun 1999 12:49:56 +0000 (12:49 +0000)]
Grammar and spelling fixes

Obtained from: OpenBSD

25 years agoFlesh out VOP_ADVLOCK.9 with a description of the available flags
kris [Wed, 30 Jun 1999 12:02:52 +0000 (12:02 +0000)]
Flesh out VOP_ADVLOCK.9 with a description of the available flags
Sync vnode.9 with the definitions in <sys/vnode.h>

Reviewed by: Doug Rabson

25 years agoCorrect reference to the obsolete vadvise() to madvise() (with appropriate
kris [Wed, 30 Jun 1999 11:50:09 +0000 (11:50 +0000)]
Correct reference to the obsolete vadvise() to madvise() (with appropriate
arguments)

PR: 11586
Submitted by: David Gilbert <dgilbert@velocet.ca>
Reviewed by: Doug Rabson

25 years agoPut radius on its proper ports.
jkh [Wed, 30 Jun 1999 06:11:14 +0000 (06:11 +0000)]
Put radius on its proper ports.

25 years agoChanged pcicmem and cardirq variables to non-dirty variables.
hosokawa [Wed, 30 Jun 1999 05:04:37 +0000 (05:04 +0000)]
Changed pcicmem and cardirq variables to non-dirty variables.

25 years agoSubmitted by: "David E. Cross" <crossd@cs.rpi.edu>
julian [Wed, 30 Jun 1999 04:29:13 +0000 (04:29 +0000)]
Submitted by: "David E. Cross" <crossd@cs.rpi.edu>
Matt missed a line..

25 years agopccard_beep parameter is not supported by rc.conf and rc.pccard of -current.
hosokawa [Wed, 30 Jun 1999 04:16:08 +0000 (04:16 +0000)]
pccard_beep parameter is not supported by rc.conf and rc.pccard of -current.
(I'll port it later...)

25 years agoAdded -z and -i option.
hosokawa [Wed, 30 Jun 1999 03:49:30 +0000 (03:49 +0000)]
Added -z and -i option.

25 years agoAn SMP-specific change: Remove unnecessary lock acquires and releases
alc [Wed, 30 Jun 1999 03:39:29 +0000 (03:39 +0000)]
An SMP-specific change: Remove unnecessary lock acquires and releases
surrounding critical sections that consist of (1) a single read or
(2) a single locked RMW operation.

(Thanks to thomma@slip.net (Tamiji Homma) for helping to test
these changes.)

25 years agoSubmitted by: Conrad Minshall <conrad@apple.com>
julian [Wed, 30 Jun 1999 02:53:51 +0000 (02:53 +0000)]
Submitted by: Conrad Minshall <conrad@apple.com>
Reviewed by: Matthew Dillon <dillon@apollo.backplane.com>

The following ugly hack to the exit path of nfs_readlinkrpc() circumvents
an Auspex bug: for symlinks longer than 112 (0x70) they return a 1024 byte
xdr string - the correct data with many nulls appended.  Without this fix
namei returns ENAMETOOLONG, at least it does on our source base and on
FreeBSD 3.0.  Note we do not (and should not) rely upon their null padding.

25 years agoFreeBSD does supports LKM now.
hosokawa [Tue, 29 Jun 1999 23:59:15 +0000 (23:59 +0000)]
FreeBSD does supports LKM now.
PR: doc/10331
Reviewed by: mpp@freebsd.org
Submitted by: Andreas Gustafsson <gson@araneus.fi>

25 years agoRemoved unused boot_crunch.conf
hosokawa [Tue, 29 Jun 1999 23:49:31 +0000 (23:49 +0000)]
Removed unused boot_crunch.conf

25 years agosscape_mss is supposed to work..
peter [Tue, 29 Jun 1999 21:53:59 +0000 (21:53 +0000)]
sscape_mss is supposed to work..

25 years agoDrop old-scsi drivers (was commented out) od0 and (not commented) sctarg0
peter [Tue, 29 Jun 1999 21:52:07 +0000 (21:52 +0000)]
Drop old-scsi drivers (was commented out) od0 and (not commented) sctarg0

25 years agoFix typo (wrong path)
peter [Tue, 29 Jun 1999 21:40:05 +0000 (21:40 +0000)]
Fix typo (wrong path)

25 years agoReconcile all the loader newvers.sh's into one common one.
peter [Tue, 29 Jun 1999 21:23:26 +0000 (21:23 +0000)]
Reconcile all the loader newvers.sh's into one common one.

25 years agoDon't use head/tail - they don't exist in a 'make world' environment.
peter [Tue, 29 Jun 1999 21:06:16 +0000 (21:06 +0000)]
Don't use head/tail - they don't exist in a 'make world' environment.
(Why do we have three copies of this script anyway?)

25 years agoDefer signals, so we will not wait for SIGCHLD after it was delivered.
dt [Tue, 29 Jun 1999 19:57:07 +0000 (19:57 +0000)]
Defer signals, so we will not wait for SIGCHLD after it was delivered.

25 years agoCorrect spelling of NMBCLUSTERS in a comment.
mph [Tue, 29 Jun 1999 19:06:16 +0000 (19:06 +0000)]
Correct spelling of NMBCLUSTERS in a comment.

Submitted by: Peter Radcliffe <pir@pir.net>

25 years agoWith asbestos suit on, make the options indenting a little more consistant
peter [Tue, 29 Jun 1999 18:58:27 +0000 (18:58 +0000)]
With asbestos suit on, make the options indenting a little more consistant
so that it doesn't screw up the alignment when commenting out an entry.
Also dequote two entries that do not need it.

25 years agoPut on my asbestos suit and attempt to tidy up and add some simple docs
peter [Tue, 29 Jun 1999 18:55:53 +0000 (18:55 +0000)]
Put on my asbestos suit and attempt to tidy up and add some simple docs
or notes to make it much more obvious what things are for people who
have not committed LINT to memory yet.

25 years agoDequote like on the i386 configs. Also remove some references to old
peter [Tue, 29 Jun 1999 18:24:09 +0000 (18:24 +0000)]
Dequote like on the i386 configs.  Also remove some references to old
i386 isa drivers that used to be order sensitive.  The probe order of
those drivers is now determined by a list in isa_compat.c and config
file order is totally irrelevant.

25 years agoidev->id_irq is an interrupt MASK, not an interrupt number.
yokota [Tue, 29 Jun 1999 17:37:44 +0000 (17:37 +0000)]
idev->id_irq is an interrupt MASK, not an interrupt number.
Thus, we need to convert the mask to the number (by ffs()) when
writing back this value to the resource in save_resource().

25 years agoKeyboard allocation/deallocation fix.
yokota [Tue, 29 Jun 1999 17:36:20 +0000 (17:36 +0000)]
Keyboard allocation/deallocation fix.

- Do not try to allocate a keyboard in pccnprobe() when probing the vt
  driver for the kernel console.  Rather, allocate a keyboard when
  initializing the vt driver in pccninit().
- Release the keyboard in pccnterm().
- Don't try to read from the keyboard, if it is not present.

25 years agoAllocate the port resource when attaching the keyboard controller,
yokota [Tue, 29 Jun 1999 17:35:09 +0000 (17:35 +0000)]
Allocate the port resource when attaching the keyboard controller,
rather than when the individual child device is attached.

25 years agoSkip the device if it is disabled, when searching for a serial port
yokota [Tue, 29 Jun 1999 17:34:16 +0000 (17:34 +0000)]
Skip the device if it is disabled, when searching for a serial port
to be used for the kernel console in sccnprobe().

25 years agoFill in tp->t_windowsize AFTER the call to (*linesw[tp->t_line].l_open)(),
yokota [Tue, 29 Jun 1999 17:30:33 +0000 (17:30 +0000)]
Fill in tp->t_windowsize AFTER the call to (*linesw[tp->t_line].l_open)(),
rather than BEFORE the call.  Otherwise the structure will be `zero'ed out
by l_open, which actually is ttyopen(), if !TS_OPEN.

PR: kern/12420

25 years agoStop signals being generated after meteor_close.
roger [Tue, 29 Jun 1999 16:45:51 +0000 (16:45 +0000)]
Stop signals being generated after meteor_close.
Update METEORSSIGNAL to disable signals by setting the signal to 0.

PR: i386/10533
Submitted by: Frode Vatvedt Fjeld <frodef@dslab7.cs.uit.no>

25 years agoClarify what happens if fd is set to -1
billf [Tue, 29 Jun 1999 16:32:22 +0000 (16:32 +0000)]
Clarify what happens if fd is set to -1

Submitted by: Alfred Perlstein <bright@rush.net>

25 years ago(mostly) fix ordering.
peter [Tue, 29 Jun 1999 16:14:20 +0000 (16:14 +0000)]
(mostly) fix ordering.

25 years agoNo longer need to set B_ASYNC flag since BUF_KERNPROC now
mckusick [Tue, 29 Jun 1999 15:57:40 +0000 (15:57 +0000)]
No longer need to set B_ASYNC flag since BUF_KERNPROC now
unconditionally sets the identity of the buffer.

25 years agoHopefully fix the remaining glitches with the BUF_*() changes. This should
peter [Tue, 29 Jun 1999 05:59:47 +0000 (05:59 +0000)]
Hopefully fix the remaining glitches with the BUF_*() changes.  This should
(really this time) fix pageout to swap and a couple of clustering cases.

This simplifies BUF_KERNPROC() so that it unconditionally reassigns the
lock owner rather than testing B_ASYNC and having the caller decide when
to do the reassign.  At present this is required because some places use
B_CALL/b_iodone to free the buffers without B_ASYNC being set.  Also,
vfs_cluster.c explicitly calls BUF_KERNPROC() when attaching the buffers
rather than the parent walking the cluster_head tailq.

Reviewed by: Kirk McKusick <mckusick@mckusick.com>

25 years agoLock buffer before calling strategy.
grog [Tue, 29 Jun 1999 04:10:05 +0000 (04:10 +0000)]
Lock buffer before calling strategy.

Idea-stolen-from: peter (sys/dev/ccd/ccd.c revision 1.49)

25 years agoclose_drive:
grog [Tue, 29 Jun 1999 04:08:51 +0000 (04:08 +0000)]
close_drive:
  Correct race condition between caller and daemon.

  Tripped-over-by: Zach Heilig <zach@uffdaonline.net>
Bernd Walter <ticso@cicely.de>
Niels Chr. Bank-Pedersen <ncbp@bank-pedersen.dk>

25 years agoCorrect type of intializer for (undocumented) cdevsw.d_parms.
grog [Tue, 29 Jun 1999 04:07:55 +0000 (04:07 +0000)]
Correct type of intializer for (undocumented) cdevsw.d_parms.

Submitted-by: peter

25 years agoMove call to umask(0) back into pw_util(), because the latter
pb [Tue, 29 Jun 1999 01:04:10 +0000 (01:04 +0000)]
Move call to umask(0) back into pw_util(), because the latter
function is also used by chpass(1) and passwd(1).

25 years agoSave common_tssd before it's loaded and the busy bit set.
luoqi [Mon, 28 Jun 1999 15:34:54 +0000 (15:34 +0000)]
Save common_tssd before it's loaded and the busy bit set.

Submitted by: bde

25 years agoFix a bug that was almost certainly making breadn() fail. BUF_KERNPROC()
peter [Mon, 28 Jun 1999 15:32:10 +0000 (15:32 +0000)]
Fix a bug that was almost certainly making breadn() fail.  BUF_KERNPROC()
was being called on the wrong bp - it should be called on the one that's
just about to be fed to VOP_STRATEGY().

25 years agoTypo: BUF_INITLOCK -> BUF_LOCKINIT and BUF_FREELOCK -> BUF_LOCKFREE.
kato [Mon, 28 Jun 1999 14:01:03 +0000 (14:01 +0000)]
Typo: BUF_INITLOCK -> BUF_LOCKINIT and BUF_FREELOCK -> BUF_LOCKFREE.

25 years agoUse the correct value for banksize so splash_pcx works in LFB modes.
des [Mon, 28 Jun 1999 13:52:29 +0000 (13:52 +0000)]
Use the correct value for banksize so splash_pcx works in LFB modes.

25 years agoSync with sys/i386/isa/clock.c revision 1.138.
kato [Mon, 28 Jun 1999 13:11:16 +0000 (13:11 +0000)]
Sync with sys/i386/isa/clock.c revision 1.138.

25 years agoSync with sys/i386/i386/userconfig.c revision 1.146.
kato [Mon, 28 Jun 1999 13:10:03 +0000 (13:10 +0000)]
Sync with sys/i386/i386/userconfig.c revision 1.146.

25 years agoSync with sys/i386/i386/machdep.c revision 1.344.
kato [Mon, 28 Jun 1999 13:08:59 +0000 (13:08 +0000)]
Sync with sys/i386/i386/machdep.c revision 1.344.

25 years agoSync with sys/i386/conf/Makefile.i386 revision 1.156.
kato [Mon, 28 Jun 1999 13:07:58 +0000 (13:07 +0000)]
Sync with sys/i386/conf/Makefile.i386 revision 1.156.

25 years agoFix a KASSERT() that was negated and lead to:
peter [Mon, 28 Jun 1999 12:34:40 +0000 (12:34 +0000)]
Fix a KASSERT() that was negated and lead to:
  nfs_strategy: buffer 0xxxxx not locked
when you attempted to write and had INVARIANTS turned on.

25 years agoFix page fault in visual userconfig's save code. (I only use normal
peter [Mon, 28 Jun 1999 12:07:36 +0000 (12:07 +0000)]
Fix page fault in visual userconfig's save code.  (I only use normal
userconfig, my original tweaks to visual mode were not well tested)

Submitted by: Peter Holm <peter@holm.cc>

25 years agoSync usage() with the manpage.
sheldonh [Mon, 28 Jun 1999 11:27:14 +0000 (11:27 +0000)]
Sync usage() with the manpage.

Approved by: mpp

25 years agoUpdate the SYNOPSIS to reflect that the -l option can be specified
mpp [Mon, 28 Jun 1999 10:50:47 +0000 (10:50 +0000)]
Update the SYNOPSIS to reflect that the -l option can be specified
more than once.

Pointed-out-by: sheldonh
25 years agomake va_fsid be of type udev_t
phk [Mon, 28 Jun 1999 10:35:07 +0000 (10:35 +0000)]
make va_fsid be of type udev_t

25 years agoAdd some extra alpha_pal_imb()'s in overkill mode. This makes this PC164SX
peter [Mon, 28 Jun 1999 09:38:09 +0000 (09:38 +0000)]
Add some extra alpha_pal_imb()'s in overkill mode.  This makes this PC164SX
run quite reliably now.  I've explicitly tagged them as /* XXX overkill? */
although one does actually check for VM_PROT_EXECUTE.

Based on a suggestion by:  Dmitrij Tejblum <tejblum@arc.hq.cti.ru>

25 years agoFix broken logic: (!wrap || log) -> (!wrap && log) .
sheldonh [Mon, 28 Jun 1999 09:28:17 +0000 (09:28 +0000)]
Fix broken logic: (!wrap || log) -> (!wrap && log) .

Reported by: David Malone <dwmalone@maths.tcd.ie>

25 years agoUse the same -UKERNEL strategy as the alpha to avoid the inlines etc.
peter [Mon, 28 Jun 1999 09:21:41 +0000 (09:21 +0000)]
Use the same -UKERNEL strategy as the alpha to avoid the inlines etc.

25 years agoDon't #include i386/isa/isa_dma.h - it's in isa/isavar.h now. This
peter [Mon, 28 Jun 1999 09:19:58 +0000 (09:19 +0000)]
Don't #include i386/isa/isa_dma.h - it's in isa/isavar.h now.  This
driver is probably not far from being MI now anyway.

25 years agoRevert back to not using -DKERNEL - it exposes inlines etc.
peter [Mon, 28 Jun 1999 09:18:44 +0000 (09:18 +0000)]
Revert back to not using -DKERNEL - it exposes inlines etc.
Zap symbols.raw and glue to make symbols.* - it's not used on the ELF-only
alpha kernel.  Symbol sorting is dead-end anyway once libkvm uses the
in-kernel linker symbol lookup.

25 years agoRevert back to not using -DKERNEL
peter [Mon, 28 Jun 1999 09:15:35 +0000 (09:15 +0000)]
Revert back to not using -DKERNEL

25 years agoMove struct prochd out of #ifdef KERNEL so the Alpha genassym can get
peter [Mon, 28 Jun 1999 09:14:18 +0000 (09:14 +0000)]
Move struct prochd out of #ifdef KERNEL so the Alpha genassym can get
at it.

25 years agoTweak include ordering so sys/systm.h is before sys/buf.h to keep buf.h's
peter [Mon, 28 Jun 1999 09:12:46 +0000 (09:12 +0000)]
Tweak include ordering so sys/systm.h is before sys/buf.h to keep buf.h's
inlines happy.

25 years agoDefine "OBSTACK_CHUNK_SIZE" as getpagesize().
obrien [Mon, 28 Jun 1999 09:05:56 +0000 (09:05 +0000)]
Define "OBSTACK_CHUNK_SIZE" as getpagesize().
Our malloc can allocte pagesized blocks efficiently and the EGCS default size
of 4072 bytes is not optimal.

Submitted by: Hidetoshi Shimokawa <simokawa@sat.t.u-tokyo.ac.jp>

25 years agoWhen requesting an exclusive lock with LK_NOWAIT, do not panic
mckusick [Mon, 28 Jun 1999 07:54:58 +0000 (07:54 +0000)]
When requesting an exclusive lock with LK_NOWAIT, do not panic
if LK_RECURSIVE is not set, as we will simply return that the
lock is busy and not actually deadlock. This allows processes
to use polling locks against buffers that they may already
hold exclusively locked.

25 years agoHmm, might as well make all of BEFORE_DEPEND first in case make depend
phk [Mon, 28 Jun 1999 07:19:51 +0000 (07:19 +0000)]
Hmm, might as well make all of BEFORE_DEPEND first in case make depend
wasn't run.

25 years agoAlso make pci_if.h early if make depend wasn't run.
phk [Mon, 28 Jun 1999 07:10:55 +0000 (07:10 +0000)]
Also make pci_if.h early if make depend wasn't run.

25 years agoMake bus_if.h and device_if.h as early as possible if make depend hasn't
phk [Mon, 28 Jun 1999 07:06:21 +0000 (07:06 +0000)]
Make bus_if.h and device_if.h as early as possible if make depend hasn't
been run.

25 years agoAmd now mounts to "/.amd_mnt" rather than "/net". "/net/<host>" now works
obrien [Mon, 28 Jun 1999 04:09:14 +0000 (04:09 +0000)]
Amd now mounts to "/.amd_mnt" rather than "/net".  "/net/<host>" now works
in addition to "/host/<host>".  This make us consistant with Sun (as we
already were with SGI).

25 years agoSyntax for user/group is changed from "user.group" to "user:group" to be
obrien [Mon, 28 Jun 1999 03:15:02 +0000 (03:15 +0000)]
Syntax for user/group is changed from "user.group" to "user:group" to be
consistant with chown(8).

25 years agoFix minor alignment problem.
obrien [Mon, 28 Jun 1999 03:03:17 +0000 (03:03 +0000)]
Fix minor alignment problem.

25 years agoNew ports/{java,irc,x11-servers} categories, Step #5 - update misc files.
billf [Mon, 28 Jun 1999 02:37:34 +0000 (02:37 +0000)]
New ports/{java,irc,x11-servers} categories, Step #5 - update misc files.

25 years agoAdd MultiTech Atlas modem
n_hibma [Sun, 27 Jun 1999 22:28:02 +0000 (22:28 +0000)]
Add MultiTech Atlas modem

25 years agoRemove superfluous semi-colon
n_hibma [Sun, 27 Jun 1999 22:24:20 +0000 (22:24 +0000)]
Remove superfluous semi-colon

25 years agoStandardize appearance of subsection headers (two spaces for section name)
jseger [Sun, 27 Jun 1999 22:12:35 +0000 (22:12 +0000)]
Standardize appearance of subsection headers (two spaces for section name)

25 years agoKirk missed a required BUF_KERNPROC(). Even though this is a non-async
peter [Sun, 27 Jun 1999 22:08:38 +0000 (22:08 +0000)]
Kirk missed a required BUF_KERNPROC().  Even though this is a non-async
transfer, the b_iodone hook causes biodone() to release it from interrupt
context.

25 years agoAn SMP-specific change: Remove an unnecessary lock acquire and release
alc [Sun, 27 Jun 1999 21:31:43 +0000 (21:31 +0000)]
An SMP-specific change: Remove an unnecessary lock acquire and release
from every system call.  (Storing a 32-bit constant is inherently
atomic.)

Reviewed by: Matthew Dillon <dillon@apollo.backplane.com>