]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years ago- Patch the 'skeleton' example driver to be more consistant with the new
nate [Sun, 26 Oct 1997 21:01:44 +0000 (21:01 +0000)]
- Patch the 'skeleton' example driver to be more consistant with the new
  state of the world.

26 years agoVFS interior redecoration.
phk [Sun, 26 Oct 1997 20:55:39 +0000 (20:55 +0000)]
VFS interior redecoration.

Rename vn_default_error to vop_defaultop all over the place.
Move vn_bwrite from vfs_bio.c to vfs_default.c and call it vop_stdbwrite.
Use vop_null instead of nullop.
Move vop_nopoll from vfs_subr.c to vfs_default.c
Move vop_sharedlock from vfs_subr.c to vfs_default.c
Move vop_nolock from vfs_subr.c to vfs_default.c
Move vop_nounlock from vfs_subr.c to vfs_default.c
Move vop_noislocked from vfs_subr.c to vfs_default.c
Use vop_ebadf instead of *_ebadf.
Add vop_defaultop for getpages on master vnode in MFS.

26 years agoRemade syscalls.master derived files.
phk [Sun, 26 Oct 1997 20:28:54 +0000 (20:28 +0000)]
Remade syscalls.master derived files.

26 years agoAdd "NOIMPL" for syscalls we know what is, but don't implement as "STD".
phk [Sun, 26 Oct 1997 20:27:51 +0000 (20:27 +0000)]
Add "NOIMPL" for syscalls we know what is, but don't implement as "STD".
Use this for getfh & nfssvc.

26 years agoSimplify the lease_check stuff.
phk [Sun, 26 Oct 1997 20:26:33 +0000 (20:26 +0000)]
Simplify the lease_check stuff.

26 years agoAlways initialize the syscall vectors for our "private" syscalls (not
phk [Sun, 26 Oct 1997 20:13:52 +0000 (20:13 +0000)]
Always initialize the syscall vectors for our "private" syscalls (not
just in the LKM case).
Plug nqnfs_vop_lease_check directly into the default_vnodeop_p table.

26 years agoIn clntudp_call(), it is possible that xdr_replymsg() might fail
wpaul [Sun, 26 Oct 1997 18:47:31 +0000 (18:47 +0000)]
In clntudp_call(), it is possible that xdr_replymsg() might fail
partway through its attempt to decode the result structure sent by
the server. If this happens, it can leave the result partially
populated with dynamically allocated memory. In this event, the
xdr_replymsg() failure is detected and RPC_CANTDECODERES is returned,
but the memory in the partially populated result struct is not
free()d.

The end result is that memory is leaked when an RPC_CANTDECODERES
error occurs. (This condition can occur if a CLIENT * handle is created
using clntudp_bufcreate() with a receive buffer size that is too small
to handle the result sent by the server.)

Fixed by setting reply_xdrs.x_op to XDR_FREE and calling
xdr_replymsg() again to free the memory if an RPC_CANTDECODERES error
is detected.

I suspect that the clnt_tcp.c, clnt_unix.c and clnt_raw.c modules
may ha a similar problem, but I haven't duplicated the condition with
those yet.

Found by: dbmalloc

26 years agoRemove yet another superfluous file.
jraynard [Sun, 26 Oct 1997 12:53:16 +0000 (12:53 +0000)]
Remove yet another superfluous file.

26 years agotypo.
wosch [Sun, 26 Oct 1997 12:51:30 +0000 (12:51 +0000)]
typo.

26 years agoUse croation week day name. This avoid a name clash
wosch [Sun, 26 Oct 1997 12:48:50 +0000 (12:48 +0000)]
Use croation week day name. This avoid a name clash
between `Subota' and `Sunday'.
         ^^           ^^

Submitted by: Slaven Rezic <eserte@cs.tu-berlin.de>

26 years agoCosmetic:
brian [Sun, 26 Oct 1997 12:42:13 +0000 (12:42 +0000)]
Cosmetic:
  Move prototypes into the correct headers.

26 years agoChange the initial amount of memory allocated for pv_entries to be proportional
dyson [Sun, 26 Oct 1997 12:33:39 +0000 (12:33 +0000)]
Change the initial amount of memory allocated for pv_entries to be proportional
to the amount of system memory.  Also, clean-up some of the new pv_entry
mgmt code.

26 years agoAdd -funsigned-char to CFLAGS (for ctype macros)
ache [Sun, 26 Oct 1997 12:20:45 +0000 (12:20 +0000)]
Add -funsigned-char to CFLAGS (for ctype macros)

26 years agoBack out mode (unsigned char)...
ache [Sun, 26 Oct 1997 12:17:33 +0000 (12:17 +0000)]
Back out mode (unsigned char)...

26 years agoBack out (unsigned char) cast, will use -funsigned-char instead
ache [Sun, 26 Oct 1997 12:14:54 +0000 (12:14 +0000)]
Back out (unsigned char) cast, will use -funsigned-char instead

26 years agoTypo.
helbig [Sun, 26 Oct 1997 10:37:35 +0000 (10:37 +0000)]
Typo.

26 years agoSynchronize usage() and man page.
helbig [Sun, 26 Oct 1997 10:33:02 +0000 (10:33 +0000)]
Synchronize usage() and man page.
Pointed out by bde.

26 years ago- Slightly change the way the border color register is updated so that
yokota [Sun, 26 Oct 1997 07:36:13 +0000 (07:36 +0000)]
- Slightly change the way the border color register is updated so that
flicker won't occur when set_border() is called.

- Properly restore the border color when switching virtual consoles.

Pointed out by: tony@dell.com
OKed by: sos

26 years ago- The daemon might go off the screen and crashed the system if the
yokota [Sun, 26 Oct 1997 07:35:18 +0000 (07:35 +0000)]
- The daemon might go off the screen and crashed the system if the
screen size was changed while the screen saver was inactive. Adjust
the positions of the daemon and the text and clip them accordingly
each time.

- Don't call set_border() too often. Some video chip may produce
flicker.

Pointed out by tony@dell.com

- Don't fill the entire screen with blank char every time the saver is
called.  Blank only the part of the screen where the daemon and the
text was previously printed.

26 years ago- 'Beep' support now happens in it's own separate file, so you can mess
nate [Sun, 26 Oct 1997 06:06:54 +0000 (06:06 +0000)]
- 'Beep' support now happens in it's own separate file, so you can mess
  around with different noises for the different events and not have it
  affect other files.

Inspired by: PAO

26 years agouse += on cflags
jmg [Sun, 26 Oct 1997 04:58:27 +0000 (04:58 +0000)]
use += on cflags

Submitted-by: Steve Price <sprice@hiwaay.net>

26 years ago- Functional changes to PCCARD support.
nate [Sun, 26 Oct 1997 04:54:16 +0000 (04:54 +0000)]
- Functional changes to PCCARD support.
 * Kill individual drivers 'suspend' routines, since there's no simple/safe
   way to suspend/resume a card w/out going through the complete probe
   at initialization time.
 * Default to using the apm_pccard_resume sysctl code, which basically
   pretends the card was removed, and then re-inserted.  Suspend/resume
   is now 'emulated' with a fake insert/removal.  (Hence we no longer
   need the driver-specific suspend routines.)

26 years ago- Do a bunch of gratuitous changes intended to make the code easier to
nate [Sun, 26 Oct 1997 04:36:24 +0000 (04:36 +0000)]
- Do a bunch of gratuitous changes intended to make the code easier to
  follow.
 * Rename/reorder all of the pccard structures, change many of the member
   names to be descriptive, and follow more closely other 'bus' drivers
   naming schemes.
 * Rename a bunch of parameter and local variable names to be more
   consistant in the code.
 * Renamed the PCCARD 'crd' device to be the 'card' device
 * KNF and make the code consistant where it was obvious.
 * ifdef'd out some unused code

26 years ago- Call 'callout_handle_init' on the two timeout channels after they are
nate [Sun, 26 Oct 1997 02:55:31 +0000 (02:55 +0000)]
- Call 'callout_handle_init' on the two timeout channels after they are
  created in the new slot controller structure.

26 years agoCosmetic (no functional changes):
brian [Sun, 26 Oct 1997 01:04:02 +0000 (01:04 +0000)]
Cosmetic (no functional changes):
o   Add missing $Id$s
o   Move extern decls from .c -> .h files
o   Staticize
o   Remove #includes from .h files
o   style(9)ify includes
o   bcopy -> memcpy
    bzero -> memset
    bcmp -> memcmp
    index -> strchr
    rindex -> strrchr
o   Move timeout.h -> timer.h (making it consistent w/ timer.c)
o   Add -Wmissing-prototypes

26 years agoChange L_SET to SEEK_SET for POSIX compliance.
jdp [Sun, 26 Oct 1997 00:41:51 +0000 (00:41 +0000)]
Change L_SET to SEEK_SET for POSIX compliance.

Submitted by: Dean Gaudet <dgaudet@arctic.org>

26 years ago- KNF the PCI stuff, no functional changes.
nate [Sat, 25 Oct 1997 21:02:44 +0000 (21:02 +0000)]
- KNF the PCI stuff, no functional changes.

26 years agoMerge changes from vendor branch.
wollman [Sat, 25 Oct 1997 18:27:25 +0000 (18:27 +0000)]
Merge changes from vendor branch.

26 years agoThis commit was generated by cvs2svn to compensate for changes in r30708,
wollman [Sat, 25 Oct 1997 18:22:53 +0000 (18:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r30708,
which included commits to RCS files with non-trunk default branches.

26 years agoNew version of the timezone database (1997i) from Arthur Olson.
wollman [Sat, 25 Oct 1997 18:22:53 +0000 (18:22 +0000)]
New version of the timezone database (1997i) from Arthur Olson.

Obtained from: Arthur David Olson <ado@elsie.nci.nih.gov>

26 years agoRecognize even more of the Znyx 314 cards.
phk [Sat, 25 Oct 1997 14:32:15 +0000 (14:32 +0000)]
Recognize even more of the Znyx 314 cards.

26 years agoFix typo in startup menu.
jkh [Sat, 25 Oct 1997 09:27:20 +0000 (09:27 +0000)]
Fix typo in startup menu.

Submitted by: max

26 years agoSomehow an error crept in during the previous commit.
dyson [Sat, 25 Oct 1997 04:49:01 +0000 (04:49 +0000)]
Somehow an error crept in during the previous commit.

26 years agoSupport garbage collecting the pmap pv entries. The management doesn't
dyson [Sat, 25 Oct 1997 02:41:56 +0000 (02:41 +0000)]
Support garbage collecting the pmap pv entries.  The management doesn't
happen until the system would have nearly failed anyway, so no signficant
overhead is added.  This helps large systems with lots of processes.

26 years agoDecrease the initial allocation for the zone allocations.
dyson [Fri, 24 Oct 1997 23:41:04 +0000 (23:41 +0000)]
Decrease the initial allocation for the zone allocations.

26 years agosleep => nointr_sleep
brian [Fri, 24 Oct 1997 22:36:31 +0000 (22:36 +0000)]
sleep => nointr_sleep
usleep => nointr_usleep
(not just a #define)
Already done by: ache

26 years ago(Null commit - the previous commit blew up in my face).
markm [Fri, 24 Oct 1997 16:35:34 +0000 (16:35 +0000)]
(Null commit - the previous commit blew up in my face).

Link this against -lcrypt. In the case where a user does not have
a ticket available on the server, this may cause chunks to be blown
when the local kinit does not find crypt(3).

26 years agoLink this against -lcrypt. In the case where the user has no key available
markm [Fri, 24 Oct 1997 16:27:46 +0000 (16:27 +0000)]
Link this against -lcrypt. In the case where the user has no key available
on the keyserver, the local kinit will blow chunks when it tries a
crypt(3) and finds it unavailable.

26 years agoMake this part of the 'krb' distribution if it is being built as the
markm [Fri, 24 Oct 1997 16:22:46 +0000 (16:22 +0000)]
Make this part of the 'krb' distribution if it is being built as the
Kerberised version.

26 years agoMake this part of the 'krb' distribution.
markm [Fri, 24 Oct 1997 16:20:25 +0000 (16:20 +0000)]
Make this part of the 'krb' distribution.

26 years ago- Recognize the Cirrus Logic PD6729/6730 PC-Card Controller behind the
nate [Fri, 24 Oct 1997 15:39:00 +0000 (15:39 +0000)]
- Recognize the Cirrus Logic PD6729/6730 PC-Card Controller behind the
  PCI bus.

[ Re-organized the code to be easier to extend, based on Michael's patch ]

Submitted by: Michael Reifenberger <root@totum.plaut.de>

26 years ago- Use the definitions found in <i386/isa/isa.h> instead of rolling our
nate [Fri, 24 Oct 1997 15:27:16 +0000 (15:27 +0000)]
- Use the definitions found in <i386/isa/isa.h> instead of rolling our
  own definition of ISA_HOLE_START.  We shouldn't need to include
  bus/processor specific code in here, but it is required.  At least by
  doing it this way it becomes more obvious where the bogusness is.

Obtained from: email with bde

26 years agoRemove the stuff we do not use from global scope. Export them again as
phk [Fri, 24 Oct 1997 14:27:23 +0000 (14:27 +0000)]
Remove the stuff we do not use from global scope.  Export them again as
needed.

26 years agoStatizice.
phk [Fri, 24 Oct 1997 14:24:55 +0000 (14:24 +0000)]
Statizice.

26 years agoPkg_add refused to process package whose packing list contains @exec/@unexec
max [Fri, 24 Oct 1997 08:32:06 +0000 (08:32 +0000)]
Pkg_add refused to process package whose packing list contains @exec/@unexec
before any file names.  This change makes pkg_add to process those
packages if @exec/@unexec commands don't contain any %[fFB].  Also
enable @exec/@unexec that have %D if the installation prefix is known to
pkg_add.
Reviewed by: jkh

26 years agofix mispelling of kld
jmg [Fri, 24 Oct 1997 05:30:12 +0000 (05:30 +0000)]
fix mispelling of kld

26 years agomake a couple functions static...
jmg [Fri, 24 Oct 1997 05:29:07 +0000 (05:29 +0000)]
make a couple functions static...

also change module_register_static to module_register_init as this
function initalizes the module for both dynamic and static modules...

26 years agoChange defaults to match reality.
hoek [Fri, 24 Oct 1997 04:13:41 +0000 (04:13 +0000)]
Change defaults to match reality.

PR: docs/3817

26 years ago- defined constant ISA_HOLE_START (would be nice to include machine/pmap.h,
nate [Thu, 23 Oct 1997 22:54:47 +0000 (22:54 +0000)]
- defined constant ISA_HOLE_START (would be nice to include machine/pmap.h,
  but it has too much baggage).
- create a new routine 'unregister_device_interrupt', which is now used
  instead of having two routines with the same code snippet.
- Minor cleanups and commenting.

[ No functional changes, just moving things around ]

26 years agoRestore back non-interruptable sleep/usleep just redefine them to not
ache [Thu, 23 Oct 1997 21:32:48 +0000 (21:32 +0000)]
Restore back non-interruptable sleep/usleep just redefine them to not
mix with standard library functions

26 years agoRemove private sleep/usleep hacks
ache [Thu, 23 Oct 1997 20:11:01 +0000 (20:11 +0000)]
Remove private sleep/usleep hacks

26 years ago- Move the untimeout code call out of the check for pwr_off_pending. It
nate [Thu, 23 Oct 1997 19:00:16 +0000 (19:00 +0000)]
- Move the untimeout code call out of the check for pwr_off_pending.  It
  shouldn't make any difference, but did on a couple of test systems.

26 years agoobey internet-draft "draft-stevens-advanced-api-04.txt" for ipv6 defs.
itojun [Thu, 23 Oct 1997 15:11:21 +0000 (15:11 +0000)]
obey internet-draft "draft-stevens-advanced-api-04.txt" for ipv6 defs.

26 years agoAdd entries for the Danpex EN-6200P2 and Megahertz XJ4336
msmith [Thu, 23 Oct 1997 14:00:40 +0000 (14:00 +0000)]
Add entries for the Danpex EN-6200P2 and Megahertz XJ4336

26 years agoFix a potentially disasterous '==' instead of '=' bug.
gibbs [Thu, 23 Oct 1997 11:32:06 +0000 (11:32 +0000)]
Fix a potentially disasterous '==' instead of '=' bug.

Submitted by: John-Mark Gurney <gurney_j@resnet.uoregon.edu>

26 years agoFix various filter problems
brian [Thu, 23 Oct 1997 10:09:35 +0000 (10:09 +0000)]
Fix various filter problems
PR: 4727
Submitted by: Chiharu Shibata <chi@bd.mbn.or.jp>

26 years agoDisallow non-root mount. If you want to allow non-root mount, change
kato [Thu, 23 Oct 1997 09:29:09 +0000 (09:29 +0000)]
Disallow non-root mount.  If you want to allow non-root mount, change
vfs.usermount into 1 with sysctl.

26 years agoSynchronize with sys/i386/isa/syscons.{c,h} revisions 1.235 and 1.35,
kato [Thu, 23 Oct 1997 09:26:30 +0000 (09:26 +0000)]
Synchronize with sys/i386/isa/syscons.{c,h} revisions 1.235 and 1.35,
respectively.

26 years agoSynchronize with sys/i386/conf/Makefile.i386 revision 1.103.
kato [Thu, 23 Oct 1997 09:20:16 +0000 (09:20 +0000)]
Synchronize with sys/i386/conf/Makefile.i386 revision 1.103.

26 years ago- Fix braino in last commit. (Slow networks should be outlawed!)
nate [Thu, 23 Oct 1997 04:53:32 +0000 (04:53 +0000)]
- Fix braino in last commit.  (Slow networks should be outlawed!)

26 years ago- Back out the last. APM_BROKEN_STATCLOCK doesn't exist in -current.
nate [Thu, 23 Oct 1997 04:18:49 +0000 (04:18 +0000)]
- Back out the last.  APM_BROKEN_STATCLOCK doesn't exist in -current.

26 years agoReject unreasonable values passed to CONS_HISTORY ioctl. It did not
yokota [Thu, 23 Oct 1997 03:23:50 +0000 (03:23 +0000)]
Reject unreasonable values passed to CONS_HISTORY ioctl. It did not
check the value and caused kernel panic when a large value was given.

- Move the configuration option SC_HISTORY_SIZE from syscons.h to
syscons.c.
- Define the maximum total number of history lines of all consoles.
It is SC_HISTORY_SIZE*MAXCONS or 1000*MAXCONS; whichever is larger.
CONS_HISTORY will allow the user to set the history size up to
SC_HISTORY_SIZE unconditionally (or the current height of the console
if it is larger than SC_HISTORY_SIZE). If the user requests a larger
buffer, it will be granted only if the total number of all allocated
history lines and the requested number of lines won't exceed the maximum.
- Don't free the previous history buffer and leave the history buffer
pointer holding a invalid pointer. Set the pointer to NULL first, then
free the buffer.

PR: bin/4592

26 years agoAdd (unsigned char) cast to ctype macros
ache [Thu, 23 Oct 1997 02:44:23 +0000 (02:44 +0000)]
Add (unsigned char) cast to ctype macros

26 years ago- When doing suspend/resume, only try to suspend those drivers on active
nate [Thu, 23 Oct 1997 02:30:39 +0000 (02:30 +0000)]
- When doing suspend/resume, only try to suspend those drivers on active
  slots.  Otherwise, we try to suspend drivers who have been disabled
  already.

[
The only reason the drivers are still on the list is because of race
conditions where the card is removed while the driver is in use.  We
leave the drivers on the slot list (leaving all of their structures in
place in case a process is using it) but set it's state to empty so that
further uses by the pccard code know not to expect active cards.
]

26 years agoBack our V1.36, due to reports of badness. (IRQ's are again allocated
nate [Thu, 23 Oct 1997 02:25:14 +0000 (02:25 +0000)]
Back our V1.36, due to reports of badness. (IRQ's are again allocated
from the 'lower' interrupts to the 'higher' interrupts.)  We need to
find a way to set the interrupt for the controller in the config file.

Determined by: handy@sag.space.lockheed.com

26 years agoAdd (unsigned char) cast to ctype macros
ache [Thu, 23 Oct 1997 02:22:51 +0000 (02:22 +0000)]
Add (unsigned char) cast to ctype macros

26 years agoAdd unsigned char cast to ctype macros
ache [Thu, 23 Oct 1997 02:08:16 +0000 (02:08 +0000)]
Add unsigned char cast to ctype macros

26 years agoAdd more (unsigned char) casts to ctype macros
ache [Thu, 23 Oct 1997 02:03:43 +0000 (02:03 +0000)]
Add more (unsigned char) casts to ctype macros
Fix casetable usage in the same manner too

26 years agoRewrote fxp_start() for better clarity and efficiency.
dg [Thu, 23 Oct 1997 01:45:15 +0000 (01:45 +0000)]
Rewrote fxp_start() for better clarity and efficiency.

26 years agoAdd (unsigned char) casts to ctype macros
ache [Thu, 23 Oct 1997 01:43:00 +0000 (01:43 +0000)]
Add (unsigned char) casts to ctype macros

26 years agoAdd (unsigned char) cast to all ctype macros
ache [Thu, 23 Oct 1997 01:29:44 +0000 (01:29 +0000)]
Add (unsigned char) cast to all ctype macros

26 years ago- Simplify code by removing a variable which was initialized once, never
nate [Thu, 23 Oct 1997 00:51:29 +0000 (00:51 +0000)]
- Simplify code by removing a variable which was initialized once, never
  modified, and was expected to have the initialized value.

26 years agoFix a few typos.
jraynard [Wed, 22 Oct 1997 23:12:27 +0000 (23:12 +0000)]
Fix a few typos.

26 years ago- If APM_BROKEN_STATCLOCK is defined, make sure the statcloock is
nate [Wed, 22 Oct 1997 23:04:07 +0000 (23:04 +0000)]
- If APM_BROKEN_STATCLOCK is defined, make sure the statcloock is
  disabled, don't rely on the flags to set it.

26 years agoBack out part of OpenGroup specs about limiting max arg since it may break
ache [Wed, 22 Oct 1997 12:04:49 +0000 (12:04 +0000)]
Back out part of OpenGroup specs about limiting max arg since it may break
compatibility.

26 years agoReflect usleep code changes:
ache [Wed, 22 Oct 1997 11:27:20 +0000 (11:27 +0000)]
Reflect usleep code changes:
Limit max arg
Change return type to int

26 years agoChange usleep return type from void to int to match OpenGroup specs.
ache [Wed, 22 Oct 1997 10:57:54 +0000 (10:57 +0000)]
Change usleep return type from void to int to match OpenGroup specs.

26 years agoChanges in spirit of OpenGroup Singe Unix specs:
ache [Wed, 22 Oct 1997 10:55:49 +0000 (10:55 +0000)]
Changes in spirit of OpenGroup Singe Unix specs:
1) Limit max allowed argument to 1000000
2) Change return type from void to int to indicate premature termination
(by signal)

26 years agoDocument EINVAL as a possible return value from open(2).
joerg [Wed, 22 Oct 1997 07:29:13 +0000 (07:29 +0000)]
Document EINVAL as a possible return value from open(2).

26 years agoReject attempts to call open() with an illegal combination of O_RDONLY,
joerg [Wed, 22 Oct 1997 07:28:51 +0000 (07:28 +0000)]
Reject attempts to call open() with an illegal combination of O_RDONLY,
O_WRONLY, O_RDWR.

26 years agoUse err(3). Add usage() and prototypes.
charnier [Wed, 22 Oct 1997 06:23:12 +0000 (06:23 +0000)]
Use err(3). Add usage() and prototypes.
Hide reference to trsp(8) that does not exist.

26 years agoUse err(3).
charnier [Wed, 22 Oct 1997 06:20:04 +0000 (06:20 +0000)]
Use err(3).
Sync man page and usage string.
Strcpy -> strncpy from OpenBSD.
-Wall cleaning.
Obtained from: OpenBSD

26 years agoFix HDLC logging (actually LogDumpBp()). It was only
brian [Wed, 22 Oct 1997 01:22:51 +0000 (01:22 +0000)]
Fix HDLC logging (actually LogDumpBp()).  It was only
outputting the first mbuf in the chain.

26 years agoPut in an initial %VERSREQ field. This will cause existing config(8)'s
peter [Wed, 22 Oct 1997 00:48:36 +0000 (00:48 +0000)]
Put in an initial %VERSREQ field.  This will cause existing config(8)'s
to give a non-fatal warning about unknown directives, so there is no
hurry to rebuild config(8) yet, apart from shutting up the warning.

26 years agoFor safety's sake, explicitly depend all objects on the configvers.h header
peter [Wed, 22 Oct 1997 00:44:08 +0000 (00:44 +0000)]
For safety's sake, explicitly depend all objects on the configvers.h header
to make sure that it's all recompiled even if there is no 'make depend'.
This is overkill, but should be one less thing that someone can do wrong.

26 years agoRecognize a %VERSREQ=nnnnn string in the system Makefile. Both config(8)
peter [Wed, 22 Oct 1997 00:38:48 +0000 (00:38 +0000)]
Recognize a %VERSREQ=nnnnn string in the system Makefile.  Both config(8)
and the kernel will have a 'config interface version number'.  If an
incompatable change is made to the kernel that requires a rebuild of
config(8) (such as the cam devtab stuff), then the version number would be
bumped in both places.  If a user neglects to rebuild config, then they
will get a nagging (but non-fatal) warning that they need to rebuild
config.

26 years agoFix the same leak as in nullfs. Now the lowervp is properly marked inactive.
roberto [Tue, 21 Oct 1997 21:08:17 +0000 (21:08 +0000)]
Fix the same leak as in nullfs. Now the lowervp is properly marked inactive.

Reviewed by: phk

26 years agoFix the file leak bug. The lower layer wasn't informed the vnode was inactive
roberto [Tue, 21 Oct 1997 21:01:34 +0000 (21:01 +0000)]
Fix the file leak bug. The lower layer wasn't informed the vnode was inactive
and kept a reference, preventing the blocks to be reclaimed.

Changed the comment in null_inactive to reflect the current situation.

Reviewed by: phk

26 years ago- Update the Adaptect SlimSCSI CIS tuple so that it works with FreeBSD's
nate [Tue, 21 Oct 1997 18:07:36 +0000 (18:07 +0000)]
- Update the Adaptect SlimSCSI CIS tuple so that it works with FreeBSD's
  default pccard daemon.

26 years ago- PCCARD support to the much maligned SCSI driver. This gives up support
nate [Tue, 21 Oct 1997 17:57:31 +0000 (17:57 +0000)]
- PCCARD support to the much maligned SCSI driver.  This gives up support
  for a couple of external CD's (notably the Sony PRD-650).

Note: In order to get my CD recognized, I had to configure the CD under
      Win95, but it seems to work now even if I turn it off.

Submitted by: PAO [minor mods by me]

26 years agoAdd const to a couple of casts to silence some of the warnings Bruce
phk [Tue, 21 Oct 1997 13:28:36 +0000 (13:28 +0000)]
Add const to a couple of casts to silence some of the warnings Bruce
has let loose on us.

26 years agoOops, comment lines can't be indented either.
bde [Tue, 21 Oct 1997 10:39:27 +0000 (10:39 +0000)]
Oops, comment lines can't be indented either.

26 years agoOops, comment lines can't be continued.
bde [Tue, 21 Oct 1997 10:36:07 +0000 (10:36 +0000)]
Oops, comment lines can't be continued.

26 years agoa few code style changes:
jmg [Tue, 21 Oct 1997 09:59:26 +0000 (09:59 +0000)]
a few code style changes:
    o start function names in column 1
    o sort order of flags in getopt and switch
    o don't try to reference progname
    o unspam some changes introduced by a 2.2.1-R build box instead of a
-current build box

doc changes:
    o document when these commands first appeared
    o put email address in angle brakets
    o minor mdoc clean up

26 years agoMerge from RELENG_2_2. Bandaid for 'lockmgr panic' on the VN device.
kato [Tue, 21 Oct 1997 09:51:47 +0000 (09:51 +0000)]
Merge from RELENG_2_2.  Bandaid for 'lockmgr panic' on the VN device.

26 years agoSynchronize with sys/i386/isa/fd.c revision 1.105.
kato [Tue, 21 Oct 1997 09:48:38 +0000 (09:48 +0000)]
Synchronize with sys/i386/isa/fd.c revision 1.105.

26 years agoSorted lists.
bde [Tue, 21 Oct 1997 08:41:15 +0000 (08:41 +0000)]
Sorted lists.

26 years agoReference the DMI table inside the SMBIOS table correctly, not using a variable
msmith [Tue, 21 Oct 1997 07:40:22 +0000 (07:40 +0000)]
Reference the DMI table inside the SMBIOS table correctly, not using a variable
that won't be initialised until a later test.
Submitted by: bde via -Wunused

26 years agoDon't install mount_msdos setuid root. Lite2's mount(2) handles
bde [Tue, 21 Oct 1997 07:26:50 +0000 (07:26 +0000)]
Don't install mount_msdos setuid root.  Lite2's mount(2) handles
permissions centrally and a setuid root mount utility just breaks
its security.  There was no new breakage in practice because
mfdosfs_mount() still checks the ruid.

26 years agoAdded -Wuninitialized to CWARNFLAGS. Warnings about uninitialialized
bde [Tue, 21 Oct 1997 06:53:24 +0000 (06:53 +0000)]
Added -Wuninitialized to CWARNFLAGS.  Warnings about uninitialialized
variables were lost when we removed -W, and 23 new ones including at
least one serious one have crept in for LINT.

Restored -Winline to CFLAGS.  This gives only 3 old warnings and 1 new
for LINT.