]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
27 years agoSend this to the Attic so there's no mixups over which kern_lock.c is in
Peter Wemm [Mon, 21 Apr 1997 13:39:56 +0000 (13:39 +0000)]
Send this to the Attic so there's no mixups over which kern_lock.c is in
use in -current.

27 years ago- Better fix for 3C589D support. Revert previous changes and add a
Nate Williams [Mon, 21 Apr 1997 13:38:32 +0000 (13:38 +0000)]
- Better fix for 3C589D support.  Revert previous changes and add a
  delay after we reset the card to allow the card to come back to life.
  It appears the newer card takes longer to reset.

Submitted by: Samuel Lam <skl@ScalableNetwork.com>

27 years agoReplace VOP_LOCK with vn_lock.
KATO Takenori [Mon, 21 Apr 1997 12:40:42 +0000 (12:40 +0000)]
Replace VOP_LOCK with vn_lock.

27 years agoSynchronize with sys/i386/isa/lpt.c revision 1.60.
KATO Takenori [Mon, 21 Apr 1997 12:33:26 +0000 (12:33 +0000)]
Synchronize with sys/i386/isa/lpt.c revision 1.60.

27 years agoSynchronize with sys/i386/isa/syscons.c.revision 1.210.
KATO Takenori [Mon, 21 Apr 1997 12:31:31 +0000 (12:31 +0000)]
Synchronize with sys/i386/isa/syscons.c.revision 1.210.

27 years agoAdd a reconnect capability directing ppp to re-establish
Brian Somers [Mon, 21 Apr 1997 01:02:02 +0000 (01:02 +0000)]
Add a reconnect capability directing ppp to re-establish
the connection after an unexpected loss of carrier:

    set reconnect timer ntries

The man page warns against using this command when your
timeout value is slightly more than the other sides :{}

Suggested by: burton@bsampley.vip.best.com (Burton Sampley)

27 years agoFix markup bogons.
John Fieber [Mon, 21 Apr 1997 00:42:54 +0000 (00:42 +0000)]
Fix markup bogons.

27 years agoMerge from 2.2.
Satoshi Asami [Mon, 21 Apr 1997 00:24:51 +0000 (00:24 +0000)]
Merge from 2.2.

27 years agoClean-up my modification of popen.c for vfork. Bruce's (this) is better.
John Dyson [Sun, 20 Apr 1997 20:17:04 +0000 (20:17 +0000)]
Clean-up my modification of popen.c for vfork.  Bruce's (this) is better.
Submitted by: Bruce Evans <bde@freebsd.org>

27 years agoBring over doc fixes.
Jordan K. Hubbard [Sun, 20 Apr 1997 19:39:23 +0000 (19:39 +0000)]
Bring over doc fixes.

27 years agoYAMF22
Jordan K. Hubbard [Sun, 20 Apr 1997 19:19:07 +0000 (19:19 +0000)]
YAMF22

27 years agoFixed the type of a timeout function and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 18:02:40 +0000 (18:02 +0000)]
Fixed the type of a timeout function and removed casts that hid the
type mismatches.  There was no problem in practice (at least on 386's).
The timeout args still get bogusly cast from int to `void *' via
caddr_t and back to int.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 17:26:55 +0000 (17:26 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  mcd and scd were/are particularly bogus.  They
used a general purpose function taking 2 args for the timeout
function and fudged varargs stuff to supply the second arg for the
timeout case.  This broke `cc -mrtd'.  Bounce through a timeout
function instead.  The timeout arg still gets bogusly cast from
int to `void *' and back.

27 years agoFix both a problem with accessing backing objects, and also release
John Dyson [Sun, 20 Apr 1997 17:12:11 +0000 (17:12 +0000)]
Fix both a problem with accessing backing objects, and also release
the process map on nonexistant pages.
PR: kern/3327
Submitted by: Tor Egge <Tor.Egge@idi.ntnu.no>

27 years agoRe-institute the efficent version of vfork. It appears to make a
John Dyson [Sun, 20 Apr 1997 16:57:12 +0000 (16:57 +0000)]
Re-institute the efficent version of vfork.  It appears to make a
difference of approx 3mins in make world on my P6!!!  This means
that vfork now has full address space sharing, so beware with
sloppy vfork programming.  Also, you really do need to apply
the previously committed popen fix in libc.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 16:54:58 +0000 (16:54 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  Not taking an arg in sequencer_timer() broke
`cc  -mrtd'.

27 years agoSync with 2.2 and clean up some tortured english that was added.
Jordan K. Hubbard [Sun, 20 Apr 1997 16:46:36 +0000 (16:46 +0000)]
Sync with 2.2 and clean up some tortured english that was added.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 16:38:22 +0000 (16:38 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  There was no problem in practice (at least on 386's).

Don't cast NULL in calls to timeout functions.  pcvt is fully prototyped
and doesn't support K&R.

Timeout functions take args of type `void *', so use magic numbers of
type `void *' for UPDATE_* to reduce the danger of wrong conversions.

Removed FreeBSD-pre-1.1-related TIMEOUT_FUNC_T macro.  It was especially
bogus for the pre-1.1 case.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 16:05:33 +0000 (16:05 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  Not taking an arg in scrn_timer() broke `cc -mrtd'.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 15:48:17 +0000 (15:48 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  There was no problem in practice (at least on 386's).

Removed NetBSD-related TIMEOUT macro.  NetBSD uses the same BSD4.4Lite
timeout interface as FreeBSD.  As a concession to portability, declare
the timeout function without using the FreeBSD timeout_t typedef.

27 years agoFixed the type of timeout functions and removed casts that hid the
Bruce Evans [Sun, 20 Apr 1997 15:36:12 +0000 (15:36 +0000)]
Fixed the type of timeout functions and removed casts that hid the
type mismatches.  There was no problem in practice (at least on 386's).

27 years agoFixed inconsistent extern declaration of a timeout function and
Bruce Evans [Sun, 20 Apr 1997 15:25:33 +0000 (15:25 +0000)]
Fixed inconsistent extern declaration of a timeout function and
removed a cast that helped hide the problem.  There was no problem
in practice (at least on 386's).

27 years agoUse separate routines for memmove() and memcpy(). This lets me drop
Bruce Evans [Sun, 20 Apr 1997 13:02:48 +0000 (13:02 +0000)]
Use separate routines for memmove() and memcpy().  This lets me drop
the (buggy) support for alternative entry points.  ALTENTRY() was only
used for memmove().  Optimizing for space was particularly silly because
memcpy() is rarely used (gcc normally inlines it).
Obtained from: NetBSD

27 years agoUse separate routines for memmove() and memcpy(). This lets me drop
Bruce Evans [Sun, 20 Apr 1997 12:46:12 +0000 (12:46 +0000)]
Use separate routines for memmove() and memcpy().  This lets me drop
the (buggy) support for alternative entry points.  ALTENTRY() was only
used for memmove().  Optimizing for space was particularly silly because
memcpy() is rarely used (gcc normally inlines it).

Obtained from: NetBSD

27 years agoMerged with the 1996/11/12 NetBSD version:
Bruce Evans [Sun, 20 Apr 1997 12:09:18 +0000 (12:09 +0000)]
Merged with the 1996/11/12 NetBSD version:
- use a slightly less bogus copyright.  This file was never contributed
  to Berkeley.  It still claims to be copright by the Regents.
- use <machine/asm.h> instead of "DEFS.h".
- use RCSID($Id$) instead of explicit assembly code and messy ifdefs.
  The rcsid won't be put into the object file until we make RCSID()
  non-null.  NetBSD uses a LIBC_SCCS ifdef here.  We used a LIBC_RCS
  instead, but I want RCSID() to be controlled directly by LIBC_RCS
  (actually by LIB_RCS).  This is the only difference with the NetBSD
  version.
- added ifdefs to support generation of memcpy() and memmove().  The
  other changes are "while I'm here" to get this.
- improved style of the copy backwards case.

27 years agoUpdating FAQ entry for X Inside.
Peter da Silva [Sun, 20 Apr 1997 11:53:53 +0000 (11:53 +0000)]
Updating FAQ entry for X Inside.
Added FAQ entries for tape changers and login names.
Fixed <code> tags.

PR: docs/3324
PR: docs/3326

27 years agoFixed linting of and compilation of the gcc byte-swapping macros
Bruce Evans [Sun, 20 Apr 1997 11:44:32 +0000 (11:44 +0000)]
Fixed linting of and compilation of the gcc byte-swapping macros
properly.  Simply don't use the gcc macros if we're not gcc, and
declare prototypes for the byte-swapping functions in case the
macro versions are not used.  The previous fix was wrong and broke
libpcap, which abuses -Dlint.

Don't pollute the namespace if _POSIX_SOURCE is defined.  This is
broken - it makes almost everything in <machine/endian.h> inaccessible
if _POSIX_SOURCE is defined, yet <machine/endian.h> is not a POSIX
header.  Other systems don't do it any better.

Removed always-false `BYTE_ORDER == BIG_ENDIAN' ifdef.
Obtained from: partly from Lite(2?) and NetBSD

27 years ago1. i82365.h: Chip identifiers should not be #ifdef PC98. Add identifier
Poul-Henning Kamp [Sun, 20 Apr 1997 07:21:12 +0000 (07:21 +0000)]
1. i82365.h: Chip identifiers should not be #ifdef PC98.  Add identifier
    (pt. unused) for TI PCI1130.

2. pccard.c: PCIC_RESUME_RESET is now (also) a sysctl.  (Never make it
    a #ifdef if it can be made a sysctl!)

3. pcic.c: make getb() and putb() proper member functions of struct
    pcic_slot.  Add a couple of missing casts.

27 years agoThe bit of the cardbus bridge support code se and I have been able to agree
Poul-Henning Kamp [Sun, 20 Apr 1997 06:57:43 +0000 (06:57 +0000)]
The bit of the cardbus bridge support code se and I have been able to agree
on so far... :-)
Reviewed by: se

27 years agoFix up the "hlt vector" change I made.
Poul-Henning Kamp [Sun, 20 Apr 1997 06:41:26 +0000 (06:41 +0000)]
Fix up the "hlt vector" change I made.
Reviewed by: bde, bde, bde

27 years agoSynchronize with sys/i386/isa/diskslice_machdep.c revision 1.25.
KATO Takenori [Sun, 20 Apr 1997 05:16:06 +0000 (05:16 +0000)]
Synchronize with sys/i386/isa/diskslice_machdep.c revision 1.25.

27 years agoTweak column headers a tad for when running in SMP mode.
Peter Wemm [Sat, 19 Apr 1997 20:28:50 +0000 (20:28 +0000)]
Tweak column headers a tad for when running in SMP mode.

27 years agoSync with original source.
Wolfram Schneider [Sat, 19 Apr 1997 20:25:36 +0000 (20:25 +0000)]
Sync with original source.

BTW, the BSD family tree has been reviewed by:
John S. Quarterman, Keith Bostic, Kirk McKusick, Peter H. Salus.

27 years agoUnderstand `--' to end processing of command options.
Wolfram Schneider [Sat, 19 Apr 1997 20:06:57 +0000 (20:06 +0000)]
Understand `--' to end processing of command options.
This means one search for a string starting with a dash.
Fix also filenames starting with a dash. Close PR #3349

27 years agoFixed long lines.
Bruce Evans [Sat, 19 Apr 1997 17:05:30 +0000 (17:05 +0000)]
Fixed long lines.

Removed unused macros CALL() and ASMSTR.

Reviewed by: jdp

27 years agoFixed #include and/or prototype bugs in synopsis.
Bruce Evans [Sat, 19 Apr 1997 15:57:20 +0000 (15:57 +0000)]
Fixed #include and/or prototype bugs in synopsis.

27 years agoUpdated comment about MANDEPEND (it was misplaced in bsd.prog.mk).
Bruce Evans [Sat, 19 Apr 1997 14:55:17 +0000 (14:55 +0000)]
Updated comment about MANDEPEND (it was misplaced in bsd.prog.mk).

27 years agoSynchronize with sys/i386/isa/sio.c revision 1.163.
KATO Takenori [Sat, 19 Apr 1997 14:54:32 +0000 (14:54 +0000)]
Synchronize with sys/i386/isa/sio.c revision 1.163.

27 years agoRemoved unused (and almost unusable) support for LDONLY. See ld/rtld
Bruce Evans [Sat, 19 Apr 1997 14:36:49 +0000 (14:36 +0000)]
Removed unused (and almost unusable) support for LDONLY.  See ld/rtld
for how to use gcc to invoke the linker only.

Removed `.if 0'ed support for not creating intermediate object files.

Removed unused include of bsd.libnames.mk (LDONLY had the last reference
to it).

Removed stale misplaced comment about MANDEPEND.

27 years agoFix a bug where a program which appended many small records to a file could
Doug Rabson [Sat, 19 Apr 1997 14:28:36 +0000 (14:28 +0000)]
Fix a bug where a program which appended many small records to a file could
wind up writing zeros instead of real data when the file is on an NFSv2
mounted directory.

While tracking this bug down, I noticed that nfs_asyncio was waking *all*
the iods when a block was written instead of just one per block.  Fixing this
gives a 25% performance improvment for writes on v2 (less for v3).

Both are 2.2 candidates.

PR: kern/2774

27 years agoAvoid division by 0 in check_part(). (It occurred when max_nsectors == 0.
Bruce Evans [Sat, 19 Apr 1997 14:14:17 +0000 (14:14 +0000)]
Avoid division by 0 in check_part().  (It occurred when max_nsectors == 0.
This case is clearly an error, but we keep calling check_part() to get
diagnostics.)

Fixed nearby indentation and commenting bugs.

27 years agoFix typo in previous commit
Andrey A. Chernov [Sat, 19 Apr 1997 11:31:38 +0000 (11:31 +0000)]
Fix typo in previous commit

27 years agoPoint people at ftp://ftp.FreeBSD.ORG/pub/FreeBSD/CTM for the deltas
Gary Palmer [Sat, 19 Apr 1997 10:40:45 +0000 (10:40 +0000)]
Point people at ftp://ftp.FreeBSD.ORG/pub/FreeBSD/CTM for the deltas
instead of ftp://freefall.cdrom.com/pub/CTM (which it hasn't been
for ages)

27 years agoUnceremoneously discard HDLC frames with less than two bytes in them.
Poul-Henning Kamp [Sat, 19 Apr 1997 09:25:58 +0000 (09:25 +0000)]
Unceremoneously discard HDLC frames with less than two bytes in them.
XXX: real fix: this could should check the number of bytes in the packet
before accessing them.

Found by: phkmalloc

27 years agoFix make depend.
Doug Rabson [Sat, 19 Apr 1997 09:15:25 +0000 (09:15 +0000)]
Fix make depend.

27 years agoAvoid `lock against myself' panic by following operation:
KATO Takenori [Sat, 19 Apr 1997 06:04:13 +0000 (06:04 +0000)]
Avoid `lock against myself' panic by following operation:

# mount -t union (or null) dir1 dir2
# mount -t union (or null) dir2 dir1

The function namei in union_mount calls union_root.  The upper vnode
has been already locked and vn_lock in union_root causes above panic.

Add printf's included in `#ifdef DIAGNOSTIC' for EDEADLK cases.

27 years agoDon't disable CPU cache in init_486dlc. If BIOS supports Cyrix 486,
KATO Takenori [Sat, 19 Apr 1997 05:25:19 +0000 (05:25 +0000)]
Don't disable CPU cache in init_486dlc.  If BIOS supports Cyrix 486,
BIOS enables CPU cache and other registers.  If BIOS does not supports
it, CPU cache is disabled at reset time.

This commit closes PR/3292.

PR: 3292

27 years agoEnable GLOBAL tags support.
Jordan K. Hubbard [Fri, 18 Apr 1997 23:44:54 +0000 (23:44 +0000)]
Enable GLOBAL tags support.

27 years agoEnable GLOBAL tags support for nvi.
Jordan K. Hubbard [Fri, 18 Apr 1997 23:36:52 +0000 (23:36 +0000)]
Enable GLOBAL tags support for nvi.
Submitted-By: Shigio Yamaguchi <shigio@wafu.netgate.net>
27 years agoDon't call timeout() for DTR wakeup or busy check if the relevant timeout
Bruce Evans [Fri, 18 Apr 1997 18:28:09 +0000 (18:28 +0000)]
Don't call timeout() for DTR wakeup or busy check if the relevant timeout
is already active.  This fixes "timeout table full" panics under 100%
interrupt load and other weird conditions.

27 years agoNo longer use AAP for queueing SCBs to the QINFIFO.
Justin T. Gibbs [Fri, 18 Apr 1997 16:34:36 +0000 (16:34 +0000)]
No longer use AAP for queueing SCBs to the QINFIFO.

Clean up the unexpected busfree handler.  We now look directly at the
message that might have caused the bus free to occur instead of looking
at an SCB flag.  This makes the handling more robust and also allows for
recovery actions that might cause an "unexpected busfree" to be performed
even if an SCB is not availible to "tag".  Most notably, this happens
when we don't find an SCB for a reconnecting target.

27 years agoRevert some changes to the selection and reselection code that were thought
Justin T. Gibbs [Fri, 18 Apr 1997 16:31:55 +0000 (16:31 +0000)]
Revert some changes to the selection and reselection code that were thought
to fix a selection timeout problem.

If we can't find an SCB for the reconnecting target, issue a bus device
reset as the SCSI2 spec suggests.

Add a missing call to "add_scb_to_free_list" in the non paging case.  In
the non-paging case, the SCBs don't really need to be on the free list,
but putting them there clears the tag field which is something the recovery
code depends on.

27 years agoUse v3 protocol by default if it is supported by the server. Allow the
Doug Rabson [Fri, 18 Apr 1997 16:23:10 +0000 (16:23 +0000)]
Use v3 protocol by default if it is supported by the server.  Allow the
user to force v2 protocol even if the server supports v3.

Obtained from: NetBSD but with a slightly different implementation

27 years agoDon't allow partial buffers to be cluster-comitted.
Doug Rabson [Fri, 18 Apr 1997 14:12:17 +0000 (14:12 +0000)]
Don't allow partial buffers to be cluster-comitted.
Zero the b_dirty{off,end} after cluster-comitting a group of buffers.

With these fixes, I was able to complete a 'make world' with remote src
and obj directories.

27 years agoGenerate mount*.[ch] and nfs_prot*.[ch] from the .x files.
Doug Rabson [Fri, 18 Apr 1997 13:23:38 +0000 (13:23 +0000)]
Generate mount*.[ch] and nfs_prot*.[ch] from the .x files.

Delete bogus local versions of mount*.[ch] and nfs_prot*.[ch].

Use v3 protocol by default for NFS mounts.  If v3 is not supported, v2 is
used automatically.  Add a new mount options for NFS, 'nfsv2' to allow the
administrator to force the use of the older protocol.

Document the new mount options and fix a couple of markup problems.

27 years agoAdd protocol definitions for NFSv3. Use -DWANT_NFS3 with rpcgen to get
Doug Rabson [Fri, 18 Apr 1997 12:31:27 +0000 (12:31 +0000)]
Add protocol definitions for NFSv3.  Use -DWANT_NFS3 with rpcgen to get
the new protocol.

Obtained from: rfc1813

27 years agoIncorporate kludge/fix from `syscons' ver 1.207 and 1.209.
Kazutaka YOKOTA [Fri, 18 Apr 1997 12:06:34 +0000 (12:06 +0000)]
Incorporate kludge/fix from `syscons' ver 1.207 and 1.209.

1) Dell Latitude XPi
This laptop has a strange, IMHO broken :-), keyboard controller which
wouldn't disable the keyboard interrupt. The kludge is to disable tty
intr. during set_keyboard(), used for changing LED and setting
typematic.

The patch also changes the function name:
    set_keyboard() -> set_keyboard_param()
Although it is a static function, the name corrides with a routine in
`syscons' and is confusing when debugging the kernel which has both
`syscons' and `pcvt' with DDB. (Suggested by Bruce)

2) doreset() bug
doreset() failed to preserve some bits in the keyboard controller's
command byte during keyboard reset. This bug may put some keyboard
controllers in old motherboards (386 and 486) in a strange state,
resulting in complete keyboard lockup or random key input.

Reviewed by: Joerg

27 years agofix up this man page
John-Mark Gurney [Fri, 18 Apr 1997 11:12:52 +0000 (11:12 +0000)]
fix up this man page

one typo was:
Submitted by: Josh Gilliam
PR#3319

27 years agoRearrange pci network card identification for easier additions.
David Nugent [Fri, 18 Apr 1997 04:44:37 +0000 (04:44 +0000)]
Rearrange pci network card identification for easier additions.
Added KTI NE2000 clone.

27 years agoBrought fix from the 2.2 branch forward (see rev 1.47.2.7): serious bugs
David Greenman [Fri, 18 Apr 1997 02:43:05 +0000 (02:43 +0000)]
Brought fix from the 2.2 branch forward (see rev 1.47.2.7): serious bugs
with reading the image header.

27 years agoFix punctuation: "it's" -> "its"
Daniel O'Callaghan [Thu, 17 Apr 1997 23:31:47 +0000 (23:31 +0000)]
Fix punctuation: "it's" -> "its"

27 years agoAmancio:
Steve Passe [Thu, 17 Apr 1997 22:33:16 +0000 (22:33 +0000)]
Amancio:
This patch fixes the problem of vic only capturing an even or odd frame plus
the my early patch for missing frames with resolutions higher than 320x240
in rgb mode.

The yuv422 patch introduces a minor bug in that a green line appears at the
bottom of the captured window . There is no easy work around for this right
now.

Reviewed by: various bt848 hackers
Submitted by: Amancio Hasty <hasty@rah.star-gate.com>

27 years agoDowngrade Slaven Rezic's name from latin2 to latin1
Wolfram Schneider [Thu, 17 Apr 1997 16:26:54 +0000 (16:26 +0000)]
Downgrade Slaven Rezic's name from latin2 to latin1

27 years agoQuick and dirty attempt to get the zp device to work with the 3c589D.
Nate Williams [Thu, 17 Apr 1997 14:33:11 +0000 (14:33 +0000)]
Quick and dirty attempt to get the zp device to work with the 3c589D.
Increase the delay in read_eeprom_data() by two orders of magnitude.

> A better fix would be to make read_eeprom_data() call
> f_is_eeprom_busy() after the DELAY().

Submitted by: Samuel Lam <skl@ScalableNetwork.com>

27 years agoFix `locking against myself' panic by multi nullfs mount of same
KATO Takenori [Thu, 17 Apr 1997 11:24:57 +0000 (11:24 +0000)]
Fix `locking against myself' panic by multi nullfs mount of same
directory pair.

27 years agoUse NULLVP instead of NULL.
KATO Takenori [Thu, 17 Apr 1997 11:17:30 +0000 (11:17 +0000)]
Use NULLVP instead of NULL.

27 years agoAdd Sean Eric Fagan and Chuck Robey for their kind monetary donations
Jordan K. Hubbard [Wed, 16 Apr 1997 17:55:22 +0000 (17:55 +0000)]
Add Sean Eric Fagan and Chuck Robey for their kind monetary donations
to the project.

27 years agoDo not set the uppervp to NULLVP in union_removed_upper. If lowervp
KATO Takenori [Wed, 16 Apr 1997 16:24:24 +0000 (16:24 +0000)]
Do not set the uppervp to NULLVP in union_removed_upper.  If lowervp
is NULLVP, union node will have neither uppervp nor lowervp.  This
causes page fault trap.

The union_removed_upper just remove union node from cache and it
doesn't set uppervp to NULLVP.  Since union node is removed from
cache, it will not be referenced.

The code that remove union node from cache was copied from
union_inactive.

27 years agoWhen "-c" is specified, don't pad the command with spaces if it is
John Polstra [Wed, 16 Apr 1997 16:08:11 +0000 (16:08 +0000)]
When "-c" is specified, don't pad the command with spaces if it is
the last field on the line.  "ps -axlc" was needlessly wrapping
around on 80-character windows.

27 years agoComment out rawcb, it is not used / not present anymore
Andrey A. Chernov [Wed, 16 Apr 1997 15:09:37 +0000 (15:09 +0000)]
Comment out rawcb, it is not used / not present anymore

27 years agoUse reset port before clearing page table in cpu_reset if PC98 is
KATO Takenori [Wed, 16 Apr 1997 12:11:37 +0000 (12:11 +0000)]
Use reset port before clearing page table in cpu_reset if PC98 is
defined.  Clearing page table could hang some new PC-98.

27 years agoFixed `make depend' and related bogons. LDFLAGS was used for
Bruce Evans [Wed, 16 Apr 1997 11:31:32 +0000 (11:31 +0000)]
Fixed `make depend' and related bogons.  LDFLAGS was used for
ld-specific flags.  LDFLAGS is really for ld-related flags for cc,
not for ld, and some flags, e.g., -Bshareable, mean completely
different things to cc and ld.  Having the wrong things in LDFLAGS
also broke the standard ${PROG} target.  This was kludged around
by using a special rule that depended on LDFLAGS being bogus.
Fixing `make depend' broke the special rule but fixed the standard
rule (except in the DESTDIR case, which was handled more strictly
here than elsewhere).

27 years agoFix typo: "oder" -> "order".
Daniel O'Callaghan [Wed, 16 Apr 1997 09:56:30 +0000 (09:56 +0000)]
Fix typo: "oder" -> "order".

27 years agoAdd FILES section to manpage.
Daniel O'Callaghan [Wed, 16 Apr 1997 09:54:50 +0000 (09:54 +0000)]
Add FILES section to manpage.
Create /var/run/ttyXn.if interface file and /var/run/slX.pid
PID file, following format of pppd(8).

27 years agoFix typo in comment; Document in an example that years beyond 1999
Daniel O'Callaghan [Wed, 16 Apr 1997 05:59:21 +0000 (05:59 +0000)]
Fix typo in comment; Document in an example that years beyond 1999
are handled; fix typo in same example.

27 years agoFix the problem in popen that makes correct vfork semantics fail.
John Dyson [Wed, 16 Apr 1997 03:26:50 +0000 (03:26 +0000)]
Fix the problem in popen that makes correct vfork semantics fail.
Specifically, popen modifies a variable "pdes[1]" in the child
in such a way that it breaks code in the parent (due to the address
space sharing.)

27 years agoUndo previous commit to avoid panic, and fix order of argument of
KATO Takenori [Wed, 16 Apr 1997 03:08:34 +0000 (03:08 +0000)]
Undo previous commit to avoid panic, and fix order of argument of
VOP_LINK().  The reason of strange behavior was wrong order of the
argument, that is, the operation

# ln foo bar

in a union fs tried to do

# ln bar foo

in ufs layer.

Now we can make a link in a union fs.

27 years agoSuggested by: David Nugent
Daniel O'Callaghan [Wed, 16 Apr 1997 02:10:22 +0000 (02:10 +0000)]
Suggested by: David Nugent
Change the name of ppp.disabled to ppp.deny for consistency with
at.deny and friends.

27 years agoSuggested by: David Nugent
Daniel O'Callaghan [Wed, 16 Apr 1997 02:07:10 +0000 (02:07 +0000)]
Suggested by: David Nugent
Use /etc/ppp/ppp.deny instead of /etc/ppp/ppp.disabled

27 years agoAdd include of <sys/types.h> no longer included by <stdio.h>.
David Nugent [Wed, 16 Apr 1997 00:05:29 +0000 (00:05 +0000)]
Add include of <sys/types.h> no longer included by <stdio.h>.

27 years agoFixed type mismatch caused by bogus prototypes. rpcgen for some reason
Bruce Evans [Tue, 15 Apr 1997 16:18:57 +0000 (16:18 +0000)]
Fixed type mismatch caused by bogus prototypes.  rpcgen for some reason
doesn't generate any prototypes for the functions to be registered.

27 years agoSecond stage of moving this to in <machine/asm.h>: just include
Bruce Evans [Tue, 15 Apr 1997 14:18:17 +0000 (14:18 +0000)]
Second stage of moving this to in <machine/asm.h>: just include
<machine/asm.h> here.

27 years agoOnly do indirections in ENTRY() if _ARCH_INDIRECT is defined.
Bruce Evans [Tue, 15 Apr 1997 14:06:34 +0000 (14:06 +0000)]
Only do indirections in ENTRY() if _ARCH_INDIRECT is defined.

27 years agoAdded -D_ARCH_INDIRECT=i387_ to CFLAGS. _ARCH_INDIRECT will soon be used
Bruce Evans [Tue, 15 Apr 1997 14:05:28 +0000 (14:05 +0000)]
Added -D_ARCH_INDIRECT=i387_ to CFLAGS.  _ARCH_INDIRECT will soon be used
to control generation of indirections in ENTRY().  Only msun needs it.

Use ${ARCH} consistently.

27 years agoQuick-hack to avoid `lock against myself' panic. It is not the real
KATO Takenori [Tue, 15 Apr 1997 12:56:57 +0000 (12:56 +0000)]
Quick-hack to avoid `lock against myself' panic.  It is not the real
fix!

The ufs_link() assumes that vnode is not unlocked and tries to lock it
in certain case.  Because union_link calls VOP_LINK after locking vnode,
vn_lock in ufs_link causes above panic.

Currently, I don't know the real fix for a locking violation in
union_link, but I think it is important to avoid panic.

A vnode is unlocked before calling VOP_LINK and is locked after it if
the vnode is not union fs.  Even though panic went away, the process
that access the union fs in which link was made will hang-up.

Hang-up can be easily reproduced by following operation:

mount -t union a b
cd b
ln foo bar
ls

27 years agoAdd global
Jordan K. Hubbard [Tue, 15 Apr 1997 12:31:09 +0000 (12:31 +0000)]
Add global

27 years agoShigio Yamaguchi's global tags package. The infrastructure support
Jordan K. Hubbard [Tue, 15 Apr 1997 12:30:38 +0000 (12:30 +0000)]
Shigio Yamaguchi's global tags package.  The infrastructure support
for this is already in place, so "make tags" (or "make tags HTML=yes")
should work after your next make world.
Submitted by: Shigio Yamaguchi <shigio@wafu.netgate.net>

27 years agoSynchronize with sys/i386/i386/trap.c revision 1.92.
KATO Takenori [Tue, 15 Apr 1997 11:49:00 +0000 (11:49 +0000)]
Synchronize with sys/i386/i386/trap.c revision 1.92.

27 years agoSynchronize with sys/i386/conf/files.i386 revision 1.159.
KATO Takenori [Tue, 15 Apr 1997 11:43:10 +0000 (11:43 +0000)]
Synchronize with sys/i386/conf/files.i386 revision 1.159.

27 years agoImplement a -r option to fsdb(8), ``read/only''.
Joerg Wunsch [Tue, 15 Apr 1997 09:02:48 +0000 (09:02 +0000)]
Implement a -r option to fsdb(8), ``read/only''.

27 years ago(1) Change comment for USE_X11 to be more explicit about what it
Satoshi Asami [Tue, 15 Apr 1997 08:36:41 +0000 (08:36 +0000)]
(1) Change comment for USE_X11 to be more explicit about what it
    means.

(2) Change MASTER_SITE_GNU etc. to use += instead of ?=, so users can
    specify a local mirror in /etc/make.conf and still get the full
    fallback in case something is missing.
Submitted by: Narvi <narvi@haldjas.folklore.ee>

(3) Skip port if USE_X11 or USE_IMAKE is defined and /usr/X11R6
    doesn't exist.
Submitted by: imp

(4) Add "-" in front of rmdir of ${DISTDIR}/${DIST_SUBDIR} in
    distclean target; the directory might not be empty because the
    subdir is shared with some other ports' distfiles.

27 years agoAdd comments about overridable MASTER_SITE_* variables.
Satoshi Asami [Tue, 15 Apr 1997 08:16:44 +0000 (08:16 +0000)]
Add comments about overridable MASTER_SITE_* variables.

Suggested by: Narvi <narvi@haldjas.folklore.ee>

27 years agoRemoved more traces of ISODEVMAP.
Bruce Evans [Tue, 15 Apr 1997 08:05:08 +0000 (08:05 +0000)]
Removed more traces of ISODEVMAP.

27 years agofix some typos, and some slight clean up...
John-Mark Gurney [Tue, 15 Apr 1997 07:41:10 +0000 (07:41 +0000)]
fix some typos, and some slight clean up...

Closes PR#3266

27 years agofix a few typos..
John-Mark Gurney [Tue, 15 Apr 1997 07:15:47 +0000 (07:15 +0000)]
fix a few typos..

Closed PR#3265

27 years agoList of users who may not use pppd in system password logins, and a
Daniel O'Callaghan [Tue, 15 Apr 1997 07:15:15 +0000 (07:15 +0000)]
List of users who may not use pppd in system password logins, and a
sample list of acceptable shells.

27 years agofix a couple typos...
John-Mark Gurney [Tue, 15 Apr 1997 07:06:15 +0000 (07:06 +0000)]
fix a couple typos...

Closer PR#3264

27 years agoReviewed by: Brian Somers <brian@freebsd.org>
Daniel O'Callaghan [Tue, 15 Apr 1997 07:00:35 +0000 (07:00 +0000)]
Reviewed by: Brian Somers <brian@freebsd.org>
pppd now creates /var/run/ttyXn.if file containing the interface name;
check that a 'login' user is not listed in /etc/ppp/ppp.disabled;
check that a 'login' user's shell is listed in /etc/ppp/ppp.shells;
make sure that passwordless 'login' logins are recorded in wtmp and
utmp.

27 years agoRemove osreldate. I got the number wrong anyway :(
Brian Somers [Tue, 15 Apr 1997 00:03:36 +0000 (00:03 +0000)]
Remove osreldate.  I got the number wrong anyway :(

27 years agoMake the next number redial ability configurable. The
Brian Somers [Mon, 14 Apr 1997 23:48:20 +0000 (23:48 +0000)]
Make the next number redial ability configurable.  The
"set redial pause [times]" command becomes
"set redial end-pause[.next-pause] [times]" and next-pause
defaults to 3 seconds.  This keeps things backwards
compatable.

Suggested by: ache