]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoComment out comments after #endif directives
Alfred Perlstein [Tue, 28 May 2002 21:06:03 +0000 (21:06 +0000)]
Comment out comments after #endif directives

22 years agoAdd pam_ksu(8), a module to do Kerberos 5 authentication and
Jacques Vidrine [Tue, 28 May 2002 20:52:31 +0000 (20:52 +0000)]
Add pam_ksu(8), a module to do Kerberos 5 authentication and
$HOME/.k5login authorization for su(1).

Reviewed by: des (earlier version)

22 years agoNetBSD have adopted our usage of the DMAADDR macro:
Josef Karthauser [Tue, 28 May 2002 20:51:43 +0000 (20:51 +0000)]
NetBSD have adopted our usage of the DMAADDR macro:
    date: 2002/05/28 12:42:39;  author: augustss;
    Change DMAADDR macro slightly.

Update the $NetBSD$ tags to reflect this and make slight changes to
usb_mem.h so that we're in sync with each other.

22 years agoHave this use 'static const char rcsid' instead of __RCSID()
Garance A Drosehn [Tue, 28 May 2002 20:37:00 +0000 (20:37 +0000)]
Have this use 'static const char rcsid' instead of __RCSID()

Discussed with: bde, obrien
MFC after: 10 days

22 years agoRemove use of __P() (actually P()) from code now that it's no longer
Alfred Perlstein [Tue, 28 May 2002 20:12:42 +0000 (20:12 +0000)]
Remove use of __P() (actually P()) from code now that it's no longer
available.

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 19:23:47 +0000 (19:23 +0000)]
Assume __STDC__, remove non-__STDC__ code.

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 18:57:20 +0000 (18:57 +0000)]
Assume __STDC__, remove non-__STDC__ code.

22 years agoAssume __STDC__, remove non-__STDC__ code. Remov...
Alfred Perlstein [Tue, 28 May 2002 18:39:53 +0000 (18:39 +0000)]
Assume __STDC__, remove non-__STDC__ code.                                      Remove private __P.

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 18:37:43 +0000 (18:37 +0000)]
Assume __STDC__, remove non-__STDC__ code.
Remove private __P.

22 years agoThis code defined a private __P, nuke it.
Alfred Perlstein [Tue, 28 May 2002 18:36:43 +0000 (18:36 +0000)]
This code defined a private __P, nuke it.

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 18:31:41 +0000 (18:31 +0000)]
Assume __STDC__, remove non-__STDC__ code.

Reviewed by: md5

22 years agoFix formatting, this is hard to explain, so I'll show one example.
Alfred Perlstein [Tue, 28 May 2002 18:15:04 +0000 (18:15 +0000)]
Fix formatting, this is hard to explain, so I'll show one example.

-       float ynf(int n, float x)       /* wrapper ynf */
+float
+ynf(int n, float x)    /* wrapper ynf */

This is because the __STDC__ stuff was indented.

Reviewed by: md5

22 years agoMFS: two more testcases highlighting recently discovered problem with version
Maxim Sobolev [Tue, 28 May 2002 18:00:44 +0000 (18:00 +0000)]
MFS: two more testcases highlighting recently discovered problem with version
comparison routine.

22 years agoRemove references to the update/noupdate flag from the description
Ian Dowse [Tue, 28 May 2002 17:53:52 +0000 (17:53 +0000)]
Remove references to the update/noupdate flag from the description
and examples of the "-s" option, since these two features operate
completely independently of each other.

Reviewed by: obrien

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 17:51:46 +0000 (17:51 +0000)]
Assume __STDC__, remove non-__STDC__ code.

Reviewed by: md5

22 years agoUse %02d in track numbers, so that 1 is 01, much easier for scripts
Søren Schmidt [Tue, 28 May 2002 17:39:17 +0000 (17:39 +0000)]
Use %02d in track numbers, so that 1 is 01, much easier for scripts

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 17:03:12 +0000 (17:03 +0000)]
Assume __STDC__, remove non-__STDC__ code.

Submitted by: keramida

22 years agoAssume __STDC__, remove non-__STDC__ code.
Alfred Perlstein [Tue, 28 May 2002 16:56:14 +0000 (16:56 +0000)]
Assume __STDC__, remove non-__STDC__ code.
Add $FreeBSD$.

Submitted by: keramida

22 years agoThis commit was generated by cvs2svn to compensate for changes in r97403,
David E. O'Brien [Tue, 28 May 2002 16:16:03 +0000 (16:16 +0000)]
This commit was generated by cvs2svn to compensate for changes in r97403,
which included commits to RCS files with non-trunk default branches.

22 years agoGcc 3.1.0 pre-release's C++ support bits from the FSF anoncvs repo
David E. O'Brien [Tue, 28 May 2002 16:16:03 +0000 (16:16 +0000)]
Gcc 3.1.0 pre-release's C++ support bits from the FSF anoncvs repo
on 9-May-2002 15:57:15 EDT.

22 years agoFixed some style bugs in recent commits.
Bruce Evans [Tue, 28 May 2002 15:24:13 +0000 (15:24 +0000)]
Fixed some style bugs in recent commits.

22 years agoAdd some checks to prevent NULL dereferences.
Dag-Erling Smørgrav [Tue, 28 May 2002 14:29:56 +0000 (14:29 +0000)]
Add some checks to prevent NULL dereferences.

Submitted by: jhay

22 years agoRemove a duplicated vfs_freeopts() that I introduced in last
Maxime Henrion [Tue, 28 May 2002 13:27:55 +0000 (13:27 +0000)]
Remove a duplicated vfs_freeopts() that I introduced in last
revision.

22 years agoThe stack is not at the top of the user struct.
Benno Rice [Tue, 28 May 2002 12:31:34 +0000 (12:31 +0000)]
The stack is not at the top of the user struct.

22 years agoRemove an assertion as to whether the current thread already had the FPU or
Benno Rice [Tue, 28 May 2002 12:28:22 +0000 (12:28 +0000)]
Remove an assertion as to whether the current thread already had the FPU or
not.  It may be desirable to put something similar back, but it's getting in
the way in it's current form.

22 years ago- Move macros that represent where syscall args are kept in a trapframe from
Benno Rice [Tue, 28 May 2002 12:24:29 +0000 (12:24 +0000)]
- Move macros that represent where syscall args are kept in a trapframe from
  trap.c to frame.h
- Use the macros in vm_machdep.c:cpu_fork() to set up the trap frame of the
  new thread.

22 years agoFixed lossage of the prototype for usage() in previous commit.
Bruce Evans [Tue, 28 May 2002 12:20:15 +0000 (12:20 +0000)]
Fixed lossage of the prototype for usage() in previous commit.

Fixed higher WARNS levels while I was here (missing prototype for main(),
and missing const poisoning).

22 years agoRemove the old prototype for kcopy. It's in cpu.h now.
Benno Rice [Tue, 28 May 2002 09:38:02 +0000 (09:38 +0000)]
Remove the old prototype for kcopy.  It's in cpu.h now.

22 years agoAdd a History section. comm(1) appeared in V4.
Tim J. Robbins [Tue, 28 May 2002 09:35:30 +0000 (09:35 +0000)]
Add a History section. comm(1) appeared in V4.

22 years agoNo need to handle '-' explicitly in getopt() loop.
Tim J. Robbins [Tue, 28 May 2002 09:26:08 +0000 (09:26 +0000)]
No need to handle '-' explicitly in getopt() loop.

Obtained from: NetBSD

22 years agoAdd support to GEOM for GUID Partition Tables (GPTs). The support
Marcel Moolenaar [Tue, 28 May 2002 09:04:48 +0000 (09:04 +0000)]
Add support to GEOM for GUID Partition Tables (GPTs). The support
is currently conditional on both the GEOM and GEOM_GPT options to
avoid getting GPT by default and having the MBR and GPT classes
clash.
The correct behaviour of the MBR class would be to back-off (reject)
a MBR if it's a Protective MBR (a MBR with a single partition of type
0xEE that spans the whole disk (as far as the MBR is concerned).
The correct behaviour if the GPT class would be to back-off (reject)
a GPT if there's a MBR that's not a Protective MBR.

At this stage it's inconvenient to destroy a good MBR when working
with GPTs that it's more convenient to have the MBR class back-off
when it detects the GPT signature on disk and have the GPT class
ignore the MBR.

In sys/gpt.h UUIDs (GUIDs) for the following FreeBSD partitions
have been defined:

GPT_ENT_TYPE_FREEBSD
FreeBSD slice with disklabel. This is the equivalent of
the well-known FreeBSD MBR partition type.
GPT_ENT_TYPE_FREEBSD_{SWAP|UFS|UFS2|VINUM}
FreeBSD partitions in the context of disklabel. This is
speculating on the idea to use the GPT to hold partitions
instead if slices and removing the fixed (and low) limits
we have on the number of partitions.

This commit lacks a GPT image for the regression suite.

22 years agoMFS r1.1.2.18 - Include vim6 rather than vim5 on disc #1.
Murray Stokely [Tue, 28 May 2002 08:46:21 +0000 (08:46 +0000)]
MFS r1.1.2.18 - Include vim6 rather than vim5 on disc #1.

22 years agoUpdate for current /usr/src.
Ruslan Ermilov [Tue, 28 May 2002 08:25:29 +0000 (08:25 +0000)]
Update for current /usr/src.

22 years agoDocument the ENABLE_SUID_NEWGRP variable.
Tim J. Robbins [Tue, 28 May 2002 08:04:50 +0000 (08:04 +0000)]
Document the ENABLE_SUID_NEWGRP variable.

Requested by: mike
PR: 36190

22 years agoImplement pmap_copy and pmap_copy_page.
Benno Rice [Tue, 28 May 2002 07:38:55 +0000 (07:38 +0000)]
Implement pmap_copy and pmap_copy_page.

22 years agoMove the kcopy() function from trap.c to machdep.c. Add a prototype.
Benno Rice [Tue, 28 May 2002 07:36:36 +0000 (07:36 +0000)]
Move the kcopy() function from trap.c to machdep.c.  Add a prototype.

22 years agoFixed the world breakage caused by my last commit.
Ruslan Ermilov [Tue, 28 May 2002 07:30:21 +0000 (07:30 +0000)]
Fixed the world breakage caused by my last commit.
NOMAN is defined when building bootstrap-tools.

Submitted by: jhay

22 years agoActually accept the default argument to SCREEN_WIDTH.
Doug Barton [Tue, 28 May 2002 07:25:44 +0000 (07:25 +0000)]
Actually accept the default argument to SCREEN_WIDTH.

Submitted by: brian

22 years agoAdd NAI copyright.
Dag-Erling Smørgrav [Tue, 28 May 2002 06:53:41 +0000 (06:53 +0000)]
Add NAI copyright.

22 years agoIf unable to retrive maxfiles / openfiles, fail rather than print garbage.
Dag-Erling Smørgrav [Tue, 28 May 2002 06:52:21 +0000 (06:52 +0000)]
If unable to retrive maxfiles / openfiles, fail rather than print garbage.
Gratuitously rename a couple of variables.
Remove unused macros.
Add NAI copyright.

Sponsored by: DARPA, NAI Labs

22 years agoDrive-by whitespace cleanup & add NAI copyright
Dag-Erling Smørgrav [Tue, 28 May 2002 06:47:32 +0000 (06:47 +0000)]
Drive-by whitespace cleanup & add NAI copyright

22 years agoDrive-by whitespace cleanup.
Dag-Erling Smørgrav [Tue, 28 May 2002 06:46:37 +0000 (06:46 +0000)]
Drive-by whitespace cleanup.

22 years agoRemove the code that was disabled in a recent commit; it is of very limited
Dag-Erling Smørgrav [Tue, 28 May 2002 06:37:27 +0000 (06:37 +0000)]
Remove the code that was disabled in a recent commit; it is of very limited
use and has been broken in -CURRENT for a long time.
Clean up unneeded entries in the nlist array.
Implement kvm-backed ttymode (which we never had before).  Incomplete as we
do not (yet?) print the correct device, sid or pgid.

Sponsored by: DARPA, NAI Labs

22 years agoWrite "FROM" heading above the hostname column, like NetBSD and GNU do.
Tim J. Robbins [Tue, 28 May 2002 06:36:46 +0000 (06:36 +0000)]
Write "FROM" heading above the hostname column, like NetBSD and GNU do.

22 years agoBack out part of previous commit; the dev_t union trick is still useful in
Dag-Erling Smørgrav [Tue, 28 May 2002 06:34:28 +0000 (06:34 +0000)]
Back out part of previous commit; the dev_t union trick is still useful in
the kvm case.

22 years agoAdd uuidgen(2) and uuidgen(1).
Marcel Moolenaar [Tue, 28 May 2002 06:16:08 +0000 (06:16 +0000)]
Add uuidgen(2) and uuidgen(1).

The uuidgen command, by means of the uuidgen syscall, generates one
or more Universally Unique Identifiers compatible with OSF/DCE 1.1
version 1 UUIDs.

From the Perforce logs (change 11995):

Round of cleanups:
o  Give uuidgen() the correct prototype in syscalls.master
o  Define struct uuid according to DCE 1.1 in sys/uuid.h
o  Use struct uuid instead of uuid_t. The latter is defined
   in sys/uuid.h but should not be used in kernel land.
o  Add snprintf_uuid(), printf_uuid() and sbuf_printf_uuid()
   to kern_uuid.c for use in the kernel (currently geom_gpt.c).
o  Rename the non-standard struct uuid in kern/kern_uuid.c
   to struct uuid_private and give it a slightly better definition
   for better byte-order handling. See below.
o  In sys/gpt.h, fix the broken uuid definitions to match the now
   compliant struct uuid definition. See below.
o  In usr.bin/uuidgen/uuidgen.c catch up with struct uuid change.

A note about byte-order:
        The standard failed to provide a non-conflicting and
unambiguous definition for the binary representation. My initial
implementation always wrote the timestamp as a 64-bit little-endian
(2s-complement) integral. The clock sequence was always written
as a 16-bit big-endian (2s-complement) integral. After a good
nights sleep and couple of Pan Galactic Gargle Blasters (not
necessarily in that order :-) I reread the spec and came to the
conclusion that the time fields are always written in the native
by order, provided the the low, mid and hi chopping still occurs.
The spec mentions that you "might need to swap bytes if you talk
to a machine that has a different byte-order". The clock sequence
is always written in big-endian order (as is the IEEE 802 address)
because its division is resulting in bytes, making the ordering
unambiguous.

22 years agoQuiet some format string warnings by marking prototypes with
Kris Kennaway [Tue, 28 May 2002 06:12:31 +0000 (06:12 +0000)]
Quiet some format string warnings by marking prototypes with
__printflike()

MFC After: 1 week

22 years agoRemove, with extreme prejudice, the notion that /tmp best needs to be a
Jeroen Ruigrok van der Werven [Tue, 28 May 2002 06:03:50 +0000 (06:03 +0000)]
Remove, with extreme prejudice, the notion that /tmp best needs to be a
softlink to /var/tmp.

This is horribly wrong since /tmp and /var/tmp serve different goals.
Even given the text that in the old days things were different doesn't help,
since our scripts clean /tmp, and we depend on /var/tmp to persist for
vi editing sessions recovery and other likewise applications.

22 years agoAdd syscall uuidgen() for generating Univerally Unique Identifiers
Marcel Moolenaar [Tue, 28 May 2002 05:58:06 +0000 (05:58 +0000)]
Add syscall uuidgen() for generating Univerally Unique Identifiers
(UUIDs). On ia64 UUIDs, aka GUIDs, are used by EFI and the firmware
among others. To create GUID Partition Tables (GPTs), we need to
be able to generate UUIDs.

22 years agoOops, don't print /dev/ twice.
Dag-Erling Smørgrav [Tue, 28 May 2002 05:45:56 +0000 (05:45 +0000)]
Oops, don't print /dev/ twice.

22 years agostruct tty -> struct xtty. Reenable some previously disable code, but
Dag-Erling Smørgrav [Tue, 28 May 2002 05:42:32 +0000 (05:42 +0000)]
struct tty -> struct xtty.  Reenable some previously disable code, but
temporarily disable some rarely-used code that needs more work.

Sponsored by: DARPA, NAI Labs

22 years agoIntroduce struct xtty, used when exporting tty information to userland.
Dag-Erling Smørgrav [Tue, 28 May 2002 05:40:53 +0000 (05:40 +0000)]
Introduce struct xtty, used when exporting tty information to userland.
Make kern.ttys export a struct xtty rather than struct tty.  Since struct
tty is no longer exposed to userland, remove the dev_t / udev_t hack.

Sponsored by: DARPA, NAI Labs

22 years agoHook newgrp(1) up to the build.
Tim J. Robbins [Tue, 28 May 2002 05:07:42 +0000 (05:07 +0000)]
Hook newgrp(1) up to the build.

PR: 36190

22 years agoAdd the newgrp(1) utility, which changes groups. This is required by
Tim J. Robbins [Tue, 28 May 2002 05:05:28 +0000 (05:05 +0000)]
Add the newgrp(1) utility, which changes groups. This is required by
the POSIX.2 UPE.

PR: 36190
Reviewed by: -standards, silence on -audit

22 years agoPrepend BIG_ENDIAN, BYTE_ORDER, LITTLE_ENDIAN, and PDP_ENDIAN with an
Mike Barcroft [Tue, 28 May 2002 04:32:25 +0000 (04:32 +0000)]
Prepend BIG_ENDIAN, BYTE_ORDER, LITTLE_ENDIAN, and PDP_ENDIAN with an
underscore.  This is the preferred form.  Also fix a preprocessor
syntax error in an error-detection section.

22 years agoTypo fix: s/IPSec/IPsec/g
Bruce A. Mah [Tue, 28 May 2002 00:16:22 +0000 (00:16 +0000)]
Typo fix:  s/IPSec/IPsec/g

PR: 38627
Submitted by:  Marc Fonvieille <marc@blackend.org>

22 years agoCoerce pid_t to long rather than int for better portability.
Brian Somers [Mon, 27 May 2002 23:19:53 +0000 (23:19 +0000)]
Coerce pid_t to long rather than int for better portability.

Suggested by: Theo de Raadt <deraadt@openbsd.org>

22 years ago o Remove unused #defines.
Alan Cox [Mon, 27 May 2002 22:10:28 +0000 (22:10 +0000)]
 o Remove unused #defines.

22 years agoThe ports collection now contains over 7000 applications and consumes
Murray Stokely [Mon, 27 May 2002 19:31:02 +0000 (19:31 +0000)]
The ports collection now contains over 7000 applications and consumes
about 165MB of space when first installed.

Submitted by:  rpratt

22 years agoCorrect a check for NUL.
Robert Drehmel [Mon, 27 May 2002 19:27:43 +0000 (19:27 +0000)]
Correct a check for NUL.

Spotted by: bde

22 years agoI missed the setting of PREFIX_INCLUDE_DIR when I visually santity checked
David E. O'Brien [Mon, 27 May 2002 17:48:28 +0000 (17:48 +0000)]
I missed the setting of PREFIX_INCLUDE_DIR when I visually santity checked
the output from a autoconfig run.

22 years agoRe-enable running of makewhatis(1).
Ruslan Ermilov [Mon, 27 May 2002 15:52:40 +0000 (15:52 +0000)]
Re-enable running of makewhatis(1).

22 years agoBootstrap makewhatis(1).
Ruslan Ermilov [Mon, 27 May 2002 15:51:38 +0000 (15:51 +0000)]
Bootstrap makewhatis(1).

22 years agoAs far as I can tell, the maximum packet size for 802.11b is 2346,
David Malone [Mon, 27 May 2002 13:39:01 +0000 (13:39 +0000)]
As far as I can tell, the maximum packet size for 802.11b is 2346,
which means that the valid range for the RTS threshold is 0-2347.
The default is definitely 2347.

22 years agoRemove a stray "t" in the usage message.
David Malone [Mon, 27 May 2002 13:33:46 +0000 (13:33 +0000)]
Remove a stray "t" in the usage message.

PR: 38605
Submitted by: KOIE Hidetaka <hide@koie.org>
MFC after: 1 week

22 years agophsyical -> physical
Bruce Evans [Mon, 27 May 2002 13:08:08 +0000 (13:08 +0000)]
phsyical -> physical

Submitted by: Andre Guibert de Bruet <andy@siliconlandmark.com>

22 years agoPrint srr1 in printtrap()
Benno Rice [Mon, 27 May 2002 11:20:19 +0000 (11:20 +0000)]
Print srr1 in printtrap()

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoGet the correct memory regions from OpenFirmware. We were getting the
Benno Rice [Mon, 27 May 2002 11:18:12 +0000 (11:18 +0000)]
Get the correct memory regions from OpenFirmware.  We were getting the
"available" ranges, not the "physical" ranges.  Clean up some of the
bootstrap code in the process.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years ago - Move the loop conditional into the "for" header.
Robert Drehmel [Mon, 27 May 2002 11:01:30 +0000 (11:01 +0000)]
 - Move the loop conditional into the "for" header.
 - Remove redundant "? :" construct.
style(9):
 - Place a space after return statements.
 - Compare pointers to NULL.
 - Do not use ! to compare a character to nul.

22 years agoChange the ">>> make release ..." output to also include ${TARGET}.
Ruslan Ermilov [Mon, 27 May 2002 10:59:41 +0000 (10:59 +0000)]
Change the ">>> make release ..." output to also include ${TARGET}.

22 years agoFixed the nasty bug I introduced in rev. 1.671 that got
Ruslan Ermilov [Mon, 27 May 2002 10:52:03 +0000 (10:52 +0000)]
Fixed the nasty bug I introduced in rev. 1.671 that got
fatal in the previous delta.  Repeat after me: exists()
is executed at parse time.

22 years agoUse correct types in [sf]uword32.
Benno Rice [Mon, 27 May 2002 10:50:47 +0000 (10:50 +0000)]
Use correct types in [sf]uword32.

22 years agoDon't risk catching a signal while handling a signal for a dying child, as we
Juli Mallett [Mon, 27 May 2002 08:10:24 +0000 (08:10 +0000)]
Don't risk catching a signal while handling a signal for a dying child, as we
can then end up not properly clearing wtmp/utmp entries.

PR: bin/37934
Submitted by: Sandeep Kumar <skumar@juniper.net>
Reviewed by: markm
MFC after: 2 weeks

22 years agoWork around a memory fault on ia64 caused by having the 1MB buffer on
Marcel Moolenaar [Mon, 27 May 2002 07:54:43 +0000 (07:54 +0000)]
Work around a memory fault on ia64 caused by having the 1MB buffer on
the stack in DoFile(). This needs some investigation. In the mean time
we do a one time malloc() for the buffer to have it on the heap instead.

22 years agoThe XCU issue 5 requirement of accepting the file operand intermingled
Tim J. Robbins [Mon, 27 May 2002 06:53:54 +0000 (06:53 +0000)]
The XCU issue 5 requirement of accepting the file operand intermingled
with options has been withdrawn in issue 6, to which nl(1) now conforms.

22 years agostyle(9): use errx() where appropriate instead of fprintf() + exit().
Tim J. Robbins [Mon, 27 May 2002 06:46:29 +0000 (06:46 +0000)]
style(9): use errx() where appropriate instead of fprintf() + exit().

22 years agostyle(9): use err() instead of perror() + exit().
Tim J. Robbins [Mon, 27 May 2002 06:37:34 +0000 (06:37 +0000)]
style(9): use err() instead of perror() + exit().

22 years agoClaim conformance to IEEE Std 1003.1-2001.
Tim J. Robbins [Mon, 27 May 2002 06:08:14 +0000 (06:08 +0000)]
Claim conformance to IEEE Std 1003.1-2001.
See also csplit(1).

22 years agoAllow "-" to be specified as an operand as well as an option.
Tim J. Robbins [Mon, 27 May 2002 05:27:10 +0000 (05:27 +0000)]
Allow "-" to be specified as an operand as well as an option.
SUSV3 requires something like "split -- -" to work. Document the "-" operand.

22 years agoDon't tsleep() with an sb_mtx held.
Dima Dorfman [Mon, 27 May 2002 05:20:15 +0000 (05:20 +0000)]
Don't tsleep() with an sb_mtx held.

22 years agoAvoid overflowing `fname' if the file name prefix given on the command
Tim J. Robbins [Mon, 27 May 2002 04:59:46 +0000 (04:59 +0000)]
Avoid overflowing `fname' if the file name prefix given on the command
line is too long.

Obtained from: OpenBSD

22 years agoExit non-zero if the tags file cannot be opened.
Tim J. Robbins [Mon, 27 May 2002 03:54:45 +0000 (03:54 +0000)]
Exit non-zero if the tags file cannot be opened.

22 years agoRemove spurious period.
Dima Dorfman [Mon, 27 May 2002 03:45:27 +0000 (03:45 +0000)]
Remove spurious period.

22 years agoFrom NetBSD:
Tim J. Robbins [Mon, 27 May 2002 03:17:28 +0000 (03:17 +0000)]
From NetBSD:
    Revision 1.10 Sat Oct 14 17:41:55 2000 UTC by bjh21
    Don't core dump with an empty format string.  Fixes PR#11218.
    Patch supplied by Launey Thomas.

Obtained from: NetBSD

22 years agoInitialise the `positions' array correctly before use.
Tim J. Robbins [Mon, 27 May 2002 02:01:25 +0000 (02:01 +0000)]
Initialise the `positions' array correctly before use.

22 years agoUse underscored variant of BYTE_ORDER and friends to allow this to
Mike Barcroft [Mon, 27 May 2002 00:55:17 +0000 (00:55 +0000)]
Use underscored variant of BYTE_ORDER and friends to allow this to
work in a !__BSD_VISIBLE environment.

22 years agoRegen.
Josef Karthauser [Mon, 27 May 2002 00:01:21 +0000 (00:01 +0000)]
Regen.

22 years agoAdd a couple of new aue ethernet adapters from NetBSD:
Josef Karthauser [Mon, 27 May 2002 00:00:48 +0000 (00:00 +0000)]
Add a couple of new aue ethernet adapters from NetBSD:
ELECOM LDUSBLTX
IODATA USBETTXS

22 years agoUse aue_lookup for looking up devices.
Josef Karthauser [Sun, 26 May 2002 23:54:37 +0000 (23:54 +0000)]
Use aue_lookup for looking up devices.

22 years agoCorrect the usage of DMAADDR in a piece of '#if 0'd code. (The compiler
Josef Karthauser [Sun, 26 May 2002 22:13:09 +0000 (22:13 +0000)]
Correct the usage of DMAADDR in a piece of '#if 0'd code. (The compiler
didn't pick it up.)

22 years agoMFNetBSD:
Josef Karthauser [Sun, 26 May 2002 22:11:34 +0000 (22:11 +0000)]
MFNetBSD:
    revision 1.124
    date: 2002/05/26 03:10:02;  author: minoura;  state: Exp;  lines: +3 -3
    Clear done_head in the HCCA *before* acknoledging the interrupt.
    Driver lost some completed transfers under heavy loads.

22 years agoMFNetBSD: ohci.c (1.124), uhci.c (1.159), usbdi.c (1.100)
Josef Karthauser [Sun, 26 May 2002 22:00:06 +0000 (22:00 +0000)]
MFNetBSD: ohci.c (1.124), uhci.c (1.159), usbdi.c (1.100)
    date: 2002/05/19 06:24:31;  author: augustss;  state: Exp;
    Update dma memory access API a little.

NetBSD have adopted our way of using the KERNADDR macro.  Update
the revision tags to show that we're in sync, and remove the casts
that they did in their adaptation.

22 years agoMore catchup MFbeds:
Udo Erdelhoff [Sun, 26 May 2002 19:41:51 +0000 (19:41 +0000)]
More catchup MFbeds:
hardware/alpha/proc-alpha.sgml: MFen 1.40 -> 1.41
hardware/common/dev.sgml: MFen 1.87 -> 1.91
installation/common/install.sgml: translation fixes
readme/article.sgml: translation fixes
relnotes/common/new.sgml: MFen 1.340 -> 1.355

22 years agoThe pre-release sweep continues:
Udo Erdelhoff [Sun, 26 May 2002 19:35:41 +0000 (19:35 +0000)]
The pre-release sweep continues:

Just update the 'original revision' markers of some files where the last
update of the english version has no impact on the translated version
(spelling checks are a premium cause of this) except making it harder
to find the changes that still need to be translated/comitted.

22 years agoAdd a proof-of-concept encryption class.
Poul-Henning Kamp [Sun, 26 May 2002 18:14:38 +0000 (18:14 +0000)]
Add a proof-of-concept encryption class.

"The only hard problem in cryptography is key-management."

All sectors are encrypted with AES in CBC mode using a constant key,
currently compiled in and all zero.

To activate this module, write the magic header on the partition:

echo "<<FreeBSD-GEOM-AES>>" | dd conv=sync of=/dev/md98

The encrypted device will be one sector shorter and have ".aes"
appended to its name.

Sponsored by: DARPA & NAI Labs.

22 years agoGive the closet-dev_t we hand to the diskdrivers a name.
Poul-Henning Kamp [Sun, 26 May 2002 17:35:14 +0000 (17:35 +0000)]
Give the closet-dev_t we hand to the diskdrivers a name.

22 years agoOnly clear the spoiled flag if the class had no spoiled method, the spoiled
Poul-Henning Kamp [Sun, 26 May 2002 17:17:25 +0000 (17:17 +0000)]
Only clear the spoiled flag if the class had no spoiled method, the spoiled
method may have deallocated the consumer already and modifying free()'ed
memory is bad style.

Sponsored by: DARPA & NAI Labs.

22 years agoAlpha has crtfastmath also.
David E. O'Brien [Sun, 26 May 2002 17:16:21 +0000 (17:16 +0000)]
Alpha has crtfastmath also.

22 years agoRev 1.12 broke `make all' w/o a previous `make depends'.
David E. O'Brien [Sun, 26 May 2002 17:09:31 +0000 (17:09 +0000)]
Rev 1.12 broke `make all' w/o a previous `make depends'.

22 years agoInitial translation
Sebastien Gioria [Sun, 26 May 2002 17:07:32 +0000 (17:07 +0000)]
Initial translation