]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoDisable SGR support in grotty(1) pending the resolution on -arch.
ru [Fri, 18 Oct 2002 09:10:44 +0000 (09:10 +0000)]
Disable SGR support in grotty(1) pending the resolution on -arch.

21 years agoWhen a user gets refused because the password is wrong, use the
markm [Fri, 18 Oct 2002 08:23:24 +0000 (08:23 +0000)]
When a user gets refused because the password is wrong, use the
older "BAD SU" syslog message that folks prefer. There is quite
a bit more tweaking that can be done with other similar messages.

Asked for by: tjr

21 years agoAdd myself
adamw [Fri, 18 Oct 2002 07:50:46 +0000 (07:50 +0000)]
Add myself

Approved by: pat (mentor)

21 years agoEveryone else is doing it! (adding their birthdays)
trhodes [Fri, 18 Oct 2002 06:21:09 +0000 (06:21 +0000)]
Everyone else is doing it! (adding their birthdays)

21 years agoNew committer's first commit(birthday).
maho [Fri, 18 Oct 2002 05:48:55 +0000 (05:48 +0000)]
New committer's first commit(birthday).

21 years agoAdd a new man page describing the mac_bsdextended policy.
chris [Fri, 18 Oct 2002 05:40:48 +0000 (05:40 +0000)]
Add a new man page describing the mac_bsdextended policy.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

21 years agoCosmetic line-wrapping change that has the side-effect of not producing
chris [Fri, 18 Oct 2002 05:31:39 +0000 (05:31 +0000)]
Cosmetic line-wrapping change that has the side-effect of not producing
the (incorrectly-spaced) output "... Network Associates Inc.  under ..."

21 years agoRemove a superfluous line containing only `.'
chris [Fri, 18 Oct 2002 05:29:39 +0000 (05:29 +0000)]
Remove a superfluous line containing only `.'

21 years agoNew committer first committion.
daichi [Fri, 18 Oct 2002 04:56:03 +0000 (04:56 +0000)]
New committer first committion.
  addition my birthday.

21 years agoBackout my previous "brain cramp" commit.
pirzyk [Fri, 18 Oct 2002 04:53:46 +0000 (04:53 +0000)]
Backout my previous "brain cramp" commit.

Requested by: Most developers
Apologies to: Most developers, with special note to <ken@kdm.org>
Collabroation in the future with: Kenneth D. Merry <ken@kdm.org>

21 years agoCover the ELF headers with the text segment so that they get loaded
marcel [Fri, 18 Oct 2002 04:46:36 +0000 (04:46 +0000)]
Cover the ELF headers with the text segment so that they get loaded
into memory. This brings us in line with the other architectures and
more easily allows us to do machine dependent processing on the ELF
file (such as scanning for unwind information).

21 years agoinstall "fast ipsec" include files
sam [Fri, 18 Oct 2002 04:21:29 +0000 (04:21 +0000)]
install "fast ipsec" include files

21 years agoOutput "human-readable" values with a non-0 precision where
dd [Fri, 18 Oct 2002 04:06:59 +0000 (04:06 +0000)]
Output "human-readable" values with a non-0 precision where
appropriate.  Before this, a 2.9 GB file was misleadingly reported as
"2G".  This mostly brings unit_adjust() in line with what is in du(1).

Reviewed by: jmallett
Approved by: nik

21 years agoExempt the "wheel group requirement" by default when su'ing to root if
rwatson [Fri, 18 Oct 2002 02:39:21 +0000 (02:39 +0000)]
Exempt the "wheel group requirement" by default when su'ing to root if
the wheel group has no explicit members listed in /etc/group.  This adds
the "exempt_if_empty" flag to pam_wheel in the default configuration;
in some environments, it may be appropriate to remove this flag, however,
this default is the same as pre-pam_wheel.

Reviewed by: markm
Sponsored by: DARPA, Network Associates Laboratories

21 years agoIntroduce 'exempt_if_empty' option to pam_wheel(8), which bypasses the
rwatson [Fri, 18 Oct 2002 02:37:29 +0000 (02:37 +0000)]
Introduce 'exempt_if_empty' option to pam_wheel(8), which bypasses the
group membership requirement if the group has no explicit members listed
in /etc/group.  By default, this group is the wheel group; setting this
flag restores the default BSD behavior from 4.x.

Reviewed by: markm
Requested by: various
Sponsored by: DARPA, Network Associates Laboratories

21 years agoHave lockinit() initialize the debugging fields of a lock
mckusick [Fri, 18 Oct 2002 01:34:10 +0000 (01:34 +0000)]
Have lockinit() initialize the debugging fields of a lock
when DEBUG_LOCKS is defined.

Sponsored by: DARPA & NAI Labs.

21 years agoWhen the number of dirty buffers rises too high, the buf_daemon runs
mckusick [Fri, 18 Oct 2002 01:29:59 +0000 (01:29 +0000)]
When the number of dirty buffers rises too high, the buf_daemon runs
to help clean up. After selecting a potential buffer to write, this
patch has it acquire a lock on the vnode that owns the buffer before
trying to write it. The vnode lock is necessary to avoid a race with
some other process holding the vnode locked and trying to flush its
dirty buffers. In particular, if the vnode in question is a snapshot
file, then the race can lead to a deadlock. To avoid slowing down the
buf_daemon, it does a non-blocking lock request when trying to lock
the vnode. If it fails to get the lock it skips over the buffer and
continues down its queue looking for buffers to flush.

Sponsored by: DARPA & NAI Labs.

21 years agoWith the revised single-lock method used in snapshots, the
mckusick [Fri, 18 Oct 2002 01:17:28 +0000 (01:17 +0000)]
With the revised single-lock method used in snapshots, the
BA_NOWAIT flag is no longer needed.

Sponsored by: DARPA & NAI Labs.

21 years agoTypo in verb: zeros -> zeroes.
keramida [Fri, 18 Oct 2002 01:05:46 +0000 (01:05 +0000)]
Typo in verb: zeros -> zeroes.

21 years agonot sure if this correct, but it compiles again
sam [Fri, 18 Oct 2002 00:26:34 +0000 (00:26 +0000)]
not sure if this correct, but it compiles again

21 years agoSeparate fiels reported by disk_err() with spaces, so that output doesn't
sobomax [Thu, 17 Oct 2002 23:48:29 +0000 (23:48 +0000)]
Separate fiels reported by disk_err() with spaces, so that output doesn't
look cryptic.

MFC after: 1 week

21 years agoFix silly typo: loacal --> local.
sobomax [Thu, 17 Oct 2002 23:46:32 +0000 (23:46 +0000)]
Fix silly typo: loacal --> local.

21 years agoChanged the scsi mode command to use 10 bytes scsi commands instead
pirzyk [Thu, 17 Oct 2002 23:37:03 +0000 (23:37 +0000)]
Changed the scsi mode command to use 10 bytes scsi commands instead
of the 6 byte ones.  This helps with using a IDE cdrom behind a USB
interface.

PR: kern/43885

21 years agoStyle: use sa_handler instead of __sigaction_u.__sa_handler.
tjr [Thu, 17 Oct 2002 23:32:44 +0000 (23:32 +0000)]
Style: use sa_handler instead of __sigaction_u.__sa_handler.

21 years agoActivate ugidfw.8 man page.
chris [Thu, 17 Oct 2002 22:43:11 +0000 (22:43 +0000)]
Activate ugidfw.8 man page.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

21 years agoReplace the conventional usage of strncpy() by using strlcpy().
robert [Thu, 17 Oct 2002 22:27:21 +0000 (22:27 +0000)]
Replace the conventional usage of strncpy() by using strlcpy().

21 years ago - Use strlcpy() rather than strncpy() to copy NUL terminated
robert [Thu, 17 Oct 2002 22:00:30 +0000 (22:00 +0000)]
 - Use strlcpy() rather than strncpy() to copy NUL terminated
   strings.
 - Pass the correct buffer size to getcredhostname().

21 years agoBe consistent and return the NUL at the end of kern.geom.conf{xml,dot}.
phk [Thu, 17 Oct 2002 21:39:06 +0000 (21:39 +0000)]
Be consistent and return the NUL at the end of kern.geom.conf{xml,dot}.

Spotted by: sam

21 years agoInstead of (sizeof(source_buffer) - 1) bytes, copy at most
robert [Thu, 17 Oct 2002 21:02:02 +0000 (21:02 +0000)]
Instead of (sizeof(source_buffer) - 1) bytes, copy at most
(sizeof(destination_buffer) - 1) bytes into the destination buffer.
This was not harmful because they currently both provide space for
(MAXCOMLEN + 1) bytes.

21 years agoFD locking is not enabled anymore, so the table which contains the owner
jmallett [Thu, 17 Oct 2002 20:26:26 +0000 (20:26 +0000)]
FD locking is not enabled anymore, so the table which contains the owner
of a file descriptor has NULL entries, so don't dereference the table entries
to get the owners ever -- don't print the owners when processing a thread_dump
request as a result of SIGINFO.

Reviewed by: deischen

21 years agoUse strlcpy() instead of strncpy() to copy NUL terminated strings
robert [Thu, 17 Oct 2002 20:03:38 +0000 (20:03 +0000)]
Use strlcpy() instead of strncpy() to copy NUL terminated strings
for safety and consistency.

21 years agoAdd Dutch ISO keymap
ache [Thu, 17 Oct 2002 19:56:00 +0000 (19:56 +0000)]
Add Dutch ISO keymap
PR: 17699
Submitted by: Edwin Groothuis <mavetju@chello.nl>

21 years agoNUL terminate sysctl kern.disks
phk [Thu, 17 Oct 2002 19:27:02 +0000 (19:27 +0000)]
NUL terminate sysctl kern.disks

21 years agoo ioctl DIOCGDINFO error wasn't checked
sam [Thu, 17 Oct 2002 18:34:32 +0000 (18:34 +0000)]
o ioctl DIOCGDINFO error wasn't checked
o memory wasn't reclaimed in certain cases
o add more msgs under #ifdef DEBUG
o rewrite tangle of for loops for clarity

NB: Open_Disk should redo how it malloc's memory so the caller can free
    everything.  Documentation says the caller can free the disk list to
    reclaim everything but this leaks the indirect strings.  Fixing this
    is simple for the sysctl case but adds complexity to the fallback,
    non-sysctl, case.

21 years agocorrect arg order to strlcpy/strlcat under #ifdef alpha
sam [Thu, 17 Oct 2002 18:23:21 +0000 (18:23 +0000)]
correct arg order to strlcpy/strlcat under #ifdef alpha

21 years agoAdd the !define(COMPILING_LINT)
pirzyk [Thu, 17 Oct 2002 18:17:28 +0000 (18:17 +0000)]
Add the !define(COMPILING_LINT)

pass the pointy hat...

Requested by: Juli Mallett <jmallett@FreeBSD.org>

21 years agoWhen entering the firmware mappings into the kernel tlb, clear all 'soft'
tmm [Thu, 17 Oct 2002 18:16:21 +0000 (18:16 +0000)]
When entering the firmware mappings into the kernel tlb, clear all 'soft'
bits that might be set in the firmware tte data field, and set the soft
flag TD_EXEC to mark the page executable. Failing to do the latter would
cause fatal instruction faults in the prom in certain situations.

Reviewed by: jake

21 years agoAllocate major 177 for the OpenFirmware control device, which I have
tmm [Thu, 17 Oct 2002 18:11:49 +0000 (18:11 +0000)]
Allocate major 177 for the OpenFirmware control device, which I have
ported from NetBSD and plan to commit soon.

21 years agoAdd mount_devfs so we can remove MAKEDEV from the fixit floppy. These
sam [Thu, 17 Oct 2002 18:04:53 +0000 (18:04 +0000)]
Add mount_devfs so we can remove MAKEDEV from the fixit floppy.  These
two changes make stuff fit in a floppy image again.

Note: tested only for i386, but (blindly) applied to all machines

21 years agoQuirk for HP 315 USB Digital Camera
njl [Thu, 17 Oct 2002 18:04:41 +0000 (18:04 +0000)]
Quirk for HP 315 USB Digital Camera

Submitted by: Keith White <kwhite@uottawa.ca>
PR: kern/41010

MFC after: 1 week

21 years agoStrip splash, tun, and COMPAT_FREEBSD4 from the config used to build
sam [Thu, 17 Oct 2002 18:02:27 +0000 (18:02 +0000)]
Strip splash, tun, and COMPAT_FREEBSD4 from the config used to build
the BOOTMFS kernel.  These help reduce the kernel size so things fit
in a floppy image.  There are more low-hanging fruit to be had here
if things fail to fit again.

21 years agoMark more drivers as going on the (nonexistent) drivers.flp floppy
sam [Thu, 17 Oct 2002 18:00:30 +0000 (18:00 +0000)]
Mark more drivers as going on the (nonexistent) drivers.flp floppy
to reduce the size of the bootstrap kernels.  This is needed to make
stuff fit in primary floppy images.  Specific drivers moved: RAID
support, tx Ethernet, and wi (wireless).  Probably want to revisit
at least the last two drivers before release.

Submitted by: jhay (mostly)

21 years agolast arg of in6?_gif_output() is not used any more.
ume [Thu, 17 Oct 2002 17:47:55 +0000 (17:47 +0000)]
last arg of in6?_gif_output() is not used any more.

Obtained from: KAME
MFC after: 3 weeks

21 years ago- drop too short IPv6 frame
ume [Thu, 17 Oct 2002 17:42:46 +0000 (17:42 +0000)]
- drop too short IPv6 frame
- NULL != 0

Obtained from: KAME
MFC after: 3 weeks

21 years agos/gifp/ifp/
ume [Thu, 17 Oct 2002 17:39:56 +0000 (17:39 +0000)]
s/gifp/ifp/

Obtained from: KAME
MFC after: 3 weeks

21 years agofix kldload error return when a module is rejected because it's statically
sam [Thu, 17 Oct 2002 17:28:57 +0000 (17:28 +0000)]
fix kldload error return when a module is rejected because it's statically
linked in the kernel.  When this condition is detected deep in the linker
internals the EEXIST error code that's returned is stomped on and instead
an ENOEXEC code is returned.  This makes apps like sysinstall bitch.

21 years agoFix a bug that showed up when trying to produce a database for
joe [Thu, 17 Oct 2002 16:29:26 +0000 (16:29 +0000)]
Fix a bug that showed up when trying to produce a database for
subdirectories, and ended up making us loop forever.

Add the username to the md5 of the commit to make it slightly more
unique.

Make the 'cvs' run quietly.

21 years agoFix compile with options SC_NO_SUSPEND_VTYSWITCH.
jhb [Thu, 17 Oct 2002 16:09:13 +0000 (16:09 +0000)]
Fix compile with options SC_NO_SUSPEND_VTYSWITCH.

21 years ago - Allocate only enough space for a temporary buffer to hold
robert [Thu, 17 Oct 2002 15:52:42 +0000 (15:52 +0000)]
 - Allocate only enough space for a temporary buffer to hold
   the path including the terminating NUL character from
   `struct sockaddr_un' rather than SOCK_MAXADDRLEN bytes.
 - Use strlcpy() instead of strncpy() to copy strings.

21 years agomy(4) requires miibus, so that move it into proper section and also add a
sobomax [Thu, 17 Oct 2002 15:32:03 +0000 (15:32 +0000)]
my(4) requires miibus, so that move it into proper section and also add a
verbose description into comment area.

21 years agoSwitch from the bs driver to the ct driver.
nyan [Thu, 17 Oct 2002 13:57:57 +0000 (13:57 +0000)]
Switch from the bs driver to the ct driver.

21 years ago1. Fix a comment. Locking _is_ needed (but not done).
iwasaki [Thu, 17 Oct 2002 13:55:39 +0000 (13:55 +0000)]
1. Fix a comment.  Locking _is_ needed (but not done).
2. Update a comment.  We now restore much more than RTC updates and
   interrupts.
3. Order change.  Stop interrupts by writing to RTC_STATUSB,
   restore rate bits for the interrupts by writing to RTC_STATUSA,
   then enable interrupts again.
   This seems to be done perfectly backwards in startrtclock().
   Otherwise, the idea for this change was obtained from
   startrtclock().
4. Don't stop the clock (RTCB_HALT).  We only program some control bits
   and don't want to stop the clock.
5. (Not really related.)  Add caveats to the comment about timer_restore().
   The update is non-atomic since locking is not done.

On locking:
6. rtcin() and writertc() are locked() adequately by splhigh() in RELENG_4,
   but this locking is null in -current.
7. Doing things in the correct order in (3) combined with (6) is probably
   enough locking for rtcrestore() in RELENG_4.  In -current, the
   writertc()'s race with rtcintr() unless the BIOS disables RTC interrupts.

Submitted by: bde (including commit message)
MFC after: 1 week

21 years agoAdded the new variable CTAGS which, if set to "ctags", reverts
ru [Thu, 17 Oct 2002 13:48:13 +0000 (13:48 +0000)]
Added the new variable CTAGS which, if set to "ctags", reverts
to creating the tags file using ctags(1).  Defaults to "gtags".
Made GTAGSFLAGS and HTAGSFLAGS overrideable, added CTAGSFLAGS.
Folded bsd.prog.mk version of `tags' into bsd.dep.mk.

PR: bin/42852

21 years agoFixed the quoting of the value of SC_CUT_SEPCHARS. The double quotes
bde [Thu, 17 Oct 2002 13:47:31 +0000 (13:47 +0000)]
Fixed the quoting of the value of SC_CUT_SEPCHARS.  The double quotes
needed to be quoted (to get a C string literal), not the value itself.

Fixed the value of SC_CUT_SEPCHARS.  Setting this value would have had no
effect even if it were used, since the value was the same as the default.

The above bugs had no effect except to set bad examples, since test
coverage of SC_CUT_SEPCHARS is broken by enabling a negative option.

Removed (unquoted) double quotes for all options.  They were all bogus
since they had no effect except to make non-strings look like strings.
Most of the non-strings were expressions.  The value of INIT_PATH is
a non-string since it is stringified later (unlike SC_CUT_SEPCHARS).

Fixed parenthesization errors inside bogus quotes (parenthesize values
if they have more than one token in them but don't parenthesize single
tokens).

21 years agoRemove unused #includes. Eliminate castings by using size_t instead of int.
charnier [Thu, 17 Oct 2002 13:19:47 +0000 (13:19 +0000)]
Remove unused #includes. Eliminate castings by using size_t instead of int.

21 years agoUse .Fl/Ar for flags and arguments.
charnier [Thu, 17 Oct 2002 13:04:49 +0000 (13:04 +0000)]
Use .Fl/Ar for flags and arguments.

21 years agoFix off-by-one error when pushing back a multibyte sequence in
tjr [Thu, 17 Oct 2002 13:04:00 +0000 (13:04 +0000)]
Fix off-by-one error when pushing back a multibyte sequence in
wide character class (%l[) and wide string (%ls) conversions.

21 years agoIntroduce option enumeration the standard way.
charnier [Thu, 17 Oct 2002 12:58:14 +0000 (12:58 +0000)]
Introduce option enumeration the standard way.

21 years agoMerged from sys/isa/syscons_isa.c revision 1.19.
nyan [Thu, 17 Oct 2002 12:54:51 +0000 (12:54 +0000)]
Merged from sys/isa/syscons_isa.c revision 1.19.

21 years agoMerged from sys/dev/syscons/syscons.c revision 1.390.
nyan [Thu, 17 Oct 2002 12:51:43 +0000 (12:51 +0000)]
Merged from sys/dev/syscons/syscons.c revision 1.390.

21 years agoMake part of the previous change clearer; check flags for SUPPRESS directly
tjr [Thu, 17 Oct 2002 12:06:29 +0000 (12:06 +0000)]
Make part of the previous change clearer; check flags for SUPPRESS directly
instead of checking whether we're using a temporary buffer.

21 years agoThe field width for single-byte string conversions (%c, %s, %[) is the
tjr [Thu, 17 Oct 2002 12:02:36 +0000 (12:02 +0000)]
The field width for single-byte string conversions (%c, %s, %[) is the
maximum number of bytes that may be stored in the array, not the maximum
number of wide characters to read. The wording of the standard unfortunately
does not make this clear.

21 years agoAdd new syscons option SC_NO_SUSPEND_VTYSWITCH.
iwasaki [Thu, 17 Oct 2002 07:04:31 +0000 (07:04 +0000)]
Add new syscons option SC_NO_SUSPEND_VTYSWITCH.
This disables vty switch during suspend/resume.

21 years agoput an #error directive when SMP and CPU_DISABLE_CMPXCHG are set
pirzyk [Thu, 17 Oct 2002 05:51:36 +0000 (05:51 +0000)]
put an #error directive when SMP and CPU_DISABLE_CMPXCHG are set
together.

Requested by: Lars Eggart <larse@isi.edu>
Enlighted how to do it by: John Baldwin <jhb@freebsd.org>

21 years agoMake kqueue(2) work for virtual terminals.
kbyanc [Thu, 17 Oct 2002 05:48:56 +0000 (05:48 +0000)]
Make kqueue(2) work for virtual terminals.

PR: 40486
Submitted by: Jilles Tjoelker <jilles+fbsd-bugs@stack.nl>

21 years agoDo not emit values as var=value while everything else looks like
marcel [Thu, 17 Oct 2002 05:41:10 +0000 (05:41 +0000)]
Do not emit values as var=value while everything else looks like
XML. Emit the values as <var>value</var> for consistency.

21 years agoSuppress the uninitialized variable warning on ia64 introduced by
marcel [Thu, 17 Oct 2002 03:45:34 +0000 (03:45 +0000)]
Suppress the uninitialized variable warning on ia64 introduced by
the previous commit by initializing i and j. These initializations
will normally be eliminated by the compiler.

21 years agoAdd a man page for ugidfw(8).
chris [Thu, 17 Oct 2002 01:54:37 +0000 (01:54 +0000)]
Add a man page for ugidfw(8).

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

21 years agoReplace old, not-working URL for an online index of DNS blacklist
keramida [Wed, 16 Oct 2002 22:52:56 +0000 (22:52 +0000)]
Replace old, not-working URL for an online index of DNS blacklist
servers with a reference to the Google directory.

Reviewed by: gshapiro
MFC after: 1 week

21 years agode-__P().
alfred [Wed, 16 Oct 2002 22:27:27 +0000 (22:27 +0000)]
de-__P().

21 years agode-__P()
alfred [Wed, 16 Oct 2002 22:26:32 +0000 (22:26 +0000)]
de-__P()

21 years agode-__P()
alfred [Wed, 16 Oct 2002 22:18:42 +0000 (22:18 +0000)]
de-__P()

21 years agoMFbed: Catch up with the latest changes in the english version
ue [Wed, 16 Oct 2002 21:11:08 +0000 (21:11 +0000)]
MFbed: Catch up with the latest changes in the english version

hardware/common/dev.sgml: 1.106 -> 1.108
relnotes/common/new.sgml: 1.432 -> 1.433

21 years agoFix three <para> that should have been </para>
ue [Wed, 16 Oct 2002 20:58:26 +0000 (20:58 +0000)]
Fix three <para> that should have been </para>

21 years agouse encapcheck.
ume [Wed, 16 Oct 2002 20:16:49 +0000 (20:16 +0000)]
use encapcheck.

Obtained from: KAME
MFC after: 3 weeks

21 years agoFix a fairly subtle bug in mbuf_init() where the reference counter
bmilekic [Wed, 16 Oct 2002 19:59:08 +0000 (19:59 +0000)]
Fix a fairly subtle bug in mbuf_init() where the reference counter
contiguous space was being allocated from the clust_map
instead of the mbuf_map as the comments indicated.  This resulted in
some address space wastage in mbuf_map.

Submitted by: Rohit Jalan <rohjal@yahoo.co.in>

21 years ago- after gif_set_tunnel(), psrc/pdst may be null. set IFF_RUNNING accordingly.
ume [Wed, 16 Oct 2002 19:49:37 +0000 (19:49 +0000)]
- after gif_set_tunnel(), psrc/pdst may be null.  set IFF_RUNNING accordingly.
- set IFF_UP on SIOCSIFADDR.  be consistent with others.
- set if_addrlen explicitly (just in case)
- multi destination mode is long gone.
- missing break statement
- add gif_set_tunnel(), so that we can set tunnel address from within the
  kernel at ease.
- encap_attach/detach dynamically on ioctls
- move encap_attach() to dedicated function in in*_gif.c

Obtained from: KAME
MFC after: 3 weeks

21 years agoFix oops in my last commit, I was calculating a new length but then not
dillon [Wed, 16 Oct 2002 19:16:33 +0000 (19:16 +0000)]
Fix oops in my last commit, I was calculating a new length but then not
using it.  (The code is already correct in -stable).

Found by: silby

21 years agoWhitespace.
jhb [Wed, 16 Oct 2002 19:11:59 +0000 (19:11 +0000)]
Whitespace.

21 years agoUse the global pcib devclass instead of our own static copy.
jhb [Wed, 16 Oct 2002 18:38:35 +0000 (18:38 +0000)]
Use the global pcib devclass instead of our own static copy.

21 years agoAdd hardware entries for trm and ubsa devices.
bmah [Wed, 16 Oct 2002 18:06:13 +0000 (18:06 +0000)]
Add hardware entries for trm and ubsa devices.

Submitted by: kan (ubsa), cognet (trm)

21 years agoNew release notes: ubsa driver, trm driver, bzgrep(1) et al.
bmah [Wed, 16 Oct 2002 18:05:20 +0000 (18:05 +0000)]
New release notes:  ubsa driver, trm driver, bzgrep(1) et al.

Modified release notes:  gdb 5.2.1.

MFCs noted:  matcd(4) removal, cvs 1.11.2.

Submitted by: kan (ubsa), cognet (trm)

21 years agoAdd "pci108e,8000" (psycho) and "pci108e,a000" (US-IIi sabre) to the list of
tmm [Wed, 16 Oct 2002 17:37:50 +0000 (17:37 +0000)]
Add "pci108e,8000" (psycho) and "pci108e,a000" (US-IIi sabre) to the list of
recognized compat properties. This should make the psycho driver attach
properly on SPARCengine Ultra AX machines.
Switch to a table-driven logic to recognize the ID's, since their number
is now large enough to justify this.

These changes are analogous to those made in NetBSD r.1.35, but
implemented a bit differently.

21 years agoCatch up to changes in acpivar.h to add support for using ACPI on
jhb [Wed, 16 Oct 2002 17:28:53 +0000 (17:28 +0000)]
Catch up to changes in acpivar.h to add support for using ACPI on
4-stable systems.

Sponsored by: The Weather Channel

21 years agostruct timecounter is defined in sys/time.h on 4-stable.
jhb [Wed, 16 Oct 2002 17:27:40 +0000 (17:27 +0000)]
struct timecounter is defined in sys/time.h on 4-stable.

Sponsored by: The Weather Channel

21 years ago- Include lock headers on current in this header instead of putting
jhb [Wed, 16 Oct 2002 17:27:11 +0000 (17:27 +0000)]
- Include lock headers on current in this header instead of putting
  #if's in all the other ACPI source files.
- Use splhigh() for the ACPI subsystem lock on 4-stable.

Sponsored by: The Weather Channel

21 years ago- In AcpiOsGetThreadId(), allow for curproc to be NULL on 4-stable systems.
jhb [Wed, 16 Oct 2002 17:25:29 +0000 (17:25 +0000)]
- In AcpiOsGetThreadId(), allow for curproc to be NULL on 4-stable systems.
  In that case use proc0's pid to return the thread ID.
- For 4-stable, use the generic swi taskqueue for ACPI events rather than
  implementing our own.

Sponsored by: The Weather Channel

21 years agoImplement the internal locks of an ACPI semaphore with splhigh() on
jhb [Wed, 16 Oct 2002 17:23:34 +0000 (17:23 +0000)]
Implement the internal locks of an ACPI semaphore with splhigh() on
4-stable.  -current uses a mutex for this purpose.

Sponsored by: The Weather Channel

21 years ago- curproc may be NULL in 4-stable. In that case use the vmspace from
jhb [Wed, 16 Oct 2002 17:22:03 +0000 (17:22 +0000)]
- curproc may be NULL in 4-stable.  In that case use the vmspace from
  proc0.
- Remove unused include.

Sponsored by: The Weather Channel

21 years agoInclude <sys/select.h> on -stable instead of <sys/selinfo.h> to get the
jhb [Wed, 16 Oct 2002 17:20:43 +0000 (17:20 +0000)]
Include <sys/select.h> on -stable instead of <sys/selinfo.h> to get the
definition of struct selinfo.

Sponsored by: The Weather Channel

21 years agowhitespace/indentation fix from phk
sam [Wed, 16 Oct 2002 17:07:41 +0000 (17:07 +0000)]
whitespace/indentation fix from phk

21 years agoUse a linked list to keep the psycho softcs instead of a statically
tmm [Wed, 16 Oct 2002 17:03:36 +0000 (17:03 +0000)]
Use a linked list to keep the psycho softcs instead of a statically
sized array.
While being there, deuglify the psycho pair detection loop which became
quite awkward in a previous code reorganization.

21 years agomdoc(7) police: markup and spelling fixes.
ru [Wed, 16 Oct 2002 16:46:58 +0000 (16:46 +0000)]
mdoc(7) police: markup and spelling fixes.

Prodded by: phk

21 years agoAdd a missing PROC_UNLOCK in ptrace() for the PT_IO case.
jhb [Wed, 16 Oct 2002 16:28:33 +0000 (16:28 +0000)]
Add a missing PROC_UNLOCK in ptrace() for the PT_IO case.

PR: kern/44065
Submitted by: Mark Kettenis <kettenis@chello.nl>

21 years agoRemove programs that no longer need a kerberisation option.
markm [Wed, 16 Oct 2002 16:12:22 +0000 (16:12 +0000)]
Remove programs that no longer need a kerberisation option.

21 years agoDekerberise. The corresponding userland stuff has been dekerberised
markm [Wed, 16 Oct 2002 16:10:46 +0000 (16:10 +0000)]
Dekerberise. The corresponding userland stuff has been dekerberised
for ages, and no-one seems to have noticed. Viva PAM!

21 years agoDekerberise. The corresponding userland stuff has been dekerberised
markm [Wed, 16 Oct 2002 16:09:05 +0000 (16:09 +0000)]
Dekerberise. The corresponding userland stuff has been dekerberised
for ages, and no-one seems to have noticed. Viva PAM!

Fix some easy/trivial warnings while I'm here.

21 years agos/rpc.mountd/mountd/
charnier [Wed, 16 Oct 2002 16:04:50 +0000 (16:04 +0000)]
s/rpc.mountd/mountd/
Add FBSDID
Reorder #if/#endif around sccsid to conform style(9)

21 years agoUpdated the CPU_DISABLE_CMPXCHG notes to warn users not to also enable
pirzyk [Wed, 16 Oct 2002 15:59:07 +0000 (15:59 +0000)]
Updated the CPU_DISABLE_CMPXCHG notes to warn users not to also enable
SMP with it.

Requested by: Lars Eggert <larse@ISI.EDU>

21 years agoArrange to not have /* inside a comment. Typo fix.
charnier [Wed, 16 Oct 2002 15:56:13 +0000 (15:56 +0000)]
Arrange to not have /* inside a comment. Typo fix.