]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoThis commit was generated by cvs2svn to compensate for changes in r89010,
sheldonh [Mon, 7 Jan 2002 14:26:12 +0000 (14:26 +0000)]
This commit was generated by cvs2svn to compensate for changes in r89010,
which included commits to RCS files with non-trunk default branches.

22 years agoImport patch from vendor (taken from nvi-1.81.5) that fixes problematic
sheldonh [Mon, 7 Jan 2002 14:26:12 +0000 (14:26 +0000)]
Import patch from vendor (taken from nvi-1.81.5) that fixes problematic
interaction between the leftright and number options.

PR: bin/23912
Reported by: "Stephen D. Spencer" <gladiatr@boneyard.lawrence.ks.us>
Obtained from: skimo@kotnet.org

22 years agoClean up G. B. Shaw quotes:
tg [Mon, 7 Jan 2002 13:47:22 +0000 (13:47 +0000)]
Clean up G. B. Shaw quotes:
- fix missing punctuation,
- remove duplicates in fortunes2,
- don't abbreviate first names.

Remove duplication of Shaw's principle in fortunes2.

22 years agoAdd exim to the list of packages that's guaranteed to be on the first
sheldonh [Mon, 7 Jan 2002 13:43:45 +0000 (13:43 +0000)]
Add exim to the list of packages that's guaranteed to be on the first
CDROM.

Approved by: steve

22 years agoAdd some missing years, fix Bach's birthday entry.
tg [Mon, 7 Jan 2002 13:35:31 +0000 (13:35 +0000)]
Add some missing years, fix Bach's birthday entry.

22 years agoAccording to jhb, the alpha bootstrap code depends on the root
sheldonh [Mon, 7 Jan 2002 12:25:30 +0000 (12:25 +0000)]
According to jhb, the alpha bootstrap code depends on the root
filesystem using a block size of 8192.  Since this seems unlikely to
be fixed soon (specifically in time for 4.5-RELEASE on the RELENG_4
branch), fall back to the old default block and frag sizes of 8192 and
1024 in sysinstall on the alpha.

Reported by: jhb

22 years agoComment out an example that was missed on first import.
sheldonh [Mon, 7 Jan 2002 08:41:55 +0000 (08:41 +0000)]
Comment out an example that was missed on first import.

Reported by: Jeroen Ruigrok <asmodai@wxs.nl>

22 years agoAdd 'R'ecover option that deletes a partition and attempts
dillon [Mon, 7 Jan 2002 07:51:24 +0000 (07:51 +0000)]
Add 'R'ecover option that deletes a partition and attempts
to recover its space into the previous partition.  Revert 'D'elete
to not attempt to recover any space.

Do not auto-create /home as per release engineers decision (though
I think this is a mistake).  However, all of this code will be
replaced later on anyway either with Jordan's stuff or with
some other sort of templater, so it isn't a big deal.

22 years agoModify Delete_Chunk() into Delete_Chunk2() which can take a flags
dillon [Mon, 7 Jan 2002 07:47:25 +0000 (07:47 +0000)]
Modify Delete_Chunk() into Delete_Chunk2() which can take a flags
argument.  Leave a compatibility shim for Delete_Chunk().

Implement DELCHUNK_RECOVER flag so sysinstall can ask libdisk
to recover space when deleting a chunk.

22 years agoFix a missing "ipfw:" in a syslog message.
cjc [Mon, 7 Jan 2002 07:12:09 +0000 (07:12 +0000)]
Fix a missing "ipfw:" in a syslog message.

MFC after: 1 day

22 years agotunefs no longer outputs a warning if one tries to set soft-updates on
dd [Mon, 7 Jan 2002 07:01:52 +0000 (07:01 +0000)]
tunefs no longer outputs a warning if one tries to set soft-updates on
an unmounted filesystem.

PR: 32266
Submitted by: Maxim Konovalov <maxim@macomnet.ru>

22 years agoIn the words of the submitter:
dd [Mon, 7 Jan 2002 06:51:19 +0000 (06:51 +0000)]
In the words of the submitter:

The first "synopsis" example has a "[/prefixlength]" which shouldn't
be there, since that stuff is part of the preceeding "address" as is
explained in the description of "address".

(The way it is now, 192.168.0.1/16/prefixlength would be a proper
operand.  Note that "prefixlength" is not mentioned by name anywhere.)

PR: 32462
Submitted by: Gary W. Swearingen <swear@blarg.net>

22 years agoLink loop(4) to lo(4). The device's name is 'loop', which is what one
dd [Mon, 7 Jan 2002 06:48:30 +0000 (06:48 +0000)]
Link loop(4) to lo(4).  The device's name is 'loop', which is what one
(at least a new one) would expect the manual page to be called (even
if the device is lo#).

PR: 32453
Submitted by: Gary W. Swearingen <swear@blarg.net>

22 years agoIn the words of the submitter:
dd [Mon, 7 Jan 2002 06:45:17 +0000 (06:45 +0000)]
In the words of the submitter:

disklabel(8)'s "Reading the disk label" section starts out "To examine
or save the label on a disk drive,...".  This is confusing.  The given
command (disklabel [-r] disk) doesn't save anything (except to standard
out, but that should go without saying).  It reads as if the command
might save something on the disk drive.

PR: 32452
Submitted by: Gary W. Swearingen <swear@blarg.net>

22 years agoAssorted grammar and other mechanical fixes.
dd [Mon, 7 Jan 2002 06:26:37 +0000 (06:26 +0000)]
Assorted grammar and other mechanical fixes.

PR: 32941
Submitted by: <setantae@submonkey.net>

22 years agoWe are munmap(2), so there's no need to list ourselves in the SEE ALSO
dd [Mon, 7 Jan 2002 06:12:25 +0000 (06:12 +0000)]
We are munmap(2), so there's no need to list ourselves in the SEE ALSO
section; instead, list our partner in crime, mmap(2).

PR: 33153
Submitted by: Faried Nawaz <fn@hungry.org>

22 years agoNuke the paragraph that says "One can obtain user connection request
dd [Mon, 7 Jan 2002 06:10:37 +0000 (06:10 +0000)]
Nuke the paragraph that says "One can obtain user connection request
data without confirming the connection by issuing a recvmsg(2) [...]".
There's no such code in the kernel.

PR: 26861
Submitted by: Richard A Steenbergen <ras@e-gerbil.net>,
Tom Rhodes <darklogik@pittgoth.com>

22 years agoSuggest using memmove(3) if src and dst may overlap.
dd [Mon, 7 Jan 2002 06:03:37 +0000 (06:03 +0000)]
Suggest using memmove(3) if src and dst may overlap.

PR: 33298
Submitted by: Tim J. Robbins <tim@robbins.dropbear.id.au>

22 years agoMFalpha 1.18: Bring this closer to reality by removing mention of wd#
dd [Mon, 7 Jan 2002 05:45:38 +0000 (05:45 +0000)]
MFalpha 1.18: Bring this closer to reality by removing mention of wd#
and adding ATA stuff.

22 years agoATAPI CDROM drives are acd#, not act#.
dd [Mon, 7 Jan 2002 05:43:31 +0000 (05:43 +0000)]
ATAPI CDROM drives are acd#, not act#.

22 years agoRemove duplicate entry for Descartes.
grog [Mon, 7 Jan 2002 05:07:03 +0000 (05:07 +0000)]
Remove duplicate entry for Descartes.

PR: 33630
Submitted by: Christopher Elkins <chrise@scardini.com>
MFC after: 2 weeks

22 years agoPut URL references in footnotes (rather than in parentheses following
bmah [Mon, 7 Jan 2002 03:51:41 +0000 (03:51 +0000)]
Put URL references in footnotes (rather than in parentheses following
the link text), and put footnotes at the bottom of each page.  This
change affects the release documentation only.

22 years agoExtend Olympus E-100RS quirk to cover entire E series of digital cameras.
kbyanc [Mon, 7 Jan 2002 03:32:56 +0000 (03:32 +0000)]
Extend Olympus E-100RS quirk to cover entire E series of digital cameras.

22 years agoRemoved <sys/queue.h> from the list of required include files and
davidc [Sun, 6 Jan 2002 23:56:12 +0000 (23:56 +0000)]
Removed <sys/queue.h> from the list of required include files and
corrected the name of the function (get_cyclecount()).

Requested by: bde

22 years agoAll information from the "BUGS" section not belonging in "SECURITY
chris [Sun, 6 Jan 2002 21:48:37 +0000 (21:48 +0000)]
All information from the "BUGS" section not belonging in "SECURITY
CONSIDERATIONS" moved to "COMPATIBILITY".

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

22 years agoAdd new "SECURITY CONSIDERATIONS" sections.
chris [Sun, 6 Jan 2002 21:16:41 +0000 (21:16 +0000)]
Add new "SECURITY CONSIDERATIONS" sections.

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

22 years ago o Add missing synchronization (splnet()/splx()) in aio_free_entry().
alc [Sun, 6 Jan 2002 21:03:39 +0000 (21:03 +0000)]
 o Add missing synchronization (splnet()/splx()) in aio_free_entry().
 o Move the definition of struct aiocblist from sys/aio.h to kern/vfs_aio.c.
 o Make aio_swake_cb() static.

22 years agos/oferride/override/
imp [Sun, 6 Jan 2002 20:33:55 +0000 (20:33 +0000)]
s/oferride/override/

submitted by: dima

22 years agoMinor doc nit: Replace contractions with their expanded forms.
chris [Sun, 6 Jan 2002 20:28:27 +0000 (20:28 +0000)]
Minor doc nit: Replace contractions with their expanded forms.

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

22 years agoAdd note about new rc.conf for networkfs. Fix typos from PR:33499
imp [Sun, 6 Jan 2002 20:18:13 +0000 (20:18 +0000)]
Add note about new rc.conf for networkfs.  Fix typos from PR:33499

22 years agoMake the utility name in the usage() diagnostic match the actual
mike [Sun, 6 Jan 2002 19:57:26 +0000 (19:57 +0000)]
Make the utility name in the usage() diagnostic match the actual
utility name.

Submitted by: 3d
PR: 33496

22 years agoAdd support for C++.
mike [Sun, 6 Jan 2002 19:04:12 +0000 (19:04 +0000)]
Add support for C++.

Submitted by: bde
PR: 33590

22 years agoUpdate length more correctly when parsing a cis info field.
imp [Sun, 6 Jan 2002 18:03:55 +0000 (18:03 +0000)]
Update length more correctly when parsing a cis info field.

Before, we were using
while (*p++ && --len > 0);
to do this.  However, len doesn't get decremented for the NUL byte, so when
we used len later to see if we still have CIS left for some optional fields,
we'd run off the end of an array and dump core.

Instead, replace it with
len -= strlen(p) + 1;
p += strlen(p) + 1;
which is more correct.  It is a little bogus to assume that p points to
a valid C string, but only a little.  The PC Card SPEC mandates that it
does, and we already depend on that with the use of strdup a few lines
earlier.  Since much of the rest of the cis parsing code isn't hyper
retentive about error checking, I'll leave that level of checking for
another time and/or another committer :-).

22 years agoMFS: Sync to 4-STABLE...wordsmithing and markup fixes, also mention
bmah [Sun, 6 Jan 2002 17:17:55 +0000 (17:17 +0000)]
MFS:  Sync to 4-STABLE...wordsmithing and markup fixes, also mention
manpages as a source for further reading.

article.sgml 1.1.2.9->1.1.2.10, 1.1.2.11->1.1.2.12

22 years agoMerged from sys/dev/sio/sio.c revisions 1.354 and 1.358.
nyan [Sun, 6 Jan 2002 09:54:50 +0000 (09:54 +0000)]
Merged from sys/dev/sio/sio.c revisions 1.354 and 1.358.

22 years agoMFi386: revisions 1.382 and 1.386.
nyan [Sun, 6 Jan 2002 09:47:36 +0000 (09:47 +0000)]
MFi386: revisions 1.382 and 1.386.

22 years agoSorted the lists.
nyan [Sun, 6 Jan 2002 09:46:30 +0000 (09:46 +0000)]
Sorted the lists.

22 years agoEnable WARNS?=2.
kbyanc [Sun, 6 Jan 2002 08:47:19 +0000 (08:47 +0000)]
Enable WARNS?=2.

22 years agoDocument sbuf_trim() and sbuf_vprintf(); add MLINKs.
kbyanc [Sun, 6 Jan 2002 08:40:42 +0000 (08:40 +0000)]
Document sbuf_trim() and sbuf_vprintf(); add MLINKs.
Clarify that the contents of a caller-supplied buffer are undefined and
should not be examined directly.

Reviewed by: des

22 years ago * Implement SBUF_AUTOEXTEND flag; sbufs created with this flag are
kbyanc [Sun, 6 Jan 2002 08:38:23 +0000 (08:38 +0000)]
 * Implement SBUF_AUTOEXTEND flag; sbufs created with this flag are
   automatically extended to prevent overflow.
 * Added sbuf_vprintf(); sbuf_printf() is now just a wrapper around
   sbuf_vprintf().
 * Include <stdio.h> and <string.h> when building libsbuf to silence
   WARNS=4 warnings.

Reviewed by: des

22 years agoReorder a calculation in sbreserve so that it does not overflow
silby [Sun, 6 Jan 2002 06:50:54 +0000 (06:50 +0000)]
Reorder a calculation in sbreserve so that it does not overflow
with multi-megabyte socket buffer sizes.

PR: 7420
MFC after: 3 weeks

22 years agoMerge the changes in the English version (1.247 --> 1.248).
kuriyama [Sun, 6 Jan 2002 04:30:46 +0000 (04:30 +0000)]
Merge the changes in the English version (1.247 --> 1.248).

22 years agoAvoid reentrantly sending on the same socket, which causes a kernel panic.
archie [Sun, 6 Jan 2002 01:08:30 +0000 (01:08 +0000)]
Avoid reentrantly sending on the same socket, which causes a kernel panic.

22 years ago- Teach SIGIO code to use cr_cansignal() instead of a custom CANSIGIO()
rwatson [Sun, 6 Jan 2002 00:54:46 +0000 (00:54 +0000)]
- Teach SIGIO code to use cr_cansignal() instead of a custom CANSIGIO()
  macro.  As a result, mandatory signal delivery policies will be
  applied consistently across the kernel.

- Note that this subtly changes the protection semantics, and we should
  watch out for any resulting breakage.  Previously, delivery of SIGIO
  in this circumstance was limited to situations where the subject was
  privileged, or where one of the subject's (ruid, euid) matched one
  of the object's (ruid, euid).  In the new scenario, subject (ruid, euid)
  are matched against the object's (ruid, svuid), and the object uid's
  must be a subset of the subject uid's.  Likewise, jail now affects
  delivery, and special handling for P_SUGID of the object is present.
  This change can always be reversed or tweaked if it proves to disrupt
  application behavior substantially.

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

22 years ago- Push much of the logic for p_cansignal() behind cr_cansignal, which
rwatson [Sun, 6 Jan 2002 00:20:12 +0000 (00:20 +0000)]
- Push much of the logic for p_cansignal() behind cr_cansignal, which
  authorized based on a subject credential rather than a subject process.
  This will permit the same logic to be reused in situations where only
  the credential generating the signal is available, such as in the
  delivery of SIGIO.
- Because of two clauses, the automatic success against curproc,
  and the session semantics for SIGCONT, not all logic can be pushed
  into cr_cansignal(), but those cases should not apply for most other
  consumers of cr_cansignal().
- This brings the base system inter-process authorization code more
  into line with the MAC implementation.

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

22 years agoRelease text vnode in exit() rather than wait(). Occasionally
dwmalone [Sat, 5 Jan 2002 21:47:58 +0000 (21:47 +0000)]
Release text vnode in exit() rather than wait(). Occasionally
fifesystem problems could prevent the release from completing and
this could result in init being blocked indefinitely.

This was looked over by Matt ages ago.

Approved by: dillon

22 years agoAdd a new `SECURITY CONSIDERATIONS' section. Sample code similar to
chris [Sat, 5 Jan 2002 20:44:34 +0000 (20:44 +0000)]
Add a new `SECURITY CONSIDERATIONS' section.  Sample code similar to
the first revision of strcpy(3)'s section is included, but should be
removed as the Security Architecture document is committed and
completed.

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

22 years agoInclude a section cross-reference to `EXAMPLES' in `SECURITY
chris [Sat, 5 Jan 2002 20:36:47 +0000 (20:36 +0000)]
Include a section cross-reference to `EXAMPLES' in `SECURITY
CONSIDERATIONS' as a source for relevant sample code.

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

22 years agoRemove the example from SECURITY CONSIDERATIONS. This is mostly
chris [Sat, 5 Jan 2002 20:24:38 +0000 (20:24 +0000)]
Remove the example from SECURITY CONSIDERATIONS.  This is mostly
duplicated code from EXAMPLES.

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

22 years agoAs discussed with Bruce Evans and Kirk McKusick:
chris [Sat, 5 Jan 2002 20:18:35 +0000 (20:18 +0000)]
As discussed with Bruce Evans and Kirk McKusick:
o Combine ufs.7 and ffs.7 into a single ffs.7 man page.
o Remove all references to `ufs' as a file system.
o Proper (lack of) capitalization for `ffs'.

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

22 years agoPut -Wno-uninitialized at the end of the warns flags as it may be enabeled
dwmalone [Sat, 5 Jan 2002 20:17:50 +0000 (20:17 +0000)]
Put -Wno-uninitialized at the end of the warns flags as it may be enabeled
by several flags (in this case -W and -Wall).

Reviewed by: bde

22 years agoBe more careful about freeing memory after parsing commands.
dwmalone [Sat, 5 Jan 2002 20:13:01 +0000 (20:13 +0000)]
Be more careful about freeing memory after parsing commands.
Hiroyuki YAMAMORI gave a patch for the EPRT command in the
PR below. Problems with the rest of the patch are my fault.

PR: 33268
Reviewed by: iedowse, sheldonh

22 years agoNew (but very overdue) release note: ata(4) 48-bit addressing.
bmah [Sat, 5 Jan 2002 18:46:03 +0000 (18:46 +0000)]
New (but very overdue) release note:  ata(4) 48-bit addressing.

MFC noted:  locale renaming.

22 years agoPre-calculate the checksum for multicast packets sourced on a
fenner [Sat, 5 Jan 2002 18:23:53 +0000 (18:23 +0000)]
Pre-calculate the checksum for multicast packets sourced on a
multicast router.  This is overkill; it should be possible to
delay to hardware interfaces and only pre-calculate when forwarding
to a tunnel.

22 years agoUpdate Judaic for year 2002. Name Josef as maintainer.
dwmalone [Sat, 5 Jan 2002 18:10:19 +0000 (18:10 +0000)]
Update Judaic for year 2002. Name Josef as maintainer.

PR: 33389
Submitted by: Josef Grosch <jgrosch@mooseriver.com>

22 years agoFix a bug where the mutex name wasn't always displayed for processes in
jhb [Sat, 5 Jan 2002 17:18:59 +0000 (17:18 +0000)]
Fix a bug where the mutex name wasn't always displayed for processes in
SMTX in utils such as ps and top.  The KI_CTTY flag was assigned to
kinfo_proc->ki_kiflag rather than or'd into the flag, thus clobbering
any flags set earlier, including KI_MTXBLOCK.

Prodding by: peter

22 years agoIn the word selection mode don't append newline if the word ends at the
sobomax [Sat, 5 Jan 2002 16:52:44 +0000 (16:52 +0000)]
In the word selection mode don't append newline if the word ends at the
screen boundary.

22 years agoWrite the "msinfo" result to stdout instead of stderr
sos [Sat, 5 Jan 2002 15:21:05 +0000 (15:21 +0000)]
Write the "msinfo" result to stdout instead of stderr

22 years agoPut the "mtxname" keyword in alphabetical order (t comes after s) so
peter [Sat, 5 Jan 2002 12:19:51 +0000 (12:19 +0000)]
Put the "mtxname" keyword in alphabetical order (t comes after s) so
that the keyword is recognized.

22 years agoConvert a bunch of 1 << PCPU_GET(cpuid) to PCPU_GET(cpumask).
peter [Sat, 5 Jan 2002 09:41:37 +0000 (09:41 +0000)]
Convert a bunch of 1 << PCPU_GET(cpuid) to PCPU_GET(cpumask).

22 years agoFix forward_roundrobin(). It was mistakenly using the cpu number as
peter [Sat, 5 Jan 2002 09:38:47 +0000 (09:38 +0000)]
Fix forward_roundrobin().  It was mistakenly using the cpu number as
though it was a mask.  As a result, we sent AST IPI's to the wrong
cpu and/or left out some.

Spotted by: jake

22 years agoAdd a per-cpu variable, cpumask, the preshifted equivalent of 1 << cpuid.
peter [Sat, 5 Jan 2002 09:35:50 +0000 (09:35 +0000)]
Add a per-cpu variable, cpumask, the preshifted equivalent of 1 << cpuid.
We use this around the place a lot.

22 years agoChange the preemption code for software interrupt thread schedules and
jhb [Sat, 5 Jan 2002 08:47:13 +0000 (08:47 +0000)]
Change the preemption code for software interrupt thread schedules and
mutex releases to not require flags for the cases when preemption is
not allowed:

The purpose of the MTX_NOSWITCH and SWI_NOSWITCH flags is to prevent
switching to a higher priority thread on mutex releease and swi schedule,
respectively when that switch is not safe.  Now that the critical section
API maintains a per-thread nesting count, the kernel can easily check
whether or not it should switch without relying on flags from the
programmer.  This fixes a few bugs in that all current callers of
swi_sched() used SWI_NOSWITCH, when in fact, only the ones called from
fast interrupt handlers and the swi_sched of softclock needed this flag.
Note that to ensure that swi_sched()'s in clock and fast interrupt
handlers do not switch, these handlers have to be explicitly wrapped
in critical_enter/exit pairs.  Presently, just wrapping the handlers is
sufficient, but in the future with the fully preemptive kernel, the
interrupt must be EOI'd before critical_exit() is called.  (critical_exit()
can switch due to a deferred preemption in a fully preemptive kernel.)

I've tested the changes to the interrupt code on i386 and alpha.  I have
not tested ia64, but the interrupt code is almost identical to the alpha
code, so I expect it will work fine.  PowerPC and ARM do not yet have
interrupt code in the tree so they shouldn't be broken.  Sparc64 is
broken, but that's been ok'd by jake and tmm who will be fixing the
interrupt code for sparc64 shortly.

Reviewed by: peter
Tested on: i386, alpha

22 years agoRemove brain damaged code in witness_lock(). We could have easily
jhb [Sat, 5 Jan 2002 08:29:54 +0000 (08:29 +0000)]
Remove brain damaged code in witness_lock().  We could have easily
just used PCPU_GET(spinlocks) w/o needing the w_mtx held.  It is more
correct to just check td_critnest now though.

22 years agoMerge the changes in the English version (1.244 --> 1.247).
kuriyama [Sat, 5 Jan 2002 08:01:41 +0000 (08:01 +0000)]
Merge the changes in the English version (1.244 --> 1.247).

22 years agoAdd a new flag '-c' to disable repeated line compression when the output
archie [Sat, 5 Jan 2002 07:34:52 +0000 (07:34 +0000)]
Add a new flag '-c' to disable repeated line compression when the output
is a pipe to another program, or, if specified twice, in all cases.

PR: bin/32420

22 years agoGC unfinished function selected_proc_ipi(). It is a duplicate of
peter [Sat, 5 Jan 2002 06:44:27 +0000 (06:44 +0000)]
GC unfinished function selected_proc_ipi(). It is a duplicate of
apic_ipi_singledest() anyway.

22 years agoMove initialization of the MKMODULESENV envorinoment to kern.pre.mk
imp [Sat, 5 Jan 2002 06:21:06 +0000 (06:21 +0000)]
Move initialization of the MKMODULESENV envorinoment to kern.pre.mk
from kern.post.mk so port makefiles can augment it.

Submitted by: nyan

22 years agoNew release notes: SA-02:01, SA-02:02.
bmah [Sat, 5 Jan 2002 06:07:17 +0000 (06:07 +0000)]
New release notes:  SA-02:01, SA-02:02.

22 years agoFully catch up to the recent critical section API change. Update the
jhb [Fri, 4 Jan 2002 23:56:17 +0000 (23:56 +0000)]
Fully catch up to the recent critical section API change.  Update the
content of the manpage and document cpu_critical_enter/exit.

22 years agoWe were right the first time...the HPT372 is an ATA133 controller.
bmah [Fri, 4 Jan 2002 21:55:32 +0000 (21:55 +0000)]
We were right the first time...the HPT372 is an ATA133 controller.

Pointed out by: Kent Stewart <kstewart@owt.com>

22 years agoOops, missed one point where $pwd_mkdb had been used.
yar [Fri, 4 Jan 2002 21:28:32 +0000 (21:28 +0000)]
Oops, missed one point where $pwd_mkdb had been used.
Now it should be @pwd_mkdb since we moved to Perl's
safe system() not using /bin/sh.

22 years agoMinor grammar and punctuation fixes
yar [Fri, 4 Jan 2002 18:17:07 +0000 (18:17 +0000)]
Minor grammar and punctuation fixes
in the SO_ACCEPTFILTER description.

22 years agoState clearly that one should call listen(2) on a socket
yar [Fri, 4 Jan 2002 18:12:38 +0000 (18:12 +0000)]
State clearly that one should call listen(2) on a socket
at first and try to set an accept_filter(9) on it only after that.
Also document errno value that will be set if installing the
filter on a non-listening socket.

22 years agoo Spelling fix in comment: tcp_ouput -> tcp_output
rwatson [Fri, 4 Jan 2002 17:21:27 +0000 (17:21 +0000)]
o Spelling fix in comment: tcp_ouput -> tcp_output

22 years agoArgh...argv->argc
guido [Fri, 4 Jan 2002 13:46:57 +0000 (13:46 +0000)]
Argh...argv->argc

22 years agoFix the coredump that occurs when, from the tfpt prompt, a 'c' command is
guido [Fri, 4 Jan 2002 13:42:14 +0000 (13:42 +0000)]
Fix the coredump that occurs when, from the tfpt prompt, a 'c' command is
issued without an argument.

22 years agoRegen.
sheldonh [Fri, 4 Jan 2002 13:11:36 +0000 (13:11 +0000)]
Regen.

22 years agoInstall nsmb.conf with mode 0600, since it may be modified to include
sheldonh [Fri, 4 Jan 2002 13:02:51 +0000 (13:02 +0000)]
Install nsmb.conf with mode 0600, since it may be modified to include
passwords for remote shares.

Reported by: Andre Albsmeier <andre@albsmeier.net>

22 years agoImport this patch to address user concerns.
darrenr [Fri, 4 Jan 2002 12:32:34 +0000 (12:32 +0000)]
Import this patch to address user concerns.

PR:  27615
Submitted by: Andria Thomas <andria@tovaris.com>
Approved by: Me.
MFC after: 7 days

22 years agoAxe a stale comment. Holding sched_lock across both setrunqueue() and
jhb [Fri, 4 Jan 2002 10:55:51 +0000 (10:55 +0000)]
Axe a stale comment.  Holding sched_lock across both setrunqueue() and
mi_switch() is sufficient.

22 years agoDon't turn on RB_CDROM if the rootdev is a BIOS cd device for now as this
jhb [Fri, 4 Jan 2002 07:59:27 +0000 (07:59 +0000)]
Don't turn on RB_CDROM if the rootdev is a BIOS cd device for now as this
breaks when cdboot is used with an MFS root.

22 years agoThrow the $FreeBSD$s back in, properly escaping them.
silby [Fri, 4 Jan 2002 05:27:47 +0000 (05:27 +0000)]
Throw the $FreeBSD$s back in, properly escaping them.

22 years agoRemove $FreeBSD$s from previous commit; perl thinks that they're
silby [Fri, 4 Jan 2002 01:40:50 +0000 (01:40 +0000)]
Remove $FreeBSD$s from previous commit; perl thinks that they're
something to be interpreted.  Urk.

22 years agoAdd the system default timer engine.
tanimura [Fri, 4 Jan 2002 01:16:18 +0000 (01:16 +0000)]
Add the system default timer engine.

22 years ago- Provide toggles to show debug messages. Set new sysctl variables
tanimura [Fri, 4 Jan 2002 01:13:49 +0000 (01:13 +0000)]
- Provide toggles to show debug messages. Set new sysctl variables
  hw.midi.debug and hw.midi.seq.debug to 1 to enable debug log.

- Make debug messages human-frendly.

- Implement /dev/music.

- Add a timer engine required by /dev/music.

- Fix nonblocking I/O.

- Fix the numbering of midi and synth devices.

22 years agoDo not derefer null.
tanimura [Fri, 4 Jan 2002 01:03:46 +0000 (01:03 +0000)]
Do not derefer null.

Reviewed by: des

22 years agounbreak -current (at least on my cvsups all today)
julian [Thu, 3 Jan 2002 23:33:44 +0000 (23:33 +0000)]
unbreak -current (at least on my cvsups all today)

22 years agoCorrect the path for the stable snapshot server.
jkh [Thu, 3 Jan 2002 23:33:21 +0000 (23:33 +0000)]
Correct the path for the stable snapshot server.

Noticed by:  "Peter Holm" <p_holm@mail.tele.dk>

22 years agoFix (I think) a typo in the hardware list (HPT372 is ATA100 instead of
bmah [Thu, 3 Jan 2002 23:24:03 +0000 (23:24 +0000)]
Fix (I think) a typo in the hardware list (HPT372 is ATA100 instead of
ATA133?).

MFC noted:  "more" chipsets supported by ata(4).

22 years agoNew release notes (and hardware list updates): dc(4) support
bmah [Thu, 3 Jan 2002 22:56:43 +0000 (22:56 +0000)]
New release notes (and hardware list updates):  dc(4) support
for Xircom 3201 and Conexant LANfinity RS7112, ifpi(4).

22 years agoSolve vnode_if.pl's identity crisis; make sure that it refers to itself
silby [Thu, 3 Jan 2002 21:53:09 +0000 (21:53 +0000)]
Solve vnode_if.pl's identity crisis; make sure that it refers to itself
as vnode_if.pl instead of vnode_if.sh.

PR: 33509
MFC after: 3 weeks

22 years agoAdd opie lib dir to includes list
ache [Thu, 3 Jan 2002 21:20:29 +0000 (21:20 +0000)]
Add opie lib dir to includes list

22 years agoImplement REDUCED INTERRUPT OPERATION usage form FC cards- this allows the
mjacob [Thu, 3 Jan 2002 20:43:22 +0000 (20:43 +0000)]
Implement REDUCED INTERRUPT OPERATION usage form FC cards- this allows the
firmware to delay completion of commands so that it can attempt to batch
a bunch of completions at once- either returning 16 bit handles in mailbox
registers, or in a resposne queue entry that has a whole wad of 16 bit handles.

Distinguish between 2300 and 2312 chipsets- if only because the revisions
on the chips have different meanings.

Add more instrumentation plus ISP_GET_STATS and ISP_CLR_STATS ioctls.
Run up the maximum number of response queue entities we'll look at
per interrupt.

If we haven't set HBA role yet, always return success from isp_fc_runstate.

MFC after: 2 weeks

22 years agoDon't build csu or libgcc on sparc64 for the time being.
jhb [Thu, 3 Jan 2002 19:57:35 +0000 (19:57 +0000)]
Don't build csu or libgcc on sparc64 for the time being.

22 years agoRemove unnecessary machine/bootinfo.h includes.
jhb [Thu, 3 Jan 2002 19:10:49 +0000 (19:10 +0000)]
Remove unnecessary machine/bootinfo.h includes.

Submitted by: jake

22 years agoRegenerate.
joe [Thu, 3 Jan 2002 18:58:08 +0000 (18:58 +0000)]
Regenerate.

22 years agoAdd a new linksys ethernet adapter: USB_PRODUCT_LINKSYS_USB10TX2.
joe [Thu, 3 Jan 2002 18:56:35 +0000 (18:56 +0000)]
Add a new linksys ethernet adapter: USB_PRODUCT_LINKSYS_USB10TX2.

22 years agoFix the footer in Japanese which is likely to mislead
hrs [Thu, 3 Jan 2002 17:42:38 +0000 (17:42 +0000)]
Fix the footer in Japanese which is likely to mislead
the readers.

22 years agoAdd some missing prototypes.
joe [Thu, 3 Jan 2002 16:14:35 +0000 (16:14 +0000)]
Add some missing prototypes.
Remove some diagnostics that aren't relevant yet.