]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoAdd an isp(4) module. sbus support is only compiled in on sparc64.
jhb [Thu, 31 Oct 2002 19:50:18 +0000 (19:50 +0000)]
Add an isp(4) module.  sbus support is only compiled in on sparc64.
pci support is always compiled in.

21 years agoplugged memory leakage in some erroneous cases
ume [Thu, 31 Oct 2002 19:45:48 +0000 (19:45 +0000)]
plugged memory leakage in some erroneous cases

Obtained from: KAME
MFC after: 1 week

21 years agoAdd a module for mpt(4).
jhb [Thu, 31 Oct 2002 19:39:23 +0000 (19:39 +0000)]
Add a module for mpt(4).

21 years agoComment out 'mpt' for the time being. This driver doesn't have a module
jhb [Thu, 31 Oct 2002 19:02:07 +0000 (19:02 +0000)]
Comment out 'mpt' for the time being.  This driver doesn't have a module
either.

21 years agoUse the strong symbol'd version of pthread_mutex_init so we don't
deischen [Thu, 31 Oct 2002 18:17:58 +0000 (18:17 +0000)]
Use the strong symbol'd version of pthread_mutex_init so we don't
accidentally call a stub or application provided version of the
same routine.

Submitted by: dfr

21 years agoInvoke 3 ACPI task threads as default if option ACPI_MAX_THREADS is
iwasaki [Thu, 31 Oct 2002 17:58:39 +0000 (17:58 +0000)]
Invoke 3 ACPI task threads as default if option ACPI_MAX_THREADS is
not defined.
To make previous default behavior (ACPI_MAX_THREADS undefined), define
option ACPI_MAX_THREADS as 0.

21 years agoGive a meaningfull diagnostic when we cannot determine the filesystem type.
phk [Thu, 31 Oct 2002 15:32:39 +0000 (15:32 +0000)]
Give a meaningfull diagnostic when we cannot determine the filesystem type.

21 years agoThere is no 'isp' module to move off to the 3rd floppy, so comment out the
jhb [Thu, 31 Oct 2002 14:38:21 +0000 (14:38 +0000)]
There is no 'isp' module to move off to the 3rd floppy, so comment out the
entry for now.

21 years agoIncorrect typecast.
emoore [Thu, 31 Oct 2002 14:10:00 +0000 (14:10 +0000)]
Incorrect typecast.

21 years agoDo not print a header line if it would be empty; required by 1003.1-2001.
tjr [Thu, 31 Oct 2002 13:41:37 +0000 (13:41 +0000)]
Do not print a header line if it would be empty; required by 1003.1-2001.

21 years agoBack out previous commit - it was not well thought out.
sobomax [Thu, 31 Oct 2002 13:05:56 +0000 (13:05 +0000)]
Back out previous commit - it was not well thought out.

Clue beaten in by: ume
Pointy hat to: sobomax

21 years agoEnable GEOM by default.
nyan [Thu, 31 Oct 2002 12:14:05 +0000 (12:14 +0000)]
Enable GEOM by default.

21 years agoAdd ATAPI/CAM man page.
thomas [Thu, 31 Oct 2002 10:52:47 +0000 (10:52 +0000)]
Add ATAPI/CAM man page.

Reviewed by: roberto

21 years agoDocument the new behavior of -C with respect to inode-change-time.
ru [Thu, 31 Oct 2002 10:43:32 +0000 (10:43 +0000)]
Document the new behavior of -C with respect to inode-change-time.
Document that -p also preserves the access time of the source.

21 years agoDo not change the target file's inode-change-time if -C is
ru [Thu, 31 Oct 2002 10:41:51 +0000 (10:41 +0000)]
Do not change the target file's inode-change-time if -C is
specified, the target file already exists and the files are
the same, and the target's file flags and mode need not to
be changed.

21 years agoFix POLA breakage in 1.29: IPv4 should be default. This makes `-4' option
sobomax [Thu, 31 Oct 2002 10:36:12 +0000 (10:36 +0000)]
Fix POLA breakage in 1.29: IPv4 should be default. This makes `-4' option
a nop, but we'll probably want to keep it for compatibility with other
KAME-based systems.

Complained about by: Andrey Lakhno <land@dnepr.net>
MFC after: 3 days

21 years agoade@ no longer has the FreeBSD account.
ru [Thu, 31 Oct 2002 08:56:34 +0000 (08:56 +0000)]
ade@ no longer has the FreeBSD account.

21 years agoSort by month/day/year/login.
ru [Thu, 31 Oct 2002 08:55:33 +0000 (08:55 +0000)]
Sort by month/day/year/login.

21 years agoKSE-enabled processes only.
davidxu [Thu, 31 Oct 2002 08:00:51 +0000 (08:00 +0000)]
KSE-enabled processes only.

21 years agoStyle(9) improvements.
phk [Thu, 31 Oct 2002 07:56:40 +0000 (07:56 +0000)]
Style(9) improvements.

21 years agoSet the sector size for the disk.
phk [Thu, 31 Oct 2002 07:55:38 +0000 (07:55 +0000)]
Set the sector size for the disk.

21 years agoRestore to pc98 support.
nyan [Thu, 31 Oct 2002 05:51:25 +0000 (05:51 +0000)]
Restore to pc98 support.

21 years agoUse ${MACHINE} variable instead of using '.if .endif' each machines.
nyan [Thu, 31 Oct 2002 05:43:49 +0000 (05:43 +0000)]
Use ${MACHINE} variable instead of using '.if .endif' each machines.

21 years agoAdd '#include <err.h>' for warn().
nyan [Thu, 31 Oct 2002 05:38:48 +0000 (05:38 +0000)]
Add '#include <err.h>' for warn().

21 years agoRename the speaker device for pc98 to 'pcspeaker'.
nyan [Thu, 31 Oct 2002 05:19:33 +0000 (05:19 +0000)]
Rename the speaker device for pc98 to 'pcspeaker'.

21 years agoMFi386: revision 1.55.
nyan [Thu, 31 Oct 2002 05:12:22 +0000 (05:12 +0000)]
MFi386: revision 1.55.
Rename from atspeaker to pcspeaker. (PC98 is not PC/AT)

21 years agoDisable COMPAT_FREEBSD4.
nyan [Thu, 31 Oct 2002 04:51:05 +0000 (04:51 +0000)]
Disable COMPAT_FREEBSD4.

21 years agoActually save the bootblock in the disk structure. Write the bootblock
jake [Thu, 31 Oct 2002 04:25:17 +0000 (04:25 +0000)]
Actually save the bootblock in the disk structure.  Write the bootblock
to the right place on the disk instead of srewn all over it.

21 years agoWrap test case in an if statement to avoid crunching a return value.
gordon [Thu, 31 Oct 2002 02:06:05 +0000 (02:06 +0000)]
Wrap test case in an if statement to avoid crunching a return value.

21 years agoMake __sF static. This can not be allowed to exist in 5.x.
alfred [Thu, 31 Oct 2002 01:54:27 +0000 (01:54 +0000)]
Make __sF static.  This can not be allowed to exist in 5.x.

21 years agoAdd tunable "hw.eisa_slots" to allow overriding the default number
peter [Thu, 31 Oct 2002 00:45:49 +0000 (00:45 +0000)]
Add tunable "hw.eisa_slots" to allow overriding the default number
of slots (10).

21 years agoRemove bogus advertising clauses.
thomas [Wed, 30 Oct 2002 22:27:54 +0000 (22:27 +0000)]
Remove bogus advertising clauses.

Reviewed by: n_hibma, roberto

21 years agoInitialize the new salt field in the lock sector.
phk [Wed, 30 Oct 2002 22:14:34 +0000 (22:14 +0000)]
Initialize the new salt field in the lock sector.

Sponsored by: DARPA & NAI Labs

21 years agoChange the kkey generation cherry-picker to use MD5.
phk [Wed, 30 Oct 2002 22:13:54 +0000 (22:13 +0000)]
Change the kkey generation cherry-picker to use MD5.

Sponsored by: DARPA & NAI Labs

21 years agoamr.c,
emoore [Wed, 30 Oct 2002 22:00:11 +0000 (22:00 +0000)]
amr.c,
amr_cam.c,
amrreg.h,
amrvar.h:
- added support for 12/16 byte cdb's, effecting CAM branch only ( non-disk support )

amrreg.h:
- increased number of scatter gather elements from 16 to 26.

amr_pci.c:
- amr_pci_free(), incorrect bus tag meant for 'amr_mailbox_dmat' was being freed

all:
- copyright change requested by scottl

Reviewed by: ps,scottl
MFC after: 1 week

21 years agoWith the recent libdisk changes, alpha doesn't need (as much) special magic.
phk [Wed, 30 Oct 2002 20:55:25 +0000 (20:55 +0000)]
With the recent libdisk changes, alpha doesn't need (as much) special magic.

21 years agoUntested alpha disk writer.
phk [Wed, 30 Oct 2002 20:54:22 +0000 (20:54 +0000)]
Untested alpha disk writer.

21 years agoWrite out 15 sectors of boot code.
phk [Wed, 30 Oct 2002 20:52:07 +0000 (20:52 +0000)]
Write out 15 sectors of boot code.

21 years agoCorrect path in previous commit. libssh lives under src/secure/lib, not
jhb [Wed, 30 Oct 2002 20:33:28 +0000 (20:33 +0000)]
Correct path in previous commit.  libssh lives under src/secure/lib, not
under src/lib.

Reported by: phk

21 years agoOK, you can use the 4.x boot loader and boot blocks to load kernels, but
imp [Wed, 30 Oct 2002 20:11:07 +0000 (20:11 +0000)]
OK, you can use the 4.x boot loader and boot blocks to load kernels, but
I'm nervous about recommending it, so note it as a footnote only for now.

21 years agoSave and restore the interrupt Edge/Level Control Registers (ELCR)
iedowse [Wed, 30 Oct 2002 19:55:06 +0000 (19:55 +0000)]
Save and restore the interrupt Edge/Level Control Registers (ELCR)
across system suspends on the Intel 82371AB PCI-ISA bridge. On a
Sony Vaio C1XD that I have, these registers are not set correctly
after an ACPI resume. The result is that after resuming, a shared
IRQ is left in edge-triggered mode so the interrupt can later become
jammed in a state where the line remains asserted, but the handler
is never called.

Reviewed by: jhb

21 years agoRevision 1.136 introduced two features that may cause undefined
iedowse [Wed, 30 Oct 2002 19:08:56 +0000 (19:08 +0000)]
Revision 1.136 introduced two features that may cause undefined
behaviour of the hardware: a possibly reserved bit of the receive
descriptor (RFA) `status' field is borrowed to record no-resource
(RNR) events, and the same status field is read and written to at
a time that may clash with the hardware updating this field.

There is no hardware documentation available to determine if these
things are safe to do; the second issue almost certainly isn't, and
the first is only safe if there is documentation saying that this
bit is free to be used by the driver. The PR referenced below
provides extremely convincing evidence that the changes cause random
crashes on some (unusual) hardware.

Since these features are only required by the DEVICE_POLLING case,
this commit makes their use conditional on that option. It does not
change the DEVICE_POLLING case, but at least people with the rare
hardware on which this code causes problems can now avoid the crashes
by not enabling DEVICE_POLLING.

PR: kern/42260
Reviewed by: luigi
Problem revision found by: Pawel Malachowski <pawmal@unia.3lo.lublin.pl>
Tested by: Pawel Malachowski <pawmal@unia.3lo.lublin.pl>
MFC after: 1 week

21 years agoMove to C99 sparse structure initialization for the mac_policy_ops
rwatson [Wed, 30 Oct 2002 18:48:51 +0000 (18:48 +0000)]
Move to C99 sparse structure initialization for the mac_policy_ops
structure definition, rather than using an operation vector
we translate into the structure.  Originally, we used a vector
for two reasons:

(1) We wanted to define the structure sparsely, which wasn't
    supported by the C compiler for structures.  For a policy
    with five entry points, you don't want to have to stick in
    a few hundred NULL function pointers.

(2) We thought it would improve ABI compatibility allowing modules
    to work with kernels that had a superset of the entry points
    defined in the module, even if the kernel had changed its
    entry point set.

Both of these no longer apply:

(1) C99 gives us a way to sparsely define a static structure.

(2) The ABI problems existed anyway, due to enumeration numbers,
    argument changes, and semantic mismatches.  Since the going
    rule for FreeBSD is that you really need your modules to
    pretty closely match your kernel, it's not worth the
    complexity.

This submit eliminates the operation vector, dynamic allocation
of the operation structure, copying of the vector to the
structure, and redoes the vectors in each policy to direct
structure definitions.  One enourmous benefit of this change
is that we now get decent type checking on policy entry point
implementation arguments.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoNew release notes: check_utility_compat(3), uuid(3).
bmah [Wed, 30 Oct 2002 18:38:04 +0000 (18:38 +0000)]
New release notes:  check_utility_compat(3), uuid(3).

Modified release notes:  Add atapicam(4) manpage cross-reference, NTP
4.1.1a, tzdata2002d.

21 years agoFix core dumps when invoking with something like ``last -w1''.
mux [Wed, 30 Oct 2002 18:19:59 +0000 (18:19 +0000)]
Fix core dumps when invoking with something like ``last -w1''.

PR: bin/44583
Reported by: Amit Chakradeo <sdbug@spam.chakradeo.net>
MFC after: 3 days

21 years agoVarious minor type, prototype tweaks -- clean up cruft due to lack of
rwatson [Wed, 30 Oct 2002 18:10:46 +0000 (18:10 +0000)]
Various minor type, prototype tweaks -- clean up cruft due to lack of
type checking on entry points (to be introduced shortly).

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoBuild Objective-C runtime support library with thr-posix.c.
kan [Wed, 30 Oct 2002 18:01:06 +0000 (18:01 +0000)]
Build Objective-C runtime support library with thr-posix.c.

Submitted by: davidc
Requested by: obrien

21 years agoWhile 'mode_t' seemed like a good idea for the access mode argument for
rwatson [Wed, 30 Oct 2002 17:56:57 +0000 (17:56 +0000)]
While 'mode_t' seemed like a good idea for the access mode argument for
MAC access() and open() checks, the argument actually has an int type
where it becomes available.  Switch to using 'int' for the mode argument
throughout the MAC Framework and policy modules.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoDocument most of bsd.doc.mk variables.
ru [Wed, 30 Oct 2002 17:03:48 +0000 (17:03 +0000)]
Document most of bsd.doc.mk variables.

21 years agoWhen we appear to want a signed value from sizeof(), use ssize_t in the cast,
jmallett [Wed, 30 Oct 2002 15:50:00 +0000 (15:50 +0000)]
When we appear to want a signed value from sizeof(), use ssize_t in the cast,
rather than int.

21 years agoAll uses of CD_HACK have been eliminated.
ru [Wed, 30 Oct 2002 15:30:06 +0000 (15:30 +0000)]
All uses of CD_HACK have been eliminated.

21 years agoFixes to get rid of CD_HACK.
ru [Wed, 30 Oct 2002 15:20:50 +0000 (15:20 +0000)]
Fixes to get rid of CD_HACK.

21 years agoUse __func__ and break a long line.
des [Wed, 30 Oct 2002 15:01:29 +0000 (15:01 +0000)]
Use __func__ and break a long line.

21 years agoUnbreak 'make release' by adding libssh to KPROGS.
des [Wed, 30 Oct 2002 14:27:12 +0000 (14:27 +0000)]
Unbreak 'make release' by adding libssh to KPROGS.

Suggested by: jhay

21 years agoFix a bug in fenner's _fetch_writev() patch (rev 1.36)
des [Wed, 30 Oct 2002 14:25:00 +0000 (14:25 +0000)]
Fix a bug in fenner's _fetch_writev() patch (rev 1.36)

Submitted by: fenner

21 years agoRegen:
sheldonh [Wed, 30 Oct 2002 13:35:48 +0000 (13:35 +0000)]
Regen:

* Hart: rev 339 of pcidevs.txt (2002-10-21)
* Boemler: vendors.txt (2002-10-28)

21 years agoFixes to get rid of CD_HACK.
ru [Wed, 30 Oct 2002 13:07:34 +0000 (13:07 +0000)]
Fixes to get rid of CD_HACK.

21 years agoAdd two additional references to the See Also section, which contain much
tjr [Wed, 30 Oct 2002 11:49:05 +0000 (11:49 +0000)]
Add two additional references to the See Also section, which contain much
better descriptions of UTF-8 and related issues.

21 years agoOops. scr2txt is in the textproc category.
murray [Wed, 30 Oct 2002 09:20:50 +0000 (09:20 +0000)]
Oops.  scr2txt is in the textproc category.

Pointed out by: jhay
MFC After: 1 day

21 years agoMan page for ATAPI/CAM.
thomas [Wed, 30 Oct 2002 08:42:24 +0000 (08:42 +0000)]
Man page for ATAPI/CAM.

Reviewed by: roberto

21 years agoDon't need to check if SO_OOBINLINE is defined.
hsu [Wed, 30 Oct 2002 08:32:19 +0000 (08:32 +0000)]
Don't need to check if SO_OOBINLINE is defined.
Don't need to protect isipv6 conditional with INET6.
Fix leading indentation in 2 lines.

21 years agoDon't pass the return address to exception_save in register b0. Use
marcel [Wed, 30 Oct 2002 07:58:11 +0000 (07:58 +0000)]
Don't pass the return address to exception_save in register b0. Use
a true scratch register. This change and future re-allocations will
eventually result in code that we can unwind to to get the preserved
registers of the process. This of course means that we cannot trash
them while saving the process context.

While re-allocating, remove the register aliases. Abstraction is in
this case disadvanteous.

21 years agoUse "deprecated" instead of "depreciated" where appropriate.
sheldonh [Wed, 30 Oct 2002 07:49:11 +0000 (07:49 +0000)]
Use "deprecated" instead of "depreciated" where appropriate.

21 years agoMake pthread_sigmask(3) operate on the thread signal mask, not the process
mini [Wed, 30 Oct 2002 07:13:27 +0000 (07:13 +0000)]
Make pthread_sigmask(3) operate on the thread signal mask, not the process
signal mask.

21 years agoDelete worthless comments.
tjr [Wed, 30 Oct 2002 06:34:52 +0000 (06:34 +0000)]
Delete worthless comments.

21 years agoUse KSE to schedule threads.
mini [Wed, 30 Oct 2002 06:07:18 +0000 (06:07 +0000)]
Use KSE to schedule threads.

21 years agoRecognize the (incorrect) error code a MediaHawk server sends in reply to
des [Wed, 30 Oct 2002 06:06:16 +0000 (06:06 +0000)]
Recognize the (incorrect) error code a MediaHawk server sends in reply to
unrecognized commands such as MDTM.

Requested by: Stephen Roome <stephen_roome@pepcross.com>

21 years agoRewrite cpu_switch(). The most notable change is the fact that we now
marcel [Wed, 30 Oct 2002 05:55:29 +0000 (05:55 +0000)]
Rewrite cpu_switch(). The most notable change is the fact that we now
have f16-f31 as part of the context. The PCB has been reorganized to
better match how we save and restore the (preserved) registers. This
commit also moves the context restoriation to its own function (named
pcb_restore), as we did with pcb_save.

Only minimal effort has been put in writing optimal assembly. The
expectation is that there will be more rounds of changes.

21 years agoCheck NULL thread mailbox pointer.
davidxu [Wed, 30 Oct 2002 05:09:29 +0000 (05:09 +0000)]
Check NULL thread mailbox pointer.

21 years agoClarify my feelings towards fetch / libfetch.
des [Wed, 30 Oct 2002 04:53:58 +0000 (04:53 +0000)]
Clarify my feelings towards fetch / libfetch.

21 years agoRecommit the non-broken parts of 1.34 and 1.37.
des [Wed, 30 Oct 2002 04:43:00 +0000 (04:43 +0000)]
Recommit the non-broken parts of 1.34 and 1.37.
Change the type and name of a variable introduced in 1.33.

21 years agoDon't forget to set the error flag when _fetch_read() returns -1.
des [Wed, 30 Oct 2002 04:42:01 +0000 (04:42 +0000)]
Don't forget to set the error flag when _fetch_read() returns -1.

21 years agoImplement DCE 1.1 compliant UUID functions. Immediate use of these
marcel [Wed, 30 Oct 2002 03:51:00 +0000 (03:51 +0000)]
Implement DCE 1.1 compliant UUID functions. Immediate use of these
functions is expected for uuidgen(1), mca(8) and gpt(8). Given the
generic use of UUIDs beyond the scope of the DCE 1.1 specification,
visibility of the data structure at all levels of the machine,
including firmware and the wish to not create a permanent build-
time FreeBSD-ism for DCE compliant applications by creating a new
library, it was decided that libc would be the least inappropriate
place. Also, because the UUID functions live in libc under IRIX as
well, we have maximized our portability and left as many options
open as possible.

This implementation introduces an extension not found in the
specification: the status parameter is allowed to be a NULL-
pointer. The reason for introducing the extension is because
the status is almost never of any use.

The manpage that's part of this commit is a minimal place-holder
and is further fleshed-out in the near future.

Approved by: re@
Contributed by: Hiten Mahesh Pandya <hiten@unixdaemons.com>
Sponsored by: marcel :-)
Tested on: alpha, i386, ia64

21 years agoOutput .lf directives.
grog [Wed, 30 Oct 2002 03:24:16 +0000 (03:24 +0000)]
Output .lf directives.

21 years agoStyle fixes.
davidxu [Wed, 30 Oct 2002 03:01:28 +0000 (03:01 +0000)]
Style fixes.

21 years agoDon't forget to set syscall result.
davidxu [Wed, 30 Oct 2002 02:39:10 +0000 (02:39 +0000)]
Don't forget to set syscall result.

21 years agoAdd an actual implementation of kse_thr_interrupt()
davidxu [Wed, 30 Oct 2002 02:28:41 +0000 (02:28 +0000)]
Add an actual implementation of kse_thr_interrupt()

21 years agoRemove extranious memory barriers, and correct the placement of a few others.
gallatin [Wed, 30 Oct 2002 01:41:44 +0000 (01:41 +0000)]
Remove extranious memory barriers, and correct the placement of a few others.
This provides a 30% reduction in system time and a 6% reduction in wallclock time
for a make buildworld on my xp1000 (one 21264).

FWIW, I've been running this for nearly 2 months without problems.

Portions submitted by: ticso, jhb
Tested by: jhb (ds20 dual 21264)

21 years agoMake the second serial port available for general use by default.
gallatin [Wed, 30 Oct 2002 01:34:41 +0000 (01:34 +0000)]
Make the second serial port available for general use by default.
We've been talking about this for years, but nobody has done it.
(and I don't think anybody has used this for debugging since Doug
and I were doing the initial bootstrapping..)

21 years agoCheck for errors and zero-length transfers in the ulpt_input() input
iedowse [Wed, 30 Oct 2002 01:18:58 +0000 (01:18 +0000)]
Check for errors and zero-length transfers in the ulpt_input() input
pipe callback function, and just return if these cases are detected.
Without these checks, the ulpt driver may cause an infinite loop
of failing USB transfers that can hang the whole machine. This makes
printing work for me on a HP DJ950C printer.

21 years agoThe future's bright: the future's Ballard...
fanf [Wed, 30 Oct 2002 01:00:48 +0000 (01:00 +0000)]
The future's bright: the future's Ballard...

21 years agoReinstate revs 1.35-36 and 1.38. Revisions 1.34 and 1.37 were specifically
imp [Wed, 30 Oct 2002 00:17:16 +0000 (00:17 +0000)]
Reinstate revs 1.35-36 and 1.38.  Revisions 1.34 and 1.37 were specifically
the root cause of the bus errors I was experiencing.

Submitted by: fenner
Tested by: obrien
Prompted by: peter

21 years agoTry again to fix the KASSERT.
rwatson [Wed, 30 Oct 2002 00:16:59 +0000 (00:16 +0000)]
Try again to fix the KASSERT.

21 years agoFix a KASSERT bug that showed up only in the LINT build, not the
rwatson [Wed, 30 Oct 2002 00:04:31 +0000 (00:04 +0000)]
Fix a KASSERT bug that showed up only in the LINT build, not the
module build, so I missed it in the merge.

Pointed out by: sam

21 years agoWhile an interface can be depreciated, we prefer deprecated.
rwatson [Tue, 29 Oct 2002 20:53:46 +0000 (20:53 +0000)]
While an interface can be depreciated, we prefer deprecated.

Submitted by: Wayne Morrison <tewok@tislabs.com>

21 years agoMinor comment typo fix.
rwatson [Tue, 29 Oct 2002 20:51:44 +0000 (20:51 +0000)]
Minor comment typo fix.

Submitted by: Wayne Morrison <tewok@tislabs.com>

21 years agoUpdate for 4.1.1a.
roberto [Tue, 29 Oct 2002 20:30:43 +0000 (20:30 +0000)]
Update for 4.1.1a.

Tested on: Sparc64 (panther), Alpha (beast) & i386

21 years agoUpdate ntpd to 4.1.1a. This contains many bug fixes made on the STABLE
roberto [Tue, 29 Oct 2002 20:20:36 +0000 (20:20 +0000)]
Update ntpd to 4.1.1a. This contains many bug fixes made on the STABLE
branch and a few new drivers. See contrib/ntp/ChangeLog for details.

Hide kernel header sys/lock.h from ntp [1]

PR: bin/33914
Submitted by: thomas, bde[1]
MFC after: 1 month

21 years agoThis commit was generated by cvs2svn to compensate for changes in r106167,
roberto [Tue, 29 Oct 2002 20:11:45 +0000 (20:11 +0000)]
This commit was generated by cvs2svn to compensate for changes in r106167,
which included commits to RCS files with non-trunk default branches.

21 years agoMerge conflicts.
roberto [Tue, 29 Oct 2002 20:04:27 +0000 (20:04 +0000)]
Merge conflicts.

MFC after: 1 month

21 years agoThis commit was generated by cvs2svn to compensate for changes in r106163,
roberto [Tue, 29 Oct 2002 19:58:12 +0000 (19:58 +0000)]
This commit was generated by cvs2svn to compensate for changes in r106163,
which included commits to RCS files with non-trunk default branches.

21 years agoVirgin import of ntpd 4.1.1a
roberto [Tue, 29 Oct 2002 19:58:12 +0000 (19:58 +0000)]
Virgin import of ntpd 4.1.1a

21 years agoHook up no-op stubs for reboot, swapon, sysctl entry points.
rwatson [Tue, 29 Oct 2002 19:57:28 +0000 (19:57 +0000)]
Hook up no-op stubs for reboot, swapon, sysctl entry points.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoImplement Biba policy entry points for mac_check_system_swapon()
rwatson [Tue, 29 Oct 2002 19:18:52 +0000 (19:18 +0000)]
Implement Biba policy entry points for mac_check_system_swapon()
and mac_check_system_sysctl(), providing additional integrity
protections relating to swap target device selection and system
management via sysctl().  Require Biba privilege for both; also
require that the target of swap operations be a high integrity
object, since swap data is high integrity.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Asssociates Laboratories

21 years agoRequire Biba privilege to relabel a network interface.
rwatson [Tue, 29 Oct 2002 19:14:16 +0000 (19:14 +0000)]
Require Biba privilege to relabel a network interface.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoFinally get around to committing Bill Paul's FEC netgraph nodes.
julian [Tue, 29 Oct 2002 19:12:44 +0000 (19:12 +0000)]
Finally get around to committing Bill Paul's FEC netgraph nodes.
These are really only partly netgraph nodes as they do not use the
netgraph interfaces for many of the functions for which they could
be used, however they represent important functionality.

Submitted by: wpaul
MFC after: 2 days

21 years agoAdd a missing newline to the end of a device_printf().
jhb [Tue, 29 Oct 2002 19:08:55 +0000 (19:08 +0000)]
Add a missing newline to the end of a device_printf().

Reported by: Michael G. Petry <petry@NetMasters.Com>

21 years agoDon't take out the rl_mtx lock in the attach routine. The only way
imp [Tue, 29 Oct 2002 19:02:19 +0000 (19:02 +0000)]
Don't take out the rl_mtx lock in the attach routine.  The only way
we'd need it is if we're interrupted.  So, register the interrupt last
in the attach routine.

21 years agoGetting closer to getting things right:
phk [Tue, 29 Oct 2002 17:12:02 +0000 (17:12 +0000)]
Getting closer to getting things right:
Always ignore the RAW_PART for BSD.
Having no quirks is not a mistake for an architecture.

Tested by: DES

21 years agoRemove Perl bits still lurking in BSD.local.dist. It is a
tobez [Tue, 29 Oct 2002 16:48:03 +0000 (16:48 +0000)]
Remove Perl bits still lurking in BSD.local.dist.  It is a
responsibility of lang/perl5 to deal with those nowadays.

Requested by: kkenn