]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoPut the checks we perform on a bio before calling ::start in their
phk [Thu, 6 Feb 2003 21:01:36 +0000 (21:01 +0000)]
Put the checks we perform on a bio before calling ::start in their
own function, handle all validation and truncation at the time we
process the bio instead of when it gets scheduled.

21 years agoMerge the following from the English version:
hrs [Thu, 6 Feb 2003 17:35:45 +0000 (17:35 +0000)]
Merge the following from the English version:

1.488 -> 1.490 relnotes/common/new.sgml

21 years agoFix comment typo.
orion [Thu, 6 Feb 2003 17:32:02 +0000 (17:32 +0000)]
Fix comment typo.

Sync with userland test framework which now deals better with pcm feeder kobj
emulation.

Reduce max rate from 96kHz to 48kHz as userland tests found a few bad
points about 90kHz and we don't care about operating up there for now.

21 years agoFix memory leak.
simokawa [Thu, 6 Feb 2003 17:23:01 +0000 (17:23 +0000)]
Fix memory leak.

21 years agoNew errata: SA-03:01.
bmah [Thu, 6 Feb 2003 17:06:57 +0000 (17:06 +0000)]
New errata:  SA-03:01.

21 years agoNew release note: SA-03:01.
bmah [Thu, 6 Feb 2003 16:35:39 +0000 (16:35 +0000)]
New release note:  SA-03:01.

21 years agoMinor structural change: It always struck me as odd that "security
bmah [Thu, 6 Feb 2003 16:21:49 +0000 (16:21 +0000)]
Minor structural change: It always struck me as odd that "security
advisories" were stuck between "kernel" and "userland".  Move this
(empty for now) section to the first subsection of "What's New".

21 years agoFix another typo in a comment which I noticed while doing the MFC.
gj [Thu, 6 Feb 2003 14:52:47 +0000 (14:52 +0000)]
Fix another typo in a comment which I noticed while doing the MFC.

21 years agoUse pam_group(8) instead of pam_wheel(8).
des [Thu, 6 Feb 2003 14:33:23 +0000 (14:33 +0000)]
Use pam_group(8) instead of pam_wheel(8).

21 years agoPass the correct, verified username to PAM instead of getlogin().
des [Thu, 6 Feb 2003 14:29:28 +0000 (14:29 +0000)]
Pass the correct, verified username to PAM instead of getlogin().

21 years agoMaybe I was a little too fast? Remove debugging code, and commit the
des [Thu, 6 Feb 2003 14:27:48 +0000 (14:27 +0000)]
Maybe I was a little too fast?  Remove debugging code, and commit the
Makefile and man page which I'd forgotten to 'cvs add'.

Sponsored by: DARPA, NAI Labs

21 years agoFix build error.
nyan [Thu, 6 Feb 2003 14:26:41 +0000 (14:26 +0000)]
Fix build error.

21 years agoReplace pam_wheel(8) with pam_group(8) which has a cleaner interface. The
des [Thu, 6 Feb 2003 14:24:14 +0000 (14:24 +0000)]
Replace pam_wheel(8) with pam_group(8) which has a cleaner interface.  The
pam_wheel(8) module was written to work in spite of a broken libpam, and
has grown organically since its inception, which is reflected in both its
functionality and implementation.  Rather than clean up pam_wheel(8) and
break backward compatibility, I've chosen to reimplement it under a new,
more generic name.

Sponsored by: DARPA, NAI Labs

21 years agoMake sure the message is only printed once.
des [Thu, 6 Feb 2003 14:19:50 +0000 (14:19 +0000)]
Make sure the message is only printed once.

21 years agoDon't blame markm for what he didn't do - writing these man pages, for
des [Thu, 6 Feb 2003 13:47:21 +0000 (13:47 +0000)]
Don't blame markm for what he didn't do - writing these man pages, for
instance.  Also bump the date since I made substantial modifications
earlier today.

21 years agoUpdate copyright.
des [Thu, 6 Feb 2003 12:56:51 +0000 (12:56 +0000)]
Update copyright.

21 years agoAdd support for escape sequences in the arguments (e.g. %u for user name)
des [Thu, 6 Feb 2003 12:56:39 +0000 (12:56 +0000)]
Add support for escape sequences in the arguments (e.g. %u for user name)

Sponsored by: DARPA, NAI Labs

21 years agoExport the PAM environment to the child process instead of the "normal"
des [Thu, 6 Feb 2003 12:40:58 +0000 (12:40 +0000)]
Export the PAM environment to the child process instead of the "normal"
environment list, which may be unsafe and / or sensitive.

Sponsored by: DARPA, NAI Labs

21 years agoThe .Fn function. Spelling.
charnier [Thu, 6 Feb 2003 11:29:40 +0000 (11:29 +0000)]
The .Fn function. Spelling.

21 years agoThe .Fn function
charnier [Thu, 6 Feb 2003 11:25:03 +0000 (11:25 +0000)]
The .Fn function

21 years agoThe .Fn function
charnier [Thu, 6 Feb 2003 11:10:32 +0000 (11:10 +0000)]
The .Fn function

21 years agoOops. Include opt_ddb.h.
benno [Thu, 6 Feb 2003 11:07:39 +0000 (11:07 +0000)]
Oops.  Include opt_ddb.h.

21 years agoThe .Fn function
charnier [Thu, 6 Feb 2003 11:04:47 +0000 (11:04 +0000)]
The .Fn function

21 years agoAdd a driver that attaches to the gpio node of macio and allows you to enter
benno [Thu, 6 Feb 2003 10:59:06 +0000 (10:59 +0000)]
Add a driver that attaches to the gpio node of macio and allows you to enter
DDB when the interrupt button (aka the "programmer's switch") is pressed.

This isn't unfortunately an NMI, but it's a handy way to get into DDB
quickly if needed.

21 years agoMinimal manual page for pam_kerberosIV(8).
des [Thu, 6 Feb 2003 10:55:11 +0000 (10:55 +0000)]
Minimal manual page for pam_kerberosIV(8).

Sponsored by: DARPA, NAI Labs

21 years agoAdd a cast to silence a warning.
benno [Thu, 6 Feb 2003 10:47:57 +0000 (10:47 +0000)]
Add a cast to silence a warning.

21 years agoIf a device tries to allocate an interrupt that's not on it's resource list,
benno [Thu, 6 Feb 2003 10:42:22 +0000 (10:42 +0000)]
If a device tries to allocate an interrupt that's not on it's resource list,
assume that the child knows what it's doing and add it to the resource list.

21 years agoTeach the CAM module how to deregister itself so it can be unloaded.
scottl [Thu, 6 Feb 2003 03:31:02 +0000 (03:31 +0000)]
Teach the CAM module how to deregister itself so it can be unloaded.

21 years agoRelocate a call to enable inetd so that it is set regardless of
rwatson [Thu, 6 Feb 2003 01:55:40 +0000 (01:55 +0000)]
Relocate a call to enable inetd so that it is set regardless of
whether the user chooses to edit inetd.conf.

PR: 39311
Reported by: Martin Faxer <gmh003532@brfmasthugget.se>

21 years agoRemove vestiges of no longer needed unp_rvnode field.
hsu [Thu, 6 Feb 2003 01:34:43 +0000 (01:34 +0000)]
Remove vestiges of no longer needed unp_rvnode field.

Approved by: phk (who originally added it in rev 1.8 of unpcb.h)

21 years agoFix use of an uninitialized pointer introduced in a previous revision.
mtm [Thu, 6 Feb 2003 01:08:19 +0000 (01:08 +0000)]
Fix use of an uninitialized pointer introduced in a previous revision.

Approved by: markm (mentor)(implicit)

21 years agoFree the container and sim objects on detach. Convert several uses of
scottl [Thu, 6 Feb 2003 00:37:37 +0000 (00:37 +0000)]
Free the container and sim objects on detach.  Convert several uses of
MALLOC to malloc.

21 years agoDon't include aac_cam.h, since it was nuked in the last commit.
scottl [Thu, 6 Feb 2003 00:07:12 +0000 (00:07 +0000)]
Don't include aac_cam.h, since it was nuked in the last commit.
Deregister the shutdown eventhandler on unload.

21 years agoVarious cleanups:
scottl [Wed, 5 Feb 2003 23:24:25 +0000 (23:24 +0000)]
Various cleanups:
 - Move the command timeout check from a separate repeating timeout to the
   kthread since the kthread is already running periodically.
 - Move printing the hardware print buffer to the kthread.
 - Properly shut down the kernel thread on detach.
 - Detach the child array devices on detach.
 - Don't issue a controller halt command on detach.  Doing so requires a PCI
   reset to wake the controller back up.  The driver can now be unloaded as
   long as CAM support is not enabled.

21 years agoSince we drop NSHUFF values now, set default seed to what it becomes
ache [Wed, 5 Feb 2003 21:28:52 +0000 (21:28 +0000)]
Since we drop NSHUFF values now, set default seed to what it becomes
after srandom(1)

21 years agoSince we drop NSHUFF values now, set default seed to what it becomes
ache [Wed, 5 Feb 2003 21:25:50 +0000 (21:25 +0000)]
Since we drop NSHUFF values now, set default seed to what it becomes
after srand(1)

21 years agoMove the DC_SETBIT to the attach function.
mbr [Wed, 5 Feb 2003 21:21:26 +0000 (21:21 +0000)]
Move the DC_SETBIT to the attach function.

Reviewed by: phk

21 years agoImplement the new "struct disk" centered API for device drivers.
phk [Wed, 5 Feb 2003 21:13:08 +0000 (21:13 +0000)]
Implement the new "struct disk" centered API for device drivers.

This commit should not change anything as no device drivers use the
new API yet.

21 years agoDocument the recent changes for vinum as root filesystem.
joerg [Wed, 5 Feb 2003 21:05:03 +0000 (21:05 +0000)]
Document the recent changes for vinum as root filesystem.

MFC after: 2 weeks

21 years agoFinally make vinum ready for the root filesystem in FreeBSD 5.x. (A
joerg [Wed, 5 Feb 2003 21:02:56 +0000 (21:02 +0000)]
Finally make vinum ready for the root filesystem in FreeBSD 5.x.  (A
similar patch has been in 4.x for a while, but is more hacky there.)

For this to work, vinum has to be loaded early (e. g. from
boot/loader), for obvious reasons.  If the kernel env variable
(aka. loader variable) "vinum.autostart" is set, vinum then asks the
sysctl kern.disks for all available disks in the system, and scans
them for possible vinum headers.

For statically compiled kernels, this behaviour can be obtained even
without boot/loader by using "options VINUM_AUTOSTART" (though this is
not the recommended way).

Alternatively, the 4.x way to specify "vinum.drives" is also supported.

No further hacks (like the 4.x "vinum.root" variable) are needed,
since in 5.x, mountroot() asks back at the drivers to have them
resolve the name of the root FS into a dev_t (using the dev_clone
eventhandler).

(The MFC reminder below is for a partial MFC for vinum.autostart, the
rest is already there in 4.x.)

Timed out on: grog
MFC after: 2 weeks

21 years agoAdd three more properties to struct disk: Stripe offset and width
phk [Wed, 5 Feb 2003 20:17:45 +0000 (20:17 +0000)]
Add three more properties to struct disk:  Stripe offset and width
and max request size.  These will be used to optimize our clustering
behaviour.

21 years agoUse a more standard error message. Add FBSDID.
charnier [Wed, 5 Feb 2003 20:08:39 +0000 (20:08 +0000)]
Use a more standard error message. Add FBSDID.

Reviewed by: ru

21 years agoThe lockmanager has to keep track of locks per thread, not per process.
julian [Wed, 5 Feb 2003 19:36:58 +0000 (19:36 +0000)]
The lockmanager has to keep track of locks per thread, not per process.

Submitted by: david Xu (davidxu@)
Reviewed by: jhb@

21 years agoFix slight disorder that broke sorting. Put in bold warning about the
sobomax [Wed, 5 Feb 2003 19:22:39 +0000 (19:22 +0000)]
Fix slight disorder that broke sorting. Put in bold warning about the
fact that in this case order matters.

Submitted by: Peter Edwards <pmedwards@eircom.net>

21 years agoMFS: 1.2.2.2
trhodes [Wed, 5 Feb 2003 19:16:18 +0000 (19:16 +0000)]
MFS: 1.2.2.2

21 years agoTake the rc4 code out of ng_mppc module so we don't fail to load when
ambrisko [Wed, 5 Feb 2003 19:11:11 +0000 (19:11 +0000)]
Take the rc4 code out of ng_mppc module so we don't fail to load when
we have the rc4 code already in the kernel (via wlan stuff or awi).
Add a dependency on the rc4 module so if it doesn't exist then load it.

Reviewed by: archie

21 years agoRemove MAKEDEV quirks, now obsolete.
alex [Wed, 5 Feb 2003 18:33:36 +0000 (18:33 +0000)]
Remove MAKEDEV quirks, now obsolete.

21 years agoFix for fwmem write part.
simokawa [Wed, 5 Feb 2003 17:19:34 +0000 (17:19 +0000)]
Fix for fwmem write part.
- fix typo
- don't need to check receive buffer.
- don't forget to free xfer on errer.

21 years agoAdd regression test for soon-to-be-committed fpclassify().
mike [Wed, 5 Feb 2003 16:40:18 +0000 (16:40 +0000)]
Add regression test for soon-to-be-committed fpclassify().

21 years agoRemove srandom():
ache [Wed, 5 Feb 2003 15:56:04 +0000 (15:56 +0000)]
Remove srandom():
1) It is already called in init_main.c:proc0_post()
2) It is called each time read_random_phony() called, because "initialized"
variable is never set to 1.

Approved by:    markm

21 years agoComment out srandom():
ache [Wed, 5 Feb 2003 15:32:24 +0000 (15:32 +0000)]
Comment out srandom():
1) Already called in init_main.c:proc0_post()
2) Seed is bad

21 years agoNODEVFS cleanup.
ru [Wed, 5 Feb 2003 15:31:43 +0000 (15:31 +0000)]
NODEVFS cleanup.

21 years agoDo not start err(3) strings with uppercase nor dot terminate them.
charnier [Wed, 5 Feb 2003 14:25:43 +0000 (14:25 +0000)]
Do not start err(3) strings with uppercase nor dot terminate them.
Standard sccsid[] area according to style(9).

21 years agoHeader was included twice. Hardcode program name in usage string.
charnier [Wed, 5 Feb 2003 14:20:27 +0000 (14:20 +0000)]
Header was included twice. Hardcode program name in usage string.

21 years agoWrap long line. Use .Ar instead of .Nm where appropriate.
charnier [Wed, 5 Feb 2003 14:15:02 +0000 (14:15 +0000)]
Wrap long line. Use .Ar instead of .Nm where appropriate.

21 years agoSpelling.
charnier [Wed, 5 Feb 2003 14:03:55 +0000 (14:03 +0000)]
Spelling.

21 years agoThe .Fn function
charnier [Wed, 5 Feb 2003 14:00:46 +0000 (14:00 +0000)]
The .Fn function

21 years agoThe .Xr utility
charnier [Wed, 5 Feb 2003 13:50:41 +0000 (13:50 +0000)]
The .Xr utility

21 years agoAdd FBSDID. udp/bootps -> bootps/udp. Use err(3).
charnier [Wed, 5 Feb 2003 13:45:25 +0000 (13:45 +0000)]
Add FBSDID. udp/bootps -> bootps/udp. Use err(3).

21 years agoPrevent uppercase after .Xr by adding ``The ... utility/system call''.
charnier [Wed, 5 Feb 2003 13:36:13 +0000 (13:36 +0000)]
Prevent uppercase after .Xr by adding ``The ... utility/system call''.

21 years agoAdd FBSDID.
charnier [Wed, 5 Feb 2003 13:23:46 +0000 (13:23 +0000)]
Add FBSDID.

21 years agoDisplay elapsed time (-o etime) using [[dd-]hh:]mm:ss, which according to
charnier [Wed, 5 Feb 2003 13:18:17 +0000 (13:18 +0000)]
Display elapsed time (-o etime) using [[dd-]hh:]mm:ss, which according to
Solaris man page is the POSIX way.

Reviewed by: jmallett

21 years agoEnclose sccsid according to style(9).
charnier [Wed, 5 Feb 2003 12:56:40 +0000 (12:56 +0000)]
Enclose sccsid according to style(9).
udp/timed -> timed/udp

21 years agoGC an unused variable.
benno [Wed, 5 Feb 2003 12:34:10 +0000 (12:34 +0000)]
GC an unused variable.

21 years agoExport the ns_per_tick variable through md_var.h rather than by declaring
benno [Wed, 5 Feb 2003 12:33:49 +0000 (12:33 +0000)]
Export the ns_per_tick variable through md_var.h rather than by declaring
it extern in cpu.c.

21 years ago- Use cpu_setup() instead of identifycpu().
benno [Wed, 5 Feb 2003 12:10:46 +0000 (12:10 +0000)]
- Use cpu_setup() instead of identifycpu().
- Remove identifycpu().

21 years agoAdd cpu.c. This contains one exported function, cpu_setup(), which handles
benno [Wed, 5 Feb 2003 12:09:53 +0000 (12:09 +0000)]
Add cpu.c.  This contains one exported function, cpu_setup(), which handles
setup of and printing information about cpus.

Obtained from: NetBSD (parts)

21 years ago- Update spr.h
benno [Wed, 5 Feb 2003 12:04:29 +0000 (12:04 +0000)]
- Update spr.h
- Add hid.h

Obtained from: NetBSD

NOTE: This undoes some changes I'd made to prefix the processor name defines
with PVR_.  This was due to my original decision to use MPC750 as a cpu name.
With this changed, the PVR_ change is no longer required.

21 years agoAdd an inline function wrapper for the mfpvf (Move From Processor Version
benno [Wed, 5 Feb 2003 11:59:27 +0000 (11:59 +0000)]
Add an inline function wrapper for the mfpvf (Move From Processor Version
Register) instruction.

21 years agoNot all cpus are MPC750s. Replace the MPC750 cpu option with OEA. This
benno [Wed, 5 Feb 2003 11:37:59 +0000 (11:37 +0000)]
Not all cpus are MPC750s.  Replace the MPC750 cpu option with OEA.  This
stands for Operating Environment Architecture and is the specification that
all of the MPC6xx, MPC7xx, MPC7xxx and IBM7xx CPUs adhere to.

21 years agoReplace the inline asm in delay() with a while loop. This may not be as
benno [Wed, 5 Feb 2003 11:26:14 +0000 (11:26 +0000)]
Replace the inline asm in delay() with a while loop.  This may not be as
efficient but it appears to actually work.  Some investigation may be
required.

21 years ago- Rename the "powerpc" timecounter to the "decrementer" timecounter.
benno [Wed, 5 Feb 2003 11:16:36 +0000 (11:16 +0000)]
- Rename the "powerpc" timecounter to the "decrementer" timecounter.
- Initialise it earlier.

21 years agoThis file has no longer any content from the original Berkeley file so
phk [Wed, 5 Feb 2003 11:11:39 +0000 (11:11 +0000)]
This file has no longer any content from the original Berkeley file so
replace the UCB copyright with a FreeBSD 2 clause thing.

Remove some no longer relevant comments.

21 years agoAllow "~/" in pathnames to work for a chrooted user.
yar [Wed, 5 Feb 2003 11:11:32 +0000 (11:11 +0000)]
Allow "~/" in pathnames to work for a chrooted user.

21 years ago* Update copyright year.
dougb [Wed, 5 Feb 2003 11:09:21 +0000 (11:09 +0000)]
* Update copyright year.
* Give back one line of ws when showing a diff.
  This was requested by several.
* Un-quote DIFF_FLAG, and add DIFF_OPTIONS. This will allow the user to
  do more creative things in a mergemaster rc file. Un-quoting is necessary
  in order to handle things like CVS Id tags that look like shell variables.
* Simplify terminal width determination.

Suggestions related to ignoring the CVS Id's were taken from several, with
dillon and gad providing particularly helpful feedback.

21 years agoAvoid undefined symbol error with an IPv4 only kernel.
ume [Wed, 5 Feb 2003 11:05:45 +0000 (11:05 +0000)]
Avoid undefined symbol error with an IPv4 only kernel.

Reported by: "Sergey A. Osokin" <osa@freebsd.org.ru>

21 years ago* Update copyright year.
dougb [Wed, 5 Feb 2003 11:03:05 +0000 (11:03 +0000)]
* Update copyright year.
* Add an example that ignores CVS Id tags for the new DIFF_OPTIONS variable.

21 years agoMoved 802.11 (device wlan) consumers to the third floppy as well.
ru [Wed, 5 Feb 2003 10:45:06 +0000 (10:45 +0000)]
Moved 802.11 (device wlan) consumers to the third floppy as well.
Shouldn't these modules have the dependency on "wlan" module?

21 years agoReduce diff to i386/isa/clock.c by unifdef -DPC98
phk [Wed, 5 Feb 2003 10:16:43 +0000 (10:16 +0000)]
Reduce diff to i386/isa/clock.c by unifdef -DPC98

21 years agoTypo in last commit
phk [Wed, 5 Feb 2003 09:35:02 +0000 (09:35 +0000)]
Typo in last commit

21 years agoMFi386: write the correct weekday back to the RTC.
phk [Wed, 5 Feb 2003 09:33:57 +0000 (09:33 +0000)]
MFi386: write the correct weekday back to the RTC.

21 years agoi386/i386/tsc.c was repo-copied from i386/isa/clock.c.
phk [Wed, 5 Feb 2003 09:20:40 +0000 (09:20 +0000)]
i386/i386/tsc.c was repo-copied from i386/isa/clock.c.

Remove all the stuff that does not relate to the TSC.

Change the calibration to use DELAY(1000000) rather than trying to check
it against the CMOS RTC, this drastically increases precision:

Using 25 samples on a Athlon 700MHz UP machine I find:

                stddev          min          max        average
CMOS             22200 Hz    -74980 Hz     34301 Hz   704928721 Hz
DELAY             1805 Hz     -1984 Hz      2678 Hz   704937583 Hz

(The difference between the two averages is not statistically significant.)

expressed in PPM of the frequency:
                stddev          min          max
CMOS             31.49 PPM  -106.37 PPM    48.66 PPM
DELAY             2.56 PPM     2.81 PPM     3.80 PPM

This code will not be used until a followup commit to sys/isa/clock.c
and sys/pc98/pc98/clock.c which will only happen after some field testing.

21 years agoMention that the CLOCK_VIRTUAL and CLOCK_PROF clocks are not implemented.
tjr [Wed, 5 Feb 2003 09:17:32 +0000 (09:17 +0000)]
Mention that the CLOCK_VIRTUAL and CLOCK_PROF clocks are not implemented.

PR: 8376

21 years agoMake get_cyclecount() use binuptime() when no tsc is available: it is cheaper.
phk [Wed, 5 Feb 2003 08:55:10 +0000 (08:55 +0000)]
Make get_cyclecount() use binuptime() when no tsc is available: it is cheaper.

21 years agoObtain the media type from the shared memory and only use the eeprom
ps [Wed, 5 Feb 2003 08:54:36 +0000 (08:54 +0000)]
Obtain the media type from the shared memory and only use the eeprom
as a fallback.

21 years agoProperly get the drive geometry from the controller. This should
ps [Wed, 5 Feb 2003 08:43:46 +0000 (08:43 +0000)]
Properly get the drive geometry from the controller.  This should
fix booting off of volumes > 255GB.

21 years agoClose a memory leak when rescanning the bus.
ps [Wed, 5 Feb 2003 07:46:08 +0000 (07:46 +0000)]
Close a memory leak when rescanning the bus.

21 years agoAdd some reference.
simokawa [Wed, 5 Feb 2003 06:41:19 +0000 (06:41 +0000)]
Add some reference.

21 years agoSupport for suspend/resume in laptops. Loosely based on the wi(4) patches
ambrisko [Wed, 5 Feb 2003 03:37:42 +0000 (03:37 +0000)]
Support for suspend/resume in laptops.  Loosely based on the wi(4) patches
but not quite.  The memory descriptors and most of the card parameters
need to be writen to the card.

MFC in: 3 days

21 years agoThe manual page lists only 2 files, however it reads as `three files' which is
trhodes [Wed, 5 Feb 2003 02:14:03 +0000 (02:14 +0000)]
The manual page lists only 2 files, however it reads as `three files' which is
obviously incorrect.

PR: 46841
Submitted by: Sakamoto Seiji <s-siji@hyper.ocn.ne.jp>

21 years agoImplement mpo_check_kld_load() and mpo_check_kld_unload() for the Biba
rwatson [Tue, 4 Feb 2003 22:38:33 +0000 (22:38 +0000)]
Implement mpo_check_kld_load() and mpo_check_kld_unload() for the Biba
MAC policy.  To load a KLD, require that the subject hold Biba privilege,
and the the kernel module be marked as high integrity.  To unload a
KLD, require that the subject hold Biba privilege.

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

21 years agoPlace more stringent checks on process credential relabeling for the Biba
rwatson [Tue, 4 Feb 2003 21:28:46 +0000 (21:28 +0000)]
Place more stringent checks on process credential relabeling for the Biba
and MLS policies: as we support both an effective (single) element and
range (available) elements, require that the single be in the range if
both the single and range are defined in the update.  Remove comments
suggesting that such a check might be a good idea.

Don't introduce a similar check for network interfaces; due to different
interpretations of the single and range elements, it's not clear that
it's useful to do so.

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

21 years agoRename the variable 'grade' to 'type' in interface parsing and
rwatson [Tue, 4 Feb 2003 21:00:51 +0000 (21:00 +0000)]
Rename the variable 'grade' to 'type' in interface parsing and
labeling for Biba.

Rename the variable 'level' to 'type' in interface parsing and
labeling for MLS.

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

21 years agoRemove another unused doc Makefile.
trhodes [Tue, 4 Feb 2003 20:46:23 +0000 (20:46 +0000)]
Remove another unused doc Makefile.

21 years agoCross ref chflags(1).
trhodes [Tue, 4 Feb 2003 20:38:51 +0000 (20:38 +0000)]
Cross ref chflags(1).

Suggested by: Craig Carey <research@ijs.co.nz) on -doc.

21 years agoAdded MI version of <sys/imgact_aout.h>, "a.out.h", which is
ru [Tue, 4 Feb 2003 18:56:34 +0000 (18:56 +0000)]
Added MI version of <sys/imgact_aout.h>, "a.out.h", which is
always compatible with the i386 version.

This fixes one of the problems I had cross-releasing i386 on
Alpha: the produced "-f aout" binaries are now identical.

21 years agoCorrect grammatical error in previous commit.
des [Tue, 4 Feb 2003 18:47:17 +0000 (18:47 +0000)]
Correct grammatical error in previous commit.

21 years agoRemove 2 more unused doc Makefiles.
trhodes [Tue, 4 Feb 2003 18:26:08 +0000 (18:26 +0000)]
Remove 2 more unused doc Makefiles.

21 years agoExtra precautions before trying to start init(8).
des [Tue, 4 Feb 2003 18:16:50 +0000 (18:16 +0000)]
Extra precautions before trying to start init(8).

21 years agoLet tilde expansion be done even if a file/directory doesn't exist yet.
yar [Tue, 4 Feb 2003 17:50:38 +0000 (17:50 +0000)]
Let tilde expansion be done even if a file/directory doesn't exist yet.
This makes such natural commands as "MKD ~user/newdir" or "STOR ~/newfile"
do what they are supposed to instead of failing miserably with the
"File not found" error.

This involves a bit of code reorganization.  Namely, the code doing
glob(3) expansion has been separated to a function; a new function
has been introduced to do tilde expansion; the latter function is
invoked on a pathname before the former one.  Thus behaviour mimicing
that of the Bourne shell has been achieved.