]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoWhitespace only. Make the style of the contributed sections consistent with
dougb [Sun, 29 Oct 2000 07:47:51 +0000 (07:47 +0000)]
Whitespace only. Make the style of the contributed sections consistent with
the rest of the file. (Read, put it back the way I like it.)

23 years agoFinish the job of conditionalizing UUCP by preventing files in /etc/uucp
dougb [Sun, 29 Oct 2000 06:57:59 +0000 (06:57 +0000)]
Finish the job of conditionalizing UUCP by preventing files in /etc/uucp
from being installed, and make rmail conditional on neither of
NO_SENDMAIL and NOUUCP.

PR: bin/21321
Submitted by: Me

23 years agoChange flags of ncv and nsp driver. Both were conflicted with the flags
non [Sun, 29 Oct 2000 06:54:32 +0000 (06:54 +0000)]
Change flags of ncv and nsp driver. Both were conflicted with the flags
used in lower layer (scsi_low.c).

The flag of ncv for KME KXLC004 was chaged from 0x1 to 0x100.
The flag of nsp for PIO mode was chaged from 0x1 to 0x100.

23 years agoChanged not to use struct isa_device in the argument of scsi_low_activate().
non [Sun, 29 Oct 2000 06:47:16 +0000 (06:47 +0000)]
Changed not to use struct isa_device in the argument of scsi_low_activate().
You will not need COMPAT_OLDISA in config file.
This interface may have to change in future.

23 years agoRemove unnecessary printf().
non [Sun, 29 Oct 2000 06:03:47 +0000 (06:03 +0000)]
Remove unnecessary printf().

23 years agoAdd about ncv, nsp, and stg SCSI drivers and card names which
non [Sun, 29 Oct 2000 05:58:27 +0000 (05:58 +0000)]
Add about ncv, nsp, and stg SCSI drivers and card names which
are supposed to work.

23 years agoAdd hints for ISA cards (such as TMC1610M) which use stg driver.
non [Sun, 29 Oct 2000 05:51:46 +0000 (05:51 +0000)]
Add hints for ISA cards (such as TMC1610M) which use stg driver.
The values has been given by Arai Mikio <m-arai@sco.bekkoame.ne.jp>.

23 years agoFix a bug that shows up when checking out files by date with the
jdp [Sun, 29 Oct 2000 03:56:49 +0000 (03:56 +0000)]
Fix a bug that shows up when checking out files by date with the
"-D date" command line option.  There is code in the original to
handle a special case.  If the date search finds revision 1.1 it
is supposed to check whether revision 1.1.1.1 has the same date
stamp, which would indicate that the file was originally brought
in with "cvs import".  In that case it is supposed to return the
vendor branch version 1.1.1.1.

However, there is a bug in the code.  It actually compares the date
of revision 1.1 for equality with the date given on the command
line -- clearly wrong.  This commit fixes the coding bug.

There is an additional bug which is _not_ fixed in this commit.
The date comparison should not be a strict equality test.  It should
allow a fudge factor of, say, 2-3 seconds.  Old versions of CVS
created the two revisions with two separate invocations of the RCS
"ci" command.  We have many old files in the tree in which the
dates of revisions 1.1 and 1.1.1.1 differ by 1 second.

Approved by: peter

23 years agoReduce kernel stack usage by not having large packets on the stack.
tegge [Sun, 29 Oct 2000 01:19:32 +0000 (01:19 +0000)]
Reduce kernel stack usage by not having large packets on the stack.
Supply correct size parameter to dhcpd.
Replace some magic numbers with macro names.
Handle more than one interface.

23 years agoCount per-address statistics for IP fragments.
joe [Sun, 29 Oct 2000 01:05:09 +0000 (01:05 +0000)]
Count per-address statistics for IP fragments.

Requested by: ru
Obtained from: BSD/OS

23 years agoNuke a bit of dead code.
truckman [Sun, 29 Oct 2000 01:00:36 +0000 (01:00 +0000)]
Nuke a bit of dead code.

23 years agofix minor oops in new data compression page definition
mjacob [Sun, 29 Oct 2000 00:54:59 +0000 (00:54 +0000)]
fix minor oops in new data compression page definition

23 years agoSync with usr.bin/telnet/telnet.c r1.9 - fix buffer overflow in DISPLAY
kris [Sun, 29 Oct 2000 00:10:14 +0000 (00:10 +0000)]
Sync with usr.bin/telnet/telnet.c r1.9 - fix buffer overflow in DISPLAY

23 years agoIgnore ``#'' properly when told to
brian [Sat, 28 Oct 2000 23:56:03 +0000 (23:56 +0000)]
Ignore ``#'' properly when told to

23 years agoMake '-i iface' optional. Assume an0 if it's not otherwise specified.
joe [Sat, 28 Oct 2000 23:33:28 +0000 (23:33 +0000)]
Make '-i iface' optional.  Assume an0 if it's not otherwise specified.

23 years agoRBG -> RGB
nik [Sat, 28 Oct 2000 23:25:41 +0000 (23:25 +0000)]
RBG -> RGB

PR: docs/21724
Submitted by: TOGAWA Satoshi <toga@puyo.org>

23 years agoUnconditionally turning on the I/O and memory enable bits in the PCI
msmith [Sat, 28 Oct 2000 23:07:13 +0000 (23:07 +0000)]
Unconditionally turning on the I/O and memory enable bits in the PCI
command register is too aggressive.  Revert to the previous behaviour, but
leave the new behaviour available as an undocumented option.  It's not
clear what the Right, Right Thing is to do here, but the more conservative
approach is safer.

23 years agoAdd 3 SCSI-3 bits for Sequential Access Device Congifuration Page.
mjacob [Sat, 28 Oct 2000 22:47:47 +0000 (22:47 +0000)]
Add 3 SCSI-3 bits for Sequential Access Device Congifuration Page.
Add Data Compression Page (SCSI-3). Add some clarifying comments
at the top of the file that say what the syntax is.

23 years agoDon't ignore VESA modes that have the NON-VGA bit set. All of the SVGA
jhb [Sat, 28 Oct 2000 22:35:57 +0000 (22:35 +0000)]
Don't ignore VESA modes that have the NON-VGA bit set.  All of the SVGA
graphical modes on some systems have this bit set causing our VESA code
to ignore them.

23 years agoAdd a brief comment telling people to retain 'device miibus' as necessary.
nik [Sat, 28 Oct 2000 22:32:17 +0000 (22:32 +0000)]
Add a brief comment telling people to retain 'device miibus' as necessary.

PR: docs/21981
Submitted by: Matthew Emmerton <matt@gsicomp.on.ca>

23 years agoAdd a new "-h" Human-friendly h/m/s output format.
obrien [Sat, 28 Oct 2000 21:48:53 +0000 (21:48 +0000)]
Add a new "-h" Human-friendly h/m/s output format.

Reviewed by: bde

23 years agoThe latest & greatest unused #include spotter. Faster, one-pass and
phk [Sat, 28 Oct 2000 21:44:05 +0000 (21:44 +0000)]
The latest & greatest unused #include spotter.  Faster, one-pass and
more precise than ever.

1230 hits on the current kernel.

23 years ago* Bring back the guts of crt{i,n}.S. This allows C++ exceptions to work
obrien [Sat, 28 Oct 2000 21:26:48 +0000 (21:26 +0000)]
* Bring back the guts of crt{i,n}.S.  This allows C++ exceptions to work
when using the egcs and gcc-devel ports, along with GCC built from stock
public FSF sources.  With out this change, FreeBSD will be removed from
the list of systems GCC 3.0 must be evaluated on before release.  With
the effort some of us put into getting FreeBSD on this list, we should
not turn this effort into a waste, else we might not be worth fighting
for in the future.  (note that Alpha and IA-64 versions of crt{i,n}.S
are needed)

* Switch from our own crt{begin,in} to those created from GCC's crtstuff.c.
This will allow us to switch to DWARF2 exceptions in the future, along with
staying in sync with any future GCC requirements.

* Break out our ELF branding bits into a seperate file.  Currently this
is now included by our crt1.c files (since this functionality was part of
our native crtbegin.c).  Later crtbrand.o will be merged in the creation
of crti.o.

23 years agoFix this my way. David had absolutely no call overriding MAINTAINER without
des [Sat, 28 Oct 2000 20:53:02 +0000 (20:53 +0000)]
Fix this my way. David had absolutely no call overriding MAINTAINER without
even giving me 24 hours to read his mail and find the bug.

23 years agoInstall per function manpages so one doesn't mistakenly think we don't
obrien [Sat, 28 Oct 2000 20:32:39 +0000 (20:32 +0000)]
Install per function manpages so one doesn't mistakenly think we don't
have manpages for libfetch's functions.

23 years agoAdd missing call to knote_fdclose() in setugidsafety() and fdcloseexec().
alc [Sat, 28 Oct 2000 20:27:32 +0000 (20:27 +0000)]
Add missing call to knote_fdclose() in setugidsafety() and fdcloseexec().

Reviewed by: jlemon

23 years agoRepeat after me, "check to see that a pointer isn't NULL before
obrien [Sat, 28 Oct 2000 20:27:11 +0000 (20:27 +0000)]
Repeat after me, "check to see that a pointer isn't NULL before
dereferenceing it".  This fixes ``pkg_add -r''.

23 years agoadd commented-out dependancies on the new sb8 and sb16 drivers
cg [Sat, 28 Oct 2000 19:35:54 +0000 (19:35 +0000)]
add commented-out dependancies on the new sb8 and sb16 drivers

23 years agomodules for the new sb8 and sb16 drivers
cg [Sat, 28 Oct 2000 19:28:49 +0000 (19:28 +0000)]
modules for the new sb8 and sb16 drivers

23 years agoadd commented-out entries for the new sb8 and sb16 drivers
cg [Sat, 28 Oct 2000 19:24:14 +0000 (19:24 +0000)]
add commented-out entries for the new sb8 and sb16 drivers

23 years agosplit up sb16 and sb/sbpro drivers
cg [Sat, 28 Oct 2000 19:20:03 +0000 (19:20 +0000)]
split up sb16 and sb/sbpro drivers

we do not support sb versions <2.00, and the sb8 driver has not been tested
yet.

these drivers are not yet enabled by default.

23 years agoStyle & grammar fixes.
mpp [Sat, 28 Oct 2000 18:37:37 +0000 (18:37 +0000)]
Style & grammar fixes.

PR: docs/22374, docs/13020

23 years agosome clarification on if_getmtu()
ume [Sat, 28 Oct 2000 17:57:38 +0000 (17:57 +0000)]
some clarification on if_getmtu()
  - avoid to use freed (by freeifaddrs) data
  - 1st try getifaddrs, then try SIOCGIFMTU as the last resort

Submitted by: JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp>
Obtained from: KAME Project

23 years agodefined HAVE_IFM_DATA for (free|net|open)bsd
ume [Sat, 28 Oct 2000 17:54:19 +0000 (17:54 +0000)]
defined HAVE_IFM_DATA for (free|net|open)bsd

Submitted by: JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp>
Obtained from: KAME Project

23 years agoWhen an interface isn't specified use wi0 as the default.
joe [Sat, 28 Oct 2000 16:25:25 +0000 (16:25 +0000)]
When an interface isn't specified use wi0 as the default.
This gets around a segmentation fault that occurs when the interface
isn't specified first before any of the options.

Inspired by PR:  bin/22241

23 years agoWhen an interface isn't specified use wi0 as the default.
joe [Sat, 28 Oct 2000 16:18:41 +0000 (16:18 +0000)]
When an interface isn't specified use wi0 as the default.
This gets around a segmentation fault that occurs when the interface
isn't specified first before any of the options.

Inspired by PR: bin/22241

23 years agoFix a blatant bug when running '/etc/pccard_ether device stop'.
joe [Sat, 28 Oct 2000 14:31:40 +0000 (14:31 +0000)]
Fix a blatant bug when running '/etc/pccard_ether device stop'.
I wasn't dropping the interface correctly, but not noticing because
the interface was going away when the card was ejected.

Submitted by: Motomichi Matsuzaki <mzaki@e-mail.ne.jp>

23 years agocleanups of my previous, quick & dirty fix of unaligned access errors on alpha.
gallatin [Sat, 28 Oct 2000 14:26:23 +0000 (14:26 +0000)]
cleanups of my previous, quick & dirty fix of unaligned access errors on alpha.
submitted by: bde

23 years agoadd PC-Card melody beep(PC Card bus, kludge version)
sanpei [Sat, 28 Oct 2000 13:35:34 +0000 (13:35 +0000)]
add PC-Card melody beep(PC Card bus, kludge version)

Original idea from: PAO3

23 years agoMention the new '40LD' AMI RAID controllers that we now support.
msmith [Sat, 28 Oct 2000 11:51:12 +0000 (11:51 +0000)]
Mention the new '40LD' AMI RAID controllers that we now support.

23 years agoRestore GDC mode to initial mode instead of 24KHz.
nyan [Sat, 28 Oct 2000 11:27:56 +0000 (11:27 +0000)]
Restore GDC mode to initial mode instead of 24KHz.

Submitted by: Tomokazu HARADA <tkhara@osk4.3web.ne.jp>

23 years agoFixed extention memory check routine.
nyan [Sat, 28 Oct 2000 11:16:42 +0000 (11:16 +0000)]
Fixed extention memory check routine.

Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

23 years agoRemoved extra calculation for X position (PC-98 only).
nyan [Sat, 28 Oct 2000 10:59:21 +0000 (10:59 +0000)]
Removed extra calculation for X position (PC-98 only).

Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

23 years agoReturn -10 from the PCI probe to allow room for an updated driver to
msmith [Sat, 28 Oct 2000 10:46:30 +0000 (10:46 +0000)]
Return -10 from the PCI probe to allow room for an updated driver to
override one built into the kernel.

23 years agoFix typo s/DE_DEVICEID_FE2500/DC_DEVICEID_FE2500/
peter [Sat, 28 Oct 2000 10:03:54 +0000 (10:03 +0000)]
Fix typo s/DE_DEVICEID_FE2500/DC_DEVICEID_FE2500/

23 years agoUse EX_USAGE in an example, as the SAME manual page describes a
alex [Sat, 28 Oct 2000 10:02:23 +0000 (10:02 +0000)]
Use EX_USAGE in an example, as the SAME manual page describes a
few lines higher.

PR: 22371
Submitted by: andrew@ugh.net.au

23 years agoAdd some more information to the AUTHORS section.
asmodai [Sat, 28 Oct 2000 09:56:34 +0000 (09:56 +0000)]
Add some more information to the AUTHORS section.

Remove unnecessary duplication of asr after the .Nm macro.

23 years agoAdd PCI IDs for some additional cardbus cards. Yes, there really is
wpaul [Sat, 28 Oct 2000 09:00:20 +0000 (09:00 +0000)]
Add PCI IDs for some additional cardbus cards. Yes, there really is
a RealTek 8139 cardbus device. Unfortunately it doesn't quite work yet
because the CIS parser barfs on it.

Submitted by msmith, with some small tweaks by me.

23 years agoWe should include <machine/stdarg.h> not <stdarg.h>
msmith [Sat, 28 Oct 2000 08:32:43 +0000 (08:32 +0000)]
We should include <machine/stdarg.h> not <stdarg.h>

Submitted by: phk

23 years agoQuick fix for namespace problems which broke many ports. Spell
bde [Sat, 28 Oct 2000 08:20:33 +0000 (08:20 +0000)]
Quick fix for namespace problems which broke many ports.  Spell
uint16_t as "unsigned short" like it used to be and uint32_t as
"unsigned int" not like it used to be (was: "unsigned long").

23 years agoMake the -d option actually work.
alex [Sat, 28 Oct 2000 07:37:57 +0000 (07:37 +0000)]
Make the -d option actually work.
Today fast terminals are usually the default.

PR: 21634
Submitted by: Eric S. Van Gyzen <eric@vangyzen.net>

23 years agoSome AML code assumes that a function without an explicit ReturnOp will
msmith [Sat, 28 Oct 2000 07:32:13 +0000 (07:32 +0000)]
Some AML code assumes that a function without an explicit ReturnOp will
return the last value returned by a nested method call.  This violates
the ACPI spec, but is implemented by the Microsoft interpreter, and thus
vendors can (and do) get away with it.

Intel's stance is that this is illegal and should not be supported.
As they put it, however, we have to live in the real world.  So go ahead
and implement it.

Submitted by: Mitsaru IWASAKI <iwasaki@jp.freebsd.org>

23 years agoFreeBSD-specific customisations:
msmith [Sat, 28 Oct 2000 07:29:08 +0000 (07:29 +0000)]
FreeBSD-specific customisations:

 - Set debugger options for kernel build
 - Define some missing functions
 - Bring in GCC defines
 - Disable the 'wbinvd' macro as it conflicts with our inline

23 years agoAdd some extra fuctions:
msmith [Sat, 28 Oct 2000 07:26:48 +0000 (07:26 +0000)]
Add some extra fuctions:

 - AcpiGetProcessorID (fetch the ACPI processor ID for a given ACPI_HANDLE)
 - AcpiSetSystemSleepState (set the Sx sleeping state, proposed by Intel
   but not actually implemented)

23 years agoAdd some verbiage about the new ACPICA code. Note that this probably
msmith [Sat, 28 Oct 2000 07:07:02 +0000 (07:07 +0000)]
Add some verbiage about the new ACPICA code.  Note that this probably
breaks building this config, as the old ACPI code will conflict.  The
old code will be going away shortly, so this should not be an issue.

23 years agoAllow PCI busses to be connected to host bridges detected by ACPI as well.
msmith [Sat, 28 Oct 2000 07:04:07 +0000 (07:04 +0000)]
Allow PCI busses to be connected to host bridges detected by ACPI as well.

23 years agoConnect the new ACPICA code to the 'acpica' device.
msmith [Sat, 28 Oct 2000 07:03:39 +0000 (07:03 +0000)]
Connect the new ACPICA code to the 'acpica' device.

23 years agoInitial FreeBSD OSPM (operating system power management) modules for
msmith [Sat, 28 Oct 2000 06:59:48 +0000 (06:59 +0000)]
Initial FreeBSD OSPM (operating system power management) modules for
ACPICA.  Most of these are still works in progress.  Support exists for:

 - Fixed feature and control method power, lid and sleep buttons.
 - Detection of ISA PnP devices using ACPI namespace.
 - Detection of PCI root busses using ACPI namespace.
 - CPU throttling and sleep states (incomplete)
 - Thermal monitoring and cooling control (incomplete)
 - Interface to platform embedded controllers (mostly complete)
 - ACPI timer (incomplete)
 - Simple userland control of sleep states.
 - Shutdown and poweroff.

23 years agoFreeBSD-specific OSD (operating system dependant) modules for the Intel
msmith [Sat, 28 Oct 2000 06:56:15 +0000 (06:56 +0000)]
FreeBSD-specific OSD (operating system dependant) modules for the Intel
ACPICA code.

23 years agoRevert two experimental changes which escaped from my devel machine.
phk [Sat, 28 Oct 2000 06:55:12 +0000 (06:55 +0000)]
Revert two experimental changes which escaped from my devel machine.

23 years agoThis commit was generated by cvs2svn to compensate for changes in r67754,
msmith [Sat, 28 Oct 2000 05:01:06 +0000 (05:01 +0000)]
This commit was generated by cvs2svn to compensate for changes in r67754,
which included commits to RCS files with non-trunk default branches.

23 years agoInitial import of the Intel-maintained ACPI Component Architecture. This
msmith [Sat, 28 Oct 2000 05:01:06 +0000 (05:01 +0000)]
Initial import of the Intel-maintained ACPI Component Architecture.  This
is Intel's reference implementation of the core operating-system ACPI
support.  (This import is on the vendor branch.)

23 years agoHook ipfilter back up as it compiles now.
obrien [Sat, 28 Oct 2000 04:51:47 +0000 (04:51 +0000)]
Hook ipfilter back up as it compiles now.

23 years agoAdd a quirk entry for Nakamichi MJ-* changers. The changer in question is
ken [Sat, 28 Oct 2000 03:14:59 +0000 (03:14 +0000)]
Add a quirk entry for Nakamichi MJ-* changers.  The changer in question is
a <NAKAMICH MJ-5.16S 1.06>.

Submitted by: Kevin Van Maren <vanmaren@fast.cs.utah.edu>

23 years agoadd Xircom CreditCard adapters and workalikes(xe driver)
sanpei [Sat, 28 Oct 2000 01:35:21 +0000 (01:35 +0000)]
add Xircom CreditCard adapters and workalikes(xe driver)
to Default (GENERIC) Configuration

23 years agoThe x86 atomic operations are already locked, so they do not need an
jhb [Sat, 28 Oct 2000 00:28:15 +0000 (00:28 +0000)]
The x86 atomic operations are already locked, so they do not need an
additional locked instruction to guarantee a write barrier for the acquire
variants.

Approved by: dfr
Pointy hat to: jhb

23 years agoFix typo in the 4.x case.
msmith [Fri, 27 Oct 2000 23:40:45 +0000 (23:40 +0000)]
Fix typo in the 4.x case.

23 years agoimplement SNDCTL_DSP_POST
cg [Fri, 27 Oct 2000 22:20:45 +0000 (22:20 +0000)]
implement SNDCTL_DSP_POST

23 years agoFix a couple of whitespace nits.
jhb [Fri, 27 Oct 2000 21:45:50 +0000 (21:45 +0000)]
Fix a couple of whitespace nits.

23 years agoRemove 86 unneeded #includes
phk [Fri, 27 Oct 2000 21:32:10 +0000 (21:32 +0000)]
Remove 86 unneeded #includes

23 years agoRemove more unused #includes.
phk [Fri, 27 Oct 2000 21:14:25 +0000 (21:14 +0000)]
Remove more unused #includes.

23 years agoTypo fix.
mpp [Fri, 27 Oct 2000 20:55:15 +0000 (20:55 +0000)]
Typo fix.

23 years agoInitialize rn_mklist in rn_newpair(). The undocumented assumption
wollman [Fri, 27 Oct 2000 20:50:14 +0000 (20:50 +0000)]
Initialize rn_mklist in rn_newpair().  The undocumented assumption
seems to be that the nodes are bzero'd beforehand, but the submitter
found that this was not always the case, and in any event defensive
programming here costs epsilon squared.

PR: 22244
Submitted by: Dave Gillam <daveg@chiaro.com>

23 years agoremove a stddef.h missed by phk
gallatin [Fri, 27 Oct 2000 18:57:31 +0000 (18:57 +0000)]
remove a stddef.h missed by phk

23 years agoExplicitly initialize _pw_passwd.
nectar [Fri, 27 Oct 2000 18:27:07 +0000 (18:27 +0000)]
Explicitly initialize _pw_passwd.

23 years agoAdd usage of M_ZERO to malloc calls where the result was just bzeroed.
mjacob [Fri, 27 Oct 2000 16:40:57 +0000 (16:40 +0000)]
Add usage of M_ZERO to malloc calls where the result was just bzeroed.

PR: 22186
Submitted (partially) by: josh@zipperup.org

23 years agoAdd M_ZERO to malloc calls where the result is just bzeroed.
mjacob [Fri, 27 Oct 2000 16:36:26 +0000 (16:36 +0000)]
Add M_ZERO to malloc calls where the result is just bzeroed.
PR: 22188
Submitted (partially) by: josh@zipperup.org

23 years agoupdate man page to reflect new support
mjacob [Fri, 27 Oct 2000 16:12:53 +0000 (16:12 +0000)]
update man page to reflect new support

23 years agoUse FreeBSD.org, not freebsd.org, as has been the precedent.
asmodai [Fri, 27 Oct 2000 16:05:39 +0000 (16:05 +0000)]
Use FreeBSD.org, not freebsd.org, as has been the precedent.

23 years agoPlease welcome the asr manpage [no, not alt.sysadmin.recovery], but
asmodai [Fri, 27 Oct 2000 15:54:34 +0000 (15:54 +0000)]
Please welcome the asr manpage [no, not alt.sysadmin.recovery], but
the Adaptec SCSI RAID cards.

Submitted by: msmith, scottl, Mark Salyzyn

23 years agoComment out the entry for Workbit Ninja SCSI series in PIO mode.
non [Fri, 27 Oct 2000 13:59:16 +0000 (13:59 +0000)]
Comment out the entry for Workbit Ninja SCSI series in PIO mode.
This confilcted with the entry of non PIO mode.

If you want to use them in PIO mode comment out entry of the non PIO mode
and uncommnet the PIO mode entry.

23 years agoInclude sys/param.h for `__FreeBSD_version' rather than the non-existent
obrien [Fri, 27 Oct 2000 12:53:31 +0000 (12:53 +0000)]
Include sys/param.h for `__FreeBSD_version' rather than the non-existent
osreldate.h.

Submitted by: dougb

23 years agoRemove ipfilter and mark it as broken until it will be fixed by maintainer.
ache [Fri, 27 Oct 2000 12:45:31 +0000 (12:45 +0000)]
Remove ipfilter and mark it as broken until it will be fixed by maintainer.
Not compiles under -current

23 years agoMake RES_OPTIONS=inet6 work.
ume [Fri, 27 Oct 2000 12:34:52 +0000 (12:34 +0000)]
Make RES_OPTIONS=inet6 work.
Basically PR22196, but slightly modified.

PR: bin/22196

23 years agoConvert all users of fldoff() to offsetof(). fldoff() is bad
phk [Fri, 27 Oct 2000 11:45:49 +0000 (11:45 +0000)]
Convert all users of fldoff() to offsetof().  fldoff() is bad
because it only takes a struct tag which makes it impossible to
use unions, typedefs etc.

Define __offsetof() in <machine/ansi.h>

Define offsetof() in terms of __offsetof() in <stddef.h> and <sys/types.h>

Remove myriad of local offsetof() definitions.

Remove includes of <stddef.h> in kernel code.

NB: Kernelcode should *never* include from /usr/include !

Make <sys/queue.h> include <machine/ansi.h> to avoid polluting the API.

Deprecate <struct.h> with a warning.  The warning turns into an error on
01-12-2000 and the file gets removed entirely on 01-01-2001.

Paritials reviews by:   various.
Significant brucifications by:  bde

23 years agoMFS: don't expect result code until you close the data connection
des [Fri, 27 Oct 2000 11:37:21 +0000 (11:37 +0000)]
MFS: don't expect result code until you close the data connection

23 years agoExpand on the reasons for and against defining NO_F00F_HACK. This is one of
dougb [Fri, 27 Oct 2000 10:57:33 +0000 (10:57 +0000)]
Expand on the reasons for and against defining NO_F00F_HACK. This is one of
those options that is frequently misunderstood, and ends up on -questions.

PR: 21852

23 years agoBackout revision 1.7 which was a bad idea since it would force people
dfr [Fri, 27 Oct 2000 09:36:34 +0000 (09:36 +0000)]
Backout revision 1.7 which was a bad idea since it would force people
to reinstall boot1 after a 'make world'.

Unfortunately this means that people who have already installed a new
boot1 from a 'make world' after 2000/09/18 *must* reinstall it after
their next build using something like:

        # disklabel -B /dev/da0c

23 years agoRemove bogus undocumented macros used to control conditional assembly.
phk [Fri, 27 Oct 2000 08:39:41 +0000 (08:39 +0000)]
Remove bogus undocumented macros used to control conditional assembly.

23 years agoRemove #if DO_DEFLATE
phk [Fri, 27 Oct 2000 08:38:26 +0000 (08:38 +0000)]
Remove #if DO_DEFLATE
Remove #if DO_BSD_COMPRESS

They are the wrong way to enable/disable features and undocumented to boot.

23 years agoDeclare or #define per-cpu globals in <machine/globals.h> in all cases.
bde [Fri, 27 Oct 2000 08:30:59 +0000 (08:30 +0000)]
Declare or #define per-cpu globals in <machine/globals.h> in all cases.
The i386 UP case was messily different.

23 years agoFetch the protocol header (TCP, UDP, ICMP) only from the first fragment
ru [Fri, 27 Oct 2000 07:19:17 +0000 (07:19 +0000)]
Fetch the protocol header (TCP, UDP, ICMP) only from the first fragment
of IP datagram.  This fixes the problem when firewall denied fragmented
packets whose last fragment was less than minimum protocol header size.

Found by: Harti Brandt <brandt@fokus.gmd.de>
PR: kern/22309

23 years agoMerge from i386: Add MSP ioctl
roger [Fri, 27 Oct 2000 07:07:24 +0000 (07:07 +0000)]
Merge from i386: Add MSP ioctl

23 years agoAdd Melco LPC2-T/LPC3-TX PCMCIA Ethernet cards.
tanimura [Fri, 27 Oct 2000 06:22:40 +0000 (06:22 +0000)]
Add Melco LPC2-T/LPC3-TX PCMCIA Ethernet cards.

23 years agoAs the blocking model has seems to be troublesome for many, disable
markm [Fri, 27 Oct 2000 06:06:04 +0000 (06:06 +0000)]
As the blocking model has seems to be troublesome for many, disable
it for now with an option.

This option is already deprecated, and will be removed when the
entropy-harvesting code is fast enough to warrant it.

23 years agoAdd PCI device ID for an upcoming revision of the product.
msmith [Fri, 27 Oct 2000 05:58:45 +0000 (05:58 +0000)]
Add PCI device ID for an upcoming revision of the product.

23 years agoAdd a missing newline to a diagnostic.
msmith [Fri, 27 Oct 2000 05:58:16 +0000 (05:58 +0000)]
Add a missing newline to a diagnostic.

23 years agoDefining usage() "__dead" is just ridiculous. It also won't compile with
obrien [Fri, 27 Oct 2000 05:51:59 +0000 (05:51 +0000)]
Defining usage() "__dead" is just ridiculous.  It also won't compile with
a June 2000 GCC 2.96 snapshot.

Reviewed by: bde

23 years agoReport the type of mbuf that couldn't be allocated before
brian [Fri, 27 Oct 2000 05:34:40 +0000 (05:34 +0000)]
Report the type of mbuf that couldn't be allocated before
aborting ppp.

23 years agoAdd and document the WITNESS_SKIPSPIN and WITNESS_DDB kernel options.
jhb [Fri, 27 Oct 2000 03:00:28 +0000 (03:00 +0000)]
Add and document the WITNESS_SKIPSPIN and WITNESS_DDB kernel options.