]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoGC some KSE compatiblity macros that were somehow still here.
jhb [Mon, 26 Nov 2001 23:52:35 +0000 (23:52 +0000)]
GC some KSE compatiblity macros that were somehow still here.

22 years agoGC non-FreeBSD code that didn't work anyways.
jhb [Mon, 26 Nov 2001 23:45:12 +0000 (23:45 +0000)]
GC non-FreeBSD code that didn't work anyways.

22 years agoSome CVS surgery has been done to tip(1) to clean out a commit that I
markm [Mon, 26 Nov 2001 23:18:21 +0000 (23:18 +0000)]
Some CVS surgery has been done to tip(1) to clean out a commit that I
should have done differently.

Restore tip(1) to the build. I will return to this with a clean
import to tip in due course.

22 years agoGrammar police (system console terminal type menu).
bmah [Mon, 26 Nov 2001 23:14:21 +0000 (23:14 +0000)]
Grammar police (system console terminal type menu).

22 years agoAfter some CVS surgery to remove an unwise commit, restore some
markm [Mon, 26 Nov 2001 23:13:33 +0000 (23:13 +0000)]
After some CVS surgery to remove an unwise commit, restore some
post-UUCP-removal stuff:

1) Link cu(1) to tip(1).
2) Force removal of system immutable bit. (Was needed by UUCP's
   insecure cu(1).

22 years agoTemporarily remove tip/cu from the build while repo-surgery is done.
markm [Mon, 26 Nov 2001 22:35:10 +0000 (22:35 +0000)]
Temporarily remove tip/cu from the build while repo-surgery is done.

22 years agoAdd LOMAC options (the "Z" flag in both cases) to display extra information
green [Mon, 26 Nov 2001 22:21:15 +0000 (22:21 +0000)]
Add LOMAC options (the "Z" flag in both cases) to display extra information
in ls(1) and ps(1).

Sponsored by: DARPA, NAI Labs

22 years agoMFS: I was confused. This code wasn't in -current after all.
imp [Mon, 26 Nov 2001 21:25:03 +0000 (21:25 +0000)]
MFS: I was confused.  This code wasn't in -current after all.

Merge in the irq 0 detection.  Add comment about why.

If we have irq 0, ignore it like we do irq 255.  Some BIOS writers aren't
careful like they should be.

22 years agoNew release note: TCP syncache.
bmah [Mon, 26 Nov 2001 20:06:25 +0000 (20:06 +0000)]
New release note:  TCP syncache.

22 years agoBring the ucred(9) manpage in sync with the source.
ru [Mon, 26 Nov 2001 18:37:50 +0000 (18:37 +0000)]
Bring the ucred(9) manpage in sync with the source.

Specifically, document the crshared() function and
fix the prototype and description of the crcopy()
function.

PR: docs/32275
Submitted by: Chad David <davidc@acns.ab.ca>
Reviewed by: jhb

22 years ago- Fix some poor signal handler usage.
arr [Mon, 26 Nov 2001 17:53:02 +0000 (17:53 +0000)]
- Fix some poor signal handler usage.

Reviewed by: -audit (and their silence), jhb, maintainer's silence

22 years agoAdd entries for GDM (GNOME Display Manager).
sobomax [Mon, 26 Nov 2001 16:55:40 +0000 (16:55 +0000)]
Add entries for GDM (GNOME Display Manager).

MFC after: 3 days

22 years agoBack out 1.88 (NO_SYNC_CACHE quirk for Infortrend IFT-3102). Since
iedowse [Mon, 26 Nov 2001 16:54:37 +0000 (16:54 +0000)]
Back out 1.88 (NO_SYNC_CACHE quirk for Infortrend IFT-3102). Since
this device properly reports that the sync cache command is
unsupported, the bug is that we still complain about it on the
console.

Noticed by: gibbs

22 years agoSplit the sio driver for pc98 into bus front end and back end.
nyan [Mon, 26 Nov 2001 12:29:53 +0000 (12:29 +0000)]
Split the sio driver for pc98 into bus front end and back end.
(merged from the files in sys/dev/sio)

22 years agoRestore the ability to use IP_FW_ADD with setsockopt(2) that got
ru [Mon, 26 Nov 2001 10:05:58 +0000 (10:05 +0000)]
Restore the ability to use IP_FW_ADD with setsockopt(2) that got
broken in revision 1.86.  This broke natd(8)'s -punch_fw option.

Reported by: Daniel Rock <D.Rock@t-online.de>,
setantae <setantae@submonkey.net>

22 years agobde suggests that sio really wants to manage its own softc. This
imp [Mon, 26 Nov 2001 07:39:51 +0000 (07:39 +0000)]
bde suggests that sio really wants to manage its own softc.  This
allows us to move the sio softc data structure back into sio.c and
reduce the complexity of the non sio.c sio files.

Submitted by: bde

# I didn't fix the locking issues that bruce also submitted.

22 years agoMake the start memory address that we use configurable via a boot
imp [Mon, 26 Nov 2001 07:17:09 +0000 (07:17 +0000)]
Make the start memory address that we use configurable via a boot
loader parameter.  This allows us to more easily boot on big memory
configuration machines.  hw.pccbb.start_mem.  Reflect this in a sysctl
so we can read it from userland.

# Note: we need a TUNABLE_ULONG to do this right.  I'll add that to
# kernel.h soon.

22 years agoOnly call pccard_function_disable when we have a configured config
imp [Mon, 26 Nov 2001 07:14:00 +0000 (07:14 +0000)]
Only call pccard_function_disable when we have a configured config
entry.  Otherwise we panic on boot for cards that we can't allocate a
config entry for (my 3.3V wavelan card currently fails thusly).

22 years agoExperimental patch to try to properly clip the range of the memory
imp [Mon, 26 Nov 2001 07:12:35 +0000 (07:12 +0000)]
Experimental patch to try to properly clip the range of the memory
request to one that's supported by the bridge.  I'm not 100% sure this
is correct, but it makes it easier for the cardbus bridge to allocate
its memory.

Similar code is needed for the I/O range.  Also, I'm not sure if I
should be doing this based on memory or pmemory (but likely should do
it based on some flag that tells us to prefetch or not).

Talked about a long time ago with: msmith

22 years agoNew release note: sh(1) now implements test as a built-in command, also
bmah [Mon, 26 Nov 2001 06:00:28 +0000 (06:00 +0000)]
New release note:  sh(1) now implements test as a built-in command, also
no longer has printf as a built-in.

Submitted by: knu

22 years agostyle(9) fixes:
imp [Mon, 26 Nov 2001 05:42:45 +0000 (05:42 +0000)]
style(9) fixes:
o declare variables at the start of the function, not the start of
  a block.
o return (retval);
o minor {} tweaks.

22 years agoo Add #include <sys/systm.h> to fix a warning about printf() being
rwatson [Mon, 26 Nov 2001 04:06:47 +0000 (04:06 +0000)]
o Add #include <sys/systm.h> to fix a warning about printf() being
  used without a prototype.

22 years agoo Add 'kern.maxusers' to the list of commented out sample tunable
rwatson [Mon, 26 Nov 2001 03:49:59 +0000 (03:49 +0000)]
o Add 'kern.maxusers' to the list of commented out sample tunable
  variables, since that can now be tuned at boot-time without a
  recompile.

22 years agopeter pointed out that I missed a hard coded location for the memory.
imp [Mon, 26 Nov 2001 02:11:22 +0000 (02:11 +0000)]
peter pointed out that I missed a hard coded location for the memory.
Also add a define for I/O stuff (for cardbus).

22 years agoFirst part of patches to make sio grok 16-bit serial cards under
imp [Mon, 26 Nov 2001 00:39:33 +0000 (00:39 +0000)]
First part of patches to make sio grok 16-bit serial cards under
NEWCARD.  Other patches may be reqiured to sio to prevent a hang on
eject.  Also add commented out entries for sio_pccard.c in files.pc98
to match other architectures.

Submitted by: yamamoto shigeru-san

22 years agoMake the address we kludily allocate in pccbb be a #define for easier
imp [Mon, 26 Nov 2001 00:33:04 +0000 (00:33 +0000)]
Make the address we kludily allocate in pccbb be a #define for easier
changing.  Also change it from 0x44000000 to 0x84000000 for large memory
machines.

# the PCI bus code should do this for us.  This is a bandaide, not a
# solution.

22 years agoLose the line listing CD-{R,RW} drives. Listing these few devices is
bmah [Mon, 26 Nov 2001 00:26:10 +0000 (00:26 +0000)]
Lose the line listing CD-{R,RW} drives.  Listing these few devices is
not particularly helpful, since there are many drives that work with
FreeBSD.

Submitted by: wilko

22 years agoFirst alpha cleanup-patch to try to reduce the amount of hardware
bmah [Mon, 26 Nov 2001 00:24:04 +0000 (00:24 +0000)]
First alpha cleanup-patch to try to reduce the amount of hardware
that shows up on non-i386 hardware lists "by default" (because no
set of hardware platforms was explicitly listed).

Reviewed by: -alpha

Contributions from: Rolf Neugebauer <neugebar@dcs.gla.ac.uk>,
wilko, jhb

22 years agoAdd attribution for lomac(4) feature.
bmah [Mon, 26 Nov 2001 00:17:50 +0000 (00:17 +0000)]
Add attribution for lomac(4) feature.

Requested by: rwatson

22 years agoFix POLA - when selecting line into syscons' cut'n'paste buffer (double
sobomax [Sun, 25 Nov 2001 22:51:30 +0000 (22:51 +0000)]
Fix POLA - when selecting line into syscons' cut'n'paste buffer (double
click) do not include newline into the buffer. This is exacly how
things worked before my recent changes to the cut'n'paste code and
how they work in 4-STABLE.

22 years agoDEVFS has resurfaced.
dd [Sun, 25 Nov 2001 21:06:21 +0000 (21:06 +0000)]
DEVFS has resurfaced.

22 years agoAddress two minor issues: implement the _PC_NAME_MAX and _PC_PATH_MAX
dd [Sun, 25 Nov 2001 21:00:38 +0000 (21:00 +0000)]
Address two minor issues: implement the _PC_NAME_MAX and _PC_PATH_MAX
pathconf() variables for directories, and set st_size and st_blocks
(of struct stat) for directories as appropriate.  Note that st_size is
always set to DEV_BSIZE, since the size of the directories is not
currently kept.

Reviewed by: phk, bde

22 years agoSince kqueue support was added to tail, the "-F" option (check for
iedowse [Sun, 25 Nov 2001 18:03:28 +0000 (18:03 +0000)]
Since kqueue support was added to tail, the "-F" option (check for
renames/rotations) only detected cases where the file itself was
moved or deleted. If part of the path to the file (or a symlink
in the path) was changed instead, tail would not notice.

Fix this by ensuring that we stat the path at least once every
second in the -F case to check for changes. We still use kqueue
when possible to inform us quickly when the file has changed.

PR: bin/24955
Submitted by: Maxim Konovalov <maxim@macomnet.ru>
MFC after: 1 week

22 years agoSet a BOOTABLE variable to "-b" if MACHINE is not pc98, and use it for
nyan [Sun, 25 Nov 2001 16:43:45 +0000 (16:43 +0000)]
Set a BOOTABLE variable to "-b" if MACHINE is not pc98, and use it for
mkisoimages.sh instead of "-b" directly.

MFC after: 1 week

22 years agoStyle clean-up.
sheldonh [Sun, 25 Nov 2001 12:49:32 +0000 (12:49 +0000)]
Style clean-up.

Submitted by: bde

22 years agoContext:
bmilekic [Sun, 25 Nov 2001 04:42:54 +0000 (04:42 +0000)]
Context:
For an object type, we maintain a variable mb_mapfull. It is 0 by default
and is only raised to 1 in one place: when an mb_pop_cont() fails for
the first time, on the assumption that the reason for the failure is
due to the underlying map for the object (e.g. clust_map, mbuf_map) being
exhausted.

Problem and Changes:
Change how we define "mb_mapfull." It now means: "set to 1 when the first
mb_pop_cont() fails only in the kmem_malloc()-ing of the object, and
only if the call was with the M_TRYWAIT flag." This is a more conservative
definition and should avoid odd [but theoretically possible] situations
from occuring. i.e. we had set mb_mapfull to 1 thinking the map for the
object was actually exhausted when we _actually_ failed in malloc()ing
the space for the bucket structure managing the objects in the page
we're allocating.

22 years agoUse 'GET' method instaed of 'HEAD', since some proxy server doesn't work
matusita [Sun, 25 Nov 2001 00:50:57 +0000 (00:50 +0000)]
Use 'GET' method instaed of 'HEAD', since some proxy server doesn't work
with 'HEAD' method.

Actually, when http.c was born, it used 'GET' method.  This was changed
with revision 1.4 (which was submitted as PR: 21449).  I've confirmed
to Philipp Mergenthaler <philipp.mergenthaler@stud.uni-karlsruhe.de>,
the submitter of PR: 21449, and it's absolutely OK that we can use
GET method.

Add missing 'FreeBSD' tag, and copyright notice.  This file is originally
submitted by PR: 11316; I've contacted to the PR originator to submit it.

PR: 32238
Submitted by: Christoph Weber-Fahr <christoph.weber-fahr@arcor.de> (patch),
and Philipp Mergenthaler <un1i@rz.uni-karlsruhe.de> (copyright)
MFC after: 1 week

22 years agoSpelling police: sucessful -> successful.
dd [Sat, 24 Nov 2001 23:41:32 +0000 (23:41 +0000)]
Spelling police: sucessful -> successful.

22 years agoSpelling police: sucessful -> successful.
dd [Sat, 24 Nov 2001 23:38:17 +0000 (23:38 +0000)]
Spelling police: sucessful -> successful.

Submitted by: Anders Andersson <anders@codefactory.se>

22 years agoFix typo.
dd [Sat, 24 Nov 2001 21:22:16 +0000 (21:22 +0000)]
Fix typo.

22 years agoRemove a stale list of supported filesystems. If we do need a list of
dd [Sat, 24 Nov 2001 21:19:46 +0000 (21:19 +0000)]
Remove a stale list of supported filesystems.  If we do need a list of
filesystems somewhere, it probably shouldn't be here, and we should
make a better effort to keep it updated.

22 years agoWhoops... left some junk in previous commit.
luigi [Sat, 24 Nov 2001 20:24:36 +0000 (20:24 +0000)]
Whoops... left some junk in previous commit.

22 years agoLet "make" use the .mk files from the correct source tree instead
luigi [Sat, 24 Nov 2001 20:22:20 +0000 (20:22 +0000)]
Let "make" use the .mk files from the correct source tree instead
of the default ones (which could be incorrect when doing cross builds).

Also, try to be backward compatible when compiling wmk (the goal being
able to use the most recent "picobsd" script on a wide range of
source trees).

22 years agoSpelling police: "more then" - "more than" where appropriate.
dd [Sat, 24 Nov 2001 19:47:12 +0000 (19:47 +0000)]
Spelling police: "more then" - "more than" where appropriate.

22 years agoconvert holdsock() to fget(). Add XXX reminder for future socket locking.
dillon [Sat, 24 Nov 2001 18:28:22 +0000 (18:28 +0000)]
convert holdsock() to fget().  Add XXX reminder for future socket locking.

22 years agoAdd suspend/resume support.
iwasaki [Sat, 24 Nov 2001 18:00:33 +0000 (18:00 +0000)]
Add suspend/resume support.

MFC after: 1 week

22 years agoRemove the stuff about only the superuser being able to use this (see
dd [Sat, 24 Nov 2001 17:06:20 +0000 (17:06 +0000)]
Remove the stuff about only the superuser being able to use this (see
snp.c 1.64), and add a note to the BUGS section warning the reader
that this manual page is stale and may be inaccurate.

22 years agoTalk about what the user needs to do to get the snp devices, and how
dd [Sat, 24 Nov 2001 17:02:58 +0000 (17:02 +0000)]
Talk about what the user needs to do to get the snp devices, and how
watch(8) will try to help them by loading the module.

PR: 25420

22 years agoChange 'superuser' to 'user' which was forgotten in the previous
dd [Sat, 24 Nov 2001 16:56:43 +0000 (16:56 +0000)]
Change 'superuser' to 'user' which was forgotten in the previous
delta.  Remove fake SCCS id while I'm here.

22 years agoPatch to allow the ed driver interrupt routine to terminate if the
imp [Sat, 24 Nov 2001 16:15:18 +0000 (16:15 +0000)]
Patch to allow the ed driver interrupt routine to terminate if the
card is ejected while we're in this routine.

yamamoto-san's original patch had a small race window for AX88190
chips, which I corrected by limiting the number of iterations we'd try
to reset the bits to be about 15ms rather than forever.  This seems to
work for me, but I don't have a large collections of cards based on
this chipset.

Submitted by: YAMAMOTO Shigeru

22 years ago(Forced commit to list actual problems fixed / PRs affected).
ru [Sat, 24 Nov 2001 16:12:04 +0000 (16:12 +0000)]
(Forced commit to list actual problems fixed / PRs affected).

Overview of problems fixed:

- fix support for saving and restoring filter/NAT state information
  (across reboots for example);

- ipmon(8) is started before loading any filter/NAT rules;

- ipmon(8) and ipfs(8) do not solely depend on ipfilter_enable anymore,
  they now also work when only ipnat_enable is true;

- the multiple occurrences of code loading the ipfilter kernel module
  have been removed;

- the options have been removed from the _program variables in
  defaults/rc.conf and the comments in that file have been updated to
  reflect (possibly new) reality;

- the rc.conf.5 manual page has been updated to reflect the changes.

Submitted by: Arjan de Vet <devet@devet.org>
PR: conf/25223, kern/25344, conf/25809,
conf/26275, bin/27016, conf/31482

22 years agoCheck the return value of tsleep() in snpread(). This may be
dd [Sat, 24 Nov 2001 15:59:46 +0000 (15:59 +0000)]
Check the return value of tsleep() in snpread().  This may be
important, since our sleep is interruptable (PCATCH flag).

PR: 30985
Discovered by: Vladimir Jakovenko <vovik@lucky.net>
Submitted by: Valentin Nechayev <netch@segfault.kiev.ua>

22 years agoRemove the sentence about only root being able to run watch(8). It
dd [Sat, 24 Nov 2001 15:51:34 +0000 (15:51 +0000)]
Remove the sentence about only root being able to run watch(8).  It
was never technically true (it's snp(4) that required root, not
watch(8)), and after snp.c 1.64, isn't even effectively true, since
who can run watch(8) depends on the permissions of the snp device(s).

Sort options in SYNOPSIS and DESCRIPTION while I'm here.

22 years agoAdd an -f option which allows one to specify a snp device to use.
dd [Sat, 24 Nov 2001 15:41:38 +0000 (15:41 +0000)]
Add an -f option which allows one to specify a snp device to use.
Previously, watch would always use the first device it could
successfully open, but this isn't always desired.  Specifically, it
may not be desired during debugging (of snp), or if a particular snp
device has different permissions (which makes since after snp.c 1.64).

22 years agos/processes/files/ inherited from kvm_getprocs(3).
ru [Sat, 24 Nov 2001 15:38:28 +0000 (15:38 +0000)]
s/processes/files/ inherited from kvm_getprocs(3).

PR: docs/32252
Submitted by: tobez

22 years agosecond part of the patches to complete ipf changes to rc
darrenr [Sat, 24 Nov 2001 15:36:30 +0000 (15:36 +0000)]
second part of the patches to complete ipf changes to rc

PR: multiple
Submitted by: Arjan de Vet <devet@devet.org>

22 years agoReturn EBUSY if we try to attach to a tty that is already being
dd [Sat, 24 Nov 2001 15:34:18 +0000 (15:34 +0000)]
Return EBUSY if we try to attach to a tty that is already being
snooped on.  This causes all kinds of Bad Things(tm) to happen since
closing one session will clobber state that's needed for the other
one.  This could theoretically be supported if the code was careful,
but until somebody implements that, preventing this will stop people
from unknowingly shooting themselves in the foot.

22 years agoCreate a snpbasedev variable which holds a reference to the first snp
dd [Sat, 24 Nov 2001 15:10:53 +0000 (15:10 +0000)]
Create a snpbasedev variable which holds a reference to the first snp
device cloned, and assign all further devices to depend on it.  This
allows us to call dev_depends() on it at module unload time to get rid
of /dev/snp* (in the devfs case, anyway).  For this to work, we must
not destroy the device at close time.  [Idea stolen from if_tun.]

The above has the following sideaffects: (a) The snp device used by
watch(8) will remain after watch(8) exits.  This is probably how it
should have been all along, and how it was before devfs came along.
(b) Module unload doesn't panic if there are any /dev/snp* devices
which haven't been used (and thus previously destroyed).  Thus, we can
reenable the unload functionality disabled in rev. 1.65.

PR: 32012

22 years agoRevert incorrect KSEfication: realitexpire expects a struct proc *, not a
des [Sat, 24 Nov 2001 14:09:50 +0000 (14:09 +0000)]
Revert incorrect KSEfication: realitexpire expects a struct proc *, not a
struct thread *.

22 years agoResolve all the ipfilter startup issues in rc.network with one big patch
darrenr [Sat, 24 Nov 2001 13:48:30 +0000 (13:48 +0000)]
Resolve all the ipfilter startup issues in rc.network with one big patch
to get it all right, allowing ipnat to be enabled independantly of ipfilter
in rc.conf (among other things).

PR: multiple
Submitted by: Arjan de Vet <devet@devet.org>
Reviewed by: Giorgos Keramidas <keramida@FreeBSD.org>

22 years agoSubmitted by:David E. O'Brien
groudier [Sat, 24 Nov 2001 12:35:48 +0000 (12:35 +0000)]
Submitted by:David E. O'Brien
MFC after:3 days
- Add memory barrier definition for sparc64.
  Patch sent by David E. O'Brien, approved by maintainer.
- Fix an endianization error of a bus physical address used from SCRIPTS
  that made the driver fail on big endian machines as sparc64.

22 years agoAllow an ``at now'' specification.
brian [Sat, 24 Nov 2001 10:43:53 +0000 (10:43 +0000)]
Allow an ``at now'' specification.

PR: 32242
Submitted by: Alan E <alane@geeksrus.net>
MFC after: 3 weeks

22 years agoSince we used '#ifdef __i386__', don't close with '#endif /* !__alpha__ */'
dfr [Sat, 24 Nov 2001 10:11:14 +0000 (10:11 +0000)]
Since we used '#ifdef __i386__', don't close with '#endif /* !__alpha__ */'

22 years agoPC98 boxes cannot boot from CD-ROM.
nyan [Sat, 24 Nov 2001 05:21:53 +0000 (05:21 +0000)]
PC98 boxes cannot boot from CD-ROM.

22 years agoBe more explicit about the fact that realloc() might return a
keramida [Sat, 24 Nov 2001 03:05:40 +0000 (03:05 +0000)]
Be more explicit about the fact that realloc() might return a
different pointer than the one passed to it.

PR: docs/31925
Submitted by: Andrew <andrew@ugh.net.au>

22 years agoWhitespace change - replace leading spaces with tabs.
luigi [Sat, 24 Nov 2001 01:47:50 +0000 (01:47 +0000)]
Whitespace change - replace leading spaces with tabs.

22 years agoRemove the use of _PATH_DEV in the example.
obrien [Sat, 24 Nov 2001 01:34:12 +0000 (01:34 +0000)]
Remove the use of _PATH_DEV in the example.

The kernel certainly doesn't use _PATH_DEV or even /dev/ to find the device.
It cannot, since "/" has not been mounted.  Maybe the only affect of using
/dev/ is that it gets put in the mounted-from name for "/", so that mount(8),
etc., display an absolute path before "/" has been remounted.  Many have
never bothered typing the full path, and code that constructs a path in
rootdevnames[] never bothered to construct a full path, so the example
shouldn't have it.

Submitted by: bde

22 years agoFix decode table
ache [Fri, 23 Nov 2001 22:27:03 +0000 (22:27 +0000)]
Fix decode table

PR: 32233
Submitted by: Basileios Anastasatos <B.Anastasatos@MyRealBox.com>

22 years agoRemove the hack that ensures that rt_Update() works on FreeBSD. Now
brian [Fri, 23 Nov 2001 19:20:45 +0000 (19:20 +0000)]
Remove the hack that ensures that rt_Update() works on FreeBSD.  Now
that the ncpaddr code doesn't create default routes with non-zero
masks, everything works as it should.

22 years agoThe Olympus C-1 digital camera uses a non-standard BBB command-status
iedowse [Fri, 23 Nov 2001 17:35:35 +0000 (17:35 +0000)]
The Olympus C-1 digital camera uses a non-standard BBB command-status
signature, but otherwise behaves just like a normal USB mass-storage
device. Add a new quirk to cover this case, and enable it for C-1
cameras. The quirk enables translation from the C-1 signature to
the normal CSWSIGNATURE value.

Reviewed by: n_hibma

22 years agoWhen writing messages to the routing socket, round sockaddr sizes
brian [Fri, 23 Nov 2001 17:19:36 +0000 (17:19 +0000)]
When writing messages to the routing socket, round sockaddr sizes
up in the same way that we expect them to be when we read them.

This is a no-op on i386 and probably on alphas, as we currently
only support AF_INET and AF_INET6.

22 years agoBe paranoid about non-zero netmasks being associated with INET addresses
brian [Fri, 23 Nov 2001 17:19:27 +0000 (17:19 +0000)]
Be paranoid about non-zero netmasks being associated with INET addresses
of 0.0.0.0.

The OpenBSD PF_ROUTE/NET_RT_DUMP sysctl is sending back routes with
RTAX_NETMASK set, but the corresponding sockaddr being 4 zero bytes
(with an address family of zero).  ppp was getting confused by this
and ending up interpreting it as a 0.0.0.0/32 routing table
destination and subsequently failing to do anything with the route.

Specifically, after this fix, ppp under OpenBSD can successfully
change and delete the default route again !

22 years agoRegen.
iedowse [Fri, 23 Nov 2001 16:36:28 +0000 (16:36 +0000)]
Regen.

22 years agoAdd Olympus C-1.
iedowse [Fri, 23 Nov 2001 16:35:40 +0000 (16:35 +0000)]
Add Olympus C-1.

22 years agoRemove an unused variable (oops)
brian [Fri, 23 Nov 2001 15:47:04 +0000 (15:47 +0000)]
Remove an unused variable (oops)

22 years agoThe RETURN VALUES section is not appropriate for section 1 manual
dd [Fri, 23 Nov 2001 14:40:35 +0000 (14:40 +0000)]
The RETURN VALUES section is not appropriate for section 1 manual
pages; rename it to DIAGNOSTICS.  Also use the .Ex macro while I'm
here.

22 years agoFix grammar in AUTHORS section.
dd [Fri, 23 Nov 2001 14:37:27 +0000 (14:37 +0000)]
Fix grammar in AUTHORS section.

22 years agoAdd suspend/resume code mostly merged from fxp driver.
iwasaki [Fri, 23 Nov 2001 14:27:33 +0000 (14:27 +0000)]
Add suspend/resume code mostly merged from fxp driver.

22 years agoSearch ${MACHINE}/drivers.conf instead of ${MACHINE_ARCH}/drivers.conf.
nyan [Fri, 23 Nov 2001 14:08:25 +0000 (14:08 +0000)]
Search ${MACHINE}/drivers.conf instead of ${MACHINE_ARCH}/drivers.conf.
The previous commit is not completely.

22 years agoDon't adjust_linklocal() when pulling a sockaddr out of an ncpaddr or
brian [Fri, 23 Nov 2001 12:39:20 +0000 (12:39 +0000)]
Don't adjust_linklocal() when pulling a sockaddr out of an ncpaddr or
ncprange structure.

Don't write() the netmask for IPv6 sockaddrs to the routing socket if
the prefixlen is 128.

It seems that messages written to the routing socket with the scopeid
set for link local addresses are not understood.  Instead, we have to
put the scopeid in the 5th and 6th bytes of the address (see
adjust_linklocal() in ncpaddr.c).  I think this may be a bug in the
KAME implementation - it should really understand both forms.

22 years agoFixed a buffer overrun. In my kernel configuration, tcp_syncache happens
bde [Fri, 23 Nov 2001 12:31:27 +0000 (12:31 +0000)]
Fixed a buffer overrun.  In my kernel configuration, tcp_syncache happens
to be followed by nfsnodehashtbl, so bzeroing callouts beyond the end of
tcp_syncache soon caused a null pointer panic when nfsnodehashtbl was
accessed.

22 years agoRegen.
sheldonh [Fri, 23 Nov 2001 12:03:50 +0000 (12:03 +0000)]
Regen.

This file is now generated using src/tools/tools/pciid/mk_pci_vendors.pl,
which merges the Boemler and Heckenbach lists used for rev 1.2.

For now, mk_pci_vendors.pl is called with the -l option, which uses
the entry with the longer description where the same device or vendor
is found in both lists.

If it turns out that this causes to much "back-and-forth" in future
deltas, we can drop the use of the -l option.

22 years agoRemove a bogus log_Id() decl
brian [Fri, 23 Nov 2001 11:15:35 +0000 (11:15 +0000)]
Remove a bogus log_Id() decl

22 years agoAdd us-ascii_to_cp437 screenmap
ache [Fri, 23 Nov 2001 11:15:21 +0000 (11:15 +0000)]
Add us-ascii_to_cp437 screenmap

22 years agoFix usage for the log command
brian [Fri, 23 Nov 2001 11:07:43 +0000 (11:07 +0000)]
Fix usage for the log command

22 years agoWhitespace tweak
brian [Fri, 23 Nov 2001 10:46:28 +0000 (10:46 +0000)]
Whitespace tweak

22 years agoAdd fonts,screenmaps,console types
ache [Fri, 23 Nov 2001 07:40:54 +0000 (07:40 +0000)]
Add fonts,screenmaps,console types

22 years agoAdd support for the `restrict' type qualifier, new in C99.
mike [Fri, 23 Nov 2001 06:55:41 +0000 (06:55 +0000)]
Add support for the `restrict' type qualifier, new in C99.

Reviewed by: wollman

22 years agoValidate requested sleep state in acpi_SetSleepState() to avoid reentry
iwasaki [Fri, 23 Nov 2001 05:57:03 +0000 (05:57 +0000)]
Validate requested sleep state in acpi_SetSleepState() to avoid reentry
during wakeup procedure.

22 years agoAdd table for iso-8859-4 vga9 fonts
ache [Fri, 23 Nov 2001 01:39:42 +0000 (01:39 +0000)]
Add table for iso-8859-4 vga9 fonts

22 years ago1) Map all unused characters into 'space'.
ache [Fri, 23 Nov 2001 01:25:56 +0000 (01:25 +0000)]
1) Map all unused characters into 'space'.
2) iso-8859-4_to_cp437 not do what its name tells at all, remove it

22 years agoStandard conformance cleanup
ache [Fri, 23 Nov 2001 00:10:17 +0000 (00:10 +0000)]
Standard conformance cleanup

22 years agoExpand the comment on the layout of softc, arpcom and ifnet structures,
luigi [Thu, 22 Nov 2001 23:59:56 +0000 (23:59 +0000)]
Expand the comment on the layout of softc, arpcom and ifnet structures,
and list the places where the assumption is used.

22 years agoStandards conformance cleanup
ache [Thu, 22 Nov 2001 23:39:06 +0000 (23:39 +0000)]
Standards conformance cleanup

22 years agoFix typo introduced with last commit.
orion [Thu, 22 Nov 2001 23:38:26 +0000 (23:38 +0000)]
Fix typo introduced with last commit.

22 years agoSync with main
ache [Thu, 22 Nov 2001 21:43:43 +0000 (21:43 +0000)]
Sync with main

22 years agoBack out part of prev. commit for CP437 case, there are lower controls
ache [Thu, 22 Nov 2001 21:41:43 +0000 (21:41 +0000)]
Back out part of prev. commit for CP437 case, there are lower controls
graphics defined in codepage

22 years agoCopy from main termcap
ache [Thu, 22 Nov 2001 21:05:57 +0000 (21:05 +0000)]
Copy from main termcap

22 years agoDon't use lower controls for console ACS. It makes screen or cut copy charset
ache [Thu, 22 Nov 2001 21:04:44 +0000 (21:04 +0000)]
Don't use lower controls for console ACS. It makes screen or cut copy charset
adequate.

22 years agoMerge with main variant
ache [Thu, 22 Nov 2001 19:39:36 +0000 (19:39 +0000)]
Merge with main variant