]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoI neglected to notice that the change text had already been added in a
imp [Tue, 5 Jun 2001 16:43:51 +0000 (16:43 +0000)]
I neglected to notice that the change text had already been added in a
different place. Back out what I added since the other text is better.

Noticed by: nectar

23 years agoForced commit; CVS didn't prompt for a log message for the original
bmah [Tue, 5 Jun 2001 15:33:30 +0000 (15:33 +0000)]
Forced commit; CVS didn't prompt for a log message for the original
commit.

Add a README file where the old *.TXT files used to live, redirecting
developers to the doc directory and giving a URL for a Web-accessible
version of RELNOTESng.

Submitted by: dd (with some editing)

23 years agoPR: i386/26347
hm [Tue, 5 Jun 2001 15:31:47 +0000 (15:31 +0000)]
PR: i386/26347
Submitted by: Nicola Vitale <nivit@libero.it>
Reviewed by: hm
Fix screen number display in HP mode in case PCVT_NSCREENS > 10

23 years agoRevert previous commit. I have no idea how CVS let me do this; it
bmah [Tue, 5 Jun 2001 15:31:35 +0000 (15:31 +0000)]
Revert previous commit.  I have no idea how CVS let me do this; it
didn't even ask me for a log message.

23 years ago*** empty log message ***
bmah [Tue, 5 Jun 2001 15:28:41 +0000 (15:28 +0000)]
*** empty log message ***

23 years agoPR: bin/26817
hm [Tue, 5 Jun 2001 15:16:45 +0000 (15:16 +0000)]
PR: bin/26817
Submitted by: Clement Ballabriga <clement@asso.ups-tlse.fr>
Reviewed by: hm
Check for read errors from the controlling tty when in fullscreen mode.

23 years agoTalk a bit about how cloning works with devfs(5).
brian [Tue, 5 Jun 2001 14:26:17 +0000 (14:26 +0000)]
Talk a bit about how cloning works with devfs(5).
Make it clearer about what's going on with TUNSIFHEAD and TUNSLMODE.
Tidy up a little.

23 years agoUse WITHOUT_X11, not NO_X, as the standardized way of turning off doc
bmah [Tue, 5 Jun 2001 14:16:50 +0000 (14:16 +0000)]
Use WITHOUT_X11, not NO_X, as the standardized way of turning off doc
ports' dependencies on X Windows during release builds.

23 years agoAdd a kqueue filter for writing to ufs filesystems which always returns
jlemon [Tue, 5 Jun 2001 13:52:37 +0000 (13:52 +0000)]
Add a kqueue filter for writing to ufs filesystems which always returns
true.  This permits better interoperability with programs which register
filters on their stdin/stdout handles.

Submitted by: Niels Provos <provos@citi.umich.edu>

23 years agoCorrect cross-references:
sobomax [Tue, 5 Jun 2001 12:50:33 +0000 (12:50 +0000)]
Correct cross-references:
  setsockopt.3 --> setsockopt.2
  syslog.8 --> syslogd.8
  tcpdump.8 --> tcpdump.1

MFC after: 1 week

23 years agoCorrect cross-references:
sobomax [Tue, 5 Jun 2001 12:40:03 +0000 (12:40 +0000)]
Correct cross-references:
  ng_bpf.8    --> ng_bpf.4
  ng_ether.8  --> ng_ether.4
  ng_iface.8  --> ng_iface.4
  ng_pppoe.8  --> ng_pppoe.4
  ng_socket.8 --> ng_socket.4
  ng_tty.8    --> ng_tty.4
  ng_{type}.4 --> /dev/null

MFC after: 1 week

23 years agoFix cross-references:
sobomax [Tue, 5 Jun 2001 12:23:22 +0000 (12:23 +0000)]
Fix cross-references:
  ipnat.8 --> ipnat.1
  environ.5 --> environ.7
  isssetugid.2 --> issetugid.2

MFC after: 1 week

23 years agoDon't use tabs here!
ru [Tue, 5 Jun 2001 07:07:52 +0000 (07:07 +0000)]
Don't use tabs here!

23 years agoFix type-o's, update page
dillon [Tue, 5 Jun 2001 05:59:21 +0000 (05:59 +0000)]
Fix type-o's, update page

23 years agoClose the line displine on detach. Lots of folks have submittd this, and
imp [Tue, 5 Jun 2001 05:58:57 +0000 (05:58 +0000)]
Close the line displine on detach.  Lots of folks have submittd this, and
I think bde even reviewed it once.

Also, change the name of ActionTEC pat to more generic Lucent Kermit
chip.  Add stub for Xircom card.  Add cardbus attachment too.

23 years agoAdd a line discipline close routine which restores some functionality
dd [Tue, 5 Jun 2001 05:07:53 +0000 (05:07 +0000)]
Add a line discipline close routine which restores some functionality
I accidently nuked in rev. 1.54.  Also rework the error handling in
snplwrite a little.

23 years agoStyle and cosmetic cleanups. This driver is now reasonably stlye(9)
dd [Tue, 5 Jun 2001 05:00:17 +0000 (05:00 +0000)]
Style and cosmetic cleanups.  This driver is now reasonably stlye(9)
compliant.  All the variable definitions and function names are
reasonably consistent, and the functions which should be static (i.e.,
all of them) are.  Other assorted fixes were made.  The majority of
the delta is indentation fixes.

Partially reviewed by: bde

23 years agoDocument what happens when the real id is changed.
imp [Tue, 5 Jun 2001 04:35:42 +0000 (04:35 +0000)]
Document what happens when the real id is changed.

I may have obtained this from NetBSD, but I don't recall.  I do know that
I checked FreeBSD's implementation to make sure that it matched what I wrote.

23 years agoRemove #if 0'd code. It turned up on my MAXPATHLEN + 1 sweep.
imp [Tue, 5 Jun 2001 04:29:55 +0000 (04:29 +0000)]
Remove #if 0'd code.  It turned up on my MAXPATHLEN + 1 sweep.

23 years agoOnly build i82365_isa attachment when we have isa bus.
imp [Tue, 5 Jun 2001 04:26:12 +0000 (04:26 +0000)]
Only build i82365_isa attachment when we have isa bus.

23 years agoMinor improvements:
imp [Tue, 5 Jun 2001 04:22:30 +0000 (04:22 +0000)]
Minor improvements:
o Explain snprintf's return value better.
o Document snprintf, et al, were defined in C-99
o Warn against %n.

23 years agoThere seems to be a problem that the order of disk write operation being
obrien [Tue, 5 Jun 2001 01:49:37 +0000 (01:49 +0000)]
There seems to be a problem that the order of disk write operation being
incorrect due to a missing check for some dependency.  This change
avoids the freelist corruption (but not the temporarily inconsistent
state of the file system).

A message is printed as a reminder of the under lying problem when a
pagedep structure is not freed due to the NEWBLOCK flag being set.

Submitted by: Tor.Egge@fast.no

23 years agoAdd a missing word.
dd [Tue, 5 Jun 2001 00:25:18 +0000 (00:25 +0000)]
Add a missing word.

Obtained from: OpenBSD

23 years agoAdd a missing word.
dd [Tue, 5 Jun 2001 00:11:08 +0000 (00:11 +0000)]
Add a missing word.

23 years agoInvert the meaning of the -d option (i.e. default to *not* list dynamic rules,
des [Mon, 4 Jun 2001 23:56:26 +0000 (23:56 +0000)]
Invert the meaning of the -d option (i.e. default to *not* list dynamic rules,
but list them if -d was specified).

Avoid listing expired dynamic rules unless the (new) -e option was specified.

If specific rule numbers were listed on the command line, and the -d flag was
specified, only list dynamic rules that match the specified rule numbers.

Try to partly clean up the bleeding mess this file has become.  If there is
any justice in this world, the responsible parties (you know who you are!)
should expect to wake up one morning with a horse's head in their bed.  The
code still looks like spaghetti, but at least now it's *properly intented*
spaghetti (hmm?  did somebody say "tagliatelle"?).

23 years agoCorrect a spelling nit (a -> an).
dd [Mon, 4 Jun 2001 23:33:02 +0000 (23:33 +0000)]
Correct a spelling nit (a -> an).

23 years agoUse the l_nullioctl exported from tty_conf.c rather than rolling our own.
dd [Mon, 4 Jun 2001 23:31:21 +0000 (23:31 +0000)]
Use the l_nullioctl exported from tty_conf.c rather than rolling our own.

23 years agoUnstaticize l_nullioctl; it is needed elsewhere (like in tty_snoop.c).
dd [Mon, 4 Jun 2001 23:30:47 +0000 (23:30 +0000)]
Unstaticize l_nullioctl; it is needed elsewhere (like in tty_snoop.c).

Suggested by: bde

23 years agoAdded the -l option to df, so to be compatable with other unicies.
pirzyk [Mon, 4 Jun 2001 23:07:15 +0000 (23:07 +0000)]
Added the -l option to df, so to be compatable with other unicies.

PR: bin/27240
Reviewed by: GAWollman
MFC after: 2 weeks

23 years agoAdd sprog.7 to the list.
eric [Mon, 4 Jun 2001 22:15:44 +0000 (22:15 +0000)]
Add sprog.7 to the list.

23 years agoWhile in the interrupt loop, check for a bogus interrupt value of 0xff.
jlemon [Mon, 4 Jun 2001 22:01:44 +0000 (22:01 +0000)]
While in the interrupt loop, check for a bogus interrupt value of 0xff.
This may be returned if the underlying hardware is a pc-card which has
been ejected.

Reviewed by: warner

23 years agoFixed pkg_add so when it does not find a package, it returns != 0.
pirzyk [Mon, 4 Jun 2001 21:35:08 +0000 (21:35 +0000)]
Fixed pkg_add so when it does not find a package, it returns != 0.

PR: bin/23083
Submitted by: Patch submitted by Maxime Henrion <mux@qualys.com>
MFC after: 2 weeks

23 years agoNow that we've got the ne765.h file installed under <dev/ic/>, use it
joerg [Mon, 4 Jun 2001 21:16:28 +0000 (21:16 +0000)]
Now that we've got the ne765.h file installed under <dev/ic/>, use it
here instead of the old "/sys/isa/ic/..." hack.  This should fix
building the world for machines that don't have a /sys symlink,
finally.

23 years agoRound #2 of the sys/isa/ic/ => sys/dev/ic/ move: install sys/dev/ic
joerg [Mon, 4 Jun 2001 21:15:05 +0000 (21:15 +0000)]
Round #2 of the sys/isa/ic/ => sys/dev/ic/ move: install sys/dev/ic
as /usr/include/dev/ic.

23 years agoMove out the files from src/sys/isa/ic/ to src/sys/dev/ic/, so they
joerg [Mon, 4 Jun 2001 21:04:14 +0000 (21:04 +0000)]
Move out the files from src/sys/isa/ic/ to src/sys/dev/ic/, so they
can be made userland-visible as <dev/ic/...>.  Also, those files are
not supposed to contain any bus-specific details at all, so placing
them under .../isa/ has been a misnomer from the beginning.

The files in src/sys/dev/ic/ have been repo-copied from their old
location (this commit is a forced null commit there to record this
message).

23 years agoIn the same manner as getpass(3), block SIGINTR and SIGSTP while we are
markm [Mon, 4 Jun 2001 20:59:49 +0000 (20:59 +0000)]
In the same manner as getpass(3), block SIGINTR and SIGSTP while we are
actually reading the line.

23 years agoReturn an empty string rather than an error at EOF.
markm [Mon, 4 Jun 2001 20:08:55 +0000 (20:08 +0000)]
Return an empty string rather than an error at EOF.

23 years agoInstall this SUID root. Heck, it is su, after all.
markm [Mon, 4 Jun 2001 19:54:35 +0000 (19:54 +0000)]
Install this SUID root. Heck, it is su, after all.

23 years agoExtend the su(1) example to take into account recent options to give
markm [Mon, 4 Jun 2001 19:52:36 +0000 (19:52 +0000)]
Extend the su(1) example to take into account recent options to give
a more BSD flavour.

23 years agoBig module cleanup.
markm [Mon, 4 Jun 2001 19:47:56 +0000 (19:47 +0000)]
Big module cleanup.

Move common stuff into Makefile.inc, and tidy up all the Makefiles
as a result.

Build new modules.

Put a commented-out dependancy on libpam for the (shared) modules.
I can't bring this in just yet, as the dependancy (modules->libpam)
is reversed for the static case (libpam->modules).

23 years agoNull file to bring back a file from the dead. This allows the real commit
markm [Mon, 4 Jun 2001 19:25:41 +0000 (19:25 +0000)]
Null file to bring back a file from the dead. This allows the real commit
to happen remotely. Damn CVS bugs :-(

23 years agoAdd the "nullok" option that causes this module to succeed if the Unix
markm [Mon, 4 Jun 2001 19:16:57 +0000 (19:16 +0000)]
Add the "nullok" option that causes this module to succeed if the Unix
password is empty/null.

23 years agoTidy up the options list (and make it more extendable), and add some
markm [Mon, 4 Jun 2001 19:12:08 +0000 (19:12 +0000)]
Tidy up the options list (and make it more extendable), and add some
extra "standard" options.

23 years agoNew release notes: ti(4) and xl(4) VLAN fixes, RFC 1323/1644 workaround
bmah [Mon, 4 Jun 2001 18:57:54 +0000 (18:57 +0000)]
New release notes:  ti(4) and xl(4) VLAN fixes, RFC 1323/1644 workaround
for old terminal services, IP multicast on VLAN devices works, IPv4
fragmentation denial-of-service mitigation, diskcheckd(8).

23 years agoAdd some new utility authenticators.
markm [Mon, 4 Jun 2001 18:44:47 +0000 (18:44 +0000)]
Add some new utility authenticators.

pam_securetty silently succeeds if the user is on a secure tty
as defined by /etc/ttys.

pam_ftp does "anonymous ftp" style authentication with options for
specifying the anonymous user(s).

23 years agoUse bitmasks of the KTR_* constants instead of hexidecimal values for
jhb [Mon, 4 Jun 2001 18:26:02 +0000 (18:26 +0000)]
Use bitmasks of the KTR_* constants instead of hexidecimal values for
the KTR_COMPILE and KTR_MASK examples.

23 years agoAdd a new psuedo-KTR trace level KTR_ALL which is a mask of all currently
jhb [Mon, 4 Jun 2001 18:24:07 +0000 (18:24 +0000)]
Add a new psuedo-KTR trace level KTR_ALL which is a mask of all currently
used KTR levels.

23 years agofirst blush at some FC path inquiry settings
mjacob [Mon, 4 Jun 2001 18:23:49 +0000 (18:23 +0000)]
first blush at some FC path inquiry settings

23 years agoDo NOLUNS dance for oddball Exabyte. We *really* need to do this as hints.
mjacob [Mon, 4 Jun 2001 18:20:52 +0000 (18:20 +0000)]
Do NOLUNS dance for oddball Exabyte. We *really* need to do this as hints.

Correct match for A5000 SES instance.

PR: 19887
MFC after: 2 weeks

23 years agofirst blush at some FC path inquiry settings
mjacob [Mon, 4 Jun 2001 18:08:30 +0000 (18:08 +0000)]
first blush at some FC path inquiry settings

23 years agoIf the chip isn't in power state D0, put it in power state D0. I
imp [Mon, 4 Jun 2001 17:14:28 +0000 (17:14 +0000)]
If the chip isn't in power state D0, put it in power state D0.  I
elected to do this in the probe rather than the attach so that we don't
disturb things which this might reset.  different cards have different
quirks, according to their datasheets.

This should fix the "I booted in windows and rebooted to FreeBSD and
now things don't work" problem.

PR: 4847, 20670

23 years agoRecommend what to do in a single user boot.
imp [Mon, 4 Jun 2001 16:17:06 +0000 (16:17 +0000)]
Recommend what to do in a single user boot.

PR: 25028
Submitted by: Mark Andrews <marka@nominum.com>

23 years agoAdd sprog.7, the beginnings of the secure programming guide.
eric [Mon, 4 Jun 2001 15:32:44 +0000 (15:32 +0000)]
Add sprog.7, the beginnings of the secure programming guide.

Reviewed by: ru

23 years agoFix the default number of virtual consoles in the usage.hlp file: there
olgeni [Mon, 4 Jun 2001 15:16:41 +0000 (15:16 +0000)]
Fix the default number of virtual consoles in the usage.hlp file: there
were 3 virtual consoles in older releases, but now they are 8.

MFC after: 1 week

23 years agoAdd BSD-style copyright headers
brian [Mon, 4 Jun 2001 15:09:51 +0000 (15:09 +0000)]
Add BSD-style copyright headers

Approved by: Charles Mott <cmott@scientech.com>

23 years agoRemove configuration file for pim6[ds]d.
ume [Mon, 4 Jun 2001 14:53:52 +0000 (14:53 +0000)]
Remove configuration file for pim6[ds]d.

Submitted by: sumikawa

23 years agoChange to a standard BSD-style copyright
brian [Mon, 4 Jun 2001 14:52:17 +0000 (14:52 +0000)]
Change to a standard BSD-style copyright

Approved by: Atsushi Murai <amurai@spec.co.jp>

23 years agoRenamed to README.nat
brian [Mon, 4 Jun 2001 14:40:06 +0000 (14:40 +0000)]
Renamed to README.nat

23 years agoAdd BSD style copyrights (with permission from Charles Mott where appropriate)
brian [Mon, 4 Jun 2001 14:38:29 +0000 (14:38 +0000)]
Add BSD style copyrights (with permission from Charles Mott where appropriate)
Deprecate -alias further (after a repo-copy)

23 years agoWhen looking for an interface appropriate for the (new or changing)
ru [Mon, 4 Jun 2001 14:13:15 +0000 (14:13 +0000)]
When looking for an interface appropriate for the (new or changing)
route in ifa_ifwithroute(), as the last resort, look up the route to
the gateway, not destination (to derive the interface from).

PR: kern/27852
Submitted by: Iasen Kostoff <tbyte@tbyte.org>
MFC after: 2 weeks

23 years agomdoc(7) police: minor markup cleanup.
ru [Mon, 4 Jun 2001 13:21:46 +0000 (13:21 +0000)]
mdoc(7) police: minor markup cleanup.

23 years agomsdos -> msdosfs
ru [Mon, 4 Jun 2001 13:08:54 +0000 (13:08 +0000)]
msdos -> msdosfs
null -> nullfs
umap -> umapfs
union -> unionfs

Approved by: obrien

23 years agoProperly initialize a temporary FILE that is allocated off the stack.
deischen [Mon, 4 Jun 2001 12:36:07 +0000 (12:36 +0000)]
Properly initialize a temporary FILE that is allocated off the stack.
The change to reuse _up from FILE (to allow FILE to grow without changing
size) overlooked FILE being allocated off the stack.

Approved by: sobomax

23 years agoCorrect a comment - the time service returns seconds since 1900 not 1970.
dwmalone [Mon, 4 Jun 2001 11:47:08 +0000 (11:47 +0000)]
Correct a comment - the time service returns seconds since 1900 not 1970.

Submitted by: ru

23 years agoThis patch cleans up the ident stuff in inetd. The code which has
dwmalone [Mon, 4 Jun 2001 11:43:29 +0000 (11:43 +0000)]
This patch cleans up the ident stuff in inetd. The code which has
been patched so many times it was a bit of a mess. There are style,
code and man page cleanups. The following are the functional changes:

The RFC only permits the returning of 4 possible error
codes, make sure we only return these (PR 27636).

Use MAXLOGNAME to determine the longest usernames.

Add a -i flag, which returns the uid instead of the username
(this is from a PR 25787, which also contained alot of the
cleanups in this patch).

PR: 25787, 27636
Partially Submitted by: Arne.Dag.Fidjestol@idi.ntnu.no
Reviewed by: Arne.Dag.Fidjestol@idi.ntnu.no, green
MFC after: 3 weeks

23 years agoResetting using COR bit 7 appears to be bad on pci based systems.
imp [Mon, 4 Jun 2001 07:41:05 +0000 (07:41 +0000)]
Resetting using COR bit 7 appears to be bad on pci based systems.
#ifdef it out for now.

23 years agoSet up us the pccard bomb.
imp [Mon, 4 Jun 2001 07:10:19 +0000 (07:10 +0000)]
Set up us the pccard bomb.

23 years agoThe CL-PD6832 support in the last commit was done by Takeshi
imp [Mon, 4 Jun 2001 07:05:48 +0000 (07:05 +0000)]
The CL-PD6832 support in the last commit was done by Takeshi
Shibagaki-san (shiba@freebsd.org).  I neglected to mention that.

23 years agoAdd new pci attachment for pcic. This supports pci cards as well as
imp [Mon, 4 Jun 2001 06:49:46 +0000 (06:49 +0000)]
Add new pci attachment for pcic.  This supports pci cards as well as
card bus bridges.

We now always use pci interrupts for pci cards.  This will allow us to
more easily configure things.  You must change your IRQ lines in
/etc/pccard.conf to match what we've probed.  I'm not sure the right
way to deal with this right now.

Development of pci pcmcia has been funded by Monzoon Networks AG.  I
am grateful for their generosity.

23 years agoYou can't free a string and then use it as the error message to
paul [Mon, 4 Jun 2001 04:22:57 +0000 (04:22 +0000)]
You can't free a string and then use it as the error message to
errx. Instead use warn to print the string, then cleanup and exit
normally.

This fixes a core dump if the executable to be run doesn't exist.

23 years agoThe pipe_write() code was locking the pipe without busying it first in
dillon [Mon, 4 Jun 2001 04:04:45 +0000 (04:04 +0000)]
The pipe_write() code was locking the pipe without busying it first in
certain cases, and a close() by another process could potentially rip the
pipe out from under the (blocked) locking operation.

Reported-by: Alexander Viro <viro@math.psu.edu>
23 years agoS_IFCHR is not a bit mask, it's just a value in a field. The correct
paul [Mon, 4 Jun 2001 03:39:14 +0000 (03:39 +0000)]
S_IFCHR is not a bit mask, it's just a value in a field. The correct
way to clear that field is to use S_IFMT.

Pointed out by BDE.

23 years ago#defines for pci way interrupt routing.
imp [Mon, 4 Jun 2001 03:36:22 +0000 (03:36 +0000)]
#defines for pci way interrupt routing.

23 years agoMove the pcic interrupt from pcic.c to pcic_isa.c. The ISA handling
imp [Mon, 4 Jun 2001 03:29:06 +0000 (03:29 +0000)]
Move the pcic interrupt from pcic.c to pcic_isa.c.  The ISA handling
for card change interrupts is different than the pci stuff that's
coming soon.  Set the management irq in different ways.  If
pci_parallel interrutp routing, then use the PCI way of getting
interrupts.  Move polling mode into pcic_isa since when we're routing
via pci polling doesn't work because many bridges (systems hang solid).

If we're routing interrupts via pci, they can be shared, so flag them
as such.

Note, this doesn't actually change anything since the pci attachment
isn't quite ready to be committed.

23 years agoMinor style(9) nit. a|b -> a | b.
imp [Mon, 4 Jun 2001 03:13:45 +0000 (03:13 +0000)]
Minor style(9) nit.  a|b -> a | b.

23 years agoUse correct flag in test whether this interrupt handler is fast or not.
mjacob [Mon, 4 Jun 2001 00:52:09 +0000 (00:52 +0000)]
Use correct flag in test whether this interrupt handler is fast or not.

PR: 27866
Submitted by: Peter Jeremy <peter.jeremy@alcatel.com.au>

23 years agoNew release note: pim6dd(8) and pim6sd(8) removed.
bmah [Sun, 3 Jun 2001 23:46:40 +0000 (23:46 +0000)]
New release note:  pim6dd(8) and pim6sd(8) removed.

MFCs noted:  Ukranian console support, O_DIRECT.

23 years agoPrevent denial of service using bogus fragmented IPv4 packets.
jesper [Sun, 3 Jun 2001 23:33:23 +0000 (23:33 +0000)]
Prevent denial of service using bogus fragmented IPv4 packets.

A attacker sending a lot of bogus fragmented packets to the target
(with different IPv4 identification field - ip_id), may be able
to put the target machine into mbuf starvation state.

By setting a upper limit on the number of reassembly queues we
prevent this situation.

This upper limit is controlled by the new sysctl
net.inet.ip.maxfragpackets which defaults to 200,
as the IPv6 case, this should be sufficient for most
systmes, but you might want to increase it if you have
lots of TCP sessions.
I'm working on making the default value dependent on
nmbclusters.

If you want old behaviour (no upper limit) set this sysctl
to a negative value.

If you don't want to accept any fragments (not recommended)
set the sysctl to 0 (zero).

Obtained from: NetBSD
MFC after: 1 week

23 years agoFix my email address. I accidentally cut'npasted the wrong (old)
joerg [Sun, 3 Jun 2001 20:41:21 +0000 (20:41 +0000)]
Fix my email address.  I accidentally cut'npasted the wrong (old)
hostname laste time.

23 years agoAdd diskcheckd to /etc/rc with a knob in rc.conf.
phk [Sun, 3 Jun 2001 20:15:13 +0000 (20:15 +0000)]
Add diskcheckd to /etc/rc with a knob in rc.conf.

Make the default setting YES for now to get some experience with it.

Note: If people starts seeing disk errors because of this then it
should not be backed.

23 years agoAdd diskcheck-daemon.
phk [Sun, 3 Jun 2001 20:02:03 +0000 (20:02 +0000)]
Add diskcheck-daemon.

With a small disk being 20GB these days, chances are pretty good that
an ailing sector will not be read while still being recoverable by
the drive.

Diskcheck daemon will read disks in the background at a low rate and
that way give the diskdrive a chance to detect and correct soft read
errors before they become hard errors.

Idea by: phk
Written by:     ben

23 years agoRemove pim6[ds]d from the tree. The software had a restrictive license
ume [Sun, 3 Jun 2001 18:16:39 +0000 (18:16 +0000)]
Remove pim6[ds]d from the tree.  The software had a restrictive license
than we can handle.  pim6[ds]d are available as ports instead.

23 years agoFirst, wrap the if_up() call into splimp()/splx() because
yar [Sun, 3 Jun 2001 17:31:11 +0000 (17:31 +0000)]
First, wrap the if_up() call into splimp()/splx() because
if_up() must be called at splnet or higher.
Second, set the IFF_RUNNING flag on an interface after its
resources (i.e. tunnel source and destination addresses)
have been set. Note that we don't set IFF_UP because it is
if_up()'s job to do that.

PR: kern/27851
Submitted by: Horacio J. PeÓa <horape@compendium.com.ar>

23 years agoFix a grammar bogon and removed whitespace at EOL.
schweikh [Sun, 3 Jun 2001 16:08:25 +0000 (16:08 +0000)]
Fix a grammar bogon and removed whitespace at EOL.
MFC after: 1 week

23 years ago- Optimise current cursor position displaying routines, to avoid unnecessary
sobomax [Sun, 3 Jun 2001 12:42:19 +0000 (12:42 +0000)]
- Optimise current cursor position displaying routines, to avoid unnecessary
  cursor movements;
- minor cleanups in cursor position displaying routines;
- properly resent "file modified" flag when editing several files.

MFC after: 1 week

23 years agoMove gif_interfaces from an IP6 option to a regular IP option.
brian [Sun, 3 Jun 2001 12:26:56 +0000 (12:26 +0000)]
Move gif_interfaces from an IP6 option to a regular IP option.

PR: 26543
Submitted by: Brooks Davis <brooks@one-eyed-alien.net>
MFC after: 3 weeks

23 years agoFix reversed arguments to pci_write_config()
peter [Sun, 3 Jun 2001 09:45:40 +0000 (09:45 +0000)]
Fix reversed arguments to pci_write_config()

PR: kern/9408
Submitted by: Philipp Mergenthaler <philipp.mergenthaler@stud.uni-karlsruhe.de>

23 years agoRemove unused includes, use *min() inline functions rather than a
dd [Sun, 3 Jun 2001 05:17:39 +0000 (05:17 +0000)]
Remove unused includes, use *min() inline functions rather than a
home-grown macro, rewrite a confusing conditional in snpdevtotty(),
and change ibuf to 512 bytes instead of 1024 bytes in dsnwrite().

Reviewed by: bde

23 years agoWhen tring to find out if this is a request for a write in
dd [Sun, 3 Jun 2001 04:58:51 +0000 (04:58 +0000)]
When tring to find out if this is a request for a write in
kernel_sysctl and userland_sysctl, check for whether new is NULL, not
whether newlen is 0.  This allows one to set a string sysctl to "".

23 years agoMerged from sys/kern/subr_diskmbr.c revision 1.45.
nyan [Sun, 3 Jun 2001 04:44:26 +0000 (04:44 +0000)]
Merged from sys/kern/subr_diskmbr.c revision 1.45.

23 years agoMerged from sys/kern/subr_diskmbr.c revision 1.52.
nyan [Sun, 3 Jun 2001 03:36:57 +0000 (03:36 +0000)]
Merged from sys/kern/subr_diskmbr.c revision 1.52.

23 years agoPut the originator's full name in the From and Reply-to fields, and
dd [Sun, 3 Jun 2001 02:40:37 +0000 (02:40 +0000)]
Put the originator's full name in the From and Reply-to fields, and
stick their username (which sendmail will make into an e-mail address)
inside '<>'.  Sendmail will still DTRT with this, and it conveniently
puts the submitter's name and e-mail address on one line, just like it
should be after "Submitted by" in a commit message.

23 years agoInclude sys/mutex.h to silence a warning.
dd [Sun, 3 Jun 2001 02:19:07 +0000 (02:19 +0000)]
Include sys/mutex.h to silence a warning.

23 years agoRemove unused VNDEVICE variable.
dd [Sun, 3 Jun 2001 01:36:06 +0000 (01:36 +0000)]
Remove unused VNDEVICE variable.

23 years agoUse the new create-link procedure instead of (make element gi: "A" ...).
dd [Sun, 3 Jun 2001 00:08:43 +0000 (00:08 +0000)]
Use the new create-link procedure instead of (make element gi: "A" ...).

23 years agoCorrect varuable name: WITHOUT_X --> WITHOUT_X11.
sobomax [Sat, 2 Jun 2001 23:56:11 +0000 (23:56 +0000)]
Correct varuable name: WITHOUT_X --> WITHOUT_X11.

Reviewed by: ports, portmgr
Approved by: will, obrien

23 years ago"$FreeBSD $" -> "$FreeBSD$" (the space causes CVS to think this isn't
dd [Sat, 2 Jun 2001 23:12:23 +0000 (23:12 +0000)]
"$FreeBSD $" -> "$FreeBSD$" (the space causes CVS to think this isn't
a keyword, so it wasn't expanding it)

23 years agoOnly touch the PCR register in order to set bits for the fxp driver.
jlemon [Sat, 2 Jun 2001 19:51:02 +0000 (19:51 +0000)]
Only touch the PCR register in order to set bits for the fxp driver.
The 3C509-TX card apparently had a slightly different version of the
chip, and has problems when this register is set.  The problem does
not appear on the 3C509{BC} cards, but since only the fxp driver needs
specific bits set, conditionalize on that.

23 years agoRemove double //'s in path.
obrien [Sat, 2 Jun 2001 19:09:37 +0000 (19:09 +0000)]
Remove double //'s in path.

23 years agoSupport GCC 3.0
obrien [Sat, 2 Jun 2001 17:35:50 +0000 (17:35 +0000)]
Support GCC 3.0