]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoIf a driver probe fails, unset it from the device. This fixes a problem
Doug Rabson [Tue, 29 Feb 2000 09:36:25 +0000 (09:36 +0000)]
If a driver probe fails, unset it from the device. This fixes a problem
with certain multiport cards.

Approved by: jkh

24 years agoRevise this for the brave new world of "crypto"
Jordan K. Hubbard [Tue, 29 Feb 2000 09:09:36 +0000 (09:09 +0000)]
Revise this for the brave new world of "crypto"

Submitted by: markm
Approved by: me!

24 years ago* Fix the memory probing code so that it sees all memory clusters, not
Doug Rabson [Tue, 29 Feb 2000 08:48:10 +0000 (08:48 +0000)]
* Fix the memory probing code so that it sees all memory clusters, not
  just the first one.
* Don't reserve extra memory for the prom console unless the platform
  actually uses it.
* Fix some historical confusion and a minor bug in the message buffer
  initialisation.

Submitted by: gallatin for the prom console part
Approved by: jkh

24 years agoAdd the new ses example code directories, unbreak world.
Bill Fumerola [Tue, 29 Feb 2000 08:45:22 +0000 (08:45 +0000)]
Add the new ses example code directories, unbreak world.

Submitted by: cpiazza, who...
Actually tested: make world
Desired by: jkh, mjacob

24 years agoFix a nit in my previous commit: make SIGHUP and SIGCHLD restartable
Joerg Wunsch [Tue, 29 Feb 2000 08:02:29 +0000 (08:02 +0000)]
Fix a nit in my previous commit: make SIGHUP and SIGCHLD restartable
as they ought to be.  The description of SA_RESTART was a little
unobvious to me in the man page, so i missed it.  Thanks to Bruce for
spotting this.

Submitted by: bde

24 years agoAdd definitions for vinum_raid[45] and checkupdates.
Greg Lehey [Tue, 29 Feb 2000 06:22:09 +0000 (06:22 +0000)]
Add definitions for vinum_raid[45] and checkupdates.

Approved-by: jkh
24 years agoDon't print any error message if we can't open the history file. This
Greg Lehey [Tue, 29 Feb 2000 06:21:38 +0000 (06:21 +0000)]
Don't print any error message if we can't open the history file.  This
replaces an older attempt to silence vinum(8) when started in
single-user mode.

Add entries for vinum_raid[45].

Replace the preprocessor variable name CDEV_MAJOR with
VINUM_CDEV_MAJOR.

continue_revive: Set the proc title to 'reviving <sdname>' so that it
can be recognized in a ps list.

Approved-by: jkh
24 years agovinum_ldi: Honour the 'recurse' flag and show the subdisks on this
Greg Lehey [Tue, 29 Feb 2000 06:20:57 +0000 (06:20 +0000)]
vinum_ldi: Honour the 'recurse' flag and show the subdisks on this
drive if it is set.

vinum_lvi: Change line spacing.  This is still not right.

Change checks for striped or RAID-5 plexes to a macro 'isstriped',
which now also includes RAID-4.

Change checks for RAID-5 plexes to a macro 'isparity', which now also
includes RAID-4.

vinum_lsi: If the subdisk is initializing or reviving, don't wait a
whole second to check whether it's advancing; wait in increments of 50
ms up to 1 second.  This makes for a barely perceptible delay.

Approved-by: jkh
24 years agoNew function checkupdates: check whether configuration updates are
Greg Lehey [Tue, 29 Feb 2000 06:19:42 +0000 (06:19 +0000)]
New function checkupdates: check whether configuration updates are
enabled and print a warning message in varous places if they are not.

Fix typos in comments.

Change some indents to approach style(9).

initvol: If init fails, place the subdisks in a "down" state.

vinum_start(): If we're starting a plex, try doing it directly.  It's
possible that the state of the plex differs from that of the subdisks,
for example if somebody has used 'setupstate'.

New functions vinum_raid4 and vinum_raid5 to create RAID-4 and RAID-5
volumes, analagous to vinum_mirror and vinum_stripe.

vinum_checkparity: Don't try to check the parity of a non-parity plex,
print a rude remark instead.

Approved-by: jkh
24 years agoDocument NODESCRYPTLINKS
Kris Kennaway [Tue, 29 Feb 2000 06:17:33 +0000 (06:17 +0000)]
Document NODESCRYPTLINKS

24 years agoReplace the preprocessor variable names BDEV_MAJOR and CDEV_MAJOR with
Greg Lehey [Tue, 29 Feb 2000 06:17:26 +0000 (06:17 +0000)]
Replace the preprocessor variable names BDEV_MAJOR and CDEV_MAJOR with
VINUM_BDEV_MAJOR and VINUM_CDEV_MAJOR respectively.

Set DRIVE_MAXACTIVE and VINUM_MAXACTIVE to 30000, effectively
disabling the request limitation code.  This code was added as an
attempt to escape from a bug which seems to have gone away, and it's
very likely I'll remove the code Real Soon Now, but I don't want to do
it just yet.

struct drive: Remove references to vnode pointers, including debug
output.  Vinum now talks directly to the device driver.  Instead, add
a dev_t.

enum plexorg: Add an instance for RAID-4.

Change checks for striped or RAID-5 plexes to a macro 'isstriped',
which now also includes RAID-4.

Change checks for RAID-5 plexes to a macro 'isparity', which now also
includes RAID-4.

Approved-by: jkh
24 years agoplex_org: Add case for RAID-4.
Greg Lehey [Tue, 29 Feb 2000 06:16:44 +0000 (06:16 +0000)]
plex_org: Add case for RAID-4.

Approved-by: jkh
24 years agoset_drive_state: update subdisk states more correctly.
Greg Lehey [Tue, 29 Feb 2000 06:16:06 +0000 (06:16 +0000)]
set_drive_state: update subdisk states more correctly.

set_sd_state: update the state of a subdisk in a multi-plex volume
more correctly.

update_plex_state: Bring the plex up correctly when the last subdisk
comes up.

checksdstate: Update comments.

vpstate: Don't return an "up" state on a degraded, unattached plex.

start_object: Return a sensible error message when trying to revive a
subdisk whose drive is down.  Previously it returned EBUSY.

Approved-by: jkh
24 years agoFix horrible, embarrassing breakage which caused occasional panics and
Greg Lehey [Tue, 29 Feb 2000 06:15:26 +0000 (06:15 +0000)]
Fix horrible, embarrassing breakage which caused occasional panics and
data corruption.  It's a wonder it worked at all.

Led-on-the-right-path-by: dillon
revive_block: Add treatment for RAID-4.

Add function parityrebuild, called by revive_block and parityops.

Approved-by: jkh
24 years agolaunch_requests: If too many requests are active, include PCATCH in
Greg Lehey [Tue, 29 Feb 2000 06:14:44 +0000 (06:14 +0000)]
launch_requests: If too many requests are active, include PCATCH in
the tsleep call flags.

Submitted-by: Bernd Walter <ticso@cicely.de>
Remove references to vnode pointers, including debug output.  Vinum
now talks directly to the device driver.

bre: Add case for RAID-4.

sdio: Don't try to write to a down drive.  Set the sd state instead.

Approved-by: jkh
24 years agoAdd support for RAID-4. This is the only change which is needed for
Greg Lehey [Tue, 29 Feb 2000 06:14:06 +0000 (06:14 +0000)]
Add support for RAID-4.  This is the only change which is needed for
RAID-4 support: the rest is just configuration management.

Requested-by: peter
Approved-by: jkh

24 years agoMalloc: Save the caller's file name in the correct entry.
Greg Lehey [Tue, 29 Feb 2000 06:13:24 +0000 (06:13 +0000)]
Malloc: Save the caller's file name in the correct entry.

24 years agoAdd keywords 'partition' and 'raid4'.
Greg Lehey [Tue, 29 Feb 2000 06:13:02 +0000 (06:13 +0000)]
Add keywords 'partition' and 'raid4'.

Approved-by: jkh
24 years agoModify comments to indicate that RAID-5 functions also apply to
Greg Lehey [Tue, 29 Feb 2000 06:12:34 +0000 (06:12 +0000)]
Modify comments to indicate that RAID-5 functions also apply to
RAID-4.

Approved-by: jkh
24 years agoopen_drive: open drives directly rather than going via namei and
Greg Lehey [Tue, 29 Feb 2000 06:11:13 +0000 (06:11 +0000)]
open_drive: open drives directly rather than going via namei and
vn_open.  This is necessary in order to be able to open drives before
the root file system is mounted.  This also involves restructuring the
drive struct, which no longer contains a vnode pointer.  Instead,
open_drive sets an open flag.  It's a horrible kludge, and I'll gladly
borrow a Danish axe and hack it in little pieces when devfs comes.

read_drive, write_drive, drive_io_done: Replace with driveio.  The
function names are now macros.

driveio: Fix horrible, embarrassing breakage which was the reason why
read_drive and write_drive existed in the first place.

Code-torn-to-shreds-by: dillon
format_config: Don't save config of objects in referenced state.  They
get rebuilt automatically.

Change checks for striped or RAID-5 plexes to a macro 'isstriped',
which now also includes RAID-4.

Change checks for RAID-5 plexes to a macro 'isparity', which now also
includes RAID-4.

Replace the preprocessor variable names BDEV_MAJOR and CDEV_MAJOR with
VINUM_BDEV_MAJOR and VINUM_CDEV_MAJOR respectively.

vinum_scandisk: Don't free memory twice on error, once is enough.

Approved-by: jkh
24 years agoRemove references to vnode pointers, including debug output. Vinum
Greg Lehey [Tue, 29 Feb 2000 06:10:21 +0000 (06:10 +0000)]
Remove references to vnode pointers, including debug output.  Vinum
now talks directly to the device driver.

Approved-by: jkh
24 years agoRemove declarations of read_drive and write_drive and replace them
Greg Lehey [Tue, 29 Feb 2000 06:09:41 +0000 (06:09 +0000)]
Remove declarations of read_drive and write_drive and replace them
with macros.

Add declaration for parityrebuild, a new function.

Approved-by: jkh
24 years agoAdd support for RAID-4, even though it is inferior in almost every way
Greg Lehey [Tue, 29 Feb 2000 06:08:33 +0000 (06:08 +0000)]
Add support for RAID-4, even though it is inferior in almost every way
to RAID-5.  peter claims that it might be faster for sequential
reading, since the drive caches don't trip over the parity blocks.  I
have seen no evidence to support this, but it's a trivial change.

Requested-by: peter

Change checks for striped or RAID-5 plexes to a macro 'isstriped',
which now also includes RAID-4.

Change checks for RAID-5 plexes to a macro 'isparity', which now also
includes RAID-4.

atoi(): Remove, nobody was talking to it.

give_sd_to_drive: If no space is available, make the subdisk down,
don't delete it.

Change the manner in which the subdisk count was maintained to avoid
cases where the count was not adjusted correctly.

config_drive: Check if we have subdisks referencing us, and add them
if so.  This fixes problems which arose when a drive is replaced in a
running system.

config_sd: Add support for a keyword 'partition', whose meaning will
   be revealed in the fullness of time.

Cosmetic: Shorten some console messages.

Approved-by: jkh
24 years agoChange the startup sequence of the vinum volume from SI_SUB_DRIVERS
Greg Lehey [Tue, 29 Feb 2000 06:07:39 +0000 (06:07 +0000)]
Change the startup sequence of the vinum volume from SI_SUB_DRIVERS
to SI_SUB_VINUM, thus making it possible for Vinum to access I/O
devices and start.

Replace the preprocessor variable names BDEV_MAJOR and CDEV_MAJOR with
VINUM_BDEV_MAJOR and VINUM_CDEV_MAJOR respectively.

Style fixes: replace NULL with 0 where appropriate.

Submitted-by: Charlie Root <root@sms-1.follo.net> (yup, that's all I
      have to go on).

Approved-by: jkh

24 years agoCosmetic changes.
Greg Lehey [Tue, 29 Feb 2000 06:07:01 +0000 (06:07 +0000)]
Cosmetic changes.

Approved-by: jkh

24 years agoAdd in mailbox return codes for failed fabric logins (port_id_used,
Matt Jacob [Tue, 29 Feb 2000 05:54:48 +0000 (05:54 +0000)]
Add in mailbox return codes for failed fabric logins (port_id_used,
loop_id_used, etc...)

Do a more precise structure for Get All Next name server responses.

Approved: jkh

24 years agoMinor non-FreeBSD changes (keeping source sync'd).
Matt Jacob [Tue, 29 Feb 2000 05:53:41 +0000 (05:53 +0000)]
Minor non-FreeBSD changes (keeping source sync'd).
Approved: jkh

24 years agoPrettier print of fabric devices being attached- say what kind of
Matt Jacob [Tue, 29 Feb 2000 05:53:10 +0000 (05:53 +0000)]
Prettier print of fabric devices being attached- say what kind of
port they are (e.g., F_Port vs. N_Port).

Approved: jkh

24 years agoSlightly cleaner fabric support (whiter whites! redder reds!).. No,
Matt Jacob [Tue, 29 Feb 2000 05:52:14 +0000 (05:52 +0000)]
Slightly cleaner fabric support (whiter whites! redder reds!).. No,
seriously- only attempt to logout a previously logged in fabric device.

Fix a longstanding bug for aborting overtime commands- handle halves
have always been reversed.

Clean up some error messages to indicate channel number.

Approved:jkh

24 years agoClean up defines for correct 12160/1080 exclusion. Final 4.0.
Matt Jacob [Tue, 29 Feb 2000 05:48:46 +0000 (05:48 +0000)]
Clean up defines for correct 12160/1080 exclusion. Final 4.0.
approved: JKH

24 years agoAdd NODESCRYPTLINKS knob to prevent spamming of libcrypt -> libscrypt
Kris Kennaway [Tue, 29 Feb 2000 05:47:52 +0000 (05:47 +0000)]
Add NODESCRYPTLINKS knob to prevent spamming of libcrypt -> libscrypt
symlinks. The name is against my better judgement, but I defer to ancient
tradition here because I'm a nice guy.

Reviewed by: -current

24 years agoSome minor fixing - final for 4.0. Debugging messages toned down a bit.
Matt Jacob [Tue, 29 Feb 2000 05:45:50 +0000 (05:45 +0000)]
Some minor fixing - final for 4.0. Debugging messages toned down a bit.
Approved: jkh

24 years agoUpdate SES userland examples to something out of the toy stage.
Matt Jacob [Tue, 29 Feb 2000 05:44:19 +0000 (05:44 +0000)]
Update SES userland examples to something out of the toy stage.
Approved: jkh

24 years agoAdd SI_SUB_VINUM startup sequence for Vinum. This is part of Vinum
Greg Lehey [Tue, 29 Feb 2000 01:57:33 +0000 (01:57 +0000)]
Add SI_SUB_VINUM startup sequence for Vinum.  This is part of Vinum
root file system support.

Approved-by:  jkh

24 years agocorrect wording in BUGS section; it's not possible to dismount swap
Gregory Sutter [Mon, 28 Feb 2000 21:29:37 +0000 (21:29 +0000)]
correct wording in BUGS section; it's not possible to dismount swap
devices, but it's certainly possible to make use of them.

PR: 17013
Submitted by: Giorgos Keramidas <keramida@ceid.upatras.gr>

24 years agounderline "pattern" in the synopsis.
Gregory Sutter [Mon, 28 Feb 2000 21:23:27 +0000 (21:23 +0000)]
underline "pattern" in the synopsis.

PR: 17042
Submitted by: Edwin Groothuis <mavetju@chello.nl>

24 years agoLimit the maximum permissible TCP window size to 65535 octets if
Paul Saab [Mon, 28 Feb 2000 21:18:21 +0000 (21:18 +0000)]
Limit the maximum permissible TCP window size to 65535 octets if
window scaling is disabled.

PR: kern/16914
Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
Reviewed by: wollman
Approved by: jkh

24 years agoA lot of tools need to be built before compilation proper can happen.
Mark Murray [Mon, 28 Feb 2000 21:04:38 +0000 (21:04 +0000)]
A lot of tools need to be built before compilation proper can happen.

24 years agoDon't try to build k5 PAM; it ain't ready yet.
Mark Murray [Mon, 28 Feb 2000 21:00:50 +0000 (21:00 +0000)]
Don't try to build k5 PAM; it ain't ready yet.

24 years agoYesterday I had to fix a badly broken disk, and found that fsck kept dying:
Kirk McKusick [Mon, 28 Feb 2000 20:02:41 +0000 (20:02 +0000)]
Yesterday I had to fix a badly broken disk, and found that fsck kept dying:

  DIR I=64512 CONNECTED. PARENT WAS I=4032
  fsck: cannot find inode 995904

fsdb found the inodes with no problem:

  fsdb (inum: 64512)> inode 995904
  current inode: directory
  I=995904 MODE=40777 SIZE=512
        MTIME=Feb 14 15:27:07 2000 [0 nsec]
        CTIME=Feb 14 15:27:07 2000 [0 nsec]
        ATIME=Feb 24 10:31:58 2000 [0 nsec]
  OWNER=nobody GRP=nobody LINKCNT=4 FLAGS=0 BLKCNT=2 GEN=38a41386
  Direct blocks: 8094568 0 0 0 0 0 0 0 0 0 0 0
  Indirect blocks:  0 0 0

The problem turns out to be a program logic error in fsck.  It stores
directory inodes internally in hash lists, using the number of
directories to form the hash key:

        inpp = &inphead[inumber % numdirs];

Elsewhere, however, it increments numdirs when it finds unattached
directories.  I've made the following fix, which solved the problem in
the case in hand.

Submitted by: Greg Lehey <grog@lemis.com>
Reviewed by: Matthew Dillon <dillon@apollo.backplane.com>
Approved by: Kirk McKusick <mckusick@mckusick.com>

24 years agoGet the order of things right; the keys need to be generated
Mark Murray [Mon, 28 Feb 2000 19:54:06 +0000 (19:54 +0000)]
Get the order of things right; the keys need to be generated
early to allow entropy to replenish.
sshd must start late to catch the full effects of ldconfig.

24 years agoReset the hardware debug registers when exec'ing a new image.
Brian S. Dean [Mon, 28 Feb 2000 19:48:51 +0000 (19:48 +0000)]
Reset the hardware debug registers when exec'ing a new image.

Reviewed by: bde,jlemon
Approved by: jkh

24 years agoThis fixes a problem where the SIOCGIFCONF ioctl goes wrong. This
Guido van Rooij [Mon, 28 Feb 2000 19:30:25 +0000 (19:30 +0000)]
This fixes a problem where the SIOCGIFCONF ioctl goes wrong.  This
is triggered when qmail is used with INET6 enabled.  The bug
manifests itself in that the space variable can become negative
and that in the comparison in the guards of the 2 loops, this was
not noticed because sizeof() returns an unsigned and thus the signed
variable gets promoted to unsigned. I decided not to make space
unsigned because I think we should guard against this from happening.
Thus panic() in case space becomes negative.

Approved by: jkh

24 years agoNew distribution names.
Mark Murray [Mon, 28 Feb 2000 19:25:34 +0000 (19:25 +0000)]
New distribution names.

24 years agoNew distribution name.
Mark Murray [Mon, 28 Feb 2000 19:24:33 +0000 (19:24 +0000)]
New distribution name.

24 years agoGenerate new sshd host key when necessary. I'm tired of
Jordan K. Hubbard [Mon, 28 Feb 2000 19:21:05 +0000 (19:21 +0000)]
Generate new sshd host key when necessary.  I'm tired of
waiting for someone to commit this. :)

24 years agoRemove largescale evidence of crack-smoking.
Mark Murray [Mon, 28 Feb 2000 19:15:32 +0000 (19:15 +0000)]
Remove largescale evidence of crack-smoking.

Where a k4 applet has a k5 namesake, rename the k5 version
from k<app> to k5<app>. (Repo copy done).

Do some repairs to dependancies to support make world properly.

24 years ago1) Add kerberos5 functionality.
Mark Murray [Mon, 28 Feb 2000 19:03:50 +0000 (19:03 +0000)]
1) Add kerberos5 functionality.
   by Daniel Kouril <kouril@informatics.muni.cz>
2) Add full LOGIN_CAP capability
   by Andrey Chernov

24 years agoFix accept(2) behavior in that accepted sockets don't inherit the
Marcel Moolenaar [Mon, 28 Feb 2000 18:58:59 +0000 (18:58 +0000)]
Fix accept(2) behavior in that accepted sockets don't inherit the
parents flags.

Note on the PR:
The PR contains another patch that's not being committed without
further background information. The PR stays open for now.

PR: 16946 (Victor A. Salaman <salaman@teknos.com>)
Prompted by: msmith
Indirect/implicit approval: jkh (shoot me if I'm wrong :-)

24 years agoDon't put truncated hostnames in utmp
Brian Somers [Mon, 28 Feb 2000 18:51:30 +0000 (18:51 +0000)]
Don't put truncated hostnames in utmp

Approved by: jkh

24 years agoPrint scope id for scoped addrs.
Yoshinobu Inoue [Mon, 28 Feb 2000 18:38:19 +0000 (18:38 +0000)]
Print scope id for scoped addrs.

Approved by: jkh

24 years agoPrevious fix was incomplete and has problem in reverse lookup failed dest,
Yoshinobu Inoue [Mon, 28 Feb 2000 18:37:07 +0000 (18:37 +0000)]
Previous fix was incomplete and has problem in reverse lookup failed dest,
so fix it.

Approved by: jkh

Submitted by: Ian West <ian@niw.com.au>

24 years agoFix a serious bug in syslogd regarding the handling of pipes. The bug
Joerg Wunsch [Mon, 28 Feb 2000 17:49:43 +0000 (17:49 +0000)]
Fix a serious bug in syslogd regarding the handling of pipes.  The bug
would cause syslogd to eventually kill innocent processes in the
system over time (note: not `could' but `would').  Many thanks to my
colleague Mirko for digging into the kernel structures and providing
me with the debugging framework to find out about the nature of this
bug (and to isolate that syslogd was the culprit) in a rather large
set of distributed machines at client sites where this happened
occasionally.

Whenever a child process was no longer responsive, or when syslogd
receives a SIGHUP so it closes all its logging file descriptors, for
any descriptor that refers to a pipe syslogd enters the data about the
old logging child process into a `dead queue', where it is being
removed from (and the status of the dead kitten being fetched) upon
receipt of a SIGCHLD.  However, there's a high probability that the
SIGCHLD already arrives before the child's data are actually entered
into the dead queue inside the SIGHUP handler, so the SIGCHLD handler
has nothing to fetch and remove and simply continues.  Whenever this
happens, the process'es data remain on the dead queue forever, and
since domark() tried to get rid of totally unresponsive children by
first sending a SIGTERM and later a SIGKILL, it was only a matter of
time until the system had recycled enough PIDs so an innocent process
got shot to death.

Fix the race by masking SIGHUP and SIGCHLD from both handlers mutually.

Add additional bandaids ``just in case'', i. e. don't enter a process
into the dead queue if we can't signal it (this should only happen in
case it is already dead by that time so we can fetch the status
immediately instead of deferring this to the SIGCHLD handler); for the
kill(2) inside domark(), check for an error status (/* Can't happen */
:) and remove it from the dead queue in this case (which if it would
have been there in the first place would have reduced the problem to a
statistically minimal likelihood so i certainly would never have
noticed the bug at all :).

Mirko also reviewed the fix in priciple (mutual blocking of both
signals inside the handlers), but not the actual code.

Reviewed by: Mirko Kaffka <mirko@interface-business.de>
Approved by: jkh

24 years agoA huge rewrite of the manual page (mostly -mdoc related).
Ruslan Ermilov [Mon, 28 Feb 2000 15:21:12 +0000 (15:21 +0000)]
A huge rewrite of the manual page (mostly -mdoc related).

Reviewed by: luigi, sheldonh

24 years agoFixed configuration of fast interrupts for the pci cy driver. They were
Bruce Evans [Mon, 28 Feb 2000 08:12:24 +0000 (08:12 +0000)]
Fixed configuration of fast interrupts for the pci cy driver.  They were
an early newbus casualty.  The fix in rev.1.28 didn't work because the
most important part of it used a wrong macro name.

Approved by: jkh

24 years agoFixed configuration of fast interrupts for the isa cy driver. They were
Bruce Evans [Mon, 28 Feb 2000 07:52:22 +0000 (07:52 +0000)]
Fixed configuration of fast interrupts for the isa cy driver.  They were
an early newbus casualty.  The isa compatibility cruft turned out to be
well suited for this fix.

Approved by: jkh

24 years agoUpdate the description of NOCRYPT and NOSECURE to match reality.
Kris Kennaway [Mon, 28 Feb 2000 07:07:26 +0000 (07:07 +0000)]
Update the description of NOCRYPT and NOSECURE to match reality.

24 years agoUpdate a comment in elf_coredump to reflect that if you madvise
Paul Saab [Mon, 28 Feb 2000 06:36:45 +0000 (06:36 +0000)]
Update a comment in elf_coredump to reflect that if you madvise
with MADV_NOCORE, its address space is also excluded from a core
file.

Pointed out by: alc

24 years agoAdd MAP_NOCORE to mmap(2), and MADV_NOCORE and MADV_CORE to madvise(2).
Paul Saab [Mon, 28 Feb 2000 04:10:35 +0000 (04:10 +0000)]
Add MAP_NOCORE to mmap(2), and MADV_NOCORE and MADV_CORE to madvise(2).
This
This feature allows you to specify if mmap'd data is included in
an application's corefile.

Change the type of eflags in struct vm_map_entry from u_char to
vm_eflags_t (an unsigned int).

Reviewed by: dillon,jdp,alfred
Approved by: jkh

24 years agoFix a repetition typo about the settings the settings.
Brian Feldman [Mon, 28 Feb 2000 03:02:17 +0000 (03:02 +0000)]
Fix a repetition typo about the settings the settings.

Submitted by: Kris Dow <kris@vilnya.demon.co.uk>

24 years agoAdd OpenSSH blurb and some other minor changes.
Kris Kennaway [Mon, 28 Feb 2000 01:57:15 +0000 (01:57 +0000)]
Add OpenSSH blurb and some other minor changes.

24 years agoFix diagnostic printing test condition (was always true)
Andrey A. Chernov [Mon, 28 Feb 2000 01:48:50 +0000 (01:48 +0000)]
Fix diagnostic printing test condition (was always true)

Approved by: jkh

24 years agoFix a typo, so that the Aironet 4500 really does have an entry here.
Steve Price [Mon, 28 Feb 2000 01:44:07 +0000 (01:44 +0000)]
Fix a typo, so that the Aironet 4500 really does have an entry here.

PR: 17028
Submitted by: David Malone <dwmalone@maths.tcd.ie>
Approved by: jkh

24 years ago-it do, among other things, clear out any
Alfred Perlstein [Mon, 28 Feb 2000 00:31:18 +0000 (00:31 +0000)]
-it do, among other things, clear out any
+it does, amongst other things, clear out any

The old sentance didn't seem to make sense.

24 years agoChange RETTOKEN from '\n' to '\r'; it didn't work under some or all
Matthew Hunt [Sun, 27 Feb 2000 23:02:47 +0000 (23:02 +0000)]
Change RETTOKEN from '\n' to '\r'; it didn't work under some or all
terminals before.  This change tested on FreeBSD syscons, XFree86 xterm,
and Solaris xterm.

24 years agoDocument the following buildworld knobs:
Kris Kennaway [Sun, 27 Feb 2000 22:56:27 +0000 (22:56 +0000)]
Document the following buildworld knobs:

NO_FORTRAN NOCRYPT NOGAMES NOINFO NOLIBC_R NOSECURE NOSHARE

24 years agoDocument NOPERL for disabling building of PERL altogether.
Kris Kennaway [Sun, 27 Feb 2000 22:44:46 +0000 (22:44 +0000)]
Document NOPERL for disabling building of PERL altogether.

24 years agoCompleted start/join code (well only infrastructure mode)
Duncan Barclay [Sun, 27 Feb 2000 19:52:29 +0000 (19:52 +0000)]
Completed start/join code (well only infrastructure mode)

Added receive code and support for Webgear encapsulation.

More debugging macros/functions.

conditionalised timeout for start/join network

conditonalised attribute/common memory hacks

identified tracking code with XXX_TRACK

sorted out initialistion of instance structure to some extent

finished docuementing the start/join sequence

24 years agoInitial import
Duncan Barclay [Sun, 27 Feb 2000 19:47:06 +0000 (19:47 +0000)]
Initial import

24 years agoAdd libipsec.
Yoshinobu Inoue [Sun, 27 Feb 2000 18:41:27 +0000 (18:41 +0000)]
Add libipsec.

Approved by: jkh

24 years agoAdd IPv6 services into inetd.conf.
Yoshinobu Inoue [Sun, 27 Feb 2000 18:39:34 +0000 (18:39 +0000)]
Add IPv6 services into inetd.conf.
Also enable some standard IPv6 apps by default.
These entries will be simply ignored on systems with no INET6 defined.

Approved by: jkh
Suggested by: peter

24 years agoWrap if_up() by splnet.
Yoshinobu Inoue [Sun, 27 Feb 2000 18:36:30 +0000 (18:36 +0000)]
Wrap if_up() by splnet.

Approved by: jkh

Submitted by: peter

24 years agoAt detaching IPv6 raw socket, also finish IPv6 multicast router.
Yoshinobu Inoue [Sun, 27 Feb 2000 18:35:10 +0000 (18:35 +0000)]
At detaching IPv6 raw socket, also finish IPv6 multicast router.

Approved by: jkh

Submitted by: fenner

24 years agoFix style bugs I introduced in the last revision.
Brian Feldman [Sun, 27 Feb 2000 16:40:39 +0000 (16:40 +0000)]
Fix style bugs I introduced in the last revision.

Brucified by: bde

24 years agoFix the problem that caused the boot to fail when modules were loaded.
Søren Schmidt [Sun, 27 Feb 2000 12:41:07 +0000 (12:41 +0000)]
Fix the problem that caused the boot to fail when modules were loaded.

Real braino, confuses two different softc types, I wonder how this
could ever work :(

24 years agoThe best way to remove IPv6 things is to remove lines with the pattern
Brian Feldman [Sun, 27 Feb 2000 07:36:34 +0000 (07:36 +0000)]
The best way to remove IPv6 things is to remove lines with the pattern
'IPv6', and it's not necessary to do it explicitly per entry.

Approved by: jkh

24 years agoDo some cleanups of the IPv6 stuff. This is a non-functional change.
Brian Feldman [Sun, 27 Feb 2000 07:35:42 +0000 (07:35 +0000)]
Do some cleanups of the IPv6 stuff.  This is a non-functional change.

Approved by: jkh

24 years agoRemove option's description that has gone with PAM.
Ruslan Ermilov [Sun, 27 Feb 2000 07:14:33 +0000 (07:14 +0000)]
Remove option's description that has gone with PAM.

24 years agoCrank __FreeBSD_version for OpenSSH
Kris Kennaway [Sun, 27 Feb 2000 06:42:14 +0000 (06:42 +0000)]
Crank __FreeBSD_version for OpenSSH

Approved by: jkh

24 years agopkg_create doesn't use more(1) for displaying +DISPLAY files, it
Kris Kennaway [Sun, 27 Feb 2000 06:32:25 +0000 (06:32 +0000)]
pkg_create doesn't use more(1) for displaying +DISPLAY files, it
just cats it to stdout.

24 years agoAdd an eleventh-hour gimick... But since it's only in `games' and
Joerg Wunsch [Sun, 27 Feb 2000 01:21:28 +0000 (01:21 +0000)]
Add an eleventh-hour gimick...  But since it's only in `games' and
Jordan OK'ed it, i think it's a nice one for those who need it.  I
once mentioned in Usenet that i've hacked morse(6) so it can control
an external device like a tone generator or even a ham radio
transceiver, so you could actually use that `game' for really
transmitting morse code through the air.  A couple of people then
bugged be to send them my hack (which it was by that time).

So i've now finally found some hours to clean it up (like freaking out
with the transmitter no longer keyed in case someone ^c's out of the
program...), and to write the man page addition.

While i was at it, i also cleaned up some minor nits in Lyndon's
/dev/speaker code, mainly #ifdef related stuff like handling getopt()
if SPEAKER is not defined etc.

enjoy & 73

Approved by: jkh, the restless...

24 years agoEnable IPv6 options
Jordan K. Hubbard [Sat, 26 Feb 2000 22:13:21 +0000 (22:13 +0000)]
Enable IPv6 options

24 years agoElide INET6 related options for boot floppies.
Jordan K. Hubbard [Sat, 26 Feb 2000 22:11:27 +0000 (22:11 +0000)]
Elide INET6 related options for boot floppies.

24 years agoEnable IPv6 support by default.
Jordan K. Hubbard [Sat, 26 Feb 2000 22:09:31 +0000 (22:09 +0000)]
Enable IPv6 support by default.

24 years agoAdd iseek= and oseek= aliases for the preexisting skip= and seek=
Brian Feldman [Sat, 26 Feb 2000 21:29:44 +0000 (21:29 +0000)]
Add iseek= and oseek= aliases for the preexisting skip= and seek=
operands.  Can _YOU_ tell skip= and seek= apart with 100% accuracy
every time?

This also seems to make us option-for-option compatible with the
Solaris dd(1).

Approved by: jkh
Suggested by: peter

24 years agoAdd IPv6 related docs.
Yoshinobu Inoue [Sat, 26 Feb 2000 19:44:12 +0000 (19:44 +0000)]
Add IPv6 related docs.

Reviewed by: phantom

24 years agoSpelling: ``adress'' -> ``address''
Chris Costello [Sat, 26 Feb 2000 18:59:34 +0000 (18:59 +0000)]
Spelling:  ``adress'' -> ``address''

24 years agoIntroduce ``add'' and ``remove'' as alternatives to ``alias'' and ``delete''.
Robert Watson [Sat, 26 Feb 2000 18:55:39 +0000 (18:55 +0000)]
Introduce ``add'' and ``remove'' as alternatives to ``alias'' and ``delete''.
Improve compatibility with BSD/OS, and also more accurately reflect that
IP aliases aren't really any different than the primary IP address on an
interface.

Reviewed by: dcs
Approved by: jkh

24 years agoImplement queue-limiting based on bandwidth.
Poul-Henning Kamp [Sat, 26 Feb 2000 18:28:24 +0000 (18:28 +0000)]
Implement queue-limiting based on bandwidth.

Approved by: jkh

24 years agoSync with internat.freebsd.org; weak symbols vs static libs == trouble
Peter Wemm [Sat, 26 Feb 2000 16:57:17 +0000 (16:57 +0000)]
Sync with internat.freebsd.org; weak symbols vs static libs == trouble

24 years agoUse 'following' instead of 'follow' in a couple of places for a slightly
Steve Price [Sat, 26 Feb 2000 16:55:21 +0000 (16:55 +0000)]
Use 'following' instead of 'follow' in a couple of places for a slightly
better read.

24 years agoSnap shot check in.
Duncan Barclay [Sat, 26 Feb 2000 14:49:24 +0000 (14:49 +0000)]
Snap shot check in.

The code doesn't compile (I'll do that in a minute), but functionally
it gets to receveing the interrupt from the start or join net command.

There is a lot of code to deal with the limitations of the PCCard memory
mapping code - I'm changing some of it hence the check in to preserve something
that worked.

24 years agoMerge from internat.freebsd.org; move VERBOSE_STUBS to a better spot.
Peter Wemm [Sat, 26 Feb 2000 14:20:18 +0000 (14:20 +0000)]
Merge from internat.freebsd.org; move VERBOSE_STUBS to a better spot.

24 years agoMerge from internat.freebsd.org repo, minus change to rsa_eay.c (missing)
Peter Wemm [Sat, 26 Feb 2000 13:19:18 +0000 (13:19 +0000)]
Merge from internat.freebsd.org repo, minus change to rsa_eay.c (missing)

Reorganize and unify libcrypto's interface so that the RSA implementation
is chosen at runtime via dlopen().

This is a checkpoint and may require more tweaks still.

24 years agoSuggest ppp -nat, not ppp -alias
Brian Somers [Sat, 26 Feb 2000 13:13:16 +0000 (13:13 +0000)]
Suggest ppp -nat, not ppp -alias

24 years agoMerge from internat.freebsd.org repo, minus change to rsa_eay.c (missing)
Peter Wemm [Sat, 26 Feb 2000 13:13:03 +0000 (13:13 +0000)]
Merge from internat.freebsd.org repo, minus change to rsa_eay.c (missing)

Reorganize and unify libcrypto's interface so that the RSA implementation
is chosen at runtime via dlopen().

This is a checkpoint and may require more tweaks still.

24 years agoAt great personal risk (to my already fragile sanity), reorganize
Peter Wemm [Sat, 26 Feb 2000 13:06:55 +0000 (13:06 +0000)]
At great personal risk (to my already fragile sanity), reorganize
the rsa stubs for libcrypto.  libcrypto.so now uses dlopen() to
implement the backends for either the native or rsaref implemented
RSA code.
This involves:
- unifying the libcrypto and openssl(1) source so there is no
  #ifdef RSAref variations.
- using weak symbols and dlopen()/dlsym() routines to access the
  rsa method vectors.

Releases will enable the user to choose International, US (rsaref) or
no RSA code at install time.
'make world' will DTRT depending on whether you have the international
or US source.  For US users, you must either install rsaref (the port
or package) or (if you don't fear RSA Inc) use the (superior)
International rsa_eay.c code.

This has been discussed at great length by the affected folks and even
we have a great deal of confusion.  This is a checkpoint so we can tune
the results.  This works for me in all permutations I can think of and
should result in a CD/ftp 'release' just about doing the right thing now.

24 years agoBe neater with the screen contents.
Jordan K. Hubbard [Sat, 26 Feb 2000 12:43:01 +0000 (12:43 +0000)]
Be neater with the screen contents.

24 years agoFix benign but stupid variable typo.
Jordan K. Hubbard [Sat, 26 Feb 2000 12:35:00 +0000 (12:35 +0000)]
Fix benign but stupid variable typo.

24 years agoMake all menu items more consistent and deal with a few remaining
Jordan K. Hubbard [Sat, 26 Feb 2000 12:33:49 +0000 (12:33 +0000)]
Make all menu items more consistent and deal with a few remaining
bogus accelerators.

Based largely on feedback from: Basit Syed <syed@pinenut.nosc.mil>