]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoDrop login.c and logwtmp.c from the library build. None of the apps
peter [Fri, 4 Oct 2002 00:24:31 +0000 (00:24 +0000)]
Drop login.c and logwtmp.c from the library build.  None of the apps
that we build use it, and any that do should be using the official
login()/logwtmp() in libutil.

21 years agoUpdate stubs for post-kseIII.
peter [Fri, 4 Oct 2002 00:19:58 +0000 (00:19 +0000)]
Update stubs for post-kseIII.

21 years agoUpdate for post-kseIII
peter [Fri, 4 Oct 2002 00:18:21 +0000 (00:18 +0000)]
Update for post-kseIII

21 years agoCorrect a problem wherein NFS servers running NFSv2 would not return
rwatson [Thu, 3 Oct 2002 21:50:37 +0000 (21:50 +0000)]
Correct a problem wherein NFS servers running NFSv2 would not return
certain classes of failure responses to the client during a failed
remove operation.

Submitted by: Ian Dowse <iedowse@maths.tcd.ie>

21 years agoFix all whitespace style bugs where spaces have been used
schweikh [Thu, 3 Oct 2002 20:39:29 +0000 (20:39 +0000)]
Fix all whitespace style bugs where spaces have been used
where tabs should have been. Cut & paste disease?

Spotted by: vi(1) tabstop=4
MFC after: 2 weeks

21 years agoMove swblk_t from sys/types.h to sys/dmap.h, the only include that uses
peter [Thu, 3 Oct 2002 20:30:09 +0000 (20:30 +0000)]
Move swblk_t from sys/types.h to sys/dmap.h, the only include that uses
it.  It's also only used in vm/vm_swap.c, but that is also the only source
file that #include's <sys/dmap.h>.  sys/dmap.h could probably be embedded
entirely in vm_swap.c since that is the only consumer of it.

21 years agoFix style problems in the quirk entry section.
ken [Thu, 3 Oct 2002 19:59:24 +0000 (19:59 +0000)]
Fix style problems in the quirk entry section.

MFC after: 2 weeks

21 years agoAdd stpcpy(3).
obrien [Thu, 3 Oct 2002 19:51:04 +0000 (19:51 +0000)]
Add stpcpy(3).

21 years agoAdd stpcpy(3).
obrien [Thu, 3 Oct 2002 19:39:20 +0000 (19:39 +0000)]
Add stpcpy(3).

21 years agoActually implement gzip's -c option. minigzip now also doubles
johan [Thu, 3 Oct 2002 19:30:22 +0000 (19:30 +0000)]
Actually implement gzip's -c option. minigzip now also doubles
as zcat properly.

PR: 13043
Submitted by: Toshihiko ARAI <toshi@jp.freebsd.org>
Jonathan Towne <jontow@twcny.rr.com>

Approved by: peter, sheldonh (mentor)
MFC After: 2 weeks

21 years agoFix various style(9) bugs:
mike [Thu, 3 Oct 2002 18:34:13 +0000 (18:34 +0000)]
Fix various style(9) bugs:
o Unusual order of #ifndef _FOO_H_, followed by license.
o Missing tabs in typedef and #define.
o Reversed comment for #endif.

21 years agoCope with a GCC bug by using an alternative, but equivalent function
mike [Thu, 3 Oct 2002 18:12:03 +0000 (18:12 +0000)]
Cope with a GCC bug by using an alternative, but equivalent function
prototype for regexec().

Noticed by: robert

21 years agoAllocate 64K recieve buffer for DNS responses.
ume [Thu, 3 Oct 2002 17:25:33 +0000 (17:25 +0000)]
Allocate 64K recieve buffer for DNS responses.
Though res_query.c also defines and refers MAXPACKET, it is not
related to ansbuf.  So, I didn't touch res_query.c.

21 years agoOops, the previous version was a last minute test version with off_t
bde [Thu, 3 Oct 2002 17:19:36 +0000 (17:19 +0000)]
Oops, the previous version was a last minute test version with off_t
replaced by int instead of size_t.

Spotted by: fanf

21 years agoConnected boot2.
nyan [Thu, 3 Oct 2002 16:21:40 +0000 (16:21 +0000)]
Connected boot2.

21 years agoAdded some header files from -stable and fixed the boot[12] programs.
nyan [Thu, 3 Oct 2002 16:20:14 +0000 (16:20 +0000)]
Added some header files from -stable and fixed the boot[12] programs.

21 years agoUse a non-bogus type for representing the values of offsets in structs.
bde [Thu, 3 Oct 2002 15:00:07 +0000 (15:00 +0000)]
Use a non-bogus type for representing the values of offsets in structs.
off_t is for offsets in files, and it is signed so it was no better
than the original type of int for avoiding warnings from broken lints,
except accidentally on machines like i386's where size_t is smaller
than off_t.

21 years agoFix some twelve hour clock braindamage. The range for hours is 0-23 not 1-12.
fanf [Thu, 3 Oct 2002 14:11:29 +0000 (14:11 +0000)]
Fix some twelve hour clock braindamage. The range for hours is 0-23 not 1-12.

MFC After: 1 week

21 years agoImprove three instances of questionable or confusing grammar.
tjr [Thu, 3 Oct 2002 14:09:06 +0000 (14:09 +0000)]
Improve three instances of questionable or confusing grammar.

21 years agoAdd an example.
tjr [Thu, 3 Oct 2002 14:07:26 +0000 (14:07 +0000)]
Add an example.

21 years agoUpdated text for the MAC framework, UFS2.
rwatson [Thu, 3 Oct 2002 13:46:09 +0000 (13:46 +0000)]
Updated text for the MAC framework, UFS2.

21 years agoDocument towlower() and towupper() in separate manual pages instead of
tjr [Thu, 3 Oct 2002 11:23:06 +0000 (11:23 +0000)]
Document towlower() and towupper() in separate manual pages instead of
trying to confusingly document both on the same page. The new manual pages
are based on tolower(3) and toupper(3) instead of the old towlower(3).

21 years agoPoint out that although toupper() and tolower() really accept rune_t's
tjr [Thu, 3 Oct 2002 11:14:00 +0000 (11:14 +0000)]
Point out that although toupper() and tolower() really accept rune_t's
and not just unsigned char's, callers should use towupper() and towlower()
instead when working with wide characters if portability is a concern.

21 years agoHTTP authentication got broken in rev. 1.49.
ru [Thu, 3 Oct 2002 10:42:19 +0000 (10:42 +0000)]
HTTP authentication got broken in rev. 1.49.

21 years agotowlower() appeared twice in the synopsis; one of the occurrences should
tjr [Thu, 3 Oct 2002 10:40:01 +0000 (10:40 +0000)]
towlower() appeared twice in the synopsis; one of the occurrences should
have been towupper(). Add towupper() to the Name section while I'm at it.

Obtained from: NetBSD (junyoung)

21 years agoAdd an Examples section with an example of how to use the functions.
tjr [Thu, 3 Oct 2002 08:49:29 +0000 (08:49 +0000)]
Add an Examples section with an example of how to use the functions.

21 years agoStatic'ify a variable.
alfred [Thu, 3 Oct 2002 06:44:01 +0000 (06:44 +0000)]
Static'ify a variable.

Submitted by: Matt Emmerton <matt@gsicomp.on.ca>

21 years agoAdd missing const qualifier in tfind().
mike [Thu, 3 Oct 2002 06:33:33 +0000 (06:33 +0000)]
Add missing const qualifier in tfind().

21 years agoFix various style(9) bugs:
mike [Thu, 3 Oct 2002 06:31:16 +0000 (06:31 +0000)]
Fix various style(9) bugs:
o Source ID's in wrong location.
o Space used, instead of tab, after typedef.
o Unaligned function prototype for twalk().

Other changes:
o Add missing const qualifier in tfind().
o Add comment about missing functions.

21 years agoRemove namespace pollution by not including <sys/types.h> and
mike [Thu, 3 Oct 2002 06:27:50 +0000 (06:27 +0000)]
Remove namespace pollution by not including <sys/types.h> and
<time.h>, instead provide a local typedef for pid_t and forward
declare struct timespec.

21 years agoCorrect an indentation.
anholt [Thu, 3 Oct 2002 06:10:54 +0000 (06:10 +0000)]
Correct an indentation.

Noticed by: phk

21 years agohashinit() calls MALLOC(), so release the filedesc lock in knote_attach()
truckman [Thu, 3 Oct 2002 06:03:26 +0000 (06:03 +0000)]
hashinit() calls MALLOC(), so release the filedesc lock in knote_attach()
before calling hashinit() and relock afterwards, taking care to see that
we don't lose a race.

21 years agoIn lieu of a good way to prevent every possible looping in make(1), stop
jmallett [Thu, 3 Oct 2002 04:32:47 +0000 (04:32 +0000)]
In lieu of a good way to prevent every possible looping in make(1), stop
there from being more than 500 processes forked by make(1), to prevent a
forkbomb from happening, in a dumb and mechanical way.

PR: alane
Submitted by: bin/42772
MFC after: 2 weeks

21 years agoXXX Add a check for p->p_limit being NULL before dereferencing it. This is
jmallett [Thu, 3 Oct 2002 04:09:00 +0000 (04:09 +0000)]
XXX Add a check for p->p_limit being NULL before dereferencing it.  This is
totally bogus but will hide the occurances of access of 0xbc(NULL) which
people have run into lately.  This is not a proper fix, just a bandaid, until
the cause of this happening is tracked down and fixed.

Reviewed by: rwatson

21 years agoIn an SMP environment post-Giant it is no longer safe to blindly
truckman [Thu, 3 Oct 2002 02:13:00 +0000 (02:13 +0000)]
In an SMP environment post-Giant it is no longer safe to blindly
dereference the struct sigio pointer without any locking.  Change
fgetown() to take a reference to the pointer instead of a copy of the
pointer and call SIGIO_LOCK() before copying the pointer and
dereferencing it.

Reviewed by: rwatson

21 years agoset ke_bound to NULL when kse owner thread becomes runnable.
davidxu [Thu, 3 Oct 2002 01:22:05 +0000 (01:22 +0000)]
set ke_bound to NULL when kse owner thread becomes runnable.

Reviewed by: julian (mentor)

21 years agoFix various style(9) bugs:
mike [Wed, 2 Oct 2002 23:25:38 +0000 (23:25 +0000)]
Fix various style(9) bugs:
o Unusual order of #ifndef _FOO_H_, followed by license.
o Missing tab in struct sched_param between type and member name.
o Space used, instead of tab, after #define.
o Reversed comment for #endif.
o Irregular comment block.
o Space used, instead of tab, to seperate return value type from
  function name.
o Unordered function prototypes.

21 years agoWhitespace fix only
julian [Wed, 2 Oct 2002 23:12:01 +0000 (23:12 +0000)]
Whitespace fix only

21 years agoFix rare and probably inconsequential memory leak.
phk [Wed, 2 Oct 2002 20:34:23 +0000 (20:34 +0000)]
Fix rare and probably inconsequential memory leak.

Spotted by: FlexeLint

21 years agoCatch up to SMTX -> SLOCK changes.
jhb [Wed, 2 Oct 2002 20:33:52 +0000 (20:33 +0000)]
Catch up to SMTX -> SLOCK changes.

21 years agoRename the mutex thread and process states to use a more generic 'LOCK'
jhb [Wed, 2 Oct 2002 20:31:47 +0000 (20:31 +0000)]
Rename the mutex thread and process states to use a more generic 'LOCK'
name instead.  (e.g., SLOCK instead of SMTX, TD_ON_LOCK() instead of
TD_ON_MUTEX())  Eventually a turnstile abstraction will be added that
will be shared with mutexes and other types of locks.  SLOCK/TDI_LOCK will
be used internally by the turnstile code and will not be specific to
mutexes.  Making the change now ensures that turnstiles can be dropped
in at a later date without affecting the ABI of userland applications.

21 years agoFix a probably rare and inconsequential memory leak.
phk [Wed, 2 Oct 2002 20:29:54 +0000 (20:29 +0000)]
Fix a probably rare and inconsequential memory leak.

Spotted by: FlexeLint

21 years agoHook rpc.rstatd back up to the build now that it compiles.
mike [Wed, 2 Oct 2002 18:37:11 +0000 (18:37 +0000)]
Hook rpc.rstatd back up to the build now that it compiles.

21 years agoRename local function havedisk() to haveadisk() to avoid conflict a
mike [Wed, 2 Oct 2002 18:29:50 +0000 (18:29 +0000)]
Rename local function havedisk() to haveadisk() to avoid conflict a
with another function by the same name in a system header.

21 years agoAccess td->td_kse inside sched_lock.
jmallett [Wed, 2 Oct 2002 18:25:09 +0000 (18:25 +0000)]
Access td->td_kse inside sched_lock.

Submitted by: julian

21 years agoAdd a man page for the KSE system calls.
archie [Wed, 2 Oct 2002 18:01:51 +0000 (18:01 +0000)]
Add a man page for the KSE system calls.

Reviewed by: julian, ru

21 years agoAdd 2 Ids for new ServerWorks host to PCI bridge chipset.
iwasaki [Wed, 2 Oct 2002 17:50:38 +0000 (17:50 +0000)]
Add 2 Ids for new ServerWorks host to PCI bridge chipset.
These are still unknown name but these are working as well
as the other ServerWorks chipset.
Description strings should be corrected when the chipsets
are known.

MFC after: 1 week

21 years ago-Wall'ify (mostly).
archie [Wed, 2 Oct 2002 17:07:39 +0000 (17:07 +0000)]
-Wall'ify (mostly).

21 years agoLet kse_wakeup() take a KSE mailbox pointer argument.
archie [Wed, 2 Oct 2002 16:48:16 +0000 (16:48 +0000)]
Let kse_wakeup() take a KSE mailbox pointer argument.

Reviewed by: julian

21 years agoDe-obfuscate local use of members of 'struct thread', for which we have
jmallett [Wed, 2 Oct 2002 16:39:39 +0000 (16:39 +0000)]
De-obfuscate local use of members of 'struct thread', for which we have
local variables, and group assignment.

21 years agoUnhook rpc.rstatd from the build until it actually compiles.
jhb [Wed, 2 Oct 2002 16:14:16 +0000 (16:14 +0000)]
Unhook rpc.rstatd from the build until it actually compiles.

21 years agoCode that has been #if0'd with the comment "WHY" since revision 1.1 can
jmallett [Wed, 2 Oct 2002 16:06:39 +0000 (16:06 +0000)]
Code that has been #if0'd with the comment "WHY" since revision 1.1 can
disappear.

21 years agoRemove 6-years-stale #if0, the behaviour isn't ever going to be limited to
jmallett [Wed, 2 Oct 2002 16:03:26 +0000 (16:03 +0000)]
Remove 6-years-stale #if0, the behaviour isn't ever going to be limited to
compatible mode, as far as I know, since we use it...

21 years agoMFi386: fork1 apparently takes 4 args now
gallatin [Wed, 2 Oct 2002 14:30:14 +0000 (14:30 +0000)]
MFi386: fork1 apparently takes 4 args now

Forgotten by: scottl

21 years agoAdd an example showing how to use wcstok(). Fix ordering of See Also section.
tjr [Wed, 2 Oct 2002 14:28:26 +0000 (14:28 +0000)]
Add an example showing how to use wcstok(). Fix ordering of See Also section.

21 years agoPC98 is not a true architecture, but a subarchitecture of i386, so we need
phk [Wed, 2 Oct 2002 13:52:36 +0000 (13:52 +0000)]
PC98 is not a true architecture, but a subarchitecture of i386, so we need
to use a slightly differnt command to PC98::buildworld.

21 years agoTeach ncal(1) about month names specified on the command line.
roam [Wed, 2 Oct 2002 12:31:10 +0000 (12:31 +0000)]
Teach ncal(1) about month names specified on the command line.

Reviewed by: -audit
Approved by: silence on -audit
MFC after: 3 weeks

21 years agoAdd the 'restrict' type qualifier to the prototypes of `sigaction',
robert [Wed, 2 Oct 2002 10:53:44 +0000 (10:53 +0000)]
Add the 'restrict' type qualifier to the prototypes of `sigaction',
`sigprocmask', `sigaltstack', and `sigwait' as well as to the
prototypes of the apparantly unimplemented functions `sigtimedwait'
and `sigwaitinfo'.  This complies with IEEE Std 1003.1-2001.

21 years agoUse %d in a printf() format string and cast the argument to int instead of
tjr [Wed, 2 Oct 2002 09:57:54 +0000 (09:57 +0000)]
Use %d in a printf() format string and cast the argument to int instead of
using %td when we know that the number is between 0 and 9; mksyntax is a
build tool and needs to work on -stable.

21 years agoSince bpf is no longer an optional component, remove associated ifdef's.
sobomax [Wed, 2 Oct 2002 09:38:17 +0000 (09:38 +0000)]
Since bpf is no longer an optional component, remove associated ifdef's.

Submitted by: don't quite remember - the name of the sender disappeared
with the rest of my inbox. :(

21 years agoAbsorb <sys/bus_private.h> into kern/subr_bus.c to prevent misunderstandings.
phk [Wed, 2 Oct 2002 09:34:29 +0000 (09:34 +0000)]
Absorb <sys/bus_private.h> into kern/subr_bus.c to prevent misunderstandings.

Suggested by: bde
Approved by: dfr

21 years agoMark two places where an unsigned number is checked "if (foo < 0)" with
phk [Wed, 2 Oct 2002 09:11:18 +0000 (09:11 +0000)]
Mark two places where an unsigned number is checked "if (foo < 0)" with
an XXX comment.

Somebody[TM] should look at this in some detail.

Spotted by: FlexeLint

21 years agoFix mis-indentation.
phk [Wed, 2 Oct 2002 09:09:25 +0000 (09:09 +0000)]
Fix mis-indentation.

Spotted by: FlexeLint

21 years ago#include the right thing for PC98
phk [Wed, 2 Oct 2002 08:59:28 +0000 (08:59 +0000)]
#include the right thing for PC98

Sponsored by: DARPA & NAI Labs.

21 years agoFurther improve --fast-read mode - when reading archive from stdin ensure that
sobomax [Wed, 2 Oct 2002 08:42:06 +0000 (08:42 +0000)]
Further improve --fast-read mode - when reading archive from stdin ensure that
SIGTERM we are sending to decompressor when there is nothing left to do is
actually delivered to decompressor, not to the another copy of tar, which
does unblocking.

MFC after: 2 weeks

21 years agostyle(9):
mike [Wed, 2 Oct 2002 07:55:33 +0000 (07:55 +0000)]
style(9):
o Align members of struct if_nameindex.
o Align and sort function prototypes.

21 years agoMove GEOM's sysctls under kern.geom.
phk [Wed, 2 Oct 2002 07:51:02 +0000 (07:51 +0000)]
Move GEOM's sysctls under kern.geom.

Sponsored by: DARPA & NAI Labs.

21 years agoAdd restrict type-qualifier.
mike [Wed, 2 Oct 2002 07:49:35 +0000 (07:49 +0000)]
Add restrict type-qualifier.

21 years agoPut some failing ioctl related printfs under a suitable debug flag.
phk [Wed, 2 Oct 2002 07:48:39 +0000 (07:48 +0000)]
Put some failing ioctl related printfs under a suitable debug flag.

Sponsored by: DARPA & NAI Labs.

21 years agoo Add typedef for size_t.
mike [Wed, 2 Oct 2002 07:48:13 +0000 (07:48 +0000)]
o Add typedef for size_t.
o Fix regoff_t typedef to not require <sys/types.h> as a prerequisite
  (specifically use `__off_t' instead of `off_t').
o Add restrict type-qualifier to function prototypes.

21 years agoUse standards visibility conditionals to conditionalize most of this
mike [Wed, 2 Oct 2002 07:44:50 +0000 (07:44 +0000)]
Use standards visibility conditionals to conditionalize most of this
header (details on how the visibility conditionals work are available
in <sys/cdefs.h>).  Use standard types instead of BSD specific ones,
so that this header compiles in the standards case (specifically this
means changing `u_int' to `unsigned int').

21 years agoSome kernel threads try to do significant work, and the default KSTACK_PAGES
scottl [Wed, 2 Oct 2002 07:44:29 +0000 (07:44 +0000)]
Some kernel threads try to do significant work, and the default KSTACK_PAGES
doesn't give them enough stack to do much before blowing away the pcb.
This adds MI and MD code to allow the allocation of an alternate kstack
who's size can be speficied when calling kthread_create.  Passing the
value 0 prevents the alternate kstack from being created.  Note that the
ia64 MD code is missing for now, and PowerPC was only partially written
due to the pmap.c being incomplete there.
Though this patch does not modify anything to make use of the alternate
kstack, acpi and usb are good candidates.

Reviewed by: jake, peter, jhb

21 years agoAdd LIBEXPAT definition.
phk [Wed, 2 Oct 2002 07:36:08 +0000 (07:36 +0000)]
Add LIBEXPAT definition.

21 years agoBuild libexpat out of src/contrib/expat.
phk [Wed, 2 Oct 2002 07:35:35 +0000 (07:35 +0000)]
Build libexpat out of src/contrib/expat.

Not connected to the build yet.

21 years agoThis commit was generated by cvs2svn to compensate for changes in r104349,
phk [Wed, 2 Oct 2002 07:16:04 +0000 (07:16 +0000)]
This commit was generated by cvs2svn to compensate for changes in r104349,
which included commits to RCS files with non-trunk default branches.

21 years agoVirgin import (trimmed) of eXpat version 1.95.5
phk [Wed, 2 Oct 2002 07:16:04 +0000 (07:16 +0000)]
Virgin import (trimmed) of eXpat version 1.95.5

21 years agoToo strict error checking in rev. 1.22 broke pwd_mkdb(8) in NIS
maxim [Wed, 2 Oct 2002 07:02:46 +0000 (07:02 +0000)]
Too strict error checking in rev. 1.22 broke pwd_mkdb(8) in NIS
environment. An empty UID and GID are valid there.

Spotted by: rwatson

21 years agoCast malloc() assignments.
mdodd [Wed, 2 Oct 2002 06:24:19 +0000 (06:24 +0000)]
Cast malloc() assignments.

21 years agosize_t is not a struct (fix mislabelling in a comment).
dd [Wed, 2 Oct 2002 05:15:34 +0000 (05:15 +0000)]
size_t is not a struct (fix mislabelling in a comment).

21 years agoUse more standard variable names for modevent functions.
mdodd [Wed, 2 Oct 2002 04:47:52 +0000 (04:47 +0000)]
Use more standard variable names for modevent functions.

21 years agoCreate/destroy memory zones on module load/unload, not on device
mdodd [Wed, 2 Oct 2002 04:42:24 +0000 (04:42 +0000)]
Create/destroy memory zones on module load/unload, not on device
attach/detach.

21 years agoInclude <sys/cdefs.h> so the visibility conditionals are available.
mike [Wed, 2 Oct 2002 04:22:34 +0000 (04:22 +0000)]
Include <sys/cdefs.h> so the visibility conditionals are available.
(This should have been included with the previous revision.)

21 years agoUse visibility conditionals. Only TCP_NODELAY ends up being defined
mike [Wed, 2 Oct 2002 04:19:47 +0000 (04:19 +0000)]
Use visibility conditionals.  Only TCP_NODELAY ends up being defined
in the standards case.

21 years agoo Update a visibility conditional.
mike [Wed, 2 Oct 2002 04:13:32 +0000 (04:13 +0000)]
o Update a visibility conditional.
o Adjust some comments in keeping with the header's local style.
o Change some typedefs to use types that don't require namespace
  pollution or deprecated types.
o Move some macros to the "does not belong in this header" section.

21 years agoRemoved bits related to MBRs now that they have a proper home in
bde [Wed, 2 Oct 2002 03:29:00 +0000 (03:29 +0000)]
Removed bits related to MBRs now that they have a proper home in
<sys/diskmbr.h>.

21 years agoAdd an -o option to exit after receiving one reply. This can be used
dd [Wed, 2 Oct 2002 03:24:27 +0000 (03:24 +0000)]
Add an -o option to exit after receiving one reply.  This can be used
to test whether a link is live.

PR: 38573
Submitted by: David Taylor <davidt@yadt.co.uk>
Obtained from: NetBSD

21 years agoAdd a new MAC entry point, mac_thread_userret(td), which permits policy
rwatson [Wed, 2 Oct 2002 02:42:38 +0000 (02:42 +0000)]
Add a new MAC entry point, mac_thread_userret(td), which permits policy
modules to perform MAC-related events when a thread returns to user
space.  This is required for policies that have floating process labels,
as it's not always possible to acquire the process lock at arbitrary
points in the stack during system call processing; process labels might
represent traditional authentication data, process history information,
or other data.

LOMAC will use this entry point to perform the process label update
prior to the thread returning to userspace, when plugged into the MAC
framework.

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

21 years agoDon't crash when the user feeds us nonesense in the form:
dd [Wed, 2 Oct 2002 02:17:59 +0000 (02:17 +0000)]
Don't crash when the user feeds us nonesense in the form:
ifconfig IF ether WHATEVER -alias

PR: 42544
Submitted by: Mike Makonnen <makonnen@pacbell.net>

21 years agoCorrect the exports(5) example to use -ro in the paragraph that talks
dd [Wed, 2 Oct 2002 02:07:08 +0000 (02:07 +0000)]
Correct the exports(5) example to use -ro in the paragraph that talks
about how exporting / and /usr works even if they're read-only.

PR: 42076
Submitted by: Marco Molteni <molter@tin.it>

21 years agoRemove options NO_MODULES=true. Lots of modules build, but may not load
jake [Wed, 2 Oct 2002 01:33:50 +0000 (01:33 +0000)]
Remove options NO_MODULES=true.  Lots of modules build, but may not load
or work properly.  MAC modules work at least :).

21 years agoIn a diskless setup, rc.d/initdiskless may overwrite parts of /etc on
dd [Wed, 2 Oct 2002 00:59:14 +0000 (00:59 +0000)]
In a diskless setup, rc.d/initdiskless may overwrite parts of /etc on
a per-machine or per-cluster (with different ways of expressing what's
part of a cluster) basis.  In order for this to be effective, rc.conf
has to be reread after initdiskless is finished.  Implement this by
adding a hook to etc/rc which rereads rc.conf by request.  This can
also be implemented by renaming initdiskless to initdiskless.sh and
sourcing rc.conf there manually, but it was decided that, that would
be uglier than a hook in etc/rc.

Developed in concert with: gordon

21 years agoRemove some #if 0'd code. After this, a "make unifdef" here produces
dd [Wed, 2 Oct 2002 00:37:01 +0000 (00:37 +0000)]
Remove some #if 0'd code.  After this, a "make unifdef" here produces
no changes (against crypto telnet).

Reviewed by: MD5

21 years agoMerge from crypto telnet: permit the argument to -s to be a hostname.
dd [Wed, 2 Oct 2002 00:29:34 +0000 (00:29 +0000)]
Merge from crypto telnet: permit the argument to -s to be a hostname.

21 years agoPermit the argument to the -s option to be a hostname. I see no
dd [Wed, 2 Oct 2002 00:27:14 +0000 (00:27 +0000)]
Permit the argument to the -s option to be a hostname.  I see no
reason to restrict this to a numeric address.

PR: 41841
Submitted by: Dmitry Pryanishnikov <dmitry@atlantis.dp.ua>,
Maxim Maximov <mcsi@agava.com>

21 years agoRemove the "special processes" section. It has rotted, and the idea
dd [Wed, 2 Oct 2002 00:09:24 +0000 (00:09 +0000)]
Remove the "special processes" section.  It has rotted, and the idea
ceased to be useful when the number of "special processes" went from 3
to one per device.  I considered replacing it with a "kernel threads"
section, but this seemed like the wrong place for that.

PR: 40969

21 years agoDon't claim to return the fileid which we unloaded. The kernel
dd [Tue, 1 Oct 2002 23:53:51 +0000 (23:53 +0000)]
Don't claim to return the fileid which we unloaded.  The kernel
doesn't do this, and it wouldn't be very useful if it did, since the
caller supplies us with that number.

PR: 41329
Submitted by: Michael Galassi <nerd@xyz.com>

21 years agoCorrect minor errors in the code examples.
dd [Tue, 1 Oct 2002 23:35:33 +0000 (23:35 +0000)]
Correct minor errors in the code examples.

PR: 41993
Submitted by: David Lay <dsl@foozle.org>

21 years agochoosen -> chosen
dd [Tue, 1 Oct 2002 23:15:32 +0000 (23:15 +0000)]
choosen -> chosen

PR: 42962
Submitted by: Scott D Friedemann <sdf@expertune.com>

21 years agoActually use the exitstatus value that we maintain.
dd [Tue, 1 Oct 2002 22:59:11 +0000 (22:59 +0000)]
Actually use the exitstatus value that we maintain.

PR: 43562
Submitted by: Marc Olzheim <marcolz@ilse.nl>

21 years agoFix a typo in BGE_BIGENDIAN_INIT.
jake [Tue, 1 Oct 2002 22:33:40 +0000 (22:33 +0000)]
Fix a typo in BGE_BIGENDIAN_INIT.