]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoAdd a 'reset' command. This is useful for panics really early before
Peter Wemm [Sat, 3 Nov 2001 04:55:48 +0000 (04:55 +0000)]
Add a 'reset' command.  This is useful for panics really early before
any symbols are loaded.  Especially for unattended machines.

22 years agoDeclare struct uio so consumers don't have to #include <sys/uio.h> unless
Dag-Erling Smørgrav [Sat, 3 Nov 2001 03:17:16 +0000 (03:17 +0000)]
Declare struct uio so consumers don't have to #include <sys/uio.h> unless
they intend to call proc_rwmem().

22 years agoReduce the number of #include dependencies by declaring some of the structs
Dag-Erling Smørgrav [Sat, 3 Nov 2001 03:07:09 +0000 (03:07 +0000)]
Reduce the number of #include dependencies by declaring some of the structs
used in pseudofs.h as opaque structs.

22 years agoallocrqg: Don't ever try to write to a NULL pointer.
Greg Lehey [Sat, 3 Nov 2001 02:39:15 +0000 (02:39 +0000)]
allocrqg: Don't ever try to write to a NULL pointer.

Reported by: Jeff Roberson <jeff@midstream.com>

22 years agoTighten up the random seeding a little bit more. getpid() alone only
David E. O'Brien [Sat, 3 Nov 2001 02:14:10 +0000 (02:14 +0000)]
Tighten up the random seeding a little bit more.  getpid() alone only
affect only lower bits of seed which would resulte in the same seed for
sequences of fast-started awk's resulting the same random sequence.

Submitted by: ache

22 years agoRevert rev 1.3 which moved us away from POSIX character classes.
David E. O'Brien [Sat, 3 Nov 2001 01:35:07 +0000 (01:35 +0000)]
Revert rev 1.3 which moved us away from POSIX character classes.
The community feels our base AWK must handle them.

22 years agoRevert rev 1.7 which moved us away from POSIX character classes.
David E. O'Brien [Sat, 3 Nov 2001 01:33:12 +0000 (01:33 +0000)]
Revert rev 1.7 which moved us away from POSIX character classes.
The community feels our base AWK must handle them.

22 years agoRemove a debugging warnx() that got committed by mistake.
Dag-Erling Smørgrav [Sat, 3 Nov 2001 01:31:46 +0000 (01:31 +0000)]
Remove a debugging warnx() that got committed by mistake.

22 years agoImplement i386/i386/pmap.c 1.292 for alpha, ia64 (avoid free
Matthew Dillon [Sat, 3 Nov 2001 01:08:55 +0000 (01:08 +0000)]
Implement i386/i386/pmap.c 1.292 for alpha, ia64 (avoid free
page exhaustion / kernel panic for certain madvise() scenarios)

22 years agoWe have a _SIG_VALID() macro, so use it instead of duplicating the test all
Dag-Erling Smørgrav [Fri, 2 Nov 2001 23:50:00 +0000 (23:50 +0000)]
We have a _SIG_VALID() macro, so use it instead of duplicating the test all
over the place.  Also replace a printf() + panic() with a KASSERT().

Reviewed by: jhb

22 years agoBring Gawk back. There just isn't any other POSIX compliant AWK out there.
David E. O'Brien [Fri, 2 Nov 2001 23:48:32 +0000 (23:48 +0000)]
Bring Gawk back. There just isn't any other POSIX compliant AWK out there.
The biggest thing missing from Bell-Labs AWK is the character class regexes.

22 years agoUpdate to Gawk 3.1.0.
David E. O'Brien [Fri, 2 Nov 2001 23:46:57 +0000 (23:46 +0000)]
Update to Gawk 3.1.0.

22 years agoBring Gawk back. There just isn't any other POSIX compliant AWK out there.
David E. O'Brien [Fri, 2 Nov 2001 23:43:07 +0000 (23:43 +0000)]
Bring Gawk back. There just isn't any other POSIX compliant AWK out there.
The biggest thing missing from Bell-Labs AWK is the character class regexes.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r85919,
David E. O'Brien [Fri, 2 Nov 2001 23:41:29 +0000 (23:41 +0000)]
This commit was generated by cvs2svn to compensate for changes in r85919,
which included commits to RCS files with non-trunk default branches.

22 years agoI'd put down $10 says the author did a last minute tweak to awkgram.y (which
David E. O'Brien [Fri, 2 Nov 2001 23:41:29 +0000 (23:41 +0000)]
I'd put down $10 says the author did a last minute tweak to awkgram.y (which
had an error in it), and applied it by hand to awkgram.c (getting it right)
rather than really generating an new awkgram.c properly using Bison/YACC...
Fix his mistake here.

22 years agoAdd the as-yet-unused S_ALLSTOPS which I forgot in previous commit.
Dag-Erling Smørgrav [Fri, 2 Nov 2001 23:15:13 +0000 (23:15 +0000)]
Add the as-yet-unused S_ALLSTOPS which I forgot in previous commit.

22 years agoFix a bug I introduced yesterday. People who built world since the
Crist J. Clark [Fri, 2 Nov 2001 22:00:17 +0000 (22:00 +0000)]
Fix a bug I introduced yesterday. People who built world since the
previous commit yesterday may wish to check /var/run for junk.

22 years agoWe use the stock version of this file now, but since someone foolishly
David E. O'Brien [Fri, 2 Nov 2001 21:57:00 +0000 (21:57 +0000)]
We use the stock version of this file now, but since someone foolishly
took about 15 files off the vendor branch for what turned out to be
good reason a Gawk update takes an order of magnitude more effort than
it should...

22 years agoArrgh. A clean pc98 build failed due to bogons on my part :-(.
Warner Losh [Fri, 2 Nov 2001 21:50:15 +0000 (21:50 +0000)]
Arrgh.  A clean pc98 build failed due to bogons on my part :-(.

Fix it by putting back the link of machine to sys/i386/include rather
than ../../include (aka sys/pc98/include).  I had a stale machine link
on my first test.

Not sure what the "right" fix is, but this unbreaks things.

22 years agoWe use the stock 3.1.0 file.
David E. O'Brien [Fri, 2 Nov 2001 21:44:42 +0000 (21:44 +0000)]
We use the stock 3.1.0 file.

22 years agoThe bug rev 1.6 fixed was fixed in the vendor sources.
David E. O'Brien [Fri, 2 Nov 2001 21:40:02 +0000 (21:40 +0000)]
The bug rev 1.6 fixed was fixed in the vendor sources.

22 years agoWe use the stock 3.1.0 file now.
David E. O'Brien [Fri, 2 Nov 2001 21:39:18 +0000 (21:39 +0000)]
We use the stock 3.1.0 file now.

22 years agoWe use the stock 3.1.0 file (other than needing $FreeBSD$).
David E. O'Brien [Fri, 2 Nov 2001 21:37:05 +0000 (21:37 +0000)]
We use the stock 3.1.0 file (other than needing $FreeBSD$).

22 years agoFactor the common parts of the Makefile.foo files. This introduces two
Warner Losh [Fri, 2 Nov 2001 21:34:20 +0000 (21:34 +0000)]
Factor the common parts of the Makefile.foo files.  This introduces two
new files: kern.pre.mk, which contains most of the definitions, and
kern.post.mk, which contains most of the rules.

I've tested this on i386 and pc98.  I have had feedback on the sparc64
port, but no reports from anybody on alpha, ia64 or powerpc.  I
appologize in advance if I've broken you.

Reviewed by: jake, jhb, arch@

22 years agoMerge revision 1.2 (unspam l10n ranges check) into Gawk 3.1.0.
David E. O'Brien [Fri, 2 Nov 2001 21:34:12 +0000 (21:34 +0000)]
Merge revision 1.2 (unspam l10n ranges check) into Gawk 3.1.0.

22 years agoMerge revs 1.9 (don't choke on long lines),
David E. O'Brien [Fri, 2 Nov 2001 21:26:14 +0000 (21:26 +0000)]
Merge revs 1.9 (don't choke on long lines),
rev 1.3 (use srandom(time() ^ getpid())) into Gawk 3.1.0.

22 years agoBetter error messages for the cases where device_add_child fails. We
Warner Losh [Fri, 2 Nov 2001 21:26:07 +0000 (21:26 +0000)]
Better error messages for the cases where device_add_child fails.  We
should also whine if the old pccardd is used, but that's a little
harder than it sounds.

This also has the effect of fixing a typo that was in the last
version I committed.

22 years agoWell spell it "gnuregex.h".
David E. O'Brien [Fri, 2 Nov 2001 21:20:18 +0000 (21:20 +0000)]
Well spell it "gnuregex.h".

22 years agoo Remove (struct proc *p = td->td_proc) indirection in ipcperm(),
Robert Watson [Fri, 2 Nov 2001 21:20:05 +0000 (21:20 +0000)]
o Remove (struct proc *p = td->td_proc) indirection in ipcperm(),
  as suser_td(td) works as well as suser_xxx(NULL, p->p_ucred, 0);
  This simplifies upcoming changes to suser(), and causes this code
  to use the right credential (well, largely) once the td->td_ucred
  changes are complete.  There remains some redundancy and oddness
  in this code, which should be rethought after the next batch of
  suser and credential changes.

22 years agoBack out the -w, option strict and our($...). They don't work for me and
Warner Losh [Fri, 2 Nov 2001 21:14:17 +0000 (21:14 +0000)]
Back out the -w, option strict and our($...).  They don't work for me and
have broken the kernel build.

22 years agoo Remove the local temporary variable "struct proc *p" from vfs_mount()
Robert Watson [Fri, 2 Nov 2001 21:11:41 +0000 (21:11 +0000)]
o Remove the local temporary variable "struct proc *p" from vfs_mount()
  in vfs_syscalls.c.  Although it did save some indirection, many of
  those savings will be obscured with the impending commit of suser()
  changes, and the result is increased code complexity.  Also, once
  p->p_ucred and td->td_ucred are distinguished, this will make
  vfs_mount() use the correct thread credential, rather than the
  process credential.

22 years agoRemove an extra " that crept into a string.
Mike Silbersack [Fri, 2 Nov 2001 21:11:36 +0000 (21:11 +0000)]
Remove an extra " that crept into a string.

22 years agoUpdate vendor branch to gawk-3.1.0.
David E. O'Brien [Fri, 2 Nov 2001 21:06:08 +0000 (21:06 +0000)]
Update vendor branch to gawk-3.1.0.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r85898,
David E. O'Brien [Fri, 2 Nov 2001 21:06:08 +0000 (21:06 +0000)]
This commit was generated by cvs2svn to compensate for changes in r85898,
which included commits to RCS files with non-trunk default branches.

22 years agoArgh!
Poul-Henning Kamp [Fri, 2 Nov 2001 19:12:06 +0000 (19:12 +0000)]
Argh!

patch added the nmount at the bottom first time around.

Take 3!

22 years agoo Introduce group subset test, which limits the ability of a process to
Robert Watson [Fri, 2 Nov 2001 18:44:50 +0000 (18:44 +0000)]
o Introduce group subset test, which limits the ability of a process to
  debug another process based on their respective {effective,additional,
  saved,real} gid's.  p1 is only permitted to debug p2 if its effective
  gids (egid + additional groups) are a strict superset of the gids of
  p2.  This implements properly the security test previously incorrectly
  implemented in kern_ktrace.c, and is consistent with the kernel
  security policy (although might be slightly confusing for those more
  familiar with the userland policy).
o Restructure p_candebug() logic so that various results are generated
  comparing uids, gids, credential changes, and then composed in a
  single check before testing for privilege.  These tests encapsulate
  the "BSD" inter-process debugging policy.  Other non-BSD checks remain
  seperate.  Additional comments are added.

Submitted by:   tmm, rwatson
Obtained from:  TrustedBSD Project
Reviewed by:    petef, tmm, rwatson

22 years agoAdd empty shell for nmount syscall (take 2!)
Poul-Henning Kamp [Fri, 2 Nov 2001 18:35:54 +0000 (18:35 +0000)]
Add empty shell for nmount syscall (take 2!)

22 years agoo Add new header <sys/stdint.h>.
Mike Barcroft [Fri, 2 Nov 2001 18:05:43 +0000 (18:05 +0000)]
o Add new header <sys/stdint.h>.
o Make <stdint.h> a symbolic link to <sys/stdint.h>.
o Move most of <sys/inttypes.h> into <sys/stdint.h>, as per C99.
o Remove <sys/inttypes.h>.
o Adjust includes in sys/types.h and boot/efi/include/ia64/efibind.h
  to reflect new location of integer types in <sys/stdint.h>.
o Remove previously symbolicly linked <inttypes.h>, instead create a
  new file.
o Add MD headers <machine/_inttypes.h> from NetBSD.
o Include <sys/stdint.h> in <inttypes.h>, as required by C99; and
  include <machine/_inttypes.h> in <inttypes.h>, to fill in the
  remaining requirements for <inttypes.h>.
o Add additional integer types in <machine/ansi.h> and
  <machine/limits.h> which are included via <sys/stdint.h>.

Partially obtain from: NetBSD
Tested on: alpha, i386
Discussed on: freebsd-standards@bostonradio.org
Reviewed by: bde, fenner, obrien, wollman

22 years agoAdd nmount() stub function and regenerate the syscall-glue which should
Poul-Henning Kamp [Fri, 2 Nov 2001 17:59:23 +0000 (17:59 +0000)]
Add nmount() stub function and regenerate the syscall-glue which should
not need to check in generated files.

22 years agoReserve 378 for the new mount syscall Maxime Henrion <mux@qualys.com>
Poul-Henning Kamp [Fri, 2 Nov 2001 17:58:26 +0000 (17:58 +0000)]
Reserve 378 for the new mount syscall Maxime Henrion <mux@qualys.com>
is working on.  (This is to get us more than 32 mountoptions).

22 years agoRather than just change the arguments to suser() change the function to
Mike Barcroft [Fri, 2 Nov 2001 17:42:03 +0000 (17:42 +0000)]
Rather than just change the arguments to suser() change the function to
suser_xxx() as well.

Pointy hat to: rwatson

22 years agoDon't hide the failure to allocate device behind boot verbose. It is
Warner Losh [Fri, 2 Nov 2001 17:33:06 +0000 (17:33 +0000)]
Don't hide the failure to allocate device behind boot verbose.  It is
still telling us of real problems so should remain until it stops
doing that.

Submitted by: OGAWA Takaya <t-ogawa@triaez.kaisei.org>

22 years agoPrint a warning when device_add_child returns NULL. This used to be
Warner Losh [Fri, 2 Nov 2001 17:31:01 +0000 (17:31 +0000)]
Print a warning when device_add_child returns NULL.  This used to be
impossible at this point, but now it apparently is.  Grump.

Submitted by: OGAWA Takaya <t-ogawa@triaez.kaisei.org>

22 years agoAlways set unit number to -1 unless some other unit is specified in
Warner Losh [Fri, 2 Nov 2001 17:28:43 +0000 (17:28 +0000)]
Always set unit number to -1 unless some other unit is specified in
the config file.  This fixes the breakage caused by the recent change
in the behavior of device_add_child for ata (which shows soren's
reservations were well founded).

Submitted by: OGAWA Takaya <t-ogawa@triaez.kaisei.org>

22 years ago + Fix another possible vn_close race, in the same fashion as r1.95.
Jonathan Lemon [Fri, 2 Nov 2001 17:04:32 +0000 (17:04 +0000)]
 + Fix another possible vn_close race, in the same fashion as r1.95.
 + Check that the cached vnode type != VBAD before calling devsw(),
   this can happen if the vnode has been revoked.

22 years agoo Add a comment to p_candebug() noting that the P_INEXEC check should
Robert Watson [Fri, 2 Nov 2001 16:41:06 +0000 (16:41 +0000)]
o Add a comment to p_candebug() noting that the P_INEXEC check should
  really be moved elsewhere: p_candebug() encapsulates the security
  policy decision, whereas the P_INEXEC check has to do with "correctness"
  regarding race conditions, rather than security policy.

  Example: even if no security protections were enforced (the "uids are
  advisory" model), removing P_INEXEC could result in incorrect operation
  due to races on credential evaluation and modification during execve().

Obtained from: TrustedBSD Project

22 years agoThe sound drivers live in sound/driver, not sound/drivers
Warner Losh [Fri, 2 Nov 2001 16:40:06 +0000 (16:40 +0000)]
The sound drivers live in sound/driver, not sound/drivers

submitted by: Alex Kapranoff <alex@kapran.bitmcnit.bryansk.su>

22 years agoMerge from POSIX.1e Capabilities development tree:
Robert Watson [Fri, 2 Nov 2001 15:22:32 +0000 (15:22 +0000)]
Merge from POSIX.1e Capabilities development tree:

o Reorder and synchronize #include's, including moving "opt_cap.h" to
  above system includes.
o Introduce #ifdef'd kern.security.capabilities sysctl tree, including
  kern.security.capabilities.enabled, which defaults to 0.

The rest of the file remains stubs for the time being.

Obtained from: TrustedBSD Project

22 years agoMerge from POSIX.1e Capabilities development tree:
Robert Watson [Fri, 2 Nov 2001 15:16:59 +0000 (15:16 +0000)]
Merge from POSIX.1e Capabilities development tree:

o POSIX.1e capabilities authorize overriding of VEXEC for VDIR based
  on CAP_DAC_READ_SEARCH, but of !VDIR based on CAP_DAC_EXECUTE.  Add
  appropriate conditionals to vaccess() to take that into account.
o Synchronization cap_check_xxx() -> cap_check() change.

Obtained from: TrustedBSD Project

22 years agoFixed sioreg.h path.
Yoshihiro Takahashi [Fri, 2 Nov 2001 15:15:09 +0000 (15:15 +0000)]
Fixed sioreg.h path.

22 years agoo Capabilities cap_check() interface revised to remove _xxx, so rename
Robert Watson [Fri, 2 Nov 2001 15:08:08 +0000 (15:08 +0000)]
o Capabilities cap_check() interface revised to remove _xxx, so rename
  in p_cansched().  Also, replace '0' with 'NULL' for the ucred * pointer.

Obtained from: TrustedBSD Project

22 years agoTypo correction: detailled -> detailed.
Giorgos Keramidas [Fri, 2 Nov 2001 14:59:51 +0000 (14:59 +0000)]
Typo correction: detailled -> detailed.

Approved by: nik

22 years agoo Since kern_acl.c uses #ifdef CAPABILITIES to control
Robert Watson [Fri, 2 Nov 2001 14:53:04 +0000 (14:53 +0000)]
o Since kern_acl.c uses #ifdef CAPABILITIES to control
  capability-specific semantics, #include "opt_cap.h".

Obtained from: TrustedBSD Project

22 years agoo Merge recent changes from the POSIX.1e capability tree:
Robert Watson [Fri, 2 Nov 2001 14:45:22 +0000 (14:45 +0000)]
o Merge recent changes from the POSIX.1e capability tree:
    - Introduce cap_check_td() to reflect suser_td()
    - Introduce non-portable cap_from_text2_np(), which converts a
      text-form capability to an internal form relative to an existing
      capability, and non-portable cap_to_text2_np(), which accepts
      a flag argument specifying more about the text output format.
    - Introduce CTT_* flags controlling the behavior of
      cap_to_text2_np().

Submitted by: tmm
Obtained from: TrustedBSD Project

22 years agoDocument support for VIA 82C596 ATA-33 and ATA-66.
Jeroen Ruigrok van der Werven [Fri, 2 Nov 2001 12:22:17 +0000 (12:22 +0000)]
Document support for VIA 82C596 ATA-33 and ATA-66.

22 years agophkmalloc->evilchecks++;
Poul-Henning Kamp [Fri, 2 Nov 2001 11:32:28 +0000 (11:32 +0000)]
phkmalloc->evilchecks++;

If zero bytes are allocated, return pointer to the middle of page-zero
(which is protected) so that the program will crash if it dereferences
this illgotten pointer.

Inspired & Urged by: Theo de Raadt <deraadt@cvs.openbsd.org>

22 years agoAdd notice about early use from malloc.c forbidding malloc use from
Poul-Henning Kamp [Fri, 2 Nov 2001 11:30:42 +0000 (11:30 +0000)]
Add notice about early use from malloc.c forbidding malloc use from
this function.

22 years agoCall ast() from exception_restore when we are restoring to user mode.
Doug Rabson [Fri, 2 Nov 2001 10:24:44 +0000 (10:24 +0000)]
Call ast() from exception_restore when we are restoring to user mode.

22 years agoUse static storage for the unwind state so that we can still get backtraces
Doug Rabson [Fri, 2 Nov 2001 10:04:22 +0000 (10:04 +0000)]
Use static storage for the unwind state so that we can still get backtraces
when the VM system is hosed.

22 years ago#ifdef KTRACE a variable to silence a warning.
Poul-Henning Kamp [Fri, 2 Nov 2001 09:55:01 +0000 (09:55 +0000)]
#ifdef KTRACE a variable to silence a warning.

Submitted by: Maxime "mux" Henrion <mux@qualys.com>

22 years agoExplicitely mention this driver supports the RealTek 8029.
Jeroen Ruigrok van der Werven [Fri, 2 Nov 2001 09:43:45 +0000 (09:43 +0000)]
Explicitely mention this driver supports the RealTek 8029.
This should lessen some questions.

22 years agoFix alignment of the chipsets to the largest one described.
Jeroen Ruigrok van der Werven [Fri, 2 Nov 2001 09:37:35 +0000 (09:37 +0000)]
Fix alignment of the chipsets to the largest one described.

22 years agoIn order to perform faster when doing "head -n", use a pair of
Alfred Perlstein [Fri, 2 Nov 2001 09:27:16 +0000 (09:27 +0000)]
In order to perform faster when doing "head -n", use a pair of
fgetln/fwrite instead of getc/putchar, this seems about five times
faster.

22 years agostyle(9) cleanup.
Poul-Henning Kamp [Fri, 2 Nov 2001 09:23:34 +0000 (09:23 +0000)]
style(9) cleanup.

Submitted by: j mckitrick <jcm@freebsd-uk.eu.org>
Reviewed by: phk, /sbin/md5

22 years agochange a global into a local, misc style fixes
Alfred Perlstein [Fri, 2 Nov 2001 09:21:11 +0000 (09:21 +0000)]
change a global into a local, misc style fixes

22 years agoTurn the symlinks around, instead of ad0s1 -> ad0s1c, make it ad0s1c -> ad0s1.
Poul-Henning Kamp [Fri, 2 Nov 2001 09:16:25 +0000 (09:16 +0000)]
Turn the symlinks around, instead of ad0s1 -> ad0s1c, make it ad0s1c -> ad0s1.

Requested by: peter

22 years agoRemember to actually free the pv_entry in pmap_remove_entry().
Doug Rabson [Fri, 2 Nov 2001 08:56:58 +0000 (08:56 +0000)]
Remember to actually free the pv_entry in pmap_remove_entry().

22 years agoEmit a warning if the mbuf or mbuf cluster allocation failed.
Jonathan Lemon [Fri, 2 Nov 2001 05:10:40 +0000 (05:10 +0000)]
Emit a warning if the mbuf or mbuf cluster allocation failed.

Requested by: Mike Barcroft

22 years agoCosmetic: Don't output a blank line if a driver provides
Yaroslav Tykhiy [Fri, 2 Nov 2001 02:20:21 +0000 (02:20 +0000)]
Cosmetic: Don't output a blank line if a driver provides
no media status.

PR: bin/30587
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
MFC after: 1 week

22 years agoargh! cut/paste typo. :-(
Peter Wemm [Fri, 2 Nov 2001 01:45:11 +0000 (01:45 +0000)]
argh! cut/paste typo. :-(
(committed on a different machine to what I was testing it on)

22 years ago"Fix" a problem that got copied from alpha to ia64 and broke there.
Peter Wemm [Fri, 2 Nov 2001 00:41:00 +0000 (00:41 +0000)]
"Fix" a problem that got copied from alpha to ia64 and broke there.
When we truncate the msgbuf size because the last chunk is too small,
correctly terminate the phys_avail[] array - the VM system tests
the *end* for zero, not the start.  This leads the VM startup to
attempt to recreate a duplicate set of pages for all physical memory.

XXX the msgbuf handling is suspiciously different on i386 vs
alpha/ia64...

22 years agoDocument the /var/run/named directory.
Crist J. Clark [Thu, 1 Nov 2001 23:55:46 +0000 (23:55 +0000)]
Document the /var/run/named directory.

Suggesteded by: brian

22 years agoAdd a directory in /var/run to store ppp(8) command sockets.
Crist J. Clark [Thu, 1 Nov 2001 23:44:12 +0000 (23:44 +0000)]
Add a directory in /var/run to store ppp(8) command sockets.

PR: bin/29966
Approved by: brian
MFC after: 4

22 years agoMFS: Stop reaping children. It makes evil things happen to the extraction
Jordan K. Hubbard [Thu, 1 Nov 2001 23:32:46 +0000 (23:32 +0000)]
MFS: Stop reaping children.  It makes evil things happen to the extraction
loop.

22 years agoOops, deleted wrong BIF acquisition timeout invocation by mistake
Mitsuru IWASAKI [Thu, 1 Nov 2001 22:20:41 +0000 (22:20 +0000)]
Oops, deleted wrong BIF acquisition timeout invocation by mistake
in my previous commit.
acpi_cmbat_get_bif() from taskqueue calls acpi_cmbat_timeout() so it's
better for startup.

22 years agoo Update copyright dates.
Robert Watson [Thu, 1 Nov 2001 21:37:07 +0000 (21:37 +0000)]
o Update copyright dates.
o Add reference to TrustedBSD Project in license header.
o Update dated comments, including comment in extattr.h claiming that
  no file systems support extended attributes.
o Improve comment consistency.

22 years agoo Move suser() calls in kern/ to using suser_xxx() with an explicit
Robert Watson [Thu, 1 Nov 2001 20:56:57 +0000 (20:56 +0000)]
o Move suser() calls in kern/ to using suser_xxx() with an explicit
  credential selection, rather than reference via a thread or process
  pointer.  This is part of a gradual migration to suser() accepting
  a struct ucred instead of a struct proc, simplifying the reference
  and locking semantics of suser().

Obtained from: TrustedBSD Project

22 years agoo Permit osf1-emulated programs to modify uid/gid under jail by
Robert Watson [Thu, 1 Nov 2001 20:16:03 +0000 (20:16 +0000)]
o Permit osf1-emulated programs to modify uid/gid under jail by
  switching suser() to suser_xxx() and adding PRISON_ROOT flag.

22 years agoSubstitute "[:space:]" with the character constants it expands to.
David E. O'Brien [Thu, 1 Nov 2001 19:14:16 +0000 (19:14 +0000)]
Substitute "[:space:]" with the character constants it expands to.
This was a silent 'failure' when using Bell-Labs awk.

Submitted by: David Wolfskill <david@catwhisker.org>

22 years agoAlpha now has a TGA-capable console driver.
Wilko Bulte [Thu, 1 Nov 2001 19:07:33 +0000 (19:07 +0000)]
Alpha now has a TGA-capable console driver.

22 years agoTGA based display adapters are now supported.
Wilko Bulte [Thu, 1 Nov 2001 19:01:11 +0000 (19:01 +0000)]
TGA based display adapters are now supported.

22 years agoAlways compile in bridge hooks, or this code will not work (efficiently) with
Luigi Rizzo [Thu, 1 Nov 2001 17:10:52 +0000 (17:10 +0000)]
Always compile in bridge hooks, or this code will not work (efficiently) with
the loadable bridge module.

22 years agoAdd entries for "dumb" and "dialup" terminals, as the shell in HEAD
Luigi Rizzo [Thu, 1 Nov 2001 16:52:34 +0000 (16:52 +0000)]
Add entries for "dumb" and "dialup" terminals, as the shell in HEAD
refuses to start if it does not find a matching entry for the terminal
type.

My impression is that this is a problem in the shell, because
at least on PicoBSD images, the shell itself coredumps.
Anyways, this is the quickest fix i can come up with.

22 years agoSome fix for the recent apm module changes.
Mitsuru IWASAKI [Thu, 1 Nov 2001 16:34:07 +0000 (16:34 +0000)]
Some fix for the recent apm module changes.
 - Now that apm loadable module can inform its existence to other kernel
   components  (e.g. i386/isa/clock.c:startrtclock()'s TCS hack).
 - Exchange priority of SI_SUB_CPU and SI_SUB_KLD for above purpose.
 - Add simple arbitration mechanism for APM vs. ACPI.  This prevents
   the kernel enables both of them.
 - Remove obsolete `#ifdef DEV_APM' related code.
 - Add abstracted interface for Powermanagement operations.  Public apm(4)
   functions, such as apm_suspend(), should be replaced new interfaces.
   Currently only power_pm_suspend (successor of apm_suspend) is implemented.

Reviewed by: peter, arch@ and audit@

22 years agoProvide a new picobsd option, "--modules", to also build modules
Luigi Rizzo [Thu, 1 Nov 2001 16:17:37 +0000 (16:17 +0000)]
Provide a new picobsd option, "--modules", to also build modules
(modload and friends are sufficiently small to consider their inclusion
in PicoBSD images).

22 years agoTidy up the variable declarations and switch on warnings and strict.
Josef Karthauser [Thu, 1 Nov 2001 12:46:08 +0000 (12:46 +0000)]
Tidy up the variable declarations and switch on warnings and strict.

Reviewed by: diffing the generated files from before and after the change.

22 years agoModify the way host.conf and nsswitch.conf are treated at boot time:
Dag-Erling Smørgrav [Thu, 1 Nov 2001 12:39:01 +0000 (12:39 +0000)]
Modify the way host.conf and nsswitch.conf are treated at boot time:

 - if nsswitch.conf exists, host.conf is auto-generated for compatibility
   with legacy applications and libraries.

 - if host.conf exists but nsswitch.conf does not, nsswitch.conf is auto-
   generated as usual.

22 years agoJust typo corrections (mmunmap->munmap etc), no content changes.
Jens Schweikhardt [Thu, 1 Nov 2001 11:36:04 +0000 (11:36 +0000)]
Just typo corrections (mmunmap->munmap etc), no content changes.

MFC after: 1 week

22 years agoUpdate to the latest vendor's file.
David E. O'Brien [Thu, 1 Nov 2001 10:07:55 +0000 (10:07 +0000)]
Update to the latest vendor's file.

22 years agoBring the binutils_2_11_cvs_20011031 version of this to the HEAD branch.
David E. O'Brien [Thu, 1 Nov 2001 10:07:26 +0000 (10:07 +0000)]
Bring the binutils_2_11_cvs_20011031 version of this to the HEAD branch.

22 years agoThis merges in the changes from the FSF 19-July-2001 snapshot to the
David E. O'Brien [Thu, 1 Nov 2001 10:06:32 +0000 (10:06 +0000)]
This merges in the changes from the FSF 19-July-2001 snapshot to the
2001 Halloween snapshot.

22 years agoThis merges in the changes from the FSF anoncvs 19-July-2001 snapshot to
David E. O'Brien [Thu, 1 Nov 2001 10:03:10 +0000 (10:03 +0000)]
This merges in the changes from the FSF anoncvs 19-July-2001 snapshot to
the 2001 Halloween snapshot.

22 years agoUpdate to the latest vendor's file.
David E. O'Brien [Thu, 1 Nov 2001 09:55:05 +0000 (09:55 +0000)]
Update to the latest vendor's file.

22 years agoUpdate to the latest vendor file.
David E. O'Brien [Thu, 1 Nov 2001 09:54:43 +0000 (09:54 +0000)]
Update to the latest vendor file.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r85815,
David E. O'Brien [Thu, 1 Nov 2001 09:24:29 +0000 (09:24 +0000)]
This commit was generated by cvs2svn to compensate for changes in r85815,
which included commits to RCS files with non-trunk default branches.

22 years agoHalloween import of Binutils from the FSF 2.11 branch.
David E. O'Brien [Thu, 1 Nov 2001 09:24:29 +0000 (09:24 +0000)]
Halloween import of Binutils from the FSF 2.11 branch.
This fixes some IA-64 related bugs.

22 years agoFix a typo in a format string, and fix error checking for missing
Luigi Rizzo [Thu, 1 Nov 2001 08:45:02 +0000 (08:45 +0000)]
Fix a typo in a format string, and fix error checking for missing
masks in "limit" rules.

22 years agoRemove an extra splimp() call.
Luigi Rizzo [Thu, 1 Nov 2001 08:30:38 +0000 (08:30 +0000)]
Remove an extra splimp() call.

Spotted-by: diff(1)
22 years agoAdd the TGA video driver. This is a great accomplishtment and will help
David E. O'Brien [Thu, 1 Nov 2001 08:26:30 +0000 (08:26 +0000)]
Add the TGA video driver.  This is a great accomplishtment and will help
us a lot on older Alphas.
Andrew Gallatin, Thomas V. Crimi, and Peter Jeremy contributed to this
work along with the submitter.

Submitted by:  Andrew M. Miklic <miklic@home.com>

22 years agoRepocopied from sys/dev/syscons/scvgarndr.c
David E. O'Brien [Thu, 1 Nov 2001 08:22:49 +0000 (08:22 +0000)]
Repocopied from sys/dev/syscons/scvgarndr.c