]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoIn gcc 3.1 Pentium/MMX now has its own -march=XXX option.
Maxim Sobolev [Thu, 13 Jun 2002 08:35:36 +0000 (08:35 +0000)]
In gcc 3.1 Pentium/MMX now has its own -march=XXX option.

22 years agoSupport the "--" end of options marker, useful to disambiguate between
Tim J. Robbins [Thu, 13 Jun 2002 08:07:20 +0000 (08:07 +0000)]
Support the "--" end of options marker, useful to disambiguate between
signal numbers and process groups (negative pid arguments), and
required by SUSv3.

22 years agoDon't list shell builtins when the hash command is used (SUSv3)
Tim J. Robbins [Thu, 13 Jun 2002 07:50:50 +0000 (07:50 +0000)]
Don't list shell builtins when the hash command is used (SUSv3)

22 years agoUse a cv rather than tsleep and friends to do the sleep/wakeup
Warner Losh [Thu, 13 Jun 2002 07:19:31 +0000 (07:19 +0000)]
Use a cv rather than tsleep and friends to do the sleep/wakeup
handshake between the ISR and the worker thread.  Move the mutex lock
so that it only protects the cv_wait.  This elimiates the not sleeping
with pccbb1 held messages some people were seeing.

Reviewed by: jhb (at least an early version)

22 years agoRemote pci.h/NPCI usage from i4b code.
Brooks Davis [Thu, 13 Jun 2002 06:04:28 +0000 (06:04 +0000)]
Remote pci.h/NPCI usage from i4b code.

Approved by: hm

22 years agoBah humbug. Fix another typo on the same comment line. It also had
Garance A Drosehn [Thu, 13 Jun 2002 02:37:50 +0000 (02:37 +0000)]
Bah humbug.  Fix another typo on the same comment line.  It also had
an option-space instead of a space...

MFC after: 10 days

22 years agoFix a typo in a comment from the previous commit. I had a bullet-character
Garance A Drosehn [Thu, 13 Jun 2002 02:32:42 +0000 (02:32 +0000)]
Fix a typo in a comment from the previous commit.  I had a bullet-character
instead of an asterisk.

Noticed by: keramida
MFC after: 10 days

22 years agoAlmost complete rewrite of the lpc commands 'abort', 'enable', 'disable',
Garance A Drosehn [Thu, 13 Jun 2002 01:55:48 +0000 (01:55 +0000)]
Almost complete rewrite of the lpc commands 'abort', 'enable', 'disable',
'restart', 'start', 'stop' and 'up'.  These are commands which mainly
just alter the access bits on the lock-file of a queue, and they all
now use a central routine to do that.  This reduces the amount of code
that is run as the priv userid, and eliminates a number of cases where
error messages were written while that priv uid was in effect.

As far as users are concerned, there should be no noticable difference
in the new versions.  In case there *is*, the previous implementations
are still there as 'xabort', 'xenable', etc, so they are available for
instant fallback.  If no one reports a problem after a few weeks, then
a later update will remove those x-commands.

Reviewed by: freebsd-audit and freebsd-print@bostonradio.org
MFC after: 10 days

22 years agoIf a RAD_FILTER_ID is supplied by the RADIUS server, treat it as an
Brian Somers [Wed, 12 Jun 2002 23:45:15 +0000 (23:45 +0000)]
If a RAD_FILTER_ID is supplied by the RADIUS server, treat it as an
additional label from ppp.linkup & ppp.linkdown to load.

Suggested and mostly submitted by: andrew pavlov <and@kremenchug.net>

22 years agoBump the version number to reflect the recent RADIUS commits
Brian Somers [Wed, 12 Jun 2002 23:00:12 +0000 (23:00 +0000)]
Bump the version number to reflect the recent RADIUS commits

22 years agoDon't forget to process the Ident field on the front of
Brian Somers [Wed, 12 Jun 2002 21:36:07 +0000 (21:36 +0000)]
Don't forget to process the Ident field on the front of
RAD_MICROSOFT_MS_CHAP_ERROR and RAD_MICROSOFT_MS_CHAP2_SUCCESS
messages, and remove the hack in chap.c to ignore that ident field
on the client side.

This anomoly was hacked around during development, and I forgot to
go back and fix it properly.

Spotted by: Sergey Korolew <ds@rt.balakovo.ru>

22 years agoAdd PCI bus enumeration and latency timer setup to the sparc64 MD PCI
Thomas Moestl [Wed, 12 Jun 2002 19:20:57 +0000 (19:20 +0000)]
Add PCI bus enumeration and latency timer setup to the sparc64 MD PCI
code. Both tasks are not always performed completely by the firmware.
The former is required to get some e450 models to boot; the latter fixes
the repeated fifo underruns with hme(4)s and gem(4)s observed on some
machines (and probably performance problems with other peripherals as
well).

22 years agoThe UDP head was unlocked too early in one unicast case.
Jeffrey Hsu [Wed, 12 Jun 2002 15:21:41 +0000 (15:21 +0000)]
The UDP head was unlocked too early in one unicast case.

Submitted by: bug reported by arr

22 years ago- Add sanity check for argument.
Mitsuru IWASAKI [Wed, 12 Jun 2002 14:55:19 +0000 (14:55 +0000)]
- Add sanity check for argument.
- Add new entry to stripdirs.  The `generate' directory appeared in
  acpica-unix-20020517.

22 years agoIf trap() is called when ddb is active, then go directly to trap_fatal();
Bruce Evans [Wed, 12 Jun 2002 13:30:52 +0000 (13:30 +0000)]
If trap() is called when ddb is active, then go directly to trap_fatal();
do not blunder around enabling interrupts and running trap handlers.
trap_pfault() will normally pass control to ddb's fault handler which
will normally do the right thing.

This bug is very old. but in old versions of FreeBSD it is probably only
serious for trap handling that involves sleeping.  In -current, attempting
to examine unmapped memory while stopped at a breakpoint at mi_switch()
was always fatal.

22 years agoReorder LDADD to fix static linking.
Dag-Erling Smørgrav [Wed, 12 Jun 2002 12:46:50 +0000 (12:46 +0000)]
Reorder LDADD to fix static linking.

22 years agoReorder libraries to fix static linking (libssl depends on libcrypto, so
Dag-Erling Smørgrav [Wed, 12 Jun 2002 12:45:24 +0000 (12:45 +0000)]
Reorder libraries to fix static linking (libssl depends on libcrypto, so
it must come before libcrypto in LDADD)

Reviewed by: bde

22 years ago o Acquire and release Giant in vm_map_unlock_and_wait().
Alan Cox [Wed, 12 Jun 2002 08:15:52 +0000 (08:15 +0000)]
 o Acquire and release Giant in vm_map_unlock_and_wait().

Submitted by: tegge

22 years agoAs a stop-gap measure, add one INP_LOCK_DESTROY() to in6_pcbdetach() to
Jeffrey Hsu [Wed, 12 Jun 2002 06:01:22 +0000 (06:01 +0000)]
As a stop-gap measure, add one INP_LOCK_DESTROY() to in6_pcbdetach() to
get kernel compiled with INET6 to boot.

22 years agode __P
Warner Losh [Wed, 12 Jun 2002 04:28:15 +0000 (04:28 +0000)]
de __P
ANSI functions
minor knf

22 years agoKNF.
Warner Losh [Wed, 12 Jun 2002 04:18:35 +0000 (04:18 +0000)]
KNF.
Use ANSI functions rather than old K&R style.

22 years agoString lengths and sizeof()s are size_t not int. Mark an unused parameter
Juli Mallett [Wed, 12 Jun 2002 04:18:31 +0000 (04:18 +0000)]
String lengths and sizeof()s are size_t not int.  Mark an unused parameter
of ReadMakefile as __unused, it's there because this function is used by
the abstracted list interface which normally deals with item handlers which
take two arguments.  Add a missing static prototype.

22 years agoKill __P.
Warner Losh [Wed, 12 Jun 2002 04:13:09 +0000 (04:13 +0000)]
Kill __P.

22 years agoThe error functions take constant pointers to strings for their format.
Juli Mallett [Wed, 12 Jun 2002 04:11:01 +0000 (04:11 +0000)]
The error functions take constant pointers to strings for their format.

22 years agoFix logic which resulted in missing a call to INP_UNLOCK().
Jeffrey Hsu [Wed, 12 Jun 2002 03:11:06 +0000 (03:11 +0000)]
Fix logic which resulted in missing a call to INP_UNLOCK().

22 years agoFix typo where INP_INFO_RLOCK should be INP_INFO_RUNLOCK.
Jeffrey Hsu [Wed, 12 Jun 2002 03:08:08 +0000 (03:08 +0000)]
Fix typo where INP_INFO_RLOCK should be INP_INFO_RUNLOCK.
  Submitted by: tegge, jlemon

Prefer LIST_FOREACH macro.
  Submitted by: jlemon

22 years agoMake nselcol, the number of select collisions since boot, unsigned as
Kelly Yancey [Wed, 12 Jun 2002 02:08:18 +0000 (02:08 +0000)]
Make nselcol, the number of select collisions since boot, unsigned as
negative collisions simply doesn't make sense.

PR: (one small part of) 19720
Approved by: alfred

22 years agoUnderstand the following Microsoft Vendor Specific RADIUS attributes:
Brian Somers [Wed, 12 Jun 2002 00:33:17 +0000 (00:33 +0000)]
Understand the following Microsoft Vendor Specific RADIUS attributes:

  RAD_MICROSOFT_MS_MPPE_ENCRYPTION_POLICY
  RAD_MICROSOFT_MS_MPPE_ENCRYPTION_TYPES
  RAD_MICROSOFT_MS_MPPE_RECV_KEY
  RAD_MICROSOFT_MS_MPPE_SEND_KEY

These attributes may be supplied by a RADIUS server when MSCHAPv2 is
used to authenticate.

It *should* now be possible to build ppp with -DNODES and still support
CHAP/MSCHAP/MSCHAPv2/MPPE via a RADIUS server, but the code isn't yet
smart enough to do that (building with -DNODES just looses these
facilities).

Sponsored by: Monzoon

22 years agoAdd the following functions:
Brian Somers [Wed, 12 Jun 2002 00:21:07 +0000 (00:21 +0000)]
Add the following functions:

  rad_request_authenticator()
    Returns the Request-Authenticator relevant to the most recently received
    RADIUS response.

  rad_server_secret()
    Returns the Shared Secret relevant to the most recently received
    RADIUS response.

Neither of these functions should be necessary, however, the
MS-MPPE-Recv-Key and MS-MPPE-Send-Key Microsoft Vendor Specific
attributes are supplied in a mangled (encrypted) format, requiring
this information to demangle.

It's not clear whether these functions should be replaced with a
rad_demangle() function or whether these attributes are one-offs.

Sponsored by: Monzoon

22 years agoAdd role="historic" attributes to all release notes that were MFC-ed
Bruce A. Mah [Tue, 11 Jun 2002 23:06:41 +0000 (23:06 +0000)]
Add role="historic" attributes to all release notes that were MFC-ed
as of 4.6-RELEASE (or earlier).  This commit shouldn't have any immediate
effect, but we'll eventually use some stylesheet magic to remove
the historic release notes from the output stream.  This will have
the effect of making the 5.0-RELEASE release notes contain material
only relevent to -CURRENT *or* recently MFC-ed to 4-STABLE.

The stylesheet fixes will follow later, once I work out a couple more
details.  I wanted to get this commit done now, before anything gets
MFC-ed in the post-4.6 world (so we don't need to go back later and
figure this out).

If a historic release note gets modified, it may very well be
appropriate to remove its historic attribute.

22 years agoNew release notes: GNU sort (textutils 2.0.21), NetBSD stat 20020605,
Bruce A. Mah [Tue, 11 Jun 2002 21:43:54 +0000 (21:43 +0000)]
New release notes:  GNU sort (textutils 2.0.21), NetBSD stat 20020605,
GNU tar 1.13.25.

Updated release notes:  IPFilter 3.4.28.

22 years agoThis actually compiles under lint now, by effectively #if 0'ing it
Alfred Perlstein [Tue, 11 Jun 2002 21:22:32 +0000 (21:22 +0000)]
This actually compiles under lint now, by effectively #if 0'ing it
when compiling LINT, linking LINT was broke, so unbreak by removing
the preprocessor directives.

22 years agosendmail 8.12.4 has been imported.
Gregory Neil Shapiro [Tue, 11 Jun 2002 21:18:42 +0000 (21:18 +0000)]
sendmail 8.12.4 has been imported.

22 years agosendmail 8.12.4 has been imported
Gregory Neil Shapiro [Tue, 11 Jun 2002 21:17:35 +0000 (21:17 +0000)]
sendmail 8.12.4 has been imported

22 years agoResolve conflicts from sendmail 8.12.4 import
Gregory Neil Shapiro [Tue, 11 Jun 2002 21:16:51 +0000 (21:16 +0000)]
Resolve conflicts from sendmail 8.12.4 import

22 years agocatch up with ktrace changes, KTRPOINT takes a 'struct thread' not
Alfred Perlstein [Tue, 11 Jun 2002 21:14:02 +0000 (21:14 +0000)]
catch up with ktrace changes, KTRPOINT takes a 'struct thread' not
'struct proc' now.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r98121,
Gregory Neil Shapiro [Tue, 11 Jun 2002 21:12:04 +0000 (21:12 +0000)]
This commit was generated by cvs2svn to compensate for changes in r98121,
which included commits to RCS files with non-trunk default branches.

22 years agoImport sendmail 8.12.4
Gregory Neil Shapiro [Tue, 11 Jun 2002 21:12:04 +0000 (21:12 +0000)]
Import sendmail 8.12.4

22 years agoTime counter stats are unsigned, advertise them to sysctl(8) that way.
Kelly Yancey [Tue, 11 Jun 2002 19:47:44 +0000 (19:47 +0000)]
Time counter stats are unsigned, advertise them to sysctl(8) that way.

PR: (one small part of) 19720
Approved by: phk

22 years ago o Properly handle a failure by vm_fault_wire() or vm_fault_user_wire()
Alan Cox [Tue, 11 Jun 2002 19:13:59 +0000 (19:13 +0000)]
 o Properly handle a failure by vm_fault_wire() or vm_fault_user_wire()
   in vm_map_wire().
 o Make two white-space changes in vm_map_wire().

Reviewed by: tegge

22 years agoSO_PRIVSTATE has been commented out for long enough now....
Garrett Wollman [Tue, 11 Jun 2002 18:23:11 +0000 (18:23 +0000)]
SO_PRIVSTATE has been commented out for long enough now....

22 years agoAdd a reference count to struct fetchconn so we don't prematurely close and
Dag-Erling Smørgrav [Tue, 11 Jun 2002 11:27:28 +0000 (11:27 +0000)]
Add a reference count to struct fetchconn so we don't prematurely close and
free a cached FTP connection.

22 years agoUse the constants from <limits.h> for the sizes of integral C types
Robert Drehmel [Tue, 11 Jun 2002 11:27:20 +0000 (11:27 +0000)]
Use the constants from <limits.h> for the sizes of integral C types
rather than defining them ourselves.

22 years agoRemember to initialize the control block head mutex.
Jeffrey Hsu [Tue, 11 Jun 2002 10:58:57 +0000 (10:58 +0000)]
Remember to initialize the control block head mutex.

22 years agoFix typo.
Jeffrey Hsu [Tue, 11 Jun 2002 10:56:49 +0000 (10:56 +0000)]
Fix typo.

Submitted by: Kyunghwan Kim <redjade@atropos.snu.ac.kr>

22 years ago#include <sys/disklabel.h> to get BBSIZE.
Poul-Henning Kamp [Tue, 11 Jun 2002 10:19:59 +0000 (10:19 +0000)]
#include <sys/disklabel.h> to get BBSIZE.

22 years agopaste(1) appeared at least as early as 32v.
Tim J. Robbins [Tue, 11 Jun 2002 08:13:03 +0000 (08:13 +0000)]
paste(1) appeared at least as early as 32v.

22 years agocut(1) appeared at least as early as System III.
Tim J. Robbins [Tue, 11 Jun 2002 08:11:12 +0000 (08:11 +0000)]
cut(1) appeared at least as early as System III.

22 years agoMake the alpha architecture use the common ufsread().
Poul-Henning Kamp [Tue, 11 Jun 2002 06:56:31 +0000 (06:56 +0000)]
Make the alpha architecture use the common ufsread().

Submitted by: ticso

22 years ago o Teach vm_map_delete() to respect the "in-transition" flag
Alan Cox [Tue, 11 Jun 2002 05:24:22 +0000 (05:24 +0000)]
 o Teach vm_map_delete() to respect the "in-transition" flag
   on a vm_map_entry by sleeping until the flag is cleared.

Submitted by: tegge

22 years agoEvery array elt is initialized in the following loop, so remove
Jeffrey Hsu [Mon, 10 Jun 2002 23:48:37 +0000 (23:48 +0000)]
Every array elt is initialized in the following loop, so remove
unnecessary M_ZERO.

22 years agoNote that this appeared at least as early as PWB UNIX.
Juli Mallett [Mon, 10 Jun 2002 22:59:59 +0000 (22:59 +0000)]
Note that this appeared at least as early as PWB UNIX.

Use the literal string 'PWB UNIX', as we still have no .At macro for it.

22 years agops(1) appeared in v4
Juli Mallett [Mon, 10 Jun 2002 22:48:51 +0000 (22:48 +0000)]
ps(1) appeared in v4

Submitted by: grog

22 years agoConvert hit and miss counters to unsigned values. Surely negative values
Kelly Yancey [Mon, 10 Jun 2002 22:40:26 +0000 (22:40 +0000)]
Convert hit and miss counters to unsigned values.  Surely negative values
for either does not make sense.

PR: (one small part of) 19720

22 years agoDillon's recent commits to the dynamic linker without running them
John Polstra [Mon, 10 Jun 2002 21:51:16 +0000 (21:51 +0000)]
Dillon's recent commits to the dynamic linker without running them
by me first have given me a good excuse to drop my MAINTAINERship.

MFC after: 1 week

22 years agoCorrect a bug in the last commit. The whole point of creating a 'done:'
Matthew Dillon [Mon, 10 Jun 2002 21:15:50 +0000 (21:15 +0000)]
Correct a bug in the last commit.  The whole point of creating a 'done:'
goto target was so the cache could be freed.  So free the cache after
done: rather then before done: (!)

Submitted by: Gavin Atkinson <gavin@ury.york.ac.uk>

22 years agoLock up inpcb.
Jeffrey Hsu [Mon, 10 Jun 2002 20:05:46 +0000 (20:05 +0000)]
Lock up inpcb.

Submitted by: Jennifer Yang <yangjihui@yahoo.com>

22 years ago- Whitespace only: use return statement consistentlt (return (foo), not
Maxim Sobolev [Mon, 10 Jun 2002 19:25:21 +0000 (19:25 +0000)]
- Whitespace only: use return statement consistentlt (return (foo), not
  return(foo)), kill extra blank names between function names;
- fix format string in printf(): devtoname() returns string, not pointer.

22 years agoIn tracking down an installation seg fault with then openoffice port
Matthew Dillon [Mon, 10 Jun 2002 18:52:31 +0000 (18:52 +0000)]
In tracking down an installation seg fault with then openoffice port
Martin Blapp determined that the elf dynamic loader was at fault.  In
particular, the loader uses alloca() to allocate a symbol cache on the
stack.  Normally this would work just fine, but if the loader is called
from a threaded program and the object being loaded is fairly large the
alloca() can blow away the thread stack and effect other nearby thread
stacks as well.  My testing showed that the symbol cache can be as large
as 250KBytes during the openoffice port build and install sequence.  Martin
was able to work around the problem by disabling the symbol cache
(cache = NULL;).  However, this solution is not adequate for commit because
it can cause an enormous cpu burden for applications which do a lot of
dynamic loading (e.g. like konqueror).

The solution is to use anonymous mmap() to temporarily allocate space to
hold the symbol cache.  In testing I found that replacing the alloca()
with mmap() has no observable degredation in performance.

It should be noted that this bug does not necessarily cause an immediate
crash but can instead result in long term corruption and instability in
applications that load modules from threads.  The bug is almost certainly
responsible for some of the instabilities found in konqueror, for example,
and possibly netscape too.

Sleuthing work by: Martin Blapp <mb@imp.ch>
X-MFC after: Before or after the 4.6 release depending on the release engineers

22 years agoPut geom_gpt.c under the GEOM option instead of having a special GEOM_GPT
Poul-Henning Kamp [Mon, 10 Jun 2002 18:49:41 +0000 (18:49 +0000)]
Put geom_gpt.c under the GEOM option instead of having a special GEOM_GPT
option for it.

22 years agoFix up the wording thoughout, and document locking.
Chad David [Mon, 10 Jun 2002 17:36:05 +0000 (17:36 +0000)]
Fix up the wording thoughout, and document locking.

22 years agoFix wording and spelling in the trapsignal() description.
Chad David [Mon, 10 Jun 2002 17:08:42 +0000 (17:08 +0000)]
Fix wording and spelling in the trapsignal() description.

22 years agocvs -j -j gave me two copies of TARGET_VERSION and I didn't notice
David E. O'Brien [Mon, 10 Jun 2002 15:48:23 +0000 (15:48 +0000)]
cvs -j -j gave me two copies of TARGET_VERSION and I didn't notice
it before committing.

22 years agoRemove -j from synopsis. The form given in the synopsis was incorrect,
Tim J. Robbins [Mon, 10 Jun 2002 10:39:33 +0000 (10:39 +0000)]
Remove -j from synopsis. The form given in the synopsis was incorrect,
and it should not have been there at all since it is a deprecated option.

22 years agoWhen stat(2) fails, put the name of the component it failed on in the
Tim J. Robbins [Mon, 10 Jun 2002 10:03:46 +0000 (10:03 +0000)]
When stat(2) fails, put the name of the component it failed on in the
warning message as well as the path argument.

22 years agoSort sections, use Ex macro in Diagnostics section, point out that the first
Tim J. Robbins [Mon, 10 Jun 2002 09:33:15 +0000 (09:33 +0000)]
Sort sections, use Ex macro in Diagnostics section, point out that the first
example only works with csh(1).

22 years agoAdd History section. sleep appeared in v4.
Tim J. Robbins [Mon, 10 Jun 2002 09:07:22 +0000 (09:07 +0000)]
Add History section. sleep appeared in v4.

22 years agoAdd another example for using paste(1) since tjr seemed to want more examples
Juli Mallett [Mon, 10 Jun 2002 09:03:49 +0000 (09:03 +0000)]
Add another example for using paste(1) since tjr seemed to want more examples
on IRC.  This one is to create a colon seperated list of directories from
find(1), suitable for use in the shell's PATH.

22 years agoAdd an EXAMPLES section with three examples. Document the problem with
Tim J. Robbins [Mon, 10 Jun 2002 08:50:56 +0000 (08:50 +0000)]
Add an EXAMPLES section with three examples. Document the problem with
multibyte characters and the -d option in a BUGS section.

22 years agoRefer to environ(7) for description of COLUMNS instead of describing it
Tim J. Robbins [Mon, 10 Jun 2002 07:55:13 +0000 (07:55 +0000)]
Refer to environ(7) for description of COLUMNS instead of describing it
in-line. Document effects of locale environment variables in the same way.

22 years agoThe 10 occurrence limit for the `s' command documented here has not existed
Tim J. Robbins [Mon, 10 Jun 2002 07:44:04 +0000 (07:44 +0000)]
The 10 occurrence limit for the `s' command documented here has not existed
for a very long time (or never did exist).

22 years agoNewlines are not escaped anymore.
Tim J. Robbins [Mon, 10 Jun 2002 07:27:32 +0000 (07:27 +0000)]
Newlines are not escaped anymore.

22 years agoDon't write escape newlines with `l' command (SUSv3)
Tim J. Robbins [Mon, 10 Jun 2002 07:25:35 +0000 (07:25 +0000)]
Don't write escape newlines with `l' command (SUSv3)

22 years agoGo back to taking the user's umask into account. Thanks to Alfred for
Doug Barton [Mon, 10 Jun 2002 07:16:42 +0000 (07:16 +0000)]
Go back to taking the user's umask into account. Thanks to Alfred for
hatching the idea of using dc, and Giorgos (keramida) for incubating it.

This also reverses most of the previous commit which took out or
modified the text about umask stuff.

22 years ago o In vm_map_entry_create(), call uma_zalloc() with M_NOWAIT on system maps.
Alan Cox [Mon, 10 Jun 2002 06:11:45 +0000 (06:11 +0000)]
 o In vm_map_entry_create(), call uma_zalloc() with M_NOWAIT on system maps.
   Submitted by: tegge
 o Eliminate the "!mapentzone" check from vm_map_entry_create() and
   vm_map_entry_dispose().  Reviewed by: tegge
 o Fix white-space usage in vm_map_entry_create().

22 years agoCorrectly handle global substitutions where the pattern is only "$", the
Tim J. Robbins [Mon, 10 Jun 2002 06:06:20 +0000 (06:06 +0000)]
Correctly handle global substitutions where the pattern is only "$", the
EOL anchor, when the last input line does not end in a newline character.
Picked up by the GNU sed test suite.

22 years agoDocument the cred_update_thread() function, and fix the RETURN VALUES.
Chad David [Mon, 10 Jun 2002 05:18:00 +0000 (05:18 +0000)]
Document the cred_update_thread() function, and fix the RETURN VALUES.

22 years agoNote early appearence of some commands. These actually appeared in PWB, but
Juli Mallett [Mon, 10 Jun 2002 05:03:20 +0000 (05:03 +0000)]
Note early appearence of some commands.  These actually appeared in PWB, but
it seems we don't have a macro for that yet, so list them in the first UNIX
release since then that we have a .At for: v7.

22 years agoChange our default XF86Config location from /etc/ to /etc/X11/,
David E. O'Brien [Mon, 10 Jun 2002 04:47:26 +0000 (04:47 +0000)]
Change our default XF86Config location from /etc/ to /etc/X11/,
following the lead of The XFree86 Project's default.

Approved by: Murray

22 years agoThe moduledata_t argument is not a pointer, and sysstem_sub_id should
Chad David [Mon, 10 Jun 2002 04:37:11 +0000 (04:37 +0000)]
The moduledata_t argument is not a pointer, and sysstem_sub_id should
be sysinit_sub_id.

PR: docs/34583
Approved by: murray

22 years agoCorrect the logic for determining whether the per-CPU locks need
Ian Dowse [Mon, 10 Jun 2002 03:25:23 +0000 (03:25 +0000)]
Correct the logic for determining whether the per-CPU locks need
to be destroyed. This fixes a problem where destroying a UMA zone
would fail to destroy all zone mutexes.

Reviewed by: jeff

22 years agoAdd signal.9 and friends.
Chad David [Mon, 10 Jun 2002 03:01:36 +0000 (03:01 +0000)]
Add signal.9 and friends.

22 years agoNew man page that documents many of the kernel related signal functions.
Chad David [Mon, 10 Jun 2002 03:00:24 +0000 (03:00 +0000)]
New man page that documents many of the kernel related signal functions.

Requested by: alfred

22 years agoCheck the run dependencies, not the build dependencies
Brian Somers [Mon, 10 Jun 2002 00:46:29 +0000 (00:46 +0000)]
Check the run dependencies, not the build dependencies

Spotted by: steve
Pointy hat: brian

22 years ago o Add vm_map_wire() for wiring contiguous regions of either kernel
Alan Cox [Sun, 9 Jun 2002 20:25:18 +0000 (20:25 +0000)]
 o Add vm_map_wire() for wiring contiguous regions of either kernel
   or user vm_maps.  This implementation has two key benefits when compared
   to vm_map_{user_,}pageable(): (1) it avoids a race condition through
   the use of "in-transition" vm_map entries and (2) it eliminates lock
   recursion on the vm_map.

Note: there is still an error case that requires clean up.

Reviewed by: tegge

22 years agoo Remove include of <sys/types.h>, it adds too much pollution;
Mike Barcroft [Sun, 9 Jun 2002 19:39:18 +0000 (19:39 +0000)]
o Remove include of <sys/types.h>, it adds too much pollution;
  instead, add typedefs of only gid_t, time_t, and uid_t.
o Remove an unneeded conditional that hid macros that are in the
  implementation namespace and therefore don't need to be hidden when
  _POSIX_SOURCE is defined.
o Adjust some conditionals for compatibility with X/Open and POSIX.
o Note missing functions getpwnam_r() and getpwuid_r().

22 years agoDon't depend on <pwd.h> to bogusly include <sys/types.h>.
Mike Barcroft [Sun, 9 Jun 2002 19:29:55 +0000 (19:29 +0000)]
Don't depend on <pwd.h> to bogusly include <sys/types.h>.

22 years agoFix driver to re-enable sound output on AD1816 based cards caused by an
Hellmuth Michaelis [Sun, 9 Jun 2002 14:20:17 +0000 (14:20 +0000)]
Fix driver to re-enable sound output on AD1816 based cards caused by an
obviously bogous return value of ad1816chan_setformat().
PR:             37932
Submitted by:   Martin Kaeske <Martin.Kaeske@Stud.TU-Ilmenau.DE>
Reviewed by:    hm
MFC after:      10 days

22 years agoCatch up with kernel.
Poul-Henning Kamp [Sun, 9 Jun 2002 10:59:04 +0000 (10:59 +0000)]
Catch up with kernel.

22 years agoImprove some on the naming.
Poul-Henning Kamp [Sun, 9 Jun 2002 10:57:34 +0000 (10:57 +0000)]
Improve some on the naming.

Submitted by: iedowse

22 years agoStamp out Danglish: Spelling, grammer and other nitpicking.
Poul-Henning Kamp [Sun, 9 Jun 2002 10:25:51 +0000 (10:25 +0000)]
Stamp out Danglish: Spelling, grammer and other nitpicking.

Submitted by: "Steven G. Kargl" <kargl@troutmask.apl.washington.edu>

22 years agoPer previous discussion, and with Mark's blessing, update the value
Doug Barton [Sun, 9 Jun 2002 09:28:02 +0000 (09:28 +0000)]
Per previous discussion, and with Mark's blessing, update the value
of this knob to reflect (-)current reality.

22 years agoif you have taken the mbuf out of the message object, then if you pass
Julian Elischer [Sun, 9 Jun 2002 07:28:35 +0000 (07:28 +0000)]
if you have taken the mbuf out of the message object, then if you pass
the object to someone else, you need to put the mbuf back into it first..

22 years agoUse socklen_t for the length of a socket structure instead of `int'.
Giorgos Keramidas [Sun, 9 Jun 2002 04:18:45 +0000 (04:18 +0000)]
Use socklen_t for the length of a socket structure instead of `int'.

Reviewed by: mike

22 years agoReplace <strings.h> with <string.h>. No functions from the former are
Giorgos Keramidas [Sun, 9 Jun 2002 04:15:40 +0000 (04:15 +0000)]
Replace <strings.h> with <string.h>.  No functions from the former are
used in this file, and strlen() needs to be prototyped by the latter,
for this to compile without warnings.

Reviewed by: mike

22 years agoFix bug which has been there since rev 1.1 where && was used instead of &.
Jeffrey Hsu [Sun, 9 Jun 2002 03:57:34 +0000 (03:57 +0000)]
Fix bug which has been there since rev 1.1 where && was used instead of &.

22 years agoRenamed the idempotency identifier to match the file name. Cleaned up
Bruce Evans [Sun, 9 Jun 2002 02:52:40 +0000 (02:52 +0000)]
Renamed the idempotency identifier to match the file name.  Cleaned up
indentation and comments.

22 years agoANSIfy a few prototypes, thus fixing a few warnings.
Giorgos Keramidas [Sun, 9 Jun 2002 02:20:58 +0000 (02:20 +0000)]
ANSIfy a few prototypes, thus fixing a few warnings.

Reviewed by: grog

22 years agoCheck the return value of getcwd() to avoid printf()ing a NULL. Mark
Mike Barcroft [Sun, 9 Jun 2002 00:46:24 +0000 (00:46 +0000)]
Check the return value of getcwd() to avoid printf()ing a NULL.  Mark
usage() as __dead2 to avoid a GCC warning.

Spotted by: keramida

22 years agoAllow one to profile FORTRAN77 programs.
David E. O'Brien [Sun, 9 Jun 2002 00:03:56 +0000 (00:03 +0000)]
Allow one to profile FORTRAN77 programs.