]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoFixed some style bugs (mainly misalignment of backslashes).
Bruce Evans [Wed, 19 May 2004 16:04:26 +0000 (16:04 +0000)]
Fixed some style bugs (mainly misalignment of backslashes).

20 years agoMoved most of the "MI" definitions and declarations from <machine/profile.h>
Bruce Evans [Wed, 19 May 2004 15:41:26 +0000 (15:41 +0000)]
Moved most of the "MI" definitions and declarations from <machine/profile.h>
to <sys/gmon.h>.  Cleaned them up a little by not attempting to ifdef
for incomplete and out of date support for GUPROF in userland, as in
the sparc64 version.

20 years agoInclude <sys/gmon.h> instead of <machine/profile.h> for the declaration
Bruce Evans [Wed, 19 May 2004 14:36:38 +0000 (14:36 +0000)]
Include <sys/gmon.h> instead of <machine/profile.h> for the declaration
of kmupetext().  The declaration is misplaced in <machine/profile.h>
since it is not MD and not related to the lowest level of profiling.
It will be moved, but getting it via <sys/gmon.h> already works.

20 years agoWhitespace nit.
Ruslan Ermilov [Wed, 19 May 2004 11:35:03 +0000 (11:35 +0000)]
Whitespace nit.

20 years agoMaintain statistics about the received frames.
Ruslan Ermilov [Wed, 19 May 2004 11:26:33 +0000 (11:26 +0000)]
Maintain statistics about the received frames.

20 years agoFix integer overflow in the file size output when dealing with
Lukas Ertl [Wed, 19 May 2004 11:07:30 +0000 (11:07 +0000)]
Fix integer overflow in the file size output when dealing with
large files (i.e. DVD images).

Reviewed by:  des@

20 years agoAdded pgrep to the SYNOPSIS.
Ruslan Ermilov [Wed, 19 May 2004 10:46:56 +0000 (10:46 +0000)]
Added pgrep to the SYNOPSIS.
Fixed SYNOPSIS.
Fixed markup nits.

20 years agoClean up language and markup.
Ruslan Ermilov [Wed, 19 May 2004 10:45:40 +0000 (10:45 +0000)]
Clean up language and markup.

20 years agoLanguage and markup polishing.
Ruslan Ermilov [Wed, 19 May 2004 10:44:24 +0000 (10:44 +0000)]
Language and markup polishing.

20 years agoBumped the document date.
Ruslan Ermilov [Wed, 19 May 2004 09:54:14 +0000 (09:54 +0000)]
Bumped the document date.

20 years agoBumped the document date.
Ruslan Ermilov [Wed, 19 May 2004 09:53:41 +0000 (09:53 +0000)]
Bumped the document date.
Fixed the grammar nit.

20 years agoBumped document date.
Ruslan Ermilov [Wed, 19 May 2004 09:53:03 +0000 (09:53 +0000)]
Bumped document date.
Fixed markup nit.

20 years agoBumped document date.
Ruslan Ermilov [Wed, 19 May 2004 09:52:40 +0000 (09:52 +0000)]
Bumped document date.
Fixed the grammar nit.

20 years agoMarkup nits.
Ruslan Ermilov [Wed, 19 May 2004 09:51:58 +0000 (09:51 +0000)]
Markup nits.

20 years agoBumped document date.
Ruslan Ermilov [Wed, 19 May 2004 09:51:31 +0000 (09:51 +0000)]
Bumped document date.
Dealt with hard sentence breaks and whitespace at EOL.

20 years agoFixed document date.
Ruslan Ermilov [Wed, 19 May 2004 09:49:54 +0000 (09:49 +0000)]
Fixed document date.

20 years agoBumped document date.
Ruslan Ermilov [Wed, 19 May 2004 09:49:37 +0000 (09:49 +0000)]
Bumped document date.
Fixed a bunch of hyphen misspellings.
Fixed one warning.

20 years agoRestore part of the word that got (accidentally) removed.
Ruslan Ermilov [Wed, 19 May 2004 09:46:37 +0000 (09:46 +0000)]
Restore part of the word that got (accidentally) removed.
Removed whitespace at EOL.

20 years agoMarkup nits.
Ruslan Ermilov [Wed, 19 May 2004 09:45:46 +0000 (09:45 +0000)]
Markup nits.

20 years agoBumped document date.
Ruslan Ermilov [Wed, 19 May 2004 09:44:31 +0000 (09:44 +0000)]
Bumped document date.
Removed hard sentence break.

20 years agoAllow the location of the INDEX file to specified to pkg_version.
Josef Karthauser [Wed, 19 May 2004 09:18:08 +0000 (09:18 +0000)]
Allow the location of the INDEX file to specified to pkg_version.
This is particularly convenient on a cluster of machines to prevent
having to rebuild the INDEX file on each.

Reviewed by: portmgr

20 years agoI've recently been looking at the Seventh Edition source
Tim Kientzle [Wed, 19 May 2004 06:38:38 +0000 (06:38 +0000)]
I've recently been looking at the Seventh Edition source
code available at tuhs.org, and found out that my chronology
is a bit off.  In particular, Seventh Edition already used
the "linkflag" and "linkname" fields.  Also, it appears that
there was no tar in Sixth Edition, contrary to what an earlier
tar.1 manpage claimed.

A few mdoc fixes also crept in here.

20 years agoRefine the heuristic used to determine whether or not to obey
Tim Kientzle [Wed, 19 May 2004 06:35:47 +0000 (06:35 +0000)]
Refine the heuristic used to determine whether or not to obey
the size field for a hardlink entry.  Specifically, ensure that
we do obey the size field for archives that we know are pax interchange
format archives, as required by POSIX.

Also, clarify the comment explaining why this is necessary and explain
the (very unusual) conditions under which it might fail.

20 years agoMerge parser fix from vendor branch (rev 1.1.1.16)
Dag-Erling Smørgrav [Wed, 19 May 2004 06:21:56 +0000 (06:21 +0000)]
Merge parser fix from vendor branch (rev 1.1.1.16)

20 years agotrap_pfault() shouldn't be acquiring Giant. Found to blow up
Peter Grehan [Wed, 19 May 2004 06:05:42 +0000 (06:05 +0000)]
trap_pfault() shouldn't be acquiring Giant. Found to blow up
with MUTEX_PROFILING.

Submitted by:  Suleiman Souhlal <refugee@segfaulted.com>

20 years agoAdd explicit list of SiS AGP chipsets based on Linux kernel's list.
Eric Anholt [Wed, 19 May 2004 05:25:31 +0000 (05:25 +0000)]
Add explicit list of SiS AGP chipsets based on Linux kernel's list.

Prompted by: i386/59503

20 years agoRemove a long obsolete paragraph from the BUGS section.
Alan Cox [Wed, 19 May 2004 03:25:17 +0000 (03:25 +0000)]
Remove a long obsolete paragraph from the BUGS section.

20 years agoReplace the lame big endian crc with wpaul's standard big endian crc
Warner Losh [Wed, 19 May 2004 02:16:46 +0000 (02:16 +0000)]
Replace the lame big endian crc with wpaul's standard big endian crc
algorithm, supplied by wpaul himself.  The lame one has an origin
that's been called into question, so rather than argue about that (one
could make an excellent fair use argument), replace it with better
code since that's what FreeBSD is about.

Submitted by: wpaul[1], Klaus Klein

[1] Bill called this a silly bikeshed.  Maybe his is not incorrect.

20 years agoUnbreak builds without DDB. Bad Bruce! No cookie! :-)
Peter Wemm [Wed, 19 May 2004 01:23:48 +0000 (01:23 +0000)]
Unbreak builds without DDB.  Bad Bruce! No cookie! :-)

20 years agoDefault is to restore timestamps, even when operating as root. <sigh>
Tim Kientzle [Wed, 19 May 2004 01:21:11 +0000 (01:21 +0000)]
Default is to restore timestamps, even when operating as root. <sigh>

20 years agosyncache broke rev 1.23 which was done to fix the "thundering herd"
Paul Saab [Wed, 19 May 2004 00:22:10 +0000 (00:22 +0000)]
syncache broke rev 1.23 which was done to fix the "thundering herd"
problem in Apache.  Fix it.

Reviewed by: peter

20 years agoFix a segfault when looking up a non-existent group.
Tim Kientzle [Tue, 18 May 2004 23:40:25 +0000 (23:40 +0000)]
Fix a segfault when looking up a non-existent group.

20 years agoThe 'call mcount' hooks that gcc inserts when profiling are in a place that
Peter Wemm [Tue, 18 May 2004 22:52:32 +0000 (22:52 +0000)]
The 'call mcount' hooks that gcc inserts when profiling are in a place that
cannot handle the scratch registers being trashed.  So we have to preserve
them ourselves.

20 years agoFor amd64, explicitly compile mcount.po, rather than copying mcount.o. We
Peter Wemm [Tue, 18 May 2004 22:49:15 +0000 (22:49 +0000)]
For amd64, explicitly compile mcount.po, rather than copying mcount.o.  We
need to compile it with -fno-omit-frame-pointers since the mcount code
depends on that, and by default it omits them without -pg.

20 years agoBump document date on behalf of the -e option.
Ruslan Ermilov [Tue, 18 May 2004 20:45:32 +0000 (20:45 +0000)]
Bump document date on behalf of the -e option.
Spell "ID" correctly.

20 years agoBump document date on behalf of the -E option.
Ruslan Ermilov [Tue, 18 May 2004 20:42:36 +0000 (20:42 +0000)]
Bump document date on behalf of the -E option.

20 years agoAdded -u to the SYNOPSIS.
Ruslan Ermilov [Tue, 18 May 2004 20:40:37 +0000 (20:40 +0000)]
Added -u to the SYNOPSIS.

20 years agoBump document date on behalf of the -u option.
Ruslan Ermilov [Tue, 18 May 2004 20:38:43 +0000 (20:38 +0000)]
Bump document date on behalf of the -u option.

20 years agoBump document date on behalf of the -M option.
Ruslan Ermilov [Tue, 18 May 2004 20:36:54 +0000 (20:36 +0000)]
Bump document date on behalf of the -M option.

20 years agoImprove the delay algorithm used in bgfsck. From the author:
Scott Long [Tue, 18 May 2004 19:51:41 +0000 (19:51 +0000)]
Improve the delay algorithm used in bgfsck.  From the author:

 shuffles the timing and sleep calls in bgfsck from:

 sleep timer_on io timer_off io io io io io io io
   to
 sleep io io io io io io io timer_on io timer_off

 The original method basically guaranteed that the timed I/O included a
 disk seek every time, which made bgfsck sleep for much longer than
 necessary.

Submitted by:   Dan Nelson
Reviewed by:    kirk

20 years ago"in in" is not always wrong (backout 1.18).
Daniel Harris [Tue, 18 May 2004 18:17:25 +0000 (18:17 +0000)]
"in in" is not always wrong (backout 1.18).

Keep the onion metaphor consistent.

20 years agoBe smarter about hardlink sizes: some tar programs write
Tim Kientzle [Tue, 18 May 2004 18:16:30 +0000 (18:16 +0000)]
Be smarter about hardlink sizes:  some tar programs write
a non-zero size but no body, some write a non-zero size and include
a body.  To distinguish these cases, look for a valid tar header immediately
following a hardlink header with non-zero size.

20 years agoBuild/install gprof on amd64 as well, after repocopying the MD bits.
Peter Wemm [Tue, 18 May 2004 18:05:19 +0000 (18:05 +0000)]
Build/install gprof on amd64 as well, after repocopying the MD bits.
Not that it has anything to process yet, but that comes next.

20 years agoEffect the correct use of "affect".
Daniel Harris [Tue, 18 May 2004 17:50:08 +0000 (17:50 +0000)]
Effect the correct use of "affect".

Use em dashes instead of " - ".

Use .Em instead of *emphasis*.

Note that securing root indirectly (by securing staff accounts) works
only if direct root access has been limited. [1]

s/hacker/attacker, as done in the handbook. (inspired by [1])

PR: 52878 [1]
Submitted by: Brian Minard <bminard@flatfoot.ca> [1]

20 years agoAccept (but ignore) GNU tar's -B option. bsdtar never
Tim Kientzle [Tue, 18 May 2004 17:16:55 +0000 (17:16 +0000)]
Accept (but ignore) GNU tar's -B option.  bsdtar never
needs it, but it does no harm for people to specify it.

20 years agoUse the simpler __BUS_ACCESSOR macros for ivars instead of defining them
Nate Lawson [Tue, 18 May 2004 16:53:29 +0000 (16:53 +0000)]
Use the simpler __BUS_ACCESSOR macros for ivars instead of defining them
ourselves.

20 years agoPressing 's' in the initial menu should result in selecting the 'Standard'
Ken Smith [Tue, 18 May 2004 16:18:04 +0000 (16:18 +0000)]
Pressing 's' in the initial menu should result in selecting the 'Standard'
installation as far as most people are concerned but both 'Standard' and
'Select' begin with S and 'Select' is winning.  This makes it so 'Select'
is not select-able using a keystroke but that is probably for the best
and the text on the screen adequately describes how to move back and forth
between 'Select' and 'Exit'.

Adapted from work by: josef@
PR: i386/37999
MFC after: 1 week

20 years ago<stdint.h> should define WINT_M{AX,IN} independent from whether WCHAR_MIN is
Stefan Farfeleder [Tue, 18 May 2004 16:04:57 +0000 (16:04 +0000)]
<stdint.h> should define WINT_M{AX,IN} independent from whether WCHAR_MIN is
defined.  Otherwise first including <wchar.h> and then <stdint.h> leads to no
WINT_M{AX,IN} at all.

PR: 64956
Approved by: das (mentor)

20 years agoDon't depend on NULL's expansion being a pointer, cast it before it is passed
Stefan Farfeleder [Tue, 18 May 2004 15:53:58 +0000 (15:53 +0000)]
Don't depend on NULL's expansion being a pointer, cast it before it is passed
to variadic functions.

Approved by: das (mentor)

20 years agoo Fix comments: s/m_tag_allocate/m_tag_alloc/.
Maxim Konovalov [Tue, 18 May 2004 14:13:23 +0000 (14:13 +0000)]
o Fix comments: s/m_tag_allocate/m_tag_alloc/.

Spotted by: Gleb Smirnoff
MFC after: 1 week

20 years agoRemove spurious semicolons.
Stefan Farfeleder [Tue, 18 May 2004 13:57:37 +0000 (13:57 +0000)]
Remove spurious semicolons.

Approved by: das (mentor)
Reviewed by: ipfw@

20 years agoAn access count change is a prerequisite for doing I/O, so
Lukas Ertl [Tue, 18 May 2004 12:37:47 +0000 (12:37 +0000)]
An access count change is a prerequisite for doing I/O, so
change 'or' to 'and'.

Pointed out by:  pjd@

20 years agoThese hyphens don't need to be escaped.
Lukas Ertl [Tue, 18 May 2004 12:36:12 +0000 (12:36 +0000)]
These hyphens don't need to be escaped.

Pointed out by:  pjd@

20 years agoFix spelling errors (I forgot that one in my last commit of GEOM man pages).
Lukas Ertl [Tue, 18 May 2004 12:12:07 +0000 (12:12 +0000)]
Fix spelling errors (I forgot that one in my last commit of GEOM man pages).

20 years agoThe cp driver may be useful for pc98.
Yoshihiro Takahashi [Tue, 18 May 2004 12:02:39 +0000 (12:02 +0000)]
The cp driver may be useful for pc98.

20 years agoMFi386: revision 1.1160.
Yoshihiro Takahashi [Tue, 18 May 2004 11:56:56 +0000 (11:56 +0000)]
MFi386: revision 1.1160.

20 years agoMFi386: Add NETGRAPH_CRONYX.
Yoshihiro Takahashi [Tue, 18 May 2004 11:56:24 +0000 (11:56 +0000)]
MFi386: Add NETGRAPH_CRONYX.

20 years agoMFi386: revision 1.492.
Yoshihiro Takahashi [Tue, 18 May 2004 11:34:04 +0000 (11:34 +0000)]
MFi386: revision 1.492.

20 years agoUpdate several GEOM related manual pages:
Lukas Ertl [Tue, 18 May 2004 11:33:37 +0000 (11:33 +0000)]
Update several GEOM related manual pages:

  *) Fix spelling errors.
  *) Better english.
  *) Re-phrasing.

Not Reviewed by:  pjd (one week timeout)
Approved by:      grog (mentor)

20 years agoNote about ng_sppp (4) driver.
Roman Kurakin [Tue, 18 May 2004 07:38:46 +0000 (07:38 +0000)]
Note about ng_sppp (4) driver.

20 years agoFix panic which occurs when given sector size for memory-backed device
Pawel Jakub Dawidek [Tue, 18 May 2004 07:30:04 +0000 (07:30 +0000)]
Fix panic which occurs when given sector size for memory-backed device
is less than DEV_BSIZE (512) bytes.

Reported by: Mike Bristow <mike@urgle.com>
Approved by: phk

20 years agoMarkup overhaul.
Ruslan Ermilov [Tue, 18 May 2004 07:23:37 +0000 (07:23 +0000)]
Markup overhaul.

20 years ago1. Remove '-' sign between model an manufacturer.
Roman Kurakin [Tue, 18 May 2004 07:12:06 +0000 (07:12 +0000)]
1. Remove '-' sign between model an manufacturer.
2. Add information about cp(4) driver (Cronyx Tau-PCI).

20 years agoAdd information about cp(4) driver (Cronyx Tau-PCI).
Roman Kurakin [Tue, 18 May 2004 06:48:27 +0000 (06:48 +0000)]
Add information about cp(4) driver (Cronyx Tau-PCI).

20 years agoct(4) now has manual page, so add reference to it.
Roman Kurakin [Tue, 18 May 2004 06:36:17 +0000 (06:36 +0000)]
ct(4) now has manual page, so add reference to it.

20 years agoFixed DDB_NOKLDSYM on amd64's:
Bruce Evans [Tue, 18 May 2004 05:30:06 +0000 (05:30 +0000)]
Fixed DDB_NOKLDSYM on amd64's:

machdep.c:
Initialize the symbol table pointers, not quite like for other arches.

db_elf.c:
Don't claim to be an i486 in the fake ELF header.

20 years agoIf a symbol has section+offset definitions provided, always use instead
Peter Wemm [Tue, 18 May 2004 05:15:43 +0000 (05:15 +0000)]
If a symbol has section+offset definitions provided, always use instead
of doing a name lookup for global symbols.  This fixes the snd_pcm module.

20 years agoRemove leftover padding variables.
Peter Wemm [Tue, 18 May 2004 05:14:19 +0000 (05:14 +0000)]
Remove leftover padding variables.
Convert some silent 'ignore programmer error' cases into panics
Remove 'align' field from section table (no longer needed)

20 years agoClarify an error message.
Tim Kientzle [Tue, 18 May 2004 00:13:06 +0000 (00:13 +0000)]
Clarify an error message.

20 years agoClarify and extend paragraphs on interoperation
Yaroslav Tykhiy [Mon, 17 May 2004 23:09:10 +0000 (23:09 +0000)]
Clarify and extend paragraphs on interoperation
of fcntl(2), flock(2), and lockf(3) advisory locks.
Add such a paragraph to the flock(2) manpage for the
sake of consistency.

Reviewed by: Cyrille Lefevre and Kirk McKusick on -arch
MFC after: 2 weeks

20 years agoOops, I left a duplicate 'relocbase' declaration.
Peter Wemm [Mon, 17 May 2004 22:26:17 +0000 (22:26 +0000)]
Oops, I left a duplicate 'relocbase' declaration.

Submitted by:  Koop Mast <kwm@rainbow-runner.nl>

20 years agogetgrent() and friends should set errno if there is an error.
Tim Kientzle [Mon, 17 May 2004 22:15:49 +0000 (22:15 +0000)]
getgrent() and friends should set errno if there is an error.
Also, clarify the manpage description of when errno is set and
explain that clients should set errno=0 first if they want useful
error information.

20 years agoTurn on modules for amd64. Fear.
Peter Wemm [Mon, 17 May 2004 22:13:14 +0000 (22:13 +0000)]
Turn on modules for amd64.  Fear.

20 years agoAdd rman_get_device(9) to SYNOPSIS, .Nm and MLINK it to
Hiten Pandya [Mon, 17 May 2004 21:25:02 +0000 (21:25 +0000)]
Add rman_get_device(9) to SYNOPSIS, .Nm and MLINK it to
rman(9).

20 years agoSince we go to the trouble of compiling the kobj ops table for each class,
Peter Wemm [Mon, 17 May 2004 21:24:39 +0000 (21:24 +0000)]
Since we go to the trouble of compiling the kobj ops table for each class,
and cannot handle it going away, add an explicit reference to the kobj
class inside each linker class.  Without this, a class with no modules
loaded will sit with an idle refcount of 0.  Loading and unloading
a module with it causes a 0->1->0 transition which frees the ops table
and causes subsequent loads using that class to explode.  Normally, the
"kernel" module will remain forever loaded and prevent this happening, but
if you have more than one linker class active, only one owns the "kernel".

This finishes making modules work for kldload(8) on amd64.

20 years agoMarkup nits.
Ruslan Ermilov [Mon, 17 May 2004 21:24:06 +0000 (21:24 +0000)]
Markup nits.

20 years agoClean up the code some more. Unify the text/data (progbits) and bss
Peter Wemm [Mon, 17 May 2004 21:20:23 +0000 (21:20 +0000)]
Clean up the code some more.  Unify the text/data (progbits) and bss
(nobits) tables to simplify some code.  Try and shorten some of the very
wide lines.  Somewhere along the way, I think I fixed the memory
corruption that caused panics after going multiuser.

20 years agoDeal with REL records that have the addend embedded variable sized targets
Peter Wemm [Mon, 17 May 2004 21:16:49 +0000 (21:16 +0000)]
Deal with REL records that have the addend embedded variable sized targets
rather than the RELA table.  I dont know if bintutils will ever generate
REL records, but just in case.....

20 years agoCorrect existing usage lines, add two more.
Ruslan Ermilov [Mon, 17 May 2004 21:11:35 +0000 (21:11 +0000)]
Correct existing usage lines, add two more.

20 years agoMarkup nits.
Ruslan Ermilov [Mon, 17 May 2004 21:05:39 +0000 (21:05 +0000)]
Markup nits.

20 years agoMarkup and grammar nits.
Ruslan Ermilov [Mon, 17 May 2004 20:58:42 +0000 (20:58 +0000)]
Markup and grammar nits.

20 years agoMarkup overhaul.
Ruslan Ermilov [Mon, 17 May 2004 20:45:21 +0000 (20:45 +0000)]
Markup overhaul.

20 years agoNow that I learned about how all this is working together, quickly
Joerg Wunsch [Mon, 17 May 2004 20:17:39 +0000 (20:17 +0000)]
Now that I learned about how all this is working together, quickly
document it for anyone else who might be interested in it.

MFC after: 2 weeks

20 years agoSwitch from using the vnode interlock to a private mutex in fifo_open()
Don Lewis [Mon, 17 May 2004 20:16:40 +0000 (20:16 +0000)]
Switch from using the vnode interlock to a private mutex in fifo_open()
to avoid lock order problems when manipulating the sockets associated
with the fifo.

Minor optimization of a couple of calls to fifo_cleanup() from
fifo_open().

20 years agoNG_ONE2MANY_XMIT_ALL algorithm should also be listed here.
Pawel Jakub Dawidek [Mon, 17 May 2004 19:35:40 +0000 (19:35 +0000)]
NG_ONE2MANY_XMIT_ALL algorithm should also be listed here.

Submitted by: Michel Lavonds <fox@vader.aacc.cc.md.us>
Approved by: archie

20 years agoFix my own style(9) bugs:
Joerg Wunsch [Mon, 17 May 2004 19:19:08 +0000 (19:19 +0000)]
Fix my own style(9) bugs:

. forward declare all static functions
. add a couple of redundant parens in return statements where they've
  been missing
. remove the space after exit since it's a function

20 years agoFix various style(9) bugs. This includes the removal of wrong
Joerg Wunsch [Mon, 17 May 2004 18:55:45 +0000 (18:55 +0000)]
Fix various style(9) bugs.  This includes the removal of wrong
reimplementations of enodev() (for the smbread() and smbwrite()
functions), as well as fixing various errno values to conform to
errno(3).

Bruce also points out that a number of the pointer == NULL tests
are probably nonsense because the respective checks are already
done at upper layers.

(Mostly) submitted by: bde

20 years agoFix typo in comment. While here, end the sentence with a period and
Marcel Moolenaar [Mon, 17 May 2004 18:36:14 +0000 (18:36 +0000)]
Fix typo in comment. While here, end the sentence with a period and
remove the empty line between the fdc and sio devices. The empty
line suggests that the comment applies to fdc only while it applies
to all following devices and options.

Typo spotted by: ru@

20 years agoSet errno to 0 before calling getpwuid/getgrgid, so that error
Tim Kientzle [Mon, 17 May 2004 18:29:12 +0000 (18:29 +0000)]
Set errno to 0 before calling getpwuid/getgrgid, so that error
messages will be displayed only when there is, in fact, an error.

20 years agoPOSIX prohibits any library function from setting errno to 0.
Tim Kientzle [Mon, 17 May 2004 18:27:05 +0000 (18:27 +0000)]
POSIX prohibits any library function from setting errno to 0.
Correct my previous commit and add a comment to the manpage
indicating that the user must set errno to 0 if they wish to
distinguish "no such user" from "error".

Pointed out by: Jacques Vidrine (nectar@)

20 years agoMinor language fixes.
Yaroslav Tykhiy [Mon, 17 May 2004 17:55:19 +0000 (17:55 +0000)]
Minor language fixes.

20 years agoRip out the too verbose "spurious interrupt" printf's, they dont serve
Søren Schmidt [Mon, 17 May 2004 17:53:12 +0000 (17:53 +0000)]
Rip out the too verbose "spurious interrupt" printf's, they dont serve
a purpose any longer.

20 years agoNew release notes:
Yaroslav Tykhiy [Mon, 17 May 2004 17:43:40 +0000 (17:43 +0000)]
New release notes:
cron(8) -j and -J options.

20 years agoTurn SCSI pre-fetch ON. This is mainly for 64XX and 64X based
Paul Saab [Mon, 17 May 2004 17:27:38 +0000 (17:27 +0000)]
Turn SCSI pre-fetch ON.  This is mainly for 64XX and 64X based
controllers and allows the controller to prefetch 1-2k on certain
PCI memory reads to the host.  The spec says this should only be
used for IA32 based systems.

Informed of feature by: John Cagle <first.last@hp.com>

20 years agoUndid scottl's recent changes.
Vinod Kashyap [Mon, 17 May 2004 17:16:58 +0000 (17:16 +0000)]
Undid scottl's recent changes.

20 years agoAllow buildworld and friends to complete when make(1) is called
Ruslan Ermilov [Mon, 17 May 2004 16:19:51 +0000 (16:19 +0000)]
Allow buildworld and friends to complete when make(1) is called
with some debug flags (-d).

Noticed by: Gleb Smirnoff

20 years agoAdd Cronyx Tau-PCI sync WAN adapters family entry.
Roman Kurakin [Mon, 17 May 2004 14:42:59 +0000 (14:42 +0000)]
Add Cronyx Tau-PCI sync WAN adapters family entry.

20 years agoSpell Cronyx Tau and Sigma families correctly.
Roman Kurakin [Mon, 17 May 2004 14:35:32 +0000 (14:35 +0000)]
Spell Cronyx Tau and Sigma families correctly.

MFC after: 3 days

20 years agoConnect Cronyx Tau-PCI to the system.
Roman Kurakin [Mon, 17 May 2004 14:24:52 +0000 (14:24 +0000)]
Connect Cronyx Tau-PCI to the system.

20 years agoUpdate man page date for yesterday's change.
Archie Cobbs [Mon, 17 May 2004 14:21:00 +0000 (14:21 +0000)]
Update man page date for yesterday's change.