]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoAdd my birthday.
philip [Wed, 21 Jan 2004 18:11:46 +0000 (18:11 +0000)]
Add my birthday.

Approved by: njl (mentor)

20 years agoDon't leak softc's when destroying interfaces.
brooks [Wed, 21 Jan 2004 18:05:38 +0000 (18:05 +0000)]
Don't leak softc's when destroying interfaces.

Init the softc list when loaded.

Noticed by: Maurycy Pawlowski-Wieronski <maurycy at fouk dot org>

20 years agoFix style(9) of my previous commit.
cperciva [Wed, 21 Jan 2004 18:03:54 +0000 (18:03 +0000)]
Fix style(9) of my previous commit.

Noticed by: nate
Approved by: nate, rwatson (mentor)

20 years agoRevert previous commit, ru has ideas how to better document this
schweikh [Wed, 21 Jan 2004 17:58:33 +0000 (17:58 +0000)]
Revert previous commit, ru has ideas how to better document this
elsewhere.

20 years agoReduce gratuitous includes: don't include jail.h if it's not needed.
rwatson [Wed, 21 Jan 2004 17:10:47 +0000 (17:10 +0000)]
Reduce gratuitous includes: don't include jail.h if it's not needed.
Presumably, at some point, you had to include jail.h if you included
proc.h, but that is no longer required.

Result of: self injury involving adding something to struct prison

20 years agoNew errata: IPsec bugs. Source material for this errata note was...
bmah [Wed, 21 Jan 2004 17:03:13 +0000 (17:03 +0000)]
New errata:  IPsec bugs.  Source material for this errata note was...

Submitted by: Bjoern A. Zeeb < bzeeb at Zabbadoz dot NeT >

20 years agoDisable constant merging in isdnd and isdnmonitor. This is a "make my
cperciva [Wed, 21 Jan 2004 16:49:05 +0000 (16:49 +0000)]
Disable constant merging in isdnd and isdnmonitor.  This is a "make my
life easier" patch: I doubt this will affect anyone else, but the FreeBSD
Update build code was getting very confused by this.

Approved by: rwatson (mentor)
PR: bin/61087

20 years agoTheo de Raadt has brought to our attention that he wrote at least part
imp [Wed, 21 Jan 2004 16:44:49 +0000 (16:44 +0000)]
Theo de Raadt has brought to our attention that he wrote at least part
file and that he has not released it under a free license.  Until the
matter can be resolved, honor his claim and remove it.

PR: misc/61676

20 years agoAllow devfs path rules to work on directories. Without this fix,
cperciva [Wed, 21 Jan 2004 16:43:29 +0000 (16:43 +0000)]
Allow devfs path rules to work on directories.  Without this fix,
devfs rule add path fd unhide
is a no-op, while it should unhide the fd subdirectory.

Approved by: phk, rwatson (mentor)
PR: kern/60897

20 years agoAfter setting a ruleset on a new devfs mount, apply it.
cperciva [Wed, 21 Jan 2004 16:36:17 +0000 (16:36 +0000)]
After setting a ruleset on a new devfs mount, apply it.

Approved by: phk, rwatson (mentor)
PR: bin/61047

20 years agoMerge security fix from vendor branch (revision 1.1.1.14).
nectar [Wed, 21 Jan 2004 16:29:16 +0000 (16:29 +0000)]
Merge security fix from vendor branch (revision 1.1.1.14).

20 years agoMFi386: revisions from 1.1116 to 1.1119.
nyan [Wed, 21 Jan 2004 16:28:19 +0000 (16:28 +0000)]
MFi386: revisions from 1.1116 to 1.1119.
Remove NEWCARD related devices.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r124793,
nectar [Wed, 21 Jan 2004 16:27:56 +0000 (16:27 +0000)]
This commit was generated by cvs2svn to compensate for changes in r124793,
which included commits to RCS files with non-trunk default branches.

20 years agoIncorporate two security fixes from later versions of CVS.
nectar [Wed, 21 Jan 2004 16:27:56 +0000 (16:27 +0000)]
Incorporate two security fixes from later versions of CVS.
From the NEWS file of cvs 1.11.11:

  * pserver can no longer be configured to run as root via the
    $CVSROOT/CVSROOT/passwd file, so if your passwd file is
    compromised, it no longer leads directly to a root hack.  Attempts
    to root will also be logged via the syslog.

  * Malformed module requests could cause the CVS server to attempt
    to create directories and possibly files at the root of the
    filesystem holding the CVS repository.  Filesystem permissions
    usually prevent the creation of these misplaced directories, but
    nevertheless, the CVS server now rejects the malformed requests.

Obtained from: ccvs.cvshome.org

20 years agoBring ps.1 up to date with changes in the past seven years:
cperciva [Wed, 21 Jan 2004 16:25:36 +0000 (16:25 +0000)]
Bring ps.1 up to date with changes in the past seven years:
* Remove mention of '>', 'A', and 'S' states
* Mention 'W' state.
* List 'J' state in the correct location.
* Sync with flags in sys/proc.h

Approved by: rwatson (mentor)
MFC after: 7 days

20 years agoMerged from sys/dev/sio/sio.c revision 1.418.
nyan [Wed, 21 Jan 2004 16:12:29 +0000 (16:12 +0000)]
Merged from sys/dev/sio/sio.c revision 1.418.

20 years agoSync to pccarddevs 1.81.
ume [Wed, 21 Jan 2004 16:03:40 +0000 (16:03 +0000)]
Sync to pccarddevs 1.81.

20 years agoThere is PANASONIC KXLC005 which has another product id.
ume [Wed, 21 Jan 2004 16:02:20 +0000 (16:02 +0000)]
There is PANASONIC KXLC005 which has another product id.

20 years agoThis is Sparc64-only.
ru [Wed, 21 Jan 2004 14:06:25 +0000 (14:06 +0000)]
This is Sparc64-only.

20 years agoThis is i386-only.
ru [Wed, 21 Jan 2004 14:01:29 +0000 (14:01 +0000)]
This is i386-only.

20 years agoMark APM manpages as i386-only.
ru [Wed, 21 Jan 2004 13:38:17 +0000 (13:38 +0000)]
Mark APM manpages as i386-only.

20 years agoMark this manpage as i386-only.
ru [Wed, 21 Jan 2004 13:24:38 +0000 (13:24 +0000)]
Mark this manpage as i386-only.

20 years agoA couple of minor clarifications.
yar [Wed, 21 Jan 2004 13:17:24 +0000 (13:17 +0000)]
A couple of minor clarifications.

20 years agoCreate section 4 and 8 man subdirectories for all Tier 1 architectures.
ru [Wed, 21 Jan 2004 13:15:19 +0000 (13:15 +0000)]
Create section 4 and 8 man subdirectories for all Tier 1 architectures.

20 years agoRemoved unused {man,cat}9/i386.
ru [Wed, 21 Jan 2004 13:09:16 +0000 (13:09 +0000)]
Removed unused {man,cat}9/i386.

20 years agoReplace home-grown dup2() implementation with actual dup2() calls. This
des [Wed, 21 Jan 2004 12:50:01 +0000 (12:50 +0000)]
Replace home-grown dup2() implementation with actual dup2() calls.  This
should slightly reduce the number of system calls in critical portions of
the shell, and select a more efficient path through the fdalloc code.

Reviewed by: bde

20 years agoAdd a description for the hardware-assisted checksumming
yar [Wed, 21 Jan 2004 12:45:37 +0000 (12:45 +0000)]
Add a description for the hardware-assisted checksumming
facilities based on http://people.freebsd.org/~jlemon/csum.txt
and my own observations.

20 years agoSetting pccb->ccb_h.status to CAM_REQ_CMP one time is enough.
cognet [Wed, 21 Jan 2004 10:49:42 +0000 (10:49 +0000)]
Setting pccb->ccb_h.status to CAM_REQ_CMP one time is enough.

20 years agoUse the reinstallkernel.debug target, now that we have it.
ru [Wed, 21 Jan 2004 08:24:44 +0000 (08:24 +0000)]
Use the reinstallkernel.debug target, now that we have it.

20 years agoThe gdbinit.${MACHINE_ARCH} script may not exist.
ru [Wed, 21 Jan 2004 08:23:46 +0000 (08:23 +0000)]
The gdbinit.${MACHINE_ARCH} script may not exist.

20 years agoAdd syscons options and enable USB, since there is no conflict between
grehan [Wed, 21 Jan 2004 05:22:53 +0000 (05:22 +0000)]
Add syscons options and enable USB, since there is no conflict between
the OpenFirmware console and the syscons console when using a USB
keyboard.

20 years agoAdd required GFB options as well as one for ofw/syscons.
grehan [Wed, 21 Jan 2004 05:20:58 +0000 (05:20 +0000)]
Add required GFB options as well as one for ofw/syscons.

20 years agoAdd syscons files, and also generate a syscons font since Apple
grehan [Wed, 21 Jan 2004 05:19:59 +0000 (05:19 +0000)]
Add syscons files, and also generate a syscons font since Apple
SVGA adapters don't have one available in their ROMs.

20 years ago- Catch up with panic __LINE__/__FILE__ changes by moving panic calls
grehan [Wed, 21 Jan 2004 05:18:08 +0000 (05:18 +0000)]
- Catch up with panic __LINE__/__FILE__ changes by moving panic calls
out of asm.
- remove some long-dead code from machdep.c

20 years agoA syscons implementation using the 8-bit framebuffer set up by
grehan [Wed, 21 Jan 2004 05:16:23 +0000 (05:16 +0000)]
A syscons implementation using the 8-bit framebuffer set up by
OpenFirmware. Not at all optimized, but provides a PC-style
user-experience.

Tested on revA imac, B&W G3, 2k iBook, and G4 eMac.

20 years ago__powerpc__ conditional code for the syscons OpenFirmware/PPC framebuffer.
grehan [Wed, 21 Jan 2004 05:08:51 +0000 (05:08 +0000)]
__powerpc__ conditional code for the syscons OpenFirmware/PPC framebuffer.
Took the opportunity to reduce

 __i386__ || __ia64__ || __amd64__ || __sparc64__ || __powerpc__

 to

 !__alpha__

reviewed by: gallatin

20 years agoInclude <stdlib.h> to get abort() prototype.
grehan [Wed, 21 Jan 2004 05:04:25 +0000 (05:04 +0000)]
Include <stdlib.h> to get abort() prototype.

20 years agoUpdate 128-bit long double constants to match what is expected
grehan [Wed, 21 Jan 2004 04:56:39 +0000 (04:56 +0000)]
Update 128-bit long double constants to match what is expected
by libc

20 years agoUpdated manh shift constant type to 'ULL' for PPC to fix
grehan [Wed, 21 Jan 2004 04:51:50 +0000 (04:51 +0000)]
Updated manh shift constant type to 'ULL' for PPC to fix
shift-too-large compile error

reviewed by: das

20 years agoMake sure pid file creation doesn't complain if it can't create the file.
njl [Wed, 21 Jan 2004 03:03:40 +0000 (03:03 +0000)]
Make sure pid file creation doesn't complain if it can't create the file.
This removes a warning message when suspending is called by a non-root
user.

20 years agoFix ipip_output() to always set *mp to NULL on failure, even if 'm'
sam [Tue, 20 Jan 2004 22:45:10 +0000 (22:45 +0000)]
Fix ipip_output() to always set *mp to NULL on failure, even if 'm'
is NULL, otherwise ipsec4_process_packet() may try to m_freem() a
bad pointer.

In ipsec4_process_packet(), don't try to m_freem() 'm' twice; ipip_output()
already did it.

Obtained from: netbsd

20 years agofix build after KAME changes
sam [Tue, 20 Jan 2004 22:44:21 +0000 (22:44 +0000)]
fix build after KAME changes

20 years agoAdd ports-accessibility and ports-x11-themes to the example supfiles.
marcus [Tue, 20 Jan 2004 22:39:57 +0000 (22:39 +0000)]
Add ports-accessibility and ports-x11-themes to the example supfiles.

20 years agoMerge from DragonFlyBSD rev 1.10:
hsu [Tue, 20 Jan 2004 21:40:25 +0000 (21:40 +0000)]
Merge from DragonFlyBSD rev 1.10:

date: 2003/09/02 10:04:47;  author: hsu;  state: Exp;  lines: +5 -6
Account for when Limited Transmit is not congestion window limited.

Obtained from: DragonFlyBSD

20 years agoUse getenv_int() rather than reimplementing it.
jhb [Tue, 20 Jan 2004 21:38:48 +0000 (21:38 +0000)]
Use getenv_int() rather than reimplementing it.

20 years agoSlightly rearrange wording to deconfuse our man.cgi.
schweikh [Tue, 20 Jan 2004 21:33:45 +0000 (21:33 +0000)]
Slightly rearrange wording to deconfuse our man.cgi.

PR: 60743
Submitted by: Lucas <aspiesrule@mcleodusa.net>
MFC after: 2 weeks

20 years agoImport sdpd(8) sources. This is Bluetooth Service Discovery Protocol daemon.
emax [Tue, 20 Jan 2004 20:48:26 +0000 (20:48 +0000)]
Import sdpd(8) sources. This is Bluetooth Service Discovery Protocol daemon.
Extend libsdp(3) API to allow service registration and removal.
Fix uninitialized variable bug in sdpcontrol(8).

Reviewed by: imp (mentor)
No objection: ru

20 years agoPay attention to mac_portacl_enabled.
rwatson [Tue, 20 Jan 2004 18:33:02 +0000 (18:33 +0000)]
Pay attention to mac_portacl_enabled.

Submitted by:   simon

20 years agoUpdate the support for the VIA 8237 southbridge.
sos [Tue, 20 Jan 2004 16:51:02 +0000 (16:51 +0000)]
Update the support for the VIA 8237 southbridge.

20 years agoAdd /var/db/ports/ (support directory necessary for
eivind [Tue, 20 Jan 2004 13:51:23 +0000 (13:51 +0000)]
Add /var/db/ports/ (support directory necessary for
ports that use the new OPTIONS infrastructure)

20 years agoThe <bsd.files.mk> API seems the best to use here.
ru [Tue, 20 Jan 2004 13:31:35 +0000 (13:31 +0000)]
The <bsd.files.mk> API seems the best to use here.

20 years agoRemoved duplicate y.tab.h from SRCS and CLEANFILES.
ru [Tue, 20 Jan 2004 13:13:40 +0000 (13:13 +0000)]
Removed duplicate y.tab.h from SRCS and CLEANFILES.

20 years agoNormalize.
ru [Tue, 20 Jan 2004 13:09:29 +0000 (13:09 +0000)]
Normalize.

20 years agoGrammar.
ru [Tue, 20 Jan 2004 12:57:28 +0000 (12:57 +0000)]
Grammar.

20 years agomdoc(7): Fix title.
ru [Tue, 20 Jan 2004 10:31:52 +0000 (10:31 +0000)]
mdoc(7): Fix title.

20 years agoUse a single style for share/man/ makefiles:
ru [Tue, 20 Jan 2004 10:27:23 +0000 (10:27 +0000)]
Use a single style for share/man/ makefiles:

- Sort MAN and MLINKS in "dictionary" order ignoring case.

- For multi-value MAN and multi-pair MLINKS, put each value/pair
  on its own line, for easier sorting and so that further diffs
  are easier to see.

20 years agoUse a larger field for the size in blocks; the current width (4 digits)
des [Tue, 20 Jan 2004 09:27:03 +0000 (09:27 +0000)]
Use a larger field for the size in blocks; the current width (4 digits)
is only good for 5 MB.

20 years agoAlso sort ignoring case -- makes finding things by browsing easier.
ru [Tue, 20 Jan 2004 08:55:06 +0000 (08:55 +0000)]
Also sort ignoring case -- makes finding things by browsing easier.

20 years agoCorrect instances of mtx_lock()/mtx_unlock() that should have been
wpaul [Tue, 20 Jan 2004 08:19:42 +0000 (08:19 +0000)]
Correct instances of mtx_lock()/mtx_unlock() that should have been
mtx_pool_lock()/mtx_pool_unlock().

20 years agoReduce latency when using the SNDCTL_DSP_RESET ioctl by calling
matk [Tue, 20 Jan 2004 05:30:09 +0000 (05:30 +0000)]
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling
chn_resetbuf().

Submited by: Pyun YongHyeon <yongari@kt-is.co.kr>
Approved by: tanimura (mentor)

20 years agoFix sort misordering.
obrien [Tue, 20 Jan 2004 04:37:07 +0000 (04:37 +0000)]
Fix sort misordering.

20 years agoDiscard the first 1024 bytes of output as suggested by
das [Tue, 20 Jan 2004 04:22:47 +0000 (04:22 +0000)]
Discard the first 1024 bytes of output as suggested by
http://citeseer.nj.nec.com/fluhrer01weaknesses.html and
http://citeseer.nj.nec.com/531224.html .

PR: 61126
Submitted by: Jeff Ito <jeffi@rcn.com>

20 years agoFix a panic when kldloading a sound driver. Do this by replacing the
matk [Tue, 20 Jan 2004 03:58:57 +0000 (03:58 +0000)]
Fix a panic when kldloading a sound driver.  Do this by replacing the
link-list of dev_t's with named variables.  Remove used code.

Approved by: tanimura (mentor)

20 years agoFix handling of FP traps:
marcel [Tue, 20 Jan 2004 03:29:24 +0000 (03:29 +0000)]
Fix handling of FP traps:
o  For traps, the cr.iip register points to the next instruction to
   execute on interrupt return (modulo slot). Since we need to get
   the bundle of the instruction that caused the FP fault/trap, make
   sure we fetch the previous bundle if the next instruction is in
   fact the first in a bundle.
o  When we call the FPSWA handler, we need to tell it whether it's
   a trap or a fault (first argument). This was hardcoded to mean a
   fault.

Also, for FP faults, when a fault is converted to a trap, adjust the
cr.iip and cr.ipsr registers to point to the next instruction. This
makes sure that the SIGFPE handler gets a consistent state.

20 years agoUse 'uint32_t' instead of 'long' when a 32-bit integer is intended.
das [Tue, 20 Jan 2004 03:02:18 +0000 (03:02 +0000)]
Use 'uint32_t' instead of 'long' when a 32-bit integer is intended.
This results in no functional change, aside from fixing a data
corruption bug on LP64 platforms.  The code here could still use a
significant amount of cleanup.

PR: 56502
Submitted by: hrs (earlier version)

20 years agos/framep/tf/g -- this normalizes on the use of tf to point to the
marcel [Tue, 20 Jan 2004 02:35:46 +0000 (02:35 +0000)]
s/framep/tf/g -- this normalizes on the use of tf to point to the
trapframe and improves grep-ability.

20 years agopread/pwrite:
ache [Tue, 20 Jan 2004 01:27:42 +0000 (01:27 +0000)]
pread/pwrite:
follow lseek spirit - return EINVAL on negative offset for non-VCHR

20 years agoSimplify mpool_get() and mpool_write() by using pread() and pwrite()
das [Tue, 20 Jan 2004 00:40:35 +0000 (00:40 +0000)]
Simplify mpool_get() and mpool_write() by using pread() and pwrite()
instead of lseek()/_read() and lseek()/_write().

PR: bin/54276
Submitted by: <dnelson@allantgroup.com>

20 years agoIf we have a working link again after connectivity loss, or if we need
mbr [Mon, 19 Jan 2004 22:07:59 +0000 (22:07 +0000)]
If we have a working link again after connectivity loss, or if we need
to renew a lease, contact the dhcp-server directly instead of using
INADDR_BROADCAST all the time. This should fix some brain-dead dhcp
server implementations which give you all the time a new IP if the
lease has not yet expired.

Instead of using ICMP to check if the server is alive, we just check
the return value of sendto() and additionally have a timeout there.

20 years agoDocument __MAKE_CONF and its default, /etc/make.conf.
schweikh [Mon, 19 Jan 2004 21:44:58 +0000 (21:44 +0000)]
Document __MAKE_CONF and its default, /etc/make.conf.

MFC-After: 2 weeks

20 years agoAdd linenumber and source filename to panic(9) output.
phk [Mon, 19 Jan 2004 21:27:11 +0000 (21:27 +0000)]
Add linenumber and source filename to panic(9) output.

Ideally a traceback should be printed too, any takers ?

20 years ago"The Aged Aged Man" was not a word doubling.
schweikh [Mon, 19 Jan 2004 21:22:21 +0000 (21:22 +0000)]
"The Aged Aged Man" was not a word doubling.

Spotted by: grog

20 years agoOne more instance of magic number used in place of IO_SEQSHIFT.
kan [Mon, 19 Jan 2004 20:45:43 +0000 (20:45 +0000)]
One more instance of magic number used in place of IO_SEQSHIFT.

Submitted by: alc

20 years agoAdd WDM major/minor #defines.
wpaul [Mon, 19 Jan 2004 20:45:27 +0000 (20:45 +0000)]
Add WDM major/minor #defines.

20 years agoSpell magic '16' number as IO_SEQSHIFT.
kan [Mon, 19 Jan 2004 20:03:43 +0000 (20:03 +0000)]
Spell magic '16' number as IO_SEQSHIFT.

20 years agoImplement IofCompleteRequest() and IoIsWdmVersionAvailable().
wpaul [Mon, 19 Jan 2004 19:57:00 +0000 (19:57 +0000)]
Implement IofCompleteRequest() and IoIsWdmVersionAvailable().
Correct IofCallDriver(): it's fastcall, not stdcall.
Add vector to vsprintf().

20 years agoClean up makefiles.
ru [Mon, 19 Jan 2004 19:26:02 +0000 (19:26 +0000)]
Clean up makefiles.

Reviewed by: harti

20 years agoImplement atoi() and atol(). Some drivers appear to need these. Note
wpaul [Mon, 19 Jan 2004 19:21:25 +0000 (19:21 +0000)]
Implement atoi() and atol(). Some drivers appear to need these. Note
that like most C library routines, these appear to be _cdecl in Windows.

20 years agoEliminate some code duplication: since ndis_runq() and ndis_intq() were
wpaul [Mon, 19 Jan 2004 18:56:31 +0000 (18:56 +0000)]
Eliminate some code duplication: since ndis_runq() and ndis_intq() were
basically the same function, compact them into a single loop which can
be used for both threads.

20 years agolibc is now WARNS=2 clean with the exception of the gdtoa bits (which
nectar [Mon, 19 Jan 2004 16:16:53 +0000 (16:16 +0000)]
libc is now WARNS=2 clean with the exception of the gdtoa bits (which
are now not built with warnings enabled at all).

20 years agoAdd prototypes for the three syscall stubs that are invoked here,
nectar [Mon, 19 Jan 2004 16:14:58 +0000 (16:14 +0000)]
Add prototypes for the three syscall stubs that are invoked here,
in order to quiet warnings.

20 years agoRemove files which have been removed in the 3.01RC12 release
mbr [Mon, 19 Jan 2004 15:31:57 +0000 (15:31 +0000)]
Remove files which have been removed in the 3.01RC12 release
of isc-dhcpd.

20 years agoFix breakage on timeout/retries. The bug cause a sema to be leaked so
sos [Mon, 19 Jan 2004 15:20:00 +0000 (15:20 +0000)]
Fix breakage on timeout/retries. The bug cause a sema to be leaked so
that the calling process would newer wakeup.

20 years agoRefactor _pthread_mutex_init
mtm [Mon, 19 Jan 2004 15:00:57 +0000 (15:00 +0000)]
Refactor _pthread_mutex_init
o Simplify the logic by removing a lot of unnecesary nesting
o Reduce the amount of local variables
o Zero-out the allocated structure and get rid of
  all the unnecessary setting to 0 and NULL;

Refactor _pthread_mutex_destroy
o Simplify the logic by removing a lot of unnecesary nesting
o No need to check pointer that the mutex attributes points
  to. Checking passed in pointer is enough.

20 years agoImplement reference counting of read-write locks. This uses
mtm [Mon, 19 Jan 2004 14:51:45 +0000 (14:51 +0000)]
Implement reference counting of read-write locks. This uses
a list in the thread structure to keep track of the locks and
how many times they have been locked. This list is checked
on every lock and unlock. The traversal through the list is
O(n). Most applications don't hold so many locks at once that
this will become a problem. However, if it does become a problem
it might be a good idea to review this once libthr is
off probation and in the optimization cycle.
This fixes:
o deadlock when a thread tries to recursively acquire a
  read lock when a writer is waiting on the lock.
o a thread could previously successfully unlock a lock it did not own
o deadlock when a thread tries to acquire a write lock on
  a lock it already owns for reading or writing [ this is admittedly
  not required by POSIX, but is nice to have ]

20 years agoRemoved custom "load" and "unload" targets.
ru [Mon, 19 Jan 2004 14:00:26 +0000 (14:00 +0000)]
Removed custom "load" and "unload" targets.

Removed "deinstall" targets -- the idea is to provide the standard
"deinstall" target.

20 years agoUse the standard <bsd.files.mk> API to install files.
ru [Mon, 19 Jan 2004 13:40:51 +0000 (13:40 +0000)]
Use the standard <bsd.files.mk> API to install files.

20 years agoFix "5.x release on 4.x box" support.
matusita [Mon, 19 Jan 2004 13:35:46 +0000 (13:35 +0000)]
Fix "5.x release on 4.x box" support.

It is required to check /etc/rc.d/ldconfig before use, because
${CHROOT} sandbox is sometimes 4.x world.  If /etc/rc.d/ldconfig
is not there, run ldconfig(8) directly.

Discussed with: imp, obrien, ru

20 years agoSync to 1.159 of usbdevs
sanpei [Mon, 19 Jan 2004 12:53:23 +0000 (12:53 +0000)]
Sync to 1.159 of usbdevs

20 years agoAdd support IO-data DVD Multi-plus unit iU-CD2
sanpei [Mon, 19 Jan 2004 12:51:40 +0000 (12:51 +0000)]
Add support IO-data DVD Multi-plus unit iU-CD2

PR: kern/61578
Submitted by: Masaharu FUJITA <m@fjts.org>
MFC after: 1 week

20 years agoadd cross-reference to clock_gettime(2)
phk [Mon, 19 Jan 2004 12:41:39 +0000 (12:41 +0000)]
add cross-reference to clock_gettime(2)

20 years agoNormalize SUBDIR.
ru [Mon, 19 Jan 2004 12:36:08 +0000 (12:36 +0000)]
Normalize SUBDIR.

20 years agoBring the *printf(3) documentation up to date with the code:
das [Mon, 19 Jan 2004 08:28:30 +0000 (08:28 +0000)]
Bring the *printf(3) documentation up to date with the code:

- Update and improve the documentation for %[aA]
  o Like %[eE], %[aA] may round the result if a precision is specified.
  o Grammar police: Fix a split infinitive.
  o The FreeBSD implementation does better than the minimum required
    by C99 (literal translation of the mantissa).  The digit before
    the hexadecimal-point is never 0 unless the number itself is 0.
  o Clarify that the exponent field represents a decimal exponent of 2.
  o Discuss the fact that multiple valid representations are possible.
  o Remove the entry in the BUGS section claiming that %[aA] is not
    implemented.

- Remove the entry in the BUGS section claiming that the ' flag for
  printing thousands separators is unimplemented for floating-point.

- Remove the entry in the BUGS section claiming that the L modifier
  reduces the precision to "double" before conversion.

20 years agoProperly program the multicast filter in ndis_setmulti(),
wpaul [Mon, 19 Jan 2004 07:03:46 +0000 (07:03 +0000)]
Properly program the multicast filter in ndis_setmulti(),
and fix promisc mode in ndis_ioctl().

20 years agoAdd regression tests for printf's %a/%A formats.
das [Mon, 19 Jan 2004 05:59:07 +0000 (05:59 +0000)]
Add regression tests for printf's %a/%A formats.

While here, disable some of the long double tests on i386, since
FreeBSD/i386 is the only port that doesn't evaluate long doubles in
their full precision (due to constant folding bugs in gcc).

20 years agoAdd regression tests for some of the bugs recently discovered in the
das [Mon, 19 Jan 2004 05:30:56 +0000 (05:30 +0000)]
Add regression tests for some of the bugs recently discovered in the
vendor's strtod() implementation.

While here, disable some of the long double tests on i386, since
FreeBSD/i386 is the only port that doesn't evaluate long doubles in
their full precision (due to constant folding bugs in gcc).

20 years agoChange WI_RID_SCAN_RES compatibility interface to return the result after
onoe [Mon, 19 Jan 2004 05:25:43 +0000 (05:25 +0000)]
Change WI_RID_SCAN_RES compatibility interface to return the result after
active scan is completed just as WI_RID_READ_APS.
This fixes wicontrol -L for ath(4) and awi(4) to have results even if
the driver cannot associate any APs.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r124703,
das [Mon, 19 Jan 2004 05:14:12 +0000 (05:14 +0000)]
This commit was generated by cvs2svn to compensate for changes in r124703,
which included commits to RCS files with non-trunk default branches.

20 years agoImport gdtoa 20040118. This revision addresses some corner cases in
das [Mon, 19 Jan 2004 05:14:12 +0000 (05:14 +0000)]
Import gdtoa 20040118.  This revision addresses some corner cases in
denormal and locale handling in strtod().

20 years agoFixed misplacement and bitrot in the cy driver's configuration in
bde [Mon, 19 Jan 2004 02:18:34 +0000 (02:18 +0000)]
Fixed misplacement and bitrot in the cy driver's configuration in
rev.1.1040.  It is a miscellaneous isa+pci driver, but came back
described as a pci-only driver and placed in an i4b pci subsection
after its migration to /sys/conf/NOTES.  Put it back where it used to
be, fully unsorted in the `Miscellaneous hardware' section.  Reduced
nearby disorder in this section by moving configuration of the digi
driver to where it was for the old digiboard drivers, so that the
order at least matches the order in the table of contents.