]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoMake the following name changes to KSE related functions, etc., to better
Archie Cobbs [Wed, 25 Sep 2002 18:10:42 +0000 (18:10 +0000)]
Make the following name changes to KSE related functions, etc., to better
represent their purpose and minimize namespace conflicts:

kse_fn_t -> kse_func_t
struct thread_mailbox -> struct kse_thr_mailbox
thread_interrupt() -> kse_thr_interrupt()
kse_yield() -> kse_release()
kse_new() -> kse_create()

Add missing declaration of kse_thr_interrupt() to <sys/kse.h>.
Regenerate the various generated syscall files. Minor style fixes.

Reviewed by: julian

21 years ago<sys/types.h> is no longer needed.
Mike Barcroft [Wed, 25 Sep 2002 16:49:09 +0000 (16:49 +0000)]
<sys/types.h> is no longer needed.

21 years agoDo away with AAC_COMPAT_LINUX option entirely. The functionality will
Scott Long [Wed, 25 Sep 2002 15:21:50 +0000 (15:21 +0000)]
Do away with AAC_COMPAT_LINUX option entirely.  The functionality will
automatically be enabled if the kernel is compiled with COMPAT_LINUX.

Submitted by: jhb
MFC after: 3 days

21 years agoTry not to lint(1) contrib'ed sources. This Is Very Messy.
Mark Murray [Wed, 25 Sep 2002 14:19:51 +0000 (14:19 +0000)]
Try not to lint(1) contrib'ed sources. This Is Very Messy.

21 years agoFix a declaration that is actually supposed to be a macro definition.
Mark Murray [Wed, 25 Sep 2002 13:46:23 +0000 (13:46 +0000)]
Fix a declaration that is actually supposed to be a macro definition.

Submitted by: marius@alchemy.franken.de

21 years agoRound up instead of towards 0 in clock_getres() so that a resolution of
Bruce Evans [Wed, 25 Sep 2002 12:00:38 +0000 (12:00 +0000)]
Round up instead of towards 0 in clock_getres() so that a resolution of
0 is never returned.

PR: 41781
MFC after: 3 days

21 years agoDo not dump core on 'ipfw add unreach': handling null strings in
Maxim Konovalov [Wed, 25 Sep 2002 11:22:36 +0000 (11:22 +0000)]
Do not dump core on 'ipfw add unreach': handling null strings in
fill_reject_code(). Please note ipfw/ipfw2.c is not affected.

PR: bin/42304
Submitted by: Andy@wantpackets.com
MFC after: 1 day

21 years agoDon't lint contrib'ed sources, even if the builder gas asked for it.
Mark Murray [Wed, 25 Sep 2002 09:59:54 +0000 (09:59 +0000)]
Don't lint contrib'ed sources, even if the builder gas asked for it.
Its Just Too Noisy.

21 years agoDon't lint contrib'ed sources, enven if the builder has asked for it.
Mark Murray [Wed, 25 Sep 2002 09:59:01 +0000 (09:59 +0000)]
Don't lint contrib'ed sources, enven if the builder has asked for it.
Its Just Too Noisy.

21 years agoDon't lint contrib'ed sources, even if the builder has asked for linting.
Mark Murray [Wed, 25 Sep 2002 09:58:00 +0000 (09:58 +0000)]
Don't lint contrib'ed sources, even if the builder has asked for linting.
Its Just Too Noisy.

21 years agoAvoid linting GNU contrib'ed stuff, even if the build engineer asked
Mark Murray [Wed, 25 Sep 2002 09:55:20 +0000 (09:55 +0000)]
Avoid linting GNU contrib'ed stuff, even if the build engineer asked
for it (via WANT_LINT). It's Just Too Noisy.

21 years agoDisqualify UID/GID with non-numeric character.
Maxim Konovalov [Wed, 25 Sep 2002 08:49:19 +0000 (08:49 +0000)]
Disqualify UID/GID with non-numeric character.

PR: bin/41721
Reviewed by: tjr, silence on -audit
MFC after: 2 weeks

21 years agoUnifdef(1) the crypto telnet sources.
Mark Murray [Wed, 25 Sep 2002 07:53:29 +0000 (07:53 +0000)]
Unifdef(1) the crypto telnet sources.

21 years agoCatch up with "base" telnet.
Mark Murray [Wed, 25 Sep 2002 07:28:04 +0000 (07:28 +0000)]
Catch up with "base" telnet.
s/FALL THROUGH/FALLTHROUGH/ for lint(1).

21 years agoCatch up with "base" telnet.
Mark Murray [Wed, 25 Sep 2002 07:26:25 +0000 (07:26 +0000)]
Catch up with "base" telnet.

s/FALL THROUGH/FALLTHROUGH/ for lint(1).
s/Usage/usage/ for consistency.

21 years agoFrom the requestor:
Mark Murray [Wed, 25 Sep 2002 07:24:01 +0000 (07:24 +0000)]
From the requestor:

"Could you do me a favor and fix sys_bsd.c to get the howmany() macro
from <sys/param.h>, instead of <sys/types.h>?  This will save me from
having to worry about the unsync'd bits before making the change."

Requested by: mike

21 years agoThe AAC_COMPAT_LINUX option was really annoying, since it made the
Scott Long [Wed, 25 Sep 2002 05:00:25 +0000 (05:00 +0000)]
The AAC_COMPAT_LINUX option was really annoying, since it made the
aac driver dependent on the linux emulation module.  This was
especially bad for the release engineers who tried to move the
aac driver from the kernel onto the drivers floppy.  The linux
compat bits for this driver are now in their own driver, aac_linux.
It can be loaded as a module or compiled into the kernel.  For
the latter case, the AAC_COMPAT_LINUX option is needed, along with
the COMPAT_LINUX option.

I've tested this in every configuration I can think of.  This is an
MFC candidate for 4.7.

Idea from: rwatson
MFC after: 3 days

21 years agoAccomodate change in the output of "make package-depends" that came
Bruce A. Mah [Wed, 25 Sep 2002 04:36:07 +0000 (04:36 +0000)]
Accomodate change in the output of "make package-depends" that came
with ports/Mk/bsd.ports.mk rev. 1.424.  (It now prints the origin
after the package name.)

21 years agoUse the standardized CHAR_BIT constant instead of NBBY in userland.
Mike Barcroft [Wed, 25 Sep 2002 04:06:37 +0000 (04:06 +0000)]
Use the standardized CHAR_BIT constant instead of NBBY in userland.

21 years agoAdd a pkg_version variable so that it's possible to run portsversion instead
Brian Somers [Wed, 25 Sep 2002 03:01:42 +0000 (03:01 +0000)]
Add a pkg_version variable so that it's possible to run portsversion instead
of pkg_version in periodic/weekly/400.status-pkg.

21 years ago - Remove unused LK_NOOBJ.
Jeff Roberson [Wed, 25 Sep 2002 02:49:50 +0000 (02:49 +0000)]
 - Remove unused LK_NOOBJ.

21 years ago - Convert locks to use standard macros.
Jeff Roberson [Wed, 25 Sep 2002 02:49:48 +0000 (02:49 +0000)]
 - Convert locks to use standard macros.
 - Lock access to the buflists.
 - Document broken locking.
 - Use vrefcnt().

21 years ago - Document broken locking.
Jeff Roberson [Wed, 25 Sep 2002 02:47:49 +0000 (02:47 +0000)]
 - Document broken locking.
 - Use vrefcnt().

21 years ago - Lock accesses to v_usecount.
Jeff Roberson [Wed, 25 Sep 2002 02:45:50 +0000 (02:45 +0000)]
 - Lock accesses to v_usecount.
 - Convert interlock locks to use standard macros.

21 years ago - Don't use the interlock to protect v_writecount.
Jeff Roberson [Wed, 25 Sep 2002 02:44:55 +0000 (02:44 +0000)]
 - Don't use the interlock to protect v_writecount.

21 years ago - Use vrefcnt() instead of v_usecount.
Jeff Roberson [Wed, 25 Sep 2002 02:42:43 +0000 (02:42 +0000)]
 - Use vrefcnt() instead of v_usecount.

21 years ago - Hold the vn lock over vm_mmap().
Jeff Roberson [Wed, 25 Sep 2002 02:42:04 +0000 (02:42 +0000)]
 - Hold the vn lock over vm_mmap().

21 years ago - Use incore() instead of gbincore() so we don't have to acquire the
Jeff Roberson [Wed, 25 Sep 2002 02:39:39 +0000 (02:39 +0000)]
 - Use incore() instead of gbincore() so we don't have to acquire the
   vnode interlock.

21 years ago - Lock access to the buf lists.
Jeff Roberson [Wed, 25 Sep 2002 02:38:43 +0000 (02:38 +0000)]
 - Lock access to the buf lists.
 - Use vrefcnt() where appropriate.
 - Add some locking asserts.

21 years ago - Lock access to the buf lists.
Jeff Roberson [Wed, 25 Sep 2002 02:34:56 +0000 (02:34 +0000)]
 - Lock access to the buf lists.
 - Use vrefcnt() where appropriate.

21 years ago - Use vrefcnt() instead of directly accessing v_usecount.
Jeff Roberson [Wed, 25 Sep 2002 02:33:29 +0000 (02:33 +0000)]
 - Use vrefcnt() instead of directly accessing v_usecount.

21 years ago - Use vrefcnt() where it is safe to do so instead of doing direct and
Jeff Roberson [Wed, 25 Sep 2002 02:32:42 +0000 (02:32 +0000)]
 - Use vrefcnt() where it is safe to do so instead of doing direct and
   unlocked accesses to v_usecount.
 - Lock access to the buf lists in the various sync routines.  interlock
   locking could be avoided almost entirely in leaf filesystems if the
   fsync function had a generic helper.

21 years ago - Lock access to the buf lists in spec_sync()
Jeff Roberson [Wed, 25 Sep 2002 02:29:49 +0000 (02:29 +0000)]
 - Lock access to the buf lists in spec_sync()
 - Fixup interlock locking in spec_close()

21 years ago - Hold the vp lock while accessing v_vflags.
Jeff Roberson [Wed, 25 Sep 2002 02:28:07 +0000 (02:28 +0000)]
 - Hold the vp lock while accessing v_vflags.

21 years ago - Lock down the syncer with sync_mtx.
Jeff Roberson [Wed, 25 Sep 2002 02:22:21 +0000 (02:22 +0000)]
 - Lock down the syncer with sync_mtx.
 - Enable vfs_badlock_mutex by default.
 - Assert that the vp is locked in VOP_UNLOCK.
 - Use standard interlock macros in remaining code.
 - Correct a race in getnewvnode().
 - Lock access to v_numoutput with interlock.
 - Lock access to buf lists and splay tree with interlock.
 - Add VOP and VI asserts.
 - Lock b_vnbufs with the vnode interlock.
 - Add vrefcnt() for callers who want to retreive the vnode ref without
   holding a lock.  Add a comment that describes when this is safe.
 - Add vholdl() and vdropl() so that callers who already own the interlock
   can avoid race conditions and unnecessary unlocking.
 - Move the VOP_GETATTR() in vflush() into the WRITECLOSE conditional case.
 - Hold the interlock before droping the mntlist_mtx in vflush() to avoid
   a race.
 - Fix locking in vfs_msync().

21 years ago - Properly lock v_vflags in getdirents().
Jeff Roberson [Wed, 25 Sep 2002 02:13:38 +0000 (02:13 +0000)]
 - Properly lock v_vflags in getdirents().

21 years ago - Use incore() where no other interlock locking is necessary.
Jeff Roberson [Wed, 25 Sep 2002 02:12:32 +0000 (02:12 +0000)]
 - Use incore() where no other interlock locking is necessary.
 - Lock access to numoutput.

21 years ago - Lock accesses to v_numoutput.
Jeff Roberson [Wed, 25 Sep 2002 02:11:37 +0000 (02:11 +0000)]
 - Lock accesses to v_numoutput.
 - Lock calls to gbincore.

21 years agoBack down to 1003.2-1992 for the time being -- it is causing too many
Garrett Wollman [Wed, 25 Sep 2002 01:58:56 +0000 (01:58 +0000)]
Back down to 1003.2-1992 for the time being -- it is causing too many
headaches for common but deprecated uses of standard utilities.

21 years ago - Don't protect mountedhere with the vn interlock.
Jeff Roberson [Wed, 25 Sep 2002 01:44:21 +0000 (01:44 +0000)]
 - Don't protect mountedhere with the vn interlock.
 - Protect mountedhere with the vn lock.

21 years ago - Use the standard vp interlock macros.
Jeff Roberson [Wed, 25 Sep 2002 01:42:24 +0000 (01:42 +0000)]
 - Use the standard vp interlock macros.

21 years ago - Finish the struct vnode lock annotation.
Jeff Roberson [Wed, 25 Sep 2002 01:37:25 +0000 (01:37 +0000)]
 - Finish the struct vnode lock annotation.
 - Order fields by what lock is required to access them.

21 years ago - Get rid of the unused LK_NOOBJ.
Jeff Roberson [Wed, 25 Sep 2002 01:24:58 +0000 (01:24 +0000)]
 - Get rid of the unused LK_NOOBJ.

21 years ago - Lock access to numoutput on the swap devices.
Jeff Roberson [Wed, 25 Sep 2002 01:24:17 +0000 (01:24 +0000)]
 - Lock access to numoutput on the swap devices.

21 years ago - Add a ASSERT_VOP_LOCKED in vnode_pager_alloc.
Jeff Roberson [Wed, 25 Sep 2002 01:23:43 +0000 (01:23 +0000)]
 - Add a ASSERT_VOP_LOCKED in vnode_pager_alloc.
 - Lock access to v_iflags.

21 years agoRemoved debug code.
Jake Burkholder [Wed, 25 Sep 2002 01:13:46 +0000 (01:13 +0000)]
Removed debug code.

21 years agoPass the function to call (trap or syscall) to tl0_trap and tl1_trap in %o2.
Jake Burkholder [Wed, 25 Sep 2002 00:26:04 +0000 (00:26 +0000)]
Pass the function to call (trap or syscall) to tl0_trap and tl1_trap in %o2.

21 years agoRearrange tl1_trap slightly, also save and restore the out registers so
Jake Burkholder [Tue, 24 Sep 2002 23:22:42 +0000 (23:22 +0000)]
Rearrange tl1_trap slightly, also save and restore the out registers so
that instruction emulation is possible in kernel mode.

21 years agoAdd some PCI id's for fxp. I've only tested 0x103B, but the Intel
Paul Saab [Tue, 24 Sep 2002 23:13:38 +0000 (23:13 +0000)]
Add some PCI id's for fxp.  I've only tested 0x103B, but the Intel
Linux driver defines 0x103[B-E] so add those as well.

Obtained from: Intel Linux e100 driver
MFC: Immediately if re@ allows it, otherwise after 4.7-RELEASE

21 years agoGet howmany() macro from <sys/param.h>, instead of depending on its
Mike Barcroft [Tue, 24 Sep 2002 22:30:56 +0000 (22:30 +0000)]
Get howmany() macro from <sys/param.h>, instead of depending on its
existence in <sys/types.h>.

21 years agoAllocate stack space for the trapframe along with the normal register
Jake Burkholder [Tue, 24 Sep 2002 22:19:20 +0000 (22:19 +0000)]
Allocate stack space for the trapframe along with the normal register
frame in the save instruction, rather than doing a separate sub.

21 years agoCode cleanup: use mpt_prt instead of device_printf.
Matt Jacob [Tue, 24 Sep 2002 21:33:43 +0000 (21:33 +0000)]
Code cleanup: use mpt_prt instead of device_printf.

21 years agoRemove all mention of MTX_NOSWITCH. It was obsoleted a long time ago.
John Baldwin [Tue, 24 Sep 2002 20:45:58 +0000 (20:45 +0000)]
Remove all mention of MTX_NOSWITCH.  It was obsoleted a long time ago.

21 years agoUse POSIX-style sort syntax in example
Andrey A. Chernov [Tue, 24 Sep 2002 20:15:11 +0000 (20:15 +0000)]
Use POSIX-style sort syntax in example

21 years agoMake it works with POSIX sort (POS arg)
Andrey A. Chernov [Tue, 24 Sep 2002 20:10:59 +0000 (20:10 +0000)]
Make it works with POSIX sort (POS arg)
All old sorts understand -k too.

21 years agoMake it works with POSIX sort (POS arg)
Andrey A. Chernov [Tue, 24 Sep 2002 20:05:01 +0000 (20:05 +0000)]
Make it works with POSIX sort (POS arg)
All old sorts understand -k too.

21 years agoSet _POSIX2_VERSION=199209 before calling MKkey_defs.sh as workaround of
Andrey A. Chernov [Tue, 24 Sep 2002 19:55:37 +0000 (19:55 +0000)]
Set _POSIX2_VERSION=199209 before calling MKkey_defs.sh as workaround of
obsolete +POS sort syntax.

21 years agoThe spec says that unifdefall removes all #ifs, and this now requires
Tony Finch [Tue, 24 Sep 2002 19:50:03 +0000 (19:50 +0000)]
The spec says that unifdefall removes all #ifs, and this now requires
the -k flag to unifdef.

21 years agoThe readability is improved by reducing the use of the passive voice.
Tony Finch [Tue, 24 Sep 2002 19:48:39 +0000 (19:48 +0000)]
The readability is improved by reducing the use of the passive voice.
No changes to the meaning are made.

21 years agoIt is probably a bad idea to unconditionally process directives with
Tony Finch [Tue, 24 Sep 2002 19:27:44 +0000 (19:27 +0000)]
It is probably a bad idea to unconditionally process directives with
constant controlling expressions: in particular, removing #if 0 sections
is considered "rude". This commit changes the default so that such
things are passed through unchanged, and the old behaviour can be had
with the -k "kill konsts" flag.

Suggested by: markm
MFC after: 3 weeks

21 years agoMake it work with POSIX sort (POS arg).
Andrey A. Chernov [Tue, 24 Sep 2002 19:05:40 +0000 (19:05 +0000)]
Make it work with POSIX sort (POS arg).
All old sorts understand -k too.

21 years agoCorrect the description of what fortunes get installed (by default, also the
Jens Schweikhardt [Tue, 24 Sep 2002 19:00:52 +0000 (19:00 +0000)]
Correct the description of what fortunes get installed (by default, also the
offensive ones). Make it clear that datfiles/Makefile needs to be edited in
case someone wants it otherwise.

PR: misc/36153
Submitted by: Nick Jennings <nick@namodn.com>
X-MFC after: 4.7 is out the door.

21 years agoMake it work with POSIX sort (POS arg).
Andrey A. Chernov [Tue, 24 Sep 2002 18:53:46 +0000 (18:53 +0000)]
Make it work with POSIX sort (POS arg).
All old sorts understand -k too.

21 years agoMake it work with POSIX sort (POS arg)
Andrey A. Chernov [Tue, 24 Sep 2002 18:16:36 +0000 (18:16 +0000)]
Make it work with POSIX sort (POS arg)
All old sorts understand -k too.

21 years agoConvert most printf()s to if_printf()s.
Brooks Davis [Tue, 24 Sep 2002 17:37:32 +0000 (17:37 +0000)]
Convert most printf()s to if_printf()s.

21 years agoAdd a new helper function if_printf() modeled on device_printf(). The
Brooks Davis [Tue, 24 Sep 2002 17:35:08 +0000 (17:35 +0000)]
Add a new helper function if_printf() modeled on device_printf().  The
function takes a struct ifnet pointer followed by the usual printf
arguments and prints "<interfacename>: " before the results of printf.
Since this is the primary form of printf calls in network device drivers
and accounts for most uses of the ifnet menber if_unit, this
significantly simplifies many printf()s.

21 years agoAdd forgotten <sys/_types.h> include. Fix some style bugs while I'm
Mike Barcroft [Tue, 24 Sep 2002 17:28:12 +0000 (17:28 +0000)]
Add forgotten <sys/_types.h> include.  Fix some style bugs while I'm
here.

PR: 43326

21 years agoMention about how to build sendmail with SASLv2.
Hajimu UMEMOTO [Tue, 24 Sep 2002 17:07:37 +0000 (17:07 +0000)]
Mention about how to build sendmail with SASLv2.

21 years agoSplit user trap processing out into a separate routine so that traps which
Jake Burkholder [Tue, 24 Sep 2002 16:42:18 +0000 (16:42 +0000)]
Split user trap processing out into a separate routine so that traps which
never result in user traps don't have to plow through it.

21 years agoCorrected license in the source files. It should say "MUST" instead of "MAY".
Prafulla Deuskar [Tue, 24 Sep 2002 16:27:59 +0000 (16:27 +0000)]
Corrected license in the source files. It should say "MUST" instead of "MAY".

MFC after: 2 days

21 years agoFix some bitrotted documentation.
Mark Murray [Tue, 24 Sep 2002 13:34:36 +0000 (13:34 +0000)]
Fix some bitrotted documentation.

21 years agoWarn when setinvalidrune() is referenced for consistency with the rest
Tim J. Robbins [Tue, 24 Sep 2002 09:25:37 +0000 (09:25 +0000)]
Warn when setinvalidrune() is referenced for consistency with the rest
of the rune functions (except sgetrune() and sputrune(), which are really
macros).

21 years agoAdd cross-references between wide character and single-byte character
Tim J. Robbins [Tue, 24 Sep 2002 09:22:47 +0000 (09:22 +0000)]
Add cross-references between wide character and single-byte character
versions of printf() and scanf().

21 years agoRemove an unneeded call to _sfrefill() that was missed in the conversion
Tim J. Robbins [Tue, 24 Sep 2002 09:18:32 +0000 (09:18 +0000)]
Remove an unneeded call to _sfrefill() that was missed in the conversion
from vfscanf() to vfwscanf(). It doesn't hurt to have it there, but it's
redundant since __fgetwc() will refill the buffer if it needs to.

21 years agoBack out last commit. Linux uses the old 4.3BSD sockaddr format.
Jonathan Mini [Tue, 24 Sep 2002 07:03:01 +0000 (07:03 +0000)]
Back out last commit.  Linux uses the old 4.3BSD sockaddr format.

21 years agoAdd support to fsck_ffs to account for storage for extended
Kirk McKusick [Tue, 24 Sep 2002 05:18:17 +0000 (05:18 +0000)]
Add support to fsck_ffs to account for storage for extended
attributes.

Sponsored by: DARPA & NAI Labs.

21 years agoMFi386: revisions 1.127, 1.128 and 1.129.
Yoshihiro Takahashi [Tue, 24 Sep 2002 03:17:38 +0000 (03:17 +0000)]
MFi386: revisions 1.127, 1.128 and 1.129.

21 years agoOLDCARD -> NEWCARD
Warner Losh [Tue, 24 Sep 2002 02:55:33 +0000 (02:55 +0000)]
OLDCARD -> NEWCARD
pccbb -> cbb

PR: 43263

21 years agoTwo new majors. devd and watchdog
Warner Losh [Tue, 24 Sep 2002 02:46:46 +0000 (02:46 +0000)]
Two new majors.  devd and watchdog

21 years agoMove the com_emr register definition to sioreg.h.
Yoshihiro Takahashi [Tue, 24 Sep 2002 02:35:57 +0000 (02:35 +0000)]
Move the com_emr register definition to sioreg.h.

21 years agoMerged from sys/dev/sio/sio.c revision 1.379.
Yoshihiro Takahashi [Tue, 24 Sep 2002 02:26:55 +0000 (02:26 +0000)]
Merged from sys/dev/sio/sio.c revision 1.379.

21 years agoMFi386: revisions 1.536 and 1.538.
Yoshihiro Takahashi [Tue, 24 Sep 2002 02:22:35 +0000 (02:22 +0000)]
MFi386: revisions 1.536 and 1.538.

21 years agoMFi386: revision 1.419
Yoshihiro Takahashi [Tue, 24 Sep 2002 02:18:41 +0000 (02:18 +0000)]
MFi386: revision 1.419

21 years agoFixed to build after removing a.out suppot.
Yoshihiro Takahashi [Tue, 24 Sep 2002 02:17:13 +0000 (02:17 +0000)]
Fixed to build after removing a.out suppot.

21 years agoUse the new va_copy macro to copy variable argument lists instead of
Tim J. Robbins [Tue, 24 Sep 2002 00:47:27 +0000 (00:47 +0000)]
Use the new va_copy macro to copy variable argument lists instead of
assignment. This is needed on powerpc but is also more correct for the
other ports.

Submitted by: grehan
Tested on: alpha, i386, sparc64

21 years agoIf the peer gives us 0.0.0.0 as his IP number, NAK it rather than accepting
Brian Somers [Mon, 23 Sep 2002 22:40:43 +0000 (22:40 +0000)]
If the peer gives us 0.0.0.0 as his IP number, NAK it rather than accepting
it as being in range.

  set ifaddr 1.2.3.4/0 5.6.7.8/0

no longer allows 0.0.0.0 as a valid IP.

Reported/tested by: Bohdan Horst <nexus@hoth.amu.edu.pl>
MFC after: 3 days

21 years agoAdd an NBBY define to eventually replace the one in <sys/types.h>.
Mike Barcroft [Mon, 23 Sep 2002 22:01:10 +0000 (22:01 +0000)]
Add an NBBY define to eventually replace the one in <sys/types.h>.

21 years agoOk, make this compile for real this time. recvfrom_args doesn't have a
John Baldwin [Mon, 23 Sep 2002 21:05:41 +0000 (21:05 +0000)]
Ok, make this compile for real this time.  recvfrom_args doesn't have a
fromlen member, instead it has a fromlenaddr pointer member.  Set it to
NULL.

21 years agoUse correct variable name so that previous commit actually compiles.
John Baldwin [Mon, 23 Sep 2002 20:58:59 +0000 (20:58 +0000)]
Use correct variable name so that previous commit actually compiles.

21 years agoParameterize MPT_MAX_REQUESTS based upon device type (FC has Global Credits
Matt Jacob [Mon, 23 Sep 2002 19:41:10 +0000 (19:41 +0000)]
Parameterize MPT_MAX_REQUESTS based upon device type (FC has Global Credits
of 1024- Ultra4 256). Rename 'requests' tag to 'request_pool' for clarity.
Make sure we do correct xpt_freeze_simq/CAM_RELEASE_SIMQ if we run out
of chip resources.

MFC after: 6 days

21 years agouse __packed.
Alfred Perlstein [Mon, 23 Sep 2002 18:54:32 +0000 (18:54 +0000)]
use __packed.

21 years agoNow that we only probe host-PCI bridges once, we no longer have to check to
John Baldwin [Mon, 23 Sep 2002 18:14:31 +0000 (18:14 +0000)]
Now that we only probe host-PCI bridges once, we no longer have to check to
see if we have been probed before by checking for a pciX bus device.

21 years agoPut verbose printf's in the PCI BIOS interrupt routing code under
John Baldwin [Mon, 23 Sep 2002 18:13:42 +0000 (18:13 +0000)]
Put verbose printf's in the PCI BIOS interrupt routing code under
if (bootverbose).

21 years agoo Move select() helper macros from <sys/types.h> to <sys/select.h>.
Mike Barcroft [Mon, 23 Sep 2002 17:45:51 +0000 (17:45 +0000)]
o Move select() helper macros from <sys/types.h> to <sys/select.h>.
o Include <sys/select.h> from <sys/types.h> in the __BSD_VISIBLE case,
  so applications and base software can be slowly updated.
o Prototype select() in <sys/select.h>.  It was previously only
  prototyped in <unistd.h>.
o Add some XXX's to <sys/types.h>.

Reviewed by: -standards

21 years agoUpdate the nexus driver for the addition of the legacy driver:
John Baldwin [Mon, 23 Sep 2002 16:03:00 +0000 (16:03 +0000)]
Update the nexus driver for the addition of the legacy driver:
- nexus no longer has PCI bridges as direct children, so the PCI bus
  ivar is no longer used and is removed.
- Don't attach default EISA, ISA, or MCA busses.  Instead, if we do not
  have an acpi0 device after bus_generic_probe(), add a legacy0 child
  device.
- Remove machine/nexusvar.h.

21 years agoHang apm off of the legacy device instead of the nexus.
John Baldwin [Mon, 23 Sep 2002 15:56:09 +0000 (15:56 +0000)]
Hang apm off of the legacy device instead of the nexus.

21 years agoChange the nexus_pcib driver (eventually to be renamed to legacy_pcib) to
John Baldwin [Mon, 23 Sep 2002 15:52:30 +0000 (15:52 +0000)]
Change the nexus_pcib driver (eventually to be renamed to legacy_pcib) to
hang off of the legacy driver instead of the nexus.

21 years agoAdd a new legacy(4) device driver for use on machines that do not have
John Baldwin [Mon, 23 Sep 2002 15:50:06 +0000 (15:50 +0000)]
Add a new legacy(4) device driver for use on machines that do not have
ACPI or for when ACPI support is disabled or not present in the kernel.
Basically, the nexus device is now split into two with some parts
(such as adding default ISA, MCA, and EISA busses if they aren't found
as well as support for PCI bus device ivars) being moved to the legacy
driver.

21 years agoDo not add a start offset time to the track duration time.
Maxim Konovalov [Mon, 23 Sep 2002 15:00:23 +0000 (15:00 +0000)]
Do not add a start offset time to the track duration time.

Submitted by: Dmitry Afanasiev, <KOT@MATPOCKuH.SPb.Ru>
MFC after: 2 weeks

21 years agoALLLANG is long dead, garbage collect it. Instead, make it possible
Ruslan Ermilov [Mon, 23 Sep 2002 14:33:47 +0000 (14:33 +0000)]
ALLLANG is long dead, garbage collect it.  Instead, make it possible
to use DOC_LANG to specify which languages the doc/ is built for.

Note: be aware that the DOC_LANG setting in /etc/make.conf will now
take the effect on "make release".  (This is probably the desired
behavior anyway.)

PR: docs/42924