]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoRemove GCC-specific flags and commented out cruft.
markm [Sun, 28 Apr 2002 12:14:10 +0000 (12:14 +0000)]
Remove GCC-specific flags and commented out cruft.

22 years agoSome easy const fixes.
markm [Sun, 28 Apr 2002 12:12:49 +0000 (12:12 +0000)]
Some easy const fixes.

22 years agoANSIfy, VCS inclusions and some very minor style.
markm [Sun, 28 Apr 2002 12:09:33 +0000 (12:09 +0000)]
ANSIfy, VCS inclusions and some very minor style.

22 years agoMinor reorder for style.
markm [Sun, 28 Apr 2002 12:08:18 +0000 (12:08 +0000)]
Minor reorder for style.

22 years agoFix some easy warnings; ANSIfy, constify.
markm [Sun, 28 Apr 2002 12:04:20 +0000 (12:04 +0000)]
Fix some easy warnings; ANSIfy, constify.
Fix VCS inclusion and deregister.

22 years agoANSIfy functions, and move extern declarations into a shared header
markm [Sun, 28 Apr 2002 11:59:43 +0000 (11:59 +0000)]
ANSIfy functions, and move extern declarations into a shared header
where ${DEITY} intended them to be.

22 years agoFix a number of warnings. A gcc preprocessor bug relating to macros
iedowse [Sun, 28 Apr 2002 11:59:30 +0000 (11:59 +0000)]
Fix a number of warnings. A gcc preprocessor bug relating to macros
with variable numbers of arguments made this slightly harder than
it should be. Avoid the bug by not doing string concatenation within
the macros, and instead add a new function to syslog or print the
error messages.

22 years agoAdd the Dynalink IS64PPH+ as a supported card.
gj [Sun, 28 Apr 2002 11:52:53 +0000 (11:52 +0000)]
Add the Dynalink IS64PPH+ as a supported card.

Reported by: znerd

22 years agoFix VCS inclusions.
markm [Sun, 28 Apr 2002 11:52:43 +0000 (11:52 +0000)]
Fix VCS inclusions.

22 years agoMore low hanging warnings fruit; ANSIfy, de-register, constify.
markm [Sun, 28 Apr 2002 11:51:03 +0000 (11:51 +0000)]
More low hanging warnings fruit; ANSIfy, de-register, constify.

22 years agoChange instances of avma1pp2- to ifpi2- in printf's since the name
gj [Sun, 28 Apr 2002 11:47:10 +0000 (11:47 +0000)]
Change instances of avma1pp2- to ifpi2- in printf's since the name
of the driver should be emitted.

This was already changed in the code committed to RELENG_4.

22 years agoFix some low-hanging fruit in the warnings department; ANSIfy functions,
markm [Sun, 28 Apr 2002 11:44:30 +0000 (11:44 +0000)]
Fix some low-hanging fruit in the warnings department; ANSIfy functions,
sort out some const issues.

22 years agoEasy warns fixes; use const, mark unused function parameters. Remove
markm [Sun, 28 Apr 2002 11:16:43 +0000 (11:16 +0000)]
Easy warns fixes; use const, mark unused function parameters. Remove
unused includes.

22 years agoRemove GCC-specific flags.
markm [Sun, 28 Apr 2002 11:15:47 +0000 (11:15 +0000)]
Remove GCC-specific flags.

22 years agoTrivial warnings fixes; staticise and use unsigned.
markm [Sun, 28 Apr 2002 11:13:20 +0000 (11:13 +0000)]
Trivial warnings fixes; staticise and use unsigned.

Use __FBSDID().

22 years agoUse __FBSDID, ANSIfy functions, fix const string.
markm [Sun, 28 Apr 2002 11:00:34 +0000 (11:00 +0000)]
Use __FBSDID, ANSIfy functions, fix const string.

22 years agoUse __FBSDID(), ANSIfy headings, use const/unsigned in minor but
markm [Sun, 28 Apr 2002 10:56:09 +0000 (10:56 +0000)]
Use __FBSDID(), ANSIfy headings, use const/unsigned in minor but
appropriate places.

22 years agoStyle only; format function declarations, sort header includes, and
markm [Sun, 28 Apr 2002 10:49:15 +0000 (10:49 +0000)]
Style only; format function declarations, sort header includes, and
use __FBSDID().

22 years agoOops, remove references to NLOCKED and NWANTED, now that they no
iedowse [Sun, 28 Apr 2002 10:24:38 +0000 (10:24 +0000)]
Oops, remove references to NLOCKED and NWANTED, now that they no
longer exist.

22 years agoRemoved unused forward struct declaration.
bde [Sun, 28 Apr 2002 09:51:45 +0000 (09:51 +0000)]
Removed unused forward struct declaration.

22 years agoAdd mca.c.
marcel [Sun, 28 Apr 2002 08:43:47 +0000 (08:43 +0000)]
Add mca.c.

22 years agoNuke -g
scottl [Sun, 28 Apr 2002 06:38:38 +0000 (06:38 +0000)]
Nuke -g

22 years ago o Introduce and use vm_map_trylock() to replace several direct uses
alc [Sun, 28 Apr 2002 06:07:54 +0000 (06:07 +0000)]
 o Introduce and use vm_map_trylock() to replace several direct uses
   of lockmgr().
 o Add missing synchronization to vmspace_swap_count(): Obtain a read lock
   on the vm_map before traversing it.

22 years agoHook the DRM up to the build and add it to NOTES.
anholt [Sun, 28 Apr 2002 04:58:40 +0000 (04:58 +0000)]
Hook the DRM up to the build and add it to NOTES.

Approved by: des

22 years agoAdd DFE-530TX+ to rl(4) list.
bmah [Sun, 28 Apr 2002 04:49:19 +0000 (04:49 +0000)]
Add DFE-530TX+ to rl(4) list.

Submitted by: Karl Heller <kheller2@mac.com>

22 years agoAdd code to emulate arithmetic, comparison and conversion operations
jake [Sun, 28 Apr 2002 02:36:36 +0000 (02:36 +0000)]
Add code to emulate arithmetic, comparison and conversion operations
on long double, which are not implemented in hardware on any UltraSPARC
chip that I know of.  This just calls into the existing floating point
emulator, which is still needed to emulate other floating point operations
in certain conditions.  Without this gcc has to generate the quad floating
point instructions directly, which sometimes causes internal compiler
errors.

Reviewed by: tmm

22 years agoAlphabetize descriptions and remove the "PCI" from the desciptions.
imp [Sun, 28 Apr 2002 02:04:28 +0000 (02:04 +0000)]
Alphabetize descriptions and remove the "PCI" from the desciptions.

Suggested by: brooks

22 years agoAdd partition type for IBM's JFS.
grog [Sun, 28 Apr 2002 01:52:17 +0000 (01:52 +0000)]
Add partition type for IBM's JFS.

Submitted by: Hiten Pandya <hiten@uk.FreeBSD.org>

22 years agoAdd many prototypes and ANSI-ize 'patch. Not a complete job of it, but it gets
gad [Sun, 28 Apr 2002 01:33:45 +0000 (01:33 +0000)]
Add many prototypes and ANSI-ize 'patch.  Not a complete job of it, but it gets
our patch to look a little more like NetBSD's, and has the nice characteristic
that the object code is exactly the same after the change as before it (even in
patch.c and pch.c, which have pesky 'assert' statements in them).

Reviewed by: /sbin/md5 on i386, alpha, sparc64
MFC after: 3 days

22 years agoAdd FreeBSD id's to these files, in advance of a separate update.
gad [Sun, 28 Apr 2002 01:26:40 +0000 (01:26 +0000)]
Add FreeBSD id's to these files, in advance of a separate update.

MFC after: 3 days

22 years agoOops. Previous commit was to fix the problem which was noticed by tmm.
peter [Sun, 28 Apr 2002 00:17:26 +0000 (00:17 +0000)]
Oops. Previous commit was to fix the problem which was noticed by tmm.

22 years agoWe do not necessarily need to map/unmap pages to zero parts of them.
peter [Sun, 28 Apr 2002 00:15:48 +0000 (00:15 +0000)]
We do not necessarily need to map/unmap pages to zero parts of them.
On systems where physical memory is also direct mapped (alpha, sparc,
ia64 etc) this is slightly harmful.

22 years agoNew release notes: isdn4bsd Q.931 subaddressing.
bmah [Sun, 28 Apr 2002 00:09:42 +0000 (00:09 +0000)]
New release notes:  isdn4bsd Q.931 subaddressing.

Updated release notes:  IPFilter 3.4.27 (+ MFC), isdn4bsd 1.0.2.

Fix up some confusion about ifpi(4) and ifpi2(4).

22 years agoAdd devices supported by aue(4), cue(4), kue(4), an(4), wi(4)
bmah [Sun, 28 Apr 2002 00:01:12 +0000 (00:01 +0000)]
Add devices supported by aue(4), cue(4), kue(4), an(4), wi(4)
drivers.  Minor fix for ray(4).

Based on:

src/etc/defaults/pccard.conf 1.236
src/sys/dev/usb/if_aue.c 1.57
src/sys/dev/usb/if_cue.c 1.28
src/sys/dev/usb/if_kue.c 1.40
src/sys/dev/usb/usbdevs 1.88

22 years agoAvoid the user-visible effect of setting SA_NOCLDWAIT when the
iedowse [Sat, 27 Apr 2002 22:41:41 +0000 (22:41 +0000)]
Avoid the user-visible effect of setting SA_NOCLDWAIT when the
SIGCHLD handler is SIG_IGN. This is a reimplementation of the
problematic revision 1.131 of kern_exit.c. To avoid accessing process
UPAGES, we set a new procsig flag when the SIGCHLD handler is SIG_IGN
and use that instead.

22 years agoFinish fixing hints. Remember the use_kenv state for the next run.
peter [Sat, 27 Apr 2002 22:32:57 +0000 (22:32 +0000)]
Finish fixing hints.  Remember the use_kenv state for the next run.
Otherwise we fall back to using the static hints the next time around.
We still have the leftover fallback code there which meant that we skipped
the use_hints checking on the second and subsequent calls.  Also, be a bit
more careful about walking off the end of the envp array.

I've extracted this from a larger diff.  I hope I didn't miss anything...

22 years agoPartial fix for hints
peter [Sat, 27 Apr 2002 22:25:13 +0000 (22:25 +0000)]
Partial fix for hints

Obtained from:  mux

22 years agoRemove a stale comment saying that the vnode lock must be the first
iedowse [Sat, 27 Apr 2002 22:20:33 +0000 (22:20 +0000)]
Remove a stale comment saying that the vnode lock must be the first
element in the structure pointed to by vp->v_data; the vnode lock
is now within the vnode structure itself.

22 years agoRemove the nfs_{lock,unlock,islocked} functions and the associated
iedowse [Sat, 27 Apr 2002 22:10:16 +0000 (22:10 +0000)]
Remove the nfs_{lock,unlock,islocked} functions and the associated
definitions; they have been unused and #if 0'd out since the Lite/2
merge and we are unlikely to want them in the future.

22 years ago o Begin documenting the (existing) locking protocol on the vm_map
alc [Sat, 27 Apr 2002 22:01:37 +0000 (22:01 +0000)]
 o Begin documenting the (existing) locking protocol on the vm_map
   in the same style as sys/proc.h.
 o Undo the de-inlining of several trivial, MPSAFE methods on the vm_map.
   (Contrary to the commit message for vm_map.h revision 1.66 and vm_map.c
   revision 1.206, de-inlining these methods increased the kernel's size.)

22 years agoObliterate the new password as well as the old, even if though it's hashed.
des [Sat, 27 Apr 2002 21:58:23 +0000 (21:58 +0000)]
Obliterate the new password as well as the old, even if though it's hashed.

22 years agoEmulate ldq and stq (load/store long double) instructions. GCC has started
jake [Sat, 27 Apr 2002 21:56:28 +0000 (21:56 +0000)]
Emulate ldq and stq (load/store long double) instructions.  GCC has started
using these to load long doubles, but they aren't implemented in hardware
on (at least) UltraSPARC I and II machines.
Emulate popc in the user trap handler as well.
Re-arrange slightly to make support functions more accessible.

Reviewed by: tmm

22 years agoAdd makefiles for DRM modules
anholt [Sat, 27 Apr 2002 20:55:03 +0000 (20:55 +0000)]
Add makefiles for DRM modules

Approved by: des

22 years agoAdd the code for the DRM, based on the code from the drm-kmod port.
anholt [Sat, 27 Apr 2002 20:47:57 +0000 (20:47 +0000)]
Add the code for the DRM, based on the code from the drm-kmod port.
This is not hooked up yet, that will come later.

Approved by: des

22 years agoMove ID tag.
eric [Sat, 27 Apr 2002 19:57:22 +0000 (19:57 +0000)]
Move ID tag.

22 years agoFor what it's worth, fix the compilation of an I386_CPU-only kernel
alc [Sat, 27 Apr 2002 18:13:35 +0000 (18:13 +0000)]
For what it's worth, fix the compilation of an I386_CPU-only kernel
now that certain warnings are fatal.

22 years agoDon't call vm_map_growstack() from trapwrite() as vm_fault() now performs
alc [Sat, 27 Apr 2002 17:07:15 +0000 (17:07 +0000)]
Don't call vm_map_growstack() from trapwrite() as vm_fault() now performs
this automatically.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r95567,
darrenr [Sat, 27 Apr 2002 17:01:32 +0000 (17:01 +0000)]
This commit was generated by cvs2svn to compensate for changes in r95567,
which included commits to RCS files with non-trunk default branches.

22 years agoImport 3.4.27
darrenr [Sat, 27 Apr 2002 17:01:32 +0000 (17:01 +0000)]
Import 3.4.27

22 years agoThis commit was generated by cvs2svn to compensate for changes in r95565,
darrenr [Sat, 27 Apr 2002 17:01:31 +0000 (17:01 +0000)]
This commit was generated by cvs2svn to compensate for changes in r95565,
which included commits to RCS files with non-trunk default branches.

22 years agoImport 3.4.27
darrenr [Sat, 27 Apr 2002 17:01:31 +0000 (17:01 +0000)]
Import 3.4.27

22 years agoMFi386 1.222: Remove vm_map_growstack() and acquisition and release of Giant
alc [Sat, 27 Apr 2002 17:00:28 +0000 (17:00 +0000)]
MFi386 1.222: Remove vm_map_growstack() and acquisition and release of Giant
around vm_fault() in trap_pfault().

22 years agoMerge updates from 3.4.26 - 3.4.27.
darrenr [Sat, 27 Apr 2002 16:56:25 +0000 (16:56 +0000)]
Merge updates from 3.4.26 - 3.4.27.

22 years agoThis patch fixes my breakage of ssid matching. I introduced the
imp [Sat, 27 Apr 2002 16:03:25 +0000 (16:03 +0000)]
This patch fixes my breakage of ssid matching.  I introduced the
breakage when I tried to merge OpenBSD wi_hostap changes into the
tree.  Skibo found the problem and submitted these patches.  Thanks!

Submitted by: skibo@pacbell.net

22 years agoBetter names for the PCI cards. The biggest change is that we now
imp [Sat, 27 Apr 2002 16:01:26 +0000 (16:01 +0000)]
Better names for the PCI cards.  The biggest change is that we now
identify the Intersil Prism 2.5 PCI native card as that, rather than
Linksys, the first folks to get it to market.

22 years agoIntroduce CVSCMDARGS make variable to set command-line options for cvs.
matusita [Sat, 27 Apr 2002 14:23:32 +0000 (14:23 +0000)]
Introduce CVSCMDARGS make variable to set command-line options for cvs.
You may want set '-D data-spec' to this variable.

PR: 31218
Submitted by: Alexandr Listopad <laa@laa.zp.ua>
MFC after: 5 days
(if re@ permits)

22 years agoMFi386 1.222: Remove vm_map_growstack() and acquisition and release of Giant
alc [Sat, 27 Apr 2002 09:34:15 +0000 (09:34 +0000)]
MFi386 1.222: Remove vm_map_growstack() and acquisition and release of Giant
around vm_fault().

22 years agoFix the code fragment clobbered in my last commit.
tanimura [Sat, 27 Apr 2002 09:33:49 +0000 (09:33 +0000)]
Fix the code fragment clobbered in my last commit.

22 years agoAdd a global sx sigio_lock to protect the pointer to the sigio object
tanimura [Sat, 27 Apr 2002 08:24:29 +0000 (08:24 +0000)]
Add a global sx sigio_lock to protect the pointer to the sigio object
of a socket.  This avoids lock order reversal caused by locking a
process in pgsigio().

sowakeup() and the callers of it (sowwakeup, soisconnected, etc.) now
require sigio_lock to be locked.  Provide sowwakeup_locked(),
soisconnected_locked(), and so on in case where we have to modify a
socket and wake up a process atomically.

22 years agoExplain magic number.
phk [Sat, 27 Apr 2002 07:28:54 +0000 (07:28 +0000)]
Explain magic number.

Add magic date no explanation.

Add a delta which was lost in transit yesterday which prevented
other timecounters from actually being used.

22 years agoFix a {} bug which doesn't have any effect yet.
phk [Sat, 27 Apr 2002 07:07:37 +0000 (07:07 +0000)]
Fix a {} bug which doesn't have any effect yet.

Spotted by: jake

22 years agoMake the dummy timecounter actually tick or we will never get anyhere.
phk [Sat, 27 Apr 2002 07:06:52 +0000 (07:06 +0000)]
Make the dummy timecounter actually tick or we will never get anyhere.

22 years agoRemove reference to the TCP_RESTRICT_RST option, which was removed
dougb [Sat, 27 Apr 2002 06:24:58 +0000 (06:24 +0000)]
Remove reference to the TCP_RESTRICT_RST option, which was removed
over a year ago.

Small ws twiddle while I'm here.

22 years agoKill one exact duplicate of a Hunter S. Thompson quote. Put on peril-sensitive
jmallett [Sat, 27 Apr 2002 06:14:32 +0000 (06:14 +0000)]
Kill one exact duplicate of a Hunter S. Thompson quote.  Put on peril-sensitive
sunglasses for great evil, too.

22 years agoDeath to duplicate fortunes.
jmallett [Sat, 27 Apr 2002 05:46:13 +0000 (05:46 +0000)]
Death to duplicate fortunes.

22 years agoFix handling of weak references to undefined symbols on ia64:
marcel [Sat, 27 Apr 2002 05:32:51 +0000 (05:32 +0000)]
Fix handling of weak references to undefined symbols on ia64:
o  Set st_shndx for sym_zero to SHN_UNDEF instead of SHN_ABS.
   This gives us something to reliably test against.
o  For weak references to undefined sysmbols (as indicated by
   having st_shndx equals SHN_UNDEF) in the context of OPDs,
   the address of the OPD is to be zero, not the address of
   the function it contains.
o  For weak references to undefined symbols in all other cases
   (only DIR64LSB at this time), the actual relocated value is
   to be zero, not the value prior to relocating.

Roughly speaking, weak references to undefined symbols are no-ops.

Tested on: i386, ia64

22 years agoWhitespace bogon.
jhb [Sat, 27 Apr 2002 04:48:36 +0000 (04:48 +0000)]
Whitespace bogon.

22 years agoRedirect stdin from the input file, rather than passing the input file to
jmallett [Sat, 27 Apr 2002 03:11:33 +0000 (03:11 +0000)]
Redirect stdin from the input file, rather than passing the input file to
uuencode(1), and set a umask, so that the mode in the header is predictable.

If it varies, then the test is right to fail.

Remove the note about this test falsely failing, with that in mind.

22 years agoInsert a semi-colon between label 'skip:' and the closing brace
marcel [Sat, 27 Apr 2002 02:58:18 +0000 (02:58 +0000)]
Insert a semi-colon between label 'skip:' and the closing brace
of the FOREACH loop to silence GCC 3.

22 years agoNow that local symbols aren't looked up with the symbol hash table,
marcel [Sat, 27 Apr 2002 02:53:31 +0000 (02:53 +0000)]
Now that local symbols aren't looked up with the symbol hash table,
binding works for local symbols. Remove the workaround...

22 years agoDon't do symbol lookups for local symbols. The symbol index in the
marcel [Sat, 27 Apr 2002 02:48:29 +0000 (02:48 +0000)]
Don't do symbol lookups for local symbols. The symbol index in the
relocation identifies the symbol to which we need to bind. This
solves a problem seen on ia64 where the symbol hash table does not
contain local symbols and thus resulted in unresolved symbols.

Tested on: alpha, i386, ia64

22 years agoPrefix tests with PASS and FAIL, to make grepping easier, and note this in
jmallett [Sat, 27 Apr 2002 02:26:23 +0000 (02:26 +0000)]
Prefix tests with PASS and FAIL, to make grepping easier, and note this in
the README.

This affects only the base-system regression tests, of course.

22 years agoAdd a CAM interface to the aac driver. This is useful in case you should
scottl [Sat, 27 Apr 2002 01:31:17 +0000 (01:31 +0000)]
Add a CAM interface to the aac driver.  This is useful in case you should
ever connect a SCSI Cdrom/Tape/Jukebox/Scanner/Printer/kitty-litter-scooper
to your high-end RAID controller.  The interface to the arrays is still
via the block interface; this merely provides a way to circumvent the
RAID functionality and access the SCSI buses directly.  Note that for
somewhat obvious reasons, hard drives are not exposed to the da driver
through this interface, though you can still talk to them via the pass
driver.  Be the first on your block to low-level format unsuspecting
drives that are part of an array!

To enable this, add the 'aacp' device to your kernel config.

MFC after: 3 days

22 years agoMake this driver compile for -stable also to prep for MFC
imp [Fri, 26 Apr 2002 23:11:23 +0000 (23:11 +0000)]
Make this driver compile for -stable also to prep for MFC

22 years agoMove the new byte order function prototypes from <sys/param.h> to
mike [Fri, 26 Apr 2002 22:48:23 +0000 (22:48 +0000)]
Move the new byte order function prototypes from <sys/param.h> to
<sys/endian.h>.  This puts us in line with NetBSD and OpenBSD.

22 years ago o Control access to the vm_page_buckets with a mutex.
alc [Fri, 26 Apr 2002 22:44:15 +0000 (22:44 +0000)]
 o Control access to the vm_page_buckets with a mutex.
 o Fix some style(9) bugs.

22 years agoRename the file used to specify the nextboot to make it clear that this
wes [Fri, 26 Apr 2002 22:32:15 +0000 (22:32 +0000)]
Rename the file used to specify the nextboot to make it clear that this
is a loader configuration file and can be used for more than just a
kernel name.

Submitted by: Gordon Tetlow <gordont@gnf.org>

22 years agoNow that the private parts of timecounters are no longer being fingered
phk [Fri, 26 Apr 2002 21:51:08 +0000 (21:51 +0000)]
Now that the private parts of timecounters are no longer being fingered
by other bits of code, split struct timecounter into two.

struct timecounter contains just the bits which pertains to the hardware
counter and the reading of it.

struct timehands (as in "the hands on a clock") contains all the ugly bit
fidling stuff.  Statically compile ten timehands.

This commit is the functional part.  A later cosmetic patch will rename
various variables and fieldnames.

22 years agoHide the private parts of timecounter from a couple of places that don't
phk [Fri, 26 Apr 2002 21:31:44 +0000 (21:31 +0000)]
Hide the private parts of timecounter from a couple of places that don't
really need to know the gory details.

22 years agoRedo the pps bit to avoid digging into the private bits of the timecounter.
phk [Fri, 26 Apr 2002 21:22:02 +0000 (21:22 +0000)]
Redo the pps bit to avoid digging into the private bits of the timecounter.

22 years agoCatch up ata(4) controllers, fix a typo (s/Cyrex/Cyrix).
bmah [Fri, 26 Apr 2002 21:12:51 +0000 (21:12 +0000)]
Catch up ata(4) controllers, fix a typo (s/Cyrex/Cyrix).

22 years agoA long, long time ago, msmith introduced vfs.root.mountfrom
dcs [Fri, 26 Apr 2002 20:52:59 +0000 (20:52 +0000)]
A long, long time ago, msmith introduced vfs.root.mountfrom
loader variable, which let users specify the root mount point
the exact way one does after booting the kernel.

Let's take this opportunity to document it...

22 years agoSimplify the RFC2783 and PPS_SYNC timestamp collection API.
phk [Fri, 26 Apr 2002 20:24:28 +0000 (20:24 +0000)]
Simplify the RFC2783 and PPS_SYNC timestamp collection API.

22 years agoNew release notes: at(1) -r/-t, ls(1) -h, pwd(1) -L, split(1) -a,
bmah [Fri, 26 Apr 2002 20:13:13 +0000 (20:13 +0000)]
New release notes:  at(1) -r/-t, ls(1) -h, pwd(1) -L, split(1) -a,
sysctl(8) -d, unexpand(1) -t.

MFC noted:  tftp(1)/tftpd(8) IPv6 support.

22 years agoIf we are making a "release", we can obviously use -DNOCLEAN buildworld,
ru [Fri, 26 Apr 2002 19:40:45 +0000 (19:40 +0000)]
If we are making a "release", we can obviously use -DNOCLEAN buildworld,
to save some time, because ${CHROOTDIR}/usr/obj does not even exist.

22 years agoInitialize MCA in cpu_startup() so that it's ready before we wake-up
marcel [Fri, 26 Apr 2002 19:36:25 +0000 (19:36 +0000)]
Initialize MCA in cpu_startup() so that it's ready before we wake-up
the application processors. This allows us to collect unconsumed AP
specific error records as part of the wake-up.

22 years agoMCA specific code has been moved to a seperate file. It is expected
marcel [Fri, 26 Apr 2002 19:33:20 +0000 (19:33 +0000)]
MCA specific code has been moved to a seperate file. It is expected
to grow enough to be in the way here.

22 years agoMachine Check Architecture (MCA) support code. Error records are
marcel [Fri, 26 Apr 2002 19:30:11 +0000 (19:30 +0000)]
Machine Check Architecture (MCA) support code. Error records are
collected at boot and made available through sysctl(8). At the
moment, the following MIB names are created:

hw.mca.count - The number of error records collected.
hw.mca.first - The lowest sequence number present.
hw.mca.last - The highest sequence number present.
hw.mca.<X> - The error record with sequence number <X>.

Using sysctl(8) allows us to easily detect and analyze the records,
which is very helpful during development of MCA but can also be used
in production as a way to collect machine health statistics.

22 years agoDon't ask root for the old password, except in the NIS case.
des [Fri, 26 Apr 2002 19:28:17 +0000 (19:28 +0000)]
Don't ask root for the old password, except in the NIS case.

Sponsored by: DARPA, NAI Labs

22 years agoMachine Check Architecture (MCA) structures and constants.
marcel [Fri, 26 Apr 2002 19:16:59 +0000 (19:16 +0000)]
Machine Check Architecture (MCA) structures and constants.

22 years agoModify our 'patch' utility to recognize the
gad [Fri, 26 Apr 2002 19:01:19 +0000 (19:01 +0000)]
Modify our 'patch' utility to recognize the
   \ No newline at end of file
line that some versions of diff print out if the last line of the two files
are different, and one of the two files does not have a newline character
on that last line.

This change is still somewhat under discussion in -arch and -standards, but I
want to commit it to -current today so I'd have the chance to MFC it to -stable
before the code freeze for 4.6-release (which would be May 1st).

Note: the related change to 'diff' (so it might *generate* that line) is NOT
expected to be included in 4.6-release.  We can debate that change later.

Obtained from: NetBSD (1.13 of basesrc/usr.bin/patch/pch.c, by kristerw)
MFC after: 4 days

22 years agoMilestone #1 in cross-arch make releases.
ru [Fri, 26 Apr 2002 17:55:27 +0000 (17:55 +0000)]
Milestone #1 in cross-arch make releases.

Do not install games and profiled libraries to the ${CHROOTDIR}
with the initial installworld.

Eliminate the need in the second installworld.  For that, make sure
_everything_ is built in the "world" environment, using the right
tool chain.

Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1.  Split the
buildworld process into stages, and skip some stages when
SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5
dists).

Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running
makewhatis(1) at the end of installworld (used when making crypto,
krb4, and krb5 dists).

In release/scripts/doFS.sh, ensure that the correct boot blocks are
used.

Moved the creation of the "crypto" dist from release.5 to
release.2.

In release.3 and doMFSKERN, build kernels in the "world"
environment.  KERNELS now means "additional" kernels, GENERIC is
always built.

Ensure we build crunched binaries in the "world" environment.
Obfuscate release/Makefile some more (WMAKEENV) to achieve this.

Inline createBOOTMFS target.

Use already built GENERIC kernel modules to augment mfsfd's
/stand/modules.  GC doMODULES as such.

Assorted fixes:

Get rid of the "afterdistribute" target by moving the single use
of it from sys/Makefile to etc/Makefile's "distribute".

Makefile.inc1: apparently "etc" no longer needs to be last for
"distribute" to succeed.

gnu/usr.bin/perl/library/Makefile.inc: do not override the
"install" and "distribute" targets, do it the "canonical" way.

release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and
catpages appear in the right dists.  Note that because Perl does
not respect the MANBUILDCAT (and NOMAN), this results in a loss of
/usr/share/perl/man/cat* empty directories.  This will be fixed
soon.

Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it
means "make KerberosIV"), as documented in the make.conf(5)
manpage.  Most of the userland makefiles did not test it for "YES"
anyway.

XXX Should specialized kerberized libpam versions be included into
the krb4 and krb5 dists?  (libpam.a would be incorrect anyway if
both krb4 and krb5 dists were choosen.)

Make sure "games" dist is made before "catpages", otherwise games
catpages settle in the wrong dist.

Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>

22 years agoThis commit was generated by cvs2svn to compensate for changes in r95504,
mikeh [Fri, 26 Apr 2002 16:51:03 +0000 (16:51 +0000)]
This commit was generated by cvs2svn to compensate for changes in r95504,
which included commits to RCS files with non-trunk default branches.

22 years agoImport the FreeBSD differences Luke patched in lukemftp:
mikeh [Fri, 26 Apr 2002 16:51:03 +0000 (16:51 +0000)]
Import the FreeBSD differences Luke patched in lukemftp:
       *) Large file ASCII support
       *) -4/-6 options to force IPv4/IPv6 respectively

Also, fix for buffer overflow on PASV from malicious server.

22 years agoNew release notes: kenv(2), kenv(1) can set/delete kernel environment
bmah [Fri, 26 Apr 2002 15:36:37 +0000 (15:36 +0000)]
New release notes:  kenv(2), kenv(1) can set/delete kernel environment
variables, mv(1) -n.

MFCs noted:  last(1) -y.

22 years agoAdd ac97_reset function that polls ready bits in power register. One
orion [Fri, 26 Apr 2002 15:27:56 +0000 (15:27 +0000)]
Add ac97_reset function that polls ready bits in power register.  One
component in tweaks required for the ICH3 controller in the thinkpad
x22 reported by Colin Perkins on -multimedia.

22 years agoAdd script to install the compat4x distribution, just like other
matusita [Fri, 26 Apr 2002 13:33:46 +0000 (13:33 +0000)]
Add script to install the compat4x distribution, just like other
compat*-install.sh scripts.

PR: 37485
MFC after: 5 days

22 years agoMove the winding of timecounters out of hardclock and into a normal
phk [Fri, 26 Apr 2002 12:37:36 +0000 (12:37 +0000)]
Move the winding of timecounters out of hardclock and into a normal
timeout loop.

Limit the rate at which we wind the timecounters to approx 1000 Hz.

This limits the precision of the get{bin,nano,micro}[up]time(9)
functions to roughly a millisecond.

22 years agoCorrect indent.
ume [Fri, 26 Apr 2002 12:27:55 +0000 (12:27 +0000)]
Correct indent.

22 years agoBackout copyright notice I broke by my previous commit.
ume [Fri, 26 Apr 2002 12:16:34 +0000 (12:16 +0000)]
Backout copyright notice I broke by my previous commit.