]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoThe toggle carry bit is stored in the headp not the tailp.
Nick Hibma [Sat, 29 Jan 2000 14:18:31 +0000 (14:18 +0000)]
The toggle carry bit is stored in the headp not the tailp.

24 years agoFix for (amongst other things) the cross-compile case where a too-old
Mark Murray [Sat, 29 Jan 2000 14:18:02 +0000 (14:18 +0000)]
Fix for (amongst other things) the cross-compile case where a too-old
version of libperl was being found.

Tested on: beast

24 years agocosmetic changes only.
Søren Schmidt [Sat, 29 Jan 2000 14:10:59 +0000 (14:10 +0000)]
cosmetic changes only.

24 years agoRemove #include "ppc.h" and #if NPPC > 0 - this is always true as it is
Peter Wemm [Sat, 29 Jan 2000 14:02:30 +0000 (14:02 +0000)]
Remove #include "ppc.h" and #if NPPC > 0 - this is always true as it is
'optional ppc' in conf/files*

24 years agoSynced with sys/kern/subr_diskmbr.c rev 1.43.
KATO Takenori [Sat, 29 Jan 2000 14:00:58 +0000 (14:00 +0000)]
Synced with sys/kern/subr_diskmbr.c rev 1.43.

24 years agoFixed IEXTEN handling in raw mode again. The effect of IEXTEN when
Bruce Evans [Sat, 29 Jan 2000 13:55:02 +0000 (13:55 +0000)]
Fixed IEXTEN handling in raw mode again.  The effect of IEXTEN when
ICANON is off is implementation-defined.  Under BSD, IEXTEN is
independent of ICANON, so it must be turned off to get "raw" mode.

This was first fixed in rev.1.4 (1995/10/21) of libncurses/lib_raw.c,
but the change was lost in the downgrade to the contrib version.

The fix here is the same as in the old rev.1.4, less style bugs.  A
better fix would use cfmakeraw(3) to actually handle all of the
complications for switching to raw mode.

24 years agoAdd ip6fw.
Yoshinobu Inoue [Sat, 29 Jan 2000 13:54:44 +0000 (13:54 +0000)]
Add ip6fw.
Yes it is almost code freeze, but as the result of many thought, now I
think this should be added before 4.0...

make world check, kernel build check is done.

Reviewed by: green
Obtained from: KAME project

24 years agoRemove a workaround for a gas bug. It couldn't assemble a certain
Peter Wemm [Sat, 29 Jan 2000 13:51:17 +0000 (13:51 +0000)]
Remove a workaround for a gas bug.  It couldn't assemble a certain
lgdt instruction, but the binutils based one is fine and has been
for ages.

24 years agoSynced with sys/boot/i386/btx/btx/btx.s rev 1.14.
KATO Takenori [Sat, 29 Jan 2000 13:48:36 +0000 (13:48 +0000)]
Synced with sys/boot/i386/btx/btx/btx.s rev 1.14.

24 years agofix breakage of make release.
Yoshinobu Inoue [Sat, 29 Jan 2000 13:45:44 +0000 (13:45 +0000)]
fix breakage of make release.

Confirmed by: German Tischler <tanis@gaspode.franken.de>

24 years agoSynced with sys/isa/ppc.c rev 1.25.
KATO Takenori [Sat, 29 Jan 2000 13:40:17 +0000 (13:40 +0000)]
Synced with sys/isa/ppc.c rev 1.25.

Reminded by: nyan

24 years agoFix our -mprofiler-epilogue code.
David E. O'Brien [Sat, 29 Jan 2000 13:06:33 +0000 (13:06 +0000)]
Fix our -mprofiler-epilogue code.

"The problem is that egcs/gcc-2.95's reorganisation of the prologue and
epilogue code to use rtl instead of output_asm_insn() completely broke our
hooks.  rtl is emitted in a different order, only after optimisation, while
output_asm_insn() is emitted immediately.  rtl is presumably used so that
the prologue and epilogue can be optimised.

I couldn't find any good examples to copy.  gcc's own
FUNCTION_BLOCK_PROFILER still uses output_asm_insn() and seems to be
completely broken.  One of the XXX comments points to this.

IIRC, the hacks here basically arrange to emit magic label names; then when
the magic names are output, they are transformed into prologue and epilogue
code."

Submitted by: bde

24 years agoUndo the ill-conceived breakage of the previous commit and really fix:
Jason Evans [Sat, 29 Jan 2000 12:50:47 +0000 (12:50 +0000)]
Undo the ill-conceived breakage of the previous commit and really fix:

For libc_r renamed syscalls, correct symbol naming from
_thread_sys_foo() <-- _foo() to _thread_sys_foo() <-- _foo() <-- foo().  This
is necessary for system calls which libc_r doesn't define foo().

Some weak symbols such as poll() are defined twice.  From what I understand,
depending on one weak symbol or the other to be used is a bad idea.  All
such weak symbols defined in the libc_r-specific code should therefore be
made strong (non-weak?).

Simplify PSEUDO() to not define any weak symbols, since they aren't ever
needed.

alpha/SYS.h:

Correct reversed usage of WEAK_ALIAS(), which has reversed arguments from
__weak_reference().  Also, fix reversal of symbols, so that syscall foo()
is a weak alias for _foo().

Add WEAK_ALIAS() call to PRSYSCALL(), which unlike the i386 version of
PRSYSCALL(), is not defined in terms of PSYSCALL().

Make PSEUDO() equivalent to the i386 version.

24 years agoRegen after pointing to the real setres[ug]id(). Remove dummy function.
Peter Wemm [Sat, 29 Jan 2000 12:45:36 +0000 (12:45 +0000)]
Regen after pointing to the real setres[ug]id().  Remove dummy function.
(Note: getres[ug]id is still a dummy as it returns via pointers to a
 different sized integer and does need real wrappers.)

24 years agoUse the real setres[ug]id(2) rather than dummy stubs. Newer glibc's
Peter Wemm [Sat, 29 Jan 2000 12:43:29 +0000 (12:43 +0000)]
Use the real setres[ug]id(2) rather than dummy stubs.  Newer glibc's
call this.

24 years ago"Rev.1.45 of cc_tools/Makefile was a wrong fix for a problem introduced
David E. O'Brien [Sat, 29 Jan 2000 12:38:21 +0000 (12:38 +0000)]
"Rev.1.45 of cc_tools/Makefile was a wrong fix for a problem introduced
in rev.1.44 (the egcs to gcc switch).  The problem is that print-rtl.o
is now needed to build some tools, but it wasn't added to the list of
objects which are specially handled because they are prerequisites for
tools."

Submitted by: bde

24 years agoRevert to rev 1.6 until post 4.0-R. As probably 98% of i386 installs
David E. O'Brien [Sat, 29 Jan 2000 12:18:03 +0000 (12:18 +0000)]
Revert to rev 1.6 until post 4.0-R.  As probably 98% of i386 installs
happen with a keyboard and monitor the console change was not as needed
in the i386 case as the Alpha case.  IMO >50% of Alpha installs are using
a serial console, the change matching rev 1.7 should not be backed out.

24 years agoMove the header installation down into `readline' which is the subsystem
David E. O'Brien [Sat, 29 Jan 2000 12:01:52 +0000 (12:01 +0000)]
Move the header installation down into `readline' which is the subsystem
the headers are part of.

24 years agoMask off the last two bits before comparing. It might just be that some
Nick Hibma [Sat, 29 Jan 2000 11:59:31 +0000 (11:59 +0000)]
Mask off the last two bits before comparing. It might just be that some
hardware might leave those bits in the wrong state.

24 years agoTripmine for bad hardware.
Nick Hibma [Sat, 29 Jan 2000 11:50:44 +0000 (11:50 +0000)]
Tripmine for bad hardware.

24 years agoSorry in this just befor code freeze commit.
Yoshinobu Inoue [Sat, 29 Jan 2000 11:49:07 +0000 (11:49 +0000)]
Sorry in this just befor code freeze commit.
This is fix to usr.sbin/trpt and tcp_debug.[ch]
I think of putting this after 4.0 but,,,

 -There was bug that when INET6 is defined,
  IPv4 socket is not traced by trpt.

 -I received request from a person who distribute a program
  which use tcp_debug interface and print performance statistics,
  that
    -leave comptibility with old program as much as possible
    -use same interface with other OSes

  So, I talked with itojun, and synced API with netbsd IPv6 extension.

makeworld check, kernel build check(includes GENERIC) is done.

But if there happen to any problem, please let me know and
I soon backout this change.

24 years agoRemove MAINTAINER=
Brian Somers [Sat, 29 Jan 2000 11:08:00 +0000 (11:08 +0000)]
Remove MAINTAINER=

I don't claim to own the code and certainly don't want to discourage
people from fixing or updating it.

[I know it's the 29th, but the FREEZE hasn't yet been posted to committers]

24 years agoFixed chgrp lossage in rev.1.233. Most floppy devices and all slices for disk
Bruce Evans [Sat, 29 Jan 2000 10:32:49 +0000 (10:32 +0000)]
Fixed chgrp lossage in rev.1.233.  Most floppy devices and all slices for disk
devices because accessible by group wheel instead of group operator.

Didn't fix fishy group for rsa*.ctl.  This device should have group operator
if that is safe, or mode 600 and group wheel.

Removed ssc and uk*.

Removed bogus (redundant) chmod's to 600.

24 years agoBack out 1.134. It works from 3.3R, but breaks cross compilation for too
Warner Losh [Sat, 29 Jan 2000 08:39:07 +0000 (08:39 +0000)]
Back out 1.134.  It works from 3.3R, but breaks cross compilation for too
little gain.  I'll work out the issues after 4.0R is out.

24 years agoSimplify messages of Pentium II, Pentium II Xeon, Celeron, Pentium III
KATO Takenori [Sat, 29 Jan 2000 07:49:02 +0000 (07:49 +0000)]
Simplify messages of Pentium II, Pentium II Xeon, Celeron, Pentium III
and Pentium III Xeon CPUs.  If a CPU is one of Pentium II, Pentium II
Xeon and Celeron, the message is always "Pentium II/Pentium II
Xeon/Celeron".   If a CPU is one of Pentium III and Pentium III Xeon,
the message is always "Pentium III/Pentium III Xeon".

24 years agoCosmetic changes.
KATO Takenori [Sat, 29 Jan 2000 07:36:46 +0000 (07:36 +0000)]
Cosmetic changes.
 - Fixed order of include files.
 - Fixed white spaces.

24 years agoUpdate the /usr/include/ hierarchy for the cam/ and cam/scsi/ directories.
Chris Costello [Sat, 29 Jan 2000 06:42:23 +0000 (06:42 +0000)]
Update the /usr/include/ hierarchy for the cam/ and cam/scsi/ directories.

PR: 16345
Submitted by: Nick Hibma (henny) <n_hibma@webweaving.org>

24 years agochanges to osf1_wait4()
Andrew Gallatin [Sat, 29 Jan 2000 06:31:27 +0000 (06:31 +0000)]
changes to osf1_wait4()
- only allocate rusage struct when caller wants rusage info
- fix a stupid paren mismatch bug that was causing EPERM to get returned
  to callers rather then ECHILD

24 years agoSynced with sys/isa/sio.c rev 1.287.
KATO Takenori [Sat, 29 Jan 2000 04:47:22 +0000 (04:47 +0000)]
Synced with sys/isa/sio.c rev 1.287.

24 years agoSynced with sys/i386/isa/wd.c rev 1.218.
KATO Takenori [Sat, 29 Jan 2000 04:42:49 +0000 (04:42 +0000)]
Synced with sys/i386/isa/wd.c rev 1.218.

24 years agoFix various typos and mdoc style issues.
Mike Pritchard [Sat, 29 Jan 2000 04:18:51 +0000 (04:18 +0000)]
Fix various typos and mdoc style issues.

Reviewed by: rwatson

24 years agoAdd a manual page for the ELF dynamic linker. I initially created
John Polstra [Sat, 29 Jan 2000 03:16:54 +0000 (03:16 +0000)]
Add a manual page for the ELF dynamic linker.  I initially created
rtld.1 by means of a repository copy from "src/libexec/rtld-aout/rtld.1".
Then I edited it to make it (more) accurate for the ELF dynamic
linker.

24 years agoMove the man pages for the a.out dynamic linker into the 1aout
John Polstra [Sat, 29 Jan 2000 03:13:49 +0000 (03:13 +0000)]
Move the man pages for the a.out dynamic linker into the 1aout
section.  I created rtld.1aout earlier with a repository copy.

This clears the way for the ELF dynamic linker man page, which I
will commit next.

24 years ago"Completed" the previous fix. Return ENOMEM on memory allocation failure
Bruce Evans [Sat, 29 Jan 2000 03:02:55 +0000 (03:02 +0000)]
"Completed" the previous fix.  Return ENOMEM on memory allocation failure
in sioattach(), not ENXIO.  Free resources before returning early in
sioprobe() and sioattach().

24 years agoMinor tweak: the D-Link 10/100 USB ethernet adapter is apparently using
Bill Paul [Sat, 29 Jan 2000 02:16:47 +0000 (02:16 +0000)]
Minor tweak: the D-Link 10/100 USB ethernet adapter is apparently using
the same design as the LinkSys adapter and needs the same special handling
to enable its PHY.

24 years agoEnable '=' character in the filename.
Boris Popov [Sat, 29 Jan 2000 02:10:37 +0000 (02:10 +0000)]
Enable '=' character in the filename.

Submitted by: Dan Nelson <dnelson@emsphone.com>

24 years agoGrammar fixes:
Chris Costello [Sat, 29 Jan 2000 02:04:31 +0000 (02:04 +0000)]
Grammar fixes:
     - "In result" generally sounds better as "As a result".
     - "faith interface ..." was changed to "The faith interface" at the
       beginning of a sentence.

24 years agoAdd three entries:
Warner Losh [Sat, 29 Jan 2000 02:04:16 +0000 (02:04 +0000)]
Add three entries:
# Apollo PCMCIA Ethernet Adapter
# Olicom OC2220
# National Semiconductor InfoMover NE4100

I forgot who submitted the first two, but the third one was submitted
by Jim Bloom.

24 years agoGrammar fix: ``Different than'' should really be ``different from''.
Chris Costello [Sat, 29 Jan 2000 01:54:59 +0000 (01:54 +0000)]
Grammar fix:  ``Different than'' should really be ``different from''.

24 years agoAdd a #define for RTLD_LOCAL as required by the Single Unix
John Polstra [Sat, 29 Jan 2000 01:53:35 +0000 (01:53 +0000)]
Add a #define for RTLD_LOCAL as required by the Single Unix
Specification.

24 years agoRevive the warning that dllockinit() is experimental and subject to
John Polstra [Sat, 29 Jan 2000 01:33:21 +0000 (01:33 +0000)]
Revive the warning that dllockinit() is experimental and subject to
change.

24 years agoWhen a threads package registers locking methods with dllockinit(),
John Polstra [Sat, 29 Jan 2000 01:27:04 +0000 (01:27 +0000)]
When a threads package registers locking methods with dllockinit(),
figure out which shared object(s) contain the the locking methods
and fully bind those objects as if they had been loaded with
LD_BIND_NOW=1.  The goal is to keep the locking methods from
requiring any lazy binding.  Otherwise infinite recursion occurs
in _rtld_bind.

This fixes the infinite recursion problem in the linuxthreads port.

24 years agoForgot to setup disc2 (the livefs) for booting also.
David E. O'Brien [Sat, 29 Jan 2000 01:15:52 +0000 (01:15 +0000)]
Forgot to setup disc2 (the livefs) for booting also.

24 years agoRestore to version 1.14. Only opt_inet.h is required again.
Brian Somers [Sat, 29 Jan 2000 01:11:03 +0000 (01:11 +0000)]
Restore to version 1.14.  Only opt_inet.h is required again.

24 years agoRemove unused includes
Brian Somers [Sat, 29 Jan 2000 01:10:24 +0000 (01:10 +0000)]
Remove unused includes

24 years ago* Use `console' rather than `ttyv0' so there will always be a getty on
David E. O'Brien [Sat, 29 Jan 2000 00:50:21 +0000 (00:50 +0000)]
* Use `console' rather than `ttyv0' so there will always be a getty on
  the console reguardless of the type of console.
* Make the Alpha version match the i386 version.

24 years agoSetup cdrom/disc1/ for booting in the Alpha case.
David E. O'Brien [Sat, 29 Jan 2000 00:26:41 +0000 (00:26 +0000)]
Setup cdrom/disc1/ for booting in the Alpha case.

24 years agodifferentiate cmi8330 and als100 pnp cards based on their vendor id. this
Cameron Grant [Sat, 29 Jan 2000 00:18:29 +0000 (00:18 +0000)]
differentiate cmi8330 and als100 pnp cards based on their vendor id.  this
is a kludge for 4.0

24 years agoFor syscalls that are renamed to _thread_sys_foo, create a weak alias
Jason Evans [Fri, 28 Jan 2000 22:47:21 +0000 (22:47 +0000)]
For syscalls that are renamed to _thread_sys_foo, create a weak alias
called _foo, not _thread_sys_foo.

24 years agoGet the MEDIA_CHANGED status right here too.
Søren Schmidt [Fri, 28 Jan 2000 22:17:29 +0000 (22:17 +0000)]
Get the MEDIA_CHANGED status right here too.

24 years agoUse _fcntl() (not fcntl()) inside of fcntl().
Jason Evans [Fri, 28 Jan 2000 22:10:27 +0000 (22:10 +0000)]
Use _fcntl() (not fcntl()) inside of fcntl().

Reported by: green

24 years agoUse atapi_test_ready not atapi_wait_ready, I need more sleep :(
Søren Schmidt [Fri, 28 Jan 2000 21:30:31 +0000 (21:30 +0000)]
Use atapi_test_ready not atapi_wait_ready, I need more sleep :(

24 years agoRemove Fortran from build tools. It isn't needed to build the system
Warner Losh [Fri, 28 Jan 2000 21:15:24 +0000 (21:15 +0000)]
Remove Fortran from build tools.  It isn't needed to build the system
and costs us an extra 2% to build it for no reason.  It may break
building cross compilation environments for fortran, but that isn't
officially supported at this time anyway (also, the % of our user base
that would use that is < .001% imho).  This does't break fortran (it
is built again later anyway).

Reviewed by: obrien
Tested by: make buildworld and make buildworld -DNOCLEAN

24 years agoGet the MEDIA_CHANGED status right.
Søren Schmidt [Fri, 28 Jan 2000 21:03:25 +0000 (21:03 +0000)]
Get the MEDIA_CHANGED status right.

24 years agorename disk_delete() to disk_destroy().
Poul-Henning Kamp [Fri, 28 Jan 2000 20:49:43 +0000 (20:49 +0000)]
rename disk_delete() to disk_destroy().

24 years agoFix a bug that could crash the system if you press ^T while a slower
Brian Feldman [Fri, 28 Jan 2000 20:40:29 +0000 (20:40 +0000)]
Fix a bug that could crash the system if you press ^T while a slower
system is slowed down and in the right spot (a race condition in fork()).

The "previous time" fields have moved from pstat to proc.  Anything which
uses KVM needs to be recompiled with a new libkvm/headers.

A couple wacky u_quad_t's in struct proc are now u_int64_t (the same, but
according to lack of 'quad's in proc.h and usage in kern_resource.c).
This will have no effect on code.

This has been make-world-and-installed-new-kernel-which-works-fine-tested.

Reviewed by: bde (previous version)

24 years agoFix typo's.
Jeroen Ruigrok van der Werven [Fri, 28 Jan 2000 20:20:28 +0000 (20:20 +0000)]
Fix typo's.

PR: 16430
Submitted by: Adam.Kranzel@teru.dyndns.org (shade@dnai.com)

24 years agoFix the atapi-fd.c (afd device) so that fx MSDOS formatted ZIP
Søren Schmidt [Fri, 28 Jan 2000 20:18:19 +0000 (20:18 +0000)]
Fix the atapi-fd.c (afd device) so that fx MSDOS formatted ZIP
disks can be mounted. Use b_pblkno instead of b_blkno

Found by: phk

Protect (proberly) against setting modes on nonexisting devices.

24 years ago- Fix an issue that paniced the machine in data overrun
Gerard Roudier [Fri, 28 Jan 2000 20:11:32 +0000 (20:11 +0000)]
- Fix an issue that paniced the machine in data overrun
  condition.
- 1 line change that allows to balance chips between ncr
  and sym using pci compat option (not compiled by default
  in 4.0 but maintains the driver source 3.4 compatible).

24 years agoCount AF_INET6 attachement to routing socket.
Yoshinobu Inoue [Fri, 28 Jan 2000 20:10:51 +0000 (20:10 +0000)]
Count AF_INET6 attachement to routing socket.

Obtained from: KAME project

24 years agoIntroduce ACL man pages en masse for library calls, and general introduction.
Robert Watson [Fri, 28 Jan 2000 20:07:00 +0000 (20:07 +0000)]
Introduce ACL man pages en masse for library calls, and general introduction.

Introduce ACL man pages en masse for library calls, and general introduction.

Also, fix acl_valid.c non-portable calls to include _np in their names,
making them standard-happy as well as consistent with acl.h

24 years agoFix inconsistent debug output. (syslog -> warnx)
Yoshinobu Inoue [Fri, 28 Jan 2000 20:06:15 +0000 (20:06 +0000)]
Fix inconsistent debug output. (syslog -> warnx)

Specified by: sheldonh

Reviewed by: des

24 years agoFix rshd coredump when AF_INET socket is used.
Yoshinobu Inoue [Fri, 28 Jan 2000 20:02:02 +0000 (20:02 +0000)]
Fix rshd coredump when AF_INET socket is used.

Confirmed by: F. Heinrichmeyer <fritz.heinrichmeyer@fernuni-hagen.de>

24 years agoThere were so far only 42 different conversion specifications in
Joerg Wunsch [Fri, 28 Jan 2000 17:40:42 +0000 (17:40 +0000)]
There were so far only 42 different conversion specifications in
strftime(3), add another one. :)  %z yields the local timezone's offset
in hours and minutes, as used in RFC822 headers.  There's a precedence
for this in Lunux' libc, and Internet software (like Perl scripts)
start using it.

OKed by (wrt. the code freeze): jkh

24 years agoBack out previous commit; it was premature.
Archie Cobbs [Fri, 28 Jan 2000 17:11:07 +0000 (17:11 +0000)]
Back out previous commit; it was premature.

24 years agoFix compilation on ISA only setups...
Søren Schmidt [Fri, 28 Jan 2000 15:57:13 +0000 (15:57 +0000)]
Fix compilation on ISA only setups...

24 years agoFix man page structures/constants to match recent changes in acl.h
Robert Watson [Fri, 28 Jan 2000 15:22:51 +0000 (15:22 +0000)]
Fix man page structures/constants to match recent changes in acl.h
Fix a few more namespace messes and whitespace curiosities in acl.h
Fix comments in acl.h
Clean up some function prototypes from acl.h that won't be committed
    before the code freeze.

Some of this kindly pointed out by: the ever patient bde

24 years agoAllow for cdldr to be built in ${.OBJDIR}/../cdldr/cdldr instead of
John Baldwin [Fri, 28 Jan 2000 15:19:22 +0000 (15:19 +0000)]
Allow for cdldr to be built in ${.OBJDIR}/../cdldr/cdldr instead of
only looking for it in ${.CURDIR}/../cdldr/cdldr.  This fixes world.

24 years agoUpdate for ntp 4.0.99b.
Ollivier Robert [Fri, 28 Jan 2000 15:13:45 +0000 (15:13 +0000)]
Update for ntp 4.0.99b.

24 years agoUpdate to 4.0.99b. According to Harlan Stenn, the things holding the
Ollivier Robert [Fri, 28 Jan 2000 15:05:59 +0000 (15:05 +0000)]
Update to 4.0.99b. According to Harlan Stenn, the things holding the
release of 4.1.0 are not FreeBSD-related so importing this one is
not a problem.

24 years agoMerge conflicts with the import of 4.0.99b.
Ollivier Robert [Fri, 28 Jan 2000 15:02:35 +0000 (15:02 +0000)]
Merge conflicts with the import of 4.0.99b.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r56746,
Ollivier Robert [Fri, 28 Jan 2000 14:55:50 +0000 (14:55 +0000)]
This commit was generated by cvs2svn to compensate for changes in r56746,
which included commits to RCS files with non-trunk default branches.

24 years agoVirgin import of ntpd 4.0.99b
Ollivier Robert [Fri, 28 Jan 2000 14:55:50 +0000 (14:55 +0000)]
Virgin import of ntpd 4.0.99b

24 years agoFixed world breakage (wrong path to pim6stat).
Bruce Evans [Fri, 28 Jan 2000 13:53:11 +0000 (13:53 +0000)]
Fixed world breakage (wrong path to pim6stat).

Fixed other bugs in special install rule (2 cosmetic, 2 non-cosmetic).

24 years agoCleanup the ata_dmainit function a bit.
Søren Schmidt [Fri, 28 Jan 2000 13:35:43 +0000 (13:35 +0000)]
Cleanup the ata_dmainit function a bit.
Also allow BIOS setup DMA on unknown controllers.

24 years agoBackout diffs which should not be included.
Yoshinobu Inoue [Fri, 28 Jan 2000 13:16:34 +0000 (13:16 +0000)]
Backout diffs which should not be included.

24 years agoNull commit. This is only to record that the problems mentioned in
Joerg Wunsch [Fri, 28 Jan 2000 13:05:55 +0000 (13:05 +0000)]
Null commit.  This is only to record that the problems mentioned in
the message of rev 1.20 only seem to happen after the recent upgrade
to groff v 1.15.  Anyway, using comments instead of \c seems to be a
safe way (IMHO), so it should stay.

24 years agoCorrect the list of error messages. It was incomplete.
Nick Hibma [Fri, 28 Jan 2000 13:05:26 +0000 (13:05 +0000)]
Correct the list of error messages. It was incomplete.

24 years agoChanged setflags() to set_flags(). This fixes part of the world breakage
Bruce Evans [Fri, 28 Jan 2000 12:44:50 +0000 (12:44 +0000)]
Changed setflags() to set_flags().  This fixes part of the world breakage
due to recently incremented namespace pollution in <unistd.h>.

24 years agoRemove the i386 bit from the manpage which was left there due
Jeroen Ruigrok van der Werven [Fri, 28 Jan 2000 12:24:53 +0000 (12:24 +0000)]
Remove the i386 bit from the manpage which was left there due
to my misunderstanding.

24 years ago#This is a null commit to give correct description for the previous change.
Yoshinobu Inoue [Fri, 28 Jan 2000 12:17:49 +0000 (12:17 +0000)]
#This is a null commit to give correct description for the previous change.
#Please forget the strange log message of the previous commit .

IPv6 multicast routing.
  kernel IPv6 multicast routing support.
  pim6 dense mode daemon
  pim6 sparse mode daemon
  netstat support of IPv6 multicast routing statistics

  Merging to the current and testing with other existing multicast routers
  is done by Tatsuya Jinmei <jinmei@kame.net>, who writes and maintainances
  the base code in KAME distribution.

  Make world check and kernel build check was also successful.

Obtained from: KAME project

24 years agoFixed a memory leak for slices with an (unsupported) bad sector table.
Bruce Evans [Fri, 28 Jan 2000 11:51:08 +0000 (11:51 +0000)]
Fixed a memory leak for slices with an (unsupported) bad sector table.
Broken in: rev.1.80.

24 years agoUpdate this with the additional nets recomended by reading
Rodney W. Grimes [Fri, 28 Jan 2000 11:30:28 +0000 (11:30 +0000)]
Update this with the additional nets recomended by reading
draft-manning-dsua-01.txt.

Stop using public addresses as samples and use the recommended
192.0.2.0/24 netblock that has specifically been set aside for
documentation purposes.

Reviewed by: readers of freebsd-security did not respond to a request
                for review

24 years agoAdd ${DEBUG_FLAGS} to CFLAGS in bsd.kmod.mk, not in scattered module
Bruce Evans [Fri, 28 Jan 2000 11:26:46 +0000 (11:26 +0000)]
Add ${DEBUG_FLAGS} to CFLAGS in bsd.kmod.mk, not in scattered module
makefiles.  Bad examples in fxp/Makefile keep getting copied to new
makefiles.

24 years agoRegen.
Nick Hibma [Fri, 28 Jan 2000 10:31:12 +0000 (10:31 +0000)]
Regen.

24 years agoCorrect the entry for the Kodak DC290.
Nick Hibma [Fri, 28 Jan 2000 10:26:29 +0000 (10:26 +0000)]
Correct the entry for the Kodak DC290.

Submitted By: Oliver Fromme <olli@dorifer.heim3.tu-clausthal.de>

24 years agoDon't permit generation of non-physical disk addresses.
Bruce Evans [Fri, 28 Jan 2000 10:22:07 +0000 (10:22 +0000)]
Don't permit generation of non-physical disk addresses.

subr_diskmbr.c:
Don't "helpfully" enlarge our idea of the disk size to cover all the
primary slices.  Instead, truncate or discard slices that don't seem
to be on the disk.  The enlargement was a hack for disks that don't
report their size (e.g., MFM disks).  It is just wrong in general.

wd.c:
In CHS mode, limit the disk size so that cylinder numbers >= 65536
cannot occur.  This normally only affects disks larger than 33.8GB.
CHS mode accesses to addresses above the limit are now properly broken
(an error is returned instead of garbage for reads and disk corruption
for writes).

PR: 15611
Reviewed by: readers of freebsd-bugs did not respond to a request
             for review

24 years agoFix English, mdoc and layout of the previous commit, as requested by
Sheldon Hearn [Fri, 28 Jan 2000 10:21:19 +0000 (10:21 +0000)]
Fix English, mdoc and layout of the previous commit, as requested by
the committer (shin).  While I don't have permission for this change
from the inetd maintainer (des), I assume that shin has permission
and I'm just fixing his contribution up for him.

Okay, I couldn't resist, I made some extra changes:

* Replace ".Tn FreeBSD" with .Fx
* Make the illegal TCPMUX and IPSEC sections legal subsections
  of the IMPLEMENTATION NOTES section.

Requested by: shin

24 years agoTrack internal(!) libc symbol naming change.
Jason Evans [Fri, 28 Jan 2000 08:49:23 +0000 (08:49 +0000)]
Track internal(!) libc symbol naming change.

24 years agoFix misunderstanding that ata(4) is i386 only. Since this file has
Jeroen Ruigrok van der Werven [Fri, 28 Jan 2000 08:41:04 +0000 (08:41 +0000)]
Fix misunderstanding that ata(4) is i386 only.  Since this file has
no history, it does not warrant repo action.

24 years agoOops, fix compile error due to the new way of storing the transfer mode.
Søren Schmidt [Fri, 28 Jan 2000 08:03:21 +0000 (08:03 +0000)]
Oops, fix compile error due to the new way of storing the transfer mode.

24 years agoMerged with sys/i386/conf/GENERIC rev 1.238.
KATO Takenori [Fri, 28 Jan 2000 07:15:23 +0000 (07:15 +0000)]
Merged with sys/i386/conf/GENERIC rev 1.238.

24 years agoInstall setflags.3 and its link to getflags.3.
Bruce Evans [Fri, 28 Jan 2000 07:14:52 +0000 (07:14 +0000)]
Install setflags.3 and its link to getflags.3.

24 years agoChanged setflags() to set_flags(). This fixes world breakage due to
Bruce Evans [Fri, 28 Jan 2000 07:12:03 +0000 (07:12 +0000)]
Changed setflags() to set_flags().  This fixes world breakage due to
recently incremented namespace pollution in <unistd.h>.

24 years agoMitigate the stream.c attacks
Warner Losh [Fri, 28 Jan 2000 06:13:09 +0000 (06:13 +0000)]
Mitigate the stream.c attacks

o Drop all broadcast and multicast source addresses in tcp_input.
o Enable ICMP_BANDLIM in GENERIC.
o Change default to 200/s from 100/s.  This will still stop the attack, but
  is conservative enough to do this close to code freeze.

This is not the optimal patch for the problem, but is likely the least
intrusive patch that can be made for this.

Obtained from: Don Lewis and Matt Dillon.
Reviewed by: freebsd-security

24 years agoSorry I didn't commit these files at the commit just a few minutes before.
Yoshinobu Inoue [Fri, 28 Jan 2000 05:27:14 +0000 (05:27 +0000)]
Sorry I didn't commit these files at the commit just a few minutes before.
(IPv6 multicast routing)
I think I mistakenly touched TAB and the last arg sys/netinet6 to
the cvs commit changed to sys/netinet6/in6_proto.c.

24 years agoIPv6 multicast routing.
Yoshinobu Inoue [Fri, 28 Jan 2000 05:10:56 +0000 (05:10 +0000)]
IPv6 multicast routing.
  kernel IPv6 multicast routing support.
  pim6 dense mode daemon
  pim6 sparse mode daemon
  netstat support of IPv6 multicast routing statistics

  Merging to the current and testing with other existing multicast routers
  is done by Tatsuya Jinmei <jinmei@kame.net>, who writes and maintainances
  the base code in KAME distribution.

  Make world check and kernel build check was also successful.

24 years agoAdd support for detection of the ESS 1688 found on my Dell Latitude LM
Andrew Gallatin [Fri, 28 Jan 2000 04:28:11 +0000 (04:28 +0000)]
Add support for detection of the ESS 1688 found on my Dell Latitude LM

24 years agoFixed sign and overflow bugs that caused the allocation size of the kernel
David Greenman [Fri, 28 Jan 2000 04:04:58 +0000 (04:04 +0000)]
Fixed sign and overflow bugs that caused the allocation size of the kernel
malloc region (kmem_map) to be wrong and semi-random on systems with more
than 1GB of RAM. This is not a complete fix, but is sufficient for
machines with 4GB or less of memory. A complete fix will require some
changes to the getenv stuff so that 64bit values can be passed around.

NOT FIXED: machines with more than 4GB of RAM (e.g. some large Alphas)
since we're still using ints to hold some of the values.

Reviewed by: bde

24 years agoImplement a machine-independent (word size-independent) FICL_TRUE.
Daniel C. Sobral [Fri, 28 Jan 2000 03:59:12 +0000 (03:59 +0000)]
Implement a machine-independent (word size-independent) FICL_TRUE.