]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoThis commit was generated by cvs2svn to compensate for changes in r91100,
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:24:02 +0000 (01:24 +0000)]
This commit was generated by cvs2svn to compensate for changes in r91100,
which included commits to RCS files with non-trunk default branches.

22 years agoVendor import of OpenPAM Cantaloupe
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:24:02 +0000 (01:24 +0000)]
Vendor import of OpenPAM Cantaloupe

22 years agoThis commit was generated by cvs2svn to compensate for changes in r91097,
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:23:20 +0000 (01:23 +0000)]
This commit was generated by cvs2svn to compensate for changes in r91097,
which included commits to RCS files with non-trunk default branches.

22 years agoVendor import of OpenPAM Caliopsis
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:23:20 +0000 (01:23 +0000)]
Vendor import of OpenPAM Caliopsis

22 years agoThis commit was generated by cvs2svn to compensate for changes in r91094,
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:22:51 +0000 (01:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r91094,
which included commits to RCS files with non-trunk default branches.

22 years agoVendor import of OpenPAM Calamite
Dag-Erling Smørgrav [Sat, 23 Feb 2002 01:22:51 +0000 (01:22 +0000)]
Vendor import of OpenPAM Calamite

22 years agoAdd some DIAGNOSTIC code.
Julian Elischer [Fri, 22 Feb 2002 23:58:22 +0000 (23:58 +0000)]
Add some DIAGNOSTIC code.
While in userland, keep the thread's ucred reference in a shadow
field so that the usual place to store it is NULL.
If DIAGNOSTIC is not set, the thread ucred is kept valid until the next
kernel entry, at which time it is checked against the process cred
and possibly corrected. Produces a BIG speedup in
kernels with INVARIANTS set. (A previous commit corrected it
for the non INVARIANTS case already)

Reviewed by: dillon@freebsd.org

22 years agoANSIfy.
Mark Murray [Fri, 22 Feb 2002 23:31:57 +0000 (23:31 +0000)]
ANSIfy.

22 years agoFix warnings inspired by lint, a commercial lint and WARNS=4.
Mark Murray [Fri, 22 Feb 2002 21:24:14 +0000 (21:24 +0000)]
Fix warnings inspired by lint, a commercial lint and WARNS=4.

22 years agoFix a boatload of warnings inspired by lint, a commercial lint
Mark Murray [Fri, 22 Feb 2002 21:21:37 +0000 (21:21 +0000)]
Fix a boatload of warnings inspired by lint, a commercial lint
and WARNS=4.

22 years agoWarnings fixes.
Mark Murray [Fri, 22 Feb 2002 21:13:31 +0000 (21:13 +0000)]
Warnings fixes.

22 years agoFix warnings inspired by lint, a commercial lint and WARNS=4.
Mark Murray [Fri, 22 Feb 2002 21:11:03 +0000 (21:11 +0000)]
Fix warnings inspired by lint, a commercial lint and WARNS=4.

22 years agoPartially fix (well, work around) warnings inspired by lint, a
Mark Murray [Fri, 22 Feb 2002 21:02:58 +0000 (21:02 +0000)]
Partially fix (well, work around) warnings inspired by lint, a
commercial lint and WARNS=4.

22 years agoFix warnings inspired by lint, a commercial lint and WARNS=4.
Mark Murray [Fri, 22 Feb 2002 21:00:16 +0000 (21:00 +0000)]
Fix warnings inspired by lint, a commercial lint and WARNS=4.

22 years agoWarnings fixes inspired by lint, a commercial lint and WARNS=4.
Mark Murray [Fri, 22 Feb 2002 20:57:53 +0000 (20:57 +0000)]
Warnings fixes inspired by lint, a commercial lint and WARNS=4.

22 years agoWarning fix. Enter into the spirit of getopt(3) a bit more.
Mark Murray [Fri, 22 Feb 2002 20:55:04 +0000 (20:55 +0000)]
Warning fix. Enter into the spirit of getopt(3) a bit more.

22 years agoFix warnings inspired by lint, a commercial lint and WARNS=4.
Mark Murray [Fri, 22 Feb 2002 20:51:00 +0000 (20:51 +0000)]
Fix warnings inspired by lint, a commercial lint and WARNS=4.

22 years agoFix warnings inspired by lint, a commercial lint and WARNS=4
Mark Murray [Fri, 22 Feb 2002 20:45:09 +0000 (20:45 +0000)]
Fix warnings inspired by lint, a commercial lint and WARNS=4

22 years agoAlso blindly attempt to fix broken world with respect to proc.p_runtime
Brian Feldman [Fri, 22 Feb 2002 19:10:09 +0000 (19:10 +0000)]
Also blindly attempt to fix broken world with respect to proc.p_runtime
changes.

22 years agoThe existing bazaar and site-specific policy in rc.diskless1 is Just Wrong;
David E. O'Brien [Fri, 22 Feb 2002 19:05:22 +0000 (19:05 +0000)]
The existing bazaar and site-specific policy in rc.diskless1 is Just Wrong;
and looks like no other Unix diskless configuration I've ever seen.
Thus allow a more traditional /etc.

Note, the use of an MFS /var should also be settable.
Otherwise installing ports(packages) is just a total PITA.

22 years agoUpdate for binutils_2_12_anoncvs_20020221.
David E. O'Brien [Fri, 22 Feb 2002 18:49:37 +0000 (18:49 +0000)]
Update for binutils_2_12_anoncvs_20020221.

22 years ago(Belatedly) add the required EXPORT_SYMS. I'm not sure the list is complete,
Dag-Erling Smørgrav [Fri, 22 Feb 2002 18:21:20 +0000 (18:21 +0000)]
(Belatedly) add the required EXPORT_SYMS.  I'm not sure the list is complete,
but at least linprocfs works (I haven't had the opportunity to test other
stuff that depends on the linux module, like aac or tdfx)

22 years agoModified release note: Binutils update. While I'm here, fix a minor
Bruce A. Mah [Fri, 22 Feb 2002 16:26:38 +0000 (16:26 +0000)]
Modified release note:  Binutils update.  While I'm here, fix a minor
grammatical mismatch ("Binutils" is the name of a package and
therefore is singular, not plural).

22 years ago- Whitespace fixes leftover from previous commit.
Andrew R. Reiter [Fri, 22 Feb 2002 13:43:56 +0000 (13:43 +0000)]
- Whitespace fixes leftover from previous commit.

Submitted by: bde

22 years ago- Whitespace fixup left over from previous commit.
Andrew R. Reiter [Fri, 22 Feb 2002 13:33:10 +0000 (13:33 +0000)]
- Whitespace fixup left over from previous commit.
- Remove bogus cast.

Submitted by: bde

22 years agoConvert p->p_runtime and PCPU(switchtime) to bintime format.
Poul-Henning Kamp [Fri, 22 Feb 2002 13:32:01 +0000 (13:32 +0000)]
Convert p->p_runtime and PCPU(switchtime) to bintime format.

22 years agoUse better scaling factor for NTPs correction.
Poul-Henning Kamp [Fri, 22 Feb 2002 12:59:20 +0000 (12:59 +0000)]
Use better scaling factor for NTPs correction.
Explain the magic.

22 years agoAllow PCI_ALLOW_UNSUPPORTED_IO_RANGE to leave broken setups broken enough
Poul-Henning Kamp [Fri, 22 Feb 2002 11:21:02 +0000 (11:21 +0000)]
Allow PCI_ALLOW_UNSUPPORTED_IO_RANGE to leave broken setups broken enough
to work.

22 years agoGC: BIO_ORDERED, various infrastructure dealing with BIO_ORDERED.
Poul-Henning Kamp [Fri, 22 Feb 2002 09:26:35 +0000 (09:26 +0000)]
GC: BIO_ORDERED, various infrastructure dealing with BIO_ORDERED.

22 years agoGC: BIO_ORDERED going away.
Poul-Henning Kamp [Fri, 22 Feb 2002 09:18:46 +0000 (09:18 +0000)]
GC: BIO_ORDERED going away.

22 years agoDisksort will not "munge" requests, BIO_ORDERED or not, so remove
Poul-Henning Kamp [Fri, 22 Feb 2002 09:14:06 +0000 (09:14 +0000)]
Disksort will not "munge" requests, BIO_ORDERED or not, so remove
use of BIO_ORDERED.

22 years agoReplace bowrite() with BUF_WRITE in ufs.
Poul-Henning Kamp [Fri, 22 Feb 2002 09:03:00 +0000 (09:03 +0000)]
Replace bowrite() with BUF_WRITE in ufs.

Remove bowrite(), it is now unused.

This is the first step in getting entirely rid of BIO_ORDERED which is
a generally accepted evil thing.

Approved by: mckusick

22 years agoGC: bufqueues are not used under -current, we use bioqueues.
Poul-Henning Kamp [Fri, 22 Feb 2002 07:45:16 +0000 (07:45 +0000)]
GC: bufqueues are not used under -current, we use bioqueues.

22 years agoUpdate for binutils_2_12_anoncvs_20020221.
David E. O'Brien [Fri, 22 Feb 2002 05:35:18 +0000 (05:35 +0000)]
Update for binutils_2_12_anoncvs_20020221.

22 years agoExplain how I did the binutils_2_12_anoncvs_20020221 import.
David E. O'Brien [Fri, 22 Feb 2002 05:11:49 +0000 (05:11 +0000)]
Explain how I did the binutils_2_12_anoncvs_20020221 import.

22 years agoMerge rev 1.2 (teach `ld' how to access FreeBSD's ld ELF hints)
David E. O'Brien [Fri, 22 Feb 2002 04:52:00 +0000 (04:52 +0000)]
Merge rev 1.2 (teach `ld' how to access FreeBSD's ld ELF hints)
into Binutils 2.12_20020221_snap.

22 years agoBring the binutils_2_12_20020221_snap version of this to the HEAD branch.
David E. O'Brien [Fri, 22 Feb 2002 04:51:15 +0000 (04:51 +0000)]
Bring the binutils_2_12_20020221_snap version of this to the HEAD branch.

22 years agoMerge rev 1.2 (FreeBSD a.out configuration support) into
David E. O'Brien [Fri, 22 Feb 2002 04:49:49 +0000 (04:49 +0000)]
Merge rev 1.2 (FreeBSD a.out configuration support) into
Binutils 2.12_20020221_snap.

22 years agoUse the stock (2.12_20020221_snap) version of this.
David E. O'Brien [Fri, 22 Feb 2002 04:48:12 +0000 (04:48 +0000)]
Use the stock (2.12_20020221_snap) version of this.

22 years agoUse the stock version of this.
David E. O'Brien [Fri, 22 Feb 2002 04:47:47 +0000 (04:47 +0000)]
Use the stock version of this.

22 years agoUse the vendor's 2.12_20020221_snap version of this file as in rev 1.4.
David E. O'Brien [Fri, 22 Feb 2002 04:47:16 +0000 (04:47 +0000)]
Use the vendor's 2.12_20020221_snap version of this file as in rev 1.4.

22 years agoMerge rev 1.6 (only define "ELF_DYNAMIC_INTERPRETER" if it isn't defined
David E. O'Brien [Fri, 22 Feb 2002 04:46:20 +0000 (04:46 +0000)]
Merge rev 1.6 (only define "ELF_DYNAMIC_INTERPRETER" if it isn't defined
elsewhere) into Binutils 2.12_20020221_snap.

22 years agoMerge rev 1.8 (only define "ELF_DYNAMIC_INTERPRETER" if it isn't defined
David E. O'Brien [Fri, 22 Feb 2002 04:45:47 +0000 (04:45 +0000)]
Merge rev 1.8 (only define "ELF_DYNAMIC_INTERPRETER" if it isn't defined
elsewhere) into Binutils 2.12_20020221_snap.

22 years agoMerge our ELF branding into Binutils 2.12_20020221_snap.
David E. O'Brien [Fri, 22 Feb 2002 04:44:56 +0000 (04:44 +0000)]
Merge our ELF branding into Binutils 2.12_20020221_snap.

22 years agoUse make -V VARIABLE | xargs ... to pass argument lists to program
Luigi Rizzo [Fri, 22 Feb 2002 04:44:28 +0000 (04:44 +0000)]
Use make -V VARIABLE | xargs ... to pass argument lists to program
so that this is safe even if VARIABLE is longer than kern.argmax.

There is another instance of CFILES which might need the same treatment,
and might be noticed when doing a "make links".

The same has to be done in RELENG_4 (on some different file).

Noticed-by: picobsd cross-compiling LINT
Suggested-by: Alfred (bright@mu.org), des@freebsd.org
MFC-after: 3 days

22 years agoFix a bug where a short write to a non-blocking socket would
John Birrell [Fri, 22 Feb 2002 04:26:54 +0000 (04:26 +0000)]
Fix a bug where a short write to a non-blocking socket would
leave the descriptor locked, causing other threads to hang
if they happened to access the socket.

MFC after: 5 days

22 years agoThis commit was generated by cvs2svn to compensate for changes in r91041,
David E. O'Brien [Fri, 22 Feb 2002 04:18:42 +0000 (04:18 +0000)]
This commit was generated by cvs2svn to compensate for changes in r91041,
which included commits to RCS files with non-trunk default branches.

22 years agoImport of Binutils from the FSF 2.12 branch (pre-.0 release).
David E. O'Brien [Fri, 22 Feb 2002 04:18:42 +0000 (04:18 +0000)]
Import of Binutils from the FSF 2.12 branch (pre-.0 release).

These bits are taken from the FSF anoncvs repo on 21-Feburary-2002 20:15 PST.

22 years ago- Massive style fixup.
Andrew R. Reiter [Fri, 22 Feb 2002 04:14:49 +0000 (04:14 +0000)]
- Massive style fixup.

Reviewed by: mike
Approved by: dfr

22 years agorecovery from new binutils part II: fix the alpha kernel so that
Andrew Gallatin [Thu, 21 Feb 2002 23:34:09 +0000 (23:34 +0000)]
recovery from new binutils part II: fix the alpha kernel so that
it handles R_ALPHA_RELATIVE relocs with addends properly.

tested by: ticso

22 years agoReorder some of the ioctls and add a few new ones.
Matt Jacob [Thu, 21 Feb 2002 23:30:05 +0000 (23:30 +0000)]
Reorder some of the ioctls and add a few new ones.

MFC after: 1 day

22 years agoo style(9) and consistency fix:
Chris D. Faulhaber [Thu, 21 Feb 2002 23:18:04 +0000 (23:18 +0000)]
o style(9) and consistency fix:
  - if (!var) -> if (var == NULL)
o spelling fix (althouh -> although)

Reviewed by: rwatson
Obtained from: TrustedBSD Project

22 years agoo style(9) and consistency fixes:
Chris D. Faulhaber [Thu, 21 Feb 2002 23:17:19 +0000 (23:17 +0000)]
o style(9) and consistency fixes:
  - if (!var) -> if (var == NULL)
  - return val; -> return (val);

Reviewed by: rwatson
Obtained from: TrustedBSD Project

22 years agoAdd more argument checking
Chris D. Faulhaber [Thu, 21 Feb 2002 23:13:06 +0000 (23:13 +0000)]
Add more argument checking

Reviewed by: rwatson
Obtained from: TrustedBSD Project

22 years agostatic'ize and declare functions
Chris D. Faulhaber [Thu, 21 Feb 2002 23:12:25 +0000 (23:12 +0000)]
static'ize and declare functions

Reviewed by: rwatson
Obtained from: TrustedBSD Project

22 years agoRevert wchan functionality. Add 'mwchan' to supply new duel mutex/msleep
Matthew Dillon [Thu, 21 Feb 2002 18:27:16 +0000 (18:27 +0000)]
Revert wchan functionality.  Add 'mwchan' to supply new duel mutex/msleep
functionality and make it the default.

With additional improvements by: Mark Peek <mp@FreeBSD.org>

22 years agostaticize functions and variables
Bill Fumerola [Thu, 21 Feb 2002 18:13:31 +0000 (18:13 +0000)]
staticize functions and variables
change function declarations to ANSI
change a variable that stores sizeof() values to size_t
use return to escape the end of main(), not exit(3)

22 years agouse the typedef 'ubig' to declare pr_limit, not what ubig is typedef'd
Bill Fumerola [Thu, 21 Feb 2002 17:33:56 +0000 (17:33 +0000)]
use the typedef 'ubig' to declare pr_limit, not what ubig is typedef'd
to be.

PR: misc/35181
Submitted by: Stefan Farfeleder <e0026813@stud3.tuwien.ac.at>
MFC after: 3 days

22 years agoAdd support for iovcnt greater than 1. This should resolve problems
Boris Popov [Thu, 21 Feb 2002 16:23:38 +0000 (16:23 +0000)]
Add support for iovcnt greater than 1. This should resolve problems
with some applications.

Obtained from: Darwin project
MFC after: 2 weeks

22 years agoRemove redundant checks for iovcnt > 1. This should be handled properly
Boris Popov [Thu, 21 Feb 2002 16:18:39 +0000 (16:18 +0000)]
Remove redundant checks for iovcnt > 1. This should be handled properly
in the subr_mchain.

Obtained from: Darwin project
MFC after: 2 weeks

22 years agoAdd unicode related definition for future use. Descibe Samba bug.
Boris Popov [Thu, 21 Feb 2002 16:13:19 +0000 (16:13 +0000)]
Add unicode related definition for future use. Descibe Samba bug.

Obtained from: Darwin project
MFC after: 2 weeks

22 years agoAdd subfunction definition for future use.
Boris Popov [Thu, 21 Feb 2002 16:10:42 +0000 (16:10 +0000)]
Add subfunction definition for future use.

MFC after: 2 weeks

22 years agoFix build when MAKE_KERBEROS4 is not requested. This was broken with
Jacques Vidrine [Thu, 21 Feb 2002 15:54:20 +0000 (15:54 +0000)]
Fix build when MAKE_KERBEROS4 is not requested.  This was broken with
the last Heimdal import.

Reported by: jhay, roam
(both the same day!)

22 years agoBring rc.firewall{,6} more in line with the word and spirit of
Crist J. Clark [Thu, 21 Feb 2002 13:14:19 +0000 (13:14 +0000)]
Bring rc.firewall{,6} more in line with the word and spirit of
rc.conf(5) and the files' inline documentation.

  - Add the "closed"-type, documented in both places, but which did not
    exist in the code.

  - When provided a ruleset, the system should not make any assumptions
    about the sites's policy and should add no rules of its own.

  - Make the "UNKNOWN" (documented in-line) actual work as advertised,
    load no rules.

Prodded by: Igor M Podlesny <poige@morning.ru>
MFC after: 1 week

22 years agoFixed printf format errors. In printgprof.c, also convert the scale
Bruce Evans [Thu, 21 Feb 2002 12:07:21 +0000 (12:07 +0000)]
Fixed printf format errors.  In printgprof.c, also convert the scale
without possibly losing lots of precision, and print the scale using
%g instead of %d in case it is non-integral.  %g might not be the best
format for this.

22 years agoAdded missing copyright. Obtain one and a vendor id from gprof.c. aout.c
Bruce Evans [Thu, 21 Feb 2002 12:03:06 +0000 (12:03 +0000)]
Added missing copyright.  Obtain one and a vendor id from gprof.c.  aout.c
was split off from gprof.c in rev.1.7 of the latter.  elf.c is mostly new,
but the old copyright sort of applies to it and is better than none.

Use __FBSDID() for the FreeBSD id.

22 years ago- Add support for Simplified Direct Access Device, mostly for
Hidetoshi Shimokawa [Thu, 21 Feb 2002 11:58:47 +0000 (11:58 +0000)]
- Add support for Simplified Direct Access Device, mostly for
Firewire/SBP-II devices.

- Add quirk for Logitec USB/Firewire HDD.

MFC after: 3 days.

22 years agoFixed a missing variable declaration so that gprof compiles with -DDEBUG.
Bruce Evans [Thu, 21 Feb 2002 10:34:52 +0000 (10:34 +0000)]
Fixed a missing variable declaration so that gprof compiles with -DDEBUG.

22 years agoObviate the need to set the COMPILER_PATH and LIBRARY_PATH in Makefile.inc1
Ruslan Ermilov [Thu, 21 Feb 2002 10:23:41 +0000 (10:23 +0000)]
Obviate the need to set the COMPILER_PATH and LIBRARY_PATH in Makefile.inc1
to pick up the correct cross-tools (the compiler executables and binutils)
and special linker files (crt*.o).  This is now controlled by a single knob,
TOOLS_PREFIX, when building cross-tools.

Fixed regression in Makefile.inc1,v 1.203 (-nostdinc).  This clobbered target
architecture's CFLAGS with building host's CPUTYPE setting in /etc/make.conf,
and had a nice but nasty side effect of exposing some (normally hidden) bugs
in system headers.

(Attempt to move the "-nostdinc -I..." part of CFLAGS into the new CINCLUDES
(modeled after a similar CXXINCLUDES) eventually failed because hard-coding
${WORLDTMP}/usr/include to be the first in the include list does not always
work, e.g. lib/libbind.)

Compensate the -nostdinc removal by making cpp(1) built in the cross-tools
stage to not look for <> header files in the building host's /usr/include
(already committed as gnu/usr.bin/cc/cc_tools/freebsd-native.h, revisions
1.10-1.12, STANDARD_INCLUDE_DIR).

: $ /usr/obj/usr/src/i386/usr/bin/cpp -v /dev/null
:
: Before:
:
: #include <...> search starts here:
:  /usr/obj/usr/src/i386/usr/include
:  /usr/include
: End of search list.
:
: After:
:
: #include <...> search starts here:
:  /usr/obj/usr/src/i386/usr/include
:  /usr/obj/usr/src/i386/usr/include

(Disabling the use of GCC_INCLUDE_DIR in the FREEBSD_NATIVE case would fix
the duplicate above.)

Get rid of the (now unneeded) -I${DESTDIR}/usr/include magic in bsd.prog.mk
and bsd.lib.mk.  Finish the removal of LDDESTDIR in bsd.lib.mk,v 1.55 -- we
no longer have users of it.

The required changes to gcc were already committed as contrib/gcc.295/gcc.c,
revisions 1.23 and 1.24.

Basically, this allows for the changes above plus makes gcc(1) persistent
about path configuration, whether it's configured as a native or a cross
compiler:

: $ /usr/obj/usr/src/i386/usr/bin/cc -print-search-dirs
: install: /usr/obj/usr/src/i386/usr/libexec/(null)
: programs: /usr/obj/usr/src/i386/usr/libexec/elf/:/usr/obj/usr/src/i386/usr/libexec/
: libraries: /usr/obj/usr/src/i386/usr/lib/
:
: $ /usr/obj/alpha/usr/src/i386/usr/bin/cc -print-search-dirs
: install: /usr/obj/alpha/usr/src/i386/usr/libexec/(null)
: programs: /usr/obj/alpha/usr/src/i386/usr/libexec/elf/:/usr/obj/alpha/usr/src/i386/usr/libexec/
: libraries: /usr/obj/alpha/usr/src/i386/usr/lib/

Reviewed by: bde, obrien

22 years agoMoved the definition of the machine-independent macro UNITS_TO_CODE
Bruce Evans [Thu, 21 Feb 2002 07:12:57 +0000 (07:12 +0000)]
Moved the definition of the machine-independent macro UNITS_TO_CODE
from <number of machines> machine-dependent headers to the one
non-header here it is used so that it is easier to fix.  This macro
just divides the machine-dependent offset OFFSET_OF_CODE by the
machine-independent scale factor sizeof(UNIT), as required for bug
for bug compatibility with the scaling of pc's in gprof.c.  UNIT is
the type of a profiling counter, and its size has nothing to do with
the correct scale factor except both are usually 2.

22 years agoReserved one of the spare fields in struct gmon to record the history
Bruce Evans [Thu, 21 Feb 2002 05:52:49 +0000 (05:52 +0000)]
Reserved one of the spare fields in struct gmon to record the history
counter type, as threatened in rev.1.8 (the density doesn't need to
be recorded since it can be derived from other fields).  This doesn't
affect binary compatibility, but new utilities won't be able to depend
on the contents of this field because libc/gmon/gmon.c was broken --
it wrote garbage to the spare fields.

Added a history counter type field to struct gmonparam.  This breaks
binary compatibility a little, since kgmon wanted to read the whole
struct.  Fixed kgmon to only depend on reading the critical earlier
parts of the struct.  This should also fix 6+ year old breakage of
binary compatibility when the profrate field was added.

Only initialize the new field in struct gmon for now, so that the
compatibility code for this (in kgmon) gets tested.  The compatibility
code has to guesstimate the value.  The new field in struct gmonparam
is for the kernel to initialize so that kgmon doesn't have to guess.

22 years agoDocument the hw.physmem kernel environment variable.
Chad David [Thu, 21 Feb 2002 05:15:52 +0000 (05:15 +0000)]
Document the hw.physmem kernel environment variable.

Reviewed by: alfred

22 years agoFixed some style bugs. Added a comment about a bug in PT_SSTEP.
Bruce Evans [Thu, 21 Feb 2002 04:47:38 +0000 (04:47 +0000)]
Fixed some style bugs.  Added a comment about a bug in PT_SSTEP.

Approved by: des

22 years agoRecover bits that were lost in transition in rev.1.76:
Bruce Evans [Thu, 21 Feb 2002 04:37:55 +0000 (04:37 +0000)]
Recover bits that were lost in transition in rev.1.76:
- P_INMEM checks in all the functions.  P_INMEM must be checked because
  PHOLD() is broken.  The old bits had bogus locking (using sched_lock)
  to lock P_INMEM.  After removing the P_INMEM checks, we were left with
  just the bogus locking.
- large comments.  They were too large, but better than nothing.

Remove obfuscations that were gained in transition in rev.1.76:
- PROC_REG_ACTION() is even more of an obfuscation than PROC_ACTION().

The change copies procfs_machdep.c rev.1.22 of i386/procfs_machdep.c
verbatim except for "fixing" the old-style function headers and adjusting
function names and comments.  It doesn't remove the bogus locking.

Approved by: des

22 years agoDon't define NOADDITIONAL in CFLAGS again, it is already in named.h
Andrey A. Chernov [Thu, 21 Feb 2002 04:20:47 +0000 (04:20 +0000)]
Don't define NOADDITIONAL in CFLAGS again, it is already in named.h
This commit unbreak 'make depend'

22 years agoUse new ID scheme.
Mike Barcroft [Thu, 21 Feb 2002 02:17:57 +0000 (02:17 +0000)]
Use new ID scheme.

Fix env(1)'s exit status to conform with SUSv3.

Submitted by: Tim Robbins <tim@robbins.dropbear.id.au>
MFC after: 2 weeks

22 years agoFix a problem where a local loop disk logs out- and we get a PORT LOGGED
Matt Jacob [Thu, 21 Feb 2002 01:56:08 +0000 (01:56 +0000)]
Fix a problem where a local loop disk logs out- and we get a PORT LOGGED
OUT status. We are, apparently, required to force the f/w to log back in
if we want to try and talk to that disk again. This means either issuing
a LOGIN LOCAL LOOP PORT mailbox command, or by issuing a LIP. I've elected
to issue a LIP because this has a better chance of waking up the disk which
clearly just crashed and burned.

These should not occur at all. If they do, they should be darned rare.

MFC after: 1 week

22 years agoCommit some infrastructure for turning on -Werror for kernel compiles.
Peter Wemm [Wed, 20 Feb 2002 23:35:56 +0000 (23:35 +0000)]
Commit some infrastructure for turning on -Werror for kernel compiles.
It doesn't actually do it yet though.  This adds a flag to config so
that we can exclude certain vendor files from this even when the rest
of the kernel has it on.  make -DNO_WERROR would also bypass all of it.

22 years agoMFen (1.281 --> 1.283).
Jun Kuriyama [Wed, 20 Feb 2002 22:41:10 +0000 (22:41 +0000)]
MFen (1.281 --> 1.283).

22 years agoOops, used wrong error value for unimplemented syscalls.
Julian Elischer [Wed, 20 Feb 2002 22:27:09 +0000 (22:27 +0000)]
Oops, used wrong error value for unimplemented syscalls.

22 years agoPass me the pointy hat please. Be sure to return a value in a non-void
Peter Wemm [Wed, 20 Feb 2002 22:25:54 +0000 (22:25 +0000)]
Pass me the pointy hat please.  Be sure to return a value in a non-void
function.  I've been running with this buried in the mountains of compiler
output for about a month on my desktop.

22 years agoTidy up some unused variables
Peter Wemm [Wed, 20 Feb 2002 21:25:44 +0000 (21:25 +0000)]
Tidy up some unused variables

22 years agoPrefix structure members to protect them against clashes with eg.
Alfred Perlstein [Wed, 20 Feb 2002 20:47:21 +0000 (20:47 +0000)]
Prefix structure members to protect them against clashes with eg.
c++ keywords.

This keeps us in sync with NetBSD because they actually committed
my delta first.

Ok'd by: lennard

22 years agoCleanup of nmdm device
Julian Elischer [Wed, 20 Feb 2002 20:13:13 +0000 (20:13 +0000)]
Cleanup of nmdm device

22 years agoAllow for better control over the GCC front-end when building a cross
David E. O'Brien [Wed, 20 Feb 2002 19:41:54 +0000 (19:41 +0000)]
Allow for better control over the GCC front-end when building a cross
compiler.

* Undo the diking out of cross compiler logic from gcc.c rev 1.16.
* Add the `CROSS_STARTFILE_PREFIX' knob.
* Add our own definition of `STANDARD_INCLUDE_DIR'.  This should have been
  included in freebsd-native.h rev 1.5.
* Minimize a little bit more, things we dike out in the FREEBSD_NATIVE case.

Submitted by: ru & obrien

22 years agoMF4S: Fix typo in mention of sbniconfig port.
Bruce A. Mah [Wed, 20 Feb 2002 19:25:03 +0000 (19:25 +0000)]
MF4S:  Fix typo in mention of sbniconfig port.

22 years agoAdd missing m_free() so we actually drain the send buffer in monitor mode.
Brooks Davis [Wed, 20 Feb 2002 18:23:59 +0000 (18:23 +0000)]
Add missing m_free() so we actually drain the send buffer in monitor mode.

Submitted by: Kenjiro Cho <kjc@csl.sony.co.jp>

22 years agoDelete a needless rule for DAD. An unspecified address is never used
Hajimu UMEMOTO [Wed, 20 Feb 2002 18:05:44 +0000 (18:05 +0000)]
Delete a needless rule for DAD.  An unspecified address is never used
as a destination address of IPv6 packets.

Submitted by: cjc
MFC after: 1 week

22 years agoBUGFIX: make use of the pointer to the target of skipto rules,
Luigi Rizzo [Wed, 20 Feb 2002 17:15:57 +0000 (17:15 +0000)]
BUGFIX: make use of the pointer to the target of skipto rules,
so that after the first time we can follow the pointer instead
of having to scan the list.
This was the intended behaviour from day one.

PR: 34639
MFC-after: 3 days

22 years agoMake sortdeps() working with dependency lists containing one or no elements.
Maxim Sobolev [Wed, 20 Feb 2002 17:15:17 +0000 (17:15 +0000)]
Make sortdeps() working with dependency lists containing one or no elements.

Submitted by: roam
MFC in: 1 week

22 years ago(forced)
Maxim Sobolev [Wed, 20 Feb 2002 17:11:33 +0000 (17:11 +0000)]
(forced)

Previous delta was a subject of:

MFC in: 1 week

22 years agoProvide protection against incorrect input (dependency list).
Maxim Sobolev [Wed, 20 Feb 2002 17:07:19 +0000 (17:07 +0000)]
Provide protection against incorrect input (dependency list).

Submitted by: roam

22 years agofix file descriptor leak.
Alfred Perlstein [Wed, 20 Feb 2002 17:06:37 +0000 (17:06 +0000)]
fix file descriptor leak.

Submitted by: Mark Santcroos <marks@ripe.net>

22 years agoWhen expanding a syncache entry into a socket, inherit the socket options
Jonathan Lemon [Wed, 20 Feb 2002 16:47:11 +0000 (16:47 +0000)]
When expanding a syncache entry into a socket, inherit the socket options
from the current listen socket instead of the cached (and possibly stale)
TCB pointer.

22 years agoCheck the status of the card bridge first thing, rather than last in
Warner Losh [Wed, 20 Feb 2002 16:20:27 +0000 (16:20 +0000)]
Check the status of the card bridge first thing, rather than last in
the loop.  This fixes the "my card is in the laptop on boot, but
doesn't attach" problem.

22 years ago- Fix style further by adding parentheses around return values so that
Andrew R. Reiter [Wed, 20 Feb 2002 16:05:30 +0000 (16:05 +0000)]
- Fix style further by adding parentheses around return values so that
  they look like:
return (val);  instead of:  return val;

22 years agoDocument NGM_PPPOE_ACNAME.
Brian Somers [Wed, 20 Feb 2002 15:55:02 +0000 (15:55 +0000)]
Document NGM_PPPOE_ACNAME.

22 years agoDocument NGM_PPPOE_ACNAME handling.
Brian Somers [Wed, 20 Feb 2002 15:53:47 +0000 (15:53 +0000)]
Document NGM_PPPOE_ACNAME handling.

22 years agoHandle NGM_PPPOE_ACNAME messages.
Brian Somers [Wed, 20 Feb 2002 15:52:20 +0000 (15:52 +0000)]
Handle NGM_PPPOE_ACNAME messages.

Submitted by: Andre Albsmeier <andre@albsmeier.net>
Approved by: julian

22 years agoSend a NGM_PPPOE_ACNAME message to userland when a node is connected.
Brian Somers [Wed, 20 Feb 2002 15:51:11 +0000 (15:51 +0000)]
Send a NGM_PPPOE_ACNAME message to userland when a node is connected.

Submitted by: Andre Albsmeier <andre@albsmeier.net>
Shuffled about by: brian
Approved by: julian