]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoAdd a comment indicating that the vnode locking in this section of the
Robert Watson [Sun, 10 Feb 2002 21:29:02 +0000 (21:29 +0000)]
Add a comment indicating that the vnode locking in this section of the
kernel linker code may be wrong: it fails to hold a lock across the
call to VOP_GETATTR(), and vn_rdwr() with IO_NODELOCKED.

22 years agoDon't make it seem like vm.max_proc_mmap only affects MAP_FIXED.
Dima Dorfman [Sun, 10 Feb 2002 21:07:56 +0000 (21:07 +0000)]
Don't make it seem like vm.max_proc_mmap only affects MAP_FIXED.

PR: 34005
Submitted by: Steven Grady <grady@digitaldeck..com>,
Hiten Pandya <hitmaster2k@yahoo.com>

22 years agoReinforce locking requirement for VOP_REVOKE: the lock must not be held.
Robert Watson [Sun, 10 Feb 2002 20:48:22 +0000 (20:48 +0000)]
Reinforce locking requirement for VOP_REVOKE: the lock must not be held.

Discussed with: phk

22 years agoMake sure to grab vnode lock on a vnode before calling VOP_GETATTR()
Robert Watson [Sun, 10 Feb 2002 20:45:43 +0000 (20:45 +0000)]
Make sure to grab vnode lock on a vnode before calling VOP_GETATTR()
to perform an ownership test in revoke().  This is also required for
MAC hooks so that the vnode lock is held during a call to the MAC
framework.  Release the lock before calling VOP_REVOKE().

Discussed with: phk, mckusick

22 years agoNice set of fixes to use SCRIPT instead of PROG, thus fixing up
Mark Murray [Sun, 10 Feb 2002 19:48:19 +0000 (19:48 +0000)]
Nice set of fixes to use SCRIPT instead of PROG, thus fixing up
a lot of nasty STRIP= problems. This has the added side effect
if neatening up some leaf makefiles very nicely.

Submitted by: ru

22 years agoRevert 1.14: VOP_GETATTR() really does require a vnode lock.
Robert Watson [Sun, 10 Feb 2002 19:28:34 +0000 (19:28 +0000)]
Revert 1.14: VOP_GETATTR() really does require a vnode lock.

Confirmed by: mckusick

22 years agoStyle consistency fix for MAC ioctls added previously.
Robert Watson [Sun, 10 Feb 2002 16:55:35 +0000 (16:55 +0000)]
Style consistency fix for MAC ioctls added previously.

22 years agoFill in the uhci_dump_ii function (from NetBSD).
Josef Karthauser [Sun, 10 Feb 2002 15:38:47 +0000 (15:38 +0000)]
Fill in the uhci_dump_ii function (from NetBSD).

22 years agoAdd this FreeBSD standard header.
David E. O'Brien [Sun, 10 Feb 2002 14:27:20 +0000 (14:27 +0000)]
Add this FreeBSD standard header.

22 years agoAdd some system code names. Explicitely list a few unsupported hardware
Wilko Bulte [Sun, 10 Feb 2002 13:21:18 +0000 (13:21 +0000)]
Add some system code names. Explicitely list a few unsupported hardware
types that gave rise to emailed questions. Add quick note on Miata CPU
upgradeability.

22 years agoUpdate this to 5.006.
David E. O'Brien [Sun, 10 Feb 2002 12:37:52 +0000 (12:37 +0000)]
Update this to 5.006.

22 years agoCosmetic changes:
KATO Takenori [Sun, 10 Feb 2002 11:23:14 +0000 (11:23 +0000)]
Cosmetic changes:
- Collected i486 identification codes in one place like
  586 and 686.
- Merged two cases (0x470 and 0x490) for `Enhanced Am486DX4
  Write-Back.'
- Replaced `unknown' into `Unknown'.

Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

22 years agoCosmetic changes.
Yoshihiro Takahashi [Sun, 10 Feb 2002 10:27:37 +0000 (10:27 +0000)]
Cosmetic changes.

22 years agoAdd needed include.
Yoshihiro Takahashi [Sun, 10 Feb 2002 10:16:22 +0000 (10:16 +0000)]
Add needed include.

22 years ago- Refine the iskanji1 function.
Yoshihiro Takahashi [Sun, 10 Feb 2002 10:15:41 +0000 (10:15 +0000)]
- Refine the iskanji1 function.
- Print continuous ascii characters at a time.

Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

22 years agoFixed mouse cursor on a console.
Yoshihiro Takahashi [Sun, 10 Feb 2002 10:14:39 +0000 (10:14 +0000)]
Fixed mouse cursor on a console.

Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

22 years agomemrange.h is in sys/.
Dima Dorfman [Sun, 10 Feb 2002 08:52:25 +0000 (08:52 +0000)]
memrange.h is in sys/.

PR: 34460
Submitted by: Bruce Dang <bruce@research.teamxor.org>, roam

22 years agocrdup(9) is not a protocol.
Dima Dorfman [Sun, 10 Feb 2002 08:19:58 +0000 (08:19 +0000)]
crdup(9) is not a protocol.

PR: 34624
Submitted by: John Nielsen <nielsenj@cs.byu.edu>,
Hiten Pandya <hiten@uk.FreeBSD.org>

22 years agoUpdate license.
Robert Watson [Sun, 10 Feb 2002 06:14:03 +0000 (06:14 +0000)]
Update license.

22 years agoTeach getextattr to query the EA size and allocate appropriate sized
Robert Watson [Sun, 10 Feb 2002 06:13:14 +0000 (06:13 +0000)]
Teach getextattr to query the EA size and allocate appropriate sized
buffers before reading the memory.  Arguably, the failure modes here
are poor, but we can now read >2k EAs.  Also, update the copyrights
and licenses while I'm here.

22 years agoUse the getprogname(3) function instead of directly accessing
Mike Barcroft [Sun, 10 Feb 2002 05:56:36 +0000 (05:56 +0000)]
Use the getprogname(3) function instead of directly accessing
`__progname'.

Submitted by: dd

22 years agoRemove a stray 'const' that slept into extattr_set_vp(), and could
Robert Watson [Sun, 10 Feb 2002 05:31:55 +0000 (05:31 +0000)]
Remove a stray 'const' that slept into extattr_set_vp(), and could
result in compiler warnings.

22 years agoNull interrupt handlers should be OK, so if we don't have a function,
Warner Losh [Sun, 10 Feb 2002 05:04:20 +0000 (05:04 +0000)]
Null interrupt handlers should be OK, so if we don't have a function,
just return.

22 years agoMinor style tweaks.
Robert Watson [Sun, 10 Feb 2002 04:57:08 +0000 (04:57 +0000)]
Minor style tweaks.

Remove an unneeded comment and commented out code that won't be
needed.

22 years agoCopyright + license update.
Robert Watson [Sun, 10 Feb 2002 04:50:24 +0000 (04:50 +0000)]
Copyright + license update.

22 years agoUpdate userland tools to reflect extattr API changes.
Robert Watson [Sun, 10 Feb 2002 04:48:26 +0000 (04:48 +0000)]
Update userland tools to reflect extattr API changes.

Note that getextattr has not yet been updated to dynamically allocate
a read buffer, although that can now be done.

Obtained from: TrustedBSD Project
Sponsored by: NAI Labs

22 years agoPart III: Update extended attribute system call interface documentation.
Robert Watson [Sun, 10 Feb 2002 04:46:28 +0000 (04:46 +0000)]
Part III: Update extended attribute system call interface documentation.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years agoPart II: Update system calls for extended attributes. Rebuild of
Robert Watson [Sun, 10 Feb 2002 04:44:37 +0000 (04:44 +0000)]
Part II: Update system calls for extended attributes.  Rebuild of
         generated files.

22 years agoPart I: Update extended attribute API and ABI:
Robert Watson [Sun, 10 Feb 2002 04:43:22 +0000 (04:43 +0000)]
Part I: Update extended attribute API and ABI:

o Modify the system call syntax for extattr_{get,set}_{fd,file}() so
  as not to use the scatter gather API (which appeared not to be used
  by any consumers, and be less portable), rather, accepts 'data'
  and 'nbytes' in the style of other simple read/write interfaces.
  This changes the API and ABI.

o Modify system call semantics so that extattr_get_{fd,file}() return
  a size_t.  When performing a read, the number of bytes read will
  be returned, unless the data pointer is NULL, in which case the
  number of bytes of data are returned.  This changes the API only.

o Modify the VOP_GETEXTATTR() vnode operation to accept a *size_t
  argument so as to return the size, if desirable.  If set to NULL,
  the size will not be returned.

o Update various filesystems (pseodofs, ufs) to DTRT.

These changes should make extended attributes more useful and more
portable.  More commits to rebuild the system call files, as well
as update userland utilities to follow.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years agoThis commit was generated by cvs2svn to compensate for changes in r90446,
Mark Peek [Sun, 10 Feb 2002 04:40:26 +0000 (04:40 +0000)]
This commit was generated by cvs2svn to compensate for changes in r90446,
which included commits to RCS files with non-trunk default branches.

22 years ago- Add support for autodetection of *.euc?? and *.EUC locales.
Mark Peek [Sun, 10 Feb 2002 04:40:26 +0000 (04:40 +0000)]
- Add support for autodetection of *.euc?? and *.EUC locales.
  - ja_JP.eucJP, ja_JP.EUC
  - ko_KR.eucKR, ko_KR.EUC
  - zn_CN.eucCN, zn_CN.EUC
- Add support for zh_TW.Big5
- Add ja_JP.Shift_JIS alias for ja_JP.SJIS

Submitted by: phantom
Approved by: Christos Zoulas <cristos@zoulas.com> - TCSH maintainer

22 years agoo Use bus_generic_setup_intr instead of bus_setup_intr.
Warner Losh [Sun, 10 Feb 2002 03:34:44 +0000 (03:34 +0000)]
o Use bus_generic_setup_intr instead of bus_setup_intr.
o Call bus_generic_setup_intr and check its return value.  Don't setup
  func until we successfully get the interrupt from our parent.
o Add comments about some maybe questionable stuff so I can check later
  to make sure that it really is that way.

22 years agoo Use INTR_TYPE_AV (the highest possible) rather than INTR_TYPE_BIO
Warner Losh [Sun, 10 Feb 2002 03:28:37 +0000 (03:28 +0000)]
o Use INTR_TYPE_AV (the highest possible) rather than INTR_TYPE_BIO
o Don't allow INTR_TYPE_FAST.  Since we are sharing the interrupt between
  CSC and the functions, they can't be FAST because fast interrupts can't
  be shared.
o Add the same workaround for resume that we have in OLDCARD.
o Also, return the error from bus_generic_resume rather than ignoring it.

22 years agoIt's OK to be pedantic, but long long literals are OK.
Poul-Henning Kamp [Sun, 10 Feb 2002 02:30:53 +0000 (02:30 +0000)]
It's OK to be pedantic, but long long literals are OK.

22 years agoo No longer mount /proc by default on newly installed systems. Almost
Robert Watson [Sun, 10 Feb 2002 01:34:04 +0000 (01:34 +0000)]
o No longer mount /proc by default on newly installed systems.  Almost
  all facilities that previously relied on /proc have been rewritten
  to use ptrace().  procfs has presented a substantial security
  hazard for years, with several user->root compromises in the last
  few years.  Procfs will continue to be available but will require
  administrator intervention to use.

Reviewed by: scottl, jedgar, mike, tmm

22 years agoRemove di_inumber since LFS is long gone.
Poul-Henning Kamp [Sun, 10 Feb 2002 00:55:49 +0000 (00:55 +0000)]
Remove di_inumber since LFS is long gone.

22 years agoBoot verbosify printing the CIS, since we don't normally need to do that.
Warner Losh [Sat, 9 Feb 2002 21:34:06 +0000 (21:34 +0000)]
Boot verbosify printing the CIS, since we don't normally need to do that.

22 years agoMake unsupported memory range message bootverbose only
Warner Losh [Sat, 9 Feb 2002 21:32:02 +0000 (21:32 +0000)]
Make unsupported memory range message bootverbose only

22 years agoFix MAKEDEV for RocketPort (rp(4)) cuaR* and ttyR* to work with the
John Baldwin [Sat, 9 Feb 2002 21:16:54 +0000 (21:16 +0000)]
Fix MAKEDEV for RocketPort (rp(4)) cuaR* and ttyR* to work with the
updated driver.  The newer driver in current outputs a version string
that contains a space, so we need to eat two words in between RocketPortX
and the number of ports on the board.

22 years agoThis has been sitting in my local tree long enough. Remove the use
Daniel Eischen [Sat, 9 Feb 2002 19:58:41 +0000 (19:58 +0000)]
This has been sitting in my local tree long enough.  Remove the use
of an alternate signal stack for handling signals.  Let the kernel
send signals on the stack of the current thread and teach the threads
signal handler how to deliver signals to the current thread if it
needs to.  Also, always store a threads context as a jmp_buf.  Eventually
this will change to be a ucontext_t or mcontext_t.

Other small nits.  Use struct pthread * instead of pthread_t in internal
library routines.  The threads code wants struct pthread *, and pthread_t
doesn't necessarily have to be the same.

Reviewed by: jasone

22 years agoSilence a warning by initializing the return value. This wasn't
Daniel Eischen [Sat, 9 Feb 2002 19:48:55 +0000 (19:48 +0000)]
Silence a warning by initializing the return value.  This wasn't
set in the case of a short write, and I think returning 0 for this
is what was intended.

22 years agoThis file needs <syslog.h>.
Dag-Erling Smørgrav [Sat, 9 Feb 2002 14:12:09 +0000 (14:12 +0000)]
This file needs <syslog.h>.

Sponsored by: DARPA, NAI Labs

22 years agoRecognize VIA C3 Samuel 2.
KATO Takenori [Sat, 9 Feb 2002 05:18:01 +0000 (05:18 +0000)]
Recognize VIA C3 Samuel 2.

MFC after: 3 days

22 years agoMFC noted: an(4) LEAP, Home key, and Linux utility support.
Bruce A. Mah [Sat, 9 Feb 2002 01:38:23 +0000 (01:38 +0000)]
MFC noted:  an(4) LEAP, Home key, and Linux utility support.

22 years agoReplace accidentally removed setrunqueue()
Julian Elischer [Sat, 9 Feb 2002 01:38:16 +0000 (01:38 +0000)]
Replace accidentally removed setrunqueue()
solves problem with machines failing to sync in booting.
Submitted by: Tor.Egge@cvsup.no.freebsd.org

22 years agoUse the mtx_owner() macro in one spot in _mtx_lock_sleep() to make the
John Baldwin [Sat, 9 Feb 2002 00:12:53 +0000 (00:12 +0000)]
Use the mtx_owner() macro in one spot in _mtx_lock_sleep() to make the
code easier to read.

22 years agoAdd missing "nullok" option to pam_unix.
Dag-Erling Smørgrav [Fri, 8 Feb 2002 23:27:22 +0000 (23:27 +0000)]
Add missing "nullok" option to pam_unix.

22 years agoRemove NO_WERRORs and WARNS=n's. To be revisited after GCC3.
Mark Murray [Fri, 8 Feb 2002 23:07:37 +0000 (23:07 +0000)]
Remove NO_WERRORs and WARNS=n's. To be revisited after GCC3.

22 years agoRemove leaf node WARNS?=2 (that mainly I added). This should
Mark Murray [Fri, 8 Feb 2002 22:31:43 +0000 (22:31 +0000)]
Remove leaf node WARNS?=2 (that mainly I added). This should
help the GCC3 transition and CURRENT in general.

22 years agoDon't hardcode paths. This is generally silly, and in this case
Mark Murray [Fri, 8 Feb 2002 21:28:07 +0000 (21:28 +0000)]
Don't hardcode paths. This is generally silly, and in this case
breaks custom installworlds.

Pointed out by: bde

22 years agoFix a bug introduced in r. 1.28: when copy{in,out} would fail for an
Thomas Moestl [Fri, 8 Feb 2002 20:19:44 +0000 (20:19 +0000)]
Fix a bug introduced in r. 1.28: when copy{in,out} would fail for an
iovec that was not the last one in the uio, the error would be ignored
silently.

Bug found and fix proposed by: jhb

22 years agoCorrect email address for bug reports and patch submissions.
David E. O'Brien [Fri, 8 Feb 2002 19:45:58 +0000 (19:45 +0000)]
Correct email address for bug reports and patch submissions.

Submitted by: Marco Molteni <molter@tin.it>

22 years agoApparently during the KSE M2 commit bzero() on the i386 was changed so that
John Baldwin [Fri, 8 Feb 2002 19:16:47 +0000 (19:16 +0000)]
Apparently during the KSE M2 commit bzero() on the i386 was changed so that
it's first parameter was volatile.  Catch i486_bzero() and i586_bzero()'s
prototypes up to this to quiet warnings.

22 years agoDon't grab the ICU lock while reading the current pending interrupts and
John Baldwin [Fri, 8 Feb 2002 18:30:36 +0000 (18:30 +0000)]
Don't grab the ICU lock while reading the current pending interrupts and
current masked interrupts from the AT PIC.

Requested by: bde

22 years agoTurn SMP back on by default in GENERIC on the Alpha.
John Baldwin [Fri, 8 Feb 2002 18:25:07 +0000 (18:25 +0000)]
Turn SMP back on by default in GENERIC on the Alpha.

22 years agoTurn interrupt thread preemption back on on the alpha.
John Baldwin [Fri, 8 Feb 2002 18:23:46 +0000 (18:23 +0000)]
Turn interrupt thread preemption back on on the alpha.

Tested by: mjacob

22 years agoNow that cross-tools ld(1) has been fixed to look for dynamic
Ruslan Ermilov [Fri, 8 Feb 2002 13:42:58 +0000 (13:42 +0000)]
Now that cross-tools ld(1) has been fixed to look for dynamic
dependencies in the correct place, record the fact that -lssh
depends on -lcrypto and -lz.

Removed false dependencies on -lz (except ssh(1) and sshd(8)).
Removed false dependencies on -lcrypto and -lutil for scp(1).

Reviewed by: markm

22 years agoReword the ``Please remove the FreeBSD fixit CDROM now.'' to
Brian Somers [Fri, 8 Feb 2002 13:42:35 +0000 (13:42 +0000)]
Reword the ``Please remove the FreeBSD fixit CDROM now.'' to
``Please remove the FreeBSD fixit CDROM/DVD now.''.

MFC after: 2 weeks

22 years agopeter points out that we probably should not mess with the sysctl(8)
Crist J. Clark [Fri, 8 Feb 2002 13:25:33 +0000 (13:25 +0000)]
peter points out that we probably should not mess with the sysctl(8)
values at all if they are not purposefully set. What if the
administrator messed with them in /etc/sysctl.conf? We don't want to
overwrite them.

If 'log_in_vain' is zero, do not force the issue. If it is non-zero,
set it.

22 years agoFreeBSD 4.1 bootstrapping aid (setproctitle(3) is in -lutil).
Ruslan Ermilov [Fri, 8 Feb 2002 13:09:08 +0000 (13:09 +0000)]
FreeBSD 4.1 bootstrapping aid (setproctitle(3) is in -lutil).

This backs out (sort of) delta 1.18 to perl/miniperl/Makefile.
Update to the ld(1) comment by peter in this revision:

ld(1) built as part of the cross-tools stage of buildworld has
been fixed to look for dynamic dependencies in the right place,
${WORLDTMP}/usr/lib, effective binutils/ld/Makefile,v 1.20.

Approved by: markm

22 years agoFix the code that selects the default binary type if the actual type can't
Dag-Erling Smørgrav [Fri, 8 Feb 2002 12:42:55 +0000 (12:42 +0000)]
Fix the code that selects the default binary type if the actual type can't
be determined.

PR: bin/34698
Submitted by: (in part) Stefan Farfeleder <e0026813@stud3.tuwien.ac.at>
MFC after: 1 weeks

22 years agoFixed RCSLOCALID buffer overflow (by making an overflow condition fatal).
Ruslan Ermilov [Fri, 8 Feb 2002 11:57:43 +0000 (11:57 +0000)]
Fixed RCSLOCALID buffer overflow (by making an overflow condition fatal).

Spotted by: John Johnson <bio.metrix@gte.net>

22 years agoMFi386: revisions 1.493 and 1.494.
Yoshihiro Takahashi [Fri, 8 Feb 2002 11:14:16 +0000 (11:14 +0000)]
MFi386: revisions 1.493 and 1.494.

22 years agoRegen.
Sheldon Hearn [Fri, 8 Feb 2002 11:04:50 +0000 (11:04 +0000)]
Regen.

22 years agoFreeBSD 4.1 bootstrapping aid (``:C'' make(1) modifier used
Ruslan Ermilov [Fri, 8 Feb 2002 09:48:34 +0000 (09:48 +0000)]
FreeBSD 4.1 bootstrapping aid (``:C'' make(1) modifier used
by gnu/usr.bin/binutils/as/Makefile.inc0 is not there).

XXX: make(1) should return with a non-zero exit status here.

22 years agoFreeBSD 4.1 bootstrapping aid (HALT and PDWN are not defined there).
Ruslan Ermilov [Fri, 8 Feb 2002 09:37:12 +0000 (09:37 +0000)]
FreeBSD 4.1 bootstrapping aid (HALT and PDWN are not defined there).

22 years agoFreeBSD 4.1 bootstrapping aid (__FBSDID is not there).
Ruslan Ermilov [Fri, 8 Feb 2002 09:34:17 +0000 (09:34 +0000)]
FreeBSD 4.1 bootstrapping aid (__FBSDID is not there).

22 years agoFix broken Giant locking protocol introduced in rev 1.114. You cannot
Peter Wemm [Fri, 8 Feb 2002 09:16:57 +0000 (09:16 +0000)]
Fix broken Giant locking protocol introduced in rev 1.114.  You cannot
unlock Giant if it is not locked in the first place.  This make the
nfstat(2) syscall (#278) a nice panic(2) implementation.

22 years agoBah, I managed to turn cosmetic things into real bugs. Fix shadowed
Peter Wemm [Fri, 8 Feb 2002 08:56:01 +0000 (08:56 +0000)]
Bah, I managed to turn cosmetic things into real bugs.  Fix shadowed
variable declarations. :-(  Definately not my day today.

22 years agoGRR. Unbreak WARNS=2.
Peter Wemm [Fri, 8 Feb 2002 07:52:51 +0000 (07:52 +0000)]
GRR.  Unbreak WARNS=2.
cc1: warnings being treated as errors
usr.bin/finger/lprint.c: In function `lprint':
usr.bin/finger/lprint.c:191: warning: precision is not type int (arg 2)
*** Error code 1

Pointy hat to: markm

22 years agoGRR. This was not WARNS=2 clean. You cannot printf a 'long' with
Peter Wemm [Fri, 8 Feb 2002 07:49:13 +0000 (07:49 +0000)]
GRR.  This was not WARNS=2 clean.    You cannot printf a 'long' with
'%qd'.

usr.bin/du/du.c:288: warning: long long int format, long int arg (arg 2)

Pointy hat to: markm

22 years agoRemove bogus range restrictions that attempted to restrict the range
Warner Losh [Fri, 8 Feb 2002 07:31:02 +0000 (07:31 +0000)]
Remove bogus range restrictions that attempted to restrict the range
of I/O in 1.5.  It looks like I got it right only for some of the
cases.  Instead, allow ISA addresses as a special case.  Most PCI
bridges decode this range.  I need to investigate PCI bridges better
to know if this is always true or not, but for now assume that it is
since that seems to be the most common case.

# We need to allocate addresses better for the pccard stuff...

Submitted by: phk, mitsunaga-san

22 years agoo Merge various recent fixes from the MAC branch relating to extattrctl():
Robert Watson [Fri, 8 Feb 2002 05:58:41 +0000 (05:58 +0000)]
o Merge various recent fixes from the MAC branch relating to extattrctl():
- Fix null-pointer dereference introduced when snapshotting
  was introduced.  This occured because unlike the previous code,
  vn_start_write() doesn't always return a non-NULL mp, as
  filesystems may not support the VOP_GETWRITEMOUNT() call.  For
  now, rely on two pointers, so that vn_finished_write() works
  properly.
- Fix locking problems on exit, introduced at some past time,
  some when snapshots came in, where a vnode might not be
  unlocked before being vrele'd in various error situations.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years agoAdd missing &
Peter Wemm [Fri, 8 Feb 2002 04:17:33 +0000 (04:17 +0000)]
Add missing &

22 years agoInspect ELF header and reject any non-FreeBSD shared objects.
Maxim Sobolev [Fri, 8 Feb 2002 04:08:57 +0000 (04:08 +0000)]
Inspect ELF header and reject any non-FreeBSD shared objects.

MFC after: 2 weeks

22 years agoFix world breakage (compiling libstdc++):
Peter Wemm [Fri, 8 Feb 2002 03:55:37 +0000 (03:55 +0000)]
Fix world breakage (compiling libstdc++):
sys/time.h:137: integer constant out of range
sys/time.h:137: warning: decimal integer constant is so large that it is unsigned
sys/time.h:153: integer constant out of range
sys/time.h:153: warning: decimal integer constant is so large that it is unsigned

22 years agoMake this part compile and try and use the p_threads tailq. It might work
Peter Wemm [Fri, 8 Feb 2002 03:19:56 +0000 (03:19 +0000)]
Make this part compile and try and use the p_threads tailq.  It might work
but I am not sure, I haven't been able to compile the rest of gdb yet.

22 years agoadd missing <string.h> to give strcpy() a prototype
Andrew Gallatin [Fri, 8 Feb 2002 03:01:03 +0000 (03:01 +0000)]
add missing <string.h> to give strcpy() a prototype

22 years agoFix a fatal trap when using ksched_setscheduler() (eg: mozilla, netscape
Peter Wemm [Fri, 8 Feb 2002 02:56:10 +0000 (02:56 +0000)]
Fix a fatal trap when using ksched_setscheduler() (eg: mozilla, netscape
etc) which use:  td->td_last_kse->ke_flags |= KEF_NEEDRESCHED;

22 years agoAdd a helper script used to take an Intel-supplied ACPI CA drop and make it
Mike Smith [Fri, 8 Feb 2002 02:16:45 +0000 (02:16 +0000)]
Add a helper script used to take an Intel-supplied ACPI CA drop and make it
ready for importing.  This should save me from my own over-cleanliness.

22 years agoAdd -i option (SUSv3) to patch(1). This allows one to specify a
Mike Barcroft [Fri, 8 Feb 2002 02:05:32 +0000 (02:05 +0000)]
Add -i option (SUSv3) to patch(1).   This allows one to specify a
patch file on command line instead of stdin.

Submitted by: Joseph Mallett <jmallett@xMach.org>
MFC after: 1 month

22 years agoremove superfluous blank line
Julian Elischer [Fri, 8 Feb 2002 01:38:32 +0000 (01:38 +0000)]
remove superfluous blank line

22 years agoo __P removal
Warner Losh [Thu, 7 Feb 2002 23:57:01 +0000 (23:57 +0000)]
o __P removal
o register removal
o use new style prototypes and function definitions

22 years agoFix a couple of style bugs introduced (or touched by) previous commit.
Peter Wemm [Thu, 7 Feb 2002 23:06:26 +0000 (23:06 +0000)]
Fix a couple of style bugs introduced (or touched by) previous commit.

22 years agoFix a whole bunch of long lines introduced by previous commit by using
Peter Wemm [Thu, 7 Feb 2002 23:05:40 +0000 (23:05 +0000)]
Fix a whole bunch of long lines introduced by previous commit by using
td = FIRST_THREAD_IN_PROC(p) once, after we have identified the process
that we are operating on.

22 years agoFix a long line touched in previous commit (but not caused by previous
Peter Wemm [Thu, 7 Feb 2002 23:03:41 +0000 (23:03 +0000)]
Fix a long line touched in previous commit (but not caused by previous
commit)

22 years agoAttempt to patch up some style bugs introduced in the previous commit
Peter Wemm [Thu, 7 Feb 2002 22:40:34 +0000 (22:40 +0000)]
Attempt to patch up some style bugs introduced in the previous commit

22 years agoAttempt to unmangle some code touched in the previous commit.
Peter Wemm [Thu, 7 Feb 2002 22:36:19 +0000 (22:36 +0000)]
Attempt to unmangle some code touched in the previous commit.

22 years agoFix a long line introduced in previous commit.
Peter Wemm [Thu, 7 Feb 2002 22:29:41 +0000 (22:29 +0000)]
Fix a long line introduced in previous commit.

22 years agoFix long line added in previous commit.
Peter Wemm [Thu, 7 Feb 2002 22:28:24 +0000 (22:28 +0000)]
Fix long line added in previous commit.

22 years agoPatch up some long lines introduced in previous commit.
Peter Wemm [Thu, 7 Feb 2002 22:26:51 +0000 (22:26 +0000)]
Patch up some long lines introduced in previous commit.

22 years agoFix style bug introduced in previous commit.
Peter Wemm [Thu, 7 Feb 2002 22:26:07 +0000 (22:26 +0000)]
Fix style bug introduced in previous commit.

22 years agoOccationally background fsck would cause a spurious ``freeing free
Kirk McKusick [Thu, 7 Feb 2002 22:13:56 +0000 (22:13 +0000)]
Occationally background fsck would cause a spurious ``freeing free
inode'' panic. This change corrects that problem by setting the
fs_active flag when the inode map changes to notify the snapshot
code that the cylinder group must be rescanned.

Submitted by: Robert Watson <rwatson@FreeBSD.org>

22 years agoReplace a few more lines of code orphaned by MFREE going away.
Matt Jacob [Thu, 7 Feb 2002 22:08:32 +0000 (22:08 +0000)]
Replace a few more lines of code orphaned by MFREE going away.

MFC after: 1 day

22 years agoYou cannot take the address of a variable with the 'register' qualifier.
Matt Jacob [Thu, 7 Feb 2002 22:06:23 +0000 (22:06 +0000)]
You cannot take the address of a variable with the 'register' qualifier.

MFC after: 1 day

22 years agoRemove the inline assembly optimized versins of bintime_*() functions,
Poul-Henning Kamp [Thu, 7 Feb 2002 21:57:28 +0000 (21:57 +0000)]
Remove the inline assembly optimized versins of bintime_*() functions,
they have no measurable effect and they don't belong in an MI header
even if they did.

22 years agoRevise timercounters to use binary fixed point format internally.
Poul-Henning Kamp [Thu, 7 Feb 2002 21:21:55 +0000 (21:21 +0000)]
Revise timercounters to use binary fixed point format internally.

The binary format "bintime" is a 32.64 format, it will go to 64.64
when time_t does.

The bintime format is available to consumers of time in the kernel,
and is preferable where timeintervals needs to be accumulated.

This change simplifies much of the magic math inside the timecounters
and improves the frequency and time precision by a couple of bits.

I have not been able to measure a performance difference which was not
a tiny fraction of the standard deviation on the measurements.

22 years agoPre-KSE/M3 commit.
Julian Elischer [Thu, 7 Feb 2002 20:58:47 +0000 (20:58 +0000)]
Pre-KSE/M3 commit.
this is a low-functionality change that changes the kernel to access the main
thread of a process via the linked list of threads rather than
assuming that it is embedded in the process. It IS still embeded there
but remove all teh code that assumes that in preparation for the next commit
which will actually move it out.

Reviewed by: peter@freebsd.org, gallatin@cs.duke.edu, benno rice,

22 years agopre-emptively fix a KSE/M3 problem.
Julian Elischer [Thu, 7 Feb 2002 20:28:25 +0000 (20:28 +0000)]
pre-emptively fix a KSE/M3 problem.

Make a slight change so that libkvm reaches the main thread via the
linked list, rather than assuming it is in the proc structure. Both
conditions are true in -current but only the first will be true in
the KSE M3 world.

22 years agoComplete '_' hack.
Alexey Zelkin [Thu, 7 Feb 2002 19:32:40 +0000 (19:32 +0000)]
Complete '_' hack.