]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoUserland changes to go with vnode->v_tag change.
Nate Lawson [Sat, 14 Sep 2002 18:51:10 +0000 (18:51 +0000)]
Userland changes to go with vnode->v_tag change.

21 years agoUse the macros from machine/fsr.h; some minor cleanups.
Thomas Moestl [Sat, 14 Sep 2002 18:07:03 +0000 (18:07 +0000)]
Use the macros from machine/fsr.h; some minor cleanups.

21 years agoAdd implementations of fpgetmask(), fpgetround(), fpgetsticky(),
Thomas Moestl [Sat, 14 Sep 2002 18:06:21 +0000 (18:06 +0000)]
Add implementations of fpgetmask(), fpgetround(), fpgetsticky(),
fpsetround(), fpsetsticky(), obtained from NetBSD and tweaked a little
to use definitions from machine/fsr.h instead of magic numbers.

21 years agoUse the definitions in machine/fsr.h instead of duplicating these magic
Thomas Moestl [Sat, 14 Sep 2002 18:00:44 +0000 (18:00 +0000)]
Use the definitions in machine/fsr.h instead of duplicating these magic
numbers here (the values need to correspond to the %fsr ones for some
libc functions to work right).

21 years agoClean up a bit, and add some more macros to access %fsr fields.
Thomas Moestl [Sat, 14 Sep 2002 17:58:54 +0000 (17:58 +0000)]
Clean up a bit, and add some more macros to access %fsr fields.

21 years agoLimit UDP payload size for EDNS0 to 0xffff, not use lower 16bit value.
Hajimu UMEMOTO [Sat, 14 Sep 2002 17:28:40 +0000 (17:28 +0000)]
Limit UDP payload size for EDNS0 to 0xffff, not use lower 16bit value.

Obtained from: KAME

21 years agokthread.h is not needed
Scott Long [Sat, 14 Sep 2002 16:54:46 +0000 (16:54 +0000)]
kthread.h is not needed

21 years agoRemoved a debugging printf() left here in peter's
Maxime Henrion [Sat, 14 Sep 2002 15:14:29 +0000 (15:14 +0000)]
Removed a debugging printf() left here in peter's
last commit.

Spotted by: kris

21 years agoRemove a conditional #include <sys/kernel.h>, it is already
Maxime Henrion [Sat, 14 Sep 2002 14:44:41 +0000 (14:44 +0000)]
Remove a conditional #include <sys/kernel.h>, it is already
included unconditionally before.

Submitted by: Olivier Houchard <cognet@ci0.org>

21 years agoReturn an error if a symbol is not found in reloc_jmpslots() instead of
Thomas Moestl [Sat, 14 Sep 2002 12:14:24 +0000 (12:14 +0000)]
Return an error if a symbol is not found in reloc_jmpslots() instead of
crashing.

21 years agoRemove all use of vnode->v_tag, replacing with appropriate substitutes.
Nate Lawson [Sat, 14 Sep 2002 09:02:28 +0000 (09:02 +0000)]
Remove all use of vnode->v_tag, replacing with appropriate substitutes.
v_tag is now const char * and should only be used for debugging.

Additionally:
1. All users of VT_NTS now check vfsconf->vf_type VFCF_NETWORK
2. The user of VT_PROCFS now checks for the new flag VV_PROCDEP, which
is propagated by pseudofs to all child vnodes if the fs sets PFS_PROCDEP.

Suggested by:   phk
Reviewed by:    bde, rwatson (earlier version)

21 years agoReference the socket we're accepting.
Benno Rice [Sat, 14 Sep 2002 08:56:10 +0000 (08:56 +0000)]
Reference the socket we're accepting.

21 years agoApparently something down in the guts of vm/uvm still needs giant
Julian Elischer [Sat, 14 Sep 2002 06:23:43 +0000 (06:23 +0000)]
Apparently something down in the guts of vm/uvm still needs giant

Obtained from:  mini via P4 KSE tree.

21 years agos/VOP_NULL/VATTR_NULL/
Robert Watson [Sat, 14 Sep 2002 03:32:57 +0000 (03:32 +0000)]
s/VOP_NULL/VATTR_NULL/

21 years agoPut firewire under architectures it actually might conceivably
Matt Jacob [Sat, 14 Sep 2002 02:05:33 +0000 (02:05 +0000)]
Put firewire under architectures it actually might conceivably
build on, let alone actually do some useful work on real hardware.

Namely, put it in i386/pc98. There is no bus_dma.h on the other architecures,
tra la.

21 years agoRelax checking of incoming PPTP GRE packets a bit: ignore a bogus payload
Archie Cobbs [Sat, 14 Sep 2002 00:00:49 +0000 (00:00 +0000)]
Relax checking of incoming PPTP GRE packets a bit: ignore a bogus payload
length field when there's no payload indicated by the header 'S' bit.
This works around semi-brokenness in the Mac OS X PPTP client.

21 years agoWhen using res_send/res_query/res_search, the caller must either
Jacques Vidrine [Fri, 13 Sep 2002 20:31:29 +0000 (20:31 +0000)]
When using res_send/res_query/res_search, the caller must either
insure enough space is available for the response, or be prepared
to resize the buffer and retry as necessary.

Do the conservative thing and make sure enough space is available.

Reviewed by: silence on freebsd-audit

21 years agoThe nfsroot code should be conditional on NFSCLIENT, not NFS (which no
Peter Wemm [Fri, 13 Sep 2002 20:29:42 +0000 (20:29 +0000)]
The nfsroot code should be conditional on NFSCLIENT, not NFS (which no
longer exists as an option)

21 years agoBegin of MFen 1.423
Sebastien Gioria [Fri, 13 Sep 2002 20:23:26 +0000 (20:23 +0000)]
Begin of MFen 1.423

21 years agomlockall() and munlockall() are unimplemented; remove their prototypes.
Mike Barcroft [Fri, 13 Sep 2002 18:29:15 +0000 (18:29 +0000)]
mlockall() and munlockall() are unimplemented; remove their prototypes.

21 years agoPromote gcore from i386-specific to the MI section.
Peter Wemm [Fri, 13 Sep 2002 18:29:08 +0000 (18:29 +0000)]
Promote gcore from i386-specific to the MI section.

Compiled on: alpha, sparc64, ia64

21 years agoAdd #include <sys/queue.h> for sparc64's benefit. The MD includes there
Peter Wemm [Fri, 13 Sep 2002 18:24:59 +0000 (18:24 +0000)]
Add #include <sys/queue.h> for sparc64's benefit.  The MD includes there
have less polution.

21 years agoUse a crowbar to move the a.out code out of gcore.c and into its own
Peter Wemm [Fri, 13 Sep 2002 16:33:35 +0000 (16:33 +0000)]
Use a crowbar to move the a.out code out of gcore.c and into its own
file so that we have a chance of using gcore on non-i386 platforms.  Use
linker sets to reduce the registration glue.  Remove md-sparc.c, we do not
have an a.out sparc32 port.  aoutcore.c was repocopied from gcore.c.

21 years agoFor boot verbose, print resources assigned in a more asthetically pleasing
Warner Losh [Fri, 13 Sep 2002 16:23:41 +0000 (16:23 +0000)]
For boot verbose, print resources assigned in a more asthetically pleasing
way.

PR: 42745 (although done in a different way)

21 years agoAdd an implementation of fabs() (which is quite trivial).
Thomas Moestl [Fri, 13 Sep 2002 16:01:26 +0000 (16:01 +0000)]
Add an implementation of fabs() (which is quite trivial).
When it is called directly, gcc is smart enough to generate inline
code for it, which is why it wasn't noticed before that it was missing.

fabs() would probably better fit into libm, but it has traditionally been
in libc on FreeBSD, so there is probably software around that makes
assumptions about this by now.

21 years agoFix fcntl(..., F_GETOWN, ...) and fcntl(..., F_SETOWN, ...) on sparc64
Thomas Moestl [Fri, 13 Sep 2002 15:15:16 +0000 (15:15 +0000)]
Fix fcntl(..., F_GETOWN, ...) and fcntl(..., F_SETOWN, ...) on sparc64
by not passing a pointer to a register_t or intptr_t when the code in
the lower layers expects one to an int.

21 years agoOverhaul the kernel-simulation bits here to use pthreads instead.
Poul-Henning Kamp [Fri, 13 Sep 2002 13:13:00 +0000 (13:13 +0000)]
Overhaul the kernel-simulation bits here to use pthreads instead.

If somebody wanted to, this could for the beginning of a "libkernel"
which could be used to run kernel code in userland.

Sponsored by: DARPA & NAI Labs.

21 years agoAdd firewire module
Katsushi Kobayashi [Fri, 13 Sep 2002 12:42:57 +0000 (12:42 +0000)]
Add firewire module

21 years agoInitial import for IEEE1394 OHCI chipdet device driver and SBP-2 (Serial
Katsushi Kobayashi [Fri, 13 Sep 2002 12:31:56 +0000 (12:31 +0000)]
Initial import for IEEE1394 OHCI chipdet device driver and SBP-2 (Serial
Bus Protocol 2:SCSI over IEEE1394) support for CAM.

21 years ago"Fix" printf format issues by using %j
Poul-Henning Kamp [Fri, 13 Sep 2002 11:41:25 +0000 (11:41 +0000)]
"Fix" printf format issues by using %j

Sponsored by: DARPA & NAI Labs.

21 years agoUse biowait() rather than DIY.
Poul-Henning Kamp [Fri, 13 Sep 2002 11:39:11 +0000 (11:39 +0000)]
Use biowait() rather than DIY.

Sponsored by: DARPA & NAI Labs

21 years agoFix a bug were a struct bio was freed and certain members subsequently
Poul-Henning Kamp [Fri, 13 Sep 2002 11:37:38 +0000 (11:37 +0000)]
Fix a bug were a struct bio was freed and certain members subsequently
examined.  Use biowait() instead of DIY.

21 years agoOops, broke the build there. Uninline biodone() now that it is non-trivial.
Poul-Henning Kamp [Fri, 13 Sep 2002 11:28:31 +0000 (11:28 +0000)]
Oops, broke the build there.  Uninline biodone() now that it is non-trivial.

Introduce biowait() function.  Currently there is a race condition and the
mitigation is a timeout/retry.  It is not obvious what kind of locking (if any)
is suitable for BIO_DONE, since the majority of users take are of this
themselves, and only a few places actually rely on the wakeup.

Sponsored by: DARPA & NAI Labs.

21 years agoMake biodone() default to wakeup() on the struct bio if no bio_done
Poul-Henning Kamp [Fri, 13 Sep 2002 11:00:48 +0000 (11:00 +0000)]
Make biodone() default to wakeup() on the struct bio if no bio_done
method was specified.

21 years agoAdd a couple more of the big/little-endian conversion routines and make
Poul-Henning Kamp [Fri, 13 Sep 2002 10:33:10 +0000 (10:33 +0000)]
Add a couple more of the big/little-endian conversion routines and make
them visible from userland, if need be.

I wish that the C language contained this as part of struct definintions,
but failing that, I would settle for an agreed upon set of functions for
packing/unpacking integers in various sizes from byte-streams which may
have unfriendly alignment.

This really belongs in <sys/endian.h> I guess.

21 years agoFix another two printf() format errors which weren't warned
Maxime Henrion [Fri, 13 Sep 2002 09:37:44 +0000 (09:37 +0000)]
Fix another two printf() format errors which weren't warned
about because the bio_blknos were bogusly casted to long long.

21 years agoDrop the proc lock while calling fdcheckstd() which may block to allocate
Don Lewis [Fri, 13 Sep 2002 09:31:56 +0000 (09:31 +0000)]
Drop the proc lock while calling fdcheckstd() which may block to allocate
memory.

Reviewed by: jhb

21 years agoFix another printf() format error which wasn't warned about
Maxime Henrion [Fri, 13 Sep 2002 09:19:55 +0000 (09:19 +0000)]
Fix another printf() format error which wasn't warned about
because the bio_blkno was bogusly casted to an int.

21 years agoFix a printf() format error on 64 bits architectures.
Maxime Henrion [Fri, 13 Sep 2002 09:16:06 +0000 (09:16 +0000)]
Fix a printf() format error on 64 bits architectures.
Also fix some style bugs on the same line.

21 years agoDo not risk using the kernel pgtok() which assumes the page size is
Peter Wemm [Fri, 13 Sep 2002 07:13:33 +0000 (07:13 +0000)]
Do not risk using the kernel pgtok() which assumes the page size is
constant.

21 years agoRestore original behaviour of recursion preventer.
Maxim Sobolev [Fri, 13 Sep 2002 06:24:27 +0000 (06:24 +0000)]
Restore original behaviour of recursion preventer.

Submitted by:   sumikawa

21 years agoFix an i830/i845 test that shouldn't get hit in normal use and remove a printf.
Eric Anholt [Fri, 13 Sep 2002 04:17:28 +0000 (04:17 +0000)]
Fix an i830/i845 test that shouldn't get hit in normal use and remove a printf.

Submitted by: David Dawes <dawes@XFree86.Org>

21 years agoNew release notes: sendfile(2) bugfix + MFC.
Bruce A. Mah [Fri, 13 Sep 2002 03:55:31 +0000 (03:55 +0000)]
New release notes:  sendfile(2) bugfix + MFC.

Modified release notes:  newsyslog(8) now uses bzip2(1) by default.

Move two TCP-related release notes from the general kernel section to
the networking protocols section.  Minor wording adjustment in one
item, note MFC of the other.

21 years agoStyle: tab between #define and macro name.
Tim J. Robbins [Thu, 12 Sep 2002 23:33:17 +0000 (23:33 +0000)]
Style: tab between #define and macro name.

21 years agoAdd a blank line to make this more consistent with the rest of the file.
John Baldwin [Thu, 12 Sep 2002 21:07:06 +0000 (21:07 +0000)]
Add a blank line to make this more consistent with the rest of the file.

21 years agoPass the physical instead of the logical block number in
Joerg Wunsch [Thu, 12 Sep 2002 20:53:50 +0000 (20:53 +0000)]
Pass the physical instead of the logical block number in
bp->bio_pblkno.  Otherwise, all formats not using 512 byte
physical sectors got screwed.

21 years agoFix sendmail to use checkyesno since rc.conf is no longer re'sourced from the
Gordon Tetlow [Thu, 12 Sep 2002 17:31:09 +0000 (17:31 +0000)]
Fix sendmail to use checkyesno since rc.conf is no longer re'sourced from the
debug statement.

21 years agoUse bzip2 instead of gzip for those logs we compress.
David E. O'Brien [Thu, 12 Sep 2002 17:28:07 +0000 (17:28 +0000)]
Use bzip2 instead of gzip for those logs we compress.

Basic idea agreed to by: rwatson

21 years agoRemove the sourceing of rc.conf from the debug subroutine. This caused
Gordon Tetlow [Thu, 12 Sep 2002 17:27:36 +0000 (17:27 +0000)]
Remove the sourceing of rc.conf from the debug subroutine. This caused
problems when you try to override variables locally.

21 years agoEmacs decides between v8 and v9 based on "__arch64__". So define it.
David E. O'Brien [Thu, 12 Sep 2002 17:19:09 +0000 (17:19 +0000)]
Emacs decides between v8 and v9 based on "__arch64__".  So define it.

Desired by: tmm

21 years agoRemove some debugging cruft I accidently committed with rev 1.4.
David E. O'Brien [Thu, 12 Sep 2002 16:10:21 +0000 (16:10 +0000)]
Remove some debugging cruft I accidently committed with rev 1.4.

21 years agoTry to detect support for the `long long' type so that ANSI-C[89] clean
David E. O'Brien [Thu, 12 Sep 2002 16:05:56 +0000 (16:05 +0000)]
Try to detect support for the `long long' type so that ANSI-C[89] clean
code will know not to try to use `long long'.
Unfortunately the GCC spec parser will not allow us to properly detect the
"iso9899:1990" and "iso9899:199409" forms of the acceptable -std= arguments,
because of the ':' in the -std argument.  :-(  I have left them in the spec
as a place holder in hopes someone knows a way to make the detection of
them work.

Desired by: wollman

21 years agoWe're going to have to use a TARGET_ARCH of "x86_64" rather than "x86-64"
David E. O'Brien [Thu, 12 Sep 2002 15:56:21 +0000 (15:56 +0000)]
We're going to have to use a TARGET_ARCH of "x86_64" rather than "x86-64"
to make building GCC happy.

21 years agoFix the GENERIC build. Don't refer to the non-existant fw_one_pass.
David E. O'Brien [Thu, 12 Sep 2002 15:36:13 +0000 (15:36 +0000)]
Fix the GENERIC build.  Don't refer to the non-existant fw_one_pass.

21 years agoAdd preliminary mostly untested support for the Silicon Image Sil680 chip.
Søren Schmidt [Thu, 12 Sep 2002 15:25:59 +0000 (15:25 +0000)]
Add preliminary mostly untested support for the Silicon Image Sil680 chip.

21 years agoEnd of initial effort of the translation of the CURRENT release notes. Now
Sebastien Gioria [Thu, 12 Sep 2002 15:21:02 +0000 (15:21 +0000)]
End of initial effort of the translation of the CURRENT release notes. Now
need to update :-(((

21 years agoChange a couple of splcam() calls that do not make sense in atapi-cam
Thomas Quinot [Thu, 12 Sep 2002 15:15:33 +0000 (15:15 +0000)]
Change a couple of splcam() calls that do not make sense in atapi-cam
(what we want to mask is the ATA interrupts, which run at level splbio),
and add interrupt masking around the statement that modifies the
request queue for a SIM.

This resolves problems reported by users under heavy ATAPI load.

Reviewed by: roberto

21 years agoEnd of translation of section "Contributed Software"
Sebastien Gioria [Thu, 12 Sep 2002 14:45:50 +0000 (14:45 +0000)]
End of translation of section "Contributed Software"

21 years agoRearrange where to get the ATAPI magic sequence.
Søren Schmidt [Thu, 12 Sep 2002 14:32:33 +0000 (14:32 +0000)]
Rearrange where to get the ATAPI magic sequence.

Submitted by: Benjamin Close <cisbjc@cs.unisa.edu.au>

21 years agoUse the proper fields for security.
Søren Schmidt [Thu, 12 Sep 2002 14:08:28 +0000 (14:08 +0000)]
Use the proper fields for security.

PR: 41870

21 years agoWait for drive to become ready if its in progress of loading..
Søren Schmidt [Thu, 12 Sep 2002 13:55:24 +0000 (13:55 +0000)]
Wait for drive to become ready if its in progress of loading..
Based on patch in PR-42659, but done somewhat differently.

PR: 42659

21 years agoDont hang in atprq on poll_dsc command.
Søren Schmidt [Thu, 12 Sep 2002 13:53:33 +0000 (13:53 +0000)]
Dont hang in atprq on poll_dsc command.
This is a temporary fix until I get proper locking done.

Submitted by: iedowse@maths.tcd.ie

21 years agoFixed style bugs in resource_list_add_next().
Bruce Evans [Thu, 12 Sep 2002 13:45:38 +0000 (13:45 +0000)]
Fixed style bugs in resource_list_add_next().

21 years agoProtect arguments to the putwc and putwchar macros with parens.
Tim J. Robbins [Thu, 12 Sep 2002 10:27:48 +0000 (10:27 +0000)]
Protect arguments to the putwc and putwchar macros with parens.

21 years agoCorrect type of second argument: it is wchar_t ** restrict,
Tim J. Robbins [Thu, 12 Sep 2002 09:25:27 +0000 (09:25 +0000)]
Correct type of second argument: it is wchar_t ** restrict,
not wchar_t * restrict.

21 years ago- Fix two obvious locking bugs; 1) returning with lock held when it needed
Andrew R. Reiter [Thu, 12 Sep 2002 05:00:32 +0000 (05:00 +0000)]
- Fix two obvious locking bugs; 1) returning with lock held when it needed
  to be dropped, 2) attempting to lock acct_mtx while already holding it.
  Sorry to those who experienced pain.
- Added two comments referring to two areas in which acct_mtx is held over
  vnode operations that might sleep.  Patch in the works for this.

21 years agoAdd AGP support for Intel i830M and i845 thanks to patches from moto kawasaki
Eric Anholt [Thu, 12 Sep 2002 04:23:18 +0000 (04:23 +0000)]
Add AGP support for Intel i830M and i845 thanks to patches from moto kawasaki
<kawasaki@mbg.sphere.ne.jp> and David Dawes <dawes@XFree86.org>.

21 years agoMake bridging and layer2-ipfw obey net.inet.ip.fw.one_pass.
Luigi Rizzo [Thu, 12 Sep 2002 01:05:46 +0000 (01:05 +0000)]
Make bridging and layer2-ipfw obey net.inet.ip.fw.one_pass.
I should have committed this ages ago.

The MFC for if_ethersubr.c could be done in the usual few days (only
ipfw2 uses it), the one for bridge.c should probably wait until
after 4.7 because it changes an existing though mostly undocumented
behaviour (on which i hope nobody relies). All in all, i'll wait for
both things unless there is demand.

MFC after: 35 days

21 years agoStore the port number in "fwd" rules in host format, same as ipfw1
Luigi Rizzo [Thu, 12 Sep 2002 00:45:32 +0000 (00:45 +0000)]
Store the port number in "fwd" rules in host format, same as ipfw1
has always done.

Technically, this is the wrong format, but it reduces the diffs in
-stable. Someday, when we get rid of ipfw1, I will put the port number
in the proper format both in kernel and userland.

MFC after: 3 days
(with re@ permission)

21 years agoUpdate to reflect reality.
Archie Cobbs [Wed, 11 Sep 2002 21:40:02 +0000 (21:40 +0000)]
Update to reflect reality.

Reviewed by: mini
MFC after: 3 days

21 years agoAdd man pages for getcontext()/setcontext(), makecontext()/swapcontext(),
Archie Cobbs [Wed, 11 Sep 2002 21:39:21 +0000 (21:39 +0000)]
Add man pages for getcontext()/setcontext(), makecontext()/swapcontext(),
and ucontext_t.

Reviewed by: mini
MFC after: 3 days

21 years agoOnly probe one PHY on the D-Link 580 version of the card (ie rev 0x12).
Doug Ambrisko [Wed, 11 Sep 2002 21:26:22 +0000 (21:26 +0000)]
Only probe one PHY on the D-Link 580 version of the card (ie rev 0x12).
The 550 version is location at address 1 but since it works right we
let the code find whatever PHY it can.

Fix a fragment issue on TX.  If the number of frags are more then the
driver has allocated then bring all the frags together into one packet
and send it out.  Code derived from the fxp driver.

Tested and found by: Francois Tigeot <francois.tigeot@nic.fr>
Hellmuth Michaelis <hm@kts.org>

MFC after: 1 week

21 years ago- Change utrace ktrace events to malloc the work buffer before getting a
John Baldwin [Wed, 11 Sep 2002 21:00:56 +0000 (21:00 +0000)]
- Change utrace ktrace events to malloc the work buffer before getting a
  request structure.
- Re-optimize the case of utrace being disabled by doing an explicit
  KTRPOINT check instead of relying on the one in ktr_getrequest() so that
  we don't waste time on a malloc in the non-tracing case.
- Change utrace() to return an error if the copyin() fails.  Before it
  would just ignore the request but still return success.  This last is
  a change in behavior and can be backed out if necessary.

21 years agoRemove support for synchronous ktrace requests now that none exist anymore.
John Baldwin [Wed, 11 Sep 2002 20:58:10 +0000 (20:58 +0000)]
Remove support for synchronous ktrace requests now that none exist anymore.
They were an ugly, gross hack.

21 years ago- Change ktrace genio events to only copy up to ktr_geniosize bytes of a
John Baldwin [Wed, 11 Sep 2002 20:56:05 +0000 (20:56 +0000)]
- Change ktrace genio events to only copy up to ktr_geniosize bytes of a
  transfer to a malloc'd buffer and use that bufer for the ktrace event.
  This means that genio ktrace events no longer need to be synchronous.
- Now that ktr_buffer isn't overloaded to sometimes point to a cached uio
  pointer for genio requests and always points to a malloc'd buffer if not
  NULL, free the buffer in ktr_freerequest() instead of in
  ktr_writerequest().  This closes a memory leak for ktrace events that
  used a malloc'd buffer that had their vnode ripped out from under them
  while they were on the todo list.

Suggested by: bde (1, in principle)

21 years ago- Add a kern.ktrace sysctl node.
John Baldwin [Wed, 11 Sep 2002 20:49:55 +0000 (20:49 +0000)]
- Add a kern.ktrace sysctl node.
- Rename kern.ktrace_request_pool tunable/sysctl to
  kern.ktrace.request_pool.
- Add a variable to control the max amount of data to log for genio events.
  This variable is tunable via the tunable/sysctl kern.ktrace.genio_size
  and defaults to one page.

21 years agoChange namei and syscall ktrace events to malloc work buffers before
John Baldwin [Wed, 11 Sep 2002 20:46:50 +0000 (20:46 +0000)]
Change namei and syscall ktrace events to malloc work buffers before
obtaining a ktr_request structure from the free pool so we can avoid
starving other threads of ktr_request structures.

21 years agoSomebody forgot to turn out the lights.
Peter Wemm [Wed, 11 Sep 2002 20:20:38 +0000 (20:20 +0000)]
Somebody forgot to turn out the lights.
(ie: These files are long since obsolete)

21 years agoCosmetic cleanup to a message.
John Baldwin [Wed, 11 Sep 2002 19:49:43 +0000 (19:49 +0000)]
Cosmetic cleanup to a message.

21 years agoSometime after binutils-2.9 (ie: ages ago), the handling of __start_set_*
Peter Wemm [Wed, 11 Sep 2002 19:08:09 +0000 (19:08 +0000)]
Sometime after binutils-2.9 (ie: ages ago), the handling of __start_set_*
and __stop_set_* became friendlier.  This trick to force the export
of the section delimiter symbols is no longer necessary.  We needed to
force this so that we could look up the symbols via a symbol table search
when the module was loaded at run time.  ld used to eliminate them if
they were not referenced at compile/link time.

21 years agoCommit "unrelated style fixes" part of Bruce's patch (regarding
Peter Wemm [Wed, 11 Sep 2002 18:16:40 +0000 (18:16 +0000)]
Commit "unrelated style fixes" part of Bruce's patch (regarding
bcopy/memcpy) seperately.

Submitted by:  bde

21 years agoIndentation does not make a block.. need curly braces too.
Julian Elischer [Wed, 11 Sep 2002 18:15:26 +0000 (18:15 +0000)]
Indentation does not make a block.. need curly braces too.
Submitted by: Eagle-eyes evans <bde@freebsd.org>

21 years agoModify previous commit to solve the real problem that made gcc think
Peter Wemm [Wed, 11 Sep 2002 18:12:29 +0000 (18:12 +0000)]
Modify previous commit to solve the real problem that made gcc think
the timestamp was aligned.  ie: Use a void * instead of struct timeval *
which gcc assumes will be aligned.  Go back to memcpy().

Submitted by: bde

21 years agoI completely fubared this. An empty EXPORT_SYMS= is not valid. I know I
Peter Wemm [Wed, 11 Sep 2002 18:03:03 +0000 (18:03 +0000)]
I completely fubared this.  An empty EXPORT_SYMS= is not valid.  I know I
build this several times, and even tested loaded the module.  I must have
changed something right before commit.

Pointed out by: bde

21 years agoWhitespace consistency fix from addition of IAHD_REG_PRETTY_PRINT: use
Robert Watson [Wed, 11 Sep 2002 16:51:48 +0000 (16:51 +0000)]
Whitespace consistency fix from addition of IAHD_REG_PRETTY_PRINT: use
tabs not spaces.

21 years agoIn kvm_openfiles/kvm_open, mark the file descriptors as close-on-exec.
Jacques Vidrine [Wed, 11 Sep 2002 16:41:39 +0000 (16:41 +0000)]
In kvm_openfiles/kvm_open, mark the file descriptors as close-on-exec.
Applications can not do this themselves, as the descriptors are hidden
behind the opaque `kvm_t' type.

21 years agoCorrect a usage of fnctl that could not be right and results in a
Jacques Vidrine [Wed, 11 Sep 2002 16:38:33 +0000 (16:38 +0000)]
Correct a usage of fnctl that could not be right and results in a
no-op.  I assume it was meant that the close-on-exec flag be set here.

21 years agoIt's bad enough people can't figure out to use the same code, or in
Poul-Henning Kamp [Wed, 11 Sep 2002 12:44:58 +0000 (12:44 +0000)]
It's bad enough people can't figure out to use the same code, or in
this case, ugly macros, but the data tables can be reused:

Put one copy of the software HDLC tables in its own file.

21 years agoFix spelling of Bismarck's name and remove duplicate
Thomas Gellekum [Wed, 11 Sep 2002 11:21:36 +0000 (11:21 +0000)]
Fix spelling of Bismarck's name and remove duplicate
Bertrand Russell quotations.

21 years agocorrect another spammage.. sorry bruce.. not exactly sure how my patch
Julian Elischer [Wed, 11 Sep 2002 09:08:53 +0000 (09:08 +0000)]
correct another spammage.. sorry bruce.. not exactly sure how my patch
reverted out your change but hopefully that's it..

21 years agorevert a line that was not part of my change..
Julian Elischer [Wed, 11 Sep 2002 09:04:02 +0000 (09:04 +0000)]
revert a line that was not part of my change..
I think it was a part of someone else's commit that
somehow got reverted by my patch.

21 years agoAdd an implementation of wcsftime() (wide character version of strftime()).
Tim J. Robbins [Wed, 11 Sep 2002 08:57:11 +0000 (08:57 +0000)]
Add an implementation of wcsftime() (wide character version of strftime()).

21 years agoComment and whitespace changes.
Julian Elischer [Wed, 11 Sep 2002 08:25:35 +0000 (08:25 +0000)]
Comment and whitespace changes.

21 years agoCompletely redo thread states.
Julian Elischer [Wed, 11 Sep 2002 08:13:56 +0000 (08:13 +0000)]
Completely redo thread states.

Reviewed by: davidxu@freebsd.org

21 years agoInclude <vm/uma.h> instead of depending on namespace pollution in
Bruce Evans [Wed, 11 Sep 2002 07:13:28 +0000 (07:13 +0000)]
Include <vm/uma.h> instead of depending on namespace pollution in
<sys/malloc.h>.

Sorted includes as much as possible.  Removed banal comment(s) attached to
includes.

21 years agoFix LINT build on alpha by completing move of cy and apm_saver to
Nate Lawson [Wed, 11 Sep 2002 05:33:15 +0000 (05:33 +0000)]
Fix LINT build on alpha by completing move of cy and apm_saver to
i386/conf/NOTES rather than the global conf/NOTES.

Suggested by: bde

21 years agoAdd `restrict' type-qualifier.
Mike Barcroft [Wed, 11 Sep 2002 05:05:48 +0000 (05:05 +0000)]
Add `restrict' type-qualifier.

21 years agoConditionalize some objects to match the functions that they are used
Mike Barcroft [Wed, 11 Sep 2002 05:04:06 +0000 (05:04 +0000)]
Conditionalize some objects to match the functions that they are used
with.  Enable `restrict' type-qualifier.

21 years agoAdd support for the AMD x86-64 Hammer platform.
David E. O'Brien [Wed, 11 Sep 2002 04:57:09 +0000 (04:57 +0000)]
Add support for the AMD x86-64 Hammer platform.

21 years agoSome uses of the variable n needed to be int, others needed to be
Mike Barcroft [Wed, 11 Sep 2002 04:26:09 +0000 (04:26 +0000)]
Some uses of the variable n needed to be int, others needed to be
size_t.  Add a new variable to cope.