]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoFix an external declaration that was causing telnetd to core dump.
Mark Murray [Mon, 6 May 2002 09:46:29 +0000 (09:46 +0000)]
Fix an external declaration that was causing telnetd to core dump.

MFC after: 1 week
PR: 37766

22 years agoUse hw.ata.ata_dma instead of hw.atamodes, which doesn't exist in anymore.
Dima Dorfman [Mon, 6 May 2002 09:40:52 +0000 (09:40 +0000)]
Use hw.ata.ata_dma instead of hw.atamodes, which doesn't exist in anymore.

Reviewed by: maxim

22 years agoAdd another Askey ISA modem ID.
David Malone [Mon, 6 May 2002 09:02:55 +0000 (09:02 +0000)]
Add another Askey ISA modem ID.

PR: 35813
Submitted by: Chris Knight <chris@aims.com.au>

22 years agoinet_aton returns 0 on failure, not -1.
David Malone [Mon, 6 May 2002 08:39:43 +0000 (08:39 +0000)]
inet_aton returns 0 on failure, not -1.

PR: 37680
Submitted by: Scott Bertilson <scott@nts.umn.edu>
MFC after: 2 weeks

22 years agoMake argument to printf match format.
Marcel Moolenaar [Mon, 6 May 2002 08:27:21 +0000 (08:27 +0000)]
Make argument to printf match format.

22 years agoPC98 still uses wd#, so resurrect the wd# entry and mark is "pc98 only".
Dima Dorfman [Mon, 6 May 2002 07:48:36 +0000 (07:48 +0000)]
PC98 still uses wd#, so resurrect the wd# entry and mark is "pc98 only".

Noticed by: bde (four months ago)

22 years ago o Header files shouldn't depend on options: Provide prototypes
Alan Cox [Mon, 6 May 2002 06:20:04 +0000 (06:20 +0000)]
 o Header files shouldn't depend on options: Provide prototypes
   for uiomoveco(), uioread(), and vm_uiomove() regardless
   of whether ENABLE_VFS_IOOPT is defined or not.

Submitted by: bde

22 years ago o Condition the compilation and use of vm_freeze_copyopts()
Alan Cox [Mon, 6 May 2002 05:45:57 +0000 (05:45 +0000)]
 o Condition the compilation and use of vm_freeze_copyopts()
   on ENABLE_VFS_IOOPT.

22 years agoHold the currently selected vnode's lock across the call to VOP_GETVOBJECT.
Jeff Roberson [Mon, 6 May 2002 04:47:43 +0000 (04:47 +0000)]
Hold the currently selected vnode's lock across the call to VOP_GETVOBJECT.

Don't try to create a vm object before the file system has a chance to finish
initializing it.  This is incorrect for a number of reasons.  Firstly, that
VOP requires a lock which the file system may not have initialized yet. Also,
open and others will create a vm object if it is necessary later.

22 years agoAffect the first of stdin, stdout and stderr that is attached to a terminal.
Tim J. Robbins [Mon, 6 May 2002 04:33:04 +0000 (04:33 +0000)]
Affect the first of stdin, stdout and stderr that is attached to a terminal.
Write status information to stdout instead of stderr.
Exit status when an error occurs musn't be 1, that is reserved for
indicating that messages are disabled.

These changes bring mesg(1) up to SUSv3 conformance.

Reviewed by: mike

22 years ago-I and -J both set the same variable, replstr, to the string to replace with
Juli Mallett [Mon, 6 May 2002 03:38:28 +0000 (03:38 +0000)]
-I and -J both set the same variable, replstr, to the string to replace with
a line of input, and both work differently, so prevent them from both being
passed to xargs(1).

22 years ago o Some improvements to the page coloring of vm objects, particularly,
Alan Cox [Mon, 6 May 2002 03:34:17 +0000 (03:34 +0000)]
 o Some improvements to the page coloring of vm objects, particularly,
   for shadow objects.

Submitted by: bde

22 years agoInclude <sys/queue.h> so that this file provides its own namespace
Bruce Evans [Mon, 6 May 2002 03:13:08 +0000 (03:13 +0000)]
Include <sys/queue.h> so that this file provides its own namespace
pollution which is required for its includes of <sys/_lock.h> and
<sys/_mutex.h> to work.

22 years agos/sysctl -w/sysctl/
Chris D. Faulhaber [Mon, 6 May 2002 00:45:27 +0000 (00:45 +0000)]
s/sysctl -w/sysctl/

22 years ago- Properly initialize the count_win when the user turns the info window
Mark Peek [Mon, 6 May 2002 00:25:36 +0000 (00:25 +0000)]
- Properly initialize the count_win when the user turns the info window
  back on (and count_win is recreated).
- Create info_win as the same size for all invocations.

PR:             37552
Submitted by:   Mark Valentine <mark@thuvia.demon.co.uk>
MFC after:      1 week

22 years ago o Move vm_freeze_copyopts() from vm_map.{c.h} to vm_object.{c,h}. It's plainly
Alan Cox [Mon, 6 May 2002 00:12:47 +0000 (00:12 +0000)]
 o Move vm_freeze_copyopts() from vm_map.{c.h} to vm_object.{c,h}.  It's plainly
   an operation on a vm_object and belongs in the latter place.

22 years agoRestored db_elf.c from the Attic. This will be used for a quick fix for
Bruce Evans [Mon, 6 May 2002 00:05:44 +0000 (00:05 +0000)]
Restored db_elf.c from the Attic.  This will be used for a quick fix for
the longstanding brokenness of symbols in ddb at boot time.  It doesn't
compile and is not attached to the build yet.

22 years agoDocument the lchflags(2) syscall.
Maxime Henrion [Sun, 5 May 2002 23:51:32 +0000 (23:51 +0000)]
Document the lchflags(2) syscall.

22 years agoAdd the lchflags(2) syscall.
Maxime Henrion [Sun, 5 May 2002 23:47:41 +0000 (23:47 +0000)]
Add the lchflags(2) syscall.

Reviewed by: rwatson

22 years agoAdd an entry for the lchflags(2) syscall. It's useful to prevent
Maxime Henrion [Sun, 5 May 2002 23:37:44 +0000 (23:37 +0000)]
Add an entry for the lchflags(2) syscall.  It's useful to prevent
a symlink deletion.

Reviewed by: rwatson

22 years agoMove a KASSERT() in open() prior to unlocking the vnode. It's not safe to
Jeff Roberson [Sun, 5 May 2002 23:17:13 +0000 (23:17 +0000)]
Move a KASSERT() in open() prior to unlocking the vnode.  It's not safe to
call VOP_GETVOBJECT without a lock.

22 years ago o Condition the compilation of uiomoveco() and vm_uiomove()
Alan Cox [Sun, 5 May 2002 22:42:40 +0000 (22:42 +0000)]
 o Condition the compilation of uiomoveco() and vm_uiomove()
   on ENABLE_VFS_IOOPT.
 o Add a comment to the effect that this code is experimental
   support for zero-copy I/O.

22 years agoBegin to turn some of the code in this man page into documentation.
Chris Costello [Sun, 5 May 2002 22:09:12 +0000 (22:09 +0000)]
Begin to turn some of the code in this man page into documentation.
Move the code that I have not yet finished documenting into the
`IMPLEMENTATION NOTES' section.

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

22 years agoFix a couple of problems which could cause panics at runtime:
Luigi Rizzo [Sun, 5 May 2002 21:34:10 +0000 (21:34 +0000)]
Fix a couple of problems which could cause panics at runtime:

 + setting a bandwidth too large for a pipe (above 2Gbit/s) could
   cause the internal representation (which is int) to wrap to a
   negative number, causing an infinite loop in the kernel;

 + (see PR bin/35628): when configuring RED parameters for a queue,
   the values are not passed to the kernel resulting in panics at
   runtime (part of the problem here is also that the kernel does
   not check for valid parameters being passed, but this will be
   fixed in a separate commit).

These are both critical fixes which need to be merged into 4.6-RELEASE.

MFC after: 1 day

22 years agoIndentation and comments cleanup, no functional change.
Luigi Rizzo [Sun, 5 May 2002 21:27:47 +0000 (21:27 +0000)]
Indentation and comments cleanup, no functional change.

MFC after: 3 days

22 years agoNew feature: allow origins of all dependencies be recorded into package list
Maxim Sobolev [Sun, 5 May 2002 21:03:25 +0000 (21:03 +0000)]
New feature: allow origins of all dependencies be recorded into package list
using new `@comment DEPORIGIN:...' directive. This would allow us to make
many neat things including:

- easier binary upgrades;
- source upgrades without using external tools by simply extending
  bsd.port.mk and pkg_install tools;
- mixed-mode upgrades (source + binary);
- depreciate and deorbit silly +REQUIRED_BY files in the near future.

This feature is no-op until appropriate bsd.port.mk patch is committed, and
even when it is already committed packages generated will remain 100%
compatible with old set of pkg_install tools (module all those neat
features, of course).

MFC after: 6 days

22 years agoIn a origin lookup mode (-O) don't bail out when encountering empty package
Maxim Sobolev [Sun, 5 May 2002 20:50:27 +0000 (20:50 +0000)]
In a origin lookup mode (-O) don't bail out when encountering empty package
directory, because this prevent this option from being used from the
package-depends target of bsd.port.mk since it creates such empty dir
during its normal operation.

MFC after: 6 days

22 years agoBring in a slightly older version of usr/bin/passwd which does not
Luigi Rizzo [Sun, 5 May 2002 20:43:27 +0000 (20:43 +0000)]
Bring in a slightly older version of usr/bin/passwd which does not
require libpam and all the libraries that libpam brings in.
We have no room for those on picobsd images.

22 years agoExpand the one-line function pbreassignbuf() the only place it is or could
Poul-Henning Kamp [Sun, 5 May 2002 20:37:08 +0000 (20:37 +0000)]
Expand the one-line function pbreassignbuf() the only place it is or could
be used.

22 years agoMove some UFS related stuff home where it belongs.
Poul-Henning Kamp [Sun, 5 May 2002 20:04:33 +0000 (20:04 +0000)]
Move some UFS related stuff home where it belongs.

22 years agoFix make(1) behaviour regarding SysV-style substitution when given a nil
Juli Mallett [Sun, 5 May 2002 19:04:58 +0000 (19:04 +0000)]
Fix make(1) behaviour regarding SysV-style substitution when given a nil
left-hand-side.

PR: bin/5297
Submitted by: "Matthew Emmerton" <matt@gsicomp.on.ca>
MFC after: 1 week

22 years agoAdd a regression test for bin/5297, regarding sysv substitution with a nil
Juli Mallett [Sun, 5 May 2002 19:02:30 +0000 (19:02 +0000)]
Add a regression test for bin/5297, regarding sysv substitution with a nil
left-hand-side.

22 years agoNo longer attempt to power off the ACLINK during suspend - luigi
Orion Hodson [Sun, 5 May 2002 15:37:09 +0000 (15:37 +0000)]
No longer attempt to power off the ACLINK during suspend - luigi
reports this causes his ich machine to hang.

22 years agoAdd a KERNELDUMPMAGIC_CLEARED macro to unbreak savecore. Since
Maxime Henrion [Sun, 5 May 2002 13:47:21 +0000 (13:47 +0000)]
Add a KERNELDUMPMAGIC_CLEARED macro to unbreak savecore.  Since
it is a "magic" value, what it expands to is not really important.
I set it to "Cleared Kernel Dump", but that can be changed later
if someone thinks it's not good enough.

Pointy hat to: fenner

22 years agoUpdate usage() and manpage to match reality (add -O into usage and adjust
Maxim Sobolev [Sun, 5 May 2002 13:44:32 +0000 (13:44 +0000)]
Update usage() and manpage to match reality (add -O into usage and adjust
both to make clear that -W and -O are stand-alone options).

MFC after: 6 days

22 years agoUse PLIST_ORIGIN instead of rolling own version using PLIST_COMMENT.
Maxim Sobolev [Sun, 5 May 2002 13:09:04 +0000 (13:09 +0000)]
Use PLIST_ORIGIN instead of rolling own version using PLIST_COMMENT.

MFC after: 6 days

22 years agoHandle PLIST_ORIGIN properly.
Maxim Sobolev [Sun, 5 May 2002 12:54:11 +0000 (12:54 +0000)]
Handle PLIST_ORIGIN properly.

Poked by: kris
Pointy hat grabbed by: sobomax
MFC after: 6 days

22 years agoFix broken entity
Udo Erdelhoff [Sun, 5 May 2002 12:05:08 +0000 (12:05 +0000)]
Fix broken entity

Approved by: re@

22 years agoFix the fallback to mount(2) with RELENG_4 kernels and -CURRENT
Maxime Henrion [Sun, 5 May 2002 10:08:04 +0000 (10:08 +0000)]
Fix the fallback to mount(2) with RELENG_4 kernels and -CURRENT
utilities by catching SIGSYS.

Submitted by: bde

22 years agoo Add struct mca_guid
Marcel Moolenaar [Sun, 5 May 2002 08:26:25 +0000 (08:26 +0000)]
o  Add struct mca_guid
o  Add currently known GUIDs
o  Slight restyling

22 years agoo Include md_var.h
Marcel Moolenaar [Sun, 5 May 2002 08:18:42 +0000 (08:18 +0000)]
o  Include md_var.h
o  Remove definition of struct ia64_fdesc
o  Remove prototype of os_boot_rendez
o  Use the FDESC_FUNC and FDESC_GP abstractions

22 years agoRemove definition of struct ia64_fdesc. It's been moved to md_var.h
Marcel Moolenaar [Sun, 5 May 2002 08:16:10 +0000 (08:16 +0000)]
Remove definition of struct ia64_fdesc. It's been moved to md_var.h

22 years agoo Move definition of struct ia64_fdesc here to remove duplication.
Marcel Moolenaar [Sun, 5 May 2002 08:14:32 +0000 (08:14 +0000)]
o  Move definition of struct ia64_fdesc here to remove duplication.
o  Add prototype of os_boot_rendez.

22 years agoModify the -p implementation to use a user's locale, so they can respond to
Juli Mallett [Sun, 5 May 2002 06:42:44 +0000 (06:42 +0000)]
Modify the -p implementation to use a user's locale, so they can respond to
the prompt in their native language.

Also make the prompt fit what POSIX asks for (?...).

This should not affect use of -p with yes(1) [as every locale I know of matches
'y' as YESEXPR as well], but that's what -t is for anyway.  -p is meant to be
really used interactively.

Submitted by: tjr, jmallett

22 years ago o Remove GIANT_REQUIRED from vm_map_lookup() and vm_map_lookup_done().
Alan Cox [Sun, 5 May 2002 05:36:28 +0000 (05:36 +0000)]
 o Remove GIANT_REQUIRED from vm_map_lookup() and vm_map_lookup_done().
 o Acquire and release Giant around vm_map_lookup()'s call
   to vm_object_shadow().

22 years agoFix a typo.
Juli Mallett [Sun, 5 May 2002 05:15:16 +0000 (05:15 +0000)]
Fix a typo.

Submitted by: Carl Schmidt <cschmidt@slackerbsd.org>

Wrap an obscenely long line while I'm here.

22 years agoReturn the correct error code (ENOSYS, not EINVAL) from nosys(). Getting
Bruce Evans [Sun, 5 May 2002 04:50:47 +0000 (04:50 +0000)]
Return the correct error code (ENOSYS, not EINVAL) from nosys().  Getting
killed by SIGSYS for unimlemented syscalls is bad enough.

Obtained from: Lite2 branch

The Lite2 branch has some other interesting unmerged (?) bits in this
file.  They are well hidden among cosmetic regressions.

22 years agoCheck for possible overflow from sysconf _SC_ARG_MAX and error out in a
Juli Mallett [Sun, 5 May 2002 04:42:50 +0000 (04:42 +0000)]
Check for possible overflow from sysconf _SC_ARG_MAX and error out in a
correct manner.  Revert my incorrect change to use err(3) for malloc(3)
failing.  Use a size_t variable to store the size of the argument buffer
we allocate, and remove silly casts as the result of having this around.
Modify the math in some of the paranoid checks for buffer overflow to
account for the fact we now are dealing with the actual size of the
buffer.  Remove the static qualifier for arg_max, and the bogus setting
of it to -1.

Include <limits.h> for the definitions we use to check for possible
overflows.

Submitted by: bde

22 years agoFixed breakage of binary compatibility of the kern.clockrate sysctl in
Bruce Evans [Sun, 5 May 2002 04:33:09 +0000 (04:33 +0000)]
Fixed breakage of binary compatibility of the kern.clockrate sysctl in
sys/time.h rev.1.53, etc.  Zero out the entire struct clkinfo and not
just the new spare part of it so that there is no possibility of leaking
kernel stack context to userland.

22 years agoFixed checking for VM86 mode in doreti which I broke in rev.1.30. Only
Bruce Evans [Sun, 5 May 2002 03:19:48 +0000 (03:19 +0000)]
Fixed checking for VM86 mode in doreti which I broke in rev.1.30.  Only
the case of VM86 calls from the kernel was broken, so this bug was not
a security hole.

PR: 36710
Submitted by: David Xu <davidx@viasoft.com.cn> (version for RELENG_4)
MFC after: 3 days

22 years agoIn an effort to make this utility easier to work with at a source level, move
Juli Mallett [Sun, 5 May 2002 02:46:53 +0000 (02:46 +0000)]
In an effort to make this utility easier to work with at a source level, move
out the parse loop to a seperate function, and move local variables around as
is needed.  To keep the state of some variables and keep from having huge arg
lists to functions, make a bunch global.

Change use of err(3) in situations where malloc(3) will be setting errno to
errx(3) since assuming malloc(3) sets errno is non-portable, and provides no
more useful information in the context of FreeBSD.

Submitted by: bde (err/errx), Carl Schmidt <cschmidt@slackerbsd.org> (some
of the movement of the input loop to a function)
Approved by: src/tools/regression/usr.bin/xargs

22 years ago- revert back to vmcore.#
Bill Fenner [Sun, 5 May 2002 01:04:00 +0000 (01:04 +0000)]
- revert back to vmcore.#
- reimplement -z
- use syslog()
- improve consistancy of messages
- allow -f to recover cleared dumps
- return bufsize to 1024 * 1024
- return the ability to write sparse files
- update man page
- fix minfree to require 2k for info file instead of the kernel size
- include Berkeley copyright too due to amount of old code copied

Submitted by: Chad David <davidc@acns.ab.ca>

22 years agoRemove leading whitespace from the setuid file lists.
Crist J. Clark [Sun, 5 May 2002 00:59:37 +0000 (00:59 +0000)]
Remove leading whitespace from the setuid file lists.

Due to the way we run ls(1), through xargs(1), the leading whitespace
can change even when the setuid files haven't. To avoid displaying
these lines, we currently run diff(1) with the '-w' option. However,
this is probably not the ideal way to go; there is a very, very small
possibility for diff(1) to miss things is shouldn't. So, with the
leading space cleaned, we can revert to the '-b' option which is
"safer."

PR: conf/37618
Reviewed by: brian
MFC after: 3 days

22 years ago* Fix a small typo, thanks to olgeni
Doug Barton [Sat, 4 May 2002 22:53:03 +0000 (22:53 +0000)]
* Fix a small typo, thanks to olgeni
* Document the COMP_CONFS variable in the .mergemasterrc example

22 years ago* Add some niceties to make the -p option quieter, and better behaved
Doug Barton [Sat, 4 May 2002 22:45:12 +0000 (22:45 +0000)]
* Add some niceties to make the -p option quieter, and better behaved
* Take a simple idea about setting the screen width using stty (for sdiff's
purposes) thoughtfully suggested by Joerg to an absurd conclusion.

22 years agoUse pages instead of uz_maxpages, which has not been initialized yet, when
Jeff Roberson [Sat, 4 May 2002 21:49:29 +0000 (21:49 +0000)]
Use pages instead of uz_maxpages, which has not been initialized yet, when
creating the vm_object.  This was broken after the code was rearranged to
grab giant itself.

Spotted by:     alc

22 years agoTweak a data type from char * to u_char *
Brian Somers [Sat, 4 May 2002 21:47:43 +0000 (21:47 +0000)]
Tweak a data type from char * to u_char *

22 years ago o Make _vm_object_allocate() and vm_object_allocate() callable
Alan Cox [Sat, 4 May 2002 20:23:48 +0000 (20:23 +0000)]
 o Make _vm_object_allocate() and vm_object_allocate() callable
   without holding Giant.
 o Begin documenting the trivial cases of the locking protocol
   on vm_object.

22 years agoUse -ffreestanding for kernel bits unconditionally.
David E. O'Brien [Sat, 4 May 2002 20:07:33 +0000 (20:07 +0000)]
Use -ffreestanding for kernel bits unconditionally.

22 years agoFix a typo.
Maxime Henrion [Sat, 4 May 2002 19:50:09 +0000 (19:50 +0000)]
Fix a typo.

Submitted by: dwmalone

22 years agoShake unused stuff out of the flags in struct buf->b_flags.
Poul-Henning Kamp [Sat, 4 May 2002 19:40:34 +0000 (19:40 +0000)]
Shake unused stuff out of the flags in struct buf->b_flags.

22 years agoMake ``set mru'' require a context. In multi-link mode, there's no
Brian Somers [Sat, 4 May 2002 19:38:43 +0000 (19:38 +0000)]
Make ``set mru'' require a context.  In multi-link mode, there's no
point in being allowed to ``set mru'' for the MP lcp layer.

Spotted by: Richard Browne <richb@timestone.com.au>
MFC after: 1 month

22 years agoRemove a six year old undocumented #ifdef : NO_B_MALLOC.
Poul-Henning Kamp [Sat, 4 May 2002 19:24:55 +0000 (19:24 +0000)]
Remove a six year old undocumented #ifdef : NO_B_MALLOC.

22 years agoThe struct buf->b_act was not used anywere.
Poul-Henning Kamp [Sat, 4 May 2002 19:06:32 +0000 (19:06 +0000)]
The struct buf->b_act was not used anywere.

22 years agoRestore the ability interrupt dumps on i386, based on
Bill Fenner [Sat, 4 May 2002 17:45:48 +0000 (17:45 +0000)]
Restore the ability interrupt dumps on i386, based on
the old kern_shutdown.c .  Other archs might be able to
use similar code but I don't have anything to test on.

22 years agoRemove obsolete code (that was already #if 0'd out).
Matthew Dillon [Sat, 4 May 2002 17:10:15 +0000 (17:10 +0000)]
Remove obsolete code (that was already #if 0'd out).
Requested by: Hiten Pandya <hitmaster2k@yahoo.com>

22 years agoWhen a program launched on the local console exits, the shell will
Maxime Henrion [Sat, 4 May 2002 15:42:38 +0000 (15:42 +0000)]
When a program launched on the local console exits, the shell will
call read() to get the next command, and scread() disables the
screensaver.  We don't want this behaviour in the sc_saver_keybonly
case.

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

22 years agoDocument the fact that at(1) is run through cron(8) and that this affects
Jeroen Ruigrok van der Werven [Sat, 4 May 2002 15:33:12 +0000 (15:33 +0000)]
Document the fact that at(1) is run through cron(8) and that this affects
the granularity of the at queue runs.

Submitted by: mux

22 years ago(forced)
Maxim Sobolev [Sat, 4 May 2002 14:51:07 +0000 (14:51 +0000)]
(forced)

Previous delta (pkg_info -O) is subject of:

MFC after: 1 week

22 years agoAdd ability to query installed packages based on origin.
Maxim Sobolev [Sat, 4 May 2002 14:49:49 +0000 (14:49 +0000)]
Add ability to query installed packages based on origin.

22 years agoUse region 7 addresses for the slabs in the PV and PT zones so that we
Doug Rabson [Sat, 4 May 2002 12:21:32 +0000 (12:21 +0000)]
Use region 7 addresses for the slabs in the PV and PT zones so that we
don't confuse the zone allocater by translating region 5 addresses to
region 7 addresses (which is unavoidable for PTEs).

22 years agoDo not call nge_tick() if we get an PHY interrupt, nge_tick() calls
Poul-Henning Kamp [Sat, 4 May 2002 11:15:33 +0000 (11:15 +0000)]
Do not call nge_tick() if we get an PHY interrupt, nge_tick() calls
mii_tick() which should only be called once per second.

Our current MII/PHY state-engine is not able to deal with PHY
interrupts as far as I can tell, and most net drivers don't seem
to use the link/status change call-back mechanism.  It seems that
MII/PHY was orphaned before it grew up.

22 years agoIncrease gigE negotiation timeout to 17 seconds.
Poul-Henning Kamp [Sat, 4 May 2002 11:08:49 +0000 (11:08 +0000)]
Increase gigE negotiation timeout to 17 seconds.

10 seconds is not enough to negotiate a gigE link with a cisco switch which
holds carrier off for several seconds between tries.

22 years agoClean up mii/phy drivers: Remove the MIIF_DOINGAUTO which doesn't really
Poul-Henning Kamp [Sat, 4 May 2002 11:00:30 +0000 (11:00 +0000)]
Clean up mii/phy drivers:  Remove the MIIF_DOINGAUTO which doesn't really
do anything at the end of the day except bloat the drivers which has
copy&pasted it.

22 years agoRestore the minfree functionality into savecore. This patch
Maxime Henrion [Sat, 4 May 2002 10:36:35 +0000 (10:36 +0000)]
Restore the minfree functionality into savecore.  This patch
restores the check_space() function with small modifications
to make it work with the current code.  The patch was slightly
modified by Bill Fenner to handle error messages better.

Reviewed by: fenner

22 years agoMake sure we don't index the pm_rid array out of bounds in
Marcel Moolenaar [Sat, 4 May 2002 08:04:28 +0000 (08:04 +0000)]
Make sure we don't index the pm_rid array out of bounds in
pmap_ensure_rid(). This can happen because the function is
called for both user and kernel addresses, while the rid array
only has room for user addresses. This bug got exposed by rev
1.58 of ia64/ia64/pmap.c and rev 1.8 of ia64/include/pmap.h.

22 years agostyle(9): 'if' and 'while' need a space after them.
Alfred Perlstein [Sat, 4 May 2002 07:40:49 +0000 (07:40 +0000)]
style(9): 'if' and 'while' need a space after them.

22 years agoAdd myself.
Jonathan Mini [Sat, 4 May 2002 06:34:11 +0000 (06:34 +0000)]
Add myself.

Approved by: alfred

22 years agoGrouping changes.
Juli Mallett [Sat, 4 May 2002 06:00:19 +0000 (06:00 +0000)]
Grouping changes.

Submitted by: Carl Schmidt <cschmidt@slackerbsd.org>

22 years agoDon't use pointers as booleans.
Juli Mallett [Sat, 4 May 2002 05:53:02 +0000 (05:53 +0000)]
Don't use pointers as booleans.

22 years agoTry to de-confuse ad-hoc, demo ad-hoc and ibss:
Warner Losh [Sat, 4 May 2002 04:59:20 +0000 (04:59 +0000)]
Try to de-confuse ad-hoc, demo ad-hoc and ibss:
o Don't ever refer to ad-hoc mode in the raw.  Instead, refer to it as
  demo ad-hoc mode or lucent legacy demo ad-hoc or some variation on this
  theme.
o Talk about point-to-point modes rather than ad-hoc modes.  Use ibss where
  appropriate.
o Fix type IBBS -> IBSS

22 years agoInclude systm.h so panic(9) is defined when doing DEBUG_ALL_VFS_LOCKS.
Jeff Roberson [Sat, 4 May 2002 02:40:37 +0000 (02:40 +0000)]
Include systm.h so panic(9) is defined when doing DEBUG_ALL_VFS_LOCKS.

22 years agoInclude systm.h for panic(9) so that DEBUG_ALL_VFS_LOCKS compiles.
Jeff Roberson [Sat, 4 May 2002 02:37:00 +0000 (02:37 +0000)]
Include systm.h for panic(9) so that DEBUG_ALL_VFS_LOCKS compiles.

22 years agoinclude systm.h in vnode_if.c so that panic is defined when we're doing
Jeff Roberson [Sat, 4 May 2002 02:35:13 +0000 (02:35 +0000)]
include systm.h in vnode_if.c so that panic is defined when we're doing
DEBUG_ALL_VFS_LOCKS.

22 years ago o Remove GIANT_REQUIRED from vm_map_lookup_entry() and
Alan Cox [Sat, 4 May 2002 02:07:36 +0000 (02:07 +0000)]
 o Remove GIANT_REQUIRED from vm_map_lookup_entry() and
   vm_map_check_protection().
 o Call vm_map_check_protection() without Giant held in munmap().

22 years agoReplace calloc(3) with malloc(3), and set the tail of the
Juli Mallett [Sat, 4 May 2002 00:08:06 +0000 (00:08 +0000)]
Replace calloc(3) with malloc(3), and set the tail of the
argument list to NULL, to terminate the arguments passed
to execvp(2).

Thanks to: bde (for pointing out some missing parens)

And with apologies to Ozzy Osbourne:

On his way to dinner
It took him by surprise
When an email from bde
Said his code was full of lies

Style (indentation)
And his bracing were all wrong
He casted to size_t
When he should have cast to long

I don't mind

Single letter identifiers
Unwrapped Line
Over 80 chars
Far over 80 chars

Who can we get to send diffs
We need Bruce D. Evans
Shows you comparason to his
We need Bruce D. Evans again

22 years agoem(4) updates.
Bruce A. Mah [Fri, 3 May 2002 22:41:35 +0000 (22:41 +0000)]
em(4) updates.

22 years agoNew release notes: PCI_ENABLE_IO_MODES (+MFC), drm drivers,
Bruce A. Mah [Fri, 3 May 2002 22:18:30 +0000 (22:18 +0000)]
New release notes:  PCI_ENABLE_IO_MODES (+MFC), drm drivers,
csplit(1), mk_cmds(1) removal (+MFC), sysinstall(8) multi-volume
support (+MFC).

Modified release notes:  m4 20020426, OpenPAM Cinnamon.

Other MFCs noted:  CPU_ATHLON_SSE_HACK.

Un-MFC-ed:  Emphemeral port range change.

22 years agostyle(9): sort includes, remove EOL spaces.
Maxim Konovalov [Fri, 3 May 2002 21:03:37 +0000 (21:03 +0000)]
style(9): sort includes, remove EOL spaces.

MFC after: 1 week

22 years agoI have no idea how I managed to get two lines about geom in there...
Poul-Henning Kamp [Fri, 3 May 2002 21:02:42 +0000 (21:02 +0000)]
I have no idea how I managed to get two lines about geom in there...

22 years agowait(2) for compress_log() processes to complete before exit.
Maxim Konovalov [Fri, 3 May 2002 20:53:37 +0000 (20:53 +0000)]
wait(2) for compress_log() processes to complete before exit.

Spotted by: Grigoriy Strokin <grg@isabase.philol.msu.ru>
Reviewed by: -current
MFC after: 1 month

22 years agoRename `eqrelval' which was renamed from `eqrel' to quiet a shadow warning
Juli Mallett [Fri, 3 May 2002 20:46:10 +0000 (20:46 +0000)]
Rename `eqrelval' which was renamed from `eqrel' to quiet a shadow warning
to `op' as suggested by Bruce, as this is used to store the operator between
vl and vr.

Submitted by: bde

22 years agoVarious updates to VOP_ATTRIB(9): mention locks being shared/exclusive
Robert Watson [Fri, 3 May 2002 20:28:38 +0000 (20:28 +0000)]
Various updates to VOP_ATTRIB(9): mention locks being shared/exclusive
for VOP_GETATTR() and VOP_SETATTR(), reference VOP_NULL() to suggest
clearing all of *vap with it before setting specific values.  Cross
reference VFS(9).  Indicate that failure modes are possible from
VOP_GETATTR() and that an errno value is returned.

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

22 years agoFix a bug whereby we were getting ~0 and comparing it to maxsize, i.e. if
Juli Mallett [Fri, 3 May 2002 19:45:41 +0000 (19:45 +0000)]
Fix a bug whereby we were getting ~0 and comparing it to maxsize, i.e. if
s1 was 0 length, and replstr was 0 length, etc., we would end up subtracting
one from zero and seeing if it was greater than the size_t (unsigned) var
maxsize...  This would cause us to return a string consisting of essentially
only match, which is not the right behaviour if we have 0 length inpline.

22 years agoAdd a test for the -R option.
Juli Mallett [Fri, 3 May 2002 18:44:43 +0000 (18:44 +0000)]
Add a test for the -R option.

22 years agoHPFS picks up the vop_stdgetpages and vop_stdputpages member functions
Poul-Henning Kamp [Fri, 3 May 2002 18:23:29 +0000 (18:23 +0000)]
HPFS picks up the vop_stdgetpages and vop_stdputpages member functions
via the default entry and the default vop vector.

22 years agoMinor tweaks:
Warner Losh [Fri, 3 May 2002 18:11:17 +0000 (18:11 +0000)]
Minor tweaks:
o go ahead and document ibss-master and ibss modes, since there are
  patches in the pipeline to support them.
o Note that they aren't implemented yet.
o Note that different regulatory domains have different default
  channels.
o Note that Lucent cards prior to firmware 6.0.4 do not support
  ibss mode, and only support the older demo ad-hoc mode.
o Note that PRISM2 chipsets do not support WDS mode of operation
  (the mysterious -p 2 option).

22 years agoTypo fix: detects -> detect.
Juli Mallett [Fri, 3 May 2002 17:59:25 +0000 (17:59 +0000)]
Typo fix: detects -> detect.

Reviewed by: phk

22 years agoFix a memory-leak when configuring a vnode backed md(4) device fails.
Poul-Henning Kamp [Fri, 3 May 2002 17:55:10 +0000 (17:55 +0000)]
Fix a memory-leak when configuring a vnode backed md(4) device fails.

Submitted by: Martin FaxĂ©r <gmh003532@brfmasthugget.se>
MFC after: 4 weeks

22 years agoMove substitution out to a function, prerun(), and replace use of run() with
Juli Mallett [Fri, 3 May 2002 17:41:54 +0000 (17:41 +0000)]
Move substitution out to a function, prerun(), and replace use of run() with
it.  It handles everything (right now) that needs done before run(), except
the -J case, because while that would be easy, I don't quite understand -J.

Reviewed by: src/tools/regression/usr.bin/xargs