]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoMerge local changes, add new files and remove obsoleted ones.
Mike Smith [Wed, 31 Oct 2001 02:34:45 +0000 (02:34 +0000)]
Merge local changes, add new files and remove obsoleted ones.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r85756,
Mike Smith [Wed, 31 Oct 2001 02:32:29 +0000 (02:32 +0000)]
This commit was generated by cvs2svn to compensate for changes in r85756,
which included commits to RCS files with non-trunk default branches.

22 years agoImport the Intel ACPI CA 20011018 snapshot
Mike Smith [Wed, 31 Oct 2001 02:32:29 +0000 (02:32 +0000)]
Import the Intel ACPI CA 20011018 snapshot

22 years agoThis commit was generated by cvs2svn to compensate for changes in r85754,
Mike Smith [Wed, 31 Oct 2001 02:32:28 +0000 (02:32 +0000)]
This commit was generated by cvs2svn to compensate for changes in r85754,
which included commits to RCS files with non-trunk default branches.

22 years agoImport the Intel ACPI CA 20011018 snapshot
Mike Smith [Wed, 31 Oct 2001 02:32:28 +0000 (02:32 +0000)]
Import the Intel ACPI CA 20011018 snapshot

22 years agoSheldon asked me to make a note of the Message-ID of the e-mail from
Nick Sayer [Wed, 31 Oct 2001 02:04:33 +0000 (02:04 +0000)]
Sheldon asked me to make a note of the Message-ID of the e-mail from
Dave approving the copyright statement.

The message was like this:

Date: Mon, 29 Oct 2001 13:18:36 -0500
From: David Safford <safford@watson.ibm.com>
Message-Id: <20011029181842.0DCAAD7C59@safford.watson.ibm.com>
Subject: Re: SRA copyright?
In-Reply-To: Your message of "Mon, 29 Oct 2001 09:13:14 PST."
     <3846.66.126.254.34.1004375594.squirrel@medusa.kfu.com>

looks fine.

dave safford

> >> Some time ago we spoke about SRA and importing it into FreeBSD. I
> >> forgot to ask if you had a prefered license boilerplate for the top of
> >> the files. It has come up recently, and the SRA code in FreeBSD
> >> doesn't have one.
> >>
> >>
> > I really have no preference - use whatever is most convenient in the
> > FreeBSD environment.
> >
> > dave safford
>
> Here is a representative diff. Let me know if you feel any changes are
> necessary.
>
> ===================================================================
> RCS file: /c/ncvs/src/crypto/telnet/libtelnet/pk.h,v
> retrieving revision 1.3
> retrieving revision 1.4

and so on and so on

22 years agoAdd a long-overdue nail to the deprecated /dev/urandom interface
Mark Murray [Tue, 30 Oct 2001 21:26:50 +0000 (21:26 +0000)]
Add a long-overdue nail to the deprecated /dev/urandom interface
by asking some things that need unpredictable numbers to read
/dev/random instead.

22 years agoPut a Band-Aid over the asr driver so that it hopefully won't cause
Scott Long [Tue, 30 Oct 2001 21:13:39 +0000 (21:13 +0000)]
Put a Band-Aid over the asr driver so that it hopefully won't cause
panics until the author comes up with a real fix.

22 years agoThreads sit on condition variable wait queue's, not proceses (sic).
John Baldwin [Tue, 30 Oct 2001 20:43:45 +0000 (20:43 +0000)]
Threads sit on condition variable wait queue's, not proceses (sic).

22 years agoResequence network protocols section.
Bruce A. Mah [Tue, 30 Oct 2001 20:25:33 +0000 (20:25 +0000)]
Resequence network protocols section.

Some ipfw(8) and ip6fw(8) items were mistakenly put here; move them
to the userland area.

No content or markup changes.

22 years agoNew release notes: gprof(1) -K, tip(1) update and cu(1) substitute,
Bruce A. Mah [Tue, 30 Oct 2001 20:19:31 +0000 (20:19 +0000)]
New release notes:  gprof(1) -K, tip(1) update and cu(1) substitute,
awk from Bell Labs replaces GNU awk.

Updated release note:  UUCP removal.

Delete release note:  GNU awk update.

22 years agoImplement -e option. It modifies the output produced by sysctl(8) in
Anton Berezin [Tue, 30 Oct 2001 20:15:32 +0000 (20:15 +0000)]
Implement -e option.  It modifies the output produced by sysctl(8) in
such a way that the name and the value of the variable(s) are separated
with `=' instead of the usual `: '.  This is useful for producing output
that can be fed back to the sysctl utility (pasted to sysctl.conf, for
example).

Reviewed by: rwatson
Approved by: markm
MFC after: 2 weeks

22 years agoFix a bug where restore(8) segfaults while trying to restore on a
Anton Berezin [Tue, 30 Oct 2001 20:06:59 +0000 (20:06 +0000)]
Fix a bug where restore(8) segfaults while trying to restore on a
read-only FS.

Reviewed by: audit silence
Approved by: markm
MFC after: 2 weeks

22 years agoGroup ed(4) entries together.
Bruce A. Mah [Tue, 30 Oct 2001 19:49:17 +0000 (19:49 +0000)]
Group ed(4) entries together.

22 years agoInstall libssh and libssh_pic. These are needed when building
Mark Murray [Tue, 30 Oct 2001 19:45:00 +0000 (19:45 +0000)]
Install libssh and libssh_pic. These are needed when building
statically, and when building things (like login(8)) standalone.
libssh_pic is needed for libpam and modules.

Requested by: peter

22 years agotip(1) can do cu(1). We don't need this anymore.
Mark Murray [Tue, 30 Oct 2001 19:36:57 +0000 (19:36 +0000)]
tip(1) can do cu(1). We don't need this anymore.

22 years agoFix a (long standing?) bug in ip_output(): if ip_insertoptions() is
Bill Paul [Tue, 30 Oct 2001 18:15:48 +0000 (18:15 +0000)]
Fix a (long standing?) bug in ip_output(): if ip_insertoptions() is
called and ip_output() encounters an error and bails (i.e. host
unreachable), we will leak an mbuf. This is because the code calls
m_freem(m0) after jumping to the bad: label at the end of the function,
when it should be calling m_freem(m). (m0 is the original mbuf list
_without_ the options mbuf prepended.)

Obtained from: NetBSD

22 years agoMake sure the netmask always has an address family. This fixes Linux
Dag-Erling Smørgrav [Tue, 30 Oct 2001 15:57:20 +0000 (15:57 +0000)]
Make sure the netmask always has an address family.  This fixes Linux
ifconfig, which expects the address returned by the SIOCGIFNETMASK ioctl
to have a valid sa_family.  Similar changes may be necessary for IPv6.

While we're here, get rid of an unnecessary temp variable.

MFC after: 2 weeks

22 years agoAdd -K support to gprof(1), which enables dynamic symbol resolution from
Brian Feldman [Tue, 30 Oct 2001 15:54:09 +0000 (15:54 +0000)]
Add -K support to gprof(1), which enables dynamic symbol resolution from
the currently-running kernel (and supercedes an executable file argument
given).  With this change, properly-compiled KLD modules are now
able to be profiled.

Obtained from: NAI Labs CBOSS project
Funded by: DARPA

22 years agoStyle(9) fix, mainly white spaces.
Mitsuru IWASAKI [Tue, 30 Oct 2001 15:51:52 +0000 (15:51 +0000)]
Style(9) fix, mainly white spaces.

22 years agoRestore the main BAR for the bridge on resume. Some machines don't save
Warner Losh [Tue, 30 Oct 2001 15:31:49 +0000 (15:31 +0000)]
Restore the main BAR for the bridge on resume.  Some machines don't save
this accross suspend/resume events and this was causing the dreaded false
positive hit on my "static bug" test.

Note: the PCI bus code should do this for us.
Note2: We don't do the same for I/O based pci devices since it is
more code and doesn't appear to be necessary.

Submitted by: Toshiyuki Kawashima-san <tos@fa2.so-net.ne.jp>
Obtained from: bsd-nomads:16012

22 years agoAdd the sysctl "kern.function_list", which currently exports all
Brian Feldman [Tue, 30 Oct 2001 15:21:45 +0000 (15:21 +0000)]
Add the sysctl "kern.function_list", which currently exports all
function symbols in the kernel in a list of C strings, with an extra
nul-termination at the end.

This sysctl requires addition of a new linker operation.  Now,
linker_file_t's need to respond to "each_function_name" to export
their function symbols.

Note that the sysctl doesn't currently allow distinguishing multiple
symbols with the same name from different modules, but could quite
easily without a change to the linker operation.  This will be a nicety
to have when it can be used.

Obtained from: NAI Labs CBOSS project
Funded by: DARPA

22 years agoAlso, machine/profile.h should be necessary for the function prototype
Brian Feldman [Tue, 30 Oct 2001 15:10:16 +0000 (15:10 +0000)]
Also, machine/profile.h should be necessary for the function prototype
of kmupetext().

22 years agoUse kmupetext() for ELF KLDs to allow for increased text segment size.
Brian Feldman [Tue, 30 Oct 2001 15:08:51 +0000 (15:08 +0000)]
Use kmupetext() for ELF KLDs to allow for increased text segment size.

Obtained from: NAI Labs CBOSS project
Funded by: DARPA

22 years agoAdd kmupetext(), a function that expands the range of memory covered
Brian Feldman [Tue, 30 Oct 2001 15:04:57 +0000 (15:04 +0000)]
Add kmupetext(), a function that expands the range of memory covered
by the profiler on a running system.  This is not done sparsely, as
memory is cheaper than processor speed and each gprof mcount() and
mexitcount() operation is already very expensive.

Obtained from: NAI Labs CBOSS project
Funded by: DARPA

22 years agoWhen dropping a packet because there is no room in the queue (which itself
Jonathan Lemon [Tue, 30 Oct 2001 14:58:27 +0000 (14:58 +0000)]
When dropping a packet because there is no room in the queue (which itself
is somewhat bogus), update the statistics to indicate something was dropped.

PR: 13740

22 years agoOLDCARD isn't supported on alpha or ia64, so don't pretend that it is by
Warner Losh [Tue, 30 Oct 2001 14:53:36 +0000 (14:53 +0000)]
OLDCARD isn't supported on alpha or ia64, so don't pretend that it is by
including it.

22 years agoDon't try to use sio with NEWCARD 16 bit yet. It eats all pccards :-)
Warner Losh [Tue, 30 Oct 2001 14:49:29 +0000 (14:49 +0000)]
Don't try to use sio with NEWCARD 16 bit yet.  It eats all pccards :-)

Reported by: Marcell Moolenaar

22 years agoSome improvements of control method battery driver.
Mitsuru IWASAKI [Tue, 30 Oct 2001 14:24:26 +0000 (14:24 +0000)]
Some improvements of control method battery driver.
 - Add a new MIB for battery info expire time in order to make it changeable.
   Battery info expire time can be specified by
   hw.acpi.battery.info_expire in sec.
 - Add own MALLOC type and fix some potential memory leakages.
 - Change some frequent printings to verbose printing.
 - Stop timeout during acpi_cmbat_get_bst() too.  This should reduce
   the races with BIF evaluation.
 - Remove acpi_cmbat_get_bif() invocation from acpi_cmbat_attach().
   This was redundant because this should be called from
   acpi_cmbat_timeout() now.

22 years agoUse -ffreestanding instead of -fno-builtin. That's the officially blessed
Doug Rabson [Tue, 30 Oct 2001 09:37:00 +0000 (09:37 +0000)]
Use -ffreestanding instead of -fno-builtin.  That's the officially blessed
way of saying that its not linking with libc.

Submitted by: peter

22 years agoAdjust for building with gcc 3.0.1.
Doug Rabson [Tue, 30 Oct 2001 09:09:13 +0000 (09:09 +0000)]
Adjust for building with gcc 3.0.1.

22 years agoSwitch to the One True AWK from Bell-Labs, away from GNU AWK.
David E. O'Brien [Tue, 30 Oct 2001 08:56:50 +0000 (08:56 +0000)]
Switch to the One True AWK from Bell-Labs, away from GNU AWK.

22 years agoFix typos.
Julian Elischer [Tue, 30 Oct 2001 08:37:45 +0000 (08:37 +0000)]
Fix typos.

22 years agoA node that allows ethernet type packets to be filtered to different
Julian Elischer [Tue, 30 Oct 2001 07:28:17 +0000 (07:28 +0000)]
A node that allows ethernet type packets to be filtered to different
hooks depending on ethertype. Great for prototyping protocols.
connects to the lower and upper hooks of an ethernet type of node.

Obtained from: Monzoon Networks.
Thanks to Andre Oppermann, May 2001.

22 years agoUse the thread we have instead of finding another
Julian Elischer [Tue, 30 Oct 2001 07:15:46 +0000 (07:15 +0000)]
Use the thread we have instead of finding another
that may be the wrong one.

22 years agoAdd code to copy the enironment and loader metadata into kernel space.
Jake Burkholder [Tue, 30 Oct 2001 06:37:36 +0000 (06:37 +0000)]
Add code to copy the enironment and loader metadata into kernel space.

22 years agoAdd definitions for network support, doesn't work yet.
Jake Burkholder [Tue, 30 Oct 2001 06:31:45 +0000 (06:31 +0000)]
Add definitions for network support, doesn't work yet.
Pass the right arguments to the kernel.
Replace magic numbers with symbolic constants.
Pass the real openfirmware entry point to OF_init.

22 years agoUse ENTRY() for defining functions in asm.
Jake Burkholder [Tue, 30 Oct 2001 06:27:34 +0000 (06:27 +0000)]
Use ENTRY() for defining functions in asm.
Remove asm functions to call the openfirmware and kernel entry points;
we can just call them directly.
Don't use the stack pointer for an intermediate result in setx.
Put the stack in the bss.

22 years agoMake the openfirmware entry point function pointer non-static so that it
Jake Burkholder [Tue, 30 Oct 2001 06:23:32 +0000 (06:23 +0000)]
Make the openfirmware entry point function pointer non-static so that it
can be passed to the kernel.

22 years agoMFGeneric: ->1.327 Sync with GENERIC
Warner Losh [Tue, 30 Oct 2001 06:11:15 +0000 (06:11 +0000)]
MFGeneric: ->1.327 Sync with GENERIC

22 years agoMove device lnc to isa section, since it no longer uses the compat shims.
Warner Losh [Tue, 30 Oct 2001 06:08:56 +0000 (06:08 +0000)]
Move device lnc to isa section, since it no longer uses the compat shims.
Add comment about lnc.
Remove probe order comment from isa_compat.c.  That appears to no longer
be the case.

22 years agoLink Slovak with Czech
Andrey A. Chernov [Tue, 30 Oct 2001 01:55:19 +0000 (01:55 +0000)]
Link Slovak with Czech

PR: 31598
Submitted by: Vladimir Guzma <vg@cs.tut.fi>

22 years agoMFi386: sys/i386/i386/machdep.c revisions 1.481 and 1.482
Yoshihiro Takahashi [Tue, 30 Oct 2001 00:53:19 +0000 (00:53 +0000)]
MFi386: sys/i386/i386/machdep.c revisions 1.481 and 1.482

22 years agoFix a typo in comment and #ifdef fixes: GRAP_PRIO -> GRAB_PRIO so that
John Baldwin [Tue, 30 Oct 2001 00:19:42 +0000 (00:19 +0000)]
Fix a typo in comment and #ifdef fixes: GRAP_PRIO -> GRAB_PRIO so that
x86 SMP kernels actually boot again to single user mode.

Pointy hat to: jhb
Noticed by: jlemon

22 years agoMore reorganization of 3Com devices. While I'm here, add some
Bruce A. Mah [Mon, 29 Oct 2001 22:31:22 +0000 (22:31 +0000)]
More reorganization of 3Com devices.  While I'm here, add some
devices (gleaned from comments in driver source files) supported
by xl(4).

22 years agoUpdate tuning so that it mentions maxusers, nmbclusters, and nsfbufs as
Mike Silbersack [Mon, 29 Oct 2001 22:29:01 +0000 (22:29 +0000)]
Update tuning so that it mentions maxusers, nmbclusters, and nsfbufs as
tunables in loader.conf rather than just kernel options.

MFC after: 3 days

22 years agoWhen scanning for control messages, don't process the data mbufs.
David Malone [Mon, 29 Oct 2001 20:04:03 +0000 (20:04 +0000)]
When scanning for control messages, don't process the data mbufs.
This could cause hangs if a unix domain socket was closed with data
still to be read from it.

Tested by: Andrea Campi <andrea@webcom.it>

22 years agoFix world by trimming an extra comment terminator.
John Baldwin [Mon, 29 Oct 2001 19:22:38 +0000 (19:22 +0000)]
Fix world by trimming an extra comment terminator.

22 years agoSome small improvements of ACPI thermal driver.
Mitsuru IWASAKI [Mon, 29 Oct 2001 18:09:43 +0000 (18:09 +0000)]
Some small improvements of ACPI thermal driver.
 - Give a guaranteed minimum cooling run time to avoid too frequent
   cooling system On/Off switching.  The minimum cooling run time can be
   specified by hw.acpi.thermal.min_runtime in sec.
 - Refine message printing (_AC-1 -> NONE).
 - Add verbose mode enable/disable capability by hw.acpi.verbose in bool.

Reviewed by: acpi-jp@ folks

22 years agoDon't set CR0_NE in cpu_setregs() for the SMP case, since setting it
Bruce Evans [Mon, 29 Oct 2001 16:31:15 +0000 (16:31 +0000)]
Don't set CR0_NE in cpu_setregs() for the SMP case, since setting it
is npx.c's job and setting it here breaks the edit-time option of not
setting it in npx.c.  (It is not set in the right places for the SMP
case, but always setting it here is harmless because there isn't even
an edit-time option to not set it.)

22 years agoSystem wide policy should be returned when no policy found in the SPD.
Hajimu UMEMOTO [Mon, 29 Oct 2001 16:29:41 +0000 (16:29 +0000)]
System wide policy should be returned when no policy found in the SPD.
The packet was rejected in ipsec[46]_tunnel_validate().

Obtained from: KAME
MFC after: 1 week

22 years agoMake CD-ROM distributions at the pc98 release target.
Yoshihiro Takahashi [Mon, 29 Oct 2001 16:25:13 +0000 (16:25 +0000)]
Make CD-ROM distributions at the pc98 release target.

MFC after: 3 days

22 years agoDon't assume fixit media is only either a floppy disk or 2nd CDROM disc
Brian Somers [Mon, 29 Oct 2001 16:17:19 +0000 (16:17 +0000)]
Don't assume fixit media is only either a floppy disk or 2nd CDROM disc

Reviewed by: murray
MFC after: 1 week

22 years agoDon't assume fixit media is a floppy disk.
Brian Somers [Mon, 29 Oct 2001 16:13:29 +0000 (16:13 +0000)]
Don't assume fixit media is a floppy disk.

Reviewed by: murray
MFC after: 1 week

22 years agoAdd Berkeley copyright to SRA.
Nick Sayer [Mon, 29 Oct 2001 16:12:16 +0000 (16:12 +0000)]
Add Berkeley copyright to SRA.

This is by the kind permission of Dave Safford, formerly of TAMU who wrote the
original code. Here is an excerpt of the e-mail exchange concerning this
issue:

Dave Safford wrote:
>Nick Sayer wrote:
>> Some time ago we spoke about SRA and importing it into FreeBSD. I forgot to
>> ask if you had a prefered license boilerplate for the top of the files. It
>> has come up recently, and the SRA code in FreeBSD doesn't have one.

>I really have no preference - use whatever is most convenient in the
>FreeBSD environment.

>dave safford

This is the standard BSD license with clause 3 removed and clause 4
suitably renumbered.

MFC after: 1 day

22 years agoA few more style changes picked up whilst working on an MFC to -stable.
Josef Karthauser [Mon, 29 Oct 2001 15:09:07 +0000 (15:09 +0000)]
A few more style changes picked up whilst working on an MFC to -stable.

22 years agoFix some whitespace, and a comment that I missed in the last commit.
Josef Karthauser [Mon, 29 Oct 2001 14:08:51 +0000 (14:08 +0000)]
Fix some whitespace, and a comment that I missed in the last commit.

22 years agoAdd unwind.c.
Doug Rabson [Mon, 29 Oct 2001 12:04:51 +0000 (12:04 +0000)]
Add unwind.c.

22 years ago* Factor out common code for manipulating the RSE backing store.
Doug Rabson [Mon, 29 Oct 2001 12:04:23 +0000 (12:04 +0000)]
* Factor out common code for manipulating the RSE backing store.
* Implement a fairly simplistic parser for unwinding stack frames.
* Use unwind records for DDB's 'trace' command. Also add support for
  tracing past exceptions to the context which generated the exception.

The stack unwind code requires a toolchain based on binutils-2.11.2 or
later and gcc-3.0.1 or later.

22 years agoMake the various bits of SMP code conditional on SMP so that I can still
Doug Rabson [Mon, 29 Oct 2001 11:57:12 +0000 (11:57 +0000)]
Make the various bits of SMP code conditional on SMP so that I can still
build non-SMP kernels.

22 years agoPut symbols at the start and end of the unwind section so that we can
Doug Rabson [Mon, 29 Oct 2001 11:40:14 +0000 (11:40 +0000)]
Put symbols at the start and end of the unwind section so that we can
find it at runtime.

22 years agoVarious fixes to make stack traces using the unwind tables work properly.
Doug Rabson [Mon, 29 Oct 2001 11:30:54 +0000 (11:30 +0000)]
Various fixes to make stack traces using the unwind tables work properly.

22 years agoFix disassembly of 'add a=b,c,1' and make the disassembly of the various
Doug Rabson [Mon, 29 Oct 2001 11:26:10 +0000 (11:26 +0000)]
Fix disassembly of 'add a=b,c,1' and make the disassembly of the various
break and nops consistent.

22 years agoThe size of the ELF hash table changed from 64 bits in the prototype
Peter Wemm [Mon, 29 Oct 2001 10:20:19 +0000 (10:20 +0000)]
The size of the ELF hash table changed from 64 bits in the prototype
toolchains to 32 bits in 2.11.2.

Obtained from: dfr

22 years agoMostly cosmetic. Use indentation that is reasonably close to other
Peter Wemm [Mon, 29 Oct 2001 10:18:58 +0000 (10:18 +0000)]
Mostly cosmetic.  Use indentation that is reasonably close to other
ia64 *.S files in our tree (eg: locore.s).  Put the rest of the args
in the .section lines.

Reviewed by: dfr

22 years agoUpdate for the new toolchain. ld doesn't provide _GLOBAL_OFFSET_TABLE_
Peter Wemm [Mon, 29 Oct 2001 10:14:51 +0000 (10:14 +0000)]
Update for the new toolchain.  ld doesn't provide _GLOBAL_OFFSET_TABLE_
in 2.11.2, so use a relocatable method of calculating gp.

Reviewed by: dfr

22 years agoUpdate rtld for the "new" ia64 ABI. In the old toolchain, the
Peter Wemm [Mon, 29 Oct 2001 10:10:10 +0000 (10:10 +0000)]
Update rtld for the "new" ia64 ABI.  In the old toolchain, the
DT_INIT and DT_FINI tags pointed to fptr records.  In 2.11.2, it points
to the actuall address of the function.  On IA64 you cannot just take
an address of a function, store it in a function pointer variable and
call it.. the function pointers point to a fptr data block that has the
target gp and address in it.  This is absolutely necessary for using
the in-tree binutils toolchain, but (unfortunately) will not work with
old shared libraries.  Save your old ld-elf.so.1 if you want to use
old ones still.  Do not mix-and-match.

This is a no-op change for i386 and alpha.

Reviewed by: dfr

22 years agoFix a dependency violation (branch after alloc)
Peter Wemm [Mon, 29 Oct 2001 10:05:32 +0000 (10:05 +0000)]
Fix a dependency violation (branch after alloc)

22 years agoFix fragmented packet handling.
Munechika SUMIKAWA [Mon, 29 Oct 2001 07:55:57 +0000 (07:55 +0000)]
Fix fragmented packet handling.

Obtained from: KAME
MFC after: 3 weeks

22 years agoo Send a test IPI from the BSP to itself at the same time APs
Marcel Moolenaar [Mon, 29 Oct 2001 07:34:48 +0000 (07:34 +0000)]
o  Send a test IPI from the BSP to itself at the same time APs
   are woken up.
o  Make IPIs synchronuous by default. If we want asynchronuous
   IPIs, we may want to make the memory fence controllable.

22 years agoAdd an IPI used for testing proper operation of delivering IPIs.
Marcel Moolenaar [Mon, 29 Oct 2001 07:30:37 +0000 (07:30 +0000)]
Add an IPI used for testing proper operation of delivering IPIs.

22 years agoUse a more sophisticated check that works better with older versions
Mike Barcroft [Mon, 29 Oct 2001 07:29:03 +0000 (07:29 +0000)]
Use a more sophisticated check that works better with older versions
of GCC when setting __func__ to NULL.  This also better supports GCC
when using C++.  Move the definition to a better location.

Submitted by: bde

22 years agoMake the output from assert() look more like the example in the C99
Mike Barcroft [Mon, 29 Oct 2001 07:07:25 +0000 (07:07 +0000)]
Make the output from assert() look more like the example in the C99
standard.

Requested by: bde

22 years agoMake the clock vector 255 instead of 240. On Lion boxes, 240 is
Marcel Moolenaar [Mon, 29 Oct 2001 05:17:36 +0000 (05:17 +0000)]
Make the clock vector 255 instead of 240. On Lion boxes, 240 is
the AP wake-up vector. We probably want a more dynamic approach
to assigning vectors in the future...

22 years agoForced commit: Previous diff removed the UserConfig section from
Bruce A. Mah [Mon, 29 Oct 2001 05:17:21 +0000 (05:17 +0000)]
Forced commit:  Previous diff removed the UserConfig section from
the alpha hardware notes.

22 years agoAdd infrastructure for an ia64 hardware list. It's a long way from
Bruce A. Mah [Mon, 29 Oct 2001 05:08:52 +0000 (05:08 +0000)]
Add infrastructure for an ia64 hardware list.  It's a long way from
being usable at this point, so it's not even hooked up to the rest
of the RELNOTESng build; however it will be helpful for testing the content
as it approaches reality.

22 years agoSmall correction in the LOCAL_SAPIC structure. The Flags field
Marcel Moolenaar [Mon, 29 Oct 2001 04:59:35 +0000 (04:59 +0000)]
Small correction in the LOCAL_SAPIC structure. The Flags field
starts at offset 8; not 6. Hence the structure is 12 bytes and
not 10 bytes. Adjust the definition so that the ProcessorEnabled
flag is moved from bit 15 to bit 31 in the Flags field.

The definition now matches ACPI 2.0 Errata 1.5.

22 years agoStart to unravel some of the mess in the Ethernet NIC section. The
Bruce A. Mah [Mon, 29 Oct 2001 04:45:08 +0000 (04:45 +0000)]
Start to unravel some of the mess in the Ethernet NIC section.  The
original hardware list was mostly organized by driver except for the
case of NICs by Intel, DEC, and 3Com, whose NICs tended to be
organized by manufacturer.  Most of these entries are now organized by
driver, except for a few 3Com NICs; also the multitude of ed(4)
drivers needs to be coalesced.

While I'm here, fix a markup glitch in the ISDN section that caused
any non-i386 hardware list to be incorrectly numbered.

22 years agoClean up the style of this header file.
Josef Karthauser [Mon, 29 Oct 2001 04:41:28 +0000 (04:41 +0000)]
Clean up the style of this header file.

22 years agoClarify a release note entry about fe(4)-supported PC Cards.
Bruce A. Mah [Mon, 29 Oct 2001 04:14:40 +0000 (04:14 +0000)]
Clarify a release note entry about fe(4)-supported PC Cards.

22 years agoMore white space changes.
Josef Karthauser [Mon, 29 Oct 2001 03:46:28 +0000 (03:46 +0000)]
More white space changes.

22 years agoMore stylistic tidying.
Josef Karthauser [Mon, 29 Oct 2001 03:25:49 +0000 (03:25 +0000)]
More stylistic tidying.

22 years agolog routine called w/ %ld and int argument, cast argument to long
Matthew Dillon [Mon, 29 Oct 2001 02:22:49 +0000 (02:22 +0000)]
log routine called w/ %ld and int argument, cast argument to long

22 years agofix int argument used in printf w/ %ld (cast to long)
Matthew Dillon [Mon, 29 Oct 2001 02:19:19 +0000 (02:19 +0000)]
fix int argument used in printf w/ %ld (cast to long)

22 years agopromote tv_sec in printf to make it type agnostic
Matthew Dillon [Mon, 29 Oct 2001 02:17:41 +0000 (02:17 +0000)]
promote tv_sec in printf to make it type agnostic

22 years agoo Do not parse the MADT as a side-effect in AcpiOsGetRootPointer,
Marcel Moolenaar [Mon, 29 Oct 2001 02:16:02 +0000 (02:16 +0000)]
o  Do not parse the MADT as a side-effect in AcpiOsGetRootPointer,
   do it as a side-effect of probing for MP hardware. This allows
   us to scan for local SAPICs early (especially before MBUF
   initialization).
o  Fix the Local SAPIC structure so that matches the Local SAPIC
   table entry. Now that the Local SAPIC info is the same as the
   Local APIC info, stop dumping the Local APIC entries.
o  For every Local SAPIC entry in the MADT that's not disabled,
   let the SMP code know about it. They represent actual CPUs.
o  Register the OS_BOOT_RENDEZ entry point and provide a (bogus)
   implementation for the entry point.
o  Provide a mapping for internal IPI numbers to ExtINT vectors.
o  In a MP system, announce the CPUs and start them by sending
   IPI_AP_WAKEUP to each of them. Not that it makes a difference
   at this time :-)
o  Miscellaneous style fixes and other adjustments.

22 years agoMake ttyprintf() of tv_sec value type agnostic.
Matthew Dillon [Mon, 29 Oct 2001 01:23:28 +0000 (01:23 +0000)]
Make ttyprintf() of tv_sec value type agnostic.

22 years agocast dev_t to int in printf to fix warning.
Matthew Dillon [Mon, 29 Oct 2001 01:22:15 +0000 (01:22 +0000)]
cast dev_t to int in printf to fix warning.

22 years agoSet RB_MULTIPLE (multiple console support) if the kernel is booted
Marcel Moolenaar [Mon, 29 Oct 2001 01:11:40 +0000 (01:11 +0000)]
Set RB_MULTIPLE (multiple console support) if the kernel is booted
with the -D flag.

22 years agoFix ctime(&int) due to change to lastlog->ll_time
Matthew Dillon [Mon, 29 Oct 2001 00:40:05 +0000 (00:40 +0000)]
Fix ctime(&int) due to change to lastlog->ll_time

22 years agoRemove training white spaces, and some other style violations.
Josef Karthauser [Mon, 29 Oct 2001 00:37:24 +0000 (00:37 +0000)]
Remove training white spaces, and some other style violations.

22 years agolocaltime() was being called with an int-pointer due to recent utmp
Matthew Dillon [Mon, 29 Oct 2001 00:36:01 +0000 (00:36 +0000)]
localtime() was being called with an int-pointer due to recent utmp
fixes.

22 years agolocaltime() was being called on an int pointer (due to recent
Matthew Dillon [Mon, 29 Oct 2001 00:34:20 +0000 (00:34 +0000)]
localtime() was being called on an int pointer (due to recent
utmp-related commits).  fixed.

22 years agoFix long format mismatch with int arg
Matthew Dillon [Mon, 29 Oct 2001 00:32:58 +0000 (00:32 +0000)]
Fix long format mismatch with int arg

22 years ago1) In devclass_alloc_unit(), skip duplicated wired devices (i.e. with fixed
Andrey A. Chernov [Sun, 28 Oct 2001 23:32:35 +0000 (23:32 +0000)]
1) In devclass_alloc_unit(), skip duplicated wired devices (i.e. with fixed
number) instead of allocating next free unit for them.  If someone needs
fixed place, he must specify it correctly. "Allocating next" is especially bad
because leads to double device detection and to "repeat make_dev panic" as
result.  This can happens if the same devices present somewhere on PCI bus,
hints and  ACPI.  Making them present in one place only not always
possible, "sc" f.e.  can't be removed from hints, it results to no console at
all.

2) In make_device(), detect when devclass_add_device() fails, free dev and
return. I.e. add missing error checking. This part needed to finish fix in 1),
but must be done this way in anycase, with old variant too.

22 years agoo Restore previous inconsistent style to login_fbtab.c, following its
Robert Watson [Sun, 28 Oct 2001 22:54:16 +0000 (22:54 +0000)]
o Restore previous inconsistent style to login_fbtab.c, following its
  modification to add glob support.  The submitted patch used a degraded
  form of KNF.

Reviewed by: bde

22 years agoAdjust printfs to be time_t agnostic.
Matthew Dillon [Sun, 28 Oct 2001 22:53:45 +0000 (22:53 +0000)]
Adjust printfs to be time_t agnostic.

22 years agoDo not assume that sizeof(tv_sec) is sizeof(time_t) - it will be, but
Matthew Dillon [Sun, 28 Oct 2001 20:53:17 +0000 (20:53 +0000)]
Do not assume that sizeof(tv_sec) is sizeof(time_t) - it will be, but
it isn't yet because our tv_sec is currently a long (so 64 bit architectures
are already broken to a degree).

22 years agoFix int vs time_t, convert int to time_t
Matthew Dillon [Sun, 28 Oct 2001 20:40:58 +0000 (20:40 +0000)]
Fix int vs time_t, convert int to time_t

22 years agoConvert time_t to/from 32 bit representations for transmission over
Matthew Dillon [Sun, 28 Oct 2001 20:33:07 +0000 (20:33 +0000)]
Convert time_t to/from 32 bit representations for transmission over
a network and storage.